\ (. V r Division of fishes, U. S. National Museum % w A^r53 % *1 U. S. DEPARTMENT OF COMMERCE DANIEL C. ROPER, Secretary BULLETIN OF THE UNITED STATES BUREAU OF FISHERIES VOL. XL VIII FRANK T. BELL COMMISSIONER UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1940 - : .c . r?t p— ^ -r- | r .. • y • ?.-■ ■. -1- ;i •Jr CONTENTS & Page Preliminary report on the life history of the common shrimp, Penaeus setiferus (Linn.). By F. W. Weymouth, Milton J. Lindner and W. W. Anderson. (Bulletin No. 14, issued Sept. 29, 1933.) 1-26 The homing instinct and age at maturity of pink salmon ( Oncorhynchus gorbuscha) . By Frederick A. Davidson. (Bulletin No. 15, issued Aug. 26, 1934.) 27-39 Reproduction and development of whitings or kingfishes, drums, spot, croaker, and weakfisheb or sea trouts, family sciaenidae, of the Atlantic coast of the United States. By Samuel F. Hildebrand and Louella E. Cable. (Bulletin No. 16, issued Sept. 19, 1934.) 41-117 Races of herring, Clupea pallasii, in Southeastern Alaska. By George A. Rounsefell and Edwin H. Dahlgren. (Bulletin No. 17, issued May 23, 1935.) 119-141 Effects of crude oil pollution on oysters in Louisiana waters. By Paul S. Galtsoff, Herbert F. Prytherch, Robert O. Smith, and Vera Koehring. (Bulletin No. 18, issued Nov. 29, 1935.) 143-210 Age and growth of the cisco, Leucichthys artedi, (Le Sueur), in the lakes of the northeastern highlands, Wisconsin. By Ralph Hile. (Bulletin No. 19, issued Feb. 27, 1936.) 211-317 Supplemental notes on the fishes of the Gulf of Maine. By Henry B. Bigelow and William C. Schroeder. (Bulletin No. 20, issued Nov. 17, 1936.) 319-343 Adaptation of the feeding mechanism of the oyster, Ostrea gigas, to changes in salinity. By A. E. Hopkins. (Bulletin No. 21, issued Dee. 15, 1936.) 345-364 Detection and measurement of stream pollution. By M. M. Ellis. (Bulletin No. 22, issued Apr. 7, 1937.) 365-437 Experimental observations on spawning, larval development, and setting in the Olympia oyster, Ostrea lurida. By A. E. Hopkins. (Bulletin No. 23, issued Oct. 19, 1937.) 439-503 Further notes on the development and life history of some teleosts at Beaufort, N. C. By Samuel F. Hildebrand and Louella E. Cable. (Bulletin No. 24, issued Mar. 15, 1938.) ■- 505-642 The migrations of pink salmon ( Oncorhynchus gorbuscha) in the Clarence and Sumner Straits regions of Southeastern Alaska. By Frederick A. Davidson and Leroy S. Christey. (Bulletin No. 25, issued May 17, 1938.) 643-666 The geographic distribution and environmental limitations of the Pacific salmon (genus Oncorhynchus) . By Frederick A. Davidson and Samuel J. Hutchinson. (Bul- letin No. 26, issued June 11, 1938.) 667-692 The salmon and salmon fisheries of Swiftsure Bank, Puget Sound, and the Fraser River. By George A. Rounsefell and George B. Kelez. (Bulletin No. 27, issued Oct. 17, 1938.) 693-823 The life history of the striped bass, or rockfish, Roccus saxatilis (Walbaum). By John C. Pearson. (Bulletin No. 28, issued Oct. 26, 1938.) 825-851 m ERRATA Page 164, line 4 should read “of an automatic siphon arrangement similar to that shown in figure 4, received an” Page 174, first 2 lines should be deleted. Page 177, last line should read “duration of test.” Page 232, last line of second paragraph should read “(pp. 237-249).” Page 588, line 8 should read “of the eye as in the adult.” xv U.S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner PRELIMINARY REPORT ON THE LIFE HISTORY OF THE COMMON SHRIMP PENAEUS SETIFERUS (LINN.) By F. W. WEYMOUTH, MILTON J. LINDNER and W. W. ANDERSON From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 14 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1933 For sale by the Superintendent of Documents, Washington, D. C. Price 5 cents PRELIMINARY REPORT ON THE LIFE HISTORY OF THE COMMON SHRIMP PENAEUS SETIFERUS (LINN.) By F. W. Weymotjtii, Ph.D., Milton J. Lindner, and W. W. Anderson, United States Bureau of Fisheries J- CONTENTS Page Introduction 1 Production and value of shrimp 2 Previous work on life history 6 Life history 8 Nature of data 8 Interpretation of data 9 Recognition of age groups 9 Spawning 11 Sex-ratio 14 Larvae 15 Life history- — Continued Page Interpretation — Continued Young 15 Growth 18 Fate of adults 19 Habits 21 Depletion and protection 22 Summary 23 Bibliography 25 INTRODUCTION Tlie present report is concerned with the salient features of the life history of the common or so-called "lake shrimp” ( Penaeus setiferus) of the South Atlantic and Gulf coasts and the bearing of these facts on the problems of the shrimp industry. The information has been obtained in the carrying out of a program of cooperative shrimp investigation in which the Bureau of Fisheries has been supported by the States of Louisiana, Georgia, and Texas. The loyalty and industry of the staff and of our associates have made possible the substantial results here recorded and the members deserve the particularized credit for which our space is too limited.1 Although the study is far from complete, it seems best to place the information obtained on record, in part because it is the most complete life history available for any species of shrimp and in part to give the purposes and needs of the investigational program. A brief statement will first be given of the extent and importance of the shrimp fishery in the United States, and next the state of knowledge at the beginning of the present work; then the life history of Penaeus setiferus will be outlined very briefly, after which the evidence from which these facts were drawn will be given in such detail as is now possible (this will constitute the great bulk of the paper) ; and finally, the bearing of this information upon the problems of depletion and protection will be discussed briefly. 1 The staff of the shrimp investigations includes F. W. Weymouth, Milton J. Lindner, and Gordon Gunter with headquarters in New Orleans, La., W. W. Anderson at Brunswick, Ga., J. S. Gutsell at Beaufort, N.C., and Kenneth H. Mosher at Aransas Pass, Tex. J. N. Gowanloeh and Forrest Durand, of the Bureau of Research and Statistics of the Louisiana Department of Con- servation, have been so closely associated in the cooperative program as scarcely to be differentiated from the staff members. We gladly acknowledge the aid of Dr. Waldo Schmitt, TJ.S. National Museum, in the identification of material; and of Dr. R. Von Ihering, of the Instituto Biologico, for information concerning the shrimp fishery of Sao Paulo, Brazil. W'e are also greatly indebted to many men in the shrimp industry, in particular to the late John Dymond, Jr., former president of the Southern Canners Exchange; R. R. Rice, of Aransas Pass, Tex., and Senator Jules Fisher, of New Orleans, La., who have aided in the collection of data and have placed at our disposal their records for statistical analysis. Approved for publication. Mar. 22, 1933. i 2 BULLETIN OF THE BUREAU OF FISHERIES PRODUCTION AND VALUE OF SHRIMP The shrimp fishery of the United States produced a total of 113,263,000 pounds in 1929 and 92,327,000 pounds in 1930, ranking in each year ninth in volume among all the fisheries. The value to the fishermen of the shrimp taken in 1929 was $4,575,000; this placed the fishery fifth in value. It was exceeded only by the salmon, oyster, haddock, and halibut. In 1930, due to a decline in price and in catch, the value was $3,134,000, so that it fell to tenth rank. As may be seen from table 1, more than 95 percent of the shrimp taken were from the South Atlantic and Gulf States, slightly over 40 percent coming from Louisiana alone. Here occur the following species: Penaeus setiferus, Penaeus brasiliensis , Macrobrachium sp., Xiphopenaeus kroyeri, Trachypenaeus constridus Figure 1. — Side view of the common shrimp ( Penaeus setiferus) with some of the more conspicuous structures labeled; a, antennule; b, rostrum or “spine”; c, eye; d, cephalothorax or “head”; e, carapace or “head shell”;/, abdomen or “tail”; g, telson or “tail spine”; h, uropod or “tail fin”; i, first pleopod or “swimming leg”;/, fifth pereiopod or “walking leg”; k, chela or “pincer” of third periopod; l, third maxilliped; m, antenna or “whisker”; n, antennal scale or acicle. and Sicyonia sp. They may be distinguished as follows: Macrobrachium is easily recognized from its presence in fresh or slightly brackish water. Only the first two pair of walking legs have pincers, the first being relatively very large. It supports a distinct river fishery of small volume not here considered. All of the other species belong to the Penaeidae and agree in having the first three pairs of walking legs, instead of only two, armed with pincers, a family character. These shrimp are characteristically salt-water forms, although they may be found in brackish bays at certain times of year. The two species of Penaeus may be told from Trachypenaeus and Xiphopenaeus by the fact that the rostrum is armed with spines both above and below. P. setiferus differs but little from P. brasiliensis. The most easily recognized distinguishing characteristic is to be found in the groove on either side of the rostrum. In P. setiferus these grooves, conspicuous along the side of the rostrum, become shallow as they reach the carapace and are soon lost. In P. brasiliensis, however, the grooves continue to the back margin of the carapace. (See figs. 3 and 4.) LIFE HISTORY OF THE COMMON SHRIMP 3 STATES MILLIONS OF POUNDS 5 10 15 20 25 30 35 LOUISIANA FLORIDA TEXAS GEORGIA MISSISSIPPI ALABAMA N. CAROLINA S. CAROLINA Figure 2. — Shrimp catch for the eight South Atlantic and Gulf States in 1930. A. B. C. Figure 3.— Dorsal view of the carapaces of A, the common shrimp ( Penaeus setiferus)-, B, the grooved shrimp ( P . brasiliensis ); and C, the sea bob ( Xiphopenaeus kroyeri) showing the presence of rostral grooves in the grooved shrimp and their absence in the common shrimp and the sea bob. 4 BULLETIN OF THE BUREAU OF FISHERIES Table 1. — Shrimp taken in the United States in 1930 according to districts 1 Section Catch Value Pounds 88, 117, 000 2, 785, 000 932, COO 341, 000 Percent 95.43 Dollars 2, 995, 000 49, 000 42, 000 29, 000 15, 000 4,000 Percent 95. 56 3. 02 1. 56 1.01 1.34 Middle Atlantic States... .. .... _ _ . .. .37 .93 147, 000 5, 000 . 16 . 48 .01 .13 92, 327, 000 100. 00 3, 134, COO 100. 00 1 No catch was reported in the Chesapeake Bay States or in the States bordering on the Great Lakes. Trachypenaeus shows a family resemblance to Penaeus but is much smaller and the rostrum, although armed with 7 to 9 spines above, is smooth below. Xiphope- naeus differs from all the other penaeids in having a rostrum as long or longer than the carapace and the fourth and fifth pairs of walking legs much elongated and slender. These four long legs, together with the two antennae or “feelers” projecting beyond the “head” gave rise to the term “six barb” among the French fishermen. This has been corrupted to “sea bobs”, and it is by this name that they are gener- ally known. Sicyonia is an uncommon small form read- ily distinguished by the very short rostrum and the crest which continues from the rostrum down the middle of the body. Of the penaeids, Sicyonia and Trachypenaeus are inci- dental only and of no eco- nomic importance. Xipho- penaeus may contribute 2 or 3 percent to the catch. It is too small to be used in canning but finds its way int° the fresh markets and dry- ing platforms. P. brasiliensis may at times be abundant but forms less than 5 percent of the total. The single species Penaeus setiferus accounted for about 90 percent of all the shrimp taken in the United States, or 100 of the 113 million pounds caught in 1929. Figure 4.— Lateral view of the carapaces of A, the common shrimp; B, the grooved shrimp; and C, the sea bob. In both the common shrimp and the grooved shrimp teeth appear on both the upper and lower surfaces of the rostrum, while in the sea bob they are absent from the lower surface. LIFE HISTORY OF THE COMMON SHRIMP 5 Table 2. — Shrimp taken in the South Atlantic and Gulf States for 1929 and 1930, arranged by States 1929 1930 Catch Value Catch Value Louisiana, , ... ... .. Florida Mississippi Georgia Texas. Alabama North Carolina South Carolina . Total Pounds 49, 455, 982 18, 618, 564 13, 101,450 12, 377, 619 9,415,317 4, 396, 400 897, 495 287, 711 Percent 45. 56 17. 15 12.07 11.40 8. 67 4.05 .83 .27 Dollars 2, 025, 336 879, 192 421, 491 581,015 327, 008 154, 139 30, 560 16, 625 Percent 45. 66 19. 82 9. 50 13. 10 7. 37 3. 48 .69 .37 Pounds 38, 664, 487 16, 848, 576 8, 489, 050 8, 852, 712 10, 189, 318 2, 982, 200 1, 298, 610 792, 733 Percent 43. 88 19. 12 9. 63 10.05 11.56 3. 38 1.47 .90 Dollars 1, 159, 626 635, 506 318, 871 334, 576 377, 016 97, 219 40, 752 31,814 Percent 38.71 21.22 10.65 11.17 12. 59 3.25 1.36 1.06 108, 550, 538 100. 00 4, 435, 366 99.99 88, 117, 686 99. 99 2, 995, 380 100.00 Figure 5.— Shrimp catch for the South Atlantic and Gulf States from 1880 to 1930. Another phase deserves emphasis. While the shrimp ranks high among the fisheries of the United States, it is preeminent among those of the South Atlantic and Gulf States. Here among eight coastal States it ranks first in value, being followed by the oyster and the mullet; and second in volume, being exceeded only by the menhaden, which, however, ranks only fourth or fifth in value. In volume the shrimp is followed by the oyster and the mullet. 6 BULLETIN OF THE BUREAU OF FISHERIES A fishery of this magnitude, the most important in the South, giving employ- ment to thousands of fishermen and factory operatives, deserves careful attention if it is to be maintained in this valuable state. The rapid expansion of the fishery during the past 40 years, throughout which time it has doubled in volume every 8 or 9 years, has suggested caution, and recently the men in the industry have taken the initiative in asking for investigation, fearing, apparently, that some of the fluctuations in the catch were the forerunners of serious depletion. The prime requisites for successful handling of the problems of the shrimp, as of other fishery resources not adapted to cultivation, are information as to the abundance and knowledge of the life history. The first is needed to tell when protection is required and the second to guide such protection into intelligent and effective channels. Unfortunately data for even the roughest analysis of abundance is lacking. State and Federal data give only the total catch and except for the last year or so no indication of the effort that produced the catch. Some private records alone serve to show the boat catches by which the effort could be analyzed, but these records are too scattered and cover too short a time to solve the problem. The outlook in this direction is, however, now more hopeful, for in Louisiana, a law, originally drafted with the assistance of members of our staff, in the future will give figures from which depletion may be detected as it appears. In consequence of these conditions we are unable to present in this report any data on the important question of abundance. On the other hand, we must have an adequate knowledge of the life history and habits of the shrimp in order that, should depletion become apparent, intelligent pro- tective measures may be applied at once. The only purpose of this preliminary report is to make available what information has been obtained against the time when it may be necessary to frame protective measures. Data on many points are far from conclusive but still may be useful for legislative action ; the paper must be considered solely as such a progress report in spite of the fact that the data at hand establish a more complete life history than has hitherto been available for any shrimp of economic importance. PREVIOUS WORK ON LIFE HISTORY At the time the present investigation was started the information available con- cerning the American penaeids, and especially Penaeus setiferus, other than that in purely systematic papers was very meager. Fritz Muller in 1863 had reported from Brazil the occurrence of a nauplius larva presumably belonging to the genus Penaeus. In 1883 Brooks obtained penaeid larvae at Beaufort, N.C., and was able to trace the main stages of the larval history. He assumed them to be the young of Penaeus brasiliensis, but neither he nor Muller established the specific characters for the larvae. This work was solely morphological, interest being centered in the occurrence of the primitive nauplius larvae previously found only in lower Crustacea. In this connection it may be well to point out a peculiarity of the Penaeidae which has made study of the life history difficult. Unlike the majority of the higher Crus- tacea, the female penaeid does not carry the developing eggs attached to the abdomi- nal legs. In consequence, although eggs of the other shrimp and crabs, as for example Crago vulgaris of the North Sea, may readily be obtained for study, no one has ever reported a fertilized egg of Penaeus. The mature ovarian egg is very small, measur- ing from one fourth to one third of a millimeter in diameter, and the unique larval history of Penaeus is related to this fact. The minute larvae hatching as nauplii pass through 6 or 8 distinct stages, in contrast to the condensed development of most deca- LIFE HISTORY OF THE COMMON SHRIMP 7 pods in which the eggs, well supplied with stored yolk, hatch into larvae correspond- ing to stages occurring late in the larval history of Penaeus. No further work was done until more than 20 years later at the Gulf Biologic Station, Cameron, La. In 1908 Spaulding and Guilbeau and in 1910 Gates reported on the shrimp. Facilities were not available at the Gulf Biologic Station for an inten- sive study of the shrimp; consequently these preliminary papers contain only brief observational notes on the behavior and a few scattered length measurements. Spaulding, whose observations were the most extensive, reports that larval shrimp, less than one half inch in length, were obtained in Calcasieu Pass in August and Sep- tember. These were similar in form to the adult, but a series of larval stages was not obtained nor was identification of the species possible. No females with devel- oped ovaries were taken, but from a study of the males and the time of appearance of the larvae he concludes that there is a single breeding season extending from the first of May to the last of July. He thought it possible that the eggs were laid in the Gulf but presents no direct evidence. The size at maturity is not established. In 1918 the Louisiana Department of Conservation began a more intensive study of the shrimp. This work was done by Percy Viosca and published partly under his name (1920) and partly under that of Tulian, who was then commissioner (1920, 1923, 1926). Viosca’s findings are more extensive than his predecessor’s but it is to be regretted that none of the data upon which his conclusions were based is given in any of the publications. As a result it is impossible to evaluate his interpretations of the life history. He states that Penaeus setijerus spawns in the Gulf, chiefly on the evi- dence that sexually mature shimp are found only in outside waters. The young are said to live in the plankton of the Gulf until a size of 1% inches is reached. “By May reasonable numbers of baby shrimp appear in the shallow waters near the coast line and a large proportion gradually migrate into brackish water, all growing rapidly throughout the summer.” This is all the information toward a life history of Penaeus setijerus available in 1930. The following is a bald summary of the life history as determined by the present work. The eggs are laid from March or April to August or September in the outside waters of the ocean or Gulf. The post-larval young beginning at a length of 7 mm are found in bays, creeks, bayous, and lakes in warm, shallow, brackish water with mud bottoms. The young grow rapidly and with increasing size gradually migrate to deeper water of greater salinity. During July at an average length of about 90 mm they enter the commercial catch, appearing first in the bays, creeks, and other “inside” waters and later outside. They continue in the fishery, furnishing all of the fall catch, with its peak in October, until the following spring and summer, when they spawn and disappear at the age of 1 year. By late fall they have reached a length of about 120 mm, which they maintain during the winter. Resuming growth in the spring, they show a rapid and striking differentiation in the size of the sexes and spawn at lengths of 130 to 170 mm for the males and 135 to 190 mm for the females. Their fate is unknown, but their complete disappearance from the commercial catch is undoubted. The breeding season is characterized by (1) a development of the gonads; (2) a rapid differentiation in size between the sexes; (3) a difference in the behavior of the 175006—33—2 8 BULLETIN OF THE BUREAU OF FISHERIES sexes, so that the proportion of sexes, uniform during the winter, shows wide fluctu- ation. Far more mature shrimp of both sexes are found outside than inside. The shrimp is most abundant in shallow coastal waters near river mouths or deltas. It is omnivorous, feeding on a variety of animals and plants. LIFE HISTORY NATURE OF DATA First in importance are the length frequencies obtained by measuring shrimp taken by the standard commercial gear, the otter trawl. All of the shrimp measured in Texas were samples from the catch of fishermen; the same is true of part of the data from Louisiana. The remainder of the shrimp were obtained with standard gear from boats operated by members of the staff in Louisiana, Georgia, and North Carolina. This experimental fishing differs from the commercial fishing solely in the localities fished. The fishermen obviously cannot afford to fish in localities yielding small catches, while in the experimental fishing these localities give information fully as valuable as those where many shrimp are taken. Of these length-frequency series the most important are those for Penaeus setiferus, although similar but far less complete data are available for Penaeus brasiliensis, Xiphopenaeus kroyeri, and Trachypenaeus constrictus. The series from Georgia is, because started earlier and unmarred by breaks, the most valuable. The total length measured from the tip of the rostrum to the tip of the telson has been used as the standard measurement. This measurement was chosen over other linear measurements such as carapace length and over weight and volume after a study of the variability of measurements obtained in various ways. Weight and volume are subject to great variation because of the adhering water on a body as irregular in shape as the shrimp. These measurements consume much time and require apparatus not easily used aboard small boats. In obtaining the total length the shrimp is placed, ventral surface down, on the scale of a special measuring board with the rostrum in contact with a fixed block. The measurement is read and recorded to the nearest millimeter. For plotting and analysis the measurements are first tabulated, then grouped into 5-mm classes in such a way that the mid-point always falls on a 3 or an 8. For example, all shrimp measuring from 131 to 135 mm (both figures inclusive) are grouped together in a class interval with a mid-point of 133 mm and so plotted in a frequency polygon. Similarly, all from 136 to 140 mm are grouped with a mid-value of 138. Measurements other than total length have been made for special purposes and in smaller numbers; most notable are those for racial studies now in progress. The measurements from the four localities, covering, in the case of Georgia, over 2 years, now number more than 150,000 and therefore represent an adequate basis for statistical treatment. (See table 3.) LIFE HISTORY OF THE COMMON SHRIMP 9 Table 3. — Shrimp ( Penaeus setiferus) measured during semimonthly periods in the three major regions of investigation Period Feb. 1. Feb. 2_. Mar. 1. Mar. 2.. Apr. 1.. Apr. 2.. May 1_. May 2.. June 1_. June 2_. July 1.. July 2. _ Aug. 1.. Aue. 2.. Sept. 1. Sept. 2. Oct. 1.. Oct. 2__ Nov. 1_. Nov. 2.. Dec. 1.. Dec. 2.. 1931 Subtotal. Georgia Males 547 1,021 476 559 819 732 1,222 1,012 604 81 453 427 918 1,730 1.551 1.552 664 1,329 632 1, 158 1,737 19, 227 Fe- males 638 1, 111 563 627 830 821 1,314 1,287 1,010 411 411 373 521 1,398 1,467 1,720 752 1,389 679 1,323 1, 862 20,507 11,775 Louisiana Males 155 37 530 969 964 1,996 1, 972 1, 151 1,339 1,994 319 349 Fe- males 139 50 476 788 889 2,004 2, 028 1,308 1, 261 2,206 362 451 11,962 Texas Males 495 575 311 494 69 222 433 315 488 598 588 394 463 439 348 Fe- males 536 525 289 506 131 178 367 310 512 602 612 406 537 561 452 6, 232 6, 524 Period Georgia Louisiana Texas Males Fe- males Males Fe- males Males Fe- males 1932 Jan. 1 456 528 328 317 178 222 Jan. 2 476 561 172 228 Feb. 1 301 343 777 823 Feb. 2 275 338 193 207 Mar. 1 507 514 485 607 291 309 Mar. 2 457 526 361 378 186 214 Apr. 1 1,260 1,322 272 265 420 480 Apr. 2_ _ . _ 1,348 1,592 126 82 101 99 May 1 _ 125 212 326 285 118 282 May 2_ 662 931 233 248 393 407 June l 349 596 73 49 177 223 June 2 314 688 248 290 101 99 July 1 373 356 237 270 314 286 July 2 963 683 255 287 306 294 Aug. l._. - _ 1,017 704 217 204 792 608 Total 28, 110 30, 401 14, 936 15, 244 10, 751 11,305 Beside length measurements, all observers have recorded the conditions of the gonads, permitting these specimens to be classified not only by length but also by stage of sexual maturity. The field examination has been by naked eye, but in addition all the observers have made microscopic examination of part of the specimens and preserved material for future study as a check on the field classification. Smaller shrimp than those taken by the commercial gear have been seined or taken by small beam trawls or "try nets” with bobbinet casings. These shrimp, ranging from 7 to 20 mm, have been measured under the low power of the microscope with an eyepiece micrometer. Of a different nature is the material obtained by plankton nets. The study of the larval stages thus obtained is not yet complete. INTERPRETATION OF DATA RECOGNITION OF AGE GROUPS The interpretation of the length-frequency data from Georgia upon which, as our most complete series, most of our knowledge of the life history is based, may con- veniently begin with those for July 1931. Here there is a well-marked bimodality of the curve due to the presence of two completely separated groups differing markedly in size. This is plainly shown in figure 6, in which the smaller group with a mode at about 95 mm extends from 65 to 125 for both males and females, while the larger group, centered about 153 mm for the males and 175 for the females, extends for the males from 140 to 170 mm and for the females from 150 to 190 mm. There is thus a gap of 20 mm between the two groups. At no other time of the year can two com- pletely separated groups be found. Examination of the other records show an essentially similar condition during July in Louisiana and during August in Texas. 10 BULLETIN OF THE BUREAU OF FISHERIES Let us first trace these groups backward to see whence they come. In the case of the larger group this is easily done. Neglecting minor changes of size, this group can be traced back through the spring and winter to the previous year, where we may leave them for the present. (See fig. 10.) The smaller group is less easily followed. They are not present in the commer- cial catch in the earlier months. An extensive search for this group of young has been carried on by seining in shallow water, representing many degrees of salinity, LENGTH IN INCHES Figure 6. — Half-month frequency distributions of the common shrimp from the outside waters of Georgia, July 1 to August 31, 1931, showing the first appearance of the young shrimp of the year and the disappearance of the mature shrimp from the commercial catch. from ocean beaches to creeks and lakes far inland. Many localities yielded nothing. So far no young have been found on ocean or Gulf beaches either in Georgia, Louisiana, or Texas. In the marshes near the grass small shrimp were taken but proved to be young and adult palaemonids. Finally, in the summer of 1931, the young were found in the slightly brackish waters of Dover Creek at Lampadozier, Ga., 10 or 12 miles from the sea. Here the young from 7 or 8 to 40 mm have been systematically collected during the proper seasons. In Louisiana during 1932 the young have been obtained under a variety of conditions ranging from the slightly brackish waters of Bayou Rigaud in lower Bara- taria Bay to the almost fresh waters of Little Lake. The latter is more than 20 miles LIFE HISTORY OF THE COMMON SHRIMP 11 distant from the Gulf in a direct line and much farther by the winding bayous and canals through which the shrimp would have to pass if, as seems probable, they were spawned at sea. Also during 1932 these young shrimp have been found in similar localities near Beaufort, N.C., and Aransas Pass, Tex., consequently records are now on hand for these young from the four States in which the investigations are in progress. The habitat of the post-larval young up to about 50 mm seems to be on the inside and appears to be characterized by shallow water, muddy bottoms, lugh temperature, and moderate to very low salinity. A more detailed study of the physical characteristics of the localities in which young shrimp are found is in prog- Figuke 7.— Frequency distributions of the immature and maturing females of the common shrimp in Georgia waters during the spawning season of 1931. Solid lines represent distribution of female shrimp with ovaries composed exclusively of undeveloped eggs and broken lines indicate distribution of shrimp with maturing eggs. ress. These bayous, creeks, and lake margins are also nursery grounds for many species of fish, the larvae and young of which are taken while seining for the young shrimp. While Crago, the shrimp of the North Sea, differs in many respects from Penaeus, the young also make their early growth far from the sea (Ehrenbaum 1890). SPAWNING What is the origin of these very young, apparently just emerged from the larval state, some of which can be found from April through September? We naturally turn to the larger group which we have said was present at this time. If we classify the large shrimp by the observed state of the gonad, we have the following picture. In Georgia, in 1931 from the middle of April to the end of July, some degree of matu- rity was observable in the ovaries. The members of the small group appearing in July were without exception immature. Some scattering representatives of the 12 BULLETIN OF THE BUREAU OF FISHERIES larger mature group were found in August and even September, but the score of individuals taken are in strildng contrast with the hundreds obtained in earlier months. Although we have distinguished four or five stages of ovary development in our records, it is difficult to separate these sharply or to determine how long each lasts. Judging from the appearance of the ripest stages that we have recognized and of the young, some shrimp are spawning within 2 or 3 weeks from the date on which ovary development is first noticed. We may say, therefore, that in Georgia in 1931 the larger group was spawning from the last of April to the middle of August and that the young entering the commercial catch in July came from the earlier spawners among these large shrimp. The intervening time had been spent by the young probably in the smaller and more brackish creeks in feeding and growing. Having thus established the previous history and relations of the two groups found in the commercial catch of July, we may profitably complete the presentation of all the data on breeding and spawning that we have obtained. Like Ehrenbaum in his work on Crago vulgaris, we must confess to serious gaps in this phase of the life history. The Penaeidae as a group are characterized by marked sexual dimorphism, including certain elaborate and well-marked secondary sexual characters associated with sperm transfer. In Penaeus setiferus, in contrast to the crabs, the female is much larger than the male. Thus, in Georgia, in June 1931, males and females of a comparable degree of maturity averaged 144.5 and 156.5 mm, respectively, the females thus ex- ceeding the males by 8.31 percent. There are other differences in general form; according to Alcock (1906) the rostrum of the female is proportionally larger, a per- sistent juvenile character. The most interesting differences for the question that we are now considering are those of the secondary sexual characters. In Penaeus the males produce sperma- tophores, often of considerable complexity. The inner ramus of the first pair of pleopods is modified, forming a structure called the petasma or andricum. By means of the petasma the spermatophore is presumably transferred to the female, although the process as far as we know has never been observed. In Penaeus setiferus the endopodite of the second pair of pleopods is also modified to some extent and probably also plays a part in the transference of the spermatophore. In most species the female has on the ventral surface of the thorax between the bases of the last pair of pereiopods a structure of plates partially enclosing a space into which the spermatophore is placed. This thelycum varies greatly in complexity, in some species ( Penaeus brasiliensis and Xiphopenaeus kroyeri ) being well formed, in others less developed. In Penaeus settferus the thelycum is vestigial in character and does not function as a secondary sexual organ. The spermatophore is attached, by a gluelike secretion of the male, to the ventral surface of the female between the bases of the third and fourth pereiopods. Two winglike processes on the anterior end of the attached spermatophore fit securely into the grooves between the third and fourth pereiopods and form an inverted funnel into which the "glue” flows, fastening the spermatophore to the female. The spermatophore so attached is quite easily dislodged. This may, to a certain extent, explain why in Penaeus setiferus spermatophore-bearing females are much less abundant in the catches than are similar females of Penaeus brasiliensis and LIFE HISTORY OF THE COMMON SHRIMP 13 Xiphopenaeus kroyeri. In fact, only 20 such females have been obtained from the thousands of shrimp examined. Considering only females handled during those months in which spermatophores have been found, we have the relations shown in table 4. Table 4. — Penaeus setiferus spermatophore records April to August 1981 and April to July 1932 State Females examined Sperma- tophores found Percent Georgia 11,015 3 0. 027 Louisiana - 3, 561 7 .20 Texas - 3,911 10 .26 All these 20 spermatophore-bearing females were in the ripest of the stages dis- tinguished by examination of the ovary. Their lengths, which range from 141 to 186 mm with a mean at 166.3, agree well with the lengths of females showing the last stages of maturity. The period in which these females are found is from April to August, inclusive, which covers almost the entire season during which the ovaries appear ripe. All were taken in fairly deep water, either in open sounds or the sea or Gulf, in localities where the salinity is the highest found in the range of the shrimp for this season of the year. The next question is as to the time during which the male is capable of furnishing spermatophores. The spermatophores when formed are easily recognized in, or expressed from, the lower end of the vasa deferentia of the male, and records of their presence or absence are at hand for part of our series of length-frequency data. In 1932 in Louisiana maturing males were first present in the latter part of March, thus appearing slightly in advance of maturing females. They usually appear slightly earlier and last as late in the season as mature females. We may conclude, therefore, that during the period that females are capable of spawning, breeding males are also present. To summarize: Ripe females are common during April, May, June, July, and present but scarce in August and September (Georgia, 1931). Spermatophore- bearing females have been obtained from April to August (these must be within a few hours of spawning). Mature males are present throughout this period. Young in the first post-larval stages are found from at least late April to August. We feel justified in assuming, on the basis of these data, that in 1931 spawning occurred during April, May, June, July, and to a reduced extent in August and September, even with- out having as yet observed spawning or having obtained fertilized eggs. Of course, the dates of spawning will vary from season to season. It may next be inquired, Where are the shrimp spawning grounds? On this important question our evidence is again indirect, consisting of four types of data — the relative distribution of males and females, the relative distribution of the various stages of maturity, the distribution of that stage of maturity represented by the sper- matophore-bearing females, and the distribution of larvae. The plankton material containing penaeid larvae has not yet been analyzed and therefore cannot throw light on the location of the spawning grounds. As stated, all spermatophore-bearing females have been obtained in fairly deep water of high salinity. Since we have good reason to think that the spermatophores 14 BULLETIN OF THE BUREAU OF FISHERIES are soon dislodged and the eggs must be laid and fertilized before this happens, we may infer that spawning occurs in these localities. There remains to consider the distribution of the sexes and the stages of their maturity. As the simplest we shall first consider the distribution of the stages of sexual maturity. If the numbers of immature and maturing males or females from different localities are compared, the greater proportion of maturing individuals will be found in outside waters. (See fig. 8.) This is more marked with the later stages of maturity. Thus, in Georgia, in 1932, an intermediate stage of ovary development was over twice as common in outside waters as in the creeks, while the last stage of ovary development recognized was six and one half times as common. As will be recalled, the spermato- phore-bearing females, which rep- resent the last stage of maturity, were all found in the Gulf or ocean with a single exception taken in an exposed sound. A satisfactory analysis of these data will require the inclusion of figures from another breeding season and must be presented at a later date. The present data can be interpreted in but one way — that by far the greater proportion of spawning takes place in the Gulf or ocean and that only a slight amount, if any, occurs on the inside. The spawning seems to be correlated with high salinity. SEX RATIO Figure 8.— Percent of maturing females in the inside and outside waters of Georgia from March to July 1931. A comparison of the propor- tion of the two sexes in the catch at different times of year and in different localities shows fluctuations associated with spawning. Since additional data will be required for a satisfactory analysis of these fluctuations, only brief mention of the facts will here be made. As may be seen from figure 9, in Georgia from February 1931 to July 1932 there is a well-marked annual cycle. During the winter (from the last half of September 1931 to the last half of April 1932, inclu- sive) the proportion is constant, the females constituting slightly over 52 percent of the 25,601 shrimp measured. During the similar part of the preceding year the females furnished essentially the same proportion of the 11,283 shrimp. In contrast, the period from the last half of May to the first half of September 1931, inclusive, and again from the first half of May to the first half of August 1932 (where, at this time, our record ends) there are sudden and remarkable fluctuations of the proportion, the females rising above 80 percent and falling nearly to 30. This period covers the time during which the adults mature, spawn, and disappear. A few weeks after spawning begins and when about half the females are maturing the LIFE HISTORY OF THE COMMON SHRIMP 15 fluctuations begin and only end with disappearance of the adults from the fishing. The rise in proportion of the females succeeded by a fall appear to be the result of differential behavior of males and females associated with spawning. We have found no indication of a sex reversal in Penaeus such as has been claimed by Berkeley for Pandalus. Further data is being collected on this interesting question. Figure 9.— Percent of females appearing in catches from inside and outside waters of Georgia from February 1, 1931, to August 15, 1932, during which period 58,521 shrimp were measured and examined. LARVAE The study of the larvae of Penaeus by Muller (1863, 1864), Brooks (1883), and Kishinouye (1900b) have shown the main features of the unusual larval history be- ginning with the nauplius, not found elsewhere among the higher Crustacea. There are a large number of stages (6 to 8) including nauplius, metanauplius, protozoea, zoea, metazoea, and mysis. In no case has the species of the larvae been known with cer- tainty, and in consequence there exists no data for the identification of the larvae. This is a serious lack where 4 species of penaeids appear in the commercial catch and 15 or 20 have been recorded, the larvae of which may be encountered. Other facts indispensable to this phase of the life history, such as the duration of the stages and their behavior, are also lacking. The extensive plankton collections of the investigation have not yet been studied in detail, and although larvae are present we are as yet unable to give a satisfactory summary of their distribution. YOUNG From the distribution of the mature adults, spawning must take place predomi- nantly, if not exclusively, in outside waters of high salinity. What little is known of the distribution of the larvae does not conflict with this. The distribution of the post- larval young is, however, different. These forms, essentially similar to the adult except in size, the smallest that we have obtained measuring 7 mm in length, give up the swimming habitat of the larvae and seek the bottom, thus resembling the adult in habits as in form. The young, as has been stated, are found on muddy bottoms in shallow water of high temperature and low salinity exclusively in bays or inside waters and often far from the ocean or Gulf. The upper limit of size of the young in 1G BULLETIN OF THE BUREAU OF FISHERIES this habitat is not sharply defined; some shrimp as large as 100 mm have been en- countered, but the majority are below 40 mm. We have no data on the growth in this earliest stage, but they first appeared in the Georgia commercial catch at an average size of 90 mm (70 to 105) in July 1931, and again in July 1932, 2 or 3 months after the beginning of spawning. The reactions that lead the larvae from the outside more saline waters to the brackish muddy-bot- tom waters favored by the young, or lead the young back to the waters of higher sa- linity where they first enter the commercial catch, have not been studied, although they must hold the key to an understanding of one of the most interesting phases of the shrimp life history. To compare with this picture just given of the spawning, larvae and young of Penaeus setiferus, we may review the findings of Ehrenbaum (1890) on Crangon ( Crago ), Mortensen (1897) on Palaemon, and Kishinouye (1900a), Spaulding (1908), Gates (1910), and Yiosca (1920, Tulian 1920, 1923, and 1926) on Penaeus. Ehren- baum found that in the North Sea the ovigerous females of Crago and the small free swimming larvae are found only in the strongly salt water of the flats and about the offshore islands of the North Sea. Only very exceptionally are larvae or ovigerous females taken in the brackish waters of the estuaries. In sharp contrast, the young, measuring 5 to 10 mm, are found in great numbers from spring to late summer in the Dollart and Jade Rivers, far upstream in brackish water. From the percentage of ovigerous females found at different seasons, Ehren- baum concluded that there were two spawning seasons and that hatching occurred primarily in July and to a less extent in March. No additional support for the double spawning season has been presented by the later writers. In Palaemon fabricii, studied by Mortensen, the females spawn in deep water in May, the eggs hatching chiefly in June. As in Crago the larvae are pelagic and never appear in the brackish water of fjords or creeks. After a pelagic life of 3 to 5 weeks, the first young appear early in July. The young are found in the rapidly growing vegetation of the shallow water of little creeks and fjords. The larger females are said to lay a second batch of eggs shortly after the hatching of the first. In neither of these species has fertilization been observed. It will be seen that the distribution of the ripest females, the larvae, and the young in both these species agree with what we have observed in Penaeus. The data of Kishinouye (1900a) are of more interest because they deal with various Japanese species of Penaeus closely related to the forms we are studying. Unfortunately, the basis of his statements are not given in the English summary of his paper. He gives tables for sexual maturity in five species. The period of May to September, inclusive, contains all these mature seasons, although most are individual- ly shorter than this. In three species the males are said to be sexually mature throughout the year. The eggs ripen in spring and spawning takes place in summer and autumn. Kishinouye describes the spermatophores in several species. Eggs are said to be discharged from time to time as they ripen. He mentions finding the larvae in shallower water than that in which the spawning shrimp live but does not specifically describe the habitat of the young. The breeding season given by Kishinouye agrees in general with our findings. It is not true of the male of Penaeus setiferus that mature individuals are present at all times of the year, as he claims for certain species. We are very much inclined to doubt that in Penaeus setiferus spawning takes place from time to time as the eggs ripen; it is more likely that all the eggs are spawned at one time. LIFE HISTORY OF THE COMMON SHRIMP 17 Only those findings of Spaulding, Gates, and Viosca relating to the spawning and young will be given here. Spaulding thought that the breeding began about the middle of June and that spawning took place “in the deeper water of the larger bays or even in the Gulf.” He interpreted his data as showing either a long breeding season during the summer or two breeding seasons a year. Two years later (1910) Gates reported that there was only one breeding season and that “this extends from, approxi- mately, the first of May to the last of July.” Our investigations show that the period given by Gates covers the most intensive portion of the spawning season and that the entire season may be longer by a month or two on either end. Contrary to Gates’ assertion that “the smaller the shrimp the earlier it spawns”, we find that the larger shrimp mature first, consequently they also probably spawn first (fig. 7). Viosca states that Penaeus setiferus spawns in the Gulf, chiefly on the evidence that sexually mature shrimp are found only in outside waters. The young are said to live in the plankton of the Gulf until a size of 1% inches is reached. “By May, reasonable numbers of baby shrimp appear in the shallow waters near the coast line and a large proportion gradually migrate into brackish waters, all growing rapidly throughout the summer.” That the sexually mature shrimp are more abundant in the Gulf than in the inside waters agrees with our experience, but that no mature are taken in the bays is too sweeping a statement. Although always strikingly less abundant, especially in the more mature stages, we have records of many shrimp in advanced stages of ovary development from the inside waters both of Louisiana and of Georgia. Viosca was led, apparently, to infer that the larvae live in the plankton up to a length of 1 % inches by his inability to obtain smaller young. The largest larvae found by Muller were 4.5 mm long, the smallest post-larval young about 5 mm. We have obtained young as small as 7 mm. Obviously the larvae do not reach a length of iy4 inches (31.75 mm). Either this size was inferred but not actually seen, or, if seen, some other crustacean was mistaken for Penaeus, We have been unable to find young of this size on the outside, either in the plankton, the trawl material, or on the ocean beaches, although we have obtained them in inside waters in all four States in which we have worked. Although a gradual inward migration of young such as pictured by Viosca would be a reasonable process, the presence of many young of 7 mm and up in the creeks, bayous, and lakes, and their complete absence on ocean beaches show that the larvae often pass directly to shallow brackish waters. Like Viosca we have found that the larger young of 20 to 50 mm move, in general, seaward through the summer and fall, so that there is always a gradient of decreasing size from the waters of greater salinity toward fresh water. Owing to the extended spawning season, the advent of the young into the com- mercial catch is followed week by week by new contingents which move into the fishing zone and become large enough to be taken by the trawls. As a result, the lower limit of size in the commercial catch is approximately constant through the summer and into the fall, as might be expected, since the limit is set by the selective action of gear and the movements of the young and not by growth. It may be affected, how- ever, in a minor way by changes in temperature or by changes in the fishing areas which are not always strictly comparable from month to month. 18 BULLETIN OF THE BUREAU OF FISHERIES GROWTH Because of the fact that young are constantly being added to the population sampled by the fishing, it is extremely difficult to approximate the growth. As the earliest spawned shrimp continually increase in length, and small of a relatively uniform size are constantly appearing, the range of sizes, of course, increases ; consequently the growth is not represented by the mean or modal lengths, which each month are based on a new assortment of ages. It is clear that the increase of the means is much slower than the growth of shrimp of any particular age. The advancing upper size limit of the entire group should measure the growth of the older shrimp but is difficult in practice to determine. No statistical measure exists which will be comparable for all these types of compound frequency curves. In discussing the series of length frequencies, we have used the mode, or most abundant size as determined by inspection, for the average, and for the upper and lower limits a point where the ordinate has fallen to 1 percent of the total number of measure- ments. Admittedly inaccurate, this may serve, however, as a first approximation for the purposes of the present discussion. The length-frequency data obtained with a standard trawl in the ordinary fishing grounds is ill adapted to show growth. For the present, therefore, only the most general features will be sketched and a more detailed analysis postponed to a future time when another season’s records are available. The average size of the young at the time of appearance in July 1931 was about 90 mm, with a range from 70 to 105. (See fig. 10.) At this time there is no significant difference in size between the sexes; in fact a clear differentiation does not occur until the following April. Through July and August growth is rapid; and by the 1st of September the most common size is about 130 mm, while the largest exceed 160 mm. The lower limit rises slightly and remains at approximately 80 mm until the first of the following year. We may consider that with the present gear about 70 mm is the minimum size entering the commercial catch to any extent. During September, October, and November the average length fluctuates between 120 and 130 mm. During this period the females seem slightly the larger. This is the time of the fall maximum of abundance and the peak of the average commercial catch. From a value of 130 mm early in November the average size falls slightly, reaching a low point of about 120 mm near the first of the year. The difference in size between the sexes also disappears. This corresponds in a general way with the low period of the commercial catch. Obviously these decreases of average size must result from behavior, which is differential with regard to size — the large and the small shrimp must frequent dif- ferent localities or react differently in other ways. It is suggestive that the winter decrease of size is accompanied by a very marked decrease of abundance. During the low temperatures of the winter, movements or habits of the shrimp are changed in such a way as to remove the majority of them from the field of the commercial fishery on the Georgia coast. Following the first of the year (1932) the lower limit begins slowly to rise and by the 1st of April has increased from 80 to 100 mm. It is apparent that by January the last shrimp spawned in the preceding summer have entered the commercial catch and that the increase of the lower limit now reflects the growth. This seems the more probable, since the average advances at the same rate from 120 to 140 mm during this time. 10 0 20 10 0 30 20 10 0 30 20 10 0 30 20 10 0 30 20 10 0 (0 0 10 0 10 0 10 0 10 0 10 0 10 O' .—Frequency distribution of the population of common shrimp in hnlf-month periods for Qeorgia from February 1, 1931, to August 15, 1032. The it tho measurements of 58,621 shrimp, With but fow variations the distributions for Louisiana and Texas follow those for Georgia very closely . Theso curves repre- 175000—33. (Face p. 18.) LIFE HISTORY OF THE COMMON SHRIMP 19 During the following months of April, May, June, and July 1932, accompanying a steady increase of water temperature from 56° F. in March to 86 in July there is a remarkable exhibition of growth. This is manifested in three ways. The steady ensuing growth is differential, being more strongly expressed in the females than the males; there is a surprising reduction of the size range; and the sex products mature rapidly. As will be recalled, this period coincides with the time during which mature shrimp are taken. These first appear among the largest; but soon the smallest have exceeded the minimum size for sexual maturity, maturing becomes general, and by the 1st of July no immature of any size can be found in this group of shrimp. The progress of growth is very interesting. During the spring, in each sex a dis- tribution of wide range without a sharp mode is converted into a compact group with little variation of size and a well-marked mode. This may be caused by either or both of two processes — first, the more rapid growth of the smaller and younger members; second, the disappearance of the larger individuals that have spawned. Whatever the process, its orderly progress is striking. In fact, a frequency distribu- tion from May or June may readily be identified by its characteristic shape. Although up to April there has been no significant difference in size between the sexes as evidenced from the frequency curves (except for the short period in the fall of 1931 already noted), the growth of the females, beside producing a characteristi- cally compact and homogenous group of this sex, rapidly outstrips that of the males until in June, as previously stated, the sexes differ in length by 8 percent, the males averaging 144 and the females 156 mm. By this time the range has been so reduced that the ninth decil of the males (the largest males remaining after the upper 10 percent have been cut off) is below the average of the females and the first decil of the females is above the average length of the males. FATE OF ADULTS A question of great interest and importance both from the theoretical and the practical standpoint is the fate of the larger group present in July. The smaller group of shrimp, traced into earlier months, was found to have been spawned by the larger, and the larger to have overwintered from young of the previous spring or summer. Traced onward from July the smaller group persists through the winter and spawns the following spring and summer. Does the group of larger shrimp also survive the winter and take part in a second spawning? Again our evidence is indirect but impressive. We shall consider the length fre- quency and sexual maturity data in the light of the abundance of the shrimp. No accurate measure of the abundance of the shrimp in the waters of Georgia at different seasons is available. An approximation, however, is furnished by the number of shrimp taken in experimental trawling. By these useful, although imperfect, data let us follow the abundance of the group of small shrimp from the time of their entrance into the commercial catch until their disappearance. The numbers increase rapidly, being constantly augmented by new young of later and later hatchings. The maxi- mum of abundance is reached in the fall, usually September or October, agreeing in time with the fall peak of the commercial catch. (See fig. 11.) Although the fall peak of the total catch tends to be overemphasized by the intensity of the fishing at this season and in general economic factors prevent a complete correspondence, nevertheless the fall catch rests on a period of marked abundance of shrimp. From 20 BULLETIN OF THE BUREAU OF FISHERIES this peak the abundance declines and reaches a low point in late winter and early spring, coinciding with the low point of the catch which falls in February, March, or April. The numbers obtained by experimental trawling then again increase, reaching a second peak in April or Maj7 at the time of the spring catch of large, mature shrimp so prized by the canners. The crest is more prominent and earlier than the spring catch would suggest. From this peak of abundance there is a rapid and steady decline. In the first half of July, when the young of the year make their first appearance in the commer- cial catch, the two groups are approximately equal in abundance. By the latter part of the month the abundance of the group of larger shrimp has already fallen >- ■< 2 Ol. uJ tO Figure 11— Average monthly shrimp catch in Georgia from 1926 to 1931, inclusive. The average monthly polygons for Louisiana and Texas follow the Georgia polygon very closely. below that of the growing young. The group of large shrimp can be more or less clearly recognized in August by their greater length, but the rapid growth of the young (see fig. 10) soon brings about an overlap, so that size alone will not suffice to identify them. The group of large shrimp are at this time sexually mature, and by this criterion a few may be found in September, but the number is very small — two or three out of hundreds. A careful study of the size frequency curves and those of sexual maturity fail to disclose further trace of them, and we are forced to conclude that they disappear from the fishery in Georgia. We have also been unable to find spent individuals such as should be present after the spawning season if the adults remained within the range of the fishery. Among the thousands of immature shrimp no undoubtedly spent female has been found, although some males examined may possibly belong to this category; the number of these, however, is small. LIFE HISTORY OF THE COMMON SHRIMP 21 Kishinouye (1900a) believes that certain species of Penaeus die after spawning while others live through a second year. He does not indicate the basis of this belief. Without evidence, it is idle to speculate whether the process of spawning is fatal to the shrimp — whether, like the salmon, they are weakened and soon succumb to enemies and disease or whether they retire to some new and as yet undiscovered habitat. All we know is that they play no further part in the fishery. The shrimp is therefore an “annual”, being spawned in the spring or summer, spawning at the same season of the following year, and then passing out of the fishery. Since it is probable that those spawned early furnish the early spawners of the next year, the life span is very close to 1 year. This is the condition in Georgia. Whether or not in Louisiana and Texas the life span will prove to be the same is not yet settled, but the data so far available give no evidence of an essentially different life history. It should again be emphasized that the above account refers only to P. setiferus. How many of these findings may apply to P. brasiliensis, Xiphopenaeus kroyeri, and Trachy penaeus constrictus remains to be determined. A consideration of these forms has been deferred because of their lesser importance and because data on these less common species has accumulated far more slowly. HABITS Observation of the shrimp under natural conditions is so difficult that only the most obvious habits can be thus determined. Systematic observations in aquaria have not yet been carried out. In spite of the paucity of direct observation it may be well to assemble what general knowledge is available. As already stated, Penaeus setiferus, like other penaeids, is a shrimp of subtropical littoral waters. The range of temperature under which we have taken it is 9° to 31° C. (48° to 88° F.). In the colder waters of the North Sea in Europe or of Alaska no penaeids are taken, the fishery depending rather on pandalids or cragonids. As with the Indian species observed by Alcock, P. setiferus appears to favor the deltas of large rivers like the Mississippi, although it is not confined to such localities. For example, shrimp are taken in large numbers off Cape Canaveral, Fla., in a region where there are no large streams. Perhaps the relation to rivers rests on the fact that, as we have seen, the young for part of their development frequent brackish or almost fresh water. It appears to be more common on muddy than on sandy bottoms, but it must be remembered that mud is more common than sand in the neighborhood of large deltas. Like other shrimp, Penaeus setiferus swims in two ways. Usually it swims for- ward by means of the pleopods or abdominal legs. When moving rapidly it swims or rather leaps backward by flexing the powerful muscular abdomen and sweeping the large tail fin under the body. In this way it may leap out of the water like a fish or out of an uncovered aquarium. Although the shrimp spend most of their time on the bottom, there are a few well-authenticated cases where a school has been seen swimming near the surface. At times they bury themselves in the mud. No systematic examination of the stomach contents has been carried out, but a few observations indicate that it is a voracious and well-nigh omnivorous feeder. This is supported by the observation of Viosca (Tulian 1920), and of Alcock (1906) and Kishinouye (1900a) on other species. Worms, Crustacea (not excluding shrimp of the 22 BULLETIN OF THE BUREAU OF FISHERIES same species), small mollusks, and plant debris are all eaten. Often it appears to eat the mud or sand for the organic matter which it contains. The presence of consider- able amounts of mud or sand in the intestine appears to be the rule. In aquaria, fish or other shrimp are readily and successfully attacked and eaten. Convincing evidence is at hand that the behavior of the shrimp is influenced by temperature and salinity, but as yet no observations have been made on its reac- tions to these stimuli or to others such as light, hydrogen-ion concentration, and oxygen tension. Early work on these important features of the shrimp’s behavior occupies a prominent place in the program of the cooperative investigation. DEPLETION AND PROTECTION The shrimp investigation was initiated by the Federal and State Governments to supply the biological information necessary to guide analysis of the state of the fishery and to permit the framing of effective protective legislation when such is needed. It is not possible here to discuss in detail the bearing of the facts presented in this report upon the questions of depletion and protection; the intention is to do this in a subsequent publication. Here we may only indicate the lines which such a discussion must take. First, it should be emphasized that depletion can only be detected by a careful analysis of the abundance of shrimp and that knowledge of abundance requires ade- quate statistical data. Existing catch statistics are inadequate, since they do not show the effort by which the indicated total catches were obtained. In addition we must know the amount of gear and number of men, or the individual boat catches must be recorded for analysis. Improvements in the method of gathering statistics have recently been made by Louisiana, and it is to be hoped that all of the eight South Atlantic and Gulf States will so modify their regulations covering the reporting of fish taken as to make possible the future analysis of abundance. Fortunately it is not necessary to take an alarmist attitude, as we have obtained no evidence of serious depletion. At the same time common prudence should make impossible a complacent inaction until depletion is easy to see and hard to remedy. While there is yet time a concentrated effort should be made to institute the collection of statistics which may be used effectively in the immediate future. In the absence of evidence of serious depletion at the present time we are unwill- ing to urge increased stringency of existing regulatory measures. We shall therefore merely suggest the bearing of the facts of the life history of the shrimp on the prob- lems of protection and indicate the types of restriction which may be employed when analysis of the catch statistics indicates that additional protection is needed. The outstanding features of the life history of the shrimp which may affect the question of its resistance to overfishing are: (1) Its short life span of 1 year, (2) its extended breeding season of about 5 months, (3) the very large number of eggs pro- duced, and (4) its extensive habitat in the littoral waters of the South Atlantic and Gulf. The shrimp’s life span of 1 year, shorter than that of any other important eco- nomic animal, is clearly an unfavorable factor. The shrimp fishery must obviously lack the stability shown, for example, by the halibut fishery, where animals from 7 to 20 years of age appear in the catch and the entire failure of the young of any particular year would not seriously reduce the total. When the pressure of over- LIFE HISTORY OF THE COMMON SHRIMP 23 fishing begins to be felt the shrimp will show sudden and violent fluctuations probably disastrous to the fishery. Its extended breeding season of at least 4 % months, longer than that of the majority of animals of economic importance, and the large number of eggs produced are factors very favorable to the shrimp, since they render almost impossible the failure of an entire breeding season. The South Atlantic and Gulf coasts, cut by a multitude of tidal rivers and bayous, furnish an unequaled extent of shallow, littoral waters of all degrees of salinity in which the shrimp flourishes. This extensive favorable habitat and the wide range extending from Massachusetts south to Brazil is an obviously advantageous condition. Without it, the multitude of shrimp would not have developed in this region. But while it may furnish a breeding reserve in places not readily fished, it affords no com- plete protection. The shrimp seems to be confined wholly to a narrow coastal strip favorable for fishing, since, as we have seen, Penaeus setiferus appears to be absent from deeper water. Certain of the creeks and bayous in which the young are found are unsuitable for fishing; but since these young all move to the larger bays or to outside waters during the winter, at some time all shrimp on our coast must run the gauntlet of trawls and seines. The number that escape depends wholly on the intensity of the fishery. It is clear that under natural conditions the favorable factors were so far in the ascendancy as to produce an amazing abundance of shrimp. Even the great numbers taken by man have not sufficed to produce an alarming depletion. We must look forward, however, to a time when the increasingly intense fishery will turn the balance against the shrimp. Then the catch, maintained by constantly increasing effort, will begin to show great variations from year to year, some of them ruinous to fisher- men and canners. When additional protection becomes necessary it may take one or more of four lines- — limitation of sizes taken, closure of certain seasons, closure of certain areas, and regulation of gear. All of these have been tried; the regulation of gear and the closure of certain areas, as, for example, nursery grounds and inside waters furnishing a preponderance of young, and certain seasons promise the most satisfactory results. Experiments now in progress will, it is hoped, indicate how gear may advantageously be modified. SUMMARY The present paper contains the results so far attained in certain phases of a cooperative program of study of the shrimp initiated by the United States Bureau of Fisheries and participated in by the States of Louisiana, Georgia, and Texas. The shrimp supports the most valuable fishery in the South. Little has been known of its abundance, life history, and habits, consequently it warrants an investi- gation of some magnitude. The importance of the shrimp fishery may be seen from the fact that in the United States in 1929, 113,263,000 pounds were caught, the value of which to the fishermen was $4,575,000. Of this, 95 percent was produced by the eight South Atlantic and Gulf States, Louisiana alone contributing 43 percent. In the South the shrimp exceeds the combined value of the two fisheries next below it. Otter trawls operated from gas-driven boats take about 90 percent of the shrimp, the remaining 10 percent being caught by seines and cast nets. 24 BULLETIN OF THE BUREAU OF FISHERIES Of the six species of shrimp caught in the South the common shrimp ( Penaeus setiferus ) is the most valuable. This species comprises about 95 percent of the catch in the South Atlantic and Gulf States, or about 90 percent of the entire amount of shrimp taken in the United States. P. setiferus spawns from March or April to August or September apparently in the outside waters. The post-larval young, measuring from 7 or 8 mm up, are first encountered in bays, bayous, and “lakes”, sometimes far inland, but not on ocean or Gulf beaches. Their habitat is warm, shallow, brackish waters with muddy bottoms. The young grow rapidly, and as they grow they seek deeper waters of greater salinity. During July, at an average size of about 90 mm, they enter the commercial catch, first on the inside, later outside. They continue in the fishery until the fol- lowing spring and summer, when they spawn and disappear at the age of 1 year. During this time they grow to a length of about 120 mm; remain at this size through the winter, resume growth in the spring; and, after a rapid differentiation of the size of the sexes, spawn at length about 130 to 170 mm for the males and 135 to 190 mm for the females. Their fate is unknown, but that they disappear from the commercial catch following spawning is undoubted. During the breeding season in the inside waters of Georgia there is a much greater proportion of females than males, while in the outside waters there is a greater propor- tion of males. In Georgia at the beginning of the breeding season there is evident a differential behavior of the males and females ; a differential growth rate between the sexes, with the females outstripping the males; and a change in the length frequency groupings from a wide to a very narrow range. The common shrimp is most abundant in the coastal waters near river mouths or delta regions. It is omnivorous in its feeding habits. At present the shrimp fishery does not appear to show serious depletion, consequently increased stringency of exist- ing laws is not imperative. In the life history of the shrimp there are two factors favoring depletion — the short life span and exposure to fishing at all times. On the other hand, there are two factors opposing depletion- — the extended breeding season and the large number of eggs produced. The most recent and at present one of the most decisive factors of the environment is the fishing of man. To date the favoring factors have maintained a vast shrimp population. With a more and more intensive fishery, the unfavorable factors must at some time become the more powerful. Because of the short life of the common shrimp, depletion, when it appears, will probably run a disastrously rapid course; consequently vigilance in safeguarding the industry is necessary. Against this time, methods of protection must be carefully planned. There are four types of regulation applicable to the shrimp fishery; namely, closed seasons, closed areas, size limits, and restriction of gear. The fishery records of the various States are entirely inadequate to permit an analysis of abundance. For the safety of the shrimp industries it is imperative that this lack of catch records should speedily be remedied. It is recommended that all States collect on a standard form uniform records of the daily catch of each fisherman; certain of the States are now working on this problem. It is only by such records that any accurate indication of threatened depletion may be detected. LIFE HISTORY OF THE COMMON SHRIMP 25 BIBLIOGRAPHY Alcock, A. 1906. The prawns of the Peneus group. Catal. Indian Mus., Decapod Crustacea, part III, Macrura, fasc. 1, 55 pp., 9 pis. Calcutta. Bate, C. Spence. 18S8. Report on the Crustacea Macrura collected by H.M.S. Challenger. Challenger Reports, vol. XXIV, xc + 942 pp., 157 pis. London. Berkeley, Alfreda A. 1929. Sex reversal in Pandalus danae. Amer. Nat., vol. 63, pp. 571-573. Berkeley, Alfreda, A. 1930. The post-embryonic development of the common pandalids of British Columbia. Contr. Canad. Biol., n.s., vol. 6, pp. 81-163, 30 figs. Toronto. Bonnot, Paul. 1932. The California shrimp industry. Fish Bull. no. 38, Sacramento, Calif., 20 pp., 11 figs. Brooks, W. K. 1883. The metamorphosis of Penaeus. Johns Hopkins Univ. Circ., vol. 2, Baltimore. Also Ann. Mag. Nat. Hist., vol. II, no. 19, p. 6. de Man, J. G. 1911. Penaeidae. Decapoda of the Siboga Expedition, pt. I, vol. 39. Leiden. Ehrenbaum, E. 1889. Garneelen Fischerei an der Nordsee. Mitt. f. Kusten-und Hochseefisclierei, Deutschen Fischerei-Vereins Bd. 3, no. 3, pp. 61-78. Berlin. Ehrenbaum, E. 1890. Zur Naturgeschichte von Crangon vulgaris. Beilagen, Mitt, der Sektion f. Kusten-und Hochseefischerei des Deutschen Fischerei-Vereins. Berlin. Fiedler, R. H. 1929. Fishery industries of the United States, 1928. Appendix IX, Report, U.S. Com. Fish., 1929, pp. 401-625. Washington. Fiedler, R. H. 1931a. Fishery industries of the United States, 1929. Appendix XIV, Report, U.S. Com. Fish., 1930, pp. 705-1068, 29 figs. Washington. Fiedler, R. H. 1931b. Fishery industries of the United States, 1930. Appendix II, Report, U.S.Com.Fish., 1931, pp. 109-552, 23 figs. Washington. Galtsoff, Paul S. 1930. Destruction of oyster bottoms in Mobile Bay by the flood of 1929. Appendix XI, Report, U.S.Com.Fish, 1929, pp. 741-758, 3 figs. Washington. Gates, Wm. H. 1910. Shrimp. Fifth Bienn. Report, Director Gulf Biol., Sta., Cameron, La., pp. 8-12. Baton Rouge, La. Gowanloch, James Nelson. 1931. The probable cause of “Iodine Shrimp.” La. Conser. Rev., vol. II, pp. 31-32. New Orleans. Guilbeau, Braxton H. 1908. Shrimp. Fourth Bienn. Report, Director Gulf Biol. Sta., Cam- eron, La., pp. 11-13. Baton Rouge, La. Having, B. 1930. Der Granat ( Crangon vulgaris Fabr.) in den hollandischen Gewassern, Cons. Perm. Internat. Explor. Mer, Journ. Counseil, vol. 5, pp. 57-87, 8 Abb. Copenhague. Hay, W. P., and C. A. Shore. 1918. The decapod crustaceans of Beaufort, N.C., and the sur- rounding region. Bull. U.S. Bur. Fish., vol. XXXV, 1915-16. Washington. Herdman, W. A. 1904. An outline of the shrimp question. Trans. Liverpool Biol. Soc., vol. 18, pp. 157-167. Liverpool. Herrington, W. C., and Dan Merriman. 1932. The advantages of the new savings trawl. Fish. Gaz., vol. 49, pp. 8-10. New York. Higgins, Elmer. 1931. Federal Bureau and State governments cooperate in shrimp investiga- tions. La. Conser. Rev., vol. I, no. 7, pp. 3-5. New Orleans. Hynes, Frank W. 1929. Shrimp fishery of southeast Alaska. Appendix I, Report, U.S. Com. Fish., 1929, pp. 1-18, 3 figs. Washington. Kishinouye, K. 1900a. Japanese species of the genus Penaeus. J. Fish. Bur., Tokyo, vol. 8, no. 1, pp. 1-29, pis. I-IX. Kishinouye, K. 1900b. The nauplius stage of Penaeus. Zool. Anz., Leipzig, Bd. 23, no. 607, pp. 73-75. Korschelt, E., and K. Heider. 1899. Textbook of embryology of invertebrates, vol. II, pp. 267- 271. Macmillan, New York. Muller, Fritz. 1863. Die Verwandlung der Garneelen. Arch. Naturgesch., Berlin, Bd. 29, pp. 8-23. Muller, Fritz. 1864. (Translation of above.) Ann. Mag. Nat. Hist., London, series 3, vol. 14, p. 104-115. Muller, Fritz. 1878. Ueber die Naupliusbrut der Garneelen. Z. wiss. Zool., Leipzig, Bd. 30, pp. 163-166. 26 BULLETIN OF THE BUREAU OF FISHERIES Myers, Hu B., and Gowanloch, J. N. 1932. Report of Bureau of Scientific Research and Statistics. La. Conser. Rev., vol. II, no. 9, pp. 7-15. New Orleans. Orr, Arthur. 1921. Description of shrimp canning methods. Mem. S-165, U.S.Bur.Fish. Washington. Orr, Arthur. 1921. Otter trawls used in the shrimp industry of the South Atlantic and Gulf States. Mem. S-166, U.S.Bur.Fish. Washington. Rathbun, Richard. 1883. Notes on the shrimp and prawn fisheries of the United States. Bull. U.S. Fish. Com., 1882, vol. II, 139-152. Washington. Rathbun, Richard. 1887. The crab, lobster, crayfish, rock lobster, shrimp, and prawn fisheries. In Fishery Industries of the United States. By G. Brown Goode and associates. Section V, vol. 2, part XXI, 1887, pp. 627-810. Washington. Schmitt, W. T. 1926. Report on Crustacea Macrura ( Peneidae , Campylonolidae, and Pandalidae ) obtained by the F.I.S. Endeavour in Australian Seas. Biological Results of Endeavour Expedition, vol. V, pt. 6. Sydney. Scofield, N. B. 1919. Shrimp fisheries of California. Calif. Fish Game, vol. 5, no. 1, pp. 1-12, 5 figs. Sacramento. Senoo, Hidemi. 1910. Kuruma-ebi no sechado [Growth rate of Penaeus]. Dobutsugaku Zasshi, pp. 91-93. Tokyo. [In Japanese.] Spaulding, M. Herrick. 1908. Preliminary report on the life history and habits of the “Lake Shrimp” Penaeus setiferus. Gulf Biol. Sta., Cameron, La., Bull. no. 11, pp. 1-24, 6 pis. Baton Rouge. Tressler, D. K. 1923. The American shrimp industry. In Marine Products of Commerce, pp. 548-556. Chemical Catalogue Co., New York. Tulian, E. A. 1920. Louisiana — greatest in production of shrimp— Penaeus setiferus. Fourth Bien. Report, La. Dept. Conser., 1918-1920, pp. 106-114. New Orleans. Tulian, E. A. 1923. The present status of the Louisiana shrimp industry. Trans. Amer. Fish. Soc., vol. 56, pp. 169-174. Hartford, Conn. Viosca, Percy. 1920. Report of the Biologist. Fourth Bien. Report, La. Dept. Conser., 1918- 1920, pp. 120-130. New Orleans. Weymouth, F. W. 1931. Shrimp investigations on the South Atlantic and Gulf coasts. La. Conser. Rev., vol. I, no. 13, pp. 11-13. New Orleans. Weymouth, F. W. 1932. Federal and State cooperative shrimp investigations in Louisiana. La. Conser. Rev., vol. II, no. 9, pp. 39-40. New Orleans. Yost, Bartley, F. 1922. Shrimp fishing industry on Pacific coast of Mexico. Mem. S-226, U.S.Bur.Fish. Washington. U.S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner THE HOMING INSTINCT AND AGE AT MATURITY OF PINK SALMON (ONCORHYNCHUS GORBUSCHA) By FREDERICK A. DAVIDSON From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVI1I Bulletin No. 15 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1934 For sale by the Superintendent of Documents, Washington, D.C. Price 5 cents I THE HOMING INSTINCT AND AGE AT MATURITY OF THE PINK SALMON ( ONCORHYNCHUS GORBUSCHA)1 By Frederick A. Davidson, Pli.D., Associate Aquatic Biologist, United States Bureau of Fisheries CONTENTS Introduction Marking pink salmon fry Interpretation of results from marking experiments Homing instinct Age of pink salmon at maturity Pink salmon marking experiment in British Columbia Summary Literature cited Page 27 28 29 29 34 36 38 39 INTRODUCTION The pink salmon, Oncorhynchus gorbuscha (Walbaum), like the other species of Pacific salmon spend part of their life cycle in the sea and part in the rivers and creeks where they spawn and die. In southeastern Alaska, the pink salmon begin their spawning migration from the sea in the latter part of June and continue until late in September. Although the salmon migrate into the streams during the early summer months spawning does not begin until the second or third week in August. In March and April of the following year, the fry which have developed in size to a little over an inch in length (see fig. 1) and weigh less than 0.008 of a pound, emerge from their nests in the gravel and migrate directly to the sea. During their sojourn in the sea the fry develop into adult salmon weighing from 3 to 8 pounds and upon reaching maturity return to the streams to spawn. The population of mature pink salmon that returns to spawn in any stream varies in size from year to year. This is due not only to the influence of changes in the natural elements of the habitats (stream and ocean) in which the population develops but likewise, and to a far greater extent, to the changes in the intensity of the commercial fishery that is imposed upon it during its migration from the sea. In view of this fact the Bureau of Fisheries is endeavoring to regulate the intensity of the commercial salmon fishery in Alaska so as to provide for an adequate run of pink salmon in every stream each season.2 In order to secure an accurate count of the number of mature pink salmon migrating each season into each of several impor- tant salmon streams in southeastern Alaska, the Bureau has constructed weirs (see fig. 7) through which the salmon are counted on their way to the spawning grounds. 1 Bulletin No. 15. Approved for publication, Feb. 21. 1934. 2 On June 6, 1924, Congress approved an act which provided that not less than 50 percent of all the adult salmon returning to any stream should be permitted to ascend the stream to spawn. 27 28 BULLETIN OF THE BUREAU OF FISHERIES The operation of these weirs each season is providing records of the size of the pink- salmon runs in each of the several streams on successive years. Hence it is possible, by means of these records, to make a study of the returns in numbers of adults resulting from the spawning of pink-salmon populations of varying size in an individ- ual stream. However, in making such a study it is essential to know the extent to which the pink-salmon fry, after reaching maturity, return to spawn in the streams from whence they came and likewise the age at which they mature and return to spawn. In life history studies of other species of Pacific salmon3 it has been shown that their homing instinct and age at maturity can be determined by marldng the fry (removing two of their fins) as they leave a particular stream and then examining the subsequent runs of mature salmon for individuals bearing the marks. In this way the extent to which the adult salmon return to spawn in their parent stream and likewise the age at which they return may be readily determined. Two exper- iments of this type have been carried on with pink-salmon fry, one at the Federal hatchery on the Duckabush River in the State of Washington and one in Snake Creek at Olive Cove, Alaska. MARKING PINK-SALMON FRY The marking experiment at the Duckabush River hatchery was carried on in the spring of 1930 and is the first known attempt to mark pink-salmon fry at the time they normally migrate from the streams. The fry in this experiment, as well as in the Snake Creek marking experiment, were marked by the removal of both their dor- sal and adipose fins. This double mark was used because mature salmon are quite frequently found with one fin missing due to natural causes.4 Furthermore it is not only necessary to remove 2 fins but likewise 2 fins that are widely separated on the body. In this way, the chances are remote of finding a mature salmon with an identical mark which is due to natural causes. The 36,000 fry that were marked in this experiment were taken from the hatchery tanks where they had previously hatched under artificial conditions. The dorsal and adipose fins were removed by clipping them close to the backs of the fry with a pair of straight-bladed finger-nail clippers. The two fry pictured in figure 1 show the appearance of normal and marked individuals. Owing to the small size of the fry it was necessary to select apparatus that would facilitate their marking and minimize their mortality during the marking operation. The apparatus shown in figure 2 was found to give satisfactory results. The shock due to the operation in removing the fins from the fry did not appre- ciably affect their mortality. In fact, the mortality of the fry during confinement in the tanks was practically the same in both the unmarked and marked stocks. Under normal conditions when there was an abundant supply of fresh water in the tanks the mortality during a 24-liour period did not exceed 2 percent of the total numbers in either of the stocks. However, if the supply of fresh water was reduced, the mor- tality would, in extreme cases, mount to 25 percent of their total number within a few hours. The marking experiment in Snake Creek at Olive Cove, Alaska, was carried on in the spring of 1931. The 50,000 fry that were marked in the experiment were s See the following references: Gilbert (1913) for coho salmon, Rich and Holmes (1928) for Chinook salmon, Snyder (1921 to 1924) for Chinook salmon, and Foerster (1929) for soekeye salmon. 1 See page 37 for a discussion of the errors resulting from the use of only single fin marks in salmon marking experiments. Bull., U.S.B.F. (No. 15.) Figure 1. — Normal and marked pink-salmon fry at time of seaward migration. Figure 2. — Apparatus used in marking pink-salmon fry. Operator stands in front of shallow square net that rests on sides of tank just at surface of water. The small amount of water in the center of the net is sufficient to maintain fry in a semiasphyxi- ated state when they are placed there preparatory to marking. The two reading lenses, 4 inches in diameter, suspended above near side of net more than double image of a fry when it is held under them while being marked. The gas lamp beside net provides a direct light for operator and incidently radiates heat upon the hand that is used to catch the fry and hold them while being marked. Tanks are supplied wdth running water from stream. In order to keep an accurate account of number of fry marked the tally counter mounted on side of tank near frame that supports the reading lenses, is tripped every time operator marks a fry. Upper compartment of each tank is used to provide temporary storage for unmarked fry. After marking , fry are released in lower compartments where they are held from 12 to 24 hours before liberation in stream. ' HOMING INSTINCT OF PINK SALMON 29 caught in the stream during their seaward migration from the spawning grounds where they had hatched under natural conditions. The fry were caught with a small screen wire trap net (see fig. 6) that was anchored in the middle of the stream where the current was the swiftest. Although thousands of fry were caught in the trap, not one was caught during the day, the fry migrating only at night. Many trips were made to the spawning grounds throughout the period the fry were migrat- ing, but invariably no pink-salmon fry could be found during the daytime, either in the stream or in the pools along the sides of the stream. With the aid of a lantern they could he found in the stream any time during the night, but at dawn they dis- appeared, and those that did not reach the bay hid under the rocks or in the gravel. The actual marking of the fry did not involve any difficulties for the method of marking had been worked out at Duckabush the year before. The only change in the routine from that used at Duckabush was that the marking was done at night so that the fry could be liberated at the time they normally migrated. INTERPRETATION OF RESULTS FROM MARKING EXPERIMENTS HOMING INSTINCT The 36,000 pink-salmon fry that were marked at the Duckabush River hatchery in the spring of 1930 came from the spawn of 1929. Working upon the supposition that the pink salmon mature at 2 years of age,5 the 1931 run of mature salmon in the Duckabush River was examined for individuals bearing the marks. In order to determine the extent to which the Duckabush River pink salmon returned to their parent stream to spawn, the 1931 runs in all the streams along the south shore of the Strait of Juan de Fuca, east of Port Angeles (see fig. 3) were likewise examined for individuals bearing the marks. The hatchery operations on the Duckabush River made it possible to examine each one of the 3,800 mature pink salmon that composed the 1931 run. This was done at the time they were removed from the river and put in a retaining pond where they were held until fully mature and ready to spawn. The runs in the other streams along the canal and the south shore of the strait were examined by frequent observations of the schools of salmon as they migrated into the shallow waters on the spawning grounds. Where possible the salmon were collected in the shallow areas by means of a large net and carefully examined for individuals bearing the marks. The hatchery attendants at the State hatchery on the Dungeness River examined the 40,000 pink salmon in that stream that were spawned artificially. The observation of the schools of salmon on the spawning grounds began in the latter part of August and continued until late in September. Out of the 3,800 pink salmon examined in the Duckabush River 5 females and 3 males were found bearing distinct adipose and dorsal scars. One female bearing both * * a dorsal and adipose scar was observed in the Hamma Hamma River in which approx- imately 1,500 pink salmon were spawning. One male bearing both a dorsal and adi- pose scar was found dead along the banks of the Dosewallips River in which approxi- mately 5,000 pink salmon were spawning.6 The dorsal scars on most of the marked 5 Gilbert (1913) from a study of the markings on the scales of pink salmon taken in various localities came to the conclu- sion that the pink salmon invariably mature at 2 years of age. Furthermore, since the runs of pink salmon in the Ducka- bush River and neighboring streams occur only on alternate years, the pink salmon in these streams must either mature at 2 or 4 years of age. A further discussion of Gilbert’s work is given in this paper under the section “Age at maturity of pink salmon.” • The Hamma Hamma, Duckabush, and Dosewallips Rivers are the only streams on Hood Canal in which pink salmon spawn in any numbers. The pink-salmon runs in these streams usually occur at the same time during the season. 30 BULLETIN OF THE BUREAU OF FISHERIES salmon were roughened by the partial regeneration of a few spines. A marked salmon of this type is shown in figure 4. Only a few of the dorsal scars were perfectly smooth, showing no signs of regeneration, as seen on the marked salmon in figure 5. The adi- Figure 3.— Hood Canal and the south shore of the Strait of Juan de Fuca, Wash. The large dots mark the location of fish hatcheries. The pink salmon spawn only in the Hamma Hamma, Duckabush, Dosewallips, and Dungeness Rivers, and Morse Creek. pose scars, with a few exceptions, were perfectly smooth and showed no signs of regeneration. A number of pink salmon with only adipose fin scars were also recovered in the streams during the survey in the fall of 1931. Thirteen were found in the Duckabush Bull., U.S.B.F. (No. 15.) Figure 4.— Normal (upper) and marked (lower) adult pink salmon. The dorsal fin on the marked individual has partially regenerated. Figure 5. — Normal (upper) and marked (lower) adult pink salmon. Neither the dorsal nor adipose fin on the marked individual show signs of regeneration. Bull., U.S.B.F. (No. 15.) Figure 6. — Screen-wire trap and net used in catching the pink-salmon fry in Snake Creek. Figure 7. — Fish trap built in weir at Snake Greek. The large dip nets held by the operators were used to lift the salmon out of the trap. The sliding gate in the weir just to the right of the trap was opened when the salmon were counted through the weir. HOMING INSTINCT OF PINK SALMON 31 River, 1 in the Dosewallips River, 3 in the Dungeness River, and 1 in Morse Creek. In a controlled experiment wherein 500 fry were kept for a number of months after being marked it was found that their dorsal fins had a greater tendency to regenerate than did their adipose fins. It may be that some of the 13 salmon with only adipose scars that were found in the Duckabush River were marked individuals whose dorsal fins had regenera ted. However, owing to the frequent occurrence of salmon whose adipose fins are missing due to natural causes,7 it is exceedingly illogical as well as hazardous to consider any of the individuals with only adipose scars as marked salmon whose dorsal fins regenerated. Especially is this true of the 3 pink salmon with only adipose scars that were found in the Dungeness River, for out of the 40,000 pink salmon that were examined in that stream when they were spawned artificially not one was found with both its dorsal and adipose fins missing. The 50,000 pink-salmon fry, marked in Snake Creek at Olive Cove, Alaska, in the spring of 1931, hatched from the eggs spawned in the stream in the fall of 1930. Since marked pink salmon returned to the Duckabush River at 2 years of age, it was be- lieved that they would likewise return to Snake Creek at 2 years of age. Hence the run of mature pink salmon in Snake Creek and the runs in the neighboring streams were examined during the summer of 1932 for individuals bearing both dorsal and adipose fin scars. The runs of pink salmon in Snake Creek are usually rather large, ranging be- tween 50,000 to 100,000, so that it was impractical to try to catch and observe each fish in the run. By making use of the counting weir that is operated in the creek it was possible to devise a method whereby a large portion of the individuals com- posing the run could be examined. An enclosure or trap was built above the weir, as shown in figure 7, into which the salmon could pass through a V-shaped opening in the weir. The opening into the enclosure was never blocked so that the salmon could pass through it at all times. The salmon that were trapped in the enclosure were removed by means of large dip nets, examined for missing fins, and then thrown into the stream above the weir so that they could continue on their way to the spawning grounds. At times during the day the salmon collected in such large numbers in the stream that it was necessary to open the gates of the weir (see fig. 7) and count them as they passed through on their way up stream. An accu- rate account was kept each day of the number of pink salmon lifted from the trap and the number counted through the weir. Although only a portion of the entire run of pink salmon in the stream was examined for marked individuals as it passed through the trap, it was possible, as will be shown later, to calculate the number of marked individuals in the entire run. The 1932 run of pink salmon in Anan Creek, a neighboring stream (see fig. 8), was examined for marked individuals in a similar manner by the operation of a trap that was built in its weir. In addition to the examination of the 1932 pink-salmon runs in Snake Creek and Anan Creek a survey, similar to that made in the streams along Hood Canal in 1931, was made of the other pink-salmon streams in the vicinity of Olive Cove. An examination of the 1932 run at Olive Cove shows that out of a total of 7,944 pink salmon lifted out of the trap, 23 (10 males and 13 females) had both dorsal and adipose fin scars, 5 had only dorsal fin scars, and 10 had only adipose fin scars. Although the marked individuals (those with dorsal and adipose fin scars) are the 7 In both marking experiments it was not uncommon to find a fry whose adipose fin had never developed. By marking the fry under the reading lenses it was not difficult to determine the presence or absence of the fins. 32 BULLETIN OE THE BUREAU OF FISHERIES only individuals that can be considered as returns, it is interesting to note that the dorsal scars on the 5 salmon having only these scars were identical to the dorsal scars on the 23 marked salmon. It is not at all unlikely that these 5 individuals resulted from the fry whose adipose fins were missed when they were marked. Regardless of this close resemblance, these 5 individuals cannot be considered as returns, for by so doing, the 10 salmon with only adipose scars would likewise have to be considered as returns. This as previously pointed out 8 would be a hazardous thing to do. Figure 8. — Snake Creek at Olive Cove, Alaska, and vicinity. All of the streams shown on the map support a large population of pink salmon each season. The method used in calculating the number of marked salmon in the entire run of pink salmon in Snake Creek in the summer of 1932 is given in the following equation: 9 23 + (l0,640X7-|U)_54 where 23 equals the number of marked salmon found in the trap, 10,640 equals the total number of salmon counted through the gates in the weir, 7,944 equals the 8 See discussion on page 29. > This equation is based upon the assumption that there was the same proportion of marked salmon in the numbers counted through the weir as in the numbers lifted out of the trap. There is no reason to believe that the trap was selective in regard to the marked salmon; i. e., that more of the marked salmon in the run passed through the trap than through the gates in the weir. On days when the water in the creek was low and the visibility good, marked salmon were observed passing through the gates in the weir. During these times a total of 9 marked salmon were seen passing through the gates. HOMING INSTINCT OF PINK SALMON 33 total number of salmon lifted out of the trap, and 54 equals the number of marked salmon in the entire run of pink salmon. Out of the 132,351 pink salmon that composed the 1932 run at Anan Creek, 13,965 were examined as they were lifted out of the trap. Two individuals hearing fin scars were found among the salmon caught in the trap, one had only a dorsal fin scar and the other had only an adipose fin scar. No marked salmon were either caught in the trap or observed in the stream. Although the salmon with only the dorsal fin scar might have originated from a fry whose adipose fin was missed when marked, the absence of any marked individuals (those with both dorsal and adipose fin scars) in the salmon caught in the trap makes it rather improbable that this individual was a Snake Creek salmon. The stream located nearest to Snake Creek in which pink salmon spawn in large numbers is Thoms Place Creek, approximately 10 miles south of Olive Cove and on the opposite side of Zimovia Strait. This stream was visited a number of times dur- ing the summer when the pink salmon were migrating into Snake Creek. During these visits only a few pink salmon were seen in the stream and none of these was marked. Toward the last of August when the run in Snake Creek was practically completed, a large run of pink salmon migrated into Thoms Place Creek. A survey of this stream was again made at this time, but no marked pink salmon were observed in the run. A survey was made of the streams along Eastern Passage, the streams in Brad- field Canal other than Anan Creek, and the streams along the shores of Stikine Strait. Owing to the unusual lateness of the pink-salmon runs throughout the whole district it was not until late in August that the pink salmon collected in the shallower regions of the streams where they could be observed. About 200,000 pink salmon in all were observed as they swam about in the shallow waters of the streams. No individuals with both their dorsal and adipose fins missing were found in any of the streams that were visited. In view of the returns from these marking experiments, it is conceivable that the extent to which the pink salmon return to their parent streams to spawn may be dependent upon the proximity of other pink-salmon streams in the vicinity. That is to say, the pink salmon composing the runs in streams that are more or less isolated from other pink-salmon streams may show very little or no tendency towards straying whereas the pink salmon in streams flowing into bays and in close proximity to other pink-salmon streams may stray more or less into the neighboring streams. The stray- ing of the marked Duckabush River pink salmon into the Hamma Hamma and Dose- wallips Rivers may be due to the close proximity of these streams to the Duckabush River. The Dosewallips River is located 4 miles north of the Duckabush River and the Hamma Hamma River is located 9 miles south of the Duckabush River. The apparent lack of straying of the Snake Creek marked pink salmon might well be due to its isolation from the other large pink salmon streams in the district. Although Thoms Place Creek is only 10 miles south of Olive Cove, the main run of pink salmon in this stream is usually much later than the run in Snake Creek. No marked individuals were found either among the pink salmon in this stream during the early part of the summer or in the large run that occurred late in August. The other pink- salmon streams in the district are all located more than 20 miles distant from Olive Cove. Anan Creek, in which a weir trap was operated in a manner similar to the operation of the trap in Snake Creek, is 25 miles distant from Olive Cove. The 54190—34 2 34 BULLETIN OF THE BUREAU OF FISHERIES run of pink salmon in Anan Creek also appears at practically the same time during the season as the run in Snake Creek. If the Snake Creek marked salmon strayed any distance from their parent stream in perceptible numbers, it is highly probable that they would have been picked up by the trap in the Anan Creek weir. However no marked pink salmon were either found in the trap or observed in the stream. This, together with lack of recovery of marked pink salmon in any of the other more distant streams in the vicinity, makes it rather improbable that the Snake Creek pink salmon strayed in perceptible numbers from their parent stream. AGE OF PINK SALMON AT MATURITY The scales of the pink salmon like the scales of the other species of Pacific salmon, tend to grow at a rate proportional to the growth of the fish. This is accomplished by the deposition of new material around the border of the scale. Delicate ridges appear on the surface of the scale at intervals during its growth which form concentric rings separated by spaces of varying width. (See fig. 9.) The variation in the width of the spaces between the rings may be attributed to a corresponding variation in the rate of growth of the fish. That is during the spring and summer when the fish grows rapidly the rings on the scale are widely separated whereas during the fall and winter when growth is greatly retarded the rings are crowded together. In this way the surface of the scale is marked by bands of widely spaced rings followed by bands of closely spaced rings. Since the wide band of rings is formed in the spring and summer and the nar- row band in the fall and winter, the two together represent a year’s growth. C. H. Gilbert (1913) in his paper “Age at maturity of the Pacific coast salmon” pointed out that the pink salmon all mature at 2 years of age as judged by the age reading of the growth rings on their scales. He described the pink-salmon scale as having three definite bands of growth rings. A central band of widely spaced rings of the sea type 10 followed by a band of closely set rings, which is in turn followed by another band of widely spaced rings. The first two bands of rings represent the growth of the fish during the first spring and summer in the sea and the first and only winter in the sea. The second band of widely spaced rings represents the growth of the fish during the second spring and summer in the sea. Since the second summer’s growth band is never followed by a winter band of rings, Gilbert came to the conclusion that the pink salmon always mature at the close of the second year of their life. Owing to the rapid growth the pink salmon must make, during their 16 or 17 months’ sojourn in the sea, in order to mature at 2 years of age, some doubt has been expressed in regard to their age at maturity as determined from the bands of growth rings on their scales. In view of this fact a study was made of the scales on the marked pink salmon that were known to return to the Duckabush River and to Snake Creek at 2 years of age. The scale pictured in figure 9 is from a marked salmon that re- turned to Snake Creek and the scale pictured in figure 10 is from a marked salmon that returned to the Duckabush River. These scales were chosen because they are representative of the scales of the other marked pink salmon found in these streams. The growth rings on the scale in figure 9 may be divided into the three character- istic bands described by Gilbert (1913). The first band of widely spaced rings, i. e., those laid down in the life of the salmon during its first spring and summer in the sea, i° Since the pink salmon fry in most cases leave the streams before their scales have appeared, the only rings that are formed in their scales are those that form during their life in the sea. Bull., U.S.B.F. (No. 15.) Figure 9. — The scale of a marked pink salmon recovered in the 1932 ran of pink salmon in Snake Creek at Olive Cove, Alaska. The letters “W.B.” indicate the winter band of growth rings. Figure 10.— The scale of a marked pink salmon recovered in the 1931 run of salmon in the Duckabush River. Wash. The letters “W.B.’ ’indicate the winter band of growth rings and the letters “I.B.” indicate the incidental band of growth rings. ■ HOMING INSTINCT OF PINK SALMON 35 occupies the central area of the scale. This band is terminated at its outer border by a narrow band of closely set growth rings or the winter band (W.B.). The second band of widely spaced growth rings, i. e., those laid down in the life of the salmon dur- ing its second spring and summer in the sea, immediately follows the winter band (IV. B.) and terminates at the margin of the scale. The growth rings on the Duckabush River pink-salmon scale shown in figure 10 likewise may be divided into the three characteristic bands. The first band of widely spaced growth rings is, however, not very distinct since the rings near its center are in such close proximity that they form a rather definite incidental band of closely spaced rings ( I.B. ). The winter band of closely spaced growth rings (W.B.) and the succeeding second band of widely spaced growth rings are on the other hand very distinct and well marked around the circumference of the scale. This incidental band of closely spaced rings (I.B.) that occurred in the life of the salmon during its first spring and summer in the sea is no doubt due to a temporary retardation in its rate of growth. The occurrence of such aberrations on the scales of the Pacific salmon was recognized by Gilbert (1913, p. 5), who says: Thus it comes that the surface of the scale is mapped out in a definite succession of areas, a band of widely spaced rings always followed by a band of closely crowded rings, the two together constituting a single year’s growth. That irregularities occur will not be denied, and this is natural, inasmuch as growth may be checked by causes other than the purely seasonal one. Had not the scale in figure 10 been known to have come from a pink salmon that matured at 2 years of age, it would still be illogical to assume the incidental band of closely spaced rings (I.B.) to be a true winter band. Such an assumption would mean that the salmon bearing this scale spent 3 summers and 2 winters in the sea and ma- tured at 3 years of age. This, however, would be impossible since the pink salmon runs in the streams on Hood Canal and along the Strait of Juan de Fuca occur only on alternate years. Hence the salmon composing these runs must mature at 2 or some multiple of 2 years of age. Since the pink-salmon runs in the Duckabush River occur only on the odd years, it is impossible for any of the pink salmon in this stream to mature at 3 years of age. In order to check the possibility of any of the pink salmon in this stream maturing at 4 years of age, the run in this stream was examined for individuals bearing the marks in the summer of 1933. No marked salmon were found in the 1933 run and every fish was examined carefully at the time it was spawned artificially by the hatcherymen. The runs of pink salmon in Snake Creek occur every year, hence the 1933 run in Snake Creek was examined in order to check the possibility of the pink salmon in this stream maturing at 3 years of age. The salmon run in this stream was examined in the same way that it was examined during the summer of 1932. No marked pink salmon was found in the 1933 run. There was, however, a number of pink salmon with deformed and missing adipose fins found in this run, which is further evidence that pink salmon with missing adipose fins due to natural causes are not infrequently found in the runs. There is still other evidence which indicates that the pink salmon mature at 2 years of age. In 1913 and 1915 large shipments of eyed pink-salmon eggs were sent from the Pacific coast to the fish hatcheries in the New England States. These eggs were hatched artificially and in the spring of 1914 and the spring of 1916 the fry were liberated in a number of streams along the coast. In the summer and fall 36 BULLETIN OF THE BUREAU OF FISHERIES of 1915 a large number of mature pink salmon returned to the streams in which the fry were liberated. A number of ripe eggs were removed from some of the females and after fertilizing them they were taken to the hatchery where they later developed into normal fry. In the summer and fall of 1917 another large run of pink salmon appeared in the streams in which the fry were liberated. Some of these salmon were sent to Dr. C. H. Gilbert who, after examining their scales, claimed that they had retained their original habit of migrating directly to the sea upon leaving their nests in the gravel and returning to the rivers to spawn and die at 2 years of age.11 The scales of the pink salmon that have appeared in Snake Creek and in other streams in southeastern Alaska on different years have been examined and all show bands of growth rings similar to those on the scale in figure 9. The scales of the unmarked as well as those of the marked salmon that appeared in the 1931 run in the Duckabush River were also similar to the scale shown in figure 10. In fact all the evidence thus far collected indicates that the pink salmon mature at 2 years of age and until contradictory evidence is found it may be assumed with relative cer- tainty that they mature consistently at the close of their second year of life. PINK SALMON MARKING EXPERIMENT IN BRITISH! COLUMBIA • In the spring of 1931 Dr. A. L. Pritchard, of the Canadian Pacific Biological Station, working at McClinton Creek in Massett Inlet, British Columbia, marked 185,000 pink-salmon fry by the removal of only their adipose fins. The returns from this marking experiment are reported in table l.12 In discussing the significance of these returns Pritchard (1932, p. 10) makes the following statements: The return of 95 marked fish (52 percent of the recoveries) to McClinton Creek definitely establishes that there is a tendency on the part of the pink salmon of this locality to return to its native spawning area and that maturity is reached at the end of 2 years. It is not unlikely that the 22 fish taken in Massett Inlet and 16 in Naden Harbour and Otard Bay would have ultimately appeared in McClinton Creek had they not been caught by the com- mercial nets, in which case the return to McClinton Creek would have been 73 percent of the recoveries. The capture in other localities of 50 fish lacking the adipose fin (27 percent of the recoveries) is evidently indicative of a certain degree of wandering. Although not a single fin abnormality was discovered among 310,000 pink-salmon fry handled at McClinton Creek and the Tlell River, reports from other areas indicate the possibility that such abnormalities may exist to a very small extent. It is felt, however, that the indications shown by the large returns from some of the out- lying districts should not be considered insignificant. According to these statements Pritchard is apparently of the opinion that little hazard was involved by the use of only the adipose-fin mark for the future indenti- fication of the salmon, and that all of the salmon reported in table 1 were originally McClinton Creek fry. 11 For a detailed discussion of the results from this transplantation of the pink salmon on the Atlantic coast see Reports of the Commissioner of Fisheries, 1916 and 1917, pp. 30 and 75, respectively. u See A. L. Pritchard (1932) for data on recoveries of the adipose-marked salmon in the various localities as given in this table. HOMING INSTINCT OF PINK SALMON 37 Table 1. — Returns from the pink-salmon marking experiment at McClinton Creek, British Columbia [As reported by A. L. Pritchard (1932)] Locality of recovery Num- ber of recov- eries 1 Approxi- mate distanco from McClin- ton Creek Locality of recovery Num- ber of recov- eries 1 Approx- mate distance from McClin- ton Creek English English Alaska: miles Southern Birtish Columbia, Johnstone miles 8 1,000 Strait 24 400 Snake Creek, Olive Cove 3 10 180 Queen Charlotte Islands area: Anan Creek 3 1 180 Naden Harbour 14 40 Northern British Columbia (exclusive of ( Hard Bav-- 2 no Queen Charlotte Islands): Massett Inlet. 22 10 1 120 McClinton Creek - 95 0 Naas River area.. . . . ... . 1 140 Skeena River. . 4 130 Chatham Sound 1 110 1 Pritchard assumed that pink salmon with only adipose fin scars recovered in the summer and fall of 1932 originated from the McClinton Creek marking experiment. 2 The data from Karluk Beach were reported by J. T. Barnaby of the U.S. Bureau of Fisheries. 3 The data from Snake Creek at Olive Cove and Anan Creek were reported by F. A. Davidson of the U.S. Bureau of Fisheries. One of the causes of error in the earliest marking experiments was the use of single fin marks for the future indentification of the salmon. Rich and Holmes (1928, p. 217) in reviewing these experiments make the following statements: The greatest cause of error in the earlier experiments was the failure of the investigators to realize that salmon occasionally lose one or more of their fins in other ways, and that as a result, if only one fin is removed experimentally, the mark may be duplicated accidently. For example, Hubbard removed the adipose fin from Chinook fingerlings at the Clackamas hatchery in Oregon in 1895. The reported returns from this marking are so greatly opposed to the known facts of the life history and growth of Chinook salmon that they are obviously in error, and there can be no question that they included fish not marked by Hubbard. That the adipose fins of pink salmon are likewise missing due to natural causes was pointed out by the author in discussing the returns from the Duckabush River marking experiment. (See pp. 29-31 of text.) In reporting the 10 pink salmon with only adipose fin scars from Snake Creek in Olive Cove, Pritchard failed to mention the fact that 50,000 pink-salmon fry were marked in this stream at the same time the fry were marked in McClinton Creek. Since the fry in Snake Creek were marked by the removal of both their dorsal and adipose fins, these 10 salmon were probably native to Snake Creek and consisted of salmon whose adipose fins were naturally missing or marked salmon whose dorsal fins had regenerated. (See discussion on p. 31 of the text.) The 8 pink salmon with only adipose fin scars reported by Barnaby from Karluk Beach were found during a 2-day examination of approximately 45,000 pink salmon composing part of the run into the Karluk River. Had Barnaby examined the entire run of 3,500,000 pink salmon in the Karluk River it is conceivably possible that, if the above proportion prevailed, he would have found in the neighborhood of 600 pink salmon with deformed or missing adipose fins. This together with the great distance that separates these 2 streams makes it highly improbable that these 8 pink salmon were of McClinton Creek origin. Furthermore if all the pink salmon with only adipose fin scars recovered in the Alaskan streams were of McClinton Creek origin, then it is conceivably possible that pink salmon with both dorsal and adipose fin scars, those of Snake Creek origin, should likewise have been found in the British Columbia streams. However no pink salmon bearing both dorsal and adipose fin scars were reported from Canadian waters. 38 BULLETIN OF THE BUREAU OF FISHERIES In view of the returns from the Duckabush River and Snake Creek marking experiments, it is doubtful if the adult pink salmon often stray into streams other than those in the close proximity of their parent stream. The Dungeness River is approximately 60 miles distant from the Duckabush River, and in the examination of the 40,000 pink salmon at the Dungeness hatchery no salmon were found that could be considered as originating in the Duckabush River. Anan Creek is approxi- mately 25 miles from Snake Creek in Olive Cove and during the examination of the 13,965 trap-caught salmon at Anan no pink salmon were found that could be said to have originated in Snake Creek. In fact no pink salmon were found in any of the streams in the vicinity of Olive Cove that could be said to have come from the Snake Creek marked fry. It is not improbable that the 22 adipose marked pink salmon picked up by the commercial fishermen in Massett Inlet were of McClinton Creek origin. However for Pritchard to assume that all the salmon with adipose fin scars recovered in the other localities were likewise of McClinton Creek origin is to make an assumption too broad to remain within the realm of probability. SUMMARY In the spring of 1930, 36,000 pink-salmon fry were marked at the Duckabush River hatchery, Washington, by the removal of both their dorsal and adipose fins. These fry were hatched artificially from the spawn of 1929. In the summer and fall of 1931, 3,800 pink salmon returned to the Duckabush River, 8 of which were found to have both their dorsal and adipose fins missing. The Dosewallips River is located 4 miles north of the Duckabush River, and the Hamma Hamma River is located 9 miles south of the Duckabush River. One pink salmon with both its dorsal and adi- pose fins missing was found in each of these streams. Since the pink-salmon run in these streams occurs only on alternate years no run appeared in the summer and fall of 1932. In the spring of 1931, 50,000 pink-salmon fry were marked in Snake Creek at Olive Cove, Alaska, by the removal of their dorsal and adipose fins. These fry hatched in the stream under natural conditions and came from the spawn of 1930. In the summer and fall of 1932, 18,584 pink salmon returned to Snake Creek, 54 of which had both their dorsal and adipose fins missing. The nearest pink-salmon stream in the vicinity of Snake Creek is Thoms Place, a small stream about 10 miles south of Snake Creek. The pink-salmon run in this stream, however, does not occur until the run in Snake Creek is practically completed. The other pink-salmon streams in the vicinity are all more than 20 miles distant from Snake Creek. No pink salmon with both their dorsal and adipose fins missing were found in any of the streams in the vicinity. In view of the returns from these marking experiments is conceivable that the extent to which the pink salmon return to their parent streams to spawn may be dependent upon the proximity of other pink-salmon streams in the vicinity. That is to say, that the pink salmon composing the runs in streams that are more or less isolated from other pink-salmon streams may show very little or no tendency toward straying, whereas the pink salmon in streams flowing into bays and in close proximity to other pink-salmon streams may stray more or less into the neighboring streams. The marked pink salmon that have thus far returned from these marking experi- ments have all returned at 2 years of age. Their age at maturity, as determined by HOMING INSTINCT OF PINK SALMON 39 the number of bands of growth rings on their scales, was in every case consistent with the age at which they actually matured and returned to spawn. The scales of the pink salmon that have appeared in Snake Creek and in other streams in southeastern Alaska on different years have been examined and all show bands of growth rings on their scales similar to those on the scales of the marked Snake Creek salmon. The scales of the unmarked pink salmon that appeared in the 1931 run in the Duckabush River were likewise similar to the scales of the marked salmon that were recovered in the run. Additional evidence in regard to the age at which the pink salmon mature is found in the original establishment of pink-salmon runs in a number of streams in the New England States. Large shipments of eyed pink-salmon eggs were sent to the hatcheries in the New England States in 1913 and 1915. The fry that developed from these eggs in the spring of 1914 and the spring of 1916 were liberated in a number of streams along the coast. In the summer and fall of 1915 a substantial run of adult pink salmon returned to the streams in which the fry were liberated. A similar run of pink salmon occurred in the summer and fall of 1917. Dr. C. H. Gilbert, an author- ity on the Pacific salmon, examined some of the scales taken from the adult pink salmon in these streams in 1917 and claimed that, in spite of their new environment, the pink salmon retained their original habit of returning to the rivers to spawn and die at 2 years of age. All of the evidence thus far collected indicates that the pink salmon mature at 2 years of age and until contradictory evidence is found, it may be assumed with relative certainty that they consistently mature at the close of their second year of life. LITERATURE CITED Foerster, R. E. 1929. An investigation of the life history and propagation of the sockeye salmon ( Oncorhynchus nerka ) at Cultus Lake, British Columbia. No. 3. The down-stream migration of the young in 1926 and 1927. Contr. Canad. Biol, and Fish., N.S., vol. 5, no. 3, Oct. 12, 1929. Toronto. Gilbert, Charles H. 1913. Age at maturity of Pacific coast salmon of the genus Oncorhynchus. Bull., U.S. Bur. Fish., vol. XXXII, 1912 (1913), pp. 1-22, 29 figs. Washington. Pritchard, A. L. 1932. Return of marked pink salmon in 1932. Prog. Reports, Pacific Biol. Sta. Nanaimo, B.C., and Fish. Exper. Sta., Prince Rupert B.C., no. 15, 1932, pp. 10-11. Prince, Rupert, B.C. Rich, Willis H., and Harlan B. Holmes. 1928. Experiments in marking young chinook salmon on the Columbia River, 1916 to 1927. Bull., U.S. Bur. Fish., vol. XLIV, 1928 (1929), pp. 215- 264, 85 figs. Washington. Snyder, J. O. 1921. Three California marked salmon recovered. Calif. Fish and Game, vol. 7, no. 1, Jan. 1921, pp. 1-6, figs. 1-4. Sacramento. Snyder, J. O. 1922. The return of marked king salmon grilse. Calif. Fish and Game, vol. 8, no. 2, Apr. 1922, pp. 102-107, figs. 40-50. Sacramento. Snyder, J. O. 1923. A second report on the return of king salmon marked in 1919, in Klamath River. Calif. Fish and Game, vol. 9, no. 1, Jan. 1923, pp. 1-9, figs. 1-5. Sacramento. Snyder, J. O. 1924. A third report on the return of king salmon marked in 1919 in Klamath River. Calif. Fish and Game, vol. 10, no. 3, July 1924, pp. 110-114, pis. 1-2. Sacramento. ■ I ■ '.i ■' : y : ' ■■ ■ ' , r I V: r ’ ■ U.S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner Reproduction and Development of Whitings or Kmgfishes Drums, Spot, Croaker, and Weakfishes or Sea Trouts Family Sciaenidae, of the Atlantic Coast of the United States By SAMUEL F. HILDEBRAND and LOUELLA E. CABLE From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 16 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1934 For sale by the Superintendent of Documents, Washington, D.C. Price 20 cents REPRODUCTION AND DEVELOPMENT OF WHITINGS OR KINGFISHES, DRUMS, SPOT, CROAKER, AND WEAKFISHES OR SEA TROUTS, FAMILY SCIAENIDAE, OF THE ATLANTIC COAST OF THE UNITED STATES1 By Samuel F. Hildebrand, Senior Ichthyologist, United States Bureau of Fisheries, and Louella E. Cable , formerly Junior Aquatic Biologist, United States Bureau of Fisheries Introduction Artificial keys to the eggs and young as far as known The whitings or kingfishes ( Menticirrhus americanus, M. saxatilis and M. littoralis) Characters of the adults Key to the species Menticirrhus americanus (Linnaeus) . Spawning Descriptions of the young Distribution of the young Growth Menticirrhus saxatilis (Block and Schneider) 64 Spawning 65 Description of the eggs and young 65 Distribution of the young 70 Growth 70 Menticirrhus littoralis (Holbrook) 70 Spawning 71 Descriptions of the young 71 Distribution of the young 75 Growth 75 Star drum ( Stellifer lanceolatus (Hol- brook)) 75 Spawning 76 Descriptions of the young 76 Distribution of the young 83 Growth 83 Page Banded drum ( Larimus fasciatus Hol- brook) 84 Spawning 84 Descriptions of the young 85 Distribution of the young 91 Growth 91 The weakfishes or sea trouts ( Cynoscion nebulosus, C. regalis, and C.nothus)-- 91 Key to the species 91 Cynoscion nebulosus (Cuvier and Valenciennes). Spotted trout; speckled trout; spotted weakfish-.. 92 Spawning 93 Descriptions of the young 94 Distribution of the young 101 Growth 102 Cynoscion regalis (Bloch and Schnei- der). Gray trout; gray weakfish _ 102 Spawning 103 Descriptions of the eggs and young 104 Distribution of the young 107 Growth 108 Cynoscion nothus (Holbrook). Bas- tard trout 110 Spawning 111 Descriptions of the young 112 Distribution of the young 115 Growth 116 Bibliography 116 CONTENTS Page 41 42 51 52 52 53 53 54 62 62 INTRODUCTION The account of the Sciaenidae that follows is based mostly on specimens collected on the coast of North Carolina. However, some specimens from the coast of New Jersey, from Chesapeake Bay, and from the coast of Texas have also been studied. In the preparation of the keys to the eggs and young, specimens were used as far as 1 Bulletin no. 16. Approved for publication, Apr. 3, 1934. 41 42 BULLETIN OF THE BUREAU OF FISHERIES available, and the last six papers listed in the bibliography were drawn upon freely. The keys form a sort of summary of tlie present state of knowledge of the characters of the eggs and young of the Sciaenidae of the Atlantic coast of North America. It is hoped that they will prove useful in the further study of the eggs and the young of this family of large fishes, most of the members of which are of much economic importance. It is evident from the present treatise that considerable gaps remain to be filled to make known fully the embryology and larval development of the Sciaenidae, though quite a number of stages in the development of several species, heretofore unknown, are described and figured. It has been particularly difficult to identify the eggs; and female fish, ripe enough for stripping, have been taken seldom. In a few instances the eggs were secured by confining nearly ripe females in tanks. How- ever, this procedure was successful only if the females were ripe enough to spawn the first or second night after capture. If not, the fish retained the eggs. It was shown by Hildebrand and Cable (1930, p. 418) that spots ( Leiostomus xanthurus), though containing advanced roe, retained it for several months after capture, that is, long after the spawning season had ended. The principal hope of securing the eggs for certain identification, therefore, lies in the capture of fully, or almost fully, ripe fish. The illustrations accompanying this report, unless otherwise stated, were prepared by the junior author, who also made many of the preliminary identifications and assisted in various ways in the study of the specimens. The senior author is respon- sible for the final identifications and the composition of the report. ARTIFICIAL KEYS TO THE EGGS AND YOUNG SO FAR AS KNOWN The following series of keys are entirely artificial, that is, they are designed to identify the fish rather than to show the natural relationship. The eggs of only three species of Sciaenidae from the Atlantic coast of the United States are known to date. A key by which the eggs of these three species may be identified is offered. The other keys are intended to identify the young of various sizes. The “Key to the newly hatched }roung” is based on live specimens hatched in the laboratory. It is limited, therefore, to the three species for which the eggs are known. All the other keys are based on preserved specimens. The length of the newly hatched young given in some instances exceeds that of the specimens treated in the next key, not- withstanding that the “smaller” specimens (1.7 to 1.8 millimeters long) were older fish, as shown by their development. The apparent decrease in length no doubt is caused by shrinkage in the preservatives. Some of the species included in the keys for the larger size groups are missing in those for the smaller fish, because the smaller stages are not yet known. The size groups chosen for the keys are thought to be close enough together to make it possible to identify specimens of intermediate lengths. Specimens of lengths intermediate of those on which the keys are based should be identified by the key covering a range nearest their size. For example, specimens 11 millimeters long could be identified most readily by using the key for specimens 8 to 10 millimeters long, whereas specimens 14 millimeters long should be identified by using the key prepared for specimens 15 to 20 millimeters in length. Specimens exactly half way between two groups, in general, should be identifiable as readily by the key for the smaller specimens as by the one for the larger ones. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 43 No key is offered for specimens more than 35 millimeters long because it is believed that such specimens have sufficient adult characters developed to be identi- fied by keys prepared from adult fish, such as appear in various taxonomic treatises. More complete descriptions of the eggs and young than are contained in the keys presented occur in the following works. For descriptions of the eggs and newly hatched young of Menticirrhus saxatilis see Welsh and Breder (1924). For accounts of the egg and larval development of Bairdiella chrysura consult Kuntz (1914) and Hildebrand and Cable (1930). For descriptions of young Sciaenops ocellatus and Pogonias crornis refer to Pearson (1929). For accounts of the development of the eggs and the young of Cy noscion regalis see Welsh and Breder (1923) and Pearson.2 And for descriptions of young Leiostomus xanthurus and Micropogon undulatus refer to Hildebrand and Cable (1930). Accounts of the other species included in the keys are contained in the present paper. KEY TO THE EGGS 3 [Sciaenid eggs so far as known are pelagic and spherical in form] a. Eggs generally with only 1 oil globule, rarely with more than 2. b. Range in diameter 0.66 to 0.77 millimeter, average 0.69; yolk with only a few pigment spots (dark green in color) in advanced state of development, none in younger stages; embryo and oil globule in advanced stages with somewhat scattered dark dots. Bairdiella chrysura bb. Range in diameter probably about 0.8 to 0.92 millimeter, average 0.84; yolk profusely dotted (with dark green granules) in advanced state of development, some dots appearing during the cleavage stages; embryo and oil globule profusely spotted -Cynoscion regalis aa. Eggs generally with more than 1 oil globule, often with 3 or 4, sometimes with as many as 13 to 18, reduced to 1 in advanced embryonic stages. Range in diameter 0.76 to 0.92 millimeter, average about 0.83. Yolk without pigment spots, the embryo with rather prominent scattered grayish dots Menticirrhus saxatilis KEY TO NEWLY HATCHED YOUNG 1 a. Length at hatching 1.5 to 1.8 millimeters, average about 1.6; vent nearer end of snout than tip of tail (notochord) by a distance equal to almost twice the diameter of eye. Caudal portion of body with a congregation of dark green chromatophores forming a suggestion of a cross bar at midcaudal length Bairdiella chrysura aa. Length at hatching 1.75 to 2.0 millimeters, average about 1.9; vent less than an eye’s diameter nearer end of snout than tip of tail (notochord). Caudal portion of body with two cross bars composed of dark green chromatophores, one at about midcaudal length and the other one (less distinct) about midway between the vent and the first mentioned bar. Cynoscion regalis aaa. Length at hatching 2.0 to 2.25 millimeters; vent about an eye’s diameter nearer end of snout than tip of tail (notochord) ; caudal portion of body with 3 equally spaced black and golden bars, the first one being situated immediately behind the vent Menticirrhus saxatilis KEY TO SPECIMENS 1.7 TO 1.8 MILLIMETERS LONG a. Caudal portion of body long and slender, much longer than rest of body; distance from end of snout to vent about 2.75 times in length to tip of tail (without finfold); depth immediately behind vent notably less than diameter of eye. 3 Unpublished manuscript in the files of the Bureau of Fisheries entitled “Seasonal Distribution and Abundance of Pelagic Marine Fish Eggs and Young Fishes at the entrance to Chesapeake Bay, Va.’’ 3 The key is based on live eggs. • The key is based on live larvae. Keys to all other young are founded on preserved specimens. 44 BULLETIN OF THE BUREAU OF FISHERIES b. Head and body rather deep, greatest depth about 4.0 in length to tip of tail (notochord); mouth well developed, very strongly oblique, the gape anteriorly well above middle of eye. Color plain or with a row of minute dark chromatophores along ventral outline of the tail, none in advance of vent, sometimes with one dark chromatophore on middle of side above vent, occasionally a few on the head Leiostomus xanthurus bb; Head and body more slender, greatest depth about 4.5 in length to tip of tail (notochord) ; mouth less perfectly developed, the gape anteriorly at about middle of eye; the imper- fectly developed maxillary apparently reaching only slightly past anterior margin of eye. Ventral outline of the trunk as well as the tail with distinct, well separated black chromatophores, also a few definite ones on the head, none on side of the trunk and tail. Cynoscion regalis aa. Caudal portion of body proportionately shorter and deeper; distance from snout to vent about 2.2 to 2.5 times in length to tip of tail (without finfold) ; depth immediately behind vent equal to or greater than diameter of eye. c. Head and trunk rather short and deep, the greatest depth about 3.1 times in length without caudal finfold; preanal distance 2.1 to 2.4 times, and postanal distance without finfold 1.75 to 1.9 times; depth immediately behind vent about equal to diameter of eye. Ventral outline of chest and abdomen with dark markings, with a large prominent spot just in advance of vent; anterior half to two-thirds of ventral outline of tail with a close-set series of dark dots; base of primitive pectoral pale Cynoscion nebulosus cc. Head and trunk very short and deep, the greatest depth about 2.6 times in length without caudal finfold; preanal distance 2.25 times; and postanal distance without finfold 1.88 times in length; depth immediately behind vent notably greater than diameter of eye. Caudal portion of body with an almost continuous dark longitudinal line along middle of side, and another one on the ventral outline; no prominent dark spot in advance of vent; base of primitive pectoral pale Menticirrhus americanus ccc. Head and trunk rather short and deep, the greatest depth about 3.1 times in length without caudal finfold; preanal distance 2.4 times and postanal distance without finfold 1.6 times; depth immediately behind vent about equal to diameter of eye. Median line of abdomen with a small black spot just before vent, generally with an elongated black spot on ventral outline at about midcaudal length, sometimes with several small dark spots; inner surface of base of primitive pectoral black Stellifer lanceolatus cccc. Head and trunk moderately long and deep, greatest depth 3.1 to 3.4 in length without caudal finfold; preanal distance 1.9 to 2.0 times and postanal distance 1.9 to 2.2 times; depth immediately behind vent somewhat greater than diameter of eye. Ventral out- line of tail general^ with a row of black spots; a black spot at angle of opercle; fresh specimens with a black vertical band at shoulder, which generally fades in preserv- ative; inside of base of primitive pectoral pale Bairdiella chrysura KEY TO SPECIMENS 2.5 TO 3.0 MILLIMETERS LONG a. Caudal portion of body long, slender, generally longer (much longer in Cynscion regalis and Leiostomus xanthurus) than rest of body; distance from snout to vent about 2.0 to 2.5 times in total length (without caudal finfold). b. An abrupt and pronounced break present in contour at vent; caudal portion of body very slender; depth immediately behind vent less than 1.5 times diameter of eye; distance from tip of snout to vent about 2.2 to 2.6 in total length without caudal finfold. c. Caudal portion of body moderately slender, depth immediately behind vent about 1.1 times the diameter of eye; mouth almost vertical; the gape anteriorly near level of upper margin of eye; a few dark chromatophores on chest or abdomen; an elongate one posterior to hind-gut at its juncture with the tail; a series of widely spaced chromatophores along the ventral outline of the tail; no black lateral stripe or spots present Leiostomus xanthurus cc. Caudal portion of body very slender, depth immediately behind vent about equal to diameter of eye; mouth moderately oblique; the gape anteriorly at or below middle of eye; ventral outline of chest and abdomen generally with a few chromatophores, a prominent one just in advance of vent; ventral outline of tail with well separated dark chromatophores, one at about midcaudal length enlarged and usually with long radiating lines; no black lateral stripe or spots present ..Cynoscion regalis REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 45 bb. The break in contour at vent less pronounced; caudal portion of body proportionately deeper; depth immediately behind vent about 1.7 to 1.85 times diameter of eye; distance from tip of snout to vent about 2.0 in total length without caudal finfold; mouth strongly oblique, the gape anteriorly generally above middle of eye; ventral outline of trunk and tail with very small dark chromatophores, none on side or back Micropogon undulalus aa. Caudal portion of body shorter, about equal to or shorter (without finfold) than rest of body ; distance from snout to vent about 1.75 to 1.9 in total length (without caudal finfold). d. Caudal portion of body anteriorly with a black lateral stripe, continuous or composed of more or less separated elongated spots; ventral outline behind vent with close-set black chromatophores, often forming a nearly continuous line. e. Body moderately deep, the greatest depth about 2.6 in total length without caudal fin- fold; vent situated almost exactly at midbody length; depth immediately behind vent about 1.25 times diameter of eye; mouth moderately oblique, the gape anteriorly being somewhat below middle of eye; black lateral stripe on anterior half to two-thirds of tail composed of somewhat disconnected dashes; no black on the dorsal outline. Cynoscion nebulosus ee. Body very deep, the greatest depth about 2.4 times in total length without caudal fin- fold; vent situated well in advance of midbody length; depth immediately behind vent about 1.5 times diameter of eye; mouth strongly oblique, the gape anteriorly being about on a level with upper margin of eye; black lateral stripe on anterior half to two-thirds of tail continuous; dorsal outline of anterior part of tail with an almost continuous black line Menticirrhus americanus dd. Caudal portion of body without a black lateral stripe; ventral outline behind vent with a few to a series of several well-separated black chromatophores. /. Inner surface of base of primitive pectoral black; no black cross bar at shoulder, nor a black spot at lower angle of opercle; an elongate black spot on ventral surface at about midcaudal length. g. Body quite deep and robust, greatest depth 2.3 to 2.5 in total length without caudal finfold; mouth large, very broad, rather strongly oblique, the gape ante- riorly being about on a level with the middle of the eye; maxillary reaching almost to middle of eye; preopercle with a few spines Stellifer lanceo'atus gg. Body more slender, the greatest depth 2.8 to 3.2 in total length without caudal finfold; mouth somewhat smaller, narrower, rather less strongly oblique, the gape anteriorly being about on level with lower margin of pupil; maxillary reaching about to anterior margin of pupil; preopercle without evident spines. Larimus fasciatus //. Inner surface of base of pectoral not black; a broad black cross bar at shoulder (at least in fresh specimens), arid a black spot at lower angle of opercle; ventral surface of tail with a series of dark dots, none of them especially enlarged; body quite slender, the greatest depth 2.8 to 3.0 in the total length without the caudal finfold. Bairdiella chrysura KEY TO SPECIMENS 3.5 TO 4.0 MILLIMETERS LONG [Caudal fin generally more or less developed, and the notochord is bent upward posteriorly] a. Body moderately slender, the greatest depth more than 3 (about 3.3 to 3.8) times in length to end of notochord. b. Body without a black longitudinal stripe on side; mouth strongly oblique, the gape anteriorly above middle of eye; vent near midbody length, the distance from snout to vent being contained 2.0 to 2.1 times in length to end of notochord. c. Body abruptly more slender posterior to vent; greatest depth of body 3.3 to 3.4, and the greatest depth behind vent 6.2 to 7.5 in length to end of notochord; eye large, equal to length of snout, its diameter 1.35 to 1.4 in greatest depth behind vent. Leiostomus xanthurus cc. Body proportionately deeper behind vent; greatest depth of body 3.4 to 3.8, and the greatest depth behind vent 4.8 to 5.6 in length to end of notochord; eye smaller, shorter than snout, 1.9 to 2.2 in greatest depth behind vent Micropogon undulatus 46 BULLETIN OF THE BUREAU OF FISHERIES bb. Body with a black longitudinal stripe on side; ventral outline of tail with a rather close-set series of dark spots, none of the spots especially enlarged; mouth less strongly oblique, the gape anteriorly about on level with middle of e3re; vent situated behind midbody length, the distance from snout to vent being contained 1.75 to 1.8 times in length to end of notochord; body moderately elongate, the greatest depth 3.4 to 3.6 in length to end of notochord Cynoscion nebulosus aa. Body deeper, the greatest depth generally less than 3 (about 2.2 to 3.0) times in the length to the end of the notochord. d. Inside of base of primitive pectoral pale, never black; body moderately elongate; greatest depth generally more than 2.5 (about 2.6 to 3.0) times in length to end of notochord; vent at or near midbody length, the distance from snout to vent being contained about 1.75 to 2.0 times in the length to end of notochord. e. Body without a dark longitudinal stripe on side; ventral outline without a continuous stripe, with or without a series of dark spots; vent almost at midbody length, the dis- tance from snout to vent being contained about 1.9 to 2.0 times in the length to end of notochord. /. No black cross bar at shoulder; ventral outline of tail with a series of well-separated black spots, one of the spots at about midcaudal length enlarged, generally with branching lines; chest and abdomen with a few large branching chromatophores; greatest depth 2.7 to 3.0 in length to end of notochord Cynoscion regalis //. Shoulder region with a broad ill-defined black cross bar (at least in fresh specimens) ; ventral outline of tail with several indefinite dark spots, also two on the back, none with branching lines; greatest depth about 2.6 in length to end of notochord. Bairdiella chrysura ee. Body with a dark longitudinal stripe on the side, one along the ventral outline of the tail and another one on the back; vent somewhat behind midbody length, the distance from snout to vent being contained about 1.75 times in the length to end of noto- chord; greatest depth about 2.7 in length to end of notochord. Menticirrhus americanus dd. Inside of base of primitive pectoral black; body quite deep; greatest depth generally less than 2.5 (about 2.2 to 2.5) times in length to end of notochord; vent notably behind midbody length, the distance from the snout to vent being contained about 1.5 to 1.7 times in the length to the end of the notochord. g. Body robust anteriorly; a large vacant space between brain and cranium; preopercle with spines; mouth moderately oblique, the gape anteriorly not much above lower margin of e3’e; maxillary reaching about to pupil Stellifer lanceolatus gg. Body more compressed anteriorly; no vacant space between brain and cranium; no preopercular spines; mouth strongly oblique, the gape anteriorly on level with middle of eye; maxillary reaching under middle of eye Larimus fasciatus KEY TO SPECIMENS 5 TO 6 MILLIMETERS LONG [The vertical fins, and sometimes the pectoral fins, are sufficiently developed to permit the enumeration of at least the soft rays] a. Anal fin long, with II, 9 to 13 rays. (Spines not always developed.) b. Body moderately deep, the greatest depth about 2.7 to 3.0 in length to base of caudal fin; mouth strongly oblique, the gape anteriorly only slightly below upper margin of eye; anal with 11 or 12 soft rays; body without dark stripes or bars; a prominent black chro- matophore present below posterior half of base of anal Cynoscion regalis bb. Body rather deep, greatest depth about 3.1 in length to base of caudal fin; mouth rather strongly oblique, the gape anteriorly about on level with middle of eye; anal with 12 or 13 soft rays; body without dark stripes or bars; ventral outline behind vent with a series of small dark spots of about uniform size Leiostomus xanthurus bbb. Body quite deep, greatest depth 2.6 to 2.8 in length to base of caudal fin; mouth moderately oblique, the gape anteriorly usually on level with lower margin of pupil; anal with 9 or 10 soft rays; a broad black ill-defined bar across chest and extending to shoulder; ventral outline behind vent with a single prominent black ehromatophore situated below last anal rays Bairdiella chrysvra aa. Anal fin shorter, with I or II, 6 to 8 ra\s. (Spines not always developed.) c. Body moderately slender, the greatest depth more than 2.2 times in the length to base of caudal; vent at or near midbod3r length, never more than an eye’s diameter in advance of or behind the midpoint between tip of snout and base of caudal. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 47 d. Anal fin with only 1 spine and 7, rarely 8, soft rays; body rather deep, compressed, depth about 2.75 in length to base of caudal; mouth moderately oblique, the gape wholly below level of lower margin of eye; caudal portions of body usually with a dark stripe along middle of side and another along ventral edge, these sometimes faint or broken up into spots Menticirrhus americanus dd. Anal fin with 2 spines; body without stripes, or at most with an indication of a stripe along middle of side of caudal portion, the color of the stripe when present being subsurface. e. Eye small, shorter than snout, the diameter also less than depth of caudal peduncle. /. Mouth strongly oblique, the gape anteriorly about on level with middle of eye; maxillary reaching to or slightly past anterior margin of eye; anal with 8 soft rays; color plain, consisting principally of dark chromatophores along ventral outline, none on head or side Micropogon undulatus ff. Mouth slightly oblique, the gape anteriorly about on level with lower margin of eye; maxillary reaching about under middle of eye; anal with 6 or 7 soft rays (sometimes insufficiently developed to be enumerated accurately) ; color consisting principally of dark chromatophores on head, along sides and along ventral outline behind vent Pogonias cromis ee. Eye larger, equal to or longer than snout, the diameter also equal to or greater than depth of caudal peduncle; anal fin with II, 8 rays. g. Body moderately slender, depth 3.3 to 3.5 in length to base of caudal; vent somewhat nearer base of caudal than tip of snout; distance from snout to vent 1.6 to 1.85 in length to base of caudal; a broken dark band on middle of side under second dorsal sometimes present, the color being subsurface; a large black chromatophore with branching lines present below base of posterior anal rays and generally another one on upper part of side, some- what in advance of origin of second dorsal Sciaenops ocellatus gg. Body rather deep, greatest depth 2.6 to 2.8 in length to base of caudal; vent generally nearer tip of snout than base of caudal; distance from snout to vent 2.0 to 2.3 in length to base of caudal; no dark lateral band; a few dark points on chest and abdomen; a slight dark spot at origin of anal generally present and another definite one at end of anal; a slight shoulder spot present Stellifer lanceolalus cc. Body very deep, depth about 2.15 times in length to base of caudal; vent notably more than an eye’s diameter nearer base of caudal than tip of snout; distance from snout to vent 1.4 times in length to base of caudal; mouth strongly oblique; the gape anteriorly some- what above middle of eye; anal with II, 6 to 8 rays; inside of base of pectorals black. Larimus fasciatus KEY TO SPECIMENS 8 TO 10 MILLIMETERS LONG [AH the fins are developed and an accurate enumeration of the rays is now possible] Anal fin with a single weak spine and 7 or 8 soft rays; body wholly or largely covered with large black chromatophores; no dark band on snout in front of eye; base of pectoral fins not black. b. Body moderately compressed and fairly deep, greatest depth 3.6 to 3.8 in length to base of caudal fin; pupil of eye almost perfectly round; caudal fin nearly symmetrical, long and pointed; anal fin usually with I, 7 (rarely 8) rays; spinous dorsal often partly black; ventrals colorless Menticirrhus americanus bb. Body strongly compressed and deep; depth 2.8 to 3.0 in length to base of caudal; pupil of eye elliptical, that is, vertically elongate; caudal fin asymmetrical, more or less rounded, never sharply pointed, the longest rays in lower half of fin; anal fin usually with I, 8 (rarely 9) rays; spinous dorsal and ventrals usually wholly black _ _ Menticirrhus saxatilis bbb. Body somewhat compressed, rather broad and low; depth 3.6 to 4.0 in length to base of caudal; pupil of eye elliptical, that is, vertically elongate; caudal fin asymmetrical, more or less rounded, never sharply pointed, the longest rays in lower half of fin; anal fin with I, 7 rays; fins all colorless Menticirrhus littoralis 48 BULLETIN OF THE BUREAU OF FISHERIES aa. Anal fin with 2 spines, usually rather strong, and 6 to 13 soft rays; body generally with few or no black chromatophores on sides and back, largely unspotted, c. Dorsal fin with X to XII-I, 20 to 22 rays (the! spines sometimes not all developed). d. Body quite deep, depth 2.6 to 2.75 in length to base of caudal; an abrupt decrease in depth posterior to vent; vent more than an eye’s diameter in advance of anal; mouth strongly oblique, the gape anteriorly only a little below middle of eye; anal fin with II, 8 rays; caudal fin long, pointed; an elongate black spot or line in advance of upper anterior angle of gill opening; no black chromatophores at origin of first dorsal. Stellifer lanceolatus dd. Body moderately deep, depth about 2.5 in length to base of caudal; no abrupt decrease in depth posterior to vent; vent not more than an eye’s diameter in advance of origin of anal; anal fin with II, 9 or 10 rays; caudal fin round; no black spot or line in advance of upper anterior angle of gill opening; some black chromatophores on upper part of side near origin of first dorsal Bairdiella chrysura ddd. Body somewhat more slender, depth about 3.0 in length to base of caudal; no abrupt decrease in depth posterior to vent; vent less than an eye’s diameter in advance of origin of anal; mouth rather strongly oblique, the gape anteriorly about on level with lower margin of pupil; anal fin with II, 6 or 7 rays; no black spot or line in advance of upper anterior angle of gill opening; upper surface of head and sides variously dotted with black chromatophores Pogonias cromis cc. Dorsal fin with X or XI— I, 23 to 34 rays (the spines not always developed) . e. Dorsal and anal fins long, the former with 30 to 34, and the latter with 12 or 13 soft rays; mouth moderately oblique, the gape anteriorly slightly above level of lower margin of eye; maxillary Teachings almost to middle of eye; caudal fin rather short, slightly rounded or straight; a few dark chromatophores present along ventral outline only Leiostomus xanthurus ee. Dorsal and anal fins shorter, the former with 23 to 29 rays, and the latter with 6 to 12 rays (rarely with 13 in Cy noscion regalis). f. Body very deep, strongly compressed, depth about 1.9 in length to base of caudal; mouth rather strongly oblique, the gape anteriorly a little above lower margin of eye; anal fin short with II, 6 rays; caudal fin rather long pointed; sides with large branching chromatophores, variable in number among specimens, partly at least arranged in longitudinal series; lower half of pectorals and ventrals black. Larimus fasciatus ff. Body more slender, the depth more than 2.5 times in length to base of caudal; anal fin longer, with more than 7 soft rays; basal half of pectorals and ventrals not black. g. Anal generally with 8 or 9 soft rays ( Cynoscion nothus rarely with 10 rays). h. Second dorsal moderately short, with 23 to 25 soft rays; eye large, about as long as spout, its diameter nearly equal to least depth of caudal pe- duncle; a broken lateral stripe present posteriorly, the color subsurface; body variously marked above and below with black chromatophores; a large branching chromatophore at base of posterior anal rays, and usually another prominent one on upper part of side, somewhat in advance of second dorsal Sciaenops ocellatus hh. Second dorsal rather long, with 28 or 29 soft-rays; eye small, shorter than snout, its diameter shorter than least depth of caudal peduncle; ventral surface with small black chromatophores; sides and back plain without black markings Micropogon undulatus hhh. Second dorsal moderately long, generally with 25 to 27 rays (sometimes 24 to 28) ; eye rather small, somewhat shorter than snout, its diameter notably shorter than least depth of caudal peduncle; body moderately deep; depth 3.0 to 3.25 in length to base of caudal; vent far in advance of origin of anal, the distance notably longer than diameter of eye; black lateral band wanting; 2 elongate dusky spots at base of anal, 1 at anterior rays and another at the posterior ones Cynoscion nothus gg. Anal generally with 10 to 12 soft rays ( Cynoscion regalis sometimes with 13). i. Body moderately deep; depth 2.95 to 3.0 in length to base of caudal; vent more than an eye’s diameter in advance of origin of anal; snout REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 49 moderately blunt; lower jaw scarcely projecting; no black lateral stripe, sides generally with groups of black chromatophores; a large black branching chromatophore below base of last anal rays. _ Cynoscion regalis ii. Body more elongate; depth 3.6 to 4.0 in length to base of caudal; vent less than an eye’s diameter in advance of origin of anal; snout pointed; lower jaw projecting prominently; a black lateral stripe present on posterior part of body; a black band in front of eye, sometimes extend- ing through eye to opercle; no prominent black chromatophores at base of anal Cynoscion nebulosus KEY TO SPECIMENS 15 TO 20 MILLIMETERS LONG [Scales usually are developed at this range in size and sometimes the body is fully covered at a length of 20 millimeters] Anal fin with a single weak spine and 7 to 9 soft rays; mouth more or less inferior, nearly hori- zontal; lower jaw with a rather prominent knob at tip (which develops into a short thick barbel as the fish grows). b. Body moderately deep, greatest depth 3.6 to 3.8 in length to base of caudal; pupil almost perfectly round; caudal fin long and pointed, the longest rays being longer than head; anal with I, 7 (rarely 8) rays; spinous dorsal and ventrals with black chromatophores, not wholly black Menticirrhus americanus bb. Body deeper, greatest depth 3.25 to 3.4 in length to base of caudal; pupil elliptical; caudal fin broadly pointed, the longest rays in lower half of fin, notably shorter than head; anal with I, 8 (rarely 9) rays; spinous dorsal and ventrals black. Menticirrhus saxatilis bbb. Body quite elongate, greatest depth 3.75 to 4.2 in length to base of caudal; pupil elliptical; caudal fin asymmetrically rounded, the longest rays in lower half of fin, shorter than head; anal with I, 7 rays; spinous dorsal and ventrals not pigmented. Menticirrhus littoralis . Anal fin with 2 rather strong spines and 6 to 13 soft rays; lower jaw without a knob at tip. c. Snout more or less pointed; lower jaw projecting prominently; anal with IT, 9 to 13 rays (rarely II, 8 in Cynoscion nothus). d. Anal usually with 9, occasionally with 8 or 10, soft rays; body rather deep, the greatest depth 3.2 to 3.6 in length to base of caudal; caudal fin evidently very long and pointed, the longest rays longer than head; chromatophores small, forming indefinite blotches, blotches on back and in lateral line if present separate, not forming cross bars Cynoscion nothus dd. Anal usually with 11 or 12, occasionally with 13, soft rays; body rather deep as in C. nothus; caudal fin rather long and pointed, the longest rays about equal to length of head; chromatophores large, forming rather distinct blotches, those on back and in lateral line connected, forming more or less definite cross bars on anterior part of body Cynoscion regalis ddd. Anal with 10 or 11 soft rays; body slender, the greatest depth 3.9 to 4.1 in length to base of caudal; caudal fin short, broadly pointed, the longest rays notably shorter than head; chromatophores very small, arranged so as to form an indefi- nite lateral band and another band along base of dorsal and generally extending forward on head; no blotches or cross bars present Cynoscion nebulosus cc. Snout not pointed, usually short and blunt; lower jaw not projecting (except in Larimus fasciatus ) equal to or more usually shorter than upper jaw; anal with II, 6 to 8 rays, exclusive of Leiostomus xanthurus which has II, 12 or 13. e. Anal long, with 12 or 13 soft rays; second dorsal with 30 to 34 soft rays; body mod- erately deep, the greatest depth about 3.4 to 3.8 in length to base of caudal; margin of caudal fin distinctly concave Leiostomus xanthurus ee. Anal shorter, with 6 to 10 soft rays; second dorsal shorter, with 19 to 29 soft rays; margin of caudal rounded or pointed. /. Second dorsal with 19 to 22 soft rays; first dorsal with 10, 11, or 12 spines. g. First dorsal with 11 or 12 spines; anal with 8 to 10 soft rays. h. Head broad, spongy; eye small, shorter than the snout; interorbital notably broader than eye; anal with 8 soft rays; caudal fin long, pointed Stellifer lanceolatus 50 BULLETIN OF THE BUREAU OF FISHERIES hh. Head narrow, compressed, not spongy; eye larger, fully as long as snout, about as wide as interorbital; anal with 9 or 10 soft rays; caudal fin rather short, the margin round Bairdella chrysura gg. First dorsal with 10 spines; anal with 6 or 7 soft rays; head deep, compressed; mouth inferior; a series of short barbels generally visible on lower jaw; dark cross bars present Pogonias cromis ff. Second dorsal with 23 to 29 soft rays; first dorsal with 10 spines only. i. Body very deep, compressed, the depth 2.4 in length to base of caudal; mouth strongly oblique, the gape anteriorly above lower margin of eye; second dorsal with 24 to 27 soft rays; anal with 6 to 8 soft rays; caudal fin long, pointed; body with very large chromatophores, forming an indefinite band on anterior part of body; pectorals and ventrals largely black; spinous dorsal with big black spots Larimus fasciatus ii. Body more elongate, the depth being contained more than three times in length to base of caudal; mouth only slightly oblique, the gape wholly below the eye; caudal fin more or less rounded, not long and pointed. j. Second dorsal with 28 or 29 soft rays; caudal fin long, pointed, as long as head; a row of short, slender barbels visible on chin in some specimens. Micropogon undulatus jj. Second dorsal with 23 to 25 soft rays; caudal fin short, angulate, shorter than head; no barbels on chin Sciaenops ocellatus KEY TO SPECIMENS 30 TO 35 MILLIMETERS LONG [At this size the body generally is fully pigmented and frequently the color pattern of the adult is present] a. Anal fin with a single weak spine and 7 to 9 soft rays; mouth notably inferior, nearly or quite horizontal, the conical snout usually projecting beyond it; lower jaw with a very short thick barbel at tip. b. Ventral fins small, shorter than pectorals, not reaching beyond tips of pectorals; scales on chest not much smaller than on rest of body; sides with dark blotches or more usually with black bands. c. Body slender, greatest depth 3.8 to 4 in length to base of caudal; pupil round; anal rays typically 7, rarely 8; caudal fin long and pointed, the longest rays fully as long as head; sides with dark blotches, usually forming indefinite cross bars, not forming a V on the side under spinous dorsal Menticirrhus americanus cc. Body deeper, greatest depth 3.5 to 3.65 in length to base of caudal; pupil eliptical; anal rays typically 8, sometimes 9; caudal fin not long and pointed, angulate, the longest rays shorter than the head; sides usually with black cross bars, the second and third nearly or quite meeting on middle of side, forming a V under spinous dorsal. Menticirrhus saxatilis bb. Ventral fins large, longer than pectorals and reaching far beyond them; scales on chest notably reduced in size; body slender, greatest depth 4.05 to 4.3 in length to base of caudal; pupil eliptical; anal rays typically 7; caudal fin short, the lower lobe longest, notably shorter than head; sides with a few black dots, no large blotches or bars, color mostly silvery Menticirrhus littoralis aa. Anal fin with 2 rather strong spines; mouth superior, terminal, or only slightly inferior, the snout not projecting prominently; lower jaw with or without barbels, numerous if present. d. Snout pointed; lower jaw projecting prominently; barbels wanting; upper jaw anteriorly with enlarged canines, curved backward; body elongate, greatest depth 3.2 to 4.2 in length to base of caudal; anal with 9 to 13 soft rays (rarely only 8 in Cynoscion nothus). e. Body rather deep, greatest depth 3.2 to 3.6 in length to base of caudal; snout only moderately pointed; mouth oblique, the gape anteriorly somewhat above lower mar- gin of eye; lower jaw projecting moderately; caudal fin long, pointed, the longest rays longer than head; sides usually with dark blotches or dark bars, no longitudinal black bands present. /. Anal fin usually with 9, occasionally with 8 or 10 soft rays; sides generally with dark blotches in two series, one along the back and the other on the middle of the side Cynoscion nothus REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 51 ff. Anal fin usually with 11 or 12, occasionally with 13, rays; sides with more or less definite dark cross bars Cynoscion regalis ee. Body more slender, greatest depth 3.95 to 4.2 in length to base of caudal; snout sharply pointed; mouth less oblique, the gape anteriorly below eye; lower jaw projecting strongly; caudal fin not very long nor sharply pointed, the longest rays shorter than head; anal fin with 10 or 11 rays; side and back each with a black longitudinal band, no blotches or cross bars present Cynoscion nebulosus dd. Snout not pointed; lower jaw equal to or shorter than the upper, projecting slightly in Larimus fasciatus; no enlarged canine teeth in upper jaw; body sometimes deep, sometimes elongate; anal with 6 to 8 soft rays, exclusive of Leiostomus xanthurus which has 12 or 13. g. Anal long, with 12 or 13 soft rays; second dorsal with 30 to 34 soft rays; body moder- ately deep, greatest depth about 3.3 times in length to base of caudal; margin of caudal fin distinctly concave Leiostomus xanthurus gg. Anal shorter, with 6 to 10 soft rays; second dorsal shorter, with 19 to 29 soft rays; margin of caudal round or pointed, not concave. h. Second dorsal with 19 to 22 soft rays; first dorsal with 10, 11, or 12 spines. i. Body deep, compressed, depth 2.5 to 2.65 in length; head narrow, compressed; mouth horizontal, inferior; lower jaw distinctly shorter than the upper one, with numerous small barbels; anal with 6 or 7 soft rays; body with about 5 black cross bars Pogonias cromis ii. Body not quite as deep, depth about 2.9 to 3.1 in length to base of caudal; mouth somewhat oblique, terminal, the jaws of about equal length; no bar- bels on lower jaw; anal with 8 to 10 soft rays; no cross bars on body. j. Head, narrow, compressed; interorbital not broader than large eye; skull hard, not spongy; anal with 9 or 10 soft rays; caudal fin round, not long; no dark blotches on back Bairdiella chrysura jj. Head broad, scarcely compressed; interorbital about two times as wide as the small eye; skull soft, spongy, cavernous; anal with 8 soft rays; caudal fin long and pointed; back with a series of black blotches. Stellifer lanceolatus hh. Second dorsal with 23 to 29 soft rays; first dorsal with 10 spines only. k. Body deep, compressed, greatest depth 2.6 to 2.8 in length to base of caudal; mouth strongly oblique, the lower jaw projecting somewhat; anal with 6 to 8 soft rays; caudal fin long and pointed; body with a broad cross bar under spinous dorsal and 4 or 5 narrower ones behind it Larimnus fasciatus kk. Body elongate, the depth about 3.3 to 3.6 in length to base of caudal; mouth horizontal, inferior, lower jaw included; anal with 8 soft rays; no cross bars on body. I. Second dorsal with 28 to 29 soft rays; a row of short slender barbels usually visible on lower jaw; caudal fin long and pointed, the longest rays equal to length of head; body with small dark spots on sides and back. Micropogon undulatus II. Second dorsal with 23 to 25 soft rays; barbels wanting; caudal fin short, only slightly angulate, the longest rays notably shorter than head; body with dark spots placed as in M. undulatus, but larger and more distinct. Sciaenops ocellatus THE WHITINGS OR KINGFISHES (Menticirrhus americanus, M. saxatilis and M. littoralis) Three species of whiting, namely, Meticirrhus americanus, M. saxatilis, and M. littoralis, occur on the coast of North Carolina. The species resemble each other so closely that the fishermen generally fail to distinguish them, and they are not separated in the market. The most widely used common names, in books at least, are whiting and king whiting. Other local names are king-fish, roundhead, sea mullet, and sea mink. The fish are known at Beaufort as “sea mullet”, and much less commonly as “sea mink.” Since the species are not separated in the market, the relative abundance of each cannot be ascertained from commercial records. It is 52 BULLETIN OF THE BUREAU OF FISHERIES known from personal observation, however, that americanus is by far the most numer- ous and commercially the most important species in the vicinity of Beaufort. The other two are about equally common, but not abundant. The whitings are choice food fishes and generally command a good price. One or more species is common enough to enter into the commercial catches along the coast all the way from Massachusetts to Texas. The States having the largest catches in 1929 are the following: 6 New Jersey, 52,408 pounds; Virginia, 54,650; North Carolina, 387,168; South Carolina, 100,754; Georgia 51,500; Florida, 664,943; and Louisiana, 41,829 pounds. It is evident from the foregoing records that the whitings are important food fishes over a wide range of the eastern and southern shores of the United States. The writers are pleased, therefore, to offer some new information in regard to the life histories of these useful fishes. No young under 9 or 10 millimeters in length have been recognized either as ■saxatilis or littoralis. Since the local species of the genus are readily distinguishable at the size mentioned, as pointed out in the descriptions of the young, it is improbable that the smaller specimens (unless it be the very smallest ones) consist of more than one species, and all seem referable to americanus. The three local species apparently all spawn simultaneously at Beaufort, the reproductive period occurring in the spring and early summer. A rapid rate of growth is indicated during the first several months of life. The habitat of the smaller young, ranging from about 10 to 60 millimeters in length, appears to be identical in saxatilis and littoralis, as these fish were taken only in the surf along the outer shores of the “banks”, while americanus occurs further off shore and also in the inside protected waters. The adults of americanus and saxatilis are found both in the inside and outside waters, while littoralis evidently is confined almost entirely to the outside open waters. CHARACTERS OF THE ADULTS The body is long and rather low in the three species under consideration. The back is notably narrower than the abdomen, and much more strongly curved than the ventral outline. The head is low, and the conical snout projects well beyond the horizontal mouth. A short, thick barbel is present at the chin. The number of vertebrae is about the same in each species. In one specimen of each species counted, americanus and littoralis each had 10 body and 15 caudal vertebrae, and saxatilis had 10 body and 16 caudal vertebrae. The pupil of the eye is large and nearly or quite round in americanus at all ages, while it is smaller and vertically quite elongate in the young of about 6 inches and less in length of saxatilis and littoralis. In large preserved specimens of the last-mentioned species the pupil does not always appear elongate. Dorsal with 10 or 11 spines; anal with 1 weak spine. The chief diagnostic characters of the adults are included in the following key. KEY TO THE SPECIES a. Scales on chest not especially smaller than on sides; pectoral fins long, reaching to or past tips of the ventrals. b. Sides plain or with obscure bars, not forming a V; anal usually with 7 soft rays; none of the dorsal spines especially produced, and none reaching far, if at all, beyond origin of second dorsal; scales 86 to 90, counting vertical series above lateral line americanus » The statistics are taken from the “ Fisheries Industries of the United States in 1930”, by R. H. Fiedler. Appendix II, Report, U.S. Commissioner of Fisheries, 1931 (1932), pp. 109-552, 23 figs. Washington. The fish are listed under "king whiting” or “king- fish.” REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 53 bb. Sides usually with black bars, the one on the nape and the one below the spinous dorsal meet- ing on the side forming a V; anal usually with 8 soft rays; the longest dorsal spine pro- duced, reaching far past origin of second dorsal; scales 91 to 96, counting vertical series above lateral line saxatilis aa. Scales on chest much smaller than on sides; pectoral fins short, failing conspicuously to reach tips of ventrals; dorsal spines short, none of them produced; anal usually with 7 soft rays; scales 70 to 75, counting vertical series above lateral line; coloration normally plain silvery littoralis MENTICIRRHUS AMER1CANUS (Linnaeus) Menticirrhus americanus ranges from New York to Texas and apparently is commercially the most important species of the genus from Chesapeake Bay south- ward. The maximum weight reported is 2% pounds. However, the average weigh probably does not exceed a half pound per fish. This species is taken throughout the year in the vicinity of Beaufort. It becomes quite scarce, however, during the winter and especially during cold spells. It is most numerous in the spring when large catches are made with “sink nets”,6 operated chiefly from Beaufort Inlet to Cape Lookout. During comparatively recent years, since the otter trawl has come into use locally, it is taken in the deeper waters with that gear also. In the inside waters it is taken chiefly with drag nets or seines. It takes the hook now and then, but it scarcely holds a place at Beaufort as a sport or game fish. SPAWNING Fish with well-developed roe have been seen at Beaufort from time to time from April to June. However, mature eggs of this species have not been secured. It is evident from the capture of two larvae 4.0 and 5.5 millimeters long in the tow on April 23 and 26 (1927) that spawning, some years at least, commences as early as April. The capture of young under 10 millimeters, and some under 5.0 millimeters in length during each succeeding month until about the middle of September shows almost certainly that the spawning season at Beaufort extends from April through August and possibly into the early part of September. (See table 1.) The principal spawning period, judging mainly from the abundance of the larvae in the tow, seems to extend from the latter part of June through July and August. Fish with large roe were seen most commonly during May and June. Spawning in the vicinity of Beaufort probably occurs chiefly along the outside shores of the “banks”, although it seems probable that some spawning may take place also within the inside waters. While the eggs if taken (in the tow) were not recognized, it is known that adult fish are common along the outer shores of the banks during the spawning season, where they no doubt congregate to carry out reproduc- tive activities. Furthermore, 15 of the 18 young, 5.0 millimeters and under in length in the collection were caught along the outside shores. The larger young, up to 20 millimeters in length, of which many were taken, too, were much more numerous there than in the inside waters. It seems highly probable, therefore, that spawn- ing in the vicinity of Beaufort takes place chiefly in the ocean along the shores of the “banks.” Welsh and Breder (1923, p. 187) suggest that this species may have two spawning seasons. No evidence of more than one spawning period was secured at Beaufort. (See table 1.) However, the spawning season, as pointed out in a preceding para- « Sink nets are gill nets, weighted and sunk to the bottom (hence sink nets) in several fathoms of water. 54 BULLETIN OF THE BUREAU OF FISHERIES graph, evidently is a long one. Therefore, the range in the size of the young of any one season is very great. The range in length in September, for example, according to examples measured extends from 2.75 to 156 millimeters. (See table 2.) Speci- mens of early and late spawnings, without intermediate ones, might readily lead to the conclusion that they represented the product of two distinct spawning periods. It is possible that the authors mentioned might have been misled because of insuffi- cient material, that is, by the absence of specimens in their collections from the entire spawning period. Spawning evidently takes place about simultaneously in Chesapeake Bay and Beau- fort, as Hildebrand and Schroeder (1928, p. 293) report the observation of ripe fish from Cape Charles, Va., in May and other ripe ones in the Norfolk (Ya.) market in June. These writers, also, report the capture in Chesapeake Bay of “young, three fourths inch or more in length, early in the summer.” Welsh and Breder (1923, p. 186), working at Atlantic City, N.J., on the other hand, had found no ripe fish and no spent ones there as late as August (1920), although many females with well-developed but hard, row were seen. Nor do these investigators report the capture of young in that locality. Therefore, it would seem that the fish spawn much later in New Jersey than in North Carolina. Figure l.—Menticinhus americanus. From a specimen 1.7 millimeters long. DESCRIPTIONS OF THE YOUNG Specimen 1.7 millimeters long. — The head and body are rather robust, somewhat compressed, greatest depth only a little less than length of body to vent. The tail is moderately slender, it tapers gradually and ends in a sharp point as usual in recently hatched teleosts. Myomeres are too indistinct to be definitely enumerated, about 16 evident. The mouth is very strongly oblique and moderately large, and prominent teeth are present on the jaws. The eye is prominent, being fully half as long as the head, and it has a round pupil. The vent is situated somewhat in advance of midbody length. The vertical finfold, in the only specimen of this size at hand, is largely torn away, being present, however, around the tip of the tail. The distal portion of the notochord is straight and in line with the axis of the body. The color is mostly pale, with dark markings (specks) around the mouth, along most of the ventral outline and along the middle of the tail. These markings are close enough together to form a longitudinal dark line along the ventral outline of the tail and another one along the middle line of the anterior portion of the tail. These dark lines are quite characteristic and are an aid in identifying the very small larvae with larger ones in which the dark lines are more strongly emphasized (fig. 1). REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 55 The specimen described, which was taken in the tow, has been compared with preserved specimens of newly hatched larvae of saxatilis. The latter were hatched by the late William W. Welsh from eggs stripped from fish taken at Atlantic City, N.J., on July 27, 1920. The larvae (preserved in formalin) although quite as long as the specimen described, evidently are much younger, as shown by the state of development. The resemblance is not striking, as the specimen of americanus is much deeper anteri- orly, and the mouth, jaws, and teeth are much further advanced in development. The specimens of saxatilis retain no color markings. It is evident, however, from the description and illustrations by Welsh and Breder (1923, pp. 190-193) that rather definite markings were present when the larvae were alive. A slight resem- blance is evident in the markings of the most advanced stage figured by Welsh and Breder of saxatilis and the 1.7 millimeter specimen of americanus already described. This resemblance consists chiefly in a series of dark dots along the ventral outline of the tail. The pupil of the eye is round in the specimens at hand of both species, although in larger examples (10.0 millimeters and upward in length) of saxatilis the pupil is vertically elongate, whereas it is always nearly or quite round in americanus. The comparison of the small fry at hand of americanus and saxatilis, because of the comparatively great difference in age, even though they are of about equal length, does not aid greatly in determining the likenesses and differences that exist between the early larval stages of the two species. Since no newly hatched larvae of americanus have been secured, no young of saxatilis under 10.0 millimeters (exclu- sive of the newly hatched fry), and none of littoralis less than 9.0 millimeters in length, the study of the relationship of the smaller fry of the local species of the genus must await the collection of specimens of the proper sizes and ages. Specimens 2.9 millimeters long. — The body is very deep, somewhat compressed, with an abruptly more slender and rapidly tapering tail which ends in a sharp point. A very sharp break in the ventral outline occurs at the vent, where the depth dimin- ishes abruptly. The greatest depth of the body is only slightly less than the length of the body to the vent. Myomeres are mostly indistinct, about 23 evident. The mouth is somewhat less oblique than in the smaller specimen described in the fore- going paragraph, the margin of the upper jaw being about on the same level as the upper margin of the eye, and the gape reaches nearly to the pupil. The teeth in the jaws apparently are much smaller than in the smaller specimen described in the preceding section. The vent is situated notably nearer the tip of the tail (notochord) than the tip of the snout. The vertical finfold remains continuous, with slight indications of rays at the distal part of the tail. Pectoral finfolds are rather promi- nent, but the ventrals are not evident. The general color is pale, with faint ill- defined dark spots about the mouth and along the ventral outline of the abdomen. Two prominent longitudinal black stripes are situated on the anterior two thirds of the tail, one follows the middle line and the other the ventral outline. Several minute dark dots forming an indefinite line are present also on the dorsal outline of the anterior part of the tail (fig. 2). The very deep body, with the abruptly more slender tail and the two evident black stripes on the caudal portion of the body are the chief recognition marks at this size. These characters also identify smaller specimens, as shown in the descrip- tion of a 1.7 millimeter fish, but they are much more prominent and pronounced in somewhat larger specimens. 67094—34 2 56 BULLETIN OF THE BUREAU OF FISHERIES Specimens 3.8 millimeters long. — The body is deep and compressed, and the tail has become proportionately much deeper and less sharply tapering since a length of about 2.9 millimeters was attained. The break in the ventral contour described in smaller fish remains, but is less pronounced. The greatest depth is now equal to only about two-thirds the length of the body to the vent. The mouth, although still strongly oblique, has become somewhat lower, the gape anteriorly being only slightly above the lower margin of the eye, and the maxillary (which is now well developed) reaches a little beyond the middle of eye. The pupil is nearly or quite round, as in smaller specimens, and the vent remains situated nearer the tip of the notochord than the tip of the snout. The bases of the soft dorsal and anal fins are evident and indications of rays are present. The notochord has become curved upward, and below it the caudal fin is partly developed with rather definite rays. The pectoral fins are short and broad, with indications of rays, and the ventral fins are not evident. The color has changed little since a length of 2.9 millimeters was attained. The dark markings about the mouth, along the ventral outline of the abdomen, and on the tail remain virtually as described in the smaller fish, except that the dark markings on the dorsal outline of the tail are a little more prominent and form a somewhat more defi- Figure 2.—Menticirrhus americanus. From a specimen 2.9 millimeters long. nite stripe. Several indefinite dark markings now are present on middle of side, extending from the eye to opposite the vent (fig. 3). The principal recognition marks are the same at this size as in 2.9 millimeter specimens previously described. The chief changes have taken place in the develop- ment of the tail, which has become much deeper posteriorly, the mouth is less strongly oblique, and some of the fins are becoming differentiated. Specimens about 5.8 millimeters long. — The fish is much more shapely than it was when smaller. Yet, the body is quite unlike that of the adult, as it is deep and com- pressed. The break in the ventral outline, behind the vent, pronounced in smaller fish, is no longer prominent, and the caudal portion of the body has become propor- tionately much deeper. The greatest depth of the body is now scarcely equal to two- thirds of the preanal length and is contained about 2.75 times in the total length to the base of the caudal fin. Myomeres are indistinct anteriorly and again poste- riorly, about 25 evident. The mouth remains moderately oblique, and the lower jaw is a little shorter than the upper. The gape is wholly below the level of the lower margin of the eye and the maxillary reaches nearly opposite the anterior margin of the eye. The vent is situated a little nearer the tip of the snout than the end of the caudal fin. Considerable advancement in the development of the fins has been made since a length of 3.8 millimeters was attained, most of the soft rays being fairly well REPRODUCTION AND DEVELOPMENT OP SCIAENIDAE 57 differentiated in the second dorsal and the anal, and a fairly accurate count of the anal rays may be made, the number being 7 or 8. The caudal fin is better developed than the dorsal and anal; it is round in outline and the longest rays are about equal to the depth of the body just posterior to the vent. The pectorals apparently have made little advancement and the ventral fins are not evident. The dark longitudinal stripes on the caudal portion of the body, described in smaller fish, generally are not evident. The stripes, having broken up into large chromatophores, remain in rather definite rows in some specimens, but are more scattered in others. In still other specimens of this size, and even larger ones, the stripe along the lateral line, however, remains well defined. Considerable variation in color, that is, with respect to the number, intensity, and arrangement of the spots present, is evident. The distal portion of the tail remains pale and almost transparent (fig. 4). The chief recognition marks at this size are the deep compressed body, the large oblique mouth, the short anal fin, and the rather large, dark markings on the caudal portion of the body. However, these markings are faint in some specimens, as shown in the foregoing description, while they are large and well defined in others. Specimens 8 millimeters long. — The body is proportionately more slender than in specimens 5.8 millimeters long and tlfe head is much broader. The rest of the body, Figure i.—Menticinhus americanus. From a specimen 5.8 millimeters long. however, remains quite strongly compressed. The greatest depth is contained about 3.0 times in the total length to the base of the caudal fin. The mouth is only slightly oblique, the lower jaw is a little shorter than the upper, and the maxillary reaches nearly to the posterior margin of the eye. Considerable advancement in the develop- ment of the fins has been made. The spinous dorsal is evident now, and the ventral fins are rather long and prominent. The pectoral fins have increased greatly in length and reach the vent. The caudal fin has become produced and pointed, the longest rays being nearly as long as the head. The middle portion of the body is largely spotted with dark chromatophores of various sizes, which are present also on 58 BULLETIN OF THE BUREAU OF FISHERIES the head and nape. Dusky color is present around the mouth and across the middle of the mandible. The distal portion of the tail remains pale and more or less trans- parent, the upward curved notochord within it being plainly visible under the micro- scope with transmitted light. Slight dark color has appeared on the anal and on the margin of the spinous dorsal (fig. 5). The fish at this size bears about the same recognition marks as in somewhat smaller specimens. However, the head is less strongly compressed, the mouth is less oblique, the spinous dorsal and ventral fins are now developed, the caudal fin has be- come somewhat produced and pointed, and the dark markings on the body have in- creased greatly in number and cover a larger area. Specimens 10 millimeters long. — The body is moderately deep, compressed, the greatest depth being contained about 3.6 to 3.8 times in the length to base of caudal fin. The head is moderately compressed, and its length is greater than depth of body, being contained about 2.75 to 3.1 in length to base of caudal. The interorbital is strongly convex; the eye is only a little longer than the snout, and it has a rather large, perfectly round pupil. The mouth is moderately oblique, the upper lip ante- riorly being slightly below the level of the lower margin of the eye; the maxillary reaches nearly to posterior margin of eye; and the upper jaw projects beyond the lower one. The fins are all well developed, and the rays may be enumerated accurately. The longest dorsal spines reach only slightly past the origin of the second dorsal when deflexed ; the caudal fin is nearly symmetrical and quite pointed ; and the ventral and pectoral fins are long, generally coterminal, reaching nearly or quite to the vent. The fish has dark chromatophores almost everywhere, except on its ventral surface. They are most numerous on the side from opercular margin to end of base of anal, and few on the head and caudal peduncle. The spinous dorsal and the anal fin generally are partly black; the caudal fin has at most a few dark points on the base, and the other fins are without color (fig. 6). REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 59 The present species, at a length of 10 millimeters, is intermediate of the other local species of the genus in the shape of the body, that is, the body is deeper and more strongly compressed (especially the head) anteriorly than in littoralis, and less so than in saxatilis. It also has a larger and an almost perfectly round disk-shaped pupil, whereas the pupil is vertically elongate and somewhat eliptical in littoralis and saxatilis. (The shape of the pupil often becomes somewhat distorted during preser- vation and, therefore, the difference in shape is not always as plainly evident in pre- served specimens as in fresh material.) The caudal fin is quite long and pointed and nearly symmetrical in americanus. In the other species this fin is quite broadly and asymmetrically rounded, the longest rays being in the lower half of the fin. In americanus the spinous dorsal often is partly black, while the ventral fins are color- less, or at most with only a few dark points. In saxatilis the spinous dorsal and the ventral fins are wholly black, whereas in littoralis these fins are colorless. Specimens 13 to 15 millimeters long. — The body has become less strongly com- pressed since a length of 10 millimeters was reached and has become deeper in the region of the vent, causing a notably less pronounced tapering in the depth from the head to the vent. The length and greatest depth of the body, however, remains proportionately about the same as in the 10-millimeter fish. The head is less com- pressed; the mouth is more nearly horizontal, and the snout now projects slightly beyond the upper jaw. The caudal fin has become proportionately longer and more strongly pointed. No other changes in the fins that appear to be worthy of note are evident. The color has changed little. The general tendency is toward fewer large branching cliormatophores and more numerous small dark points. The spinous dorsal and the anal are at least partly black; the second dorsal, the pectorals, and the caudal are unmarked, or the caudal at most may have a few dark markings on the base; and the ventral fins may have only a few dark dots or be solidly black at the base. The fish obviously has made fair headway toward the adult form in the shape of the head, the body, and the mouth. This species at the size just described differs from littoralis about as in the smaller fish previously described, that is, the body is deeper, the pupil is round, the caudal is longer and more pointed, the spinous dorsal is at least partly black, the soft dorsal is unmarked, and the ventral fins have at least some dark points if not partly black. No specimens of saxatilis between 11 and 17 millimeters in length are at hand. Therefore, the exact differences between specimens of americanus arid saxatilis of this size cannot be stated at this time. Specimens 18 to 20 millimeters long. — The body lias continued to grow rather more elongate and rounder. The head remains somewhat deeper than wide, however, and the greatest depth (which occurred at the posterior part of the head in smaller specimens, now falls under the spinous dorsal) is contained 3.6 to 3.8 times in the length to the base of the caudal fin. The mouth is nearly horizontal and inferior, and the conical projecting snout is only a little shorter than the eye. A slight knob has appeared at the tip of the lower jaw, which is the beginning of the development of the characteristic mandibular barbel of the adults of the genus. Scalation is nearly complete (although not shown in the illustration). The caudal fin remains long and pointed. The longest rays now definitely occur in the lower half of the fin, and are somewhat longer than the head. Pigmentation has progressed rather rapidly. While variation is evident among specimens, most usually the entire body is spotted with black or dark brown. The spinous dorsal and the anal fins are largely black, as in smaller fish; the ventral fins bear dark spots; and generally two dark blotches have appeared on the base of the caudal fin (fig. 7). 60 BULLETIN OF THE BUREAU OF FISHERIES Considerable headway toward acquiring the shape and form of the adult has been made, and the conical projecting snout and the rudimentary mandibular barbel at once identify the fish as a Menticirrhus. The shape of the body in the present species remains intermediate of the other local species of the genus, that is, the body is slightly deeper than in littoralis, but not as deep as in saxatilis. The pupil, of course, remains round, whereas it is definitely elliptical in the other two species. The tail is longer and more strongly pointed, the longest rays being longer than the head, whereas in littoralis and saxatilis the longest rays are shorter than the head. Figure 7. — Menticirrhus americanus. From a specimen 20 millimeters long. Specimens SO to 35 millimeters long. — The body has become somewhat rounder anteriorly since a length of 18 to 20 millimeters was attained, and the greatest depth now is contained in the length to the base of the caudal fin 3.8 to 4 times, which are proportions prevailing in the adult. The mouth is inferior and horizontal, and the snout projects rather prominently, as in the adult. The mandibular barbel is very short, a mere knob, and scalation is complete. None of the dorsal spine are produced and the longest ones reach to the origin of the second fin when deflexed. The caudal fin is long and pointed, the longest rays being slightly below the middle of the fin and quite as long as the head. The pectoral fins are long and reach almost to the tips of the ventrals, or nearly to the vent. The body is fully pigmented, some specimens being much darker than others. The lower parts generally are silvery and the upper parts brownish. The ground colors are overlaid with irregular dark specks or spots, which follow the rows of scales more or less definitely. Some specimens have dark blotches and others bear suggestions of dark cross bars. The spinous dorsal is largely REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 61 black. A large black spot generally is present on the base of the caudal, preceded by a pale crossline, the rest of the fin being plain translucent. The anal and ventral fins usually are dusky to black, and the pectoral fins generally are plain translucent, or sometimes with dusky points (fig. 8). The round pupil and the long pointed tail remain as dominant characters in distinguishing the present species from the other two local species of the genus. The long pectorals and the large scales (which are not notably reduced on the chest) also aid in distinguishing the present species from littoralis, while it differs further from saxalitis in the more elongate body. Specimens 50 to 60 millimeters long. — The fish has acquired virtually the shape of the adult, the body being only a little more strongly compressed anteriorly. The proportions of the depth to the length are the same as in fish 30 to 35 millimeters long, and these proportions also occur in the adults. The snout is more sharply conical, projects beyond the mouth more prominently than in smaller fish, and is now fully as long as the eye. The mandibular barbel is developed and is about half as long as the pupil. No important changes have taken place in the development of the Figure Q.—MenticirThus americanus. From a specimen 59 millimeters long. fins. The spines in the first dorsal are somewhat shorter, and the lower lobe of the caudal remains notably more pointed and proportionately longer, than in the adult. The color remains about as in 35-millimeter fish. Some specimens retain indefinite dark blotches on the sides, and others have more or less definite oblique bars. | In the majority of specimens at hand the base of the caudal is dusky, and this color ex- tends on the lower lobe of the fin. The anal and ventrals are injpart white, as in the adult (fig. 9.) Specimens 50 to 60 millimeters long are enough like the adult to be recognized readily. Specimens of this size differ more strongly from full-grown fish in the shape of the caudal fin than in any other one character, and this member does not acquire the adult shape, namely, a slightly concave upper lobe and a moderately short, rather sharply rounded lower lobe, until the fish reach a length of about 100 millimeters. At a length of 50 to 60 millimeters the spinous dorsal has become quite pointed, though none of the rays are especially produced, nor do any of them reach far beyond the origin of the second dorsal. The fish may be distinguished from the other species of the genus at this size by about the same characters which identify the adults, namely, the rather small scales (86 to 90 vertical series above the lateral line), which are not 62 BULLETIN OF THE BUREAU OF FISHERIES reduced on the chest; the short anal, which typically has only 7 rays; the long pectorals, which reach nearly or quite to the tips of the ventrals; and by the presence of obscure bars and blotches on the sides and back. The large roundish pupil remains conspicuous and readily distinguishes this species from the other local forms. DISTRIBUTION OF THE YOUNG It has been shown elsewhere (p. 53) that the young under 20 millimeters in length were taken more abundantly off Beaufort Inlet than in the inside waters. The larger ones, however, were caught equally as often and in equally large numbers within the harbor and adjacent waters as off the inlet. This species was not taken in the surf along the outer shores of the “banks”, which apparently is a favorite habitat of the young of the other local species of the genus. It seems evident, furthermore, that the young, like the adults, are chiefly bottom dwelling. Larvae under 10 millimeters in length appeared in the surface tow only 3 times, whereas in about an equal number of hauls on the bottom they were taken 22 times. Individuals over 10 millimeters in length were not taken at the surface with the gear used, namely, 1-meter tow nets, and only a few exceeding a length of 10 millimeters were taken with this apparatus on the bottom. The failure to capture the larger young with meter tow nets on the bottom where they certainly were present, as shown by large catches made with an especially constructed otter trawl, no doubt was caused by their ability either to avoid the meter nets or to escape from them. It is unfortunate that no satisfactory net for catching fish, except small larvae, at the surface has become available. Since the surface of the waters could not be properly sampled for young of about 10 millimeters and upward in length, it cannot be stated definitely that these larger young do not occur there. However, according to the data secured (which may not be entirely reliable, as already shown) the habitat of the young is almost identical with that of the adult; that is, fish of near]}7 all ages inhabit both inside and off-shore waters and apparently are almost wholly bottom dwelling. GROWTH While the measurements secured are not numerous, it is believed, nevertheless, that they show in a general way the growth of the young during the first several months of life. All the measurements made, both of young and adults, are tabulated by months in 5-millimeter groups in table 1. It is evident from this table that the young of the 0-class are readily distinguishable from the older fish until about the end of September. In October and November this class is less clearly distinct. During the winter months, that is, from December to March, an insufficient number of fish was taken to give reliable data. In table 2 the range in size of the 0-class and the arithmetical average of the specimens measure are shown for each month. The upper extreme may not be quite accurate for the fall months. However, the few larger specimens that may have been wrongly assigned to either year class would not affect the average length greatly. It is evident from the tables presented that the range in size of the 0-class is very great toward the end of the summer and in the fall. In September, for example, when spawning probably has just ended, the range in length of the specimens of the 0-class measured extends from 2.5 to 156 millimeters. In November the difference in size of the specimens of this same class (if the data were correctly interpreted) is REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 63 even greater, as the range extends from 15 to 209 millimeters. This large range in size, no doubt, is caused in part by a difference in the rate of growth ; more particularly, however, to the very long spawning season, which apparently extends from April through August. Table 1. — Length frequencies of 1,779 Menticirrhus americanus [Measurements to nearest millimeter; in 5-millimeter groups] Total length April May June July August Sep- tember October No- vember Decem- ber Janu- ary Febru- ary March 0-4 - 1 2 9 5 3 5-9 1 1 6 132 6 10 io-i4 11 32 14 1 15-19 2 2 5 3 1 20-24 - 5 4 3 1 25-29 7 13 15 2 3 30-34 9 35 8 4 1 35-39 9 37 2 3 7 40-44 15 41 4 7 7 45-49 12 22 1 16 5 50-54 18 29 4 15 7 55-59 16 20 8 20 13 60-64 14 26 12 16 11 1 65-69 4 24 22 8 9 70-74 14 16 20 8 9 75-79 10 20 15 10 4 80-84 11 39 10 5 10 85-89 6 7 9 2 3 90-94 - 2 16 2 2 8 i 95-99 - 2 7 9 3 5 100-104 1 7 6 3 2 105-109 4 3 2 1 1 3 110-114 5 3 2 3 115-119 . 1 1 i 5 3 2 120-124 - 3 2 3 2 6 1 125-129 1 2 2 2 9 3 1 130-134 1 2 2 5 2 4 135-139 3 6 1 3 140-144 9 1 10 2 3 145-149 14 1 1 1 5 1 150-154 13 1 2 1 13 7 2 155-159 16 1 2 1 12 3 2 160-164 2 15 2 17 3 l 65-1 69 1 21 2 8 6 i 1 70-1 74 1 13 17 8 175-179 9 2 30 i 2 180-184 1 5 1 1 22 1 185-189 1 5 2 13 2 1 190-194 2 4 2 2 20 i 1 195-199 1 2 i 10 4 200-204 2 8 i 205-209 i i 5 210-214 2 1 1 1 1 215-219 2 2 220-224 1 i 2 1 1 1 1 225-229 1 1 2 i 5 3 230-234 i 1 2 1 2 235-239 1 3 2 2 1 3 240-244 1 1 2 2 245 249 1 1 1 5 250-254 1 1 1 1 1 2 1 1 1 255—259 1 1 1 1 2 1 260 264 1 i 1 4 1 2 265-269 1 1 270-274 1 1 275-279 1 4 2 2 i 1 1 1 1 28.5-289 1 290-294 1 2 1 29.5-299 1 300-304 1 1 305-309 1 310-314 31.5 319 320-324 32.5 329 330-334 33.5 339 1 1 340-344 34.5 349 1 64 BULLETIN OF THE BUREAU OF FISHERIES The average rate of growth as shown by the tables, is quite regular and rapid during the first several months of life. The data are not sufficient to show how it progressed during the winter nor the sec- ond summer. However, since the average length during November of various samples taken over a period of several years, consisting of a total of 327 fish, is 135.01 millimeters ( 5 % inches) with an apparent maximum length of 209 millimeters (8K inches), it seems evident that some of the earliest and fastest growing individuals of this year class reach a marketable size, and probably maturity, during their sec- ond summer. Others may require a year longer. Welsh and Breder (1923, p. 188) produce limited data which they believed tended to show that the fish reached a length of only 140 to 240 millimeters (5 % to 9 ){ inches) at Fernandina, Fla., by the second winter, the modal length being around 170 milli- meters (6 % inches). In New Jersey the rate of growth was thought to be a little slower, as individuals apparently in their second winter averaged only about 160 millimeters (6 % inches) in length. The authors conclude that maturity is reached in about 3 years. We agree that some in- dividuals no doubt require 3 years to reach a marketable size and maturity. However, it seems certain, as already shown, that at Beaufort not an inconsiderable number becomes mature and marketable before an age of 3 years is attained. Table 2. — Monthly summaries of length measurements of 1,519 Menticirrhus americanus during their first year of life ( based on table 1 ) APR. my JURE MY AU6. SEPT. OCT. NOV DEC JAN. FEB. tlAR. Figure 10. — Growth of Menticirrhus americanus during first year or so of life. Solid line, average size; dot and dash line, maximum size; broken line, minimum size. (Graph based on table 2.) Month Fish meas- ured Smallest milli- meter Largest milli- meter Average milli- meter Month Fish meas- ured Smallest milli- meter Largest milli- meter Average milli- meter A 1 2 4 5.5 4.75 October.. . 140 21 194 71. 2 TVTfl v 3 2. 75 6 3.91 November 328 15 209 135. 01 Tpjnp 28 3 17 8.0 December 49 63 1 186 153. 2 326 1.7 102 31.42 January 32 106 i 192 i 133. 28 August. 406 5 156 57. 48 February 6 120 223 183. 33 September — 183 2.5 156 63. 05 March ... ... 16 191 237 216.6 i The apparent decrease in length, of course, shows that the larger individuals of the year class either were not present or not properly represented. MENTICIRRHUS SAXATILIS (Bloch and Schneider) The range of Menticirrhus saxatilis extends from Maine to Florida. Therefore, its range overlaps considerably that of americanus and also to a lesser extent that of littoralis. However, it ranges farther northward and apparently not as far southward as the other local species. In Chesapeake Bay and southward saxatilis is less numer- i REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 65 ous than americanus, but northward (the northern limit of the range of americanus being New York) it is more numerous. This species occurs in the vicinity of Beau- fort, generally in small numbers, virtually throughout the year. It is taken both in the sounds and estuaries and along the outer shores of the “banks.” Its local com- mercial value, however, is rather small. It is reported to reach a maximum weight of 3 pounds, but it probably does not average over a half a pound in the markets of North Carolina. It is caught in the same way as, and often in company with, americanus. SPAWNING Ripe fish were not obtained at Beaufort. It is possible, however, to determine the time and place of spawning in a general way from the collection of a rather small number of young fish. The smallest young of this species obtained were 10 milli- meters long, and were caught on June 20 (1932). On this same date a total of 29 specimens, ranging from 10 to 62 millimeters in length, was caught. Two days later 2 specimens 91 and 93 millimeters long were taken. These specimens no doubt belong to the same year class and are the product of the current year’s spawning. During July, of several years, 14 specimens, ranging more or less gradually from 20 to 116 millimeters in length, besides some larger ones, which quite certainly are older fish, were secured. During August, 5 young, ranging from 70 to 95 millimeters in length, were taken. Fish 91 to 93 millimeters long quite probably are a few months old, while specimens only 10 millimeters long may be only a few weeks old. The specimens collected in June, supported in a measure by those taken in July and August, appear to show that spawning begins not later than April and that it extends through May. Additional data quite probably would show that spawning extends over an even longer period of time. The smaller specimens, 70 millimeters and less in length, were all taken in the surf on the outer shores of the “banks.” The larger ones were taken there in part also and in part in the estuary of Newport River. The presence of the small fish on the outer shores of the banks suggests that spawning may take place near there, as it quite surely does in the other local species of the genus. Spawning takes place about the same time in Chesapeake Bay as at Beaufort, beginning probably a little later, for Hildebrand and Schroeder (1928, p. 291) report the capture of fish 16 millimeters long in Chesapeake Bay late in June. These inves- tigators also report the collection of specimens ranging from 35 to 154 millimeters in length late in September and others in October 50 to 185 millimeters long, all of which they believe to belong to the 0-class. Welsh and Breder (1923, p. 190) state that spawning (presumably at Atlantic City, N.J., where they carried on investigations) commences in June and continues until August, and reaches its maximum in late June or early July. They took ripe fish in July which were stripped, and the eggs were incubated and hatched in the labor- atory. The same authors report spawning early in June at Woods Hole, Mass. The information available, then, indicates that spawning begins a month or two later in New Jersey and Massachusetts than it does at Beaufort, and that it probably ends correspondingly later. DESCRIPTIONS OF THE EGGS AND YOUNG The egg and its development. — The following account of the egg and its develop- ment is based on the description given by Welsh and Breder (1923, p. 190, figs. 46-49). The eggs are spherical, 0.76 to 0.92 millimeter in diameter, and lighter than sea water. 66 BULLETIN OF THE BUREAU OF FISHERIES They are almost colorless, having only a slight yellowish tinge, and they contain from 1 to 18 oil globules. If only one globule is present! it is larger than if several are present. During the development of the egg the oil globules become amalgamated until only one remains at hatching. The incubation period has a duration of 46 to 50 hours in water temperatures of 68° to 70° F. Segmentation and development proceeded as in Bairdiella (Kuntz, 1914). About 18 hours after fertilization grayish chromatophores appeared on the dorso-lateral aspects of the embryo and on the sur- face of the oil globule. Six hours later the chromatophores had become black on the oil globule, and the embryo was dotted with black. Some black chromatophores were scattered also over the yolk near the embryo. Newly hatched young. — The following account of the newly hatched young is based on descriptions and figures by Welsh and Breder (1923, pp. 190-193, figs. 50- 53). The fry are 2.0 to 2.5 millimeters long at hatching, and they float in an inverted position. The head is slightly deflected, the vent is notably in advance of midbody length, and the oil globule lies in the posterior part of the yolk sac. Pigmentation consists of three vertical bands, each consisting of black and dull gold chromato- phores, the first one being above the vent and two posterior to it. A patch of chromatophores of the same colors lies in the dorsal finfold anteriorly, and similar ones are scattered over the yolk sac. None of the larvae lived over 7 days. They lost the gold pigment by the second day and all markings had become less conspicuous. By this time the pectoral fins had become plainly visible. On the fourth day only traces of dark bands remained, and a row of black chromatophores had appeared along the ventral surface posterior to the vent. The blotch in the dorsal finfold remained conspicuous, the eye had become pigmented, the pectoral fins had become dotted with black and gold chromato- phores, and the abdomen had a golden tinge. The yolk sac was almost completely absorbed, and the mouth was open. On the fifth day the larvae, when at rest, floated with the head downward. With the unaided eye the fry appeared dark brown anterior to the vent, and the tail was transparent. On the sixth day the eye had a steel-blue luster, and no trace of rudimentary fins was evident. On the seventh day, shortly before the longest survivors died, a few of the fry had attained a length of 2.8 millimeters. Therefore, very little growth had been gained. The larvae described by Welsh and Breder were preserved, in part, and were compared by us with the smallest fry of americanus taken at Beaufort. Although specimens of the last-mentioned species, quite as small as the larvae described in the preceding paragraphs, were taken at Beaufort, the resemblance is very remote, as stated elsewhere (p. 55), which apparently is caused partly by a difference in age and partly by the differences of the two species. The account of the development of this species must remain incomplete for the present, as the stages between the newly hatched young and specimens 10 millimeters in length have not been secured. Sjjecimens 10 millimeters long. — The body is rather deep, compressed, the greatest depth being contained about 2.8 to 2.9 in length to the base of caudal fin. The head is quite narrow, compressed, and its length is equal to or a little longer than the greatest depth of the body. The interorbital is convex. The eye is longer than the snout, and it has a very small vertically slightly elongate pupil. The mouth is large and moderately oblique, the upper lip anteriorly being nearly on a level with the lower margin of the pupil. The maxillary reaches somewhat past the middle of the eye, and the upper jaw projects beyond the lower one. The fins are all well developed, REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 67 permitting an accurate enumeration of the rays. The longest spines in the first dorsal reach past the origin of the second dorsal when deflexed; the caudal fin is asymmetrically rounded, the longest rays being in the lower half of the fin; and the ventral and pectoral fins are rather long and coterminal, not quite reaching vent. The body almost everywhere is dotted with prominent black chromatophores. An indefinite brownish band is present on the back below the base of the dorsal fins, and another one extends along the lateral ventral edge from the origin of the anal to the base of the lower rays of the caudal. The spinous dorsal is almost wholly black; the second dorsal is colorless, except for an indefinite elongate dark band on the base of about the middle third of the fin; the caudal fin is colorless, with a white base, and sometimes with one or a few large black chromatophores; the anal fin is colorless, except for dark dots on its base; the ventral fins are wholly black; and the pectoral fins are plain translucent (fig. 11). All the specimens at hand were caught among black, partly suspended vegetable debris, which they resemble in color. It seems probable that these fish are darker than they would have been had they been taken in a different environment. This species is characterized at this size chiefly by the deep, strongly compressed body; the narrow, well-compressed head; the small elliptical pupil; the broad, asym- metrically rounded caudal; and by the black spinous dorsal and ventral fins. Figure 11. — Menticnrhus saiatilis. From a specimen 10 millimeters long. (Drawing by Miss Nell Henry.) The specimens described in the foregoing paragraphs are the smallest ones taken. The confusion of this species with smaller specimens of the genus, identified and described as americanus is cpiite improbable, because of the rather pronounced differences among the species in specimens 10 millimeters in length, as shown by the descriptions and illustrations offered. Certainly some of the differences would be evident, also, in somewhat smaller specimens. Specimens 18 to 20 millimeters long. — The body has become somewhat more slender since a length of 10 millimeters (no specimens between 10.5 and 17 millimeters in length having been collected) was attained, and it remains rather strongly com- pressed, the greatest depth being contained in the length to the base of the caudal about 3.25 to 3.4 times. The head has become broader, but it remains much deeper than broad. The mouth is slightly oblique and nearly terminal, the snout scarcely projecting beyond the upper jaw. A slight knob, the beginning of the characteristic barbel of the adult, is evident at the symphysis of the lower jaw. Scalation is nearly complete, and the lateral line is developed anteriorly to about the middle of base of second dorsal. The caudal fin is rather broadly pointed; the longest rays, which are shorter than the head, are in the lower half of the fin. The general color is 68 BULLETIN OF THE BUREAU OF FISHERIES (of specimens taken in black vegetable debris) almost uniform dark brown, with a slight indication of a broad vertical bar, darker than the ground color, on the side under the spinous dorsal, and another one under the middle of the base of the second dorsal. The spinous dorsal and the ventral fins are black; the pectoral hns are colorless; the second dorsal and anal have at least a partly black base, the rest of these hns being colorless; and the caudal is colorless, except for a black blotch on the base (fig. 12). The present species is deeper than either of the other two local species of the genus. Its elliptical pupil and rather bluntly pointed tail, which is shorter than the head, at once distinguish this species from americanus. The dark cross bars on the body, the black spinous dorsal, and the black ventral hns separate it from littoralis. Specimens 80 to 35 millimeters long. — The body has continued to become more elongate and less strongly compressed. The depth remains proportionately a little greater than in the adult, it being contained in the length to the base of caudal 3.5 to 3.65 times. The mouth is nearly horizontal, inferior, and the snout projects moderately beyond it. The mandibular barbel is short, but plainly evident, and scalation is complete. None of the dorsal spines are produced, the longest reaching Figure 12. — Menticirrhus saxatilis. From a specimen 20 millimeters long. (Drawing by Miss Nell Henry.) opposite the base of the hrst or second ray of the second dorsal when dehexed. The caudal hn is slightly angular, and the longest rays, which are in the lower half of the fin, are notably shorter than the head. The pectoral hns scarcely reach the tips of the ventrals, and the latter do not quite reach the vent. The body is quite fully pigmented; the ground color is silvery, brightest on lower parts of body, overlaid almost everywhere with dark brown dots. Dark bars are evident in the majority of the specimens at hand; the hrst one is on the posterior part of the head and runs obliquely downward and backward on the opercle; the second one crosses the nape and is parallel with the hrst one; the third one lies under the spinous dorsal and bends forward slightly, nearly or quite joining the second one on the middle of the side. The two together form a V, which is a recognition mark in the adult. Posterior to the bars described are dark blotches suggestive of bars. The spinous dorsal and the ventrals remain almost wholly black; the second dorsal and anal are black at the base or at least are dotted with black; the caudal hn bears two irregular dark spots on its base and is plain ti^nslucent elsewhere; and the pectoral hns are plain, more or less dotted with black at the base (hg. 13). This species, like littoralis, continues to differ from americanus in the eliptical pupil and the short angulate caudal hn. It differs from both the other local species REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 69 in the deeper body; and from littoralis in the presence of large scales (which are not notably reduced) on the chest, and in the longer pectoral fins. This species, according to the specimens at hand, alone has definite dark bars (on anterior part of the body) at this size. Specimens 50 to 60 millimeters long. — The body has continued to grow less compressed and somewhat more elongate, the depth now being contained in the length 3.8 to 4.1 times, which are the dominating proportions in the adult. The Figure 13. — Menticirrhus saxatilis. From a specimen 30 millimeters long. (Drawing by Miss Nell Henry.) snout is conical, it projects much more strongly than in the smaller fish described in the foregoing section, and is somewhat longer than the eye. Although none of the dorsal spines are notably produced, the longest one reaches well past the origin of the second dorsal. The caudal fin lias a slightly concave margin and the lower lobe remains notably longer and somewhat angulate. The color varies greatly among specimens. Some are dark brown, others silvery gray. However, all have rather definite oblique dark bars on the anterior part of the body and a few blotches poste- Figure 14. — Menticirrhus saxatilis. From a specimen 63 millimeters long. riorly. The spinous dorsal and the ventrals remain almost wholly black in the darker specimens, but are only partly dusky in the lighter ones (fig. 14). The characters of the adult are perhaps less fully developed in this species than in the other local forms at the length described. The third dorsal spine becomes notably produced when the fish reaches a length of about 85 millimeters, although not yet as long as in larger specimens. The caudal fin does not acquire fully the shape it has in full-grown fish until a length of about 120 millimeters is attained. However, 70 BULLETIN OF THE BUREAU OF FISHERIES the fish resemble the adult sufficiently to be recognized readily. The characters that distinguish the species at this size are largely the same as those used in identi- fying the adult; namely, the small scales (91 to 96 vertical series above the lateral line), which are not notably reduced in size on the chest; the rather high spinous dorsal (the third ray later becoming conspicuously produced) with the longest spines reaching well past the origin of the second dorsal; the longer anal, typically with 8 soft rays; the moderately long pectorals, which usually reach to tips of ventrals; and the presence of rather definite black bars on the back and sides, the one crossing the nape and the one under the spinous dorsal nearly or quite meeting on the side to form a V above the pectoral. The small, vertically elongate, eliptical pupil remains conspicuous, as in smaller specimens, and readily distinguishes this species from americanus, although not from littoralis. DISTRIBUTION OF THE YOUNG ■ '. ; ~ Young under 10 millimeters in length were not taken. Those from 10 to 70 millimeters in length were all collected in the surf along the outer shores of the “banks. ” The larger ones were also taken there in part and in part in the estuary of Newport River. So far as known, the young dwell on or near the bottom like the adults. GROWTH The measurements obtained are inadequate to cast much light upon the rate of growth. However, specimens of the 0-class, 113 to 116 millimeters long, taken in July when very probably not over 3 months old, suggest rapid growth. Hildebrand and Schroeder (1928, p. 291) report specimens, 35 to 154 millimeters in length, from Chesapeake Bay taken late in September, and others 50 to 185 milli- meters long taken in October. If the larger fish are correctly assigned to the 0-class, a plienominally rapid growth must take place. Welsh and Breder (1923, p. 192) state, “The growth of Menticirrhus saxatilis the first summer is exceedingly rapid.” These authors record specimens of the 0-class from Woods Hole, Mass., as much as 100 millimeters long on September 1 ; others from Cape May, N.J., 90 millimeters long on August 8; and still others from Chesa- peake Bay, 140 millimeters long on September 12. All the data that are available, therefore, point to a very rapid growth in this species during the first summer. Very little is known about the growth after the first summer. The present investigation, being concerned almost wholly with the development of the young, has yielded virtually nothing on this phase of the life history. Welsh and Breder (1923, p. 194) state that from the examination of the scales of a small series of examples from New Jersey it appeared that a modal length of about 120 millimeters was attained by the first winter, the majority of the fish being 100 to 150 millimeters long. In the second winter the modal length was about 250 millimeters, and the third winter 350 millimeters. Then, the writers conclude that maturity is reached during the third or fourth summer. The earlier spawning season and the very rapid growth during the first summer further south, that is, in Chesapeake Bay and at Beaufort, suggests that maturity may be attained there a year earlier. MENTICIRRHUS LITTORALIS (Holbrook) Menticirrhus littoralis ranges from Chesapeake Bay southward to the Gulf coast, the exact limits of its southern distribution being undetermined. Its range north- ward does not extend as far as that of either of the other two local species of the REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 71 genus. Southward it probably extends further than that of saxatilis and equally as far as americanus. This species as a whole is of much less commercial importance than the others. At Beaufort it is much less numerous than americanus and about equally as common as saxatilis, its commercial value being small. This species evidently seldom enters Beaufort Inlet, as only one record (Bird Shoal, 1906) of its capture in the inside waters is at hand. The writers did not find it in the more or less enclosed waters of the vicinity. The young, ranging upward of about 15 milli- meters, are common during the summer in the surf along the outer shores of the “banks” where they may be taken with small collecting seines in company with young pompano, spot, and occasionally with its congener, saxatilis. SPAWNING Smith (1907, p. 324) says, “At Cape Lookout spawning fish were found by the writer in June, and at Beaufort ripe eggs have been taken on several occasions be tween June 1 and 10.” Unfortunately, the eggs are not described. If the eggs were taken in the tow during the present investigation, they were not recognized, and ripe fish were not seen. Neither did the larvae appear in the tow. The smallest specimens secured are 9 and 10 millimeters long, and were taken on June 20 (1932) with a bobbinet seine in the surf on Bogue Banks, opposite Fort Macon. Fish ranging from 15 to about 120 millimeters long are abundant there throughout the summer. Some of the fish taken in June that certainly belonged to the 0-class had attained a length as great as 75 to 80 millimeters. Examples evidently of the 0-class, caught during July, ranged from 1 1 to 58 millimeters (the larger individuals of the class being missing), those taken during August ranged from 15 to 98 millimeters, and those caught in September from 15 to 120 millimeters. A collection made on June 10 includes specimens 53 to 57 millimeters long, and one made on June 20 contains specimens 75 to 80 millimeters long. It seems reasonable to assume that these rather large, well-developed specimens are 6 to 8 weeks old. If that be true then spawning probably begins not later than the 1st of May. Young only 15 millimeters long were taken as late as September 14. Fish of such a size probably were hatched in August. Although these data are not as complete as desirable, they do indicate rather strongly that spawning begins not later than the 1st of May and that it continues into August. Although the eggs and small fry under 9 millimeters in length were not taken, it seems almost certain that spawning occurs only in the open outside waters. This conclusion is arrived at for the reason that the adults rarely enter the inside waters about Beaufort and, furthermore, the young of 9 millimeters and upward in length were found only in the surf. DESCRIPTIONS OF THE YOUNG Specimens 10 millimeters long. — The body is quite elongate, compressed, the greatest depth being contained about 3.6 to 4.0 in the length to base of caudal fin. The head is rather broad, more or less quadrate in cross section. Its width is equal to about three fourths its depth, and its length is somewhat greater than the depth of the body, being contained about 3.2 in the length to base of caudal. The inter- orbital is rather flat and broad. The eye is longer than the snout, and it has a very small, vertically slightly elongate pupil. The mouth is moderately large and oblique, the upper lip anteriorly is about on a level with the lower margin of the eye, and the 57094—34 3 72 BULLETIN OF THE BUREAU OF FISHERIES upper jaw projects beyond the lower one. The fins are all well developed and an accurate enumeration of the rays is obtainable. The longest spines in the first dorsal reach somewhat beyond the origin of the second dorsal when deflexed; the caudal fin is asymmetrically rounded, the longest rays being in the lower half of the fin ; the ventral and pectoral fins are rather long and co terminal, not quite reaching the vent. The body is densely dotted with dark chromatophores, pigmentation in some specimens being almost general, and dark brown in color as seen with the unaided eye. The spinous dorsal, the ventrals, and the pectorals are colorless. The caudal fin, too, is colorless or at most with one to a few dark dots on the base. The posterior half of the base of the soft dorsal and most of the base of the anal are black, the amount of black present varying among specimens, the rest of these fins being plain colorless (fig. 15). All specimens at hand of the size described were taken among black, partly suspended vegetable debris, which they resemble in color. Somewhat larger speci- mens caught in the same environment, too, are black, while specimens taken on white sand are pale silvery. It seems probable, therefore, that small fish, if caught in a different environment, too, would be paler in color. This species is recognized, when about 10 millimeters long, chiefly by the slender body; the broad head, which is somewhat quadrate in cross section; the small elliptical pupil; the broad asymmetrically rounded caudal; and by the absence of black on the spinous dorsal and ventral fins. The smallest specimen of this species at hand is 9 millimeters long. It does not differ noticeably from the 10-millimeter ones described in the foregoing para- graphs. It is quite unlikely that a confusion of this species has taken place with the smaller specimens of the genus identified and described as americanus. The very evident differences, existing at a length of 10 millimeters, as shown by the descriptions and illustrations offered, surely would be evident also in part in smaller specimens. Specimens 13 to 15 millimeters long. — The body has become rather deeper and heavier posteriorly, although the length and the greatest depth remain proportionately about the same as in 10-millimeter fish. The head is more compressed and less evidently quadrate in cross section. The mouth is only slightly oblique, and the snout projects somewhat beyond the upper jaw. No pronounced changes in the fins have taken place. The caudal fin apparently has become slightly more pointed, but retains an asymmetrical shape. The body in the specimens at hand remains dark 1 4 REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 73 brown. The spinous dorsal, the ventral, and pectorals remain colorless. The amount of black on the base of the soft dorsal and on the anal varies among specimens from a small spot on the base of each fin to a long band or blotch involving a considerable portion of each fin. The caudal fin is colorless, except for a pair of more or less tri- angular-shaped black spots on its base. The specimens described were taken among black vegetable debris and, therefore, may be generally darker in color than others would be if taken in a different environment. In this species, as in americ anus, moderately pronounced headway has been made toward acquiring the adult form, especially in the shape of the head, body, and mouth. The slender body, the elliptical pupil, the asymmetrically rounded caudal, and the absence of black on the spinous dorsal and the ventral fins serve to characterize littoralis, as in the 10-millimeter fish previously described. Specimens 18 to 20 millimeters long. —The body has become only slightly more rounded and elongate since a length of about 15 millimeters was attained, the greatest depth being contained in the length to the base of the caudal about 3.75 to 4.2 times. The head remains low and wide, its depth being only a little greater than its width. The mouth is nearly horizontal and inferior, and the rather conical snout, which is scarcely shorter than the eye, projects slightly. The development of the character- istic barbel of the adult at the symphysis of the lower jaw is indicated by a minute knob. Scalation is nearly complete (although not shown in the illustration), and the lateral line is developed anteriorly, extending to middle of base of soft dorsal. The caudal fin has become rather more broadly asymmetrically rounded, and the longest rays are shorter than the head. Pigmentation is very variable, depending upon the environment in which the specimens were taken. Some specimens are almost uniformly brown, while others are silvery with scattered brownish chromato- phores. The spinous dorsal (except for a few black dots), the pectorals, and the ventrals in all specimens at hand remain plain translucent. The caudal fin is void of color except for a pair of dark blotches on the base, which may be separate or united. The dorsal and anal are largely black at the base in some specimens, whereas they bear only comparatively few dark points in others (fig. 16). Specimens of littoralis of the size described remain more slender than either of the other two local species of the genus. The difference in this respect, however, is slight between littoralis and americanus, but comparatively great between littoralis and saxatilis. The strongly eliptical pupil and the short asymmetrically rounded caudal fin at once separate littoralis from americanus, although not from saxatilis. The 74 BULLETIN OF THE BUREAU OF FISHERIES almost total absence of black on the spinous dorsal and ventral fins, also, aids in dis- tinguishing littoralis from the other local species. Specimens 30 to 35 millimeters long. — The body has become more rounded and has acquired virtually the form of the adult, the depth being contained in the length to the base of the caudal 4.05 to 4.3 times, which are the proportions found also in the adult. The mouth is inferior and horizontal, and the conical snout, which is now a little longer than the eye, projects notably beyond it. The mandibular barbel, although plainly visible, still has the appearance of a knob. Scalation is complete, and the scales on the chest are smaller than those on the sides and on the abdomen. The dorsal spines are short, the longest ones failing to reach the first soft ray of the second dorsal when deflexed. The caudal fin is slightly angular and the longest rays, which are in the lower half of the fin, are notably shorter than the head. The pec- toral fins are short, failing by fully an eye’s diameter to reach the tips of the ventrals, and the latter do not quite reach the vent. The ground color is silvery, overlaid on sides and back with grayish dots. Indications of dark blotches are present on the sides of some specimens, while others appear uniform silvery gray. The dorsal, caudal, and anal fins are more or less dotted wdth black, while the paired fins usually are plainly translucent (fig. 17). The attainment of virtually the adult form at so small a size is quite striking. While the body remains somewdiat more slender than in the other two local species of the genus, the difference is slight between the present species and americanus. How- ever, the eliptical pupil and the short angulate caudal fin have been retained and are useful in separating this species and americanus. The very short pectorals and the reduced scales on the chest distinguish the present species from both americanus and saxatilis. Specimens 50 to 60 millimeters long. — The differences between specimens of this size and those described in the preceding section are not pronounced. The body has continued to become more rounded anteriorly and apparently proportionately not quite as deep, the greatest depth being contained in the length to the base of the caudal 4.3 to 4.4 times. These proportions are, also, common to the adult. The caudal now is shaped virtually as in the full grown fish, that is, its margin is concave and the lower lobe is longer and somewhat more sharply rounded than the upper one. The color is plain silvery gray, as seen with the unaided eye. Under magnification REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 75 dark points are evident on the sides and back, but they have become smaller and less numerous since a length of 35 millimeters was attained (fig. 18). Specimens of the size described are virtually young adults and are readily recog- nized. The body remains slightly more compressed, however, and the caudal fin is less deeply concave and the lower lobe is more sharply rounded. The principal characters that distinguish the adults of this form from the other local species of the genus are developed at this size, and are usable in making identi- fications. They are the rather large scales (72 to 74 vertical series above the lateral line), which are conspicuously reduced in size on the chest; the low dorsal spines, none being produced; the short anal, which typically has only 7 soft rays; the very short pectorals, which fail conspicuously to reach the tips of the ventrals; and the plain silvery gray coloration. The small, vertically elongate pupil still is con- spicuous and readily distinguishes this species from americanus, although not from saxatilis. DISTRIBUTION OF THE YOUNG Young under 9 millimeters in length were not taken. Their habitat and distribu- tion, therefore, remain unknown. Young ranging upward of 9 millimeters were found only in the surf, at Cape Lookout and on the outer shores of the “banks.” So far as known the young, like the adults, dwell on or near the bottom. GROWTH The measurements of 180 young taken during June, July, August, and September are entirely inadequate to cast much light upon the rate of growth. However, speci- mens of the 0-class, 98 millimeters long taken in August, and others 120 millimeters in length caught in September, when probably only about 3 or 4 months old, suggest rapid growth. Nothing is known concerning the rate of growth after the first several months. STAR DRUM (Stellifer lanceolatus (Holbrook)) This small drum has no common name at Beaufort. The two names, “star drum” and “bullhead” occur in the literature. The species ranges from Chesa- peake Bay to Texas and possibly to Mexico, its southern range of distribution not having been definitely determined. The species is recorded from Chesapeake Bay (Hildebrand and Schroeder, 1928, p. 282) from a single specimen. It is recorded 76 BULLETIN OF THE BUREAU OF FISHERIES from North Carolina (Beaufort) by Smith (1907, p. 315), who refers to it as a “rare drum.” However, during recent years it has proven to be rather common off Beau- fort Inlet in water ranging from about 2 to 7 fathoms in depth. Here the star drum is sometimes taken in considerable numbers in trawls. Since specimens have been taken almost throughout the year, it obviously is a permanent resident, although quite evidently more numerous during the summer than during the winter. Only 3 small specimens were secured in the inside waters throughout several years of intensive collecting, indicating that this drum seldom crosses Beaufort Bar. It is reported as very common on the South Atlantic and Gulf coasts from South Carolina southward. The species has no direct commercial value because of its small size, the maximum length attained being only about 165 millimeters (6)^ inches). It no doubt enters into the food of other fishes to some extent, as it occupies areas frequented by such predatory species as the bluefish, weakfish, croaker, king whitings (or “sea mullet”), and flounders. The star drum is the only species of the genus Stellifer recorded from the coast of the United States. The genus is represented by several species in tropical America. The adult star drum is recognized readily by the rather low head, broad interorbital, blunt snout, the slightly concave dorsal outline of the head, and the terminal or slightly inferior mouth. The spongy skull, with bony ridges and large caverns (readily perceptible to the touch), also is quite distinctive. The body is rather robust and elongate, the depth being contained in the standard length 3.0 to 3.3 times, and the length of the head is about equal to the depth. The caudal fin is pointed at all ages. The dorsal has X or XI-I, 21 or 24 rays', and the anal II, 7 or 8. The scales in the lateral series number about 47 to 50. The color is plain bluish gray above and silvery below. SPAWNING The star drum spawns during the summer. Young under 5 millimeters in length were taken at Beaufort in July, August, and September. The largest specimens of the 0-class taken in July (on the 25th day of the month) are 60 to 69 millimeters long. These large young quite probably are more than a month old. Only one ripe fish, a female taken August 4, 1914, was observed. It seems reasonable to assume from these data that spawning begins certainly not later than June, and probably as early as May, and that it extends through August. The eggs have not been studied. The larvae were taken from Beaufort Inlet to Cape Lookout, that is, along Shackleford Banks, some rather near the shore and others several miles off shore, in the general vicinity inhabited more or less permanently by the adults. Besides the off-shore collections only one small lot of larvae was caught in the inside waters where they may have been carried by winds and tides. The indications, therefore, are that the young are hatched on the grounds regularly occupied by the adults and that no migration for the purpose of spawning takes place. DESCRIPTIONS OF THE YOUNG Specimens 1.8 to 2 millimeters long.— The head and trunk are short and moderately deep, compressed, and the tail is long and slender, an abrupt decrease in depth occur- ring at the vent. The distance from the snout to the vent is contained in the length to the end of the notochord about 2.4 times, distance posterior to vent (without finfold) 1.6 times, and the greatest depth about 3.1 times. The mouth is moderately large and oblique, and the snout is very short. The myomeres are too indistinct to enumerate accurately; about 25 can be counted, and only about 5 of them are in REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 77 advance of tlie vent. A small black spot is present on the median line of the abdomen immediately in advance of the vent, generally another elongate black spot is situated on the ventral outline of the tail at about midcaudal length. However, sometimes several small black spots occur on the ventral outline of the tail. The inner surface of the base of the primitive pectoral is black, and dark points are present on the sides along the upper margin of the abdomen. Three dark dots also are situated on the dorsal outline. The first is placed slightly in advance of the anterior margin of the eye, the next one is somewhat behind the posterior margin of the eye, and the third one is fully a half an eye’s diameter behind the second one (fig. 19). Specimens of this size differ notably from somewhat larger ones in the long slender tail, which becomes proportionately much shorter and stouter quite abruptly. The deep body, the moderately large oblique mouth, and the color markings are useful in identifying these small larvae with somewhat larger ones. Specimens 2.5 millimeters long. — The fish has increased greatly in depth since a length of about 1.8 to 2.0 millimeters was attained, and the tail has become propor- tionately much shorter and stockier. The distance from the snout to the vent is contained 1.75 to 1.9 times in the length to end of the notochord, distance posterior to vent (without finfold) 2.0 to 2.3 times, and the greatest depth 2.3 to 2.5 times. The cavernous nature of the skull, characteristic of the adult, already is indicated by an apparently vacant space over the brain. (The outline of the brain is plainly visible because of the transparency of the coverings of the head.) The mouth is moderately large and oblique; the gape anteriorly is about on a level with the middle of the eye and the maxillary reaches under the pupil. Distinct teeth are present on the jaws, and a few spines have appeared on the preopercular margin. The myomeres are partly indistinct; about 8 can be counted in advance of the vent and about 17 behind it. A slight thickening of the vertical finfold is noticeable in those places where the soft dorsal, the caudal, and anal fins are developing. Slight indications of rays are present in the caudal fin (that is, below the end of the tail). The color markings of the smaller specimens described remain, though somewhat modified. A few dark points (more distinct in some specimens than in others) persist on the median line of the abdomen. The dark spot at about midlength of the ventral outline of the tail has become more distinct and quite elongate, and some specimens have a smaller spot in front of this one and another one behind it. The inner surface of the base of the primitive pectoral remains black. A few dark points persist along the upper margin of the abdomen, ending in a small dark spot dorsally of the base of the pectoral. These dark markings, in part at least, consist of dark membranes, visible through the body wall. The three dark markings on the head, noticed in smaller specimens, now have the appearance of dark cross partitions, lying below the surface of the skull (fig. 20). 78 BULLETIN OF THE BUREAU OF FISHERIES The indications are that the fish when about 2.0 millimeters long (preserved specimens) virtually ceases to increase in length for a while. During the time it fails to grow in length it appears to increase greatly in robustness, and the caudal portion of the body becomes proportionately much shorter, as shown by the propor- tional measurements given in the foregoing description and by figures 19 and 20. Specimens 2.5 millimeters long have the appearance of being considerably older than those 2.0 millimeters long. Specimens 3.0 to 3.5 millimeters long. — The fish has increased further in depth, the tail particularly having become much deeper, and the body, although rather robust, is compressed throughout. The decrease in depth immediately behind the vent, abrupt in younger fish, has become much less pronounced. The distance from the snout to the vent is contained 1.55 to 1.65 times in the length to the end of the noto- chord, distance posterior to vent (without finfold) 2.0 to 2.2 times, and the greatest depth 2.2 to 2.5 times. The notochord remains nearly straight in a specimen 3.0 millimeters long, but is bent upward quite sharply distally in a fish 3.3 millimeters long. A large vacant space exists between the brain and the cranium, and a few rather prominent spines are present on the preopercular margin. The mouth remains moderately oblique, the gape anteriorly being only a little higher than the lower Figure 20. — Stellifer lanceolatus. From a specimen 2.5 millimeters long. margin of the eye, and the maxillary reaches nearly opposite the middle of the eye. The caudal and pectoral fins contain well developed rays; the soft dorsal and anal bases are rather well outlined, but the rays are not fully differentiated; the spinous dorsal is not evident; and the ventral fins if present appear as mere tufts of membrane. Color markings on the chest and abdomen have increased in number and intensity. A short dark cross line, preceded by a median black spot, usually is present on the chest; a few indistinct dark points appear at the base of the ventrals; and a rather distinct short dark bar (sometimes consisting of a few dark points only) crosses the abdomen behind the ventrals. The elongate dark spot at about midlength on the ventral outline of the tail, mentioned in descriptions of smaller larvae, now is situated at or near the end of the base of the anal fin, and occasionally a black point is present directly over it on the dorsal outline. In some specimens, at least, a small dark spot has appeared also at the origin of the anal. The inner surface of the base of the pectorals remains black, and a dark shoulder spot is present as in the younger fish (fig. 21). The most prominent advancement in the development since a length of about 2.5 millimeters was attained consists in the further deepening of the body, especially the caudal portion, causing a less pronounced break in the ventral outline at the vent. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 79 Specimens J^.5 to 5.5 millimeters long. — The fish has become much more shapely than it was at a length of about 3.5 millimeters. It is quite deep and strongly com- pressed throughout, and the break in the ventral outline at the vent has almost dis- appeared, a space exceeding in length the diameter of the eye between the vent and origin of the anal being occupied by a transparent membrane. The distance from the snout to the vent is contained 2.0 to 2.3 times in the length to the base of the caudal, distance from vent to base of caudal 1.9 to 2.1 times, and the greatest depth 2.6 to 2.8 times. The cavernous nature of the skull is quite evident from the large vacant spaces occurring above the brain and around the snout. The eye and snout are of Figure 21 .—Slellifcr lanceolatus. From a specimen 3.3 millimeters long. about equal length. The preopercular margin has several prominent spines at its lower angle. The mouth is oblique, the gape anteriorly is about on a level with the lower margin of the pupil, and the maxillary reaches below the middle of the eye. The fins, exclusive of the spinous dorsal and the ventrals which remain rudimentary, are quite well developed. About 22 rays can be counted in the soft dorsal and 10 in the anal (the spines and the soft rays not being differentiated in the latter). The caudal fin is moderately long and somewhat pointed. A few dark points remain on the median line of the chest and abdomen, but are less evident than in younger fish. The elongate black spot at the end of the base of the anal persists and is variable in size and intensity among specimens. A slight dark spot remains evident at the origin of the anal in some individuals, though missing in others. The black on the inner surface of the base of the pectoral, prominent in the younger stages described, has disappeared entirely according to the specimens at hand. A slight shoulder spot and a few dark markings behind it, which extend to the vent, still persist. It is evident that the last-mentioned color markings are below the surface of the body, and apparently consist of a dark membrane on the dorsal wall of the abdominal cavity. A faint dark bar has appeared at the base of the caudal fin (fig. 22). 80 BULLETIN OF THE BUREAU OF FISHERIES Specimens 7 to 8 millimeters long. — No pronounced changes in the proportions of the body have taken place since a length of about 4.5 to 5.5 millimeters was attained. The distance from the snout to the vent is contained 2.0 to 2.2 times in the length to the base of caudal; distance from vent to base of caudal 1.8 to 2.1 times; and the greatest depth 2.65 to 2.8 times. A slight concavity remains in the ventral outline of the body between the vent and origin of anal, which is occupied by a thin mem- brane, apparently a remanent of the finfold. Above this concavity in the ventral outline, the body is very thin and semitransparent. The snout remains short and narrow, its length being equal to the diameter of eye, and the profile ascends abruptly from its tip. Considerable advancement has been made in the development of the fins. The spinous dorsal is fairly well differentiated, although rather more closely connected with the soft dorsal than in the adult. About 12 spines and 24 soft rays may be counted. The anal fin consists definitely of 2 spines and 7 or 8 soft rays. The caudal fin is frayed in the specimens at hand, but it evidently is long and pointed. The ventral fins remain small. The dark markings present vary in size and number among specimens. Generally a few dark dots are present on the median line of the chest and abdomen. The spot at the end of base of anal, present in all the smaller larvae at hand, persists. Some specimens now have two spots at and behind the anal base, and each one may have a narrow vertical projection. Some specimens also have a small black spot near the origin of the anal. A rather prominent dark spot is present above the vent, which often is more or less connected with other spots, reaching to an elongate, finely branched shoulder spot. A narrow dark bar at the base of the caudal is evident (fig. 23). Specimens 10 to 13 millimeters long. — The proportions of the body remain about as in somewhat smaller specimens. The distance from the snout to the vent is contained 1.85 to 2.0 times in the length to the base of the caudal fin, distance from vent to base of caudal 2.05 to 2.15 times, and the greatest depth 2.6 to 2.75 times. A slight concavity remains evident in the ventral outline between the vent and the origin of the anal (which is occupied by a thin membrane as in smaller fish), and the distance between these two points exceeds the diameter of the eye. The snout has become broader and rounder; the mouth is less oblique and wholly below the lower margin of the eye ; and the maxillary reaches nearly or quite to the posterior margin of the eye. In all the characters about the head, just mentioned, the fish is approach- ing the shape and form of the adult. The eye is unusually small. The cavernous nature of the skull, with its bony ridges, is quite prominent. The preopercular spines are smaller and more numerous than in younger fish. The fins are all well developed, the dorsal formula is X or XI-I, 23 or 24, and that of the anal II, 7 or 8. The caudal remains long and pointed, the longest rays being longer than the head. Little REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 81 advancement in pigmentation has taken place. The few dark markings usually present are shown in figure 24. Some specimens, however, have a series of 3 to 7 (instead of 1) dark spots on the side below the spinous dorsal, as well as several on the posterior part of the head and on the nape (fig. 24). Specimens 18 to 20 millimeters long. — The fish has made considerable headway toward acquiring the form of the adult. Although the head has increased in width, it remains proportionately much more compressed than in the adult, its length is contained 2.55 to 2.75 in the total length without the caudal fin. The interorbital space has become relatively broad and is about 1.5 times as wide as the small eye. The snout is slightly longer than the eye and is contained 4.6 to 5.5 times in the head. The preopercular margin has 3 or 4 small spines, and 3 bony stays are partly embedded in the interopercle. A row of bony serrae are present at the shoulder. A slight concavity remains present between the vent and the anal fin, the membrane (rem- anent of the finfold) which occupied the concavity in smaller fish has disappeared, and the distance between the vent and origin of anal is now scarcely as long as the Figure U.—SleUifer lanceolalus. From a specimen 13 millimeters long. eye. The proportionate depth of the body remains the same as in fish 10 to 13 millimeters long. Much advancement in pigmentation has taken place. The dark spot near the end of the base of the anal, present in all younger fish, remains evident and is preceded (along the base of the anal) as well as followed (on the ventral outline of the peduncle) by a row of dark dots and spots. The characteristic elongate, slightly arched spot on the opercle, immediately in advance of the upper angle of the gill opening, remains prominent. The mouth is margined with black and numer- ous dusky markings occur on the head and nape, and on the upper parts of the sides. The markings are variable in size and intensity among specimens. One row of spots follows the base of the dorsal and another one parallels it about an eye’s diameter lower on the side. Other scattered markings are present, and a dark bar is situated on the base of the caudal. Specimens 25 to 30 millimeters long. — Although advancement toward the adult form has been rather rapid since a length of 18 to 20 millimeters was attained, the fish remains more compressed anteriorly and the snout is more pointed. The head is contained 2.8 to 2.9 times in the length to the base of the caudal, and the depth 2.7 to 3.1 times. These proportions are close to those of adult fish. The small eye is surrounded by a ridge which is not yet fully ossified, its longest diameter is contained 4.5 to 5.2 in the head; snout 4.25 to 4.5 times; interorbital 2.7 to 3.2 times; maxillary 1.8 to 1.95 times. The concavity between the vent and the anal, present in younger fish has disappeared, and the body in this region has thickened and no longer is semitransparent. Outlines of scales are present on most of the body. If scales 82 BULLETIN OF THE BUREAU OF FISHERIES actually were developed they have been lost. Pigmentation has advanced rather rapidly. The chief color markings are shown in the accompanying illustration (fig. 25). Specimens 40 to 50 millimeters long. — The head is lower and notably broader than in fish about 30 millimeters long. Although the snout is more rounded than in youngei fish, it still remains narrower and rather more pointed than in the adult. The small eye, though lateral, is near the dorsal outline; the interorbital is broad, its width being contained 2.4 to 2.8 times in the head. Several bony ridges are present on the head among which caverns occur. The preopercle has a few small spines at the angle and the interopercular margin is strongly serrate, the spines at the angle being largest. The scapular spines have become much smaller. The caudal fin apparently remains rather longer and more pointed than in the adult, the longest rays being more than an eye’s diameter longer than the head. Pigmentation is general. The Figure 25. — Stellifer lanceolatus. From a specimen 29 millimeters long. (After Welsh and Breder.) back of preserved specimens is fight brown and the lower parts are silvery. The elongate dark spot on the upper part of the opercle and the dusky color around the mouth remain as in smaller specimens. Dark blotches, present in somewhat younger fish, along the back generally persist, but are less distinct, the color being more generally distributed as dusky points. The spinous dorsal is dusky, and all the other fins have dark dots, being very few, however, on the ventrals and pectorals. A dusky bar remains present on the base of the caudal fin. Specimens 75 to 85 millimeters long. — Fish of this size are young adults, having the body proportions of full-grown fish. The head is somewhat depressed and its dorsal outline is slightly concave over and behind the posterior margin of the eye, as in the adult. The snout is broad, bluntly conical, and almost as broad as deep at anterior margin of the eye. The mouth is oblique, but wholly below the level of the lower margin of eye; the maxillary reaches about under posterior margin of the pupil. The small eye is shorter than the snout, and not as wide as the interorbital, its diameter being contained 4.4 to 4.8 times in the head. The preopercular spines are small, and those on the interopercle are strong, as in full grown fish. The scapular spines, described in younger fish, are now represented as serrations on an enlarged scale. The caudal fin remains rather pointed and is about as long as the head. The pectoral fins are long, being only a little shorter than the head. The color remains about the same, as in 40 to 50 millimeter fish. The dark points on the fins have become more numerous, however, and the anal and the tips of ventrals are quite dusky. Specimens of this size have virtually all the characters of the adult and are easily recognized. Figure 26 illustrates the adult without indicating the scales. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 83 DISTRIBUTION OF THE YOUNG It has been shown (p. 76) that the young of the star drum in the vicinity of Beaufort apparently are hatched and live on the grounds occupied by the adults as a permanent residence, that is, the larvae were taken from Beaufort Inlet to Cape Lookout, as far as 7 to 12 miles off shore, and only one small lot was caught in the inside waters. Furthermore, all the young collected were taken on the bottom, showing that the young, like the adults, dwell on the bottoms. The star drum, therefore, appears to dwell at or near the bottom throughout life. GROWTH The number of young taken is insufficient to determine accurately the rate of growth even during the first several months of life. During July 122 young of the 0-class were caught, ranging in length from 2.5 to 69 millimeters, having an average length of 21.3 millimeters. Although no young were taken earlier in the season, it is believed that the larger specimens, as already explained (p. 76), are more than a month old and probably were hatched in May or June. In August 617 young were caught, ranging in length from 2.3 to 90 millimeters, with an average length of 49.3 Figure 26 .—StellifeT lanceolatus. From a specimen 164 millimeters long. Scales not indicated. (After Welsh and Breder.) millimeters. Only 11 larvae, all less than 10 millimeters in length, were taken in September. During October 62 young, ranging in length from 38 to 112 millimeters, averaging 90.4 millimeters, were caught. The smaller young of the season obviously are missing in this catch, as the smallest specimen taken next month is 26 millimeters long. The largest one in the 0-class caught in November is only 91 millimeters, and the average length of 34 specimens taken is only 58 millimeters. It is obvious, therefore, that although the smaller ones are present the larger ones of the 0-class are missing in the catch for November. The number of specimens taken during the other months is too small to be worthy of consideration. From the data presented it is evident that some of the young reach a length as great as 100 to 125 millimeters (4 to 5 inches) during their first summer. Since the maximum length attained by the star drum is only about 6 inches, it seems probable that the fastest growing individ- uals reach maturity when about 1 year old. Others may require a year longer to reach that stage. The foregoing analysis of the rate of growth is in general agreement with the data presented by Welsh and Breder (1923, p. 175), who found individuals of the 0-class varying from 10 to 40 millimeters in length (the smallest larvae of the season evi- 84 BULLETIN OF THE BUREAU OF FISHERIES dently being missing) in a collection of fish made in Winyah Bay, S.C., in July (1915). From a collection from Fernandina, Fla., these writers determined that the fish reach a length of 50 to 90 millimeters by the first winter. These writers state that the 1-year-old fish are 70 to 100 millimeters long in July and that the same year class reaches a length of 80 to 140 millimeters by the second winter, and they conclude that maturity is reached at the age of 1 year. BANDED DRUM (Larimus fasciatus Holbrook) The banded drum, also known as “bastard drum”, “bastard perch”, and “chub ”, ranges from Cape Cod to Texas. According to published records, it occurs only as a straggler from Virginia northward, and is common from Cape Hatteras southward. The limit of its southward range apparently has not been determined definitely. The species is common off Beaufort Inlet, where it is a year-round resident, and where it sometimes is taken in large numbers in shrimp trawls in water ranging from a few to several fathoms in depth. According to the records at the Fisheries laboratory at Beaufort, the species does not enter the harbor nor adjacent sounds. Neither has it appeared in collections made with seines along the outer shores of the “banks.” The banded drum was first recorded from North Carolina by Smith (1907, p. 314), who reported it from “several specimens” with the notation, “It is not common anywhere.” However, Welsh and Breder (1923, p. 170) state “* * * although stragglers have been taken as far north as Woods Hole, it is not found in abundance north of Cape Hatteras. South of this point and on the shores of the Gulf of Mexico it is one of the most abundant fishes, being taken in large numbers in the trawls of the shrimp fishermen.” The species evidently did not appear often in the collections at Beaufort until the otter trawl came into use in about 1913 or 1914. We have no record of its capture with seines, probably because it does not frequent areas near Beaufort in which such nets commonly are operated. The banded drum does not grow large enough to be of much direct commercial value. Comparatively few individuals attain a length as great as 8 inches. The largest one seen at Beaufort was 206 millimeters (8% inches) long. The species probably enters into the food of other fishes to some extent, as it occupies areas fre- quented by such predatory species as the bluefish, weakfish, croaker, king whiting (“sea mullet”), and flounder. The banded drum is the only species of the genus occurring north of Florida. The adult is readily recognized by the deep compressed body (depth in standard length 2.6 to 2.8), the strongly oblique mouth, long dorsal fin (X-I, 24 to 27), short anal fin (II, 6 to 8), and by the 7 to 9 black crossbars. The skull is firm, without conspicuous caverns, and no barbels are present about the mouth. SPAWNING The banded drum evidently spawns throughout the summer. Young under 5 millimeters in length were taken in July, August, September, and October, and one specimen 12 millimeters long was taken on May 29 (1930). This specimen, according to measurements made of collections, belongs to a new year class. Furthermore, the largest young of the 0-class taken in July, as shown elsewhere (p. 91), were as much as 70 millimeters long and the average length of 87 specimens measured was 34 milli- meters. Unless the rate of growth in this species were unusually rapid, the larger individuals certainly would be as much as 2 months old. It seems quite probable, REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 85 therefore, that spawning begins as early as May, even though no larvae were taken in June, and it extends through October, as some of the smallest larvae (1.9, 2.1, and 3.0 millimeters long) in the collection were taken on October 22 (1928). The eggs, if taken, were not recognized. The larvae were caught in the tow in the same general vicinity where the adults are found throughout the year, that is, from Beaufort Inlet to Cape Lookout and as much as 12 miles offshore, beyond which collecting was not extended. It seems improbable, therefore, that a migration for the purpose of spawning takes place. DESCRIPTIONS OF THE YOUNG Specimens 1.9 to 2.0 millimeters long. — The body is moderately deep and robust for such a small fish, and the decrease in depth at the vent is not abrupt. The caudal portion is rather shorter than the rest of the body, the distance from vent to tip of notochord being contained 2.25 times in length without caudal fin membrane, and the greatest depth of body 4.15 times. Myomeres are partly indistinct; about 25 may be counted. The mouth is moderately large and oblique, the gape anteriorly being at or slightly above middle of eye, and the maxillary reaches about to the pupil. The eye is moderately small for such a young fish and does not exceed the snout in length. The color markings of preserved specimens consist principally of a dark spot at the vent, a larger and more distinct one at midcaudal length, and a third and smaller one about half-way between the last-mentioned spot and the tip of the noto- chord. The smaller larvae of this species differ from related forms quite notably in having the caudal portion of the body rather deep anteriorly, thereby eliminating an abrupt decrease in the depth at the vent. Specimens 2.3 to 2.6 millimeters long. — The head and trunk are deep, compressed, and the decrease in depth just posterior to the vent is less abrupt than in related species. The caudal portion of the body remains somewhat shorter than the rest of the fish, being contained in the total length without the caudal fin membrane 2.1 to 2.6 times, and the greatest depth is contained in the length 2.8 to 3.2 times. The mouth is moderately large, oblique, with the lower jaw projecting in advance of the upper one, the gape anteriorly being on or above the level of lower margin of pupil, and the maxillary reaches nearly or quite under anterior margin of pupil. Slight indications of rays are discernible in the vertical fin membranes, as well as in the pectorals. The color of preserved specimens is brownish. The air bladder may be seen as a small clear area, with a dark margin, situated above the base of the pectorals. The ventral outline of the abdomen is slightly dusky, with a small black spot just behind the gill membranes and another one usually present just behind the vent. Two or three black spots occur on the ventral outline of the tail. The first one, if present, is a short distance behind the vent; the second and more distinct one (sometimes quite elongate) is near midcaudal length, and the third one is about midway between the last-mentioned spot and tip of notochord. The inner surface of the base of the pectoral is black (fig. 27). The head and trunk become abruptly deeper while the fish grows in length from about 1.9 to 2.5 millimeters. Specimens 3 to 3.5 millimeters long. — The fish has become proportionately deeper and more robust, the greatest depth now being contained in the total length without the caudal fin membrane about 2.3 times. Although the tail remains much 86 BULLETIN OP THE BUREAU OF FISHERIES more slender than the rest of the body, it has become proportionately deeper and rather shorter, its length being contained in the length of the body about 2.3 times. The mouth remains strongly oblique and the maxillary reaches about under the middle of the orbit. The eye is small, scarcely longer than the snout. The cranium is very transparent, permitting the lobes of the brain to be seen clearly. The fin membranes remain about as in the younger fish already described. A slight thick- ening of the membranes is discernible, however, in the places that will be occupied by the bases of the dorsal and anal fins. The color markings have not changed per- ceptibly, the most prominent ones being the dark spot on about midlength of the ventral outline of the tail, and the black inner surface of the base of the pectorals (fig. 28). The dark color markings described, together with the very deep compressed body, serve as the principal diagnostic characters in the smaller larvae of this species. Specimen ^.5 millimeters long. — The body is very deep and moderately compressed anteriorly, tapering rapidly behind the head. The break in the ventral outline behind the vent, very abrupt in the smaller specimens, is no longer pronounced. The caudal portion is notably shorter than the rest of the body, its length being contained about 2.6 times in the total length without the caudal fin, and the greatest depth is con- tained about 2.15 times in the length. The mouth remains strongly oblique, and the maxillary reaches under the middle of the orbit. The eye is comparatively small, being only a little longer than the snout, and is situated somewhat nearer to the dorsal than the ventral outline. The fins are fairly well developed, all of them, exclusive of the ventrals, having fairly well-developed rays. About 36 rays may be counted in the dorsal, and 7 in the anal. A structural differentation is not evident 87 REPRODUCTION AND DEVELOPMENT OF SCI AENID AE between spines and soft rays. However, the rays in the anterior part of the dorsal, which will form spines, are shorter than the others. The color in spirits remains dark brown. The anterior part of the abdomen is dusky and the dark spot on the ventral outline of the tail, described in smaller specimens, is situated immediately behind the base of the anal. The inner surface of the base of the pectorals remains black, as in the smaller larvae (fig. 29). The very deep compressed body, moderately large oblique mouth, the black spot behind the base of the anal, and the black inner surface of the base of the pectorals are the principal recognition marks. Specimen 7 millimeters long. — The head and trunk remain deep and compressed, as in smaller specimens. The tail has become deeper, causing the fish to appear more shapely. It has decreased still further in proportionate length, the distance from the vent to the base of the caudal fin being contained in the total length without the caudal fin 3.8 times, and the greatest depth of the body is contained 2.35 times in the length. The mouth remains strongly oblique, the gape anteriorly being about on a level with the middle of the eye, and the maxillary reaches slightly beyond the middle of the orbit. The fins remain about as in the 4.5-millimeter fish already described, except that some of the anterior rays of the dorsal and the first one of the anal now resemble spines somewhat. The caudal fin is quite fully developed, but the exact shape cannot be determined as it is frayed at the tip in the single specimen of this size at hand. However, it very probably is pointed, as indicated in the accompanying illustration. The pectoral fins are long and reach the origin of the anal. Many black chromatophores are now present, as shown in figure 30. The characteristic black spot just posterior to the base of the anal, present in all the smaller specimens examined, persists. The inner surface of the base of the pectoral remains black, and the base of the ventral is dusky (fig. 30). Specimens 9 to 10.5 millimeters long. — The body has become somewhat deeper, its greatest depth being contained about 1.9 times in the length. The dorsal outline is much more strongly curved than the ventral one. It ascends sharply anteriorly and descends strongly posterior to the origin of the dorsal. The mouth has become less strongly oblique, as the gape anteriorly is only a little above the lower margin of the eye. The spines and soft rays in the vertical fins are fully differentiated. The first dorsal consists of 10 spines, the second one (which is scarcely separated from the first one) has 1 spine and 26 or 27 soft rays, and the anal has 2 spines and 6 soft rays. The caudal fin is nearly as long as the head and is pointed. The ventral fins reach 57094—34 4 88 BULLETIN OF THE BUREAU OF FISHERIES almost to the origin of the anal, and the long pectorals reach to or past the middle of the anal base. Black pigment spots have increased greatly in number, as shown in figure 31. However, the spots vary in size and number among individuals, some examples having more and larger spots than the one illustrated. The black spot immediately behind the anal, present in the smallest specimens taken, persists, though it has become much more elongate and branched. The basal half of both the ventrals and pectorals is black (fig. 31). The deep compressed body, oblique mouth, the number of rays in the dorsal and anal fins, and the prominent black spots on the body distinguish specimens of this size from other local forms. Specimens 15 to 17 millimeters long. — The body remains deep and compressed, although proportionately more slender than at a length of about 10 millimeters. The greatest depth, which falls under the origin of the dorsal, is contained about 2.4 times in the length. The vent is well posterior to midbody length, the tail (without the fin) being contained in the length of the body about 3.15 times. The head is large and compressed. The preopercular margin has small serrae and three long spines at the angle. The eye is small, being scarcely longer than the short snout. The gape anteriorly is about on a level with the lower margin of the eye, and the maxillary reaches nearly opposite the posterior margin. Scales are present on most of the body, although not shown in figure 32. The dorsal and anal fins are rather high, the longest rays of the former reaching past the base of the caudal when deflexed, and those of REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 89 the latter about opposite it. The caudal fin is long and pointed (not round as shown by Welsh and Breder, 1923, p. 171, fig. 22), being longer than the head. The ventral fins, which have a strong spine, reach to or a little beyond the vent, and the pectorals reach opposite the anal base. Black pigment spots have increased in number as shown in figure 32. A concentration of spots, suggesting a broad black band, has taken place on the side below the base of the spinous dorsal. The black spots extend on the spinous dorsal, and the ventrals and pectorals are mostly black. A black spot at the end of the anal base, which serves as a recognition mark in the younger stages, is not evident in some specimens, though it persists in others (fig. 32). Specimens 20 to 25 millimeters long. — The body is slightly more elongate than in the smaller fish described in the foregoing section, and has acquired proportions near those prevailing among adult fish, the depth being contained in the standard length 2.5 times. The head remains rather longer than in adults and is contained 2.7 times in the length. The eye is small, lateral, near the dorsal profile, and is a little longer than the very short snout, being contained in the head 3.2 to 3.6 times. The mouth virtually has the position it occupies in the adult; the gape anteriorly is a little above the level of the lower margin of the eye ; and the maxillary reaches nearly to a vertical from the posterior margin of the eye, its length being contained 2.0 to 2.2 times in the head. The margin of the preopercle remains strongly serrate, but the long spines at the angle, present in somewhat smaller fish, have disappeared. The body is now fully covered with large scales. The fins remain about as in the smaller fish described in the preceding section. The caudal fin is notably longer than the head, and remains strongly pointed. The ventral and pectoral fins are long, the former reach a little beyond the vent and the latter opposite the anterior part of the anal base. Pigmenta- tion has increased greatly, as shown in figure 33, although it is not yet general and complete. Indications of three dark crossbars are present. The anterior one, situated under the spinous dorsal, is broad and moderately distinct. The second one is situated under about the end of the anterior third of the soft dorsal, and the third one is under the posterior third of that fin. The posterior bars are narrower and much less distinct than the anterior one. A concentration of dark chromatophores occurs near the base of the caudal, forming a dark blotch. The ventral fins are almost wholly black, and the inner surface of the basal two-thirds of the pectorals also is black (fig. 33). Fish about 20 to 25 millimeters long resemble the adults sufficiently to make identification comparatively easy. The chief diagnostic characters are the rather 90 BULLETIN OF THE BUREAU OF FISHERIES short compressed body, the blunt head, short snout, strongly oblique mouth, the fin formulae (D. X-I, 24 to 27; A. II, 6 or 7), and the dark crossbars. Although only 3 bars are developed, whereas adults have 7 to 9, they are an aid in distinguishing the species. Specimens 35 to J+0 millimeters long. — The most evident advancement in the devel- opment since a length of about 25 millimeters was attained is in color. Pigmentation is general, and about seven dark crossbars are present. The broad one under the base of the spinous dorsal, present in smaller fish, shows indications of dividing into two narrower bars. The caudal fin remains nearly as long and pointed as in fish 25 milli- meters long. The second anal spine has become very strong, as in the adult, and is only a little shorter than the snout and eye. The pectoral fins have become propor- tionately shorter and extend only to the tips of the ventrals, which scarcely reach the origin of the anal. The serrations on the preopercular margin have decreased in size, but remain somewhat more prominent than in the adult. Specimens 60 to 65 millimeters long. — The proportionate depth (2.6 to 2.8 in the length) is about the same as in the fully matured fish, but the body remains more Figure 33. — Larimus fascialus. From a specimen 24 millimeters long. compressed. Specimens of this size differ from the adult conspicuously in the long pointed tail, the longest rays of which exceed the length of the head by at least an eye’s diameter. The color pattern of the adult is quite fully developed. The ground color is silvery, the back in preserved specimens being slightly brownish. The sides have from about 7 to 9 dark bands. The spinous dorsal, the anal, ventrals, and pectorals are largely dusky to black, and the other fins are plain or somewhat punctulate with dusky. The caudal fin generally does not acquire the characteristic shape it has in the adult (the upper lobe slightly concave and the lower one rounded) until the fish reaches a length of about 100 millimeters. In the adult, as in the young, the middle rays are the longest, but they do not exceed the head in length. When the fish reaches a length of about 100 millimeters the body has increased considerably in robustness, and this increase continues until the fish is full grown. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 91 DISTRIBUTION OF THE YOUNG It has been shown elsewhere (p. 85) that the young were taken in the general vicinity where the adults are permanent residents. The larvae, like the adults, apparently dwell at or near the bottom, for no young were taken in towings at the surface, whereas they appeared in bottom hauls 9 times (16 specimens, 1.9 to 15 millimeters long), notwithstanding that many more surface than bottom hauls were made. The eggs, although not taken, probably are pelagic, as in other sciaenids so far as known. If that be true the larvae must descend to the bottom very early, as already indicated. The banded drum, therefore, appears to be bottom dwelling almost throughout life. GROWTH The number of young fish measured is too small to cast much light upon the rate of growth. The range in length of 87 specimens of the 0-class, taken in July is from 3 to 70 millimeters, the average length being 34 millimeters. The range of 54 speci- mens taken in August is from 3 to 77 millimeters, with an average length of 54 milli- meters. The only other month for which a considerable number of measurements is available is November. However, the larger young of the season obviously are missing, as the 215 fish measured range in length from 15 to only 63 millimeters, with an average length of 44.5 millimeters. The largest specimen, which evidently belongs to the 0-class, taken in October is 91 millimeters long, the largest one in Februar}^ (no large ones of the 0-class having been taken in November, December, and January) is 113 millimeters, and the largest one in March is 123 millimeters long. While the data are quite incomplete, they do suggest a fairly rapid rate of growth, indicating that the largest young may reach a length of 120 to 125 milli- meters (4% to 5 inches) when 1 year of age, although the average must be much smaller. THE WEAKFISHES OR SEA TROUTS (Cynoscion nebulosus, C. regalis, and C. nothus) Three species of sea trouts, namely Cynoscion nebulosus, C. regalis, and C. nothus occur on the coast of North Carolina. The two species named first are very important food fishes, as shown in the subsequent pages, whereas the last-mentioned one is of little value. The adults may be distinguished as shown by the following key. KEY TO THE SPECIES a. Soft dorsal and anal scaleless; gill rakers short, 8 on lower limb of first arch; body with round black spots nebulosus aa. Soft dorsal and anal covered with small scales; body without round black spots. b. Anal fin usually with 12 soft rays, sometimes 11, infrequently 13; lower limb of first arch most frequently with 17 gill rakers, often with 16 or 18, infrequently with 15 or 19; caudal fin definitely emarginate in specimens 300 millimeters (12 inches) and more in length; upper parts of sides usually with dark greenish wavy oblique stripes or reticulations. regalis bb. Anal fin usually with 9 soft rays, sometimes with 8 or 10; lower limb of first arch usually with 13 gill rakers, sometimes 12 or 14; caudal fin never definitely emarginate, the upper lobe in large specimens somewhat concave and the lower one rounded, pointed in fish under 200 millimeters (8 inches) in length; coloration generally plainer, seldom with dark wavy oblique lines or reticulations nothus The distinguishing characters of young C. nebulosus and C. regalis are pointed out in a separate paragraph following the descriptions of the different stages in the 92 BULLETIN OF THE BUREAU OF FISHERIES development of C. nebulosus. Similarly, the distinguishing characters of C. regalis and C. nothus are shown in a paragraph following the descriptions of the various stages in the development of C. nothus. CYNOSCION NEBULOSUS (Cuvier and Valenciennes) SPOTTED TROUT; SPECKLED TROUT; SPOTTED WEAK FISH The spotted trout ranges from New York to Texas and is commercially an impor- tant fish from Virginia southward to Texas. According to the statistical records 7 of this Bureau, the following catches were made by States: Virginia, 198,000 pounds; North Carolina, 694,309; South Carolina, 10,900; Georgia, 48,450; Florida, 2,790,566; Alabama, 105,981 ; Mississippi, 125,112; Louisiana, 387,101; and Texas, 1,043,353. This weakfish reaches a fairly large size. Individuals weighing 5 to 10 pounds are not unusual on the coast of North Carolina. The maximum weight attained, so far as known to the writers, is 16 pounds (Hildebrand and Schroeder, 1928, p. 298). The spotted trout is a fish of superior flavor, and it always commands a good price. Its importance in North Carolina is greatly increased because it is caught in the shallow waters of the estuaries and sounds throughout the year. The catches made during the winter are of particularly great importance to the fishermen who depend on haul-net fishing. A catch of speckled trout during the winter, which is a very lean season with the haul-seine fishermen, often is a real “life saver.” In fact, the speckled trout is of first importance to these fishermen in the vicinity of Beaufort. Even though the total annual catch is smaller than for the gray trout, the annual money value is greater, because of the higher price it commands in the market. Furthermore, much of the return, as already indicated, comes during the winter when it is badly needed. Although the fish inhabit the shallow waters of the sounds and estuaries during the winter, they do at times become numb from the cold. For example, on January 7, 1926, fishermen are reported to have made good wages picking up numb and floating fish in North River. The weather had been rather unusually cold for a week or more, the air temperature having dropped to 12° F. on December 28. The water temperature at 4:30 p.m. at the laboratory pier, just off a moderately deep channel where the water undoubtedly was not as cold as in the shallow estuary of North River, was 55°, 57°, 53° F., respectively, on January 5, 6, and 7. Again on January 4, 1928, many numb floating speckled trout were seen in the vicinity of Beaufort. The senior author picked up several along the shores of Pivers Island, and commercial fishermen found it profitable to “gather” the fish. Perhaps many more fish were eaten by water fowl than were gathered by man. At least, the activities of the sea gulls were great around Pivers Island. January 4 was preceded by several days of exceptionally cold weather (for that vicinity), the air temperature reaching the low mark of 14° F. on January 2. The water temperature at 4:30 p.m. was 42°, 41°, and 39° F., respectively, on January 2, 3, and 4. The numbed fish became quite active upon being taken into a heated room. It is thought, therefore, that they might have revived in nature with the return of warmer weather, if they had been left unmolested. i Fisheries Industries of the United States, 1931. By R. H. Fiedler. Appendix II, Report, Commissioner of Fisheries, 1932 (1933), pp. 97-440. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 93 Sometimes during cold weather the trout, although not numb enough to float, became quite helpless. The fish school more or less and if once found they are easily surrounded by a net. Fisherman often make large catches at such times. According to Coker (in Smith, 1907, p. 313) speckled trout have not always been present in the shallow waters around Beaufort, N.C., during the winter. However, they have occurred there with a fair degree of regularity since 1914 when the senior author first began to make observations at Beaufort. Pearson (1929, p. 190) reports that speckled trout in Texas mostly depart from the grassy areas during cold weather and seek the deeper holes and channels. At Beaufort the movement seems to be in the reverse direction. It appears to be of interest to report a remarkable case of recovery from a very serious injury. This case is of special interest, as the speckled trout and other species of the genus Cynoscion in some localities are known as weakfisli, presumably because they do not stand handling well and die quickly when caught. On June 27, 1927, a specimen of speckled trout that had had its tail severed from the body slightly behind the anal fin, was caught off Pivers Island. The wound had healed completely, and the stump of the tail, which was entirely without a fin, had become scaled over in part. The fish appeared to be healthy and in a well-fed condition. The descriptions of the young fish, under 10 millimeters in length, that follow are based on specimens from Beaufort. In drawing up those of the larger fish, collections from the Gulf coast, made by Isaac Ginsburg and John C. Pearson, were used in addition to the specimens from Beaufort. SPAWNING Ripe fish, or fish with well-developed gonads, were not seen during this study Yarrow (Smith, 1907, p. 312) observed females with large roe in April at Beaufort. J. H. Potter, a local fish dealer of long experience, reported to the senior author that he saw a ripe female, from which eggs were ‘'running”, on June 10, 1915, the first and only fish of the genus he had ever seen with “roe of any size.” Pearson (1929, p. 180), working on the coast of Texas says, “The spawning season of the trout begins in early spring (not before March) and continues as late in the summer as October. The spotted trout * * * are found in all stages of development throughout the spring and summer and probably spawn for weeks. The height of the spawning season occurs in April and May, however.” The young apparently are not numerous at Beaufort. Only 47 specimens 80 millimeters and less in length were secured during several years of rather intensive collecting, and generally only one to several specimens were taken at a time. Only 17 individuals under 5 millimeters in length were secured, and of these 6 were taken during June of 4 different years, namely, 1927, 1928, 1931, and 1932, and the other 11 were caught in one haul on August 26, 1929. No specimen less than 10 millimeters in length was caught during July. The largest specimen of the 0-class taken in June is 11 millimeters long, the largest secured during July is 71 millimeters, and the largest one taken in August is 150 millimeters in length. The size range of the limited number of specimens taken at Beaufort, therefore, suggests that spawning probably begins there in May and extends into August. Pearson (1929, p. 180) has stated: “The spotted trout spawns largely, if not entirely, within the bays and lagoons along the coast of Texas * * Twelve of the 17 specimens under 5 millimeters in length, caught at Beaufort, are from stations 94 BULLETIN OF THE BUREAU OF FISHERIES 5 and 6 miles offshore, 1 was taken in the bight at Cape Lookout, and the remaining 3 were caught in the estuary of Newport River. Such small, comparatively helpless, young almost certainly had not voluntarily traveled far. However, they may at times be carried comparatively long distances by wind and tide. It is impossible to determine the exact spawning ground or grounds from the limited material secured at Beaufort. It seems probable that spawning may take place both in the inside and outside waters. The general scarcity of the young suggests that there is no important spawning ground in the vicinity of Beaufort. DESCRIPTIONS OF THE YOUNG Specimens 1.8 millimeters long. — The head and trunk are deep, and the caudal portion of the body is very slender, an abrupt break in the ventral contour of the body occuring at the vent. The vent in these small specimens is situated in advance of midbody length, the preanal distance being contained in the length to the tip of the notochord, 2.1 to 2.4 times, and the postanal distance 1.75 to 1.9. The greatest depth of the body is contained in the length about 3.1 times, and the depth behind the vent is scarcely greater than the diameter of the eye. The mouth is moderately Figure 34. — Cynoscion nebulosus. From a specimen 2.0 millimeters long. large and strongly oblique, the gape anteriorly being somewhat above the level of middle of eye, and it extends backward somewhat under the eye. The myomeres are indistinct anteriorly and posteriorly; about 25 may be counted. The vertical finfold is uninterrupted and is without indications of fin rays. The pectoral fin membranes are prominent, but the ventrals are not evident. Dark markings are present on the ventral outline of the chest and abdomen, with a prominent spot immediately in advance of the vent. A series of close-set black spots occupies the anterior half to two-thirds of the ventral outline of the tail. A few indefinite dark, probably subsurface, markings also are present above the abdominal mass (fig. 34). The specimens described in the foregoing paragraph were taken in the same tow- net haul in which the somewhat larger ones, described in the next section were taken. The identification is somewhat uncertain because insufficient specimens are available and because of the darkened condition of some of the preserved ones, obliterating color markings. However, the larvae of nearly all the other local species of sciaenids are known. Therefore, the identification may be fairly definitely established through elimination. The larvae of C. nebulosus at a length of 1 .8 millimeters differ from those of C. regalis of the same length, so far as can be judged from the material available, in the rather deeper head and trunk, and apparently more slender tail, the decrease in depth at the vent being greater and more abrupt. The greatest depth of the body is contained in the total length to the tip of the notochord about 3.1 times in C. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 95 nebulosus, whereas the greatest depth is contained about 4.0 to 4.5 times in the length in C. regalis. Another difference apparently exists in color, as the dark spots on the ventral outline of the tail are more prominent and in a much closer set series in C. nebulosus than in C. regalis. Specimens 2.5 millimeters long. — The head and trunk are rather deep, and the caudal portion of body, although proportionately deeper than in specimens 1.8 millimeters long, remains moderately slender. The vent now is situated almost exactly at midbody length, and the greatest depth (measured somewhat behind the head) is contained about 2.6 times in the length to the tip of the notochord. The depth immediately behind the vent now is notably greater than the diameter of the eye. The mouth is large and strongly oblique, the gape anteriorly is only slightly below the level of the middle of eye, and the maxillary reaches nearly to the vertical from the posterior margin of the pupil. The myomeres are indistinct posteriorly, about 27 may be counted. The finfold remains continuous. However, a thickening of the tissues has taken place below the distal part of the notochord, and also farther forward, constituting respectively the primitive bases of the caudal and anal fins. The pectoral fin membranes are prominent, but the ventrals are not yet evident. Figure 35—Ciinoscion nebulosus. From a specimen 2.5 millimeters long. A black lateral stripe, consisting of nearly connected dashes, begins about at the vertical from the vent and extends nearly half the distance to the tip of the tail. The closely approximated dots, forming an almost continuous dark stripe along the ventral outline of the tail, although somewhat less distinct than in smaller specimens, remain present. Dark dots also remain evident on the ventral outline of the chest and abdomen (fig. 35). The body proportions, according to the specimens at hand, are almost identical in C. nebulosus and C. regalis at this size. The chief diagnostic character is the dark lateral stripe, present in C. nebulosus, but wanting in C. regalis. The first-mentioned species, also, has more numerous and more closely approximated black spots of nearly uniform size, on the ventral outline of the tail. In C. regalis the spot at about midcaudal length already is somewhat larger than the others, a distinction that becomes more pronounced in somewhat larger specimens. Specimens 8.0 to 3.6 millimeters long. — The body is quite elongate and compressed, the break in the ventral outline at the vent, abrupt in smaller specimens, no longer remains pronounced. The head and trunk now exceed the rest of the body in length, the preanal distance being contained in the length to the tip of the notochord 1.75 to 1.8 times, and the postanal distance 2.2 to 2.3 times, and the greatest depth is contained 3.4 to 3.6 times in the length. The large mouth remains strongly oblique, 96 BULLETIN OF THE BUREAU OF FISHERIES the gape anteriorly being about at the level of the middle of the eye and the maxillary reaches to middle of pupil. The notochord is curved slightly upward distally, and rather definite rays (caudal fin) are present below the curved portion. The bases of the soft dorsal and the anal fins are rather definitely shown by the thickening of the tissues, though no rays are present. The pectoral fin membranes remain rather prominent as in smaller specimens, but no ventrals are evident. The black lateral stripe already present at a length of 2.5 millimeters, has become more prominent, and it extends from the shoulder nearly to the base of the primitive caudal fin, which (as already stated) is situated under the ventral side of the distal part of the notochord. The stripe is rather more continuous than in smaller specimens, but the dashes of which it is composed still are visible in some specimens. Furthermore, the line has slight vertical projections, making its edges somewhat ragged. The line extends forward faintly across the opercle and on the snout. Both the upper and lower lips are dusky. More or less black is present over the abdominal mass, which probably is subsurface. The dark markings along the ventral outline of the tail, distinct in the smallest specimens at hand, remain, and are nearly uniform in size. A few very small and indefinite dark points are evident on the ventral surface of the head and trunk (fig. 36). Figure 36. — Cynoscion nebu'.osus. From a specimen 3.2 millimeters long. This species, at the length described, is distinguished from C. regalis principally by the black lateral line, which C. regalis does not possess. The first-mentioned species also has much more numerous black markings along the ventral outline of the tail, none of which are especially enlarged, whereas in C. regalis the spot lying at the posterior end of the base of the primitive anal is considerably larger than the others. C. nebulosus, at the size described, appears to be somewhat more slender, the greatest depth being contained 3.4 to 3.6 in length, compared with about 2.7 to 3.0 in C. regalis. Insufficient specimens are available, however, to determine definitely the relationship in this respect. Specimen 7 millimeters long. — The body is quite elongate, compressed, and shapely for such a small fish. The dorsal profile anteriorly ascends rather gently and is moderately convex. The vent is situated much nearer the base of the caudal than the tip of the snout (preanal distance 1.6 times in standard length; postanal distance to base of caudal, 2.6 times), and the greatest depth is contained 3.3 times in the standard length. The snout is rather pointed and longer than the eye. The mouth is large, less strongly oblique than in smaller specimens; the gape anteriorly is almost on a level with the lower margin of the pupil ; and the maxillary reaches about to the vertical from the posterior margin of the pupil. Although the caudal fin is well developed and apparently rather sharply rounded (the rays are broken in part and the exact shape cannot be determined), the upward curved tip of the notochord REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 97 remains visible. The anal and the soft dorsal are quite fully developed, but the spinous dorsal remains rudimentary. The anal consists of 1 1 soft rays, and the soft dorsal has 26 rays, formulae prevailing in adults. The pectoral fins, too, are developed , and have differentiated rays, but the ventrals are very rudimentary, that is, mere tufts of membrane less than half the length of the eye. A black lateral stripe persists, although rather less distinct than in smaller fish. A definite black band is present on the snout in advance of the eye, and indefinitely on the opercle behind the eye. Both lips are black. The black on the ventral outline, described in smaller specimens, now is confined almost wholly to the ventral surface of the caudal peduncle. Sub- surface dark markings over the abdominal mass remain visible, although less distinct than in younger fish. Small dusky spots are scattered over the cranium, and a row of more or less definite black spots is present on the back and extends along the base of the dorsal fins (fig. 3 7). 8 Only one specimen of the size described is at hand. This species is more slender and more shapely than C. regalis at this size (depth in standard length 3.3 times in the former and 2.8 in the latter), and the snout is more pointed. C. regalis is much plainer in color, as it has no lateral stripe and no black markings on the head or back, and only a few on the ventral outline. The enlarged black spot near the middle of the base of the anal is quite distinctive, as no such spot is present in C. nebulosus. The enlarged spot on the anal base in C. regalis is present at a length of about 3.5 millimeters, and it serves as a mark of distinction in specimens ranging from that length to 7 millimeters and upward. Specimens 10 to 12 millimeters long. — The fish is moderately slender, compressed, and shapely. The head is rather low, and pointed, its length is contained 2.7 to 2.85 times in the standard length, and the greatest depth 3.8 to 4.2 times. The snout is rather long, its length to tip of upper jaw being contained 3.5 times in the head, and the eye 4.2 to 4.4 times. The mouth is large, moderately oblique; the lower jaw projects rather prominently; the gape anteriorly is about on a level with the lower margin of the pupil; and the maxillary reaches slightly past middle of eye, its length being contained 2.25 to 2.9 times in the head. Small spines are present on the margins of the preopercle and interopercle. Two stained and cleared specimens have respec- tively 25 and 26 vertebrae. The fins are all developed, but the spinous dorsal remains quite low, the longest spines being shorter than the eye, and the ventrals, although having rays, also remain short, being equal to or scarcely longer than the eye. The * The figure by Pearson (1929, p. 179, fig. 24) based on a specimen 7.8 millimeters long, presumably taken on the coast of Texas shows the body as being much more compressed and less shapely than in the Beaufort fish. Furthermore, the mouth is shown as. less strongly oblique and the caudal fin as much more broadly rounded. Marked differences in color, also, are evident. If the figure is correct, much variation must exist among specimens. The figure certainly does not agree as to the shape of the body with a faded specimen from Texas examined by us. 98 BULLETIN OF THE BUREAU OF FISHERIES caudal fin is moderately long and somewhat pointed, the longest rays being in the lower half of fin and much shorter than the head. Pigmentation has advanced little since a length of 7 millimeters was attained. A few additional dusky markings have appeared along the black lateral stripe, and the dusky markings on the head and back have increased somewhat in number and intensity (fig. 38). 9 This species remains notably more slender than C. regalis, and the head is lower, and the snout more pointed (the depth in the former is contained 3.8 to 4.2 in the length and in the latter 2.95 to 3.0 times). The caudal fin is less sharply pointed and the rays are shorter than in C. regalis. The differences in color remain about as in 7-millimeter specimens. Although C. regalis now has a few dusky spots on the sidp and the back, they are not arranged in definite series. The characteristic dusky spot near the middle of the base of the anal remains in that species, and in some specimens a smaller one precedes it and another one follows it. No black spots occur on the base of the anal in C. nebulosus. Specimens 16 to 20 millimeters long. — The body is quite slender and moderately compressed, the shape being close to that of the adult. The head is long and low, its length being contained 2.7 to 3.0 times in the standard length, and the greatest depth of the body 3.9 to 4.15 times. The snout is rather long and pointed, its length is contained 3.5 times in the head, and the eye 4.0 to 5.0 times. The mouth has ac- quired virtually the shape and position it has in adult fish. It is somewhat oblique, the gape anteriorly is scarcely above the lower margin of the pupil, and the maxillary reaches nearly opposite the posterior margin of the eye, its length being contained 2.2 to 2.8 times in the head. Scales are evident on the middle of the side from the shoul- der nearly to the base of the caudal at a length of 16 millimeters; at a length of 20 millimeters the body is almost fully scaled. The spinous dorsal has increased greatly in height since a length of 12 millimeters was attained, and the longest spines are as long as the snout, therefore, proportionately about as long as in the adult. The caudal fin remains somewhat pointed and the longest rays are about as long as the head without the snout. The ventral fins have increased greatly in length since the fish attained a length of 12 millimeters, and now are fully twice as long as the eye. When the fish reach a length of about 16 millimeters the black lateral line, present in smaller fish, disappears or becomes laid over gradually by an indefinitely outlined dark band composed of numerous minute brownish or dark markings which extend forward on the side of the head and snout (quite indefinite on the head in some speci- mens), and backward on about the basal half of the caudal fin. Dark dots on the head and back have become much more numerous and those on the back form more — 8 The criticisms pertaining to Pearson’s figure 24 (1928, p. 179), set forth in footnote 8, in general apply to his figure 25, with the addition that the origin of the spinous dorsal, which is shown as having its origin over the pre-opercular margin, actually has its origin slightly behind the vertical from the base of the pectoral, according to specimens from North Carolina as well as from Texas, which the writers were able to examine. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 99 or less definite longitudinal bands than in smaller specimens. At a length of about 19 millimeters dark dots sometimes develop on the spinous dorsal (fig. 39). This species continues to differ notably from C. regalis in the much more slender body (depth in standard length 3.9 to 4.15 times in C. nebulosus, 3.3 to 3.4 times in C. regalis ), much lower head and more pointed snout, and in the much more promi- nently projecting lower jaw. The species also differ notably in color, as C. nebulosus is marked chiefly by dark longitudinal bands, whereas C. regalis has no longitudinal bands, but is marked anteriorly with indefinite broad dusky crossbars on the back, and posteriorly with lateral quadrate blotches. It is obvious that C. nebulosus acquires the shape and form of the adult at a smaller size than C. regalis. Specimens 25 to 80 millimeters long. — No change of importance has taken place in the proportions of the body since a length of 16 to 20 millimeters was attained. The head is contained 2.7 to 2.9 times in the standard length and the depth 3.95 to 4.2 times.10 The eye is small, being contained 4.25 to 5.0 in the head, the snout 3.25 to 4.5, and the maxillary 2.25 to 2.4 times. Gill rakers are well developed, 8 on lower limb of first arch. The body is fully scaled, and a sheath of scales is evident along Figure 39. — Cynoscion nebulosus. From a specimen 20 millimeters long. the bases of the dorsal and anal. The dorsal fins are moderately high, the longest rays and the longest spine being nearly equal in length and each is contained in the head about 2.5 to 2.8 times. The shape of the caudal remains unchanged, and the longest rays are about as long as the head. The ventral and pectoral fins are virtually equal in length and are contained 2.15 to 2.4 times in the head. A definite dark brownish band, with broken edges, extends from the snout along the side and on the caudal fin. The band is broken from the eye to the opercular margin where it forms more or less disconnected blotches. At the base of the caudal it is crossed by a pale line, and on the base of that fin it is darker and becomes somewhat pointed posteriorly. It is not definitely outlined, however, as most of the fin bears dark dots. The upper surface of the head is mostly brownish, and tills color extends on the back as two indefinite bands, one on each side of the base of the dorsal fins. Posterior to the spi- nous dorsal the bands are more or less broken up into blotches. In some specimens the bands are more nearly continuous than in others. A few dark dots are present on the dorsal fins, being most numerous on the anterior margin of the spinous dorsal. Indications of scattered dark points also are present on the anal, but the pectorals and ventrals remain unmarked (fig. 40). 10 Attention is called to figure 15 in Welsh and Breder (1923, p. 167) which is based on a specimen 28 millimeters long. This illustration appears to have been drawn from an abnormally deep fish. At least the body is shown as much deeper than in any of the specimens now at hand. Furthermore, the lower jaw projects too prominently, and the caudal fin has the longest rays in the lower half of the fin, and not in the middle as shown in the figure. 100 BULLETIN OF THE BUREAU OF FISHERIES C. nebulosus remains more slender than C. regalis (depth in standard length of the former 3.95 to 4.2 and the latter 3.3 to 3.5), the head is lower and more pointed, and the mouth is less strongly oblique (the gape in C. nebulosus being wholly below the eye, whereas in C. regalis it is only a little below the middle of the eye anteriorly). The caudal fin is rather short and only moderately pointed in C. nebulosus, the longest rays being about as long as the head, whereas in C. regalis the fin is long, sharply pointed, the longest rays being notably longer than the head. The differences in color remain about as indicated for specimens 16 to 20 millimeters long. Specimens 85 millimeters and upward in length. — The larger young of this species have been described adequately by Welsh and Breder (1923, pp. 164 to 169) and the descriptions by these authors were augmented with remarks on smaller fish by Pearson (1929, pp. 178 and 179). It does not appear essential, therefore, to draw up further descriptions. However, it does seem desirable, in the light of the study of large series of specimens, to include a discussion of the relationship of this species and C. regalis, as development proceeds. The species are most readily separated by color, as the patterns are very different. However, evident structural differences also are present. In C. nebulosus the dark longitudinal bands, present in smaller specimens, begin to break up into spots when the fish are about 30 millimeters long. At a length of 35 to 40 millimeters, the band on the back is divided into more or less definite quadrate blotches, but the lateral band remains almost continuous. When the fish reach a length of about 60 to 70 millimeters the lateral band also becomes divided into spots, and the caudal fin, which previously was almost solidly black at the base, has become spotted. At about the same size spots develop between the bands, and at a length of 70 to 80 millimeters the entire upper half of the side bears indefinitely outlined spots, which are irregular in shape and variable in size. The spots (in preserved specimens) are rather more brownish in color than in younger fish. At a length of 110 to 120 millimeters, the characteristic small, roundish black spots of the adult, which occupy the upper two-thirds of the sides as well as the dorsal and caudal fins, are quite fully developed. Specimens of C. regalis, 35 to 40 millimeters long, on the other hand, have broad dark crossbands on the upper part of the side, which extend to or somewhat below the lateral line. These bars disappear when the fish reach a length of about 70 to 80 millimeters. Specimens ranging from about 80 to 100 millimeters in length are almost plain grayish above and silvery below. The characteristic dark wavy lines REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 101 or reticulations of the adult begin to appear when the fish reach a length of about 100 millimeters, and at a length of about 120 to 130 millimeters the fish have acquired almost fully the color pattern of the adult. C. nebulosus is more slender and more shapely than C. regalis throughout the younger stages. However, at a length of about 175 millimeters the differences in the shape of the body virtually have disappeared. It is evident from table 3 that C. nebulosus acquires the proportionate depth of the adult very early, that is, at a length of about 10 to 12 millimeters, whereas C. regalis is much deeper throughout the younger stages and does not have the proportionate depth of the adult until a length of about 175 millimeters is attained. The shape of the caudal fin differs quite notably in the two species throughout the younger stages, that is, from a length of about 20 to 125 millimeters. During this entire period of development this fin, although pointed in both species, is notably longer in C. regalis than in C. nebulosus, the longest rays in the former being the middle ones, which are more or less filamentous in the younger stages, and equal to or longer than the head, whereas the longest rays in the latter are in the lower half of the fin and are not longer than the head. For some time during the course of development the upper lobe of the caudal fin in both species is slightly concave and the lower lobe more or less roundish. Although the fin must undergo a greater change in C. regalis to acquire the broadly concave outline which obtains in adults of both species, it is acquired somewhat earlier, that is, at a length of about 250 to 280 millimeters (there being much variation in the development among specimens), whereas this shape is not acquired by C. nebulosus until the fish is about 300 millimeters long. The two species differ also in the number of gill rakers, C. regalis having 11 or 12 on the lower limb of the first arch, whereas C. nebulosus has only 7 or 8. The gill rakers are developed and can be counted readily when a length of about 26 milli- meters is attained. The species differ, furthermore, in that C. nebulosus has no scales on the fins throughout life, whereas C. regalis has minute scales on about the basal half of the soft dorsal, the caudal and the anal fins. The scales first appear when the fish are about 45 to 55 millimeters long. Table 3. — -Proportionate depths of specimens of various size of Cynoscion nebulosus and C. regalis C. nebulosus C. regalis C. nebulosus C. regalis Total length in millimeters Depth in standard length Number of specimens measured Depth in standard length Number of specimens measured Total length in millimeters Depth in standard length Number of specimens measured Depth in standard length Number of specimens measured 7 3.3 1 2.8 1 70-90 3. 7-4. 0 7 3. 4-3. 6 3. 5-3. 6 3. 5-4. 1 4. 1-4.4 7 5 4 6 10-12 3. 8 -4. 2 3 2. 95-3. 0 3 120-130- 3. 8-4. 1 5 16-20 3. 9 -4. 15 3 3. 3 -3. 4 3 150-175- 3. 9-4. 25 4 25-30 3. 95-4. 2 6 3. 3 -3. 5 6 210-240. 4. 1-4. 3 6 40-50 4. 0 -4. 25 6 3. 3 -3. 4 6 DISTRIBUTION OF THE YOUNG It has been stated elsewhere (p. 93) that the young under 5 millimeters in length, in the collection mostly were taken at off-shore stations. The larger ones of the same year class, ranging from 5 millimeters upward, however, were all caught in inside waters. The indications are, therefore, that the young, if actually hatched in part off shore, enter the inside waters soon afterward. 102 BULLETIN OF THE BUREAU OF FISHERIES Only a few of the smaller larvae at hand were caught in surface towings, and all the larger ones were taken in bottom hauls. Therefore, the indications are that the young of this species are chiefly bottom dwelling. GROWTH The number of young of the 0-class, consisting of only 126 specimens taken and measured, is entirely too small to cast much light on the rate of growth. However, the measurements parallel those of C. regalis rather closely, and indicate that the fish grow rapidly during the first summer of life, attaining a modal length of about 170 millimeters (6% inches) when 7 or 8 months old. The size at which sexual matur- ity is reached remains unknown. CYNOSCION REGALIS (Bloch and Schneider) GRAY TROUT; GRAY WEAKFISH The gray trout ranges from Massachusetts to the east coast of Florida11. It is a commercially important species from Rhode Island to Florida. The States having the largest catches are New Jersey, with an average annual catch from 1928 to 1930 inclusive of 9,160,346 pounds;12 Virginia, 12,710,389 pounds; and North Carolina, 4,415,059 pounds. This trout is reported to reach a maximum weight of 30 pounds. In Chesapeake Bay fish weighing 10 pounds are not particularly uncommon. How- ever, at Beaufort fish weighing 6 pounds are considered large ones and 9-pound fish are exceptional. The gray trout is a well-flavored fish and always commands a good price. This species and the spotted trout constitute the chief support of the haul-seine fishermen who generally go in pairs and fish the shallower inside waters in the vicinity of Beaufort. Their equipment consists of a small motor boat, which generally has a small cabin forward over the engine, a rowboat, and a 300 yard straight dragnet. Frequently two crews “pair”, that is, they unite their gear and power by lashing their nets together and making long drags, using the power of their motor boats. The gray trout is taken commercially at Beaufort from about March to Decem- ber, the length of the fishing season varying according to the annual fluctuation in the mildness, or severity, and duration of winter weather, for the large fish leave when cold weather comes and return as early in the spring as the temperature of the water again becomes agreeable. The large fish evidently are more sensitive to temperature than the smaller ones. Whereas, the larger ones leave the shallow waters, as already stated, upon the approach of cold weather, the smaller individuals remain there, except during the brief cold snaps that occur locally. Small gray trout were taken many times during the winter in Newport River, for example, during the several years of the present investigation. They were common also off Bogue and Shackleford Banks in a few to several fathoms of water. However, they pre- sumably withdraw to deeper and warmer water during cold snaps, but return within a few days after the temperature rises. Therefore they may make several migra- tions during one winter. The large fish, on the other hand, make only one migration 11 In a study of the sea trouts of the Gulf coast, Ginsburg (1929, p. 83) found that the fish from that coast, formerly considered identical with the gray trout of the Atlantic coast, is a distinct species, which he named C. arenarius. Therefore, the range of the gray trout, previously thought to extend to Texas, is now restricted to the Atlantic coast. 12 The statistics offered are taken from “Fisheries Industries of the United States”, for the years 1929, 1930, and 1931. The “average annual catch” is the arithmetical average for the 3 years, 1928, 1929, and 1930. The gray trout and the spotted trout are listed separately only for 1930. The spotted trout is unimportant, from a statistical standpoint, in all the States named, exclusive of North Carolina in which this species constituted 22.7 percent of the catch in 1930. However, to make the figures comparable the entire catch of trout is considered for each State. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 103 as already indicated, leaving the shallow waters in the autumn or early in the winter and returning early in the spring when low temperatures are not likely to recur. Although the smaller gray trout generally migrate away from the shallow waters during cold snaps, they apparently do not always go early enough, for on January 14, 1927, quite a few dead and numb fish were seen along the shores of Pivers Island. A sample of 6 fish was gathered having a range in length of 122 to 182 millimeters. Several days of rather unusually cold weather preceded January 14. The water temperature at the laboratory pier, taken daily at 4:30 p.m., dropped to 41° F. on January 9, and it remained there until January 14 when it came up to 48° F. A temperature of about 41° F., therefore, probably is close to lethal. SPAWNING The spawning season of the gray trout is a moderately long one. Fish with well-developed roe (although not ripe) were observed as early as April 16 (1914), and as late as August 17 (1914). The young were taken first on May 25 (1932). However, they are rather large ones, ranging from 6.5 to 9 millimeters in length, and therefore may have been a few weeks old when caught. Larvae under 5 milli- meters in length were caught during June, July, and as late as August 27 (1930). According to these data the spawning season at Beaufort begins in May and extends through June, July, and August. The conclusion relative to the spawning period at Beaufort is in general agree- ment with statements by Welsh and Breder (1923, pp. 150-158), who say that the season is a protracted one, commencing in May and continuing until September. The great majority of the fish are reported, however, to spawn between the middle of May and the middle of June. These authors state, furthermore, that the spawning season is little affected by the latitude and that it occurs at about the same time from Cape Cod to the Carolinas. Higgins and Pearson (1927, p. 57) found virtually all mature females taken in pound nets in Pamlico Sound, N.C., in a spawning condition in the second week of June (1925) when they began their investigation, and by August 10 spawning had been completed. It was noticed at Beaufort that the large fish develop roe earlier than the smaller ones. Among thousands of fish examined only large ones had gonads in an advanced state of development in April, whereas in August those fish that still contained roe were small ones, generally 15 inches and somewhat less in length. In the vicinity of Beaufort spawning apparently takes place principally, if not wholly, at sea. This conclusion is arrived at from a study of the movements of adult fish and from a study of the distribution of the young. It has been stated elsewhere that the large fish return to the inside waters early in the spring from as yet a rela- tively unknown winter home. For a time these fish are quite abundant. However, within a month or two after their arrival, or about the latter part of May, the large fish become scarce and remain so until about July when they again increase in abun- dance. During two seasons (1914 and 1915) of almost ceaseless effort no females quite ripe were found among thousands of fish examined. Virtually all contained roe early in the spring. Later, when the large fish returned after a period of absence they had spawned. The explanation appears to be that the fish go out to sea to spawn and return to the estuaries and sounds when the process is completed. The foregoing explanation of the scarcity of large fish in the inside waters during a period, which undoubtedly is the height of the spawning season, is substantiated in 57094—34 5 104 BULLETIN OF THE BUREAU OF FISHERIES a measure by the distribution of the young 10 millimeters and less in length, as shown by the collections. Young 5 millimeters and less in length were taken quite sparingly, as only 25 were secured, including 17 taken outside. The young ranging upward of 5 millimeters in length were taken much more abundantly. The collection contains 270 specimens 6 to 10 millimeters long, including 253 individuals taken off Bogue and Shackleford Banks and only 17 from Beaufort Harbor and adjacent waters. In explanation it may be stated that the foregoing results were secured, notwithstanding that many more hauls, using identical gear, were made in the inside than in the out- side waters. It seems to be evident, therefore, that the young, 10 millimeters and less in length, are much more numerous off Bogue and Shackleford Banks than within the harbor and adjacent sounds and estuaries. The bulk of the young undoubtedly remain in the general vicinity where they are hatched until the power to swim is developed. Occasionally the floating eggs and the helpless larvae no doubt are carried considerable distances by the wind and tide. It is possible that the comparatively few larvae taken in the inside waters were carried there involuntarily. The fins are entirely undeveloped until a length of about 6 millimeters is attained, and imperfectly until a length of at least 10 milli- meters is reached. Prior to the rather full development of the fins the young fish probably cannot swim far, nor in a definite direction. Therefore, the majority of the young 10 millimeters and less in length would be expected to be found chiefly near the spawning ground. The conclusion that in the vicinity of Beaufort spawning takes place principally, if not wholly, at sea, therefore, seems to be justified from the information given relative to the movements of the adults during the spawning season, and from the data presented concerning the distribution of the young. DESCRIPTIONS OF THE EGGS AND YOUNG The reader is referred to Welsh and Breder (1923, pp. 150-153) for descriptions of the eggs and their development, as well as for accounts of the newly hatched young. Descriptive notes and illustration of the development of the young are offered by Pearson.13 They have been compared with specimens of similar sizes in our collection and found to be essentially correct. Accordingly, descriptions and illustrations are omitted in this paper. However, the diagnostic characters are shown in the accompanying keys. Some remarks relative to the eggs of the gray trout seems to be in order, because of the extremely great variation in their diameter shown by Welsh and Breder and also by Pearson. It is believed that the eggs were taken numerous times off Beau- fort Inlet, but were confused with those of the white perch, Bairdiella chrysura. How- ever, near the close of the investigation it was noticed that the larger eggs, identified as perch eggs, had more numerous pigment spots on the yolk and these were particu- larly prominent on the embryo. It was noticed, also, that the newly hatched fish not only were larger as would be expected, but contained more pigment spots than those hatched from the smaller eggs. Furthermore, the larvae hatched from the larger eggs have two bars composed of dots slightly greenish in color in life on the caudal portion of the body; one of the bars being at about midcaudal length, and the other one, which is rather less distinct, about midway between the vent and the bar already described. The larvae hatched from the smaller eggs, on the other hand, have only one bar which is situated at about midcaudal length. n Unpublished manuscript in the files of the Bureau of Fisheries. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 105 The numerous dots on the egg and embryo are in general agreement with the illustrations given by Welsh and Breder (1923, p. 152) of the eggs of C. regalis, and the illustrations of the recently hatched young presented on page 155 of the work already cited contain suggestions of two crossbars on the caudal portions of the body, though less distinct than in the specimens observed by us. Insufficient eggs of the larger type were secured, after the differences in pig- mentation were noticed, to obtain a sufficiently large series of measurements to show the exact relationship in size. A sample of 205 eggs was taken at random from a towing made on June 3, 1930. Among this lot 25 eggs of the type described were found. These eggs ranged in diameter from 0.8 to 0.92 millimeter and the average was 0.84 millimeter. The smaller eggs, namely, those of the white perch, Bairdiella chrysura, had a range in diameter of 0.66 to 0.76 millimeter and an average of 0.69 millimeter. The last-mentioned proportions quite certainly are correct for Bair- diella, as the range of 97 eggs, spawned in the aquarium on 2 different dates, ranged in diameter from 0.66 to 0.72 millimeter and averaged 0.68 millimeter. Although the number of the larger eggs measured is quite limited, enough of the smaller ones were measured to show that two distinct groups were present.14 Kuntz (1914, pp. 4-10) states that the diameter of the eggs of Bairdiella ranges from 0.7 to 0.8 millimeter, thereby indicating that he may have measured eggs of two species. However, his description and illustrations of the development appear to be correct for Bairdiella. Welsh and Breder (1923, p. 151, table 3) give measurements of eggs taken from 2 specimens of C. regalis, which show, as already stated, a comparatively large difference in the size of the eggs produced by the 2 examples. One of the fish had eggs ranging in diameter, according to 8 eggs measured, of 0.98 to 1.03 millimeters and, therefore, all larger than any measured at Beaufort. The other one had eggs ranging in diameter, according to 10 eggs measured, from 0.8 to 0.84 millimeter and, therefore, near the size of the eggs taken at Beaufort. Pearson (MS.) working with specimens collected at the entrance of Chesapeake Bay, found even a larger range among eggs which he believed to be those of C. regalis. He remarks, however, that the eggs fall into two groups as to size. One of the groups ranged in diameter from 0.7 to 0.9 millimeter, with a mode at 0.83 millimeter, and the other one had a range in diameter of 0.9 to 1.13 millimeters, with a mode at 1.0 millimeter. Therefore, the range of the two groups of eggs combined is greater even than that given by Welsh and Breder (loc. cit.), as it extends from 0.7 to 1.13 millimeters. The time during which Pearson took the eggs of both groups (the larger ones from April 24 to May 23 and the smaller ones from May 6 to July 24) quite probably covers nearly the entire spawning period of C. regalis. However, the pigfish, Ortho- pristis chrysopterus, and the perch, Bairdiella chrysura, too, spawn during this period, and their eggs faff into the range of the eggs measured, and are extremely similar in appearance. It has been stated already that the eggs of Bairdiella measured at Beaufort ranged in diameter from 0.66 to 0.76 millimeter and that they are similar in appearance during the cleavage stages to somewhat larger eggs believed to be the eggs of the gray trout. The eggs of the pigfish, according to Hildebrand and Cable (1930, p. 399), range in diameter from 0.7 to 0.8 millimeter, and therefore are more or 14 The oil globule, too, was measured in each egg. There is definite overlapping and only an average difference appears to exist, the average in 100 perch eggs being 0.18 millimeter and in 25 trout eggs 0.2 millimeter. 106 BULLETIN OF THE BUREAU OF FISHERIES less intermediate in size of the perch and the trout, but overlap somewhat with both species. At Beaufort the eggs of the three species named often were taken in the same towing, assuming of course that the largest ones were trout eggs. The pigfish eggs, like those of the gray trout, are very similar to the perch eggs and are scarcely dis- tinguishable (although their average size is larger), until the embryos become well developed. In the advanced embryonic stages the oil globule in the perch egg acquires greenish specks, whereas the oil globule generally remains clear in the pigfish, or at most it acquires only a few minute specks. Furthermore, the position of the oil globule in relation to the embryo is characteristic. In the perch (also the trout) it lies far behind the head, whereas in the pigfish it lies at or near the ventral surface of the head. The newly hatched fish, too, may be distinguished by the location of the oil globule within the yolk sac, as it retains the approximate position it has in the em- bryo, that is, it lies in the posterior part of the yolk sac in the perch (and trout), and in the anterior part of the yolk sac or under the head in the pigfish. Pearson’s (MS.) measurements, in the absence of descriptions of color marking and statements in regard to the position of the oil globule with respect to the embryo, suggest that he may have grouped the eggs of the pigfish, the white perch (also known as the yellow-tail perch), as well as those of the gray trout, all under C. regalis. All three species no doubt spawn in Chesapeake Bay or off the month of the bay during the period in which the eggs were taken by Pearson, as Hildebrand and Schroeder (1928, pp. 258 and 280) took spawning pigfish during June, and ripe perch during May in Chesapeake Bay, and Radcliffe (in Welsh and Breder, 1923, p. 151) took trout eggs there during the spring. Since Pearson does not list the eggs of the pigfish and the perch (both very common species in Chesapeake Bay), although they most probably were present among the trout eggs, lends support to the supposition that they may not have been distinguished. Too much faith apparently should not be placed in the measurements of eggs, unless it is known definitely that the eggs were spawned in water of about even density. If preserved eggs are measured, it would seem necessary to use one preservative of uniform strength. Delsman (1931, p. 403) has shown that a comparatively large difference in size exists in the eggs of one species ( Cybium guttatum), depending upon the density of the water in which they are taken. Eggs collected in the mouths of rivers where the water was brackish were larger, varying in diameter from 1.24 to 1.36 millimeters, than those taken in salt water, which had a range in diameter of 1.05 to 1.26 millimeters. Delsman concludes: So we easily came to the conclusion that the size of the egg depends on the salinity of the water, increasing or decreasing in proportion to the latter getting lower or higher. The same phenomenon has been observed in other pelagic eggs, the eggs of the Baltic Sea fishes, e.g., being bigger than the corresponding eggs of the North Sea. Kuntz (1914, pp. 4-10), presumably Welsh and Breder (1923, p. 151), and the present writers measured live eggs, whereas Pearson measured preserved eggs. The present investigators are not aware of data that show how the eggs are affected by the commonly used preservatives. Pearson apparently assumes that some shrinkage takes place and, therefore, left the eggs “in preservative at least 6 months in order to obtain nearly the minimum shrinkage.” However, he does not name the preserva- tive, nor the strength at which it was used. The measurements made by Pearson and the other investigators, therefore, may not be directly comparable, although in general they are in agreement, as already stated. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 107 The measurements made by us, which are shown in preceding pages, were all made in the laboratory with an eyepiece micrometer and are based on eggs taken at sea (exclusive of some perch eggs spawned in a tank, as stated in the text) and brought to the laboratory alive in water dipped up where the towings were made. The water used in the tank in which some of the perch eggs were spawned was pumped from underneath the laboratory pier at high tide when the salinometer readings generally average about 1.025, as compared with 1.03 off Beaufort Inlet. Therefore, all the measurements made at Beaufort are based on eggs spawned in salt water of rather high density. It has been shown that the eggs of two examples of gray trout measured by Welsh and Breder (1923, p. 151) differed markedly in size. The writers assume that some fish simply have larger eggs than others, and do not state whether the eggs of both fish were spawned in water of about even density. According to our measurements, the eggs of any one species, when spawned in strictly salt water, are quite uniform in size. Our experience, and the information gained by Delsman (1931, p. 403) and others, relative to the difference in the size of the eggs of one species according to the density of the water in which the eggs are spawned, suggest that Welsh and Breder may have used water of very unequal density in which the eggs of the two samples of trout mentioned presumably were artificially spawned. Certainly we must con- sider the large variation in the diameter of the trout eggs given by the writers men- tioned as unusual. In fact such a wide variation, under identical conditions, would suggest the presence of two races. DISTRIBUTION OF THE YOUNG It has been stated in the section of this paper dealing with the spawning (p. 104) of tins species that the young, 10 millimeters and less in length, are much more numer- ous off Beaufort Inlet than in the harbor and adjacent sounds and estuaries. Some- what larger fish, too, appear to be rather more numerous in the outside than the inside waters, although they tend to congregate and become abundant in certain restricted areas in the estuaries. For example, the young ranging from about 15 to 100 milli- meters in length are very numerous above “cross rock” in Newport River during the summer or after the middle of June. The water is only a foot or two deep above cross rock at low tide and the bottom is very muddy. The water generally is brackish, but the density fluctuates greatly, depending upon the stage of the tide and the rain- fall, as this mud flat is only a short distance from the mouth of the narrow channel of Newport River where this stream discharges fresh water into its estuary. Pearson (MS.), too, found that young gray trout when about 8 to 10 millimeters long “settle on the bottom” in Chesapeake Bay in coves and creeks at Lynnhaven Roads and elsewhere. Soft or muddy bottom seems to be preferred by the young during the first sum- mer. Those that remain in the ocean, as well as those that enter the harbor, seem to seek muddy bottom. For that reason, presumably, the young are numerous in the vicinity of the sea buoy off Beaufort Inlet where the bottom is soft. It is presumed that the young fish find food abundant on the mud. However, the food require- ments of the young under about 40 millimeters in length have not been studied. Larger ones, according to Welsh and Breder (1923, pp. 159-164), feed on a variety of forms, including copepods, amphipods, isopods, shrimps, crabs, worms, and fish. 108 BULLETIN OF THE BUREAU OF FISHERIES The number of larvae, 5 millimeters and less in length, in the collection is too small to show the vertical distribution. Since the eggs are pelagic, the larvae would be expected to remain at or near the surface, at least, until the yolk is absorbed. The yolk sac contains an oil globule which persists until virtually all the yolk is absorbed (Welsh and Breder, 1923, p. 153, fig. 10). The oil globule almost cer- tainly would keep the newly hatched fish at or near the surface. However, the indi- cations are that the young fish in part, at least, go to the bottom very early, as the 25 specimens at hand having a range in length of 1.5 15 to 5.0 millimeters were all taken in bottom hauls. The larger ones undoubtedly stay principally on the bottom, as all except 2 of a total catch of 270 specimens, 6 to 10 millimeters long, were caught in bottom hauls. It is stated elsewhere (p. 102) that the large fish leave the shallow waters during cold weather, whereas the smaller ones are more apt to remain. The young of the 0- class were taken throughout the winter. It seems significant, however, that the larger individuals are missing during the coldest months (see table 5), indicating that the largest young of the year probably follow the adults to a more agreeable habitat. GROWTH The study of the rate of growth was limited chiefly to the 0-class. The tables and graph presented herewith show quite conclusively that the rate of growth is very rapid during the first summer. The distinction between the 0-class and the 1- class is not very distinct after July, as some overlapping takes place, and it is prob- able that a few specimens may have been wrongly assigned in table 5. However, the specimens improperly placed are so few that the averages cannot be affected importantly, and therefore may be accepted as approximately correct. The largest individuals of the 0-class, as shown by the accompanying tables and graph, already had attained a length of 215 to 220 millimeters (8)4 to 8% inches) in October, and the average length of 914 fish measured was 150.4 millimeters (6 inches). In December a few individuals, apparently belonging to the 0-class, had attained a length as great as 220 to 233 millimeters (8% to 9% inches), and the average 15 Specimens only 1.5 millimeters long in the present collection do not retain a yolk sac. Much shrinkage takes place in larval fishes during the hardening process if alcohol and formalin is used, as with the specimens at hand. Judging from the degree of shrinkage known to occur in some other species, as in the pigfish ( Osthopristes chrysopterus), for example, it seems probable that the specimens of trout 1.5 millimeters long may have had a length of 2.5 millimeters before preservation. Welsh and Breder (1923, p. 153, fig. 10) show that the yolk sac is very small in live or fresh larvae when 2.2 millimeters long, and it would be expected to disappear very soon afterward. Figure 41. — Growth of Cynoscion regalis during first year or so of life. Solid line, average size; dot and dash line, maximum size; broken line, minimum size. (Graph based on table 5.) REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 109 length of 464 specimens measured was 168.8 millimeters (6% inches). The larger fish of the 0-class appear to be missing in the catches made during the winter, as the maximum and the average lengths of the specimens caught during January, February, and March were less than for December. (See table 5.) In April the largest indi- viduals were 225 to 238 millimeters (9 to 9F inches) long, and the average length of 1,455 fish measured was 167.2 millimeters (6% inches). Therefore, the fish caught in April were about the same size as those taken in December. The data presented appear to show that some of the larger individuals of the 0-class, like virtually all the adults, leave the shallower water, in which the collections were made, during cold weather. The data appear to show, also, that fish grow little, if at all, during their first winter, for in April even the adult fish have returned, and the samples taken should represent the class correctly. After April growth appears to be fairly rapid, for in June (when many of the fish were a year old) the average length of 615 examples measured was 175.8 millimeters (7 inches) and the extremes in length extended from 97 to about 253 millimeters (4 to 10 inches). The length attained in June shows that some of the fish undoubtedly reach a marketable size 16 during their second summer. The size attained by the young fish at Beaufort, as shown by the accompanying tables, is not in disagreement with published accounts. Eigenmann (1901, p. 47), working with fish at Woods Hole, Mass., states, “It seems very probable that the fish reaches a marketable size in about a year from birth.” Welsh and Breder (1923, p. 157) have calculated the average length of fish hatched on June 1, based on “meas- urements of 10 large samples of young fish, taken at various times from July to March” to he 30, 80, 130, 170, and 180 millimeters (1 %, ?>}%, 5’s, 6%, and 7 inches), respectively, on July 1, August 1, September 1, October 1, and November 1. These writers unfortunately do not name the localities in which the fish were caught. Therefore, the data cast no light on the rate of growth in different localities. No calculations were made by us that are directly comparable. However, the lengths given by the authors cited, in general, fall in with the larger individuals of the 0-class caught at Beaufort, N.C., on approximately the same dates. Table 4. — Length frequencies of 8,382 gray trout ( Cynoscion regalis) under 380 millimeters in length Millimeters May June July Aug- ust Septem- ber October Novem- ber Decem- ber Janu- ary Febru- ary March April 0-4 1 4 5 10 5-9 8 25 164 25 10-14 10 123 20 1 15-19 7 29 5 20-24 3 14 8 25-29 .. 1 23 10 30-34 .. 12 14 2 35-39 11 7 40-44 2 14 2 1 45-49 17 2 3 3 1 50-54 24 4 1 55-59 30 2 2 2 60-64 29 1 4 7 65-69 27 5 6 9 70-74 34 11 6 12 75-79 55 11 2 2 80-84 38 24 4 1 1 85-89 47 23 12 3 1 1 1 1 90-94 38 47 9 2 1 2 1 95-99 1 33 51 19 1 1 2 100-104 32 54 21 6 1 6 1 2 1 105-109 1 21 81 11 9 6 2 2 1 3 1 110-114 18 58 30 21 9 5 2 2 2 3 115-119 11 49 34 16 3 3 1 6 1 13 120-124. i 12 31 37 16 9 6 1 11 i 26 19 The legal minimum size limit in North Caroliift in 1925 is given as 9 inches by Higgins and Pearson (1927, p 49, fig. 8). 110 BULLETIN OF THE BUREAU OF FISHERIES Table 4. — Length frequencies of 8,382 gray trout ( Cynoscion regalis) under 380 millimeters in length — Continued Millimeters May June July August Septem- ber October Novem- ber Decem- ber Janu- ary Febru- ary March April 125-129 1 1 5 37 33 17 8 8 1 7 2 29 130-134 3 4 25 42 19 13 12 4 11 1 32 135-139 3 4 1 16 30 22 25 13 9 11 2 32 140-144 3 14 2 25 33 62 34 13 25 25 2 60 145-149 2 21 29 28 102 63 11 38 26 9 76 150-154 7 41 3 21 24 139 76 30 60 32 27 125 155-159 5 23 5 20 22 121 87 26 68 26 19 181 160-164 4 47 23 34 18 109 75 41 75 40 33 228 165-169 9 40 18 26 21 66 76 38 57 21 30 184 170-174 14 55 18 18 20 42 73 56 42 26 32 147 175-179 9 46 12 6 11 21 33 43 28 27 25 no 180-184 7 43 8 5 23 27 23 40 18 16 26 82 185-189 4 42 12 3 14 15 23 28 26 12 17 40 190-194 4 47 19 5 14 16 20 28 11 3 19 32 195-199 4 53 14 4 6 6 10 10 7 6 9 11 200-204 5 41 21 3 4 6 3 20 4 3 8 14 205-209 --- 2 25 25 14 3 6 5 10 3 3 210-214 _ 4 19 23 9 3 2 2 4 4 10 215-219 . 1 17 25 6 6 4 2 5 1 7 220-224 2 13 17 6 8 2 1 2 2 1 3 225-229 _ . _ _ __ 5 17 7 13 1 1 2 230-234 _ 4 7 3 10 5 1 1 1 235-239 . 5 4 10 2 1 240-244 _ - __ 3 2 9 10 2 1 5 245-249 - 1 1 9 7 3 1 2 250-254 . . 3 1 9 -7 5 1 1 2 255-259 _ - 2 9 6 1 2 260-264 . . 2 10 6 3 1 1 265-269 1 2 4 2 270-274 1 4 6 3 1 275-279 _ - - 2 4 1 4 1 1 280-284 _ _ 10 5 2 3 1 285-289 - - - 1 290-294 . _ . 10 2 2 1 295-299 4 4 1 2 1 1 1 300-304 _ 16 6 1 305-309 _ 1 9 310-314 . _ 12 1 315-319 - 12 9 2 2 320-324 . 15 1 325-329 . - 3 1 1 330-334 13 1 335-339 5 340-344 12 3 345 349 2 350-354 7 3 355-359 - 1 2 360-364 13 1 365-369 --- --- 5 3 370-374 _ 16 .175 379 3 Table 5. — Monthly summaries of length measurements of 7,883 gray trout ( Cynoscion regalis) during the first 12 to 14 months of life ( based on table 3) Month May June July August September. October. .. November. December. Fish measured Smallest Largest Average Month Fish measured Smallest Largest 8 6.5 9 8. 1 January 486 78 204 52 2.3 40 11.5 February. 311 106 220 873 4.4 143 50.4 March 281 94 229 820 1.7 185 105.4 April 1,455 87 238 548 10.5 209 140.3 May 90 125 222 914 46 220 150.4 June 615 97 253 685 48 231 164. 1 July 281 150 265 464 85 233 168.8 Average 161.5 152.5 165.5 167.2 174.6 175.8 197.9 CYNOSCION NOTHUS (Holbrook) BASTARD TROUT The bastard trout ranges from Chesapeake Bay to the southwestern coast of Texas. Although it was described (from South Carolina) as early as 1860, its exact status as a species was not well understood until recently. For a time its validity was doubted by some writers (Coles, 1916, p. 30, and Welsh and Breder, 1923, p. 169), probably because plain colored C. regalis were thought to be G. nothus. However, REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 111 Hildebrand and Schroeder (1928, p. 299) obtained two specimens of true C. nothus from Chesapeake Bay, and thereby were enabled to show definitely that this is a valid species. Finally Ginsburg (1929, p. 71-85), who examined many specimens, shows the relationship of C. nothus , C. regalis, and C. arenaris (a Gulf coast species) in great detail. The fishermen and fish dealers of Beaufort and Morehead City recognize C. nothus as distinct from C. regalis, giving it the name “bastard trout” because it is supposed to be a cross of C. regalis and some other species. The diag- nostic characters of the bastard trout are given in the key to the species. It has been found during recent years that the bastard trout is rather common off Beaufort Inlet during the summer. However, no specimens were taken during the winter, probably because the fish migrate to warmer water. This trout certainly is not a regular inhabitant of the inside waters of the vicinity, as no adult specimen and only 4 young were taken there during 5 years of rather systematic and extensive collecting. The young, all taken in 1 tow-net haul, may have drifted in with the tide, as explained elsewhere. The bastard trout appeared rather frequently during the summer, sometimes in considerable numbers, in a collecting trawl hauled in a few to several fathoms of water off Shackleford and Bogue Banks. At times, particularly during the early part of summer, catches of about 100 pounds a day were taken in a commercial pound net operated for a few years off Bogue Banks. It cannot be said, however, that this species is of much commercial importance on the coast of North Carolina. Further southward, and particularly on the Gulf coast, it seems to be more important. The bastard trout apparently does not grow large. The largest individual seen at Beaufort was only 228 millimeters (9 }{ inches) long. This species is closely related to C. regalis, and the separation of the young prior to the development of the anal fin, that is, specimens under about 6 millimeters in length, may be difficult. It is thought, however, that the position of the color markings, which differ somewhat in specimens 9.5 millimeters long (the smallest C. nothus recognized), as pointed out on page 113, would aid in separating the smaller young. The position of the color markings in all the specimens at hand of tills group under 9.5 millimeters in length agree with C. regalis. It is concluded, therefore, that specimens of C. nothus under 9.5 millimeters in length are missing. The descriptions of the young that follow are based on specimens collected at Beaufort, exclusive of the account of fish 75 millimeters and upward in length, which is based in part on specimens from Beaufort and in part on specimens from the Gulf coast. SPAWNING No ripe fish, so far as known, have been observed, and if the eggs have been taken in the tow they were not recognized, therefore, the only clue to the time and place of spawning is furnished by the collection of young fish. The smallest young taken were secured on August 8, 1930, and the smallest specimen in that collection is 9.5 millimeters long. The largest specimen taken the same month, which certainly belongs to the 0-class, is 88 millimeters long. No specimen of this year class less than 39 millimeters and none exceeding a length of 95 millimeters were taken dur- ing September. However, in October 30 specimens 17 to 25 millimeters in length were secured, and the largest one of this year class is only 80 millimeters long, show- ing that the larger young are missing in the collection. In November, when nu- merous specimens of the 0-class were taken, the range in length extends from 19 to 112 BULLETIN OF THE BUREAU OF FISHERIES 147 millimeters. It seems reasonable to conclude from the range in size of specimens taken during four summer and fall months, when the rate of growth no doubt is rapid, that spawning very probably begins in May or June (for specimens 80 to 88 mil- limeters long, such as were taken on Aug. 14, 1930, quite probably are 2 to 3 months old), and that it extends at least through August (as 30 specimens 17 to 25 milli- meters long, were taken on Oct. 30, 1929). It probably may be assumed that spawning takes place at sea at Beaufort in the general vicinity inhabited by the adults. At least, that is where all the young in the collection, exclusive of one small collection consisting of 4 specimens 9.5 to 12 millimeters long, were taken. It is thought that the 4 specimens of young, too, were hatched at sea, but that they had drifted in part with the tides and probably in part had swum to the place of capture, namely, near the mouth of Core Creek in Newport River. The presence of these four fish in the inside waters is interesting, not only because no other young were taken there, but because no adults were caught in the inside waters. DESCRIPTIONS OF THE YOUNG Specimens 9.5 to 11.0 millimeters long. — The body is rather deep and com- pressed, the depth being contained 3.0 to 3.25 in the standard length. The head particularly is short and deep, being contained 2.85 to 3.0 times in the length. The snout, too, is rather short and blunt, and is about as long as the eye, 3.7 to 4.4 times in the head. The mouth is large, quite strongly oblique; the gape anteriorly is scarcely above the lower margin of the eye; the lower jaw projects only slightly; and the maxillary reaches nearly opposite the posterior margin of the eye, 1.65 to 1.85 in the head. The fins are all developed and a definite enumeration of the rays is obtainable. The anal consistently has 9 articulated rays in the 4 specimens of this size at hand (although in larger specimens only 8 rays sometimes are present), and the dorsal fin has 28 or 29 rays. The caudal fin is frayed in the specimens at hand, and its exact shape cannot be determined. However, it evidently is quite long and pointed. The ventral fins are small, scarcely as long as the eye. Pre- served specimens are very plain in color. A few dusky markings are evident along the ventral outline of the caudal peduncle, and an elongate dusky spot is present at the base of the anterior rays of the anal and similar spots at the end of that fin. A dusky spot is situated just in advance of the spinous dorsal, another one is under the base of the spinous dorsal, a third one is situated under the anterior rays of the soft dorsal, and a fourth one is near the end of the base of this fin. The jaws ante- riorly are slightly dusky (fig. 42). This species is closely related to C. regalis from which it is distinguished, at the size described, chiefly by the shorter anal fin, which has 9 (sometimes 8) rays, where- REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 113 as C. regalis has 11 to 13. According to the preserved specimens at hand, C. nothus is somewhat plainer in color and some of the markings present are a little differently placed. C. regalis at a length of 9 to 11 millimeters has numerous dusky dots on the side, which C. nothus does not possess. C. regalis generally has more numerous spots on the back also. On the other hand, C. nothus has two elongate dusky spots at the base of the anal, one being below the anterior rays and the other one at the base of the last rays, whereas only one spot placed near the middle of the anal base is present in C. regalis. The specimens described are the smallest ones of the species recognized. It is possible that this species and C. regalis are so similar in the younger stages that they were not recognized as distinct. Identification apparently would be somewhat un- certain until the anal fin is developed sufficiently to permit the enumeration of the rays. However, it is thought that the markings on the ventral surface of the tail might be useful in separating the species, since the spots appear to be a little dif- ferently placed in the two species, as already stated. In the younger specimens the markings are quite uniformly placed, and the large spot near midcaudal length (being the one that is situated below the base of the anal when that member be- comes developed) is steadfastly present and is thought to be diagnostic for C. regalis. Figure 43.— Cynoscion nothus. From a specimen 26 millimeters long. Specimens 24 to 26 millimeters long. — The fish has become more shapely since a length of 9 to 11 millimeters was attained (no specimens of intermediate sizes are at hand), that is, the head and trunk are not as disproportionately deep in comparison with the tail as in the smaller fish. The greatest depth is contained 3.2 to 3.6 times in the standard length, and the head now is considerably longer than deep, its length being contained 2.5 to 2.7 times in the standard length. The snout is a little sharper than in smaller fish; its length remains about equal to the diameter of the eye, and is contained 3.7 to 4.3 times in the head. The mouth is moderately oblique; the lower jaw projects little; and the maxillary reaches to or a little beyond the pos- terior margin of the pupil, 2.0 to 2.2 in the head. Gill rakers are developed, 11 were counted on the lower limb of the first arch in each of two specimens. Scales are evident on the side, though the sides are not completely covered. The caudal fin is sharply pointed and the middle rays are much longer than the head. The ventral fins have increased greatly in length and are about twice as long as the eye. Dark color markings have become much more numerous; the markings (chromatophores) mostly remaining small and rather scattered, and are present on the sides, on the head, on the back, and along the ventral outline. In some specimens congregations of dark marks are present along the lateral line, forming indefinite quadrate spots. In some specimens a concentration of dark markings, also, has occurred on the back 114 BULLETIN OF THE BUREAU OF FISHERIES where they form saddle-like blotches. A dark cross line is present on the base of the caudal fin and a few dark dots have appeared on the spinous dorsal (fig. 43). This species is very close to C. regalis, differing (in addition to having fewer anal rays) chiefly in color. In general, the individual markings (chromatophores) are smaller and form less definite blotches. Furthermore, blotches on the back and in the lateral line, if present, are separate, whereas in C. regalis, blotches somewhat similarly placed, are connected and form more or less definite crossbands, at least on the anterior part of the body. The indications are that C. nothus has the middle rays of the caudal more strongly produced, though this cannot be determined defi- nitely because this fin is more or less damaged in all the specimens at hand. Specimens 40 to 46 millimeters long. — The body has become deeper posterior to the ventral fins, and the head is rather more pointed. The proportions, measured as in the smaller specimens, have not changed greatly. The head is contained 2.5 to 2.75 in the standard length, and the depth 3.2 to 3.6 times. The eye and snout are of about equal length and are contained 3.7 to 4.3 in the head. The mouth remains quite oblique, the gape anteriorly being somewhat above the lower margin of the eye, and the maxillary reaches about opposite the posterior margin of the pupil, it being contained 1.9 to 2.2 in the head. The body is fully covered with scales and small scales also are evident on the base of the soft dorsal and on the anal. The caudal fin remains long and pointed, the middle rays being much produced and nearly an eye’s diameter longer than the head. The body has become more definitely blotched, the blotches consisting chiefly of a series along the lateral line and another on the back. In the majority of specimens examined the blotches are separate, but in some examples the ones on the anterior part of the back tend to unite with those on the side to form cross bars. Dark dots are present on both dorsals, the caudal, and occasionally on the anal. This species continues to differ from C. regalis in color, the body generally being blotched, or if suggestions of crossbars occur they are present anteriorly only, and are less distinct than in C. regalis. The caudal fin is notably longer and more sharply pointed in C. nothus, the longest rays being nearly an eye’s diameter longer than the head, whereas in C. regalis none of the rays exceed the length of the head. In the proportional measurements given in the foregoing description the two species are almost identical at a range in length of 40 to 46 millimeters. Specimens 75 millimeters and upward in length. — The depth is contained 3.4 to 4.0 times in the standard length in specimens 75 to 90 millimeters long, and these proportions prevail in adult fish, also. However, the body remains notably more compressed, its greatest width being somewhat less than the length of the maxillary. The fish apparently increase in robustness very gradually as long as growth continues. At a length of 150 millimeters the width of the body is about equal to the length of the maxillary, and in larger fish the width exceeds the length of the maxillary. The head is contained 2.9 to 3.0 times in the length in fish 75 to 90 millimeters long, as compared with 3.2 to 3.4 in specimens ranging from 150 to 200 millimeters in length. Therefore, the head remains proportionately longer, as usual, in the smaller fish. The mouth virtually has attained the position and nearly the pro- portions of the adult in specimens only about 40 millimeters long, described in the foregoing section, and the two large recurved canines in the upper jaw, character- istic of this genus, slightly evident at a length of 45 millimeters, are prominent in specimens 75 millimeters long. Small scales are present on the basal portion of the fins, exclusive of the spinous dorsal, at a length of 75 millimeters (and earlier in some REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 115 specimens), and they are extended almost all over the fins by the time the fish reach a length of about 125 millimeters, apparently increasing in density with age. The caudal fin remains pointed in specimens 75 millimeters long and the longest rays (although proportionately shorter than in smaller fish) are about equal to the length of the head. This fin is still moderately pointed in fish 150 millimeters long, but the longest rays then are notably shorter than the head. The caudal does not acquire the shape it has in adults, that is, with the upper lobe slightly concave and the lower one rather sharply rounded, until the fish reach a length of about 200 millimeters.17 The blotches present in the smaller fish described in the preceding section become less distinct at a length of about 60 millimeters. In the preserved specimens exam- ined only traces are left in fish 75 millimeters long, and soon afterward they disappear and the fish are plain grayish above and silvery below. Only large specimens (of the preserved material examined) 200 millimeters and upward in length have traces of oblique lines running along the rows of scales above the lateral line. The dark cross line on the base of the caudal, present in small fish, has disappeared in specimens about 60 millimeters long. Dark dots on the dorsal and caudal fins increase rapidly in number as growth proceeds. In specimens about 100 millimeters long the caudal fin is quite dusky and distally almost black. The dorsal fins, too, are densely dotted and soon become dusky in color, and the spinous dorsal has a black margin (fig. 44). This species and C. regalis resemble each other closely. Beside the difference in the number of anal rays, the slight difference in the number of gill rakers, and vertebrae, pointed out elsewhere, it is evident that almost throughout life (that is, after the caudal fin becomes developed) C. nothus has a longer and more sharply pointed caudal fin, which never becomes truncate as in C. regalis. Although the fins, exclusive of the spinous dorsal, become covered with small scales in both species, those of C. nothus appear to be more densely scaled in adult specimens. DISTRIBUTION OF THE YOUNG It has been pointed out elsewhere (p. Ill) that no young under 9.5 millimeters in length were taken. Therefore, the early stages remain unknown. It has been stated, also (p. 1 12),, that all the young collected, exclusive of four specimens, were caught in the The shape of the caudal, because the fin was frayed in the specimen drawn, is incorrectly shown as round for the adult in Hilde- brand and Schroeder (1928, p. 299, fig. 175). It is correctly indicated in Ginsburg’s figure (1929, p. 81, fig. 5), which shows the upper lobe to be slightly concave and the lower one rather sharply rounded. 116 BULLETIN OF THE BUREAU OF FISHERIES general vicinity (off Beaufort Inlet) inhabited by the adults during the summer. It is quite certain, therefore, that at least the larger young occupy the same grounds with the adults. Both young and adults were taken only with an especially adapted otter trawl, which, of course, was hauled on the bottom. If the larvae of this species behave like those of the gray and the spotted trout, as would be expected, they may occur at the surface when very small, but soon descend to the bottom. GROWTH An insufficient number of young fish was caught to determine accurately from length measurements the rate of growth of the 0-class. Specimens that definitely belong to this year class were taken only during August, September, October, and November. In August 54 young, ranging in length from 9 to 88 millimeters, with an average length of 61.1 millimeters, were measured. In September only 15 speci- mens, 31 to 95 millimeters in length, averaging 57.2 millimeters, were measured. The smaller young obviously are missing in the collection for September, as the fish caught during October consisting of 179 specimens, contain fish ranging in length from 17 to 80 millimeters, averaging only 46.5 millimeters in length. The larger young of the 0-class obviously are missing in the collections made in October. In November, only, a sufficient number of specimens was taken and measured to give reliable information. The 486 specimens of the 0-class measured, range in length from 19 to 147 millimeters, and have an average length of 75.8 millimeters. Since this fish quite certainly is a smaller species than the other local species of this genus (as pointed out on p. 1 1 1) a slower rate of growth would be expected, and that is what the limited number of measurements appears to indicate. Instead of attaining a length of about 170 millimeters (6% inches) during the first 7 or 8 months of life, as in C. regalis, and probably in C. nebulosus, this species appears to reach a length of only about 75 millimeters (3 inches). The size at which sexual maturity is reached remains unknown. BIBLIOGRAPHY Coles, Russell J. 1916. Is Cynoscion nothus an abnormal regalis ? Copeia, no. 30, April 24, 1916, pp. 30-31. New York. Delsman, H. C. 1931. Fish eggs and larvae from the Java Sea. Treubia, vol. XIII, no. 3-4 December 1931, pp. 401-410, 11 figs. Buitenzorg. Java. Eigenmann, Carl H. 1901. Investigations into the history of the young squeteague. Bull., U.S. Fish Com., vol. XXI, 1901 (1902), pp. 45-51, 9 figs. Washington. Ginsburg, Isaac. 1929. Review of weakfishes (Cynoscion) of the Atlantic and Gulf coasts of the United States, with a description of a new species. Bull., U.S. Bur. Fish., vol. XLV, 1929 (1930), pp. 71-85, 7 figs. Washington. Higgins, Elmer, and John C. Pearson. 1927. Examination of the summer fisheries of Pamlico and Core Sounds, N.C., with special reference to the destruction of undersized fish and the protection of the gray trout, Cynoscion regalis (Bloch and Schneider). Appendix II, Report, U.S. Com. Fish., 1927 (1928), pp. 29-65, 15 figs. Washington. Hildebrand, Samuel F., and Louella E. Cable. 1930. Development and life history of four- teen teleostean fishes at Beaufort, N.C. Bull., U.S. Bur. Fish., vol. XLVI, 1930, pp. 383-488, 101 figs. Washington. Hildebrand, Samuel F., and William C. Schroeder. 1928. Fishes of Chesapeake Bay. Bull., U.S. Bur. Fish., vol. XLIII, part 1, 1927 (1928), 388 pages, 211 figs. Washington. Kuntz, Albert. 1914. The embryology and larval development of Bairdiella chrysura and Anchovia mitchelli. Bull., U.S. Bur. Fish., vol. XXXIII, 1913 (1915), pp. 1-19, 46 figs. Washington. REPRODUCTION AND DEVELOPMENT OF SCIAENIDAE 117 Pearson, John C. 1929. Natural history and conservation of the redfish and other commercial Sciaenids on the Texas coast. Bull., U.S. Bur. Fish., voi. XLIV, 1928 (1929), pp. 129-214, 44 figs. Washington. Smith, Hugh M. 1907. The fishes of North Carolina. North Carolina Geol. and Econ. Survey, vol. II, 1907, xi, 453 pp., 21 pis., 187 figs. Raleigh. Welsh, W. W., and C. M. Breder, Jr. 1923. Contributions to the life histories of Sciaenidae of the eastern United States coast. Bull., U.S. Bur. Fish., vol. XXXIX, 1923-24 (1924), pp. 141-201, 60 figs. Washington. U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner RACES OF HERRING, CLUPEA PALLASII IN SOUTHEASTERN ALASKA By GEORGE A. ROUNSEFELL and EDWIN H. DAHLGREN From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 17 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON :'1935 For sale by the Superintendent of Documents, Washington, D. C. Price 10 cents RACES OF HERRING, CLUPEA PALLASU, IN SOUTHEASTERN ALASKA1 & By George A. Rounsefell, Ph. D., Associate Aquatic Biologist, and Edwin H. Dahlgren, Junior Aquatic Biologist, United States Bureau of Fisheries * CONTENTS Fast Introduction 119 Spawning and feeding localities 120 Analysis of vertebral counts 123 Discussion of factors influencing vertebral count distribution within a population 123 Existence of races proven by heterogeneity of samples from all localities 124 Homogeneity of material from individual localities 129 Segregation of races 129 Analysis of growth rates 133 Analysis of year classes 138 Tagging 138 Summary 140 Literature cited 140 INTRODUCTION The reasons underlying this attempt to study the individuality and distribution of each population of herring are many, and for the most part, rather obvious. When a locality where herring have been abundant fails to produce its wonted supply, a question always arises as to the causes of such a failure. Aside from natural fluc- tuations in abundance or unusual unavailability to the fishermen, the apparent causes are migration or depletion. Without an intimate knowledge of the herring stocks either explanation is possible. For example, Whale Bay (Rounsefell, 1930, p. 238) on the outer coast of Baranof Island, produced a tremendous run of herring in 1925 but failed the following year. No herring have been caught in this bay since that time. It is now practically certain that this temporary run was caused by the summer herring schools that normally congregate about Cape Ommaney shifting farther north. On the other hand, the failure of the once important fishery in Kootznahoo Inlet, on Admiralty Island, may be fairly ascribed to depletion of the stocks from overfishing (Rounsefell, 1931, p. 35-36). Only by an intimate knowledge of the areas inhabited by each stock of herring is it possible to know whether fishing in one area is affecting the supply in another. It may now be said with certainty, for instance, that the heavy fishing in lower Chatham Strait can in no wise be blamed for the scarcity of herring tvat has been noticeable for several years in the “inside” waters of Behm Canal, Ernest Sound, Zimovia Strait, and upper Frederick Sound (Rounsefell, 1930, p. 236, 237 and 307; > Bulletin no. 17. Approved for publication, Oct. 4, 1934. 119 120 BULLETIN OF THE BUREAU OF FISHERIES Rounsefell, 1931, p. 35-36). The herring of these “inside” waters belong to popula- tions quite distinct from those of Chatham Strait. It is well known that different populations may exhibit different structural peculiarities owing to differences in environment or to differences in heredity arising during long periods of isolation. The study of the individuality of the populations has been based largely on these structural differences. Whether the differences in the characters chosen are due to heredity or to environment has not been considered as being of great importance, as long as the characters are fairly stable within each population so that significant differences indicate very slight intermingling, if any, between adjacent stocks of herring. Success has finally been achieved for the direct method of tracing migrations by the release and recovery of tagged herring (Rounsefell and Dahlgren, 1933). This method may be called the direct method of racial investigation in contradis- tinction with the indirect method in which the movements or lack of movement of a population are inferred from the statistical analysis of morphological characters. Owing to the newness of this method which was first attempted in 1932, only a few results have been attained. Yet these few results offer such corroboration of our racial work as to inspire confidence in our results. In the determination of the individuality of populations by indirect methods it was deemed advisable, profiting by the experience gained in the preliminary racial work (Rounsefell, 1930), to concentrate on vertebral counts. This was the more necessary, owing to the difficulty of securing enough samples of fresh herring from various localities in the nearly perfect condition necessary for body measurements. The rates of growth and relative abundance of year classes have also been employed as indicators of populational differences. SPAWNING AND FEEDING LOCALITIES At the present time there are 3, or perhaps 4, spawning areas in southeastern Alaska, where the herring may be counted upon to spawn in abundance each spring. (See fig. 1.) Of these, the spawning grounds in Sitka Sound, on the outside of Baranof Island are probably the largest. Those at the entrance to Klawak Inlet in San Alberto Bay are undoubtedly a close second, and those centering near Juneau in Stephens Passage are easily third. The spawning grounds in Kootznahoo Inlet were once of great importance but have declined. As indicated in figure 1, there are a number of minor spawning grounds, a few of which were considered of importance in the past. There are certainly additional localities, not noted, where a few herring occasionally spawn. It may be of interest to note here the distances between the four major spawning grounds. These distances, measured approximately from the centers of spawning, are as follows: Sitka to Craig (Klawak Inlet), 120 miles; Sitka to Kootznahoo Inlet, 70 miles by Peril Strait; Sitka to Kootznahoo Inlet, 140 miles by Cape Ommaney; Sitka to Juneau, 150 miles by Icy Strait; Sitka to Juneau, 120 miles by Peril Strait; Juneau to Kootznahoo Inlet, 65 miles; Craig to Kootznahoo Inlet, 140 miles; and Craig to Juneau, 210 miles. It is difficult to theorize as to the significance of the considerable distance between any major spawning grounds. It may mean that some of the minor spawning grounds are used merely by occasional schools straying from the main body of herring of any particular race. On the other hand it may indicate that there are two kinds or types RACES OF HERRING, SOUTHEASTERN ALASKA 121 of herring populations. One type would be those races in- habiting the major spawning grounds, and, by inference, some of the minor spawning grounds of importance in the past. This type of race might be rather migratory in its hab- its, thus accounting for the distances between major spawning grounds. Many of the minor spawning grounds might then be inhabited by herring of local character, rather nonmigratory in habits. Such a type is suggested by the herring found spawning at the head of Gut Bay (see fig. 2) in the middle of June, at least 6 weeks after the normal cessation of spawning at Craig and Sitka. Such small bodies of herring might seldom stray from a single inlet or fiord. Possibly a herring population may change gradually from a nonmigratory to a migratory habit, or vice versa, accord- ing to its abundance, as spatial considerations are known to affect the migrations of mam- mals. There are two quite dis- tinct herring fisheries in south- eastern Alas k a — the minor fishery for halibut bait, carried on during the halibut fishing season throughout the whole area; and, secondly, the major fishery of the herring plants which produces salt herring and fish meal and oil. The second fishery operates only from June 1 to September 30 and its fishing operations are confined to the western por- tion of southeastern Alaska. Figure 1.— Southeastern Alaska, showing spawning grounds and feeding grounds of herring. Cross-hatched areas indicate spawning grounds; horizontally barred areas indicate feeding grounds of importance to summer fishery of salt herring and fish oil and meal industry. The importance of both types of areas is roughly proportional to size of circles. Black dots show location of 7 herring plants operating in 1934; small circles give location of 4 herring plants remaining inactive in 1934. 122 BULLETIN OF THE BUREAU OF FISHERIES * £ i,w §>. I S I OT ft O X3 rl M h (5) u o P D u. O « 'O a ^ a "35056 . « g 3 3 3 *-• £ •a .2 cq o a> '3 ^ ° ft P ft a " as ® 2 2 .a a -g ’3 'C ^ »3 ,H 5 ^ N co c - co a ■'T k-* 03 ^ ca x] •- O ti a O' QJ 5 - „ J3 H§0 © ° ~- % aj ■ 55 3 i o O S g >. a O 8s oaf 3 ~ ^ " ■" o s 0 *'1' o g-3! o g< © © g p= *3 TO n o g 5 {- fl a> *- - ft P _ b£ K. ft 5 L* cO w © P t 8 3 iwg & ■ 3 O ft +0 o £ 2 >> ari v/J 4J z; '•'« -*-* O C3 . „ o .5 3 8 o o ■o ° « “ a 3 « a M o 2 3 T « g o 3 w « "3 (£ 03 *" rrt ft P ^ ® £ I ° | 5 -g 8 3 O C w ~ ? 'S M ® a A 5.®5o|fe u. P ,9 □ ® „ .. <~> © o © »•« S 6ft £ o -a a H - — P a.' to P .;=; © © © « £ £ < P © ft P P T3 ^ - P W p 'S ^ > of o - “s « ^ S 5 ^ co -d S a> . . U M ” Is g ° Issa^isr.sgfc. 'Qp^ofl®2”^2g = ^ > § 2 ’3 o p ® a 2 -2 w .S-S*** 8 jfgS^f o o p 3 to CO i - £ co 5 § ’1 1 £ bfl t3 S c 2 C3 ft "3 J3 ° '"' « Id - >,33 !>> 2 * -3 ■e a a " - o 3 3 rn O g « m & £ g « > 03 U3 C3 . - a « ■“ bC « M o QU v, n rj ® g £ W 51 O n *y £ 2 2 ” p ■« . ® '3 5 rt a £ O W •9 - ,® ft 03 © Ph I .5 £ .ts . e-i '-' s S M * tn S’ . ® 3 ^ ft P< o ce 0 O ■B 3 g -s .3 « o c M ® - ■ I— ►> P o ^ O a> uj -O 10 P . o 5 p . „ o : 02 3 c TO ® *3 — 1 »>» o >» . 71 5 » o ^ p a ^ a .P ° Tg „ ^•ol C^j « tJ3 TO T3 .a « > TO O P

> •3W^-S I S § 0 P tg 6 c0 P ® C/2 P 2 O w ‘En ® w o W M TO t>» , ^ § ! to a ►> o ^ ^ o ® .2 © o o $ fc b3 ^ g w -g w o ® P Ph S ScoW -M- O fl ^ ' !S ^ ® l; ® -0 cO P p W” c3 « c0 TO 2 o CO (1, '|H O- (L m o «- o SaS^3_ fe P > w 71 ‘rr ! « _ a ft- . § £ a 2 Jj o “ “ CO lis -9 Ph ■ p jo CO ® >— » -*-3 |^g 0 fe g ” S CO o 1 S' M "- -tj . . cO - « g';£ tn □ C3 -t-< " -a® •t= a to TO P-< Q TO M c0 p OT s O Jg P 2-^2 o *f 03 a § 2 co M O I § co ^ E- *j ® * g 'E ® g o •a S 3 3 § Q a = “O o 3 " a a 3 .2 ® w> o* " 0 .a © 2 S3 ft - hfl ^ S C I OT I © a w a § 1 >» to _p ; CO CO cO CO M o O « s ft- s I® I E3 ^ >> 3" & 11 = -3 >, W 3 a a ° a S g d ^ c a •; o ■= a -3 ? ^ H W > CO CC hP CO hH CO ,1 RACES OF HERRING, SOUTHEASTERN ALASKA 123 This fishery has not in recent years been conducted eastward of a line from Juneau to Klawak, or south of Noyes Island. So closely is this fishery identified with Chatham Strait, the great waterway that extends with its continuation, Lynn Canal, for two- thirds the length of southeastern Alaska, that if a straight line is drawn down Chatham Strait and extended southward it will be noted that the summer fish- ery operates on both sides of this line from Noyes Island to Juneau, a distance of 180 miles, yet practically the entire catch is made within 35 miles of this line, the only exception being occasional fishing in Sitka Sound 65 miles from the line through Peril Strait. These summer feeding grounds fished by the herring plants are shown in figure 1. The importance of each to the fishery is roughly indicated by the size of the circles. The most important fishing ground, by far, is the area surrounding Cape Ommaney, the southern tip of Baranof Island. This is due largely to the abundance of herring around the cape, but also, in some measure, to the proximity of the herring plants. Of the herring plants now in operation the one farthest from the cape is that at Wash- ington Bay, on Kuiu Island, 35 miles distant; all of the others are on the eastern side of Baranof Island, 25 miles being the greatest distance any of them are from Cape Ommaney. (See fig. 1.) The area along the north shore of Noyes Island, including the waters surrounding the Maurelle Islands, is another great herring feeding ground. Tebenkof Bay, Coronation Island, and Warren Island also contribute a share of the c.atca, but their importance fluctuates, some years being practically blanks (Rounsefell, 1931). The feeding grounds at the juncture of Frederick Sound and Chatham Strait were once important fishing grounds but have declined tremendously (Rounsefell, 1931, pp. 33 and 34) and are now of minor importance. The feeding grounds in Icy Strait and near Juneau in Stephens Passage are heavy producers of herring on occasion, but are too distant from the plants to warrant fishing when herring are abundant elsewhere. Analyses of the catch statistics to determine the relative abundance on these various spawning and feeding grounds cannot be accurately made without knowledge of the interrelationships of the populations inhabiting different areas. The next section takes up a discussion of the methods of determining the individuality of these populations by means of the vertebral count. ANALYSIS OF VERTEBRAL COUNTS DISCUSSION OF FACTORS INFLUENCING VERTEBRAL COUNT DISTRIBUTIONS WITHIN A POPULATION In an earlier report on the herring of Prince William Sound (Rounsefell and Dahl- gren, 1932) it was shown that a high negative correlation (in that case —0.85) exists between the average temperature during the developmental period and the average number of vertebrae in different year classes or “brood years” of herring from the same locality. Comparison of the means of samples of herring without division into year classes is thus shown to introduce variation other than that expected in random sampling. Therefore all of our samples have been divided into year classes so that only vertebral counts of herring hatched during the same spring are compared. The vertebral count of samples of a year class caught in any one year could not be compared with samples of the same year class taken during ensuing years without showing that there was not, due to selection, a tendency for the mean verte- bral count to rise or fall with advancing age. To determine this point, samples of 124 BULLETIN OF THE BUREAU OF FISHERIES the 1926 and 1927 year classes from eight localities, caught in their fourth summer, were compared with samples of the same year classes from the same localities taken during their fifth summer (neglecting samples of less than 25). Between the two series the average difference in vertebral count was 0.042, which was obviously of no significance as the chances were 1 in 8 that the means were the same.* 1 2 (See table 1.) Having thus failed to note any correlation between age and number of vertebrae in our samples, the means of the 4-year-olds of each year class were compared with the means of the 5-year-olds for the same year classes in each of the eight localities. Of these eight comparisons, none showed a significant difference,3 although one might be regarded as doubtful, the chances being 1 in 25 that the two means are the same. Table 1. — Means of vertebral counts of 1926 and 1927 year classes from various localities compared at 4 and at 5 years of age Locality Year class Mean Differ- ence Locality Year class Mean Differ- ence Age 4 Age 5 Age 4 Age 5 Inside Cape Ommaney Larch Bay Do 1926 1927 1926 1926 1926 52.411 52. 573 52. 478 52. 404 52. 472 52. 408 52. 750 52. 433 52. 482 52. 464 0. 003 -. 177 .045 -. 078 .008 Hoonah Petersburg Do. 1926 1927 1926 52. 364 52. 488 52. 147 52. 474 52. 408 52. 253 -0. 110 .080 -.106 Coronation Island Average.. 52. 417 52. 459 -.042 To determine if sex had any effect on the number of vertebrae, the mean vertebral count for both sexes was determined for each of a series of 24 samples of the 1926 year class caught during the summer of 1930 at Larch Bay, containing 491 males and 493 females. The mean of the 24 unweighted means for males was 52.431; for females, 52.435. The difference between these means was of no statistical significance. In making this comparison, the means of the males and of the females were not weighted because the presence of more than one population amongst the samples would then cause a weighting of the data according to the number of individuals in the samples. EXISTENCE OF RACES PROVEN BY HETEROGENEITY OF SAMPLES FROM ALL LOCALITIES Proof of the existence of independent stocks of herring is furnished by an analysis of the averages of all samples of vertebral counts of herring of the 1926 year class in southeastern Alaska. (See table 3 for total samples of each locality.) The object of this analysis is to prove whether or not all of the samples could have been drawn from the same population. From localities where many samples are available some of the samples are statistically different from others. This is to be expected according to the laws of probability. Before comparing the samples of one locality with those from another, it is essential that it be known whether any differences found may be due merely to such expected random variation or may be ascribed to a difference between the populations from which the samples are drawn. Therefore, it was neces- sary to test the data as a whole to determine if all of the samples could have been drawn from one population. 2 Quoting from Fisher (1930, p. Ill), “In cases in which each observation of one series corresponds in some respects to a particu- lar observation of the second series, it is always legitimate to take the differences and test them * * In this method the test shows whether the mean difference differs significantly from 0, which is taken successively as each mean of the first series. t is the mean difference divided by its standard error, which in this case was 1.54 which yields a probability of 0.12 or approximately 1 to 8. 2 This method of comparison is explained by Fisher (1930, p. 107), see section on “ Segregation of races.” Probability was 0.04 or 1 to 2§. RACES OF HERRING, SOUTHEASTERN ALASKA 125 The method used in testing the homogeneity of the means of all of the samples is merely an extension of the method of comparison of two means to the comparison of several means. This method is called the “Analysis of variance” by Fisher (1930, p. 196). Wollaston (1933) in an article in the Journal du Conseil, with a foreword by R. A. Fisher, expounds the use of this method and its applicability to herring race problems. Quoting from Wollaston: There are two fundamentally different ways of approaching observational scientific data. The first is to lay out the data as graphically as possible and see what they suggest; the second, to formu- late, without examining the data in detail, hypotheses which the data may be expected to prove or disprove, and then to test the agreement of the data with the hypotheses. The first result remains but a suggestion, and the actuality of the suggested phenomena cannot be stated in terms of proba- bility. The second allows definite statement of probability that the hypotheses are true or not true. The great majority of fishery workers, including Dr. Schnakenbeck in his work criticized in this paper, have adopted the first way. His conclusions may be right, but his method of approach, which I will call the a posteriori method, includes no test whatever as to the probability of his being right. ******* Sound statistical tests of probability can only be applied to data treated in the second way. It is even better to formulate hypotheses to be tested by the data before these are collected than to do so before they are worked up. The research can then be given the exactness of pure experimental science, giving equally definite positive or negative results. ******* The main object of this paper is to introduce into fishery research some of the most important methods developed by R. A. Fisher, of Rothamsted Laboratory. These are offered as alternative and far preferable to empirical methods which take no account of the variability which occurs be- tween samples drawn from a larger population. For the purpose in question I have used the data collected by Dr. Schnakenbeck from the North Sea, and I propose to show that Dr. Fisher’s methods are perfectly adequate to deal with such data and to extract all the information from them which they are capable of giving. As I have not had access to subsidiary data, collected by other workers and used by Dr. Schnakenbeck in his report, it cannot be said definitely whether these would have modified my conclusions and brought them more nearly into line with Schnakenbeck ’s. It is hoped that all available data bearing on the herring race question will eventually be combined in a com- plete statistical analysis on the lines laid down here. This must be considered merely as an introduc- tion to such analysis. ******* Though the mathematical theory on which the present paper is founded is somewhat advanced, the methods introduced herein are very simple in application. The first part deals almost exclu- sively with Fisher’s methods for finding the best-fitting Curve of Error to fit to highly grouped data. Readers who have not to deal with such data may prefer to omit this part. The second part (from p. 23 on) deals with Fisher’s method of the Analysis of Variance, which is of almost universal application in testing the significance of variations in any phenomenon under different conditions. There is no other method so ideally fitted for this kind of test. This second part will be therefore found worth reading by anyone who is engaged in fishery research and who is not familiar with Fisher’s work. Every step in the application of the method to the present prob- lem is shown in detail, and described as far as possible in nonmathematical language. ******* We have then an ideal set of conditions for the application of Fisher’s Analysis of Variance,4 which is a powerful weapon for distinguishing between real differences between samples and those which are probably due to variation “within samples.” « The term “Analysis of Variance” is somewhat misleading. It is the total sum of squares which is analyzed. If, however, each estimate of variance is considered as weighted by its degrees of freedom, the term Analysis of Variance is quite correct, as will be seen later. 103465°— 35 2 126 BULLETIN OF THE BUREAU OF FISHERIES This is not the place to give a full description of the Analysis of Variance, but I propose to discuss shortly in nonmathematical language the assumptions on which it is founded, since I did inot myself find Fisher’s own description (Fisher (4)) very easy to follow, nor do I consider that the ogical bases of the method were sufficiently emphasized in the work cited. Supposing then for the moment that every single one of our vertebral counts is a sample from a strictly normal population of vertebral counts, we can take at random from the whole set of counts any given number of counts n/, and calculate for the set the mean number £i, and the variance, S(xi— x{)2 wjiere ni = ni'_ i. This variance, which we will call SI, is an estimate of the variance a2 «1 of the whole population, founded on the number of degrees of freedom n\. The term degrees of freedom means the number of ways in which the frequencies of the counts may be varied at will, provided that given fixed relations between the data are adhered to. In calculating S\ we have used xi as an estimate of the true mean, x\ being calculated from the data themselves. As £h or the sum of all the n/ z/s, is fixed, only n\ — 1 of them may be varied at will, the last being fixed by the sum of the n/ — 1, whatever its value. We can take any other random sample of n'T counts and obtain another estimate s2r of the variance a 2, based on nr degrees of freedom. We can also obtain s2, from our total set of N' counts and this is an estimate of a 2 based onV-1 degrees of freedom. All these values of s2 may be shown to be efficient estimates of 47 48 49 50 51 52 53 54 55 1923 1 2 4 75 446 334 45 2 909 52. 370 541. 802 1923 1 21 38 8 68 52. 779 29. 691 1924 1 7 31 31 4 74 52. 392 61. 635 1924 1 12 32 13 2 60 52. 050 36. 850 1924 2 24 33 3 62 52. 597 24. 919 1924 7 51 24 5 87 52.311 42. 621 1925 1 1 15 61 54 4 1 137 52. 328 92. 219 1925 16 61 43 4 124 52. 282 65. 121 1925 1 27 104 50 5 187 52. 166 95. 861 1925 1 5 29 44 7 86 52. 570 85. 081 1925 25 42 3 70 52. 686 21. 086 1926 1 37 183 150 18 389 52. 378 207. 450 1926 9 53 40 2 104 52. 337 45. 221 1926 4 37 26 4 71 52. 423 33. 324 1926 8 27 18 2 55 52. 255 30. 436 1926 2 27 25 3 57 52. 509 24. 246 1926 10 89 81 9 189 52. 471 85. 090 1926 1 3 64 274 144 7 1 494 52. 178 250. 324 1926 3 37 67 35 1 143 51.958 87. 748 1926 1 8 37 36 4 1 87 52. 414 63. 103 1926 21 84 81 7 193 52. 384 101. 627 1926 9 48 26 3 86 52. 267 40. 849 1926 1 7 66 37 5 116 52. 328 55. 552 1926 4 36 44 84 52. 476 28. 952 1926 3 24 197 171 16 411 52. 421 198. ISO 1926 3 10 141 1,013 877 97 2 2, 143 52. 423 1, 107. 123 1926 22 221 186 19 1 449 52. 457 199. 403 1926 1 1 1 21 190 145 8 367 52. 357 156. 371 1926 1 9 36 27 1 74 52. 230 45. 095 1926 1 25 100 58 2 186 52. 188 88. 414 1926 8 34 11 1 54 52. Ill 27. 333 1926 2 20 25 4 51 52. 608 24. 157 1927 1 1 11 77 64 11 1 166 52. 440 108. 898 1927 6 28 26 7 1 68 52. 544 48. 868 1927 1 3 20 97 83 15 1 220 52. 395 158. 595 1927 1 7 32 27 1 68 52. 294 36. 118 Todd 1927 1 13 75 62 9 160 52. 394 102. 194 1927 1 1 29 245 212 36 2 526 52. 487 291. 407 1927 6 30 23 5 64 52. 422 37. 609 1927 2 29 24 4 59 52. 508 26. 746 1927 13 23 18 6 60 52. 283 50. 183 1927 1 5 46 21 5 78 52. 295 48 218 1927 1 n 134 193 22 1 362 52. 627 162. 655 1927 i 37 41 3 1 83 52. 590 34. 072 1927 1 4 27 21 2 55 52. 346 30. 436 1927 17 104 81 3 205 52. 342 86 098 1927 6 33 19 i 59 52. 254 25. 186 1928 2 24 47 a 84 52. 798 39. 560 1928 20 113 63 4 200 52. 255 85. 995 1928 13 80 40 4 137 52. 255 60. 058 1928 11 71 32 2 116 52. 216 45 612 1928 4 26 24 4 58 52. 483 30. 483 1929 10 28 26 64 52. 250 32. 000 1929 1 16 107 63 7 194 52. 304 93. 057 i Excluding the hypural. In using this method, as pointed out by Wollaston, the distribution of the variances of the samples must approach normality. Accordingly, before analyzing these samples their variances were tabulated, table 4, and the 4 samples indicated were discarded, their variances obviously being far outside of the normal range. A test of the homogeneity of the remaining 158 samples, comprising 5,964 vertebral counts showed definitely that these samples could not all have been drawn from the same population, since the observed z of 0.3387 far exceeds 0.1256, the value of z calculated at a probability of 0.01. Therefore, it must be concluded that the herring of southeastern Alaska are composed of more than one population. RACES OF HERRING, SOUTHEASTERN ALASKA 129 Table 4. — Frequency of variances of 162 individual samples of the vertebral count from various localities in southeastern Alaska Variance Number Variance Number Variance Number Variance Number 0.1900-0.2300 .. . 1 0. 5600-0. 6300 ... 19 0. 9600-1 . 0300 1.3600-1. 4300 l 2 0. 2400-0. 3100 16 0. 6400-0. 7100 15 1. 0100-1. 1100 1. 4400-1.5100 i 1 0 3200-0. 3900 25 0. 7200-0. 7900 9 1. 1200-1. 1900 1. 5200-1. 6000 0. 4000-0. 4700 38 0 8000-0. 8700 3 1.2000-1. 2700 1 1 0. 4800-0. 5500 30 0. 8800-0. 9500 2 l. 2800-1. 3500 i These samples omitted in analysis of variance. HOMOGENEITY OF MATERIAL FROM INDIVIDUAL LOCALITIES Evidence tending to prove a lack of admixture in the samples from the individual localities is shown in table 5, in which a test of the homogeneity of samples of the vertebral count of the 1926 year class has been made for each of the 7 localities from which sufficient data were available for such a test. In contrast to the results ob- tained when the southeastern Alaska material was considered as a whole, none of the 7 localities have an observed z exceeding the value of the calculated z at a prob- ability of 0.05, indicating the homogeneity of the population sampled in each locality. However, the observed z for the Warren Island samples, 0.2768, is only slightly less than that of the calculated, 0.3448, for a probability of 0.05. Therefore the Warren Island samples must be regarded with some suspicion, especially as the proportion the observed 2 forms of the calculated z is larger in the Warren Island samples than in those from the other localities. Table 5. — Analysis of variance of vertebral count samples of the 1926 year class from various localities in southeastern Alaska taken from 1929 to 1931, inclusive Number of samples Mean square Calculated z for prob ability of 0.05 ‘ Locality Mean Number Between arrays Within arrays Observed z 52. 423 2, 143 411 60 0. 6465 0. 5247 0. 1044 0. 1414 52. 421 19 .4215 .4862 .0714 .3279 52. 457 449 12 .3828 .4467 .0785 .4409 52. 357 367 8 .8397 .4829 .2768 .3448 62. 471 189 6 .6426 .4474 . 1810 .4052 52. 188 186 9 .5262 .4756 .0507 .3322 52. 378 389 n .6977 .5304 .1370 .3054 i The observed value of z is less than the calculated value at a probability of 0.05 in every case, thus showing that the popula- tion of each locality is homogeneous. SEGREGATION OF RACES Since the above evidence supports the hypothesis that the population of each locality is homogeneous, the vertebral counts from each locality have been compared to those of adjacent localities. (See fig. 3 and tables 3 and 6.) Only counts of fish of the same year class have been compared, necessitating the limiting of samples to those containing 50 or more counts. Smaller samples were not used, owing to the probability that occasional errors enter into our age determinations. PERCE /VP 130 BULLETIN OF THE BUREAU OF FISHERIES Figure 3. — Percentage vertebral count distributions ol 1926 and 1927 year classes from some of chief localities. PERCE, RACES OF HERRING, SOUTHEASTERN ALASKA 131 Table 6. — Comparisons of the means of the vertebral counts of each year class in southeastern Alaska Localities compared Year class Difference between means Summa- tion of popula- tions Standard error of difference between means Difference between means divided by standard error Juneau and Point Augusta 1927 0. 104 234 0. 119 0. 87 Do 1926 .041 493 .079 .52 Juneau and Hoonah 1927 045 Do. 1926 045 460 .094 .48 Juneau and Petersburg 1927 .047 692 .068 .69 Do 1926 .200 883 .049 1 4. 08 Do 1925 .046 261 .097 .47 Juneau and Point Gardner 1926 .093 578 .062 1. 50 Hoonah and Point Adolphus 1927 . 101 288 .114 .89 Do 1926 .168 126 .129 1.30 Hoonah and Point Augusta 1927 . 149 288 .118 1. 26 Do 1926 .086 175 . 104 .83 Favorite Bay and Point Augusta 1926 . 172 161 .109 1. 58 Favorite Bay and Point Gardner 1926 .038 246 . 101 .38 Todd and Point Augusta 1927 .150 228 .118 1.27 Todd and Gut Bay 1927 .099 238 .110 .90 Point Gardner and Petersburg 1926 .293 683 .060 l 4.88 Point Gardner and Meade Point 1928 .137 240 . 107 1. 28 Point Gardner and Deep Cove 1926 .204 275 .088 3 2. 32 Point Gardner and Tebenkof Bay 1926 .050 600 .060 .83 Meade Point and Petersburg 1926 .430 545 .104 ' 4. 13 Meade Point and Tebenkof Bay 1926 . 187 462 . 103 1. 82 Meade Point and Deep Cove 1926 .341 137 . 122 > 2. 80 Tebenkof Bay and Deep Cove 1926 . 154 497 .082 1.88 Tebenkof Bay and Port Herbert 1926 .093 527 .072 1.29 Tebenkof Bay and Big Port Walter 1926 .055 495 .081 .68 Tebenkof Bay and Cape Ommaney 1926 .002 Do 1925 .106 156 .134 .79 Tebenkof Bay and Coronation Island 1926 .036 860 .046 .78 Deep Cove and Port Herbert 1926 .061 202 .099 .62 Port Herbert and Gut Bay 1927 .012 138 . 146 .08 Port Herbert and Big Port Walter 1926 .148 200 .094 1. 57 Big Port Walter and Cape Ommaney 1926 .053 2,227 .079 .67 Cape Ommaney and Port Herbert 1927 .344 422 .098 i 3. 51 Do 1926 .095 2, 259 .068 1.40 Cape Ommaney and Deep Cove 1926 . 156 2,229 .079 2 1.97 Cape Ommaney and Coronation Island 1926 .034 2,590 .037 .92 Cape Ommaney and Warren Island 1927 .037 445 .080 .46 Do 1926 .066 2,510 .040 1. 65 Cape Ommaney and Meade Point 1926 .185 2, 194 .102 1.81 Cape Ommaney and Point Gardner 1926 .048 2,332 .053 .91 Cape Ommaney and Jamestown Bay 1926 .009 Cape Ommaney and Cape Edgecumbe 1926 .039 2,336 .053 .74 Cape Ommaney and Gut Bay 1927 .332 440 .087 i 3.82 Jamestown Bay and Cape Edgecumbe.. 1926 .030 280 .099 .30 Jamestown Bay and Favorite Bay 1926 .095 144 .134 .71 Jamestown Bay and Point Adolphus 1926 .159 142 . 141 1.13 Cape Edgecumbe and Favorite Bay 1926 .125 250 .107 1. 17 Cape Edgecumbe and Point Adolphus 1926 .129 248 . 112 1. 15 Coronation Island and Kell Bay. 1926 .227 523 .086 1 2.64 Coronation Island and Warren Island. 1926 .100 816 .047 2 2. 13 Coronation Island and Noyes Island 1926 .269 635 .059 I 4. 56 Warren Island and Kell Bay 1926 . 127 441 .086 1.48 Warren Island and Culebra Island 1927 .244 138 . 120 2 2.03 Warren Island and Noyes Island 1927 .248 288 .084 2 2.95 Do 1926 . 169 553 .060 1 2. 82 Warren Island and Wrangell 1927 . 178 147 . 117 1.52 Do 1926 .399 510 .068 i 5. 87 Warren Island and Meade Point 1926 .251 418 .103 2 2.44 Warren Island and Petersburg 1927 . 103 609 .086 1. 20 Do.... 1926 .179 861 .046 1 3.89 Culebra Island and Noyes Island... 1927 .004 260 .102 .04 Culebra Island and Klawak 1927 .092 114 .132 .70 1927 .088 264 .096 .92 Noyes Island and Port Estrella.. 1926 .077 240 .108 .71 Wrangell and Petersburg 1927 .065 590 .098 .66 Do 1926 .220 637 .069 i 3. 19 Do 1925 .116 311 .084 1.38 Wrangell and Anita Bay 1927 .086 123 .132 .65 1928 .000 Santa Ana Inlet and Frances Cove 1928 .039 253 .082 .48 > Statistically significant. 2 Approaching statistical significance. 132 BULLETIN OF THE BUREAU OF FISHERIES Any two means are compared by dividing their difference by the standard error estimated by the formula IS(x-x)2+S(x'-x')2 ( 1 " 1 \ ° V n1+n2 VH + l+w2 + iy if Xi, x2 , a:n2 + l and x'u x'2 x'n2 + l be two samples, and 1 7H+ 1 Six), x' 1 n2Jrl Six') Of the 71 comparisons between the mean vertebral counts given in table 6, 36 are between localities not over 25 miles apart, and the remaining 35 are from localities over 25 miles apart. The results of the comparisons are listed in table 7. That there is such close agreement between the results of the two group of com- parisons is surprising considering the difference in distance. The median distance apart in the close group is only 17 miles as contrasted with 52 miles in the group of distant comparisons. These results tend to indicate that ordinarily there is probably little intermingling between herring of different races, the boundaries between racial areas being quite abrupt. Table 7. — Summarized comparisons of vertebral counts Distance apart in miles 0-25 2G and over. Probability of means being the same 0.99-0.05 0.05-0.01 0.01 or less 27 3 6 26 2 7 The mean vertebral counts of herring from Petersburg, in the northern entrance to Wrangell Narrows, for instance, differ by over 4 standard errors from those of Juneau, Point Gardner, or Meade Point, by 3.89 standard errors from Warren Island, and by 3.19 from those of Wrangell. This is rather definite evidence that herring do not migrate through Wrangell Narrows or Dry Strait, and that migrations in Freder- ick Sound must be largely confined to that body of water. The samples taken in the area south of Sumner Strait and east of Clarence Strait, including Wrangell, Anita Bay, Santa Ana Inlet, and Frances Cove, do not differ amongst themselves, so that, until additional evidence is collected, it can only be assumed that the herring of these localities may intermingle. That this group of herring does not intermingle with herring of the outer coast through Sumner Strait is clearly indicated by a difference between the Warren Island and Wrangell averages of 5.87 standard errors. On the west coast of Prince of Wales Island the means of the samples from 4 closely adjacent localities, Klawak, Port Estrella, Noyes Island, and Culebra Island, do not differ amongst themselves, but differ from those of the localities to the north. The agreement of the herring of the adjacent localities (captured during the summer months) with those taken while spawning, on the important spawning grounds near Klawak, is quite in keeping with expectations. Noyes Island differs from Coronation Island by 4.56 standard errors and from Warren Island by 2.82 and 2.95 standard errors in the 2 available comparisons. However, Warren Island and Culebra Island differ by only 2.03 standard errors which gives a probability of 0.04 of the populations RACES OF HERRING, SOUTHEASTERN ALASKA 133 being the same. This is not sufficiently great odds to be able to state definitely that Warren Island and Culebra Island represent distinct populations. Thus while the significant differences between Warren Island and Noyes Island and between Corona- tion and Noyes Islands tend to indicate the lack of migration across Iphigenia Bay, the lack of a statistically significant difference between Warren Island and Culebra Island averages does not confirm this view. A test of the homogeneity of the Prince of Wales Island samples with the Warren Island data included, made by above-described methods, yielded an observed z of 0.3247 which happens to be exactly the same as the calculated z for a probability of 0.01, thus indicating that this group of samples is not all drawn from one population. The same test made without the Warren Island samples gives an observed 2 of 0.0248 which is many times less than the calculated z of 0.4420 for a probability of 0.05. Therefore, it must be concluded that the data point to the lack of migration between localities lying north and those lying south of Cape Ljmch (Iphigenia Bay). Samples from Warren Island, at the mouth of Sumner Strait, differ from those of Meade Point, near the mouth of Frederick Sound, by 2.44 standard errors which yields a probability of 0.014. Even if this difference were significant, it could not be assumed to give definite information on migration through Iveku Strait as neither locality is very close to the entrance to this channel. It does, however, suggest that the herring at the mouth of Frederick Sound do not migrate to the ocean by tills route. A sample from Kell Bay, in Affleck Canal, differs significantly in vertebral count from Coronation Island, but does not show a significant difference from Warren Island, which is about the same distance as Coronation Island from the mouth of the canal. The vertebral count comparisons of table 6 also indicate differences between the Cape Ommaney herring and those from Port Herbert and Gut Bay, but not from those caught in Big Port Walter. The Big Port Walter herring also do not show differences from Port Herbert. Therefore these localities cannot be classified without more material. ANALYSIS OF GROWTH RATES In analyzing the growth rates from the various localities for racial purposes no data were used except those from freshly caught purse-seined specimens, all of which were obtained during the summer months. Many of the body length distributions are slightly skewed, and in addition cover a wide range, with a tendency in a few cases for slight modes to form near the upper or lower range of the distribution. These disturbing factors are probably caused in large measure by errors in age reading whereby a length distribution may contain a few fish belonging to younger or older age groups. Since these doubtful measures near the extremes of the range exert a large influence in the determination of the arithmetic mean, whereas, being of doubtful authenticity they should not carry as much weight as the more centrally located items, it was decided not to use the arithmetic mean but to employ the median for the measure of central tendency. In keeping with the use of the median the interquartile range has been used as the measure of dispersion. To gain an insight into the growth increments during the summer months the data have been grouped by 10-day periods. (See table 8.) For Larch Bay 5-year-olds 134 BULLETIN OF THE BUREAU OF FISHERIES (herring in their fifth summer) taken during 1930, a consecutive series of 8 periods shows no consistent changes in length during the first 5 periods (from June 21 up to and including Aug. 10). There is an abrupt increase in length, however, between the fifth and sixth periods, the fish of the last three periods averaging about a half centimeter greater in body length. Such a sudden increase in length can scarcely be ascribed to growth but is probably due to an influx of new schools of herring onto the fishing grounds. That such a sudden change in body length is not due to growth is supported by the Noyes Island data, in wnich both the 4- and 5-year-olds taken toward the end of July were considerably smaller than those taken during the last of June, the largest difference, that between the 5-year-olds, being 6 millimeters. Table 8. — Body lengths in various localities for 1929 and 1930 by 10-day periods Locality Age Date Number of speci- mens Median Qi Qz 1929 4 June 11-20 226 199. 2 193 7 204 6 4 June 1-10 212 193 5 188 0 198 7 Do 4 June 1 1—20_ 190 194. 5 189. 7 200 1 Do 4 June 21-30 68 198 7 193 8 204 5 4 do 226 199 6 195 1 204 6 Do . - 4 July 1-10 131 200 9 196 6 206 O Do 4 July 11-20 51 203. 4 199 3 207 6 Do 4 July 21-31 39 202 4 198 3 206 1 4 May 31 85 191 7 185 1 198 3 4 June 1-10 60 194 5 187 7 202 0 4 June 11-20 87 190 4 185 1 197 1 4 do 38 194 2 187 4 198 0 4 do 147 203 4 199 2 208 5 4 June 1-10 74 186 9 183 7 191 0 4 June 8-12 95 185 9 178 8 190 9 4 July 1-10 59 200 0 195 8 204 9 Do . 4 July 11-20. 65 201 2 196 3 205 1 Do 4 July 21-31 37 201 6 198 1 208 ' 9 4 do 57 201 8 192 8 205 3 4 do 44 184. 5 175 5 189 5 4 ....do 36 187.5 179.5 191.0 1930 4 July 21-31 65 204. 6 198 1 208 9 Do 4 Sept. 1-10 132 206. 4 201 5 211 5 4 June 21-30 34 194 8 191 1 200 0 4 July 11-20 47 198 4 194 3 208 1 4 July 1-10 37 197. 8 194 8 204 4 Do - 4 July 21-31 - 28 196.0 191 5 202 0 4 June 21-30 121 192.8 188 3 197 2 "Do 4 July 1 1-20 53 187. 6 181 3 190 7 Do 4 July 21-31 34 188. 5 181. 3 190 4 Todd 4 June 1-10 160 176. 0 169 5 181 9 June 21-30 80 210. 8 204 2 214 8 Do 5 July 1-10 98 213. 0 206. 8 217 4 Do - 5 July 11-20 157 212. 1 207. 4 218 0 Do 5 July 21-31 284 211. 5 206 1 218 1 Do 5 Aug. 1-10 46 212.9 208 8 217 3 Do 5 Aug. 21-31. 67 219. 1 213. 2 223 0 Do 5 Sept. 1-10. 334 216.0 210. 7 222 ’0 Do 5 Sept. 11-20 38 216. 5 207 3 222 7 5 June 11-20 70 211.8 206 1 214 8 Do 5 June 21-30 145 209.9 200 1 214 7 Do 5 Aug. 21-31 43 213. 3 209 8 219 6 5 July 11-20 63 207. 1 201 7 210 8 5 do 68 215. 2 209. 5 218 8 Do 5 July 21-31 188 214. 7 209.3 220 4 Do 5 Aug. 11-20 48 212.5 204.8 217 2 5 June 21-30. 38 212. 2 209 5 218 9 Do 5 July 1-10... 119 209. 1 203 9 212 8 Do 5 July 11-20 162 211. 1 205. 3 215 5 Do 6 July 21-31 51 213.0 208. 4 217 3 5 June 21-30 _ . 134 204.5 199. 6 210 8 Do 5 July 11-20 39 197. 1 193.9 201 8 5 July 1-10 29 214.0 209.8 217 8 5 do... 195 213.6 208.0 218.8 Q2—Q\ 10.9 10.7 10.4 10.7 9.5 9.4 8.3 7.8 13.2 14.3 12.0 10.6 9.3 7.3 12.1 9. 1 8.8 10.8 12.5 14.0 11.5 10.8 10.0 8.9 13.8 9.6 10.5 8.9 9.4 9. 1 12.4 10.6 10.6 10.6 12.0 8.5 9.8 11.3 15.4 8.7 14.6 9.8 9.1 9.3 11.1 12.4 9.4 8.9 10.2 8.9 11.2 7.9 8.0 10.8 Because of these large variations in the size of herring of the same age taken from the same locality, it was adjudged unwise to attempt any careful statistical com- RACES OF HERRING, SOUTHEASTERN ALASKA 135 parisons between localities. The medians and quartiles are given for each month in table 9. Figures 4,5, and 6 give the length distributions for some of the localities. In these figures the lengths have been grouped by 3-millimeter categories and smoothed once by three’s. The figures and the table reveal at once that the herring from 4 Figure 4. — Percentage length distributions, grouped by 3-millimeter categories and smoothed once by three’s for herring of 1926 year class caught in 1929 (their fourth summer). localities: The Noyes Island area (including Culebra Island and Port Estrella), the Douglas Island-Icy Strait area, Affleck Canal (Kell Bay) and Peril Strait (Todd) are all much slower growing than those of the other localities. These differences are so great that we have considered these 4 localities to represent groups of fish separate from the neighboring stocks or populations. The Peril Strait herring appear to be the slowest growing of any we have so far encountered in Alaska, the median of the 4-year-olds taken in June 1930 being only 176.0 millimeters. 136 BULLETIN OF THE BUREAU OF FISHERIES Figure 5. — Percentage length distributions, grouped by 3-millimeter categories and smoothed once by three’s for herring of 1927 year class caught in 1930 (their fourth summer). Table 9. — - Body lengths in various localities for 1929 and 1930 by months Locality Age Date sampled Num- ber of speci- mens Median Qi Qi Qi Qi Year Month 4 1929 June... ... ... 226 199.2 193.7 204.6 10.9 4 1929 do _ 470 194.2 189. 4 200. 1 10.7 4 1929 do_ 226 199.6 195. 1 204. 6 9.5 Do - --- -- 4 1929 July 221 202.0 197.2 206.3 9. 1 4 1929 May31__ __ _ 85 191.7 185. 1 198. 3 13.2 4 1929 June. _ 60 194.5 187.7 202. 0 14. 3 4 1929 do. _ . 87 190.4 185. 1 197. 1 12.0 4 1929 ... do_ - 38 194.2 187. 4 198.0 10.6 4 1929 do_ __ 147 203.4 199. 2 208. 5 9.3 4 1929 do. _ . ... 74 186.9 183. 7 191.0 7.3 4 1929 95 185. 9 178.8 190. 9 12. 1 4 1929 July 161 200.8 196. 4 205.7 9.3 Favorite Bay 4 1929 do 57 201.8 192.8 205.3 12.5 4 1929 do. .. 49 185.8 176. 1 191. 4 15. 3 4 1929 do. 71 188.3 182. 4 198. 1 15.7 4 1930 do. . .. 96 203.5 197.3 207.9 10.6 Do - 4 1930 September. _ .. . 152 206.4 201. 4 211. 6 10.2 4 1930 June 45 196.8 191.8 201.9 10. 1 4 1930 July 47 198. 4 194. 3 208. 1 13.8 4 1930 do 80 197. 7 193. 8 204. 5 10. 7 4 1930 June ._ . . 121 192.8 188.3 197. 2 8.9 iWJpo 4 1930 July 87 187.8 181.3 190.5 9. 2 4 1930 June.. .. ... 160 176. 0 169. 5 181.9 12. 4 5 1930 _ do 80 210.8 204. 2 214.8 10. 6 5 1930 July. . .. . . . 539 212.0 206.8 217. 9 11. 1 Do - - - — -- 5 1930 August 113 216.3 211. 5 221.0 9.5 Do - -- - -- -- --- 5 1930 September 372 216. 0 210.6 222. 1 11.5 5 1930 Jun9 ._ _ 215 210.4 204. 2 214.8 10. 6 ~~ Do - 5 1930 August . . . 43 213.3 209.8 219.6 9.8 5 1930 July 63 207. 1 201. 7 210.8 9. 1 5 1930 do 256 214.8 209. 3 219. 8 10. 5 5 1930 August— . . . 48 212.5 204.8 217. 2 12. 4 5 1930 June... 38 212.2 209. 5 218.9 9.4 5 1930 July __ 332 210.4 204.9 214.9 10. 0 6 1930 June ... 134 204.5 199. 6 210.8 11. 2 5 1930 July 54 198.5 194.7 206.0 11.3 Point Gardner 5 1930 do 29 214.0 209.8 217.8 8.0 5 1930 do. 195 213.6 208.0 218.8 10.8 RACES OF HERRING, SOUTHEASTERN ALASKA 137 That the northern slow-growing group of herring found in Icy Strait and the vicinity of Juneau all belong to the same race cannot be assumed from growth rates. A test of the homogeneity of the vertebral count distributions in this area (see A, fig. 2) gives an observed z (see illustration in section on vertebrae) of 0.4069. The Figure 6. — Percentage length distribu; ;e- • grouped by 3-millimeter categories and smoothed once by three’s, for herring of the 192fi > 'ar class caught in 1930 (in their fifth summer). calculated z for a probability of 0.05 is 0.3388, and for a probability of 0.01 it is 0.4909. Therefore, in this case, the observed z is large enough to be of doubtful significance. In this test the mean square between the arrays 0.2268 is less than that within the arrays 0.5117 which may indicate a difference in variance. It must be realized that no definite conclusions as to the racial homogeneity of this area can be reached without further data. 138 BULLETIN OF THE BUREAU OF FISHERIES ANALYSIS OF YEAR CLASSES Differences in the relative proportions of different year classes present in the populations may be of value as an indication of the extent of intermingling. The relative abundance of any particular year class in the catch is influenced chiefly by three factors: First, the relative number of larvae of that year class hatching and sur- viving through the juvenile period until of an age or size to enter the catch; second, by the rate of natural mortality; and third, by the increased rate of mortality induced by the fishery. However, the relative abundance of any particular year class in the catch may change during the season. Such a progressive change from younger to older age groups is apparent as the season advances, for instance, in Prince William Sound. Owing to such fluctuations in the relative proportions of each age group present at any particular time, only the major differences between localities can be emphasized without further data. In figure 10 are given histograms of the percentage age distributions for the major fishing grounds during the 1929 and 1930 seasons. Several features are worthy of em- phasis. These are (1) the approximate equality of the 1926 and 1927 year classes at Noyes Island as contrasted with the overwhelming dominance of the 1926 year class at most of the other localities, (2) the great dominance of the 1927 year class at Todd, (3) the dominance of the 1923 year class at Douglas Island and Favorite Bay, (4) the large percentage of the catch older than the 1923 year class at Douglas Island. These salient facts all support the indications given by the analyses of the vertebral counts and of the growth rates which separate the Noyes Island area, Todd, and the Juneau-Icy Strait area (including Douglas Island) as independent of neighboring areas. Whether Favorite Bay is independent of Point Gardner we hesitate to say without further data. TAGGING Curtailment of fishing operations during 1932 and 1933, due to economic condi- tions, and an increased abundance in the Cape Ommaney tribe of herring, owing to the accession to the catch of certain dominant year classes, caused fishing during 1932 to be carried on almost wholly in the area between Noyes Island and the lower end of Baranof Island. During 1933 the boats did not fish farther south than Warren Island. Tins condition, which was in contrast to the widespread fishing conducted for several years previously, made our tagging work of less value than it would have been under normal conditions, as obviously the presence of tagged herring cannot be detected where no fishing operations are being conducted. However the results are presented for what they are worth. For a full discussion of tagging methods and manner of recovery of the tags, see Rounsefell and Dahlgren (1933). During the fishing season of 1933 (June 1 to September 30) 101 belly tags and 7 opercle tags were recovered from 2,499 of the former and 1,470 of the latter affixed to herring released at Jamestown Bay (Sitka) between April 21 and April 25, 1933. All of these tags were recovered around Cape Ommaney, between Larch Bay and Port Alexander, giving the first definite proof of a migration of some length, as it is approximately 66 miles by water from Jamestown Bay to Port Alexander. On the other hand, out of 996 belly tags and 824 opercle tags affixed to herring released at Cape Bendel, just under 60 miles from Port Alexander, on August 17, 1932, not a tag has been recovered. This may be considered rather definite evidence of a lack of migration between Cape Bendel and Cape Ommaney. U. S, Bureau of Fisheries, 1935 Bulletin No. 1 7 FIGURE 7. — BLUESTONING A SEINE. At frequent intervals throughout the season a seine is dipped into a solution of copper sulphate to kill the marine growths that cause the web to rot. Figure 8. — The Scientific Vessel ■•Heron" at anchor in Port Conclusion. Note the live box used in tagging operations moored astern. The acquisition of this motor launch in 1932 permitted the carrying out of the herring tagging program. U. S. Bureau of Fisheries, 1935 Bulletin No. I 7 Figure 9. — Herring Tagging operations. The tagging live box is filled with herring which the man with the net dips a few at a time into the half barrel of sea water under the shelter. From this half barrel the herring are removed one at a time by hand and placed on damp cheesecloth stretched over a frame. A sharp arrow-pointed knife is used to puncture a small hole through the wall of the belly and the flat metal tag is inserted within the body cavity of the herring, the small wound healing completely within a few days. RACES OF HERRING, SOUTHEASTERN ALASKA 139 In another tagging experiment at Auke Bay near Juneau, 800 belly tags and 772 opercle tags were affixed to spawning herring released on May 3, 4, and 5, 1933. Figure 10.— Percentage age distributions of herring taken during 1929 and 1930 in some of the principal fishing grounds. No recoveries have been made, supporting the previous conclusion of a lack of migra- tion between Juneau and Cape Ommaney. 140 BULLETIN OP THE BUREAU OF FISHERIES SUMMARY The results so far attained in determining the herring populations and the areas inhabited by them serve to show the immense complexity of the problem. Although much remains to be done, the present work has set apart some of the chief areas and answered many pressing questions. It must be borne in mind that where morpho- logical differences have not been shown we can only assume that the populations are the same, until such time as we obtain evidence to the contrary. Such evidence may come from tagging experiments. It must be further noted that the area bound- aries shown in figure 2 are merely lines across which neighboring populations do not migrate according to positive evidence. They are not intended to convey the impression that each area necessarily contains only one race. Indeed, the section on tagging shows that area C very probably does contain more than one race. With these restrictions in mind the following results are listed: 1. Differences in vertebral count between Wrangell and Warren Island herring indicate that there is no migration through Sumner Strait. 2. Differences in vertebral count between Petersburg and Wrangell herring show a lack of intermingling through Dry Strait or Wrangell Narrows. 3. Differences in vertebral count between Point Gardner and Petersburg show no migration through Frederick Sound. 4. Differences in growth rate and vertebral count indicate a lack of migration along the outer coast between localities lying north and those lying south of Cape Lynch (Heceta Island). 5. The presence of very slow-growing herring in Peril Strait indicates a distinct race and shows a lack of migration through this waterway. 6. Differences in growth rate and vertebral count indicate that the herring of Icy Strait and the vicinity of Juneau do not migrate down Chatham Strait or through Stephens Passage. Failure to recover any of the herring tagged at Auke Bay in the Cape Oinmaney fishery tends to confirm this view. 7. A difference of 2.44 standard errors, giving a probability of 0.014 between the mean vertebral counts of Warren Island and Meade Point indicates that there is probably no migration through Keku Strait. 8. Failure to recover any of the tagged herring released at Cape Bendel in the Cape Ommaney fishery indicates a lack of migration between lower Chatham Strait and Frederick Sound. 9. Recovery of tagged herring proves that the great spawning grounds in Sitka Sound are the mainstay of the tremendous herring fishery at Cape Ommaney. LITERATURE CITED Buchanan-Wollaston, H. J. 1933. Some modern statistical methods: their application to the solution of herring race problems. With foreword by R. A. Fisher. Jour, du Cons., Cons. Perm. Intern. Explor. Mer. vol. VIII, no. 1, pp. 7-47, 4 figs. Copenhague. Fishee, R. A. 1930. Statistical methods for research workers. No. V. Biological monographs and manuals. 3d edition, 283 pp., 12 figs. Edinburgh. Rounsefell, George A. 1930. Contribution to the biology of the Pacific herring, Clupea pallasii, and the condition of the fishery in Alaska. Bull., U. S. Bur. Fish., vol. XLV, 1929 (1930), pp. 227-320, 53 figs. Rounsefell, George A. 1930a. The existence and causes of dominant year classes in the Alaska herring. Contr., Marine Biol., 1930, pp. 260-270, 5 figs. Stanford University, Calif. RACES OF HERRING, SOUTHEASTERN ALASKA 141 Rounsefell, George A. 1931. Fluctuations in the supply of herring, Clupea pallasii, in south- eastern Alaska. Bull., U. S. Bur. Fish., vol. XLVII, 1931 (1932), pp. 15-56, 26 figs. Rounsefell, George A. 1934. Several distinct races of herring are found in southeastern Alaska. Pacific Fisherman, vol. 32, no. 3, February 1934, p. 31. Seattle. Rounsefell, George A., and Edwin H. Dahlgren. 1932. Fluctuations in the supply of herring, Clupea pallasii, in Prince William Sound, Alaska. Bull., U. S. Bur. Fish., vol. XLVII, 1932, pp. 263-291, 15 figs. Rounsefell, George A., and Edwin H. Dahlgren. 1933. Tagging experiments on the Pacific herring. Jour, du Cons., Cons. Perm. Intern. Explor. Mer., vol. VIII, no. 3, pp. 371-384, 6 figs. Copenhague. o U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner EFFECTS OF CRUDE OIL POLLUTION ON OYSTERS IN LOUISIANA WATERS By Paul S. Galtsoff, Herbert F. Prytherch Robert 0. Smith, and Vera Koehring From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 18 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1935 For sale by the Superintendent of Documents, Washington, D. C. Price 20 cents . EFFECTS OF CRUDE OIL POLLUTION ON OYSTERS IN LOUISIANA WATERS 1 By Paul S. Galtsoff, Ph. D., Herbert F. Prytherch, Ph. D., Robert O. Smith, and Vera Koehring, Ph. D., United States Bureau of Fisheries CONTENTS Page Foreword 144 Introduction (P. S. Galtsoff and H. F. Prytherch) 144 Preliminary field investigations, 1933 (H. F. Prytherch) 146 Survey of oyster bottoms in areas affected by oil-well pollution, 1934 (R. O. Smith) 150 Methods 150 General conditions 150 Lake Barre 152 Lake Felicity and Lake Chien 152 Terrebonne Bay 153 Timbalier Bay 153 Lake Raccourci 154 Lake Pelto and Pelican Lake 154 Examination of oyster beds at mouth of Bayou Grey and Little Lake 155 Conclusions 156 Experimental studies of the effect of oil on oysters 158 Review of the literature (P. S. Galtsoff) 158 Survival of oysters in oil-polluted water (H. F. Prytherch) 159 Experiments with surface film of oil 160 Survival of oysters in sea water passed through oil 161 Immersion of oysters in oil 162 Effect of oil on glycogen content of oysters 163 Experiments with brine 164 Effect of brine on glycogen content of oysters 167 Effect of oil on feeding of oysters (P. S. Galtsoff and R. O. Smith) 167 Effect on the adductor muscle 167 Effect of oil and oil well bleed-water on the rate of feeding of oysters 170 Carmine cone method 171 Drop counting apparatus 172 Preparation of water-soluble fraction of oil 174 Results obtained with the cone method 174 Results obtained with the drop-counting method 183 Effect of bleed water on the rate of feeding 188 Effect of consecutive treatments 191 Effect of crude oil on diatoms (P. S. Galtsoff and V. Koehring) 193 Method 194 Effect of heavy surface layer of oil on Nitzschia culture 196 Effect of oil held on the bottom 197 Effect of water-soluble fraction of oil 199 Effect of oil held in collodion bags 200 Effect of brine on Nitzschia 200 Discussion and conclusions (P. S. Galtsoff) 204 Bibliography 208 i Bulletin no. 18. Approved for publication, June 3, 1935. 143 144 BULLETIN OF THE BUREAU OF FISHERIES FOREWORD The mortality of oysters in Louisiana waters in 1932-33, coincident with the development of oil wells in the coastal areas of the State, brought up again a question as to the possible effect of oil on marine life. In the spring of 1933, at the request of the Louisiana Conservation Department, Dr. H. F. Prytherch was detailed by the United States Bureau of Fisheries to make an investigation in the Terrebonne Parish and adjacent territory with the view of determining the probable cause or causes of the mortality. In 1934, in an attempt to carry out a more comprehensive study, the Bureau obtained from the Civil Works Administration approval of a project to investigate the oil-pollution problem in Louisiana and to carry out both laboratory and field experiments in order to determine the effect of oil on oysters. Unfortunately, out of the $42,000 allotted for this project only an amount of $3,000 was made avail- able, and after the completion of a preliminary hydrographic survey of Timbalier and Terrebonne Bays and adjacent bodies of water the field work was discontinued. Laboratory experiments on the effect of oil on oysters and oyster food were carried out, however, at Beaufort, N. C., Woods Hole, Mass., and Washington, D. C. Although the exact cause of the mortality of oysters has not been determined, the reports of the field and laboratory investigations throw considerable light on the conditions of oyster beds in Louisiana waters and on the possible effect of oil pollu- tion on oysters. INTRODUCTION By P. S. Galtsoff and H. F. Prytherch The fact that the discharge of oil into natural waters may be detrimental to aquatic life has been recognized for a long time and was a subject of lengthy discussion before the numerous governmental committees (Oil Pollution in Navigable Waters, 1926; Pollution of Navigable Waters, 1930) attempting to remedy the situation by proper legislative action. In view of the widespread oil pollution of coastal waters, espe- cially in the vicinity of large cities and industrial centers, and the interest in this problem shown by many Federal and State agencies, it is surprising to learn that there has been very little direct experimentation on the effect of crude oil or its derivatives on fresh-water or marine organisms and that most of the statements appearing in the minutes of official hearings are based primarily on field observations and frequently represent opinions and assumptions not corroborated by direct evidence. It is true that in the case of gross pollution conditions in the affected body of water may be such as to make any deeper investigation superfluous. However, when pollution is light or only temporary, the lack of knowledge of the toxic properties of oil and of the manner in which it may affect various organisms, constitutes a serious handicap in developing efficient methods of protection. In the case of the location of oil fields in the coastal area, the question arises as to whether or not the coastal fishery and the oil industry can coexist. The problem is of particular importance in south Louisiana waters where the production of oil has increased rapidly from a total output of 5,032,400 barrels in 1927 to 15,540,341 barrels in 1933, according to Dabney (1934). Drilling operations have been extended to within a few miles of the Gulf of Mexico, particularly in Terrebonne Parish, where, according to the statistics of the Louisiana Department of Conservation, there were 17 wells in opera- EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 145 tion in 1933, having a total average daily output of 6,961 barrels. In this parish most of the wells are not located on land, but are situated in the open waters of Lake Barre, Lake Pelto, and vicinity, which constitute one of the most important oyster- producing regions of the State. Pollution of the water has occurred as the result of oil-well operations, and coincident with this condition there has been a high mortality of oysters particularly during the winter of 1932-33 and to a lesser degree during the previous winter. The aggregate losses of the oyster planters, alleged to have been caused by oil-well operations, have been estimated at several hundred thousand dollars. In January 1933 the Louisiana Department of Conservation received reports of an extensive and serious mortality of oysters in the Lake Barre and Lake Pelto region, which were corroborated by immediate field surveys of its Bureau of Research and the State Department of Health. The chemical studies of pollution and subse- quent field and laboratory experiments conducted by these departments are reported briefly by Gowanloch (1934). Similar chemical and biological field investigations were also made by chemists of the Texas Co. and by Dr. C. E. Coates, Dr. A. R. Cboppin, and Dr. W. H. Gates of the chemistry and zoology departments of Louisiana State University. In May 1933 the cooperation of the U. S. Bureau of Fisheries was requested by the Louisiana Department of Conservation, following which, field studies were made by Dr. H. F. Prytherch in May, June, and September of the oyster mortality in Terrebonne Parish in relation to oil-well operations in this region. At the Bureau’s laboratories in Washington, D. C., and Beaufort, N. C., a series of experiments have been conducted by the authors to determine the effect of different grades of crude petroleum and accompanying brine waters on the survival, feeding, and food of the oyster. There is no doubt that unavoidable pollution of water by oil and bleed water incidental to drilling operations constitutes a serious danger to local oyster, shrimp, and fishing industries from the point of view of a fisherman, it being immaterial whether pollution has actually destroyed the stock of fish and shellfish or rendered them unmarketable on account of oily taste and emaciated condition of the flesh. In both cases the industry sustains economic losses. Of course, in case of oyster or other mollusks there is a possibility of transplanting the stock to other areas unaffected by pollution. So far as the Louisiana coastal section is concerned, this appears to be only a palliative, for it is but a question of time until the development of oil fields will spread all along the coast and most of the oystermen in the State sooner or later will face the problem which at present confronts the industry in Timbalier and Terrebonne Bays and Lake Pelto, the sites of the present extensive drilling operations. The questions to which the oyster industry desires to receive a competent answer can be formulated as follows: Whether the unusual mortality which occurred in Louisiana in 1932-33 is attributable to the discharge of oil and bleed water, and how further development of the oil resources of the coastal section may affect the cultiva- tion of oysters. The marine biologist called to provide an answer finds himself in a difficult situation. As often happens in the case of an unusual mortality among fish or oysters, he is requested to investigate the cause or causes of it several weeks or months after death has occurred, and when the conditions responsible for the mor- tality have already changed or disappeared. There is a general and well understand- able tendency on the part of a layman to attribute his troubles to unusual activities or events that occurred in the affected area. The concurrence of the two phenomena, however, does not constitute in itself a proof that one is the cause of the other. 146 BULLETIN OF THE BUREAU OF FISHERIES There are two aspects in the study of mortality problems in relation to pollution of water. One is to determine the cause of the mortality, that occurred some time ago, the second is to find out how the pollutant agent may affect marine life and what are the expectations of the fishing industry if the pollution is permitted to continue. From the point of view of conservation, the second problem is of far greater impor- tance, while the parties involved in the controversy regarding the causes of mortality are primarily interested in the first one. Although the present report fails to give a definite answer to the first question, it supplies sufficient data regarding the possible dangers of oil pollution to oysters. One must bear in mind that actual conditions in the sea may be much more complex than they appear to a casual observer. There is a possibility, for instance, that because of unfavorable meteorological conditions, attacks of parasites, or other unknown factors, oysters already have become weakened or diseased. In that case an additional adverse factor, as for instance a small concentration of a toxic sub- stance in the water, may have been responsible for the mortality, although under more favorable circumstances the oysters might have been strong enough to withstand it. Being ignorant of the past history of the case and not being able to duplicate condi- tions that existed at the time of greatest mortality, the biologist is unable to determine with certainty the cause of it. He can provide, however, sufficient evidence regarding the toxicity of the suspected pollutant, the manner in which it affects the organism, its fate in the ocean, and from all this information provide a substantial basis for outlining the methods and policies of future conservation. These considerations and the lack of funds to carry out field observations on the large scale contemplated in the original project, made it necessary for the authors to concentrate their attention on the experimental studies of the effect of oil on the behavior of the oyster and on the growth of diatoms, which constitute the principal part of the oyster diet. It was thought desirable, however, to present first the preliminary field observa- tions made by H. F. Prytherch in 1933 and the result of the survey made by R. 0. Smith in March and April 1934. PRELIMINARY FIELD INVESTIGATIONS, 1933 By H. F. Prytherch Terrebonne Parish includes practically the whole oyster-producing region be- tween Barataria Bay and the Atchafaiaya River and is the westernmost section in which good oysters are obtained in considerable quantities. According to the Bureau’s fisheries statistics for 1930 (Fiedler, 1932), this parish exceeded all others in oyster production and furnished over 360 thousand bushels, valued at $156,213, or approximately 30 percent of the total State crop. Virtually all of the oysters produced here come from private beds, leased from the State, most of which formerly were natural beds that had become depleted through overfishing and the destruction of small seed (spat) by natural enemies. Through the leasing and cultivation of these areas production has been maintained and a better grade of oysters produced. The usual procedure is to stock the leased areas with seed obtained from natural beds in other coastal sections such as Sister Lake, Bay Junop, Barataria Bay, etc. The seed are planted quite densely at reported concentrations of 700 to 900 barrels per acre which does not allow a great deal of room for subsequent growth and increase in volume. This practice, however, has been generally successful for many years and is apparently caused by the fact that the oyster beds are elevated to some extent EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 147 above the surrounding bottom and occupy but a very small percentage of the total acreage of those inshore waters. When the oysters reach marketable size they are removed from the growing areas, culled, and either shipped directly to market or temporarily stored in convenient protected areas near the oyster camps. It was during these operations that the oysters died in greatest numbers, as a result pri- marily of a weakened condition of the adductor muscle and its failure to maintain closure of the shell in air or water. During shipment the oysters opened quickly and spoiled because of evaporation and the loss of shell liquor, while a high per- centage of those transplanted to storage areas soon succumbed to the attacks of crabs, small fish, etc., because of their inability to close the shell. The mortality was greatest during the winter months and occurred on beds located at distances ranging from approximately one-half mile to 9 miles from the Barre oil wells and 1 to miles from the Pelto wells as shown in figure 1. During the periods from May 23 to June 1, and from August 31 to September 2, 1933, field observations were made of the oyster mortality in Lake Barre, Lake Pelto, and vicinity in relation to existing hydrographical and biological conditions and particularly in respect to pollution of these waters by oil-well operations. These studies were made several months after the mortality was at its peak and after pol- lution of the water by oil, hydrogen sulphide, and natural gas had been considerably reduced by action of the Louisiana Conservation Department. Oil production at Lake Barre and Lake Pelto began in June and September 1929, respectively. The most important wells are those at Lake Barre, 14 of which were in operation on June 1, 1933, with a production on that day of 9,987 barrels of oil according to a report received from the Texas Co. These wells were also pro- ducing at that time from 6,500 to 7,000 barrels of brine or “bleed water” daily which after being combined and chlorinated (since February 1933) for removal of hydrogen sulphide was emptied into the bay. Samples of this effluent were found to have a salt concentration of 123 parts per thousand as compared with 15 to 22 per thousand for the surrounding waters. They became quite turbid soon after exposure to air or chlorine through the formation of colloidal suspensions of iron and sulphur, and gradually were covered with a thin film of oil after coalescence of finely divided particles. The daily output of the Lake Pelto wells was much less, amount- ing to only 57 barrels of oil and 650 to 750 barrels of brine, the latter having a salt concentration of 98 parts per thousand. Studies were made at 27 stations throughout this region as to the condition of the water and its bearing on the reproduction, growth, and mortality of oysters on both natural and planted beds, located in the immediate vicinity and at varying dis- tances from the oil wells. Particular attention was also paid to the abundance and activities of natural enemies of the oyster and to other marine organisms, particu- larly mollusks, which might be affected by oil-well pollution. The following observations were made in respect to the oyster-oil well situation in Terrebonne Parish. 1. Random samples of oysters from planted beds in Lake Barre, Lake Pelto, Timbalier Bay, and vicinity, showed that a high percentage of the oysters had died previously on all but three of these leased areas. The mortality on the various beds ranged from approximately 50 to 95 percent as shown in figure 1 and in virtually every case included only the larger and older oysters of marketable size. Many of those surviving were in poor condition and exhibited retarded and abnormal shell growth during the preceding period. 148 BULLETIN OF THE BUREAU OF FISHERIES 2. No evidence of unusual mortality or retarded growth was found on the beds at stations 23 and 24 in Lake Felicity, on 2 adjacent plantings at station 16 in Lake Pelto, or on a small natural oyster bed, at station 25, which is located at a distance of approximately 500 yards from the Barre wells. Adult oysters were also growing on the piling of these wells and showed a general vertical distribution ex- tending from 1 to 2 inches above the bottom to nearly the surface of the water. The shells of the latter measured from 2% to 3J4 inches and indicated continuous rapid growth over a preceding period of at least 1 year. 3. The oyster beds in Lake Barre and Timbalier Bay are located at distances varying from approximately 500 yards to 9 miles from the numerous Barre wells. Those in Lake Pelto lie within a radius of from 1 to 5% miles from wells in that region. The previous observations failed to show any direct relation between the degree of oyster mortality on these beds and their distance from the oil wells. 4. Oysters in all localities contained a considerable amount of ripe spawn. A heavy spawning had already occurred several weeks prior to June 1 and was still in progress at that time. 5. Tow-net collections of microscopic marine life revealed an abundance of healthy oyster larvae as well as those of other bivalve mollusks. In the immediate vicinity of the oil wells the free-swimming larval stages of the oyster were plentiful at all ages from 1 day old to setting size. 6. Examination of old shells and those of live oysters showed that intensive setting or attachment of oyster larvae had already occurred (1) on the natural beds, (2) on the planted beds where the mortality occurred, and (3) on the piling of the oil wells and other submerged objects nearby. Heavy setting was still in progress at the time of the investigation as shown by maximum concentrations ranging from 200 to 500 spat per shell on those collected at the Lake Barre and Lake Pelto wells. 7. The spat or minute seed oysters found throughout Terrebonne Parish were of varying ages and size, ranging from 1 day old or recently attached specimens with a diameter of one seventy-fifth of an inch to those 2 or 3 weeks of age having a diameter of approximately one-fourth to one-lialf inch. It was evident that metamorphosis of the oyster from the larval to spat stages had been successfully completed and that the growth obtained subsequently under existing conditions was rapid and to all appear- ances normal. No indications of unusual mortality of spat were observed. Spat collected at the Lake Pelto oil well at depths varying from a fraction of an inch to 2 feet above the bottom were in good condition, as were those found on oysters in the test boxes at the Lake Barre oil wells and on the adjacent natural bed. Though set- ting and spat production is heavy throughout this region only a small percentage survive because of the attacks of natural enemies such as crabs and the borer {Purpura haemostoma). 8. Three natural enemies of the oyster, the borer, Purpura; the boring clam, Martesia; and the boring sponge, Cliona, were found to be abundant on many of the planted beds. There was no evidence of the destruction of adult oysters by the borer. The shells of approximately 50 to 75 percent of the dead oysters examined showed heavy infestation of the boring clam and boring sponge. The numerous small tunnels and perforations made by these organisms had caused partial disintegration and weakening of the shells, interfered with normal shell growth, and, apparently, had been a serious drain upon the vitality of the oj^ster. Serious sponge attack and per- foration of the shell by the boring clam were found in many instances at the point of U. S. B. F. Bulletin No. 18. 143870—35. (Face p. 148) U S B. F. Bulletin No. 18. Figure 1. Terrebonne and Timbalier Bays, Louisiana. Lines indicate isohalines in February-March, 1934. Black circles show percent of dead oysters at various stations observed by H. F. Pry therch in 1933. Squares indicate location of oil wells. 143870-35. (Face p. 148) ' ■. £ it: ' . : * . EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 149 muscle attachment. There was no evidence of a mortality of these natural enemies or of impairment of tlieir growth and reproduction on many areas where a high per- centage of the oysters had died previously. 9. A small percentage of the dead oysters and those in a weakened and dying condition showed no serious shell injury by the boring sponge and boring clam. There was one case in particular at station 6 where 95 percent of the oysters had died on a temporary storage bed and were free of shell infestations of these organisms. This bed is located approximately 5}i miles south of the Barre wells. 10. The field studies at the 27 stations in Terrebonne Parish failed to disclose any direct evidence of the destruction of marine animals or plants by oil-well pollution. As indicated previously, conditions throughout tins region and in the immediate vicinity of the oil wells were found to be decidedly favorable for reproduction of adult oysters and for the development, attachment, and growth of oyster larvae and spat. Several other mollusks such as the boring clam, shipworm, mussel, Crepidula, Anomia, and the common borer showed no evidence of unusual mortality and were found to be reproducing and growing throughout this region in an apparently normal manner. Barnacles and green algae were abundant on the piling of the oil wells and were thriving on submerged cross beams at the Barre well at a distance of 5 feet from the bleed- water discharge where they were subjected continually to a mixture of this effluent and sea water. Blue crabs were observed feeding on these forms at the latter location during the period that the effluent and oyster samples were being collected. Many species of fish such as trout, croaker, alligator gar, bonnet head shark, channel bass, and bluefish were caught or observed at the Pelto well within 100 feet of the bleed-water discharge. At the Barre well many schools of immature fish, measuring from approximately one-half inch to 3 inches were actively swimming and feeding at distances ranging from 3 to 25 feet from the effluent discharge. 11. Hydrographical observations. — Determinations of water temperature, specific gravity, and hydrogen-ion concentration were made at all stations during the period, May 23 to June 1. The records of water temperature give an average of 28.1° C. (86.6° F.) and range from 27° to 29° C. Measurements of specific gravity at surface and bottom, corrected to 17.5° C., range from 1.0118 to 1.0174 with an average figure of 1.0143. When converted into terms of salinity (grams of salt per 1,000 grams of sea water) the concentration of salts shows a variation of 15.41 to 22.77 per mille, with an average of 18.69 per mille. These salinities are favorable for oyster growth and reproduction and correspond to those found on some of the best oyster-producing areas in this country. The present records of specific gravity have been compared with those obtained in 1906 and 1907 by the United States Bureau of Fisheries (Moore and Pope, 1910) and the Gulf Biologic Station (Cary, 1907) and show that the salt content of the water in Terre- bonne Parish was essentially the same in 1933 as at that time. The observations of hydrogen-ion concentration (as expressed in pH) show an average pH of 8.3 and range from 8.2 to 8.6. These also compare favorably with the Bureau’s records of this factor in other oyster-producing regions. Water samples, collected in the immediate vicinity of the Barre and Pelto wells (100 and 300 feet from brine discharge pipes) showed no appreciable difference in liydrogen-ion concentration and salinity at surface and bottom, and no noticeable increase in these factors as com- pared with conditions on the various oyster beds. 143870—35 2 150 BULLETIN OF THE BUREAU OF FISHERIES SURVEY OF OYSTER BOTTOMS IN AREAS AFFECTED BY OIL WELL POLLUTION 1934 By Robert 0. Smith METHODS The survey vessel was a standard oyster and shrimp lugger, 36 feet in length, with 24 inches draft, and a maximum speed of 10 miles per hour. The vessel had been reconstructed to carry scientific equipment and to provide living accommodations for 4 persons. A total of 423 stations approximately half a mile apart were made. At each station a sample of bottom water was taken for specific gravity; the depth, bottom temperature, and character of the bottom were observed. Occasional surface samples were taken for comparison. An average of 3 minutes was required for these observa- tions at each station. The water samples were taken with a Galtsoff sampler, using a 1-liter Wolff bottle. Depths were measured with a 16-foot sounding pole marked in conformity with Coast and Geodetic Survey practice. Bottom temperatures were read from a Bureau of Fisheries surface thermometer in brass cup case. Bottom water samples at each station were placed in citrate of magnesia bottles and the specific gravities of all were measured by hydrometer at the end of the day’s run. Conversion of specific gravity to salinity was made from Knudsen’s table. The stage of the tide throughout the hydrographic work is given as at Wine Island unless otherwise stated. All locations refer to the Coast and Geodetic Survey progress sketch of Terre- bonne and Timbalier Bays prepared February 1934, under the direction of Lf. W. D. Patterson, chief of party. Prior to this work no accurate chart of the region existed. Without this chart, and the signals erected in preparing it, this survey could not have been made, for there are no natural landmarks, no trees, and only occasional human habitation in the form of fishing camps. GENERAL CONDITIONS The area covered by the survey extended from Timbalier Bay on the east to Pelican Lake on the west, including approximately 400 square miles (fig. 1). The examination of beds and hydrographic survey began February 19, 1934, and continued to March 15. Two days, March 27 and 28, were spent in Barataria Bay and Lake Washington. The bottom over this entire area is exceedingly level. There are very few gullies or reefs except where islands are in the process of being broken down. The bottom, composed of soft black mud and mixed with broken shell, was devoid of vegetation at the time of the survey. On account of the shallowness of the water, rarely over 6 feet, moderate winds churn the bays from top to bottom, so that the water is seldom clear except at the passes during flood tide. The oyster reefs are limited to sections where the bottom is comparatively firm. This condition occurs usually only about the margin of the bays or around islands. The mean range of tide at Wine Island, near the center of the area, is 1.3 feet. Usually there is only 1 high and 1 low tide daily. However, the actual change in EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 151 water level is largely determined by the direction and velocity of the wind. Northerly winds drive the water out, while southerly winds pile it up inside. In either case the difference caused by wind may exceed 1 foot under ordinary conditions and as much as 10 feet during a hurricane. Since the land elevation above mean liigh water seldom exceeds 2 feet, it is evident that even moderate winds result in considerable wash from the marshes. On many of the beds the oysters were found to be quite variable in fatness. In general, it appears that accumulation of glycogen occurred very late during the winter of 1933-34. Many oystermen stated that their oysters were in better condition dur- ing the latter part of March than at any time previously in the season. It is possible that the customary planting level of 600 to 800 barrels (1,500 to 2,000 bushels) per acre is excessive at times for the amount of food available. Two previous surveys of the area made, by Moore in 1898 (1899) and by Moore and Pope from 1906 to 1909 (1910), were not sufficiently detailed to permit direct comparison of hydrographic data. However, it seems safe to say that except for the elimination of extraordinary fluctuations in density resulting from crevasses, the salinity observed in 1934 appears to be much the same as during the 1909 survey. In 1898 the water in Terrebonne Bay was found to be fresher than in Timbalier Bay, an observation not borne out by our determinations, as the salinity now is approximately the same in both bays. It is said that owing to the freshness of the water no oysters were found in Terrebonne Bayou above Lake Lagraisse prior to 1883. It was also stated in 1898 that considerable changes in topography were occurring in Terrebonne Bay. Such changes consisted in the tearing down of large amounts of marsh land then present in Terrebonne and Timbalier Bays, and separating Terre- bonne Bay from Lake Barre. It was at one time possible to go from Houma to Tim- balier Island by land. Undoubtedly this continued destruction of land area has been one of the most important factors bearing on oyster culture in the region, for the absence of obstructing land permits the rapid mixing of Gulf water with the fresh water from the bayous and serves to maintain a relatively high salinity (16 parts per million, February 1934) to the upper parts of Lake Barre and Lake Felicity. At the time of the 1898 survey market oysters were no longer produced in Bara- taria Bay, which includes Lake Washington (Grand Ecaille), and only a few dead reefs existed there as evidence of previous importance. The chief oyster-producing region was Terrebonne Bay, which included Lake Pelto. Since then the beds in Barataria Bay have been rehabilitated and extended so that at present some of the finest oysters are produced here. Although Terrebonne Parish still is foremost in quantity, the quality is in general inferior to Barataria Bay. Insofar as shape, growth, and quality are concerned, observations just made are in almost complete agreement with the early survey, and such discrepancies as exist may be caused in part by confusion in identifying localities from the local names in use at that time. An effort was made to obtain a definite record of sudden and unexplained losses of oysters in the past. No such occurrences were found other than the occasional killing of oysters said to have been caused by freshets due to breaks in the levee. 152 BULLETIN OF THE BUREAU OF FISHERIES LAKE BARRE The lake has an area of about 40 square miles and contains relatively few reefs, mostly on the southern side. Mortality of oysters was reported in 1932-33 only from the southeast side of the bay at Lafont’s camp (fig. 1), about a mile from Barre operations of the Texas Co. The loss was estimated by the fishermen to have been about 75 percent. When visited during February and March 1934, oysters were still dying as evidenced by many clean paired shells and occasional dying or newly dead oysters. The dying individuals were very thin and watery as though they had slowly starved to death. The majority of oysters have the shells honeycombed by boring sponge, boring clam or worm, and the interior of the shell usually has from 1 to 6 mud inclu- sions covered by a thin layer of shell. Out of 40 stations made on 2 days, February 19, from 11a. m. to 6 p. m. and February 20 from 9 to 10 a. m., 14 stations were made on flood tide and 20 on ebb. Salinities varied from 16 parts per thousand at Signal Odor to 32 parts per thousand at the Barre wells of the Texas Co. Isohaline contours (fig. 1) indicate that fresh water from Bayou Terrebonne enters the lake at Seabreeze and is deflected down along the islands formerly paralleling the course of the bayou. Salt water from the Gulf is crowded somewhat to the western side of the bayou by fresh water from Lake Felicity. Temperatures on the bottom at stations on February 19 ranged from 61° F. at 11 a. m. to 60° F. at 5:30 p. m. (16.1°-15.5° C.). On February 20, the temperature had fallen to 55° F. (12.7° C.). LAKE FELICITY AND LAKE CHIEN Depths in this section varied from 4 to 6 feet, with a depth of 7 feet in the channel at Seabreeze. Lake Felicity has an area of about 20 square miles, and Lake Chien, which opens into it on the north, covers about 4 square miles. Although the reefs entirely surround these lakes, production of oysters is mostly limited to their northern and southern shores. Oysters were examined at three places: at the point where Lake Chien joins Lake Felicity; on the northwest shore of Felicity; and on the southeast side of Felicity where it joins Bay Jacko. The oysters observed were not high-grade shell stock, being more suitable for seed or steaming, according to size. No mor- tality was reported as having occurred in these waters, and samples taken in February contained no shells from recently dead individuals. Owing to the low specific gravity, drills are not bothersome, so that the area is valuable for seed production. The entrance to the lake is about 4 miles from the nearest oil wells. Eighteen stations were made in the 2 lakes, from 9:30 a. m. to 3:15 p. m. February 20. Although the survey was carried on mostly during flood tide, the prevailing north wind very probably prevented a normal high tide. The salinity varied from 10 parts per thousand at the head of Lake Chien, to 20 parts per thousand at Grand Pass Felicity on the southwest side of Felicity. The depth varied from 5.5 to 7 feet, but the depth over the reefs around the shore is 2.5 to 4 feet. A depth of 25 feet with a strong current was found in Grand Pass Felicity on the southwest side where Felicity joins Lake Barre. The bottom temperature varied from 54° to 57°F. (12.2°-13.9° C.). EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 153 A number of reefs are said to be present in Bay Jacko, but the water was too shallow for the survey boat. TERREBONNE BAY This bay covers an area of about 100 square miles, exclusive of bayous, small lakes, and bays. Natural reefs are limited by soft bottom to bayous and the shores of islands in the bay. The section affected by mortality lies between Lake Barre and Terrebonne Bay. The heaviest losses in this region occurred in the vicinity of Lafont’s camp, as mentioned above. At the time of this survey, in February 1934, the better oysters were watery. Whether or not this is the usual condition during late winter could not be ascertained. This area is said to have produced large numbers of high quality oysters in the past. At present (1934) most of the oysters are used for steaming. At the 99 stations which were occupied on February 26-27 and March 5 and 8, salinities varied from 22 parts per thousand south of the Texas Co.’s wells, to over 32 parts per thousand near the west end of Timbalier Island, and less than 30 at the east end of Wine Island. The opening between these islands is known locally as Cat Island Pass and is by far the largest pass in this region, being about 5.5 miles across. Isohalines show that the western side of the bay is slightly fresher than the east side, due apparently to drainage from Bayous Terrebonne and Little Caillou. Bottom temperatures on February 26-27 varied from 51° to 59° F. (10.6°-15°C.) ; on March 5 they averaged 64.5° F. (18.0° C.); on March 8 the average was 70° F. (21.1° C.). The depth varied from 4 to 10 feet. The lower part of the bay will average 8 feet although there is a channel up to the Texas Co.’s wells through which 10 feet may be carried. The upper part of the bay is mostly 5 to 6 feet deep. TIMBALIER BAY This bay has an area of approximately 230 square miles, bisected by a string of islands lying in an east to west direction. Practically all of these islands have oyster beds around them. The bottom, except about Philobruis, is soft mud. The quality of the oysters varies greatly in the different sections of the bay. On the eastern side, including Devils Bay north through Jacks Camp Bay to Little Lake, excellent shell stock occurs, though on March 24, at the mouth of Bayou Grey and approximately 5 miles from the Leesville wells, the oysters had a pronounced oily taste. At Philobruis, 9 miles from the Leesville wells, oysters are of good quality for shell stock, no oily flavor could be detected, and no mortality was reported or observed. Practically the entire area between Philobruis and the eastern shore of the bay is covered with oyster reefs, all leased. Sponge, boring clam, and conchs are common. Most of the marketable oysters are between 2 and 3 years of age. The greatest mortality was reported in the vicinity of the islands Castete and Bull, about 14 miles from the Leesville wells, and 9 from the Barre wells. Dredgings made on March 13 consisted of about 90 percent shell. Numerous clean paired shells found at this station indicated that oysters continued to die. As a result of the mortality during the winter of 1932-33, no fresh plantings were made by the lease- holders in 1933. Local setting does not survive, so that seed must be brought from Lake Felicity or from the Louisiana marshes to the east of the Mississippi River. One hundred fifty-five stations were occupied during the 5 days of March 8, 12, 13, 14, and 15. The salinity varied from 22 parts per thousand at Philobruis. 154 BULLETIN OF THE BUREAU OF FISHERIES to 32 parts per thousand at Timbalier Lighthouse (fig. 1). All but 25 samples were taken on flood tide. The bay is very shallow. On the western and southern sides, the depth ranges from 5.5 to 7 feet. The center and northeast sections are shallower, ranging from 3 to 6.5 feet, but much of it on the eastern side does not exceed 4.5 feet. On March 8, the botton temperature averaged 71.5° F. (21.7° C.); March 12, 61° F. (16.1° C.); March 13, 65° F. (18.3° C.); March 14, 62° F. (16.6° C.); March 15, 64.5° F. (18.0° C.). LAKE RACCOURCI Lake Raccourci, which lies north of Timbalier Bay, has an area of approximately 25 square miles. The depth varies from 2.5 to 4.5 feet. The bottom is soft mud mixed with small clam shells. There are no reefs in the center of the lake. A few oysters are to be found about the entire shore line, but the chief beds are at the north- ern end and at Philobruis on the southern boundary. At the time of the survey, March 9, 1934, the beds at the upper end of Lake Raccourci were not being worked. No mortality was observed. The salinity varied from 18 parts per thousand in Bay Courant above Lake Raccourci, to 22 parts per thousand at Philobruis. Boring sponge and boring clam were present. Thirty stations on flood tide and 8 on ebb tide were occupied on March 2, 9, and 14. The bottom temperature on March 2 averaged 63° F. (17.2° C.); on March 9 and 14, the average temperature was 61.5° F. (16.4° C.); and 64.5° F. (18.0° C.), respectively. LAKE PELTO AND PELICAN LAKE Lake Pelto extends along an east and west line from Wine Island on the east where it connects with Terrebonne Bay, to Pelican Lake on the west. It connects with the Gulf at the southwest through Whiskey Island Pass. The lake is about 5 miles wide and covers an area of 50 square miles. Oil wells are located at the eastern end of the lake about equidistant from the north and south shores. The few oysters found in this lake were of good appearance and flavor, and some from the western end were equal to any produced in the region. Extensive losses have been reported throughout the lake except along the northern shores and at the extreme western end. The greatest damage, amounting in certain cases to complete destruction of the beds, has occurred about the centrally located islands and along the southern shore within a radius of 5 miles of the oil wells. Of natural enemies, the boring sponge was most abundant, there were some boring clams and some drills. Most of the oysters in this region have very dark gills. Sixty-two stations were made on Lake Pelto on March 6 and 7. Salinity varied from 26 parts per thousand along the northern shore, to 30 parts per thousand along the southern shore. Bottom temperature varied from 64° to 71° F. (20.0°-21.6° C.) on March 7. The average depth through the center of the lake is about 7 feet. The greatest depth measured was 12.5 feet at Wine Island Pass. Pelican Lake is situated at the northwest end of Lake Pelto. The area is about 9 square miles. The oysters were fairly fat, well flavored and of medium size. No mortality had occurred here according to the oystermen, and no clean shells were found on the reefs. The bayous leading off from the lake are well populated with coon oysters. The planted beds are limited to firm bottom around the shore and about the islands. Boring sponge was abundant, boring clams common, and there EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 155 were a few drills, though the actual abundance of the latter is difficult to determine until they begin to assemble for the spring spawning around the first of April. On March 7, 9 stations, all on flood tide, were made from Bay Bound to the head of Pelican Lake. The salinity varied from less than 24 parts per thousand at the head of the lake to 28 parts per thousand where Bay Round joins Lake Pelto. The average depth in Pelican Lake at three-quarters flood was 3 to 4 feet. In Bay Round, depths of 7 to 9 feet were found. The bottom temperatures on March 7 in Pelican Lake varied from 67° to 69° F. (19.4-19.6° C.). These high temperatures ended March 8, and by the morning of March 9 the temperature was down to 61° F. (16.1° C.) and only rarely exceeded 65° F. (18.3° C.) during the remainder of the month. Production of oil at Lake Pelto was discontinued April 1, 1934. The total pro- duction from this dome is not known, but at the time the survey began (February 1934) only about 40 barrels per day were reported. No oil was produced in Bay Saint Elaine until the first of April 1934, but gas was piped from here to the Lake Pelto wells for fuel. Bay Saint Elaine lies to the north of Lake Pelto at the eastern end. The area is about 4 square miles, but the shape is so irregular and there are so many marsh}7 islands within that the actual extent of the bay is difficult to determine. Few, if any, oysters were being marketed from this bay at the time of the survey. Coon oysters were abundant in places and seemed to grow rapidly. Four stations were made on March 7. Salinities varied from 18 parts per thou- sand, at the head of the bay where bayou Little Caillou empties into it above Coon Road, to 24 parts per million at signal Elaine. The bottom temperature ranged from 69° to 72° F. (20.5°-22.2° C.), an increase of about 2 degrees per mile from the lower to the upper end. The depth was generally about 5 feet, but at station 225 east of signal Elaine the depth was 16 feet. EXAMINATION OF OYSTER BEDS AT MOUTH OF BAYOU GREY AND LITTLE LAKE On March 24, 1934, a trip to the beds was made by oyster lugger from Leesville. Thirteen wells were then in operation along the bayou below Leesville. The surface of the bayou was covered with oil for a distance of 3 miles below the wells. There was a flood tide, with a strong wind from the southeast. Adult oysters, which had been on the bedding grounds 3 months, had a strong oily flavor. There had been some mortality as evidenced by recently dead paired shells. All shells were covered with a brownish black coating of a tarry consistency. When the bottom was stirred by tongs, an oily patch appeared on the surface. The oysters were of good shape but watery. There was very little gonad development. It was stated that ordinarily the oysters were very milky at that time. On March 28, 1934, an examination was made of water conditions in Lake Wash- ington, with reference to their bearing on oyster culture. At the time of the survey, there were 44 sulphur wells in operation, each producing 12 tons of crude sulphur per hour. Very little effluent found its way into the lake. The bottom over a short radius around the point of entry of the effluent was covered with a thick brownish layer of diatoms of undetermined species. There was a strong odor of hydrogen sulphide in the air, but the mud samples had very little odor, though they were quite oily. 156 BULLETIN OF THE BUREAU OF FISHERIES Based on conditions existing on March 28, there was no evidence that the opera- tion of the sulphur wells had an injurious effect on oysters. On July 4, 1933, an oil well was out of control for 36 hours, during which time it was estimated that some 3,000 barrels of oil flowed into the lake. No appreciable quantity of oil was lost subsequently. One of the leases was reported to have been heavily covered with oil, and until November oysters were unmarketable on account of the oily flavor. On March 28, 1934, a combined sample of three-quarters bushel was tonged from these beds. Some of the oysters had a slight oily taste. The tonging caused patches of oil to appear on the surface of the water, indicating that some oil still was held by the mud. CONCLUSIONS The purpose of the survey described above was to supply a knowledge of the local and general factors in the environment of the oyster beds in Terrebonne Parish and adjacent territory which might have a bearing on the problem of pollution. The hydrographic data show that conditions of salinity, current, and tempera- ture were, at the time of the survey, suitable for growing oysters throughout the area covered, and it has not been possible to assign the mortality to any known disturb- ance of the natural conditions on the oyster beds. Bearing this in mind, special attention was given to several factors whose combined effect would tend to magnify the action of any polluting substance. Among these may be mentioned the shallow- ness of the water. Even moderate winds stir the bays from top to bottom so that the water carries much suspended matter. Any polluting substance is quickly and thoroughly mixed with the water and is adsorbed by suspended matter; it may be transported over wide areas and deposited on the bottom far from the source of pollution. In general, mortality has been higher on soft, muddy bottom than on hard ground or reefs. The significance of tins is not known, but there is no evidence that silting is directly responsible for the mortality observed. Probably because of the decomposition of the organic matter, a muddy bottom presents a less favorable habitat for oysters than that found on hard ground. In 1934, oysters on many of the beds throughout the region did not become fat until February or March, which points to a possible scarcity of food organisms during the fall and winter, or to a disturbance in the functioning of the oysters’ organs of feeding. Overcrowding would tend to aggravate this situation. The oystermen state that from 600 to 800 barrels, i. e., 1,500 to 2,000 bushels, are planted per acre. While this quantity may be supported so long as conditions remain normal and the food supply adequate, it is obvious that should mortality begin in such a concentra- tion it is likely to result in the loss of a considerable number of the oysters. No direct evidence was found that the mortality had been caused by any known natural enemies, although they may have an indirect effect by increasing calcium metabolism or competing for food, and, in some cases, possibly a direct injury by attack or the secretion of poisonous substances, for boring clams, sponges, and worms are abundant in parts of Lake Barre, Timbalier Bay, Terrebonne Bay, and Lake Pelto. The borer has not been included in the above consideration of enemies be- cause its habits, range, and destructiveness are well known, the damage done by it is fairly constant, and its depredations can be eliminated as a cause of the unusual mortality. EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 157 Crude oil pollution has been suspected as the chief cause of the mortality. No information is now available upon which to base an opinion as to the validity of this belief. Light films of oil were observed to be generally present in the vicinity of the Lake Barre wells, and below Leesville to the mouth of Bayou Grey. Whether or not this oil was harmful to oysters could not be determined in the short time allotted to the survey. But regardless of any alleged toxicity of oil to oysters, two facts should be borne in mind. First, oil in water is quickly taken up by oysters, imparting an oily taste to the flesh which renders the meat unsalable. Second, the effect of oil pollution will last over a long period, for the oil is carried to the bottom by suspended mud particles and released from time to time by storms, tonging, or dredging. 143870 — 35- •3 158 BULLETIN OF THE BUREAU OF FISHERIES EXPERIMENTAL STUDIES OF THE EFFECT OF OIL ON OYSTERS REVIEW OF THE LITERATURE By Paul S. Galtsoff So far as the author was able to ascertain, the literature on the subject is rather meager, comprising less than a dozen papers. On the basis of a small number of experiments with water-gas tar carried out in 1912 at the United States Bureau of Fisheries station at Woods Hole, Mitchell (1914) arrived at the conclusion that in constantly renewed sea water tar shows no noticeable effect on oysters. Fatal effects are produced however, when considerable quantities of water-gas tar are in intimate contact with oysters kept in stagnant water. Orton (1924) when studying the causes of the unusual mortality among oysters in England during 1920 and 1921, made a few experiments with oil and arrived at the conclusion that the petroleum residue is not seriously poisonous to oysters, as all of them kept in a jar covered with a thick film of oil survived at least 7 weeks. Similar results were obtained in the laboratory of the Scottish Biological Association (Orton, 1924). Examination of water made by the Government chemist (Orton, 1. c., p. 42) showed that the sample of water in which the oysters were kept contained traces of the original petroleum waste in solution and in addition small quantities of substances of an acidic character. The latter, probably naphtenic acids which exist in certain petroleums, may have been present originally in the petroleum waste used in Orton’s experiments or may have been derived from the waste by some chemical or biological action of the sea water. Contrary results were obtained by Leenhardt (1925) who showed that the mortality of both European and Portuguese oysters ( Ostrea edulis and Gryphea angulata ) kept in water covered with petroleum increases in proportion to the amount of the latter. Thus, no mortality was observed in the oysters kept for a month in water containing from 0.01 to 0.05 percent petroleum. When the quantity of petroleum was increased to 0. 1-0.5 percent only 2 or 3 oysters out of a dozen survived. In the water con- taining 2 percent petroleum all oysters died within 1 week. Although Leenhardt states that water was changed and presumably the oysters did not suffer from lack of oxygen it is not clear how often the water was renewed. From the observations that the same mortality occurred in the oysters that never were in direct contact with petroleum as in those which were periodically covered with it, simulating the conditions on tidal flats, Leenhardt concluded that oil contains a substance soluble in sea water, and poisonous to oysters. Unfortunately both Orton’s and Leenhardt reports are very brief and fail to give all the details of the experiments. More comprehensive investigation on the effect of oil on fish was carried out by Roberts (1926) who demonstrated that oil extracts (gas oil, Diesel oil, 600 seconds and 1,500 seconds oil) prepared by shaking 100 cc of oil with 2,000 cc of boiled and filtered river water are toxic to brown trout. He attributed the toxicity to both soluble toxic substances and emulsions. Mention should be made also of the work of Stephen (quoted from Roberts, 1926) who described the adverse effect caused by a film of paraffin 0.03 mm in thickness on pelagic larvae of plaice and flounders; and of the experiments of Jee and Roberts (1923) with fresh-water shrimp killed by contact with fuel oil, and with caddis larvae and fresh-water shrimp killed by EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 159 immersion in a fuel oil extract. Gardiner (1927) found that the aqueous extracts of tar were completely without effect on freshly fertilized eggs and on “eyed” ova of trout, and that the resistance of trout alevins to the toxic action of phenol decreased with age. Alevins 60-66 days old were quite unaffected by 2-hour immersion in a solution of 40/100,000 phenol, whilst fry 110-115 days old were unable to withstand more than 15 minutes immersion. Elmhirst (1922) found that oil in water had no effect on a number of bottom organisms (Purpurea, Serpula, Authozoa, Ascidia, Mollusca) and many planktonic forms although the latter were lulled when they came in direct contact with oil. According to James (1926), under conditions approaching those prevailing in nature, a 0.1 percent dilution of light or heavy bunker oil exerts no effect on the development of cod eggs, but toxic effect of the same oils becomes apparent under conditions of limited circulation and aeration. Slight injurious effect is exerted by the same oils upon larval flounders. He rightly states that further research should be based upon more specific knowledge of the physical and chemical phenomena accompanying oil pollution. Considerable discrepancies in the results of the experiments mentioned in this brief revue may be attributed to various causes, namely, differences in the chemical composition of oils used, defects in experimental technique and small number of observations. It has been the purpose of the present experiments to obtain a better understanding of the effect of oil and brine on oysters and to carry them on a more comprehensive scale than has been done before. The experimental investigation comprises two distinct parts ; a study of the effect of oil and bleed water on mortality, rate of feeding and behavior of the adductor muscle of the oyster; and a study of the effect of oil on the growth of a diatom. For the latter experiments the culture of a single species of Nitzchia clostearia E. was used. It is believed that the results obtained with this planktonic species which occurs in the normal habitat of the oyster, are applicable to other planktonic diatoms. Most of the experiments on oysters were carried out at the United States Fisheries laboratory at Beaufort, N. C. Studies on diatom growth were made at the Bureau of Fisheries laboratory in Washington and Woods Hole, Mass. Since the experi- ments carried out by the authors were performed at different times and under differ- ent conditions, the results of their investigations are presented separately under their respective names. SURVIVAL OF OYSTERS IN OIL-POLLUTED WATER By Herbert F. Prythercii The plan and purpose of these experiments was to expose oysters to high con- centrations of the principal polluting substances from the oil wells in order to determine their effect on growth and survival. The oysters were subjected to crude petroleum, sludge, brine-effluent water, and hydrogen sulphide, samples of which, with the exception of H2S, were collected previously at the wells by the author and stored in glass containers. These studies were conducted at the United States Fisheries Biological Station at Beaufort, N. C., and involved the use of several 160 BULLETIN OF THE BUREAU OF FISHERIES different methods and apparatus, a description of which is furnished with each series of experiments. Special studies were also made of the changes in glycogen content of oysters after continuous exposure to three different grades of crude petroleum and varying concentrations of brine water from the Barre wells. The glycogen analyses were performed by P. S. Galtsoff. EXPERIMENTS WITH SURFACE FILM OF OIL The shell movements of the oyster indicate whether it may or may not be feeding and also reflect quite accurately its reactions to environmental conditions. Graphic records were obtained of the shell movements of one representative oyster in each of the experimental and control jars. For this purpose adult oysters of approximately the same age and size were immobilized in a horizontal position by cementing the lower valve to a brick or piece of slate. The upper valve was free to move, and, by means of a simple arrangement of levers and pen, its movements were transmitted and recorded continuously on a revolving smoked drum as shown in figure 2. Figure 2.— Diagram showing arrangement of apparatus for obtaining comparative records of shell movements of oysters in sea water (control) and in sea water passed through oil (experimental). A, light muscle lever; B, aluminum pen; C, counterbal- ance weight; D, constant level arrangement; E, rubber universal joint and light rod connecting upper valve of oyster to recording pen; F, oyster cemented to brick (control); G, kymograph; H, layer of oil; and I, experimental oyster. A series of experiments was arranged, employing 5 glass jars of 6-liter capacity, in each of which 2 small and 2 large adult oysters were placed, 1 of the latter being attached to the recording apparatus. Four of the jars were covered with a heavy surface layer (50 cc) of the following oils and sludge — grade A and grade B crude petroleum from the Barre wells, a composite sample of crude petroleum from the Pelto wells, and basic sludge collected from the storage tanks at the Barre wells. Each experimental jar was supplied with running sea water, which was introduced above the oil film at the rate of 8 liters per hour as shown in figure 2. This series of experiments was conducted from July 3 to September 4, 1933, during which time the surface layer of oil and sludge was renewed weekly. The temperature of the water ranged from 24.8° C. to 28.5° C., and the salinity from 32.7 to 33.4. The results of these studies are shown in table 1. All of the large adult oysters survived, and those attached to the recorder and exposed to oil and sludge showed essentially the same behavior as the control in respect to shell movements and ability EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 161 to maintain closure when kept in air for over 72 hours after completion of the ex- periment. Records of shell movements of these oysters made for 2 days before exposure to oil and sludge showed no noticeable differences in comparison with those obtained during and at the end of the experiment. Examination of table 1 shows that the experimental oysters were open on an average of 10 to 13.6 hours daily as compared with 11.2 hours for the control specimens. This difference, however, is not significant if allowance is made for the individual variations in oysters in respect to duration of open periods. For example, in other experiments 16 control oysters in running sea water (temperature 22° to 30° C.) showed an average daily open period ranging from 7.5 to 14.2 hours. Hopkins (1931) states that oysters at Beaufort, N. C., averaged between 10 and 14 hours per day open in running water. Table 1. — Length of time oysters remained open in running sea water passed through surface layers of oil and sludge [July 3 to Sept. 4, 1933] Specimen Medium Total hours open Average number of hours per day open Percent of time open 719 11.2 46.8 Running sea water — Barre oil A 783 12.2 51. 0 Running sea water — Barre oilB 664 10.3 43.2 641 10. 0 41. 7 Running sea water — Barre sludge 870 13.6 56. 6 During the last 2 weeks of the experiments 7 of the small adult oysters died, the losses occurring as follows: Control, 2; Barre oil A, 1; Barre oil B, 2; Pelto oil, 2; Barre sludge, 0. Similar losses have occurred in previous experiments not involving oil pollution and are believed to be due to the fact that these oysters were collected from a densely populated bed near the laboratory and consequently were not in as good condition as the larger oysters which were obtained from planted beds. SURVIVAL OF OYSTERS IN SEA WATER PASSED THROUGH OIL According to Gowanlocb (1934) the toxicity of Louisiana crude petroleum can be demonstrated by continuously exposing oysters to sea water passed through a heavy layer of oil. A similar series of experiments was conducted at the Beaufort laboratory using a large wooden tank having 4 watertight compartments in each of which were placed 25 small, adult “coon” oysters from a bed adjacent to the laboratory and 25 large oysters obtained from planted beds in Newport River. The arrangement of the equipment in one of experimental compartments is shown diagrammatically in figure 3. One compartment was used as a control and the o'ther 3 supplied with the same amount of running water at the rate of 20 liters per hour, which was passed through 1 liter samples of the following oils: Barre grade A, Barre grade B, and Pelto composite. The experiment was in operation from March 26 to July 27, 1934. At the completion of the experiment it was found that there was little difference in the survival of the large adult oysters in the control and experimental jars, though the small oysters exposed to oil showed a slightly greater mortality than the controls. The results of this series of experiments are given in the following table (table 2). In reviewing the results, consideration must be given to the possibility that certain compounds in the oil may reduce the oxygen content of the inflowing water and that the oil itself may 162 BULLETIN OF THE BUREAU OF FISHERIES act as a filter and prevent many food organisms from reaching the oysters in the experi- mental tanks. Figure 3.— Cross section of apparatus used to expose oysters to sea water passed through oil and smudge. A, sea water intake; B, glass cylinder containing 1 liter of oil; C, siphon tube; D, oysters; and E, overflow. Table 2. — Survival of oysters in sea water passed through oil Total number of live oysters Percent survival ac- cording to size Apr. 30 May 30 June 30 July 27 Large Small Tank no. 1 (control— no oil) 50 39 29 23 (12 large, 11 small) 48 44 50 37 24 18 (11 large, 7 small) 44 28 Tank no. 3, Barre oil B 50 38 29 21 (11 large, 10 small).. _ _ 44 40 Tank no. 4, Pelto oil.. 50 30 25 do 44 40 IMMERSION OF OYSTERS IN OIL The purpose of these experiments was to immerse oysters in crude petroleum at regular intervals in order to determine if they would be killed by direct exposure to oil or would subsequently recover from such severe treatment. By means of an artificial Figure 4.— Diagram of apparatus used to immerse oyster in oil at regular intervals. A, control jar; B, automatic siphon jar; C-F. experimental jars containing different grades of petroleum and sludge; Q, sea water intake; H, oysters (4 per jar); I, vertical movement of oil layer during each artificial tidal cycle. tidal arrangement a heavy surface film of oil was made to cover completely the oysters at intervals of 1 and 2 hours. The equipment used, as shown in figure 4, consisted of EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 163 5 tall hatching jars of 5 liter capacity, in each of which 4 adult oysters were elevated above the bottom by a 3-incli layer of clean oyster shell. An additional jar containing an automatic siphon arrangement was connected with each hatching jar by glass and rubber tubing so as to bring about the filling and emptying of these jars at regular intervals. One control and 4 experimental jars were used in each series of experiments, each of the latter containing a heavy surface layer (50 cc) of the following: Barre oil grade A, Barre oil grade B, Pelto oil (composite sample), and Barre sludge. After the experiments had been in progress for approximately 3 weeks, 50 cc of each of the above oils and sludge were added to each experimental jar, respectively. Four separate and complete series of experiments were conducted during the period from September 20, 1933, to January 3, 1934, in which a total of 64 oysters were com- pletely immersed in oil at intervals of 1 and 2 hours over a period of 6 to 8 weeks. The results of these studies are given in table 3. In the first series (1 hour interval), which was conducted in the early fall at fairly high water temperatures, there was a loss of 31.25 percent of the oysters exposed to oil and 50 percent of the control oysters. In the second series the losses in the control and experimental jars were the same amounting to 25 percent. In the third and fourth series (2 hours exposure interval) not a single specimen died in either the control or experimental jars or during the subsequent period of 3 months when they were returned to natural conditions in the harbor. During the course of the experiments it was frequently observed that the oysters kept the shell open when covered with oil and that the formation of new shell was much greater in the experimental oysters than in the controls. In order to test further the toxicity of water contaminated with oil and sludge, the overflow from all jars used in series 1 and 2 was passed into a tank containing 22 seed oysters, 22 clams (3 species), 4 gastropods (2 species), and 6 anemones, none of which died during the course of the experiments. In the third and fourth series the overflow water was passed into a tank containing approximately the same number and kind of marine organisms as the former and in addition 6 small fish (Fundulus, Hypso- blennius), all of which survived and appeared to be in a healthy condition. Table 3. — Survival of oysters after immersion in oil and sludge at regular intervals of 1 and 2 hours over periods of 6 to 8 weeks Conditions Series no. 1 (Sept. 20 to Oct. 31) Series no. 2 (Sept. 20 to Oct. 31) Series nos. 3 and 4 (Nov. 2 to Jan. 3) Total results (Sept. 20 to Jan. 3) Control Experi- mental Control Experi- mental Control Experi- mental Control Experi- mental Number of oysters used... 4 16 4 16 8 32 16 64 Number of times ovsters immersed in oil 0 1,008 0 1,008 0 720 0 720-1,008 Number oysters alive at end of experiment-. 2 11 3 12 8 32 13 55 Percent survival by jar: 50 75 100 81. 2 75 75 100 87. 5 75 50 100 81. 2 50 75 100 81. 2 75 100 100 93.7 Percent survival, total 50 68.7 75 75 100 100 81.2 85.9 EFFECT OF OIL ON GLYCOGEN CONTENT OF OYSTERS This series of experiments was conducted to determine the effect of 3 different grades of crude petroleum on the glycogen content of oysters over a period of 4 weeks (Dec. 6, 1933, to Jan. 4, 1934). Two hundred and ten oysters of uniform age and size were used, of which a representative sample of 10 oysters was taken for glycogen 164 BULLETIN OF THE BUREAU OF FISHERIES i analysis at the beginning of the experiment. The experiments were conducted in a specially constructed wood tank having 4 compartments of equal size in each of which 50 oysters were placed. Each compartment had a capacity of 135 liters and, by means of an automatic siphon arrangement similar to that shown in figure 7, received an equal amount of sea water, which was completely renewed every 3 hours during the course of the experiment. One compartment was used as a control and the other 3 covered with a heavy layer of crude petroleum (2 liters each) of the following grades: Barre oil, grade B; Barre oil, grade A; and Pelto oil composite sample. The oil completely covered the surface of the 3 experimental tanks and during each low-water interval was approximately one-half inch above the oysters. At the end of each week 10 oysters were taken from each compartment and analyzed for glycogen content by P. S. Galtsoff. The results of this experiment, given in the following table, indicate a slight de- crease in the glycogen content of oysters kept in oil-polluted water. Table 4. — Effect of heavy surface layer of oil on glycogen content of oysters [Experiment G 1] Percent glycogen, fresh basis Dec. 6 Dec. 13 Dec. 20 Dec. 28 Jan. 4 Average 1.72 2.81 2.68 2. 67 1.62 2. 30 (1.72) (1.72) (1.72) 1.89 1.82 2. 56 2. 54 2. 10 1.62 1. 57 1.98 2. 22 1.82 1.69 1.42 2.29 1.80 EXPERIMENTS WITH BRINE The principal effluent from the oil wells in Lake Barre and Lake Pelto is the brine water that is separated from the petroleum. The quantity discharged daily, at the time of the oyster mortality, varied from approximately 5,000 to 7,000 barrels for the Barre wells and from 650 to 750 barrels for the Pelto wells. The composition of pe- troleum brines, according to Clark (1924), has been found in manj^ instances to be quite similar to that of ocean water although modified by local conditions and differ- ing in concentration. Such waters have been variously interpreted, sometimes as fossil sea water which was entrapped in the original sediments, and sometimes as derived by leaching from beds of salt. Analyses of the brines from the Barre and Pelto wells are given in table 5, together with those of 3 samples of sea water which were collected from the following localities: (1) Ctyster bed of St. Pierre, west of Barre walls, where a high mortality of oysters occurred, (2) oyster bed at Sea Breeze where no mortality occurred, and (3) Gulf of Mexico in ship channel at Buoy No. 3. These analyses, prepared by the Port Arthur laboratory of the Texas Co., show that the brines are quite similar to the sea water found in this region in respect to the presence and relative amounts of the different salts but differ as to their concentration. The samples of Barre brine that were used in the subsequent experiments showed a salt concentration of from 122.8 to 123.5 parts per thousand, while those from the Pelto wells varied from 97.0 to 98.5 parts per thousand. The salt content of the water over the oyster beds is considerably lower and was found, during the field survey in May 1933, to vary from 15.41 to 22.77 parts per thousand, with an average of 18.69. EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 165 Table 5. — Analyses of brines from oil wells at Lake Barre and Lake Pelto, and sea water from oyster beds and Gulf of Mexico Constituents Barre brine Pelto brine Oyster bed, Lake Barre Oyster bed, Sea Breeze Gulf of Mexico 7.0 7.5 8. 1 7.9 8.7 Specific gravity 60° F./600 F.1 1. 109 1. 0735 1.016 1.009 1. 016 Suspended matter l__ Parts per 100,000 5.6 Parts per 100,000 3. 70 Parts per 100,000 1.0 Parts per 100,000 1.2 Parts per 100,000 1.3 Ash of suspended matter 1 1.5 .40 .4 .7 Iodine adsorption as hydrogen sulphide (H2S)1 .29 5. 34 .046 .039 . 12 Alkalinity as calcium carbonate (CaCOa)1 33.5 31.2 10.8 11.2 9.9 Silica (Sib2)2 ' 1 14.0 6.4 5. 0 2.5 1. 5 Iron and aluminum oxides (R2O3)2 20.9 68.0 8.3 4.9 1.7 Manganese oxide (MnsOO2 1. 1 .7 .6 .3 Calcium oxide (CaO)2„ _ 474.0 395. 0 34.0 24.0 38.0 Magnesium oxide (MgO)2 225.0 320.0 123.8 86.4 125.3 Alkalies calculated as sodium oxide (Na20)a„ 7, 190. 0 24.8 4, 460. 0 27.0 735.5 570.6 873.4 Sulphur trioxide (SO3)2 130.0 88.2 131.5 Chlorides as chlorine (Cl)2 9, 831. 5 9, 526. 0 1,151.3 786.6 1,150.9 Silica (SiOa)3 14.0 6.4 5.0 2.5 1.5 Iron and aluminum oxides (R2O3)3 20.9 68.0 8.3 4.9 1.7 Manganese oxide (MmOi)3 1. 1 .7 .6 .3 Calcium chloride (CaCh)3. 938.2 744.6 67.3 47.5 75.2 Calcium sulphate (CaCO-i)3--- 44.9 Magnesium chloride (MgCh)3 532.0 756.0 292.4 204.1 296.0 Sodium chloride (NaCl)3'. 13, 523. 7 36.3 9, 526. 0 1, 197. 2 947.3 1, 455. 0 233.1 Sodium sulphate (Na2S04)3 230.6 156.4 632.4 164.4 29.5 Total of calculated solids 15, 698. 6 11, 145.9 1, 965. 9 1, 392. 8 2, 062. 8 Residue on evaporation (160° C.) 16, 370. 0 15, 804. 0 19, 484. 0 11, 124.0 2, 120. 0 1, 908. 0 2, 427. 0 1,419.0 2, 050. 5 Residue after gentle ignition 10, 238. 0 12, 380. 0 1,281.5 1,914.5 Mineral matter as sulphates 1, 672. 5 2, 471. 5 ' On sample as received. 2 On sample after boiling, filtering, and making to original volume with distilled water. 3 Probably combined as indicated. The purpose of the present experiments was to determine the effect of high con- centrations of brine on the shell movements and survival of oysters. Nine large adult oysters were used, each of which was cemented to a brick by the lower valve and connected with a graphic recording apparatus as shown in figure 2, in order that an accurate record of their reactions might be obtained. Each oyster was placed in a shallow glass jar and covered with 4,000 cc of either sea water or a brine-sea water mixture, which was renewed daily. The experiments were conducted for a period of over 3 months — from January 12 to April 16, 1934. In the experimental jars the oysters were subjected to different mixtures of brine and sea water which contained 12.5, 25, 50, and 100 parts per thousand of Barre brine and 25 and 50 parts of Pelto brine. The former mixtures contained, 1.53, 3.07, 6.15, and 12.30 grams of brine salt per liter and the latter 2.45 and 4.90 grams. During the course of the experiments the average salinity of these solutions was respectively as follows: 33.30, 34.44, 36.71, and 41.25 for those containing Barre brine, and 33.81 and 35.46 for those containing Pelto brine. In the control jars the oysters were kept in normal sea water having an average salinity of 32.17. The temperature of the water in the experimental and control jars was the same and ranged from 4.2° to 23.6° C., while these studies were in progress. Temperature fluctuations of as much as 5 to 13 degrees occurred daily as each new supply of cold harbor water became warmed to room temperature. The results of the experiments are shown in table 6. None of the oysters died while exposed to brine or during a subsequent period of 1 month when they were kept in running sea water. An examination of the data presented in table 6 shows that oysters remain open for longer periods of time in sea water to which brine has been 143S70 — 35 1 166 BULLETIN OF THE BUREAU OF FISHERIES added and that this reaction can be correlated with the amount of brine present and salt content of the water. The three control specimens were open an average of 7.26 hours daily as compared with daily open periods ranging from 8.5 to 11.8 hours in experimental specimens subjected to brine in amounts varying from 12.5 to 100 parts per thousand. In many instances the brine was added when the oysters were open and had been feeding for a short period, which caused one or two partial contractions of the adduc- tor muscle but did not produce closure of the shell or appear to have any noticeable effect upon its subsequent movements. With the exception of specimen no. 7 all of the experimental oysters showed normal records as to the number of shell move- ments or muscular contractions per hour as compared with their respective records prior to the first addition of brine, and those of the control specimens. In the case of specimen no. 7 there were frequent, nearly complete, contractions of the muscle which might be attributed not only to the higher concentration of brine but also to the high salinity of the water which was considerably greater than that found in the natural environment of the oyster. In spite of such abnormal muscular activity this oyster was able to maintain shell closure for periods ranging from 24 to over 96 hours toward the end of the experiment and subsequently recovered completely in running water. Table 6. — Comparison of length of time oysters remained open in normal sea water and brine-sea water mixture Specimen Medium Average salinity Total hours open Average number of hours per day Percent of time open CONTROL No. 1 Normal sea water 32. 17 648 7.2 30.0 No. 2 do 32. 17 806 8.9 37.0 No. 3 do 32. 17 513 5.7 23.7 EXPERIMENTAL BARRE BRINE 12.5 parts per thousand-. 33.30 927 10.3 42.9 25 parts per thousand 34. 44 1,026 1,053 11.4 47.5 50 parts per thousand 36.71 11.7 48.7 No. 7 - 100 parts per thousand 41.25 1,062 11.8 49.1 PELTO BRINE 25 parts per thousand. 33. 81 765 8.5 35.4 50 parts per thousand 35. 46 819 9.1 37.9 One purpose of the experiments with brine was to determine its effect upon the holding power of the oyster muscle, since loss of this function was apparently the immediate cause of the oyster mortality in Terrebonne Parish. Consequently all of the specimens were kept in air for a period of 96 hours, at an average temperature of 19° C., after completion of the experiments during which time they were able to keep the shells tightly closed and later showed no serious effects of such treatment when returned to running water. This test was again repeated 10 days later with the same result. The previous experiments indicate that the brine waters from the Lake Barre and Lake Pelto oil wells do not affect the muscular mechanism of the oysters in relatively high concentrations provided the quantity present does not increase the salinity beyond the limits favorable for the growth of tiffs shellfish. Since these effluents are greatly diluted before reaching the oyster beds, and since no significant EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 167 increase in the salt content of the water was found on these beds during previous field investigations, it is probable that the oysters in this region were exposed to much lower concentrations of brine than those used in the experiments. This is also indicated by the fact that the Barre brine, in concentrations of 50 parts per thousand, was found to be toxic to the boring sponge, whereas a prolific growth of this organism was found on a high percentage of the oysters in the vicinity of the oil wells. EFFECT OF BRINE ON GLYCOGEN CONTENT OF OYSTERS During the period from February 7 to March 8, 1934, studies were made of the changes in glycogen content of oysters kept in sea water to which brine water from the Barre wells had been added in varying amounts. Two hundred and ten oysters were used, of which a representative sample of 10 oysters was taken to determine their glycogen content at the beginning of the experiment. For this experiment a 4-compartment wooden tank was arranged as follows: Compartment 1, 50 oysters as control in 135 liters of sea water; compartment 2, 50 oysters in a mixture of 130 liters of sea water plus 5 liters of brine; compartment 3, 50 oysters in a mixture of 125 liters of sea water plus 10 liters of brine; and compartment 4, 50 oysters in a mixture of 120 liters of sea water plus 15 liters of brine. The water in all compartments was changed once each week and continuously aerated during the course of the experiment. The salinity of the water in the different compartments was as follows: Compartment 1 (control), salinity 31; compartment 2, salinity 34.5; compartment 3, salinity 38.7; and compartment 4, salinity 42.8. At the end of each week 10 oysters were taken from each compartment and analyzed for glycogen content by P. S. Galtsoff. The results obtained are given in the following table: Table 7. — Effect of different concentrations of Barre brine on glycogen content of oysters [Experiment G 2] Date Percent glycogen, fresh basis Feb. 7 Feb. 14 Feb. 23 Mar. 1 Mar. 8 Average Control 2. 16 3. 33 2. 74 1.93 4. 36 2. 90 Brine, 5 liters ... (2. 16) (2. 16) (2. 16) 2. 66 4. 77 2. 92 3. 24 3. 15 Brine, 10 liters 3. 06 3. 38 5. 26 2. 92 3. 35 1.71 3. 15 2. 65 2. 25 2. 38 THE EFFECT OF OIL ON FEEDING OF OYSTERS By Paul S. Galtsoff and R. O. Smith EFFECT UPON THE ADDUCTOR MUSCLE It has been established that oysters have a well-developed chemical sense and are sensitive to a wide variety of chemical substances (Hopkins, 1932). When an irritating chemical solution is brought in contact with the tentacles, situated along the free border of the mantle at the edge of the shell, they retract sharply. The reaction may spread to the mantle, which contracts, and to the adductor muscle, the response of which to stimulation according to Hopkins (1932) “bears a relation- ship to concentration similar to that of the tentacular reaction, but the reaction time is longer.” 168 BULLETIN OF THE BUREAU OF FISHERIES Sharp contraction of the adductor muscle caused by a sufficiently strong con- centration of an irritating substance produces a twofold effect: The oyster snaps its valves to expell the irritating substance from the inhalent chamber and then keeps the valves tightly closed to protect itself from further irritation or injury. Since feeding can take place only when the muscle is relaxed and the valves are open, the number of hours the oyster remains open determines to a certain extent the dura- tion of feeding. One must bear in mind that the principal organs of feeding of an oyster are the gills, the function of which consists of filtering large quantities of water and in carrying microscopic food particles toward the mouth. (For a detailed de- scription of the function of gills see Galtsoff, 1928.) Obviously the gills can function only when the valves are open. The presence of a toxic or irritating substance may affect the adductor muscle which will cause the oyster to close, therefore reducing the number of feeding hours and the quantity of food consumed by it, or it may have direct harmful effect on the delicate ciliary mechanism of the gill. In both cases the feeding of the organism is impaired. The effect of oil on ciliary activity of the oyster is discussed later. At present we are interested only in its effect on the muscular activity. Records were obtained of the behavior of the adductor muscle of oysters sub- jected to Pelto oil. Immobilized oysters were attached to 24-hour recording instru- ments in such manner that every shell movement was transmitted and reproduced graphically upon charts which were divided into hours. The number of hours of activity and the number of closed hours were counted for each day. The running water, supplied to 4-liter dishes containing 2 mounted oysters, flowed through a layer of oil and out from a siphon. Fifty cc of Pelto oil was used to form the surface layer. As the force of the inflow caused globules to be constantly forced down into the water, much of the oil driven into the water gradually adhered to the sides of the dish and to the mounted oyster. Some oil was also lost occasion- ally through the outflow, and fresh oil, therefore, frequently was added so that a heavy surface film was always maintained. The oysters were so placed in the dishes of running water that the upper valve was less than an inch below the surface of the oil. Table 8. — Average number of hours oysters remain open IN RUNNING SEA WATER UNDER OIL Oyster no. Dates Days Hours Temper- ature range Oyster no. Dates Days Hours Temper- ature range 1933 0 C. 1933 0 C. I3y 12 8.6 23-26 406... Aug. 18-25. 7 10.8 25-28 Do July 12-23... 11 12. 1 22-25 Do Aug. 25-29 4 8.9 28-30 Do July 23- Aug. 5 13 8. 5 27-30 Do Aug. 29-Sept. 9 7 11.0 26-29 Do... Aug. 5-12 7 12.2 24-27 Do Sept. 9-15. 6 11.6 28-30 13x June 31-July 12 11 10.8 23-26 Do Sept. 15-20 5 12.5 25-28 Do July 12-19 7 9.5 22-25 Do Sept. 20-Oet. 2 12 8. 5 24-27 Do._ . July 19-Aug. 2 14 16.6 25-28 169. Aug. 8-17 9 8.2 26-29 354.. July 10-23..”. 13 7.3 23-26 Do Aug. 17-24 7 10.5 24-27 Do 12 9.9 26-29 Do Aug. 24-Sept. 5 8 11.0 26-29 Do Aug. 4-15. 12 8.5 24-27 Do Sept. 5-13 8 10.2 24-27 Do.. . Aug. 16-27 11 8.6 25-28 Do Aug. 18-24 6 10.2 24-27 Do Aug. 27-Sept. 14___ 14 10.3 26-29 Do Aug. 24-Sept. 7 n 7. 1 26-29 Do Sept. 14-261 10 11.5 24-27 Do Sept. 7-14. 7 8.3 27-30 150__ Juiy 18-24 6 12.4 25-28 Do. Sept. 14-29 13 8.2 24-27 Do Do Aug. 3-17 14 13! 1 24-27 7 292 1 10. 5 22-30 408 Aug. 13-18 5 8.2 26-29 1 Average. EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 169 Table 8 — Average number of hours oysters remain open — Continued IN RUNNING SEA WATER Oyster no. Dates Days Hours Temper- ature range Oyster no. Dates Days Hours Temper- ature range 1931 ° C. 1931 ° C. 119 Julv 19-27 .. 8 11. 2 28-30 84 Sept. 2-10 5 10. 0 22-25 Do . 8 12.4 27-30 Do Sept. 10-21 7 9.0 23-26 Do 8 9.0 27-30 Do Sept. 21-29 6 7.0 22-25 Do 8 9. 1 27-30 121 Sept. 2-10.. 8 12.0 22-25 Do.... 6 6. 2 27-30 Do __ Sept. 10—18 _ 8 11.5 23-26 Do.. . 7 7.0 28-30 Do Sept. 19-27.. 8 10.0 22-25 Do ... 4 7.0 22-25 311 Sept. 2-10 8 10.0 22-25 Sept. 9-18 8 6.0 22-25 Do Sept. 10-18 8 6.4 23-26 Do Sept. 18-26 8 6.0 22-25 Do Sept. 19-27 8 6.0 22-25 170 May 31-Juno 5 s 14.0 22-2.5 Do June 5-13 8 10.4 24-27 1932 Do.... 9 11. 1 24-27 68 July 25-Aug. 1 8 12.0 27-30 Do... 5 9. 2 26-29 Do Aug. 1-9 8 12.4 27-30 Do . . 6 9.0 27-30 Do Aug. 9-17 8 12. 1 27-30 Do 6 12. 1 28-30 Do Aug. 17-25 8 9.0 28-30 P Aug. 21-25 4 10. 7 24-27 121 June 20-27 7 8.3 25-28 R 4 11.9 24-27 Do 7 9.3 27-30 84 . July 19-27 8 8.5 28-30 Do July 4-10 6 11. 7 28-30 Do . 8 13.0 27-30 124 June 25-July 2 7 15. 0 27-30 Do . 8 14. 0 27-30 Do.... July 2-10...’ 8 14.3 28-30 Bn \ np 19 19 7 10 0 9.7 30 Do Aug. 19-2G 7 10.0 27-30 10. 292 i 9. 6 22-30 Do. Aug. 26-Sept. 2 7 10.4 28-30 1 Average. The behavior of the oysters left in running water under oil was compared with that of the specimens kept in clear sea water under the same conditions — mounting, attachments to recording apparatus, volume of dishes, rate of flow of water, and temperature. In view of the finding by Hopkins (1931) that temperature is one of the factors determining the length of time during which oysters remain open, the comparison between normal and experimental oysters was made for the same ranges of tem- perature. In presenting the results of the experiments (table 8) the average number of hours of activity of oysters during June, July, August, and September of 1931 and 1932 was compared with the activity of oysters left under oil in July and September 1933. By examining table 8 one can detect no significant difference in muscle behavior of the two groups of oysters. The number of hours of activity per day varied between 8.9 and 12.2 in the oysters kept under oil and between 8.8 and 11 in the untreated oysters. No significant differences are apparent in the average of 292 days of untreated oysters compared with the averages of 292 days of casters kept under oil. The average numbers of hours of activity per day for the whole range of temperature 22°-30° C. were 9.6 hours for the untreated and 10.5 hours for the experimental oysters. The question may arise that the difference in the muscular activity of oysters kept under oil and under normal laboratory conditions was unnoticeable because of the wide temperature fluctuations during the experiment. That this is not the fact can be seen from table 9 in which the average number of hours the oysters were open are computed for various temperature ranges. On account of daily temperature fluctua- tions in the laboratory water supply it was necessary to group the results of the observations into six overlapping classes. The results indicate very clearly that under the conditions of the experiments oil had no effect on the number of hours the oysters were open. 170 BULLETIN OF THE BUREAU OF FISHERIES Table 9. — Average number of hours of activity of the oysters Temperature ranges, 0 C. Item 22-25 23-26 24-27 25-28 26-29 27-30 In running sea water, under oil - 10.8 8.9 10.3 12.2 10.5 10.2 In running sea water 8.8 9.0 11.0 8.3 9.2 10.2 In another set of experiments, simultaneous observations were taken on oysters kept under oil and in the control tanks. In these experiments water was supplied at the uniform rate of flow of 6 liters per hour to each of the oysters kept in glass aquaria tanks. Experimental tanks contained enough oil to make a layer 1 centimeter thick. The amount of oil varied from 300 to 350 cc. Water was not permitted to pass through the oil layer, both the intake and out-take tube of the siphon being kept under it. The results presented in table 10 are similar to other experiments. There was no significant difference between the behavior of the oysters under oil and in the controls, the average number of hours of activity being 11.2 and 11.8 respectively. Table 10. — Average number of hours of activity of oysters under oil and in controls Experiment number Date Number of days Temperature range (° C.) Average number hours per day open Kind of oil Experimental Control 1 Oct. 11-30 20 16-22 5.75 7. 25 Pelto. 2 Oct. 19-24 6 18-22 11.9 9.5 Do. 3 Oct. 26-31 6 16-20 12.2 16.0 Do. 4 Nov. 1-7 7 18-21 11.6 12.8 Do. 5 Nov. 11-16 6 11-18 16.6 10.6 Do. 6 Nov. 19-25 7 11-17 9.6 10. 6 Barre. 7 __ Dec. 7-16 10 10-15. 5 13.8 17.2 Do. 8 12 10-17 8.5 10.5 Barre sludge. i 74 11.24 11.83 ' Total days. The results of these observations show that presence of oil in the water failed to interfere with the muscular activity of the oyster and did not reduce the duration of their feeding. EFFECT OF OIL AND OIL WELL BLEED WATER ON THE RATE OF FEEDING OF OYSTERS The purpose of the experiments here described was to determine the effect of various concentrations of the water soluble fraction from crude oil and oil well bleed water on the rate of feeding of oysters. The importance of this matter from a practical standpoint can scarcely be overestimated. Growth and fattening of oysters depend on three major conditions: 1. Abundance of food organisms in the water; 2, percentage of time the shell of the oyster is open; and 3, rate of flow of food-bearing water through the gills. It is obvious that no matter how abundant food may be in the water, or how long the shell is open, very little food will be available to the oyster unless a current is maintained through the gills. Consequently, any substance which slows down the filtering activity of the gills acts to reduce the quantity of nourishment available to the oyster. In an extreme condition, a substance which reduces the rate of flow might eventually cause the death of oysters without ever being directly toxic. EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 171 All experiments reported below were made during November and December 1933 and from June through September 1934. It was thus possible to observe the reaction of oysters to the experimental fluids over a temperature range of from 10 to 30 degrees centigrade and water salinity varying from 27.5 to 36.6 parts per million. The crude oil, collected by the Louisiana Department of Conservation from the Lake Barre and Lake Pelto wells of the Texas Co. in Terrebonne Parish, was shipped to the laboratory and kept in glass containers. The feeding activity of the oyster was determined by measuring the rate, and quantity of water pumped through the gills by the ciliated epithelium. CARMINE CONE METHOD The carmine-cone and the drop-counting methods developed by P. S. Galtsoff were used. The data obtained by these methods may be expressed as volume of water pumped or work performed. The carmine cone method, while not so accurate as the drop counting, has the advantage of being simpler and is better suited for experiments extending over several days. It has been fully described by Galtsoff (1928). Only a brief resume follows. The oyster is prepared by carefully forcing the valves apart and inserting a short piece of glass rod to hold them open. A 4-inch piece of gum rubber tubing of % inch inside diameter is pushed a short distance into the outlet of the gill chamber. Cotton is then used to close the opening between the valves so that no water can escape from the exhalent chamber except through the tube. The oyster is placed in an enameled rectangular tray approximately 14 inches by 20 inches by 2% inches, having a capacity of 7 1. A constant level arrange- ment maintains the level at 5 1 (5.25 quarts). An inverted T-tube is supported so that one of the arms may be connected to the prepared oyster and the other with a 20 cm piece of glass tubing of 6 mm bore, marked off with a 10-cm interval. The shank of the inverted T is connected to a glass funnel by a short length of rubber tubing. The funnel contains a suspension of carmine in sea water. In operation, laboratory supply sea water flows through the tray at a rate of about 200 cc per minute. This flow is stopped at the beginning of an experiment. Circulation and aeration are accomplished by a stream of air bubbling through an inclined piece of glass tubing about 18 inches in length and l/2 inch in inside diameter. After a preliminary period of half an hour to allow the oyster to adjust itself to the new conditions, it is considered ready for the experiment. A small amount of freshly made carmine suspension is admitted from the funnel and forms at the axis of the tube a cone which is carried out by the current from the oyster. By means of a stop watch, a measurement is made of the time, in seconds and tenths, required for the tip of the cone to traverse the graduated 10-cm. distance. As the mean velocity of the whole cross-sectioned area of the tube is one-half of the velocity at the axis (Galtsoff, 1928), the rate of discharge of water can be easily computed. No such computation was regarded necessary for the purpose of the present experi- ments, the results of which are presented as velocities of current in millimeters per second. Each time 10 or more readings were made and the average taken as representative of the velocity of current of water at the axis of the tube. Observations are made 172 BULLETIN OF THE BUREAU OF FISHERIES at 30- to 60-minute intervals, and are continued for from 2 to 4 hours, the average rate of flow for this period being considered as the normal rate of pumping. The flow of sea water is then cut off, if it has not already been stopped, and a predeter- mined amount of sea water is siphoned off and immediately replaced by the same amount of oil extract, prepared in the manner described below. The solution is allowed to remain in the tray from 1 to 24 hours, during which time measurements of the rate of flow are made, first at 15-minute intervals, later at 30- and 60-minute intervals. Measurements on controls run for the same period of time. The total elapsed time for these experiments, from preparation of the oyster to removal from the tray, averages 26 hours. Treatment with soluble fraction is followed by fresh sea water, measurements being made hourly to indicate the rate and extent of recovery. The plugs, tubing, and glass rod are then removed; and the oyster, suitably marked for future identi- fication, is returned to the large laboratory aquarium. cell connected to the chamber; E, overflow; P, drop counter; S, key switch connecting the signal magnet (Hi); Si, key switch connecting the circuit in the drop counter and 45 volt “ B ” battery (B2); K, kymograph; T, electriaclock; B, storage battery; Bj, 2-volt dry cells for operation of the signal magnet (Mi); and M, signal magnet connected with a drop counter P. DROP-COUNTING APPARATUS The apparatus is illustrated in figure 5. There are three main parts: 1. Supply ( w ), experimental chamber (Ch) and recording apparatus ( k ). A three-way stopcock (A) supplies either laboratory sea water or a test solution. A constant level device maintains equal pressure for both liquids, delivering 200 cc per minute to the experimental chamber. Temperature fluctuations are kept within 1° C. of the laboratory sea water. 2. Experimental chamber (Ch) consists of a celluloid box of 1,500 cc capacity with a small cell ( D ) of the same material attached to its wall and connected through the side by a short piece of glass tubing. An overflow tube (E) in the center of the small cell drains through the bottom. The height of this tube must be so adjusted that when the two chambers are connected the water in both of them remains in EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 173 equilibrium. A constant level is maintained in the large chamber ( Ch ) by means of an overflow which is so adjusted that the water content of the chamber after the oyster is placed in it will be approximately 1 1. Before beginning an experiment the stopcock ( A ) supplying sea water is turned on and the levels are carefully checked. After the equilibrium has been established a single drop added to the small chamber immediately overflows through the tube ( E ). The oyster, prepared in the same manner as for the cone method, is then placed in the large chamber; and the rubber tubing introduced in the exhalent chamber of its gills is connected to the small horizontal glass tube. Water pumped by the gills will immediately overflow through tube ( E ). An electric stirrer and thermometer are kept in the large chamber. 3. Recording apparatus. Two platinum contact wires sealed in the glass tubing (P) are placed under the overflow tube ( E ) so that each drop makes a contact which activates a small signal magnet registering on the smoked revolving drum of a spring kymograph ( K ). Current for operating the magnet is supplied by a 45-volt radio B battery. An electric clock (T) operates another signal magnet which records on the drum time intervals of 1 second. Into this circuit is connected a key switch (5) and a 2-volt dry cell to mark the time of changing from laboratory supply to test solution and vice versa. Another switch ((S'!) disconnects the circuit in the drop counter when the kymograph is not in operation. The kymograph was set to make 1 revolution in 5 minutes. Since the pumping activity of the oyster transfers liquid (sea water or test solution as the case may be) from the large chamber to the smaller, it is necessary to have a continuous flow into the large chamber to maintain the level, otherwise additional work will have to be done by the gill cilia in raising water from the large chamber into the smaller. The change from sea water to test solution therefore is made instantaneously by a twist of the three-way stopcock (A). As soon as the shift is made the oyster is subjected to a gradually increasing proportion of test solution. This solution was allowed to flow into the experimental chamber for 5, 10, or 15 minutes as noted in the tables under the heading “Duration of test solution.” At the rate of 200 cc per minute, from 12 to 15 minutes were required to change the liquid completely in the experimental chamber, so that the oyster was rarely if ever subjected to the full concentration of test solution shown in the tables under the heading “Percent soluble fraction or bleed water.” The percentage given in this column represents the concentration of the test solution, not the percentage to which the oyster was subjected. In experiment 55, table 17, for example, it should be understood that the figures do not show the effect of 10-percent bleed water remain- ing on the oyster for 10 minutes. It does show the effect of gradually replacing sea water with a 10-percent brine solution, the maximum concentration reached being unknown, but somewhat less than 10 percent. The same remarks apply to drop- counting experiments in which water soluble fraction of crude oil was used instead of bleed water. In the experiments using the cone method the specific gravity of the test solution was brought, as nearly as possible, within ±0.0002 of the laboratory supply sea water at the time of the experiment. This was not practicable when using the drop-counting apparatus, but in all cases the difference in specific gravity between the laboratory sea water and the test solution was kept as low as possible. Specific gravity determina- tions were made by Ivnudsen hydrometers certified by the National Bureau of Stand- ards. Hydrogen-ion concentration was checked with a Hellige-Klett color disk. 143870—35 5 174 BULLETIN OF THE BUREAU OF FISHERIES No change in pH of sea water was found after stirring with oil, the average both before and after stirring being 7.6. As in the case of crude oil, bleed water from the Barre and Pelto wells of the Texas Co. was furnished through the courtesy of the Louisiana Department of Conservation. When received, the bleed water had a disagreeable oily odor, was slightly brown in color, had a pH of 7. 1-7.6, and a specific gravity of 1.1064 (17.5°C). A small amount of brown flocculent precipitate present in the bottom of the bottle was left undisturbed when taking a sample of brine for the experiments. The brine was mixed with distilled water, laboratory sea water, or both just before the beginning of an experiment. The percentage of dilutants was varied in order to keep the specific gravity of the test solution as near that of the laboratory sea water as possible and avoid confusing the results through change in density. PREPARATION OF WATER SOLUBLE FRACTION OF OIL The soluble fraction solution was made by placing 6 liters of crude oil and 3 liters of laboratory supply sea water in a glass jar and stirring violently for 30 minutes. The mixture was allowed to stand overnight so that the two fluids would separate as completely as possible. However, only 2,500 to 2,900 cc of soluble fraction solution were recovered from the 3 liters of sea water added. This preparation was regarded as a stock solution from which all other dilutions were made. In the following dis- cussion and in the tables the undiluted solution is called 100 percent soluble fraction. The authors were not in a position to make a chemical analysis of the extracts they used, but due care was exercised in using exactly the same method of preparation. There is a possibility that because of the variation in temperature and salinity of water, the actual amount of substances extracted from the crude oil varied in dif- ferent samples. RESULTS OBTAINED WITH THE CONE METHOD Experiments using the cone method were begun in the early part of June 1934, and completed in the middle of September of that year. The percentages of soluble fraction used were 1, 5, 10, 20, 40, 50, 80, and 100. Since the Pelto wells were no longer producing when this work began, Barre crude oil was used. During the course of investigations, 62 experiments were performed with the cone method. The re- sults of these experiments are summarized and presented in table 11, which contains the essential information regarding the conditions under which the experiments were carried out. The hydrogen-ion concentration of the oil extract was slightly higher than that of the natural sea water. At Beaufort the pH of the water was 7.8. After dilution with distilled water to adjust the salinity, the pH of the sample was as low as 7.6. In Beaufort experiments the observed differences between the sea water and soluble fraction did not exceed 0.3 pH. In a test made at Woods Hole the pH value of the oil extract was 7.4 as compared with the pH 8.3 of the natural sea water. This extract was obtained by stirring oil and water for 24 hours and permitting them to separate in 3 days. Sea water, the pH of which was reduced by the addition of 0.1 N HC1 to 7.4, decreased the ciliary motion by 13.7 percent. The oil extract of the same pH completely inhibited the ciliary mechanism. EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 175 Table 11. — -The effect of water-soluble fraction of Lake Barre oil on the rate of pumping of water by the gills of the oyster [Cone method, Beaufort, 1934] Experiment no. Date Average ve- locity, current Effect of treat- Dura- tion of Specific gravity, 17.5° C. Increase in spe- cific Number of times Tempera- ture (° C.) Before Dur- ing ment (percent) test in hours Labora- tory sea water Soluble fraction gravity due to washing oil washed Begin- ning End Q2 CONTROL June 5 27 38 140.0 59 1.0211 25. 0 25. 4 98 .. June 13 29 36 124. 1 24 1. 0243 25. 4 23. 7 115 June 27 44 43 97.7 25 1. 0257 27.9 28.8 121 July 9 July 25 July 31 Aug. 9 Aug. 29 34 30 88. 2 43 1. 0272 27.6 25.8 139 51 54 105.8 25 1. 0272 27. 2 26. 0 151 44 62 140.9 24 1. 0270 27. 2 26. 4 161 77 75 97. 4 24 1. 0274 27. 4 25. 6 170 44 45 102.3 48 1.0270 25.7 23.9 43.7 47.8 112.0 135 1 PERCENT SOLUBLE FRACTION July 19 July 27 Aug. 2 47 58 123.4 22 1. 0279 1. 0275 0. 0002 2 27.7 28.9 146 30 42 116.6 24 1. 0274 1.0281 .0009 5 26.9 27.3 154 54 46 85.2 24 1. 0273 1.0274 .0000 21 27.6 27.0 158 Aug. 7 Aug. 13 54 48 88.8 24 1.0273 1. 0275 .0003 17 26. 1 26.7 164 41 28 68.3 24 1. 0277 1.0277 . 0004 25 27.7 28.0 168 Aug. 15 Aug. 18 58 57 98.3 18 1.0280 1.0281 .0002 19 28.4 28. 1 174 83 81 97.6 18 1.0275 1.0277 .0004 20 28.7 27.7 53.3 51.4 90.9 106. 5 PERCENT SOLUBLE FRACTION June 21 67 71 105. 9 24 1. 0259 1. 0259 .0002 4 27.4 27.5 107 ...do 41 33 80.5 24 1. 0259 1.0259 .0002 4 27.2 27.6 138 July 23 July 26 July 31 Aug. 8 Sept. 5 46 49 106.5 24 1. 0273 1. 0275 .0003 1 26.8 27. 4 144 36 41 113.9 18 1.0272 1. 0279 .0003 4 27.2 28.0 150 46 34 73.9 18 1.0270 1.0274 . 0006 16 26.8 26. 1 157 61 49 80.3 18 1.0273 1. 0275 .0003 17 26.8 26.5 182 _ 38 45 118.4 18 1. 0231 1. 0232 .0003 29 25.5 25.3 47.8 46.0 97.0 104. 10 PERCENT SOLUBLE FRACTION June 20 59 35 59.3 18 1.0258 1.0259 .0001 27.5 27.2 105 ..do 28 16 57. 1 18 1.0258 1.0259 .0001 26.9 26.8 112.... June 26 45 44 97.7 14 1.0259 1.0259 27.9 27.0 113 ...do 62 42 67. 7 12 1. 0259 1. 0259 27.9 27.6 116 June 27 47 44 93.6 18 1. 0257 1. 0259 2 28.0 28.3 117 June 28 40 31 77. 5 24 1.0249 1. 0250 28.2 28.5 170 Aug. 15 Aug. 30 59 50 84.7 24 1.0280 1.0281 .0004 19 27.9 27.4 176 39 28 71.8 24 1. 0269 1. 0267 .0007 28 23.0 22.0 47.4 36.2 76.2 99.. 20 PERCENT SOLUBLE FRACTION June 14 47 39 83.0 20 1.0239 1.0251 .0004 1 25. 1 26.4 122 July 9 ...do 64 42 65.6 18 1. 0274 1.0270 .0004 15 27.6 26.7 123 52 10 19.2 18 1.0274 1.0270 .0004 15 27.7 27. 1 137. July 23 Aug. 13 Aug. 17 Aug. 29 43 13 30.2 24 1. 0273 1. 0275 .0003 1 27.5 25.8 163 36 24 66.6 24 1.0277 1.0277 .0002 25 28.2 27.5 175 40 13 22.5 18 1. 0275 1.0277 .0004 20 28.6 27. 4 178 41 9 21.9 18 1.0270 1. 0267 .0002 22 25.3 20.4 46. 1 21.4 45.6 125. 40 PERCENT SOLUBLE FRACTION July 12 ...do 74 22 29.7 24 1. 0276 1.0274 16 26.4 27.9 126. 46 21 45.7 18 1. 0276 1. 0274 16 26.5 27.7 129 July 16 July 25 Aug. 2 Aug. 13 37 13 35. 1 23 1. 0275 1. 0276 .0003 17 28.4 29.6 142 51 26 51.0 12 1.0272 1. 0281 .0005 4 26.7 26.3 153 74 41 55.4 24 1.0273 1.0274 .0006 16 27.7 27.6 166 52 11 21. 1 18 1.0277 1. 0276 .0003 10 27.7 27.9 55.7 22.3 39.7 m 50 PERCENT SOLUBLE FRACTION July 16 Aug. 1 Aug. 7 Aug. 17 Sept. 4 58 33 56.8 24 1. 0275 1. 0276 .0003 17 28.5 29. 1 152 58 22 37.9 24 1.0277 1. 0275 .0006 21 27.8 26.3 160 112 19 16.9 24 1. 0273 1. 0275 .0002 17 27.3 26.9 173 74 15 20.2 18 1.0275 1. 0277 .0004 20 28. 1 26.9 183 34 12 35.3 24 1. 0230 1. 0232 .0003 29 24.9 26. 1 67.2 20.2 33.4 176 BULLETIN OF THE BUREAU OF FISHERIES Table 11. — The effect of water-soluble fraction of Lake Barre oil on the rate of pumping of water by the gills of the oyster — Continued [Cone method, Beaufort, 1934] Average ve- locity, current Effect of treat- ment (percent) Dura- tion of test in hours Specific gravity, 17.5° C. Increase in spe- cific gravity due to washing Number of times oil washed Tempera- ture (° C.) Experiment no. Date Before Dur- ing Labora- tory sea water Soluble fraction Begin- ning End 143 80 PERCENT SOLUBLE FRACTION July 26 Aug. 13 Aug. 15 63 0 0 18 1. 0272 1.0271 0. 0004 13 27.4 28.2 165 66 27 40.9 24 1. 0277 1.0276 .0003 10 27.4 27.0 171 85 1 1. 2 24 1.0280 1.0281 .0002 19 28.2 28.2 172 Aug. 16 Aug. 29 Sept. 6 53 2 3.8 18 1. 0277 1.0279 .0001 26 28.7 28.2 177 62 0 0 18 1. 0270 1. 0267 .0002 22 25.3 21.6 185 115 n 9.5 24 1. 0238 1. 0241 .0003 24 25.1 22.8 74.0 6.8 9.2 131 100 PERCENT SOLUBLE FRACTION July 18 July 23 July 25 July 30 Aug. 3 Aug. 8 Aug. 14 Sept. 4 76 0 0 1 1. 0277 1.0274 . 0001 18 27.9 27.8 136 61 0 0 1 1. 0273 1. 0275 .0003 12 27.8 28.2 140 59 0 0 2 1. 0272 1. 0272 .0000 20 27. 1 27.2 147 64 0 0 4 1. 0275 1. 0274 .0006 6 27.4 27. 1 156 75 0 0 4 1. 0275 1. 0275 .0000 7 27.7 27.0 162 51 28 54.9 2 1. 0274 1.0276 .0002 23 27. 1 25.9 167 54 11 20.4 6 1. 0276 1.0278 18 28.6 28.0 181 95 23 24.2 24 1.0230 1.0233 .0002 23 24.5 25.2 66.9 7.8 12.4 One of the difficulties encountered in these experiments lies in the seasonal fluctuations in the laboratory sea water. In preparing soluble fraction there was a slight increase in the specific gravity of water owing to washing with oil. How- ever, the solution used in various experiments was not always heavier than the laboratory sea water used in the same experiment. This was caused by daily fluctu- ations in the salinity and the necessity of preparing soluble fraction solution in ad- vance of the experiment. The specific gravity of the laboratory sea water used during each experiment and the specific gravity of soluble fraction solution are given in separate columns. They correspond to the measurements of velocity of current before and after treatment. In making up various concentrations of soluble fraction, efforts were made in each experiment to adjust the specific gravity of the solution, by diluting it with sea water of high or low salinity, as close as possible to that of the laboratory sea water. However, this was not alwaj^s feasible. A complete record of several experiments with 20-, 50-, and 100-percent soluble fraction are given in figures 6 and 7. The vertical lines in these figures indicate time when the laboratory sea water was replaced with soluble fraction of oil. The figures of the velocity of the current produced by the gills are the averages of 10 or more readings in millimeters per second. They all represent the velocity at the axis of the tube. In the case of the controls, there is no acutal distinction between the periods marked “Before treatment” and “During treatment.” Separa- tion has been made aribitrarily on a time basis. To provide a basis for comparison, at least 3 readings were made during a period of 2 or more hours before treatment. The change in efficiency of the ciliary motion, designated in table 11 as “Effect of treatment”, is expressed by the ratio of the average velocity of the current during treatment divided by the average velocity before treatment and multiplied by 100. The effect of changes in specific gravity on the activity of gill cilia has not been studied. However, there was relatively little change in the salinity of the laboratory EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 177 sea water from day to day, and this factor is not believed to have affected the results. In most experiments the specific gravity of the test solution was brought within ±0.0002 of the specific gravity of the laboratory sea water so as to avoid stimula- tion by salinity changes during the course of an experiment. Complete data of the specific gravity of the laboratory water and of the solution of soluble fraction used in the experiments are given in the table. Fluctuations in temperature may be disregarded insofar as they might influence the results obtained with different concentrations of soluble fraction. The maxi- mum difference in average temperature between the several groups of experiments is 0.9° C. This comparatively slight variation was made possible by using 4 sets of apparatus at a time so arranging to have a differ- ent concentration of soluble fraction in each tray, the experiments at various con- centration levels were car- ried along together so that seasonal changes in tem- perature did not affect one group of experiments more than another. Tempera- ture fluctuations during the course of individual experi- ments were usually less than 2° C. In 5 of the controls (table 11) the rate of flow increased slowly during the course of the experiments. Two others showed a very slight decrease in rate, and in 1 experiment (no. 121) in which there was no aera- tion or circulation of water the decrease amounted to 11.8 percent. The charac1 teristic gradual increase in rate of pumping is attri- buted to recovery from the mechanical stimulation of the plugging operation. The controls cover the period from June 5 to August 29, and except for the first experiment, no. 92, all were sub- jected to the same range of salinities and temperatures as the experimental oysters. On the average for all the 8 controls there was an increase in the velocity of cur- rent which amounted to 12.0 percent. The average duration of test for controls was 34 hours; for experimental oysters, 20.5 hours. The total elapsed time during which each oyster was plugged averaged approximately 6 hours more than the duration test. that 4 experiments could be run simultaneously. By Figure 6.— Effect of water soluble fraction on rate of flow of water. Exp. 178. 178 BULLETIN OF THE BUREAU OF FISHERIES In experiments extended from July 19 to August 18, 7 oysters were subjected to a 1 -percent soluble fraction solution for from 18 to 24 hours. During this time the specific gravity of the laboratory sea water was very high, ranging from 1.0273 to 1.0280 (17.5°). Water temperature averaged 27.6° C. The average rate of pumping during treatment, as compared with the rate before treatment, showed a reduction of 3.1 percent (table 11). This is insignificant by itself insofar as indicating any effect of 1-percent soluble fraction solution is concerned. However, this reduction in rate of flow is 15.6 percent lower than the control average during the treatment period. The greatest decrease in rate of flow resulting from this treatment was found in experiment 164, in which the reduction in rate amounted to 31.7 percent. This maximum effect was obtained with soluble fraction from a sample of oil washed 25 times, the increase in specific gravity caused by washing being 0.0004. If the No urs after start Figure 7.— Effect of water soluble fraction on rate of flow of water. effect of the soluble fraction was owing to mineral salts* taken up from the oil by the sea water, it would be expected that the inhibiting power of the soluble fraction on rate of flow would gradually diminish as the salts were exhausted, and that the increase in specific gravity of the sea water after washing with oil would likewise decline. However, this was not the case and the inhibiting effect of the oil extract did not diminish with subsequent washing of the sample. Except for experiment 164, the effect of treatment with 1 percent soluble fraction gives percentages closely comparable with similar figures for the control experiments. However, considering the orderly decrease in rate of flow caused by treatment with increasing percentages of soluble fraction, it is believed that a 1 percent soluble fraction exerts a definite though slight inhibiting effect on the ciliated epithelium of the gills. Seven experiments were made with 5 percent soluble fraction solution between June 21 and September 5. The average temperature was 26.8° C., and the specific EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 179 gravity ranged from 1.0231 to 1.0273 (17.5° C.). The average rate of flow during treatment is 97.0 percent of the rate before treatment. It woidd appear that 5 percent soluble fraction is no more toxic to oysters than 1 percent, over a period of 24 hours. However, on closer examination we find an interesting situation in this group. Of the seven experiments, four show an average increase of 11.2 percent during treatment, comparable to 12.0 percent shown by controls. The remaining three experiments have an average rate of 78.2 percent during treatment, comparable to the rate obtained with 10 percent solutions. In considering factors which might be responsible for the effect, or lack of it, a comparison was made between the conditions of the two groups of experiments. The average temperature of the four experiments in which an increased rate of flow resulted from treatment is 26.7°. The average for the three experiments whose flow was reduced during treatment is 26.9°. This difference of 0.2° between the. averages for the two groups is less than the fluctuations generally found during the course of an experiment. It is obvious that temperature differences could have no connection with the disparity in effect of the soluble fraction solution. Also, the effect of the soluble fraction has no apparent relationship with the specific gravity of the laboratory sea water. In experiment 182, the low specific gravity of 1.0231 was accompanied by an increase of 18 percent in rate of flow during treatment. In experiment 144, the high specific gravity of 1.0272 was associated with an increase in rate of flow of 13.9 percent during treatment. Consequently, there is no evidence that an increase or a decrease in rate of flow during treatment has any connection with a high or low specific gravity of the sea water, though this statement is intended to apply only to the limits of salinity occurring in these experiments. The difference in specific gravity between the laboratory sea water and the soluble fraction was, in most experiments, not over 0.0002. Where the difference exceeded this figure, notably in experiment 144, in which the difference was 0.0007, the effect is negligible compared with fluctuations in experiments where differences in specific gravity were not present. For example, experiments 106 and 107 were carried on simultaneously. They received soluble fraction prepared from the same sample of oil and sea water, were treated for the same length of time and the soluble fraction was adjusted to the same specific gravity as the laboratory sea water. Yet oyster 106 increased the rate of flow during treatment by 5.9 percent, while oyster 107 reduced its flow 19.5 percent. Since there is no intergrading effect between the two groups of oysters used in these experiments, the inference can be drawn that 5 percent solution exerted a definite effect on one group of them which comprised specimens more sensitive than the others. In other words, this concentration may be regarded as a threshold of inhibitory action. These results are based on treatments extending only for 18 or 24 hours. It is possible that the same depression of the efficiency of ciliary motion may be reached with smaller concentrations acting over a longer period of time. Ten percent soluble fraction solution very definitely inhibited ciliary activity and resulted in an average decrease in rate of pumping during treatment of 23.8 percent. (Table 11.) There is no relationship between the increase in the specific gravity of the undiluted oil extract and the effect on rate of flow during treatment. In experiment 180 BULLETIN OF THE BUREAU OF FISHERIES 104, the increase in specific gravity caused by washing was 0.0001, and the rate of flow during treatment was only 59.3 percent of the normal rate before treatment, while in experiment 176, the increase in specific gravity was 0.0007, and the rate of flow during treatment was 71.8 percent of the normal. The difference in specific gravity between the laboratory sea water and the test solution was held within close limits, not exceeding 0.0002 in any experiment in this group. The specific gravity of the sea water was high in all experiments, ranging from 1.0249 to 1.0280. There is no apparent relationship between a high specific gravity and effect of the soluble fraction, for the greatest reduction in rate of flow, 42.9 percent, occurred in experiment 105, in which the specific gravity was considerably lower than in experiment 170, specific gravity 1.0280, the reduction in flow during treatment for the latter experiment amounting to only 15.8 percent. Fluctuations in temperature may be disregarded as factors influencing the rate of flow during the experiments, the maximum difference in temperature at the beginning and end of an experiment in the 10 percent group not exceeding 1.0° C. The time of treatment in this group varied from 12 to 24 hours. Within these limits the duration of treatment does not appear to be a factor in determining the effect. Three experiments, 104, 105, and 116, were carried on for 18 hours. The first two showed the greatest, the last one next to the least effect for the group. Seven experiments were made with 20 percent soluble fraction solution. These were well distributed over the period from June 14 to August 29. The average temperature of the water in these experiments was 27.1° C. Specific gravities ranged from 1.0239 to 1.0277. The average reduction in rate of flow resulting from this treatment amounted to 54.4 percent. In the experiments with 20 percent solution (figs. 6 and 7) the difference in specific gravity between the laboratory sea water and the test solution was not held to as low limits in all cases as was done in the experiments with 10 percent solution. The maximum difference in specific gravity occurred in experiment 99, where the test solution was higher by 0.0012. This increase in specific gravity of the soluble fraction solution apparently was not of serious proportions in this case, for the reduction in rate of flow during treatment was only 17 percent, the lowest for the group. Con- sideration was given to the possibility that the comparatively slight effect of soluble fraction in experiment 99 might be due, in part, to the relatively low salinity of the laboratory sea water, 1.0239, for the other experiments in this group were made when the specific gravity was at or above 1.0270. However, an examination of the control and other experiments does not support this contention. While most of our experi- ments were made with sea water of higher salinity than is found over many oyster beds, nevertheless it may be stated that within the range of specific gravities used the effect of the soluble fraction is independent of the salt content of the sea water. Increase in specific gravity of the sea water after washing with oil has no bearing on the effect of the 20 percent soluble fraction. In experiments 122 and 123 (table 11), two oysters were treated with portions of the same soluble fraction for 18 hours. The increase in specific gravity following washing with oil was 0.0004. The rate of flow of oyster 122 was reduced 34.4 percent during treatment. The reduction in rate for oyster 123 was 80.8 percent. Oyster 99 was treated for 20 hours with a different sample of soluble fraction having the same increase in specific gravity after washing and the reduction in rate of flow was 17 percent, comparable to the effect in experiment 122. Oyster 163, treated for 24 hours with soluble fraction having an increase in EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 181 specific gravity of 0.0002, half that of experiment 122, showed a decrease in rate of flow of 33.4 percent, almost exactly the same as no. 122. Finally, oyster no. 178 (fig. 6), treated for 18 hours with soluble fraction having an increase in specific gravity of less than 0.0002, had the rate of flow reduced 78.1 percent, a figure of the same order of magnitude obtained in experiment 123. Fluctuations in temperature are not significant insofar as their effect on the rate of flow is concerned. Experiments 122 and 123, as mentioned above, were carried on simultaneously. The temperatures differ little in the two experiments, but the effect of the soluble fraction is not at all comparable in amount. It might appear that the temperature drop of nearly 5° C. in experiment 178 is partially responsible for the relatively large decrease in rate of flow. Actually, this is not the case as can be seen by examining figure 6 which presents a complete record of the experiment. Six experiments were made with 40 percent soluble fraction, from July 12 to August 13. The average temperature of the water in these experiments was 27.2° C. Specific gravities of the laboratory sea water were very high, with little fluctuation, the range being from 1.0272 to 1.0277 (17.5° C.). Except in experiment 142 (table 11), in which the decrease in rate of flow during treatment was almost the smallest in the group, the difference in specific gravity between the sea water and the soluble fraction was kept within 0.0002. The greatest increase in specific gravity of the sea water due to washing (0.0006) occurred in experi- ment 153, which was least affected by treatment. The maximum reduction in rate of flow following treatment was found in experiment 166, and the soluble fraction used here was a mixture from two oils which had been washed 10 and 24 times respectively. Five experiments were made with 50 percent soluble fraction solution from July 16 to September 4. The average temperature of the laboratory sea water during this period was 27.3° C. The specific gravity of the sea water varied but little in July and August, 1.0273 to 1.0277; but in the last experiment on September 4 it had fallen to 1.0230. The average rate of flow during treatment dropped 66.6 percent. There is relatively little difference in the effect of 40 and 50 percent soluble fraction. The experiments in the two groups have been separated chiefly to emphasize the close agreement between them. The main interest of the experiments with 50 percent soluble fraction is the lack of any definite relationship between the effect of the soluble fraction and specific gravity (table 11, fig. 7). The reduction in rate of flow was practically the same in experiment 152, with the high specific gravity of 1.0277, and in experiment 183, with the low specific gravity of 1.0230. Six experiments were made with 80 percent soluble fraction solution from July 26 to September 6. The average temperature of the laboratory sea water during this time was 27.0° C. Five of the experiments were carried on during July and August, when the specific gravity of the laboratory sea water ranged from 1.0270 to 1.0280. At the time of the last experiment on September 6, the specific gravity had fallen to 1.0238. The maximum difference between the specific gravity of the laboratory sea water and the soluble fraction was 0.0003. The effect of 80 percent soluble fraction was to reduce the activity of the ciliated epithelium 90.8 percent. The drop in pumping activity is 24.2 percent greater than was found with 50 percent soluble fraction. 182 BULLETIN OF THE BUREAU OF FISHERIES In experiments 143 and 177 of this group we find, for the first time, the pumping immediately stopped on addition of the soluble fraction, and not beginning again until the test solution was replaced with fresh sea water. As in previous experi- ments, an extreme effect is probably attributable to poor condition of the oysters, for the difference between the specific gravities of the laboratory sea water and the soluble fraction was very small, 0.0001 in experiment 143 and 0.0003 in experiment 177. Nor can the stoppage of flow be attributed either to an increase in specific gravity of the sea water after washing with oil, or to the number of times the oil was washed. The least effect of 80-percent solution occurred in experiment 165, in which the reduction in flow amounted to 59.1 percent. This relatively small effect is doubtless also due to the condition of the oyster, no. 165, being for some reason better able to withstand the test solution than any of the others in the group. Eight experiments were made with 100-percent soluble fraction (table 11, fig. 7). The average tem- perature of the laboratory sea water for all experi- ments was 27.3° C. The specific gravity was high, 1.0272 to 1.0277, except in the last experiment, where it was 1.0230. The average reduction in ciliary activity amounted to 87.6 percent. In five ex- periments there was a com- plete cessation of pumping. The results of all the experiments with the cone method show, without any doubt, that the presence of the water-soluble fraction of crude oil exerts an inhib- iting effect on the efficiency of the ciliary motion of the gill epithelium, the work of which is interfered with in such a way that less water is pumped through the gills, and consequently the rate of feeding of an oyster is decreased. The decrease in efficiency of the epithelium is directly proportional to the increase in concentration of the soluble fraction. This is clearly seen from an examination of figure 8, which shows the percentage of depression caused by various concentrations. The values plotted in the figure are the averages, recalculated from the data given in table 11, based on the efficiency of the ciliary motion in the controls as 100 percent. A relatively high degree of depression caused by the 20-percent solution may be considered as a breaking point in the curve, but so small a number of observations is not sufficient to determine this point with certainty. Regardless of the graphical interpretations that can be made concerning the shape of the depression curve, one Figure 8.— Depression of rate of pumping of gills caused by various concentrations of water soluble fraction of oil. EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 183 fact is clear: That a concentration between 20 and 30 percent soluble fraction will, on the average, reduce the rate of feeding of the oyster to one-half its normal value. RESULTS OBTAINED WITH THE DROP COUNTING METHOD Two series of experiments were made with the water-soluble fraction of crude oil, using the drop counting technique previously described. The first group of 12 experiments (table 12) represents winter conditions, as they cover the period of November and December 1933. Two Louisiana crude oils were used, one from the Lake Barre wells, the other from the Pelto wells. Both fields are located in Terrebonne Parish and are operated by the Texas Co. In the second series (table 13) are 19 experiments completed during May 1934, using only Lake Barre oil. The results of all experiments are summarized in these tables, which give the rate of pumping, in drops per minute, before, during, and after treatment. A column marked “Effect of treatment” gives the percent of the normal rate of flow obtained by dividing the average rate of flow during treatment by the average rate before treatment, making it possible to evaluate the inhibiting effect of oil extract on various oysters. In table 14 the results of the experiments are presented in a more detailed manner, showing for every 5-minute interval, the average number of drops of water per minute passed by the gills. Heavy type indicates observations made when the soluble-fraction solution was running through the experimental chamber. In both series of experiments a full-strength solution was used. There was a slight increase in the specific gravity of the sea water after it was stirred with oil, but there was no correlation between the specific gravity of the extract and its toxicity. Table 12. — The effect of soluble fraction of crude oil of Lake Pelto and Lake Barre wells on the rate of pumping of water by the gills of the oyster [Drop counting method, winter experiments, Beaufort, 1933] Experiment no. Date Drops per minute Effect of treat- ment Dura- tion test in min- utes Specific gravity, 17.5“ C. Tempera- ture, “ C. pH sol- uble frac- tion Increase in spe- cific gravity of sol- uble fraction Recovery time Source of oil Be- fore treat- ment Dur- ing treat- ment After treat- ment Labora- tory sea water Soluble frac- tion Be- gin- ning End Per- cent Min- utes Percent fit) Nov. 28 92.0 109. G 95.2 119. 1 5 1. 0261 1. 0277 13.6 128 5 60 A 65.8 41. 3 69.7 62. 8 5 1. 0261 1. 0262 15. 0 15. 1 7 3 0 0001 1 16 30 Dn 62 Dec. 1 31.2 19.0 27.0 61. 1 5 1. 0261 1. 02638 15.4 15.2 7.3 .00028 93 15 Do. 63 A Dec. 4 112.2 92.0 108.2 82.0 5 1.0262 1.0262 17. 4 17. 3 7. 3 107 15 Dn 64 A Dec. 5 77.2 30.0 71.5 38. 9 1. 0260 1.0260 16.5 16. 3 97 15 Do 65A Dec. 7 41.4 46. 6 39.0 112.7 5 1. 0265 14.8 14.8 123 5 66 Dec. 11 36.0 35.0 39. 5 97. 2 10 1. 0267 13. 2 13. 5 145 5 Dn 67 ...do. 60.5 54.7 58.2 90.4 10 1. 0267 13.8 14. 1 103 20 Do. 68 Dec. 12 67.2 52.0 57.2 77. 4 10 1. 0267 11.8 11. 6 77 10 Dn 69 Dec. 15 77.0 60.4 79.6 78.4 in 1. 0267 1. 0279 12.3 12.5 7.5 .0012 109 15 Do. 70 Dec. 21 1 120. 7 0 36.5 i 30. 2 15 1.0256 1. 0260 15.6 16. 5 7.6 Do. 70 A Dec. 22 i 83.5 0 64.9 > 77.7 15 1. 0256 15.0 15.4 Sea water. Average.. 66.0 54.06 62.2 82.0 14. 48 109.8 13.5 1 Numbers omitted in making average. 184 BULLETIN OP THE BUREAU OF FISHERIES Table 13. — The effect of soluble fraction of Lake Barre oil on the rate of pumping of water by the gills of the oyster [Beaufort, 1934, spring experiments] Experiment no. Date Average number drops per minute Effect of treat- ment Dura- tion of test in min- utes Specific gravity, 17.5/17.5° C Increase in specific gravity of soluble fraction Tempera- ture, ° O. Number of wash- ings Recovery time Before treat- ment Dur- ing treat- ment After treat- ment Labora- tory sea water Soluble frac- tion Be- gin- ning End Per- cent Min- utes CONTROLS Percent 73 May 9 44 40 38 91 78 1.0234 21. 7 22.0 74 121 128 134 106 71 1. 0237 21. 6 21.5 84 May 16 139 141 137 101 92 1. 0242 20.9 21.15 86 115 114 115 99 55 1.0242 21. 3 21. 45 SOLUBLE FRACTION 71 May 7 109 49 73 45.0 5 1. 0223 1.0177 0. 0007 21. 5 21. 5 1 72 35 72. May 9 128 71 119 55.0 4 1. 0234 1. 0177 .0007 20.9 20.9 1 87 30 75 May 10 132 69 64 52.0 9 1. 0237 1. 0177 .0007 21. 6 21.7 1 42 10 76 __ do 124 98 122 79.0 10 1.0237 1.0122 .0115 21. 7 21.9 92.3 20 77 __ do 129 85 99 66.0 8 1. 0237 1. 0244 .0007 21.9 22. 1 1 62.9 10 78 May 14 141 13 129 9. 2 8 1. 0242 1. 0247 .0008 22. 1 22. 4 1 96 30 79 __ do 136 19 106 14.0 8 1. 0242 1. 0249 .0007 22.4 22.7 2.. 102 30 80 146 27 105 18.5 8 1. 0242 1. 0247 .0005 23.3 23. 3 3 75 30 81 156 94 132 60.3 8 1. 0242 1. 0250 .0008 22.9 22.5 3 87.9 20 82 ___do 133 85 119 64. 0 8 1. 0242 1. 0246 .0004 22.5 22.4 4 91 15 83 126 33 100 26.2 20 1. 0242 1. 0245 .0003 22.3 22.6 93 10 85 May 16 133 31 107 23.3 23 1. 0242 1. 0244 .0002 21. 1 21.3 5 and 6_. 79. 1 20 87 May 17 87 27 55 31.0 24 1. 0241 1. 0243 .0001 19.1 19. 1 6 and 7.. 64. 1 30 88 May 21 154 42 103 27.3 38 1. 0212 1. 0245 .0005 24.2 24.3 8 and 9_. 70 20 89 — . May 22 126 38 93.5 30.2 12 1. 0230 1. 0237 .0007 24.8 24.8 9 and 10. 68 15 Average 130. 6 52.0 101. 6 40. 0 22. 1 78.7 21.6 Note.— Averages do not include controls or experiment no. 76, winch was made with dilute sea water. The column in table 13, marked “Number of washings” refers to the number of times the same oil has been stirred with sea water. In the winter experiments, the oil was used only once, then discarded. In the spring experiments as many as 10 extractions were obtained from a single sample of oil without exhausting its toxicity. One must bear in mind that because the oysters in these experiments are kept in running water, which is gradually replaced by the solution, the duration of each test lasted only a short period of time, varying from 4 to 38 minutes, depending on the amount of solution available. In figure 9, representing the results of three experi- ments, the time when the treatment began is marked with a vertical line. It has been estimated that the sea water in the experimental chamber (fig. 5, Ch) was com- pletely replaced in 12 minutes, and that the same period of time was required to replace the oil extract with normal sea water, but there was no means of checking the exact concentration of solution in the chamber at the beginning and end of the test. A decrease in the rate of pumping of water, caused by the oil extract, is notice- able almost immediately upon the beginning of treatment, and continues as long as the toxic substances are present in the water. There was a wide difference in the sensitivity of various oysters. While in some of them the pumping was completely inhibited, in others it was only slightly depressed (table 14). EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 185 too - /so - 140 130 120 no 100 90 80 70 CO 50 40 I , 30 ^.20 jk i 1 1 1 1 1 1 1 1 1 r Tlie average rates of flow of water, in drops per minute, were computed from the kymograph data. The time during which the record was taken varied for the period “Before treatment”, from 20 to 45 minutes; “During treatment”, from 15 to 45 minutes; and “After treatment”, from 15 to 90 minutes. It is believed that these periods are of sufficient duration to eliminate the effects of accidental stimula- tion and other possible dis- turbances. Although both the win- ter and spring experiments show a depression in the rate of flow following treatment, the effect is much greater in the spring series. The average depression in rate of flow resulting from treat- ment for the winter series is 18 percent. For the spring series it is 60 percent. This difference in the levels of depression for the two groups is presumably a function of temperature, and is due to the direct relationship exist- ing between temperature and rate of motion of the lateral gill cilia. It will be noted (table 12) that the average rate of flow before treatment in the winter series is 66 drops per minute, while in the spring series the average 0 is 130 drops per minute. The average temperature at the beginning of the experi- ments for the winter series is 14° C., while the spring average is 22° C. Thus an average increase of 8° C. in temperature approxi- mately doubles the rate of pumping while the depressing effect of the soluble fraction solution is trebled. Sol. fraction on Fxjot. 84 Control -• 12 Sol. Fr. o 79 Sol. Fr. o 81 Sol. Fr. a _L _L 5 15 25 85 45 Time m m/nutes 55 65 75 £>5 95 105 Figure 9. — Effect of water soluble fraction of Lake Barre oil on the activity of the ciliated epithelium. Drop counting method. 186 BULLETIN OF THE BUREAU OF FISHERIES Table 14. — The effect of soluble fraction of crude oil from Lake Barre and Lake Pelto wells on the rate of 'pumping of water by the gills of the oyster [Detailed record of the experiments. Figures printed in heavy type represent observations made when oysters are in solution of soluble fraction] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Time Experiment Date Crude oil Of treat- Drops per minute, average by 5-minute intervals ment 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 1933 Min. 60 Nov. 28 Barre 5 94 92 91 87 S3 110 118 101 104 98 91 88 60A __ do._- _ _ do 5 61 64 64 65 75 29 37 58 61 69 73 76 62 Dee. 1 _do __ 5 34 35 31 30 24 13 18 26 29 27 25 63 A Dec. 4 do 5 115 114 113 118 101 85 90 101 120 106 100 107 64A Dee. 5 do 5 80 81 74 75 76 14 14 62 75 75 68 68 65 A Dec. 7 Pelto 5 44 42 40 39 42 42 51 47 44 40 34 38 66 Dec. 11 do 10 35 35 40 34 30 22 27 52 45 38 38 37 67 __ do do 10 63 59 64 56 57 52 57 61 62 54 56 68 Dee. 12 do _ 10 70 72 59 68 81 31 44 52 51 54 56 68 69 Dec. 15 do 13 65 81 80 82 81 73 40 35 73 84 83 72 70 Dec. 21 do 15 119 122 121 121 1 1 1 0 0 29 42 39 52 70A Dec. 22 Sea water . 15 80 83 85 80 2 2 2 25 25 28 75 91 91 73 . 1934 May 9 May 10 May 16 Control 46 49 46 36 45 42 42 43 40 36 35 32 42 38 41 39 74 do __ 127 128 118 126 106 127 126 120 134 117 142 144 116 141 126 144 84 do 141 136 144 140 141 138 136 143 143 141 136 140 144 139 139 131 136 135 86 _ do _ __do 123 122 104 113 117 117 109 112 113 114 116 117 71 May 7 May 9 May 10 do... Barre 5 118 115 in 107 104 103 49 40 58 63 64 67 71 78 70 77 72 do 5 125 116 133 137 74 32 88 91 100 108 111 108 118 107 112 93 75 __ do 10 122 54 45 55 54 45 43 47 52 56 48 44 44 56 65 69 107 76 do 12 124 129 100 71 88 104 114 105 136 114 138 140 77 ___do do 10 123 12.5 72 64 SO 76 83 67 80 94 66 100 107 110 120 110 118 78 May 14 ___do do 10 148 136 77 149 123 53 0 0 0 119 220 131 136 129 114 140 133 137 79 - ___do 10 146 0 0 0 0 42 130 139 133 149 148 80 _ do do 10 153 103 4 0 0 0 6 78 109 105 133 139 147 143 144 149 81 .. May 15 ___do __do . 8 151 161 156 155 101 G7 84 125 130 135 129 125 134 134 133 131 82 do _ 8 133 131 135 132 61 44 117 120 121 117 117 120 83.. --_do do 20 120 127 126 130 80 27 30 21 i 108 117 102 89 100 98 95 97 85 May 16 May 17 May 21 May 22 do 24 137 133 129 134 104 39 31 12 0 0 71 102 105 99 104 112 119 87 do 24 87 87 87 85 64 45 32 21 1 2 37 52 54 55 56 45 60 88 _do 38 151 155 154 156 120 69 54 52 49 34 2 0 0 29 82 107 108 89. __ do 12 118 130 131 128 127 127 121 125 85 15 16 37 84 86 84 90 95 The effect of temperature on the rapidity of recovery is equally striking. The oysters in both winter and spring experiments show, after treatment with soluble fraction solution, a fairly steady rise in the rate of pumping but this initial rise is followed by a slight drop (fig. 9). The significance, if any, of this peak in the recovery curve is not known. However, it is useful in demonstrating the relation of temper- ature to rate and extent of recovery. In table 12, the column headed “Recovery time”, shows the percent of recovery at the peak as compared with the average rate before treatment. The time in minutes extends from the moment of turning off the test solution to the point where the peak occurs. For the winter series, the average recovery at the peak was 109 percent in an average time of 13.5 minutes. The spring series averaged only 78 percent recovery at the peak in an average time of 21.6 minutes (table 13). Results of the experiment carried out at Woods Hole, Mass., in which a solution of water-soluble substances present in Pelto oil was prepared by allowing them to diffuse through a collodion sack suspended in the sea water are similar to those ob- tained at Beaufort with soluble fraction prepared by shaking oil with the sea water. The sack, about 1 inch in diameter and 7 inches high, was made by dipping a glass test tube of corresponding size into a solution of Merck’s collodion (reagent quality) and drying it for 1 minute. The membrane was carefully examined under a micro- scope and tested against leakage. Fifty cubic centimeters of oil were poured into the sack and the latter was kept suspended for 4 days in 4 liters of sea water which was gently agitated by means of an electric stirrer with glass rod. The specific EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 187 gravity of tlie water (at 17.5° C.) at the beginning and at the end of this period was 1.02455 and 1.02461, respectively. This slight increase in the specific gravity may be attributed either to the evaporation of water or to the dialysis of substances present in the oil. An oyster was mounted in the usual manner and placed in a celluloid tank of 1,830-cc capacity used for measuring the rate of flow of water through the gills. Water was supplied at the rate of 100 cc per minute and the entire volume of the tank could be renewed in 20 minutes. Drops of water filtered by the oyster gills were recorded by means of an electric drop counter. The record of this experiment is presented in table 15 which shows the average number of drops per minute deliv- ered by the gills before, during, and after the addition of the water soluble fraction of the oil. During this period the temperature varied between 21.5 and 21.8° C. Table 15. — The effect of water soluble fraction of Pelto oil obtained by dialysis through a collodion membrane on the rate of flow of water through the gills of the oyster Time Drops per minute Remarks Time Drops per minute Remarks 12:05-2:25 •208 Sea water. 4:07-4:08 . 105 Sea water replaced by solution — 2:30 Solution turned on. Continued 2:50 96 Sea water replaced by solution. 4:08-7:10 2:53-2:54 96 Do. 7:10-7:11 53 Solution turned on. 3:20-3:21 90 Do. 7:11-7:12 52 Do. 3:22-3:23 66 Do. 7:12-7:13 60 Do. 3:23-3:24 90 Do. 7:15 3:24-3:25 __ 108 Do. 7:50-7:51 99 Do. 4:00-4:01 10 Do. 4:01-4:02. 14 Do. NEXT DAY 4:03-4:04 72 Do. 4:05-4:06 106 Do. 9:15-9:16.. 66 4:06-4:07 107 Do. 1 Average. By examining this table one notices that the presence of the water soluble fraction of Pelto oil greatly interferes with the normal activity of the gill epithelium. The rate of flow of water becomes less regular and decreases about 50 percent. A 3-hour exposure in this solution causes further inhibition of the ciliary motion of the epithelium, the activity of which was not fully restored even when the oyster was placed back in natural sea water. On account of the limited amount of crude oil available in the laboratory experi- ments, it was necessary in preparing the water soluble fraction to use the same sam- ple of oil several times. It was noticed that subsequent 30-minute stirrings of 2 volumes of crude oil with 1 volume of sea water did not exhaust the potency of the sample in inhibiting the work of the ciliated epithelium. Table 15 shows the per- centage of depression in rate of flow of water caused by 10-, 20-, and 40-percent solutions of the extract. It will be observed that after 28 washings, there was no decrease in the toxicity of the extract. Table 16.- — Showing no decrease in toxicity of a sample of oil due to repeated washings with sea water Number of washings Percent de- pression Number of washings Percent de- pression Number of washings Percent de- pression 10 PERCENT 20 PERCENT 40 PERCENT 2 6.7 1 17.0 4 50.0 19 10.3 1 69.5 9 54.3 28 29. 1 8 34.0 9 70.5 8 80.6 10 79.0 13 .. 68.0 16 45.0 22 77.0 25 35.0 188 BULLETIN OF THE BUREAU OF FISHERIES This observation has an important bearing on the problem of oil pollution of natural waters, indicating that oil floating on water or absorbed in mud will, for a long time, remain a source from which toxic substances diffuse into the water. It seems probable that the toxic effect is not due to the mineral salts which occur in the crude oil and which doubtless will be leached out by subsequent washings, but to the organic compounds which gradually dissolve in the sea water. Table 17. — The effect of oil well bleed water on the rate of 'pumping of water by the gills of the oyster [Drop counting method] Experiment no. Date Drops per minute Effect of treat- ment Bleed water on — Per- cent bleed water No cur- rent for— Specific gravity 17.5/17.5 Temperature, “0. pH bleed water solu- tion Source bleed water Before treat- ment Dur- ing treat- ment After treat- ment Sea water Bleed water Begin- ning End Min. Min. 55 Nov. 24 61.0 58.2 65.7 95 10 10 1.0258 1.0261 15.5 15.8 7.3 Barre. 63 Dec. 4 125.8 113.0 108.7 90 5 10 1. 0262 1. 0267 17.3 17.0 Do. C3B 41.2 30.3 40.0 73 5 10 1. 0260 1. 0265 15.2 15.6 Do. 64 __do 66.0 73.0 80.7 111 5 10 1. 0260 1. 0264 16.2 16.6 Do. 65 C Dec. 8 80.7 95.6 87.3 106 10 10 1. 0268 1. 0263 14.8 15.2 7.5 Pelto. Average. 95 54 Nov. 23 78.7 57.5 69.0 73 10 20 1. 0255 1. 0259 16.4 16.6 7.3 Barre. 64 B Dec. 6 77.5 80.2 82.5 103 5 20 1. 0260 1. 0257 16.2 16.6 Do. 64C . 95.0 74.8 96.3 79 10 20 1. 0260 1. 0258 17. 1 17.3 Do. C5D_ Dec. 8 97.0 89.6 91.6 91 10 20 1. 0268 1. 0268 15.2 15.2 Pelto. Average. 86.5 65 Dec. 7 74.2 70.2 89.0 94 10 33 1. 0268 1. 0302 15.0 15.0 7.3 Barre. 65 B Dec. 8 86.2 35.0 100.5 41 10 40 3 1. 0268 1. 0397 14.1 14.4 7.3 Do. 53 Nov. 21 71.2 55.0 68.0 77 10 33 1. 02535 1. 0379 12.8 13.2 7.3 Do. 53A 97.0 70.7 96.0 81 10 33 1. 02535 1. 0379 15.9 Do. 53B Nov. 22 130.0 56.5 97.4 44 10 40 1. 02559 1. 0436 16.6 7.7 Do. Average. 67. 4 67A Dec. 11 57.0 16.0 36.0 28 14 50 10 1. 0267 1. 0532 14. 1 14.3 Barre. 68A Dec. 12 71.0 10.0 32.0 14 10 60 12 1. 0267 1.0638 11.6 11.7 Do. 68B Dec. 13 91.0 21.5 18.0 24 10 80 29 1. 0267 1. 0851 12.0 12.0 7.5 Do. 69 A Dee. 15 84.0 4.5 24.5 5 10 100 120 1. 0267 1. 1064 14.4 15.4 7.5 Do. EFFECT OF BLEED WATER ON THE RATE OF FEEDING Eighteen experiments were made with “bleed water” or brine from the Barre and Pelto wells during November and December 1933, using the drop counting technique. An adjustment period of from 10 to 30 minutes preceded the beginning of the measurements. A summary of all the experiments is given in table 17. Column 3, “Before treatment,” shows the rate of pumping in drops per minute, averaged by 5-minute intervals, from the time the kymograph was started until the test solution was turned on. The elapsed time for this period is usually 20 minutes. Column 4, “During treatment”, shows the average rate of flow in drops per minute while the test solution was flowing over the oyster, and for 10 minutes after the test solution was turned off, as this time is required to replace it with sea water. The test solution was on for 5 or 10 minutes in most cases (col. 7). Column 5, “After treatment”, shows the average rate of pumping in drops per minute for a period about 20 minutes immediately after treatment. The value is a measure of recovery occurring during this period. There was not always a sufficient time for the rate of flow to return to the level established before treatment, but a EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 189 study of recovery was not considered to be of sufficient importance to the problem to warrant additional time. The effect of treatment obtained by dividing the average rate during treatment (col. 4) by the average rate before treatment (col. 3) multiplied by 100 is a measure of the toxicity of the test solution. However, this value sometimes masks the actual effects of the test solution, and for this reason each experiment is presented com- pletely in table 18, in which the average rate of flow in drops per minute is shown for each 5-minute interval. The intervals in which the test solution was flowing are indicated in bold face type, the figure in the last column giving the actual number of minutes during which the bleed water was running into the experimental chamber. An example of the masking of brine effects as shown in column 6, table 17, is experiment 55. In this table the reduction in rate of flow due to a 10-minute treat- ment with 10 percent bleed water solution is 5 percent. An examination of the actual figures in table 18, shows that during the first 5 minutes of treatment there was an accelerating effect of about 16 percent; during the second 5-minutes of treatment the rate fell to nearly normal ; in the 5-minute period immediately succeeding treatment, the flow decreased 36 percent, and returned to normal in the next 5-minute interval. The effect of bleed water in experiment 55, therefore, consists of an initial acceleration in ciliary activity followed by a considerable inhibition. Of the 5 experiments using approximately 10 percent bleed water solution, 3 (nos. 55, 63, and 63B) show a reduction in rate of flow during treatment, the other 2 having an increased rate of pumping. It was to be expected that similar to the action of various poisons the low concentrations of bleed water may have a stimulating effect at least in some of the individuals. Owing to large individual differences in condition and resistance of oysters, it is impossible to establish an exact concentration level for bleed water solutions, above which the rate of pumping would always be reduced, and below which an increase in rate would occur. However, the rate of pumping in the 10 percent bleed water group is reduced an average of 5 percent, which seems to indicate that this concentration slightly exceeds the limit of tolerance. Four experiments were made using 20 percent bleed water. These show a decidedly greater depressing effect, the average rate of flow during treatment for the group being only 86.5 percent of normal. This increase in depressing effect of the 20 percent bleed water probably is due to a greater proportion of ions affecting the gill cilia, for there was no increase in the specific gravity of the solution as compared with laboratory sea water. On the contrary it will be seen (table 17) that the specific gravity of the test solution in experiments with 20 percent bleed water is closer to that of the laboratory sea water than is the case in the experiments with 10 percent bleed water. The effect of 33 percent bleed water solution is shown by three experiments (nos. 65, 53, and 53 A). In two experiments (65B and 53B) 40-percent bleed water was used. The average reduction in rate of flow resulting from a 10-minute treatment with these concentrations of bleed water was 32.6 percent. One experiment each was made with bleed water concentrations of 50, 60, 80, and 100 percent. There is comparatively little difference in the effects of these con- centrations during treatment for 10 minutes. However, the length of time during which no water was pumped through the gills increases steadily, from 10 minutes in the case of 50 percent bleed water to more than 2 hours for 100 percent. The 190 BULLETIN OF THE BUREAU OF FISHERIES exact extent of the nonflow period in the latter experiment is not known as the oyster began recovery sometime during the night. Table 18. — The effect of oil well bleed water on the rate of pumping of water by the gills of the oyster [Drop-counting method. Records of two experiments made in Beaufort, 1933] Experiment No. Date Drops per minute, average by 5-minute intervals Time treat- ment (min- utes) Percent. bleed water 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 55 24 10 65 59 57 63 71 63 39 62 66 70 61 66 10 63 ___ 4 10 130 130 123 125 121 119 101 119 104 109 115 107 5 03 B Dec. 5 10 41 43 41 41 45 35 29 27 43 42 36 39 5 64 --do_ 10 61 60 70 74 65 63 79 76 83 83 72 85 5 65C. . Dec. 8 10 96 92 89 86 88 101 105 96 88 85 86 91 10 23 20 74 81 77 83 84 54 25 67 71 68 68 10 6 20 74 76 82 78 89 83 78 71 78 91 77 84 5 04 C ..do. 20 99 95 96 90 88 81 57 69 79 85 92 112 10 8 20 90 95 92 91 88 89 86 87 98 94 94 91 10 53 Nov. 21 33.3 67 76 73 69 64 59 38 59 60 70 70 71 10 do_ 33.3 88 89 91 86 87 91 80 90 96 98 102 92 21 68 96 11 53 B Nov. 22 40 119 116 126 116 123 135 56 81 114 130 120 21 32 53 59 10 65 7 33. 3 69 75 76 77 77 62 77 85 87 90 94 10 653 __ 8 40 89 88 80 88 84 37 1 18 86 99 111 106 10 67A Dec. n 50 52 61 56 59 58 19 20 0 2 18 34 39 41 47 14 63A Dec. 12 60 52 69 77 81 76 38 2 0 0 9 26 31 37 37 38 10 68 B Dec. 13 80 93 89 77 8 1 0 0 0 0 0 0 0 11 19 22 10 69 A Dec. 15 100 79 89 85 9 7 2 0 0 0 0 0 0 0 0 0 10 Undiluted bleed water as it comes from the wells, has, as stated above, a specific gravity of about 1.1064 (17.5° C.). A 20-percent solution of brine has approximately the same specific gravity as the lab- oratory sea water during November and December. Higher concentra- tions of brine undoubtedly owe part of their inhibiting action on the gill cilia to an increase in density, but this effect is probably much less pronounced than that caused by its chemical composition. It has been found, for instance, that in sea water, the specific gravity of which was increased by the addition of salt to 1.0361, the reduction in rate of pumping was 2 percent, much less than the depression caused by a 20- percent solution of brine. Sea water having a specific gravity of 1.0422 (increased by adding NaCl) caused a decrease in the rate of flow by 20 percent, almost equal to that of the 20-percent solution of brine (spe- cific gravity 1 .0258). An inference from these observations is that the inhibitory effect is due to the chemical composition of brine rather than to the increased density of water. The percentage of brine in these experiments is beyond any possibility of occur- rence in nature, except under most unusual conditions, and the oyster itself is unable to endure so high a salinity for more than a few hours, so that the results of the experi- Percent concentrat/on of b/eed water Figure 10.— Depression of the rate of pumping of the gills caused by various concentrations of “bleed water” (brine). EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 191 ments are not considered to be significant insofar as the oil pollution problem is concerned, and are included here only for the sake of completeness. The experiments with bleed water show that it contains a substance or substances inhibiting the work of the ciliated epithelium. The action becomes apparent at a concentration of 10 percent brine in sea water. The toxicity increases in proportion to the increase in concentration. This is apparent from an examination of figure 10, in which the average depression in rate of flow of water through the gills, expressed in percentage of the normal, is plotted against the concentrations. EFFECT OF CONSECUTIVE TREATMENTS In the experiments described above the oysters were treated for a veiy short period of time lasting only a few minutes. The question arises as to whether or not as a result of a more prolonged exposure, they may become adapted or develop greater tolerance to the toxic substances present in crude oil and bleed water. As stated previously, the drop counting technique is not adapted for experimenta- tion over a period of more than 2 or 3 days. Several experiments were made in which the same oyster was treated with oil extract or bleed water several times within this period. Five oysters wTere used in a total of 16 experiments. The results are summarized in table 19. Thirteen of these experiments have been analyzed to determine the effect of 2 or more treatments. Table 19. — The effect of consecutive treatments with soluble fraction of oil and bleed water on the rate of pumping by the gills of the oyster Experiment no. Date Test solution Percent of solu- tion Effect of treatment Length of treatment Temper- ature 60 100 Percent 119. 6 Minutes 5 ° c. 13. 6 60A 100 62.8 5 15. 0 63 10 90.0 5 17. 3 63A 100 82.0 5 17. 4 63 B Bleed water 10 73.0 5 15.2 65 33 94.0 10 15. 0 65A Soluble fraction _ 100 112.7 5 14.8 65 B Dec. 8 40 41.0 10 14. 1 65C do 10 106.0 10 17. 1 65D 20 91. 0 10 15. 2 78 May 14 Soluble fraction 100 9.2 8 22. 1 79 do 100 14.0 8 22.4 80 100 18.5 8 23.3 81 May 15 do__ 100 60.3 8 22.9 82... 100 64.0 8 22. 5 83 do__ .. 100 26.2 20 22.3 Oyster no. 60 was used in two experiments, 60 and 60A. The duration of treat- ment with soluble fraction from Barre oil was 5 minutes in both. Five hours elapsed between treatments. The difference in temperature at the beginning of the 2 experi- ments was 1.4° C. In experiment 60, the rate of pumping was accelerated almost 20 percent above normal, wTdle in 60A there was a reduction in rate of about 37 percent. Oyster no. 63 was used in experiments 63, 63A, and 63B. Bleed water was used in 63 and 63B, and soluble fraction from Barre crude oil in 63 A. Two hours elapsed between 63 and 63A. 63B was made the following day, 18 hours after 63A. The duration of treatment was 5 minutes in each. 192 BULLETIN OF THE BUREAU OF FISHERIES Comparing the effect of the two treatments with bleed water, it will be noted that the second treatment reduced the rate of flow 27 percent as compared with a 10-percent reduction for the first treatment. The interval between these treatments was more than 20 hours. The difference in temperature at the beginning of these experiments was 2.1° C. Oyster no. 65 was used in 5 experiments, 4 with bleed water (65, 65B, 65C, and 65D), and 1 (65A) with soluble fraction from Pelto crude oil. Nos. 65 and 65A were made on December 7, with a 3-hour rest period between. The other three experiments were made the following day. The intervals between these experiments are 65A to 65B, 18 hours; 65B to 65C, 2.5 hours; 65C to 65D, 50 minutes. Com- parison of the effects of several treatments in this case is difficult because a different concentration of bleed water was used in each. However, it will be observed that the final treatment (65D) with 20 percent bleed water was of the same order of effectiveness as the first treatment with 33 percent bleed water. The time of treat- ment was 10 minutes in both experiments, and the temperature at the beginning of each was practically the same. Experiments 78, 79, and 80 were made with the same oyster, using the soluble fraction from Barre crude oil. The duration of treatment was 8 minutes for each and the temperature difference was negligible. Twenty minutes elapsed between 78 and 79; 3 hours elapsed between 79 and 80. The average reduction in rate of flow during treatment as shown in table 19 is 90.8 percent for experiment 78, 86 percent for experiment 79, and 81 .5 for experiment 80. Thus it would appear that succeeding treatments were not so effective as the first. However, it will be remembered that the figures in this table are prepared to show only the average effect during the period of treatment, which is considered to be the actual interval during which the test solution flows over the oyster plus an additional 10 minutes for replacing the test solution with fresh sea water. A detailed analysis of the records shows that while the initial reduction in rate of flow is greater in the first experiment, and consequently tbe average reduction appears to be greater, yet the period of no current is 20 minutes in the second experiment and only 15 minutes in the first. Actually, tbe second treatment exerted a retarding action on tbe pumping activity of the cilia for a longer period than was the case in the first experiment. These remarks apply also to the third experiment, no. 80. The initial depressing effect is not so great, as in no. 78, and tbe period of no current is only 15 minutes, but this period begins after the test solution has been turned off, and subsequent recovery as evidenced by the average rate of flow will be seen to take place more slowly than in either of the preceding two experiments. Experiments 81, 82, and 83 were made the same day with the soluble fraction from Barre crude oil. The interval between treatments is approximately 2 hours for experiments 81 and 82. Two and a half hours elapsed between treatments in 82 and 83. The duration of treatment was 8 minutes in the first two and 20 minutes in tbe third experiment. Tbe latter, therefore, cannot be compared directly with the others. Oyster no. 81 used in these experiments had the highest rate of pumping in the series and had a high resistance to treatment with soluble fraction. The reduction in rate of flow caused by the test solution was 39.7 percent in experiment 81 and 36 percent in 82. The average rate at 5-minute intervals as given in table 14 does not show an appreciable difference in effectiveness of the soluble fraction for the two experiments. EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 193 The results of 13 experiments may be summarized as follows: 1. Oysters subjected to several treatments of oil extract or bleed water do not develop higher tolerance to oil extract or bleed water in concentrations used, or within the time limits studied. 2. Particularly in the case of oil extract, the second and succeeding treatments have a less immediate but more prolonged retarding effect on the rate of pumping. 3. There is no evidence of permanent injury caused by this treatment. Re- covery is usually complete. EFFECT OF CRUDE OIL ON DIATOMS By Paul S. Galtsoff and Vera Koehring Since diatoms invariably appear to be constituents of the oyster diet, their rate of growth in the medium to which various polluting substances were added, may be taken as an index of the effect of pollutant upon the food supply of the oyster. In the regions where oysters grow, diatoms are usually distributed throughout the water, in the surface layers as well as on the bottom. Oil in water may affect their growth by forming a surface film which may interfere with the gaseous exchange between air and water or by the toxic action of water soluble constituents of the oil. It has been demonstrated by Lord Raileigh (1923) and Langmuir (1916, 1917) that the spreading of an oil upon water is due to the active carboxyl group of the oil molecule which because of its great affinity to water readily goes in solution, whereas the hydrocarbons having far greater attraction for one another than for the water, remain insoluble. By spreading into a monomolecular film the carboxyl groups combine with water without causing the separation of the hydrocarbon chains. Hence, a pure hydrocarbon oil, as has been demonstrated by Hardy (1912, 1919) for benzene and cymene, fails to spread. In case of oil in the natural waters the problem is more complex because we are usually dealing with mixed oils which, upon being exposed to air, change their chemical properties. Oxidized samples of oil, according to Hacker (1925), have an increased spreading power, and chemically inert hydrocar- bons upon acquiring certain radicles, —OH, — O, —COO, and — NH2, have an at- traction for water. For a detailed discussion of this problem the reader must consult the original papers of the above-mentioned authors, whose findings are briefly men- tioned here only as an illustration of the possible changes in the solubility and behav- ior of crude oil in water, which may account either for the formation of blobs or for the unchecked spread of oil film over a great area of the sea. A comprehensive in- vestigation of the problem would require extensive chemical studies of the oil before and after its discharge into the sea, which the authors were not in a position to under- take although they fully realized the necessity of such a study for the solution of the oil pollution problem. Oil discharged into the sea does not remain on the surface. Part of it is absorbed by the colloidal particles of clay present in the water and is gradually settled on the bottom. This can be seen in both the aquaria tanks in which oil-polluted sea water is kept and in many sections of the coastal waters affected by pollution. In a com- paratively short period of time considerable quantities of oil floating in the water is found absorbed and deposited on the mud bottom from which it can be separated by squeezing and decanting. Mud contaminated with oil may be considered as a possible source from which water soluble constituents of the oil gradually go into solution even after a complete disappearance of oil from the surface. 194 BULLETIN OF THE BUREAU OF FISHERIES Experiments carried out with diatom cultures and discussed in the present paper were devised to test the effect of oil (1) as a heavy layer covering the surface of the water; (2) absorbed by some neutral substance and held on the bottom of the culture flasks; and (3) as a soluble extract. METHOD Cultures of the single diatom species, Nitzschia closterium, E., were grown in the laboratory in solutions prepared according to the Miquel formula. (Solution A, potassium nitrate 20.2 g in 100 cc distilled water; Solution B, calcium chloride 4 g, sodium phosphate (secondary crystals) 4 g, ferric chloride 2 g, and 1 cc concentrated hydrochloric acid in 80 cc of water. Add 2 cc solution A and 1 cc solution B to one liter of carefully filtered sea water; sterilize by bringing just to the boilding point, then cool and filter.) Eighty cc of this solution were poured in round pyrex flasks of 150-cc capacity and inoculated by adding 2 or 5 cc of Nitzschia stock culture. Flasks covered with inverted small beakers were placed in front of the laboratory window where they were protected from direct sunlight. In some of the experi- ments instead of Miquel solution plain filtered sea water was used. Sea water used during winter experiments was received from Woods Hole, Mass., and stored in paraffined oak barrels. It contained noticeable amounts of hydrogen sulphide. Before using, the water was aerated for at least 24 hours and the precipitated sulphur filtered off. During the summer months the sea water from the laboratory supply was used. The salinity of the water varied between 31 and 32 parts per thousand. Twice a day temperature readings were made by means of a thermometer kept in one of the flasks filled with water. Fluctuations in the light and temperature con- ditions which were not controlled, are undoubtedly responsible for certain variations in the growth rates. To avoid this difficulty a comparison was always made between the experimental and the corresponding control flasks which stood by its side. The relative position of each pair of flasks was changed every other day to compensate for a possible difference in illumination. Diatom counts were made by means of a photoelectric set-up (fig. 11) which con- sisted of Weston photronic cell ( Ph ) connected to a Weston D. C. microammeter (A) of 0-100 ranges, accurate within 0.5 percent of the full scale value at any part of it. For measuring the turbidity of a sample, 50 cc of it were poured in a cylindrical glass container C, having a diameter of 27 mm, 90 mm high, with fused bottom made of optical glass. The cylinder rested on a diaphragm, the opening of which was slightly smaller than the inside diameter of the cylinder. Both the diaphragm and the cylinder were placed directly over the photoelectric cell the surface of which was covered with thick black paper in such a manner as to cut ail the light except that which passed through the column of water in the cylinder. A 21-candlepower, 6-volt Mazda bulb ( L ) fed by a 108 ampere-hour storage battery served as a source of light. The photoelectric cell was placed in a wooden box 4 by 4 by 12 inches. The bulb was mounted at the lower end of an adjustable arm which could be moved up and down and fixed in a desired position by a set screw. A piece of fine ground glass ( Gl ) separated the two compartments of the box. The front part of the box had a door through which the glass container could be placed in position. The box was painted inside with black paint. Great care was exercised in avoiding the fluctuations in the voltage and in insuring uniformity in the intensity of light. It was noticed that there was a drop in voltage during the first 10 minutes following the EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 195 turning on of the battery and small fluctuations occurred every time the current was turned off and on. To avoid this difficulty a second Mazda bulb (Z) of the same candle power was introduced into the circuit. By means of a double switch (S) the current from the storage battery could be turned either through the bulb mounted in the box or through the second one located outside. During the observations the current was continuously passing through the spare bulb except during brief moments when measurements of the samples were made. By this arrangement the overheat- ing of the box and difficulties due to the fluctuations in the voltage of the storage battery were overcome. The battery was kept well charged. Figure 11. — Photoelectric set-up used for counting number of cells in Nitzschia culture. Ph, photoelectric cell; A, microammeter; C, glass container with fascil bottom made of optical glass; L and Li, electric bulbs; Ql, ground glass; S, double key switch; and B, storage battery. For the standardization of the apparatus the bulb was so adjusted that the light passing through the container filled with 50 cc of filtered sea water caused the needle deflection of the microammeter to stand at 30. This represented the zero point of the calibration curve. Other points were obtained by making readings with various con- centrations of Nitzschia and counting by means of Sedgwick-Rafter chamber the corresponding number of cells. Altogether 35 samples were counted and from the data obtained, a curve was plotted which permitted the conversion of microammeter readings into number of diatoms per cubic centimeter. Before withdrawing a sample for measurement each culture flask was thoroughly shaken for 2 minutes then 50 cc were taken by means of a certified volumetric pipette, and poured into container c. Care was taken to avoid air bubbles, the presence of which affected the ammeter readings. Upon measurement, which required about 30 seconds, the sample was returned to the same flask and both the container and pipette carefully washed in sterile distilled water and dried. Because of the sim- plicity and quickness of the operation it was possible to experiment with several dozens of flasks simultaneously. 196 BULLETIN OF THE BUREAU OF FISHERIES An inconsiderable bacterial population was always present in the cultures. The bacteria were examined at intervals by plating them in agar medium prepared according to Waksmann’s formula (1,000 cc sea water, 1 g peptone, 1 g glucose, 0.5 K2HP04, 2-3 drops of 10 percent FeCl3, and 15 g agar; sterilized in autoclave at 15 lbs. for 20 minutes). When the diatoms and the surrounding medium are relatively free of bacteria the Nitzsc.hia cells stay in suspension. When they tend to fall to the bottom and form loosely aggregated masses, microscopical examination shows them to be weighted with adherent bacteria. Plating under such conditions invariably shows a high bac- terial count. All subcultures from the stocks were made, therefore, by carefully decanting only top portions containing diatoms in suspension. At the beginning of each experiment all the flasks were inoculated with equal volumes of stock culture and 1 or 2 of them measured to determine the initial diatom population. Thereafter the cultures were measured every other day. At least 3 control and 3 experimental flasks were used in each test. In many experiments this number was doubled. All the figures given in tables or graphs are averages of three or more samples. EFFECT OF HEAVY SURFACE LAYER OF OIL ON NITZSCHIA CULTURE In this set of experiments Nitzsc.hia was grown in 500 cc Erlenmeyer flasks con- taming 250 cc of culture covered by 25 cc of oil. The oil was sterilized in a boiling water bath for 1 hour. There was no indication that the presence of oil kills the Nitzschia and no immediate effects on its growth were discernible even when the oil was thoroughly shaken up with the cultures several times a day. Microscopic examination showed no signs of oil sticking to the surface of the diatoms. For approximately a week the experimental cultures showed no significant difference in their growth as compared with the controls. Their further propagation was, however, markedly inhibited. Photoelectrical measurements of the cultures presented considerable difficulties as it was some times impossible to withdraw a portion of the sample free from oil globules. I11 two experiments, however, this difficulty was overcome. The first experiment, lasting 18 days, began on March 13 and was discontinued April 4, 1934. The second one, started on April 7, continued for 25 days until May 4. During the last experi- ment tests were made with purified mineral oil (Russian oil) and cod-liver oil. Cul- tures covered with cod-liver oil perished on the fourth day, those with mineral oil continued to the end of the experiment. The retarding effect of oil was noticeable in both groups, although in the cultures under Russian oil it was somewhat less pronounced than in those kept under crude Pelto oil (Fig. 12.) The adverse effect of oil can be measured by determining the percent of retardation of growth in the experimental flasks as compared with their growth in the controls. In the first experi- ment (March 13, Pelto oil) the retardation at the eighteenth day amounted to 35 percent. In the second experiment Casting 25 days) retardation due to presence of Russian oil and Pelto oil was 37 and 42 percent, respectively. An inhibitory effect of a heavy surface layer of oil on the diatom culture was quite visible in many other samples which could not be measured. This can be seen in the photographs (fig. 13) taken 30 days after the beginning of one of the experiments in February 1934. The control culture (left) had numerous diatoms and therefore EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 197 appears dark, whereas tlie flask on the right containing culture covered with oil had few diatoms, thus appearing much lighter. All the experiments indicate that a heavy, unbroken surface layer of oil inhibits diatom growth when oil remains on the surface for a week and longer. EFFECT OF OIL HELD ON THE BOTTOM In attempting to devise a method whereby oil could be held at the bottom of the flask and would not interfere with the photoelectric measurements the following pro- cedure was developed. Four grams of paraffin, melting point 52° to 55° C., was mixed with 6 drops of oil. The resulting mass, oily to the touch, and having a strong odor, Figube 12. — Growth of Nitzschia culture under heavy layer of Pelto oil. was sterilized at 150° C. and allowed to cool on the bottom of each flask and strongly adhering to it. Pure paraffin controls were prepared in the same manner. The flasks were filled with the standard Miquel medium, inoculated with Nitzschia and measured at regular intervals. The results of the experiments were, unfortunately, inconclusive. In three experiments the growth of Nitzschia in experimental flasks was retarded from 4 to 16 percent (fig. 14), as compared with their growth in pure sea water. In tln-ee other experiments there was a noticeable stimulating effect varying from 17 to 27 percent. Pure paraffin controls also produced inconsistent results. In one experiment in 13 days the number of diatoms reached exactly the same number as in the plain sea water. In other experiments the paraffin cultures 198 BULLETIN OF THE BUREAU OF FISHERIES showed a retardation by 8-9 percent, while in one experiment the growth in pure paraffin and paraffin-oil cultures was increased by 17 percent (fig. 15). Indefinite results of the experiments should probably be attributed to the action of bacteria which, as has been shown by Hopkins and Chinbal (1932), Buttner (1926), Tausson (1927), and Haas (1926), in the absence of more suitable material can grow on paraffin and utilize it as their only source of carbon. A great part of our paraffin and paraffin-oil cultures showed very abundant bacterial growth. Other attempts to incorporate oil in some suitable substance at the bottom of culture flasks were unsuccessful. EFFECT OF WATER SOLUBLE FRACTION OF OIL Two different methods were used in preparing water soluble extracts of oil. The first method consisted in stirring together oil and filtered sea water and allowing the mixture to stand for various lengths of time. In the second method measured amounts of oil were poured in collodioD bags suspended in flasks containing Nitzschia culture and the water soluble constituents of the oil gradually diffused through the membrane. The proportions of oil and water used in preparation of the extracts according to the first method as well as the duration of stirring and standing are given in table 20. Table 20. — Preparation of oil extracts in sea water Proportions oil to water Hours of stirring Days of standing Specific gravity of sea water (17.5° C./17.50 C.) Specific gravity of extract (17.5° C./17.50 C.) 11 12 12 1. 02433 1. 02400 12 2 2 1. 02461 1. 02380 12 2 3 1. 02435 1. 02434 1:2 12 10 1. 02463 1. 02434 1:2 3 7 1. 02445 1. 02448 The oil was subsequently filtered off and the resulting clear aqueous extract was heated to boiling to allow for sterilization. Before sterilizing the specific gravity of the extract was determined. In most cases it was slightly lower than that of the sea water used. The results of these experiments are summarized in table 21 which shows the percent of retardation ( — ) or stimulation (+) of growth caused by the addition of various concentrations of five different extracts. All the experiments were carried out with Lake Pelto oil. The figures are the averages of three or more samples. Table 21. — Effect of water soluble fraction of oil on the growth of Nitzschia culture [Figures in the body of the table represent percent of retardation (— ) or stimulation (+)] The figures given in table 21 represent the end results of experiments which ran from 9 to 21 days. Of the 25 experiments with various dilutions of oil extract 6 U. S. Bureau of Fisheries, 1935 Bulletin No. 18 Figure 13. — 30-day Old Nitschia Culture in Sea Water (Left), and in Sea Water Under Heavy Layer of Oil (Right). (February 1934.) EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 199 5 ^ 45 Vj ''n 40 - §>55 - C3 50 - ^ 25 20 L. '5 Q ^ 10 i 1 1 1 1 1 1 r n ; r GROWTH or N1TZSCHIA CULTURE IN 5EA WATER WITH OIL HELD IN PARAEFIN _ .... Paraffin p/us oil. Paraffin Control- Oil m paraffin plus oil extract I l I I L I J ! I 1 L 10 // 12 IJ M 15 16 n March / 934 Figube 14. 19 20 2/ Figure 15. 200 BULLETIN OF THE BUREAU OF FISHERIES showed increased growth as compared with the controls, while in 19 cases there was a retardation of the propagation or great mortality, accompanied by tremendous development of bacteria. In 14 experiments the population during some period of growth was greater than in the controls. These periods of increased rate of propa- gation occurred usually during the first few days of growth. Thereafter the controls continued to increase more rapidly and by the end of the experiment attained greater populations. These relationships may be noted by examining figure 16 representing the results of 3 experiments. In all the experiments with oil extracts, the retarding effect of the latter is appar- ent in concentrations of 25 percent and higher and when the extract is permitted to act over a considerable period of time. EFFECT OF OIL HELD IN COLLODION BAGS The method used in these experiments consisted in suspending in a flask contain- ing SO cc of culture medium, a collodion bag with 2 cc of crude Pelto oil. Bags were made by dipping a test tube in a thin solution of collodion (Merck’s reagent) and drying it for 45 seconds. Freshly made bags were sterilized by boiling in distilled water. Each bag remained suspended in a culture medium throughout the experi- ment. A more or less continuous dialysis through the collodion membrane was noticed, for water gradually collected in the bottom of the bag. The colorless and apparently somewhat volatile dialyzed substances have not been identified. Their presence was readily detected, however, by odor and irritating effect on the mucous membrane of the mouth while pipetting the sample for measurement. All the controls contained collodion bags filled with the Miquel solution or sea water. The end results of the experiments are given in table 22 the examination of which discloses that in all the cases there was a noticeable retardation in Nitzschia growth caused by oil held in collodion bags. Figure 17 presents a more complete record of one of the experiments. Table 22. — -Retardation of growth of Nitzschia culture caused by oil held in collodion bags [Each figure is an average of measurements of 6 different flasks] Beginning of experiment Duration, days Retardation of growth, percent Medium July 26 --- - 22 20 Miquel. Do. 12 24 Do 12 12 Aug. 22 14 32 Do. EFFECT OF BRINE ON NITZSCHIA Untreated brine (bleed water) separated from the oil as it comes out of the well was filtered from the heavy sediments accompanying it. The filtrate was carefully brought to the boiling point for sterilization. Sometimes this caused precipitation and the sample had to be discarded. In a series of experiments lasting from 14 to 31 days tests were made with various concentrations of brine added to sea water or Miquel solution. As can be seen by examining figures 18 and 19 representing two typical experiments, the retardation EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 201 Figure 16. Figure 17. — Growth of Nitzschia in sea water with oil kept in collodion bag suspended in flask. 202 BULLETIN OF THE BUREAU OF FISHERIES of Nitzschia growth is apparent in both media. An analysis of the results of all the experiments with brine summarized in table 23, shows that the retarding effect is more apparent in the sea water than in the Miquel solution. A 12-percent con- centration of brine causes only 10 percent retardation of Nitzschia growth in the Miquel solution, whereas in the sea water retardation averages 41 percent. Table 23. — Retarding effect of brine on the growth of Nitzschia culture [Figures in the body of the table indicate percent of retardation] Concentration of brine Dec. 23 (16 days) Jan. 8 (22 days) Feb. 2 (31 days) Feb. 2 (31 days) Apr. 2 (14 days) Apr. 2 (14 days) Average percent of retardation Percent in Miquel solution: 12 23 11 0 12 8 8 10 18.8 48 3 36 17 26 25 58 28 40 38 35 30 38 37.5 75 63 69 50 93 90 92 Percent in sea water: 12 13 66 40 46 41 18 36 74 61 57 25 31 75 79 80 66 37 76 87 82 50 77 99 88 EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 203 Both in the sea water and in Miquel solution the retarding effect is directly proportional to the concentration of brine (fig. 20). From the results of the experiments it becomes apparent that brine contains substances which interfere with the growth and rate of propagation of the diatom. F ?bruary 1934 March Figdee 19.— Effect of brine on the growth of Nitzschia in Miquel solution. The experimental work reported in this part of the report provides sufficient evi- dence of the possible adverse effect of crude oil and bleed water on the food supply of oysters and other plankton feeding animals of the sea. Probably the retarding effect of crude oil on the rate of diatom growth is primarily due to the toxic action of its water-soluble constituents. The extraction of this substance or substances is undoubtedly facilitated by the action of wind and current. Oil absorbed by mud 204 BULLETIN OF THE BUREAU OF FISHERIES and deposited on the bottom may be therefore regarded as a source of potential danger to the microscopical algae. When the bottom is stirred by passing boats or by a strong wind action, a certain amount of oil may again become released and float in Figuke 20.— Retardation in growth of Nitzschia culture in sea water and Miquel solution caused by various concentrations of brine. the water giving off water-soluble toxic substances. The adverse effect of brine is even more pronounced than that of the oil. DISCUSSION AND CONCLUSIONS By Paul S. Galtsoff Ecological and hydrographical observations presented in the first half of the report describe the conditions of the oyster bottoms affected by the oil-well pollution. Preliminary investigations carried out by Prytbercb in 1933 failed to reveal the exist- ence of a direct correlation between the intensity of mortality and the distance be- tween the affected oyster bottoms and oil wells. This is evident from an examination of the chart (fig. 1), showing the intensity of mortality on various oyster beds (black circles) and the location of active oil wells. A number of oysters, barnacles, and green algae were found growing on the piling of oil wells, and no unusual mortality was ob- served among other organisms. The presence of small numbers of oysters on piling of oil wells was also observed in 1934 by Galtsoff and Smith. Examination of oysters and plankton showed that apparently there was no interference with the development of gonads, spawning, and setting of the larvae. The diseased condition of oysters was evidenced, however, by the loss of muscular tonus and the failure of the adductor muscle to maintain closure of the shell. It is known that if such a condition continues for a long time it results in a stunted growth and abnormal shape of the shell. No unusual changes in the salinity of water and other hydrographic conditions, which might account for a great mortality, were disclosed by these observations. The oyster enemies, the borer, the boring clam, and the boring sponge, are rather abundant in Louisiana waters. Many dead oysters examined in 1933 showed heavy perforations caused by the boring clam and sponge, but on the other hand, at least EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 205 at one station (no. 6, fig. 1), 95 percent of dead shells were not infested by these pests. Oyster growers have not noticed an unusual increase in abundance of oyster enemies, and no evidence has been obtained which would indicate that such an outburst oc- curred at the time of the mortality in the winter of 1932-33. It is significant that in 1933 the mortality affected chiefly the larger and older oysters of marketable size and in several instances was especially severe among the recently transplanted oysters. Undoubtedly, the practice of overcrowding the beds by planting from 700 to 900 barrels of oysters per acre may be one of the contributing factors which aggravates the situation and in case of adverse environmental changes or the poor condition of the oysters, may materially increase their mortality. A more detailed survey of the oyster bottoms, made by R. O. Smith in 1934, failed to assign the mortality to any known disturbance of the natural conditions on oyster beds, as for instance, temperature, salinity, current, and invasion of enemies. It has been noticed that in general mortality has been higher on soft, muddy bottoms than on hard ground. At the time this survey was carried out, pollution was noticeable at the mouth of Bayou Grey where the surface of the water was covered with oil for a distance of 3 miles below the wells and there was some mortality on the oyster beds of this section. All shells were covered with a brownish-black coating of tarry con- sistency and the meats were unpalatable because of the strong oily flavor. Consider- able quantities of oil were held by mud, and oily patches appeared on the surface when the bottom wTas stirred. Light films of oil were observed also in the vicinity of the Lake Barre wells. In 1934 oysters on many beds throughout the region did not be- come fat until February or March, which points to a possible scarcity of food or to a disturbance in the functioning of the organs of feeding. The shallowness of the water throughout the oyster-producing region in Louisi- ana must be regarded as a factor which tends to magnify the action of any polluting substance. Due to stirring by wind, the water carries much suspended matter which may absorb the pollutant, transport it over wide areas, and deposit it on the bottoms far from the source of pollution. Observations in the polluted areas show that on account of the absorption by suspended clay particles, oil quickly disappears from the surface and after being deposited on the bottom, remains there for a long time. No information was obtained by the two surveys upon which to base an opinion as to the direct cause of mortality, but ample experimental evidence has been accumu- lated to show that the presence of crude oil in water produces conditions inimical to oysters. The first series of experiments designed to determine whether oysters could be killed by the presence of oil in the water or by direct contact with oil, gave negative results. Unfortunately because of the circumstances over which the investigators had no control, these experiments were carried out not in Louisiana, but in a different environment at Beaufort, N. C., with uncultivated oysters taken from oyster reefs. Samples of crude oil collected by the State Conservation Department from Louisiana oil wells were shipped to Beaufort and used throughout the experiments. It is quite possible that the results might have been different had Louisiana oysters been used. In a series of experiments lasting from 2 to 3 months, the mortality of oysters kept in running sea water under a surface layer of oil, and those kept in sea water that passed through oil was not greater than that in the controls. In the experi- ments carried out under similar but not identical conditions, Gowanlach (1934) observed considerable mortality among the oysters kept in oil-polluted water. The 206 BULLETIN OP THE BUREAU OP FISHERIES discrepency between the two sets of experiments may be due to the difference in technique or to the better condition of North Carolina oysters used in Prytherch’s experiments. In another set of laboratory experiments no higher mortality than that in con- trols was observed among the oysters which, over a period of 6 to 8 weeks, were im- mersed at regular intervals in oil (table 3, p. 163). In some of the experiments the mortality among the controls was as high as 50 percent, indicating unfavorable labo- ratory conditions under which the animals were kept. It is possible that these conditions beclouded the effect of oil on oysters. The fact that oysters survived the treatment with oil does not indicate that they were not affected by it. Analyses made by Galtsoff show slight decrease in glycogen content of oysters kept in the laboratory in the oil-polluted water (table 4). The result may be due either to the disturbance in the functioning of the feeding apparatus of the organism or to the decreased supply of food. A regular operation of the muscular mechanism involved in closing and opening of the shell is prerequisite for the normal feeding of the oyster. Two sets of experi- ments, carried out by Prytherch in 1933 and Galtsoff and Smith in 1934, gave identical results showing that the presence of oil has no effect on the mechanism of the adductor muscle. In the first set of experiments, continuous kymograph records were obtained of 5 oysters which were kept under observation for 3 months. The average number of hours per day the oysters were open was 11.2 for the controls and varied between 10.0 and 13.6 for the experimental oysters. In the second set of experiments, 6 control oysters kept under observation from 4 to 14 days, were open on the average of 10.5 hours daily, whereas the average figure for 10 experimental oysters kept under ob- servation from 4 to 8 days, was 9.6 hours. In both cases the difference is insignificant. Although the presence of oil in the sea water does not reduce the number of hours the oyster keeps its shell open, and therefore the duration of feeding of the mollusk is not decreased, the rate of feeding is easily affected by the presence of polluting substance. As the feeding of the oyster is primarily dependent upon the amount of water passed through the gills, the rate of pumping of water can be used as a measure of the rate of feeding. The results of the experiments in which the cone method, previously described by Galtsoff (1928) was used, and of those in which the drop counting technique was employed (fig. 5), show that crude oil contains substances soluble in the sea water which produces anaesthetic effect on the ciliated epithelium of the gills. The inhibiting action is not due to the mineral salts that may be leached out in preparing the water soluble fraction of the sample of oil by shaking it with sea water. It is apparent that certain organic compounds of oil are slightly soluble in sea water. This conclusion can be drawn from the observations that after 28 washings with water, the sample of oil did not lose its toxic property and yielded extract, the anaesthetic potency of which was equal to those obtained with the first washings. The inhibiting effect of the water-soluble fraction is proportional to its concentration (figs. 6, 7, 8, and 9). From a large number of experiments summarized in these figures and tables 11, 12, and 13, the inference can be drawn that a concentration between 20 and 30 percent of the soluble fraction will, on the average, reduce the rate of feeding of the oyster to one-half of its normal value (fig. 8). Under the conditions of the experiments, the recovery of the ciliary motion following the removal of the oil extract, was almost complete. Inasmuch as the EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 207 experimental oysters were kept in the extract only for a limited period of time, the result of the prolonged exposure remains to be determined. There was no indication in the present experiments of an increased tolerance in oysters due to repeated treatment. As it can be seen in tables 11, 12, 13, and 14, there was a large variation in the percentage of depression caused by a given concentration of the soluble fraction on individual oysters. Two explanations suggest themselves. First, there is a possibil- ity that in spite of the precautions taken in preparing the soluble fraction, the toxicities of individual samples were different. Second, the oysters used in the experiments may have different sensitivity and tolerance. The second explanation seems to be more plausible, for the wild oysters used in the experiments at Beaufort, coming from exposed flats, greatly varied in appearance, glycogen content, and other characteristics. From a large number of experiments with the water-soluble fraction the inference seems to be inevitable that crude oil discharged in the sea, regardless of whether it floats on the surface or, having been absorbed by mud particles, is deposited on the bottom, continuously yields water-soluble substances which narcotize the ciliated epithelium of the gills, thus reducing the rate of pumping of water and, therefore, materially decreasing the amount of food obtained by the organism. This should lead to gradual starvation and weakening of the oyster. The chemical nature of the substances and their concentration in the oil-polluted areas remains to be determined by future investigations. The effect of brine or so-called “bleed water”, which accompanies oil discharged by the wells and is usually dumped into the sea, has been studied by using the same technique as was employed in the experiments with oil. It has been found (table 6), that bleed waters of Lake Barre and Lake Pelto do not affect the muscular mechanism of the oyster in relatively high concentrations, provided the quantity present does not increase the salinity beyond the limits of tolerance. A 10-percent concentration of bleed water may exert a stimulating effect on the ciliated epithelium at least in some of the individuals. The depressing effect occurs at the concentration of 20 percent and higher. A 33-40 percent solution reduces the rate of pumping of water by the gills to 32.6 percent of its normal rate. The percentages of brine which cause this or greater depression are beyond any possibility of occurrence in nature. Experiments reported in the last section of the paper attempt to throw light on the possible effect of oil and bleed water on the production of the food of the oyster. It has been assumed that the results of the laboratory experiments with Nitzsc.hia, which occurs in the normal habitat of the oyster, and constitutes an important element in its diet, are applicable to other species of diatoms. It has been found that the presence of a heavy layer of oil on the surface of culture flasks inhibits the growth of Nitzschia (fig. 13) when oil remains on the surface for a week or longer. The soluble fraction of oil exerts a retarding effect on the growth of Nitzschia in concentration of 25 percent and higher and when the extract is permitted to act over a considerable period of time. Low concentration may have a slightly stimulating effect. In many instances the addition of the oil extract stimulated the growth of bacteria, small numbers of which were always present in cultures, and caused the death of diatoms. Water-soluble substances obtained by dialysis through a collodion membrane also exerted a retarding effect on Nitzschia, both in natural sea water and in Miquel solution (table 22). 208 BULLETIN OF THE BUREAU OF FISHERIES Bleed water retards the growth of Nitzschia, the inhibiting effect being pronounced in concentrations of 10 percent and higher. The retardation of growth is directly proportional to the concentration (fig. 20). The experimental evidence presented in the report shows that the discharge of oil into the sea produces profound changes in the normal environment of the oyster. The substances which gradually dissolve from oil in the sea water irritate the delicate ciliated mechanism. In a very dilute solution they may act as stimulants, but in higher concentrations they inhibit the activity of the ciliated epithelium and may bring about complete stoppage of the current of water through the gills. The same substances which reduce the rate of feeding of the organism affect its food supply by retarding the rate of propagation of diatoms. Obviously the presence of oil creates adverse conditions. In the light of the present investigation, it is easy to conceive that when the con- stitution of the organism is weakened by unfavorable meteorological conditions, natural changes in the environment or attacks of enemies, the pollution of water with oil may become a deciding factor which may cause irreparable injury and death of the oyster. It is obvious that from the point of view of conservation, the natural oyster resources of the sea must be protected from tliis danger. BIBLIOGRAPHY American Petroleum Institute. 1933. Pollution of surface water by oil in disposal of refinery wastes. Section 1. New York. Buttner, H. 1926. Zur Kenntnis der Mykobakterien, insbesondere ihres quantitativen Stoff- wechsels auf Paraffinnahrboden. Archiv fur Hygiene, Bd. 97, pp. 12-27. Cary, L. R. 1907. A preliminary study of the conditions for oyster culture in the waters of Terrebonne Parish, Louisiana. Bull. No. 9, Gulf Biol. Sta., Cameron, La., pp. 1-62. Baton Rouge. Clarke, Frank W. 1924. The data of geochemistry. 5th Ed. Bull. No. 770, U. S. Geol. Survey, 881 pp. Dabney, Thomas Ewing. 1934. South Louisiana produces 60 percent more oil than northern section of State. La. Conser. Rev., vol. IV, no. 3, 1934, pp. 26-32, 43, 3 figs. New Orleans. Elmhirst, R. 1922. Investigations on the effect of oil tanker discharge. Ann. Report, 1922, Scottish Marine Biological Association. Fiedler, R. H. 1932. Fishery industries of the United States, 1931. Appendix II, Report, U. S. Com. Fish., 1932 (1933), pp. 97-440. Fowler, H. C. 1933. Petroleum and natural-gas studies of the United States Bureau of Mines. Information Circular No. 6737, U. S. Bur. Mines, 50 pp. Fox, J. J., and A. J. H. Gange. 1920. Note on the determination of tar acids in drains. Drainage from tarred roads. Jour. Soc. Chem. Ind. Transactions, Aug. 16, 1920, 3 p. 260T. London. Galtsoff, P. S. 1928. Experimental study of the function of the oyster gills and its bearing on the problems of oyster culture and sanitary control of the oyster industry. Bull., U. S. Bur. Fish., vol. XLIV, 1928 (1929), pp. 1-39, 13 figs. Washington. Gardiner, A. C. 1927. The effect of aqueous extracts of tar on developing trout ova and on alevins. Min. Agri. and Fish., Great Britain. Fishery Inves., series I, vol. Ill, no. 2, 14 pp. London. Gowanloch, James Nelson. 1934. Oyster beds and oil wells, a complex conservation problem. La. Conser. Rev., vol. IV, no. 2, 1934, pp. 2, 47-48. New Orleans. Gctsell, J. S. 1921. Danger to fisheries from oil and tar pollution of waters. Appendix VII, Report, U. S. Com. Fish., 1921 (1922), 10 pp. Hacker, H. P. 1925. How oil kills Anepheline larvae. Inst. Med. Research, Federated Malay States, Malaria, Bur. Reports, vol. Ill, 1925, 62 pp. London. EFFECT OF CRUDE OIL POLLUTION ON OYSTERS 209 Haag, F. E. 1926. liber die Bedeuting von Doppelbindungen in Paraffin des Handels fur das Wachstum von Bakterien. Archiv fur Hygiene, Bd. 97, pp. 28-46. Hardy, W. B. 1912. The tension of composite fluid surfaces and the mechanical stability of films of fluid. Proc. Roy. Soc., series A, vol. LXXXVI, pp. 610-635, and series A, vol. LXXXVII, pp. 313-333. Hardy, W. B. 1919. The spreading of fluids on glass. Phil. Mag. and Jour. Science, sixth series, vol. XXXVIII, pp. 49-55. London, Edinburgh, and Dublin. Heithecker, R. E. 1932. Some methods of separating oil and water in west Texas fields, and the disposal of oil-field brines in the Hendricks oil field, Texas. Report of Inves., no. 3173, Bur. Mines, 16 pp., 7 figs. Hopkins, A. E. 1931. Temperature and the shell movements of oysters. Bull., Bur. Fish., vol. XLVII, pp. 1-14, 10 figs. Hopkins, A. E. 1932. Sensory stimulation of the oyster, Ostrea virginica, by chemicals. Bull., Bur. Fish., vol. XLVII, pp. 249-261, 11 figs. Hopkins, S. J., and A. C. Chibnall. 1932. Growth of Aspergillus versicolor on higher paraffins. Biochem. Jour., vol. XXVI, pp. 133-142. House of Representatives. 1930. Pollution of navigable waters. Hearings before the Com- mittee on Rivers and Harbors, 71st Cong., 2d sess., on H. R. 10625, 89 pp. Government Printing Office, Washington. Interdepartmental Committee. 1926. Oil pollution of navigable waters. Report to the Secretary of State by the Interdepartmental Committee, March 13, 1926, 119 pp. Govern- ment Printing Office, Washington. James, M. C. 1926. Report of the United States Bureau of Fisheries: Preliminary investigation on effect of oil pollution on marine pelagic eggs, April 1925. In “Oil pollution of navigable waters”, Report to the Secretary of State by the Interdepartmental Committee, Appendix 6, pp. 85-92. Langmuir, I. 1916-17. The constitution and fundamental properties of solids and liquids. Part I, Jour. Amer. Chem. Soc., vol. 38, 1916, pp. 2221-2295, and Part II, vol. 39, 1917, pp. 1848-1906. Lane, F. W., A. D. Bauer, H. F. Fisher, and P. N. Harding. 1924. Effect of oil pollution of coast and other waters on the public health. Public Health Reports, vol. 39, 1924, pp. 1657- 1664. Reprint No. 936, Government Printing Office, Washington. Lane, F. W., A. D. Bauer, H. F. Fisher, and P. N. Harding. 1926. Effect of oil pollution on marine and wild life. Appendix V, Report, U. S. Com. Fish. 1925 (1926), pp. 171-181. Leenhardt, H. 1925. De l’Action du Mazout sur les coquillages. Rapp, et Proc6s-Verb., Cons. Perm. Internat. Explor. Mer, vol. XXXV, pp. 56-58. Copenhague. Macadam, S. 1866. On the poisonous nature of crude paraffin oil and the products of its rectifi- cation upon fish. Reports, British Assoc. Adv. Science, vol. XXXVI, Notices and Abstracts, pp. 41-43. London. Mitchell, Philip H. 1914. The effect of water-gas tar on oysters. Bull., Bur. Fish., vol. XXXII, 1912 (1914), pp. 199-206. Moore, H. F., and T. E. B. Pope. 1910. Oyster culture experiments and investigations in Louisiana. Report, U. S. Bur. Fish., 1908 (1910), 52 pp., 8 pi. Moore, H. F. 1899. Report on the oyster beds of Louisiana. Report, U. S. Fish Com., 1898 (1899), pp. 45-100, 3 pis. Orton, J. H. 1924. An account of investigations into the cause or causes of the unusual mortality among oysters in English oyster beds during 1920 and 1921. Part 1. Reports, Min. Agric. and Fish., Great Britain. Fish Inves., series 2, vol. 6, no. 3, 199 pp., 12 pis., 9 figs. London. Rayleigh, L. 1923. Studies of iridescent colour and the structure producing it. I. The colours of potassium chlorate crystals. Proc. Roy. Soc., series A, vol. CII, pp. 668-674. 210 BULLETIN OF THE BUREAU OF FISHERIES Redeke, H. C. 1927. Report on the pollution of rivers and its relation to fisheries. Rapp, et Proces-Verb., Cons. Perm. Internat. Explor. Mer, vol. XLIII, 50 pp. Copenhague. Roberts, C. H. 1926. Oil pollution. The effect of oil pollution upon certain forms of aquatic life and experiments upon the rate of absorption, through films of various fuel oils, of atmos- pheric oxygen by seawater. Jour. Cons., Perm. Internat. Explor. Mer, vol. 1, pp. 245-275, 3 figs. Copenhague. Rushton, W., and E. C. Jee. 1923. Fuel oil and aquatic life. Salmon and Trout Mag., no. 31, pp. 89-95. London. Shaw, J. A. 1933. A brief survey of the mineral resources of Louisiana. General Bulletin Hand- book (Minerals division). Bull. No. 22, Louisiana Department of Conservation, pp. 31-125, illus. New Orleans. Schmidt, L., and J. M. Devine. 1929. The disposal of oil-field brines. Report of Investigations, Bur. Mines, no. 2945, 17 pp., 3 figs. Washington. Tausson, W. O. 1927. Uber die Oxydation der Wachse durch Mikro-organismen. Bioehem. Zeitschrift, Bd. 193, Heft 1-3, pp. 85-93. o U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner AGE AND GROWTH OF THE CISCO, LEUCICHTHYS ART ED I (LE SUEUR), IN THE LAKES OF THE NORTHEASTERN HIGHLANDS, WISCONSIN By RALPH HILE From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 19 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1936 For sale by the Superintendent of Documents, Washington, D. C. Price 30 cents AGE AND GROWTH OF THE CISCO, LEUCICHTHYS ARTED1 (LE SUEUR), IN THE LAKES OF THE NORTHEASTERN HIGHLANDS, WISCONSIN1 * By Ralph Hile, Ph. D., Assistant Aquatic Biologist, U. S. Bureau of Fisheries «3t CONTENTS Page Introduction 212 Acknowledgments 214 Materials 214 Methods 215 Gear used in collecting 215 Methods of fishing 215 Field data recorded for individual specimens 216 Treatment of preserved specimens and resulting shrinkage 217 Preparation and examination of scale material 217 Miscellaneous considerations 218 The scale method 218 Assessment of age and calculation of growth 218 Lee’s phenomenon in the Silver Lake cisco 222 Possible causes of Lee’s phenomenon in the Silver Lake cisco 224 Selection by gear 224 Selection due to dissimilar dis- tribution within the lake of the various elements of the population 224 Selection due to differential mor- tality, correlated with growth rate 225 Other possible causes 225 General growth curves for the Trout Lake, Muskellunge Lake, Silver Lake, and Clear Lake cisco populations 226 Growth in length 226 Growth in weight 230 Comparison of the growth of the Trout Lake, Muskellunge Lake, Silver Lake, and Clear Lake cisco populations with that of cisco populations in other re- gions 232 Range of length in individual age groups; maximum length and weight 234 Page Condition and the relationship between length and weight 237 Length of growing season 249 Relationship between density of popula- tion and rate of growth 253 Age composition of the samples and the relative abundance of year classes 263 Age at maturity and sex ratio 267 Annual increments of growth 271 Variation in the amount of growth in different calendar years 271 Bimodality in the calculated growth for the first year of life 280 Growth compensation 282 Growth relationships in the Trout Lake, Muskellunge Lake, Silver Lake, and Clear Lake cisco populations 286 Physical-chemical factors 287 Density of population 288 Length of the growing season 288 Parasitization 289 Condition 289 Sex ratio 292 General remarks 292 Fishes associated with the cisco 293 Selective action of gill nets 294 Review of previous investigations 294 Selective action of gill nets used in collecting samples of ciscoes from Trout Lake, Muskellunge Lake, Silver Lake, and Clear Lake 298 Trout Lake 298 Muskellunge Lake 300 Silver Lake 302 Clear Lake 304 Relative abundance of age groups in gill net samples 305 General conclusions 306 Summary 306 Bibliography 310 Appendix A. Growth of the cisco in Alle- quash and Tomahawk Lakes 316 1 Bulletin No. 19. Approved for publication, July 31, 1938. 211 212 BULLETIN OF THE BUREAU OF FISHERIES INTRODUCTION In 1927 the Wisconsin Geological and Natural History Survey and the United States Bureau of Fisheries instituted a cooperative study of the life histories of the fishes that inhabit the lakes of the highland district of northeastern Wisconsin. By the arrangements agreed upon, the survey furnished facilities for the conduct of field operations and staff members of the two organizations cooperated in the collecting and investigation of the research materials. The limnological laboratories of the survey at Trout Lake, Wis., served as headquarters for all field operations. During the summer and early autumn of 1927 Dr. Stillman Wright of the Bureau, with the aid of an assistant made collections of fishes from several lakes of the region. In 1928, when it was impossible for Dr. Wright to be in direct supervision of the field work, the collections were made by Dr. Joseph Goldsmith and an assistant. A part of the collections for the years 1927 and 1928 served as a basis for a preliminary report on the growth of the rock bass in Trout and Muskellunge Lakes (Wright, 1929). In 1929 Dr. Wright found it necessary to devote his entire time to limnological investi- gations on Lake Erie and collecting operations were suspended during that year. The collection of data for the investigation of the life histories of the fishes of the region was resumed in 1930. During this season the field work was carried on by Dr. Edward Schneberger of the survey and Dr. Ralph Hile of the Bureau. (In 1931 the collecting of materials was continued under the supervision of these 2 investigators with the aid of 2 assistants.) In 1932, when it was impossible for Dr. Hile to partici- pate in the field operations, the collections were made by Dr. Schneberger, aided by one assistant. Although the materials collected have afforded data on a number of species, the investigations have centered upon the life histories of three forms: The yellow perch, Perea flavescens Mitchill; the rock bass, Ambloplites rupestris Rafinesque; the cisco or lake herring, Leucichthys artedi (Le Sueur). Dr. Schneberger has given his attention to the study of the perch and Dr. Hile to the study of the rock bass and the cisco. The small collections of game fishes collected in the various years formed a part of the material used by Juday and Schneberger (1930, 1933) in their studies on the growth of game fishes in Wisconsin. The ultimate aim of these studies of the fishes of northeastern Wisconsin is a more precise evaluation of the various environmental factors in their relation to the growth and abundance of the different fish forms. Since 1925 the survey has been conducting extensive limnological investigations on the lakes of the region.2 The fish collections that were made for the growth studies have served also as material for study by survey staff members of food habits and parasitic infestation in the various species. The general consideration of the relationship between growth and environment awaits the termination of these several lines of investigation. In the present publication references will be made to relationship between growth and certain environmental factors, but no exhaustive discussion of the question will be undertaken. In the investigation of the ciscoes an attempt was made to select the populations for study from “type” lakes. The selection of the four lakes with which this study ispn^rily concerned was based principally on the amount of fixed CO, present in » For a history of the survey’s operations consult Juday and Birge (1930). Annual summaries of the survey’s activities appear in the Bureau’s reports on Progress in Biological Inquiries (Eich, 1926; Higgins, 192Sa, 1928b, 1930, 1931a, 1931b, 1932, and 1933). AGE AND GROWTH OF THE CISCO 213 the water. This factor was considered to be a fair index of the general productive capacity of the various waters. Size and form of basin were also taken into account in the selection of the lakes. Trout Lake is the second largest and the deepest lake of the region. In comparison with the surrounding bodies its water is relatively hard. Muskellunge is a lake intermediate as to size and as to the amount of fixed C02. Silver Lake is a small lake with intermediate conditions with respect to fixed C02. Clear Lake has an area approximately equal to that of Muskellunge Lake. Its water contains an unusually small amount of fixed C02. Table 1 gives for each lake the details of location, area, depth, and also data relative to the physical and chemical nature of the waters. The two lakes, Allequash and Tomahawk, from which small samples of scales were obtained for growth study, are included in the table. Table 1. — Description of the 6 lakes of which cisco populations were studied [The characteristics, color, pH, conductivity, fixed CO2, and organic content of plankton refer to average surface conditions. These data were furnished by the Wisconsin Geological and Natural History Survey] Lake Location Length in kilo* meters Width in kilo- meters Surface area (hec- tares) Maximum depth (meters) Volume in 1,000 cubic meters Secchi disk trans- parency (meters) Color platinum cobalt scale M a Conductivity in reciprocal meg- ohms Bound CO2 in milligrams per litre Organic matter of plankton in mil- ligrams per litre |. Town 39N, range VI IE 2. 10 1. 83 373 26.5 6.3 0 6.6 17 2. 2 0. 84 Muskellunge Town 41E, range VIIE 3. 30 1. 18 375 19.3 26, 172 9, 884 4.0 4 7.3 40 10.0 1. 16 Silver I Trout: Town 41N, range VIE 1. 69 .64 87.3 19.5 5.5 4 7.5 62 15.0 .85 North basin /Town 41 and 42N, range VI \ and VIIE. / 3. f>0 \ 4.51 2. 10 532 29.0 69,017 5.0 6 7.8 73 18.5 .66 South basin 3. 86 1,051 35.5 149, 020 4. 5 3 7.6 77 18.7 .92 Allequash___ Town 41N, range VIIE 2. 41 1. 41 142 7. 5 1. 5 30 7. 9 67 16. 9 1. 48 Tomahawk Town 38 and 39N, range VI and VIIE. 7.24 3.00 1, 476 22.5 4.2 8 7.6 69 16.7 .78 'All ciscoes collected from Trout Lake were taken in the south basin. Juday and Birge (1930) published a general description of the northeastern highland lake district. Thwaites (1929) described the glacial geology of a part of Vilas County in which all but two (Clear Lake and Tomahawk Lake) of these lakes are located. Chemical, physical, and biological data concerning the various lakes have appeared from time to time in the publications of the limnological laboratory of the survey. From time to time in the following discussion different species of fish will be mentioned by their common names. Because of the considerable local variation in the common names of fishes it is thought well to list the scientific names of the various species mentioned, along with their common names: Bluegill, Helioperca incisor (Cuvier and Valen- ciennes). Burbot, Lota maculosa Le Sueur. Ciscoes or chubs, Leucichtliys spp. Cisco, lake herring, Leucichtliys artedi (Le Sueur). Lake trout, Cristivomer namaycush (Walbaum). Largemouth black bass, Aplites salmoides (La- c6pede). Perch, yellow perch, Perea flavescens Mitchill. Pike-perch, Stizostedion vilreum (Mitchill). Rock bass, Ambloplites rupeslris Rafinesque. Sauger, Stizostedion canadense (Smith). Smallmouth black bass, Micropterus dolomieu Lac6p£de. Sucker, Catostomus commersonnii (Lacepede). Whitefish. Coregonus clupeaformis (Mitchill). 214 BULLETIN OF THE BUREAU OF FISHERIES ACKNOWLEDGMENTS I wish to acknowledge the assistance and cooperation of the members of the staff of the Wisconsin Geological and Natural History Survey. Dr. Chancey Juday offered much valuable advice, particularly in the selection of lakes for study, and sup- plied needed limnological information. Dr. Edward Schneberger was completely in charge of the 1932 collecting operations. Dr. John Van Oosten, in charge of the Great Lakes fishery investigations of the Bureau of Fisheries, gave much valuable advice (both as to field procedure and methods of analysis of data) during the execution of this work, and gave his assist- ance freely in the interpretation of difficult scales. Dr. Van Oosten also made a critical examination of the manuscript. Dr. Carl L. Hubbs, curator of fishes in the Museum of Zoology of the University of Michigan, read the original manuscript and offered valuable suggestions for its improvement. Harry C. Carver, professor of mathematics and insurance in the University of Michigan, read the section of the manuscript on the relationship between length and weight. Dr. Peter Okkelberg, secretary of the Graduate School of the University of Michigan, checked the accuracy of my translation of portions of Huitfeldt-Kaas’ (1917) publication. The reference to Olofsson’s (1932) paper is based on an abstract pre- pared by ft. W. Eschmeyer of the Michigan Institute for Fisheries Research. The determination of age and the calculation of growth of the 1928 collection of the Trout Lake cisco are entirely the work of Dr. Stillman Wright, formerly of the Bureau of Fisheries. MATERIALS The investigation of age and growth has been based upon the determination of ages for 3,882 specimens and the calculation of growth for 3,694 specimens. Speci- mens other than those whose ages were determined have been used incidentally for other purposes such as the study of vertical distribution of the cisco and the density of population. Koelz (1931) described material from each of the six lakes whose popu- lations are considered in this study. According to Koelz 3 subspecies of Leucichthys artedi are represented in these 6 populations. Clear Lake is the type locality for L. artedi clarensis Koelz. The same form is accredited to Muskellunge, Silver, and Allequash Lakes. Koelz described the Trout Lake cisco as L. artedi clemensi Koelz, and the form occurring in Tomahawk Lake as L. artedi wagneri Koelz. However, Hile (MS.) in a paper now in press has demonstrated that the ciscoes of these lakes at present are not susceptible of division into subspecies, but should be considered merely as populations of the species, L. artedi. Table 2 lists the locality and year of capture of the ciscoes used in the life-history studies. The 1931 material includes 658 preserved specimens collected for a mor- phometric study to appear in a later publication, while all the 1932 Muskellunge Lake specimens and 95 of the 1932 Clear Lake specimens were preserved. AGE AND GROWTH OF THE CISCO 215 Table 2. — Year of collection and number of specimens in the collections used in the study of age and growth of the cisco Lake Year of collection Total number 1928 1930 1931 1932 Trout 182 608 610 1,300 Muskellunge _ 281 245 016 188 1,330 135 118 378 631 249 191 440 47 66 113 68 68 Total 718 937 1, 853 379 3, 882 The relatively limited data from the small collections from Allequash and Tomahawk Lakes are not considered in the general discussions of this paper, but are treated in a special section (appendix A). METHODS GEAR USED IN COLLECTING With the exception of a small number of 0-group individuals picked up on the beaches of Trout Lake all samples used in this investigation were collected by means of gill nets. The 1928 collections were made in nets of the following sizes of mesh: 3 1%, lji, 1%, 2, 2%, and 3 inches, and in Trout Lake a 1% by 8-inch trammel. The trammel used in the capture of Trout Lake ciscoes would operate as a 1%-inch gill net. In this year no net records were kept for individual specimens. It is unlikely, however, that in any of the three populations (Trout Lake, Muskellunge Lake, Silver Lake) ciscoes were taken either in the 2%-inch- or 3-inch-mesh nets. The nets used in 1930 had also been used in 1928. However, the discarding of worn-out nets at the close of the 1928 season was so extensive that the only sizes of mesh available in 1930 for the capture of ciscoes in the three lakes named in the preceding paragraph were l}{ and 2 inches. These same nets were used in the early part of the 1931 collecting season. The depth of these nets varied from 3 to 4 % feet. The gill nets used in 1930 and in the early part of the 1931 season were in a sorry state of repair. Not only were many of the meshes broken, but holes of considerable size were of frequent occurrence. The making of repairs was not feasible as the thread had deteriorated to the extent that even ordinary handling and use of the nets fre- quently produced new tears in the webbing. The gear was completely replaced on July 22, 1931, after which date the following sizes of mesh were used: 1%, 1 }{, 1%, 2, 2]{, 2%, and 3 inches. Each net was 50 yards long and approximately 6 feet deep; 5-inch leads were used as a precaution against tangling of leads in the webbing. All Clear Lake samples (1931 and 1932) were taken in the new gear, but only a few lifts of the new gear were made in the other lakes. The question of gill net selectivity, particularly with reference to the use of gill net samples for the study of age and growth, is considered in a special section (pp. 294-306). METHODS OF FISHING The new nets purchased in 1931 were fished in gangs containing 1 net of each of the 7 mesh sizes. In sets of the old nets the larger mesh sizes, known to be ineffective on the cisco populations that were being sampled, were occasionally : Throughout this paper, sizes of mesh, unless otherwise indicated, are presented as stretched measure. 216 BULLETIN OF THE BUREAU OF FISHERIES omitted. No attempt was made to arrange the nets in any definite order as to size of mesh. In general it was customary to intersperse the larger mesh sizes with the smaller. All sets were made on the bottom. The nets were lifted and reset once each day. At least once each week they were removed from the water, stretched on a reel, and allowed to dry thoroughly. At intervals of about a month all the nets were treated with copper soap after a method suggested by Harry Hansen, formerly of the Bureau of Fisheries. The 1928 data contain only scattered information as to the depth from which the various samples were taken. During 1930 observations were made regularly as to the general depth and the nature of the bottom in the region in which the nets were set. Later it was found desirable to possess more precise information on the matter of depth. Accordingly in 1931 and 1932 soundings were made to determine the depth at which each particular net was set. Routine records were kept as to the condition of the weather (cloudiness, precipitation, and strength and direction of the wind). Catches from the different sizes of mesh were placed separately in labeled pails. After the return to the field laboratory counts of the number of fish taken in each net were added to the field notes. FIELD DATA RECORDED FOR INDIVIDUAL SPECIMENS Scale samples were taken in the field from all specimens except those preserved for morphometric studies. (The treatment of preserved specimens will be discussed later.) The scale samples were removed whenever possible from the left side of the body dorsal to the lateral line and ventral to the dorsal fin. Scales were stored in standard Bureau of Fisheries scale envelops. On each serially numbered envelop were recorded date, locality, species, length, weight, sex, state of maturity,4 and gear.5 The field numbers included in the day’s catch were cataloged. The standard length (from the tip of the snout to the end of the vertebral col- umn) was measured for each individual fish. Enough measurements of the total length (from the tip of the snout to the line connecting the tips of the extended caudal fin) were made to obtain a reliable figure for the ratio between standard and total length.6 All lengths were measured with a steel tape in a straight line between the points indicated, not along the curvature of the body, and were recorded to the nearest millimeter. Throughout this paper the unit of length can be assumed to be the millimeter. The smaller individuals were weighed on a Chatillon spring platform balance of 500-gram capacity with calibrations at 2-gram intervals. Weights of these fish were estimated to the nearest gram. Weight records above 500 grams were obtained by means of a Chatillon spring platform balance of 5-kilo capacity and with calibrations at 20-gram intervals. Weights of these larger fish were estimated to the nearest 5 grams. Toward the close of the 1931 collecting season the smaller balance devel- oped a fluctuating error that was at no point greater than 2 grams. This balance was replaced by a new instrument at the beginning of the 1932 season. 4 A fish whose sexual condition showed that it would spawn in the coming autumn was listed as mature, whether or not it may have spawned previously. ‘Sex, maturity, and gear were not recorded for the 1928 samples. ‘The ratio, total length in standard length, has the following values: Clear Lake, 0.84; Trout Lake, 0.84; Silver Lake, 0.85; Muskellunge Lake, 0.85; Allequash Lake, 0.84. The ratio was not determined for the Tomahawk Lake cisco. AGE AND GROWTH OF THE CISCO 217 TREATMENT OF PRESERVED SPECIMENS AND RESULTING SHRINKAGE Part of the fish preserved for morphometric study were weighed and measured while fresh. These specimens were provided with individual, serially numbered tin tags. Other specimens to be preserved were divided into small groups which were wrapped in cheesecloth. Within each package was included a label bearing a record of date, locality, and usually of gear. The specimens were held in a 10-percent for- malin solution until the close of the season. Upon removal from the formalin solu- tion the fish were soaked in water about 4 days and then transferred permanently to a 70-percent solution of alcohol. Since the preserved specimens — most of which were not weighed and measured in the field — constituted a part of the material used for a study of growth, it was of some importance to know the extent of the shrinkage brought about by the preserving fluids. Data on this question were obtained from 99 ciscoes from Clear Lake and from 105 ciscoes from Muskellunge Lake, all taken in 1932. These fish were weighed and their lengths measured at the time of capture. The dates of capture for the Clear Lake specimens were July 12 to July 19; for those from Muskellunge Lake, July 28 and July 29. The individuals of both collections were reweighed and remeasured October 10 and October 11, 1932. At this time they had been in alcohol about 5 weeks. The measurements and weighing of both the fresh and pre- served material were made by the same method and by the same individual (Dr. Schneberger). The ratio between preserved length and fresh length was 1.000 : 1.019 for the Muskellunge Lake cisco. This value of the shrinkage factor agrees closely with the figure, 1.016, determined by Van Oosten (1929) for the Lake Huron herring. Since the populations from Silver Lake and Trout Lake do not have greatly dissimilar length ranges (approximately 125-200 millimeters) the shrinkage factor based on the Muskellunge ciscoes has been used in correcting for shrinkage the lengths of the pre- served specimens from these two lakes. In the Clear Lake cisco, where the length range is much greater, the relative amount of shrinkage was found to vary at different lengths. The lengths of pre- served specimens from this lake were corrected according to the following shrinkage factors: Length interval: 350 250-349 150-249 Shrinkage factor 1.005 1.010 1.028 Number of specimens 16 65 18 The shrinkage factors for weight were 1.181 for the Muskellunge Lake ciscoes and 1.144 for the Clear Lake ciscoes. However, the individual shrinkage in weight varied so widely that it was not considered valid to employ a weight shrinkage factor for the purpose of making individual corrections. Only those preserved specimens that were weighed when fresh were used for the study of growth in weight. PREPARATION AND EXAMINATION OF SCALE MATERIAL Scales were soaked in water and cleaned by means of a dissecting needle and a small camePs-hair brush. Three scales from each fish were mounted on a microscope slide in a gelatin-glycerin medium prepared according to a formula presented by Van Oosten (1929). No regenerated scales or scales with abnormal or irregular shape 218 BULLETIN OF THE BUREAU OF FISHERIES were mounted. On the label of each slide were recorded the field number, laboratory number, date and locabty of collection, species, sex, maturity, length, weight, and gear. The scales were studied by means of the projection apparatus described by Van Oosten (1923). The magnifications used were Xl9 for scales of the Clear Lake ciscoes and X40.5 for scales of all other fish. MISCELLANEOUS CONSIDERATIONS It is the policy' throughout this paper to apply toward the solution of each par- ticular question all suitable data available. As a result of this procedure there appear certain discrepancies in the number of specimens listed in the tables pertaining to different phases of the general problem of the life history of the cisco. Since dis- crepancies of this sort, if unexplained, may prove a considerable annoyance to a reader, attention will be called to the causes for at least the most important of them. Table 2 lists the fish upon which determinations of age were made. These specimens, with the exception of the 1932 Muskellunge Lake collections, separated into age groups and year classes, appear in the growth tables 3 to 7. The tables of general growth in length (10 to 14) are based on fewer specimens than tables 3 to 7 since certain indicated age groups were eliminated because of selection by gear. (The 17 0-group Trout Lake ciscoes were also omitted in the computation of the general growth curve.) The tables for growth in weight (15 to 18) are based on fewer specimens than those for growth in length since part of the preserved fish used in the growth study were not weighed before preservation. Of course, the study of the length-weight relationship and of condition was based only on specimens weighed in the field. In the listing of catch per unit effort of gill nets (section on the relationship between density of population and rate of growth) it v/as, of course, valid to include fish that had not been aged. Similarly in the determination of the average length of all fish taken in a particular size of gill net (section on selective action of gill nets) it was also valid to include unaged fish. Further, the grand average length used for comparing the average length of an age group as determined from the combined catch of several different sizes of mesh with the average lengths of the same age group as based on the catch of individual sizes of mesh may include certain fish for which net records were not available. (Net records were lacking for a large part of the preserved fish.) The data of tables 40 and 73 are based in part on catches of Muskellunge Lake ciscoes taken in 1932 for morphometric investigations and included only incidentally in this study. None of the 1928 data is included in the analysis of net catches. The 1928 data also lacked records of sex and maturity and hence could not be included in the sections on sex ratio and age at maturity. Although some of these discrepancies are mentioned in the general discussions, a general statement of their origins was considered advisable. THE SCALE METHOD ASSESSMENT OF AGE AND CALCULATION OF GROWTH The treatment here of the growth data of the cisco is based upon the use of the scale method. Van Oosten (1929) established the validity of the method for this species in his study of the life history of the Lake Huron herring. It is assumed that AGE AND GROWTH OF THE CISCO 219 the method is equally valid for the populations considered in this investigation. As will appear later, the results of the analysis of the growth material presented here seem to justify the assumption. Throughout this paper the individual ages are designated as the number of years of life that have been completed. Although the cisco spawns in late autumn or early winter, the life of the individual is assumed to begin the following spring when hatching takes place. During the first year of its life the individual is a member of the 0 group, during the second year it is a member of the I group, etc. Fish hatched in the same year belong to the same year class, regardless of their age at the time of capture. Two examinations for age determination were made of the scales of each individ- ual specimen. At the second examination either a definite assignment of age was made for the troublesome scales or the slides were marked as unreadable. At the time of the final examination one of the three scales on the slide was measured. This scale was selected on the basis of clearness of markings and symmetry of form. The diameters of the different growth areas were measured along the anteroposterior axis of the scale. For this measurement the ruler was placed in such a position that its edge passed through the focus of the scale and approximately bisected the posterior field (that portion of the scale which is exposed in its natural position on the fish). Scale measurements were made with a tested millimeter ruler and recorded to the nearest millimeter (occasionally the nearest half millimeter). The assessment of age was based on the determination of the number of annuli or lines of discontinuity between growth areas of succeeding years. Two difficulties were encountered in the determination of individual ages. The later annuli of the older individuals (particularly those of the slower-growing populations) were some- times so crowded together as to make their recognition difficult or even impossible. In the latter case the scales were discarded. Accessory annuli (summer checks) were not infrequent, but their characteristically indefinite appearance and their position with respect to the true annuli were such that they were ordinarily easily recognizable. Somewhat less than 5 percent of the total number of scales examined were discarded as unsatisfactory for the determination of age. Accessory checks were found to occur regularly in all the samples of the scales of the Muskellunge Lake cisco. They could be separated from the true annuli quite easily in the growth fields of the earlier years, and consequently they presented no great difficulty in the age and growth analysis of the earlier (1928-31) collections which were composed almost entirely of I-, II-, and Ill-group fish. In 1932, however, IV- group fish were present in abundance for the first time. In this year’s collection the presence of accessory annuli, along with the IV-group’s excessively slow growth during the fourth year of life, made the separation of the III group and the IV group of that year’s collection both difficult and uncertain. Because of these difficulties and because of the abundance of data from the preceding years, the 1932 collections were not used in the computation of a general growth curve for the Muskellunge Lake population. However, certain of the data from the 1932 Muskellunge Lake samples will be used from time to time for other purposes. Van Oosten (1929) demonstrated that after the formation of the first annulus the ratio of the total scale length to body length is so nearly constant in the lake herring that the assumption of the absolute constancy of the ratio affords the most 220 BULLETIN OF THE BUREAU OF FISHERIES satisfactory means for the calculation of growth during the various years of the individual’s existence. The formida used in the calculation of growth was: T —D — J^n-Un Dt where XT= length of fish at time of capture; Dt= total diameter of scale; Xn= length of fish at end of nth year of life; Z>n— diameter of scale within the nth annulus. Tables 3 to 7 present the data for the calculated growth in length for the ciscoes of Trout, Muskellunge, Silver, and Clear Lakes. Since no difference in the growth of the two sexes was apparent in the fish of the first three lakes the data for these lakes as presented represent a combination of males and females. In the data for the Clear Lake cisco the sexes are treated separately. Table 3. — Calculated growth of Trout Lake ciscoes to the end of each year of life for the different year classes at each age; sexes combined [Groups marked with asterisks unreliable because of selective action of gear] Year class Year of capture Age Num- ber of speci- mens Length in milli- meters Year of life 1 2 3 4 5 6 7 8 9 10 11 12 1919 1931 XII i 226 80 104 117 130 148 169 180 190 199 211 218 224 1920 1931 XI 2 205 75 103 119 130 142 152 162 172 180 198 204 1922 1931 IX 2 200 83 112 127 140 154 168 ISO 190 198 1923-.. 1931 VIII 5 192 87 115 134 145 156 169 179 190 1924 f 1928 IV 17 147 85 120 133 143 \ 1931 VII 4 175 80 113 128 141 153 162 174 ( 1928 III 61 143 84 123 137 1925 •J 1930 V 9 167 82 116 135 148 160 l 1931 VI 12 165 79 116 132 145 156 164 f 1928 II 102 134 82 117 1926 \ 1930 IV 99 156 82 118 137 147 l 1931 V 79 156 81 118 135 145 154 ( 1928 *1 2 128 97 1927 l 1930 III 347 149 79 117 139 l 1931 IV 269 148 78 115 135 146 1928 / 1930 *11 36 140 91 127 t 1931 III 173 142 84 120 137 1929 1931 •II 61 136 87 127 1930 / 1930 0 17 66 \ 1931 •I 2 128 106 Table 4. — Calculated growth of Muskellunge Lake ciscoes to the end of each year of life for the different year classes at each age; sexes combined [Groups marked with asterisks unreliable because of selective action of gear] Year class Year of capture Age Number of speci- mens Length in milli- meters Year of life 1 2 3 4 1925 1928 III 10 172 105 142 162 1926 1928 II 252 160 99 147 ( 1928 •I 19 137 97 1927... | 1930 III 14 166 85 132 157 1 1931 IV 2 176 90 143 161 174 1928 / 1930 II 214 162 95 143 \ 1931 III 347 166 96 145 163 1929 / 1930 *1 17 148 103 \ 1931 II 258 149 98 138 1930 1931 *1 9 136 96 AGE AND GROWTH OF THE CISCO 221 Table 5. — Calculated growth of Silver Lake ciscoes to the end of each year of life for the different year classes at each age; sexes combined [Groups marked with asterisks unreliable because of selective action of gear] Year class Year of capture Ago Number of speci- mens Length in milli- meters Year of life 1 2 3 4 5 6 7 1923 1928 V 10 182 76 115 142 161 175 ( 1928 IV 69 174 79 125 150 166 1924 .. \ 1930 VI 3 197 72 112 134 161 177 192 1 1931 VII 1 201 77 111 135 158 172 186 196 [ 1928 III 46 165 73 127 157 1925 - •J 1930 V 25 193 74 120 145 165 185 1 1931 VI 21 194 78 119 142 158 177 189 1928 II 9 157 90 140 1926.. - { 1930 IV 58 183 78 126 152 173 ( 1931 V 108 188 80 126 150 168 182 ( 1928 *1 1 133 95 1927 . \ 1930 III 25 181 77 134 167 l 1931 IV 102 183 80 132 159 176 / 1930 II 7 173 99 151 1928 \ 1931 III 61 177 83 137 166 1929 1931 II 19 171 104 151 1930 1931 *\ 66 141 105 1 Table 6. — Calculated growth of Clear Lake ciscoes (males) to the end of each year of life for the dif- ferent year classes at each age Year class Year of capture Age Number of speci- mens Length in milli- meters Year of life 1 2 3 4 5 6 7 8 9 f 1931 VIII 3 338 118 203 248 279 297 314 326 333 \ 1932 IX 1 355 114 194 254 290 310 323 333 345 354 1924 f 1931 Vll 8 331 110 198 249 279 298 312 324 \ 1932 VIII 7 345 1C8 195 250 279 302 318 331 342 / 1931 VI 11 326 108 192 248 282 303 316 1925 \ 1932 VII 9 336 102 188 248 280 302 318 331 1926 f 1931 V 4 325 110 189 247 290 312 { 1932 VI 2 320 98 183 251 283 299 315 1927 . / 1931 IV 10 317 104 190 258 292 \ 1932 V 7 320 98 181 249 288 312 1928 / 1931 III 26 285 104 188 255 \ 1932 IV ii 314 105 191 260 302 1929 / 1931 ir 34 253 108 200 \ 1932 in 23 290 110 198 268 1930 / 1931 i 20 178 119 \ 1932 ii 20 254 113 218 1931 1932 i 20 176 122 Table 7. — Calculated growth of Clear Lake ciscoes (females) to the end of each year of life for the different year classes at each age Year class 1921. 1922. 1923. 1924. 1925. 1926. 1927. 1928. 1929. 1930. 1931. Year of capture Age Num- ber of specimens Length in milli- meters 1 2 1931 X 1932 XI 1931 IX 1932 X 1931 VIII 1932 IX 1931 VII 1932 VIII 1931 VI 1932 VII 1931 V 1932 VI 1931 IV 1932 V 1931 III 1932 IV 1931 II 1932 III 1931 I 1932 II 1932 I 2 376 1 378 1 351 1 379 13 354 2 368 21 344 8 356 12 339 12 362 2 338 1 320 4 334 7 342 25 302 14 329 35 264 21 308 18 186 13 262 11 176 111 211 110 214 114 206 1)3 235 112 199 107 200 108 190 105 194 105 183 108 190 108 188 98 183 103 191 98 177 108 199 103 195 108 207 108 205 119 114 123 219 Year of life 3 4 5 6 7 8 9 10 11 264 265 262 278 253 259 250 246 241 252 258 251 273 251 269 268 288 294 283 301 286 291 282 278 283 294 302 283 306 297 308 313 301 312 308 322 305 304 311 322 326 299 328 332 316 326 323 333 322 324 327 339 341 342 330 340 336 346 335 339 353 351 339 353 347 354 362 361 346 370 370 369 376 377 — 363 351 357 315 329 312 282 222 BULLETIN OF THE BUREAU OF FISHERIES From an examination of the tables for Trout, Muskellunge, and Clear Lakes it will be noticed that “Lee’s phenomenon” of “apparent decrease in the calculated growth” as it is determined from successively older groups of individuals, wnile pos- sibly present, is not a source of any great discrepancies, especially where large numbers of individuals are involved. On the whole the calculated growths based on different age groups in the same year’s collection or upon samples of the same year class taken at different ages, agree satisfactorily for corresponding years of life, with the exception of the large calculated growths obtained for the younger age groups (marked with asterisks in the tables) in the Trout Lake and Muskellunge Lake collections. These large growths can be explained as the result of the selective action of the gill nets used for their capture. (See section on gill net selectivity.) The small but fairly consistent discrepancies that appear in the corresponding calculated growths of older year groups of the Trout Lake fish may be partly the result of slight changes in the body scale ratio. It is probable that selective action of gear plays a role here also. At any rate the discrepancies are of such small magnitude that they affect but little any conclusions that might be drawn concerning the growth of the population as a whole. LEE’S PHENOMENON IN THE SILVER LAKE CISCO In the Silver Lake data (table 5) there appear much more pronounced disagree- ments among the corresponding calculated growths of different age groups than are present in the other three populations. Not only are these discrepancies relatively large, but the comparison of any two age groups shows that in general they tend to be cumulative with increased age. These discrepancies appear in the comparisons of calculated growths based on the same year class but taken in different seasons as well as in calculated growths based on the different age groups of a single year’s collection. To illustrate these two points there may be examined (table 5) first the calculated growths of the III, V, and VI groups of the 1925 year class and second the calculated growths of the III, IV, V, and VI groups of the 1931 collection. With certain ex- ceptions in the first year of life it will be noticed that each successively older age group tends to give smaller values for the calculated lengths for corresponding years of life. Thus it may be seen that the observed discrepancies in the calculated growths depend on the differences in the age of the groups studied. It will be noticed further that the apparent change of growth rate as it appears in the fish older than the II group differs from Lee’s phenomenon as it ordinarily occurs, in that it is the calculated growths of the later rather than the earlier years of life that are affected most. The calculated lengths of the Silver Lake ciscoes at the end of the first year of life require some special consideration. If size of sample is taken into account, the cal- culated growths for the first year of life show fairly good agreement for all age groups older than II. The calculated growths for the first year of life as based on the II group are somewhat less than those based on the I group, but those of both groups are considerably greater than the growths based on the older-age groups. The high cal- culated growths for the first year of life as based on the I group can be explained in part by the selective action of the gill nets used in collecting the samples, for there is reason to believe that even the smallest mesh of these nets may have taken only the larger individuals of this group. Such an explanation, however, does not seem to hold for the fish of the II group, since on the basis of comparisons with samples from other lakes the range of size of the Silver Lake II group seems to be such that very little AGE AND GROWTH OF THE CISCO 223 selection by gear could have taken place. Some other explanation must be sought, therefore, for the high growth values of this age group. Since disagreements in the calculated growth data of fish are generally attributed in part to changes with age in the growth relation between body and scale, it was deemed advisable to study this relationship in the Silver Lake cisco. The 1931 collec- tion was selected as best suited for this purpose. As was mentioned previously (p. 216) the scale samples used in this entire investigation were all taken from the same region of the body. Selections of scales for mounting anti for measuring were made on the basis of distinctness of the annuli and symmetry of form rather than on size. As a consequence of this procedure the relationship between the length of body and the diameters of the scales measured should be comparable in the different age groups of a single sample. As may be seen in table 8 the average ratio between the body length and the magnified diameters of the scales does not vary greatly from one age group to another. The generally made assumption of a linear relationship between scale size and body size tends to the determination for the Silver Lake cisco of the equation:7 L=10.7 mm + 1.078 D, where £=length of body in millimeters, and Zl=the (magnified) diameter of the scale in millimeters. Table 8. — The body-scale ratio for the different age groups of the Silver Lake cisco collected in 1931, and the average diameters of scales as measured at the magnification X 40.5 [Dt is total diameter. The diameters within the different growth areas of the scales are indicated by the subscript figure] Age Number of speci- mens Body- scale ratio Dt D! Da Dj D( Dj D« D7 VII... 1 21 108 102 01 19 66 1. 16 1. 16 1. 13 1.13 1. 15 1. 18 1. 17 174.0 168.3 167.9 161.9 155.7 145.8 121. 3 67.0 67.5 70.9 71.0 72.6 88.0 90. 2 96.0 104.0 112.6 117.5 119.5 128.7 117.0 123.7 133.8 140.8 145.7 137.0 139.5 150. 5 155.7 149.0 153.6 162.6 161.0 164.5 170.0 VI V IV III II I Table 9 gives a comparison of the calculated growths of the 1931 Silver Lake collection first as they were determined on the assumption of a constant body scale ratio at all ages and second as they appear after correction for the changing body- scale relationship indicated by the above equation. It will be noticed that while the "correction” produced changes in the actual values of the individual average cal- culated lengths, it did little toward the elimination of the discrepancies between the calculated growths of the different age groups. For example, it may be seen that the maximum discrepancy in the calculated growth for the first year of life (between the calculations for the I group and the VI group) was reduced from 28 millimeters in the uncorrected to 24 millimeters in the corrected data — an improve- ment in agreement of only 4 millimeters. Similarly, the improvements in the agree- ment in the calculated lengths for the later years of life are unimportant when they ; It is realized that a more precise evaluation of the body-scale relationship might have been obtained through the use of “key ” scales selected from exactly the same location in each fish. However, the complete failure of the “correction " equation to eliminate the discrepancies in the calculated growths of the Silver Lake cisco together with the fact that no corrections are needed in the other three populations made any detailed study of the body-scale relationship in these cisco populations unnecessary 224 BULLETIN OF THE BUREAU OF FISHERIES are considered in relation to the total discrepancy. Thus it may be seen that the observed discrepancies in the calculated growth of the Silver Lake cisco did not origi- nate in a change of the body-scale relationships with increasing age. Table 9. — Calculated growth of the Silver Lake cisco, collection of 1931 Age Length Uncorrected calculated lengths Corrected calculated lengths Li La La U U U Lj Li La L3 Li l5 Le Lj VII.. 201 194 188 183 177 171 141 77 78 80 80 83 104 105 111 119 126 132 137 151 135 142 150 159 166 158 168 168 176 172 177 182 186 189 196 84 84 86 86 89 108 108 116 123 129 135 139 152 138 145 152 160 167 160 160 169 176 173 178 182 187 189 196 VI V IV Ill II.... I Further light is thrown on the question by the examination of the actual average measurements of the growth areas of the scales upon which the Silver Lake growth calculations were based (table 8). The comparison of the diameter measurements of table 8 with the corresponding calculated lengths of table 9 shows clearly that the disagreements in the calculated growths of the different age groups appear also in the actual scale measurements upon which the growth calculations were made. Van Oosten (1929) also observed that Lee’s phenomenon was to be found in actual scale measurements as well as in calculated growths. The “apparent change in growth rate” in the Silver Lake cisco must, then, be considered a real change rather than an apparent change due to changing body-scale relationships. Since, as was pointed out previously, the phenomenon appears con- sistently both in comparisons between different age groups of the same year class and in comparisons between different age groups of different year classes taken in the same calendar year, it cannot be explained either on the basis of inherent differ- ences in the capacity for growth in different year classes, or on the basis of varying environmental conditions affecting growth in different years. It must rather be considered the result of some selectional factor correlated with the length of time intervening between the years for which the computations and scale measurements were made and the year when the fish was captured. Attention will be called briefly to possible sources of selection. POSSIBLE CAUSES OF LEE’S PHENOMENON IN THE SILVER LAKE CISCO SELECTION BY GEAR A more detailed consideration of the question of the selective action of the gear used will be presented later in this paper. (See section on selective action of gill nets.) It may be stated here, however, that this factor most probably does not operate on the Silver Lake ciscoes except in the I group and that therefore it cannot explain Lee’s phenomenon in these fish. SELECTION DUE TO DISSIMILAR DISTRIBUTION WITHIN THE LAKE OF THE VARIOUS ELEMENTS OF THE POPULATION Since the gill nets used in collecting the cisco samples for these studies were set directly on the bottom they fished only the lower few feet of water. If it is assumed that the larger individuals of the population tend to lead a more pelagic existence AGE AND GROWTH OF THE CISCO 225 than the smaller ones, then the more rapidly growing fish within an age group would be less subject to capture by nets set on the bottom than the slower growing individuals within the same group. Further, the effect of such segregation would be greater as age increases. The above-mentioned assumption would serve well to explain decreas- ing growth as it has been observed in the Silver Lake cisco, but data for its verification arc not available. The chief objection to such an explanation lies in the fact that it is the smaller fish not the larger that would normally be expected to lead the more pelagic existence. SELECTION DUE TO DIFFERENTIAL MORTALITY, CORRELATED WITH GROWTH RATE It has been observed repeatedly that poikilothermous organisms tend to grow more slowly and reach a greater age in northern latitudes. The phenomenon appears to depend on the relation of temperature to the rate of metabolism. It is not wholly unreasonable to assume that the individual rate of metabolism might affect the indi- vidual growth rate and the individual length of life. Were mortality greater among individuals of more rapid growth, those individuals that survived longest would be those that had actually grown most slowly. The early growth as calculated from these slow growing survivors of the older age groups would naturally be small. Fur- ther, the effect on calculated lengths of this selection through differential mortality would tend to be increased with greater age, and thus discrepancies of the type observed in the Silver Lake data would be explained. The relation between individual growth rate and individual length of life has been studied experimentally by several investigators. Osborne et al. (1917) found that a temporary preliminary stunting delayed maturity and extended the life span of rats. Titcomb et al. (1928) and McCay et al. (1931) found that trout that did not grow lived longer than those which showed growth on a similar diet. Zabinski (1929) by effecting a retardation of growth was able to prolong life in the black beetle and the roach. McCay (1933) presented a brief discussion of the general problem of the relationship of rate of growth to longevity. OTHER POSSIBLE CAUSES It is recognized that the three possible causes of Lee’s phenomenon in the Silver Lake cisco discussed in the preceding paragraphs are by no means the only possible explanations of the observed discrepancies in calculated growth as based on fish of different age. These three suggested explanations were emphasized because they appear to be the most plausible in the face of the available data. Other explanations should be mentioned briefly along with the reason for their rejection. 1. Growth may be better in some calendar years than in others. Tliis explana- tion cannot be accepted as the growth discrepancies were found to occur between members of the same year class, captured at different ages. 2. Portions of the scale may be resorbed after being laid down. The examination of the scales offered no evidence for any kind of resorption. 3. The scale fields may contract after being laid down. The nature of the structure of scales makes this explanation totally unacceptable. (See Van Oosten, 1929.) 4. More than one annulus per year may be formed. The clarity of scale markings on the Silver Lake cisco was superior to that of the Trout Lake and Muskellunge Lake ciscoes; yet the measurements of the scales of the last two populations gave 226 BULLETIN OF THE BUREAU OF FISHERIES quite consistent results in the calculation of growth. Accessory annuli were rare in the Silver Lake cisco scales. With the elimination of the more questionable scales it appears unreasonable to assume that the number of errors in the determination of ages was sufficient to account for the observed large discrepancies in the calculated growths. Certainly accessory annuli cannot account for the decrease with age in the diameter of the first growth field. 5. The individuals of the population may have segregated themselves on the basis of maturity with the result that the smaller, immature members of each age group may have been absent from the samples. This explanation lacks plausibility in the face of the strong evidence presented later (p. 268) that some if not most fish mature as members of the I group, while all II-group fish certainly can be considered mature. It should be mentioned further that segregation on basis of maturity would be expected to be most pronounced at the spawning time, not in midsummer when the cisco is confined in Silver Lake to a narrow stratum of water in the lower part of the thermocline and the upper part of the hypolimnion. Regardless of the explanation or explanations accepted as to the cause of Lee’s phenomenon in the Silver Lake cisco, it appears that the observed discrepancies do not affect the validity of the method used in this paper for the calculation of growth from scale measurements. Nevertheless, there remains the question as to why the phenomenon should be peculiar to the Silver Lake population. GENERAL GROWTH CURVES FOR THE TROUT LAKE, MUSKELLUNGE LAKE, SILVER LAKE, AND CLEAR LAKE CISCO POPULATIONS GROWTH IN LENGTH For each population considered here all collections were combined 8 to give a gen- eral growth curve (fig. 1). The indicated lengths at the end of the various years of life are in general the grand average of all calculated lengths for these years. In the Trout Lake and Clear Lake ciscoes, however, the irregularities in the later years, that result largely from the small samples, were smoothed in accordance with the observed data on the annual calculated growth increments. The Silver Lake growth data raise a question as to the value and significance of any “general” growth curve in this population. The presentation of a general growth curve involves the tacit assumption that this curve can be taken to represent the course of growth of an individual that is typical of the population as a whole. However, the Silver Lake growth curve was derived from the combination of several age groups whose growth histories were fundamentally different. It appears char- acteristic of the Silver Lake cisco that individual growth history and individual life span are definitely correlated. Consequently the combining of all age groups to obtain a general growth curve involves the lumping together of a mass of heterogeneous growth material. The typical individual that such a curve is purported to represent is probably nonexistent. Nor can the growth curve obtained from a single age group be considered homogeneous. A sample of Ill-group fish, for example, can be expected to contain fish that would normally have died within the year of collection, and along with them others that would have survived to their fifth, sixth, or seventh year, or even longer. All these different groups within a single age group would have dif- ferent types of growth. 8 Certain age groups, however, were eliminated as unreliable because of selection by gear. These were: Trout Lake, all I groups and the II groups of 1930 and 1931; Muskellunge Lake, all I groups; Silver Lake, all I groups. For Clear Lake no groups needed to be eliminated. AGE AND GROWTH OF THE CISCO 227 YEARS OF LIFE Figure 1.— General growth curves showing the average calculated standard length of the cisco (in millimeters) at the end of each year of life. Trout Lake, ; Muskellunge Lake, — . — ; Silver Lake, — ... — ; Clear Lake (males), ; Clear Lake (females), . Tables 10 to 14 give for each population the calculated length at the end of each year of life, and the increase, both absolute and percentile, during each year of life. These data are presented graphically in figures 1 to 3. Table 10. — Trout Lake cisco — calculated length in millimeters at end of year , increase in length, and percentage increase in length for each year of life [All collections combined. Sexes combined] Number of specimens Year of life Length in milli- meters AL 100AL L Number of specimens Year of life Length in milli- meters AL 100 A L L 1 12 225 6 2.7 26... 6 168 10 6. 3 3 11 219 6 2.8 114 5 158 10 6.8 3 10 213 8. 1 499 4 148 11 8. 0 5 9 197 8 4.2 1,080 . 3 137 20 17. 1 10 8 189 10 5. 6 1,132 2 117 36 44.4 14 7 179 u 6.5 1,182 .. 1 81 81 228 BULLETIN OE THE BUREAU OF FISHERIES s 2 z i i- S o a o years ok life Figure 2.— Growth curves showing the increment of growth (in millimeters) in length of the cisco for each year of life. Trout Lake, ; Muskellunge Lake, — . — ; Silver Lake, — ... — ; Clear Lake (males), ; Clear Lake (females), — . Table 11. — Muskellunge Lake cisco — calculated length in millimeters at end of year, increase in length, and -percentage increase in length for each year of life [All collections combined. Sexes combined] Number of specimens Y ear of life Length in milli- meters AL 100AL L Number of specimens Year of life Length in milli- meters AL 100AL L 2 4 174 13 8. 1 1,097 2 143 47 49.0 373 ... 3 161 18 12.6 1,097 1 96 96 Table 12. — Silver Lake cisco — calculated length in millimeters at end of year, increase in length, and percentage increase in length for each year of life [All collections combined. Sexes combined] Number of specimens Year of life Length in milli- meters AL 100AL L Number of specimens Year of life Length in milli- meters AL 1Q0AL L 1 7 196 7 3.7 529 3 155 26 20. 2 25 6 189 8 4.4 564 2 129 49 61.2 168 5 181 12 7. 1 564 . 1 80 80 397 4 169 14 9.0 AGE AND GROWTH OF THE CISCO 229 UJ H < tt X £ O cr V UJ O < »- z UJ O a UJ CL YEARS OF LIFE Figure 3.— General growth curves showing the percentage increase in length of the cisco in each year of life. Trout Lake, ; Muskellunge Lake, — . — ; Silver Lake, — ... — ; Clear Lake (males), ; Clear Lake (females), . Table 13. — Clear Lake cisco — males — calculated length in millimeters at end of year , increase in lengthy and percentage increase in length for each year of life [All collections combined] Number of specimens Year of life Length in milli- meters AL 100AL L Number of specimens Year of life Length in milli- meters AL 100AL L 1 9 354 9 2.6 73 4 287 32 12.5 11 8 345 10 3.4 122 3 255 59 30. 1 28 7 335 12 3.7 176 2 196 86 78.2 41.. 6 323 15 4.9 216 1 110 110 52 5 308 21 7.3 Table 14. — Clear Lake cisco — females — calculated length in millimeters at end of yearf increase in length, and percentage increase in length for each year of life [All collections combined] Number of specimens Year of life Length in milli- meters AL 100AL L Number of specimens Year of life Length in milli- meters AL 100AL L 1.. 11 376 7 1.9 83.... 5 312 21 7.2 4 10 369 8 2.2 101 4 291 31 11.9 7 9 361 10 2.8 147.... 3 260 61 30.7 28 8 351 11 3.2 195— 2 199 90 82. 6 61 7 340 14 4.3 224 1 109 109 74 6 326 14 4.5 230 BULLETIN OF THE BUREAU OF FISHERIES In the amount of growth the Clear Lake cisco stands far above the other three populations. The Trout Lake cisco shows the least amount of growth. The Silver Lake fish occupy a position intermediate between the Trout Lake and Muskellunge Lake populations. Examination of the curves showing annual increments and annual percentage increase throws further light on the nature of the growth of the four popula- tions. If the percentage increase is considered to represent the rate of growth, it will be seen that beyond the fourth year of life the growth rates of the four populations show only small differences and that beyond the fifth year the Trout Lake fish, although the smallest in actual size, have consistently the highest rate of growth. The annual growth increments of the populations show no great differences beyond the fifth year of life. These facts indicate that the differences in the size of the adult fish in the four populations depend upon the nature of growth during the early years of fife. There is a general convergence of growth rates at the fourth or fifth year of fife, but at that time the characteristic nature of the size composition for each popu- lation is well established. GROWTH IN WEIGHT Tables 15 to 18 present the average weight in grams for each age group in each year’s collection, together with summaries for the different years’ collections combined, of average weight of each age group, the yearly increments based on the average weights, and the yearly percentage increase in weight. In the consideration of aver- age growth in weight for the different populations the effect of gear selection should be kept in mind for the same age groups (marked with asterisks) that were eliminated in the calculation of the general growth in length (see footnote p. 226). It was found possible to combine the sexes in the Trout Lake and Muskellunge Lake samples, but not in the Clear Lake and Silver Lake collections. The females of the Clear Lake population grow in weight much more rapidly than do the males, while the males from Silver Lake tend toward a slightly better growth than do the females. Table 15. — Trout Lake cisco — Average weight in grams of each age group in each year's collection, and grand average weight for each age group, 1928, 1930, and 1931 collections combined, together with the annual increment and the percentage increase [Sexes combined. The average weights for the I-group fish are probably too high as the result of gear selection; selection probably affected the II-group weight values only slightly. Number of specimens in parentheses. Groups marked with asterisks un- reliable because of selective action of gear] Age Year of capture Grand AW 100 AW 1928 1930 1931 average w XII.. 172 (1) 119 (2) 111 (2) 60 (2) 45 (23) 39 (139) 36 (129) *33 (60) *27 (2) 172 (1) 119 (2) 111 (2) 50 (2) XI VII VI 3 6.4 V 64 (9) 44 (99) 37 (347) *32 (36) 47 (32) 41 (255) 37 (537) 31 (102) •26 (4) 6 14.6 IV 42 (17) 37 (61) 31 (102) *26 (2) 4 10. 8 Ill 6 19.4 II I.. AGE AND GROWTH OF THE CISCO 231 0 I 23456 769 10 II 12 Figure 4. — General growth curves showing the average weight (in grains) of the different age groups of cisco at time of capture in summer. Trout Lake, ; Muskellunge Lake, — . — ; Clear Lake, — ... — ; Clear Lake (males), ; Clear Lake (females), . Table 16. — Muskellunge Lake cisco — Average iveight in grams of each age group in each year's co- lection, and grand average weight for each age group, 1928, 1980, and 1931 collections combined, together with the annual increment and the percentage increase [Sexes combined. The average weights for the I-group fish are probably too high as the result of gear selection. Number of spec- imens in parentheses. Groups marked with asterisks unreliable because of selective action of gear] Year of capture Grand AW 100AW 1928 1930 1931 average w 60 (2) 51 (266) 42 (229) •31 (9) 60 (2) 51 (290) 46 (695) •35 (45) 9 17.6 55 (10) 60 (252) •33 (19) 49 (14) 45 (214) •39 (17) 5 10.9 Table 17. — Silver Lake cisco — Average weight in grams of each age group in each year's collection, and grand average weight for each age group, for the 1980 and the 1931 collections combined, together with the annual increment and the percentage increase [ Sexes separately. The average weights for the I-group fish are probably too high as the result of gear selection. Number of speci- mens in parentheses. Groups marked with asterisks unreliable because of selective action of gear] Age Year of capture Grand average AW 100AW 1928 1930 1931 w Male and female Male Female Male Female Male Female Male Female Male Female VI 106 (3) 104 (8) 104 (11) 11 11.6 V 81 (10) 98 (10) 99 (15) 95 (12) 92 (22) 96 (22) 95 (37) 10 11 11.6 13. 1 IV 68 (69) 86 (23) 87 (35) 86 (13) 80 (22) 86 (36) 84 (57) 4 11 4.9 15. 1 Ill 59 (46) 86 (13) 74 (12) 76 (10) 73 (13) 82 (23) 73 (25) 16 12 24.2 19.7 II 50 (9) 68 (3) 64 (4) 64 (4) 60 (8) 66 (7) 61 (12) I •28 (1) •32 (24) •35 (25) •32 (24) •35 (25) 232 BULLETIN OF THE BUREAU OF FISHERIES Table 18. — Clear Lake cisco — Average weight in grams of each age group in each year’s collection, and grand average weight for each age group, for the 1931 and the 1932 collections combined, together with the annual increment and the percentage increase [Sexes separately. Number of specimens in parentheses] Age Year of capture Grand average AW 100AW 1931 1932 W Male Female Male Female Male Female Male Female Male Female XI .. . 1, 190 (1) 1, 190 (1) 128 12. 1 X 1,035 (1) 1, 090 (1) 1, 062 (2) -23 -2.1 IX 915 (1) 880 (1) 1, 170 (2) 880 (1) 1,085 (3) 84 135 10.6 14.2 VIII 760 (3) 937 (11) 811 (7) 969 (8) 796 (10) 950 (19) 67 33 9.2 3.6 VII 718 (6) 859 (17) 737 (9) 1,000 (12) 729 (15) 917 (29) 56 76 8.3 9.0 VI 687 (7) 844 (8) 625 (2) 820 (1) 673 (9) 841 (9) 26 71 4.0 9.2 V 710 (2) 740 (1) 629 (7) 774 (7) 647 (9) 770 (8) 69 65 11.9 9.2 IV 603 (5) 756 (7) 567 (11) 680 (14) 578 (16) 705 (21) 160 175 38.3 33.0 III 413 (17) 525 (19) 422 (23) 634 (21) 418 (40) 530 (40) 151 230 56.6 77.7 II 272 (28) 303 (27) 259 (20) 295 (13) 267 (48) 306 (40) 196 220 276. 1 275. 0 1 75 (18) 88 (15) 68 (20) 70 (11) 71 (38) 80 (26) The weight data are based on fewer specimens than are the length data since most of the fish that were preserved for morphometric study were not weighed in the field and therefore were not used in these studies. Examination of the data bearing on growth in weight reveals how sharply the growth of the Clear Lake cisco differs from that in the other three populations (fig. 4). The differences that exist among the populations depend upon differences in the length-weight relationships in the various populations as well as upon differences in the amount of growth in length. The questions of form and changes in form, and of the relationship between length and weight are considered in a separate section (pp. 237-247.) COMPARISON OF THE GROWTH OF THE TROUT LAKE, MUSKELLUNGE LAKE, SILVER LAKE, AND CLEAR LAKE CISCO POPULATIONS WITH THAT OF CISCO POPULATIONS IN OTHER REGIONS Data on the growth of the cisco or lake herring have been published for Lake Erie by Clemens (1922), for Oconomowoc and Pine Lakes in southern Wisconsin by Cahn (1927), for Lake Huron by Van Oosten (1929), for Lake Ontario by Pritchard (1931), for the Indian Village Lakes in northern Indiana by Hile (1931), and for Hudson Bay by Dymond (1933). Tables 19 and 20 present a comparison of the growth of these populations with that of the populations of the present investigation. The lengths (table 19) are expressed in millimeters, the weights (table 20) in grams, and the ages in both tables as the number of years of life completed. The methods of presentation of the various authors were altered to conform to this uniform method of expressing length, weight, and age. Dymond did not include data on growth in weight of the Hudson Bay cisco. The largest fish in his collection was a female, 383 millimeters long and weighed 793 grams. AGE AND GROWTH OF THE CISCO 233 Table 19. — Comparison of growth in length, expressed in millimeters, of 11 cisco populations [The data for the 6 lakes at the right were adapted from various authors as indicated in the text] Age Trout ‘ Muskel- lunge 1 Silver1 Clear 1 Huron3 Ontario 3 Erie3 Indian Village3 Ocono- mowoc 4 Pine 4 Hudson Bay 3 Male Fe- male XII 225 XI 219 376 X . 213 369 345 386 345 IX 197 354 361 358 374 338 VIII 189 345 351 292 303 285 362 330 336 VII 179 196 335 340 274 297 275 374 336 314 324 VI 1C8 189 323 326 258 270 255 336 315 283 308 V 158 181 308 312 244 253 235 342 282 246 290 IV 148 174 169 287 291 235 233 215 316 223 195 284 Ill 137 161 155 255 260 218 226 190 301 174 162 214 II 117 143 129 196 199 185 196 160 260 135 125 184 I 81 90 80 110 109 127 129 125 1 Calculated lengths at the end of the year of life indicated. 1 A combination of calculated lengths and actual lengths measured in November. 3 Actual measured lengths at time of capture during the growing season. The ages of the fish are really greater than indicated here. * Actual measured lengths in midwinter after the completion of the growing season. Table 20. — Comparison of growth in weight, expressed in grams, of 10 cisco populations [The data for the 6 lakes at the right were adapted from various authors as indicated in the text] Age Trout Muskel- lunge Silver Clear Huron Ontario Erie Indian Village Ocono- mowoc Pine Male Fe- male Male Fe- male XII 172 XI 119 1, 190 X 1,062 668 696 436 IX 880 1,085 751 623 406 VIII 111 796 950 454 354 611 343 VII 729 917 488 269 810 527 262 VI 50 106 673 841 221 303 213 782 445 190 V 47 96 95 647 770 175 216 156 648 366 120 IV. 41 60 86 84 578 705 159 181 128 523 257 85 III 37 51 82 73 418 530 143 159 85 462 166 60 II - 31 46 66 61 267 306 100 105 278 105 45 I 71 80 31 It should be mentioned that the length data of table 19 are not in all cases strictly comparable from one population to another. The data for Trout, Muskellunge, Silver and Clear Lakes are all calculated lengths; the data for Lake Huron are a combination of calculated lengths and actual lengths measured in November; the data for Lake Ontario, Lake Erie, the Indian Village Lakes, and Hudson Bay represent actual measurements of fish caught during the growing season; and the data of Oconomowoc Lake and Pine Lake are actual average measurements of fish taken in the winter after the close of the growing season. These small differences in the manner of presen- tation should not affect the value of the data for purposes of comparison. Though growth in length of the cisco varies widely in different localities, certain consistent likenesses as to the manner of growth are apparent. The greatest growth in length occurs in the first year of life. The growth of the second year is large but less than that of the first. Growth decreases markedly in the third year of life. Throughout the later years the annual increment of growth in length tends to remain fairly constant. Growth in weight is too greatly complicated by the questions of differences in form and of change of form to permit a general description of the style of growth with respect to this character. 234 BULLETIN OF THE BUREAU OF FISHERIES Wilier (1929) called attention to the fact that certain fishes can be said to possess a typical species growth curve that retains a characteristic form regardless of the extent to which external factors influence the actual amount of growth. With re- spect to growth in length the cisco seems to possess such a species growth curve. It will be noticed in table 19 that the populations showing the greatest (Clear Lake, Indian Village Lakes) and the least (Trout Lake) amount of growth are found in small, inland lakes. Such a situation might be expected in view of the greater diversity of habitat furnished by the smaller bodies of water. The growth in the different populations cannot be classified according to latitude or region, but rather, the differences in growth seem to depend upon purely local conditions within the indi- vidual lake or possibly upon the genetic make-up of the local stock. A population showing a greater amount of growth in length than another does not necessarily show a greater amount of growth in weight. Two examples will illustrate this situation. (See tables 19 and 20.) At all ages beyond the seventh year the Oconomowoc Lake cisco is longer than either the males or females of the Clear Lake cisco, and yet at all ages the weights of the Oconomowoc fish are less than those of the Clear Lake fish of corresponding age. Secondly, at the age of 7 years the weights of the Clear Lake and the Indian Village Lakes ciscoes are approxi- mately equivalent. At the same time the length of the Indian Village Lakes fish at the age of 7 years is greater than that attained by the Clear Lake 9 males in the ninth year and equal to the length reached by the Clear Lake females in the tenth or eleventh year. These discrepancies between growth in length and growth in weight find their origin in the different length- weight relationships in the various populations. Populations may differ both as to general body form itself and as to the manner of change of form with increase in length. RANGE OF LENGTH IN INDIVIDUAL AGE GROUPS; MAXIMUM LENGTH AND WEIGHT Tables 21 to 25 show for the four lakes under consideration here the length distribution of the different age groups in each year’s collection. On the whole, length is a poor index of age. The amount of overlap between consecutive age groups is so great that in most instances a fish of a given length might have any of several different ages. This is particularly evident in the Trout Lake data of table 21. In the rapidly growing Clear Lake population, however, the positions of both sexes of the first three age groups (I, II, and III) stand out clearly in the length frequencies. In the 1931 Silver Lake collection the I group is well separated from the remainder of the sample. The Muskellunge Lake collection of 1931 shows a fairly distinct separation of the II and III groups, but a greater abundance of other age groups in the collections of this year might have obscured this division. 9 The average lengths of the Clear Lake fish measured at the time of capture were: IX group, male, 355 millimeters; X group female, 377 millimeters; XI group, female, 378 millimeters. AGE AND GROWTH OF THE CISCO 235 Table 21. — Length frequencies ( 5-millimeter intervals) of each age group of the Trout Lake cisco in each year’s collection Length I II III IV 1 VI VII VIII IX XI XII 1928 1930 1928 1930 1931 1928 1930 1931 1928 1930 1931 1930 1931 1931 1931 1931 1931 1931 1931 225 to 229 1 220 to 224 _ 215 to 219 1 210 to 214 205 to 209 200 to 204 1 1 195 to 199 .. 1 1 1 190 to 194 .. 2 1 185 to 189 1 1 180 to 184 __ 1 175 to 179. . 1 1 3 170 to 174 2 1 1 1 165 to 169.. 1 8 6 2 160 to 164 6 14 6 4 12 3 2 155 to 159 42 1 2 28 26 2 29 1 150 to 154.. 2 3 120 9 4 29 74 21 1 145 to 149 2 13 2 15 94 41 5 15 98 8 1 140 to 144 9 6 18 30 70 67 5 2 46 3 135 to 139. 30 6 18 12 12 42 1 16 130 to 134 1 43 4 15 1 2 12 3 125 to 129 2 16 4 8 1 | 120 to 124 1 2 1 Average 128 128 134 140 136 143 149 142 147 156 148 167 156 165 175 192 200 205 226 Table 22. — Length frequencies {5 -millimeter intervals) of each age group of the Muskellunge Lake cisco in each year’s collection [The 1932 III and IV groups are not included because of the high degree of uncertainty in their separation] Length i II III IV 1928 1930 1931 1932 1928 1930 1931 1932 1928 1930 1931 1931 180 to 184 1 2 3 175 to 179.... 3 4 1 13 2 170 to 174 14 14 2 4 73 165 to 169 28 46 3 1 2 5 126 160 to 164 2 1 88 92 20 8 3 4 111 155 to 159 4 2 78 35 24 6 i 20 150 to 154 i 2 27 18 60 7 1 145 to 149 4 2 12 5 93 1 140 to 144 6 1 3 i 50 1 135 to 139 1 4 5 5 130 to 134.... 4 4 3 125 to 129 8 i Average 137 148 136 140 160 162 149 156 172 166 166 176 Table 23. — Length frequencies ( 5-millimeter intervals) of each age group of the Silver Lake cisco in each year’s collection 236 BULLETIN OF THE BUREAU OF FISHERIES Table 24. — Length frequencies ( 10-millimeter intervals ) of each age group of the Clear Lake cisco (males) in each year’s collection Length I II III IV V VI VII VIII IX 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1932 360 to 369 2 350 to 359 1 1 1 340 to 349 3 1 5 1 1 330 to 339 2 1 2 1 5 1 2 1 320 to 329 - _ 3 1 3 2 2 2 2 2 310 to 319 . 1 2 6 4 1 i 5 i 300 to 309 . 1 2 2 2 1 290 to 299 10 10 280 to 289 8 4 1 270 to 279 .. 5 1 3 4 i 260 to 269 5 8 3 250 to 259 14 4 1 240 to 249 - 5 5 230 to 239 3 1 220 to 229. __ 1 210 to 219 1 i i 200 to 209 1 190 to 199 1 180 to 189 4 5 170 to 179 8 10 160 to 169 5 5 Average.. 178 176 253 254 285 290 317 314 325 320 326 320 331 336 338 345 355 Table 25. — Length frequencies ( 10-millimeter intervals) of each age group of the Clear Lake cisco ( females ) in each year’s collection Length I II III IV V VI VII VIII IX X XI 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1931 1932 1932 380 to 389 2 1 370 to 379 2 1 1 1 1 360 to 369 3 2 3 2 1 350 to 359 1 1 3 3 7 5 1 1 340 to 349 1 2 1 4 4 7 3 1 330 to 339 2 6 2 3 4 1 1 320 to 329 2 1 3 1 2 1 3 310 to 319. - 5 9 3 2 1 300 to 309 8 10 290 to 299 1 6 1 280 to 289 3 4 270 to 279 6 4 260 to 269 11 2 i 250 to 259 10 7 240 to 249 3 230 to 239 1 220 to 229 210 to 219 200 to 209. 4 190 to 199 1 1 180 to 189 8 2 170 to 179 5 7 160 to 169 150 to 159 1 Average 186 176 264 262 302 308 334 329 338 342 339 320 344 362 354 356 351 368 376 379 378 There are some disagreements between the length distributions of fish of the same age, but taken in different seasons. Some of these differences, particularly in the younger age groups, are due to the variation in the time of the collection of the sample within the growing season. The I groups of Clear Lake will serve as an example. (See tables 24 and 25.) In 1932 these ciscoes were collected from July 12 to July 19, while in 1931 they were taken during two periods, July 22 to July 28 and September 3 to September 5. Since the September samples averaged more in length (the average length of the September fish was about 200 millimeters) than the July samples, their presence accounts for the greater upward extension of the length distribution of the 1931 I-group Clear Lake ciscoes. However, at least some AGE AND GROWTH OF THE CISCO 237 of the other differences between the length distributions of fish of the same age may be accounted for by the actual differences in the growth histories of the year classes they represent. This appears to be true, for example, in the II groups from Muskel- lunge Lake (table 22). The 1931 II-group fish from this lake were much smaller in average size than those of the preceding 2 years. The examination of the tabula- tion of calculated lengths for these various groups (table 4) likewise shows that the 1931 II group (1929 year class) actually grew less during the first and second years of fife than did the 1929 and 1930 II groups (year classes of 1927 and 1928). In table 26 are given the data relative to the longest and the heaviest fish taken from each population. (Data concerning the lengths and weights of the oldest fish taken may be obtained from tables 3 to 7.) There is sufficient variation in individual growth that the largest individual need not necessarily be the oldest. In the Silver Lake collections and for both sexes of the Clear Lake samples the largest fish are from 1 to 4 years below the observed maximum age. It is true further that the largest fish with respect to length is not always the heaviest (for example, see the data on the longest and the heaviest fish from Silver Lake). The differences among the maximum sizes attained in the different populations are more striking if size is considered in terms of weight. The Clear Lake fish reach a weight approximately 17 times as great as those from Muskellunge Lake, 10 times as great as those from Silver Lake, and 7 times as great as those from Trout Lake. Table 26. — Data concerning the longest and the heaviest cisco in each of the f lakes Trout Muskellunge Silver Clear. Year of capture Age Length in milli- meters Weight in grams Sex 1931 XII 226 172 Female. f 1932 IV 180 70 Do. \ 1932 1 188 Male. / 1930 V 206 110 Female. 1 1930 VI 204 120 Do. | 1932 VIII 368 910 Male. ( 1931 X > 388 Female. ( 1932 VII 383 1,210 Do. * This fish was not weighed before it was preserved. The length recorded includes a correction for shrinkage. The maximum size attained depends both upon rate of growth and upon the maximum age of survival. The Trout Lake cisco is the slowest growing of the four populations, yet a few individuals survive to such an age as to reach a size greater than is attained by the fish in either Silver or Muskellunge Lakes. It is probable that further collections might change the order of arrangements of the lake with respect to the maximum size reached by the cisco, as well as the absolute maximum values for length and weight. CONDITION AND THE RELATIONSHIP BETWEEN LENGTH AND WEIGHT Weight in fishes may be considered a function of the length. If form and specific gravity 10 were constant throughout life the relationship could be expressed by the equation, W=cL\ (1) where W= weight, L= length, and c= constant. The above equation is a statement of the well-known cube law. Actually, in nature, the value of c is not constant for a species or population but is subject to a 10 Reibisch (1908) could find no appreciable seasonal variation in the specific gravity of the plaice. Keys (1928) pointed out that the hydrostatic equilibrium that exists between the fish and its environment renders great fluctuations in specific gravity unlikely. 238 BULLETIN OF THE BUREAU OF FISHERIES wide range of variation. The values of c under various definitions (coefficient of condition, condition factor, length-weight factor) have been used widely by fisheries investigators as measures of individual or average seasonal and regional differences in the condition or “degree of well-being” of fishes. Some investigators have used the coefficient as a measure of the state of sexual development. Oth erp have con- sidered condition to apply only to the state of nourishment and have removed the gonads before taking the weight for the calculation of the coefficient. Yet others have used the coefficient merely as a measure of relative heaviness and have recog- nized the effect of both the state of sexual development and of the state of nourishment on the determination of its value. In addition to its use as a means of estimating condition, the equation (1) has been employed also to describe the general length-weight relationship in populations of fishes, and thus serve as the basis for the calculation of unknown weights of fish of known length or of unknown lengths of fish of known weight. The use of the equation in this latter capacity has met with indifferent success, due to the failure of the cube law to describe accurately the relationship of length to weight in many forms of fishes. Fulton (1904) who applied the law to the relationship of length to weight in several marine species stated: The law in regard to the increase in weight according to the cube of the length, although broadly true, does not accurately apply in the case of the fishes examined. With scarcely an exception, the weight at a given length is greater than the weight calculated from the law, so that if the specific gravity of the fishes remains constant they must increase somewhat more in other dimensions than in length Although the cube law does appear to apply to the length-weight relationship in some species (Crozier and Hecht, 1914; Hecht, 1916), these instances appear to be the exceptions, for the observations of Fulton in regard to the inadequacy of the cube law in describing the length-weight relationship in fishes have been repeated by numerous investigators and on many forms of fishes. In recent years a much more satisfactory method of describing the length-weight relationship in fishes has been developed through the use of the more general equation: W=CLn, (2) where W— weight, L— length, and c= constant. In this equation the values of both C and n are determined empirically. Such a relationship has been determined by Jarvi (1920), Tjurin (1927), Clark (1928), Keys (1928), Fraser (1931), Hart (1931, 1932), Tester (1932), Walford (1932), and Schultz (1933). Some of the above authors have, however, confused the two entirely distinct issues of describing condition and expressing the length-weight relationship, and have abandoned the use of coefficients of condition based on the cube relationship in favor of those based on an empirically determined exponent ( C in W=CLn). That coeffi- cients calculated from the cube relationship and from empirically determined expo- nents are in no sense of parallel significance as measures of condition appears from the following simple illustration. AGE AND GROWTH OF THE CISCO 239 A fish with a length of 1 foot and a weight of 1 pound will show on the basis of the cube relationship a coefficient of condition of 1.00. In conformity with a rather general usage this quantity may be designated as K. If this fish doubles its length without change of form or of specific gravity its weight at the length of 2 feet will be 8 pounds, and the value of K will continue to be 1.00. If, however, the weight at 2 feet is 10 pounds instead of 8 the value of K at this greater length will be 1.25, and it will be known that a change of form 11 has occurred along with growth in length. It may be considered that 2 pounds of the weight of this 10-pound fish represent its change of form; this change is measured directly in the increased value of K. The 10-pound fish is 25 percent heavier than the 8-pound fish and is 25 percent relatively heavier than the 1-pound fish (that is, the 10-pound fish corresponds in form to a fish tfiat would have a weight of 1.25 pounds at the length of 1 foot). Thus it may be seen that values of K, by reason of their calculation from the cube relationship are direct and quantitative measures of form or relative heaviness, and in this sense are directly comparable between fishes of any length. Now if the weights of 1 pound at the length of 1 foot and of 10 pounds at 2 feet were to represent actual average conditions within a population, the corresponding length- weight equation would be: IF=1.00Z3-32193 The coefficients of condition calculated from this higher empirical exponent are 1.00 both for the fish at the length of 1 foot and weight of 1 pound and for the fish at the length of 2 feet and weight of 10 pounds. Thus it may be seen that the coefficient C fails entirely to measure in any way the change of form that occurred with increase in length. Further, the value of C for a fish that weighs 8 pounds at the length of 2 feet is 0.80, and from this value it would appear that a fish that doubles its length without change of form actually suffers a loss of condition. Since the assumption is hardly tenable that a fish can suffer such a great loss of condition without undergoing any change of form or relative heaviness it must be concluded that values of C cal- culated from empirically determined exponents fail to serve as satisfactory measures of condition. Questions concerning the use of C as a coefficient of condition will receive further consideration in connection with the presentation of the data of this investigation on condition and the length-weight relationship in the cisco populations of Trout Lake, Muskellunge Lake, Silver Lake, and Clear Lake. At the present time, however, a brief review will be presented of the observations and opinions of authors who have used or suggested the use of C in equation (2) as a coefficient of condition. The observations of Jarvi (1920) should be included in this review, although in justice to that author it should be emphasized that he was concerned primarily with the calculation of unknown lengths and weights from a general length-weight relation- ship rather than with the measure of condition. Jarvi’s data are of particular interest here since he included a comparison of coefficients based on the cube relationship and as determined from equations of the type, W=CLD, with empirically determined exponents. His coefficients for the males, the ripe females, and the spent females of the ldeine Marane ( Coregonus albula ) were, on the basis of the cube relationship, 11 The term “form” es employed throughout this section carries with it no implications as to details of shape, but refers only to relative heaviness of stature. 240 BULLETIN OF THE BUREAU OF FISHERIES 0.00848, 0.00989, and 0.00830, respectively.12 The coefficients as calculated from formulas with empirically determined exponents were, for the males, ripe females, and spent females, 0.0050, 0.0020, and 0.0056, respectively. The exponents in these last three formulas were 3.2 for the males, 3.64 for the ripe females, and 3.16 for the spent females. The comparison of the coefficients derived by the two methods shows clearly that those based on empirical exponents fail completely to reflect the relative heaviness of the fish groups to which they pertain. While the values of K based on the cube relationship show that the ripe females are on the average the relatively heaviest group, followed in order by the males and the spent females, the values of the coefficients calculated from equations with empirical exponents follow exactly the reverse order. Thus on the basis of the values of C of equation (2) the relatively heaviest group of fish would appear to be in the poorest condition. If the values of the two types of coefficients are compared in relation to the values of the exponents, n, it appears that while the values based on the cube relationship depend on the relative heaviness of the fish upon which they are based, the values of the coefficients calculated from equations with empirical exponents depend primarily not on the heaviness of the fish but rather on the value of the exponents. A large value of n is associated with a small value of the coefficient — and the reverse. Clark (1928) appears definitely to have confused the two problems of describing condition and expressing the length-weight relationship. From her study of condi- tion and the length-weight relationship in the California sardine ( Sardina caerulea) Clark concluded: “The weight of sardines increases at a rate slightly greater than the cube of the length. For the data studied the correct formula for the weight- 1000 W length factor was found to be F= ^3-"5 — But for the purpose of the present study the formula F= 1000 W U was sufficiently accurate.” Clark evidently believed that the coefficient of condition (weight-length factor), to be accurate, should tend to hold a constant value at all lengths of the fish, for she stated: “The more a species departs fiom this general weight-length relationship [cube relationship], the greater the error involved in the factor.” Concerning the changes in the value of the factor with increasing length she observed further: “Due to the error introduced from calculating F on the basis of the cube of the length, the , . „ , . 1000 w . . , , , ' curve resulting from the equation F = — ^ — rises consistently throughout the range of sizes represented in the commercial catch at San Pedro.” Clark’s conclusion that the failure of the cube law to describe the length-weight relationship makes inaccurate the use of coefficients of condition (weight-length factors) based on the cube relationship is scarcely justifiable, particularly in view of the fact that coefficients based on empirical exponents fail to reflect differences in form or relative heaviness while those based on the cube relationship offer a direct measure of relative heaviness independent of general length-weight relationships and comparable as measures of relative heaviness between fish of any length. Clark actually studied variations in the weight-length factor on the basis of values calculated from the cube relationship, but did so only because she considered that the use of these values introduced “only a minor error into the work.” i» In the data from which these coefficients were determined weights were recorded in grams and lengths in centimeters. AGE AND GROWTH OF THE CISCO 241 Clark’s data on the correlation between weight-length factors (calculated from the cube relationship) and the fat content of the sardine offers convincing evidence that measures of condition should be calculated from the cube relationship. Her data show conclusively that variations with length in the value of the weight- length factor are reflected closely in corresponding variations with length in the fat content of the sardine. Inasmuch as relative heaviness is thus showed to be de- pendent on fatness (condition), changes with length in relative heaviness must be considered also to represent changes of condition. In view of this fact it does not appear valid to measure condition in terms of a quantity that tends to be constant for fish of all lengths regardless of actual changes that may occur in relative heaviness of form with change in length. Since the quantity C in the equation W = CLD tends toward this constancy and fails to measure relative heaviness it must fail also to measure the differences of fatness (condition) upon which differences in relative heavi- ness depend. Tester (1932) likewise appears to have confused the two problems of measuring condition and describing the general length-weight relationship. Concerning his method of determining the general length-weight relationship in the smallmouth black bass ( Micropterus dolomieu ) he stated: Following the method used by Hart (1931) and others the relationship between length and weight for bass from Perch Lake [Ontario! was obtained by plotting average lengths against average weights on double logarithmic paper and deriving the natural slope of the resultant straight line drawn through the points. It was found that the weight increased by the power 3.1(7) of the length. * * * Tester assumed, however, that the failure of the cube law to describe the length- weight relationship made the use of coefficients of condition based on the cube rela- 100 W tionship invalid. Concerning the equation, K= — employed by Hile (1931) to describe condition in several species of Indiana lake fish, Tester stated: This equation is only approximate and might be better expressed — K = 100 W In the case of the Perch Lake bass the power x would have a value 3.1(7). If the power x = 3 is used, the value of K tends to increase with the length of the fish whereas it should remain constant. Tester used the value x =3. 1(7) as determined from the Perch Lake bass to compute coefficients of condition for his smaller samples from Phantom Lake, Lake Nipissing, and Georgian Bay, but observed that, “To be strictly accurate, the value x should be determined separately for each of the latter three bodies of water. * * *” Had Tester actually determined the value of the exponent for each of the bass populations he studied he would probably have become aware of the difficulties involved in the use of coefficients of condition based on empirical exponents. Schultz (1933) expressed the relationship of length to weight in the bay smelt ( Atherinops affinis oregonia ) by the equation W=FL 2-59. He also employed the quan- tity F as a measure of seasonal changes in weight, dependent on the state of develop- ment of the gonads. That Schultz considered the length-weight factor, F, identical in significance with factors calculated from the cube relationship is apparent from the following statement: “The length-weight factor, F, was used to estimate the state of development of the gonads, as has been done by D’Arcy Thompson (1917), 24535—36 3 242 BULLETIN OF THE BUREAU OF FISHERIES Van Oosten (1929), Clark (1928b), Weymouth (1918 and 1923), Reibisch (1911), Crozier and Hecbt (1915), and others.13 Of the above authors some employed the cube relationship in the study of the length-weight relationship (Thompson, Van Oosten, Weymouth (1918), Crozier and Hecht), while others used empirical exponents (Clark, Weymouth (1923)). Because of the serious objections to the use of coefficients of condition based on empirical exponents, condition in this investigation has been measured by coefficients calculated from the cube relationship. The comparison of the equations W—KL3 and W=CLn reveals an interesting connection between condition and length in popu- lations that deviate from the cube relationship. The equation, W=CLn, may be written in the form: W=f(LU(L), (3) where f(L) = CL ", (L)—L3, and m=n—3. Thus it may be seen that where weight can be expressed as a parabolic function of the type W—CLn condition can be expressed by a similar function of length. While this definition of condition as a function of length is valid only insofar as the equation (3) actually describes the length-weight relationship, the failure of equation (3) has no effect on the value of coefficients calculated from the cube relationshie as measures of relative heaviness. Where the coefficient of condition does behavp as a parabolic function of length (hyperbolic if m 61.0 1 1.0 2.8 <1931 (l932 11930 ' 42.0 1 .5 50.0 45.0 3.5 .7 Clear 11931 J1931 (1932 1 62.0 .1 # 2 > 76.0 2.3 1.9 3.7 2.4 5.5 4.3 7.5 3.8 1 This size of mesh was not represented in the gear of the preceding years. 1 The data for the 1932 Muskellunge Lake collections are presented in detail in table 40. 262 BULLETIN OF THE BUBEAU OF FISHEKIES From the examination of the data of table 44 it may be seen that the various populations show considerable differences in their relative densities at the time of capture in the depths from which the samples were taken. The Muskellunge Lake fish appear to be the most abundant although Trout Lake is not far behind. The Silver Lake data show a relative abundance slightly less than that in Trout Lake or Muskellunge, while the Clear Lake cisco population may be considered relatively sparse. The estimate of the relative densities of the four cisco populations at the time of capture is not, however, of primary importance in this investigation. Since this section is concerned chiefly with the relationship between density of population and rate of growth the most significant comparison of population densities must include a comparison of densities throughout the entire growing season rather than at some particular time within the growing season. It was pointed out in the sec- tion on the length of the growing season that growth of the Trout Lake cisco is prac- tically complete at the end of July; the Muskellunge cisco completes its growth by the end of August, possibly sooner; the season’s growth of the Silver Lake fish is two-tliirds to three-quarters complete in mid August; and the Clear Lake cisco has completed well over half its season’s growth by late July. From these facts it is apparent that the relative density of the four populations in late spring and the first half of the summer is more significant than their relative densities at the time the samples were collected. There is strong evidence that in Muskellunge Lake and Silver Lake at least the densities of the cisco popidations earlier in the growing season were less than at the time of capture in late summer. The concentration of these populations in late summer within a narrow stratum may be considered the result of the temperature and oxygen condition at that time. It is most probable that in late spring and early summer no such concentration existed, since in the early season oxygen would be expected to occur at all depths in the hypolimnion. Consequently the relative densities for the Muskellunge Lake and Silver Lake fish in table 44 must be considered too high as compared to those from Trout Lake and Clear Lake. If attention is given to the relative densities during the entire growing season the most probable arrangement of the lakes from the most dense to the least dense population is: Trout, Muskellunge, Silver, Clear. This is the same order these lakes show with respect to growth rate in weight (with slowest growth in Trout Lake and fastest growth in Clear Lake). The growth rates of the four populations are approximately in the reverse order of the productive capacities of the lakes as estimated from the amount of bound C02 in their waters (Clear Lake was estimated as the least productive, followed in order by Muskellunge, Silver, and Trout). Thus it appears that the growth rates of these populations are determined primarily by their relative densities rather than by the basic productive capacities of the waters they inhabit. Although it is recognized that crowding itself may possibly impede growth to a certain extent independently of its effect in creating competition for food, and that various physical-chemical factors may affect growth rate directly, it is believed that the differences in growth rate in these four populations depend in large measure on the varying degrees of competition for food in the different lakes. It is further probable that variation in the intensity of competition for food from lake to lake may be related to the observed differences in the length of the growing season of the different stocks. (See preceding section.) AGE AND GROWTH^OF THE CISCO 263 AGE COMPOSITION OF THE SAMPLES AND THE RELATIVE ABUNDANCE OF YEAR CLASSES Because of its importance in the study of commercial fisheries the question of age composition of the stock and of the existence of dominant year classes has in recent years received a great amount of attention. It is now well known that the success of different years’ hatches as measured in terms of the number of young produced is subject to a wide range of variation from one year to another. The hatches of some years may be so successful that that particular year class may domi- nate in the fishery over a period of 1 or several years. Yet other hatches may be so unsuccessful as to make only minor contributions to the stock. The common occur- rence of relatively successful and unsuccessful year classes has been demonstrated in many species of fish of commercial importance. The study of the age composition of the commercial catch over a period of years has served as a most valuable tool in the investigation of fluctuations in abundance. In both Europe and America the known age composition of the commercial stock, together with the observed rate of falling off of a year group from year to year, has been used by fisheries investigators as a basis of prediction of the probable yield of the fishery for the approaching season. In spite of the extensive researches that have been made on the subject of fluctua- tions little is known as to the causes that make a year class good or bad. Hjort (1914) stated that while the basic causes of fluctuations in abundance are unknown, it appears that fluctuations “have their origin in certain conditions prevailing at a very early period in the life of the fish.” He repeated this opinion in 1926. Storrow (1932) pointed out that not only are the factors that determine the success or failure of a year class complex, but also that the difficulty of observing the fish in its natural environment is great. He called attention further to the failure of attempts to simu- late natural conditions in the laboratory, a convincing proof of inadequate knowledge of optimum conditions. The most thorough-going studies on the question of fluctuations in abundance in coregonids and their causes are those of Jarvi (1920, 1924, 1930). He was able to show not only a great variation in the abundance of different year classes of the “kleine Marane” ( Coregonus albula ) but also a distinct connection between such variations in abundance and weather conditions in the spring just after the time of hatching. If there are strong winds at this critical period many of the delicate newly hatched young are destroyed by the wave action. Jarvi found, however, that strong winds at the time of spawning are less harmful. Huitfeldt-Kaas (1917) found that in some years in Lake Mj0sen almost all the spawning run individuals of Coregonus albula are taken in the commercial fishery before they have had an opportunity to spawn, and that poor year classes may result. Because of the manner (collecting with gill nets) in which the samples were obtained in the present investigation the study of the relative abundance of the various age groups in the four cisco populations and along with it the consideration of the relative abundance of the different year classes must be approached with great caution. (The limitations of gill net samples in the study of relative abundance of age groups and year classes are discussed in the section on the selective action of gill nets.) The age composition of a gill-net sample must in general be considered as descriptive of the sample rather than of the population as a whole. Exceptionally, however, the representatives of some particular year class may appear so promi- nently or be so scarce at all of the ages at which that year class appears in different 264 BULLETIN OF THE BUREAU OF FISHERIES collecting years that it may be possible to designate the year class as good or poor. The examination of the age composition of the different years’ collections in each of the four lakes (tables 45 to 48) does indicate the presence of certain year classes that can be termed relatively good or relatively poor. In the Trout Lake collections (table 45) the year classes of 1926 and 1927 appear to represent good years for the production of young. The former year class as the II group of 1928 made up more than half of that year’s collection, and 3 years later as the V group of 1931 it was still relatively abundant, comprising 13 percent of the total collection in this latter year. By reason of gear selection the 1927 year class (I group of 1928) was almost entirely lacking in the 1928 collections. This year class was, however, dominant in the collections of the later years, first as the III group of 1930 and then as the IV group of 1931. These conclusions as to the “goodness” of the 1926 and 1927 year classes are supported by the comparison of the relative abundance of corresponding age groups in the different years’ collections. The 1926 year class furnished the greatest relative abundance of any II group in any year 23 (56.0 percent of 1928 collection) and also of any V group (13.0 percent of the 1931 collection). Similarly the 1927 year class furnished the greatest relative abundance of any III group (1930) and of any IV group (1931). It should be mentioned further that the presence of these two relatively successful groups contributed toward the progressive increase from year to year in the average age and average size of the fish in the collections obtained from Trout Lake. Table 45.- — Age composition of the samples of the Trout Lake cisco [The percentages are given in parentheses] Age ture i II III IV V VI VII VIII IX X XI XII 1928 2 (1. 1) 102 (56. 0) 61 (33. 5) 17 (9.3) 1929 1930... 36 (9.2) 61 (10.0) 247 (63. 2) 173 (28. 4) 99 (25. 3) 269 (44. 1) 9 (2.3) 79 (13. 0) 1931 2 (0.3) 12 (2. 0) 4 (0.7) 5 (0. 8) 2 (0. 3) 2 (0. 3) 1 (0.2) In the Muskellunge Lake collections (table 46) the year class of 1928 (II group of 1930, III group of 1931) may be considered relatively successful. This year class, as the III group composed 55.6 percent of the 1931 collections, whereas in preceding years the III group composed a negligible portion of the total collection. The simi- larity of the age composition of the 1928 and the 1930 collections suggests the possi- bility that the 1926 year class may have been exceptionally abundant and that a collection in 1929 would have shown a high percentage of Ill-group individuals. However, the almost total absence of all age groups above the II group in the 1928 and 1930 collections suggests also the possibility that heavy mortality may regularly reduce the numbers of Muskellunge ciscoes early in life, and that the great relative abundance of the III group in 1931 may depend not only on the great relative abun- dance of the 1928 year class but also in part on the failure of this year class to suffer this customary great mortality. The scarcity of I-group individuals in all Muskel- lunge Lake collections may be considered the result of selectivity by gear. 23 The great abundance of II-group fish in 1928 may depend in part on the smaller mesh gill nets used in that year. AGE AND GROWTH OF THE CISCO 265 Table 46. — Age composition of the samples of the Muskellunge Lake cisco [The percentages are given in parentheses] Year of capture Age I II III IV 1928 19 (6. 8) 252 (89. 7) 10 (3.6) 1929. 1930 17 (6.9) 9 (1. 5) 214 (87. 3) 258 (41.9) 14 (5.7) 347 (55. 3) 1931 - 2 (0 3) The study of the age composition of the Silver Lake collections (table 47) reveals the presence of one year class (1926) which may be considered good and of one year class (1929) which may be considered poor. The 1926 year class was dominant in the collections of 2 of the 3 years, as the IV group of 1930 and as the V group of 1931. The scarcity of 1926 year class individuals as the II group in the 1928 collection can be explained on the basis of gear selectivity. The 1929 year class must be considered relatively poor because of its relative scarcity as the II group of the 1931 collections. This scarcity can hardly be the result of selection by gear as the individuals of the age groups on either side are more than three times as numerous. The great abun- dance of I-group fish in 1931 as compared with 1930 is the result of the introduction of smaller meshed nets in the first-named year. The lack of individual net records makes it impossible to determine the reason for the scarcity of I-group fish in 1928. However, the 1930 year class can safely be considered more abundant than the 1929 year class. The 1924 year class which was dominant as the 1928 IV group may pos- sibly represent a good production year. Table 47. — Age composition of the samples of the Silver Lake cisco [The percentages are given in parentheses] Year of capture Age I II III IV V VI VII 1928 1 (0.7) 9 (6. 7) 46 (34. 1) 69 (51. 1) 10 (7.4) 1929 1930 7 (5.9) 19 (5. 0) 25 (21. 2) 61 (16. 1) 58 (49. 2) 102 (27. 0) 25 (21. 2) 108 (28. 6) 3 (2. 5) 21 (5. 6) 1931 66 (17. 5) 1 (0. 3) The Clear Lake samples (table 48) present much more definite and consistent indications of the presence of successful and unsuccessful year classes than were found in the other three lakes. Figure 5 shows graphically the year class composition for the 1931 and 1932 collections. The agreement between the 2 years in the relative abundance of the different year classes is close. Since the individuals of each year class were a year older in 1932 than in 1931, and consequently of a different size range, this close agreement between the year class composition of the 2 years’ collections may be taken as strong evidence for a high degree of reliability of the Clear Lake samples both as to year class and as to age composition. 266 BULLETIN OF THE BUREAU OF FISHERIES Table 48. — Age composition of the samples of the Clear Lake cisco [The percentages are given in parentheses] Age Year of capture I II III IV V VI VII VIII IX X XI 1931 38 (15. 3) 31 (16. 2) 09 (24. 5) 33 (17. 3) 51 (20. 5) 44 (23. 0) 14 (5.6) 25 (13. 1) 6 (2.4) 14 (7. 3) 23 (9. 2) 3 (1. 6) 29 (11. 6) 21 (11.0) 16 (6. 4) 15 (7. 9) 1 (0.4) 3 (1.6) 2 (0. 8) 1 (0. 5) 1932 1 (0.5) It can be seen at once from the examination of table 48 and figure 5 that in Clear Lake the 1926 year class (V group of 1931, VI group of 1932) was very unsuccessful, 1921 1922 1923 1924 1925 1926 1927 1926 1929 1930 1931 YEAR CLASS Figure 5. — Percentage representation of the different year classes in the Clear Lake cisco collections. 1931, broken line; 1932, solid line. while the 1929 year class (II group of 1931, III group of 1932) which was dominant in the collections of both years may be termed good. It can be considered valid to state further that the 1924 and 1925 year classes (VII and VI groups of 1931) and probably the 1928 year class (III group of 1931) were more successful than those of 1926 and 1927 (V and IV groups of 1931), and that the 1929 year class (II group of 1931) was more successful than either that of 1930 (I group of 1931) or 1931 (I group of 1932). 24 In the comparison of the relative abundance of different year classes in Clear Lake some consideration should be given the matter of the age of the samples upon 24 In contrast to the other 3 populations the I-group samples of Clear Lake can be considered representative. (See the section on “The selective action of gill nets.”) AGE AND GROWTH OF THE CISCO 267 which the estimations are based. For example, the 1927 year class which as the IV group composed 5.6 percent of the 1931 collection and as the V group composed 7.3 percent of the 1932 collection must be considered to represent a much less successful year than the 1924 year class which as the VII group of 1931 composed 11.6 percent of that year’s collection and a year later as the VIII group made up 7.9 percent of the 1932 collection. Although the 1924 year class individuals in both years’ collections combined were only about one and one-half times as numerous as those of the 1927 year class, the former year class, by reason of its 3 years greater age, had suffered the reducing effect of natural mortality over a longer period of time. Consequently the 1924 year class as compared with the 1927 year class may be considered much more successful than the comparison of the representation of the two groups in the collections would indicate. A similar line of reasoning leads to the conclusion that the 1925 year class (VI group of 1931) also represents a very successful year. The same conclusion may possibly apply to the 1923 year class which was well rep- resented as the VIII group of 1931. In the preceding discussion it was shown that in each population there existed a considerable variation in the numerical abundance or successfulness of the various year classes. The years that appear to have had a production of relatively successful year classes were: Trout Lake — 1926, 1927; Muskellunge Lake — 1928; Silver Lake — 1926; Clear Lake— 1924, 1925, 1929. In Silver Lake the 1929 year class was poor, and in Clear Lake the 1926 and 1927 year classes must be considered relatively unsuccessful. Since the four populations show no agreement in the matter of goodness or poorness of the various year classes, it may be concluded that in each population the success of the individual year classes depends on the purely local conditions within the lake. This conclusion is of particular interest in view of the rather general opinion that fluctuations in the relative abundance of year classes have their origin in differences in weather conditions from one year to another. Inasmuch as these four lakes are grouped closely together (the maximum distance between any two of them is about 18 kilometers) they must all be subjected to approximately the same weather conditions. It must be remembered that relative abundance is not a wholly reliable index of absolute abundance. A poor year class may appear quite numerous in comparison to a yet poorer one. Jarvi (1920) gave an excellent illustration of this fact when he pointed out that the great relative abundance of third year fish in some years’ catches of the “kleine Marane” depended not upon the actual abundance of that particular age group but was rather the result of the scarcity of fish in their second year of life. The data of tables 45 to 48 make possible a comparison of longevity in the four populations. The Muskellunge Lake cisco has definitely the shortest average life span. The Trout Lake cisco falls second and the Silver Lake cisco third, while the average life span of the Clear Lake cisco is quite long. A possible explanation of these differences among the four populations as to the average length of life was mentioned in the section on condition in the different populations. AGE AT MATURITY AND SEX RATIO Although the collections of the cisco were made in July, August, and early September, several months before the time of spawning in November, the develop- ment of the gonads was such as to leave little doubt as to the sexual maturity of the 268 BULLETIN OP THE BUREAU OF FISHERIES individual. Fish that would have spawned the following autumn could be distin- guished easily from those which would not have spawned at that time. The for- mer were considered here as matured fish even though they may not have spawned previously. Sexual maturity occurs at an early age in each of the four populations. All the Clear Lake fish with the exception of a few I-group individuals were mature or matur- ing. Since there is good reason to believe that the I-group samples from Clear Lake are representative, it may be concluded that in this population the first spawning occurs in the autumn of the second year of life. Practically all the I-group indi- viduals in the samples from Trout, Muskellunge, and Silver Lakes would have spawned the following autumn. There is reason to believe, however, that in these populations the nets used for collecting the samples took only the larger I-group individuals. The smaller members of the age group may have been immature. In all three populations, however, all II-group individuals indicated that they would spawn in the autumn. The ciscoes of the four populations considered in this investigation attain sexual maturity at an earlier age than has been found by some investigators of this species. Clemens (1922), referring to L. artedi in Lake Erie, stated that first spawning is “probably at the end of the third summer.” The ciscoes of the inland lakes of southern Wisconsin spawn at the age of 3 years (Cahn 1927). Van Oosten (1929) found that in the Saginaw Bay herring only a few fish matured in the second year, and that the majority maturedun the third[and fourth years of life. Van Oosten’s data were based on the examination of individuals from the spawning run. In the Lake Ontario cisco Pritchard (1931) found a few females but no males mature at 2 -{-years. He stated further that several individuals of both sexes were mature at 3 + years, while spawning was general at 4 -{-years. Dymond (1933) found that in Hudson Bay the cisco does not mature “until the fourth and probably the fifth year.” The sex ratio expressed as the number of females per 100 males was determined by age groups for each of the four populations. These data are presented in table 49. Table 49. — Sex ratio according to age in each of the four populations [The ratio is expressed as the number of females per 100 males. The numbers of specimens are shown in parentheses. The age groups which were composed entirely of one sex are indicated by the sex symbols. The Clear Lake data were obtained by com- bining the 1931 and 1932 collections. The data for the remaining three lakes are based on the combination of the 1930 and 1931 collections] Lake Age I II III IV V VI Trout 100 (2) 100 (26) 110 (66) 72 (69) 162 (97) 130 (472) 160 (26) 89 (102) 204 (520) 151 (361) 126 (86) 94 (95) 354 (368) cf(2) 111 (160) 86 (39) 1, 367 (88) 1, 100 (12) Silver ... 129 (133) 82 (20) 200 (24) 100 (26) Clear Lake Age Average VII VIII IX X XI XII Trout 9(4) 400 (5) 9(2) Uf a -6° 1923 1924 1923 1920 1927 1926 1929 1930 1931 Figure 8.— Relationship in different calendar years between the deviation of growth of Clear Lake cisco from average and deviation of air temperature from normal during cisco’s growing season. Deviation from average growth, broken line; deviation from temperature, solid line. ness of growth for certain populations and over certain periods of years. For example, Trout, Silver, and Muskellunge Lakes agree in showing improvement in growth from 1928 to the relatively good year 1929, followed by a distinct drop to the relatively poor year of 1930. Similarly Trout, Muskellunge, and Clear Lakes agree in showing poor growth in 1926 and 1927. Agreements of the sort just pointed out suggest that the amount of growth in different calendar years may possibly be determined in part by factors that affect all four lakes. The failure of any general agreement of the growth deviation curves indicates, however, that local conditions that exist within a single lake also exert a powerful influence on the amount of growth in different calendar years. The most apparent influence that might affect growth similarly in all lakes is the average temperature during the growing season, while growth within a single AGE AND GROWTH OF THE CISCO 277 lake might be expected to vary somewhat with annual fluctuations in the abundance of the population wit tun that particular lake. References to the literature on the question of the change in growth rate that accompanies change in the density of population have been made previously (pp. 254-257). Some mention should be made of the much less numerous observations on the relationship between annual fluctua- tions in average temperature and annual fluctuations in growth, particularly as these observations have been made with reference to coregonids.26 Huitfedlt-Kaas (1917) demonstrated a close correlation between first-year growth and average annual temperature in the lagesild or pollan ( Coregonus albula ) of Mjpsen and Storsjp. At the same time he recognized the possible effect on growth rate of changes in the density of population from one year to another. In a later paper (Huitfedlt-Kaas, 1927) he was able to show a relationship between the amount of growth and summer temperature in the gwiniad ( Coregonus lavaretus) and the pollan ( C . albula). (He noted a similar situation in several tribes of the trout ( Salnio eriox).) Olofsson (1932) compared the growth of three species of Coregonus (C. wartmanni borealis, C. wartmanni generosus, and C. lavaretus) of Norrland in the warm summer of 1930 with that in the cold summer of 1931. In all three forms there was large growth in the warm summer and slow growth in the cold. Olofsson called attention particu- larly to the fact that occasionally growth during a cold season may be so small that the year ring formed on the scales might easily be overlooked. Some scales may even fail to form a distinct year band. The differences observed between growth of the years 1930 and 1931 were exhibited by fish of different sizes and ages. Krogius (1933) found that the curve of the deviation from average growth in different calendar years for Coregonus lavaretus of Lake Baikal followed closely the corresponding curves for deviation from average annual temperature and average annual precipitation. Van Oosten (1929) found no relation between annual fluctuations in the air temperatures during the growing season and the annual fluctuations in the first year’s growth in the Saginaw Bay herring (L. artedi). There are no records of water temperature available upon which to base a com- parison of annual deviation from normal temperature during the growing season and annual deviation from average growth in the cisco populations of this study. It is possible, however, to make a comparison between annual deviation from average growth and annual deviation from normal air temperature 27 in the region. For this purpose the records of the meteorological stations at Big St. Germaine Dam and Rest Lake Dam were taken as representative for Trout, Muskellunge, and Silver Lakes, while the records of the Minocqua station were taken as representative for Clear Lake.28 Since the lengths of the growing seasons (pp. 249-253) are not the same in the different lakes the annual deviations from normal summer temperature were calculated 18 Although his work does not deal with the coregonids, mention should be made of Segerstrale’s (1932, 1933) observations on the relationship between summer temperature and the growth of fishes in southern Finland. This work dealt with the perch ( Perea fluviatilis) and several species of cyprinids. Segerstrale found that the amount of growth during the summer is a quantity highly sensitive to the average summer temperature. 27 Climatological Data for the United States, U. S. Department of Agriculture, Weather Bureau, vols. X-XVIII, 1923-31. 28 Muskellunge Lake lies about 3 kilometers east of Trout Lake, and Silver Lake lies about 1 kilometer off the southwest corner of Trout Lake. The St. Germaine weather station is about 13 kilometers southeast of Trout Lake while the Rest Lake station lies at an approximately equal distance to the northwest of Trout Lake. The town of Minocqua is about 6 kilometers due west of Clear Lake. 278 BULLETIN OF THE BUREAU OF FISHERIES from different combinations of months. The St. Germaine and Rest Lake data were used to determine the annual deviation from normal temperature for the periods May to July and May to August, both inclusive. The former period of time may be taken to correspond approximately to the growing season of the Trout Lake cisco, and the latter to the growing seasons of the Silver Lake and Muskellunge Lake ciscoes. The annual deviation from normal temperature for the longer Clear Lake growing season (May to September, inclusive) were determined from the Minocqua station data. In these calculations of deviations from normal temperature, records for a single month were found to be occasionally lacking in the data for a station. For these few gaps in the data the corresponding records of the nearest neighboring station were substituted. This substitution is justifiable for the corresponding devia- tions from normal temperature at the different stations were almost always close to each other in value. Table 56 shows the annual deviations from normal air temperatures as based on the different combinations of months and for the weather stations mentioned in the preceding paragraph. These data are presented graphically in figures 6 to 8 along with the corresponding curves of deviation from average growth. The examination of these curves shows that there is little evidence of a connection between annual deviation from average growth and annual deviation of the air temperature from the normal during the growing season. In the Trout Lake data (fig. 6) the years (1924, 1928, and 1929) with better than average growth were years with subnormal tem- perature for the period, May to July. On the other hand, 1930 with a temperature slightly above normal was a poor growth year. The year 1929 was a good growth year in both Muskellunge Lake and Silver Lake (fig. 7), while 1930 was a year of poor growth. Yet the air temperature for the period, May to August, was below normal in 1929 and above normal in 1930. Certain other, years, however, show both sub- normal temperature and growth below average (for example, Silver Lake, 1924). There is some indication that in Clear Lake (fig. 8) there may be some slight connec- tion between annual variations in growth and annual variations in average air tem- perature over the period May to September. It may be seen that in Clear Lake all the calendar years with poorer than average growth (1924 to 1928, inclusive) showed subnormal temperatures during the growing season while 3 of the 4 years with better than average growth (1923, 1929, 1930, 1931) had temperatures above normal during the growing season. The year 1929 showed growth slightly above average, but tem- perature below normal. In general, however, the Clear Lake curves for annual devia- tion of the cisco’s growth from the average and the annual deviation of the air tem- perature during the growing season from normal show a rather conspicuous lack of parallelism. Table 56. — Average deviation of air temperatures from normal during the growing season of the cisco [Deviations are in degrees Fahrenheit; 1.8° F. = 1.0° C.] Stations Months 1923 1924 1925 1926 1927 1928 1929 1930 1931 St. Germain Rest Lake.. May, June, July 2.0 -5.8 -1.8 -1.2 -3.5 -0.7 -2.4 0.2 -0.2 Do May, June, July, August. .8 -4.8 -0.2 -0.7 -3.6 -0.5 -1.9 .8 -0.6 Minocqua May, June, July, August, September. 1.1 -4.5 -0.6 -1.6 -4.3 -0.4 -2.6 .3 .4 The failure of the data discussed in the preceding paragraph to show any clear dependence of growth in different calendar years on the temperature of the air during AGE AND GROWTH OF THE CISCO 279 the growing season should not be taken to show that in these four cisco populations the amount of growth in a particular summer does not depend in part on the tempera- ture of the lakes’ waters during the growing seasons of the several stocks. It must be recognized first that fluctuations in air temperature may not offer a perfect index of fluctuations in water temperature. A more probable explanation of the failure of annual fluctuations in temperature and in growth to show correspondence is that annual differences in growth are so closely dependent on some other factor or factors that the effect of annual differences in temperature are almost completely obscured. There is reason to believe that in the populations of this investigation the annual fluctuations in the goodness of growth may show some dependence on annual fluctua- tions in the densities of the different populations. It has been shown previously (p. 262) that the growth rates (in weight) of these four populations follow the inverse order of the relative densities of their populations. In view of this demonstrated relationship between growth rate and density of population it might well be expected that within a single population changes in the density of the cisco may be accompanied by changes in its growth rate. In the study of the age composition of the cisco samples from Trout Lake (table 45, p. 264) it was pointed out that the years 1926 and 1927 saw the production of abundant year classes, and that as the result of the abundance of these two groups the 1930 Trout Lake cisco samples were on the average composed of older fish than the 1928 samples while the 1931 samples were composed of older fish than either the 1928 or 1930 collections. This accumulation of old fish together with the contribu- tions of young in the years later than 1927 probably caused the density of the popu- lation to be rather high in 1930. This increase in the density of the population may account in part for the drop in growth from 1929 to 1930. The data for the Muskellunge Lake cisco also afford evidence for a dependence of growth variations on variations in population density. While the 1928 and 1930 collections contained only a few individuals more than 2 years old, a large number of the 1928 year class individuals were present as the III group in 1931 . (See p. 264.) The great abundance of the III group in 1931 suggests that an accumulation of the stock may have been occurring in Muskellunge Lake in 1929 and 1930. This accumu- lation may possibly account for the drop in the amount of growth from 1929 to 1930. The age and year class composition data for the Silver Lake cisco (table 47, p. 265) show that here as in Trout Lake there was probably an accumulation of the stock in 1930. In the examination of the data of table 47 it should be remembered that the large number of 1931 I-group fish causes the representation of the older age groups of the 1931 samples to appear relatively too low; the gear used in 1928 and 1930 captured few I-group individuals. If due consideration is given to the absence of the I-group fish in the 1928 and 1930 collections the relative abundance of the older age groups in the 1931 Silver Lake collections must be recognized to be higher than in 1930, while the 1930 collection in turn shows a greater abundance of old fish than the 1928 collection. In Silver Lake as in Trout Lake and Muskellunge Lake the accumulation of older fish was accompanied by a decrease in growth from 1929 to 1930. It should be mentioned that the data upon which a suggested explanation of the drop in growth from 1929 to 1930 in Trout, Silver, and Muskellunge Lakes was based fail to offer any logical explanation for the improvement of growth from 1928 280 BULLETIN OF THE BUREAU OF FISHERIES to 1929. Unfortunately there are no 1929 collections upon which to base a com- parison of the age composition of that year with the years 1928, 1930, and 1931. In the Clear Lake cisco the evidence for a dependence of annual fluctuations in the amount of growth on annual fluctuations in the density of the population is somewhat stronger than in the Trout, Muskellunge, and Silver Lake populations. The years 1923, 1 924, and 1925(p.265, table 48, fig. 5) were undoubtedly excellent years for the production of young. The presence of three successive good hatches must have led to a consider- able accumulation of the cisco stock. Corresponding with this accumulation the amount of growth per year decreased from 1923 to 1927 (fig. 8). The production of young was poor in 1926 and 1927. The occurrence of 2 poor production years com- bined with natural mortality could be expected to lessen the crowding of the popula- tion. The growth improved from 1927 to 1929, fell back a little in 1930, and im- proved again in 1931. 29 The relationship indicated in the Trout Lake, Muskellunge Lake, Silver Lake ciscoes, and especially in the Clear Lake cisco, between annual fluctuations in popula- tion density and annual fluctuations in growth is in agreement with Jarvi’s observa- tions (1920, 1924) on the kleine Marane ( Coregonus albula). The failure of variations in the amount of growth in different calendar years to show any close general dependence on either annual variations in temperature or annual variations in population density suggests that possibly these variations in growth depend closely on both factors, and that the failure of these factors to operate in the same direction in the same year tends to obscure the effect of each of them. BIMODALITY IN THE CALCULATED GROWTH FOR THE FIRST YEAR OF LIFE The examination of the frequency distributions of the calculated growth for the first year of life in the best represented age groups (tables 57 to 60) shows that some of these distributions have a distinct bimodality, which appears to be characteristic for a year class and present regardless of the age of the fish upon which the calculated growths were based. Table 57. — Trout Lake cisco — Frequency distribution by 5 -millimeter intervals of the calculated growth in length during the first year of life Year class 1926 1927 1928 1929 Length II IV V III IV II in II 1928 1930 1931 1930 1931 1930 1931 1931 no 1 1 1 105 1 1 1 1 1 1 100 4 1 1 7 4 5 6 2 95 6 2 14 4 8 15 6 90 7 8 5 17 16 6 34 13 85 9 21 15 32 24 7 31 20 80 15 30 23 74 48 5 26 12 75 30 20 27 85 78 3 31 6 70 26 11 6 83 72 26 1 65 3 6 2 30 18 3 60 5 3 55 1 28 The good growth in 1931 may have been in part due to the climatic conditions of that year as well as to the reduced number of ciscoes in the stock. The temperature over the period, May to September, inclusive, was only slightly above normal (table 56) but the autumn temperatures were exceptionally high (p. 253). It is probable that the warm autumD of 1931 gave the Clear Lake cisco a longer growing season in that year than it usually enjoys. AGE AND GROWTH OF THE CISCO 281 Table 58. — Muskellunge Lake cisco— Frequency distribution by 5-millimeter intervals of the calculated growth in length during the first year of life Length Year class Lengtli Year class 1928 1928 1929 1926 1928 1929 II II III II II II hi II 1928 1930 1931 1931 1928 1930 1931 1931 125 2 2 90 20 39 28 66 120 6 2 7 85. 35 41 64 22 115 28 8 16 3 80 29 26 60 4 110.... 24 20 46 7 75. 17 12 17 1 105.... 35 21 39 33 70 1 2 5 100 27 17 37 60 65 2 95 16 21 26 72 Table 59. — Silver Lake cisco — Frequency distribution by 5-millimeter intervals of the calculated growth in length during the first year of life Table GO. — Clear Lake cisco — Frequency distribution by 5-millimeter intervals of the calculated growth in length during the first year of life 24535—36- -6 282 BULLETIN OF THE BUREAU OF FISHERIES This bimodality of the length frequencies of the calculated growth for the first year of life may be found in the 1928 year class from Trout Lake and in the 1926 and 1928 year classes from Muskellunge Lake. All the other year classes from these two lakes, and all year classes from Silver Lake and Clear Lake, where the samples are large enough to give reliable results show unimodal distributions for the calculated growth of the first year of life. The fact that bimodality in the first year’s growth occurs only in two populations and in only one or two year classes of these populations suggests that the occurrence of the phenomenon depends on the nature of the local conditions within each lake and that these conditions vary from year to year. The most reasonable explanation for bimodality in the first year’s growth lies in the assumption that in certain years there are two hatchings rather than a single one. In early spring a period of warm weather with brilliant sunshine and no strong winds can warm the waters of the shallow littoral region to a temperature several degrees above that of the main body of the lake. At such a time the development of the eggs of the cisco would be accelerated and some might hatch. If, however, there occurred before the completion of hatching season a period of cold, windy weather the temperature of the water of the littoral region would undergo a sudden drop of several degrees.30 The development of the unhatched eggs would be retarded, and their hatching might be delayed for several days or even weeks. Such a situation would explain the observed cases of bimodality in the amount of growth during the first year of life. Eggs that develop in more exposed regions of a lake or in a lake more swept by winds would be less affected by fluctuations in weather conditions. GROWTH COMPENSATION The phenomenon of “growth compensation” — the tendency for individuals that grow relatively slowly in the early years of life to grow relatively rapidly during the later years — has been observed by numerous investigators and in several species of fish. The only study of growth compensation in the cisco was made by Van Oosten (1929) on the Saginaw Bay herring of Lake Huron. He concluded that “the large fish of an age group were the large fish in each preceding year of life * * * but that the differences between the small and large yearlings diminished each year of age- — that is, the small yearlings were rapid growers and the large yearlings slow growers.” Thus he found that compensation did occur, but that it was not sufficient to overcome completely any advantage in length which a large individual might hold at the end of the first year of life. A comparison of growth compensation in age groups with unimodal and bimodal length distributions at the end of the first year of life should yield information as to the effect of the dispersion of the length frequency in early life upon the manner of growth in later life. In the 1931 collection from Muskellunge Lake (table 58) the distribution of the calculated lengths at the end of the first year is unimodal in the II group and bimodal in the III group. Both age groups are represented by large samples (258 for the II group and 347 for the III group). For these reasons they were selected as the basis for a study of growth compensation. Table 58 shows the frequency distributions of the calculated lengths for both age groups at the end of the first year of life. The calculated length distributions of the 3° Forel (1892) pointed out that the development of high temperatures in the littoral region depends on days of great calm and brilliant sunshine, and that this warming process proceeds slowly. The negative changes in temperature proceed much more rapidly. Forel observed that with a strong south wind the temperature of the littoral waters at the port of Geneva at times dropped as much as 6° C. to 8° C. or more from one day to another. AGE AND GROWTH OF THE CISCO 283 same groups at the end of the second year of life appear in table 61. Table 62 shows the frequency distribution of the increments of growth during the second year of life. All of these data are presented graphically as percentage frequencies in figure 9. Figure 9.— Percentage frequency distribution of the 1931 samples of 1928 and 1929 year classes of Muskellunge Lake cisco with respect to calculated lengths at end of first year of life (below at left), the calculated lengths at end of second year of life (below at right), and calculated increments of growth during second year of life (above). Table 61. — Muskellunge Lake cisco, 1931 — Frequency distribution by 5-millimeter intervals of the calculated length at the end of the second year of life as based on the II group and the III group Age group Length interval Total number 120-124 125-129 130-134 135-139 140-144 145-149 150-154 155-159 160-164 II 5 16 61 82 28 29 25 8 4 258 III 4 16 53 76 92 67 31 8 347 Table 62. — Muskellunge Lake cisco, 1931 — Frequency distribution by 5-millimeter intervals of the calculated growth during the second year of life as based on the II group and the III group Age group Length interval Total number 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 II 1 8 44 68 58 44 18 13 3 1 258 Ill i 16 35 65 53 69 65 37 14 1 i 347 284 BULLETIN OF THE BUREAU OF FISHERIES Examination of the data shows that during the second year of life of the Ill-group fish there occurred sufficient growth compensation to change the length distribution of the group from the bimodal to the unimodal condition. The existence of intense com- pensatory growth is indicated in the bimodal character of the curve for the second-year growth increments of the III group. A more adequate conception is given of the growth compensation in these two age groups by the study of the correlations between the lengths attained at different times in the individual life histories, and the amounts of growth made by the individuals during different years of life. The correlation was computed for each age group for the following combinations of characters: (1) Calculated length at the end of the first year of life and calculated length at the end of the second year of life; (2) calculated growth in length during the first year of life and calculated increment of growth in length dur- ing the second year of life; (3) calculated length at the end of the first year of life and actual length at the time of capture. The results of these computations are given in table 63. Here L; is the calculated length at the end of the first year ; L2 is the calculated length at the end of the second year; ALi is the calculated growth during the second year; and LT is the actual measured length at the time of capture. Table 63.—Muskellunge Lake cisco, II and III groups of 1931 — Correlations between calculated length at the end of the first year and calculated length at the end of the second year, between the amount of growth during the first year and during the second year, and between the calculated length at the end of the first year and total length at the time of capture Data upon which the correlation is based Coefficient of correlation and its probable error II III Li and L2 0. 513±. 046 —0. 311±- 056 . 404±. 052 0. 708±. 027 -0. 826±. 017 . 399±. 045 Tables 64 to 67 show the data from which the correlations between Lj and L2 and between Li and ALi were calculated. The examination of tables 64 and 65 shows clearly that in both the II group and the III group the smaller fish at the end of the first year of life tend to be the smaller fish at the end of the second year of life, and that this tendency is the greater in the III group. Tables 66 and 67 show that in both age groups the individuals that grew least in the first year of fife tend to grow most in the second, and that here again the tendency is more pronounced in the III group than in the II group. Table 64. — ■ Relationship between calculated length in millimeters at the end of the first and second years of life in the Muskellunge Lake cisco, II group of 1931 AGE AND GROWTH OP THE CISCO 285 Table 65. — Relationship between calculated length in millimeters at the end of the first and second years of life in the Muskellunge Lake cisco, III group of 1931 Table 66. — Relationship between amount of growth in millimeters during the first year of life and the amount of growth during the second year of life in the Muskellunge Lake cisco, II group of 1931 Calculated growth during second year of life Calculated growth during first year of life Total 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 65 to 69 1 1 60 to 64 1 2 3 55 to 59 1 2 4 4 2 13 1 2 1 5 3 3 2 1 18 45 to 49 1 7 9 9 9 6 3 44 40 to 44 6 17 18 5 10 2 58 35 to 39 4 15 27 17 4 1 68 30 to 3C. 3 8 23 7 3 44 25 to 29 1 3 1 3 8 20 to 24 1 1 Total 1 4 22 56 72 60 33 7 3 258 Table 67. — Relationship between amount of growth in millimeters during first year of life and amount of growth during the second year of life in the Muskellunge Lake cisco, III group of 1931 Calculated growth during Calculated g rowth during first year of life Total second year of life 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 75 to 79 1 1 70 to 74 1 1 14 65 to 69 1 3 6 4 60 to 64 9 5 19 9 2 37 55 to 59 l 4 17 23 8 1 1 1 — 55 50 to 54 3 15 20 14 7 8 1 69 45 to 49 1 2 2 10 10 12 8 1 53 40 to 44 1 1 2 8 16 14 17 1 65 35 to 39 3 9 15 2 1 35 30 to 34 1 5 5 4 1 16 25 to 29 1 i Total 5 17 60 64 28 26 37 39 46 16 7 2 347 On the basis of the data that have been presented it is possible to draw the following conclusions concerning growth compensation as it is indicated in the two age groups considered here: (1) Growth compensation occurs in both the group with a unimodal distribution of calculated lengths at the end of the first year of life and the group with a bimodal distribution of these lengths (negative correlation of Li and ALO. (2) Growth compensation is more intense in the group with the greater disper- sion of the calculated lengths at the end of the first year of life (higher negative correlation between Lx and ALi in the III group). The tendency toward compen- sation can lead to a bimodal distribution of the growth increments for the second 286 BULLETIN OF THE BUREAU OF FISHERIES year, and can within a single season change the length distribution of the group from a bimodal to a unimodal condition. (3) When the length distribution at the end of the first year of life is unimodal and the dispersion small, individual length at the end of the first year of life exerts more effect on individual length at the end of the second year of life, than the amount of growth during the first year of life exerts on the amount of growth during the second year of life. (In the II group the positive correlation of Lj and L2 is greater than the negative correlation between Li and AL^) In the group with the greater dispersion and the bimodal distribution of lengths at the end of the first year of life the situation is reversed, that is, there is a closer connection between the amount of growth during the first year of life and the amount of growth during the second year of life than there exists between the length at the end of the first year and the length at the end of the second year. (In the III group the negative correlation between Lx and ALj is greater than the positive correlation between Li and L2.) (4) In spite of the growth compensation that occurs there is a tendency for the individual to hold throughout life a part of any advantage in length which it may hold at the end of the first year of existence. Compensation reduces individual advantage in length, but does not obliterate it (positive correlation in both groups between L! and LT). GROWTH RELATIONSHIPS IN THE TROUT LAKE, MUSKELLUNGE LAKE, SILVER LAKE, AND CLEAR LAKE CISCO POPULATIONS Throughout the preceding sections attention has been called repeatedly to the order in which the cisco populations of the four lakes arrange themselves with respect to certain characteristics such as growth rate, sex ratio, condition, and the like. In view of the apparent high degree of correspondence among certain of these orders of arrangement it is considered advisable to present the data concerning them in a summarized form, together with a brief discussion and review of the possible signifi- cance of the observed correlations. Table 68 shows that arrangement of the lakes with respect to the amount of bound carbon dioxide in their waters and to the amount of organic matter in the plankton, and also with respect to certain phases of the life history of the cisco. Although these data demonstrate a close dependence of certain phases in the life history of the cisco on the conditions of its animate and inanimate environment, as will appear in the following discussion, any attempt to describe these relationships in precise terms of cause and effect meets with serious difficulty. Table 68. — Order of the 1+ lakes with respect to the concentration of bound C02 and the abundance of organic matter in the surface plankton, and also with respect to certain phases of the life history of the cisco Lakes Item Clear Muskel- lunge Silver Trout Growth in length _ _ __ 1 2 3 4 Growth in weight _ __ _ _ 1 3 2 4 Bound COa in water _ . .. __ _ __ 4 3 2 1 Density of population. .... 4 2 3 1 Length of growing season - 1 3 2 4 Sex ratio (females per 100 males)... ... .. . 4 2 3 1 Organic matter in plankton 4 1 3 2 Condition ( K). __ . I 4 2 3 Average length of life. 1 4 2 3 AGE AND GROWTH OF THE CISCO 287 Before entering into the discussion of the relationship between growth rate and the environmental factors that may affect it, attention should be called to the fact that the order of the four lakes with respect to growth in length is not the same as their order with respect to growth in weight. Although the Muskellunge Lake cisco shows better growth in length than the Silver Lake cisco, its growth in weight is inferior to that of the Silver Lake population. The reason for this reversal of order lies in the very poor condition of the Muskellunge Lake fish. Weight un- questionably furnishes a better measure of increase in living matter than does length, but for the purposes of the present discussion it will probably be sufficient to consider both populations merely as intermediate between the extreme conditions represented by the Trout Lake cisco and the Clear Lake cisco. PHYSICAL-CHEMICAL FACTORS Certain factors may be ruled out immediately as inadequate for the observed differences in the growth rates of the four populations. Temperature fails to account for the observed differences in growth rate. Since all four lakes are located within a short distance of each other the climatic conditions that affect each of them are essentially the same. Differences in the size and form of the various basins may lead to differences in average water temperature during the summer, but here it should be pointed out that while the cisco finds its coldest summer habitat in Trout Lake and Clear Lake, the populations of these two lakes represent the extremes in growth rate. Oxygen conditions also fail to explain the differences in the growth rates of the four populations. Here, as was the case with temperature, Trout Lake and Clear Lake resemble each other most closely. Each possesses large masses of well oxy- genated water in the hypolimnion, the favorite habitat of the cisco. The abundance of bound C02 and the closely related hydrogen-ion concentration and conductivity are the only physical-chemical characteristics known to show any correlation with growth rate. Growth rate in length and the abundance of bound carbon dioxide stand in an inverse relationship to each other. It is hardly reasonable to assume that an abundance of bound C02 impedes growth directly or that a scarcity of bound C02 accelerates it. The effect of the abundance of C02 on growth rate is probably indirect and operative through its modification of the biological nature of the cisco’s environment. In general, the amount of bound C02 in a lake’s waters is roughly indicative of the biological productive capacity of that lake. In view of this fact it would hardly be expected that the poorest growth of the cisco would occur in the lake with the greatest concentration of bound C02. This apparently paradoxical situation is explained, however, if it is assumed that an abundance of bound C02 makes not only for a greater production of food organisms, but also makes for a much greater abund- ance of the ciscoes themselves, and that the abundance of the ciscoes in turn de- termines their growth rates. Such an assumption has the support of the observed fact that the cisco population is most dense in Trout Lake with the greatest amount of bound C02 and sparsest in Clear Lake with the least amount of bound C02, while Muskellunge Lake and Silver Lake with intermediate concentration of bound C03 have populations of intermediate densities. 288 BULLETIN OF THE BUREAU OF FISHERIES DENSITY OF POPULATION The section on the relationship between density of population and growth rate inclined toward the opinion that differences in the severity of competition for food related to differences in the densities of the various populations are largely responsible for differences in the growth rates of the four cisco stocks. This opinion was pre- sented as the most logical, even in the face of the facts, first that the actual existence of any competition for food in any one of the four populations remains to be demon- strated, and second that there exists a strong possibility that crowding may impede growth independently of competition for food (Willer’s “Raumfaktor”). As to the causes of the observed differences in the densities of the four populations it may be seen, as was mentioned previously, that the greatest density of the cisco population occurs in the lake with the heaviest concentration of bound C02 (Trout Lake) while the least density occurs in the lake with the lightest concentration of bound C02 (Clear Lake). The positions of Silver Lake and Muskellunge Lake are reversed with respect to the concentration of bound C02 and the density of population, but both are intermediate to the conditions found in Trout Lake and Clear Lake. Since it is generally held that the success or failure of a hatching of fish depends primarily on conditions that determine the survival of the young at a very early stage, it does not appear unreasonable to hold that differences, traceable to the concentration of bound C02, in the amount and land of plankton available to newly hatched ciscoes may account for the observed differences in the densities of the four populations. A study of the plankton cycle in each lake should throw light on the subject. There does not appear to be any complete correlation between the amount of organic matter in surface samples of plankton taken during the summer and the den- sity of the cisco populations. Although Clear Lake with the least amount of organic matter in the surface plankton has the sparsest cisco population, Trout Lake, with the densest cisco population does not have the greatest amount of organic matter in the surface plankton. It is possible further that these differences in the densities of the cisco populations may depend on yet other factors such as the availability of suitable spawning areas or the destruction of eggs and young by predators. Regardless of whether it is held that differences in growth rate depend directly on differences in population density or that both are dependent on yet other factors, it must he admitted that growth rate and density of population show a very close correlation. LENGTH OF THE GROWING SEASON In the general section under this title it was pointed out that differences in the growth rates of the four cisco populations can be explained in part by actual differences in the length of the cisco’s growing season in the various lakes. The Trout Lake cisco has definitely the shortest growing season, followed by the Muskellunge Lake cisco, the Silver Lake cisco, and the Clear Lake cisco in the order named. Thus it may be seen that in the four populations the length of the growing season follows the same order as their growth rates (in weight) and the inverse order of their densities of population. These differences in the length of the growing season cannot be accounted for on the basis of temperature and oxygen conditions for the same reasons that tempera- ture and oxygen conditions fail to account for the differences in growth rate. (See p. 287.) AGE AND GROWTH OF THE CISCO 289 It is possible that an explanation of the differences in the lengths of the growing season may be found in the study of the plankton cycles in the four lakes. Particular attention should be given to the abundance at all times of the season, and in the strata actually inhabited by the cisco, of the plankton forms most commonly taken by that species. A second possible explanation of the differences in the length of the cisco’s growing season in the various lakes is suggested by the fact that in fishes in general the termi- nation of the season’s growth in adult fish is coincident with the onset of the develop- ment of the gonads preliminary to spawning. (The determination of the length of the growing season in the four cisco populations was based almost entirely on mature and maturing individuals.) Since the average size (weight) of the spawning individuals of the four populations follows the same order as their growth rates and the lengths of their growing season, it is suggested that the development of the gonads in small spawners may begin earlier in the season than in large spawners, and that the slowness of growth of a slow-growing population may be thereby accentuated. PARASITIZATION The only published data on the incidence and severity of parasitization in the cisco populations of Trout, Muskellunge, Silver, and Clear Lakes are those presented by Dr. Chancey Juday in the Bureau of Fisheries’ report on Progress in Biological Inquiries 1931 (Higgins, 1932). Juday summarized the data then available as follows: Thirty ciscoes from Silver Lake were examined for parasites, and cestodes were found in the intestines of all of them; 80 percent of them also had Acanthocephala. In Muskellunge Lake 80 percent of the ciscoes contained cestodes and 20 percent were free of visceral parasites. In Trout Lake 16 percent were negative, 82 percent had cestodes in their intestines, and 10 percent also had Acanthocephala. The ciscoes from Clear Lake, on the other hand, were 96 percent negative; only 2 specimens out of 60 examined yielded any parasites. These fish were found to be feeding almost exclusively on Daphnia, and this may be partly responsible for the very small parasite infestation. Clear Lake also has very soft water, and the snail population, as a result, is relatively small, so that the danger of parasite infestation from this source is correspondingly small. Although the above data show that the Clear Lake cisco with the fastest growth rate has the lightest parasite infestation, the relationship between parasite infestation and growth rate in the remaining three populations is not clear. The solution of the problem concerning the relationship between growth rate and parasite infestation in the cisco awaits the examination, within each population, of the effect of individual parasitization on individual growth rate. CONDITION In the section that dealt with condition and the relationship between length and weight it was pointed out that the order of the four lakes with respect to the average condition of their cisco populations, from poorest to best, is: Muskellunge, Trout, Silver, Clear. While this arrangement does describe the order of the lakes with respect to condition as based on the samples taken, it is open to the criticism that because of the variation of K with length, additional samples taken in yet other years and showing different length distributions of the fish might possibly bring about a change in the arrangement. The determination of an average value of K within a population depends, first, on the manner of change of K with length and, second, on the length distribution of the fish used . 290 BULLETIN OF THE BUREAU OF FISHERIES It is quite possible that the most significant phase of the study of condition in these four cisco populations does not he in the determination of average values of # for the different stocks but lies rather in the study of the change of condition with increasing length. This is particularly true since the separation of the lakes with respect to the manner of change of # with increased length is much sharper than it is with respect to the average value of # in the entire population. The significance of the changes of # with length will become more apparent with the examination of 120 150 2 00 250 300 350 400 LENGTH IN MILLIMETERS Figure 10.— Theoretical values (in 1931) of the coefficient of condition (K) at different lengths, calculated from equations of the type K=CX105L“>. Trout Lake, ; Muskellunge Lake, — . — ; Silver Lake (females), — ... — ; Clear Lake (fe- males), . the equations that describe # as functions of length. These equations for 1931 (the only year in which samples were taken from all four lakes) are: 31 Muskellunge Lake: #= 141.924 L~ °-94932 Trout Lake: 7^=38.7640 Z"0-69245 Silver Lake (females): #=0.12322 X0-45372 Clear Lake (females) : #=0.04555 X0'64991 The theoretical values of # at different lengths, calculated from the above equations appear in table 69 and are presented graphically in figure 10. 31 For the purposes of this discussion one equation from each lake is sufficient. The equations for the males of Silver and Clear Lakes may be found in table 27. AGE AND GROWTH OF THE CISCO 291 Table 69. — Theoretical values of the coefficient of condition ( K ) at different body lengths, calculated from equations based on 1981 collections Length in millimeters 130 140 150 100 170 180 190 200 210 220 230 250 270 290 310 330 350 370 390 1. 332 1. 266 1. 207 1. 154 1. 107 1.064 1. 025 0. 989 0. 956 0. 926 0. 898 1.397 1. 302 1. 220 1. 147 1. 083 1.026 .975 .928 .886 .848 .813 1. 122 1. 160 1. 197 1. 232 1. 267 1.300 1.332 1.357 1.394 1.424 1.453 Clear (females) 1. 077 1. 131 1. 182 1.233 1.283 1.331 1.379 1. 426 1. 471 1. 517 1. 5G1 1. 648 1.733 1.815 1.895 1.974 2.051 2. 126 2.200 The remarks, made previously (p. 246), concerning the validity of using length- weight equations for the calculation of unknown lengths or weights outside the range of the empirical data apply likewise to the calculation of unknown values of K for lengths outside the range of empirical data. However, it may be pointed out further, on purely mathematical grounds, that the K equations cannot possibly be used for the calculation of K values in very small fish, since as length approaches zero the values of K increase without limit in the equations with negative exponents and approach zero in the equations with positive exponents. In the 1931 K equations it may be seen that the order of the four populations with respect to the value of the exponent, m, which describes the rate of change of K with change of length, is: Muskcllunge Lake, Trout Lake, Silver Lake, Clear Lake. The differences in these rates of change are reflected in the forms of the curves of figure 10. Here it may be seen that the Muskellunge Lake cisco loses condition rapidly with increase in length. The Trout Lake cisco also loses condition with increase in length, but at a slower rate than does the Muskellunge Lake cisco. In Silver Lake the condition of the cisco improves with increase in length, although this improvement is not as rapid as it is in the Clear Lake cisco. The courses of the curves indicate further that at a length of between 150 and 160 millimeters the conditions of all four populations are closely similar. If the above facts are examined in relation to the growth rates of the four popu- lations it may be seen that, while the Clear Lake cisco with the most rapid growth shows the most rapid progressive improvement in condition, the Trout Lake cisco with the least rapid growth does not show the most rapid loss of condition. Although the Muskellunge Lake cisco shows better growth than the Trout Lake cisco with respect to both length and weight, the loss of condition with increase in length pro- ceeds considerably more rapidly in the former population. This fact demonstrates at least a partial independence between the factors that determine growth rate and the factors that determine condition. Further, the fact that the factors which bring about a rapid loss of condition in one population may fail to reduce the growth rate of this population below that of a second stock with a less rapid loss of condition may be construed as a strong argument for the operation of the “space factor” in the determination of growth rate. The data of table 68 show that the arrangement of the four lakes with respect to average condition and the rate of change of condition follows the reverse order of their arrangement with respect to the average abundance of the organic matter in the surface samples of plankton. (The greater number of these plankton sam- ples were taken during the summer months.) Although the abundance of organic matter in surface samples of plankton may not serve as a wholly reliable index of the abundance of plankton forms most commonly taken by the ciscoes and in the strata of 292 BULLETIN OF THE BUREAU OF FISHERIES water inhabited by that species, the data do show that the cisco suffers the most rapid loss of condition with increase in length and shows the poorest average con- dition in the lake (Muskellunge) with the most eutrophic environment, while the most rapid gain in condition and the best average condition are found in Clear Lake with the most oligotrophic environment. While it is hardly to be inferred that a mere abundance of food causes loss of condition in the cisco or that a scarcity of food makes for better condition, it is quite probable that the cisco does not thrive in the physical and chemical conditions most conducive to a large production of food organisms. The eutrophic environment of Muskellunge Lake may, for example, force the cisco to live under such undesirable conditions of temperature and dis- solved oxygen that it fails to thrive even in the presence of abundant food, while in Clear Lake favorable physical and chemical conditions may make it possible for the cisco to reach the best of condition on a substantially smaller basic abundance of food. Average longevity and condition appear to be more closely correlated than average longevity and growth, for the four populations follow exactly the same order with respect to the two first-named characteristics. Such a relationship is logical. With particular reference to the populations whose K equations show negative ex- ponents, it appears probable that the progressive loss of condition with increase in length may bring the cisco to a point of emaciation beyond which survival is im- possible. Since length is a function of age the imposition of a limit on the length that can be attained places also a limit on age. It has been pointed out previously (p. 246) that in the Trout Lake cisco the few very old individuals taken had not followed the same length-weight relationship that held for the main body of the population. The relationship between individual condition and individual parasite infesta- tion is not known. The Clear Lake cisco with the lightest parasite infestation of any of the four populations is the population with the best condition. SEX RATIO The cisco population of the four lakes follow the same order with respect to growth in weight and the average relative abundance of males. In the section on sex ratio it was pointed out that the less viable males probably suffer greater mor- tality under adverse conditions that produce slow growth, and hence that the cor- relation between growth rate and sex ratio may be considered to result from the dependence of these two characteristics on the same environmental factors. GENERAL REMARKS The failure of certain of the growth relationships of the cisco to conform to generally accepted principles and theories suggests that, on the basis of our present knowledge, generalizations concerning these relationships are scarcely safe. A satis- factory understanding of the growth relationships of the cisco can be attained only after exhaustive studies into the biology of this form. Further, the great plasticity and adaptability of the cisco makes advisable the study of its biology in the greatest possible diversity of habitats. An illustration of the dangers of making generalizations concerning the growth and biology of the cisco is offered by the examination of the relationship between AGE AND GROWTH OF THE CISCO 293 condition ( K ) and environmental factors. The data of this study indicate that the cisco is in the best condition in the most oligotrophic environment and in the poorest condition in the most eutrophic environment. However, the situation found in the Indian Village Lake (Indiana) shows that such a relationship cannot he considered general for this plastic species. Although the lakes of northeastern Indiana are of the extreme eutrophic type and the Indian Village Lake cisco lives under what are apparently the most undesirable conditions with respect to temperature and oxygen conditions (Scott, 1931), the Indian Village Lake cisco shows excellent growth and is in better condition than any but the Clear Lake population of this study (p. 249). FISHES ASSOCIATED WITH THE CISCO There is little resemblance among the fish associations of which the four cisco populations are parts. In table 70 are recorded the numbers of individuals of all species taken in the nets used for the cisco. Fish are considered to have been taken along with the cisco if they were taken in a net of any mesh set at the same time and approximately the same depth as the nets that caught ciscoes. Table 70. — Numbers of individuals of other species of fish taken in the net catches along with the cisco samples Lakes Species Trout Muskel- lunge Silver Clear Cisco. j 1, 197 1, 863 1, 543 524 465 Perch . i 5 Pike-perch __ 37 Rock bass ... 8 1 Sucker 130 Lake trout 32 Whitefish ... ... 32 Burbot 1 Smallmouth black bass _ _ 14 In Trout Lake the fish taken with the cisco were^all typical deep-water forms. There is no evidence that the deeper regions of the lake are invaded by individuals of the shallow-water forms. The trout were mostly large individuals that were caught and held by the teeth in small mesh nets. The situation in Muskellunge Lake is unusual. The close association of the cisco with shallow-water forms is the result of the deficiency in oxygen that drives the ciscoes up from the lower, cooler strata of the lake. In Silver Lake the cisco is practically isolated, at least during the late summer. The single perch taken at a depth of 14.5 meters may be considered a straggler. In Clear Lake the pike-perch that were taken regularly along with ciscoes appear to be distributed generally throughout the region occupied by the cisco in this lake. The five perch were taken at depths between 19.5 and 24.5 meters. The rock bass was captured at a depth of 15.5 meters. Since both rock bass and perch are known to be plentiful in the shallow water of this lake their presence in the hypolimnion may be considered more or less as accidental. In all the lakes the nature of the associations of which the ciscoes are part prob- ably undergoes considerable change according to the time of year. 294 BULLETIN OF THE BUREAU OF FISHERIES SELECTIVE ACTION OF GILL NETS REVIEW OF PREVIOUS INVESTIGATIONS Since the collections upon which the present investigation is based were made by means of gill nets, it is of importance to have some measure of the selective action of the nets used upon the various populations sampled. Several investigators have presented observations on the selective action of gill nets, and some of these observa- tions have been made with specific reference to the problem of securing samples for the study of age, growth, and length distributions within a population. Much of the published material on the selective action of gill nets has been included incidentally in growth and life history studies. Other data have been presented in routine fishery reports. Because of the generally scattered nature of these data on gill- net selectivity and because of the importance to fisheries biologists of having a more adequate idea of the reliability of the samples that they take with gill nets, the presenta- tion of the data of this investigation on the question of gill-net selectivity will be preceded by a brief review of the literature 32 on the subject. The problem presented by the selective action of gill nets in the collection of materials for scientific investigations has been known for several decades. The Report on the Sea and Inland Fisheries of Ireland for 1902 and 1903 33 ; Part II — Scientific Investigations, contained in the report of the scientific adviser the statement that the investigation of herring shoals met a great difficulty in the selective action of the nets used in the commercial fishery. It was stated further, however, that the results of fishing nets of different meshes together (these experiments were made in another con- nection) indicated “ that the selection is much less in practice than it would seem to be in theory. * * *” Delsman (1914) in his study of the age and growth of the North Sea and Zuider Zee herrings presented data to show that the selective action of the gear used in obtain- ing samples can exert an important effect on the nature of the results obtained from the study of those samples. He compared samples of drift nets and of seines of differ- ent sizes of mesh. From his observations he concluded: Fange an derselben Stelle mit Netzen verschiedener Maschenweite gemacht, werden verschieden sein und die Zusammensetzung des Fanges nach Grosse und Alter wird von der Maschenweite abhan- gen. Fischt man dagegen mit demselben Netz an verschiedenen Stellen * * * so werden auch die Fange verschieden sein, und die Verschiedenheiten in ihrer Zusammensetzung werden durch die Zusammensetzung der Heringsschwarme im Meer bedingt werden. Bjerkan (1917) disagreed with Delsman as to the great importance of selectivity of gear in determining the nature of herring samples from drift nets. From the com- parison of trawl and drift-net catches he concluded that selection through the failure of drift nets to take smaller fish “cannot be very material.” He pointed out further that “drift caught samples may point to the presence of rich year classes of very different ages.” As to the effect of mesh size on the composition of drift net catches he stated further that “the size of mesh in the nets used affects the composition of the catches, but not to such an extent as might have been expected.” Further data on the action of drift nets were presented by Borley and Russell (1922). In connection with the study of the herring trawl fishery they measured the 33 No claims are made as to the completeness of this review; it should, however, bring together a sufficient mass of the miscel- laneous data on the subject to furnish a good general conception of the nature of the problem of gill-net selectivity. »> Published in 1905. AGE AND GROWTH OF THE CISCO 295 catches of herring in drift nets of three sizes of mesh.34 They summarized their results as follows: According to these samples the finer-meshed nets caught the larger fish. The conclusion to be drawn is that the size distribution of the drift-net catch is probably not determined to any great extent by the selective action of the net. Buchanan-Wollaston (1927) included with his discussion of the selective action of a trawl net certain observations as to the method of selection in drift nets. He called attention to the fact that drift nets do not capture all fish in the same manner. He pointed out that certain large individuals are retained by the maxillary barbs; some small individuals are held only because of the gill covers; while the majority are held between the gill covers and the dorsal fin. The different types of net action may lead to discontinuity in the length distribution of the catch. Hodgson (1927) presented the analysis of data from the catch of herrings in three drift nets of different-size mesh.35 Although the number of specimens was not great his results showed clearly that selective net action can lead to erroneous conclusions concerning the size and age composition of a population and also lead to inaccurate determinations of average lengths for the various age groups. Later the same author (Hodgson, 1933) presented the results of further experiments on the selective action of drift nets. These experiments were conducted with particular reference to the effect on the herring stock of the size of mesh in the drift nets em- ployed in the commercial fishery. On the basis of two separate experiments Hodgson concluded: that there is a very definite and subtle gradation in the length of the herrings caught by nets of even slightly different meshes, and it is also plain that the whole character of a fishery can be changed by the use of different nets. These investigations were all concerned with a single marine fishery — the drift- net fishery for herring. The opinions of the different investigators, however, show considerable disagreement. Of greater interest with respect to the present study are the data concerning the selective action of gill nets as they are fished in fresh water. In the following paragraphs certain of the available data with respect to the selective action of gill nets are mentioned briefly. The various investigations are treated in general in the chronological order of their appearance without any attempt at grouping according to variety or kind of fish concerned. Jarvi (1920) compared the catches of the kleine Marane ( Coregonus albula ) taken by seine with those taken by a series of seven nets having meshes ranging from 10 to 15 millimeters (bar measurement). The length compositions of the samples from the two sources agreed well. Jarnefelt (1921) noticed that his growth curves for several species of fish showed irregularities that could be traced to the selective action of the nets that he used in obtaining his samples. Concerning these irregularities Jarnefelt stated: “* * * dass es ebenso viel Senkungen der Kurve gibt wie Netze verschiedener Maschenweite benutzt werden.” However, Segerstr&le (1933) held that the irregu- larities observed by Jarnefelt and others in growth curves resulted to a large extent from inaccurate methods of calculating growth from scale measurements, and that 88 The sizes were 30, 31, and 33 meshes to the yard (bar measurement). 83 The nets used had 35, 38, and 48 meshes to the yard (bar measurement). 296 BULLETIN OF THE BUREAU OF FISHERIES irregularities in growth curves resulting from the selective action of nets are on the whole small. Creaser (1926) in connection with his study of the growth of the sunfish ( Eupo - motis gibbosus) observed that gill nets are “particularly selective in their collecting” and advocated the use of a wide range of mesh size in gathering material for biological investigation, but included no data on the selective action of gill nets. Koelz (1926) pointed out that in the Lake Ontario chub {Leucichthys spp.) fishery an increase of a quarter inch in the stretched measure of the mesh (from 2% to 2% inches) reduced the number of fish taken by more than half. Pritchard (1928) presented data on the selective action of six different sizes of mesh as that action affected the individual weight of Lake Ontario chubs {Leuci- chthys spp.). (The mesh size of his gill nets ranged from 1 % to 3 inches, stretched measure.) On the basis of his observations he stated, “The difference of one-quarter of an inch in the size of mesh may mean to the fishermen either a profitable or a ‘starvation’ industry.” In a later paper (Pritchard, 1931) he presented the numbers of chubs of all species taken in nets of 12 different sizes of mesh (1% to 5 inches). Some data were included on the average lengths of the fish taken in different mesh nets. The data of Pritchard’s 1931 paper agree with his earlier findings in indicating that a very small increase in the size of mesh may produce a great decrease in the number of fish taken. Hart (1928) published detailed data on the length distribution of pike-perch and saugers of Lake Nipigon and Lake Abitibi taken in 6 different sizes of mesh ranging from 1 % to 4% inches, stretched measure. Although Hart confined his discussion to the question of the proper legal mesh size in the commercial fishery, his data on the length distribution of the catches in different mesh size show clearly that a single size of mesh can take these spiny-rayed fish over a considerable length range. In a later paper (Hart, 1931) the same author pointed out that observed differences in the growth rate of the whitefish {Coregonus clupeaformis ) in different parts of Lake Ontario could be explained in part by the tendency for the gill nets used in collecting samples from some localities to take only the larger fish of the younger age groups while pound-net samples were largely free from such selection. A more detailed consideration of the selective action of gill nets on whitefish appeared in Hart’s (1932) study of the population statistics of Shakespeare Island Lake. (His experimental gear included 11 sizes of mesh, 1% to 5 inches.) Hart found a correlation of 0.84 ±0.01 between size of mesh and size of fish in Shakespeare Island Lake, while in Nipigon the correlation between the same variables was 0.51 ±0.02. He pointed out further that “large fish may be taken in small meshed nets although small fish are practically never taken in nets of coarse mesh.” Lechler (1929), referring to the selectivity of gill nets used in the fishery for the Reinanke {Coregonus jera) held that the selective action of the net is quite sharp. He stated: “Die Zusammensetzung der Fange nach den Jahresklassen und damit die Grosse der Fische ist eine Funktion des Umfangs und direkt von der Netzmas- chenweite abhangig. ” Further experiences caused him to modify this view slightly, for in a later publication (Lechler, 1930) he observed, “Die Fangigkeit der Netze andert sich je nach der Beschaffenheit eines Bestandes.” Wright (1929) eliminated the younger age groups from his growth data on the rock bass {Ambloplites rupestris ) as the selective action of the nets caused these early groups to have too high average values for length. AGE AND GROWTH OF THE CISCO 297 Haakh (1929) in his study of the age and growth of several species of fisli in the Bodensee found that net selection led to inaccurate determinations of the average length of some age groups. Wagler (1927, 1930a, 1930b, 1932, 1934) in his investigations of the coregonids of Bodensee and other north alpine lakes mentioned repeatedly the distorting effect of the selective action of the gill nets used in taking his samples for the study of age and growth. The effects of the selection were noticeable particularly in the high average lengths determined for the younger, smaller age groups and in the great variation in the relative abundance of the different year classes as determined from the various gill nets used. Wagler held that a close relationship exists between net selection and maximum girth of the individual fish (cf. Lechler, 1929). Elster (1934) in his study of the Blaufelchen ( Coregonus wartmanni ) expressed opinions similar to those of Wagler. Van Oosten (1929b) in his discussion of the problems of the commercial fisheries of the Great Lakes pointed out the complex nature of the question of selectivity in the gill nets of the commercial fishery and included a small amount of data on the number and size of small trout ( Cristovomer namaycush) and several species of chubs (. Leucichthys spp.) taken in 2/- and 2%-inch mesh nets. In a later paper Van Oosten (1933) published a preliminary report of experimental chub net investigations conducted by the United States Bureau of Fisheries in Lake Michigan waters, 1930-32. The data presented included average catch per net, expressed in pounds and number of fish, for the lake trout and the chub in five different sizes of mesh. The meshes used were all of commercial fishery size, and varied from 2% to 3 inches, stretched measure. The variation in the catch of chubs in nets of different size mesh was striking. In lower Lake Michigan nets of 2%-inch mesh caught more than twice as many chubs as did nets of 2 /(-inch mesh, and eight times as many as nets of 3-inch mesh. In view of the marked disagreement that the preceding review shows to exist among some authors as to the nature and scope of the selective action of gill nets, it appears probable that each species and each locality offers its own special problem of gear selection. The selective action of gear as size of mesh affects the amount of total catch (with respect both to numbers and weight) is particularly noteworthy. The existence of an intensive commercial fishery employing a more or less standard size mesh can, how- ever, exaggerate the differences of catch between two different sizes of mesh. Ordi- narily all the commercial gear of the same type used in a region for the capture of a species or group of species of similar size will meet approximately the same specifica- tions particularly with reference to the size of mesh. Consequently there tends to be a great reduction in numbers in that portion of the population most liable to capture by the standard commercial gear. The withdrawal of large quantities of commercial size fish produces an abnormal condition in size distribution of the population as a whole. The introduction of nets of other, noncommercial sizes of mesh may then lead to comparisons that do not describe accurately the fishing action of the nets in question. Data based on catches from a population unaffected by commercial fishing might be expected to show somewhat less sharp selectivity than those from a population subjected to heavy fishing. It will be seen that the data of the present investigation tend to support this view. 298 BULLETIN OF THE BUREAU OF FISHERIES SELECTIVE ACTION OF GILL NETS USED IN COLLECTING SAMPLES OF CISCOES FROM TROUT, MUSKELLUNGE, SILVER, AND CLEAR LAKES In this investigation, although other phases of selectivity will be considered briefly, the study of the selective action of gill nets will be concerned primarily with the effect of selection upon the determination of growth in the various populations. The analysis will involve particularly the comparison of average lengths as deter- mined for a single age group from samples taken in nets of different sizes of mesh. The analyses will show the basis for the elimination of certain selected groups in the computation of general growth curves (see footnote, p. 226) and will furnish a better idea of the validity of the use of gill-net samples in the study of growth. Such analysis is of especial importance here since different ranges of mesh sizes were employed in taking the different years’ collections. Attention should be called to certain criteria that are of general value in deter- mining the reliability of gill-net samples for growth study. It may be pointed out that in general a sparse representation in a sample of a young age group whose average length is near the lower limit of effectiveness of the nets used, is a source of suspicion as to the reliability of the sample of that particular group. If this same sparsely represented group gives calculated growths that are in serious disagreement with those of the older age groups it should be eliminated from the data used for the study of growth in the population as a whole. A further check on the reliability of the material that represents a particular age group lies in the comparison of the average lengths for the group as determined from samples in different sizes of mesh and as determined from the different combinations of such samples. The examination of the length distributions of the catch of nets of different size mesh may also give an idea as to the adequacy of the sampling with respect to particular age groups. As was mentioned in the introduction the 1928 data do not include records of gear for individual fish. For this reason the 1928 data cannot be used in the study of gill-net selectivity. For purposes of convenience the gear used in 1928, 1930, and the early part of the 1931 season will be designated throughout this section as "old”,36 and the gear used during the latter part of the 1931 season and through the 1932 season will be known as “new.” Descriptions of the two groups of nets were included in the section on methods. Tables 71, 72, 74, 75, 76, and 77 show the effect of size of mesh on the deter- mination of average size of the different age groups in the populations of Trout, Muskellunge, Silver, and Clear Lakes. Because of the generally small number of specimens upon which the average of lengths of the various age groups depend, attention must be given to the general trend of the differences that result from differences in gear rather than to the actual amount of the difference in any specific case. Since each population represents a distinct and separate problem in the study of selectivity, the data for each lake will be discussed separately. TROUT LAKE With the exception of a single specimen all the ciscoes collected in Trout Lake in 1930 were obtained from 1 %-inch mesh nets, while the 1928 samples were probably taken in nets of 1%-inch (trammel), 1%-inch, l^-inch, and perhaps a few in 1 %- and m Although the “old” gear as fished in 1930 and in the early part of 1931 included 5 different sizes of mesh, only the lH-inch and 2-inch nets caught ciscoes. All collections from Clear Lake were made with the new gear. The date of change in gear was July 22, 1931. AGE AND GROWTH OF THE CISCO 299 2-inch mesh nets. In 1928 the average length of 182 ciscoes was 138 millimeters, and in 1930 the average length of 490 individuals was 150 millimeters. The smaller average size of the fish of the 1928 collection may be accounted for by the use of mesh sizes smaller than 1% inches and by the relatively greater abundance of young fish in that year’s samples. The relative abundance of the different age groups as well as the average lengths of these age groups in the samples of the 2 years may be found in table 3. Table 71. — Effect of size of mesh of gill nets on the determination of the average lengths of the age groups of the Trout Lake cisco, 1931 collection [The first 2 rows show the average lengths of 5 age groups as based on samples taken July 22 in lli- and 1^-inch mesh gill nets. The third row shows the average of the 2. The fourth row shows the average lengths of the various age groups based on all fish taken in l^-inch mesh gill nets in 1931. Sexes combined. Number of specimens in parentheses] Size of mesh Age group Average length I II III IV V 1Y\ inches __ - 124 (1) 134 (22) 137 (16) 139 (14) 143 (20) 147 (18) 149 (15) 155 (2) 158 (4) 140 (61) 145 (58) 1^2 inches - ___ Averages combined 124 (1) 132 (1) 135 (38) 137 (39) 141 (34) 142 (159) 148 (33) 148 (246) 157 (6) 156 (77) Average of all 1931 lH-inch samples 147 (522) A single set of the “new” gear lifted July 22, 1931, provides the only data for Trout Lake for the comparison of growth as based on samples taken in nets with different size mesh. These data appear in table 71. In the first row appear the aver- age lengths for the various age groups as based on fish taken in the 1%-inch mesh net. The second row gives the same information for the sample from the 1^-inch mesh net. It will be seen that while the average length of all ciscoes taken in the 1 J^-inch mesh net was 5 millimeters greater than that of the ciscoes from the 1 %-inch mesh net, the differences, between samples of the same age group varied from 2 to 4 millimeters and averaged only about 3 millimeters. The third row shows the average lengths of the different age groups as based on the combination of the July 22 samples taken in the 1 /- and lK-inch mesh nets while the bottom row shows for comparison the average lengths of the various age groups as based on all the ciscoes taken in Trout Lake in the 1931 season in 1 /2-inch mesh nets. It may be seen that with the exception of the I group the averages based on the l}£-inch mesh net samples for the entire, season differ but little from those based on the combined catches from l}i- and lb-inch mesh nets. While the necessity for the elimination of the I group from the Trout Lake growth data is at once apparent, the problem presented by the II group requires further consideration. In spite of the rather close agreement between the average length of the II-group fish as determined from samples from nets of l%- and lK-inch mesh, the sparse representation of this age group in both the 1930 and 1931 collections (table 3) throws suspicion on the reliability of the samples. This suspicion is sup- ported by the high calculated lengths at the end of the first and second years of life as based on both the 1930 and 1931 II groups. Consequently both of the groups were eliminated from the Trout Lake growth data. The II group of 1928, for which year no individual net records are available, was, however, retained, first because of its abundant representation in that year’s collection, second because of the good agreement between its calculated growth and the corresponding calculated growth both of older age groups of the same year’s collection and of samples of the same (1926) year class taken in the later years, 1930 and 1931, and finally because of the 300 BULLETIN OF THE BUREAU OF FISHERIES knowledge that smaller mesh sizes were used in 1928 than in 1930 and the early part of 1931. The relationship of net size to fish size (new nets, 1931) in the Trout Lake cisco is shown in figure 1 1 . MUSKELLUNGE LAKE The old gear was used in the collection of ciscoes from Muskellunge Lake in 1928, 1930, and 1931. In the last 2 years all the fish were taken in the 1 %-incli mesh net. In 1928 the gear used included the following sizes of mesh: 1%-, 1%-, 1%-, and 2-inch. The 2-inch mesh was probably not effective. Data on the average lengths of the various age groups in these years’ collections will be found in table 4. In 1932 350 300 100 50 0 SIZE OF MESH Figure ll.— Relationship between size of mesh of gill nets (stretched measure, in inches) and the average standard length in millimeters of the ciscoes taken. Trout Lake, ; Muskellunge Lake, — . — ; Silver Lake, — . . . — ; Clear Lake, . all the specimens were taken in three (1%-, 1)2-, and 1%-inch) of the seven sizes of mesh used in the new gear. The 1932 material provides the only data for the study of the effect of mesh size on the average size of fish taken and for the comparison of the average lengths of the various age groups as based on samples from nets of different sizes of mesh. Scale examinations were made for 201 specimens of the 1932 collection, but ages could be estimated for only 189 of this number. The results of these determinations appear in table 72. 37 The table shows considerable differences between the nets of 27 These data represent all the fish taken in the 1%-inch mesh net in 1932 and lifts of the 1)4- and lH-inch mesh nets on July 28, 29, and 30, and Aug. 3, 1932. The numerical representations of the IK- and lK-inch mesh nets are comparable, while that for the lK-inch net is by comparison too high. This high representation of the lK-inch net samples has probably caused the grand aver- age length for the IV group to be too high, but it had little effect on the grand average length for the III group. AGE AND GROWTH OF THE CISCO 301 different mesh size both in the average size of the fish and in the relative abundance of the different age groups.38 Selectivity according to length appears to be greater in the Muskellunge Lake cisco, where the 1%-inch mesh net caught fish 14 millimeters longer than the l^-inch mesh sample, than in the Trout Lake cisco (table 71) where the fish from the 1%-inch mesh net were only 5 millimeters longer than those from the 1%-incli mesh net. It will be noticed further that nets of the same size mesh caught larger fish in Muskellunge Lake than in Trout Lake (fig. 11). Table 72 . — Effect of size of mesh of gill nets on the determination of the average lengths of the age groups of the Muskellunge Lake cisco, 1932 collection (Sexes combined. Number of specimens in parentheses] Size of mesh Age group Average length I II III IV V 137 (12) 150 (3) 152 (7) 158 (12) 159 (11) 162 (59) 166 (5) 168 (3) 168 (59) 177 (12) 150 (33) 164 (150) 174 (18) XVi inches . 172 (1) 139 (15) 156 (24) 162 (75) 170 (74) 172 (1) The question of the reliability of the 1930 and 1931 collections is largely a question of the reliability of samples from ljLinch mesh nets, for the collections of these 2 years were taken in nets of this size mesh. The 1928 samples may be considered as reliable as the 1932 samples. The data of table 72 show that in spite of the selection that does occur the 1%-inch mesh net samples, with the exception of the I group, give averages for the lengths of the different age groups that differ insignificantly from the corresponding averages based on the catch of nets of three different sizes of mesh. It is, therefore, a safe conclusion that with the elimination of the I groups the length data obtained from 1^-inch nets in the preceding seasons are accurate within the range of a very few millimeters. Table 73. — Length frequencies of Muskellunge Lake ciscoes taken in 1932 in nets of different mesh size [Sexes combined. The frequencies represent 18 lifts for the l!4-inch mesh net, 38 for the lH-ineh mesh net, and 37 for the Hi-inch mesh net] Length Hi-inch net 1 H-inch net Hi-inch net Total Length 1)4 -inch net lH-inch net Hi-inch net Total 185 to 189 2 1 3 145 to 149 18 48 66 180 to 184__ 6 3 9 140 to 144 24 20 44 175 to 179 27 4 31 135 to 139... 40 4 44 170 to 174 2 102 6 110 130 to 134 . 7 7 1fi.fi In IfiQ 5 2 160 to 164 13 186 1 200 154 813 18 985 150 to 154 21 91 112 Average length.. 146 160 174 The above conclusion is supported by the examination of the length frequencies of the total catch of Muskellunge Lake ciscoes in each mesh size in the entire 1932 season (table 73). In the length frequencies of the 1%-inch net samples there occurs at the 145 to 149 millimeter interval a depression that marks the separation of the I 38 It was pointed out previously (p. 219), that because of slow growth in the later years and the common occurrence of accessory annuli the separation of the later age groups above the II group in the 1932 Muskellunge Lake samples was difficult. As a conse- quence, the average lengths listed for the Ill-group and the IV-group samples may be slightly in error. Any error that exists should not, however, impair greatly the general usefulness of the data for the purpose of studying the effect of net selection on the deter- mination of the average lengths of age groups. 302 BULLETIN OF THE BUREAU OF FISHERIES group from fish of greater age. Since the bulk of the II-group fish must lie above this depression it can be seen that their length distribution lies well within the range of efficiency of the 1%-inch mesh net. The fish taken in the 1%-ineh net are so few in number that they would have little effect in the determination of growth. SILVER LAKE The “old” gear was used in the collection of all the 1928 and 1930 samples of ciscoes from Silver Lake and for the first sample taken in 1931 (July 17). A gang of the “new” nets was used in taking the only other sample of this year in this lake (Aug. 22, 1931). In the sets of the old gear in 1930 and 1931 ciscoes were taken in 1 y2- and 2-inch mesh nets only; in 1928 they were probably taken in the and 1 74- inch meshes also. In the new gear they occurred in the l}i~, 1 1%-, and 2-inch mesh nets. Table 74 shows the average length of fish taken in different mesh sizes and the effect of net selectivity on the determination of the average length of the various age groups of the 1930 collection (5 lifts of 1^-inch nets, 3 of 2-inch nets) while table 75 gives similar information for the 2 lifts made in 1931. Table 74. — Effect of size of mesh of gill nets on the determination of the average lengths of the age groups of the Silver Lake cisco, 1930 collection [Sexes combined. Number of specimens in parentheses] Size of mesh Age group Average length II III IV V VI 173 (7) 178 (16) 186 (9) 182 (34) 188 (24) 190 (4) 193 (20) 198 (1) 196 (2) 186 (66) 190 (64) 173 (7) 181 (25) 183 (58) 193 (25) 197 (3) Table 75. — Effect of size of mesh of gill nets on the determination of the average lengths of the age groups of the Silver Lake cisco, 1931 collection [The data in the upper part of the table are based on a lift of the “old” nets, July 17. The data in the lower part are based on a lift of the ‘ ‘ new ” gear, Aug. 22. The grand average for the year includes a number of preserved fish for which no net records were available. Sexes combined. Number of specimens in parentheses] Size of mesh ltt inches - 2 inches — Average — Hi inches - US inches l$i inches 2 inches Average Grand average for 1931. Age group Average length I II III IV V VI VII 143 (6) 168 (5) 174 (3) 189 (3) 173 (3) 187 (5) 163 (18) 191 (29) 192 (14) 198 (5) 143 (6) 168 (5) 182 (6) 182 (8) 192 (14) 198 (5) 141 (51) 146 (6) 172 (3) 170 (6) 174 (2) 175 (3) 176 (4) 178 (22) 174 (1) 172 (1) 173 (6) 182 (25) 184 (12) 195 (2) 187 (7) 186 (23) 190 (18) 193 (4) 192 (2) 201 (1) 146 (60) 171 (30) 183 (83) 188 (37) 142 (57) 171 (11) 177 (30) 181 (44) 188 (50) 193 (6) 201 (1) 142 (66) 171 (19) 177 (61) 183 (102) 188 (108) 194 (21) 201 (1) The data in tables 74 and 75 show rather large differences in the average size of fish taken in nets of different sizes of mesh. Although the fish taken in the 2-inch- mesh net of the old gear in 1930 were only 4 millimeters longer than those taken in the 1%-inch net, the 1931 samples taken in the same gear showed a difference of 28 millimeters, while the 2-inch-mesh net fish of the 1931 collection in the new gear had an average length 17 millimeters above the average for those from the 1%-inch net. A comparison with the Trout Lake and Muskellunge Lake data (tables 71 and 72, fig. 11) shows that the selection of the nets is sharper in the Silver Lake cisco than AGE AND GROWTH OF THE CISCO 303 in either of the former populations. The difference in average length between fish taken in l}i- and 1 ^-inch-mesh nets was for example 5 millimeters in Trout Lake, 14 millimeters in Muskellunge Lake, and 25 millimeters in Silver Lake. Further, the difference in average length between fish taken in l%- and 1%-inch-mesh nets in Muskellunge Lake was 10 millimeters while in the Silver Lake cisco this difference was 12 millimeters. Tables 74 and 75 show also that the average length for a particular age group varies considerably in samples taken by nets of different size mesh. Although cer- tain irregularities occur (as might be expected from the small size of the samples) there is a general tendency for the samples from larger mesh nets to show distinctly higher average lengths for a given age group than do the samples from nets of smaller mesh. Further examination shows, however, that this tendency does not in general affect the validity of the averages of age groups based on the combination of the samples of several nets. In the analysis of the data of tables 74 and 75 to establish the validity of the growth data for the Silver Lake cisco, it should be pointed out first that all ages above the II group in the 1931 sample taken in the new nets must be considered to have adequate representation throughout their entire length range. This is apparent from the fact that these older age groups are relatively scarce in the 1 ^-inch-net catch, while at the other extreme a 2%-inch-mesh net that was fished in the string failed to take any ciscoes at all. Since these average lengths of the older age groups from the new net sample can be considered reliable they can be used as a basis for the estima- tion of the reliability of the corresponding average lengths as determined from the old net samples. The fact that the average lengths for corresponding age groups of the 1930 and 1931 old net collections resemble each other on the whole more than either of them resembles the average lengths based on the 1931 new net collection suggests that the addition of the 1%- and 1%-inch nets may have increased the relia- bility of the Silver Lake samples. The differences between the corresponding aver- age lengths of the older age groups of the old and new net samples are not, however, of sufficient magnitude to invalidate the use of the earlier materials collected by the old gear. The II-group samples present a more difficult problem. The general scarcity of II-group individuals in the collections of the old gear would indicate that its length range may lie just below the range of efficiency of the lK-inch net, even though this size mesh was very efficient in Trout and Muskellunge Lakes in the taking of fish whose lengths were well below the observed average length (about 170 millimeters) for the Silver Lake II group. The 1931 sample with the new gear shows, however, that it is very unlikely that the Silver Lake II group has suffered serious selection by gear, as that selection affects average length, in any of the samples. The scarcity of the II-group fish in this sample taken in the new gear must be considered to show that this group in 1931 actually was less abundant than the neighboring age groups, for there is no reason to hold that the same nets that took 57 I-group and 30 Ill-group fish should fail to take II-group individuals in equal numbers provided they were equally abundant in the population. These facts together with the good agreement between the average lengths of the II-group samples as based on old and new gear samples make the retention of the II-group fish in the growth data advisable. Since there is no means of demonstrating the reliability of any of the I-group samples they must be considered, at the best, questionable. They were accordingly eliminated from the growth data. 304 BULLETIN OF THE BUREAU OF FISHERIES CLEAR LAKE All Clear Lake specimens were collected with the “new” nets. Tables 76 and 77 show the average lengths of the different age groups according to the gear by which they were captured. Because of differences in growth rate the sexes are treated separately. Table 78 and figure 11 present the length frequencies and the average lengtn of the catch according to mesh size for the total 1931 and 1932 catch of ciscoes from Clear Lake. The Clear Lake data differ from those of the three populations just considered in two important respects. First, the relationship that increase in size of mesh is correlated with an increase in the average size of the fish taken does not hold for Clear Lake as it did in the other three lakes; while the differences in the aver- age size of the fish taken in different meshes are great, the larger mesh does not always take the larger fish (table 78, fig. 11). Second, in Clear Lake, fish were taken in all seven of the sizes of mesh used, and were most abundant in nets that were totally ineffective in all the other three populations, that is, nets whose mesh size was greater than 2 inches. The circumstance that the largest mesh (3 inches) took the greatest number of fish indicates that in making these collections it would have been desirable to use additional nets of mesh size greater than 3 inches. There is, however, reason to believe that the selection resulting from failure to fish with larger mesh nets affects chiefly the relative abundance of the larger fish in the sample. The increase in size of mesh from 2% to 3 inches had little effect on the average length of the individuals captured. If this trend were to continue, nets of a mesh size greater than 3 inches would be expected to add to the number of large fish without producing any important upward extension of the size range of the sample. Table 76. — Effect of size of mesh of gill nets on the determination of the average lengths of the age groups of the Clear Lake cisco, 1931-32 Collections [(Males) Number of specimens in parentheses] Size of mesh Age group Average length I II III IV V VI VII VIII IX 173 (12) 176 (24) 190 (2) 311 (1) 336 (3) 212 (16) 185 (28) 275 (14) 296 (35) 281 (49) 307 (42) 236 (3) 245 (3) 250 (12) 256 (25) 267 (5) 254 (1) 291 (4) 286 (5) 289 (12) 291 (18) 307 (3) 328 (4) 338 (2) 310 (1) 326 (2) 328 (2) 324 (3) 341 (1) 347 (3) 334 (2) 344 (4) 2J4 inches. 2 y2 inches 314 (4) 319 (6) 324 (4) 333 (5) 329 (1) 334 (6) 355 (1) Average 176 (38) 254 (48) 289 (40) 318 (15) 323 (9) 324 (8) 334 (15) 343 (10) 355 (1) Table 77. — Effect of size of mesh of gill nets on the determination of the average lengths of the age groups of the Clear Lake cisco, 1931-32 collections [(Females) Number of specimens in parentheses] Size of mesh Age group Average length I II III IV V VI VII VIII IX X XI 1% inches 1)4 inches 1% inches 155 (1) 179 (5) 180 (12) 186 (7) 201 (1) 336 (1) 350 (1) 357 (1) 360 (3) 360 (2) 351 (7) 358 (7) 246 (2) 245 (8) 227 (17) 268 (22) 316 (29) 315 (54) 319 (65) 351 (1) 364 (1) 279 (1) 263 (7) 259 (4) 265 (14) 260 (14) 356 (3) 336 (1) 351 (9) 357 (6) 350 (11) 301 (1) 300 (9) 308 (14) 310 (15) 325 (1) 323 (3) 332 (5) 332 (8) 338 (1) 376 (1) 2 inches 371 (1)1 342 (4) 334 (4) 335 (3) 347 (4) 378 (1) 352 (1) 379 (1) Average-, 181 (26) 263 (40) 307 (39) 330 (17) 341 (8) 342 (9) 352 (30) 355 (22) 360 (3) 372 (2) 378 (1) AGE AND GROWTH OF THE CISCO 305 Table 78.- — Length frequencies of Clear Lake ciscoes taken in nets of different mesh size [The 1931 and 1932 collections and the sexes are combined] Length inches \Yi inches 1% inches 2 inches 2% inches 2>£ inches 3 inches Total 160 to 159 1 1 160 to 169 4 6 10 170 to 179 - 10 17 3 30 180 to 189 3 12 3 18 190 to 199 2 2 200 to 209 1 i 1 3 210 to 219 1 1 220 to 229 1 1 230 to 239 2 3 6 240 to 249 2 1 8 11 250 to 259 i 4 7 12 8 32 260 to 269 2 5 11 8 26 270 to 279 3 1 9 4 17 280 to 289 1 2 2 4 7 16 290 to 299 4 8 10 22 300 to 309 2 8 6 6 22 310 to 319 1 3 4 12 13 33 320 to 329 2 7 6 11 26 330 to 339 1 2 2 6 7 12 30 340 to 349 1 1 1 4 10 12 29 350 to 359 2 2 2 3 6 10 25 360 to 369 1 1 5 1 6 13 370 to 379 1 2 2 1 6 380 to 389 1 i 2 Total 3 24 44 35 64 103 108 382 Average length 259 217 201 271 306 299 315 The Clear Lake samples with the exception of the I groups do not show as close dependence between the determination of the average length of individual age groups and the size of mesh by which the sample was taken (tables 76 and 77) as was observed in the other three populations. The proof of the reliability of the Clear Lake growth data must be based, therefore, on a demonstration that the population was adequately sampled throughout its length range rather than on a comparison of the average lengths of age groups|as determined from samples from nets of different mesh size. It was pointed out in the preceding paragraph that the sampling of the larger fish was probably adequate. The scarcity of fish in the 1%-inch net can be taken to show that the younger age groups were also adequately sampled. Accordingly, no age groups were eliminated from the Clear Lake samples. RELATIVE ABUNDANCE OF AGE GROUPS IN GILL NET SAMPLES The examination of tables 71, 72, 74, 75, 76, and 77 shows that on the whole the selectivity of nets is much more sharp in its effect on the number of fish in a given age group than in its effect on the determination of the average length of that age group. The following examples will illustrate this point. In Muskellunge Lake (table 72) the average lengths of the Ill-group fish in the 1/- and 1 ^-inch-mesh nets were 159 millimeters and 162 millimeters, respectively; however, the 1%-inch net took only 11 fish of this age group while the 1%-inch net took 59. In Silver Lake (table 75, data on lift of new nets) the Ill-group sample of the l)4-inch mesh had an average length only 2 millimeters less than the Ill-group sample from the 1 /4-inch net, but was nevertheless represented by only 4 fish as compared with the 22 fish from the 1 %-inch-mesli net. In the same day’s collection the V-group samples from the same two sizes of mesh had practically identical average lengths, but were more than three times as numerous in the 1%-inch-mesh net as in the lK-inch net. In Clear Lake (tables 76 and 77) the fish of the older age groups were more abundant in the samples of the 2)4- and 3-inch-mesh nets, but the average lengths of these older age groups as determined from 2)4- and 306 BULLETIN OP THE BUREAU OF FISHERIES 3-inch-net samples failed to be consistently higher than the average lengths determined for the same age groups from samples that were taken in smaller meshes. From the above examples it may be concluded that selection by numbers is much sharper than selection by length. Apparently nets of different mesh size, which in their operation upon a population take samples that disagree only slightly as to average length, may take their samples with unequal degreees of facility, with the result that while their catches agree closely in average length they^may differ markedly as to the actual number of fish taken. Table 78 (catch ob2%-, 2/-, and 3- inch nets) shows such a situation. If small differences in the sizes of mesh are ac- companied by large differences in the number of fish captured from a certain size range, then small differences in the average sizes of different groups of fish may be expected to produce large differences in their numerical abundance in the catch of a particular size mesh. From the above conclusions it can be seen that a series of nets that may take quite reliable samples of particular age groups, as average length is concerned, may at the same time fail to capture the members of these age groups in numbers corresponding to their actual relative abundance in the population of which they are part. Conse- quently gill-net samples must be employed with extreme caution in the study of the relative abundance of age groups and year classes. GENERAL CONCLUSIONS If the collections from the four lakes are considered as a whole it may be stated that the action of a gill net of specified mesh is predictable only with reference to the specific nature of the population to be sampled. The action of a net of specified mesh depends first upon the range of length and abundance of the fish within the population and second upon those morphological characteristics that determine in what manner the fish is held captive. A demonstration of the reliability of a sample obtained by use of gill nets from one population does not indicate that the same gear will obtain an equally reliable sample from a second, different population. On the question of the use of gill-net samples for growth studies it was demon- strated that the analysis of growth data with respect to size of mesh of the gear in which samples are taken aids in the detection of age groups whose appearance in the sample is by reason of selection not representative. These age groups should be elimi- nated completely from the data. Such selected age groups were detected in three of the four populations considered here. It was demonstrated further that if these selected groups are eliminated the remaining growth data can be considered accurate and trustworthy within very narrow limits. No such high degree of reliability can be claimed for gill-net samples in the study of the relative abundance of the different age groups. SUMMARY 1 . This study of the growth of the cisco was based on the determination of age of 3,882 specimens and the calculation of growth from scale measurements of 3,694 specimens. The data for Trout, Muskellunge, Silver, and Clear Lakes are presented in the general paper while the smaller samples from Allequash and Tomahawk Lakes are treated separately in an appendix. 2. It was assumed that Van Oosten’s (1929) demonstration of the validity of the scale method in Leucichthys artedi can be considered to apply to the four cisco popu- AGE AND GROWTH OF THE CISCO 307 lations of the present investigation. The consistent results obtained appear to justify the assumption. 3. The calculation of growth from scale measurements was based on the assump- tion that body length and scale diameter show a constant ratio at allUengths beyond that at which the first annulus is laid down. In the Silver Lake cisco alone were there any discrepancies between growth as calculated from fish of different age and as de- termined empirically that might throw doubt on the validity of the assumption. 4. It was demonstrated that Lee’s phenomenon in the Silver Lake cisco does not depend on changing body-scale relationships with increase in length and age, but rather that the growth discrepancies are reflected in the actual scale measurements upon which the calculated growths were based. 5. Among the suggestions for the explanation of Lee’s phenomenon in the Silver Lake cisco were: (1) Selection by gear, (2) selection due to dissimilar distribution within the lake of the various elements of the population, and (3) selection due to differential mortality correlated with individual growth rate (greater mortality among individuals with more rapid growth). 6. The four populations (Trout Lake, Muskellunge Lake, Silver Lake, and Clear Lake ciscoes) show wide differences in the amount and rate of growth both in length and weight. Their order with respect to growth rate in length, from minimum to maximum, is: Trout Lake, Silver Lake, Muskellunge Lake, Clear Lake; the order with respect to growth in weight is: Trout Lake, Muskellunge Lake, Silver Lake, Clear Lake. 7. On the basis of data on the growth of the cisco in this and other publications a “cisco-type” of growth was described. 8. On the whole the high degree of overlap of the length distributions of consecu- tive age groups of the cisco makes length alone a poor index of age. 9. Because of individual variation in growth rate the largest fish within a popu- lation is frequently not the oldest. 10. Condition was described in terms of the quantity K in the equation, IT=.£lX10-5L3 {W— weight in grams; Z=lengtli in millimeters). It was pointed out that coefficients of condition calculated from the cube relationship describe relative heaviness independently of the general length-weight relationship, and are more satisfactory measures of condition than the quantity C in the equation W—CLn where the value of n is determined empirically. 1 1 . The four cisco populations show wide differences with respect both to aver- age condition and the manner of change of condition with change in length. The Muskellunge Lake cisco and the Trout Lake cisco show a loss of condition with increase in length (the rate of loss of condition is the more rapid in the former population). The Silver Lake cisco and the Clear Lake cisco, on the other hand, show improvement of condition with increase in length (the rate of improvement in condition is the greater in the Clear Lake cisco). The order of the four lakes with respect to the average condition of the fish in the samples, from minimum to maximum, is: Muskel- lunge Lake, Trout Lake, Silver Lake, Clear Lake. 12. In all four populations the relationship between length and weight may be described over certain length intervals by equations of the type, W=CLn, where IT=weight, Z=length, and C=a constant. In some of the populations, at least, these equations cannot be extended to lengths outside the range of the empirical data. 308 BULLETIN OF THE BUREAU OF FISHERIES 13. The approximate date at which the season’s growth of the cisco is completed in each of the lakes is: Trout Lake, end of July; Muskellunge Lake, late August; Silver Lake, early September; Clear Lake, late September or early October. Growth probably begins in all four populations shortly after the disappearance of the ice at about the 1st of May. The data show that the length of the cisco’s growing season depends on the local conditions within each lake and not on the climatic conditions that are approximately the same for all four lakes. 14. The bathymetric distribution of the cisco in middle and late summer is highly sensitive to conditions of temperature and dissolved oxygen concentration. A knowledge of the vertical distribution of the cisco is important in the comparison of the densities of the populations of different lakes. 15. The order of the four lakes with respect to the density of their cisco popula- tions, from the minimum to the maximum, is: Clear Lake, Silver Lake, Muskellunge Lake, Trout Lake. 16. The large differences observed in the relative abundance of the various year classes show that the degree of success of the hatch of the cisco in different calendar years is subject to a wide range of variation. Since the years that saw relatively successful or unsuccessful hatches of the cisco are not the same in different lakes, it must be concluded that the relative abundance of a particular year class depends (as does the length of the growing season) on the local conditions within each individ- ual lake, and not on general climatic conditions that would affect all lakes in the same manner. 17. In the Clear Lake cisco sexual maturity appears to be general at the end of the second year of life. In the other populations the cisco is known to be mature at the end of the third year of life, while some, at least, mature in the second year. 18. The order of the four lakes with respect to the sex ratio of the cisco populations (expressed as the number of females per 100 males), from minimum to maximum, is: Clear Lake, Silver Lake, Muskellunge Lake, Trout Lake. The differences among the populations with respect to the sex ratio depend on differences in the differential rates of mortality of the two sexes. 19. The available data indicate a close connection between certain phases of the life history of the cisco and the nature of its animate and inanimate environment. 20. The order of the four lakes with respect to the rate of growth of their cisco populations (in weight) is the same as their order with respect to the length of the cisco’s growing season, and the reverse of their order with respect to the density of their cisco populations. 21. The observed differences in the length of the cisco’s growing season in different lakes cannot account entirely for the observed differences in growth rate, but probably serve only to accentuate differences that already exist. 22. Suggested explanations of differences in the length of the cisco’s growing season in different lakes were: (1) That they may depend on differences from lake to lake in the nature of the plankton cycle, and (2) that the smaller spawners of slow- growing populations may begin the development of the gonads, preliminary to autumn spawning, earlier in the season than do the larger spawners of the rapidly growing populations. (The four lakes follow the same order with respect to growth rate, length of the growing season, and the average size of mature fish.) 23. The correlation between density of population and growth rate may depend on differences in the severity of the competition for food, or upon the operation of a AGE AND GROWTH OF THE CISCO 309 “space-factor”, whereby crowding alone impedes growth independently of the abun- dance of food. 24. Since, with the exception of a reversal of the positions of the closely similar Muskellunge Lake and Silver Lake populations, the four lakes arrange themselves in the same order with respect to the density of their cisco populations and with respect to the concentration of bound C02 in their waters, it was suggested that differences (traceable to the concentration of bound C02) in the kind and abundance of plankton available to newly hatched ciscoes may account for the observed differences in the densities of the four cisco populations. 25. Although the relationship between parasite infestation and growth rate has not been thoroughly investigated in the cisco it is known that the Clear Lake cisco, which has by far the best growth of any of the four populations, is practically free of parasites (96 percent negative), while the slower-growing populations from Silver, Muskellunge, and Trout Lakes have a heavy infestation of intestinal parasites (80 percent, or more, positive). 26. The best growth of the cisco occurs in Clear Lake with the smallest amount of organic matter in the surface plankton, but Muskellunge Lake, with the greatest amount of organic matter in the surface plankton does not have the slowest-growing cisco population. 27. The order of the four lakes with respect to the value of the sex ratio (females per 100 males) is the reverse of their order with respect to the rate of growth (in weight). It was suggested that the same factors that determine the rates of growth may cause the differential mortality rate of the less viable males to be greater in the slower- growing populations. 28. The Clear Lake cisco with the most rapid growth is in the best average con- dition (highest average value of K) and shows the most rapid improvement of con- dition with increase in length, but the Muskellunge Lake cisco with the poorest average condition and the most rapid loss of condition with increase in length is not the popula- tion with the slowest growth. 29. The at least partial independence of the factors that determine growth rate and the factors that, determine condition may be construed as a strong argument for the operation of the “space-factor” in the determination of growth rate. 30. The order of the four lakes with respect to the average condition of their cisco populations and the rate of change of condition with increase in length is the reverse of their order with respect to the abundance of organic matter in the surface plankton. Thus it appears that the poorest condition is found in the most eu trophic environment and the best condition in the most oligo trophic environment. 31. Longevity and condition appear to be correlated. The average life span of the cisco is shortest in Muskellunge Lake where the average condition ( K ) is poorest and increase in length is accompanied by a rapid loss of condition, and the longest in Clear Lake where the average condition is best and condition improves as length increases. 32. In a general review and summary of the growth relationships of the cisco, it was brought out that some of these relationships fail to conform to generally accepted theories. The need was mentioned for a broaderlknowledge of the biology of the cisco based on observations in a great diversity of habitats. Attention was called to the danger of premature generalizations concerning the growth relationships and the biology of this very plastic species. 310 BULLETIN OF THE BUREAU OF FISHERIES 33. In each cisco population the amount of growth made by fish in the same year of fife but in different calendar years was found to vary considerably. There is no apparent correlation between these variations in the amount of growth in different calendar years and annual variations in the average air temperature during the growing season. There is some indication that annual fluctuations in the density of each individual population may affect the amount of growth made in different calendar years, but with the possible exception of Clear Lake the data in support of such a rela- tionship are not convincing. 34. The 1928 year class of Trout Lake and the 1926 and 1928 year classes of Muskellunge Lake show a bimodal distribution of the calculated lengths at the end of the first year of life. It was suggested that the reason for these bimodal distribu- tions lies in exceptional weather conditions that bring about two hatchings in the spring. However, the effectiveness of weather conditions in bringing about two hatchings appears to depend on local conditions within each lake. 35. The 1928 year class in Muskellunge Lake which has a bimodal distribution of the calculated lengths at the end of the first year of life, and the 1929 Muskellunge Lake year class, which has an unimodal distribution of the calculated lengths at the end of the first year of life, were used as the basis of a stud}^ of the effect of the dis- tribution of lengths at an early stage in the life history on the phenomenon of growth compensation in later years. It was found that growth compensation is much more intense in the group with the greater dispersion of the length distribution at the end of the first growing season. In both groups the larger fish at the end of the first year of life tend to be the larger fish at the end of the second year of life and at the time of capture in the third or fourth summer. Growth compensation in later years reduces the advantage of the large first year fish but fails to eliminate it completely. 36. The fish associations of which the ciscoes are part vary from lake to lake. In midsummer the cisco is taken in Trout Lake only with the typical deep water forms, the whitefish, the lake trout, and the burbot; in Muskellunge Lake large numbers of perch and other shallow-water fish are taken with the cisco ; in Silver Lake the cisco spends the summer in practical isolation; and in Clear Lake the pike-perch is relatively abundant in the strata inhabited by the cisco. 37. Because of the scattered nature of the available published data, a review was made of earlier investigations concerning the selective action of gill nets. 38. The examination of the size and age composition of the gill-net samples showed that the action of a net or group of nets in taking samples of a cisco population cannot be predicted on the basis of experiences with other populations of the same species, but that each population presents its own problem of gill-net selectivity. Although gill nets are very selective in their operation, the judicious elimination of selected age groups can make gill-net samples highly reliable for the determination of growth curves. Extreme caution must be employed, however, in the use of gill- net samples for the study of age and year class composition. BIBLIOGRAPHY Birge, Edward A., and Chancey Juday. 1911. The inland lakes of Wisconsin. The dissolved gases of the water and their biological significance. Wisconsin Geol. and Nat. Hist. Sur., Bull. XXII, Scientific Series No. 7, 1911. xx+259 pp. Madison. Bjerkan, P. 1917. Age, maturity, and quality of North Sea herrings during the years 1910-13. Report, Norwegian Fish, and Mar. Inves., vol. Ill, no. 1, 1917, 119 pp. Bergen. Blegvad, H. 1933. Plaice transplantation. Jour, du Cons., Cons. Perm. Internat. Explor. Mer, vol. VIII, no. 2, 1933, pp. 161-180. Copenhague. AGE AND GROWTH OF THE CISCO 311 Borley, J. 0. 1912. Report on the experimental transplantation of plaice to the Dogger Bank carried out by the Marine Biological Association in the years 1904-08. Internat. Fish. Invest., Report IV (Southern Area Cd. 6125), 1912, 79 pp. London. Borley, J. O., and E. S. Russell. 1922. Report on herring trawling. Min. Agri. and Fish., Fishery Investigations, series II, vol. IV, no. 4, 1922, 57 pp., 12 figs. London. Bowman, Alexander. 1932. The effect on the stock of the capture of undersized fish. The haddock population of the North Sea plateau. Rapp, et Proces-Verb., Cons. Perm. Internat. Explor. Mer, vol. LXXX, no. Ill, 1932, 19 pp. Copenhague. Buchanan- Wollaston, H. J. 1927. On the selective action of a trawl net, with some remarks on the selective action of drift nets. Jour, du Cons., Cons. Perm. Internat. Explor. Mer, vol. II, no. 3, 1927, pp. 343-355. Copenhague. Buckmann, Adolf. 1932. Die Frage nach der Zweckmiissigkeit des Schutzes untermassiger Fische und die Voraussetzungen fur ihre Beantwortung. Rapp, et Proces-Verb., Cons. Perm. Internat. Explor. Mer, vol. LXXX, no. VII, 1932, 16 pp. Copenhague. Cahn, Alvin Robert. 1927. An ecological study of southern Wisconsin fishes. The brook silver- sides ( Labidesthes sicculus) and the cisco ( Leucichthys artedi) in their relations to the region. Illinois Biol. Mono., vol. XI, no. 1, 1927, 151 pp., 16 plates. Urbana. Clark, Frances N. 1928. The weight-length relationship of the California sardine ( Sardina caerulea) at San Pedro. Division of Fish and Game of California, Fish Bull. No. 12, 1928, 59 pp., XI figs. Sacramento. Clemens, Wilbert A. 1922. A study of the ciscoes of Lake Erie. University of Toronto Studies, Publications, Ontario Fish. Res. Lab., No. 2, 1922, pp. 27-37. Toronto. Creaser, Charles W. 1926. The structure and growth of the scales of fishes in relation to the interpretation of their life history, with special reference to the sunfish Eupomotis gibbosus. Museum of Zoology, University of Michigan, Misc. Pub., no. 17, Dec. 15, 1926, pp. 1-82, 12 pi., 12 figs. Ann Arbor. Crozier, William J., and Selig Hecht. 1914. Correlation of weight, length, and other body measurements in the weakfish, Cynoscion regalis. Bull., Bur. Fish. vol. XXXIII, 1913 (1915), pp. 139-147, 4 figs. Washington. Dawes, Ben. 1930. Growth and maintenance in the plaice (Pleuronecte s platessa L.). Part I. Jour., Marine Biol. Asso., United Kingdom, N. S., vol. XVII, no. 1, 1930, pp. 103-174, 19 figs. Plymouth. Dawes, Ben. 1931a. Growth and maintenance in the plaice ( P . platessa L.). Part II. Jour., Marine Biol. Asso., United Kingdom, N. S., vol. XVII, no. 3, 1931, pp. 877-947, 11 figs. Plymouth. Dawes, Ben. 1931b. A statistical study of growth and maintenance in the plaice (P. platessa L.). Jour., Marine Biol. Asso., United Kingdom, N. S., vol. XVII, no. 3, 1931, pp. 949-976, 5 figs. Plymouth. Delsman, H. C. 1914. Uber das Wachstum von Nordseehering und Zuiderseehering nach Unter- suchungen an den Schuppen. Rapporten en Verhandelingen. Uitgegeven door het Rijksinsti- tuut voor Visscherijonderzoek, Deel I, aft. 2, 1914, pp. 133-200, 4 figs., II pis. S. Gravenhage. Dymond, J. R. 1933. Biological and oceanographic conditions in Hudson Bay. 8. The coregonine fishes of Hudson and James Bays. Contrib., Canadian Biol, and Fish., N. S., vol. VIII, no. 1 (Series A, General, No. 28), 1933, pp. 1-12. Toronto. Elster, Hans-Joachim. 1934. Beitrage zur Biologie des Blaufelchen ( Coregonus wartmanni Bloch). Internationale Revue der gesamten Hydrobiologie und Hydrographic, Bd. 30, 1934, S. 181-246, 6 figs. Leipzig. Forel, F. A. 1892. Le Leman. Monographie limnologique, 3 vols., 1892. Lausanne. Fraser, J. H. 1931. On the size of Urosalpinx cinerea Say with some observations on the weight- length relationship. Proc., Malacological Soc., vol. XIX, part V, 1931, pp. 243-254. London. Fulton, T. Wemyss. 1904. The rate of growth of fishes. Twenty-second Annual Report, Fish. Bd., Scotland, 1903 (1904), part III, pp. 141-241. Edinburgh. Garstang, Walter. 1926. Plaice in the North Sea— changes in size of catch. The London Times, 1926, April 21, p. 15; April 26, p. 20. London. Geiser, S. W. 1923. Evidences of a differential death rate of the sexes among animals. Ameri- can Midland Naturalist, vol. VIII, no. 7, 1923, pp. 153-163. St. Louis. Geiser, S. W. 1924a. Sex ratios and spermatogenesis in the top-minnow, Gambusia holbrooki Grd. Biol. Buff., vol. XLVII, no. 3, 1924, pp. 175-212. Woods Hole. 312 BULLETIN OF THE BUREAU OF FISHERIES Gieser, S. W. 1924b. The differential death rate of the sexes among animals, with a suggested explanation. Washington University Studies, vol. XII, series no. 1, 1924, pp. 73-96. St. Louis. Haakh, Theodor. 1929. Studien fiber Alter und Wachstum der Bodenseefische. Archiv fur Hydrobiologie, Bd. 20, Heft 2, 1929, S. 214-295. Berlin. Hart, John L. 1928. Data on the rate of growth of pike perch ( Stizostedion vitreuvi) and sauger (S. canadense) in Ontario. University of Toronto Studies, Publications, Ontario Fish. Res. Lab., no. 34, 1928, pp. 45-55, 3 figs. Toronto. Hart, John L. 1931. The growth of the whitefish, Coregonus clupeaformis (Mitchill). Contrib., Canadian Biol, and Fish., N. S., vol. VI, no. 20 (series A, general, no. 5), April 8, 1931, pp. 429- 444, 11 figs. Toronto. Hart, John L. 1932. Statistics of the whitefish ( Coregonus clupeaformis) population of Shake- speare Island Lake, Ontario. University of Toronto Studies, Publications, Ontario Fish. Res. Lab., no. 42, 1932, pp. 1-28. Toronto. Hecht, Selig. 1916. Form and growth in fishes. Contrib., Zoological Laboratory, Museum of Comparative Zoology, Harvard College, Cambridge, no. 275. Jour. Morphology, vol. XXVII, no. 2, 1916, pp. 379-400. Philadelphia. Heincke, F., and A. Buckmann. 1926. Die Ueberfischung der Nordsee und die Wirkungen der Kriegsschonzeit auf ihren Schollenbestand. Der Fischerbote. Norddeutsche Fischerei- Zeitung, XVIII, Jahrg. 1926, S. 57-101, 13 figs. Hamburg. Higgins, Elmer. 1928a. Progress in biological inquiries, 1926. Including proceedings of the divisional conference January 4-7, 1927. Appendix VII, Report, U. S. Comm. Fish., 1927 (1928), pp. 517-681. Washington. Higgins, Elmer. 1928b. Progress in biological inquiries, 1927. Appendix VI, Report, U. S. Comm. Fish. 1928 (1929), pp. 199-247. Washington. Higgins, Elmer. 1930. Progress in biological inquiries, 1928, including extracts from the pro- ceedings of the divisional conference, January 2 to 5, 1929. Appendix X, Report, U. S. Comm. Fish., 1929 (1930), pp. 627-739. Washington. Higgins, Elmer. 1931a. Progress in biological inquiries, 1929. Appendix XV, Report, U. S. Comm. Fish., 1930 (1931), pp. 1069-1121, 11 figs. Washington. Higgins, Elmer. 1931b. Progress in biological inquiries, 1930. Appendix III, Report, U. S. Comm. Fish., 1931 (1932), pp. 553-626, 9 figs. Washington. Higgins, Elmer. 1932. Progress in biological inquiries, 1931. Appendix III, Report, U. S. Comm. Fish., 1932 (1933), pp. 441-529. Washington. Higgins, Elmer. 1933. Progress in biological inquiries, 1932. Appendix II, Report, U. S. Comm. Fish., 1933 (1934), pp. 79-147. Washington. Hile, Ralph. 1931. The rate of growth of fishes of Indiana. Investigations of Indiana Lakes, Iff No. 2. Depart. Conser., Division of Fish and Game, Indiana, Publication No. 107, 1931, pp. 9-55. Indianapolis. Hjort, Johan. 1914. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Rapp, et Prociis-Verb., Cons. Perm. Inter. Explor. Mer, vol. XX, 1914, 288 pp., 137 figs. Copenhague. Hjort, Johan. 1926. Fluctuations in the year classes of important food fishes. Jour, du Cons., Cons. Perm. Internat. Explor. Mer, vol. 1, no. 1, 1926, pp. 5-38, 18 figs. Copenhague. Hjort, Johan. 1932. Remarks on the fluctuations in number and growth in marine populations. Rapp, et Proces-Verb., Cons. Perm. Internat. Explor. Mer, vol. LXXX, no. IX, 1932, pp. 3-8. Copenhague. Hodgson, W. C. 1927. Preliminary note on experiments concerning the selective action of drift nets. Jour, du Cons., Cons. Perm. Internat. Explor. Mer, vol. II, no. 3, 1927, pp. 356-360, 3 figs. Copenhague. Hodgson, W. C. 1933. Further experiments on the selective action of commercial drift nets. Jour, du Cons., Cons. Perm. Internat. Explor. Mer, vol. VIII, no. 3, 1933, pp. 344-354. Copen- hague. Huitfeldt-Kaas, Hartwig. 1917. Mjpsens fisker og fiskerier. Det Kgl. Norske Videnskabers Selskabs Skrifter, Nr. 2, Heft 1, 1916 (1917), iv + 257 S. Trondhjem. Huitfeldt-Kaas, Hartwig. 1927. Studier over aldersforholde og veksttyper hos norske fersk- vannsfisker. Nationaltrykkeriet, 1927, S. 5-358. Oslo. AGE AND GROWTH OF THE CISCO 313 Ireland Fisheries. 1905. Report on the sea and inland fisheries of Ireland for 1902 and 1903. Part II. Scientific Investigations. Depart, of Agric. and Tech. Instruct., Ireland, 1905, 385 pp., XXXV pis. Dublin. Jarnefelt, H. 1921. Untersuchungen fiber die Fische und ihre Nahrung im Tuusulasee. Acta Societatis Pro Fauna et Flora Fennica, vol. 52, no. 1, 1921, 160 pp. Helsinki. Jarvi, T. H. 1920. Die kleine Marane ( Coregonus albula L.) im Keitelesee, eine okologische und okonomische Studie. Annales Academiae Scientiarum Fennicae, Serie A, vol. XIV, No. 1, 1920, 302 pp. Helsinki. Jarvi, T. H. 1924. Die kleine Marane ( Coregonus albula L.) im Nilakka und Pielavesi. Annales Academiae Scientiarum Fennicae, Serie A, vol. XXI, No. 2, 1924, 134 pp., 24 figs., 2 maps. Helsinki. Jarvi, T. H. 1930. On the fluctuations in abundance exhibited by the stock of lake-herring ( Coregonus albula) in Lake Keitele, Finland. Rapp, et Proees-Verb., Cons. Perm. Internat. Explor. Mer, vol. LXV, no. 20, 1930, pp. 183-188, 2 figs. Copenhague. Jensen, Aage J. C. 1928. The relation between the size of the plaice stock and the quantity of “First Class Plaice Food” in certain parts of the Limfjord. Report, Danish Biol. Sta., Bd. of Agri., vol. XXXIV, 1928, pp. 87-99, 3 figs. Copenhagen. Jensen, Aage J. C. 1932. The effect of the plaice fishery on the stock of undersized plaice and its influence on the yield of the plaice fishery in the North Sea. Rapp, et Proces-Verb., Cons. Perm. Internat. Explor. Mer, vol. LXXX, no. IV, 1932, pp. 3-6. Copenhague. Johansen, A. C. 1928. The extent of the annual transplantations of plaice in the Limfjord and the yield resulting therefrom. Report, Danish Biol. Sta., Bd. of Agri., 1928, pp. 53-86. Copenhagen. Juday, C., and Edward A. Birge. 1930. The highland lake district of northeastern Wisconsin and the Trout Lake limnological laboratory. Trans., Wisconsin Academy of Science, Arts, and Letters, vol. XXV, 1930, pp. 337-352. Madison. Juday, C., and Edward A. Birge. 1932. Dissolved oxygen and oxygen consumed in the lake waters of northeastern Wisconsin. Trans., Wisconsin Academy of Science, Arts, and Letters, vol. XXVII, 1932, pp. 415-486, 35 figs. Madison. Juday, C., and Edward Schneberger. 1930. Growth studies of game fish in Wisconsin lakes. Notes, Biol. Lab., Wisconsin Geol. and Nat. Hist. Sur., March 1930. Madison. [Mimeo- graphed.] Juday, C., and Edward Schneberger. 1933. Growth studies of game fish in Wisconsin waters — second report. Notes, Limnol. Lab., Wisconsin Geol. and Nat. Hist. Sur., April 1933, 10 pp. Madison. [Mimeographed.] Keys, Ancel B. 1928. The weight-length relation in fishes. Proc., Nat. Acad, of Science, vol. XIV, no. 12, 1928, pp. 922-925. Washington. Koelz, Walter N. 1926. Fishing industry of the Great Lakes. Appendix XI, Report, U. S. Comm. Fish., 1925 (1926), pp. 553-617. Washington. Koelz, Walter N. 1931. The coregonid fishes of northeastern America. Papers, Michigan Academy of Science, Arts, and Letters, vol. XIII, 1930 (1931), pp. 303-432. Ann Arbor. Krogius, F. 1933. Zur Kenntnis der Systematik und Biologie der Baikal Renken. Travaux de la Station Limnologique du lac Bajkal, vol. V, 1933, pp. 5-154. Lechler, Hermann. 1929. Untersuchungen fiber die Reinanke des Mondsees. (I.) Zeitschrift ffir Fischerei und deren Hilfswissenschaften, Bd. XXVII, Heft 4, 1929, S. 547-560. Neudamm und Berlin. Lechler, Hermann. 1930. Untersuchungen fiber die Reinanke des Mondsees. (II.) Zeitschrift fur Fischerei und deren Hilfswissenschaften, Bd. XXVIII, 1930, S. 485-493. Neudamm und Berlin. Lee, Rosa M. [Mrs. T. L. Williams], and George T. Atkinson. 1912. Report on plaice trans- plantation experiments to various fishing grounds in the North Sea. Marine Biol. Asso. Internat. InvestL, Report IV (Cd 6125), 1912, pp. 107-151. London. McCay, C. M. 1933. Is longevity compatible with optimum growth? Science, N. S., vol. 77, no. 2000, April 28, 1933, pp. 410-411. Lancaster, Pa. McCay, C. M., A. Tunison, Mary Crowell, D. K. Tressler, S. P. MacDonald, John W. Titcomb, and Eben W. Cobb. 1931. The nutritional requirements of trout and chemical composition of the entire trout body. Trans., American Fish. Soc., vol. 61, 1931, pp. 58-79. Hartford. 314 BULLETIN OF THE BUREAU OF FISHERIES Mottley, C. McC. 1932. The propagation of trout in the Kamloops district, British Columbia. Trans., American Fish. Soc., vol. 62, 1932, pp. 144-151. Hartford. Olofsson, Ossian. 1932. Sikens tillv&xt under en varm och en kail sommar. Ny Svensk Fiskeritidskrift, Nr. 14, 1932, S. 163-166. Goteborg. Osbobne, Thomas B., Lafayette B. Mendel, and Edna L. Ferry. 1917. The effect of retarda- tion of growth upon the breeding period and duration of life of rats. Science, N. S., vol. 45, no. 1160, March 23, 1917, pp. 294-295. New York. Pirognikoff, P. L. 1927. Materials concerning age and growth rate of rudd ( Rutilus rut. lacustris Pallas) from the Lake Chani. Reports, Ichthyological Laboratory, Siberia, vol. II, no. 5, 1927, pp. 3-15. Krasnoyarsk. Pritchard, Andrew L. 1928. A preliminary study of the genus Leucichthys in the Canadian waters of Lake Ontario. University of Toronto Studies, Publications, Ontario Fish. Res. Lab., no. 32, 1928, pp. 7-25. Toronto. Pritchard, Andrew L. 1931. Taxonomic and life history studies of the ciscoes of Lake Ontario. University of Toronto Studies, Publications, Ontario Fish. Res. Lab., no. 41, 1931, pp. 5-78. Toronto. Reibisch, J. 1908. Ein Dickenkoeffizient als Mass fur Alter und Ernahrungszustand der Fische. Die Beteiligung Deutschlands an der internationalen Meeresforschung, Jahresbericht IV/V, 1908, S. 59-65. Berlin. Reibisch, J. 1911. Biologische Untersuchungen iiber Gedeihen, Wanderung und Ort der Ent- stehung der Scholle ( Pleuronectes platessa ) in der Ostsee. Wissenschaftliche Meeresunter- suchungen, herausg. von der Kommission zur Untersuchung der deutschen Meere, neue Folge, Bd. XIII, Abteilung Kiel, Nr. 18, 1911, S. 127-204, 11 figs., 28 tables. Kiel und Leipzig. Rich, Willis H. 1926. Progress in biological inquiries 1925. Appendix I, Report, U. S. Comm. Fish., 1926 (1927), pp. 1-36. Washington. Rugde. 1923. Some notes from Norway. Salmon and Trout Magazine, No. 34, 1923, pp. 333-337. London. Russell, E. S. 1932. Is the destruction of undersized fish by trawling prejudicial to the stock? Rapp, et Proces-Verb., Cons. Perm. Internat. Explor. Mer, vol. LXXX, no. 8, 1932, pp. 2-13, 1 fig. Copenhague. Schultz, Leonard P. 1933. The age and growth of Atherinops affinis oregonia Jordan and Snyder and of other subspecies of bay smelt along the Pacific coast of the United States. Uni- versity of Washington Publications in Biology, vol. 2, no. 3, 1933, pp. 45-102. Seattle. Scott, Will. 1931. The lakes of northeastern Indiana. Investigations of Indiana Lakes, No. 3. Department of Conservation, Division of Fish and Game, Indiana, Publication No. 107, 1931, pp. 61-145. Indianapolis. SegerstrXle, Curt. 1932. Uber die jahrlichen Zuwachszonen der Schuppen und Beziehungen zwischen Sommertemperatur und Zuwachs bei Abramis brama sowie einigen anderen Cypriniden in Sud- Finland, 1911-1930. Acta Zoologica Fennica, vol. 13, 1932, 42 pp. Helsinki. Segerstrale, Curt. 1933. Uber scalimetrische Methoden zur Bestimmung des linearen Wachs- tums bei Fischen, insbesondere bei Leuciscus idus L., Abramis brama L., und Perea fluviatilis L. Acta Zoologica Fennica, vol. 15, 1933, pp. 1-168, 37 figs. Helsinki. Storrow, B. 1932. Concerning fluctuations in marine food fishes. Jour., Animal Ecology, vol. 1, no. 2, 1932, pp. 160-165. Cambridge. Strodtmann, S. and H. Langhammer. 1925. Untersuchungen fiber die Scholle in der westlichen Ostsee. Berichte der deutschen wissenschaftlichen Kommission ffir Meeresforschung, N. F., Bd. I, Nr. xi, 1925, S. 305. Berlin. Surbeck, G. 1921. Beitrag zur Kenntniss der schweizerischen Coregonen. Schweizerische Fischerei-Zeitung, Jahrgang 29, Nr. 3, 1921, S. 52-59. Bern. Tester, Albert L. 1932. Rate of growth of the small-mouthed black bass ( Micropterus dolomieu ) in some Ontario waters. University of Toronto Studies, Publications, Ontario Fish. Res. Lab., no. 47, 1932, pp. 207-221. Toronto. Thompson, D’Arcy Wentworth. 1917. On growth and form. 793 pp., 1917. Cambridge University Press. Cambridge. Thwaites, F. T. 1929. Glacial geology of part of Vilas County, Wisconsin. Trans., Wisconsin Academy of Science, Arts, and Letters, vol. XXIV, 1929, pp. 109-125. Madison. AGE AND GROWTH OF THE CISCO 315 Titcomb, John W., Eben W. Cobb, Mary F. Crowell, and C. M. McCay. 1928. The nutri- tional requirements and growth rates of brook trout. Trans., American Fish. Soc., vol. 58, 1928, pp. 205-231, 9 charts. Hartford. Tjurin, P. V. 1927. About the relation between the length of the fish and its weight. Reports, Ichthyological Lab., Siberia, vol. II, no. 3, 1927, pp. 3-21. Krasnoyarsk. Van Oosten, John. 1923. The whitefishes ( Coregonus clupeaformis). A study of the scales of whitefishes of known ages. Zoologica, vol. II, no. 17, June 19, 1923, pp. 380-412, figs. 137-144, tables I- VII. New York. Van Oosten, John. 1929a. Some fisheries problems on the Great Lakes. Trans., American Fish. Soc., vol. 59, 1929, pp. 63-85. Hartford. Van Oosten, John. 1929b. Life history of the lake herring ( Leucichthys artedi Le Sueur) of Lake Huron as revealed by its scales, with a critique of the scale method. Bull. U. S. Bur. Fish., vol. XLIV, 1928 (1929), pp. 265-428, 43 figs. Washington. Van Oosten, John. 1933. Preliminary report on investigation of chub net meshes in Lake Michigan. The Fisherman, vol. 2, no. 4, March 1933, pp. 3-4, 8. Grand Haven. Wagler, Erich. 1927. Die Blaufelchen des Bodensees ( Coregonus wartmanni Bloch). Interna- tionale Revue der gesamten Hydrobiologie und Hydrographie, Bd. 18, Nr. 3/4, 1927, S. 129-230, 13 figs., 11 plates. Leipzig. Wagler, Erich. 1930a. Die Coregonen in den Seen des Voralpengebietes. II. Die Schwebrenke des Tegernsees. Archiv fur Hydrobiologie, Bd. 21, Heft 3, 1930, S. 455-469. Berlin. Wagler, Erich. 1930b. Die Coregonen in den Seen des Voralpengebietes. III. Die Schwebrenke des Ammersees. Archiv fur Hydrobiologie, Bd. 21, Heft 3, 1930, S. 470-482. Berlin. Wagler, Erich. 1932. Die Coregonen in den Seen des Voralpengebietes. V. Der Silberfelchen des Untersees. Internationale Revue der gesamten Hydrobiologie und Hydrographie, Bd. 26, Nr. 3/4, 1932, S. 195-222, 7 figs. Leipzig. Wagler, Erich. 1934. Die Coregonen in den Seen des Voralpengebietes. VII. Der Kilch des Bodensees ( Coregonus acronius von Rapp). Internationale Revue der gesamten Hydrobiologie und Hydrographie, Bd. 30, Nr. 1/2, 1934, S. 1-48, 8 figs. Leipzig. Walford, Lionel A. 1932. The California barracuda ( Sphyraena argentea ). I. Life history of the California barracuda. II. A bibliography of barracudas (Sphyraenidae) . Division of Fish and Game of California, Fish Bull. No. 37, 1932, pp. 7-120. Sacramento. Weather Bureau, U. S. Department of Agriculture. 1923-1931. Climatological Data for the United States. Vols. X-XVIII, 1923-1931. Washington. Weymouth, Frank W. 1918. Contributions to the life history of the Pacific Coast edible crab (Cancer magister) . (No. 3.) Report, Comm. Fish., 1917 (1918), Province of British Columbia, pp. Q 81-Q 90, 2 figs. Victoria. Weymouth, Frank. 1923. The life history and growth of the Pismo clam ( Tivela stultorum Mawe). State of California Fish and Game Commission, Fish Bulletin No. 7, 1923, pp. 5-120, 15 figs., 18 graphs. Sacramento. Willer, Alfred. 1924. Die kleine Marane ( Coregonus albula L.) in Ostpreussen. Internationale Revue der gesamten Hydrobiologie und Hydrographie, Bd. XII, Heft 3/4, 1924, S. 248-265. Leipzig. Willer, Alfred. 1926. Untersuchungen fiber den Stint ( Osmerus eperlanus L.) in Ostpreussen. Zeitschrift ffir Fischerei, und deren Hilfswissenschaften Bd. XXIV, Heft 4, 1926, S. 521-558. N eudamm-Berlin. Willer, Alfred. 1929. Untersuchungen fiber das Wachstum von Fischen. Verhandlungen der Internationalen Vereinigung ffir theoretische und angewandte Limnologie, Bd. IV 1929, S. 668-684. Rome. Wright, Stillman. 1929. A preliminary report on the growth of the rock bass, Amblopliles rupestris (Rafinesque) , in two lakes of northern Wisconsin. Trans., Wisconsin Academy of Science, Arts, and Letters, vol. XXIV, 1929, pp. 581-595. Madison. Zabinski, Jan. 1929. The growth of black beetles and of cockroaches on artificial and on incom- plete diets. British Jour. Exper. Biol., vol. VI, no. 4, 1928, pp. 360-385. Edinburgh. 316 AGE AND GROWTH OF THE CISCO APPENDIX A. GROWTH OF THE CISCO IN ALLEQUASH AND TOMAHAWK LAKES Small collections of ciscoes were obtained from Allequash Lake in 1928 and in 1930 and from Tomahawk Lake in 1928. Since it is unlikely that any further collec- tions of ciscoes will be made in either of these lakes in the near future, the available growth data for each of the populations are being presented at this time. Tables 1 and 2 show the average length, average weight, average value of K, and the calculated length at the end of each year of life in each age group for each of the three collections. The most noteworthy features of the growth in these two populations are the slow growth during the first year of life in the Allequash fish, the generally excellent growth of the Tomahawk Lake cisco, and the indication that in the Allequash Lake cisco the year 1929 was one above the average in the amount of growth. (It should be pointed out that 1929 was a good growth year in the Trout Lake, Muskellunge Lake, and Silver Lake cisco populations.) The I group of the 1930 Allequash collection shows a particularly high growth for the 1929 season. It is, of course, true that the selective action of gear would be expected to cause a too high calculated growth for this group in any year but selection can hardly explain such a great growth as that which is present here in 1929, for it may be seen that in 1929 the calculated growth for the first year of life is 29 millimeters greater than that calcu- lated for 1927 from the 1928 I group. Table 1. — Allequash Lake cisco [Length, weight, value of K, and calculated length at the end of each year of life. Division of data is according to year class and year of capture. Sexes combined] Year class Year of capture Age Number of speci- mens Weight in grams K Length in milli- meters Year of life 1 2 3 4 1925 1928 III 3 54 1. 50 155 54 101 127 1926 192S II 17 45 1.51 144 59 107 1930 IV 1 140 1.67 203 49 98 138 175 1927 1928 I 27 25 1.36 123 64 1930 III 12 108 1. 56 190 58 115 165 1928 1930 II 11 71 1. 54 167 54 133 1929 1930 I 42 42 1. 43 143 93 Table 2. — Tomahawk Lake cisco, 1928 [Length, weight, value of K, and calculated length at the end of each year of life for each age group. Sexes combined] Year class Age Number of speci- mens Weight in grams K Length in milli- meters Year of life 1 2 3 4 5 1923 V 1 156 1. 53 217 81 138 172 193 210 1924 IV 16 146 1. 59 209 78 141 174 198 1925 III 39 125 1. 58 199 76 140 182 1926 IT 1 1 99 1. 61 183 81 155 1927 I 1 35 I. 42 135 83 Because of the poor growth in the first year of life the general growth of the Allequash Lake cisco differs from that of all four populations considered in the main body of this study. The closest resemblance is probably with the Silver Lake cisco. The growth of the Tomahawk Lake cisco although poorer than that of the Clear Lake cisco is distinctly better than that found in Trout, Muskellunge, and Silver Lakes. AGE AND GROWTH OF THE CISCO 317 In both Allequash and Tomahawk Lakes the high values of K indicate good condition. The exceptional conditions under which the Allequash cisco lives should be mentioned. (See table 1 of general paper.) Formerly there existed a rather general belief that the coregonids as a stenothermic form adapted to a cold-water habitat were limited in their distribution to the larger, deeper, lakes. Wilier (1924) showed that such a generalization did not apply to the distribution of the kleine Marane, Coregonus albula, in the lakes of East Prussia. Scott (1931) called attention to the exceptional conditions under which the cisco lives in Indian Village Lake (Indiana). This lake is small and shallow (6.4 meters) and in midsummer there is “only a trace” of oxygen below 4 meters. Allequash Lake is another example of a shallow-water habitat for the cisco. The cisco here must not only adjust itself to a wide annual range of temperature variation but also to unusual associates in the form of other species of fish. Along with the 70 ciscoes captured in 1930 were taken 1 blue gill, 15 suckers, 22 pike-percb, 32 rork bass, and 182 perch. ' " ■ : . : : ' : 1 " I''-;- ' v U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner SUPPLEMENTAL NOTES ON FISHES OF THE GULF OF MAINE By HENRY B. BIGELOW and WILLIAM C. SCHROEDER From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 20 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1936 For sale by the Superintendent of Documents, Washington, D. C. Price 10 cents SUPPLEMENTAL NOTES ON FISHES OF THE GULF OF MAINE0 j* By Henry B. Bigelow and William C. Schiioeder, Woods Hole Oceanographic Institution J- CONTENTS Page Introduction 320 Hagfish, Myxine glutinosa Linnaeus 321 Sea lamprey, Petromyzon marinus Lin- naeus 321 Smooth dogfish, Mustelus mustelus (Lin- naeus) 321 Great blue shark, Prionace glauca (Lin- naeus) 321 Dusky shark, Carcharhinus obscurus (Le- Sueur) 321 Shovel-head shark, Cestracion tiburo (Lin- naeus) 322 Hammerhead shark, Cestracion zygaena (Linnaeus) 322 Thresher, Alopias vulpinus (Bonnaterre) . 322 Mackerel shark, Isurus nasus (Bonna- terre) 322 Sharp-nosed mackerel shark, Isurus tigris (Atwood) 322 White shark, Carcharodon carcharias (Lin- naeus) 322 Spiny dogfish, Squalus acanthias Lin- naeus 322 Greenland shark, Somniosus microcephalus (Bloch and Schneider) 323 Key to skates and rays 323 Little skate, Raja erinacea Mitchill 324 Big skate, Raja diaphanes Mitchill 324 Prickly skate, Raja scabrata Garman 324 Brier skate, Raja eglanteria Bose 325 Smooth skate, Raja senta Garman 325 Barn-door skate, Raja stabuliforis Garman 325 Torpedo, Narcacion nobilianus (Bona- parte) 325 Chimaera, Chimaera affinis Capello 325 Common sturgeon, Acipenser sturio Lin- naeus 326 Short-nosed sturgeon, Acipenser breviros- tris LeSueur 326 Eel, Anguilla rostrata (LeSueur) 326 Page American conger, Conger oceanica (Mitch- ill) 326 Snake eel, Pisodonophis i ruentifer Goode and Bean 326 Herring, Clupea harengus Linnaeus 327 Alewife, Pomolobus pseudoharengus (Wil- son) 327 Blueback, Pomolobus aestivalis (Mitchill). 327 Thread herring, Opisthonema oglinum (Le- Sueur) 327 Round herring, Etrumeus sadina (Mitch- ill) 327 Anchovy, Anchoviella mitchilli (Cuvier and Valenciennes) 327 Striped anchovy, Anchoviella epsetus (Bon- naterre) 328 Argentine, Argentina silus Ascanius 328 Pearlsides, Maurolicus pennanti (Wal- baum) 328 Viperfish, Chauliodus sloanei Bloch and Schneider 328 Lancetfish, Alepisaurus ferox Lowe 328 Needlefish, Scomberesox saurus (Walbaum) 328 Trumpetfish, Fistularia tabacaria Lin- naeus 328 Pelagic pipefish, Syngnathus pelagicus Lin- naeus. 329 Common pipefish, Syngnathus fuscus Storer 329 Northern barracuda, Sphyraena borealis DeKay 329 Mackerel, Scomber scombrus Linnaeus 329 Tuna, Thunnus thynnus (Linnaeus) 329 Common bonito, Sarda sarda (Bloch) 330 Spearfish, marlin, Makaira albida (Poey). 330 Swordfish, Xiphias gladius Linnaeus 330 Pilotfish, Naucrates ductor (Linnaeus) 331 Rudderfish, Seriola zonata (Mitchill) 331 Mackerel scad, Decapterus macarellus (Cuvier and Valenciennes) 331 Bulletin No. 20, approved for publication May 18, 1936. Contribution No. 119 of the Woods Hole Oceanographic Institution; 80054—36 3 19 320 BULLETIN OF THE BUREAU OF FISHERIES Page Saurel, Trachurus trachurus (Linnaeus) 331 Bis-eyed scad, Trachurops crumenopthalma (Bloch) 331 Hardtail, Caranx hippos (Linnaeus) 331 Hardtail, Caranx cry sos (Mitchill) 331 Lookdown, Selene vomer (Linnaeus) 332 Leather] acket, Oligoplites saurus (Bloch and Schneider) 332 Bluefish, Pomatomus saltatrix (Linnaeus). 332 Common dolphin, Coryphaena hippurus Linnaeus 332 Opah, Lampris regius (Bonnaterre) 332 Johnson’s sea bream, Taractes princeps Johnson 332 Butterfish, Poronotus triacanthus (Peck).. 332 Harvestfish, Peprilus alepidotus (Lin- naeus) 333 Striped bass, Roccus lineatus (Bloch) 333 Sea bass, Ccntropristes striatus (Linnaeus) _ 333 Triggerfish, Balistes carolinensis Gmelin_. 333 Filefish, Monacanthus hispidus (Linnaeus) 334 Filefish, Monacanthus ciliatus (Mitchill). 334 Unicornfish, Alutera scripta (Osbeck) 334 Puffer, Spheroides maculatus (Bloch and Schneider) 334 Rosefish, Sebastes marinus (Linnaeus) 334 Black-bellied rosefish, Helicolenus dactyl- opterus (De la Roche) 334 Hook-eared sculpin, Artediellus uncinatus (Reinhardt) 335 Mailed sculpin, Triglops ommatistius Gil- bert 335 Longhorn sculpin, Myoxocephalus octo- decimspinosus (Mitchill) 335 Deep-sea sculpin, Cottunculus microps Collett 335 Sea raven, Hemitripterus americanus (Gmelin) 335 Sea snail, Neoliparis atlanticus Jordan and Evermann 336 Striped sea snail, Liparis liparis (Lin- naeus) 336 Red-winged sea robin, Prionotus strigatus (Cuvier) 336 Remora, Remora remora (Linnaeus) 336 Rock eel, Pholis gunnellus (Linnaeus) 336 Snake blenny, Lumpenus lampetraeformis (Walbaum) 336 Shanny, Leptoclinus maculatus (Fries) Arctic shanny, Stichaeus punctatus (Fabri- cius) Radiated shanny, XJlvaria subbifur cata (Storer) ... Wrymouth, Cryptacanthodes maculatus Storer Spotted wolffish, Anarrhichas minor Olaf- sen Eelpout, Zoarces anguillaris (Peck) Wolf eel, Lycenchelys verrillii (Goode and Bean) Silver hake, Merluccius bilinear is (Mitch- ill)- Pollock, Pollachius virens (Linnaeus) Cod, Gadus callarias Linnaeus Haddock, Melanogrammus aeglefinus (Lin- naeus) Long-finned hake, Urophycis chesteri (Goode and Bean) Spotted hake, Urophycis regius (Wal- baum) Four-bearded rockling, Enchelyopus cim- brius (Linnaeus) Cusk, Brosme brosme (Muller) Common grenadier, Macrourus bairdii Goode and Bean American plaice, Hippoglossoides plates- soides (Fabricius) Four-spotted flounder, Paralichthys oblon- gus (Mitchill) Rusty dab, Limanda ferruginea (Storer) __ Winter flounder, Pseudopleuronectes ameri- canus (Walbaum) Georges Bank flounder, Pseudopleuronectes dignabilis Kendall Witch flounder, Glyptocephalus cynoglossus (Linnaeus) Gulf Stream flounder, Citharichthys arcti- frons Goode American goosefish, Lophius americanus Cuvier and Valenciennes Sargassum fish, Histrio histrio (Linnaeus) . Deep-sea angler, Mancalias uranoscopus (Murray) Bibliography Page 337 337 337 337 337 337 338 338 338 338 339 339 339 339 339 340 340 340 340 340 341 341 341 342 Introduction Since the publication by the Bureau of Fisheries of “Fishes of the Gulf of Maine” (Document No. 965, U. S. Bureau of Fisheries, Bigelow and Welsh, 1925) enough new information of general interest has come to hand regarding abundance, dis- tribution, migrations, breeding habits, and food habits to warrant the issuance of a supplement to that publication. Many of these data have been obtained during the investigations carried on by the Bureau; part have been collected from corre- spondence, while part have been gleaned from published material. Brief notes 321 FISHES OF THE GULF OF MAINE and records of distribution have been taken from the Bulletin of the Boston Society of Natural History (see Firth (1931), Kendall (1931), MacCoy (1929, 1931a, 1931b, 1933), Schroeder (1931)); from Reports of the Newfoundland Fishery Research Commission (1932-1933); and from the Proceedings and Transactions of the Nova Scotian Institute of Science (Leim 1930). For the distribution of certain New England sharks in South African waters, not referred to in this paper, the reader is referred to Barnard (1925). For allowing us the use of unpublished notes we wish to thank F. E. Firth, Dr. G. W. Jeffers, Dr. A. H. Leim, Walter H. Rich, and O. E. Sette. The nomenclature used in this supplement is as in “Fishes of the Gulf of Maine.” Hagfish, Myxine glutinosa Linnaeus Recent detailed studies of the sex organs make it certain that the hag is not functionally hermaphroditic as was formerly supposed, but that in each individual either the male portion of the common sex organ matures, with the female organ remaining rudimentary, or vice versa (Conel, 1931). The fact that a 60 cm speci- men from Georges Bank contained 30 eggs, 20-25 mm long, shows that large females may produce somewhat more and slightly larger eggs than previously recorded. Sea lamprey, Petromyzon marinus Linnaeus The known range of the sea lamprey in the western Atlantic has been extended northward to the west coast of Greenland (Jensen, 1926). Smooth dogfish, Mustelus mustelus (Linnaeus) The genus Mustelus is established for this species by an opinion rendered by the International Commission on Zoological Nomenclature (Smithsonian Institution, 1926, p. 8). Smooth dogfish are taken so seldom in winter that capture of three by a trawler off Bodie Island, N. C., in 34-45 fathoms, February 1931, is of interest. Great blue shark, Prionace glauca (Linnaeus) The International Commission on Zoological Nomenclature (Smithsonian In- stitution, 1925, p. 27) has rejected Valmont’s name, Galeus; consequently the correct generic name of the species is Prionace Linnaeus. The blue shark has recently been recorded from the southwest part of the Grand Bank (Rept., Nfld. Fish. Res. Lab., 1935, p. 79). Although formerly con- sidered a stray in the Gulf of Maine, recent observations have shown the blue shark to be common there in August and September, with occasional records for July. While most often seen offshore, a number were observed and several caught by J. W. Lowes during the summer of 1935 in Massachusetts Bay. Young ones are seldom seen along our shores, but Robert Goffin reports one only 20 inches long from Menemsha Bight, near Woods Hole, Mass., August 31, 1925; while F. E. Firth records another, 38 inches long, taken 65 miles southeast of Highland Light, Cape Cod, on October 23, 1930. Dusky shark, Carcharhinus obscurus (LeSueur) The capture of an 11-foot fish on the northeast peak of Georges Bank, August 10, 1931, extends the known range to the offshore banks (Firth, 1931, p. 9). 322 BULLETIN OF THE BUREAU OF FISHERIES Shovel-head shark, Cestracion tiburo (Linnaeus) One specimen of this southern species was recorded by Garman (1913, p. 160) from Massachusetts Bay. This record was omitted from “Fishes of the Gulf of Maine” (Bigelow and Welsh, 1925). Hammerhead shark, Cestracion zygaena (Linnaeus) Captures of a 12-foot fish, in August 1928, by the swordfishing schooner Doris M. Hawes, between Browns and Georges Banks, and of a small one in Halifax Harbor, September 1932 (Vladykov, 1935, p. 8), extend the known range to the northward and eastward. Thresher, Alopias vulpinus (Bonnaterre) The International Commission on Zoological Nomenclature rejects Yalmont’s name, Vulpecula marina, consequently the next oldest name, Alopias vulpinus, must be substituted (Smithsonian Institution, 1925, p. 27). Mackerel shark, Isurus nasus (Bonnaterre) The fact that Isurus punctatus (Storer) is identical with 7. nasus (Bonnaterre) has been pointed out by Bigelow and Schroeder (1927). The range of the mackerel shark in the western side of the Atlantic is now known to extend as far north as the Grand Bank of Newfoundland (Rept., Nfld. Fish. Res. Lab., 1935, p. 79). Sharp-nosed mackerel shark, Isurus tigris (Atwood) The many recent fishery investigations in the Gulf of Maine have indicated that this species is much less common there than 7. nasus, for whereas many of the latter have been observed and captured since 1923, only one record of the sharp-nosed mackerel shark has come to our attention within that tune, a fish 8% feet long taken 10 miles northeast of Nantucket Lightship, June 22, 1930, by the schooner Linta (Firth, 1931, p. 8). White shark, Carcharodon carcharias (Linnaeus) To the few existing Gulf of Maine records of this ferocious shark are added that of a 13-foot fish taken off Portland in a gill net during October 1931 (identified by Dr. W. C. Kendall); one (identified from a tooth) which attacked a fishing boat off Digby Gut, Bay of Fundy, July 2, 1932 ; 1 a somewhat doubtful record from off Halifax, June 27, 1930; 2 and another, 15 feet long, apparently of this species, caught off Monomoy Point, Cape Cod, in the fall of 1928. Spiny dogfish, Squalus acanthias Linnaeus The spiny dogfish is now known along the American coast as far northward as the Straits of Belle Isle. It has also been taken off the west coast of Greenland at Sukkertoppen and in the vicinity of Holsteinborg (Jensen, 1914, p. 7). 1 Harry Piers, Proc., Nova Scotian Institute ol Science, vol. XVIII, pt. 3, p. 198, 1934, » Ibid., p. 196. FISHES OF THE GULF OF MAINE 323 The winter home of the spiny dogfish off the American coast has long been a subject of uncertainty, hence it is of interest to record that the schooner Victor found them plentiful about 90 miles southeast of Ambrose Channel Lightship on the tile- fisli grounds the middle of January 1928, and that the Albatross II trawled many specimens in February 1931, between Cape Hatteras, N. C., and Cape Henry, Va., in 16 to 70 fathoms of water. It appears probable, therefore, that the continental slope to the southward of New England is the chief wintering ground of this species on this side of the Atlantic. Analysis of the sizes and of the stages of development of embryos in females taken at vaiious dates and localities along the coast, and of recent captures of new born dogfish, also adds to our knowledge (previously scanty) of the breeding habits. Up until 1925 we had no record of new born dogfish within the Gulf of Maine. This, together with the facts that females containing large embryos had been often taken there in early autumn, that dogfish depart entirely from the gulf over the winter, and that new born young had been reported off Long Island in summer suggested that the area of reproduction of this species is confined to waters west and south from Cape Cod. This is not the case, however, for during the past few years, when special watch has been kept for new born dogfish, we have learned of their presence in considerable numbers on Nantucket Shoals and at various localities in the Gulf of Maine from June to August. Evidently, then, the gulf, as well as the waters off southern New England, is an important nursery. The lact that embryos, sometimes with yolk sac nearly absorbed, have repeatedly been found in females off New York in autumn and on the wintering grounds off Virginia and North Carolina in January and Feb- ruary might suggest that the coastal waters of the Middle Atlantic States also so serve. As no new born “dogs” have yet been reported to the southward of New York at any season, this question remains open, however. If it should prove that young are born in the southern wintering as well as in the northern summering grounds, the sizes of the embryos, at different localities and dates, would suggest that some are set free as early as January or February; in other words, that the season extends from midwinter right through the spring and summer. Greenland shark, Somniosus microcephalus (Bloch and Schneider) So seldom is the Greenland shark captured in the Gulf of Maine that it is of interest to record a large one taken off Portland Lightship the summer of 1926, and four others, 4 to 5 feet long, taken in* the offing of Portland from 1927 to 1933. A large one was also taken somewhere in the gulf and brought into Gloucester in Jan- uary 1929, and another, about 15 feet long, caught in an otter trawl on Jeffreys Ledge 27 miles northeast of Thatchers Island, off Cape Ann, February 16, 1931. KEY TO SKATES AND RAYS Experience has shown that existing keys are not adequate for the identification of Gulf of Maine skates and rays. The following revision is therefore offered: 1. No long dorsal spine on the tail 2 Tail with long dorsal spines (sting rays) 11 2. Two small dorsal fins, but no distinct caudal fin (includes all our common skates) 3 There is a large triangular caudal fin as well as the two dorsals Torpedo 3. Ventral surface with minute rounded tubercles Raja granulata, 3 Ventral surface smooth 4 3 Although Raja granulata Is not known from the Gulf of Maine it caD be expected there as it has been recorded from La Have Bank and from the continental edge off Halifax, Nova Scotia, from 200 fathoms. 324 BULLETIN OF THE BUREAU OF FISHERIES 4. No thorns along mid-zone of disc between eyes and ventrals. Barn-door skate, Raja stabuliforis With one or more rows of thorns along mid-dorsal zone of disc behind eyes 5 5. Posterior third of tail without any large thorns Smooth skate, Raja senta Posterior third of tail with one or more rows of large thorns 6 6. Tail with only one row of large thorns Young Raja scabrata Tail with three or more rows of thorns 7 7. Mid-row of tail thorns very much larger than any other thorns on tail. Prickly skate, Raja scabrata Mid-row of tail thorns absent or if present not much larger than other thorns on tail 8 8. Three rows of thorns on tail Brier skate, Raja eglanteria Four or more rows of thorns on tail 9 9. Length of fish more than 2J4 feet Big skate, Raja diaphanes Length of fish less than 2J4 feet 10 10. Teeth in 70 to 104 rows in each jaw; usually an eye-spot present on each pectoral. Big skate, Raja diaphanes Teeth in 38 to 60 rows in each jaw; eye-spot rarely present Little skate, Raja erinacea 11. No dorsal fins on tail Sting ray, Dasybatus marinus Tail with a dorsal fin ii: front of spine Cow-nosed ray, Rhinoptera quadriloba With regard to the relative abundance of different species of skates on the off- shore fishing banks of the Gulf of Maine, it is of interest that on a trip to Georges Bank (chiefly the northeastern part) in September 1929, aboard the otter trawler Kingfisher, 37 hauls yielded from 0 to 105 skates per haul (total 495), as follows: Baja senta, 57; B. scabrata, 325; B. stabuliforis, 42; and B. diaphanes, 71. Little skate, Baja erinacea Mitchill This skate has been described as lacking thorns along the midline; but small specimens 3% to 9 % inches long and one half-grown specimen of 13)£ inches, recently examined by us, have this row well developed. Big skate, Baja diaphanes Mitchill Recent investigations have shown that the range of this skate extends north- ward not only to the Gulf of St. Lawrence, as long known, but to the Grand Banks as well, and southward to Virginia. The big skate rarely has a median row of thorns except in the very young, so it is of interest to record a female 18 inches long taken near Jeffreys Ledge, November 1, 1927, which bears a row of large spines along the midline, from the shoulder girdle to the origin of dorsal on the tail. Young specimens of B. diaphanes and B. erinacea, especially the females, are not easily separated from each other by a casual glance, hence the number of rows of teeth which they possess has been an important means of identification. Investi- gators have given various tooth counts ranging from 80 to 110 rows for diaphanes and around 50 rows for erinacea. Several specimens in the Museum of Comparative Zoology, from New England waters, show 70 to 100 rows of teeth on the jaw of diaphanes and from 46 to about 60 rows in erinacea. Prickly skate, Baja scabrata Garman At the time of publication of “Fishes of the Gulf of Maine” the northern boundary of this skate (widespread in the Gulf of St. Lawrence) was unknown in the open Atlantic. Since then it has been found plentifully on the Grand Banks and reported from the east and north coasts of Newfoundland. FISHES OF THE GULF OF MAINE 325 Examination of a large number of prickly skates, ranging in size from young, recently hatched, to the largest recorded, allows us to add the following to previous descriptions: The pavementlike teeth are in 41 or 42 rows in the upper jaw, 40 to 44 rows in lower jaw (4 specimens) and in the male, at least, there are rather sharp cusps on those teeth situated toward the angles of the jaw. The number of large curved thorns along the midline of the tail and body was as follows on 23 specimens: 12 (2), 13 (10), 14 (8), 15 (3), with no correlation between the number of thorns and the size of the specimen. The brownish back is usually marked with small white spots. The 3roung are more spotted than the adults and have six or seven dark cross bars on the upper surface of the tail. This skate grows to 3 feet in length, or slightly larger; a 32-inch fish is about 23 inches wide. The smallest nearly mature male found was 26 inches long. Since nothing was known of the breeding habits of the prickly skate, it is worth recording that a specimen 32 inches long taken on the northern part of Georges Bank, September 22, 1929, had one egg capsule measuring 3 by 2% inches (exclusive of tendrils) in each oviduct, and that a male, 35% inches long had nearly ripe milt, but a number of other large females taken at the same time were barren. Prickly skates caught on Georges Bank in September 1929, had been feeding on fish, shrimps, spider crabs, anemones, and worms; this is the first definite information as to the diet of this skate. Brier skate, Raja eglanteria Bose An unusually large one, 37 / inches long, was taken off Woods Hole, Mass., in August 1932. Additional to the few Gulf of Maine records of this skate already reported are two specimens taken on Nantucket Shoals, near Round Shoal Buoy, by the Halcyon, one in July, the other in September 1924. This is a shoal water species, the deepest capture made by Albatross II between southern New England and the offing of Chesapeake Bay being in 38 fathoms. Smooth skate, Raja senta Garman The smooth skate, formerly believed rare in the Gulf of Maine, is now known to be quite generally distributed on our offshore fishing banks, as well as on soft bottom in the deeper parts of the gulf. We have taken it commonly on Georges Bank, in South Channel, in the deep water (80-100 fathoms) just off Cashes Ledge, near Jeffreys Ledge, and off Chatham. The slioalest capture was from 25 fathoms. The largest specimen obtained was 24 inches long. Barn-door skate, Raja stabuliforis Garman Young specimens are seldom reported, hence it is of interest to record one of 7% inches taken on the western edge of Nantucket Shoals July 14, 1930, in 28 fathoms. This fish had essentially the same characters as the adult. The range of the barn door skate is now known to extend northward to the western part of the Grand Bank of Newfoundland. Torpedo, Narcacion nobilianus (Bonaparte) As no torpedoes had been reported to the eastward of Cape Cod since 1896, the captures of a 52-incli specimen weighing 78 pounds, on the southwest part of 80054 — 36— — 2 326 BULLETIN OF THE BUREAU OF FISHERIES Georges Bank, December 8, 1930, and of another of 39 inches at Provincetown, July 28, 1931, deserve mention here. Chimaera, Chimaera ajfinis Capello It was formerly believed that the chimaera did not exceed a length of about 3 feet, but a specimen measuring 49 inches in length, 17% pounds in weight dressed, was taken October 15, 1930, 85 miles south by west of Cape Sable in a depth of about 400 fathoms (Firth, 1931, p. 9). Common sturgeon, Acipenser sturio Linnaeus Although sturgeon have seldom been reported from offshore, the recent captures of a 268-pound fish in South Channel the end of April 1928, of another of 420 pounds in April 1929, of a 335-pound fish trawled on Browns Bank in April 1936, and of a 435-pound fish on Georges Bank, latitude 41°00' N., longitude 67°45' W., on January 7, 1931, indicate that they are to be occasionally found on our outer fishing banks. Short-nosed sturgeon, Acipenser brevirostris LeSueur A 30-inch specimen, taken at Provincetown about 1907, now in the collection of the Museum of Comparative Zoology, is the only reliable record for the Gulf of Maine. This record was omitted from “Fishes of the Gulf of Maine” (Bigelow and Welsh, 1925). Eel, Anguilla rostrata (LeSueur) The known range of the American eel in northern waters has been extended to the west coast of Greenland (Jensen, 1926, p. 101). American conger, Conger oceanica (Mitchill) The American conger, long considered identical with the European, has recently been shown by Schmidt (1931) to be a distinct species, characterized by having fewer (140-149) vertebrae than the European (154-163 vertebrae); a relationship paralleling that between the American and European eels of the genus Anguilla. The American conger ranges along the continental shelf northward to Cape Cod. Its southern boundary cannot be stated until the congers of the coasts of North and South America have been critically compared. Additional to the few records of larvae already reported from the Gulf of Maine are those of two specimens (4% inches long) picked up on the beach at Newburyport, Mass., in November 1929, which were sent to us for identification. Dr. Johannes Schmidt’s discovery 4 of very young larvae in the West Indian region, but nowhere else, points to this as the chief, if not the only, spawning ground of the American conger. Snake eel, Pisodonophis cruentifer Goode and Bean Goode and Bean’s (1896, p. 147) record of this species from Jeffreys Bank — the only one for the gulf — was omitted from “Fishes of the Gulf of Maine.” A number of specimens have been taken recently between the offings of Nantucket and of Cape Henry, Va., in depths ranging from 24 to 245 fathoms by the Fish Hawk and the Albatross II. 1 See Schmidt, 1931, p. 602, for a discussion of this question. FISHES OF THE GULF OF MAINE 327 Herring, Clupea harengus Linnaeus The northern limit to the known range of the herring in the Western Atlantic has been extended to the west coast of Greenland by Jensen (1926, p. 101). Herring are so seldom taken in any large numbers on the offshore banks that it is of interest to record a catch of 2,800 pounds in South Channel and 3,000 pounds on the northern edge of Georges Bank, in October 1931. Mass destruction of young herring, cast up on the beaches has occurred from time to time in various harbors in the Gulf of Maine. A recent occurrence of this sort was reported by Dr. Austin H. Clark, who, in Manchester Harbor on the north side of Massachusetts Bay, early in August 1925, observed that the mud flats were white with stranded herring which measured 3 to 5 inches in length. Another such destruction took place in the same harbor in the summer of 1928. Alewife, Pornolobus pseudoharengus (Wilson) So little is known about the habits or migrations of the alewife while at sea that it is of interest to record the capture by Albatross II of 18 adults, 10 to 11 inches long, by otter trawl, seventy odd miles off Barnegat, N. J., on March 5, 1931. Blueback, Pornolobus aestivalis (Mitchill) The maximum length of this herring is usually given as about 1 foot but we have seen examples of it ranging up to 15 inches. The capture of seven adult specimens by Albatross II, on March 5, 1931, about 70 miles off Barnegat, N. J., suggests that, like its relative the sea herring, the blueback moves out from land, and passes the cold season near the bottom, thus throwing some light on the probable winter home of the Gulf of Maine stock. Thread herring, Opisthonema oglinum (LeSueur) The capture of a single specimen, 7 inches long, off Monomoy Point at the southern angle of Cape Cod in August 1931, extends the known range of this southern herring to the Gulf of Maine. Occasionally the thread herring is taken off southern New England; it was even reported as rather common in Buzzards Bay and Vine- yard Sound in the summer of 1885. As it is essentially a tropical fish it is not apt, however, to reach the gulf except as the rarest of strays. Round herring, Etrumeus sadina (Mitchill) This herring, recorded by Bigelow and Welsh (1925, p. 91) as Etrumeus teres DeKay, appears very rarely to stray past Cape Cod. Hence, it is of interest to record the capture of one specimen in Yarmouth River which empties into Casco Bay, and one in the bay itself, on September 15, 1924. Anchovy, Anchoviella mitchilli (Cuvier and Valenciennes) This species is listed by Bigelow and Welsh (1925, p. 124) as Anchovia mitchilli. The subgenus Anchoviella Fowler differs from the subgenus Anchovia Jordan and Evermann chiefly by having much fewer gillrakers, the former having about 35 to 50 and the latter 100 or more. 328 BULLETIN OF THE BUREAU OF FISHERIES Striped anchovy, Anchoviella epsetus (Bonnaterre) A record from off the Presumpscot River, near Portland, October 8, 1930 (Ken- dall 1931, p. 11) is the first for the Gulf of Maine. This anchovy is now known from as far northward as Halifax harbor where a number were seined September 29, 1931 (Vladykov 1935, p. 3). Argentine, Argentina silus Ascanius Until recently the argentine was considered rare in our waters, for only odd examples had been brought in from widely scattered localities. The development of otter trawling proved that argentines are in reality fairly common around the edge of Georges Bank and off Cape Cod in deep water. Thus, along the northern and northwestern slopes of the bank and to the eastward of Cape Cod, in depths of 80 to 100 fathoms, it is not unusual for a haul of the trawl to bring in from one to a dozen, and as much as 15,000 pounds has been reported by one boat during a week’s fishing (Firth 1931, p. 11). It also occurs in the deep central basin of the gulf, for the Albatross II has recently (July 1931) trawled a specimen in 115 fathoms off Mount Desert Rock. Pearlsides, Maurolicus pennanti (Walbaum) Additional Gulf of Maine records of this species include one specimen 41 mm long taken from the stomach of a cod, on Platts Bank, July 27, 1924; one 43 mm long, also from a cod’s stomach, on Cashes Ledge, August 16, 1928; and four, 32 to 39 mm long, from the stomachs of two pollock, caught in 20 fathoms, 7 miles southeast of Bakers Island, Mount Desert, Maine, July 24, 1930. Viper fish, Chauliodus sloanei Bloch and Schneider A specimen found in the stomach of a swordfish caught in the gully between Browns and Georges Banks in 1931 is the second to be definitely recorded from within the Gulf of Maine. Lancetfish, Alepisaurus jerox Lowe A record of a 5 /(-foot specimen of this rare fish caught alive in the surf on Block Island, R. I., March 12, 1928, is of especial interest even though outside the limits of the Gulf of Maine. An excellent photograph, sent in by Mrs. Elizabeth Dickens, shows the upper lobe of the caudal prolonged as a long filament, which most of the specimens so far seen have lost. This specimen had been feeding on small dogfish. Needlefish, Scomberesox saurus (Walbaum) A specimen gaffed at the surface from the Albatross II on northern Georges Bank, September 20, 1928, is the only definite offshore record for the Gulf of Maine although the needlefish has been taken in various localities there alongshore. Trumpetfish, Fistularia tabacaria Linnaeus Recent reports of the trumpetfish at Port Mouton, Nova Scotia, and on the south coast of Newfoundland, show that this tropical species may stray much farther north than previously supposed. FISHES OF THE GULF OF MAINE 329 Pelagic pipefish, Syngnathus pelagicus Linnaeus A single specimen 3K inches long, taken on Georges Bank (lat. 42°09/ N., long. 66°41' W.) September 20, 1927, by the Albatross II, is the only Gulf of Maine record. This specimen was dipped up with a mass of gulf weed (Sargassum) and was the only one found in a large amount of weed that was examined. Common pipefish, Syngnathus fuscus Storer Pipefish are rarely taken on bottom far from the immediate shore waters, hence it is of interest to report the capture of four specimens 4% to 6 inches long at a depth of 19 fathoms 10 miles south of No Mans Land, February 5, 1930. Northern barracuda, Spliyraena borealis DeKay A specimen about 2 inches long found alive in the surf at Nauset Beach, Cape Cod, September 26, 1930, by Dr. Edward P. Richardson, is the only record thus far reported for the Gulf of Maine. Young fry, a few inches long, are taken, however, from time to time in the region of Vineyard Sound and Buzzards Bay on the southern coast of New England, from July to December. Mackerel, Scomber scombrus Linnaeus The body length of the mackerel is erroneously given as about three and one-half times the depth by Bigelow and Welsh (1925, p. 188); actually, it is four to five and one-half times the depth. A small mackerel taken at Cape Lookout, N. C., in February 1925 (Coles, 1926, p. 105), extends the known range southward beyond Cape Hatteras. Recent captures of a mackerel weighing 7 K pounds,6 and of another of 7% pounds, 26 inches long, both of which we, ourselves, examined, shows that occasional giants occur, for the weight seldom exceeds 4 pounds or the length 22 inches. Tuna, Tliunnus thynnus (Linnaeus) Larger catches of tuna have been made within the Gulf of Maine in recent years partly, at least, because of an increased market demand for the fish. Thus, compared with the 69,868 pounds recorded for Massachusetts and Maine in 1919, the catch of 1934 amounted to 356,904 pounds, of which 254,076 pounds came from Cape Cod. The Nova Scotian shore of the Gulf of Maine yielded about 24,000 pounds in 1924 and 10,000 pounds in 1929. At present the annual catch for the gulf is probably between 300,000 and 400,000 pounds. Assuming an average weight of 300 pounds (probably too little, for the average weight of about 90 tuna caught off the coast of Maine in 1926 was about 540 pounds), this would represent a thousand or more fish. Off the outer coast of Nova Scotia, where tuna have been taken in larger numbers than within the Gulf of Maine, the annual catches from 1917 to 1933 have fluctuated between 152,000 and 1,550,000 pounds. The heaviest New England fish on record, taken off Rhode Island about 1913, weighed 1,225 pounds, while four or five fish have been brought into Boston that weighed approximately 1,200 pounds. Another fish weighing 1,300 pounds was shipped in 1924 from Nova Scotia to Boston (Sella, 1931, p. 61). 4 Atlantic Fisherman, August 1925. 330 BULLETIN OF THE BUREAU OF FISHERIES Small and moderate sized tuna (below 100 pounds) are comparatively rare in the Gulf of Maine. However, schools composed of individuals estimated to weigh not more than 40 to 70 pounds were observed around Boston lightship July 13 and 14, 1935.6 None below 20 pounds has been recorded within the gulf, but off southern New England, especially near Block Island, small tuna are sometimes caught, there being an unusual run of them (8 to 12 pounds) in 1928. Thus it is probable either that the lower temperatures of the Gulf of Maine are a barrier to the smaller-sized tuna, or that they find less favorable feeding grounds there than do the larger sizes. Off the New England coast the first schools are sighted late in June or early July to the southward of Block Island, over depths of about 85 fathoms, and a few days later they appear inshore. At first the fish are hungry, and there is some reason to believe that their summer migrations follow their breeding period. An example of their seasonal abundance in the shore waters of the coast of Maine may be had from the catches made in the vicinity of Casco Bay in 1926, where about 70 fish were taken in July, 17 in August, 3 in September, and 1 on October 4. Common bonito, Sarda sarda (Bloch) Two fish were reported from the mouth of Kennebec River in July and one in September 1930, and one from southern Nova Scotia (Vladykov, 1935, p. 7) in the latter month. In looking through the records of the catches made by a certain set of pound nets at Provincetown over a period of 10 years, we find the earliest catch for that locality was in July (1915), and the latest on October 4 (1919). Spearfish, marlin, Makaira albida (Poey) 7 No spearfish were reported in the Gulf of Maine from about 1880 until 1925. Since then, however, seven specimens have been brought in, all in summer, one of them from off Portland, the others from Georges, Browns, and Sable Island Banks, the last being the most northerly record for the species in the western North Atlantic. These specimens ranged from 5 feet to nearly 16 feet in length and from 21 to about 700 pounds in weight. Additional descriptive data based on two New England specimens examined by us are as follows: The first dorsal fin of one specimen has 47 stiff rays, the other fish having 48. This fin is separated from the second dorsal by a space equal to the length of the latter in the one fish, by a shorter space in the other. The first anal fin (2 spines and about 12 or 13 rays), situated below the rear part of the first dorsal, is triangular, its first rays forming a sharp angle. Swordfish, Xiphias gladius Linnaeus The largest swordfish definitely recorded from the Gulf of Maine was one, caught in the summer of 1921 by Capt. Irving King and landed at the Boston Fish Pier, that weighed 915 pounds dressed — hence, upward of 1,000 pounds alive (Fishing Gazette, September 1921, p. 13). The specimen was not measured, but the sword being more than 5 feet, the total length of the fish must have approximated 15 feet. # Data furnished by J. W. Lowes. 7 Recorded by Bigelow and Welsh (1925, p. 227) as Tetrapturus imperator (Bloch and Schneider). FISHES OF THE GULF OF MAINE 331 In 1931, another large fish was caught, 644 pounds in weight, dressed, 13 feet in length, with a sword measuring 3 feet 8 inches. Young swordfish are so rarely reported off the New England coast that it is of interest to record the capture of a 2-foot fish, weighing 7% pounds, taken by the Dacia on a trawl line September 2, 1931, on Georges Bank. Pilotfish, Naucrates ductor (Linnaeus) Up to 1925 only three definite records for the Gulf of Maine had come to hand. Since then we have learned of the capture of six more pilotfish, off Portland, in Pro- vincetown Harbor, to the southeast of Cape Cod, and on the northern edge of Georges Bank, during the summer and fall months in the years 1921, 1924, 1929, 1931, and 1933. Vladykov (1935, p. 6) reports two specimens from Sable Island Bank and one from Sambro, off Nova Scotia, in the period 1932-34. Rudderfish, Seriola sonata (Mitchill) The known range of the rudderfish has been extended northward to Halifax, Nova Scotia (Leim, 1930, p. xlvi, as S. dumerili). One fish was caught on a smelt hook off a Portland wharf in September 1921; a 5K-inch fish was taken off Boston in September 1929; another, 17)£ inches long, from South Channel the same month; and a 6-inch specimen on Nantucket Shoals in August 1930. Mackerel scad, Decapterus macarellus (Cuvier and Valenciennes) One specimen, 7 inches long, was taken in a trap at Richmond Island, off Cape Elizabeth, in September 1931, this being only the second recorded for the Gulf of Maine. Saurel, Trachurus trachurus (Linnaeus) One specimen of this fish, rare to the northward of Woods Hole, Mass., was taken in Casco Bay on August 12 and another near Castine Bay, Maine, on October 15, 1930 (Kendall, 1931, p. 11). Big-eyed scad, Trachurops crumenopthalma (Bloch) Two specimens, recently taken off Cape Cod, one at Provincetown, the other about 8 miles off the beach at Chatham, are the only positive records of this species for the Gulf of Maine. As it is caught from time to time, however, in the summer and fall as far northward as Woods Hole, it may be expected to round the cape occasion- ally. This scad has been recorded from Canso, Nova Scotia by Cornish (1907, p. 85). Hardtail, Caranx hippos (Linnaeus) A hardtail taken off Provincetown in 1933 is the second reported from the Gulf of Maine. Several specimens about 2 inches long were taken the summer of 1933 in Musquodoboit Harbor, Nova Scotia (Vladykov, 1935, p. 4). Hardtail, Caranx crysos (Mitchill) One fish was taken off Chatham in 1933. 332 BULLETIN OF THE BUREAU OF FISHERIES Lookdown, Selene vomer (Linnaeus) During the autumn of 1933 many small lookdowns were reported from traps at the mouth of Casco Bay, one also from Beverly Farms, and another from North Truro, an unusual incursion, for only three specimens had previously been recorded in the Gulf of Maine. Jones (1882 p. 89) and Honeyman (1886 p. 328) record this species (young) as occasional in the shore waters of Nova Scotia, presumably\long the east coast. Leatherjacket, Oligoplites saurus (Bloch and Schneider) A specimen taken in a trap off the outer beach at Chatham is the only record for the Gulf of Maine. Bluefish, Pomatomus saltatrix (Linnaeus) For many years no bluefish had been reported north of Cape Ann, until 1925, when one was caught off Halifax, Nova Scotia. This seems to have presaged a tempo- rary extension of range, for numbers of them visited the inner coasts of the gulf northward to Casco Bay in the summer of 1927, while in 1930 the bluefish was again reported at Halifax (two specimens) and at Port Mouton, Nova Scotia (one specimen, Leim, 1930, p. xlvi). Common dolphin, Coryphatna hippurus Linnaeus A dolphin about 3 % feet long taken 60 miles south-southwest, of Cape Sable, in the deep gully between Browns and Georges Banks by the trawler Natalie Hammond, August 15, 1930, is the first Gulf of Maine record. The specimen is now in the collection of the Boston Society of Natural History. Opah, Lampris regius (Bonnaterre) 8 A specimen about 3 feet long was taken in July 1925, on Western Bank, southwest of Sable Island, by the schooner Falmouth (Radcliffe, 1926), while another of about the same size stranded on the beach at Hyannis, Mass., on September 17, 1928. Johnson’s sea bream, Taractes princeps Johnson A fish taken on Browns Bank, off Cape Sable in January 1928 is the first record of this species for the western Atlantic. This bream previously was known only from Madeira, in the eastern Atlantic. For a detailed account and comparison with allied species see Bigelow and Schroeder (1929). Butterfish, Poronotus triacanthus (Peck) Recent records show that the northward range of this species extends to the east coast of Newfoundland, as well as to Nova Scotia as has long been known. It now seems well established that the butterfish actually withdraw from the gulf when they disappear in the autumn, as they do from the immediate shore waters farther south, and from inland waters such as Chesapeake Bay. Until very recently the winter home of the butterfish was unknown; but as they are now often taken in the winter otter trawl fishery recently established off the coast between Chesapeake s This species was given as Lampris luva (Gmelin) by Bigelow and Welsh (1925, p. 242)i FISHES OF THE GULF OF MAINE 333 Bay and Cape Hatteras, it appears that they move out to sea to winter on the outer part of the continental shelf as do several other common Gulf of Maine fishes. The illustrations of larvae 2.1 and 3.4 mm long credited by Kuntz and Radcliffe (1918) to the butterfisli and reproduced by Bigelow and Welsh (1925, fig. 116, c and d ) have since been proved to belong to one of the hakes. Harvestfish, Peprilus alepidotus (Linnaeus) Five or six specimens were reported caught in floating traps at Richmond Island, off Cape Elizabeth, Maine, in July 1929, while another was taken at the mouth of the Damariscotta River, Maine, in August 1933, the most northerly record for the species. Striped bass, Roccus lineatus (Bloch) The striped bass considerably increased in abundance along both shores of Cape Cod between 1928 and 1932, then decreased again as illustrated by the following catches reported for Barnstable County, Mass.: 1928, 8,060 pounds; 1929, 18,665 pounds; 1930, 27,385 pounds; 1931, 33,600 pounds; 1932,30,926 pounds; 1933,4,500 pounds. Anglers as well as commercial fishermen have also caught some numbers along the Eastham-Chatham Beaches and marshes during the past few years, while a 44% pound bass was caught near Brant Rock on the southern shore of Massachu- setts Bay, in November 1930. A small stock seems also to have built up in the brack- ish tributaries of Plum Island Sound north of Cape Ann, for some were taken in Parker River by anglers during the few years previous to 1930, while in that year (when fishing restrictions were relaxed) 8,700 pounds were reported thence, though smaller numbers since then. But this increase did not extend northward beyond Massachusetts waters, for the commercial reports from the States of New Hampshire and Maine did not mention bass at all in 1924, or in 1928-33.® Striped bass so rarely stray away from the immediate shoreline that it is of interest to mention the capture of a 6-pound fish in a gill net on Cod Ledge, 3 or 4 miles off Cape Elizabeth, Maine, October 15, 1931. Sea bass, Centropristes striatus (Linnaeus) Sea bass are seldom taken within the Gulf of Maine, and even on the southern New England coast are rarely caught later than early November, hence the reported capture of a 5-pound fish in December 1930, 5 miles east of Pollock Rip Lightship, in 24 fathoms, is noteworthy. Triggerfish, Balistes carolinensis Gmelin Previous to 1925, only one specimen of the trigger fish had been reported from the Gulf of Maine. Actually, this species must drift over the offshore rim of the gulf more often than the paucity of early records would suggest, for a specimen was recorded from Casco Bay in August 1931 ; another was taken in 1932 near Plymouth; a third, 15 inches long, was gaffed at the surface, on the southeast part of Georges Bank, from the fishing vessel Huntington Sanford, in July 1929; and two small fry, 2 to 3 inches in length, were picked up on the northeast part of the bank in mid- September 1927, by the Albatross II. The fact that these last were taken with gulf weed (Sargassum) suggests that triggerfish are most apt to appear on the banks with the latter. • No statistics are available for 1925-27. 334 BULLETIN OF THE BUREAU OF FISHERIES Filefish, Monacanthus hispidus (Linnaeus) The filefish appears in the inner parts of the gulf only as a stray from warmer seas, recent records being that of a fish taken off Seguin, September 12, 1929, one off Portland lightship, July 17, 1931, and a 6-inch fish at Provincetown, November 6, 1929. On the offshore banks, however, it is to be expected more frequently (which accords with its southern origin) for the Albatross II gathered 181 small fry 1 to 2 inches long, on the northeastern part of Georges Bank among floating gulf weed (Sargassum) in September 1927 ; while a larger one was picked up to the southeast of Cape Cod in that same month of 1930. Filefish, Monacanthus ciliatus (Mitchill) A 7-inch fish taken in a Provincetown trap in November 1929 is the second (and only recent) record of this species within the Gulf of Maine (Firth, 1931, p. 13). A straggler has been reported, however, from Newfoundland — far to the north of its previously known range. Unicornfish, Alutem scripta (Osbeck) Two specimens of this fish, 5 and 5 % inches long, respectively, caught on the west- ern edge of Georges Bank, constitute the first Gulf of Maine record (Mac Coy, 1931a, p. 16). Puffer, Spheroides maculatus (Bloch and Schneider) A specimen taken off Long Island, Portland Harbor, on July 24, 1933, is the first to be recorded from the northern boundary (Casco Bay) of this species since 1896. Rosefish, Sebastes marinus (Linnaeus) It is now known that rosefish may be born in the Gulf of Maine as early as the end of April, for in 1930 we saw gravid females during the last half of that month. In July 1931 the Albatross II trawled many gravid females, 10 to 13 % inches long, in the central basin of the gulf; one of these, 13 inches long, contained approximately 20,500 young 6 to 7 mm long, ready to be spawned. The fact that we obtained many young fish 2% to 5 / inches in length, off the coast of Maine from April to August, suggests that this is the approximate size attained during their first year of life. Recent catches of 75-625 rosefish per haul in a trawl by the Atlantis in 70-130 fathoms in the western and northeastern parts of the gulf are evidence of the abun- dance of this species over the soft bottoms of the basins, as well as in other parts of the gulf. The commercial importance of this species has greatly increased of late, the reported landings having risen from 1,288,000 pounds in 1934 to 14,100,000 pounds in 1935. Black-bellied rosefish, Helicolenus dactylopterus (De la Roche) A fish 13 inches long, trawled on the eastern edge of Georges Bank in 150 fathoms, October 6, 1929 (Firth, 1931, p. 13), is the first record for this species within the Gulf of Maine. In addition to previous records from outside the gulf, a number of small fish (1 / to 3 y2 inches) were trawled off southern New England in 80 to 118 fathoms during 1930. FISHES OF THE GULF OF MAINE 335 Hook-eared sculpin, Artediellus uncinatus (Reinhardt) 10 This sculpin is now known to be generally distributed in the Gulf of Maine in depths greater than 20 to 30 fathoms. Thus, in addition to the Massachusetts Bay records of many years ago, we have recently taken it repeatedly near Mount Desert, off Cape Elizabeth, near Jeffrey’s Ledge, around Cashes Ledge, along the northern slopes of Georges Bank, in the southeastern part of the basin of the gulf, and at the entrance to the deep gully between Georges and Browns Banks, in depths ranging from 20 to 150 fathoms. Individual hauls have yielded up to six or eight specimens, both on hard and on soft bottom. After examining specimens from New England waters and comparing published drawings of European fish, we can find no major differences between the hook-eared sculpins of the eastern and western Atlantic.* 11 Mailed sculpin, Triglops ommatistius Gilbert Tiffs sculpin is not as rare in the Gulf of Maine as was formerly supposed, for during the past few years we have trawled specimens near Mount Desert, in Mas- sachusetts Bay, off Cape Ann, off Cape Cod, and around the northern slope of Georges Bank, in depths of 20 to 140 fathoms, in various months from spring to autumn. The most southerly locality was about 10 miles east of Chatham. Longhorn sculpin, Myoxocephalus octodecimspinosus (Mitchill) Numerous young specimens 1% to 2 inches long taken in September, and 3 to 3% inches in February, suggest that the longhorn sculpin is about 2 to 3 inches long at 1 year of age, spawning as it does in late fall. Deep-sea sculpin, Cottunculus microps Collett A specimen, about 2 inches long, trawled by the Albatross II on the northern slope of Georges Bank, in a depth of 120 fathoms, on July 24, 1931, is the third record for the Gulf of Maine proper. Sea raven, Hemitripterus americanus (Gmelin) The fact that fish of both sexes with gonads only partially developed have recently been found on Nantucket Shoals late in June, added to previous captures of ripe females off southern New England in November and December shows this to be a late fall and early winter spawner. The sea raven is a prolific fish, for a female 20 inches long that we caught off Boothbay Harbor, Maine, in April 1925, contained about 10,000 eggs. The fact that these were definitely of two sizes, the smaller averaging 1.5 mm in diameter, the larger about 3 mm, raises the interesting question whether individual sea ravens may spawn more than once during the year. The sizes of the few young sea ravens that have been taken in the Gulf of Maine suggest that they reach a length of 2 to 4 inches by the middle of the first summer, when 6 to 8 months old; and about 6 inches by the following April, at an age of 1% years. 10 Given as Artediellus atlanlicus Jordan and Evermann by Bigelow and Welsh (1925, p. 314). 11 Jordan, Evermann, and Clark (1930, p. 377) in the Check List of Fishes placed Artediellus attauticus Jordan and Evermann in the synonymy of A. uncinatus Reinhardt 336 BULLETIN OF THE BUREAU OF FISHERIES Sea snail, Neoliparis atlanticus Jordan and Evermann The sea snail, previously unknown offshore, has recently been taken on Georges and on Browns Banks. Its range has recently been found to extend as far southward as the offing of Atlantic City, N. J. (Lat. 39°20'N.). Most of the specimens were found living in scallop shells ( Peden magellanicus), as is so often the case. Striped sea snail, Liparis liparis (Linnaeus) This sea snail was formerly known as far southward as New York but the Alba- tross II has taken it off Delaware Bay and the Grampus off Assateague, Va. (Welsh, 1915, p. 2). Red-winged sea robin, Prionotus strigatus (Cuvier) A specimen was taken off Monhegan, Maine, in 40 fathoms, in an otter trawl November 19, 1933. This is the most northerly record for this straggler in the Gulf of Maine. Remora, Remora remora (Linnaeus) Recent Gulf of Maine records of this species include one found on the bottom of a lobster trap in Portland Harbor in 1931, probably brought in by a schooner from the West Indies; one found sucking to the gills of a blue shark ( Prionace glauca) that was caught on the northeast edge of Georges Bank, August 1, 1931; one in Cape Cod Bay in September 1934, and one off Provincetown in August 1935, taken by C. W. Lowes on blue sharks; also two specimens, 6 and 17 inches long, respectively, taken on August 3, 1932, 220 miles east-southeast of Cape Ann. Previously it had been recorded only once from the Gulf of Maine. Rock eel, Pholis gunnellus (Linnaeus) Recent records show that the rock eel occurs in considerable numbers ou the offshore banks in the Gulf of Maine down to at least 40 fathoms and occasionally even to 100 fathoms (Schroeder, 1933, p. 5) as well as inshore. So many have been found in the stomachs of cod and pollock caught on Nantucket Shoals, Georges Bank, Browns Bank, Cashes Ledge, etc., that it must be an important food of these two species. The range of the rock eel recently has been found to extend south to the latitude of Delaware Bay, where in February 1930 Albatross II trawled two specimens in 23 and 38 fathoms, respectively. Snake blenny, Lumpenus lamp etraejor mis (Walbaum) Recent captures, by Albatross II, of adult snake blennies (one specimen each) off Mount Desert, off Boone Island, and on Stellwagen Bank, in depths ranging from 28 to 88 fathoms, added to earlier records from Massachusetts Bay and from the Bay of Fundy region, show that this species is generally distributed over the gulf, as records of its larvae had suggested. So slender and active is this fish that it can easily escape through the meshes of any of the nets used by commercial fishermen, hence it is seldom reported. Color notes taken from a 12-inch specimen are as follows: The body had brown markings on a whitish ground, the head being pale brown. The dorsal fin was marked obliquely with 18 pale bars, the caudal transversely with 8. The anal rays were pale brown against a colorless membrane, the ventrals white, while the pectorals were tinged with brown. FISHES OF THE GULF OF MAINE 337 One of 19 inches caught on the eastern slope of Stellwagen Bank in 42 fathoms in July 1931 is the largest on record. Shanny, Leptoclinus maculatus (Fries) One specimen of this stray from the north was trawled on the northeast part of Georges Bank in August 1926 and four (4 to 4/ inches long) were taken off Chatham, Cape Cod, in 28 fathoms, May 1, 1930, by the Albatross II. This is the most southerly record for the species. Arctic shanny, Stichaeus punctatus (Fabricius) A specimen 4% inches long of this arctic species, taken one-half mile off Little Duck Island near Mount Desert, Maine, from the stomach of a cod, on April 30, 1930, is the first record for the Gulf of Maine; the only record indeed to the southward of New- foundland. This specimen was in such good condition that it unquestionably had been living in the immediate vicinity. Radiated shanny, Ulvaria subbifurcata (Storer) This shanny was previously known to be rather common in the northeastern part of the gulf, and enough have now been found in the stomachs of cod caught on Cashes Ledge, Georges Bank, Nantucket Shoals, and other offshore grounds to show that it is widespread in other parts of the gulf as well, on hard bottom. The deepest capture was in 45 fathoms. Wrymouth, Cryptacanthodes maculatus Storer Recent captures of two specimens in the central basin of the Gulf of Maine, July 1931, in 88-95 fathoms, of three in August 1936, in 72-100 fathoms, and of another on the continental slope between 245 and 325 fathoms, shows that this species is not as closely restricted to the vicinity of the coast as previously supposed and that it reaches considerably greater depths. The locality of the capture (taken by Atlantis) last mentioned (lat. 39°31' N; long. 72°16' W.) also extends the known range somewhat farther south. Spotted wolffish, Anarrhichas minor Olafsen This Arctic species is seldom taken within the Gulf of Maine, hence the capture of a small specimen, weighing 3}i pounds, on a trawl, off Portand Lightship on April 23, 1927, is worthy of mention. On the Scotian banks, however, it is not so uncom- mon, for we have records of 2, 37 and 54 inches long, respectively, caught on Sable Island Bank in January 1934 and 5 more in March of that year. Usually about 5 to 10 from this general region are landed each 37ear at the Boston Fish Pier. Eelpout, Zoarces anguillaris (Peck) Many small specimens from 1.8 inches long upward, have recently been collected along our coast between Maine and New Jersey, including (within the gulf) Mount Desert, Stellwagen Bank, Georges Bank, and the vicinity of Chatham, suggesting that the eelpout breeds successfully throughout this range. And as all the young thus far taken have been caught in depths of 20 to 45 fathoms, probably this is the usual spawning zone. Although eelpouts have seldom been reported deeper than 50 338 BULLETIN OF THE BUREAU OF FISHERIES fathoms, Albatross II recently (July 1931) trawled a number in the basin of the gulf as deep as 90 fathoms. The sizes, in different months, of the young fry show that eelpouts in the Gulf of Maine grow to a length of about 2 inches in the first 6 months of their lives, and 3 inches in 9 months, agreeing in this respect with the growth-schedule of Bay of Fundy eelpouts derived by Clemens and Clemens (1921, p. 74) from the annual rings on the otoliths. Small specimens 5 to 6% inches long taken from February to May are probably about 1% years old. Young eelpouts, up to 3 or 4 inches long, are checkered along the sides, and irregularly blotched on the back with light and dark brown, with a small but promi- nent black spot, which fades out with growth, on the anterior part of the dorsal fin. Wolf eel, Lycenchelys verrillii (Goode and Bean) The recorded range of the wolf eel, previously known only off the coasts of New England and Nova Scotia, has now been extended southward to the offing of New York (Beebe, 1929, p. 18). The wolf eel is more common within the Gulf of Maine, in deep water, than was formerly supposed, for in the autumn and summer of 1928 and 1930 the Albatross II trawled 61 specimens, 6 to 6% inches long, in the deep basin to the westward of Jeffreys Ledge, in about 90 fathoms of water. It was also found scattered over the central basin of the gulf, in July 1931, in 95 to 123 fathoms. Silver hake, Merluccius bilinearis (Mitchill) The wintering ground of the Gulf of Maine stock of silver hake has been the sub- ject of so much speculation that the capture by the Albatross II, of many specimens between the offings of No Mans Land and off Cape Hatteras in depths ranging from 12 to 146 fathoms, in February 1930 at temperatures of 4.2° to 10.6° C. (39.5° to 51° F.), deserves mention. Such wide ranges of temperature indicate that the silver hake are well distributed on these offshore grounds during the winter. Young fish are rarely found close to shore within the gulf. Offshore, however, the Albatross II and Atlantis have trawled large numbers between 2 and about 8 inches long in widely scattered localities and in depths ranging from 20 to 115 fathoms. Measurements of young silver hake,12 recently obtained in the Gulf of Maine indicate that a length of 6-7 inches is attained at about 1 year of age. Pollock, Pollachius virens (Linnaeus) Recent tagging experiments verify the earlier view that the pollock which appear in the cold months of the year off New York and New Jersey are winter migrants from the region of Nantucket Shoals. In general the pollock in the Gulf of Maine are not migratory although occasional fish may make long journeys. Cod, Gadus callarias Linnaeus Extensive tagging experiments (Schroeder, 1930) have proved that the appear- ance of cod in winter southward along the coasts of New York and New Jersey in commercial quantities represents a regular annual mass migration from Nantucket Shoals followed by a return migration in spring. But only scattering fish join this Several hundred specimens. FISHES OF THE GULF OF MAINE 339 winter migration from tlie more northerly and easterly parts of the Gulf of Maine. It has been known that many of these cod spawn on the southern wintering grounds, but it was not until the spring of 1930 that large numbers of fry were obtained there. At that time (April) Albatross II trawled hundreds of fry 1% to 2% inches long on bottom, the most southerly catch being in latitude 36°21' N.13 Haddock, Melanogrammus aeglefinus (Linnaeus) The haddock, formerly unknown beyond the Straits of Belle Isle, in the western Atlantic, has now been reported from West Greenland (Jensen and Hansen, 1930, p. 52). From Icelandic waters comes a record of a giant haddock 44 inches long and weighing about 37 pounds (Thompson, 1929, p. 29). Long-finned hake, Urophycis chesteri (Goode and Bean) The capture of several specimens on the northern edge of Georges Bank in September 1929, in 85 to 100 fathoms, and of many to the westward and in the central basin of the Gulf of Maine the summer of 1931, in 70 to 140 fathoms, suggests that this species is more plentiful in the gulf than was previously supposed. This hake is said to be a summer spawner but very little is known concerning its rate of growth, hence we report captures of 3 fish 57 to 71 mm on April 26, 1931, and of 16 fish of 74 to 110 mm taken late in July, suggesting that a length of 4 or 5 inches is reached at 1 year of age. Spotted hake, TJrophycis regius (Walbaum) The scarcity of this hake within the Gulf of Maine is emphasized by the fact that not a single one was captured there in the numerous hauls made recently by the Albatross II. To the southward, however, many were trawled between Cape Hat- teras and the offing of Delaware Bay in 5 to 45 fathoms (chiefly in less than 20 fath- oms) from February to May 1930 and 1931. Although the spotted hake reaches a length of at least 16 inches, large fish are relatively rare. The longest of about 600 specimens taken on 14 stations by the Albatross II was only 130 mm (5 }i inches). In the largest catch (Apr. 8) the dominant size was 2 to 2 % inches. Four-bearded rockling, Enchelyopus cimbrius (Linnaeus) The rockling has recently (July 1931) been trawled in the central basin of the Gulf of Maine where it was expected, but heretofore unrecorded. The fact that one was taken in latitude 36°56' N., off Cape Charles, Va., on February 10, 1930, in only 12 fathoms, shows that in the most southerly parts of its range, it is not restricted to deep water, as previously supposed. Cusk, Brosme brosme (Muller) A fish 40 inches long and weighing 27 pounds, trawled by Albatross II in the central part of the Gulf of Maine, in 120 fathoms, is the largest definitely recorded from the Gulf of Maine. 13 These were taken during the course of 0. E. Sette’s mackerel investigations. 340 BULLETIN OF THE BUREAU OF FISHERIES Common grenadier, Macrourus bairdii Goode and Bean Recent records show that the grenadier is comparatively common on muddy bottom in the gulf, at depths greater than about 90 fathoms and that it may occasion- ally be taken shoal er, for one was reported from the slope of Jeffreys Ledge in about 50 fathoms during March 1934. The capture of a ripe male in late September verifies the earlier suggestion that the grenadier is an autumn spawner. The largest fish taken by Albatross II was 16 inches long. This grenadier has now been taken as far eastward as the Grand Banks of Newfoundland (Nfld. Rpt., 1933 (1934), p. 116). American plaice, Hippoglossoides platessoides (Fabricius) Recent trawling by Albatross II and Atlantis proves this species to be generally distributed even in the deeper parts of the central basin of the gulf, to a depth of at least 120 fathoms. A specimen 15% inches long caught off Montauk Point, N. Y., in 112 fathoms, February 6, 1930, is the most southerly and westerly record. As this flounder is a spring spawner it may be assumed that bottom stages 69 to 80 mm long trawled off Cape Cod, May 1, were about 1 year old, and 85 to 118 mm fry found at several localities in July and August were between 1% and 1% years old, those of 8-10 inches, 2% to 2% years. Four-spotted flounder, Paralichthys oblongus (Mitchill) This flounder, formerly thought rare to the east of Cape Cod, has recently been found here and there on the southern half of Georges Bank. Previously known only as far southward as New York, many have been trawled by the Albatross II south to the Virginia Capes (lat. 36°45' N.). The fact that captures were made in 23 to 112 fathoms in February (7 stations), 31 to 52 fathoms in March (two stations), 10 to 85 fathoms in April (eight stations), 15 to 35 fathoms in May (four stations), 11 to 47 fathoms in June (five stations), and 41 fathoms in July (one station) indicates that it is present and widely distributed in this general depth zone the year round. The capture of ripe specimens as late as mid- July shows that the breeding season is not limited to spring, as formerly supposed, but extends well into the summer. Rusty dab, Limanda ferruginea (Storer) Capture of a specimen, in the offing of Hog Island, Va. (lat. 37° 41' N.) consider- ably extends the known range to the southward. The captures of young dabs 2 to 4 inches long in February (17 fish), 2% to 4% inches in April (26 fish), 2% to 5% inches in May (10 fish), 3 to 5 inches in June (3 fish) and 3 to 6% inches in July (13 fish) yield the first data as to rate of growth. According to this growth schedule the rusty dab reaches a length of approximately 5 inches at 1 year of age. Winter flounder, Pseudopleuronectes americanus (Walbaum) The recovery, off Chatham and on Nantucket Shoals, of winter flounders tagged and released at Woods Hole proves that some of them, at least, may wander for longer distances than previously supposed. FISHES OF THE GULF OF MAINE 341 Georges Bank flounder, Pseudopleuronedes dignabilis Kendall This flounder, previously known only from the Georges Bank area, is now re- ported from the eastern edge of the Scotian banks and the western part of the Grand Banks of Newfoundland (Nfld. Rept., 1934 (1935), p. 79). Witch flounder, Glyptocephalus cynoglossus (Linnaeus) The witch flounder has recently been found to be generally distributed in the central basin of the Gulf of Maine where the Albatross II and Atlantis trawled it down to 140 fathoms, in July 1931 and in August 1936, respectively. Goode and Bean’s (1896, p. 433) record of it in latitude 34°39' at a depth of 603 fathoms (omitted in Bigelow and Welsh, 1925) shows that it ranges southward to the offing of Cape Hatteras in deep water. But the most southerly record of it in shoal water is a specimen taken by Albatross II in 10 fathoms off Virginia (lat. 37°50')- Many specimens from 3 to 5 inches and from 7 to 8 % inches long were taken from July to September suggesting that the witch reaches a length of about 4 inches at 1 year and about 8 inches at 2 years of age. Gulf Stream flounder, Citharichthys ardifrons 14 Goode This little flounder was formerly believed to reach a length of only about 4 inches but recently the Albatross II collected many specimens up to 7 inches long. Recent trawling experience extends knowledge of its distribution by showing that it may occur as shoal as 12 fathoms, and that it finds its northeastern boundary off the southeastern slope of Georges Bank and its southwestern boundary off Cape Hatteras. Usually only a few specimens are taken in any given trawl haul, even further to the west and south where the species appears to be most common; hence, a catch of about 100 made by the Albatross II, off Montauk Point, N. Y., in 50 fathoms, is noteworthy. Apparently, it spawns over a long season, from spring through summer, for we have found females with well-developed ovaries in February while Goode had ripe ones in September. Although the Gulf Stream flounder is not large enough and thus far has been found too scarce to be of commercial value, we can witness that it is excellent on the table. American goosefish, Lophius americanus Cuvier and Valenciennes Recent investigations by Berrill (1929) and by Procter et al. (1928) make it appear that the American goosefish, given as Lophius piscatorius in “Fishes of the Gulf of Maine” (Bigelow and Welsh, 1925, p. 524), is specifically distinct from the European. Very small goosefish are seldom reported, hence captures of 1 of 10 inches in February, 1 of 10 inches in April, 2 of 7 % and 10 inches, respectively, in May, 3 of 6% to 9 inches in July, and 3 of 4 to 4% inches in August between latitudes 43°21' N. and 37°36' N. in depths ranging from 35 to 140 fathoms, are of interest. Sargassum fish, Histrio histrio (Linnaeus) A single specimen about 4% inches long, picked up in a purse seine near the sur- face over the west central part of Georges Bank, by the schooner Old Glory on Sep- tember 15, 1930 (Firth, 1931, p. 14), extends the known range of this fish to the Gulf of Maine. ■* Parr (1931) has made a revision of the genus Citharichthys of the western Atlantic. 342 BULLETIN OP THE BUREAU OF FISHERIES Beep-sea angler, Mancalias uranoscopus (Murray) A 24^-inch specimen of this uncommon fish was trawled on Georges Bank Feb- ruary 9, 1927, by the fishing steamer Ripple; this is the only record of a member of this family (Ceratiidae) from New England waters (Parr, 1932, p. 12). BIBLIOGRAPHY Barnard, Iv. H. 1925. A monograph of the marine fishes of South Africa. Part I. Ann., So. African Museum, vol. XXI, 1925, pp. 1-418. Edinburgh. Beebe, William. 1929. Deep-sea fish of the Hudson Gorge. Zoologica, vol. XII, no. 1, 1929, 19 pp., 1 fig. New York. Berrill, N. J. 1929. The validity of Lophius americanus Val. as a species distinct from L. piscatorius Linn, with notes on the development. Contr., Canad. Biol, and Fish., N. S., vol. IV, no. 12, 1929, pp. 143-155, 7 figs. Ottawa. Bigelow, Henry B., and William C. Schroeder. 1927. Notes on northwest Atlantic sharks and skates. Bull., Museum, Comp. Zook, vol. LXVIII, no. 5, September 1927, pp. 239-251. Cambridge. Bigelow, Henry B., and W. C. Schroeder. 1929. A rare Bramid fish ( Taractes princeps John- son) in the northwestern Atlantic. Bull., Museum, Comp. Zool., vol. LXIX, no. 2, February 1929, pp. 39-50. Cambridge. Bigelow, Henry B., and William W. Welsh. 1925. Fishes of the Gulf of Maine. Bulk, U. S. Bur. Fish., vol. XL, Part I, 1924 (1925), 567 pp., 278 figs. Clemens, Wilbert A., and Lucy Smith Clemens. 1921. Contribution to the biology of the muttonfish, Zoarces anguiliaris. Cont., Canad. Biol., 1918-1920 (1921), pp. 69-83, 1 pi. Ottawa. Coles, Russel J. 1926. Notes on Cape Lookout (North Carolina) fishes — 1925. Copeia, no. 151, February 1926, pp. 105-106. Conel, J. LeRoy. 1931. The genital system of the Myxinoidea: A study based on notes and drawings of these organs in Bdellostoma made by Bashford Dean. The Bashford Dean Memorial Volume, Archaic Fishes, Article III, Amer. Museum, Nat. Hist., 1931, pp. 64-102, pi. I-IV. New York. Cornish, George A. 1907. Notes on the fishes of Canso. Further Contr., Canad. Biol., 1902-05 (1907), pp. 81-90. Ottawa. Firth, Frank E. 1931. Some marine fishes collected recently in New England waters. Bulk, Boston Soc. Nat. Hist., no. 61, October 1931, pp. 8-14. Boston. Fish, Charles J. 1927. Production and distribution of cod eggs in Massachusetts Bay in 1924 and 1925. Bulk, U. S. Bur. Fish., vol. XLIII, 1927, Part II, pp. 253-296, 16 figs. Goode, George Brown, and Tarleton H. Bean. 1896. Oceanic ichthyology. Mem., Museum, Comp. Zook, Harvard College, vol. XXII, 1896, xxxv + 553 pp. Cambridge. Also Smith. Contr. to Knowl., vol. XXX, 1895 (1896), and Spec. Bulk No. 2, U. S. Nat. Museum, 1895 (1896). Hildebrand, Samuel F., and William C. Schroeder. 1928. Fishes of Chesapeake Bay. Bulk, U. S. Bur. Fish., vol. XLIII, Part I, 388 pp., 211 figs. Honeyman, D. 1886. Nova Scotian ichthyology. Proc. and Trans., Nova Scotian Inst. Nat. Sci., vol. VI, Part IV, 1886, pp. 328-330. Halifax. Jensen, Ad. S. 1914. The selachians of Greenland. Mindeskrift for Jepetus Steenstrup, 1914, 40 pp. Copenhagen. Jensen, Ad. S. 1926. Investigations of the “Dana” in West Greenland waters. 1925. Rapp, et Proces-Verb., Cons. Perm. Inter. Explor. Mer, vol. XXXIX, 1926, pp. 85-102. Copen- hagen. Jensen, Ad. S. 1928. The fauna of Greenland. In Greenland, published by the Commission for the direction of the Geological and Geographical Investigations in Greenland, vol. I, 1928, pp. 319-355. Copenhagen. Jensen, Ad. S., and Paul M. Hansen. 1930. Underspgelser over den Grpnlandske Torsk, 1930, 55 pp. Copenhagen. Jones, J. Matthew. 1879. List of the fishes of Nova Scotia. Proc. and Trans., Nova Scotian Inst. Nat. Sci., vol. V. Part I, 1879, pp. 87-97. Halifax. FISHES OF THE GULF OF MAINE 343 Jordan, David Starr, Barton Warren Evermann, and Howard Walton Clark. 1930. Check list of the fishes and fishlike vertebrates of North and Middle America north of the northern boundary of Venezuela and Colombia. Report, U. S. Com. Fish., Part II, 1928 (1930), 670 pp. Kendall, William C. 1931. Remarks on additions to the marine fauna of the coast of Maine. Bull., Boston Soc. Nat. Hist., no. 58, Jan. 1931, pp. 9-11. Boston. Leim, A. H. 1930. Unusual fishes and other forms in Nova Scotian waters. In Proc. and Trans. Nova Scotian Inst. Nat. Sci., vol. XVII, Part 4, 1930, p. xlvi. Halifax. MacCoy, Clinton V. 1929. The mackerel in New England. Bull., Boston Soc. Nat. Hist., no. 53, Oct. 1929, pp. 3-7. Boston. MacCoy, Clinton V. 1931a. Fishes. In Museum Notes, Bull., Boston Soc. Nat. Hist., no. 58, Jan. 1931, pp. 16-18. Boston. MacCoy, Clinton V. 1931b. Fishes. In Museum Notes, Bull., Boston Soc. Nat. Hist., no. 61, Oct. 1931, p. 21. Boston. MacCoy, Clinton V. 1933. Fishes. In Museum Notes, Bull., Boston Soc. Nat. Hist., no. 69, Oct. 1933, pp. 8-9. Boston. MacCoy, Clinton V. 1934. Fishes. In Museum Notes, Bull., Boston Soc. Nat. Hist., no. 70, Jan. 1934, pp. 6-7. Boston. Newfoundland Fishery Research Commission. 1932. Annual Report Year 1931. Report, vol. 1, no. 4, 110 pp. St. Johns. Newfoundland Fishery Research Commission. 1933. Annual Report Year 1932. Report, vol. 2, no. 1, 127 pp. St. Johns. Newfoundland Fishery Research Laboratory. 1935. Annual Report Year 1934. Report, vol. 2, no. 3, 79 pp., 10 figs., 2 pis., 9 charts. St. Johns. Parr, Albert Eide. 1931. A practical revision of the western Atlantic species of the genus Citharichthys (including Etropus). Bull., Bingham Oceanographic Collection, vol IV, art. 1, 1931, 24 pp., 9 figs. New Haven. Parr, Albert Eide. 1932. On a deep-sea devilfish from New England waters and the peculiar life and looks of its kind. Bull., Boston Soc. Nat. Hist., no. 63, April 1932, pp. 3-16, 4 figs. Boston. Proctor, William, et al. 1928. A contribution to the life-history of the angler ( Lophius pisca- torius). Biological Survey of the Mount Desert Region, Part 2, Fishes, 1928, 13 pp., 5 pi. Published by the Wistar Institute of Anatomy and Biology. Philadelphia. Radcliffe, Lewis. 1926. “Opah” and “Skilligalee” landed at Boston Fish Pier. Copeia, no 151, Feb. 25, 1926, p. 112. Northampton. Schmidt, Johannes. 1931. Eels and conger eels of the North Atlantic. Nature, vol. 128, no. 3232, Oct. 10, 1931, pp. 602-604, 2 figs. London. Schroeder, William C. 1930. Migrations and other phases in the life history of the cod off southern New England. Bull., U. S. Bur. Fish., vol. XLVI, 1930 (1931), pp. 1-136, 33 figs. Schroeder, William C. 1931. Notes on certain fishes collected off the New England coast from 1924 to 1930. Bull., Boston Soc. Nat. Hist., no. 58, Jan. 1931, pp. 3-8. Boston. Schroeder, William C. 1933. Unique records of the brier skate and the rock eel from New England. Bull., Boston Soc. Nat. Hist., no. 66, Jan. 1933, pp. 5-6. Boston. Sella, Massino. 1931. The tuna ( Thunnus thynnus L.) of the Western Atlantic. An appeal to fishermen for the collection of hooks found in tunafish. Internationale Revue der gesamten Hydrobiologie und Hydrographie, Band 25, Heft 1-2, pp. 46-67, 10 figs., 1931. Leipzig. Smithsonian Institution. 1925. Opinions rendered by the International Commission on Zoolog- ical Nomenclature. Smith. Misc. Col., vol. 73, no. 3, 1925, 40 pp. Washington. Smithsonian Institution. 1926. Opinions rendered by the International Commission on Zoolog- ical Nomenclature. Smith. Misc. Coll., vol. 73, no. 4, 1926, 30 pp. Washington. Thompson, Harold. 1929. General features in the biology of the haddock ( Gadus aeglefinus L.) in Icelandic waters in the period 1903-1926. Rapp, et Proc6s-Verb., Cons. Perm. Inter. Explor. Mer., vol. LVII, 1929, 73 pp. Copenhagen. Vladykov, V. D. 1935. Some unreported and rare fishes for the coast of Nova Scotia. Proc., Nova Scotian Inst, of Sci., vol. XIX, Part 1, 1934^1935, pp. 1-8. Halifax. Welsh, W. W. 1915. Notes on the habits of the young of the squirrel hake and sea snail. Copeia, no. 18, May 15, 1915, pp. 2-3. New York. o U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner ADAPTATION OF THE FEEDING MECHANISM OF THE OYSTER ( Ostrea gigas ) TO CHANGES IN SALINITY By A. E. Hopkins From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 21 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1936 For sale by the Superintendent of Documents, Washington, D. C. Price 10 cents ADAPTATION OF THE FEEDING MECHANISM OF THE OYSTER (Ostrea gigas) TO CHANGES IN SALINITY1 By A. E. Hopkins, Ph. D., Aquatic Biologist, United States Bureau of Fisheries CONTENTS Page Introduction 345 Material and methods 346 Description of experiments 347 Series I 348 Series II 350 Series III 351 Series IV 353 Series V 355 Series VI 356 Series VII 356 Discussion 360 Summary 362 Literature cited 363 INTRODUCTION Natural beds of oysters are found characteristically in such inshore waters as the coastal estuaries and bays in which there is a considerable dilution of the ocean water by land drainage. Prytherch (1934), in an extensive monograph describing an experi- mental study of the ecology of Ostrea virginica, concluded that successful propagation of oysters depends upon the presence in the water of copper ions required by the swimming larvae before attachment, or setting, may take place. In nature the copper is introduced into sea water in solution in the fresh water of rivers, thereby locating natural oyster grounds where they are subject to frequent changes in salinity. Referring to 0. virginica, Churchill (1920) stated: Oysters are found in water ranging in density from 1.002 to 1.025 (about 2.5 to 33.0 parts per mille), but cannot withstand densities lower than 1.007 (9 parts per mille) for indefinite periods. In general they seem to thrive best in densities between 1.011 and 1.022 (14.36 to 28.80 parts per mille). This range apparently refers to areas in which natural oyster beds are located, or those areas in which propagation takes place. The feeding activities, however, may not be most efficient at the same range of salinity or density as in the case of propagation. It may frequently be observed that oysters grown on those grounds where seeds are caught most effectively do not fatten as well as those grown in different areas. Oyster growers typically catch seeds in one place and transplant them to localities which experience has shown to be favorable for * Bulletin No. 21. Approved for publication June 4, 1936. 345 346 BULLETIN OF THE BUREAU OF FISHERIES fattening. It appears possible that salinity may be one of the important factors influ- encing the feeding of oysters. The gills of oysters, by means of cilia, pump a stream of water from which food particles are filtered and passed along definite grooves to the labial palps, which convey them to the mouth (Nelson, 1923b). The rate at which water is pumped was used by Galtsoff (1928) as a criterion of the rate of feeding under different conditions of temperature. Nelson (1921, 1923a) kept records of the opening and closing of oysters immersed in the open bay along with those of temperature, salinity, turbidity, etc. He concluded that the oyster feeds most rapidly on the flood tide even when the density is approximately the same as during ebb tide. He concluded also that for 0. virginica a density of 1.008 (10.42 parts per mille) is the lowest at which feeding will go on, and that the minimum density required varies in proportion to that of the water in which the oysters have been grown. His method, however, was to study the opening or closing of oysters under different conditions, judging an open specimen as actively feeding. But, as will be shown below, an oyster may be open without feeding, though clearly when it is closed feeding is impossible. The mechanism of physiological adaptation to changes in salinity is not the question primarily discussed in this paper. This study deals with the pumping activity of the gills as influenced by salinity, although it is realized that final ex- planation of the results is a matter of cellular physiology. MATERIAL AND METHODS Because of their relatively large size, specimens of Ostrea gigas are most favorable material for experimental study. Tins species is imported from Japan as small seeds and grown in various waters of the Pacific coast. In portions of Puget Sound propagation occurs, and the resulting well-shaped, large oysters were used in these tests. The method employed has already been described (Hopkins, 1933). A simple lever, resting with the least possible weight upon the upper (right) valve of the specimen, served to record on a kymograph paper the position and movements of the valves, as affected by activity of the adductor muscle. A paper cone, thoroughly waterproofed with a solution of celluloid in acetone, attached to a system of levers made of lightweight straw, was placed so that all water pumped by the gills struck the cone. The recording tip of the straw was made of a sliver of cellophane, and this made its record on the kymograph paper directly below that of the lever recording shell movements. Fixed levers continously recorded the zero positions— the closed position of the shell and the level of cessation of pumping by the gills. Since the salinity of the water had to be maintained at a constant level, it was necessary to employ a method for continuous aeration and circulation. For this purpose small centrifugal pumps, made of celluloid (Hopkins, 1934) were used. Two of these pumps were used to circulate and aerate the water, without causing any strong currents which might disturb either the oyster or the mechanism for recording the relative rate at which the gills pump. An extra aquarium was interposed between the experimental chamber and the pumps for the control of temperature. In this aquarium either an electric-light bulb or an immersion heater was so controlled that the temperature in the experimental chamber was maintained during all tests at between 17° and 19° C. This temperature level was chosen because it had been previously determined (Hopkins, 1933, 1935) to be most favorable for stability of ADAPTATION OF OYSTER TO CHANGES IN SALINITY 347 the adductor muscle, resulting in the shells remaining at almost their optimum degree of openness. At 20° C. or above the valves tend to become more closed, although the gills themselves pump more rapidly up to between 27° and 28° C. The salinity of the running water at the laboratory varied generally between 26 and 29 parts per mille. Water of a lower salinity was prepared in a tank, large enough for several changes of water in the experimental aquaria, by mixing the laboratory water with pure spring water. Control tests showed this method to be adequate, for specimens behaved identically in bay water of salinity 28 parts per mille and in water of the same salinity made by mixing spring water and con- centrated bay water. The water level in the experimental tanks was marked so that occasional small additions of spring water prevented more than very slight changes in salinity due to evaporation. The pH of the water varied only slightly, between 7.7 and 8.0, in harmony with that of water on the oyster grounds. Specimens for experimentation were always kept in the running sea water of the laboratory for several days or weeks before use. During the first one or more days after being placed in the experimental aquarium, a specimen was tested in water of approximately the same salinity as the running laboratory sea water. In making a change of water, of either the same or a different salinity, the water was drained from the experimental tanks, which were then flushed out thoroughly with the new water before being filled to the correct level. To save time, the new water was warmed to the experimental temperature (17°-19° C.) before being siphoned in. Some specimens were mounted upon a base of plaster of paris and sand, while others were set rigidly upon a small celluloid frame. No ill effect of the plaster of paris was noted. After completion the kymograph papers were marked off exactly into 5-minute periods for analysis. Planimeter measurements were made of the area enclcsed between the record line and the reference line during each 5-minute period. These measurements are considered in the following descriptions as representing S (the degree of openness of the shell) and F (the relative rate of flow of water through the gills) during each 5-minute period. DESCRIPTION OF EXPERIMENTS These experiments were started for the purpose of determining whether there is any relationship between salinity and the rate at which water is pumped. It was thought that adaptation to a change in salinity would occur quite quickly, since in their natural environment oysters are subject to frequent and sometimes extreme salinity variations, and it was planned to make a change, allow the pumping mecha- nism to reach a stable level of activit}q then to change again. In this manner it was hoped that sufficient values might be obtained to permit graphical analysis of the relationship between salinity and the rate of feeding. It soon became evident, however, that such a procedure would require a long period of time, even with only one specimen. Adaptation was found to be extremely slow, and the variations in degree of openness of the valves and in rate of pumping are so great even under conditions of constant salinity and temperature that any test would have to be carried on for many days, or even weeks. It was, therefore, decided to attack the problem by making a study of the process of adaptation following changes in salinity with respect to shell movements and gill activity. 348 BULLETIN OF THE BUREAU OF FISHERIES The nature of the results and the different lengths of time during which specimens were subject to water of different salinities make it impossible to present the data obtained in any manner other than by description of the tests with each specimen. Seven such series are described below. SERIES I This was the first specimen on which detailed experiments were performed. The salinity at the beginning was 29.40 parts per mille and records were kept for thirty -seven 5-minute periods, or 185 minutes. Then, on the same day, the salinity was lowered to 26.96 parts per mille and the activity recorded for the next 205 minutes, or forty-one 5-minute periods. The records taken at the two salinities (table 1, series I) are closely similar, although the averages in the former case are S, 8.53; F, 6.45; and in the latter case, S, 7.13; F, 5.40. Obviously the results are inadequate for comparison because of the short duration of exposure to each medium. On the following day the salinity was lowered still further to 24.03 parts per mille and records kept for forty-seven 5-minute periods. The low F values during early treatment at this salinity, as shown in the table, indicate a marked effect of the change, though later values appear to be in harmony with those obtained at the original higher salinities. The effect of lowering the salinity is more strikingly shown in the tests of the next 2 days (Nov. 9 and 10), after changing the salinity from 24.03 to 17.76 parts per mille. Although during the first day the shell remained approximately as wide open as previously, only a feeble stream of water was pumped by the gills. In figure 1 the consecutive 5-minute values are shown graphically, to illustrate the slow Figure I.— Values of Sand F showing adaptation during 3 days to a salinity of 17.76 p. p. m. following a change from 24.03 p. p. m. Series I. 17°-19° C. adaptation of the gill mechanism (F) while the valves (S) quickly recovered their normal degree of openness. It is doubtful that adaptation was completed even after more than 2 days. However, the salinity was changed (Nov. 11) to 28.50 to test adaptation to a rise in salinity. Records were kept on the first day for fifty-four 5-minute periods. Within a very short time after the change, the values of F rose almost to normal, though not until the next day did the shell open wide enough to permit completely normal gill activity. After the specimen had been in this water for 4 days the salinity was again low- ered to 22.20 parts per mille. Adaptation to this change was relatively rapid, ADAPTATION OF OYSTER TO CHANGES IN SALINITY 349 occupying but a few hours. On the following day the medium was changed to a salinity of 29.46, and adaptation was almost immediate. During the preceding tests it appeared that adaptation to a considerable change in salinity is a very slow process, much of which occurs between the recordings on consecutive days. In order to obtain records of as many details as possible of the early stages of adaptation, the salinity was reduced to 14.60 parts per mille and during the following 12 hours records were taken which provided one hundred and twenty-five 5-minute periods. In figure 2 the results are given in detail for the Figure 2.— Adaptation to salinity of 14.60 p. p. m. following a change from salinity of 29.46 p. p. m. Series I. Values of .S and F refer to consecutive 5-minute periods. 17°-19° O. 5 days during which this salinity was used. For the first 8 hours (Nov. 18), with minor exceptions, the valves remained relatively close together, and almost no water was pumped. During the following hour the shell consistently became wider open and the gills produced a slight flow of water. Recording was continued for several hours more, but the rate of flow increased very little. However, during the follow- ing 4 days the gills steadily recovered, though large variations both in rate of flow and in degree of openness of the valves may be noted. Even after 5 days of treat- ment recovery could not be called complete, for the values of F never reached the 350 BULLETIN OF THE BUREAU OF FISHERIES previous levels. (See series I, table 1.) When the salinity was then raised to 28.56 the specimen became adapted to a comparable level within a few hours. SERIES II Preliminary tests in a salinity of 28.87 parts per mille were carried on only during 1 day (thirty-six 5-minute readings), although it later appeared that they should have been continued for several days more (series II, table 1). When the salinity was changed to 22.70 parts per mille for 2 days and then to 28.74 parts per mille, the correct level of adaptation was reached. After 3 days in the latter a relatively slight change was made to a salinity 25.08 parts per mille. A graph (fig. 3) is reproduced showing the gradual adaptation following the change. Within about 4 hours adaptation proceeded almost to completion, as may be seen in the figure by comparing the values of S and F with those of the previous day, when higher salinity was used, as well as with those of the following day. Figure 3. — Adaptation of S and F (5-minute values) following change in salinity from 28.74 to 25.08 p. p. m. Series II. 17°-19° C. After returning the specimen to a salinity 28.94 parts per mille for 1 day, the water was changed to a salinity of 17.85 parts per mille, a step considerably greater than that just described. During the first several days (series II, table 1) it appeared that adaptation was slowly occurring, but during the remainder of the 11 days there was no further obvious recovery. The valves did not remain open wide enough to permit free flow of water. Yet when the salinity was raised to 27.27 parts per mille recovery was so rapid that activity seemed to be almost at its normal level within 2 or 3 hours. A considerable difficulty in the analysis of these data lies in the fact that two activities are concerned, namely, the degree of openness of the shell (S) and the rate at which water is pumped (F). The latter is dependent upon the former, and it appears that both functions are influenced independently by changes in salinity. However, even under constant conditions of temperature and salinity, there is a ADAPTATION OF OYSTER TO CHANGES IN SALINITY 351 tremendous variation in the degree of openness. The rate of pumping of water is the activity which primarily is being studied, for on this depends the rate of feeding; but this activity may not be isolated from the influence of the adductor muscle save by propping the valves open ; in which case the results would be of little significance with respect to the reactions of the oyster as a whole to environmental factors. In a previous publication (Hopkins, 1933) it was shown that, at constant tem- perature, the rate of flow of water through the gills increases as the valves become wider open. By plotting the results of certain series to show the relationship between S and F it is possible to find whether observed effects are due to changed activity of the gills or only to the effect of the position of the valves. Such a graph is shown in figure 4, in which the results for all of the tests with a salinity of 17.85 parts per mille are given in solid points and those taken with a salinity of about 28 parts per mille in open circles. In spite of the fact that the points are considerably dispersed, they fall into relative alinement. Figure 4.— Relationship between S (degree of openness of shell) and F (rate of flow) at salinity of 17.85 p. p. m. (solid points) as compared with 28 p. p. m. (open circles). Series II. 17°-19° C. Most of the points resulting from low salinity represent low values of both S and F, but they fall into general alinement with values obtained with the higher salinity. This indicates that at least a large part of the low F readings at the lower salinity (series II, table 1) may be due to the effect of the position of the valves. Yet, also, a good many of these points are high on the S scale but much lower on the F scale as compared with the values obtained at high salinity. The trend of the low salinity curve appears to be much steeper than the other. The significance of the graph is that both the adductor muscle, by controlling shell position, and the gill mechanism were in this case responsible for the lack of complete adaptation to the lower salinity. Tests with other specimens throw further light on the subject. SERIES III The preceding two series of tests were concerned with changes in salinity between about 14 and about 29 parts per mille. It was noted that adaptation to a reduction in salinity is quite slow as compared to that following a rise. The question presented itself as to whether this difference was due to the fact that the specimens had been 352 BULLETIN OF THE BUREAU OF FISHERIES living in water of the higher salinity (28-29 parts per mille) and consequently any change from this level produced a pronounced effect, while restoration of this level merely permitted normal activity. The reactions were studied when the specimen was transferred into water of a salinity of 39.10, after 2 days of study at 28.55 and 28.27 parts per mille. Since pure ocean water has a salinity of only 35 to 36 parts per mille, there is no possi- bility that the specimen could already be adapted to water of even higher salinity. The results of the entire series are summarized in table 1, series III. It may readily be observed that the specimen became adapted quickly to the abnormally high level. In figure 5 the values of S and F are given in detail to show the reactions during the first 2 days in this high salinity. After about 3 hours the specimen was pumping water at a rate only a little below normal, and within the following 4 hours it appeared to become completely adapted, as judged by degree of openness and rate of pumping. Tests of the following day are given on the graph for comparison. Figure 5.— Adaptation of 5-minute values of S and F to salinity of 39.10 p. p. m. following change from 28.27 p. p. m. Series III. 17°-19° C. On the other hand, adaptation proceeded at the same slow rate as previously observed when the salinity was lowered to 28.04 parts per mille, as shown in figure 6. At least 3 days were required for restoration of the rate of pumping previously observed, although the valves were much wider open than at the higher salinity. The effect of lowering the salinity in this case appears to be entirely upon the gill mechanism, rather than upon the organs controlling position of the valves. Since the specimen appeared to function so well in a salinity of 39.10 parts per mille, a short test was made to give a general idea of the upper limit to which it could become adapted. A salinity of 56.32 parts per mille, twice that of the ordinary bay water, was used and records kept for 4 days (series III, table 1). During the first day the shell remained entirely closed, and on the second it opened only to a small degree. However, as the valves became sufficiently far apart the gills were ADAPTATION OF OYSTER TO CHANGES IN SALINITY 353 able to produce a slight flow of water. On the fourth day the valves were abnormally wide open, gaping, although the gills were able to function and produce a slight current. Tests would have been continued ’save ,that_gaping is generally indicative of loss of tonus of the adductor muscle and death of the organism. However, the oyster was placed in running laboratory water, where, after some days, it apparently recov- ered completely. SERIES IV The experiments described above suggest a marked sensitiv- ity of the oyster to lowered salinity, while the effect of raising the salinity is very temporary. In one case (series I) the salinity was low- ered to 14.60 parts per mille and the specimen did not become com- pletely adapted even after 5 days. It is of importance to know approx- imately the lower salinity limit at which the oyster is able to feed, for in nature the species is frequently subjected to water ranging in salin- ity from almost pure ocean water to fresh water. Before making such an experiment this speci- men (series IV, table 1) was tested for 2 days in water of a salinity of 28.06 parts per mille, after which the salinity was raised to 36 parts per mille and left for 2 days while records were kept to show the progress of adaptation. The results of this change to higher salinity are in har- mony with other similar changes already described (fig. 7), for the specimen became completely adapt- ed within a few hours, altlxmgh the initial effect of the change was to reduce the rate of pumping almost to zero. When the salinity was then reduced to 28.17 parts per mille (fig. 8) adaptation during the first few hours was rapid, but it was not until the third day Figure 6.— Graph of 5-minute values of S and F showing progressive adaptation to salinity 28.04 p. p. m. following change from 39.10 p. p. m. For reference the value, 6, is shown as broken line. Series III. 17°-19° C . 354 BULLETIN OF THE BUREAU OF FISHERIES that the gills were able to pump at the previous rate, even though the valves were well apart. These tests occupied 10 days, during which time the specimen was highly active and adaptable. At about noon on the tenth day (series IV, table 1) the salinity was reduced to 10.59 parts per mille and a few records kept during the rest of the day. However, the shell remained almost entirely closed. On the next 2 days kymograph records were not made, for the valves were only slightly open and no water was pumped. On the third day the valves had opened wider and the activity was recorded on the kymograph. Only occasionally was a little water discharged either on this day or during the entire 20 days that water of this sa- linity was used. After the first few days the valves were generally open to a normal degree, and consequently it was the gills themselves that failed to function. The tests were continued for as long as 20 days with the hope that some indication of adapta- tion of the gill mechanism would appear. In an attempt to locate as exactly as possible the minimum salinity required for functioning of these or- gans, the salinity was then raised slightly to 12.94 parts per mille and maintained for 4 days. Results being similar to those in the preceding tests, the salinity was again raised to 15.01 parts per mille for 2 days, but without any indication of increased ac- tivity. Even when the sa- linity was raised to 27.92 parts per mille the specimen did not show any sign of re- covery. It only remained abnormally wide open, as it has been during many days in water of low salinity. It was clear that subjection to a salinity as low as 10.59 parts per mille for a con- siderable period of time had done some damage which was not readily repaired even after restoration of more favorable conditions. In fact, it was only after the speci- men had been in running sea water for about 2 weeks that it appeared to pump a vigorous stream of water. Incidentally, the behavior of the oyster during all tests after the introduction of the low salinity was different from the normal in that the quick, partial closures Figure 7.— Adaptation of S and F (5-minute values) to salinity of 36 p. p. in. follow- ing change from 28.06. Series IV. 17°-19° C. ADAPTATION OF OYSTER TO CHANGES IN SALINITY 355 which occur with characteristic frequency in a normal specimen almost never were to be observed. The oyster acted almost as if anaesthetized, although the shell varied considerably in degree of openness. All movements, however, were slow. Nevertheless, secretion of shell proceeded at a rapid rate, and when the experiment was dis- continued several millimeters of thin new shell had been added to the margins of the valves. SERIES V In the series the a salinity preceding specimen was left in of 10.59 parts per mille for 20 days, resulting in damage to the pumping mechanism. It was de- sirable to know whether shorter exposure to such a salinity would have the same harmful effect. A specimen was therefore tested thoroughly for 2 days in water of high salinity (27.35 parts per mille), during which time it reacted favor- ably, pumping vigorously (series V, table 1). After being changed to a salinity of 10.64 parts per mille, the valves remained at lirst only slightly open but during the next 2 days opened abnormally wide. As in the previous series, almost no water was pumped. After 4 days of treatment, the salinity was raised to 27.39 and records kept for 5 days. Within about 2 hours after the change the gills began to pump a small stream. Beyond this, however, there was no further recovery during the 5 days. The shell opened wider and wider, but never was more than a feeble circulation of water produced by the gills. The results are similar to those described in series IV, with the Figure 8. — Five-minute values of S and F during adaptation to a salinity of 28.17 p. p. m. after change from 36 p. p. m. The value, 10, is shown as broken line for reference. Series IV. 17°-19° C. difference that in the present series exposure to low salinity had not been sufficiently prolonged to produce the harmful effect to as great a degree. 356 BULLETIN OF THE BUREAU OF FISHERIES SERIES VI With this specimen another attempt was made to obtain comparative data on the reactions to a series of salinities, the changes being made in relatively small steps. The summarized results are shown in figure 9 and table 1, series VI. After 4 days in a salinity of 28.31 parts per mille, during which there was some variability in activity, the salinity was reduced to 23.57 parts per mille and left for 4 days. Adaptation appeared to be quite rapid on the first 2 days ; but on the fourth day, for unknown reasons, the value of F was lower than before. When the salinity was changed to 20.90 parts per mille adaptation was appar- ently complete on the second day. However, reduction to 16.31 resulted in an entirely different behavior, the specimen becoming adapt- ed within a few hours to a certain low level of pumping which was maintained with- out improvement during the entire period of the test (6 days) . Examination of figure 9, as well as the results of pre- ceding tests, leads to the sug- gestion that salinities below about 20 parts per mille be- come more and more strik- ingly unfavorable in their Figure 9. — Graph of daily averages of values of S (degree of openness) and F (rela- , five rate of flow) at different salinities. Series VI. 17°-19° C. CneCt. Whether adaptation would take place in such instances to the extent that water would be pumped at the same rate as at higher salinities may not definitely be stated, but it is clear that adaptation is extremely slow. SERIES VII In those experiments which have already been described it was indicated that when specimens are placed in water of a salinity of about 15 to 17 parts per mille the rate of pumping is greatly reduced and adaptation is very slow, possibly never reaching the level observed at higher salinities. It was also shown that a salinity as low as 10 to 11 parts per mille results not only in almost complete stoppage of the flow of water through the gills but also in a harmful effect from which recovery, after return to high salinity, is extremely slow. It would appear that the lower sa- linity limit which the oyster can tolerate lies somewhere between 10 and 15 parts per mille. This experiment was performed in order to locate this limit more exactly (series VII, table 1). Preliminary tests were carried on over a period of 9 days with a salinity of 25.09 parts per mille before changing to a salinity of 13.00 parts per mille. During the first day at the lower salinity the valves did not open wide and there was no flow of water. On the next day it looked as if the specimen was going to become well adapted, but during the following days this rate of activity was not maintained. During 8 days of treatment in water of this salinity only a sluggish stream was pumped as compared with the activity observed during the preceding 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 10 19 ADAPTATION OF OYSTER TO CHANGES IN SALINITY 357 tests. The activity in this salinity was only slightly less rapid than that recorded in series I, when a salinity of 14.60 parts per mille was employed. A salinity of about 13 parts per mille appears to be but little more favorable than a salinity of about 10.5 parts per mille as studied in series IV and V. In the latter cases, however, the specimens were not able readily to recover from the harmful effect. This specimen quickly became adapted to a salinity of 25.53 parts per mille, and in spite of the fact that it had been in the water of lower salinity for 8 days it recovered completely within 2 days. The progress of adaptation in this case is one of the best examples obtained, and a graph (fig. 10) is reproduced showing all of the 5-minute values of S and F. After the first 2 hours, during which the valves were not wide open and almost no water was pumped, the shell consistently opened wider and the rate of flow of water increased. It may be noted that considerable adaptation took place overnight between recordings of the 2 days. Tins graph suggests also the close correlation between degree of openness and rate of pumping. o - - — ■ ■ — — — » — — Figure 10.— Adaptation of S and F (5-minute values) to salinity of 25.53 p. p. m. following change from 12.93 p. p. m. Series II. 17°-19° C. The value, i, is shown by a broken line for reference. In many cases of adaptation it is difficult to determine whether the rate of pumping at any time is due to the rate of activity of the gills or to the degree of openness of the valves, which thereby determine how much water may be pumped. By plotting the values obtained during the 2 days when the specimen was in water of a salinity of 25.53 parts per mille, to show the correlation between S and F (fig. 11), it was clearly shown that the points, with a few exceptions due to the initial effect of the change in salinity, fall into definite alinement; that is, the relationship between rate of flow and degree of openness was the same on the first day, when the specimen was beginning to become adapted, and on the second day, when adaptation was almost complete. This is interpreted as direct evidence that adapta- tion in this case was entirely under the control of the adductor muscle, regulating the 358 BULLETIN OF THE BUREAU OF FISHERIES valves, while the gills themselves become adapted quickly. In this respect these results are entirely different from those obtained in series IV and V, in which, after the salinity was raised following treatment in about 10.5 parts per mille, the shells Figure 11.— Relationship between S and F during the first (solid points) and second (open circles) days following change in salinity from 12.93 p. p. m. to 25.53 p. p. m. Series II. (See fig. 10.) remained wide open but the gills were inactive. Between 10.5 and 13 parts per mille, then, there is a tolerance limit, below which there appears to be a destruc- tive effect. ADAPTATION OF OYSTER TO CHANGES IN SALINITY 359 Table 1. — Daily averages of values of S ( degree of openness of shell) and F ( relative rale of flow of water) [17°-19° C.] SERIES I Date Salinity (parts per millo) Number of 5-min- ute pe- riods S (aver- age) F (aver- age) Date Salinity (parts per mille) Number of 5-min- ute pe- riods S (aver- age) F (aver- age) Nov 7 29. 40 37 8. S3 6.7,5 Nov. 14 28. 50 65 7.99 5.61 Nov. 7 26. 96 41 7. IS 6.1,0 Mean __ >162 79 7.03 7.63 6.25 7.62 Nov. 15 22. 205 Nov. 8_ 24.03 47 7.60 6.19 Nov. 16. 29. 46 71 7.66 9.08 17. 76 17. 69 54 47 7. 42 6. 98 1.21 5. 12 Nov. 18. 14. 60 125 3. 84 .09 Nov. 19-. 14. 60 72 5. 35 . 70 17.69 18 8. 15 5. 79 Nov. 20 14. 60 18 7. 18 2. 15 Mean >119 7. 36 3. 7,6 Nov. 21_ Nov. 22 14. 57 14.57 72 64 6. 43 7.28 3. 43 5.17 Nov. 11 28. 50 54 5. 62 4. 67 Mean 1351 6. 48 6.00 1.93 Nov. 12 28. 50 43 7. 35 9. 20 Nov. 23... 28. 56 51 4.85 SERIES II Nov. 28 28. 87 36 6. S3 7.29 Nov. 29 22. 70 71 7. 47 5. 69 Nov. 30 22. 70 71 8. 17 9. 34 Mean > 142 7.82 7. 51 Dec. 1 28. 74 72 8. 38 13. 31 Dec. 2 28. 74 81 8. 73 15. 48 Dec. 3 28. 74 32 8. 60 16. 47 Mean >185 8.68 16.31 Dec. 4 25.08 63 7. 47 6. 95 Dec. 5 25.08 54 7. 80 13. 90 Mean >117 7.62 10.14 Dec. 7 28. 94 63 7.73 17. 67 Dec. 8 17. 85 29 5. 07 .23 17. 85 36 6. 98 6. 06 Dec. 12 17. 83 54 8.83 6. 92 Dec. 13 17. 83 48 6. 08 4. 40 Dec. 14 17. 83 24 5. 67 3. 88 17.83 17 5. 65 2. 79 Dec. 16 17. 83 27 4.77 1.20 Dec. 17 17.83 17 5. 96 1. 69 Dec. 18 17.83 9 5. 22 1.08 1 261 6.43 7.48 3. 98 Dec. 19 27.27 53 12.42 SERIES III Jan. 4_ 28. 55 50 2.89 6.44 Jan. 5 . 28. 27 54 3.03 9.13 Jan. 6 39. 10 71 2. 01 6. 30 Jan. 7 39. 10 36 2. 58 10.01 Jan. 9. 39. 10 36 3. 54 8. 72 Jan. 10 39. 10 33 3. 57 9. 05 Jan. 11 39. 10 65 3. 95 9.10 Mean >241 3.09 8.34 Jan. 12 28.04 60 4. 15 4. 86 Jan. 13 28.04 63 4. 26 7. 32 Jan. 14 28. 04 30 4. 13 9. 73 Jan. 16 28.04 31 4. 06 11.04 Mean >184 4-08 7.53 Jan. 17 66. 32 18 0 0 Jan. 18 — 56. 32 27 1. 27 .03 Jan. 19 - . 56. 32 37 1.95 .68 Jan. 20 - 56. 32 10 7.89 .52 Mean >92 2.01 .33 SERIES IV Jan. 25 28. 06 50 5. 67 9. 14 Jan. 26 28. 06 45 7. 00 16. 70 Mean. >95 6.60 12.44 Jan. 27 36. 00 63 6. 32 10. 32 Jan. 28 36.00 46 8. 15 14. 20 Jan. 29„ 36.00 51 8. 40 12. 83 Jan. 30 36.00 64 8. 78 12. 75 Mean >224 7.89 12. 39 Jan. 31-. 28. 17 65 8. 08 6. 21 Feb. 1 28.17 64 8. 89 10. 04 Feb. 2 28. 17 64 9. 12 12. 15 Feb. 3 28. 17 18 7.81 7. 66 Mean >211 8.61 8.85 Feb. 3 . 10.59 14 .38 0 Feb. 6 10.59 36 5. 57 .40 Feb. 7 10.59 36 3. 42 0 Feb. 8 10. 59 54 6. 74 0 Feb 9 10. 59 36 7. 37 0 Feb. 11 10. 59 17 8. 37 0 Feb. 13 10. 59 23 10. 01 .46 Feb. 15, 10. 48 18 2. 86 0 Feb. 16 10. 48 17 8. 50 0 Feb. 17 10.48 51 10.63 .36 Feb. 18 10. 48 14 10. 97 .22 Feb. 20 10. 48 36 9. 46 .38 Feb. 21 10. 48 18 5. 47 0 Feb. 23 10. 48 28 7. 08 0 Mean *398 7.80 .16 Feb. 24 12. 94 54 8. 38 . 19 Feb. 25 12. 94 45 11.24 .06 Feb. 27 12.94 28 11.21 .31 Mean >12e 10.02 .17 Feb. 28 15. 01 61 11.74 .19 Mar. 1 15. 01 55 13.01 .14 Mean . . 116 12.35 .17 Mar. 2-_ 27. 92 54 11.87 .07 Mar. 3 27. 92 52 12.12 .03 Mean >106 12.00 .05 i Total. 90252—36 3 360 BULLETIN OF THE BUREAU OF FISHERIES Table 1. — Daily averages of values of S ( degree of openness of shell ) and F ( relative rate of flow of water ) — Continued SERIES V Date Salinity (parts per mille) Number of 5-min- ute pe- riods S (aver- age) F (aver- age) Mar. 7.. 27. 35 61 3.~67 7. 81 Mar. 8... 27. 35 61 3.60 7. 15 Mean ‘122 3.63 7. 48 Mar. 9_ 10. 64 27 .61 0 Mar. 10 .. 10. 61 56 5. 08 .03 Mar. 11 10. 64 16 7. 18 0 Mean '99 4.30 .02 SERIES VI Bate Salinity (parts per mille) Number of 5-min- ute pe- riods S (aver- age) F (aver- age) Mar. 13 27. 39 54 5. 46 1. 26 Mar. 14 27. 39 54 5. 37 1. 13 27. 39 63 6. 41 .68 Mar. 16 27. 32 51 6. 72 .65 Mar. 17. 27. 32 18 8.48 .65 Mean ‘240 6.18 .90 Apr. 3 28. 31 9 5. 36 7. 51 Apr. 11 20. 90 70 5. 80 5. 05 28.31 19 5. 91 7. 97 20. 90 23 5. 88 7. 07 Apr. 5 28. 31 54 4. 55 4.91 Apr. 13 20. 90 27 5. 41 5. 84 28. 31 64 4. 96 6. 82 ‘120 6.72 6.62 ‘146 23. 57 44 5. 25 2. 49 16.31 72 4. 49 1. 15 Apr. 8 23. 57 25 6. 69 6. 39 Apr. 17..- 16.31 52 5. 26 1.08 Apr. 10 23. 57 43 4. 75 2. 74 Apr. 18. 16.31 72 4. 19 1. 19 16.31 27 3.94 .67 ‘112 S. 46 ‘240 4.68 1.07 SERIES VII Nov. 24 25.09 94 2. 37 2.80 Nov. 25 25.09 15 3.54 5. 06 Nov. 27 25. 09 67 4.52 6. 57 Nov. 28 25. 09 43 4. 24 4. 97 Nov. 29 25.09 55 5. 06 6. 54 Deo. 1 25. 09 55 4.21 5. 90 Dec. 2 25. 09 55 4. 70 4. 70 Mean ‘384 3.98 6.04 Dec. 4 13. 00 80 1. 18 0 Dec. 5 13. 00 31 5. 71 2.27 Dec. 6 12.98 81 3. 90 .68 Dec. 7 12. 98 32 4.04 .71 Dec. 8 12. 98 73 3. 40 Dec. 9 12. 93 52 4. 49 .39 Dec. 10 12. 93 11 3.81 .05 Dec. 11 12. 93 80 3. 18 .24 Mean. ‘440 S.S9 .42 Dec. 12 25. 53 88 2. 73 2. 51 Dec. 13 25.53 65 4. 80 6. 95 Mean ‘153 3.61 4.40 ‘Total. DISCUSSION Organisms such as oysters which live in inshore waters, bays, and estuaries are necessarily subject to frequent changes in the density or salinity of the medium. The observations of Prytherch (1934) indicate that considerable dilution of sea water with inflowing fresh water containing copper ions is necessary for the successful propagation of the oyster. He also noted that a salinity of from 16 to 18.6 parts per mille is optimum for rate of setting, or attachment, of larvae of 0. virginica. The natural habitat is thus limited by the salinity of the water, although it may not be assumed that diluted sea water is more favorable for growth and feeding after setting and metamorphosis have taken place. Most frequently, on the other hand, is it to be observed that those grounds which are especially favorable for catching seeds are relatively unsuited for the growing and fattening of oysters for market. A difficulty of analysis encountered in this work is that even under uniform conditions of temperature and salinity there are marked variations in both rate of activity of the gills and degree of openness of the valves. Such variability is prob- ably to be expected in any study of the reactions of an organism as a whole, for ADAPTATION OF OYSTER TO CHANGES IN SALINITY 361 innumerable factors, both internal and external, influence the activity. Lack of constancy in the rate of pumping of the gills may be traced to contraction or relaxa- tion of the gill musculature, changes in size of the ostia caused by variations in diameter of blood vessels of the gill filaments (Elsey, 1935), secretion of mucus which impedes activity of the cilia, etc. However, in spite of such uncontrolled factors which render exact analysis difficult, the results are of sufficient unifoimity to warrant definite conclusions. The experimental results described above suggest that an important factor in- fluencing the quality of oyster meats grown on different grounds is the salinity conditions. The rate at which the gills pump water, from which food particles are filtered, depends upon the salinity of the medium or, more accurately stated, upon the frequency, amplitude, and duration of changes in salinity. It is well understood that rate of pumping of water does not alone determine the rate of feeding, for the abundance of food material in the water is a variable, depending upon other factors, such as the availability of nitrates and phosphates for utilization by the microscopic plant life on which oysters feed. In some instances land drainage may bring fer- tilizing materials which cause prolific development of food organisms, at the same time so diluting the sea water that oysters are not able to pump sufficient water and to feed effectively. Any significant change in salinity causes an immediate slowing or cessation in the rate of pumping. Recovery depends upon the amplitude of the change and upon whether it is to a higher or lower level. While only a few hours may be required for adaptation following a rise in salinity, several days may be necessary for recovery following the same change in the opposite direction. It was thought possible that this difference was due to the fact that the specimens had been grown in water ranging generally about 25 to 29 parts per mille and any change from this level would produce a marked effect, while restoration of this salinity resulted merely in the resumption of normal activity. However, adaptation to a further rise in salinity to 36 or 39 parts per mille, higher than oysters encounter in nature, is also rapid, but when a salinity of about 28 parts per mille is restored adaptation is very slow. Rate of adaptation following a change in salinity is also proportional to the extent of the change. Further, as the salinity becomes lower the sensitivity of the oyster to small changes increases, very much as was found (Hopkins, 1931a) with regard to temperature. It was shown that a small drop in temperature caused closure of the shell if the temperature was well below the optimum but produced no effect if near the optimum. It appears likely that endosmosis following a reduction in salinity swells the tissues and blood vessels, such as those in the gills, resulting in decrease in the size of the pores, or ostia, through which water is forced by ciliary action. An increase in salinity, on the other hand, would involve extraction of water and consequent increase in size of the ostia. The adductor muscle is probably also affected by swell- ing and shrinking. The slow rate of adaptation to a lowering of salinity is in most cases clearly traceable to diminished activity of the gills, while the valves soon open as wide as, or wider than, before the change. On the other hand, a rise in salinity involves a slow accommodation in degree of openness of the shells and an almost immediate adaptation of the gill mechanism. & In certain cases there occurred what appeared to be an actual stimulation of gill activity following rise in salinity, whereby the specimen pumped more rapidly 362 BULLETIN OF THE BUREAU OF FISHERIES than it had previously pumped in water of the same salinity. With respect to the above interpretation this would possibly be expected, for the ostia would be much wider open than normally after the osmotic pressure of the blood has reached equi- librium with the medium. Living, as oysters do, in a highly changeable environment, subject to seasonal weather conditions which may cause variations in salinity between almost pure fresh water and undiluted sea water, they must necessarily have a remarkable tolerance for salinity changes. It is well known that in time of freshet oyster beds may be covered for weeks with almost fresh water without resulting in any considerable mortality. That they are able to maintain life under such conditions is probably largely due to their ability to remain closed for long periods of time (Hopkins, 1931b), protecting the actual tissues. However, the present results indicate that pumping of water and feeding probably do not go on under such extreme conditions, so that starvation would eventually take place if the low salinity were of too great duration. It was shown above that there is a minimum limit of salinity below which pump- ing does not go on and a further, more lasting effect produced which renders the oyster capable of only very slow recovery after being returned to water of higher salinity. This limit appears to lie between 10.5 and 13 parts per mille. At the latter salinity only a small amount of water was pumped, but recovery was readily accomplished in a high salinity. These results compare favorably with Nelson’s (1921) estimate of a salinity of about 10.42 parts per mille as the lowest at which 0. virginica can feed, although his conclusion was based upon whether or not the shell remained open. Pumping activity was approximately the same at salinity ranging from about 25 to about 39 parts per mille, although the latter value is considerably h'gher than that of pure ocean water. A salinity as high as 56 parts per mille is obviously un- favorable, and, although pumping some water, the specimen gaped and would probably have soon died. From an ecological standpoint it is the lower limit that is of importance, for seldom are oysters in nature forced to adjust themselves to a salinity higher than that of the ocean. The oyster appears to be increasingly sensitive to changes in salinity below about 20 parts per mille, possibly indicating that the optimum is higher. The results do not permit one to state the optimum salinity, but it would appear to be well above 20 parts per mille, possibly as high as 30 or 35 parts per mille. It may be that the optimum depends upon the medium to which the particular oysters have been accustomed for generations, that specimen's from beds in relatively fresh bays would react differently from those grown in more saline water, as suggested by Nelson (1923a). Most likely, however, little difference of this nature is to be expected, save that a longer time would be required for adaptation of specimens from the water of higher salinity. SUMMARY 1. Adaptation of the feeding mechanism of the oyster to changes in salinity was studied by recording on the kymograph the degree of openness of the valves and the relative rate of flow of water pumped by the gills. 2. Both the activity of the gills and that of the adductor muscle, which by con- trolling the position of the valves determines the size of the inhalent and exhalent apertures, are markedly affected by any considerable change in salinity. The initial effect of such a change is to cause partial or complete contraction of the adductor muscle and slowing or cessation of the flow of water. ADAPTATION OF OYSTER TO CHANGES IN SALINITY 363 3. Recovery, or adaptation, following a rise in salinity is very rapid as compared with adaptation following the same change in the opposite direction. The former may require a few hours, while several days may be necessary in the latter case. 4. The rate of adaptation depends upon the degree of change and upon the extent of departure from optimum salinity. It is probable that, as the salinity departs further from the optimum, adaptation would never be so complete that water would be pumped at the normal rate. 5. Because of the great variability in activity of the gills and in degree of openness even under conditions of constant salinity and temperature, the results do not justify exact statement of the optimum salinity. However, the optimum is probably not greatly different from that of ocean water, for salinities between about 25 and 39 parts per mille appear to produce similar effects. 6. While the oyster will tolerate a salinity as high as 39 parts per mille, higher than that of pure ocean water, and pump at the maximum rate, a salinity of 56 parts per mille, is definitely too high for it to tolerate. 7. As the salinity is reduced below about 20 parts per mille the oyster becomes increasingly sensitive. A longer time is required for adaptation to relative stability, and it is probable that the rate of pumping would never become as high as that observed at salinities of about 28 parts per mille. 8. The lower limit of tolerance, or the minimum salinity at which water is pumped effectively, is between 10.5 and 13 parts per mille. At the former salinity almost no water is pumped, although the valves remain well open. 9. At a salinity of about 13 parts per mille little water is pumped, even after several days are allowed for adaptation, but recovery to normal activity occurs readily following restoration of a higher salinity. 10. A salinity as low as 10.5 parts per mille produces a harmful effect, after which recovery in water of high salinity is extremely slow. Tins effect appears to be within the gill mechanism rather than the adductor muscle. 11. A change to a lower salinity appears to affect the gill mechanism primarily, while following a rise in salinity the adductor muscle tends to hold the valves closer together than normally, resulting in a slower rate of pumping even though the gills may be well adapted. LITERATURE CITED Churchill, E. P., Jr. 1920. The oyster and the oyster industry of the Atlantic and Gulf coasts. Appendix VIII, Report, U. S. Com. Fish., 1919 (1921), 51 pp., 29 pis., 5 figs. Elsey, C. R. 1935. On the structure and function of the mantle and gill of Ostrea gigas (Thun- berg) and Ostrea lurida (Carpenter). Trans., R. S. C., Section V, Biological Sciences, pp. 131-160, IV plates, 1 fig. Galtsoff, Paul S. 1928. Experimental study of the function of the oyster gills and its bearing on the problems of oyster culture and sanitary control of the oyster industry. Bull., U. S. Bur. Fish., vol. XLIV, pp. 1-39, 12 figs. Hopkins, A. E. 1931a. Temperature and the shell movements of oysters. Bull., U. S. Bur. Fish., vol. XLVII, pp. 1-14, 10 figs. Hopkins, A. E. 1931b. The effect of sulphite waste liquor on the oyster ( Ostrea lurida ). In “Effects of pulp mill pollution on oysters”, by A. E. Hopkins, Paul S. Galtsoff, and H. C. McMillin. Bull., U. S. Bur. Fish., vol. XLVII, pp. 125-162, 38 figs. Hopkins, A. E. 1933. Experiments on the feeding behavior of the oyster, Ostrea gigas. -Jour., Exper. Zook, vol. 64, pp. 469-494, 10 figs. Philadelphia. 364 BULLETIN OF THE BUREAU OF FISHERIES Hopkins, A. E. 1934. A mechanism for the continuous circulation and aeration of water in small aquaria. Science, vol. 80, no. 2078, pp. 383-384, 1 fig. Hopkins, A. E. 1935. Temperature optima in the feeding mechanism of the oyster, Ostrea gigas. Jour., Exper. Zool., vol. 71, pp. 195-208, 9 figs. Philadelphia. Nelson, Thurlow C. 1921. Report of the Department of Biology, N. J. Agr. Col. Exper. Sta., year ending June 30, 1920, pp. 317-349, 6 figs., V plates. Trenton. Nelson, Thurlow C. 1923a. On the feeding habits of oysters. Proc., Soc. Exper. Biol, and Med., vol. 21, no. 2, November 1923, pp. 90-91. Nelson, Thurlow C. 1923b. The mechanism of feeding in the oyster. Proc., Soc. Exper. Biol, and Med., vol. 21, no. 3, December 1923, pp. 166-168. o U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner DETECTION AND MEASUREMENT OF STREAM POLLUTION By M. M. ELLIS From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 22 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1937 For sale by the Superintendent of Documents. Washington, D. C. Price 20 cents DETECTION AND MEASUREMENT OF STREAM POLLUTION 1 By M. M. Ellis, Ph. D., Sc. D., In Charge, Interior Fisheries Investigations, United States Bureau of Fisheries , and Professor of Physiology, University of Missouri u* CONTENTS Introduction Stream pollutants and aquatic environ- ment Physical and chemical characteristics of waters suitable for fresh-water stream fishes General field methods Water samples Mud samples Plankton Fresh-water mussels Fish Equipment Dissolved oxygen Hydrogen-ion (pH) limits Ionizable salts Specific conductance Carbon dioxide Fixed carbon dioxide (chiefly calcium and magnesium carbon- ates) Free carbon dioxide Iron Ammonia Suspensoids Depths | Stream pollutants, etc. — Continued. Physical, etc. — Continued. Ionizable salts — Continued. Temperature Bottom conditions as affected by stream pollution Action of pollutants on fishes Injuries to gills and external structures Pollutants entering the body of the fish and exerting true toxic action. Lethality of specific substances occurring in stream pollutants General consideration Test animals Goldfish Daplmia magna Water types Specific lethality tables Osmotic pressure and sodium chloride Acids Compounds of various metals Miscellaneous compounds Lethal limits of 114 substances which may be found in stream pollutants Acknowledgments Bibliography Page 365 366 366 368 369 369 369 369 370 370 370 379 383 383 386 386 388 390 391 394 396 Page 397 398 400 400 402 403 403 404 405 406 406 408 408 409 413 416 417 433 433 INTRODUCTION The menace of pollution to our inland streams and rivers is too well known to require definition. In fact, unsightly and noisome conditions due to pollution are encountered so often that they have come to be accepted by many as the usual order of things. It is true, however, that many cases of pollution could be remedied and the streams so affected restored to an acceptable state for recreation, fishing, and 1 Bulletin No. 22. Approved for publication. Sept. I, 1936. The present study is presented as the first of three dealing with pollution hazards to fresh-water fishes. The second will discuss trade wastes and chemical effluents, and the third, the cumu- lative effects of dilute pollutants. 365 366 BULLETIN OF BUREAU OF FISHERIES general use with reasonable expense, if all parties concerned would cooperate. To obtain this cooperation, it is necessary to understand the situation and to judge it fairly. Much confusion and misunderstanding has arisen in attempts to define the extent of pollution and to place the responsibility for damage to fisheries, because of the lack of available information on the conditions to be defined. In the present paper findings from the widely scattered scientific literature sup- plemented by the experimental and field work of various agencies have been brought together covering (1) the conditions which should be maintained if good fish faunae are to thrive, and (2) the specific effects of various types and components of effluents which now pollute our streams. It is hoped that the use of this information will make possible the definition of undesirable conditions with fairness both to the industrialist, who must use water and streams, and to the citizen, who is entitled to enjoy these same streams. With the limits of both the required stream conditions and of the pollutant lethalities better understood, corrective measures can be recommended intelligently, for remedial action can only be instituted when the cause and the severity of the pollution are known. The United States Bureau of Fisheries is engaged at present in such investigations based on findings presented here. STREAM POLLUTANTS AND AQUATIC ENVIRONMENT PHYSICAL AND CHEMICAL CHARACTERISTICS OF WATERS SUITABLE FOR FRESH-WATER STREAM FISHES The various effluents, municipal, industrial, and otherwise which comprise col- lectively stream pollutants may be detrimental to fishes and other aquatic life either indirectly through quantitative alterations in those substances which give fresh waters their inherent characteristics, as dissolved oxygen, carbonates, and hydrogen ions, or directly because of specific physiological and toxic effects on the aquatic organisms themselves. Many effluents are of complex composition, however, and are harmful to aquatic life through both changes in the aquatic environment and through definite toxic actions. Therefore, in determining the effects of stream pollutants on aquatic biota and particularly on fishes, it has been necessary to study both the modifications in the environment and the specific physiological actions attributable to the different pollutants. The many substances which are carried in solution and suspension by a stream, collectively determine whether the waters of that stream in themselves present con- ditions favorable or unfavorable for fishes and other aquatic organisms; and any individual fish in the stream is affected not only directly by these substances, but indirectly through their action on other forms of aquatic life which comprise in a very restricted environment the food, the enemies, and the competitors of the par- ticular individual. The definition of the amounts of these substances which should be present in water in order to maintain a suitable environment for fishes, or which may be tolerated by fishes under favorable conditions, is therefore much more involved than the designation of standards for water for human consumption, which concern but a single, air-breathing, non-aquatic animal, man. STREAM POLLUTION 367 Water standards for fishes and other aquatic organisms, moreover, are not identical with those standards which will define water as potable for human beings or satisfactory for industrial use. Water may be serviceable for many industries and yet not support fish life, or fishes may thrive in water which would be unsafe for human consumption due to the presence of particular bacteria as typhoid, or certain compounds harmful to man, as the western alkalis. The definition of waters as suitable for aquatic life is complicated still further by the fact that various species of fishes and other aquatic animals and even indi- viduals of different ages of the same species have different degrees of tolerance to deviations from the ideal environment, and to the cumulative effects of many stream pollutants. Consequently, the presence or even the survival for a time of fishes in waters suspected of pollution does not in itself constitute evidence that these waters are either satisfactory or safe for fishes. In spite of the various confusing factors which have been set forth in the pre- ceding paragraphs, it is essential in order to determine the extent and degree of pollution in any given stream to define as far as possible the limits of variation in the several components of those acpiatic complexes which desirable fishes will tolerate and in which they will still thrive. Neither minimal lethal nor arbitrary standards will suffice. The limiting values for the various substances in stream waters, with references to the effects on aquatic life, as presented here have been obtained through the correlation of data of four sorts: (a) The amounts of these substances found in natural waters where fishes were successfully maintaining themselves, (6) studies of streams which as far as could be determined were unpolluted and which, therefore, presented natural conditions, ( c ) the physiological responses of fishes and other aquatic animals to variations in the concentrations of these substances, and ( d ) the survival of aquatic forms when exposed to these substances over long periods under controlled conditions. These data have been drawn from the existing literature, and from field and laboratory studies by the staff of the Columbia (Mo.) field unit of the United States Bureau of Fisheries during the past 5 years (Ellis, 1935a). In the present consideration of water standards for fish and other acpiatic animals the dissolved and suspended substances have been divided into two groups, namely, those constituting the complex favorable to fishes in natural unpolluted waters, i. e., those substances to which the fresh-water fishes are physiologically adapted; and, those substances which are added from time to time to natural waters by man and his agencies, and to which the individual fish must adapt itself. There is, of course, some overlapping between the two groups since certain forms of pollution merely alter the amounts of specific substances normally found in streams, as in the case of the acid wastes from wire-nail mills, which effluents raise the acid ions, the iron and the sulphates, all of which occur in small quantities in most streams, to levels toxic or detrimental for aquatic forms, with, of course, disastrous results. Throughout the application of these data and standards it must be borne in mind that individual fishes and various species of fish have different degrees of resistance and tolerance. Consequently, some fishes may be found in waters where less favorable conditions than those here designated obtain, since both the minimal and maximal limits immediately compatible with life have been avoided, for these limits cannot be regarded as desirable or physiologically reasonable in determining a 368 BULLETIN OF BUREAU OF FISHERIES suitable environment for fishes any more than for man. An effort has been made, therefore, to present usable and reasonable standards of water suitability favorable to fish life; that is, standards defining waters in which a mixed fauna of fresh-water fishes of the common warm water types including desirable centrarcliids, cyprinids, catostomids, and silurids, as well as such tolerant forms as carp and gar, will thrive. Various lethal limits are also set forth. It must also be pointed out that these standards of suitability must be main- tained throughout the periods of low water, maximal temperature, and maximal liability to pollution, since a deviation in the amount of any of several substances, as dissolved oxygen, acids, or salts, to the critical level for only a few hours may so change conditions in a considerable portion of an otherwise favorable stream, that months or years may be required to reestablish the former fish fauna and normal balance of aquatic species. In a large series of field studies it has been shown that the natural, inherent water conditions of most streams can be ascertained satisfactorily for pollution studies as regards fisheries problems by determining repeatedly at different times of the night and day and at various seasons of the year the (1) dissolved oxygen, (2) pH, (3) ioniza- ble salts, (4) carbon dioxide, fixed and free, (5) total ammonia, and (6) suspensoids, since the determinations of these factors not only give specific data concerning partic- ular conditions, but also concerning several complexes which vary in even unpol- luted streams and which are definitely affected by many forms of pollution. From determinations made at many stations where good mixed fish faunae were present, it was found that the \alues from the above determinations in favorable waters, i. e., waters supporting good mixed fish faunae, fell within rather definite limits, and that deviations from these limits in our inland streams were almost always indicative of conditions unfavorable to aquatic life. However, these values alone, which cover only the more basic, inherent conditions which must be maintained in any stream if it is to support a good fish fauna and on which conditions of specific pollution are superimposed, will not suffice for the complete definition of water as favorable for aquatic life, since the absence of specifically toxic substances must also be demon- strated before the water can be finally approved as unpolluted. GENERAL FIELD METHODS Throughout these studies, both for the determination of water suitability standards and of pollution conditions, certain routine procedures were followed, in addition to the special investigations which the conditions in the particular locality required, in order that certain data from all localities could be compared fairly and without the skewing which results from haphazard sampling. The stations at which samples were taken in each locality were selected as representative of the various complexes of conditions presented. Whenever possible, samples were taken at intervals throughout the night and day and at different times during the season. Many of the stations included in this report have been visited repeatedly during the past 5 years. Although in this paper the detailed data concerning the findings on plankton, bottom organisms, fresh-water mussels, and fish population are not presented, since various general statements concerning findings on these animals have been included U. S. Bureau of Fisheries, 1937 Bulletin No. 22 Figure 1.— U. S. Quarterboat S48, the floating laboratory of the United States Bureau of Fisheries. Figure 2. — U. S. Fisheries cruiser 56, equipped with sounding and sampling booms and field-laboratory apparatus. Two such cruisers were used on the larger rivers “and streams. U. S. Bureau of Fisheries, 1 937 Bulletin No. 22 Figure 3. — Collecting bottom samples from cruiser with Peterson dredge. Figure 4. — Mussel dredge in operation from catamaran. This dredge was used in the study of bottom conditions where rocks and other obstructions made use of Peterson dredge impossible. STREAM POLLUTION 369 and since the studies of these forms constitute part of the routine background for the interpretation of the pollution hazards, the field methods employed in connec- tion with the entire pollution study operations have been given. Water Samples At each station water samples were collected by means of a brass sampler of the general type described by Kemmerer, Bovard, and Boorman (1923) and transferred at once without aeration to 300-cubic-centimeter glass-stoppered, self-sealing magnesium-citrate bottles. The first or top water samples were taken at a depth of from 1 to 2 feet, the actual surface water (the first few inches below the surface) being avoided in the general sampling because the dissolved gases in the water in immediate contact with the air do not present a true picture of the dissolved gases in the main mass of the water in the stream (this fact is discussed more fully under dissolved oxygen). Whenever the depth was greater than 3 feet a second set of water samples, bottom samples, were taken about 3 inches above the floor of the stream. Intermediate samples were collected between the top and bottom at levels from 5 to 10 feet apart wherever the depth of the water justified such sampling. Mud Samples Bottom mud samples were obtained with a Peterson dredge properly weighted to bring up about one-half cubic foot of bottom from an area approximately 1 foot square. These mud samples were divided into aliquot parts, some of which were dried (see fig. 3) or fixed with various reagents for chemical determinations, and others were sieved through a series of Monell screens for qualitative and quantitative studies of the bottom fauna. Plankton Plankton counts were made from the catch obtained by pumping slowly 1 cubic meter of the water under consideration through a standard silk bolting cloth plankton net, supported in the water and provided with a glass trap bottle of 250- cubic-centimeters capacity in which the organisms accumulated uninjured. On some of the river lakes, as Lake Pepin and Lake Keokuk, a trawl supported by a boom and operated over the side of a cruiser was used both for plankton and the larger surface animals. For purposes of quick field diagnosis of water conditions only net plankton were used, although in the detailed studies of plankton both nano- and macro- plankton were determined. Fresh- Water Mussels At the stations where fresh-water mussels were to be taken a heavy mussel dredge (see fig. 4) operated from a catamaran was used. This dredge, which was found to be very effective to a depth of 8 or 9 inches in the mud of the river floor, brought up approximately 1% cubic feet of river bottom with the contained organ- isms. Quantitative as well as qualitative studies were possible, therefore, from these dredgings. 370 BULLETIN OF BUREAU OF FISHERIES Fish The fish fauna was studied from collections made with various types of nets and seines. Several hauls were made in each of the various portions of the habitat under consideration and usually the entire catch preserved. If the entire catch were too large, representative collections from it were made and the total bulk of the catch noted. EQUIPMENT In the field studies reported in this discussion several types of equipment have been used. In the major stream surveys U. S. Quarterboat 348 (see fig. I), which is fully equipped with chemical and hydrobiological laboratories, together with living quarters for 12 people, was used as the base from which the operations were con- ducted. Attached to the quarterboat for these investigations were two 35-foot cruisers (U. S. Fisheries 53 and U. S. Fisheries 56, see fig. 2) which were specially equipped with dredging, sounding, and sampling apparatus, together with portable chemical units for such determinations and preparations as required immediate attention in the field. On certain streams where it was not feasible to move the quarterboat and cruisers, support and cooperation in equipment was given by various Government, State, and private agencies, including the loan of vessels and the use of equipment at hand. For certain stream studies automobile trucks carrying compact biochemical, biophysical, and hydrobiological apparatus were also used extensively. These mobile units were found to be very effective, as long distances could be covered quickly and the equipment taken to the specific site of the pollution problem to be investigated. (See fig. 5.) The detailed experiments, most of the bioassay work, and various analyses and tests which could not be made in the field were conducted at the United States Bureau of Fisheries laboratories at the University of Missouri. Here, through the excellent cooperation afforded by the University of Missouri, the Bureau has a research labora- tory suite of eight rooms specially equipped for biochemical, physiological, and bio- logical work, in which various pieces of apparatus devised particularly for these pol- lution studies are in operation. DISSOLVED OXYGEN The modifications of the Winkler method for oxygen determinations as described by Kemmerer, Bovard, and Boorman (1923) and as given by American Public Health Association (1933) were followed in both the field and laboratory analyses for dis- solved oxygen. The data are reported in parts per million (p. p. m.) unless other- wise specified. In figures 6, 7, 8, and 9 are presented the results of 5,809 determinations of dissolved oxygen made at 982 stations on fresh-water streams and rivers of the United States during the months of June to September, inclusive, 1930-35. Natural lakes and ponds are excluded, the data covering only conditions in stream and river waters during the warm season. This season was chosen since the oxygen carrying power of water decreases and the metabolic demands of aquatic animals for oxygen increase as the temperature of water rises. U. S. Bureau of Fisheries, 1937 Bulletin No. 22 Figure 5. — One of the field-laboratory trucks, showing stream-side operations. These trucks were equipped with apparatus and chemicals, sampling and collecting outfits, dredges, and seines. Each truck unit was able, therefore, to make detailed investigations at any stream site which could be reached by car. STREAM POLLUTION 371 From these 5,809 cases, all determinations made at stations where good, mixed, fish faunae were found at the time of sampling, were selected for a composite regardless of the actual amount of dissolved oxygen present. For purposes of this comparison good, mixed, fish faunae were defined as faunae including representatives of the fine fish group (trout, or bass, sunfish, perch, and other spiny-rayed fishes), of the rough fish group (suckers, buffalo, and catfish), and of the minnow series, in good condition Fiqure 6. — The stippled graph is a composite of dissolved oxygen values for all stations at which good mixed fish faunae were found, each river system being given equal weight to avoid skewing due to differences in actual numbers of cases. Each solid black graph presents all oxygen data for a single river unit regardless of presence or absence of fish at sampling stations. Direct evaluation of each river unit can be made in terms of the preferences of good fish faunae from the stippled graph on %vhich the black graph is superimposed. at the time the. samples were taken. Naturally, the exact species composition of these good, mixed faunae varied to some extent with the river system. This composite selected as explained, includes 1,297 determinations of dissolved oxygen from 372 stations. Figure 6 A gives the distribution of the dissolved oxygen values for this composite. To avoid skewing due to uneven number of cases from the different localities the various river systems were given equal weights in determin- ing the percentages. 99773° — 37- 372 BULLETIN OF BUREAU OF FISHERIES From figure QA it may be seen that during the ivarm season the waters at 96 percent of the good fish faunae stations carried 5 p. p. m. or more dissolved oxygen, and that in all of the 5,809 cases good, mixed fish faunae were not found in waters carrying less than 4 p. p. m. dissolved oxygen. These data collected from localities where the fish had had opportunity to choose for themselves point very strongly to 5 p. p. m. as the lower limit of dissolved oxygen, if the complex is to maintain a desirable fish faunae under natural river conditions. E. MISSISSIPPI, IMPOUNDED WATER PERCENT HASTINGS POOL (MINN.) OF CASES H STATIONS, 118 DETERMINATIONS ■*0 - 30 - F. MISSISSIPPI, IMPOUNDED WATER LAKE DAVENPORT AND MOLINE POOL (IOWA. ILL.) 13 STATIONS, 220 DETERMINATIONS SO 40 DISSOLVED OXYGEN Figure 7.— Continuation of the dissolved oxygen compj G. MISSISSIPPI, IMPOUNDED WATER PERCENT LAKE PEPIN (MINN. AND WIS.) OF CASES 22 STATIONS. 286 DETERMINATIONS •40 - H. MISSISSIPPI, IMPOUNDED WATER LAKE KEOKUK (IOWA. ILL.) 2 4 6 6 10 12 14 PARTS PER MILLION sons, stippled graph A in figure 6, being the standard. It maj7 be suggested in the absence of other data that the aggregation of the good fish faimae in waters containing 5 p. p. m. or more dissolved oxygen does not constitute proof that this amount of oxygen was required by these fishes, and that some other factor or factors delimit these complexes where good fish faunae were found. However, experimental data presented in another portion of this section support the view that dissolved oxygen is definitely one of the determining factors in these favorable complexes. A comparison of graphs C and D in figure 6, with the composite A, however, brings out the fallacy, for river waters at least, of the suggestion which has been made at times that the determination of dissolved oxygen or dissolved oxygen in terms of oxygen demand is sufficient alone to define water as suitable or unsuitable for fish life. Graph 6C gives the distribution of the dissolved oxygen value for all stations in STREAM POLLUTION 373 the Mississippi River where good fish faunae were found and graph 6 D for all stations in the Mississippi River where medium, poor, or no fish faunae were found regardless of type of pollution. In this comparison the water at over 98 percent of the stations where good fish faunae were found carried 5 p. p. m. or more dissolved oxygen, and the water at 52 percent at stations where medium, poor, or no fish faunae were found carries less than 5 p. p. m. dissolved oxygen, but the water at 48 percent of the stations where medium, poor, or no fish faunae were found because of other unfavorable con- i. K. PERCENT tW'0 J. SO 40 30 20 10 0 MISSOURI HEADWATERS 24 STATIONS 124 DETERMINATIONS 2 4 0 6 10 12 14 L. TENNESSEE 116 STATIONS 326 DETERMINATIONS 2 4 6 8 10 12 14 DISSOLVED OXYGEN IN PARTS PER MILLION Figure 8.— Continuation of dissolved oxygen comparisons, stippled graph A in figure 6, being the standard. ditions, carried 5 p. p. m. or more dissolved oxygen due to current action and other factors in spite of the polluting materials. Three other observations may be made from these studies of warm-season dis- solved-oxygen values in the various streams and river lakes. Low dissolved oxygen was found more frequently in the impounded waters of streams, particularly if the impounded waters were subject to organic pollution, than in the more rapidly flowing portions of these same streams (see. graph 7 E of Hastings Pool, a notoriously polluted portion of the Mississippi River, and graphs 7 F and 111) ; head water streams, especially those in mountainous regions usually have high dissolved oxygen, and the range of 374 BULLETIN OF BUREAU OF FISHERIES dissolved oxygen to be expected in flowing streams both polluted and unpolluted is in general from 0 to 14 p. p. m. For the final evaluation ol the various findings presented from the field studies on dissolved oxygen, experimental data covering the oxygen requirements of fresh-water fishes must also be considered, as both the amount of oxygen consumed by the fish and the minimal amount of dissolved oxygen which wall barely support life vary with the combinations of environmental factors operating at the time, and with the N. RIO GRANDE P. COLUMBIA RIVER SYSTEM DISSOLVED OXYGEN IN PARTS PER MILLION Figure 9. — Continuation of dissolved oxygen comparisons, stippled graph A figure 6, being the standard. size and species of the fish. Temperature, pH, and dissolved carbon dioxide are particularly important in this connection. Many observers have pointed out that the metabolism of, and consequently the oxygen consumed by, fishes and other aquatic animals follows in general the van’t Hoff law with reference to temperature so that the actual amount of oxygen removed from the water by the individual fish will vary with the temperature regardless of the amount of oxygen present until a near-lethal point is reached (Keyes, 1930). Ruttner (1926) states that the oxygen consumption of many aquatic animals is almost doubled with each rise of 10° C., within physiological limits, and Powers (1922 and 1932) has pointed out correlations between the utilization of oxygen by fishes at low oxygen tensions and the pH and carbonate systems. STREAM POLLUTION 375 The variation in the amount of oxygen consumed by fresh-water fishes of dif- ferent species is apparent from the work of Gardiner, King, and Powers (1922) who give the oxygen consumption of the brown trout as from 90 to 200 cubic centimeters per kilo of body weight per hour at temperatures between 4° and 20° C.; of the goldfish as 16 to 90 cubic centimeters; and of the eel 9 to 60 cubic centimeters. The lower limit for dissolved oxygen — that is, the point at which the dissolved oxygen- — is so reduced as to present a lethal condition for fishes is equally difficult to define. Kupzis (1901) reports that in general the cyprinid, Leuciscus erythrophthalmus (the European roach), could live for sometime in water containing 0.7 cubic centimeter of dissolved oxygen per liter (1 p. p. m.) but that this species of fish died from asphyxia when the dissolved oxygen was reduced to 0.4 to 0.5 cubic centimeter per liter (0.57-0.71 p. p. m.). Plelm (1924) states that trout live best in water containing 7 to 8 cubic centimeters of dissolved oxygen per liter (10-11.43 p. p. m.), but if the water be warm these fish show some discomfort when the dissolved oxygen is reduced to 5.5 cubic centimeters per liter (7.86 p. p. m.). This author also states that carp live well in water containing 5 cubic centimeters dissolved oxygen per liter (7.1 p. p. m.), but show respiratory difficulties when the dissolved oxygen is reduced to 3 cubic centimeters per liter (4.3 p. p. m.) or lower. If the water be cold, carp can live for a short time in water containing only 0.5 cubic centimeter per liter of dissolved oxygen (0.71 p. p. m.). In order of their oxygen requirements, Plehn lists first the salmonids and coregonids, then the barbe, the rutte, the pike, carp, tench, goldfish, and lowest of all the eel. Gardiner and King (1922) give asphyxia! point for trout as from 0.8 cubic centimeter of dissolved oxygen per liter (1.14 p. p. m.) at 6.5° C., to 2.4 cubic centimeters per liter (3.4 p. p. m.) at 25° C.; and for goldfish as 0.39 cubic centimeter per liter (0.56 p. p. m.) at 11° C., and 0.42 cubic centimeter per liter (0.6 p. p. m.) at 27° C. Paton (1904) found that harmful to fatal conditions for young trout developed if the dissolved oxj^gen were reduced to 2 cubic centimeters per liter (2.9 p. p. m.) or lower, although some individuals were able to live in such waters for long periods of time. Thompson (1925) states that carp and buffalo have been found living in water carrying as little as 2.2 p. p. m. of dissolved oxygen. As a rule, he found a variety of fishes only when 4 p. p. m., dissolved oxygen were present, and the greatest variety of fishes were taken from waters carrying 9 p. p. m. of dissolved oxygen. His observa- tions made at Peoria Narrows, 111., in the summer of 1923 showed that fishes died over night in waters having less than 2 p. p. m. of dissolved oxygen. Of the various species of Illinois fish discussed by Thompson the dogfish, Arnia calva, seemed to be the most sensitive to low oxygen tensions and the carp, the most resistant. From the various observations cited above, which may be taken as representative of the voluminous literature on the oxygen requirements of fishes, it may be seen that the upper limit of dissolved oxygen at which asphyxia may be expected in fresh-water fishes if there be no unusual complicating factors, is in general about 3 p. p. m. at 25° C. In evaluating the dissolved oxygen level as a lethal hazard to fish life, this upper asphyxial limit must be considered rather than the lower minimal limit of dissolved oxygen which can be tolerated by some fish for varying periods, particularly under the conditions of rapid or abrupt oxygen reduction presented in so many of the experimen- tal studies, since the actual hazard to fish life begins at the oxj^gen level where 376 BULLETIN OF BUREAU OF FISHERIES death may be expected even if some more hardy individuals survive temporarily or for some time at lower oxygen levels. It is well established for many animals and man, however, that the asphyxia! oxygen level, the oxygen level which will support life if profound compensations be made, and the oxygen level at which respiratory, cardio-vascular, and other systemic compensations begin are three quite different values. The latter — that is, the oxygen level at which respiratory and circulatory compensations are initiated — marks the lower limit as regards oxygen percentage of the favorable respiratory environment, although this oxygen level is much higher than the lethal oxygen level for these same species. It has been shown (Ellis, 1919) that although man lives in air normally containing 21-percent oxygen, and the collapse point for most human beings is reached when the oxygen in the air breathed is reduced to approximately 6 percent, that human respiratory compensation to reduced oxygen begins at about 18 percent oxygen. Applying this same principle to fishes in the present studies it has been found that individual goldfish, perch, catfish, and other species of fresh-water fishes in good condition and from favorable environments if placed in water of constant flow, favorable composition, and temperature (20° to 25° C.) may show respiratory compensations in rate and volume or both when the dissolved oxygen in this water is reduced only a little below 5 p. p. m. As there are various factors influencing the exact point at which this respiratory compensation begins, the details will be pre- sented elsewhere ; but the important finding in connection with the present discussion is that even under conditions as favorable as may be met with in fresh-water streams, respiratory compensations by fishes to oxygen reduction may begin when the dis- solved oxygen level is still almost 5 p. p. m. From these studies the case of a perch which made respiratory compensation to reduced oxygen largely by rate is presented (fig. 10) as typical. This finding that respiratory compensations by fresh-water fishes to oxygen reduction may begin when the dissolved oxygen of water is lowered only to approxi- mately 5 p. p. m. is in accord with the statement of Plebn (1924) that carp show respiratory difficulties when the dissolved oxygen is reduced to 4.3 p. p. m. and gives physiological background for the differences in fish faunae between waters carrying 4 p. p. m. or less dissolved oxygen and those carrying 5 p. p. m. or more dissolved oxygen, as reported in our field studies. Considering the data from all sources and particularly from the field and labora- tory studies presented here 5 p. p. m. of dissolved oxygen seems the lowest value which may reasonably be expected to maintain in good condition varied fish faunae of warm-water fishes in our inland streams, if the water temperature be 20° C. or above. This statement does not mean that 5 p. p. m. dissolved oxygen is the lethal point for fresh-water fishes but designates 5 p. p. m. as approximately the lower limit of favorable conditions. The fact that fish on occasion can tolerate for a period of hours or even days water carrying less than 5 p. p. m. dissolved oxygen does not justify the acceptance of such conditions as defining any stream as suitable for fishes, in view of the data presented on the natural preference by fresh-water fishes for waters containing 5 p. p. m. or more of dissolved oxygen, and the experimental evidence from various sources that vital compensations may be called for in waters carrying less than 5 p. p. m. dissolved oxygen at temperatures of 20° C. or above. STREAM POLLUTION 377 Besides it is recognized in the various fields of animal husbandry, including aquiculture and pisciculture, that if any variety of animal is to be successful and thrive merely sublethal conditions are not adequate. The application of this observation concerning sublethal conditions is particularly important in pollution investigations, for a reduction of dissolved oxygen increases the lethality of many stream pollutants, especially those injurious to the gills (v. i.). For example, it was found that certain concentrations of various metallic salts and RESPIRATION RATE PER MIN. Figure 10. — Changes in the respiration rate of a 4^-inch yellow perch, Perea flavescem, correlated with changes in the dissolved oxygen content of the surrounding water, temperature 18 degrees centigrade. acid wastes were consistently more injurious to fish when the dissolved oxygen con- tained in the water thus polluted was comparatively low, but still sublethal, than when the dissolved oxygen was high. This difference in lethality correlated with dissolved oxygen obtained in spite of the fact that these specific pollutants neither reacted with nor were removed by dissolved oxygen regardless of the amount of oxygen present. The dissolved oxygen content of unpolluted streams normally varies with at least four major sets of factors, namely, (a) physical conditions such as stream flow, stream fall, and temperature, which influence the saturation of water with oxygen from the air; ( b ) oxygen produced by aquatic plants; (c) oxygen removed by aquatic organisms both plant and animal; and (d) the oxygen demand of the organic detritus 378 BULLETIN OF BUREAU OF FISHERIES in the stream. Effluents polluting streams alter this dissolved oxygen balance chiefly through increasing the oxygen demand. Pollutants may create oxygen demand in any of three ways or combinations of these; that is, by the addition of quantities of organic matter of a putrescible sort, among which may be mentioned domestic sewage, packing-plant wastes, beet-sugar waste, and hide vat liquor from tanneries; through the action of various reducing chemicals as certain sulphite wastes from paper mills, sulphide and iron wastes from mines, and certain spent dyes from leather works; and by killing large masses of aquatic vegetation which subsequently decompose. Many effluents which create oxygen demand are also harmful to aquatic organisms because of specific toxic effects. The determination of the biochemical oxygen demand of various effluents affecting the dissolved oxygen balance in the stream does not necessarily, therefore, give a true evaluation of the pollution hazards produced by these effluents ; for various species of warm-water fishes will live in water having high oxygen demand, due to domestic sewage or industrial wastes rich in organic material but without specific toxic substances, if the dissolved oxygen level be maintained above 5 p. p. m. and the aeration be sufficient to blow off the excess of other gases as carbon dioxide, methane, and sulphur derivatives. The biochemical oxygen demands of such effluents as have high oxygen demands must be taken into account, however, if these wastes are poured into streams, in order to compute the dilution required to prevent this oxygen demand from lowering the dissolved oxygen in the stream to an unfavorable level. Two examples of the effect of dissolved oxygen on the lethality of effluents free from specific toxic substances but presenting large pollution hazards because of high oxygen demands will suffice. The tan vat liquor from a tannery on the upper Missis- sippi River in spite of the high oxygen demand of this waste was not only tolerated readily when properly aerated and unmixed with the toxic chemical effluents from the plant by various warm water fishes but materially increased the production of plankton when added to the river water. Findings from the sewage treatment system of the city of Munich, Germany, also demonstrate that high biochemical oxygen demand in itself need not be a pollution hazard to fresh-water fishes. In this system, water from the River Isar is mixed with prepared sewage and impounded for the com- mercial raising of rainbow trout and carp. This procedure was very successful over a period of years in the production of commercial quantities of both trout and carp. The following statement from a recent English review of the work of this plan (Engineering, 1935) may be quoted in this connection: The proportion of fresh water to sewage should be at least 5:1, and the proportion of oxygen not less than 5 cubic centimeters per litre (7.1 p. p. m.). If these conditions are maintained, the micro-organisms will flourish, absorbing and destroying all impurities and providing the food for the completion of the process. In such circumstances the fish will suffer no ill effects, although, as is well known, trout are most susceptible to any impurity. The oxygen balance of streams may also be affected by oils which exclude oxygen from the surface of the stream and prevent proper reaeration of the water. Oil is rarely poured into streams in such quantities as to present this difficulty, however. Again sewage and industrial sludges are often particularly harmful in reducing the dissolved oxygen through their biochemical oxygen demands during the winter season when ice covers the stream surface and interferes with the reoxygenation of the water. STREAM POLLUTION 379 HYDROGEN-ION (pH) LIMITS Hydrogen-ion concentration measurements in the laboratory and aboard Quarter- boat 348 were made electrometrically with calomel and quinhydrone electrodes (Leeds, Northrup Co., Philadelphia, type 7701-Al). In the field the Youden appa- ratus (W. M. Welch Co., Chicago, type 5270) bridging the quinhydrone electrode against standardized phthalate solution was used because of greater portability. When field conditions were such that neither of these two pieces of apparatus were available, the determinations of hydrogen-ion concentrations were made colorimet- rically by the method of Gillespie (1920). Sterile tubes, solutions, and stoppers were used in preparing the tube series for the Gillespie sets, which were calibrated against the calomel electrode and kept in a dark box except when measurements were being made. Although fractional values were recorded (i. e., hundredths), in presenting the data the value is given to the nearest tenth pH. The hydrogen-ion concentration of the inland streams of the United States, southern Canada, and northern Mexico, excepting badly polluted portions of these waters, as seen in a review of some 10,000 readings made during the past 5 years, lies in general between values of pH 6.7 and pH 8.6, with the extreme range (in our data) of pH 6.3 and pH 9.0 in streams for which no specific pollution factor affecting the hydrogen-ion concentration was readily observable. Swamp waters, bog streams, and particularly swam]) lakes not infrequently show an acidity between pH 4.5 and pH 6.0, vet at the same time may support mixed fish faunae. In some small western streams and pools containing fishes examined by the writer in North Dakota, Mon- tana, and New Mexico an alkalinity of pH 9.5 was occasionally found in alkali dis- tricts or near mineral springs. In such waters small poeciliids and cyprinids were uniformily the dominant fishes. In figures 11, 12, 13, and 14, 7,228 pH readings from inland stream waters repre- senting 1,125 localities are presented. These data were collected during the warm season, June to September, 1930-35, and are comparable with the dissolved oxygen data in the preceding section. The composite, figure 11 A, was constructed from 2,280 readings representing 409 localities where good fish faunae, as previously defined, were found. This composite covers a range between pH 6.3 and pH 9.0, with 97 percent of the cases between pH 6.7 and pH 8.6. Superimposing the graph of the composite on the graphs presenting the data from the various river systems, however, shows that, except in cases of extreme pollution, the pH values as such of stream waters, both polluted and unpolluted, do not differ materially from those of the composite. The data from the various river systems (other than the composite) are pre- sented without regard to pollution, except that no pH readings of water in the im- mediate vicinity of flumes and conduits from which effluents were escaping have been included. The extreme pH range of the flowing waters of inland streams of the United States, both polluted and unpolluted, as found in these field studies was pH 3.9 to pH 9.5, although various effluents poured into these same waters were found to range from pH 1.0 to pH 11.0 at the point of entrance into the stream. These observations show that dilution and the buffer action of various substances 99773° — 37 3 380 BULLETIN OF BUREAU OF FISHERIES PERCENT A- COMPOSITE, LOCALITIES WITH GOOD FISH FAUNA OF CASES *09 STATIONS, 2,280 DETERMINATIONS PERCENT C. MISSISSIPPI, GOOD FISH FAUNA OF CASES 114 STATIONS, 477 DETERMINATIONS 60 B. ATLANTIC COAST STREAMS 106 STATIONS, 165 DETERMINATIONS 60 40 - Figure 11. — Comparisons of pH values of various river units. D. MISSISSIPPI, MEDIUM, POOR, OR NO FISH FAUNA 105 STATIONS, 334 DETERMINATIONS Composite and black graphs as explained under figure 6. E. MISSISSIPPI, IMPOUNDED WATER PERCENT OF CASES 100 { — HASTINGS POOL f MINN.) 12 STATIONS, 136 DETERMINATIONS 80 - 60 - <0 - 20 C G. MISSISSIPPI, IMPOUNDED WATER PERCENT LAKE PEPIN (M!NN.,WIS.J OF CASES 23 STATIONS, 260 DETERMINATIONS 100 60 - 60 40 - 20 - 0 F. MISSISSIPPI, IMPOUNDED WATER LAKE DAVENPORT AND MOLINE POOL C IOWA ILL > H. MISSISSIPPI, IMPOUNDED WATER LAKE KEOKUK CIOWA ILL.) Figure 12.— Continuation of comparisons of pH values for various river units, stippled graph A figure 11, being the standard STREAM POLLUTION 381 in the river waters do change the pH values of the extremely acid and extremely alkaline wastes rather rapidly to the range of the composite pH 6.3 to pH 9.0. Reviewing all of the field pH data it may be observed within the general range of unpolluted water, pH 6.7 to pH 8.6, that mountain streams, particularly head- water streams, are in general more acid than plains or lowland streams. The Columbia River data (fig. 14P) bring out this comparison in the bimodal distribution of the pH values, since portions of the Snake River flowing through plains and desert regions and the mountainous headwaters of both the Snake and Columbia Rivers J. MISSOURI HEADWATERS 24 STATIONS, 25 DETERMINATIONS L. TENNESSEE 116 STATIONS. 316 DETERMINATIONS 100 Figure 13 —Continuation of comparisons of pH values, stippled graph A figure 11, being the standard. are included in this graph. This same distribution of pH values is also apparent in the graphs presenting the findings in the two parts of the Missouri River system (figs. 13 1 and 13J). The explanation of these distributions of pH values is clear when the carbonates and conductivities of these waters are also considered (v. i. ) , as the headwater streams contain much less dissolved solid matter than the plains, lowland, and desert streams. The effects of acid pollution are evident in the Ohio River (fig. 13 K), and of general pollution and some bog waters in the Atlantic coast streams (fig. 11 B). The field observations presented here agree in general with the statement made by Shelford (1929) that the hydrogen-ion concentration from pH 6.5 to pH 8.5 may 382 BULLETIN OF BUREAU OF FISHERIES be expected in most uncontaminated fresh-water streams and lakes; and with the observations of Powers (1921 and 1929); and of Juday, Fred, and Wilson (1924). Experimental tests have demonstrated that many species of fresh-water fishes have a great tolerance for variations in hydrogen-ion concentrations over a wide range. Creaser (1930) showed brook trout to have a voluntary toleration of liydro- gen-ion concentrations from pH 4.6 to pH 9.5. Brown and Jewell (1926) found catfish and perch living in apparently good condition in a bog lake, the water of which ranged from pH 4.4 to pH 6.4; and also in a glacial lake nearby, the water of which varied N. RIO GRANDE P. COLUMBIA RIVER SYSTEM 44 STATIONS. 153 DETERMINATIONS 3.9 ' 4.3 ' 4.7 ' 5.1 1 5.5 1 5.9 1 6.3 1 3.7 1 4.2 4.8 5.0 5.4 5.8 6.2 6-6 7.0 4.2 4.6 5.0 5.4 5.8 6.2 6.6 7.0 7.4 7.8 8.2 8.6 9.0 Figure 14. — Continuation of comparisons of pH values, stippled graph A figure 11, being the standard. 7.5 ' 7.9 ' 8.3 ' 8.7 7.8 a 2 8.6 9.0 from pH 8.2 to pH 8.7. These workers demonstrated that the fishes from the two lakes survived transfer from either lake to the other. Wiebe (1931a) reports goldfish survive rapid changes from pH 7.2 to pH 9.6; largemouth black bass, from pH 6.1 to pH 9.5; smallmouth black bass, from pH 6.6 to pH 9.3; and sunfish from pH 7.2 to pH 9.6. Powers (1930) in reviewing the problem summarizes the existing data by saying that aquatic organisms are able to withstand a wide range in pH. The writer has confirmed this statement with gammarids, daphnia, unionids, and planaria, as well as for goldfish, perch, and catfish, in connection with pollution tests. It might seem, therefore, from both the field data and laboratory findings that the pH values of stream water would be of little consequence in pollution studies and in determining standards of water suitability. Plowever, the pH of natural water is STREAM POLLUTION 383 determined by substances in solution — particularly carbonates, carbon dioxide, vari- ous salts, and a few organic substances — which collectively constitute a poor to fair buffer system; so that water more acid than pH 6.7 or more alkaline than pH 8.6 is not found generally in our inland streams, unless there be some unusual factor in the complex. Pond water, bog water, and lake water vary over a wider range, but the combination of stream flow, aeration and buffer substances holds the hydrogen-ion concentration of the larger rivers, the smaller streams, and even many brooks within the limits described. In pollution studies, therefore, it has been found advisable to view with suspicion any stream water having an hydrogen-ion concentration outside of the limits pH 6.7 to pH 8.6, until it could be definitely shown that such extra limital pH values were due to natural causes rather than to human agencies, as even badly polluted streams were usually within these limits. If the water of flowing streams were more acid than pH 6.7 or more alkaline than pH 8.6 as the result of the addition of municipal or industrial effluents the buffer and carbonate systems were usually so disturbed that conditions harmful to fishes were generally found. The determina- tion of pH, therefore, is an important aid in the study of polluted water in spite of the range of tolerance of fishes to pH changes in unpolluted waters, because excessive variation in hydrogen-ion concentrations is indicative of harmful changes in the com- plex of dissolved substances, both solids and gases, normally found in river water. Among the effluents which change the pH of stream water and break down the buffer systems are the wastes from wire-nail mills, tin-plate mills, and other metal works where acid washes are used; wastes from chemical works, particularly dye mor- dant and soda compounds; spent liquors from chrome tanning processes; whey- containing fluids from dairy products concerns; laundry waters; and some battery factory wastes. The waters from unsealed coal mines also add large quantities of acid wastes. The specific effects of high and low hydrogen-ion concentrations are dis- cussed in detail in the section on acid pollutants. (See p. 409.) IONIZABLE SALTS Specific Conductance Measurements of specific conductance were made with standard glass cells con- taining coated platinum electrodes in telephone circuit with a microlmmmer and standard variable resistance units. This apparatus was found to be very sturdy and was regularly used both in the laboratory and in the field. For convenience in pres- entation, the data are expressed as specific conductance in mhoXUU6 at 25° C. Unpolluted natural waters contain in solution small quantities of carbonates, chlorides, phosphates, and sulphates, usually some nitrates and nitrites if organic matter be present, and traces of many other salts which vary with the region through which the stream flows. The metallic ions represented are largely calcium, magne- sium, sodium, potassium, iron, and manganese, with traces of various other elements. Owing to the fact that carbon dioxide is supplied to stream water from so many sources, carbonates are the dominant salts; but because of the low solubility of most carbonates and also of most phosphates, the mineral content of river water never rises very high unless some particular substance is added to the water which will raise the solubility of these compounds or transform them into other more soluble compounds. All of the substances in solution in river water collectively exert 384 BULLETIN OF BUREAU OF FISHERIES osmotic pressure on the aquatic organisms living in the water, and many of these compounds are physiologically active, so that fresh- water fishes and other animals living in these streams have become adapted to the physical and physiological actions of this salt complex. Small variations in several of these salts may cause small variations in the species composition, particularly the invertebrates, of the faunae at any given station. Most aquatic species, however, will tolerate changes of considerable magnitude in the relative amounts of these salts normally present in flowing waters, if the very small total amount which is usually present be not exceeded. For example, the fixed carbonates in the upper Tennessee River were found to vary from 0.4 to 30 cubic centimeters per liter (computed as C02 by volume) without affecting the general composition of the aquatic fauna, and from 0.3 to 55.5 cubic centimeters in Spider Creek, a tributary to the Wabash River. The specific quantities of most of the substances comprising this salt complex are not so important as the total quantity of soluble matter involved (see section on osmotic pressure), since even the small quantities of these compounds present in ordinary soil run-off are in excess of the physiological needs of most fresh-water organisms. As in general these substances are ionizable, measurements of the spe- cific. conductance of a large number of polluted and unpolluted waters were made. The summarized data are presented in figure 15. These determinations of specific conductance of inland fresh waters show that, excepting the streams in the plains and desert regions, the specific conductance of those portions of inland streams and rivers which were supporting good, mixed fish faunae in general lay between 150 and 500 mho X 10-6 at 25° C. This uniformity of water composition in flowing streams holds even in the very deep holes of rivers. Two cases — Pan Eddy, in the Tennessee River, and the deep hole off Grand Tower, 111., in the Mississippi River — are presented as typical in table 1. Only in the very deep portions of river lakes, such as Lake Wilson in the Tennessee and Elephant Butte Reservoir in the Rio Grande, is there any marked stratification of the waters. In the deeper portions of some of these impounded waters (see table 4) definite thermal stratification, with the attendant changes in turbidity, dissolved oxygen, pH, and other physiochemical features of the water, develop in midsummer at levels variously determined by the general climatological features of the region and the level and amount of draw-off as made for the needs of industry and navigation. In such river lakes a warm, more turbid stream — the hyperlimnorrheum — flows over a colder, clearer lake — the hypolimnion — with a rather well-delimited thermocline lying between these. This condition of stratification obtains, however, only during a portion of the warm season. Specific data for one station on Lake Wilson (see table 4) show such stratification of the waters when it was at its height. During the colder portions of the year there is a rather complete mixing of the waters at all depths, even in Lake Wilson. Details of these hydrobiological data of various river lakes are presented elsewhere. The specific conductance of mountain stream waters was generally in the lower part of this range, unless excess carbon dioxide were present. However, in the streams of the western plains and desert areas, particularly those carrying the more alkaline STREAM POLLUTION 385 waters, as the Snake (Columbia River system), the Rio Grande (below headwaters), the Colorado and Gila Rivers, and the upper Missouri and Yellowstone below the mountains, a specific conductance of 2,000 mho was not unusual, the general range of specific conductance of these streams being between 200 and 2,000 mho, with SPECIFIC CONDUCTIVITY IN MHO * IQ* AT 23* C RIVER SYSTEMS 0 499 300 999 IOOO 1490 1300 1000 2000 2499 2300 2999 3000 3499 3300 3999 4 000 4499 4500 4999 COMPOSITE LOCALITIES WITH GOOO FISH FAUNA 381 DETERMINATIONS MISSISSIPPI GOOD FISH FAUNA 222 DETERMINATIONS MISSISSIPPI MEDIUM, POOR, OR NO FISH FAUNA 263 DETERMINATIONS MISSISSIPPI IMPOUNDED WATERS HASTINGS POOL (MINN.) 7 DETERMINATIONS 1 MISSISSIPPI IMPOUNDED WATERS LAKE PEPIN (MINN. WIS > 180 DETERMINATIONS | MISSISSIPPI, IMPOUNDED WATER LAKE DAVENPORT AND MOUNE POOL (IOWA, ILLINOIS) 174 DETERMINATIONS 1 MISSISSIPPI IMPOUNDED WATERS LAKE KEOKUK ( IA.. ILL.) 722 DETERMINATIONS h MISSOURI HEADWATERS 10 DETERMINATIONS .ti MISSOURI BELOW HEADWATERS 133 DETERMINATIONS OHIO 40 DETERMINATIONS b TENNESSEE 144 DETERMINATIONS 1 ATLANTIC COAST STREAMS 47 DETERMINATIONS L GULF COAST STREAMS 10 DETERMINATIONS d RIO GRANDE 23 DETERMINATIONS 1 COLORADO RIVER SYSTEM 41 DETERMINATIONS to COLUMBIA RIVER SYSTEM 39 DETERMINATIONS du 0 499 1 M T 300 999 n 1 1 IOOO 1499 nil 1300 1999 2000 2499 2300 2999 3000 3499 3300 3999 4000 *499 4300 4999 SPECIFIC CON&UCTIVITY-IN MWO * 10® AT 23° C Figure 15. — A comparison of the specific conductivity values of the waters of various river units. extreme readings of almost 5,000 mho X 1 O ' at 25° C. However, good mixed fish faunae of such species as normally inhabit these western streams were usually not found in waters with specific conductance greater than 2,000 mho X 10“6 at 25° C. These high conductivity values for certain western waters were clearly due to the natural high solids content of these waters which drain areas rich in the various soluble alkalis. 386 BULLETIN OF BUREAU OF FISHERIES From the field studies on specific conductance it was found advisable in pollution studies to look for specific pollutant action if the conductivity of the water exceeded 1.000 mho in all types of streams except those draining the more alkaline regions, or 2.000 mho in the case of the western streams discussed above, for the specific conduct- ance is very readily raised by acid and salt pollutions of several sorts. Specific conductance, therefore, offers a ready method for detection of salt and acid pollution, as produced by water from oil wells, by wastes from industries using salts or strong acids, and by soluble salts of heavy metals. This method was used very satisfactorily by the writer in determining the extent of downstream pollution from lead and zinc mines in the Coeur d’Alene district in Idaho (Ellis, 1932), and in following pollution by acid mine waters from coal mines in Tennessee and West Virginia. Carbon Dioxide Determinations of free, half-bound and fixed carbon dioxide were made by the methods of Seyler as described by Kemmerer, Bovard, and Boorman (1923) and American Public Health Association (1933). The data are expressed as cubic centi- meters of carbon dioxide per liter. FIXED CARBON DIOXIDE (CHIEFLY CALCIUM AND MAGNESIUM CARBONATES) In figures 16 and 17 the data from 6,770 determinations of fixed carbon dioxide in flowing streams of the United States and southern Canada collected during the warm season, June to September, inclusive, 1930-35, are presented. The extreme range of fixed carbonates varied from almost 0 to 70 cubic centimeters per liter, with 40 cubic centimeters per liter as the usual upper limit. A composite of determinations from waters where good fish faunae were thriving (fig. 16 A) covered the entire range from 0 to 70 cubic centimeters per liter. In this composite 96 percent of the 2,190 analyses showed the stream water to carry less than 50 cubic centimeters per liter of fixed carbon dioxide and 53 percent lay between 10 and 30 cubic centimeters per liter. Comparing these findings with the classification of lake water given by Birge and Juday (1911), it may be seen that most of the flowing waters fell in the second class, designated as medium waters; and that the large percent of the river and stream waters could be classified as either soft or medium. Hard waters were distinctly in the minority. Reviewing this composite in connection with the findings on fixed carbonates in the waters of the various river systems, it is evident that between 10 and 40 cubic centimeters of fixed carbon dioxide may be expected in most of the fresh-water streams of the United States. Fixed carbon dioxide in itself, however, does not seem to be a determining factor in classifying flowing waters as suitable or unsuitable for fresh- water fishes. Various limnologists have pointed out the importance of fixed car- bonates in maintaining food supply, and, therefore, fish faunae of various lakes (Welch, 1935), but from the data on rivers and streams presented here it must be concluded that either carbonates are not limiting factors or that they are always present in sufficient quantities to meet the needs of the stream biota. An exception to this last statement, however, must be taken at least in the case of fresh-water mus- sels, as carbonates do constitute a very vital limiting factor for these animals. STREAM POLLUTION 387 In general, it was found that the fixed carbonates vary with the character of the country drained, mountain streams in regions where igneous rocks dominate being very low in fixed carbonates and streams traversing limestone country or alkaline soil being relatively high. However, it must be added that excepting the headwater streams in mountainous regions the fixed carbonate load of flowing stream water was rarely less than 5 cubic centimeters per liter. A. COMPOSITE LOCALITIES WITH GOOD PERCENT FISH FAUNA B. ATLANTIC COAST STREAMS 61 STATIONS 109 DETERMINATIONS C. MISSISSIPPI GOOD FISH FAUNA 108 STATIONS 494 DETERMINATIONS D. MISSISSIPPI MEDIUM, POOR, OR NO FISH FAUNA 87 STA., 284 OETERMIN HASTINGS POOL C MINN. ) 8 STATIONS 112 DETERMINATIONS F. MISSISSIPPI, IMPOUNDED WATER LAKE OAVENPORT AND G. MISSISSIPPI, IMPOUNDED WATER LAKE PEPIN ( MINN. , Wl S .) 22 STATIONS H. MISSISSIPPI, IMPOUNDED WATER LAKE KEOKUK CIOWA, ILL.) 25 STATIONS Figure 16. — Fixed carbon dioxide values (as cubic centimeters of carbon dioxide per liter, bottom line) for the waters of various river units. Stippled and black graphs composed as explained under figure 6. In pollution studies, although fixed carbonates may not be important in defining particular water as suitable or unsuitable for good fish faunae, the fixed carbon dioxide is of large importance in evaluating the extent of certain types of pollution and in determining the amount of dilution required to render certain effluents relatively harmless to aquatic life. The fixed carbon dioxide of the stream water is the major buffer against which both acid and alkali pollutants act, and the fixed carbonate must be taken into consideration, therefore, when the maximal nontoxic dilution of any effluent is to be determined. Tiffs problem is discussed more fully under acid wastes, but it may be pointed out here that the toxicity to fishes and other aquatic life of many of the common acids is much greater in distilled water or water carrying very little fixed carbon dioxide than in water carrying 20 to 60 cubic centimeters per liter of fixed carbon dioxide. The reason is obvious because of the buffer and 99773° — 37 4 388 BULLETIN OF BUREAU OF FISHERIES neutralizing actions of these fixed carbonates on free acids. Again, the fixed carbon dioxide is of considerable importance in detoxifying various effluents, particularly those carrying salts of heavy metals, by precipitating various compounds, since many carbonates are relatively insoluble and, therefore, are removed at least temporarily from the waters during precipitation. This is discussed more fully in the section on metallic poisons. J. MISSOURI HEADWATERS K.OHIO L. TENNESSEE 16 STATIONS 33 STATIONS 103 STATIONS 121 DETERMINATIONS IGS DETERMINATIONS 163 DETERMINATIONS M. GULF COAST STATES 29 STATIONS 1,317 DETERMINATIONS N. RIO GRANDE ft STATIONS 129 DETERMINATIONS O. COLORADO RIVER SYSTEM 10 STATIONS Ml DETERMINATIONS 100, — GO ©Oh P. COLUMBIA RIVER SYSTEM 33 STATIONS 265 DETERMINATIONS IOO, Figure 17. — Continuation of comparisons of fixed carbon dioxide values, stippled graph A figure 16 being the standard. FREE CARBON DIOXIDE Figures 18 and 19 present the data from 3,351 determinations of free carbon dioxide in various streams of United States. From these figures, and particularly from the graph of the composite, it may be seen that river and stream waters, where good fish faunae were foimd, carried consistantly less than 5 cubic centimeters of free carbon dioxide per liter, 90 percent of the waters where good fish faunae were taken carrying less than 2 cubic centimeters of free carbon dioxide per liter. It must be pointed out in this connection that conditions in moving streams are quite different as regards free carbon dioxide from those obtaining in lakes and ponds. Due to the constant turnover of river and stream waters because of current action, there is a very uniform mixing of water from surface to bottom (see data from Grand Tower on the Mississippi and Pan Eddy in the Tennessee (table l)), so that the free carbon dioxide, as well as the other dissolved gases, is quite uniformly distributed throughout the STREAM POLLUTION 389 stream. Again, the movement of the current facilitates the constant reaeration of stream waters with a tendency for these waters to lose carbon dioxide to the air as soon as the carbon-dioxide tension in the water exceeds that of the surrounding air, which is usually quite low (2 to 4 parts per 10,000). Both of these conditions, to- gether with the usual presence of calcium and magnesium salts in stream waters, tend to keep these flowing streams more alkaline than pond, lake, and bog waters, so that in many river waters the alkalinity is raised to a point where there is little or no uncombined carbon dioxide. Table 1. — Characteristics of water at different depths in deep holes of flowing rivers, showing general uniformity of composition at any given station due to mixing action of current I. UPPER TENNESSEE RIVER, AT PAN EDDY, 18 MILES BELOW CHATTANOOGA, TENN., AUG. 21, 1931, AIR TEMPERATURE, 23° C. Depth in feet Water tem- perature, de- grees centi- grade pH Specific con- ductivity in mhoXlO-6 at 25° C. Fixed carbon dioxide, cubic centimeters per liter Dissolved oxygen parts per million Light pene- tration, mil- lionth inten- sity depth in millimeters S> 28.0 7.3 169 15.4 5. 6 .544 10 28.0 7.2 169 15.4 5. 5 561 20 27.9 7.3 169 14.9 5.7 561 30 _ 27.9 7.3 169 14. 9 5.9 561 40 . 27.8 7.3 169 14. 9 6. 0 561 50 27.8 7.3 169 14.9 5.7 561 00 ___ 27.8 7.3 109 14.9 5.2 581 70 ... 27.8 7.4 169 15. 2 5.5 552 80 27.8 7.4 169 15.2 5.7 552 90 27.8 7.4 169 15. 2 5.5 544 100 27.6 7.4 169 15.4 5.6 538 no . ... .. 27. 7 7.3 168 14. 9 5.7 528 120 27. 8 7.3 168 14. 9 6. 1 521 130 27.8 7.3 167 15. 1 5.9 513 II. MISSISSIPPI RIVER, OFF TOWER ROCK, NEAR GRAND TOWER, ILL., SEPT. 8, 1931, AIR TEMPERATURE 30.5° C. S > 28.2 7.5 340 26.4 5.3 129 10 26.0 7.5 341 26.6 5. 1 125 20 25.8 7.6 339 26.8 5.3 125 30 26.0 7.5 337 26.4 5.3 122 40 26.0 7.5 338 26.4 5.3 129 50 26.0 7.5 340 26.4 5.3 120 60 26.0 7.5 331 26. 6 5.3 122 70 20. 2 7.4 338 26.7 5.3 122 80 . . . 26.0 7.4 342 26. 7 5.3 125 90 __ 26.0 7. 4 339 26.8 5. 4 122 10(1 .. 26.0 7.4 338 26.8 5.4 119 110 •26.0 7.4 341 26.8 5.3 124 1 Surface, i. e., top 12 inches of water. The specific toxicity of carbon dioxide for aquatic organisms is well known and is discussed under acid wastes. From all of the field data and in view of the experimental findings presented, the determination of free carbon dioxide was found to be a valuable aid in pollution studies, and river and stream waters carrying more than 3 cubic centi- meters of free carbon dioxide per liter were checked carefully for some source of organic pollution. Values in excess of 3 cubic centimeters per liter usually were indicative of such pollution in our flowing inland streams. The relatively high free carbon dioxide as compared with the composite found in Hastings Pool in the upper Mississippi River and in the polluted portions of the Atlantic coast streams and other systems are evident from the graphs presented in figs. 18E and ISA’. 390 BULLETIN OF BUREAU OF FISHERIES IRON Iron determinations were made by the method of American Public Health Asso- ciation (1933), using permanent standards of cobalt and platinum salts. A limited number of total iron determinations were made in typical localities for correlation with studies of water from abandoned coal mines. Eighty-three deter- minations of total iron in waters of Mississippi, Missouri, Ohio, Tennessee, Atlantic coast, and Gulf coast systems gave the following grouping: Figure 18. — Free carbon dioxide values (as cubic centimeters of carbon dioxide per liter, bottom line) for the waters of various river units. Stippled and black graphs composed as explained under figure 6. At stations where good fish faunae were found (75 cases) the free iron ranged from a trace to 30 p. p. m., with 69 of these 75 cases below 10 p. p. m. In streams polluted with either mine waters or chemical effluents carrying iron compounds the total iron ran much higher, the maximum found in this series being 202 p. p. m. in waters of a stream of high acidity (pH 4.4) flowing from a coal mine in operation near Turley, Tenn. Iron in quantities above 100 p. p. m. was associated with acidity above average and when such conditions were found chemical or mine pollutions were suspected. In the experimental tests (see section on heavy metals) it was found that 100 p. p. m. of total iron were not immediately harmful to either fish or daplmia; and that such quantities, unless a constant flow were maintained, were usually diminished during the first few davs of experimental tests as a result of precipitation so that the STREAM POLLUTION 391 free iron in the water did not exceed 20 p. p. m. Iron as it occurs in most natural flowin<>: waters, therefore, is not a major limiting factor in the distribution of fishes in fresh-water streams. Besides in fresh-water streams and rivers the constant re- aeration of water readily offsets any important oxygen loss through iron compounds, as has been reported from bog lakes and certain deep lakes where large quantities of iron were found. In pollution studies of the inland fresh-water streams, therefore, the determina- tion of total iron is helpful in locating the source of acid in highly acid waters, as a 1. MISSOURI J. MISSOURI K. OHIO L TENNESSEE BELOW HEADWATERS PERCENT 17 STATIONS HEADWATERS 10 STATIONS 24 STATIONS 94 STATIONS OF CASES 121 DETERMINATIONS lie DETERMINATIONS 252 DETERMINATIONS 149 DETERMINATIONS M GULF COAST STREAMS N RIO GRANDE 0. COLORADO RIVER SVSTEU P COLUMBIA RIVER SYSTEM 13 STATIONS 0 STATIONS 10 STATIONS 37 STATIONS 118 DETERMINATIONS I2G DETERMINATIONS 38 DETERMINATIONS 62 DETERMINATIONS Figure 19. — Continuation of comparisons of free carbon dioxide values, stippled graph A figure 18, being the standard. combination of high iron and high acidity suggest the source of acidity as pyrite or coal beds if no specific chemical effluent is responsible for these conditions. AMMONIA Total ammonia determinations were made by the Nessler method as described by American Public Health Association (1933). Decomposing organic matter, if nitrogenous, will liberate into stream water, ammonium compounds representing a considerable portion of the total nitrogen. As ammonium compounds present at once both a hazard due to the high toxicity of ammonium carbonate for most aquatic animals, and an important source of nitrogen 392 BULLETIN OF BUREAU OF FISHERIES for the lower plants in the aquatic food chain, ammonia determinations give a signifi- cant index of the balance between stream purification through the consumption and elimination of ammonium compounds formed during the disintegration of organic detritus, and the amount of such organic wastes received by the stream. In natural unpolluted waters organic detritus consists primarily of the remains of the organisms, both plant and animal, dying in the stream and adjacent waters, together with such organic matter as may be brought into the stream by surface run-off water. In unpolluted water, therefore, the amount of ammonia and ammonium compounds (chiefly ammonium carbonate) is usually very small. The West Riding Rivers Board (1930) found the ammonia content of the River Wharfe, an unpolluted stream, to vary between 0.0 and 0.17 p. p. m.; Butcher, Pentelow, and Woodley (1927) state the ammonia content of the river Itchen, normally to be less than 0.1 p. p. m. with a maximum of 0.25 p. p. m.; Pearsall (1930) found the waters of various English lakes to carry less than 0.01 p. p. m.; Domogalla, Juday, and Peterson (1925) report the ammonia in Lake Mendota to vary between 0.0073 and 0.76 p. p. m.; and unpolluted portions of Wisconsin River above Rhine- lander, Wis. (Wisconsin State Board of Health, 1927) averaged 0.096 p. p. m. am- monia. In field studies by the Columbia, Mo., unit similar ammonia values, all i below 0.9 p. p. m., were obtained from analyses of unpolluted flowing stream waters. Polluted streams, even if carrying only a very small load of organic wastes, present quite a different picture and, as Winslow and Phelps (1906) have pointed out, from one- third to one-half of the total nitrogen of sewage will be in the form of free ammonia, largely as ammonium carbonate; and sewage will carry from 15 to 35 p. p. m. or more of total nitrogen. Wiebe (1931b) reported a maximum of 0.224 p. p. m. ammonia in the Mississippi River at Fairport, Iowa. Ellis (1935b) found from 0.36 to 1.16 p. p. m. ammonia in the Mississippi River at Davenport, Iowa, during low waters in the month of July (1934) in portions of the river that were not badly polluted and which were at the time supporting bass, catfish, and other warm-water fishes; and from 0.24 to 3.80 p. p. m. of ammonia in the badly polluted waters of the Mississippi River during September 1935 between St. Louis, Mo., and Cairo, 111. In the highly polluted Blackstone River, the Massachusetts State Board of Health (1913) reported 11.7 p. p. m. of ammonia. Similarly the writer has found 5.68 p. p. m. of ammonia in the Cache la Poudre River in Colorado at a point where there was heavy pollution with beet sugar factory wastes. In table 2 a group of ammonia data typical of the field findings by the Columbia, Mo., unit are presented. These data include no determinations from the immediate vicinity of sewers or other outlets pouring organic wastes into the streams, and no determinations from restricted local portions of streams in which unusual organic pollution was found. These data show the general range of dissolved ammonia which may be expected in our flowing streams during the warm season, and under the usual conditions of municipal and industrial pollutions. The composite series includes the ammonia values from all these stations collec- tively at which good fish faunae as previously defined were taken, and the Mississippi and coastal series include all stations in these systems regardless of fish faunae. STREAM POLLUTION 393 Table 2. — Dissolved ammonia in various stream waters during the warm season, June to October, inclusive, 1934 [For convenience of comparison, data are expressed as percents of total number of determinations for each stream unit] River system Number of stations Number of deter- minations Dissolved ammonia in parts per million 0.0-0. 9 1.0 1.9 2. 0-2. 9 3. 0-3. 9 4.0-4. 9 5. 0-5. 9 Greater than 6.0 72 178 21 58 21 (2) Entire Mississippi system. 182 478 24 49 22 3 2 (2) (2) Coastal streams 58 91 7 55 33 1 1 1 1 Derived from ammonia determinations at 72 stations throughout the entire Mississippi system and in coastal streams, at each of which stations good mixed fish faunae and associated organisms were thriving. 1 Less than 0.5 percent. It may be seen that excepting the restricted local areas where heavy pollution was found (data not presented in tlxis table) the dissolved ammonia in flowing streams in general lies below 3 p. p. m. Tliis is due to a variety of physical and chemical factors, chief among which is the constant turn-over of the moving stream waters. The good fish faunae showed a preference to waters containing less than 2 p. p. m. dissolved ammonia, as 79 percent of the cases of this group were found in water containing less than 1.9 p. p. m. dissolved ammonia. The remaining 21 percent were apparently thriving in waters containing between 2 and 3 p. p. m. dissolved ammonia, with one case (the maximal indicated by footnote 2 in the 3-3.9 column) in waters carrying 3.5 p. p. m. dissolved ammonia. Tliis maximal dissolved ammonia value among the stations where good fish faunae were taken was found at Reeds Landing, Minn., where the Mississippi River was flowing rapidly over a sandy bottom at a point just below extensive beds of aquatic vegetation in the foot of Lake Pepin. It must be noted that without exception those waters carrying 2 to 3 p. p. m. ammonia and at the same time supporting the good fish fauna were high in dissolved oxygen ; that is, 5.5 to 7 p. p. m., were of low turbidity, and were flowing over good bottoms. Toleration of dissolved ammonia above 2 p. p. m. under field conditions was always associated with otherwise good to exceptionally favorable conditions. The toxic effects of ammonia compounds have been the subject of many investi- gations; and aquatic animals have been shown to be particularly sensitive to ammo- nium carbonate, the form in which ammonia is most frequently found in inland waters, (Shelford, 1917; Belding, 1928; Steinmann, 1928; and McCay and Vars, 1931). Thirty p. p. m. of ammonia will Id 11 some tench, trout, and salmon rather rapidly (Weigelt, 1885), and 55 to 77 p. p. m. will kill sinners and carp in a few minutes to a few hours (Clark and Adams, 1913). A review of the literature on ammonia, how- ever, shows that some observers obtained toxic effects with much smaller quantities. Ellis and Chipman (1936) have repeated many of the earlier tests and extended the observations to daphnia and gammarids, as well as fish, finding that pH is a large factor in regulating the toxicity of ammonium compounds for aquatic animals, ammonium salts becoming more toxic in more alkaline media. Tliis fact explains the relatively high toxicity of ammonium carbonate to aquatic organisms as compared with other ammonium salts. From these data (presented in section on ammonia pollutants) curves were drawn showing that the toxicity of ammonium compounds increases 200 percent or more between pH 7.4 and pH 8.0. The lower limit of toxicity 394 BULLETIN OF BUREAU OF FISHERIES (death in 10 days or less, depending upon conditions of experiments) was found to be near 2.5 p. p. m. of ammonia. Some acclimatization to ammonia is possible, and it is well known that individuals of various species of fish may be found in water containing 3 to 10 p. p. m. of ammonia. However, the existing literature and the data from our experiments indicate that under average stream conditions with pH value 7.4 and pH 8.5, 2.5 p. p. m. of ammonia will be harmful to many individuals at least of the common aquatic species. Therefore, in view of the small amount of ammonia found in unpolluted natural flowing waters, 1.5 p. p. m. dissolved ammonia was considered the maximal amount of dissolved ammonia not suggestive of specific organic pollution. In flowing streams 2 to 3 p. p. m. were almost always associated with definite organic pollution and values above 3 p. p. m. in our field studies were always traceable to sewage or factory effluents. SUSPENSOIDS The amounts of finely divided suspensoids in various waters and light penetration into these waters were determined by a photoelectric apparatus described by Ellis (1934). The suspensoids — that is, particulate matter m suspension in inland fresh waters — consists normally of erosion silt, organic detritus (as discussed under ammonia), bacteria, and plankton. Each component of this mixture, with the exception of plankton, may be greatly augmented by man’s agencies, as quantities of powdered rock, cellulose pulp, sawdust, semisolid sewage, and other debris are added to natural waters. Parts of some streams, as the Yellowstone and Missouri — draining areas in which natural erosion has been proceeding rapidly — have been muddy with their loads of erosion silt since before the earliest records by man and have as a result limited fish faunae. In a large proportion of the inland streams, however, erosion silt, organic detritus, and bacteria were formerly in balance over a considerable portion of the year and conditions favorable to aquatic life maintained, although now and then floods and other unusual conditions killed many aquatic animals in these streams by inundations of silt. With the advent of civilized man and unrestricted deforestation, agriculture, and other uses of the earth’s surface, the erosion problem has become gigantic, and the effects of the loads of erosion silt carried by these once relatively clear streams overwhelming on aquatic life in many places. Erosion silt and other suspensoids (disregarding any specific toxic action of suspensoid wastes) affect fisheries directty by covering the bottom of the stream with a blanket of material which kills out the bottom fauna, greatly reduces the available food, and covers nests and spawning grounds ; and also by the mechanical and abrasive action of the silt itself which may clog and otherwise injure the gills and respiratory structures of various aquatic forms, including many fishes and mollusks (Ellis, 1936a). The mechanical action of silt and other suspensoids may not be severe or even harmful to the gills and other structures of the free-swimming fish which move about above the bottom and are not mired down by the settling deposits of the suspensoids, if the amounts of the suspensoids in the water are not too great and if the action of the suspensoids is uncomplicated by other pollutants. Normal fish and many other swimming animals secrete continuously quantities of mucus which wash away STREAM POLLUTION 395 suspensoid particles as they lodge on the gills and other exposed parts. Conse- quently, healthy and uninjured fishes can move through very muddy water or water carrying considerable quantities of pulps (Cole, 1935b), sawdust, and other suspensoids and receive little or no mechanical injury to the gills. However, as has been pointed out by Marsson (1911), small amounts of various acids, chemical wastes, and other substances which in themselves either injure the gills or alter the flow of mucus may greatly augment the mechanical action of the suspensoids, with serious results to the fish, so that through the combined action of these chemical agents and the suspensoids the abrasive action on the gills may be increased or the gills matted with deposits which under more favorable conditions would have been washed away by the mucus. Indirectly, but none the less effectively, erosion silt affects fisheries by screening out the light, by “laking down” organic wastes, and thus increasing the oxygen de- mand at the bottom of the stream, and by retaining many forms of industrial effluents as oils, chemical wastes, and pulps in beds on the floor of the stream, with disastrous results to the bottom fauna. Summarized data on turbidity of and light penetration into various stream waters are presented in table 3. Table 3. — Turbidity of stream waters as measured by light -penetration ( millionth intensity depth in meters) in various river systems during June to September, inclusive, 1932-35 [For convenience of comparison, data are expressed as percents of the total number of determinations for each stream unit] I Light penetration expressed as millionth intensity depth in meters River system Composite, streams supporting good mixed fish fauna 1 Mississippi, flowing streams, good fish fauna Mississippi, flowing streams, medium, poor, or no fish fauna Mississippi, impounded water, Hastings Pool (Minn.) Mississippi, impounded water, Lake Pepin (Minn.- Wis.)_ - Mississippi, impounded water, Lake Davenport and Moline Pool (Iowa-Ill.) Mississippi, impounded water, Lake Keokuk (Iowa-Ill.) Missouri, headwaters, clear streams (Mont.-Wyo.)_. Missouri, below headwaters (Mont., Wyo., Colo., Nebr., Kans., Mo.) Ohio (Pa., W. Va„ Ohio, Ky., Ind., 111.) Tennessee, flowing streams (Ky., Tenn., Miss., Al3.)__ Atlantic-coast streams (Maine to S. C., inclusive) Gulf-coast streams (Fla. to Tex.) Rio Grande, flowing streams (Tex., N. Mex., Ariz.)._ Colorado, flowing streams (Ariz.-Nev.) Columbia, flowing streams (Idaho, Mont., Wash., British Columbia) Num- ber of sta- tions Num- ber of deter- mina- tions Clear Cloudy Turbid Very turbid Muddy Very muddy 5.00- oo 4.90- 1.00 0.99- 0.50 0.49- 0-30 0-29-0-15 0.14-0 202 514 35 24 17 9 9 6 52 237 4 11 8 16 25 36 79 260 1 18 11 14 38 18 1 17 100 12 73 11 88 1 13 175 51 19 19 4 7 16 571 19 28 20 16 17 5 15 20 60 20 19 23 44 31 4 4 17 41 82 12 5 15 7 48 13 93 158 11 8 32 32 17 11 16 36 37 18 9 11 in 18 46 18 18 6 25 88 4 8 7 36 67 22 8 3 3 17 31 49 39 3 3 3 Grand totals: Stations, 585; light-penetrat.ion determinations, 2,344 1 Derived from light-penetration determinations made at 202 stations in the Mississippi, Missouri, Tennessee, Atlantic coast* Gulf coast, Rio Grande, Colorado, and Columbia systems, at each of which stations good mixed flish faunae and associated organ- isms were thriving. Each of these 8 stream units are given equal value in the expression of the composite. 99773° — 37 5 396 BULLETIN OF BUREAU OF FISHERIES From the fieldwork on unpolluted streams in areas where surface erosion was not materially influenced by man and where water conditions were otherwise favor- able, it has been shown by Ellis (1936a) that the millionth intensity depth (i. e., the level at which the light entering the surface of the stream would be reduced to one- millionth of its surface intensity) for clear unpolluted streams carrying little or no erosion material is 50 meters or more, and that in streams carrying a heavy load of erosion silt, like the Missouri River, the millionth intensity level may be reduced to less than 100 millimeters. Until erosion is brought under control little can be done in demanding a minimum amount of silt and consequently no standard has been suggested here, but data from over 6,000 determinations on inland streams show that the silt load of these streams should be reduced so that the millionth intensity level would not be less than 5 meters, if conditions even approximating those of times past when erosion was held in check by forest and grasslands are to be restored, in the average inland stream of the United States. The detrimental nature of erosion silt as regards fisheries is discussed more fully by Ellis (1936a). However, particulate matter introduced by man into streams can be regulated, and the detrimental action of various types of suspensoid pollutants is discussed under that heading. DEPTHS As various depths of water are selected by different species of aquatic animals, and even by different ages of the same species in many cases, there is naturally no single optimum depth of stream water for all types of aquatic life. Depth studies of streams and flowing waters, however, bring out differences between waters of natural lakes and those of flowing streams and river lakes. Current action in flowing streams mixes the water so thoroughly and continu- ously that in general the composition of the aquatic environment as presented by the water itself varies but little from a few inches below the surface to a few inches above the bottom of the stream. The temperature of the water, its turbidity, dis- solved solids content, pH, and even dissolved gases content are much the same throughout the bulk of the stream at any given station, if there be definite current action. Sloughs and other rather isolated lateral areas, particularly those contain- ing large masses of aquatic vegetation, are of course excepted, but even in such cut- off portions of streams the actual differences in water characteristics as compared with the main stream mass are often small. Of significance in pollution studies, these depth data show that pollutants may be expected to mix rather rapidly and completely with the waters in streams and even river lakes unless these river lakes be more than 50 feet deep, and even in those river lakes unless the seasonal conditions are such as to promote the stratification just described in the deeper portions of such lakes; and that the products of bottom pollution will also be distributed throughout the waters of the stream below such pollution. STREAM POLLUTION 397 Table 4. — Characteristics of water at different depths in river lakes produced by impounding flowing rivers I. LAKE WILSON 1 Depth in feet Water tem- perature, de- grees centi- grade pH Specific con- ductivity in rahox 10* at 25° C. Fixed carbon dioxide, cubic centimeters per liter Dissolved oxygen parts per million Light pene- tration, mil- lionth inten- sity depth in millimeters 8 1 30.2 7.4 177 13.4 7.5 2, 053 10 30.2 7.6 178 12.9 7.3 2, 053 20 - - 29.5 7.4 181 13.2 6. 1 1,297 30 - - 29.0 7.5 178 12.9 5.8 1,075 40 28.6 7.5 175 12.9 5.7 1, 075 50 ....... 28.2 7.5 173 12.9 5. 1 972 60 . _ _ . ....... 27.8 7.3 161 12.9 3.6 1,075 70 . 22.9 7.2 167 15.4 .2 9, 036 80 .... .... 21.9 7. 2 158 15.0 .2 9,036 90 .. 21.5 7. 1 159 15.0 .2 6,910 100 . . . .... 20.9 7.2 159 15.4 . 1 6,262 110 . 20.8 7.2 162 15.9 .2 5, 335 II. ELEPHANT BUTTE RESERVOIR 1 S* 3 6 33.0 33.0 23.8 8. 1 8. 1 8. 1 1.009 1,062 1,024 29.5 28.7 28.7 7. 1 7.2 5.9 6, 823 6,823 14,344 10 23.0 8.1 1,017 28. 1 5.5 13,414 20 22.8 7.6 1,022 28.4 6. 1 28, 874 30 . _ 21.5 7.6 1,081 29.3 5.9 30,254 40 19.8 7.6 1, 107 29. 1 6.2 63. 987 50 - 19. 6 7.6 1,068 29.2 6.5 63, 987 60 18.8 7.6 1,070 29.0 6.6 63, 987 70 _____ 18.0 7.6 996 29.6 6.8 61, 237 80 _ _ 17.8 7.6 1,067 29.8 5. 1 61,237 90 17.6 7.6 1,045 30.2 5. 1 63, 987 100 _ _ _ 17. 1 7. 6 1,059 30.7 5.3 63, 987 110 16. 9 7.6 1,052 30.4 5. 1 61,237 125 16.9 7.5 1, 123 30.7 4.7 63,987 III. LAKE KEOKUK 4 27.8 7.5 285 27.8 6.7 1, 170 27. 5 7.5 287 28.0 6.6 1, 170 27.0 7. 5 285 28.3 5.8 1, 103 27.0 7.5 290 28.3 5.4 1,065 1 Tennessee River near Florence, Ala., station 385, July 30, 1931, air temperat ure 31.3° C. A deep river lake showing midsummer stratification. ]Rio Grande River, near Hot Springs, N. Mex., station 829, June 15, 1935, air temperature 33.0° C. A deep lake showing partial stratification at time of these observations. s Surface, i. e., top 12 inches of water. ‘Mississippi River near Keokuk, Iowa, station 80, Sept. 22, 1932, air temperature 30.3° C. A medium to shallow river lake showing little or no stratification. TEMPERATURE As the temperature of the water has bearing on several factors associated with the general problems of pollution, such as dissolved oxygen carrying power of the water, the rate of bacterial decomposition of organic pollutants, and the metabolic demands of the aquatic organisms themselves, the temperature range of flowing streams and river lakes during the warm season is presented in table 5. These data include no headwaters in mountainous regions, and give, therefore, the water tem- perature range to be expected in the average inland stream of the United States during the summer months, which season as has been pointed out under the dis- cussion of dissolved oxygen presents certain specific pollution hazards. 398 BULLETIN OF BUREAU OF FISHERIES Table 5. — Water temperatures of inland streams, exclusive of headwaters in mountainous regions, during warm season, June to September, inclusive, 1930-35 River system Num- ber of stations Num- ber of cases 14.0°- 17.9° C. 18.0°- 21.9° C. Percent 22.0°- 25.9° C. of cases 26.0°- 29.9° C. 30.0°- 33.9° C. Above 34° C. Mini- mum degrees, Centi- grade Maxi- mum degrees, Centi- grade Composite1.. 726 4,545 13 16 35 31 5 m 15.0 36.6 Entire Mississippi system ... _ . 447 2, 887 12 12 46 27 3 0) 15.0 35.2 Coastal streams into Gulf of Mexico 3._ 39 1,367 7 14 32 36 9 2 16.8 36.6 1 Includes flowing streams at stations at which fish were found in Mississippi, Missouri, Tennessee, Ohio, coastal streams flowing into Atlantic Ocean and Gulf of Mexico, Rio Grande, Colorado, and Columbia system. 1 Less than 1 percent. 3 Exclusive of Mississippi system. Considering the composite in table 5, it may be seen that in spite of the high air temperatures which prevail in many parts of the United States during the summer months, often for periods of days, stream temperatures above 34° C. are not common, and that 95 percent of the 4,545 cases listed in table 5 lie below 30° C. On the other hand, 66 percent of these 4,545 cases lie between 22° and 30° C. For comparison with the composite in this table, data from the entire Mississippi system and from the coastal streams of the Gulf States, Texas to Florida, inclusive, are also presented. In the Mississippi system 73 percent of the cases fell between 22° and 30° C-., and 68 percent of the cases in the Gulf States have the same range. In evaluating biochemical oxygen demand, dissolved oxygen content, and specific toxicity values for pollution studies therefor, a water temperature range during the warm season from 15° to 36° C. may be expected in the average inland stream of the United States, with the probability of 66 percent or more that these water temperatures will lie between 22° and 30° C. During the other portions of the year — that is, outside of the warm season — the water temperatures of these streams in localities north of the freezing line will fall, with 0° C. as the limit. Our field data also show that south of the usual freezing line — that is, in streams of southern Texas, Alabama, and Florida — temperatures as low as 10° C. are not unusual during the midwinter season. The true headwater streams, particularly those in mountainous regions, present a different temperature range even during the summer months. The maximal tem- peratures for various streams which could be definitely classified as mountainous headwaters, as found in our field studies, were rarely above 10° to 15° C., with the average values lying around 10° C. or even lower. BOTTOM CONDITIONS AS AFFECTED BY STREAM POLLUTION In most stream-pollution studies, conditions at the bottom must be considered, since the bottom of the stream contributes a considerable part of the food of stream fishes through the organisms of the bottom fauna which are either eaten directly by fishes or which are included in the food chains of various species of fishes. Besides, portions of the bottom usually in the shallower parts of the stream provide nesting sites for certain species. However, as the stream bed may change from rock and coarse gravel to fine clay in the course of a few hundred yards, and as there is a wide variation in habitat preference of the many desirable organisms which inhabit stream STREAM POLLUTION 399 bottoms, it is very difficult to define all of the conditions at the bottom of a stream which may or may not be suitable for bottom inhabitants and therefore affect the fish fauna; for over this already complicated association of rock, soil, debris, and living organisms pollution may spread a blanket of silt, sludges, pulps, and poisonous fluids. Studies of the bottom faunae of polluted streams, particularly those carrying quantities of organic pollution, have shown that many of the bottom species of unpolluted streams are very sensitive to pollution conditions, so that as pollution progresses the normal bottom fauna changes giving way to certain more tolerant species of tubificid worms, chironomid midges, sphaeriid mollusks, and leeches. Con- sequently, various investigators (Marsson, 1911; Suter and Moore, 1922; Turner, 1927; Richardson, 1928; Wiebe, 1928), in connection with other data have correlated the presence of certain bottom forms with different degrees of stream pollution, and a few such species of bottom organisms have come to be used to some extent as indices of pollution. Richardson (1928) however, in discussing 49 bottom species definitely associated with pollutional or subpollutional conditions in the Illinois River points out the difficulties attending the use as indices of pollution, of even such species as the tubificid worm, Tubijex tubijex, and the midge, Chironomus plumosus, which are com- monly regarded as characteristic of pollutional conditions bordering on septic. The problem of index species is greatly complicated by the fact that the number of individ- uals of any particular index species may vary from zero over a wide range in adjacent parts of any polluted stream as the result of conditions other than those produced by pollution. In the present studies, bottom samples were taken throughout the field work as a part of the regidar routine. Reviewing the results of the examinations of many hundreds of bottom samples, there is no doubt that bottom samples are a necessary part of any pollution study, and that the findings concerning the bottom organisms are very valuable when considered with the chemical and physical data. However, the writer agrees with the statement of Richardson (1928, p. 410), in connection with his Illinois River studies, that index species are of service in pollution studies only when used with the greatest caution, and when checking with other indicators. With due regard to the limitation just discussed, three organisms — namely, the water mold, Sphaerotilus natans, the rattail maggot, Eristalis sp., and the sludge fly, Psychoda sp. — were often found helpful in delimiting the septic zone near the source of extensive organic pollution; and the tubificid worms, Tubijex sp. and Limnodrilus sp., and the larvae of the red midge, Chironomus sp. (particularly Chironomus plumosus ), if their presence or absence were supported by the chemical and physical analyses and by the data concerning the biological complex as determined from repeated samplings, were frequently used in defining the extent of bad to moderate organic pollution. In the lesser degrees of organic pollution and in most cases of chemical pollution index species as such were of little value. On the other hand, the abundance of the various species comprising the bottom fauna and particularly the species composition of the bottom fauna at various stations in the polluted area wrere usually of large value, regardless of the sort of pollution, when correlated with the physical and chemical data. Gross pollution of several kinds was quickly detected from bottom samples both by the physical and chemical conditions 400 BULLETIN OF BUREAU OF FISHERIES of the mud itself and by the paucity or absence of bottom fauna, often represented only by dead shells and other remains. Dredgings readily demonstrated such gross pollution by silt, sewage, pulps, tars and heavy oils, mine tailings, rock powders from quarry operations, and such organic wastes as starch effluents, sizings from cloth mills, and distillery slops. Downstream from a region of gross pollution the biological analyses of the bottom faunae were often particularly significant when compared with the bottom faunae in similar situations in comparable unpolluted waters (if possible of the same stream) in showing the selective effects of some pollutants, and the downstream extent of pollution. Many chemical wastes have high specific gravity and consequently tend to follow the stream bottom for some distance before complete mixing is accomplished by stream current action. Often downstream dredgings showed complete absence of bottom life of any type for some distance below the outlets of flumes carrying such chemical wastes, and the limits of the zones in which the bottom faunae had been killed out yielded valuable information concerning the points at which the effluent had been diluted to a non toxic level. ACTION OF POLLUTANTS ON FISHES Regardless of the effects on the environment, many pollutants affect fishes directly by some specific action on the living organism itself. These pollutants can be grouped, therefore, according to the locus of injury and method of action of the various active substances which the effluents contain (Ellis, 1936b). INJURIES TO GILLS AND EXTERNAL STRUCTURES The higher concentrations of almost all effluents and all lethal concentrations of some types of pollutants kill by their actions on the gills of the fish before little, if any, of the material causing the death of the fish passes beyond the gills. Such substances kill, therefore, by a combination of chemical and physical injuries rather than by true toxic action. The salts of several heavy metals, some acids, and some special chemicals as trinitrophenol combine readily with the mucus secreted by the fish’s skin, mouth, and gills, forming insoluble compounds. If the dilution of the pollutant of this type be great enough, or if the supply of the pollutant be limited so that it acts on the fish only a short time, the secretion of additional mucus may wash away the precipitated compound as fast as it is formed, oi1 may carry away the larger masses of the precipitate before serious damage to the fish results. If, however, the concentration and exposure time are large enough (very small quantities of several heavy metal salts suffice, v. i.), the precipitated insoluble compound covers the body, the lining of the mouth, and the gills. (See fig. 21.) Disastrous results follow the covering of the gills with these precipitates, and death supervenes very rapidly from the respiratory failure. In our experiments with salts of heavy metals and various other compounds producing death by precipi- tation of insoluble matter on the gills, the precipitate was found to cause death as the result of a combination of three conditions which individually or collectively disturbed the proper functioning of the respiratory and circulatory mechanisms. *** ' mm l"'1' , O STREAM POLLUTION 401 In the weaker concentrations of such effluents and harmful substances, the precipitate coated the gill filaments and filled the filament interspaces so that the water pumped through the mouth and onto the gills for the aeration of the blood could not reach the cells of the gill filaments. Consequently, the aeration of the blood with its accom- panying gas exchanges was prevented, and sooner or later death followed from a combination of anoxemia and carbon-dioxide retention. If still larger masses of flocculent precipitate were formed, as was the case when the stronger effluents were tested, the interlamellar spaces were clogged and move- ment of gill filaments became impossible. This condition affected the circulation of the blood in the gills, in that stasis of the blood in the gill capillaries usually follows cessation of movement of gill filaments in a short time. Kolff (1908) in her studies of fish hearts found that the pulsation of the heart is only one of the forces necessary to drive the blood through the gill capillaries, and that the movements of the gills are important in maintaining the gill capillary circulation. If the scales be removed from the midventral portion of the precardial region of a pithed goldfish and a tiny steel needle (the minuten nadeln on which entomologists mount very small insects) be forced through the body wall and into the pericardial cavity, when proper adjustment is made, this needle, if examined under a binocular microscope, can be seen to swing with each heart beat. Using this procedure the rate of the uninjured and unexposed heart in situ can be recorded for hours. Such fish, when the gills were perfused via the mouth with water containing salts of heavy metals or other substances producing these heavy mucoid precipitates, continued their normal respiratory movements and heart action until the precipitate clogged the gills and mechanically prevented the movement of the gill filaments. At that time an abrupt change in the activities of the heart was usually noted, the heart action dropping to about one-half its former rate. Heart block also often developed at this time. When these changes in heart rate and heart action were noted, the gill capillaries were always found to be gorged with blood on the efferent or cardiac side and showed practically complete to complete stasis of the blood in the capillaries. This same condition of blood stasis in the gill capillaries was noted in other fish in which anoxemia was produced by exposure to water containing little or no dissolved oxygen shortly after cessation of respiratory movements, so that the importance of the respiratory movements for the maintenance of capillary circulation in the gills was well established. After its rate had dropped to approximately one-half normal, the heart continued to beat for hours (in some cases for over 24 hours), although the circulation through the gills was completely blocked. In fishes which had ceased to make movements of any sort, either respiratory or general, after prolonged exposure to water low in dissolved oxygen, the hearts were found to be beating even when these fishes were opened under oil to exclude even momentary aeration of the peri- cardial fluid. The various substances producing these heavy precipitates, as well as many other pollutants (acetic acid, volatile extracts of crude oil, and others) which did not form precipitates with mucous secretions on the gills, also produce death by asphyxia through direct damage to the gill filament cells, which in turn cause stasis of the blood in the gill capillaries. Death from anoxemia and circulatory failure in the gills, therefore, was regularly found in fishes dying either in waters containing 402 BULLETIN OF BUREAU OF FISHERIES low oxygen or in the stronger hypertonic solutions of various salts and other chemicals, regardless of the formation of any sort of mucoid precipitate. In connection with gill damage, it must be pointed out that the weaker con- centrations of various pollutants which do not damage the gills rapidly enough to cause speedy respiratory and circulatory failures must also be considered as pollu- tion hazards if these solutions be even slowly toxic to the cells of the gill, for Smith (1929) has demonstrated that the gills of fresh-water fishes (carp and gold- fish) are important excretory structures, removing from 6 to 10 times as much nitrog- enous wastes as the kidneys. Keys and Willmer (1932) have found special cells associated with chloride secretion in the gills of various fishes, including fresh-water species. Damage to the gills from pollutant substances in the water can result, therefore, in the impairment of other functions, particularly those of salt balance and excretion in addition to respiration and circulation as already described, without actual toxic action on the internal vital organs of the fish. POLLUTANTS ENTERING THE BODY OF THE FISH AND EXERTING TRUE TOXIC ACTION The gills, the lining of the mouth, and the skin are the main portals of entry through which toxic substances can be absorbed into the body of the fish. However, Bond (1933) concludes that the bony fishes and many fresh- water invertebrates have little or no permeability to a variety of substances. The writer has been able to confirm this relative nonpermeability of the external structures of various fresh-water fishes, particularly the gills, for a number of substances. The active agents in vari- ous stream pollutants, therefore, may be still further subdivided into those which can enter the body of the fish through external structures as various volatile com- pounds from crude oil, dilute chlorine, ether, chloroform, methyl mercaptan, and formaldehyde, and those which do not enter or at least enter very slowly through the external structures as certain arsenic compounds. The main internal channel for the absorption of injurious and toxic substances is the gastrointestinal tract. The writer, using colloidal dye methods, has found that many fresh-water fishes refuse to swallow water containing various types of effluents for sometime after being placed in such water, for as has been pointed out by Smith (1930) most of the water required for the formation of urine in the fresh- water fish is probably taken in through the lining of the mouth. However, after varying intervals, usually 48 hours or more, fresh-water fish swallow some of the sur- rounding water even if not fed, suggesting that the water needs of the body are not entirely supplied through absorption by the external surface or the lining of the mouth. The swallowing of polluted water allows any of the injurious or toxic substances which this water may contain direct access to the mucosa of the gastrointestinal tract. Harmful substances may, therefore, either damage the lining of the intestinal tract or may be absorbed from the gastrointestinal tract, and in this way gain access to the internal vital organs. Consequently, another group of pollutants can be designated as those acting on the fish after ingestion into the gastrointestinal tract. It is through this portal that many cumulative poisons enter the body of the fish during long-time exposures to concentrations of pollutants which may at the time seem harmless. STREAM POLLUTION 403 Arsenic may be cited as an example of a substance which can be accumulated by the fish via the gastrointestinal tract after long exposure to low concentrations. To summarize, on the basis of locus of injury to the fish, stream pollutants may be classified as — (1) Those substances injurious to the gills and other external structures of the fish without any marked absorption beyond the gills. Death of the fish exposed to these pollutants results from anoxemia, through respiratory and circulatory failure, and through interference with the excretory functions of the gills. (2) Those substances killing the fish by specific toxic action after absorption through the gills, the lining of the mouth, and other external structures. (3) Those substances killing the fish by specific toxic action after absorption from the gastrointestinal tract, which region they have reached in water swallowed by the fish. LETHALITY OF SPECIFIC SUBSTANCES OCCURRING IN STREAM POLLUTANTS GENERAL CONSIDERATIONS In this section, data from the literature and from experimental assays made dur- ing the present studies are presented covering the lethality of various substances which are the active harmful agents in different types of stream pollutants. These biochemical and biophysical findings in general may be applied to pollution condi- tions ranging from acute to mild; although in the evaluation of all individual cases of pollution, caution must be exercised, as biological data of any sort are rarely suit- able for rigid application. One particular component of a given waste may consti- tute the major pollution hazard of that waste; in fact, the dangerous qualities of many wastes center largely around one substance in each effluent rather than the effluent mixture as a whole. In such cases, lethality tables are particularly helpful, but the synergistic and antagonistic actions of other substances in the stream water as well as in the waste itself must be borne in mind always. For example, the lethality of copper sulphate for fishes under certain conditions is decreased by the presence of calcium chloride and increased by sulphuric acid, and both of these compounds occur free or combined in certain industrial wastes. Again the toxic actions of small amounts of chlorine and of ammonia on aquatic life singly are different from the ac- tions of the same quantities of these substances when both are present, since together they may form either ammonium chloride or chloramine, depending upon conditions; and the toxicity of the resultant mixture may vary accordingly. From the data presented in the following tables, however, the harmless, injurious, and lethal ranges of various substances may be computed readily with reasonable accuracy for particular pollution cases if due regard be given to the chemical and physical conditions of the water of the stream involved. Tests should be made, of course, to detect the presence of unusual and unexpected substances in the local com- plex which might alter the action of the pollutant. In addition to the appraisal of the pollutant on the basis of the lethal action of its major component or components, when the pollution is acute, mild, or intermittent, 404 BULLETIN OF BUREAU OF FISHERIES under which headings a large percent of the pollution cases may be classified, the possibility of cumulative effects must be considered when determining remedial measures or dilutions. Cumulative effects from very low concentrations of some substances which seem harmless to fish life over a period of 10 to 14 days present difficult problems. Some of these are discussed in subsequent portions of this section. TEST ANIMALS In the course of the work on the lethality of stream pollutants 30 species of fresh- water animals (see table 6), including 17 species of fish, 6 species of crustaceans, and 7 species of fresh-water mussels, have been used in one connection or another as test animals. Table 6. — Species of test animals Species Common name Species Common name FISH 1. Perea flavescem 2. Micropterus salmoides ... _ 3. Lepomis incisor ... ... 4. Lepomis humilis. . .. 5. Ameiurus n ebviosas Yellow perch. Largemouth black bass. Blue gill. Orange-spotted sunfish. Common bullhead. Channel cat. AMPHIPOD8 18. Gammarus fasciatus 19. Eucrangonyx gracilis ENTOMOSTRACANS 20. Daphnia magna 21. Bosmina sp.- Gammarid. Do. Daphnia. Long-nosed dace. Dace minnow. 22. Cypris sp _ _ _ 8. Leuciscus balteatus 23. Cyclops sp._ Copepod. 9. Notemigonus chrysoleucas 10. Notropis delicious ... . 11. Hybognathus nuchalis Golden shiner. Straw-colored minnow. Silverv minnow. FRESH-WATER MUSSELS 24. Tritogonia verrucosa Buckhorn. 12. Pimephates promelas . . Fathead minnow. 25. Megalonaias gigantae Washboard. 13. Cyprinus carpio German carp. 26. Quadrula trapezoides . Small washboard. Common goldfish. Rainbow trout. 27. Fusconaia undata _ _ Pigtoe. River mucket. 28. Actinonaias carinata ... Cut-throat trout. 29. Lampsilis anodontoides . Yellow sand-shell 17. Lepisosteus platostomus _ . . . Short-nosed gar. 30. Lampsilis siliquoidea Fat mucket. From these studies and from the voluminous literature of bioassay methods and physiological and pharmacological experiments in which these and many other forms of aquatic life have been used as bioreagents, two animals — the common goldfish, Carassius auratus. and the entomostracan, Daphnia magna — were selected for the standard tests. This procedure, i. e., the use of standard test animals, is recognized in other fields of biological investigation in which the frog, the rat, and the guinea pig have established positions as assay animals, and makes possible direct comparisons of data on the actions of a variety of substances over a wide range of conditions. No single test animal combines all of the desirable qualities and in making a choice both availability and physiological suitability for the prob- lem in hand must be considered. As the variety of aquatic animals which may be affected by stream pollution is large, a lethality range for each pollutant and effluent must be established, as the specific sensitivities of various aquatic species are not the same, although they may live in the same aquatic complex. However, if the pollutant destroys any one of the major species in the complex, the entire fauna in that part of the stream may suffer as a result. The selection of goldfish and daphnia gave a combination by which the maximal and minimal limits of toxicity of any given pollutant for different stream complexes could be ascertained rapidly and with reasonable accuracy. STREAM POLLUTION 405 GOLDFISH Goldfish of any size up to 6 inches are easily obtainable in quantity from com- mercial goldfish farms, are relatively inexpensive, and stand shipment well. The practical considerations, although important in the selection of a test animal, are of course secondary to the physiological characteristics of the goldfish which make it particularly suitable for the bioassays of stream pollutants. The goldfish is tolerant of confinement and can be kept in laboratory containers and aquaria for months, making both short-time and long-time tests, uncomplicated by excitement and the attendant endocrine disturbances, possible under controlled conditions. The amount of oxygen consumed by the goldfish is not large per unit of weight, and tills fish does not succumb to low oxygen as quickly as many other fresh-water fishes. In fact, the uninjured goldfish if permitted can obtain sufficient oxygen to maintain itself for some time in water very low in or even devoid of dissolved oxygen merely by skimming and breaking the surface film of the water. Because of these facts, the toxic action of many substances which would be volatilized or oxidized by the constant aeration of water required to maintain various other species of fish can be ascertained, using goldfish as reagents; and the specific parts played by the oxygen demand, the external action, and the internal effects of the pollutant on the fish determined separately. On the other hand, if the combination of anoxemia and toxic action of the pollutant are desired, goldfish can be placed in closed containers, or merely prevented from reaching the surface layer of the water by submerged strips of glass which do not interfere with either the movement of the water or with its surface reaeration. By comparative tests the relative resistance of the goldfish to poisons, injury, and anoxemia was found to be equal to or greater than that of bass, perch, or catfish, and definitely greater than the resistance of trout or most fresh-water minnows. The minimal lethal concentration of any effluent or pollutant for goldfish established, therefore, approximately the maximal concentration of that pollutant which ought not to be exceeded if any fresh-water fish were to survive in waters polluted with such effluent. An added advantage in the use of the goldfish as a test animal is the already large literature on this fish in both physiological and pharmacological fields, in which the usefulness of this animal as a bioassay reagent is well established (Munch, 1931; Powers, 1917). The stock goldfish used for these tests were obtained in lots as needed from a commercial goldfish farm, and were held in flowing water in large hatchery tanks at a temperature ranging from 17° to 21° C. All goldfish were of the standard carp- form variety, i. e., the various ornamental varieties were excluded. The fish were fed regularly every third day a ration of prepared shrimp meal, and Myriophyllum growing in the tank was available at all times. For routine tests 3-inch goldfish (60 to 90 millimeters in length), weighing between 3 and 5 grams, were selected as standard. Larger sizes were used for special tests as required. 406 BULLETIN OF BUREAU OF FISHERIES DAPHNIA MAGNA The second test animal, Daphnia magna, which is found in the quiet backwaters of various river systems (Surber, 1936), and which is raised extensively as food for young fish, was chosen because of its high sensitivity to most stream pollutants. Comparative tests showed this animal to be more sensitive even than trout, so that the minimal lethal concentration for daphnia of a given pollutant gave approximately the maximal dilution from which damaging effects on fresh-water fishes and many other fresh-water animals could be expected (exclusive of certain cumulative effects). Consequently, between the minimal lethal concentration for daphnia on the one hand and the minimal lethal concentration for goldfish on the other, the gamut of lethal concentrations of any pollutant or effluent under conditions comparable to those of particular pollution cases could be defined rather definitely. Daphnia proved to be an excellent test animal with which to obtain a quick orientation concerning the relative toxicity of effluents of unknown lethality, since this animal is quite free from the protective mucous secretion which safeguards fish for a tune against many pollutants. Daphnia were also valuable in studying the cumulative effects of some pollutants, as under favorable controlled conditions a new parthenogenetic generation could be expected every 5 to 8 days. Like that of the goldfish, the scientific literature on daphnia is extensive, and the general physiology and reactions of this animal are well known . The use of daphnia as a bioreagent is also well established (Munch, 1931; Adams, 1927; Billiard, 1925). The daphnia used in these tests were raised from original stock secured in Texas, pedigreed strains being established and lines of parthenogenetic clones secured for the assay work. The daphnia colony is maintained in a series of large glass jars, each containing four liters of water; and the animals fed bacterial nutrient material pre- pared from cottonseed meal as described by Chipman (1934). All assay tests were made in a constant temperature cabinet (see fig. 22) at 25° C. One hundred fifty daphnia of the same age (usually 4 days old) from the same strain were placed in each jar containing 3 liters of the solution to be tested, i. e., in a known dilution of the pollutant, using water from the same source as that in which the daphnia were living as the diluent. No abrupt changes in temperature were per- mitted, and the solution was well aerated before the daphnia were placed in it. Every 48 or 72 hours, depending upon the type of test, the animals in each jar were concentrated in a pyrex glass trap devised for this work, counted, and returned to the same or fresh solution as desired. When young appeared they were removed and carried in separate jars containing the same concentration of the pollutant as that of the original jar. In this way the mortality of each succeeding generation as well as the rate of reproduction could be followed. WATER TYPES When determining the specific lethality of any given pollutant for a particular case, water from the stream into which the effluent was being poured, taken just up- stream from the point at which the pollutant entered the stream and filtered through bolting cloth to remove large masses of suspensoids and the macroplankton was used as the diluent water. In this way the pollutant was mixed with the same water as U. S. Bureau of Fisheries, 1 937 Bulletin No. 22 Figure 21. — Top, normal goldfish; bottom, goldfish killed in copper sulphate. Note thick deposit of opaque copper-mucoid precipitate covering body, fins, and eyes. This type of precipitate is produced by many wastes containing compounds of heavy metals. Figure 22. — Constant-temperature apparatus in which Daphnia magna were used as test reagents for toxicity of various stream pollutants. Assistant at the right is operating a concentrator preparatory to counting Daphnia. STREAM POLLUTION 407 in the actual case of pollution under consideration and the same chemical interre- actions were possible. When it was not feasible to use stream water at the site of the pollution, unpolluted water from a comparable stream was diluted or fortified until the salt content, the conductivity, the carbonates, and the pH were essentially the same as those of the stream water in question. This was made possible by the field and laboratory analyses of the waters of the stream. This method of compounding a synthetic stream water, although laborious at times, gave very satisfactory com- parative results. Control tests were also made in each investigation using glass- distilled water as the diluent for certain critical concentrations. In making the assays of pollutants, regardless of tbe diluent water, no material was removed after making the dilution; i. e., even if precipitates or colloidal suspen- sions were formed on the addition of the pollutant to the diluent water, the fishes or daphnia were exposed to the action of the entire mixture. Many substances which are precipitated in stream and river water are nevertheless slightly soluble in water, particularly if small quantities of carbon dioxide and other compounds be present, so that portions of the precipitate might continue to redissolve for hours or days, and insoluble substances even though precipitated to the bottom of the stream constitute supplies of various ions which may be drawn into the water as others are removed. Again, some of these precipitated materials are dissolved or chemically changed by the action of various substances in or on the body of the fish or daphnia, or by the excretory products of these animals, so that particles of precipitated material lodging temporarily on the gills might be dissolved by the carbon dioxide there, or other particles swallowed by the fish might be acted upon by the digestive fluids of the animal. In Hew of these complicating factors, the tests made in stream waters are much more valuable in obtaining the actual hazards to aquatic life than those made in distilled water. In the basic data tests (see tables 7, 8, 9, 10, 11, 12, and 13) four types of water were used as diluents, namely, glass-distilled water for the chemical controls in which the specific and uncomplicated action of the substance was studied, and soft, medium, and hard water (following the classification of Birge and Juday, 1911), i. e., with fixed carbonates less than 5 cubic centimeters per liter, 5 to 22 cubic centimeters per liter and over 22 cubic centimeters per liter, respectively, for the general lethality tests. Ranges of lethality were obtained in this way which are applicable within the limitations previously discussed to most stream waters of the United States, since the hardness of these natural stream waters is dependent almost entirely on the cal- cium and magnesium carbonates which exists in different degrees ol concentrations in the interior waters. The soft and hard natural waters were obtained at Columbia, Mo.; the hard water representing a typical limestone drainage, and the soft water a typical clay drainage. The medium water was obtained from the upper Mississippi River. In the tables the pH and specific conductance are given both for the original water and for the mixture as modified by the pollutant, so that the fluctuations in tbe diluent water itself and in the mixture as the result of the action of the pollutant on the salt and buffer balances may be followed in each case. 408 BULLETIN OF BUREAU OF FISHERIES SPECIFIC LETHALITY TABLES OSMOTIC PRESSURE AND SODIUM CHLORIDE Many effluents carry soluble substances exerting considerable osmotic pressure, and because of that property are capable when sufficiently concentrated of withdraw- ing water from the gills of fishes and from other delicate external organs of various aquatic organisms, with the attendant damage to the living cells. High concentra- tions of many sorts of pollutants present this danger quite independently of any toxic, chemical, or corrosive action they may have on aquatic life. The toxic and chemical actions of various salts in stream pollutants may be in part offset by the presence of other substances; but the osmotic action of all of the components of the mixture may be lethal regardless of their mutual antagonisms or specific toxicities, consequently the osmotic pressures of concentrated effluents must be regarded as potential hazards until these substances are sufficiently diluted to be within the limits of osmotic pressure tolerated by fresh-water fishes. Garrey (1916) finds, using the straw-colored minnow, Notropis blennius, as a reagent, that fresh-water fishes tolerate an osmotic pressure of the external medium equal to that of their own blood if the various salts and substances in the water are balanced against each other so as to exclude the specific toxic effects. As the blood of fresh-water fishes contains approximately 0. 7-percent sodium chloride, together with small quantities of calcium and potassium salts, the limit of osmotic pressure tolerated in the external medium by fresh-water fishes is therefore near 6 atmospheres, or the approximate equivalent of 7,000 p. p. m. of sodium chloride (common salt) in the surrounding water. In table 7 data are presented showing the effect of common salt in medium water (Mississippi River water) on test goldfish. This water, of course, contained small quantities of calcium and magnesium salts, which served to some extent to offset the specific toxic action which sodium chloride has in distilled water. (See sodium chloride, p. 429.) These river-water tests confirm the work of Garrey (1916), as sodium chloride killed quickly in concentrations greater than 10,000 p. p. m., and was not lethal in concentrations of 5,000 p. p. m. or less. The survivals in a concen- tration of 10,000 p. p. m. sodium chloride in river water ranged from 4 to 7 days, and the predicted maximal nonlethal concentration of sodium chloride on the basis of osmotic pressure alone in terms of the osmotic pressure of fish blood would be approxi- mately 7,000 p. p. m. Table 7. — Survival of goldfish in solutions of sodium chloride ; diluent, filtered Mississippi River water 1 Concentration ration by weight Parts per million pH Specific conductivity mhoX10-« at 25° C. Survival time * W ater Solution Water Solution 1:20 50,000 7.8 7.7 236 62, 669 30 to 40 minutes. 1:50 20, 000 7.8 7.7 235 31, 336 1 to 2 hours. 1:67 14, 925 7.8 7.7 235 25, 254 10 to 12 hours. 1:80 12, 500 7.8 7.7 235 25, 200 24 to 36 hours. 1:100. 10, 000 7.8 7.7 235 18, 772 4 to 10 days. 1:200 5, 000 7.8 7.7 235 4,156 CO 1:1,000 1.000 7.8 7.7 235 2, 328 00 1:10,000 100 7.8 7.7 235 455 CO 1:100,000 10 7.8 7.7 235 245 CO 1 Condition of experiments as described in table 8. 2 Minimal and maximal survival times as found in these experiments. Infinity sign indicates survival greater than 25 days without any apparent injury to fish. STREAM POLLUTION 409 In practical pollution work, therefore, any effluent as long as its osmotic pressure is greater than 6 atmospheres may be expected to be lethal to fresh-water fishes re- gardless of any specific toxic properties. Below this osmotic pressure the toxic, chemical, and corrosive characteristics of wastes will be the major determining fac- tors in its lethality. Brine wastes, however, often present pollution hazards largely because of their osmotic pressures. ACIDS High acidity is a characteristic of a large number of wastes, so that the action of uncombined acid is one of the major problems of stream pollution. In determining the lethality of acid wastes, both the actual acidity- — that is, the pH or hydrogen-ion concentration of the polluted water carrying acid effluent — and the specific acid in- volved must be considered, as acid wastes do not kill merely because of a particular degree of acidity. In table 8 a summary of an extensive series of experiments on the lethali t,y of 1 1 acids found in stream pollutants is presented. It is obvious, first of all, from this table that it is futile to attempt to designate a given p. p. m. value as marking the lethal point for any particular acid as the buffer substances and dissolved salts of the water into which acid waste is poured will determine the amount of free acid or hvdrogen- ions available. The actual conditions, therefore, presented to aquatic life following the addition of a given number of p. p. m. of a particular acid to soft water will be quite different from those presented by the same amount of that acid when added to hard water, with of course all sorts of intermediate possibilities between these two extremes. For example, 134 p. p. m. of sulphuric acid killed goldfish in from 6 to 90 hours when added to soft water, but were not lethal when added to hard water. Table 8.- — Survivals of 700 goldfish in various concentrations of 11 acids found in industrial wastes1 Substances Concentra- tion ratio by weight Parts per mil- Diluent water pH Specific con- ductivity mhoXlO-8 at 25° C. Survival time 2 Constituent of — lion W ater Solu- tion Water Solu- tion 1:70 14, 286 4,975 1, 585 1,000 Hard 8.0 3.0 641 936 26 to 33 minutes Do 1:201 8.0 3. 5 641 749 30 to 36 minutes. Do_ 1:631 8.0 4.0 641 673 50 to 60 minutes Do 1:1, 000 do 7.8 4.5 645 693 50 minutes to 1 hour 30 minutes. Do 1:1,005 995 do 8.0 4.5 641 652 2 hours to 5 hours 50 min- utes. Beet-sugar Do . 1:1,077 929 do_- . __ 7.9 4.6 641 662 wastes. Do 1:2, 365 423 do 8.0 5.0 641 676 4 hours 45 minutes to 20 hours. Do 1:2, 871 s 1:10, 000 3 1:100, 000 348 __ _ do.. __ _ 8.0 5. 5 641 675 Do 100 __ _do_- . __ 7.8 6.8 645 675 Do.. 10 do 7.8 7.3 645 688 Do 3 1:1,000, 000 1:1,000 1 ___ do 7.8 7. 6 645 680 co Benzoic acid 1,000 do.... 7.8 4.8 666 593 j-Ooal-tar wastes. Do Chromic acid 1:5, 000 1:1,000 200 1,000 do Very soft Hard __ _ 7.8 6. 2 5.9 1. 4 666 <50 647 647 598 3, 680 844 7 hours 16 minutes to co... Do 1:1, 250 800 7.8 5. 4 Do 1:2, 500 400 do 7.8 5.9 724 15 hours 18 minutes to 28 Do 1:5, 000 200 do 7.8 6.4 647 716 hours 24 minutes. 60 hours 24 minutes to 84 hours. [Tannery wastes. Do 1:10,000 1:10,000 1:400 100 do 7. 8 7. 3 647 689 CO Do 100 Very soft Hard 6. 2 4. 0 <50 6« 641 469 Citric acid 2, 500 1,433 8. 0 3.0 841 Do . 1:698 do 8.0 3.5 654 3 hours to 3 hours 30 min- utes. Citrus-fruit prod- Do 1:1, 119 894 __ __do. 8.0 4.0 641 542 ucts wastes. Do 1:1,600 625 do 8.0 4.5 641 494 410 BULLETIN OF BUREAU OF FISHERIES Table 8. — Survivals of 700 goldfish in various concentrations of 11 acids found in industrial wastes *• — Continued Specific con- nTT ductivity Substances Concentra- tion ratio by weight Parts per mil- Diluent water mhoXlO-® at 25° C. Survival time 2 Constituent of— lion Water Solu- tion Water Solu- tion Hydrochloric acid.. Do 3 1:1,000 1,000 196 7.8 1.8 3.0 680 3,481 1,221 1:5, 100 .do.. .. 8.0 641 1 hour 15 minutes to 1 hour 30 minutes. Do 1:5,614 178 do 8.0 3.5 641 930 1 hour 23 minutes to 2 hours 15 minutes. Do 1:6,011 166 do 8.0 4.0 641 912 4 hours 28 minutes to 6 iChemical wastes. hours 32 minutes. Do 1:6, 209 159 ... do 8.0 4. 5 641 853 CO Do 3 1:10,000 100 7.8 6.9 645 654 CO Do 3 1:100,000 10 7.8 7.5 645 650 CO Do 3 1:1, 000, 000 1:326 1 7.8 7.7 645 648 00 3,068 8.0 3.0 641 1, 100 40 to 50 minutes. _ Do 1:706 1,416 8.0 3.5 641 765 1 hour 12 minutes to 1 Dairy industries wastes. hour 45 minutes. Do 1:1,530 054 8.0 4.0 041 COO Do 1:2, 324 430 do 8.0 4.6 641 566 1 : 1, 333 750 7.8 3.4 666 1,625 818 30 to 50 minutes . . . Do 1:5, 000 200 7.8 4.9 066 •Chemical wastes. Do 1 : 10, 000 100 7.9 7.9 652 698 CO Oxalic acid 1 : 1, 000 1,000 do 7.8 2.6 666 2, 620 25 to 30 minutes... Do 1:5, 000 1:10, 000 1:17, 000 1:40, 000 1:100,000 3 1:1,000 200 7.9 5.3 652 405 CO Dye, tanning and bleaching wastes. Do . 100 7.9 6.8 652 524 00 Do 59 Very soft 6. 2 4.5 <50 <50 652 146 oo Do. _ 25 6.2 5.8 119 CD Do 10 7.9 7.6 662 co Sulphuric acid 1, 000 do _ 7.8 2.2 645 1,381 30 to 45 minutes Do 1:5, 922 169 do 7.8 3.0 640 1.203 50 minutes to 1 hour 10 minutes. Do 1:7,000 143 Soft 6.4 3.5 583 790 2 hours 30 minutes to 5 hours 17 minutes. Do. 1:7,000 143 7.8 3.5 640 1, 170 2 hours to 2 hours 20 min- utes. Do 1:7, 250 1:7, 250 138 Soft 6. 4 3.9 583 753 5 to 6 hours .. Do 138 7.8 4.0 640 868 4 hours to co _ Coal- and iron- mine waters. Do 1:7, 450 134 6.4 4.3 583 675 6 hours 12 minutes to 96 hours. Do 1:7, 450 3 1:7. 500 134 7.8 4.5 640 756 OO Do 133 7.8 5.0 6.8 645 673 OO Do 3 1:10, 000 100 7.8 645 698 CO Do 1:17, 000 59 Very soft 6.2 3.2 <50 450 1 hour to 1 hour 15 minutes. Do 1:60,000 3 1:100,000 1:120,000 31:1, 000, 000 1:1,000 17 6. 2 4. 5 <50 645 <50 645 666 93 CO Do 10 7.8 7. 5 692 OO Do 8 Very soft 6. 2 5. 6 85 00 Do 1 7.8 7. 5 695 CO Tannic acid. 1,000 do ... 7.8 6.4 631 2 hours 38 minutes to 3 hours 48 minutes. Do 1:5, 000 200 do . 7.9 7.3 652 652 7 hours 26 minutes to 9 hours 26 minutes. Tannery wastes. Do 1:10, 000 100 do 7.9 7.6 652 652 9 hours 40 minutes to 20 hours 20 minutes. Do 1:100, 000 1:1, 000 10 7.9 7. 8 652 652 CO 1,000 1, 000 Very soft 6. 2 3. 2 <50 655 900 20 to 25 minutes Do... 1:1, 000 7.8 3.9 727 3 hours to 3 hours 20 min- utes. Winery wastes. Do 1:5, 000 1:10, 000 200 7.8 4. 6 655 700 CO Do 100 Very soft 6.2 3.6 <50 205 3 hours to 3 hours 33 min- utes. Do 1:100, 000 10 6.2 4.9 <50 <50 CO 1 Standard goldfish weight 3 to 5 grams used. Each fish in individual glass container carrying 3 liters of solution. Dissolved oxygen 6 to 7 p. p. m.; temperature maintained between 18° to 23° O. Diluent waters discussed on page — . 2 Minimal and maximal survival times as found in these experiments. Infinity sign indicates survival greater than 4 days without any apparent injury to the fish. 3 Initial voiume of 3 liters per fish replaced by constant flow apparatus at rate of 1 liter per hour. Considering the actual acidity of the resultant mixture of acids and a given hard water, the most acid mixtures in which goldfish survive 4 days or more were pH 4.0, sulphuric acid; pH 4.5, hydrochloric and citric acids; pH 4.6, lactic acid; pH 4.9, nitric and tartaric acids; pH 5.5, acetic acid; pH 5.8, oxalic acid; pH 5.9, benzoic acid; pH 7.3, chromic acid; and pH 7.8, tannic acid. To this comparison, STREAM POLLUTION 411 it must be added, as previously pointed out in the section on pH of natural water, that Brown and Jewell (1926) found fish living in a bog lake where the acidity of the water was pH 4.5 due to the action of carbon dioxide and organic acids leached from the surrounding bog vegetation. It is evident from all of the data in table 8 that no matter what the diluent water lie, as far as hydrogen-ion concentration alone is concerned, the acidity near the magnitude of pH 4.0, regardless of the acid or acid- salt combinations producing this acidity, will be lethal for fresh-water fishes if that concentration of acidity be maintained. All acid concentrations more acid than pH 4.0 were lethal. In concentrations less acid than pH 4.0 the penetrative prop- erties and the lethality of the kation of the acid must also be considered, for it may be seen that at any given pH the relative lethality varies with the particular acid involved. On the basis of p. p. m. of acid present, disregarding the pH factor, the acids may be arranged in order of lethality as follows: Tannic acid, 10 p. p. m.; chromic acid, 100 p. p. m.; sulphuric acid, 130 p. p. m.; hydrochloric, 159 p. p. m.; benzoic, nitric, oxalic, and tartaric, each 200 p. p. m.; acetic acid, 348 p. p. m.; lactic, 430 p. p. m.; and citric, 625 p. p. m., in the particular hard water tested. The exact p. p. m. values are not so significant, for they will vary to some extent with the hardness of the diluent water; but the group shows the general order of magnitude of the relative lethality for these common pollutant acids. Fishes are killed by acid wastes first through the precipitation and coagulation of the mucus on the gills and by the coagulation of the gill membranes themselves. If this coagulation of gills and gill secretions does not take place, the death of the fish is attributable to the lethal action of the kation of the acid. Combinations of these two actions are not uncommon, i. e., both precipitation and specific toxic action may contribute to the death of the fish. The precipitation of the mucus and of the proteins within the gill membrane cells themselves progresses rapidly when the relative acidity of the mixture is more acid than pH 4.5 because of the acidity itself, but this precipitation may be enhanced by the action of the kation of the acid. Tannic, chromic, and nitric acids all have marked affinity for living protoplasm, forming insoluble compounds with certain protein constituents of living tissue very promptly; in fact, all three of these acids are used as histological fixing agents to kill protoplasm. In the previous comparison on the basis of pH value, the minimal lethal concentration of these acids shows that tannic, chromic, and nitric killed gold- fish in solutions less acid than the lethal solutions of sulphuric, hydrochloric, citric, and lactic acids. Acetic acid has high penetrative properties and causes a swelling of tissues which is very destructive to the living cells, so that although this acid does not coagulate the gill membrane and gill mucus like hydrochloric acid, it does disrupt the cells of gill membrane with disastrous results. Consequently, it is not surprising to find that acetic acid kills goldfish at pH of 5.5. Reviewing all of the data on acid wastes, it seems that the truly acid effects must be limited largely to those acids which kill at a hydrogen-ion concentration more acid than pH 5.0; while in the cases of those acids killing at hydrogen-ion concentration less acid than pH 5.0, lethality factors other than hydrogen-ion con- centration play the major part. Of course, as the hydrogen-ion concentration of the surrounding medium in which the fishes and other aquatic animals are living becomes more acid than pH 7.2 to 7.4 (the normal values for most living cells), 412 BULLETIN OF BUREAU OF FISHERIES detrimental effects which may act synergistically with the specific toxic actions of the substances producing such deviations in pH must be expected. In this connection, it must be pointed out that although the lethality limits of acid wastes as given here are applicable to immediate or intermittent pollution the cumulative effects must also be considered. When goldfish were kept in water at pH 4.5 produced by addition of small quantity of sulphuric acid, this amount of sulphuric acid, although tolerated for a few days without apparent injury to the fish, seemed definitely detrimental to goldfish in exposures longer than 2 weeks, even though some fishes can tolerate pH 4.5 in the natural habitat when this degree of acidity is produced by carbon dioxide and acids from decaying vegetation in bog and swamp lakes. It may be necessary, therefore, to add sulphuric and other acids to the lists of those producing specific toxic effects when long-time experiments in progress are completed. Table 9 presents data concerning the effects of acids on daphnia comparable to those given in table 8 for fishes. In the cases of daphnia which do not produce volumes of protective mucus to remove or in part buffer down the acid in the effluents to which they may be exposed, the maximal acidity of the solutions in which there was any survival by daphnia was around pH 5.4. It is significant in this connection to recall that the isoelectric point at which many proteins found in living tissue are precipitated is pH 5.5. The daphnia tests, therefore, suggest strongly that were fishes unprotected by mucus the lethal hydrogen-ion concentration for acids in general would be near pH 5.5, and explain the detrimental effects of acid wastes to the stream faunae in general even though the particular concentration may be sur- vived by some fishes. In waters less acid than pH 5.5 the same relative lethalities of the various acids were found for daphnia as for fishes. Table 9. — Survivals of l+,500 Daphnia magna in various concentrations of 5 acids found in indus- trial wastes 1 Substance Concentra- tion ratio by weight Parts per million pH Specific conduc- tivity mhoXlO"6 at 25° C. Survival time Final percent mortality W ater Solution Water Solution 1:1, 333 750 7. 4 4 0 470 516 100 Do— . 1:2, 000 500 7. 4 4 6 470 491 do— 100 Do 1:5' 333 188 7. 4 5. 0 470 461 _do 100 Do 1:8, 000 125 7. 4 5. 4 470 460 24 to 72 hours 75 Citric acid . _ 1:2, 666 375 7.6 4.0 425 532 1 to 2 hours 100 Do 1:4, 210 248 7. 6 4. 6 425 491 2 to 17 hours 100 Do 1:5, 405 185 7.6 5. 0 425 490 10 to 17 hours 100 Do 1:8, 340 120 7. 6 5.5 425 466 24 to 72 hours 40 Do 1:12, 500 80 7. 6 5. 9 425 440 OO o Hydrochloric acid. . .. 1:14, 545 69 7. 4 4.0 452 660 1 to 4 hours. . .. 100 Do 1:15, 384 65 7. 4 4. 5 452 650 do 100 Do 1:16, 667 60 7.4 5.0 452 614 4 to 17 hours 100 Do 1 : 17, 777 56 7. 4 5. 4 452 595 80 1:3’ 470 288 7. 6 4. 0 442 493 100 Do 1:4, 280 234 7.6 4.5 442 463 do 100 Do 1:5, 230 191 7. 6 5. 0 442 443 6 to 48 hours. _ 100 Do 1:5, 880 170 7.6 5. 5 442 441 26 to 72 hours ... 66 Sulphuric acid. . 1:20, 000 50 7.3 3. 5 324 610 1 to 3 hours. .. 100 Do 1:26,667 38 7. 3 4.0 324 184 24 hours 100 Do 1:33, 333 30 7. 3 4. 5 324 389 100 Do 1:34, 479 29 7. 3 5. 0 324 381 24 to 72 hours 100 Do 1 : 50, 000 20 7. 6 6. 5 375 387 0 Do 1 : 100i 000 10 7.6 7.3 375 375 do 0 Do. . 1:1, 000, 000 1 7. 6 7. 6 375 375 0 Do 1:10, 000! 000 0.1 7.6 7.6 375 375 do 0 1 Standard daphnia, 5 days old, from parthenogenie clones were used. For each dilution 150 animals were carried in glass jars containing 4 liters of the solution, and these jars held at 25° C. in constant temperature chamber. The diluent was soft water (see p. 406) from the same source as that in which the stock colonies of daphnia were living. A control of 150 animals in 4 liters of this water without test substance was maintained for each series. STREAM POLLUTION 413 COMPOUNDS OF VARIOUS METALS Compounds of many metals are rapidly lethal, because these compounds coagu- late and precipitate the mucus secreted by the gills of fishes (Carpenter, 1930) and many of the proteins in the living cells. This action has been explained in the dis- cussion of the effects of pollutants on fishes (p. 400). In addition, some metallic salts, such as ferric chloride, if present in sufficient quantities may increase the acidity of the water to a level dangerous for aquatic life and combine this hazard with that of the metallic coagulant. Since the carbonates and phosphates of many metals are relatively insoluble in water, the hardness of waters receiving effluents containing compounds of the heavier metals is an important factor in determining the immediate lethality of such effluents to fishes, as a considerable portion of these metals may be precipitated from the effluent by the salts in the water. This factor has led to much confusion concerning the absolute limits of lethality of metallic compounds for fishes, since various ob- servers have used different kinds of test water, ranging from distilled to very hard. Unless the hardness of the water in question and particularly the amounts of carbon dioxide, both fixed and free, be known the limits of lethality of any particular metallic compound in any given water are difficult to estimate even though there may be ample data on the specific toxicity of that compound. Besides, the factors of synergy and antagonism between the metallic compound itself and other compounds in the water or effluent which do not precipitate this metallic compound must be considered. Examples of this action are presented in table 10 for copper sulphate. Table 10. — Influence of other salts in solution upon the toxicity of copper sulphate to gold fish 1 Series Copper sulphate concentration Sodium nitrate concentration Calcium chloride concentration pH Specific con- ductivity mho X 10'6 at 25° C. Average survival time Ratio by weight Parts per million Ratio by weight Parts per million Ratio by weight Parts per million Water Solu- tion W ater Solu- tion A 1 : 100, 000 10 7. 1 4. 0 <50 112 B 1 : 100. 000 10 1:2,000 500 7. 1 5.7 <50 806 C .. 1:100, 000 10 1:1,000 1, 000 7. 1 5. 7 <50 1,478 D 1 : 100, 000 10 1:500 2, 000 7. 1 5.8 <50 2,749 E 1 : 100, 000 10 1:333 3, 000 7. 1 5.8 <50 4,020 F 1:100, 000 10 1:250 4, 000 7. 1 6.0 <50 5,413 G_ 1 : 100; 000 10 1:200 5, 000 7. 1 6.0 <50 6,668 H 1:100. 000 10 1 : 1G7 6, 000 7. 1 6. 4 <50 7,734 J 1:100, 000 10 1 : 143 7, 000 7. 1 6. 2 <50 9,013 K 1 : 100. 000 10 1:125 8, 000 7. 1 6. 3 <50 10, 113 L 1:100,000 10 1:2, 000 500 1:20,000 50 7. 1 6.3 <50 lj 513 5 hours 15 minutes. M 1:100,000 10 1:1,000 1,000 1:20,000 50 7. 1 6.3 <50 1,541 5 hours 30 minutes. N 1 : 100. 000 10 1:500 2,000 1:20,000 50 7. 1 6.4 <50 2,825 6 hours 30 minutes. O 1:100. 000 10 1:333 3,000 1:20, 000 50 7. 1 6.5 <50 3,984 9 hours 30 minutes. P . 1:100, 000 10 1:250 4, 000 1:20, 000 50 7. 1 6.4 <50 5,350 12 hours 30 minutes. Q 1 : 100, 000 10 1:200 5, 000 1:20,000 50 7. 1 6.4 <50 6,573 12 hours 45 minutes. R 1 : 100, 000 10 1:167 6, 000 1:20,000 50 7. 1 6.5 <50 7,866 8 hours. S 1:100,000 10 1:143 7,000 1:20, 000 50 7. 1 6. 5 <50 8,763 7 hours. T 1:100,000 10 1:125 8, 000 1:20,000 50 7. 1 6.5 <50 10,338 Do. IT 1:333 3, 000 7. 1 7. 3 <50 4,221 (2). V 1:250 4, 000 7. 1 6.7 <50 5' 350 80 hours. w 1:200 5, 000 7. 1 6.8 <50 6,620 37 hours. X 1:167 6, 000 7. 1 6.9 <50 7,798 6 hours 30 minutes. 1 Glass distilled water was used in these tests. Other conditions as described in table 8. 3 Survival greater than 4 days, fish apparently unaffected. 414 BULLETIN OF BUREAU OF FISHERIES It may be noted in this table that the addition of progressively larger quantities of sodium nitrate to the copper sulphate solution increased the survival time of fishes from 2 hours 30 minutes to 6 hours 40 minutes, up to a concentration of 1 : 200 for sodium nitrate. Beyond that concentration the osmotic action of the sodium nitrate became unfavorable (see osmotic pressure, p. 408) and the survival time deci eased with a further increase of sodium nitrate, i. e., sodium nitrate, which does not precipi- tate copper, both increased and decreased the survival time of fishes exposed to the lethal action of same amounts of copper sulphate according to the concentration of sodium nitrate present. The second half of this table shows the marked increase in survival time from 2 hours and 30 minutes in the copper sulphate alone to 12 hours 45 minutes in the same strength of copper sulphate solution to which both sodium nitrate and calcium chloride were added. The amount of calcium chloride used was very small, namely, 50 p. p. m., and in this concentration no precipitate formed, yet this quantity of calcium chloride greatly enhanced the protective action of the sodium nitrate against the lethal action of copper sulphate. These discussions of the precipitation of metallic compounds out of effluents and of the synergistic and antagonistic actions of compounds which do not preci- pitate the particular metal in question show, both specifically in the case of effluents carrying compounds of metals and in general in the case of all stream pollutants carrying chemical wastes, that field examinations of local conditions, analyses both of the effluents and the water receiving these effluents, and bioassays of the mixtures of the two must be made before the lethal limits of such wastes can be determined. In table 1 1 summarized data from a large series of tests covering the lethality limits of certain compounds of eight metals found in commercial effluents are presented. From these data the expected lethal ranges and the relative lethality of these eight compounds in various types of water, together with the changes in salt balance and relative acidity produced by these compounds, may be estimated if uncomplicated by substances other than those normally found in stream water, or by excessive amounts of those substances. Table 11. — Survivals of 850 goldfish in various concentrations of salts of 8 metals found in industrial wastes 1 Substance Concen- tration ratio by Parts per mil- Diluent water pH Specific eon- ductivity inhoXlO-6 at 25° C. Survival time 2 Constituent of— weight lion Water Solu- tion Water Solu- tion Aluminum potassium sulphate. 3 1:1,000 1,000 Hard 7.8 5. 5 678 1,247 1 to 10 hours Do. 3 1: 10,000 100 do - 7.8 6.8 678 736 12 hours to co . •Tannery wastes. Do 3 1:100, 000 10 _ do__ 7.8 7.6 678 584 Do. 1:1,000, 000 1 ___do_ 7.8 7. 7 678 677 OO Cobaltous chloride 1:1,000 1,000 Very soft 6.2 6.6 <50 1,906 28 to 29 hours Electroplating wastes. Do- 1:1, 000 1,000 Hard. 7.8 7.2 647 2,265 30 to 31 hours _ Do. 1:10, 000 3 1:1, 000 100 Very soft 6. 2 6.5 <50 <50 690 308 168 hours to <» . __ 1, 000 1,000 Glass distilled. _ 7.0 5.6 1, 123 1,348 1 to 2 hours ~Do r_ 3 1:1,000 Hard 7.8 6.8 1 hour to 2 hours 20 no 3 1:10, 000 100 _ do 7.8 7.0 671 637 minutes. 3 hours to 11 hours _ Copper and i brass indus- tries wastes. Do. 1:10, 000 3 1:100,000 100 do 7.9 7.0 635 637 3 to 41 hours Do_ 10 _ -__do 7.9 7.6 635 671 11 to 72 hours Do_ 1:100,000 3 1:1, 000, 000 10 __ do 7.9 7.7 635 700 48 hours to _ Do 1 do 7.9 7.5 635 640 72 hours to <= STREAM POLLUTION 415 Table 11. — Survivals of 850 goldfish in various concentrations of salts of 8 metals found in industrial wastes 1 — Continued Specific con- pH ductivity Concen- Parts mhoXlO-* Substance tration per Diluent at, zo Ay. Survival time 1 Constituent of— ratio by weight mil- water lion Water Solu- tion Water Solu- tion 3 1:1.000 1, 000 do 7.7 6.4 594 1,307 2 to 10 hours Do 3 1:10,000 3 1:100,000 3 1:1, 000, 000 1:1,000 1:10,000 100 do 7. 7 6.7 594 649 OO Do 10 ...do 7. 7 7.4 594 598 CO Do 1 ...do 7.7 7.6 594 600 CO Wire and tin- plate mills wastes. Ferric chloride Do 1,000 100 Very soft.. . . do 6.2 6.2 2. 4 3.4 <50 <50 2, 156 334 1 hour to 1 hour 10 minutes. 1 hour to 1 hour 20 minutes. Do 1:10, 000 1:100, 000 3 1:1,000 100 Hard 7.8 5.5 647 700 00 Do. 10 Very soft... . 6.2 5.0 <50 678 <50 00 1,000 100 Hard 7.8 6.4 2, 500 2 to 3 hours Mining and ■ smelting wastes. Do 3 1:10,000 3 1:100, 000 do 7.8 6.8 678 775 80 hours to “> Do 10 1 .. do 7.8 7. 4 678 678 CO Do 3 1:1, 000, 000 1:1,000 ...do 7.8 7.5 678 673 CD Nickelous chloride 1,000 Very soft.. 6.3 6.3 <50 1,876 6 hours to 18 hours 30 minutes. Do 1:1,000 1,000 100 Hard 7.8 7.4 647 2, 218 253 12 to 18 hours Electroplating wastes. Do... 1:10,000 Very soft 6.3 6.3 <50 19 hours 25 minutes to 50 hours 25 min- utes. Do 1:100.000 10 do 6.5 6.5 <50 95 200 to 210 hours 1: 1, 000 1, 000 Very soft 6. 4 3. 5 <50 647 2, 380 1 hour to 1 hour 30 Brass industries wastes. Do. 1:1,000 1,000 1,000 100 Hard 7.8 3.8 1,609 1,538 780 minutes. 4 to 5 hours 3 1:1,000 do 7.8 7.2 678 1 to 4 hours Mining and smelting wastes. Do 3 1:10,000 3 1:100,000 3 1:1, 000, 000 do 7.8 7.6 678 Do 10 do 7.8 7.6 678 678 CO Do.. _ 1 do 7.8 7.6 678 678 CD 1 Experimental conditions as in table 8. 3 Minimal and maximal survival times as found in these experiments. Infinity sign indicates survival greater than 4 days without any apparent injury to the fish. 3 Initial volume of 3 liters per fish replaced by constant flow apparatus at rate of 1 liter per hour. Another hazard from wastes containing compounds of various metals is the cumulative effect. Even though precipitated out, compounds carrying these metals, as long as they remain in the stream, on the bank, or on the floor of the stream, are subject to re-solution if conditions in the waters change (in many cases only slightly), and to ingestion by fish and other aquatic animals during feeding. The case of lead salts will suffice for example. Carpenter in a series of papers (1924, 1925, 1926, 1927, and 1930) has shown that the lethal action of lead compounds on fishes in cases where the more soluble lead salts are present is due to the precipita- tion of mucus and proteins in and on the gills by the lead ion, and that the absorption and, therefore, the internal poisoning of fishes by lead compounds under these condi- tions is negligible during moderately short exposures. This conclusion has been con- firmed by Behrens (1925) using a very delicate test (the radio-activity of an isotope of lead) which demonstrated the distribution of lead throughout the body of the fish. However, the writer in long-time experiments in which goldfish were kept in glass containers, the bottoms of which were covered with finely powdered lead ore (lead sulphide, which is very insoluble in water), found cumulative effects which resulted in the death of fish after 61 days of such exposure, although there were no casualties in the controls carried under identical conditions except for the presence of lead ore. Although these experiments which are being followed further do not in themselves prove that the cumulative action was that of lead alone or of lead in con- junction with some other substance (i. e., traces of other substances in the ore may 416 BULLETIN OF BUREAU OF FISHERIES have been responsible for these cumulative effects), these experiments do show that precipitated and insoluble substances on the bottom of the stream can be a definite pollution hazard over a long period of time due to cumulative action. Again, the writer has found that insoluble zinc ore (zinc sulphide) carried by a stream and de- posited on flats adjacent to the main stream bed during high water when exposed to the actions of sunlight, air, and moisture produced quantities of zinc sulphate, a freely soluble and definitely toxic compound which was leached back through into the stream subsequently by rains and high waters (Ellis, 1932). The synergistic physiological action of many compounds is well known, and the work of Macht and Leach (1930) on goldfish and other types of animals may be cited in this connection, as these writers found that two or more octylic alcohols were synergistic in their action on the respiratory and neuromuscular mechanisms of animals. MISCELLANEOUS COMPOUNDS In tables 12 and 13 the relative toxicities of 13 other substances found in indus- trial wastes or otherwise constituting stream pollution hazards are presented. The changes in hydrogen -ion concentra tion and salt balance are shown, as in the tables giving the data on acids and metallic compounds. It may be seen, however, that the substances listed in tables 12 and 13 do not kill because of changes either in pH or salt balance, i. e., in these two tables are included representatives of those substances which, if detrimental, are so because of specific toxic action after entering the body of the fish. Phenol produces paralysis of the neuromuscular mechanisms and hemolyzes the blood; potassium xanthate in high dilution acts after several days on the gastrointestinal tract; amyl alcohol quickly induces a semiparalyzed condition with marked incoordination, yet the fish may live for days in this state before dying; and sodium selenite causes a slow decline in the general activities of the fish. Table 12. — Survival of goldfish in solutions of sulphur and selenium compounds 1 Substance Concentra- tion ratio by weight Parts per mil- lion Diluent water pH Specific con- ductivity mho X 10-» at 25° C. Survival time 2 Constituent of— Water Solu- tion Water Solu- tion Ammonium sulphide.- 3 1:1,000 1, 000 Hard - 7.7 7.9 681 938 15 minutes to 1 hour Do. 3 1:10,000 100 do 7.7 7.8 681 735 30 minutes. 72 hours to oo. Sewage and or- Do 3 1:100,000 10 do 7.7 7.7 681 681 ganic wastes. Do_ 3 1:1,000,000 1 do 7.7 7.7 681 681 OO Hydrogen sulphide 3 1:1,000 1, 000 _-do 7.8 6.5 670 734 45 minutes to 1 hour Do. 3 1:10,000 100 7.8 7.3 670 676 3 to 4 hours _ Do. Do 3 1 : 100, 000 10 7.8 7.6 670 661 96 hours to oo _ Do 3 1:1,000' 000 1 7.8 7.7 670 670 Sodium sulphite 3 1:1,000 1,000 do 7.8 7.6 676 2,017 3 to 72 hours. Do 3 1:10,000 100 do _-- __ 7.8 7.6 676 864 96 hours to oo _ _ Paper-pulp-mill Do 3 1:100, 000 10 do 7.8 7.6 676 678 wastes. Do 3 1:1,000, 000 1 - .do _- 7.8 7.8 676 677 m Sodium selenite 1 : 1, 000 1,000 Very soft 6.4 7.5 <50 753 1 hour to 1 hour 30 min- utes Do 1:1, 000 1,000 Hard--- 7.8 7.4 647 1, 141 1 hour to 2 hours 10 min- utes. Do 1:10, 000 100 Very soft 6.4 7.6 <50 107 20 hours 56 minutes to Certain soils. Do 1:10,000 100 Hard 7.8 7.7 647 723 8 hours to 19 hours 30 minutes. Do 1:100, 000 10 Very soft 6.4 7.3 <50 <50 98 to 144 hours 1 Conditions of experiments described in table 8. 3 Minimal and maximal survival times as found in these experiments. Infinity sign indicates survival greater than 4 days without any apparent injury to the fish. 1 Constant flow as described in table 8- STREAM POLLUTION 417 The specifically toxic substances which do enter the body of the fish can be grouped therefore according to their specific pathological and pharmacological effects much as in the case of man; and these poisons which do enter the body of the fish constitute important pollution hazards in high dilution, since the high dilutions may be tolerated by the fish for a short time without apparent harmful results. Conse- quently, the compounds in this miscellaneous group must be considered individually in evaluating pollution hazards. Table 13. — Survivals of goldfish in solutions of various chemicals found in industrial wastes 1 Specific con- pH ductivity mhoXlO-8 at Substance Concentra- tion ratio by weight per Diluent water 25° C. Survival time 3 Constituent of lion Water Solu- tion Water Solu- tion Ammonium carbonate . Do . 3 1:1, 000 3 1:10, 000 1,000 100 Hard _ 7. 7 8. 1 718 2,644 956 1 to 2 hours- do 7.7 7.8 718 4 to 10 hours .. [Organic wastes, Do Do 3 1:100, 000 3 1:1,000, 000 1:1, 000 10 1 7. 7 7. 7 718 727 [ gashouse wastes. 7.7 7.7 718 718 CO Amyl alcohol (iso)- --- 1, ooo do 7.8 7.8 647 666 17 to 71 hours Do 1 : 10, 000 1:100, 000 1:1, 000,000 1:1, 000 100 _do - 7.8 7.8 647 647 82hours-co__ [Distillery wastes. Do 10 ___do --- -- 7.8 7.8 647 647 94 hours- Do 1 __do- . . 7.8 7.8 647 647 161 hours-oo ... 1 1, coo 400 do 8.0 6.7 646 673 20 to 40 minutes .. Do _. 1:2, 500 do 8.0 7.4 646 659 30minutes to 1 hour Do 1:10,000 100 do 8.0 7.6 646 649 10 minutes. 1 hour to 3 hours 25 minutes. [Chemical indus- tries wastes. Do 1:50,000 1:100, 000 1:10, 000 20 do 8. 0 8. 0 646 646 1 5 to 96 hours . . . Do 10 do 8. 0 8. 0 646 646 CO Chlorine. lool do 7.9 7.4 653 640 54 minutes to 1 hour 30 minutes. Do 1:100, 000 10 - __do 7.9 7.9 653 653 5 hours 15 minutes Do. to 48 hours. Do 1:500, 000 2 do 7.9 7.9 653 653 17 to 48 hours ... . Do 1:1,000,000 1:1,000 1 _ __do 7.9 7.9 653 653 Cresylic acid — - 1, 000 do 7.8 7.8 639 654 14 to 30 minutes Do 1:10, 000 100 do 7.8 7.8 639 641 5 to 31 hours Livestock disin- Do 1:100, 000 1:1, 000, 000 10 7.8 7.8 639 640 Do 1 do 7.8 7.8 645 645 93 to 120 hours tation wastes. Do.-.. 1:10,000, 000 3 1:1,000 0. 1 -__do 7.8 7.8 645 645 Phenol 1, 000 do 7.8 7.6 690 587 15 to 30 minutes Livestock disin- ■ fectants and gas- house wastes. Do 3 1:10, 000 100 do 7.8 7.6 690 593 60 to 72 hours Do - 3 1:100, 000 3 1:1,000,000 10 -_do 7.8 7. 8 690 690 Do . 1 do_ - 7.8 7.8 690 690 Potassium xanthate 3 1:3, 125 320 do 8. 1 8.8 669 900 Do 3 1:6,250 160 . . - -.do 8.1 8.6 669 776 12 to 24 hours ... Do.... ... 3 1:12, 500 80 do 8. 1 8.4 669 731 1 2 to 36 hours Mine flotation Do... 3 1:100, 000 3 1:10, 000, 000 10 -_do - 8. 1 8. 1 669 669 wastes. Do 0. 1 do 8. 1 8.1 609 669 4 to 5 davs . ... 1:1, 000 1, 000 S. 0 7. 7 645 753 Sodium fluoride 1:1,000 1, oco Very soft 6. 4 6.4 <50 2, 162 12 to 29 hours. ... Distillery wastes and certain soils. Do Do 1:1, 000 1:10, 000 1,000 100 Hard do 7.8 7.8 7.0 7.3 647 647 2, 346 943 60 to 102 hours. . CO 1 Condition of experiment described in table 8. 3 Minimal and maximal survival times as found in these experiments. Infinity sign indicates survival greater than 4 days without apparent injury to the fish. 3 Constant flow as described in table 8. LETHAL LIMITS OF 114 SUBSTANCES WHICH MAY BE FOUND IN STREAM POLLUTANTS In this section summaries of the lethal limits of individual substances, together with bibliographic references and statements concerning the test animals used and conditions of the experiments, are presented. These data have been massed from the literature and from the writer’s own experiments (M. M. E.). In these sum- maries no attempt has been made to include all existing references, but instead 418 BULLETIN OF BUREAU OF FISHERIES sufficient data to give a definite idea of the relative lethality of the substance in question under conditions which may be encountered in the streams of the United States. Only fresh-water fishes and fresh-water organisms have been considered. It is hoped that these summaries will prove helpful in evaluating specific pollution conditions when used in conjunction with the data presented in preceding sections. However, as has been pointed out repeatedly in this paper, arbitrary application of lethality data to specific pollution problems is absolutely impossible owing to the many limiting factors, various of which have been discussed in previous sections. Any student of the toxicity of the components of industrial and municipal efflu- ents to fishes and other aquatic life will find the pioneer work of Penny and Adams (1863) and of Weigelt, Saare, and Schwab (1885) starting points for the investigations of the commoner substances. However, much progress has been made in the study of the factors limiting the action of many pollutants, since these experiments and the lethality values given by these earlier observers have been quoted too often without regard to more recent findings. Important advances have been made in the measurement of stream-pollution hazards of particular effluents through the isolation from these effluents of various compounds of highly toxic nature, which compounds were formerly overlooked or disregarded because the actual amounts of these substances present in the effluent were quite small. These highly toxic compounds, some of which are volatile or may otherwise disappear from the effluent after a time, have been responsible for the wide discrepancies in lethality of the waste as a whole as reported by different observers working on the same type of effluent. In the present list of substances which may be participants in stream-pollution problems, several compounds have been included which are little known except to the professional chemist, although the harmful properties of these substances have been demonstrated in connection with certain industrial effluents. The list of these less familiar compounds will undoubtedly grow, both as investigations of particular effluents are completed and as chemical engineering produces new commercial proc- esses. The long series of dyes, many of which are highly toxic to aquatic life, used in industries and arts have not been included in this list of possible stream pollutants because the dye wastes are largely mordant, bleach and processing liquors containing only traces of the dyestuffs themselves which industry has found too costly to waste. The conservation of these expensive dye compounds by the manufacturing concerns has been very apparent to the writer in several recent investigations of plants dyeing fabric, paper, and leather. Acetic acid, CH3-COOH. Vinegar acid Beet-sugar pulp waste; some winery wastes; soured fruit wastes; and vinegar works. Penny and Adams (1863), 114 p. p. m. killed minnows in 20 hours, but 286 p. p. m. not fatal to goldfish; M. M. E., 100 p. p. m. in hard w’ater killed some goldfish, Carassius auratus, and 125 p. p. m. killed cladocerans, Daphnia magna, in 24-72 hours. Acetone, (CH3)2CO Gas and coal-tar wastes; paint and chemical industries. Shelford (1917), 14,250-15,050 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis. STREAM POLLUTION 419 Acridine, C9H4(N.CH)C„H4 Gas tar oils; dye industries; water sterilization. Adams (1927), 5 p. p. m. in Nile River water killed cladocerans, Daphnia sp., and copepods, Cyclops sp. Alcohols See amyl alcohol, butyl alcohol, ethyl alcohol, methyl alcohol, and octyl alcohol. Aluminium ammonium sulphate, A1(NH4) (S04)2- Ammonium alum Dye works and cloth printing industries. Weigelt, Saare, and Schwab (1885), 523 p. p. m. in tap water caused large trout to float on side after 10 hours’ exposure. Aluminium potassium sulphate, A1K(S04)2‘ Alum or potassium alum Dye works; tanneries; leather works. Penny and Adams (1863), 250 p. p. m. killed goldfish and minnows; Weigelt, Saare, and Schwab (1885), 544 p. p. m. in tap water killed California salmon in 6 hours, and medium trout in 15 hours; M. M. E., 1,000 p. p. m. in hard water killed goldfish, Carassius auratus in 1-10 hours, 100 p. p. m. killed goldfish in 12 96 hours, although some survived this strength over 100 hours. Alums See aluminium ammonium sulphate, aluminium potassium sulphate, and ferric potassium sulphate. Ammonia See ammonium hydroxide. Ammonium carbonate, (NH4)2C03 Gas wastes; wool- washings; tannery effluents. Clark and Adams (1913), 155-197 p. p. m. fatal to shiners and carp in a few minutes to a few hours; Shelford (1917), 600-800 p. p. m. killed orange-spotted sunfish, Lepomis humilis in 1 hour; M. M. E., 100 p. p. m. in hard water killed gold- fish in 4-10 hours, 48 p. p. m. in 6 days, 10 p. p. m. tolerated for more than 100 hours without appar- ent effect. Ammonium chloride, NH4C1. Sal ammoniac Gas wastes; various chemical wastes. Clark and Adams (1913), 180 p. p. m. in tap water, no effect on shiners or carp; Wells (1915b), 535 p. p. m. in tap water killed bluegills, Lepomis pallidus in 4 hours 45 minutes, and same amount in distilled water killed bluegills in 18 days; Shelford (1917), 700-800 p. p. m. killed orange-spotted sunfish, Lepomis humilis, in 1 hour; Powers (1917), 1,712 p. p. m. in distilled water killed goldfish, Carassius auratus, in 6-18 hours; M. M. E. 268 p. p. m. in hard water, killed goldfish in 6 days; 535 p. p. m. killed cladocerans, Daphnia magna, in 6 hours. Ammonium ferrocyanide (NH4)4Fe(CN)9 Gas wastes, in 1 hour. Shelford (1917), 150-200 p. p. m. killed orange-spotted sunfish, Lepomis humilis, Ammonium hydroxide, NH4OH Wool washings; chemical wastes; gas wastes. Since ammonia gas NH3 unites readily on mixing with water forming the hydroxide, these two compounds are considered together here. Clark and Adams (1913), 9.4 p. p. m. (ammonium hydroxide) did not kill shiners, carp, and large suckers but 13 p. p. m. killed all these fishes, 20 p. p. m. being fatal in 15 minutes; Belding (1928), 6.25 p. p. m. killed brook trout in 24 hours; Weigelt, Saare, and Schwab (1885), 250 p. p. m. (ammonia) killed trout in 2 hours; Shelford (1917), 7-8 p. p. m. (ammonia) in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour; M. M. E., 2-2.5 p. p. m. (ammonia) in hard water fatal to goldfish, Carassius auratus, and yellow-perch, Perea flavescens, in 24 hours to 4 days. 420 BULLETIN OF BUREAU OF FISHERIES Ammonium nitrate, NH4NO3 Chemical industries; manufacture of explosives; in fertilizers. Wells (1915b), 800 p. p. m. in tap water killed bluegills, Lepomis pallidus, in 3.9 hours; in distilled water lethal to bluegills in 16 days; Powers (1917), 4,545 p. p. m. in distilled water killed goldfish, Carassius auratus, in 90 hours. Ammonium sulphate, (NHdaSCh Chemical wastes; gas wastes; fertilizers. Wells (1915b), 66 p. p. m. in tap water killed bluegills, Lepomis pallidus, in 3 hours, 30 minutes, in distilled water lethal to bluegills in 17 days; Shelford (1917), 420-500 p. p. m. in tap water killed Lepomis humilis, in 1 hour; M. M. E., 264 p. p. m. in hard water killed goldfish, Carassius auratus, in 6 days or less. Ammonium sulphide, (NH^Si Chemical wastes; in wastes from some beet-sugar processes; in some gas wastes. M. M. E., 100 p. p. m. in hard water killed goldfish, Carassius auratus, in 72 hours, 10 p. p. m. in hard water, goldfish survived more than 100 hours of exposure. Ammonium thiocyanate, NH4CNS. Rhodanammonium Gas wastes; chemical industries. Shelford (1917), 280-300 p. p. m. in tap water killed orange- spotted sunfish in 1 hour; Demyanenko (1931), 200 p. p. m. lethal for fish. Amyl alcohol (iso), C5H11OH Chemical wastes; in some distillery wastes. M. M. E., 100 p. p. m. in hard water killed goldfish, Carassius auratus, in 82 hours, 1 p. p. m. killed goldfish in 161 hours. Aniline, CaHsNH^ Dye wastes; chemical industries; gas wastes. Shelford (1917), 1,020-1,122 p. p. m. killed orange- spotted sunfish, Lepomis humilis, in l hour; Demyanenko (1931), 250 p. p. m. lethal for fishes. Arsenic trioxide, white arsenic See sodium arsenate. Barium chloride, BaCb Some chemical wastes; some alkaline pools. Powers (1917), 5,000 p. p. m. in distilled water killed goldfish, Carassius auratus, in 12-17 hours. Benzoic acid, CjHs-CQOH Gas wastes. Shelford (1917), 550-570 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour; M. M. E., 200 p. p. m. in hard water killed goldfish in 7-96 hours. See cupric sulphate. Blue Vitriol Bromine, Brj Chemical industries; bromine plant wastes. Penny and Adams (1863), 29 p. p. m. killed min- nows and goldfish; M. M. E., 20 p. p. m. in hard water killed goldfish in 15-96 hours, 10 p. p. m. in soft water lethal for cladocerans, Daphnia magna. Butyl alcohol (iso), C4H9OH Paint and varnish solvents; some chemical industries. Powers (1917), 250 p. p. m. in distilled water killed goldfish, Carassius auratus, in 7-20 hours. Cadmium chloride, CdCB Pigment works; calico printing; chemical wastes. Powers (1917), 0.0165 p. p. m. in distilled water killed goldfish in 8 hours 40 minutes to 18 hours. STREAM POLLUTION 421 Cadmium sulphate, CdSO< Same sources as cadmium chloride. Carpenter (1927), 1,042 p. p. m. in distilled water killed minnows, Leuciscus phoxinus, in 3 hours. Calcium chloride, CaCL Wastes from bromine and salt works; in waters from oil wells; antidust road surfacing. Wells (1915b) 555 p. p. m. in tap water caused pathological degeneration of tail fin of rock-bass, Amblo- plites rupestris, in an exposure of 1 week; Garrey (1916), 2,775 p. p. m. in distilled water killed straw- colored minnows, Notropis blennius, in 2-4 days, but this species did not succumb to 277 p. p. m. in distilled water in 5-7 weeks; Powers (1917), 7,752 p. p. m. in distilled water killed goldfish, Carassius auratus, in 22-27 hours; Wiebe, Burr, and Faubion (1934), 5,000 p. p. m. killed golden shiners, Notemigonus crysoleucas, in 143 hours. Calcium hydroxide, Ca(OH)2- Lime Tannery-wastes; leather works. Weigelt, Saare, and Schwab (1885), 700 p. p. m. in tap water killed trout in 26 minutes; Marsh (1907), 18 p. p. m. (as calcium oxide) fatal to trout fry. Calcium nitrate, Ca(N03)2 Some chemical wastes. Powers (1917), 6,061 p. p. m. in distilled water killed goldfish, in 43-48 hours. Calcium oxide, CaO. Unslaked lime Forms calcium hydroxide, q. v., immediately on addition to water. Carbon bisulphide, CS2 Gas wastes; as a solvent in various chemical industries. Weigelt, Saare and Schwab (1885), a 7 minute exposure to 5,000 p. p. m. killed trout 2 days later; Shelford (1917), 100-127 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour. Carbon dioxide, CO2 From sewage and any decomposing organic wastes (see section on carbon dioxide in natural waters). Wells (1918), 113,925-151,900 p. p. m. in tap water killed fresh-water fishes. Carbon monoxide, CO Motorboat exhausts; in water returned to streams from cooling systems of engines using fuel oil if mixed with exhaust gases. Wells (1918), 1,160 p. p. m. in tap water killed straw-colored minnow, Notropis blennius, in 1 hour; the blunt-nosed minnow, Pimephales notatus, in 1 hour 55 minutes; the orange-spotted sunfish, Lepomis humilis, in 5 hours 40 minutes; the green sunfish, Lepomis cyanellus, in 6 hours; 11,314 p. p. m. under same conditions killed the black bullhead, Ameiurus melas, in 9 hours 55 minutes. Chloramine, NH2CI W'ater purification systems either as chloramine or formed from small quantities of ammonia and chlorine in the water. Coventry, Shelford, and Miller (1935), 0.3-0. 4 p. p. m. in tap water killed trout fry at once, and 0.06 p. p. m., trout fry in 48 hours; 0.4 p. p. m. killed sunfish and bull- heads; 0.76 p. p. m., hardy minnows; and 1.2 p. p. m., large carp and bullheads. Chloramine-T, CH3C6H4S02NNaCl Water purification systems. Adams (1927), 5 p. p. m. in Nile River water killed cladocerans, Daphnia sp., and copepods, Cyclops sp. Chlorine, CI2 Water purification systems; various chemical wastes. Adams (1927), 2 p. p. m. in Nile River water killed cladocerans, Daphnia sp., and copepods, Cyclops sp.; Davis (1934), 1 p. p. m. killed coarse fish; M. M. E., 1 p. p. m. in hard water killed goldfish, Carassius auratus, in 96 hours; and 0.5 p. p. m. in soft water killed cladocerans, Daphnia magna, in 72 hours or mpre. 422 BULLETIN OF BUREAU OF FISHERIES Chloride of lime, or bleaching powder This substance is of uncertain composition as usually found. Its lethal properties as regards aquatic life are largely dependent upon the amount of chlorine which is liberated when the bleaching powder is added to the water. The potential amount of chlorine available in the sample under question must be known before its toxicity can be estimated. Much confusion exists in the pre- vious reports because of that fact and consequently no limits are given here. Chromic acid, H2Cr04 Chrome tannery wastes. M. M. E., 100 p. p. m. in hard water did not kill goldfish, Carassius auratus, in 100 hours’ exposure; the same amount in very soft water killed goldfish in 30—35 minutes. Citric acid, (COOH)CH2C(OH)(COOH)CH2COOH Wastes from industries using citrus fruits. M. M. E., 894 p. p. m. in hard water killed goldfish, Carassius auratus , in 4-28 hours; 625 p. p. m. in hard water was not lethal to goldfish in 100 hours exposure; 120 p. p. m. in soft water killed cladocerans, Daphnia magna, in 24-72 hours. Cobaltous chloride, CoCl2 Pigment works; chemical industries. M. M. E., 1,000 p. p. m. in hard water killed goldfish, Carassius auratus, in 30-32 hours; 10 p. p. m. in soft water fatal to some goldfish in 168 hours; others survived exposure to this amount for longer periods. Copper, Cu2 See various cupric compounds. As Moore and Kellerman (1905) have pointed out and as is discussed here under the action of heavy metals, the amount of copper required to produce lethal results varies greatly with the water, and particularly the carbonates. These writers state that “in water containing carbonates, if the amount of dissolved C02 is very low the basic carbonate of copper formed may be considered insoluble; if, however, the water should contain a fair amount of C02 it would bring the copper carbonate at least partially into solution.” Copper compounds are used as algicides and occur in many industrial wastes. Cresol, general formula C6H,i(OH) (CH3) A mixture of the various isomeric compounds of this group. The mixture and the individual compounds occur in gas wastes in varying proportions and the mixture is used in various sheep dips and other preparations for the disinfecting of livestock. Both gas wastes and dipping vats have on occasion been the sources of stream pollution by this substance. This mixture and two of the component isomers are discussed below. Cresol Adams (1927), 10 p. p. m. in Nile River water killed cladocerans, Daphnia sp., and copepods, Cyclops sp.; Demyanenko (1931), 17-20 p. p. m. lethal for fishes. Orthocresol Shelford (1917), 55-65 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour; M. M. E., 10-20 p. p. m. in hard water killed goldfish, Carassius auratus, in 3-5 days. Paracresol Shelford (1917), 80-90 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour; Southgate, Pentelow, and Bassindale (1933), 6.2 p. p. m. in tap water caused trout, Salmo irideus, to float helpless on their backs in 1 hour 40 minutes, if the water carried circa 10 p. p. m. of dissolved oxygen and in 13 minutes if the water carried only 3 p. p. m. dissolved oxygen. STREAM POLLUTION 423 Cresylic acid, a mixture of isomeric cresols and xylenols Mine floatation wastes; sheep and cattle dips. M. M. E., 0.1 p. p. m. in hard water killed gold- fish, Carassius auratus, in 5 days and 1 p. p. m. in 6-48 hours; 0.1 p. p. m. in soft water killed cladocerans Daphnia magna, in 72 hours. Cupric chloride, CuCl2 Powers (1917), 0.0188 p. p. m. in distilled water killed goldfish in 3 hours 30 minutes to 7 hours; Carpenter (1927), 672 p. p. m. in distilled water killed the minnow, Leuciscus phoxinus, in 82 minutes. Cupric nitrate, Cu(N03)2 Dilling and Healey (1926), 0.0188 p. p. m. in tap water killed many tadpoles and interfered with the development of those which survived. Cupric sulphate, CuSOi. Blue vitriol Because of the wide use of this compound in both industry and in aquatic investigations, and because of the variation in limits of lethality copper sulphate as given by many writers, owing to differences in water, in carbonate content, and in associated substances, a larger number of refer- ences have been included for copper sulphate than for most compounds in this list. Penny and Adams (1863), 10 p. p. m. fatal to goldfish and minnows, but 5 p. p. m. under the conditions of their tests were not lethal to these fishes; Moore and Kellerman (1905), 0.143 p. p. m. in hatchery water (Cold Spring Harbor, N. Y.), maximum strength tolerated by brook trout, and 0.33 p. p. m. maxi- mum for carp and suckers, 0.4 p. p. m. for catfish, 0.5 p.p. m. for goldfish, 1.33 p.p. m. for sunfish, 2 p. p. m. for black bass; Carpenter (1927), 399 p. p. m. in distilled water killed minnows, Leuciscus phoxinus, in 62 minutes; Catt (1934), 0.5 p. p. m. in lake water not lethal to white perch and yellow perch in 15 hours, but 1 p. p. m. in lake water killed white and yellow perch in 1 to 10 hours; M. M. E., 2 p. p. m. in hard water killed goldfish, Carassius auratus, in 24—96 hours; catfish, Ameiurus nebulosus, in 96-200 hours; 1 p. p. m., some goldfish in 72 hours; 1.25 p. p. m., the amphipods, Gammarus fasciatus and Eucrangonyx gracilis, in 17-20 hours; 10 p. p. m., the isopod, Mancasellus macrourus, in 16-48 hours; 1 p. p. m. in distilled water killed cladocerans, Daphnia magna, in 15 minutes to 2 hours, and 0.25 p. p. m. in distilled water in 30 minutes to 3 hours. Ethyl alcohol, C2H5OH Fermented organic wastes, particularly fruit pulps, brewery and distillery wastes. Weigelt, Saare, and Schwab (1885), a 2-hour exposure to 10,000 p. p. m. in tap water survived by tench, Tinea vulgaris, without injury; Powers (1917), 250 p. p. m. in distilled water killed goldfish in 6-11 hours. Ethyl amine, C2H6NH2 Gas wastes. Shelford (1917). 400-800 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour. Ferric chloride, FeCB Dye industries; some ore milling operations; various chemical wastes. Powers (1917), 9 p. p. m. in distilled water killed goldfish in 20 hours; Carpenter (1927), 270 p. p. m. in distilled water killed minnows, Leuciscus phoxinus, in 90 minutes; M. M. E., 100 p. p. m. in very soft water killed goldfish, Carassius auratus, in 1 hour to 1 hour 30 minutes, but same quantity of ferric chloride in hard water was not detrimental to goldfish in a 96-hour exposure. Ferric potassium sulphate, FeK(SO.s)2. Ferric alum Dye mordant; calico printing. Weigelt, Saare, and Schwab (1885), 1 minute exposure to 10,000 p. p. m. not fatal to tench. Ferric sulphate, Fe2(S04)3 Chemical industries; as a coagulant for sewage precipitation. Clark and Adams (1913), 0.716 p. p. m. in distilled water killed shiners, carp, and suckers in 12-24 hours. 424 BULLETIN OF BUREAU OF FISHERIES Ferrous sulphate, FeS04. Green vitriol or copperas Waters from mines containing pyrites; in pickle liquor from industrial plants cleaning iron plate or wire. Weigelt, Saare, and Schwab (1885), 2,721 p. p. m. in tap water killed trout and California salmon in 31-66 minutes; Clark and Adams (1913), 2.9 p. p. m. in distilled water killed shiners, carp, and suckers in 4-24 hours; Carpenter (1927), 315 p. p. in. in distilled water killed minnows, Leuciscus phoxinus, in about 3 hours; M. M. E., 1,000 p. p. m. in hard water killed goldfish, Car assius auratus, in 2-10 hours, 100 p. p. m. in hard water apparently not harmful to goldfish in a 96-hour exposure. Gallic acid, C6H2(COOH)(QH)3 Dye wastes; tannery wastes; some chemical wastes. Penny and Adams (1863), 143 p. p. m killed goldfish and minnows. Glycerol, (CH,OH)2CHOH. Glycerine Soap factories. Weigelt, Saare, and Schwab (1885), tench survived 16-hour exposure at 8° C. to 100,000 p. p. m. in tap water without apparent effect. Hydrochloric acid, HC1. Muriatic acid In effluents from many chemical processes. Weigelt, Saare, and Schwab (1885), 1,000 p. p. m. in tap water caused trout to overturn helpless in 2-5 minutes; Wells (1915a), 3.6 p. p. m. in distilled water killed green sunfish, Lepomis cyanellus , in 48 hours. Standing Committee on Rivers Pollu- tion (1924), 200 p. p. m. in distilled water produced general collapse in perch and roach; M. M. E., 166 p. p. m. in hard water killed goldfish, Carassius auratus, in 4-7 hours; 157 p. p. m. in hard water apparently did not injure goldfish in over 100 hours’ exposure; 56 p. p. m. in soft water killed clado- cerans, Daphnia magna, in 17—72 hours. Hydrogen sulphide, H2S Produced by decomposition of many types of organic effluents, both municipal and industrial, and occurs in many trade wastes, chemical wastes, and gas wastes. This gas which readily dis- solves in Water is not only harmful in itself but in its decomposition may produce colloidal sulphur, which is also a pollution hazard. Weigelt, Saare, and Schwab (1885), a 3-hour exposure to 100 p. p. m. in tap water was fatal to tench, Tinea vulgaris, 8 days later, and 10 p. p. m. in tap water caused trout to float on back in 15 minutes; Shelford (1917), 4. 9-5. 3 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour; Belding (1929), 0.086 p. p. m. lethal for brook trout, Salvelinus fontinalis, 3.8 p. p. m. for the sucker, Catostomus commersonii, 4.3 p. p. m. for aquarium goldfish, Carassius auratus, and 6.3 p. p. m. for carp, Cyprinus carpio; M. M. E., 10 p. p. m. in hard water killed goldfish, Carassius auratus, in 96 hours or less, 5 p. p. m. killed some goldfish in 200 hours, and 1 p. p. m. in soft water killed cladocerans, Daphnia magna, in 72 hours or less. Iodine, I2 Rarely found free in stream water, and goldfish. Penny and Adams (1863), 28.5 p. p. m. killed minnows Iron, Fe2 See various ferric and ferrous compounds. Lactic acid, CH3CH(OH)COOH A component of various dairy industries wastes. M. M. E., 654 p. p. m., in hard water killed goldfish, Carassius auratus, in 6-43 hours, but 430 p. p. m. in hard water apparently was not harmful to goldfish in exposures of over 100 hours; 170 p. p. m. in soft water killed cladocerans, Dapghnia magna, in 26-72 hours. Lead, Pb2 Compounds of this metal are found in many industrial wastes and in effluents from various mining and ore milling operations. These compounds are highly toxic as has been discussed under STREAM POLLUTION 425 action of heavy metals on fishes and other aquatic organisms. Two general statements concerning the lethality of lead may be given here. Other references follow under the several salts of lead. Carpenter (1927) believes that as little as 0.33 p. p. m. of lead may be lethal to fresh-water fishes, and Billing, Healey, and Smith (1926) found that 4 p. p. m. of colloidal lead retarded the growth of young plaice. Lead acetate, Pb(C2H302)2. Sugar of lead Rushton (1922), 10 p. p. m. in stream water killed yearling trout; Carpenter (1925), 5 p. p. m. in distilled water killed minnows, Leuciscus phoxinus, in 4—16 hours, and 10 p. p. m. in distilled water, if renewed every other day, killed goldfish in 12 days. Lead nitrate, Pb(N03)2 Rushton (1922), 10 p. p. m. in stream water killed trout in 2 hours 15 minutes; Carpenter (1925), 250 p. p. m. in distilled water killed goldfish in 4-5 days and minnows, Leucisus phoxinus, in 2-3 hours; Dilling and Healey (1926), 1.6 p. p. m. retards growth of tadpoles and 3.3 p. p. m. lethal for tadpoles in tap water; Carpenter (1930), 165 p. p. m. in distilled water, if given sufficient exposure will kill steel-colored minnow, Notropis whipplii; common shiner, Notropis cornutus; blunt-nosed minnow, Hyborhynchus notatus; silver-mouthed minnow, Ericymba buccata; sucker-mouthed minnow, Phenacobius mirabilis; creek chub, Semotilus atromaculatus ; stoneroller, Campostoma anomalum; common sucker, Catostomus commersonii; chub sucker, Eriomyzon sucetta; Johnny darter, Boleosoma nigrum; fan-tailed darter, Etheostoma jlabellare; log perch, Percina caprodes; and bluegill, Lepomis pallidus; M. M. E., 100 p. p. m. in hard water fatal to goldfish, Carassius auratus, in 80 hours; 10 p. p. m. in hard water, without apparent injury to goldfish, in 96 hours’ exposure. Lead sulphate, PbSCh Carpenter (1925), 25 p. p. m. in distilled water killed goldfish, Carassius auratus, in 4 days, and minnows, Leuciscus phoxinus, in 2—3 hours. Lithium chloride, LiCl Found in some mineral springs. Powers (1917), 3,750 p. p. m. in distilled water killed goldfish in 22-27 hours. Magnesium chloride, MgCl2 A component of various waste waters from oil wells, and some industrial wastes. Garrey (1916), 476 p. p. m. in distilled water killed straw-colored minnow, Notropis blennius, in 4-6 days; Powers (1917), 6,757 p. p. m. in distilled water killed goldfish in from 78 hours to 21 days; Wiebe, Burr, and Faubion (1934), 5,000 p. p. m. in distilled water killed golden shiner, Notemigonus crysoleucas, in 96 hours. Magnesium nitrate, Mg(N03)2 Powers (1917), 12,500 p. p. m. killed goldfish in 14-16 hours. Mercuric chloride, HgCl2. Corrosive sublimate Weigelt, Saare, and Schwab (1885), 500 p. p. m. in tap water killed large trout in 54 minutes; Carpenter (1927), 13.6 p. p. m. in distilled water killed minnow, Leuciscus phoxinus, in 42 minutes. Methyl alcohol, CH3OH A solvent in many industrial operations. Weigelt, Saare, and Schwab (1885), a 2-hour exposure to 10,000 p. p. m. in tap water was tolerated by trout without apparent injury; Powers (1917), 250 p. p. m. in distilled water killed goldfish in 1 1-15 hours. Methyl mercaptan, CH3SH A highly toxic constituent of sulphite paper pulp waste. Cole (1935a), 1 p. p. m. in lake water killed white bass, Roccus chrysops, in 1 hour, 45 minutes, or less; yellow perch ( Perea jlavescens), largemouth black bass ( Micropterus salmoides), smallmouth black bass ( Micropterus dolomieu), and bluegill ( Lepomis pallidus), in 6-8 hours; and rock bass ( Ambloplites rupestris) , in 11 hours. 426 BULLETIN OF BUREAU OF FISHERIES Muriatic acid See hydrochloric acid. Naphthenic acids and naphthene derivatives These compounds are listed among the solutes which may occur in refinery wastes (American Petroleum Institute, 1935), and Kupzis (1902) believes the naphthenic compounds to be among the most toxic substances to fish, passing from crude oil into water. Kupzis (1. c.), using fractions containing the naphthenic acids extracted from various crude oils, gives the following lethality findings for this naphthenic acid fraction (which of course must not be considered as representing a single pure compound in this case). In tap water (hard) 3 p. p. m. killed percoid fishes, Acerina cernua, in 6-12 hours; 5 p. p. m. killed pickerel, Esox lucius, 36-48 hours; the minnow, Abramis brama, 72 hours; the red-eyed minnow, Scardinius erythrophthalmiis, 26 hours; and the perch, Perea fluviatilis, in 16-23 hours; 20 p. p. m. killed carp, Cyprinus carpio, in 26-36 hours, and goldfish, Carassius auratus, in 8-16 hours. Several other European species tested are not listed here. Naphthalene, CioHg Aniline and coal-tar industries. Demyanenko (1931), 10 p. p. m. lethal to fish. Nickelous chloride, NiCh Electroplating wastes; various industrial wastes. Thomas (1924), 8.1 p. p. m. in tap water killed the top-minnow, Fundulus heteroclitus, in a few hours. This species, which can also live in salt water, tolerated 259.2 p. p. m. when in salt water without apparent injury during a 2-week exposure. M. M. E., 100 p. p. m. in very soft water killed goldfish, Carassius auratus, in 19-50 hours, and 10 p. p. m. in very soft water killed in 200-210 hours. Nitric acid, HNOj Occurs in many wastes from chemical industries. It is easily broken into water and oxides of nitrogen. Weigelt, Saare, and Schwab (1885), trout after 34 minutes in 1,000 p. p. m. in tap water were helpless; Carpenter (1927), sufficient quantities to bring the water to pH 4.4 killed the minnow, Leuciscus phoxinus, in 7 hours, while quantity sufficient to bring the water to pH 5.2 was without apparent effect on this species of fish; M. M. E., 750 p. p. m. in hard water (see section on acids) killed goldfish, Carassius auratus, in 30 minutes to 1 hour, and 200 p. p. m. in hard water were without apparent effect on goldfish in exposures of over 100 hours. Octyl alcohols, general formula CsHnOH Macht and Leach (1930), 66.7 p. p. m. of primary oetynol produce respiratory and neuromus- cular paralysis in goldfish, Carassius auratus, in 15 minutes. The specific toxicity for fish of the individual octyl alcohols varied in the series of 23 which these observers studied, secondary octylie heptanol being among the least toxic, 200 p. p. m. producing paralysis in goldfish in 4 hours. Oxalic acid, H2C2O4 Bleaching, dying, and various chemical industries. M. M. E., 1,000 p. p. m. in hard water killed goldfish, Carassius auratus, in 25-30 minutes; 200 p. p. m. in hard water produced no apparent injury during exposures of 100 hours. This acid is readily precipitated out of waters by calcium salts. Oxygen, O2, and Ozone, O3 Although the usual problem in stream pollution is to secure enough oxygen, some of the proposed treatments for organic wastes involve high oxygenation or ozonation. It seems worth while there- fore to list here some tests on the effects of high oxygen and nascent oxygen. Wiebe (1933) found changes from 5.7 p. p. m. to 40.3 p. p. m. dissolved oxygen were tolerated by largemouth black bass, Micropterus salmoides, smallmouth black bass, Micropterus dolomieu, white crappie, Pomoxis sparoides, bluegill, Lepomis incisor, orange-spotted sunfish, Lepomis humilis, golden shiner, Notemi- gonus crysoleucas, and goldfish, Carassius auratus, for at least 24 hours. Hubbs (1930) states that minnows may be killed by 0.033 p. p. m. of nascent oxygen in the water and that fishes are irritated by ozone in quantities less than 0.01 p. p. m. Para-dichlorobenzene, C6H4CI2 Demyanenko (1931), 50 p. p. m. lethal to fishes. STREAM POLLUTION 427 Phenol, C0H5OH. Carbolic acid An important constituent of gas wastes, many chemical effluents, and even some sewage products. Phenol may also come into streams from sheep dips and other establishments where livestock are disinfected. Shelford (1917), 70-75 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour; Powers (1917), 51 p. p. m. in distilled water killed goldfish in 1 hour 30 minutes to 2 hours 20 minutes; Demyanenko (1931), 16.6-20 p. p. m. lethal to fishes, some can live in 15, p. p. in., but their flesh acquires phenolic smell; Alexander, Southgate, and Bassindale (1935), 0. 4-0.6 p. p. m. caused trout to overturn in 8 hours 20 minutes; Adams (1927), 10 p. p. m. in Nile River water killed cladocerans, Daphina sp., and copepods, Cyclops sp.; M. M. E., 10 p. p. m. in hard water killed goldfish, Carassius auralus, in 72 hours or less, 1 p. p. m. apparently not injurious to goldfish in exposures of 100 hours; 8 p. p. m. in soft water killed cladocerans, Daphnia magna. Potassium bicarbonate, KHCO3 Penny and Adams (1863), 2,000 p. p. m. killed minnows and goldfish. Potassium chloride, KC1 Garrey (1916), 373 p. p. m. in distilled water killed straw-colored minnows, Notropis blennius, in 12-29 hours; Powers (1917), 74.6 p. p. m. in distilled water lethal for goldfish in 4 hours 40 minutes to 15 hours. Potassium cyanide, KCN Ore milling operations; chemical works; electroplating; and in effluents from coke ovens. (See Tupholme, 1933.) Powers (1917), 0.78 p. p. m. in distilled water killed goldfish in 43-118 hours; Calatroni (1928), 15 p. p. m. in tap water immobilized tadpoles with fatal results; McArthur and Baillie (1929), 65 p. p. m. killed cladocerans, Daphnia magna; Southgate, Pentelow, and Bassin- dale (1933), 0.27 p. p. m. KCN (equal to 0.11 p. p. m. CN) at temperature of 7°-9° C. caused trout to float helpless on back in about 2 hours if the dissolved oxygen in the water were circa 11 p. p. m.t but in 10 minutes if the water carried only 3 p. p. m. dissolved oxygen; Alexander, Southgate, and Bassindale (1935) find 0.5 p. p. m. KCN (0.2 p. p. m. CN) will cause trout to overturn and become helpless in 15 minutes; M. M. E., 1 p. p. m. in river water caused glochidia of fresh-water mussels to close, rendering them incapable of attachment on fishes, and 0.1 -0.3 p. p. m., in hard water killed goldfish in 3-4 days. Potassium dichromate, Ix^C^O; Ore floatation processes; chemical industries. M. M. E., 100 p. p. m. in hard water apparently not harmful to goldfish in exposures of 108 hours, 500 p. p. m. lethal to goldfish in 3 days. Potassium ferricyanide, K3Fe(CN)6- Red prussiate of potash Penny and Adams (1863), 2,000 p. p. m. not lethal to minnows and goldfish. Potassium ferrocyanide, K4Fe(CN)6. Yellow prussiate of potash Penny and Adams (1863), 2,000 p. p. m. not lethal to minnows and goldfish; Weigelt, Saare, and Schwab (1885), trout survived 1 hour exposure to 8,723 p. p. m. in tap water without symptoms. Potassium hydroxide, KOH. Caustic potash Soap works; from some types of ashes. Wells (1915), 56 p. p. m. in distilled water killed bluegills, Lepomis pallidus, in 4 hours 25 minutes, 28 p. p. m. apparently not harmful to bluegills in 10-day exposures. Potassium nitrate, KNO3. Saltpeter Wells (1915b), 1,203 p. p. m. in tap water killed bluegills, Lepomis pallidus, in 15 days. Potassium permanganate, KM11O4 This powerful oxidizing agent is frequently used to disinfect hatchery tanks and can be applied with caution to fish themselves. Adams (1927), 5 p. p. m. in Nile River water killed cladocerans, Daphnia sp., and copepods, Cyclops sp.; M. M. E., 10 p. p. m. in hard water killed goldfish, Caras- sius auralus, in 12-18 hours. 428 BULLETIN OF BUREAU OF FISHERIES Potassium sulphate, K2S04 Wells (1915b), 869 p. p. m. in tap water killed bluegills, Lepomis pallidus, in 4 days. Potassium xanthate or potassium ethyldithiocarbonate C2H5-OCS-SK Used in ore floatation processes and on occasion to free soil from insect pests, and it is highly toxic to fishes. M. M. E., 10 p. p. m. in hard water killed goldfish in 48-96 hours, 0.1 p. p. m. in 4-5 days. Pyrethrum The volatile oil in this plant product is quite toxic to fishes. Bandt (1933) 5-10 p. p. m. toxic for carp. Pyridine, CsHjN In gas wastes, also in waters draining recently burned-over areas where combustion near ground has been incomplete. Shelford (1917), 1,477-1,576 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour; Powers (1917), 1,869 p. p. m. in distilled water killed goldfish in 10-30 hours; Demyanenko (1931) states 1,000 p. p. m. to have a feeble effect on fishes. Quinoline, C6H4N : CHCH : CH Gas wastes. Shelford (1917), 52-56 p. p. m. and 65 p. p. m. isoquinoline in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour. Saponin, a glucoside Found in various plants. Ebeling (1928) points out that wastes from potato starch factories carry sufficient saponin to be dangerous to fishes, and that consequently these wastes should be highly diluted. M. M. E., 10 p. p. m. of saponin in hard water produced marked distress in gold- fish in 5 hours, 100 p. p. m. rapidly fatal, killing goldfish in 7-24 hours. Sodium arsenite, commercial preparation usually a mixture of several sodium and arsenic compounds Some dye and tanning processes; removal of aquatic vegetation. Since arsenic trioxide is usually dissolved in sodium hydroxide, and since each lot of commercial sodium arsenite must be assayed for its arsenic content, arsenic, arsenic trioxide, and sodium arsenite are considered under this single heading. Wiebe (1930), 7 p. p. m. of As2C>3 in Mississippi River water was not detri- mental to largemouth black bass, Micropterus salmoides, smallmouth black bass, Micropterus dolomieu, white crappie, Pomoxis sparoides, bluegill, Lepomis pallidus, golden shiner, Notemigonus crysoleucas, bullhead, Amieurus nebulosus, and goldfish, Carassius auratus, in 148-hour exposures as shown by 3 months subsequent observation. Surber and Meehan (1931), 2 p. p. m. As2C>3 in Mississippi River water survived by the important fish food organisms, but 2.5-4 p. p. m. killed chironomid larvae, mayfly nymphs, the fresh- water shrimp, Hyalella, and odonata nymphs; the isopod Asellus sp., survived 10-21 p. p. m.; M. M. E., 1.3 p. p. m. As203 in distilled water killed Daphna magna, 8 p. p. m. hard water had no appreciable effect on the fresh-water mussel, Amblema peruviana, but 16 p. p. m. in hard water was fatal to Amblema peruviana in 3-16 days. Sodium carbonate, Na2C03 Found in many chemical effluents. Clark and Adams (1913), 250-300 p. p. m. killed shiners, carp, and large suckers in tap water in a few hours; Wells (1915b), 530 p. p. m. in tap water killed bluegills, Lepomis pallidus, in 3 days. Sodium chlorate, NaClCb In some effluents from chlorine and bromine works. Not markedly toxic. M. M. E., 1,000 p. p. m. in hard water produced no apparent injuries in goldfish during exposures of 5 days duration. STREAM POLLUTION 429 Sodium chloride, NaCl. Common salt Brine works; waste waters from oil wells; effluents from some dairy industries. Garrey (1916), 2,500 p. p. m. in distilled water killed straw-colored minnow, Notropis blennius, in 9-24 days (solu- tion, a = 0.105); Powers (1917), 11,765 p. p. m. in distilled water killed goldfish in 17 hours; Wiebe, Burr, and Faubion (1934), 5,000 p. p. m. in distilled water killed golden shiners, Notemigonus cryso- leucas, in 148 hours and largemouth black bass in 200-250 hours; M. M. E., 1 p. p. m. distilled, water killed cladoceran, Daphnia magna, in 3 hours, 5,000 p. p. m. in Mississippi River water appar- ently not harmful to goldfish in 25-day exposure, but 10,000 p. p. m. in Mississippi River water killed goldfish in 4—10 days. Sodium fluoride, NaF In certain brewery and distillery wastes; also found in some soils. M. M. E., 1,000 p. p. m. in hard water killed goldfish in 60-102 hours. Sodium hydroxide, NaOH. Caustic soda Soap factories; wood ashes. Clark and Adams (1913), 96 p. p. m. in tap water killed shiners, carp, and suckers in 2-10 minutes; Standing Committee on River Pollution (1924), 50 p. p. m. in distilled water was not fatal to perch and roach in a 2-hour exposure. Sodium nitrate, NaNCh. Chili saltpeter Fertilizers. Powers (1917), 1,282 p. p. m. in distilled water killed goldfish in 14 hours; M. M. E., 4,000 p. p. m. in hard water killed goldfish in 80 hours but 3,000 p. p. m. apparently did not injure goldfish in 100-hour exposures. Sodium selenite, Na2Se03 Component of certain soils; pottery works. M. M. E., 100 p. p. m. in hard water killed gold- fish in 8-20 hours, 10 p. p. m. in hard water killed in 98-144 hours. Sodium sulphate, Na2SO<. Glaubers salt Harukawa (1922), 500 p. p. m. in tap water not injurious to goldfish in 24 hours. Sodium sulphide Na2S In some beet-sugar factory effluents and in some paper-pulp wastes. Weigelt. Saare, and Schwab (1885), a 1-hour exposure to 1,150 p. p. m. in tap water kill tench, Tinea vulgaris, 6 days later. Sodium sulphite, Na2SOs Component of various paper pulp wastes anefsome beet sugar factory wastes. M. M. E., 100 p. p. m. in hard water killed goldfish in 96 hours. Stannous chloride, SnCl2 M. M. E., 1,000 p. p. m. in hard water killed goldfish in 4^5 hours. Strontium chloride, SrCl2 Powers (1917), 15,384 p. p. in. in distilled water killed goldfish in 17-31 hours. Strontium nitrate, Sr(NC>3)2 Powers (1917), 9,615 p. p. m. in distilled water killed goldfish in 32-146 hours. Sugar, cane, Ci2H220n. Sucrose Garrey (1916), 34—218 p. p. m. in distilled water (A = 0.198) killed straw-colored minnow, Notropis blennius, in 24-48 hours. Sulphur, colloidal, S2 Lime-sulphur mixtures and dips; following decomposition of hydrogen sulphide in water. Harukawa (1922), 1,600 p. p. m. in tap water fatal to goldfish in 3 hours 30 minutes to 5 hours 15 minutes, and 2,100 p. p. m. in 48-71 minutes. 430 BULLETIN OF BUREAU OF FISHERIES Sulphur dioxide, SO2 Bleaching works; various chemical industries. Weigelt, Sarre, and Schwab (1885), 10 p. p. m. in tap water caused trout to float helpless in 10 minutes; Shelford (1917), 16-19 p. p.m. in tapwater killed orange-spotted sunfish, Lepomis humilis, in 1 hour. Sulphuric acid, H2SO4. Oil of vitriol Pickle liquor from sheet metal and wire factories; waters from coal and iron mines; various chemical wastes. Wells (1915), 7.36 p. p. m., in distilled water killed bluegills, Lepomis paUidus, in 60 hours, but 3.68 p. p. m. apparently harmless over period of 1 month. M. M. E., 59 p. p. m. in very soft water killed goldfish, Carassius auratus, in 1 hour to 1 hour 15 minutes, 138 p. p. m. in hard water in 4 hours, 100 p. p. m. in hard water apparently not injurious to goldfish in 100-hour exposures; 29 p. p. m. in soft water killed cladocerans, Daphnia magna, in 24-72 hours. Tannic acid, C14H10O9 Tannery wastes; teachings from some barks and sawmill wastes. M. M. E., 100 p. p. m. in hard water killed goldfish in 9-20 hours, but 10 p. p. m. apparently not injurious in 100-hour ex- posures. Tartaric acid, COOH (CHOH)2COOH Dye wastes; mordant liquors; leather works; various chemical effluents; some winery wastes. M. M. E., 1,000 p. p. m. in hard water killed goldfish in 3-4 hours, 200 p. p. m. in hard water not injurious apparently in exposures of 100 hours, 100 p. p. m. in very soft water killed goldfish in 3 hours to 3 hours 30 minutes. Thiophene, C4H4S Gas wastes. Shelford (1917), 27 p. p. m. in tap water killed orange-spotted sunfish, Lepomis humilis, in 1 hour. Tin, Sn2 See stannous chloride. Trade wastes At end of this section. Zinc nitrate, Zn(NOs)2 Dilling and Healey (1926) found that tadpoles survived 3-month exposure to 1.89 p. p. m. but failed to develop limb buds, that 5.7 p. p. m. killed most tadpoles and that 94.7 p. p. m. killed tadpoles quickly. Zinc sulphate, ZnSCh. White vitriol Wastes from electrolytic refineries of zinc; incrustations developing from exposed zinc sulphide ores; mine tailings; several chemical effluents. Carpenter (1927), 404 p. p. m. in distilled water killed minnows, Leuciscus phoxinus, in 3 hours 20 minutes; M. M. E., 1,000 p. p. m. in hard water killed goldfish in 1-4 hours, 100 p. p. m. fatal to many goldfish in 5 days. Trade wastes Lethal limits which will be inclusive cannot be defined for sewage and trade wastes because sewage and even such specific effluents as Steffens house waste (from certain types of beet-sugar refineries) are not constant either in composition or concentration. The effects of colloidal sulphur or ferric chloride on fishes and other aquatic animals can be determined under a variety of condi- tions which can be duplicated, or the toxicity of any particular sample of Steffens waste, or pickle liquor from tin-plate mills, or deazoting fluid from dye works can also be ascertained with accuracy, but to attempt to define the minimal lethal concentration or the maximal dilution for all pickle liquors from samples obtained at any plant or series of plants is both unsound scientifically and unfair to the industry. Already much confusion, often to the detriment of fisheries interests, has resulted from the misuse in this connection of various statements in the literature concerning the tolerance of fishes for a given number of p. p. m. of a particular waste, when these observations were intended to apply to a specific case. STREAM POLLUTION 431 In practical tests the writer has found that wastes taken directly from their sources in the plant, i. e., not from sewers or flumes, differ widely in toxicity to aquatic life, at different plants using the same general process. Several factors contributed to these differences. For example, the process of tanning hides with chromium compounds is basically the same in all establishments using this process, but as is well known the plant chemists in the interests of economy and efficiency frequently make minor or at times even drastic changes in the process, with, of course, resultant changes in the effluents. Variations in raw material, emergencies in plant operation, and many other conditions also call for changes in procedure which result in changes in effluent composition and concentration. There is, therefore, no standard hide vat liquor waste or pickle liquor or winery effluent, even when taken at the source, which may be used with either fairness or accuracy in the estimation of the probable hazards of another plant engaged in the same work. Besides, as the waste moves away from its source through conduits and flumes, these differences become greater owing to the addition of cooling water, wash waters, and other effluents from differ- ent parts of the plant; so that the final mixture as poured into the stream varies not only in concen- tration, but also in composition as the result of the sequence in which the different effluents were mixed. However, from the studies of various wastes before mixing, the general changes which these effluents will produce in the stream complex and the types of their actions on aquatic life can be ascertained with considerable accuracy. In table 14 the usual effects of 29 common industrial wastes and of municipal sewage on the aquatic environment and the critical features of these actions as shown by a large series of field studies have been pointed out. Table 14. — Usual fisheries hazards of 30 common types of municipal and industrial effluents 432 BULLETIN OF BUREAU OF FISHERIES O fl 2 a.o^ © 2 Cl,* CO ■«o . © . . . 07: OOO Q-QQQ 03 © II 0 .2 ^2 .2 £ o3 O OOOO •2fi QQQQ OflnOdn O 11 * « © © J O O 0,0,0 •2 T3 TJ ■© *jg *55 g ’E ; ; ; o o © o : ; iphPh^: OOh 03 © 22 a $ © o Ph .y ®*3 © I fl 03 lll-S’S 0S0 ;ph •3 isa „ 8 fes «£ ’ti C m'S'l D m 'C O o® O O o oZt^g Zpk OO - _W ' rrt "rt © •-> © 1 1 fl u. H ! ItSo0® 1 i^Ph ^ <3 05 £€ O © O ' o o o ^ PhS ,.2 ca yZ °Hr- Z c^g ©jO ^ „ ! 0 :ph^Ph ; 0 . ® O •*- ©^3 ^ §2^ 1 U « O uL :z an S W'H « ® © Sis S.2g QT3 £ = •§ 2 Si Zg o oS Zdng ®« 5 S OT'O o §S f>» Eh ^ 5 a ® ^ ^2 m fl w^'Orn w g2°c§gg| -OT ^ © 2^'° ^ o a-Q.s g sy wj^Soom t-i oi co »d <6 3 £< * * g E-3 573 ® to g o p r-1 Z r-4 03 03 c3 Td 5P a fl be m.2 Sr* fl.a 9 g ^ c3 Cd t-. PhOPC OOOOhN CO *0 gQ SOS'S to g Ss fl a.H ® a j. H « n g S ® g p p.m « Qco^Im CO 00 01 O »— • (M aj rv co ^ fla n tr a-z ©^ w a I s §•§ t 0 S a c3 © © o c3 wooe « & o'O 3 9 o' a 3J ® «S ® a S a SP S o 1 In this table increases in both acidity and alkalinity are noted in some cases, due to the facts that two or more kinds of effluents are mixed, with one predominating at times and to changes which take place in the stream after the effluent is added. STREAM POLLUTION 433 From this table the general effects of a given effluent or comparable waste on the aquatic complex can be predicted. By field tests at the site of pollution covering the critical factors as indicated for the particular effluent (dissolved oxygen, pH, conductivity, carbonates, and other determinations for the polluted stream as the case may be), the degree of pollution can be estimated when these field data are com- pared with the standards given in the section on stream pollutants and aquatic en- vironment. When specific toxic action is indicated (last section, table 14) the limits for the expected detrimental substances can be found in the section on lethal limits. ACKNOWLEDGMENTS The writer is particularly indebted to Dr. H. L. Motley, assistant professor of physiology, University of Missouri; to Dr. W. A. Chipman, assistant aquatic biolo- gist, United States Bureau of Fisheries; to Mrs. M. D. Ellis, honorary fellow in physiology, University of Missouri; to Messrs. B. A. Westfall, W. G. Davis, Paul Pierce, R. O. Jones; to Mrs. Helene Pierce Chipman; and to Misses Cornealia G. Ellis, Zella von Gremp, Nelden Bickel, and Mary McConathy, bioassayists and chemists, who through their cooperative efforts as members of the staff of the Colum- bia, Mo., field unit, United States Bureau of Fisheries, have carried forward during the past 4 years many analyses, assays, and experiments required to give the mass of data for these studies; to the officers and employees of the Corps of Engineers, United States Army, in various river districts, who have aided these investigations in moving U. S. Quarterboat 848, and in supplying freely river and survey data from their own work; to the many industrial concerns which have cordially opened their plants and given valuable assistance in the collection of trade waste samples; and to the University of Missouri for continued cooperation in the maintenance of the Co- lumbia, Mo., laboratories. BIBLIOGRAPHY Adams, B. A. 1927. The lethal effect of various chemicals on Cyclops and Daphnia. Water Works Eng., vol. 29, pp. 361-364. New York. Alexander, W. B., B. A. Southgate, and R. Bassindale. 1935. The survey of the River Tees. Part II, The estuary. Chemical and Biological. Technical Paper No. 5, Water Pollution Res. Bd., England. London. American Petroleum Institute. 1935. Disposal of refinery wastes. Section III. Waste water containing solutes. Official Publication, 37 pp. New York. American Public Health Association. 1933. Standard methods for the examination of water and sewage. Ed. 7, 180 pp. New York. Bandt, H. 1933. Fish poisoning caused by pyrethrum. Wasser und Abwasser, Bd. 31, s. 255. fide. Chem. Absts., vol. 27, part 3, p. 5826, 1933. Easton, Pa. Behrens, Behrend. 1925. Untersuchungen fiber Aufnahme, Ausscheidung und Verteilung kleinster Bleimengen. Archiv ffir experimentelle Pathologie und Pharmakologie, Bd. 109, s. 332-357. Berlin. Belding, D. L. 1928. Toxicity experiments with fish in reference to trade waste pollution. Trans. Amer. Fish. Soc., vol. 57, 1927 (1928), pp. 100-114. Washington. Belding, D. L. 1929. The respiratory movements of fish as an indicator of toxic environment. Trans., Amer. Fish. Soc., vol. 59, pp. 238-245. Hartford, Conn. Billiard, G. 1925. La daphnie ( Daphnia pulex) consideree comme reactif biologique de la toxicite des liquides humoraux. Comptes Rendus Society de Biologie, Tome 92, pp. 1352 1354. Birge, E. A., and C. Juday. 1911. The inland lakes of Wisconsin. The dissolved gases of the water and their biological significance. Bull., Wisconsin Geol. and Nat. Hist. Sur., no. 22, pp. 1-259. Madison, Wis. 434 BULLETIN OF BUREAU OF FISHERIES Bond, R. M. 1933. A contribution to the study of the natural food-cycle in aquatic environ- ments. Bull., Bingham Oceanographic Coll., vol. 4, pp. 1-89. Brown, H. W., and M. E. Jewell. 1926. Further studies on the fishes of an acid lake. Trans., Amer. Micro. Soc., vol. 45, pp. 20-34. Menasha, Wis. Butcher, R. W., F. T. K. Pentelow, and J. W. A. Woodley. 1927. The diurnal variation of the gaseous constituents of river waters. Parts I and II. Biochem. Jour., vol. 21, pp. 945-957, and 1423-1435. Cambridge, England. Calatroni, R. 1928. Action du cyanure de potassium et de l’hyposulphite de sodium sfir les tetards. Comptes Rendus Societc de Biologie, Tome 99, pp. 2007. Paris. Carpenter, K. E. 1924. A study of the fauna of rivers polluted by lead mining in the Aberys- twyth district of Cardiganshire. Ann., Appl. Biol., vol. 11, pp. 1-23. Cambridge, England. Carpenter, K. E. 1925. On the biological factors involved in the destruction of river-fisheries by pollution due to lead-mining. Ann., Appl. Biol., vol. 12, pp. 1-13. Cambridge, England. Carpenter, K. E. 1926. The lead mine as an active agent in river pollution. Ann., Appl. Biol., vol. 13, pp. 395-401. Cambridge, England. Carpenter, K. E. 1927. The lethal action of soluble metallic salts on fishes. British Jour., Exper. Biol., vol. 4, pp. 378-390. Cambridge, England. Carpenter, K. E. 1930. Further researches on the action of metallic salts on fishes. Jour., Exper. Zool., vol. 56, pp. 407-422. Philadelphia. Catt, J. 1934. Copper sulphate in the elimination of coarse fish. Trans., Amer. Fish. Soc., vol. 64, pp. 276-279. Hartford, Conn. Chipman, W. A. 1934. A new culture medium for cladocerans. Science, vol. 79, pp. 59-60. New York. Clark, H. W., and G. O. Adams. 1913. Studies of fish life and water pollution. Forty-fourth Annual Report of the State Board of Health of Massachusetts, 1912 (1913), Pub. Doc. No. 34, Massachusetts State Board of Health, vol. 6, pp. 336-345. Boston. Cole, A. E. 1935a. The toxicity of methyl mercaptan for freshwater fish. Jour. Pharm. and Exper. Ther., vol. 54, pp. 448-453. Baltimore. Cole, A. E. 1935b. Water pollution studies in Wisconsin. Effects of industrial (pulp and paper mill) wastes on fish. Sewage Works Jour., vol. 7, pp. 280-302. Coventry, F. L., V. E. Shelford, and L. F. Miller. 1935. The conditioning of a chloramine treated water supply for biological purposes. Ecology, vol. 16, pp. 60-66. Brooklyn, N. Y. Creaser, C. W. 1930. Relative importance of hydrogen-ion concentration, temperature, dis- solved oxygen, and carbon dioxide tension, on habitat selection by brook trout. Ecology, vol. 11, pp. 246-262. Lancaster, Pa. Davis, H. W. 1934. Discussion. Trans. Amer. Fish. Soc., vol. 64, pp. 280. Hartford, Conn. Demyanenko, V. 1931. Poisoning of fish by waste waters from chemical factories and the fish test. (In Russian). Fide. Chem. Abst., vol. 27, part 2, p. 2746, 1933. Easton, Pa. Dilling, W. J., and C. W. Healey. 1926. Influence of lead and the metallic ions of copper, zinc, thorium, beryllium, and thallium on the germination of frogs’ spawn on the growth of tadpoles. Ann., Appl. Biol., vol. 13, pp. 177-188. Cambridge, England. Dilling, W. J., C. W. Healey, and W. C. Smith. 1926. Experiments on the effects of lead on the growth of plaice ( Pleuronectes platessa). Ann., Appl. Biol., vol. 13, pp. 168-176. Cambridge, England. Domogalla, B. P., C. Juday, and W. H. Peterson. 1925. Forms of nitrogen found in certain lake waters. Jour., Biol. Chem., vol. 63, pp. 269-285. Baltimore. Ebeling, G. 1928. Mittheilungen des Fisherei-Vereins, Bd. 32, s. 521-522. Berlin. Fide. Chem. Abs., vol. 23, p. 3042, 1929. Easton, Pa. Ellis, M. M. 1919. Respiratory volumes of men during short exposures to constant low oxygen tensions attained by rebreathing. Amer. Jour. Phys., vol. 50, pp. 267-279. Baltimore. Ellis, M. M. 1932. Pollution of the Coeur d’Alene River and adjacent waters by the mine wastes. Manuscript report to the Commissioner, U. S. Bur. Fish., 132 pp. Washington. Ellis, M. M. 1934. A photoelectric apparatus for turbidity and light penetration measurement. Science, vol. 80, pp. 37-38. New York. STREAM POLLUTION 435 Ellis, M. M. 1935a. Water purity standards for fresh water fishes. Special Report, U. S. Bur. Fish., pp. 1-14. Washington. Ellis, M. M. 1935b. A study of the Mississippi River from Chain of Rocks, St. Louis, to Cairo, 111., with special reference to the proposed introduction of ground garbage into the river by the city of St. Louis. Manuscript report to the Commissioner, U. S. Bur. Fish., 37 pp. Washington. Ellis, M. M. 1936a. Erosion silt as a factor in aquatic environments. Ecology, vol. 17, pp. 29-42. Lancaster, Pa. Ellis, M. M. 1936b. The effects of pollution on fish. In press. Proceedings of North American Wild Life Conference, February, 1936. Washington. Ellis, M. M., and W. A. Chipman. 1936. Studies of the toxicity of ammonium compounds for aquatic animals. In MS. Engineering. 1935. The sewage treatment system of the city of Munich. Engineering, vol. 140, pp. 679-682. London. Gardiner, J. A., and G. King. 1922. Respiratory exchange in fresh-water fish. Part IV. Further comparison of goldfish and trout. Biochem., Jour., vol. 16, pp. 729-735. Cambridge, England. Gardiner, J. A., G. King, and E. B. Powers. 1922. The respiratory exchange in fresh-water fish. Part III. Goldfish. Biochem. Jour., vol. 16, pp. 523-529. Cambridge, England. Garrey, W. C. 1916. The resistance of fresh water fish to changes of osmotic and chemical con- ditions. Amer. Jour. Phys., vol. 39, pp. 313-329. Baltimore. Gillespie, L. J. 1920. Colorimetric determination of hydrogen-ion concentration without buffer mixtures, with especial reference to soils. Soil Science, vol. 9, pp. 115-136. Baltimore. Harukawa, C. 1922. Preliminary report on toxicity of colloidal sulphur to fish. Trans., Amer. Fish. Soc., vol. 52, pp. 219-224. Washington. Hodgman, C. D. 1935. Handbook of Chemistry and Physics. Ed. 20, 1951 pp. Chemical Rubber Publishing Co. Cleveland, Ohio. Hubbs, C. L. 1930. The high toxicity of nascent oxygen. Phys. Zool., vol. 3, pp. 441-460. Chicago. Jtjday, C., E. B. Fred, and F. C. Wilson. 1924. The hydrogen-ion concentration of certain Wisconsin lakes waters. Trans. Amer. Micro. Soc., vol. 43, pp. 177-190. Menasha, Wis. Kemmerer, G., J. F. Bovard, and W. R. Boorman. 1923. Northwestern lakes of the United States: Biological and chemical studies with reference to possibilities in production of fish. Bull., U. S. Bur. Fish., vol. 39, 1923-24, pp. 51-140. Washington. Keys, A. B. 1930. Influence on varying oxygen tension upon the rate of oxygen consumption of fishes. Bull., Scripps Inst. Ocean., Technical Series, vol. 2, pp. 307-317. La Jolla, Calif. Keys, A. B., and E. N. Willmer. 1932. Chloride secreting cells in the gills of fishes, with special reference to the common eel. Jour. Phvs., vol. 76, pp. 368-378. Cambridge, England. Kolff, W. M. 1908. Untersuchungen fiber die Herztatigkeit bei Teleostiern. Archiv ffir die gesammte Physiologie des Menschen und der Tiere, Bd. 122, s. 37-97. Berlin. Kupzis, J. 1901. Ueber dem niedrigsten ffir das Leben der Fisohe nothwendigen Sauerstoffgehalt des Wassers und fiber die ffir dieselben giftigen Mengen im Wasser geloster Kohlensaure. Zeitscrift ffir Untersuchen der Nahrungs - und Genussmittel, sowie der Gebrauchs - gegen- stande, Bd. 4, s. 631-368. Berlin. Kupzis, J. 1902. Die Naplithafischgifte und ihr Einfluss auf Fische, andere Tiere und Bakterien. Zeitschrift ffir Fischerei und deren Hilfswissenschaften, Bd. 9, s. 144—167. Berlin. Macht, D. I., and H. P. Leach. 1930. Pharmacological studies of twenty-three isomeric octyl alcohols. Jour. Pliarm. and Exper. Thera., vol. 39, pp. 71-79. Baltimore. Marsh, M. C. 1907. The effect of some industrial wastes on fishes. Water-Supply and Irriga- tion Paper No. 192, U. S. Geol. Sur., pp. 337-348. Washington. Marsson, M. 1911. The significance of flora and fauna in maintaining the purity of natural waters, and how they are affected by domestic sewage and industrial wastes. (English transla- tion by E. Kuichling.) Engineering News, vol. 66, pp. 246—250. New York. Massachusetts State Board of Health. 1913. Forty-fourth Annual Report of the State Board of Health of Massachusetts, 1912 (1913). Public Document No. 34, Massachusetts State Board of Health, vol. 6, pp. 1-779. Boston. 436 BULLETIN OF BUREAU OF FISHERIES McArthur, M. C., and W. H. T. Baillie. 1929. Metabolic rates and their relation to longevity in Daphnia magma. Jour. Exper. Zool., vol. 53, pp. 243-268. Philadelphia. McCay, C. M., and H. M. Vars. 1931. Studies upon fish blood and its relation to water pollu- tion. In Biological survey of the St. Lawrence Watershed, Supplement to Twentieth Annual Report of the New York Conservation Department, pp. 230-233. Albany, N. Y. McDonald, M. 1885. Effect of waste products from Page’s Ammoniacal Works upon young shad fry. Bull., U. S. Fish. Com., vol. 5, pp. 313-314. Washington. Moore, G. T., and K. F. Kellerman. 1905. Copper as an algicide and disinfectant in water supplies. Bur. Plant Ind., Bull. No. 76, 55 pp. Washington. Munch, J. C. 1931. Bioassays, a handbook of quantitative pharmacology, 958 pp. Williams & Wilkins Co. Baltimore. Paton, D. N. 1904. Observations on the amount of dissolved oxygen in water required by young salmonidae. Proc. Royal Soc. Edinburg, vol. 24, pp. 145-150. Edinburgh. Pearsall, W. H. 1930. Phytoplankton in the English Lakes. I. The proportion in the waters of some dissolved substances of biological importance. Jour. Ecology, vol. 18, pp. 306-320. London. Penny, C., and C. Adams. 1863. Fourth Report, Royal Commission on Pollution of Rivers in Scotland, vol. 2, Evidence, pp. 377-391. London. Plehn, M. 1924. Praktikum der Fischkrankheiten. In Handbuch der Binnenfischerei Mittel- europas, Bd. I, s. 301-470. Stuttgart. Powers, E. B. 1917. The goldfish ( Carassius carassius) as a test animal in the study of toxicity. Illinois Biol. Mono., vol. 4, pp. 127-193. Urbana, 111. Powers, E. B. 1921. Experiments and observations on the behavior of marine fishes towards the hydrogen-ion concentration of the sea water, in relation to their migratory movements and habitat. Publications, Puget Sound Biol. Sta., vol. 3, 1921-25, (1925), pp. 1-22. Seattle, Wash. Powers, E. B. 1922. The physiology of the respiration of fishes in relation to the hydrogen-ion concentration of the medium. Jour. Gen. Phys., vol. 4, pp. 305-317. New York. Powers, E. B. 1929. Fresh water studies: The relative temperature, oxygen content, alkali reserve, the apparent carbon dioxide tension and the pH of the waters of certain mountain streams at different altitudes in Smoky Mountain National Park. Ecology, vol. 10, pp. 97-111. Lancaster, Pa. Powers, E. B. 1930. The relation between pH and aquatic animals. Amer. Nat., vol. 64, pp. 342-366. Boston. Powers, E. B. 1932. The relation of respiration of fishes to environment. Ecol. Mono., vol. 2, pp. 385-473. Durham, N. C. Richardson, R. E. 1928. The bottom fauna of the middle Illinois River, 1913-1925; its dis- tribution, abundance, valuation, and index value in the study of stream pollution. Bull., Nat. Hist. Sur., State of Illinois, vol. 17, pp. 387-475. Urbana, 111. Rushton, W. 1922. Biological notes. Salmon and Trout Magazine. London. Ruttner, F. 1926. Bermerkungen fiber Sauerstoffgehalt der Gewasser und dessen respiratori- schen Wert. Naturwissenschaften, Bd. 14, s. 1237-1239. Berlin. Shelford, V. E. 1917. An experimental study of the effects of gas waste upon fishes, with especial reference to stream pollution. Bull., Illinois State Lab. of Nat. Hist., vol. 11, pp. 381-412. Urbana, 111. Shelford, V. E. 1929. Laboratory and Field Ecology. Williams & Wilkins Co., 608 pp. Bal- timore. Smith, H. W. 1929. The excretion of ammonia and urea by the gills of fish. Jour. Biol. Chem., vol. 81, pp. 727-742. Baltimore. Smith, H. W. 1930. The absorption and excretion of water and salts by marine Teleosts. Amer. Jour. Phys., vol. 93, pp. 480-505. Baltimore. Southgate, B. A., F. T. K. Pentelow, and R. Bassindale. 1933. The toxicity to trout of potassium cyanide and p-cresol in water containing different concentrations of dissolved oxygen. Biochem. Jour., vol. 27, pp. 983-985. Cambridge, England. STREAM POLLUTION 437 Standing Committee on Rivers Pollution. 1924. River Pollution and Fisheries. Min. Agri. and Fish., 42 pp. London. Steinmann, P. 1928. Toxikologie der Fische. Handbuch der Binnenfischerei Mitteleuropas, Bd. 6, s. 289-342. Stuttgart. Surbeu, E. W. 1936. The culture of daphnia. The Progressive Fish Culturist, U. S. Bur. Fish., no. 17, pp. 1-6. Washington. Surber, E. W., and O. L. Meehan. 1931. Lethal concentrations of arsenic for certain aquatic organisms. Trans., Amer. Fish. Soc., vol. 61, pp. 225-239. Hartford, Conn. Suter, R., and E. Moore. 1922. Stream pollution studies. Bull. New York State Conser. Com., pp. 3-27. Albany, N. Y. Thomas, A. 1924. Studies on the absorption of metallic salts by fish in their natural habitat. II. The absorption of nickel by Fundulus heteroclitus. Jour. Biol. Chem., vol. 58, pp. 671-674. Thompson, D. H. 1925. Some observations on the oxygen requirements of fishes in the Illinois River. Bull., Nat. Hist. Sur., vol. 15, pp. 423-437. Urbana. Tupholme, C. H. S. 1933. Death of fish from cyanides in coke-oven effluents. Indus, and Eng. Chem., News Ed., vol. 11, p. 211. Easton, Pa. Turner, C. L. 1927. Biological survey of Fox, Wisconsin, and Flambeau Rivers, Wisconsin, with special reference to pollution. In Stream Pollution in Wisconsin, Special Report of the Conservation Commission and State Board of Health of Wisconsin, pp. 242-276. Madison. Weigelt, C., with 0. Saare and L. Schwab. 1885. Die Schadigung von Fischerei and Fisch- zucht durch Industrie und Haus Abwasser. Archiv fur Hygiene, Bd. 3, s. 39-117. Miinchen und Leipzig. Welch, P. S. 1935. Limnology. McGraw-Hill Co., *471 pp. New York and London. Wells, M. M. 1913. The resistance of fishes to different concentrations and combinations of carbon dioxide and oxygen. Biol. Bull., vol. 25, pp. 323-347. Lancaster, Pa. Wells, M. M. 1915a. Reactions and resistance of fishes in their natural environment to acidity, alkalinity and neutrality. Biol. Bull., vol. 29, pp. 221-257. Lancaster, Pa. Wells, M. M. 1915b. The reactions and resistance of fishes in their natural environment to salts. Jour. Exper. Zoo., vol. 19, pp. 243-283. Philadelphia. Wells, M. M. 1918. The reactions and resistance of fishes to carbon dioxide and carbon monox- ide. Bull., Illinois State Lab. Nat. Hist., vol. 11, pp. 557-578. Urbana. West Riding River Board, Laboratory Staff of. 1930. Biological survey of the River Wharfe. I. Dissolved substances of biological interest in the waters. Jour, of Ecol., vol. 18, pp. 274-285. London. Wiebe, A. H. 1928. Biological survey of the upper Mississippi River with special reference to pollution. Bull., Bur. Fish., vol. 43, pp. 137-167. Washington. Wiebe, A. H. 1930. Notes on the exposure of young fish to varying concentrations of arsenic. Trans., Amer. Fish. Soc., vol. 60, pp. 270-276. Hartford, Conn. Wiebe, A. H. 1931a. Notes on the exposure of several species of pond fishes to sudden changes in pH. Trans., Amer. Micro. Soc., vol. 50, pp. 380-393. Menasha, Wis. Wiebe, A. H. 1931b. Dissolved phosphorus and inorganic nitrogen in the water of the Mississippi River. Science, vol. 73, p. 652. Lancaster, Pa. Wiebe, A. H. 1933. The effect of high concentrations of dissolved oxygen on several species of pond fishes. Ohio Jour, of Science, vol. 23, pp. 110-126. Columbus, Ohio. Wiebe, A. H., J. G. Burr, and H. E. Faubion. 1934. The problem of stream pollution in Texas with special reference to salt water from oil fields. Trans., Amer. Fish. Soc., vol. 64, pp. 81-85. Hartford, Conn. Winslow, C. E. A., and E. B. Phelps. 1906. Investigations on the purification of Boston Sewage. U. S. Geol. Sur., Water Supply and Irrigation Papers No. 185, 163 pp. Washington. Wisconsin State Board of Health. 1927. Stream Pollution in Wisconsin, Special Report, a joint report of the Conservation Commission and the State Board of Health in Wisconsin, concerning activities in the control of stream pollution, from July 1, 1925, to December 31, 1926, 328 pp. Madison. . ■ - ' ■' ' - -■ ... ■ 9 ■ • . . /. ' ; ■ • •/, >=! •/. ’ ' ' , f;i ' sM Or’.: : ! ■ • : ■ ■ ' • U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner EXPERIMENTAL OBSERVATIONS ON SPAWNING, LARVAL DEVELOPMENT, AND SETTING IN THE OLYMPIA OYSTER OSTREA LURIDA By A. E. Hopkins From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 23 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1937 For sale by the Superintendent of Documents, Washington, D. C. Price 25 cents EXPERIMENTAL OBSERVATIONS ON SPAWNING, LARVAL DEVELOPMENT, AND SETTING IN THE OLYMPIA OYSTER, OSTREA LURID A1 By A. E. Hopkins, Aquatic Biologist, United States Bureau of Fisheries CONTENTS Page Introduction 4-39 Method of cultivation 441 Enemies of the oyster 441 Aims of investigation 443 Hydrographical observations 443 General description of region 443 Temperature 444 Salinity and pH 448 Spawning 456 Size of broods 458 Relation of temperature to spawning- 460 Spawning season 464 Development of larvae 467 Setting 471 Effect of angle of surface 472 Method of determining frequency of setting 475 Setting seasons, Oyster Bay 477 Season of 1931 477 Page Setting — Continued. Setting seasons, Oyster Bay — Continued. Season of 1932 479 Season of 1933 480 Season of 1934 481 Season of 1935 483 Setting seasons, Mud Bay 483 Season of 1931 484 Season of 1932 484 Season of 1933 484 Season of 1934 486 Season of 1935 486 Periodicity of setting 487 Stages of tide and setting 489 Depth of setting 493 Correlation between spawning and setting. 495 Discussion 497 Summary 500 Literature cited 502 INTRODUCTION The native oyster of the Pacific coast has never been produced in great enough abundance to reach markets all over the country. Toward the end of the nineteenth century extensive commercial use was made of the crops growing naturally on tide lands of Puget Sound and Willapa (Shoalwater) Bay, in the State of Washington, resulting in almost complete depletion in most of the favorable localities. In 1902, according to Galtsoff (1929), 154,000 bushels of oysters were produced; in 1904, 170,000 bushels reached the market; while in 1926 only about 58,000 bushels were grown. Since this time production has been at an even lower level. The native Willapa Bay oyster has been almost completely destroyed so that it is now difficult to find in the local markets. The native 03^ster is unique in the United States in that it never attains a shell length much greater than about 5 centimeters (2 inches), and for this reason is used primarily for special dishes such as cocktails and pan roasts. It is too small to serve on the half shell. Oyster growers commonly market them in 2-bushel sacks, containing about 5,000 oysters, or about 2,500 to the bushel. Most of the native oysters now grown on the 1 Bulletin No. 23. Approved for publication Oct. 14, 1936. 439 440 BULLETIN OF THE BUREAU OF FISHERIES Pacific coast are produced, in the southern portion of Puget Sound in the vicinity of Olympia, Wash. They are sold on the market as Olympia oysters, and a distinction is made between them and the same species grown in other localities. According to Stafford (1914) the species was described by Carpenter, who gave the name, as follows: “Ostrea lurida, n. s. Shape of edulis: texture dull, lurid, olivace- ous, with purple stains.” The species is known to occur in bays and estuaries from British Columbia to southern California. However, in some respects the oysters are quite different both in appearance and marketability with respect to their place of origin. Townsend (1893) hardly considered the native oyster of San Francisco Bay of commercial significance, although present in large numbers. The same species, farther north, at that time was bringing good prices in the markets. This was due in part to the difference in climate in the two localities, and in part to the fact that growers were beginning to cultivate their grounds and care for their crops systematic- ally, instead of merely harvesting the natural supply. Because of their susceptibility to the hot sunshine of summer and the freezing winds of winter, native oysters in Washington thrived only where they were relatively protected. Natural beds were found where the oysters were covered with water at low tide because of the slope of the tide land, or where seepage from underground would keep them moist in summer and relatively warm in winter. Pot holes would contain oysters while the intermediate ground, which becomes completely exposed at low tide, would be bare. At the end of the last century, a few years after the appearance of Dean’s work (1890) describing the method of oyster culture employed in France, the oystermen began to build dikes or structures on the tide lands which would keep the beds cov- ered at low tide. The dikes are, in principle, closely similar to those described and pictured by Dean as the “oyster parks” used in France. Whether or not the French system furnished the original inspiration for the mode of oyster culture that was de- veloped in Puget Sound within a few years is not known, but owners of natural ground began to build dikes around the beds so that the oysters would remain covered with water at low tide. After a few years of experimentation, during which it was demonstrated that dikes make it possible to grow oysters on ground previously unused as well as to reduce mortality due to freezing, the entire industry in Puget Sound undertook systematically to dike the natural beds and expand to other grounds. Until very recent years most of the dikes were built of concrete, set well down into the bottom. The thickness of these dikes varies from about 6 inches to nearly 12 inches, depending upon the location. Now dikes are usually built of creosoted lumber, which lasts a long time and is more readily handled (figs. 1, 2, 3). Also, breaks due to settling are less frequent and more simply repaired. Dikes have been constructed on relatively level mud flats and on sloping banks. In the latter case the dikes are arranged in terraces, involving a great amount of hand labor for leveling. In all cases, when new ground is made, the rather soft natural bottom has to be surfaced with gravel to make it hard and firm as well as to maintain a relatively constant level in spite of the swift tides. In southern Puget Sound, according to the tide tables of the U. S. Coast and Geodetic Survey, the maximum range of tide is 20 feet, from —3.8 to +16.2 feet. Most of the oyster grounds are between the —1-foot and the +3-foot tide levels, though some dikes require a tide as high as +8 feet to cover them. A few natural beds are in sloughs or shallow channels where they are never exposed. On the other hand, the natural beds of the same species in Yaquina Bay, Oreg., are covered by from 10 to 20 feet of water at low tide. In Puget Sound oyster growers have found U. S. Bureau of Fisheries, 1937 Bulletin No. 75 Figure 1.— Scowload of Japanese oyster shells near a culling house ready to be planted on the diked ground. Figure 2. — Egg-crate fillers as spread from scows at high tide on diked ground being placed so that they will be completely covered at low tide. U. S. Bureau of Fisheries, 1937 Bulletin No. 23 Figure 3. — Crew of men taking up seeds for transplantation to market grounds. SPAWNING AND SETTING OF OLYMPIA OYSTERS 441 that the higher dikes are best for catching seeds while the lower grounds produce a superior product for market. METHOD OF CULTIVATION Although Galtsoff (1929) gave a description of the Olympia oyster industry and the methods of cultivation in use, it is necessary to review these matters briefly because of their bearing upon the experimental work which is described below\ After spawning is well under way in June the oyster growers plant cultch, either shells or manufactured collectors, on the seed grounds. Until about 1930 the only cultch available was the native shells from the opening houses, but with the recent plantings of Japanese oysters a great quantity of these large shells is obtainable (see fig. 1). The development of the concrete-coated egg crate filler has also made larger plantings of cultch possible (see fig. 2). The spat which are caught are generally left on the seed ground for about 3 jrnars before they are transplanted to growing grounds. Seeds are usually moved in April and May, permitting planting of new cultch on the same ground a short while later. Generally the seeds moved in spring are culled the follow- ing winter, though only the largest oysters reach market. All oysters are taken up by hand since the grounds are exposed when the tide is low (see fig. 3). At low tide on one day a place to set a scow is cleared by forking the oysters to either side. The scow is staked in position at high tide and, when the ground is again exposed, the oysters forked onto it. As soon as depth of water permits, the scow is towed to the culling house and the oysters unloaded into a “sink float”, made of two logs and a bottom, so that the oysters are washed free of mud as well as protected from weather conditions. They are taken up in wheel barrows from the sink float and loaded onto the large table in the culling house, where by tedious hand labor the marketable oysters are separated from the mass of shells and smaller seeds which are returned to the ground. The workers also separate the “slipper shells”, or “cups”, Crepidula jornicata, and the whelk or native snail, Thais lamellosa, and spread them high upon the beach to die and dry so they may be used as cultch. Cullers are paid extra for the snails and “cups” which they remove. The culled oysters are spread in another sink float where they are frequently forked over until the water washes them thoroughly clean. As recpiired for market they are packed in 2-bushel sacks and shipped to the opening houses. The cullers take up the oysters, return the seeds to the beds, and prepare the oysters for market and are paid on the basis of the number of sacks shipped. Japanese do almost all of this work as well as the shucking in the opening houses. The small size of the Olympia oysters, in proportion to that of the Eastern and Pacific (Japanese) species, renders them much more expensive to handle. The aver- age age of the marketable oyster is about 4 years, and about 5,000 of them are required to fill a 2-bushel sack. Ordinarily about 3 gallons of meats are obtained to the sack, so that a gallon contains about 1,600 oysters, as compared with 150 to 250 Easterns and about 50 to 200 Pacifies. ENEMIES OF THE OYSTER Although during the last few years there has been no apparent large mortality due to parasites, there are various organisms taking a constant toll of the crops. Ducks have given more trouble near Olympia than any other enemy (Galtsoff, 1929) and several species of these find the small size, single, native oysters an ideal, readily accessible food supply. Combating these is most difficult; and although some years 442 BULLETIN OF THE BUREAU OF FISHERIES ago a strenuous campaign was waged against them, the growers now appear to accept the damage passively. For some years the great problem of oyster growers has been the “cup”, Crepidula fornicata, which was presumably introduced into these waters with Eastern seed oysters. Although not a parasite, the species multiplied until many of the diked beds contained far more “cups” than oysters. Since the growers first became uneasy about them, they have paid the cullers extra for separating them out, and in this way have considerably reduced their numbers. However, even now it is not unusual for equal numbers of sacks of oysters and “cups” to be culled from a bed. The species appears to thrive much better in the diked beds than on the natural seepage grounds. Several kinds of predatory snails are found on the grounds. The native whelk, Thais lamellosa, occurs in great abundance; and the writer has found that these drill some adult oysters and, in places, a great many spat. They appear to attack mussels primarily. They were previously unrecognized as an active enemy but are now culled out along with the “cups.” Also, their habit, during the breeding season in late winter and early spring, is to come together in large clusters around a shell or rock where the egg cases are deposited. During this time they may be taken up in sacks and placed on the beach to die. The moon snail, Polynices cewisii, is frequently seen on oyster grounds but is primarily a clam borer and probably seldom attacks oysters. The Eastern oyster drill, Urosalpinx cinerea, introduced with seed oysters from the Atlantic coast, may be found in some places, though only in Samish Bay where Japanese oysters are now grown is it relatively abundant. Of greater potential importance is the Japanese oyster drill, Tritonalia japonica, which has been introduced with seeds from the Orient. Few Japanese seeds have been planted near the im- portant Olympia oyster grounds and no damage to native oysters has yet been noted. However, in Samish Bay, this drill has propagated rapidly and for the last few years has been causing tremendous mortality among the Japanese oysters. After a visit to Samish Bay in 1928, Galtsoff (1929) wrote: Although at present there is no evidence that Tritonalia japonica is destructive to oysters, yet as a matter of precaution it is desirable to restrict the planting of Japanese species to the waters in the northern part of Puget Sound and not to extend them to the areas where high-priced Olympia oyster bottoms are located. When the writer first visited this ground 4 years later a great many drilled shells were found. In 1935 there was evidence of still greater mortality. The rapid propagation of the species to dangerous proportions indicates the prob- lem which Olympia oyster growers may soon face, especially since the thin-shelled, slow-growing native oysters would probably be more easily attacked than the rela- tively heavy-shelled Japanese oyster. Unfortunately, Galtsoff’s suggestion was not followed, and it is known that the drills have been introduced near some of the native beds. On one ground in Oyster Bay a number of drills have been found, introduced presumably with Japanese oyster shells from Samish Bay. At the time of writing nothing is being done to prevent rapid spread of the pest to other grounds. At times starfishes become abundant enough to destroy many oysters, but these are readily removed from the cultivated beds. One of the greatest problems of growers is to maintain their dikes against the “crawfish” or mud-shrimp, Upogehia pugettensis, (MacGinitie, 1930) which has a habit of burrowing under the dikes and opening passages which are rapidly enlarged by flow of water, SPAWNING AND SETTING OF OLYMPIA OYSTERS 443 AIMS OF INVESTIGATION A great many comprehensive experimental studies have been made on the biology of the oyster of the Atlantic coast, Ostrea virginica, hut the only significant investiga- tion on the practical phases of the biology of 0. lurida was that of Stafford (1914, 1915, 1916, 1917, and 1918). He made his observations in British Columbia, in the northern part of Puget Sound, where the system of oyster culture had not been developed to an extent comparable to that in use near Olympia. Townsend’s (1893) early paper gave the first general description of the industry on the Pacific coast. Recently Coe (1931a, 1931b, 1932a., and 1932b), Hori (1933), and Hopkins (1935, 1936) have furnished more specific information about the species. The primary purpose of this investigation, which was undertaken in the spring of 1931 and continued through 1935, was to make an analysis of spawning activities and setting habits of larvae with reference to environmental conditions. By develop- ing such information, it was hoped that oyster growers might be assisted in the catching of sufficient seed oysters to restore and expand the industry. In the following pages the more important of the results are described.2 HYDROGRAPHICAL OBSERVATIONS The usual methods were employed for the taking and testing of water samples at different depths and under different tidal conditions. Specific gravity was measured with hydrometers certified and corrected by the National Bureau of Standards. A Hellige hydrogen-ion comparator was used with phenol red to determine the pH. Temperature of water samples was tested with standard thermometers, and in addi- tion, continuous records of water temperature on the oyster grounds, at the level of the oysters, were made with a frequently checked thermograph. GENERAL DESCRIPTION OF REGION Puget Sound is an extremely irregular, deep body of water extending roughly 200 miles north and south in British Columbia and the State of Washington. It is continuous with the Pacific Ocean through the Straits of Juan de Fuca. The Sound is broken up into numerous bays and inlets which are generally quite deep except at their upper ends. Natural beds of native oysters were originally found in many of the small bays but were soon exhausted in all except a few localities where conditions necessary for successful propagation were especially favorable. The several bays near Olympia, Wash., have continued to produce oysters, while beds in other places disappeared, largely because of favorable environmental factors and because of the development of the system of diking the grounds and planting cultch employed by the growers. These bays are separated from the ocean by more than 150 miles of water, yet changes in salinity are relatively slight, due to the great depth throughout the Sound. 2 1 wish to express my thanks to Charles R. Maybury, director of the Department of Fisheries and Game of the State of Wash- ington and to Charles R. Pollock, supervisor of fisheries, for their cooperation in maintaining the laboratory and supplying an as- sistant and boats. Since the division of commercial fisheries became an independent department in December 1932, the director, B. M. Brennan, has continued to support this work under trying financial conditions and he deserves much credit for what has been accomplished. It is a pleasure to express my thanks to the growers of Olympia oysters, all of whom have willingly given every possible assistance. I am particularly indebted to J. J. Brenner, E. G. Brenner, and D. I. Ginder, of the J. J. Brenner Oyster Co.; Ole Hanson and J. S. Waldrip, of the Olympia Oyster Co.; G. W. Ingham, Olympia Oyster Investment Co.; E. N. Steele; Charles Brenner; W. J. Waidrip; J. B. Bowman; J. H. Post; and the late Mrs. Minnie Blass. A large part of the credit for this work is due to H. H. Adams, who served during 5 years as a most capable and efficient field assistant. 444 BULLETIN OF THE BUREAU OF FISHERIES In figure 4 a portion of a chart (from U. S. Coast and Geodetic Survey, chart no. 6460) is reproduced to show the general contours of the most important bays in which Olympia oysters are cultivated. All of the observations here described were made in the area illustrated. The most extensive and successful grounds are in Totten Inlet, commonly called Oyster Bay. Mud Bay (Eld Inlet) is next in importance. Oakland Bay and Little Skook- um (Skookum Inlet) also contain important grounds, but during the last few years, since a pulp mill began op- eration in the vicinity, they have been almost entirely out of produc- tion. (See Hopkins, Galtroff,and McMillin, 19151 ; Hopkins, 1931a). The location of culti- vated grounds is indi- caled on the chart. These are on the mud flats in the upper ends of the bays and on the relatively narrow beaches along the shores adjoining deep water. Altogether there are only some- thing like 400 to 500 acresof producing grounds. Budd Inlet, on which Olympia is located, originally con- tained widespread beds of natural oysters, but has been condemned on account of sewage pollution. TEMPERATURE Figure 4. — General contours of oyster-producing bays near Olympia, Wash. Numbers refer to depth in fathoms. Location of diked beds is shown by x’s. Dikes in which most obser- A Lsi’istol reCOrd- vations were made are indicated. Depth samples were taken in channels off Corters Point Jj, g. thermometer Was (C. P.), Maple Point (M. P.), and Deepwater Point (D. P.). . ° installed on a frame well above the high tide level but with the bulb fixed at the level of the oysters in the dike below. Protected though they are by a few inches of water at low tide, the oysters are nevertheless subjected to considerable variations in temperature as affected by both tides and seasons. The thermograph records were analyzed by averaging the readings on each hour of the day. This is necessarily not strictly accurate, but undoubtedly the error involved is within that inherent in the instrument itself. SPAWNING AND SETTING OF OLYMPIA OYSTERS 445 A graph (fig. 5) is reproduced to illustrate the daily maximum, minimum, and aver- age temperatures during winter (December 1932, January and February 1933) and summer (June through August 1933). This well represents the extremes, for during summer, at low tide, the water frequently reached 25° to 30° C., and during winter dropped to almost —2° C., or close to the freezing point of seawater. In the latter instance a great many oys- ters which were not well covered with water of high salinity were frozen and killed. The minimum tem- perature during summer and the maximum during win- ter show only slight fluc- tuations, since in summer the extreme low tides are during the day and in win- ter at night, the local tem- perature of the air not greatly affecting that of the water around the oys- ters at high tide. The difference between winter maximum and summer min- imum is about 10° C. In order to show in detail the changes in water tem- perature in the dikes dur- ing a 24-liour period, as in- fluenced by the range of tide, a graph (fig. 6) is given on which the continuous temperature records during 4 days are reproduced. Neap tide and spring tide temperature records are shown for typical days dur- ing both winter and sum- mer. In the record for Au- gust 2 it will be noted that during the several hours that the dike was exposed by a —1.7-foot tide the temperature rose gradually from about 19° to about 30° C., and that when the flood tide poured over the dike the temperature dropped about 5° almost instantly. The other summer record was taken a few days later when low tide occurred at about 5 o’clock of a cool morning, and although the dike was not quite exoosed there was a marked drop in temperature. The picture for temperature variations during winter is almost the reverse, the low tide occurring at night when the air is coolest. In all cases the 149604—37 2 DECEMBER JANUARY FEBRUARY Figure 5. — Daily average, maximum, and minimum temperature on an oyster bed in Oyster Bay during winter and summer, as shown by thermograph records. 446 BULLETIN OF THE BUREAU OF FISHERIES “i — i — i — i — i — i — i — i — r i i I l i I l I I l ( n, ' ~ - ' ' -J 1 1 1 I I !_ -1 1 1 1 1 1 I I I 1 1_ variation is slight except at low tide when the water is shallow and readily reacts to sunshine and atmospheric conditions. That is, it is the surface water which responds readily to weather condi- tions; and the oysters may be under 16 feet of water at high tide and 3 or 4 inches a few hours later. Seasonal variations in water temperature from year to year are relatively uniform, but the differences between successive seasons are sufficient to have a con- siderable bearing upon the spawning of oysters. In table 1 the monthly aver- ages for 4}i years are given, as calculated from thermo- graph records obtained in Oyster Bay. The highest average water temperature is usually in August, the lowest in January or Feb- ruary. The annual variation is represented graphically in figure 7 for the 2 years (1933, 1934) when the temperature values were most widely different from one another. The spring rise in the curve for 1934 occurred about a month earlier than in 1933, account- ing for a comparable difference in the time of spawning. Included on the graph are monthly averages of daily readings of maximum and minimum air temperature at Olympia. These records were sup- plied through the kindness of Charles F. Norrie, official weather observer. Water temperature is clearly correlated with air temperature. Other aspects of the tempera- ture conditions are considered in later sections referring to the comparison of Oyster and Mud Bays in salinity, pH, and temperature. Table L- — Average monthly water temperature in dike in Oyster Bay, calculated from thermograph records [Temperature °C.] Figure G. — Reproductions of portions of thermograph records showing variations in water temperature on oyster ground during four 24-hour periods, two in summer and two in winter. The most variable records refer to spring tides, the others to neap tides. Time and height (in feet) of high (H) and low (L) tides are indicated. (PI AX) ' — t — r — i — i — r— i — t — i n -1—1 — r— _ -a ! ?' ; * ,al • - 7 \ f / \ \ / yM WATCR (AV.) NJ ? / . ’/ \ V, tr • 7 *>'Ain{niri) \ €>% / . O. / t>- - o’ «. “ *V 1933 1—1 1 1 1 .. 1 1 i l— I L 1934 1 1 1— 123456789 10 II 12 123456789 10 II 12 Figure 7.— Average monthly temperature of water on oyster grounds during 2 years as related to monthly averages of maximum and min- imum daily air temperature at Olympia. 1932 1933 1934 1935 Month 1931 1932 1933 1934 1935 7. 28 7. 70 9. 38 7. 66 July 18.2 18. 10 18.71 19.27 19. 48 6.36 6. 07 10. 07 8.88 August . . . 18. 45 18.61 19. 12 19. 79 19. 45 8. 56 9. 30 11.83 9. 22 September _ 17. 04 17.09 16. 07 18.60 10.61 11.76 12. 03 13. 15 14.61 13.69 13.90 15.41 14. 02 13.06 15. 46 9. 92 11.70 11. 78 11.41 8. 95 17.00 16.64 17. 67 17. 93 December 7. 88 7.89 9.61 9.78 9. 38 Month January,. February. March April May June SPAWNING AND SETTING OF OLYMPIA OYSTERS 447 Table 2. — Comparison of dikes 5 and S in salinity, temperature, and pH Dike 5 Dike S Date Tide and depth Temper- ature Salinity pH Tide and depth Temper- ature Salinity pH 1932 ° C. ° C. F— 6 ft._ 7.4 27. 72 8.0 12 E— 10 ft 7.3 27. 17 8.0 19 F— 10 ft- 7.3 27. 36 8. 0 27 E— 10 ft 6. 2 25. 48 8.0 Feb. 4 E— 7 ft 5.0 26.34 8.0 11 E— 8 ft 6.3 27. 35 8.0 24 E— 4 ft- 7. 4 24. 27 8.0 E— 7 ft 7. 1 25. 24 8.0 8 E— 4 ft 8.0 21.98 8.0 15 E— 8 ft 8.0 25. 93 8.0 22 Exp 8. 2 24.60 8. 2 29 E— 8 ft 8. 3 24.61 8.2 Ebb 10. 1 24. 13 19 Exp 12. 1 21. 14 8.4 26 Exp 15.0 22. 75 8.0 May 3 Exp.. 18.9 24. 94 8.0 10 Exp 15.6 26.24 7.8 17 Exp.. 23.3 26. 33 8.4 19 Exp 13.9 25. 87 7.8 21 Exp 14.4 26. 82 8. 0 24 Ebb 14. 4 26. 76 8. 0 31 Exp 19. 4 26. 24 June 2 Exp 19.4 26. 17 7.8 4 Exp 15.0 27. 32 8.2 Exp_ 16. 1 27. 32 8.0 6 Ebb 17.8 26. 63 8.0 Exp . . 17.5 27. 49 8. 4 8 Exp 20.3 27. 11 8. 2 13 Exp 18.9 27. 23 7.8 15 Ebb 13.9 28. 04 8.0 Ebb 13. 6 26. 64 8.0 17 Exp 29.4 27. 74 7. 8 Exp 14. 4 27. 65 8.0 20 Exp 20.0 27. 12 7. 8 Exp 12.2 27. 47 8.2 22 Exp ... 18.9 27.31 8.0 Ebb 21. 7 26. 30 7.9 24 Exp. 19.4 27.81 7.9 Exp 20.6 27. 85 8.0 27 Exp . 18.3 27. 64 7.7 Exp 16.4 28. 03 29 Exp.— . 15.5 28. 07 July 1 Exp 24.4 27. 18 8.0 Exp. . 21.9 27. 07 8.0 4 Exp 18.3 27. 69 8.0 Exp 17.8 27. 86 8.0 6 Exp 28.9 29. 42 7.4 Exp 8 Exp 20.0 28. 15 8.0 Ebb 22.8 28. 12 7.8 11 Exp. 16. 1 27. 16 Exp 15.3 27. 97 13 Exp 16. 7 27. 60 Exp 15.5 28. 03 15 Exp ... 16.4 27. 63 Exp 15.5 27. 98 18 Exp 25.0 27. 94 7.8 Exp ... 18.6 27. 99 7.9 20 Exp ... 20. 5 28. 21 7. 8 Exp. 18.6 28. 21 7.9 22 Exp 21. 1 28. 21 7.8 Exp_. _ 21.4 28. 57 7.8 25 Exp.. 16. 1 28. 33 Exp 16. 1 28. 59 27 Exp 16. 1 28. 13 Exp 15.8 28. 59 29 Exp... 17. 2 26. 80 8. 0 Exp 16. 1 27.69 8.0 Aug. 1 Exp 17.2 28. 01 7.8 Exp _ _ . 17.8 27. 69 7.8 3 Exp. . 20.0 28. 16 7.8 Exp.. ... .... 18.3 28. 35 7.8 5 Exp.. 22.2 28.24 Exp 25.0 28. 30 7.8 8 Exp 16. 1 28. 40 Exp .. 16. 1 28.60 12 Exp 15.3 27. 61 Exp 15.5 28.01 15 Exp 16. 7 27. 89 Exp.. 16.4 28. 03 19 Exp 18.9 28. 17 7.8 Exp 18.9 28.31 7.9 22 Exp 16. 1 26. 88 Exp 16. 1 27.27 26 Exp 18.9 28. 17 7.4 Exp__ ... . 16. 4 27.54 7.6 29 Exp .. 18.0 28. 69 Exp 16.9 28. 82 31 Exp 17.8 28. 28 7.9 Exp 16. 1 27. 97 7.9 Sept. 2 Exp 16. 1 28. 30 Exp. . ... . .. 16. 7 28. 55 5 E— Surf 16. 7 28. 69 E— Surf.. 16. 7 28. 53 9 Exp.. 13.0 29. 08 Exp.. .. 13.3 26. 97 12 Exjp. 17.2 28. 56 Exp . 17.8 28. 71 16 Ebb 15.5 29.00 Ebb 16. 1 28. 36 19 Ebb 14.4 28. 94 Ebb.. . 14. 4 29. 04 23 Ebb . 16. 1 28. 78 8.0 26 Exp._ 13.3 28. 21 7.8 Exp . 13.3 28. 65 7.8 30 Exp 17.3 28. 59 7. 8 Exp . 17. 1 28. 87 7.6 Oct. 4 E— Surf 15.4 29. 28 7 F— Surf. 14.4 28. 99 F— Surf... 13.6 29.22 11 1 ft 12. 8 28. 26 7. 4 Exp . 17. 8 27. 47 7.8 14 E — 18 in.. . _ 15.0 28.04 7. 8 Exp 15.0 28. 64 7.8 17 E— 4 ft 14. 2 28 33 7. 9 E— Surf 13. 4 28. 48 7.9 21 E— 7 ft 13.3 28. 64 8.0 E— 5 ft _ 13. 3 28. 98 8.0 24 F— 3 ft 12.3 29. 13 8. 0 F— 4 ft 12. 4 23. 57 8.0 28 F— Surf . 12. 1 28. 73 8. 0 F— Surf 12.2 28.84 8.0 31 E— 6 ft. 11.4 28. 08 8 0 E — 4 ft 29. 02 8.0 Nov. 4 E— 8 ft 10. 1 28. 93 8. 0 10. 1 28. 04 7.9 7 F— 6 ft 11.0 28. 48 7. 9 F— 4 ft 11. 1 28.66 7.9 14 E — 6 ft 10.3 27. 21 7. 9 10.2 26. 62 7.9 21 F— 8 ft . 10. 4 26. 92 7. 8 E— 6 ft 10.4 26. 33 7.8 28 E— Surf. 10. 2 25. 61 7. 8 E— 6 ft 10. 1 26. 22 7.9 Dec. 5 E— 8 ft... 9.2 26. 53 7. 8 F— 8 ft .. 9.4 25. 21 7.8 19 E— 8 ft 6. 4 26. 25 7.9 E— 8 ft 6.4 26.65 7.9 Note. — F=flood; E=ebb; Exp.=exposed. 448 BULLETIN OF THE BUREAU OF FISHERIES SALINITY AND pH Because of the predominant deep water in Puget Sound and the relatively small streams flowing into the southern portion the variations in salinity are not often great, save on the surface. Samples were taken during summer in the exposed dikes, while throughout the rest of the year, when low tides occurred at night, samples were taken at surface and bottom at the same places. Description of conditions is here limited chiefly to Oyster Bay and Mud Bay, in which most of the experimental work was done. Since the two bays offer marked hydrographical and biological differences, it is neces- sary to go into some detail in describing the relative values of salinity and pH as a preliminary to the presentation of biological work. On the chart (fig. 4) it will be seen that the two bays are not markedly different in size, though Oyster Bay is somewhat longer. In both, most of the oyster beds are located at the upper ends where there are relatively level, or gently sloping, bottoms exposed at low tide. More fresh water enters Mud Bay through creeks and seepage than goes into Oyster Bay, but no large stream enters either. Low salinity probably never accounts for any mortality in these bays, though in periods of very heavy rain the creeks sometimes wash quantities of silt over some of the beds. Table 3. — Comparison of temperature, salinity, and pH, at low tide in 4 dikes in Mud Bay (Dike A adjoins shore; others in order to edge of channel) Date Dike A Dike B Dike C Dike D Temper- ature Salinity pH Temper- ature Salinity PH Temper- ature Salinity pH Temper- ature Salinity pH 1931 °C °C °C °C 19.4 27. 25 21.7 27.31 20.5 27. 18 20.5 27. 12 20.0 27. 07 20.8 27. 07 21. 1 26. 88 21. 1 26. 88 15.7 25. 90 16.0 25. 78 15.8 26. 05 17. 8 26. 27 17.9 25. 91 18.3 26.31 18.3 26. 30 16.7 26. 58 17.0 26.31 17.2 26. 45 16.7 25.28 13.5 24. 60 13.6 26. 27 13.5 25. 75 13.9 24. 43 18.0 24. 99 17.8 25.84 8.0 18.3 24. 58 7.8 17.8 25. 17 7.8 July 1 . 18.3 25. 95 18.9 25. 62 19.4 25. 84 19.7 25. 95 17. 2 26. 65 20.0 26. 08 20.8 26. 18 21.4 26. 18 July 9__ 19.4 26.26 8.0 19.7 26. 02 7.8 19.4 25. 90 7.8 19.4 25. 87 7.8 July 13 16.3 27. 06 7.8 16.7 26. 94 7.8 16.4 26.82 7.8 16.7 26. 82 7.8 July 17 21. 1 27. 81 8.0 21.7 27. 75 8.0 21.7 27. 60 8.0 23.6 27. 57 8.0 Tuly 20- 22.2 27.93 8. 0 22.2 27.48 8.0 21.7 27. 75 8.0 22.2 28. 07 8.2 luly 23 17.2 27.83 8.0 17.2 27. 83 8.0 17.2 27.83 8.0 17.2 27.83 8.0 luly 25 19.4 27. 75 8.0 19.4 27. 06 8.0 20.0 27.50 7.8 20.0 27. 48 7.8 July 29 25.8 28. 42 8.0 26. 1 28. 40 7.8 25.8 28. 27 7.8 26.9 28. 01 7.8 26.3 28. 12 8.0 26.4 27. 79 8.0 25.5 27. 75 26. 4 27. 69 8.0 Aug. 7 14.7 28. 13 7.8 15.0 27. 95 7.8 15.3 27.98 7.8 15.3 27. 98 7.8 Aug. 12 21.4 28.65 8.0 21.1 28.51 8.0 20.5 28. 36 8.0 21. 1 28. 10 7.8 Aug. 21 15.0 28. 59 7.8 15.0 28.41 7.8 15.0 28. 60 7.8 15.0 28. 30 7.8 17. 5 29. 72 17.5 28.51 17.5 28.31 17.5 28. 19 Aug. 26 18.9 28.82 18.9 28.31 7.8 18.6 28.15 7.8 19.3 28.24 7.8 Sept. 5 17.8 25.52 7.4 18.5 27. 84 7.4 18.3 27.63 7.4 18.6 26.58 7.4 15. 5 28. 65 15.5 28.35 7.8 15.5 28. 06 15.5 27. 92 16.7 28. 51 16. 1 28. 19 15.8 28. 26 15.0 27. 78 Oct. 8 11.7 28.44 7.8 n.i 27.88 7.6 10.5 27.64 7.6 ii. i 27. 65 7.7 In order to indicate the general results of tests on oyster grounds throughout the year, and the close comparison in salinity and pH of the water on various grounds, the values are given in table 2 for samples taken in two dikes in Oyster Bay, 1932. Dur- ing the summer season samples were taken at low tide when the dikes were exposed and the water quite warm, while at other times of year bottom samples were taken. The day-to-day variation in salinity is not great, and although dike 5 is well up the bay and dike S about 2 miles away (see chart, fig. 4) there is little difference to be noted. SPAWNING AND SETTING OF OLYMPIA OYSTERS 449 Table 4. — Comparison of average monthly values of salinity and pH in a dike in Oyster Bay and one in Mud Bay during 2 years Date Oyster Bay, dike 5 Mud Bay, dike B Date Oyster Bay, dike 5 Mud Bay, dike B Num- ber of sam- ples Aver- age salin- ity Aver- age pH Num- ber of sam- ples Aver- age salin- ity Aver- age pH Num- ber of sam- ples Aver- age salin- ity Aver- age pH Num- ber of sam- ples Aver- age salin- ity Aver- age pH 1932 1933 4 26. 93 8. 00 3 27. 75 8. 00 January 4 24. 02 7. 87 4 26. 49 7. 90 February 3 25. 99 8. 00 4 25. 76 7.98 February. ... .. 3 25. 68 8. 10 4 25. 25 8. 10 5 24.47 8. 08 4 25. 37 8. 00 March 3 24. 70 8. 17 2 26. 84 8. 10 3 22. 67 8. 20 3 27. 52 8. 30 April 5 25. 86 8.30 4 24.92 8. 26 May 7 26. 17 8. 00 4 24. 05 8. 10 May 11 26. 39 8.20 7 24.91 8. 12 9 27.31 7.91 9 25. 84 7. 88 June 12 27. 36 7. 97 12 25. 65 8. 05 July 13 27. 88 7. 85 12 25. 93 7.90 July 12 28. 11 7. 92 13 26. 68 7. 88 11 28. 05 7. 74 10 26. 53 7. 84 August. 11 28. 21 7. 96 13 26. 97 7. 94 8 28. 67 7. 80 10 28. 00 8. 00 September 6 28. 29 7. 93 7 26. 99 7.96 7 28. 49 7. 85 4 28. 64 7. 87 October.. 4 27. 49 7.90 4 23. 59 7. 90 4 27. 88 7. 90 5 27. 39 7. 92 November 4 24. 45 7.80 4 27. 42 7. 75 December 2 26.39 7. 85 2 27. 48 7.80 December. 3 24. 64 7.80 1 26. 36 7. 80 In Mud Bay, however, into which more fresh water flows, there is a distinct gradient (table 3) in the salinity of the water at low tide in a series of 4 dikes from the shore (dike A) to the edge of the channel (dike D). The first three dikes are on the same level but the last ( D ) is about 1 foot lower. The lower level does not account for the salinity difference. The main body of fresh water from creeks at the head of the bay follows the channel, while the contours of the bay tend to carry the more saline water at flood tide to the west side of the bay. The variation in salinity between individual samples taken at any time is relatively slight and the val- ues over a period of a year may be best indicated by monthly averages. In table 4 the average monthly sa- linity and pH are given for 2 full years, 1932 and 1933, in two typical dikes, dike 5 in Oyster Bay and dike B in Mud Bay. The summer samples refer to conditions at low tide when the dikes were exposed. During the rest of the year the values refer only to bottom samples taken at relatively high tide. The more clearly to represent seasonal varia- tions, the data are plotted in figures 8 and 9. The lowest salinity occurs generally during late winter and early spring, depending upon the time of greatest precipitation, and the annual varia- tion, expressed in this manner, is usually between about 24 and about 29 p.p.mille. The salinity on the oyster grounds in Mud Bay is more variable than in Oyster Bay, as may be seen by comparing the figures, and heavy rains affect the water more quickly in the former. The hydrogen-ion concentration varies in a more orderly Figure 8. — Average values of salinity and pH of water on oyster ground (dike 5) in Oyster Bay during 2 years. Most summer samples were taken at low tide while during the remainder of year bottom samples were taken. Compare with Mud Bay, figure 9. 450 BULLETIN OF THE BUREAU OF FISHERIES manner during the year and the two bays are similar in this respect. During late winter the pH rises rapidly from a low of about 7.8, reaching the maximum of about 8.3 in April. It then drops rapidly until midsummer, after which it is relatively stable. This is discussed further below. The mode of sampling on the oyster grounds involves some lack of constancy throughout the year because all samples were taken during the day, so that for a large part of the year the tide was relatively high while in summer the dikes were exposed. A better picture of the high-tide salinity and pH was obtained by making studies in the deeper channels a short distance below the oyster grounds, off Corters Point in Oyster Bay and Maple Point in Mud Bay. That the results may be used to indicate conditions obtain- ing on the oyster grounds is shown in table 5, in which the surface temperature, salinity, and pH are com- pared for three places in Oyster Bay during summer. The salinity off Corters Point and Deepwater Point is almost identical with that in the exposed dikes, although in the last the temperature and pH are decidedly different because of exposure to sunshine and warm air and the respira- tory activity of oysters and other organisms. To show briefly the an- nual variation in the waters of the two bays the average monthly values of salinity and pH at surface and bottom for Corters Point (Oyster Bay) and Maple Point (Mud Bay) are given in table 6 for 2 consecutive jrears. The surface salinity in Oyster Bay throughout the year is generally higher than in Mud Bay, though at the bottom the relationship is reversed and higher salinity prevails in Mud Bay. A similar difference was noted above with respect to the water over the oyster grounds of the two bays at low and high tide. The same average values are reproduced graphically in figures 10 and 11. The lowest salinity is to be found in late winter, near the end of the rainy season, and early in spring the gradual rise in bottom salinity begins. The bottom salinity in Oyster Bay varies during the year from about 26 to 29 parts per mille, in Mud Bay from about 27 to 29.5 parts per mille. The difference between bottom and surface Figure 9. — Average values of salinity and pH of water on oyster ground (dike B ) in Mud Bay during 2 years. Summer samples were taken at low tide while during the remainder of the yeat bottom samples were taken. Compare with Oyster Bay, figure 8. SPAWNING AND SETTING OF OLYMPIA OYSTERS 451 is much greater in the latter bay, indicating the extent of adaptation required of the oysters during the tidal cycles. Shown in the figures also is the monthly precipitation as recorded at Olympia by C. F. Norrie. Rain is markedly seasonal at this place, with almost no rainfall during summer. The low salinity in winter is directly correlated with precipitation, though there is considerable lag in the salinity at the bottom. Table 5. — Salinity, temperature, and pH in Oyster Bay at 3 different points during summer ( see chart, Jig. 3 ) Date Tide Corters Point Deepwater Point Dike 1 Surface Surface Low tide Temper- ature Salinity pH Temper- ature Salinity pH Temper- ature Salinity pH ° C. 0 C. ° C. F 16. 0 27. 57 15. 0 27.94 22. 2 27. 66 F 17.7 26. 64 14.4 28.28 18.9 27. 36 July 10 15.3 28. 40 8.4 15.0 28.37 8.2 16.7 28. 40 8. 2 F 18.0 28. 60 8. 4 28. 60 8. 4 27. 2 28. 75 8. 0 July 18 E 19. 1 28. 98 8.4 20.5 28.35 8.2 26. 1 28.63 8.2 July 24 F 19.6 28.64 8.4 18.3 28.69 8.4 19.4 28.56 7.8 Iuly 27 F 17. 2 29. 29 8.4 18.0 29. 13 8.4 27. 2 29. 09 7.8 July 30 F 18.9 29. 02 8.4 18.3 28. 86 8.4 27.8 28. 98 7.8 Aug. 1 E 19.4 29. 20 8.4 20. 5 28. 86 8.4 25.3 29.09 8.0 Aug. 3 E 18.0 29. 33 8.4 18.3 29. 17 8.4 23. 5 28. 99 8.0 Aug. 8 F 17.5 27. 74 8.2 17.5 29. 56 8.2 17.5 28.53 7.8 Aug. 11 F 17.5 29. 29 8.2 17.8 29. 1 1 8. 2 23.3 29. 37 7.8 Aug. 13 E 19.0 29. 54 8.2 19.7 28. 11 8.2 27.2 28.41 7.8 Aug. 15 - E 18.2 29. 01 8.2 18.9 29. 78 8.2 24.4 29. 65 7.8 Aug. 22 F 18.4 29. 13 8.2 17. 2 28.98 8. 2 17.5 29. 13 7.8 Aug. 27 F 18.7 29. 99 8.2 17.2 29. 76 8.2 23.3 29.56 7.8 Aug. 29 F 18.3 29. 90 8. 1 18.9 29.81 7.7 25.0 29.58 7.8 Sept. 8._ F 16.7 29. 52 8.0 16.7 29. 20 8.0 16.7 28.86 7.6 Averages. 17. 97 28.91 8.28 17. 72 28. 92 8.23 23.28 28.81 7.91 Note.— F= flood; E=ebb. Table 6 — Average monthly values of salinity and pH at surface and bottom off Corters Point ( Oyster Bay ) and Maple Point ( Mud Bay) during 3 years Corters Point Maple Point Date Surface 50 feet Surface 30 feet Number samples Average salinity Average PH Average salinity Average pH Number samples Average salinity Average pH Average salinity Average pH 1932 January 4 25. 45 7. 92 27. 68 8.05 4 25.06 8.0 28.91 8.0 February 3 26. 68 8.0 27. 97 8.0 4 23. 88 7. 97 28. 59 8.0 March 5 24. 27 8.02 25. 78 8. 04 4 23. 45 8.0 27. 16 8.0 April.. May 3 24. 34 8.3 26. 86 8.3 4 25. 37 8. 33 27.89 8.4 6 26. 94 8.4 27. 69 8.4 4 26. 94 8.4 28. 15 8.4 June 3 27.92 8.4 28. 10 8.4 4 27. 42 8.2 28.31 8.3 July 3 28. 34 8. 27 28.38 8. 37 2 27. 43 8. 15 28.28 8.2 August 4 28. 32 8.23 28.53 8.3 5 28.31 8. 14 28.84 8. 36 September. 4 28. 94 8.3 28. 67 8. 15 2 29. 11 8. 15 29. 43 8.2 October. 7 29.01 8. 01 29. 20 8.0 5 29.29 7.98 29.58 8.05 November 5 27. 26 7. 86 28. 18 7. 93 5 25. 45 7. 87 28. 95 7. 92 December 3 25. 79 7.73 26. 78 7.8 2 26. 74 7.8 28. 14 7. 85 1933 January 4 23. 79 7.87 26. 38 7. 87 4 26.27 7.83 27.20 7.9 February 4 25.59 8.1 26.21 8.1 4 24. 61 8.0 27.47 8. 1 March 4 24. 10 8. 13 26.26 8. 15 4 23. 56 8. 1 27.00 8. 1 April 5 26.05 8.28 26. 78 8. 34 4 24. 76 8.27 27. 46 8.3 May 12 27. 11 8. 24 27.66 8.28 8 26.81 8.23 27. 92 8.31 June 11 27.75 8. 13 28.00 8. 18 12 26. 29 8. 17 28.41 8.23 July 11 28. 30 8. 15 28. 46 8. 18 11 27. 50 8. 05 28. 70 8. 12 August... 11 28. 43 8. 12 28.57 8. 12 12 27. 84 8.03 28.51 8.04 September 6 28. 67 8. 07 28. 95 8. 13 7 28.32 8.0 28.59 7. 99 October... 4 27. 79 8.1 28.23 8. 1 4 27. 48 8. 02 28. 86 8. 02 November 4 26. 13 7.8 26. 96 7.8 4 26. 40 7.8 27. 30 7.8 December 3 20.53 7.8 25. 32 7.8 1 18.58 7.8 27. 17 7.8 452 BULLETIN OF THE BUREAU OF FISHERIES In these two figures the annual variations of hydrogen-ion concentration are very striking. The lowest pH value is generally in December, when it averages about 7.8. In late winter it rises rapid- ly, reaching a maximum of about 8.4 usually in April and May, from which it gradually drops during the rest of the year. The time of highest pH is somewhat later than the time of low- est salinity and is probably due to the prolific develop- ment of diatoms and other the water which algae m Figure 10. — Average monthly values of salinity and pH at surface and bottom (50 feet) off Corters Point (Oyster Bay) during 2 years. Total monthly precipitation (Olympia) is also shown. Compare with Mud Bay, figure 11. contains large amounts of fertilizing materials such as nitrates, brought in by the inflowing drainage water which is becoming warmer during early spring. The warming water and the brighter light, associated with the presence of necessary chemical substances, permit the active multiplication of plant life, and photosynthesis rapidly removes carbonic acid, raising the pH. Later, however, as available fertilizing materials become fixed by the algae and as the water becomes warmer the respiratory activity of marine animals, including oysters, crustaceans, and others, restores a high per- centage of carbonic acid to the water, lowering the pH. In this regard it is of in- terest to call attention to changes in the pH and sa- linity of the water in a dike during a complete tidal cycle. In figure 12 the depth of water in a dike is shown throughout a 24-hour period in summer as related to the salinity and pH of the water. During ebb tide the water level became low- er than the dike, leaving it exposed at about 1:15 p. m. At about 4:45 p. m. the flood tide came up to the dike level. During the time the dike was exposed the pH dropped from 8.0 to 7.9, because of carbonic acid excreted by oysters and other organisms, and the salinity rose slightly, due partly to evaporation and partly to stratification of the water permitting the less Figure 11. — Average monthly values of salinity and pH at surface and bottom (30 feet) off Maple Point (Mud Bay) during 2 years. Total monthly precipitation (Olympia) is also shown. Compare with Oyster Bay, figure 10. SPAWNING AND SETTING OF OLYMPIA OYSTERS 453 saline to remain at the surface. During the flood tide both salinity and pH rose. Variations during the rest of the period are relatively slight, though in accord with this interpretation. Table 7 .—Comparison of values of salinity , temperature, and pH off Coders Point during 1932 at different depths Date Tide Surface 6 feet 15 feet 30 feet 60 feet Tem- pera- ture Salin- ity pH Tem- pera- ture Salin- ity pH Tem- pera- ture Salin- ity pH Tem- pera- ture Salin- ity pH Tem- pera- ture Salin- ity pH 1932 °C. °C. °C. °C. °C. Jan. 5 E 7.3 26. 74 8.0 7.4 27. 25 8.0 7.4 27. 66 8.0 7.4 28. 35 8.0 7.4 28. 39 8.0 12 E 6.0 23. 96 7.8 6.4 24. 89 8.0 7.4 27. 86 8.0 7.4 27. 86 8.0 8.0 27.39 8.0 19 F 7.0 24. 60 7.9 7. 1 24. 78 8.0 7.2 26. 65 8.0 7. 2 27. 72 8.0 7.2 27. 86 8.2 27 E 6.2 26.51 8.0 6.3 26. 09 8.0 6.4 26. 19 8.0 7.0 27.01 8.0 7.0 27. 16 8.0 Feb. 4 E 4.3 26. 03 8.0 6. 1 27. 41 8.0 5.2 27. 79 8.0 5.3 27. 99 8.0 5.3 27. 99 8.0 11 E 5.4 27. 98 8.0 5.4 27. 66 8.0 6. 1 27. 68 8.0 6. 1 27. 68 8.0 6.0 27. 77 8.0 24 E 0.4 26.04 8.0 7.2 27. 63 8. 1 7.0 28. 10 8.0 7.0 28. 10 8.0 7.0 28. 16 8.0 Mar. 2 E 6.3 23. 59 8.0 6.4 23.53 8.0 7.2 24. 99 8.0 7.3 25. 97 8.0 7.3 26. 19 8.0 8 E 7.1 23. 75 8.0 7.3 23. 59 8.0 7.3 23. 68 8.0 7.3 25. 02 8.0 7.3 24. 85 8.0 15 F 7.3 25. 53 8.0 7.4 25.32 8.0 7.4 26.53 8.0 7.4 27. 11 8.0 7.4 26. 39 8.0 22 E 8. 1 24. 79 S. 0 8. 1 25. 79 8.0 8. 1 25. 87 8.0 8. 1 25. 91 8.0 8.1 25. 81 8.0 29 E 8.3 23. 09 8.1 8.4 23. 48 8.2 8.4 24. 65 8.2 8.2 24.66 8.2 8.2 25. 64 8.2 E 10. 3 23. 50 10. 2 24. 02 9. 1 25.81 9.0 26. 09 9. 0 26. 17 19 E 9.4 25.91 8.2 10.1 27. 89 8.2 9.4 25. 90 8.2 9.3 27. 75 8.2 9.3 27. 83 8.2 20 E 15. 1 23.41 8.4 13. 1 20. 00 8.4 10.4 28. 16 8.4 10. 2 28. 03 8.4 10. 2 26. 58 8.4 May 3 E 12.2 25.91 8.4 11. 4 26. 18 8.4 11.0 26. 58 8.4 10.3 26. 63 8.4 10 E 12.3 20. 80 8.4 12. 1 27. 30 8.4 12. 1 28. 53 8.4 11.3 27. 20 8.4 11.2 28. 84 8.4 17 F 15.0 26.67 8.4 14.3 27. 27 8.4 13.1 27. 39 8.4 13.0 27. 40 8.4 13.0 27.32 8.4 21 E 13. 2 27. 32 8.4 13. 1 27.31 8.4 13.7 26. 97 S. 4 13.0 27. 43 8.4 13.0 27. 46 8.4 24 E 13. 2 27.41 8.4 13. 2 27. 74 8.4 13.0 27. 66 8.4 13.0 27. 69 8.4 13.0 28. 30 8.4 31 E 13.4 27. 52 8.4 14. 0 27. 45 8.4 13. 2 27.41 8.4 12.4 26. 26 8.4 12.4 27. 63 8.4 June 6 E 13.4 27. 78 8.3 14.0 28. 03 8.4 14. 0 27. 79 8.4 13.3 27. 90 8.4 13.3 28. 06 8.4 15 E 15.0 27. 94 8.4 15.0 28.01 8.4 15.0 27. 95 8.4 14.4 28.01 8.4 14.4 28. 12 8.4 22 E 16.0 28. 03 8.4 16.0 27. 99 8.4 15. 4 27. 84 8.4 15.4 27. 95 8.4 15.4 28. 12 July 8 E 17.2 29. 83 8.4 17.0 28.26 8.4 17. 1 28. 21 8.4 16.3 28. 39 8.4 16.2 28.41 8.4 18 E 15.4 28. 39 8.2 16.3 28. 42 8.2 16. 1 28. 30 8.2 16.0 28.31 8.2 16.0 27. 00 8.2 20 E 17.2 28.51 8.2 17.0 28. 59 8.2 17.0 28. 30 8.2 16.4 28. 30 8.2 16.3 28. 48 8.2 29 F 16.2 26. 67 8.2 10.0 28. 59 8.4 16.0 28. 24 8.4 15.4 28. 24 8.4 15.4 28. 24 8.5 Aug. 3 F 16.3 28. 35 8.3 16.3 28. 42 8.3 16.3 28. 30 8.4 16.4 28. 35 8.4 16.4 28. 48 8.4 10 F 17.4 28. 21 8.2 17.3 28.21 8.2 17. 1 28. 57 8.3 17.0 28. 35 8.2 16.4 28. 68 8.2 17 F 18.0 28. 30 8.2 17.4 28.41 8.4 17.1 28.31 17.0 28. 31 8.4 17.0 28.37 8.4 24 F 17.2 28. 44 8.2 17.0 28. 45 8.2 16.3 28. 59 8.2 16.3 28. 66 8.2 16.3 28. 59 8.2 Sept. 23 F 15.3 28. 74 8.4 15. 2 29. 05 8.4 15. 1 28. 59 8.0 14.4 29. 07 8.0 14.4 28. 98 8.0 28 F 16.4 28.91 8.2 15.4 29. 07 8.2 15.3 29.00 15. 1 28.91 8.2 15. 1 28. 98 8.2 30 E 15.2 28. 94 8.2 15. 1 29. 04 8.2 15. 1 29. 00 8.2 15. 1 29. 07 8.3 15. 1 29. 07 8.2 Oet. 11 F 14.1 28. 94 8.2 14. 1 29. 20 8. 1 14. 1 29. 13 8. 1 14.0 29. 65 8. 1 13.4 29. 22 8.1 14 F 14.2 29.11 8.0 14.2 29. 11 8.0 14.2 29. 00 8.0 15.0 29. 28 8.0 14. 1 29. 23 8.0 17 F 14.0 28. 87 7.9 14.0 28. 85 7.9 14.0 28. 69 7.9 14.2 28.71 7.9 14.2 29. 23 7.9 21 F 13.2 29.08 8.0 13.3 28. 95 8.0 13.3 29. 16 8.0 13.3 29. 47 8.0 13.3 29. 18 8.0 24 F 12.4 29.11 8.0 12.5 29. 16 13.0 29. 32 8.0 13.0 29. 22 8.0 13.0 29. 22 8.0 28 F 12.4 29. 09 8.0 12.4 29. 28 8.0 12.3 28. 98 8.0 12.3 29. 14 8.0 12.3 29.28 8.0 31 F 28. 88 8. 0 29. 05 8.0 29. 36 8.0 29. 05 8. 0 29. 02 8.0 Nov. 4 F 11.0 29. 11 7.9 11.0 28. 98 8.0 29.29 8.0 11.0 29. 32 8.0 11.0 29. 43 8.0 7 F 10.3 26. 92 7.9 11. 1 28. 53 7.9 11. 1 29.09 7.9 11. 1 28. 95 7.9 11. 1 29. 29 7.9 14 E 10.2 27.17 7.9 10.4 27. 52 7.9 10.4 27. 79 7.9 10.4 28. 06 7.9 10. 4 27. 79 21 E 10.4 26. 62 7.8 10.4 27.00 7.8 10.4 26. 42 7.9 11.0 27. 39 7.9 11.0 27. 21 7.9 28 E 10.0 26. 58 7.8 10. 1 26. 85 7.9 10. 1 26. 85 7.9 10.2 27. 39 10.2 27. 20 7.9 Dec. 5 F 9.2 25. 39 7.7 9.4 26.59 7.8 9.4 25. 96 7.8 9.4 25. 95 7.8 9.4 27. 11 7.8 19 E 6.4 26.73 7.8 6.4 27. 16 7.9 7.1 27. 20 7.9 7.2 27. 38 7.9 7.2 27. 39 Note. — E=ebb.; F = flood. Table 8. — Temperature, salinity, and pH of water at different depths off Maple Point ( Mud Bay ) during 1 year Date Tide Surface 3 feet 10 feet 20 feet 30 feet Tem- pera- ture Salinity pH Tem- pera- ture Salinity pH Tem- pera- ture Salinity pH Tem- pera- ture Salinity pH Tem- pera- ture Salinity pH 1932 ° C. ° C. 0 C. 0 C. 0 C. F 4.3 21.49 8.0 6.4 26. 11 8.0 7.2 28. 66 8.0 7.4 28.91 8.0 7.4 29. 16 8.0 8 E 7.3 27. 64 8.0 7.4 27. 89 8.0 8.0 29. 66 8.0 8.1 28. 73 8.0 8. 1 28. 87 8.0 22 E 4.2 24. 33 8.0 7.0 26. 82 8.0 7.2 28. 30 8.0 7.3 28. 75 8.0 7.3 28. 95 8.0 30 E 3.4 26. 78 8.0 5.3 27. 79 8.0 6.4 28. 15 8.0 7.0 28. 95 8.0 7.0 28. 65 8.0 Feb. 6 E 5.3 28. 04 8.0 5.3 27. 78 8.0 5.3 27.88 8.0 6.0 28. 56 8.0 6.0 28. 48 8.0 13 E 4.0 21.46 8.0 5.1 27. 75 8.0 6.2 28. 55 8.0 6.3 28. 77 8.0 6.3 28. 73 8.0 19 F 6.1 26. 13 8.0 6.3 26. 13 8.0 6.4 27.61 8.0 6.4 27. 84 8.0 6.4 28. 57 8.0 26 E 8.3 19.89 7.9 8.3 29.82 8.0 8.0 29. 49 8.0 8.0 28. 12 8.0 8.0 28.61 8.0 Mar. 3 F 6.1 22. 95 8.0 6.4 28. 27 7.1 27.61 8. 0 7.2 27. 27 8.0 7.2 27. 36 8.0 10 E 7.2 21.28 8.0 8.0 24. 00 8.0 8.0 27.43 8.0 8.0 27. 88 8.0 8.0 28. 17 8.0 17 E 8.3 25. 62 8.0 9.0 25. 77 8.0 8.3 25. 44 8.0 8.1 27.59 8.0 27. 73 8.0 26 E 8.0 23.95 8.0 7.4 28. 94 8.0 8.0 25.01 8.0 8. 1 25. 09 8.0 8. 1 25. 37 8.0 F 8 4 25 72 8.4 9. 0 25. 95 8.3 27. 48 8.3 27. 90 8.3 27. 83 15 F 10.0 26. 96 8.4 10.1 26. 60 8.4 10.0 27. 56 8.4 9.0 27.31 8.4 9.0 28.01 8.4 21 E 10.3 22. 92 8.2 10.3 23.01 8.2 10.3 26. 09 8.3 10.2 28.59 8.4 10.1 28. 19 8.4 29 F 11.3 25.88 8.4 11.4 26. 85 8.4 11.0 26. 04 8.4 10.1 27. 83 8.4 27.64 149604—37 3 454 BULLETIN OF THE BUREAU OF FISHERIES Table 8. — Temperature, salinity, and pH of water at different depths off Maple Point (Mud Bay) during 1 year — Continued Surface 3 feet 10 feet 20 feet 30 feet Date Tide Tem- Tem- Tem- Tem- Tem- pera- Salinity pH pera- Salinity pH pera- Salinity pH pera- Salinity pH pera- Salinity pH ture ture ture ture ture 19S2 ° C. °C. 0 C. ° C. ° C. May 6 E 11.4 26.00 8.4 11.3 26. 45 8.4 11.3 26. 97 8.4 11. 1 27.41 8.4 11. l 27. 52 8.4 13 E 12. 4 27. 64 8.4 12.3 27. 88 8.4 12.3 27. 93 8.4 11.2 28. 30 8.4 11.2 28. 48 8.4 20 E 12.3 27. 03 8.4 12.3 27. 17 8.4 12.2 27. 32 8.4 12. 1 28. 13 8.4 12.1 28. 56 8.4 27 F 14.3 27. 11 8.4 14.0 27. 20 13.3 27. 54 8.4 12.3 28. 01 8.4 12.2 28. 03 8.4 June 3 F 14.0 26. 74 8.2 14.0 27. 30 8.2 13.2 27. 66 8.2 13.0 28. 19 8.2 12.4 27. 98 8.2 10 E 19.2 27. 14 8.2 18.0 27. 03 8.4 15.4 28. 15 8.4 14.0 28. 19 8.4 13.4 28. 16 8.4 16 E 15. 1 27.60 8.2 15. 2 27. 81 8.4 14.4 28. 39 8.4 14.2 28. 21 8.4 14.0 28. 65 8.4 23 E 15. 1 28. 21 8.2 15.2 28.12 15.2 28. 39 8.2 14.4 28. 51 8.2 14.0 28. 46 8.2 July 10 F 16.2 26. 58 8. 1 16.2 27. 65 8.2 16.0 27. 65 8.2 15.3 27. 92 8.2 15.2 28. 03 8.2 21 E 16. 2 28.28 8.2 16. 1 28. 45 lfi.O 28. 12 15. 2 28. 43 15. 1 28.53 Aug. 2 F 17.2 28. 06 8.2 16.4 28. 37 8.4 16.4 28. 37 8.4 16.2 28. 64 8.4 16.2 28. 89 8.4 4 F 20.0 28. 33 8.2 18.0 28. 66 8.2 17.3 28. 99 8.4 17.4 28.82 8.4 17.4 28. 91 8.4 11 F 15.3 28. 19 8. 1 15.3 28.19 8.2 15.3 28. 35 8.2 15.2 28. 87 8.2 15.1 28. 00 8.2 18 F 17.4 28.06 8.0 17.4 28. 48 8.2 17.0 28. 08 8.4 16.4 28. 87 8.4 16.4 28.74 8.4 25 F 16.4 28. 93 8.2 16.0 28. 82 8.4 15.4 29. 08 8.4 15.0 29. 22 8.4 15.0 29. 08 8.4 Sept. 22 F 15.2 29. 63 8.4 15. 0 29. 22 8.4 15.0 29. 34 8.4 14.2 29. 47 8.4 14.2 29. 69 8.4 29 F 15.2 28.59 7.9 15.2 28.80 8.0 15. 1 28.80 8.0 14.4 28.99 8.0 14.4 29. 17 8.0 Oct. 5 F 14.3 29. 08 8.0 14.3 28. 95 8.0 14.2 29.18 8.2 14.2 29. 47 8.2 14.2 29. 75 11 F 13.1 29. 63 8.2 13.1 29. 70 8. 2 13.0 29. 87 8.2 13. 1 29.87 8.2 13.1 29. 85 8.2 15 E 13.4 28. 95 7.9 13.4 29. 27 7.9 13.2 29. 13 7.9 13.2 29. 33 7.9 13.2 29.84 7.9 18 E 12.4 29.20 7.8 12.4 29. 22 7.8 12.4 29. 13 8.0 13.0 29. 38 8.0 13.0 29. 16 8.0 25 F 12.4 29. 58 8.0 13.0 29. 31 8.0 13.0 8.0 13.0 29. 23 8.0 13.0 29.28 8.0 Nov. 1 E 10.4 26. 96 7.9 11.2 27. 11 7.9 11.2 29. 70 8.0 11.2 29.23 8.0 29.36 8.0 8 F 11.0 23.53 7.9 11.0 24.71 7.9 11.0 26. 40 7.9 11.2 29. 72 7.9 11.2 29. 42 7.9 17 E 11.4 27. 12 7.9 11.4 27.17 7.9 11.4 28. 82 7.9 11.3 28. 93 7.9 11.3 29.22 7.9 22 F 8.2 22. 92 7.8 10.2 28. 69 7.9 10.2 28. 50 7.9 10. 2 28. 69 7.9 10.2 28.79 7.9 29 E 10.2 26. 73 10.2 27. 18 7.8 10.2 28. 08 7.9 10. 2 28. 17 10.2 27. 95 7.9 Dec. 6 F 8.3 26.17 7.8 9.0 26. 65 7.8 9.0 27. 84 7.8 9.1 27. 85 7.8 9.1 28. 06 7.8 20 F 7.1 27.31 7.8 7.0 27.16 7.9 7.0 28.30 7.9 6.4 28. 22 7.9 6.4 28. 22 7.9 Note.— E=ebb; F=flood. To indicate in detail the changes in temperature, salinity, and pH at different depths, tables 7 and 8 are reproduced, showing the observations made over a period of 1 year in Oyster Bay and Mud Bay, at the points previously de- scribed. In Oyster Bay samples were taken at surface, 6, 15, 30, and 50 feet; in Mud Bay at surface, 3, 10, 20, and 30 feet. These tables give in detail both the seasonal variation in the water and the effect of depth. The surface water is, in some cases, quite Figure 12. — Variation in salinity and pH in a dike in Oyster Bay during a complete tidal cycle. See also figures 34 to 37. ^different from that below. Table 9. — Comparison of temperature, salinity, and pH of water at different depths during winter and summer off Corters Point, Oyster Bay; and Maple Point, Mud Bay Winter 1 Summer 2 Winter Summer 2 Depth Tem- pera- ture Salinity pH Tem- pera- ture Salinity pH Depth Tem- pera- ture Salinity pH Tem- pera- ture Salinity pH Maple Point: Surface 0 C. 5.8 24. 75 8.0 ° C. 16.5 27.82 8.16 Corters Point: Surface ° C. 6.4 26. 06 7.9 ° C. 16.3 28. 21 8.28 3 feet. 6.7 26. 37 8.0 16. 1 28. 16 8. 27 6 feet— 6.7 27. 05 8.0 16.3 28. 29 8.31 10 feet 7.2 28. 61 8.0 15.6 28. 29 8. 32 15 feet 7.0 27. 62 8.0 16.1 28. 21 8.3 20 feet 7.4 28. 61 8.0 15. 1 28. 53 8. 32 30 feet 7.1 27.91 8.0 15.8 28.25 8. 32 30 feet 8.0 28. 99 8.0 14.9 28. 57 8. 32 50 feet 7.1 27. 93 8.0 15.7 28. 36 8.3 1 December 1931, January and February 1932. 2 June, July, and August 1932. SPAWNING AND SETTING OF OLYMPIA OYSTERS 455 temperature; °c. Some of these data are given as averages for winter (December, January, and February) and summer (June, July, and August) in table 9 to indicate stratification of the water, for it has been shown by Nelson and Perkins (1931) that the behavior of oyster larvae may be determined by the salinity at different depths. The values for Maple Point (Mud Bay) are plotted graphically in fig- ure 13. In winter the salinity at surface and 3 feet is much lower than at greater depths but in summer the difference between surface and bot- tom is not so great. However, the pH in summer becomes lower toward the surface, probably because of planktonic animals, while in winter it is uniform from surface to bottom. In Oyster Bay (fig. 14) the salinity, temperature, and pH are almost identical from surface to bottom though during the winter the surface water is less saline and of slightly lower pH. The presence of the deep waters adjacent to the oyster grounds accounts for the high degree of stability indicated by these figures. Figure 13. — Vertical distribution of salinity, temperature, and pH off Maple Point (Mud Bay) summer (S) and winter (W) . Compare Oyster Bay, figure 14. Table 10. — Comparison of water near mouths of Mud Bay and Oyster Bay at ebb ( E ) and flood ( F ) tides during summer Mouth of Oyster Bay Mouth of Mud Bay Date Tide 6 feet 30 feet 6 feet 30 feet Tem- pera- ture Salin- ity pH Tem- pera- ture Salin- ity pH Tem- pera- ture Salin- ity pH Tem- pera- ture Salin- ity pH May 24 E ° C 11.1 28. 39 8.2 ° C 11.2 28. 53 8.2 ° C 10. 1 28.69 8.3 ° C 10.2 28. 39 8.3 F 12.0 27. 86 8.0 12.2 27.98 8.2 11.4 28. 30 8.2 12.0 28. 44 8.2 May 26 E 11. 1 28. 39 8.2 11.2 28.53 8.2 10.4 28. 13 8.4 11. 1 28. 27 8.4 F 11.3 27. 79 8.3 11.4 28. 06 8.2 11. 1 27. 85 8.4 11. 1 28. 37 8.4 May 29 E 12. 1 27. 98 8.2 12.2 27. 85 8.2 11.2 28.41 8.2 11.4 28. 35 8.2 F 12.4 27. 65 8.2 12.4 27. 66 8.2 12.4 27. 85 8.2 12.2 27. 75 8.2 May 31 E 12.1 28.08 8.2 12.1 28.08 8.2 11.0 28. 16 8.2 10.3 28. 56 8.2 F 12. 2 27. 83 8.4 12.2 27.79 8.3 12.0 28. 04 8.2 12.0 28.69 8.2 June 2__ E 13.0 28. 24 8.2 12.3 28. 06 8.2 11.3 28. 60 8.2 11.3 28. 66 8.2 F 13.2 27. 90 8.2 12.3 28.00 8.2 11. 1 28. 40 8.2 11.0 28.71 8.2 June 9 E 11.4 27. 66 8.2 11.3 27. 77 8.2 10.4 28. 12 8.2 10.2 28. 50 8.2 F 12.4 27. 77 8.2 12.4 27. 95 8.2 12.4 28. 16 8.2 12.3 28. 37 8.2 June 14 E 14.0 27. 99 8.4 13.4 28. 30 8.4 12.3 28.31 8.4 12.2 28.64 8.4 F 14.0 27. 64 8.4 13.3 27. 88 8.4 12.2 28. 35 8.4 11.4 28. 68 8.4 June 28 E 15.0 28. 21 8.2 14.3 27. 88 8.2 13. 2 28. 51 8.2 12.4 28.60 8.2 F 15. 1 28. 04 8.0 15.1 28. 04 8.0 15.1 28. 45 8.4 14.0 28. 04 8.4 July 5 E 15. 1 28. 48 8.2 15.0 28. 40 8.2 13.3 28. 85 8.2 13. 2 28. 59 8. 2 F 14.4 28. 48 8.2 14.3 28. 74 8.2 13.2 28.74 8.2 13.0 28.69 8.2 Average.. E 12.8 28. 10 8. 22 12.7 28. 17 8.23 11.5 28. 42 8.24 11.4 28. 50 8.25 Average. F 13.0 27.87 8.20 12.8 28. 02 8. 21 12.3 28.24 8. 26 12.1 28. 41 8. 26 Note.— E=ebb; F=flood. Further comparison of the two bays is shown by samples taken during summer at points near their mouths at depths of 6 and 30 feet at ebb and flood tides. (See table 10.) The water entering Oyster Bay is of a lower salinity than that going into Mud Bay, 456 BULLETIN OF THE BUREAU OF FISHERIES while at the upper end of the latter more fresh water flows in. As a result, in Mud Bay the oysters are subjected to frequent changes in salinity at different stages of the tide while in Oyster Bay the changes are relatively slight. These differences are considered below in the com- parison of the spawning and setting activities of oysters in the two places. It may be noted on the chart (fig. 4) that the mouths of Oys- ter and Oakland Bays are close to- gether but that the latter bay is entered through a long, narrow channel. It is of interest to com- pare the salinity in the dikes of the two bays at low tide. In table 11 salinities are given for dike 5 in Oyster Bay, a typical dike in Skookum Inlet, and one in Oakland Bay. Larger streams flow into Oakland Bay and elimination is less readily accomplished, so that even during dry summer weather the salinity is lower than in Oyster Bay. Water entering the latter at flood tide is of lower salinity than that which goes into Mud Bay, probably because of outflow from Oakland Bay and neighboring waters. Com- plete exchange of water with tides is accomplished much more slowly than in Mud Bay, the water in which is therefore higher in salinity as well as colder and not as favorable for the propagation of oysters. Figure 14. — Vertical distribution of salinity, temperature, and pH off Corters Point (Oyster Bay) summer (S) and winter (W). Compare Mud Bay, figure 13. Table 11. — Comparison of salinity of water in dikes at low tide in Oyster Bay, Little Skookum, and Oakland Bay in summer Date Oyster Bay Little Skookum Oakland Bay 1933 Salinity Salinity Salinity May 22 27.17 25. 68 23. 30 27. 05 24.03 May 26 26.91 23. 86 23.69 May 29 26. 88 25.09 23.66 May 31 27. 35 24. 22 27. 17 24. 71 June 5 27. 50 25. 91 24. 09 June 7 27. 59 24. 23 June 9 27. 27 25.47 24. 37 June 12 26. 77 25. 61 25. 39 27. 54 25. 76 June 16 27. 21 26. 10 24.83 June 19 27. 92 25. 97 25. 10 Date Oyster Bay Little Skookum Oakland Bay 1933 — Continued June 21 Salinity 27. 49 Salinity Salinity 25. 43 June 23 27. 43 27.01 25. 10 June 26 27. 16 26. 06 25. 52 June 28. 28. 07 24. 65 June 30 27. 92 24. 94 July 3 27. 69 25. 25 24. 90 July 5 27. 77 25.46 July 7 27. 11 25. 99 July 14 ... 27. 83 26. 97 26. 15 July 17 28.01 26. 68 26.05 July 19 28. 12 25.91 July 21 28. 21 27.16 24. 25 July 24 28. 13 28. 22 26.44 SPAWNING The native oyster of the Pacific coast, Ostrea lurida, is biologically similar to the European oyster, 0. edulis, in that it is hermaphroditic and viviparous. In his original studies of the native oyster, Stafford (1913, 1914) described the hermaphro- ditism, pointing out that in the gonad, or ovotestis, eggs and sperms may be seen close together. Until the recent work of Coe (1931a, 1931b, 1932a) no further exact infor- SPAWNING AND SETTING OF OLYMPIA OYSTERS 457 mation on the mode of reproduction in this species was published. Coe studied in detail the spermatogenesis and life history of this oyster near La Jolla, Calif., and found it to be protandric. He stated that germ cells mature in the 1-year-old oyster and that at first the individual is male. During the rest of its life it is alternately female and male, although at any one time germ cells of both sexes may be found in the gonad because seldom are all of the sexual products discharged before the next phase begins. Stafford (1915) stated that spawning involves the discharge of eggs or sperm balls from the gonad into the suprabranchial or cloaca! chamber from which they reach the mantle chamber. He described this activity as follows: Eggs and sperms are liberated from the gonaducts into the suprabranchial chamber, and make their way through the water-tubes and gill-slits to the branchial chamber, which also serves as a brood chamber. In doing this they are assisted by the pressure of their mass * * *. Sections of oysters at the spawning season show eggs in the cavities of the gills. They do not pass readily through the gill-slits on account of the narrowness of the latter, but with the increasing mass and pressure the gills become stretched and the slits enlarged, and besides the gills appear in places to suffer disintegration. This explanation is obviously inadequate, for one has difficulty in understanding how the small increase in pressure due to eggs would force them through the gill apertures. Also, it is clear that considerable coordination would be required to keep the cloaca! chamber completely closed and prevent direct escape of the eggs. Galtsoff (unpublished manuscript) studied spawning in Ostrea virginica and reached the more probable conclusion that suction, created by opening of the valves during spawning, draws the eggs through the small openings in the gills. Elsey (1935) found that the openings, or ostia, in the gills of 0. lurida and 0. gigas have a diameter proportional to that of the eggs. The eggs of the former are about twice as great in diameter as those of the latter, and the ostia are about one-third larger. As has been described by Nelson (1922), Galtsoff (1930b, 1932) and others, eggs are finally discharged from the mantle chamber in Ostreavirginica but sperms are washed out through the cloaca with the water pumped by the gills. Stafford considered that in the native oyster the sperms also pass through the gills as do the eggs. While Stafford may have actually observed the discharge of sperms in this manner, the writer has frequently seen them issuing from the cloaca as in other species. At the time of spawning the sperms of 0. lurida are in clusters, or balls, made up of from 250 to 2,000 or more sperms, according to the estimate of Coe (1931b), who stated further that in contact with sea water the matrix in which the sperms are imbedded disinte- grates, permitting them to swim free. Both Stafford and Coe considered that eggs are fertilized by sperms brought into the branchial chamber where the eggs are held, with the water pumped by the gills. Stafford thought self-fertilization might occur, though according to Coe’s interpretation this is unlikely. It is uncertain whether the sperms from one individual will stimulate spawning in functionally female specimens, as described by Galtsoff (1930b, 1932) for other species, though such is probable. The eggs are held in the anterior end of the mantle or branchial chamber adjacent to the gills and labial palps. Here they develop for a considerable period. It is remarkable that they are not swept out along the “waste canals” in the walls of the mantle which normally function to eliminate particles of silt and other rejected material. Stafford (1914) estimated that in British Columbia the period of develop- ment within the maternal “brood chamber” is about 16% days, while Coe (1931a) suggested “a period of approximately 10 to 12 days, perhaps” in southern California. Stafford’s work on this species is published in a series of six papers (1913, 1914, 1915, 1916, 1917, 1918) some of which are frequently referred to below. 458 BULLETIN OF THE BUREAU OF FISHERIES SIZE OF BROODS Oysters have always aroused interest because of the very large numbers of eggs which they discharge. Various means of estimating the number of eggs spawned by an individual have been employed with highly divergent results. Galtsoff (1930a) gave a brief review of the literature on the subject and indicated that previous esti- mates for the average-size female of 0. virginica vary from about 9 millions to about 60 millions. He made what is apparently the most thorough and accurate study by causing oysters to spawn and then making statistical counts of the eggs. He found that a female of 0. virginica discharged from 15 millions to about 114 millions in a single spawning period, and that after spawning had occurred the meat still contained a large quantity of eggs. During three periods of spawning one specimen of 0. gigas discharged a total of about 92 million eggs. The number of eggs spawned by a single individual throughout an entire season would be considerably greater than these counts. These species are of the oviparous type while the Olympia oyster is viviparous. In the waters of the United States are two viviparous species of oysters, 0. lurida, on the Pacific coast, and 0. equestris, described by Gutsell (1926) on the South Atlantic coast. The latter is too small to be of commercial use. 0. edulis, the European oyster, is also of this type and although the largest of the three it is much smaller than the common American oyster. Because of their small size the viviparous oysters cannot be expected to discharge the large quantities of eggs described above, but they have a considerable biological advantage in that every specimen is presumably capable of functioning each season as a female, while in 0. virginica only about half of the individuals are female. In the latter sex-change occurs, as recently studied by Coe (1932c, d), but not with the frequency found in the native oyster. The fact that the larvae are carried in the limited space of the branchial chamber in viviparous species would also appear to set a limit to the size of the broods. Moebius (1883) estimated, by an apparently satisfactory method, that the average brood of 0. edulis consists of about one million larvae. Stafford (1918) stated that the much smaller native oyster bears broods of about the same size, although his method of estimating the number is not clear. In order to establish with reasonable accuracy the number of larvae produced, counts were made of a number of broods. A gravid individual was carefully opened and the larvae rinsed from the gills and mantle. After killing them with formalin they were shaken in a measured quan- tity of water and exactly determined samples placed in a flat-bottom dish and counts made by means of a counting plate. Specimens of various sizes were used, though most of them were of market size. Counts were made of the separate broods of 13 oysters, and on the mixed broods of 2 groups of 6 oysters. Table 12 gives the results and includes measurements of the shells of the maternal parents and the stage of development of the larvae in those broods which were separately counted. The average of all 25 broods is 214,642 larvae per oyster, although single broods varied from about 70,000 for the smallest specimen to 355,000 for one of the larger ones. These specimens were in general considerably smaller than those used in the two series of mixed broods, which represent more fairly the number of larvae pro- duced by the standard market-size oyster. The average of the Oyster Bay series is 283,273 and of the Mud Bay series, consisting of somewhat smaller specimens, 247,199. The number of larvae produced obviously depends upon the size of the maternal individual and upon the degree of “fatness”, or amount of stored nourishment, which SPAWNING AND SETTING OF OLYMPIA OYSTERS 459 is required for the maturation of the eggs. The data here described indicate that the ordinary marketable native oyster produces a brood of from 250,000 to 300,000 larvae. Table 12 .—Number and stage of development of larvae in broods of IS oysters, and number of larvae in 2 groups of 6 broods each 121 125. 126. 127. 128. 129. 130. 131. 132. 139. 140. 141. 142. Oyster specimen no. Length Width Number of larvae mm mm Stage 32.7 30.3 28.0 29.5 29.0 29.7 31.0 23.5 28.2 36.2 29.2 28.5 36.8 23.5 24.0 23.5 27.5 23.0 21.2 25.8 19.2 24.4 26.2 27.0 23.3 26.2 130, 628 113, 142 95, 667 150, 600 156, 875 184, 114 355, 500 69, 490 126, 174 293, 473 213, 781 136, 666 171,818 Morulae. Straight-hinge 160/1-170 /i. Do. Morulae. Do. Late morulae. Morulae. Straight-hinge 160/i-170/i. Do. Early gastrulae. Gastrulae. Do. Do. 1 2. 3 4. 5. 6. 1 2. 3 4. 5. 6. Average. Average. Oyster Bay 38.1 40.6 45.7 38. 1 40. 6 45.7 33.0 30.5 30. 5 30.5 33.0 35.5 Mud Bay 283, 273 38.1 38.1 38.1 38.1 43.2 35.5 30.5 30.5 33.0 25.4 25.4 27.9 247, 199 It has been observed that most specimens carrying larvae appear to have dis- charged almost all of the eggs from the gonad, for the meats are generally transparent and watery. This is in contrast to Galtsoff’s (1930a) observation on Ostrea virginica that at a single spawning only a relatively small proportion of the eggs are discharged. The difference may be related to the fact that the native spawns alternately as male and female (Coe, 1931a, b). Stafford (1914) stated that the presence both in the parent and in the bay of all swimming stages shows “that the young do not all swarm out from the brood chamber of the mother at the same time, age, or size, but filter out gradually, perhaps during the gaping condition of the shell while respiration is going on in the parent.” In table 12 the stage of development of the larvae of 13 broods is given. In proportion to the size of the parent there does not appear to be a significant difference in the number of larvae per brood with respect to stage of development, although it is true that the largest broods consist of embryos. While this evidence neither confirms nor refutes Stafford’s idea of the gradual release of the larvae, it is probable that many of the free-swimming larvae of small size found in plankton samples may be the result of abortions. Disturbing the parent oysters appears to cause them to discharge broods of “white larvae” or those which have not developed valves. In many cases oysters have been taken up from the beds and placed in dishes of clean running seawater in the laboratory, and those specimens bearing young larvae observed to discharge them by opening and closing the valves. That such abortions occur in nature is shown bv statistical sampling of the oysters on certain beds. 460 BULLETIN OF THE BUREAU OF FISHERIES RELATION OF TEMPERATURE TO SPAWNING The development and discharge of germ cells are well known to be functions of the water temperature. As was shown in table 1 and figure 7 the average temperature of the water in southern Puget Sound begins to rise in early spring from the winter minimum of about 6° to 9° C., reaching a level of 18° to 20° C. in August. During the early months of spring the gonads begin active development and the stored nour- ishment of the bodies is used in the maturation of the eggs and sperms. The extensive researches of Stafford (1913), Churchill (1920), Gutsell (1924), Nelson (1928a, b, c), Prytlierch (1929), Galtsoff (1930b, 1932), and others have demonstrated that there is a certain minimum, or critical temperature below which little or no spawning occurs. In Ostrea virginica this minimum is 20° C. for spawning by the female, although the male is able to spawn at a lower temperature. GaltsofFs observations are particularly sig- nificant, for he was able to show by laboratory experimentation that at a temperature of 20° C. or above, the sexually mature female may be induced to spawn by adding either sperms or sperm extract to the water. Below 20° C. sperms would not stimulate spawning. It is of importance to note that GaltsofFs work confirmed the conclusion of previous investigators based upon ecological observations. In the same manner he was able to induce spawning in the Japanese oyster, 0. gigas, by addition of sperms when the temperature was 25° C. or higher, although more recently Elsey (1933) wrote that spawning could be stimulated at 22° C. Orton (1920) stated that the European oyster, Ostrea edulis, spawns when the water temperature reaches about 60° F. (15° to 16° C.). In 0. lurida Coe (1931a) found specimens bearing larvae whenever the water temperature was as high as 16° C., while Hori (1933) set the minimum at 14° and the maximum at 20° C. The exactness of such observations is not always clear, however, for it is questionable whether the temperatures given are actually averages or merely approximations, and maxima and minima are not stated. Prytherch (1929), in his thorough study of various phases of the spawning and setting behavior of Ostrea virginica near Milford, Conn., concluded that spawning begins when the temperature at high tide reaches 20° C. He considered that the lower pH (7.2) of the water at low tide, when the temperature was much higher, was the factor preventing spawning. At high tide, on the other hand, when the water reached about 20° C. and the pH was about 8.2 the oysters spawned. It is obvious that some factor in addition to temperature is necessary to stimulate spawning, and the suggestion that a high pH is required appears, from his results, to be well founded. Judging from these investigations, it is to be concluded that it is not the maximum tempera- ture on any particular day which must be up to the critical level, but the minimum, or high-tide temperature. It is clear, also, that the gonads must be at the required state of maturity before spawning will take place, regardless of temperature, for it was noted (Hopkins, 1931b) near Galveston, Tex., that the oysters were not ready to spawn during spring until the water temperature, after a rapid rise, reached about 25° C. At the same time, it has been well demonstrated that, with other factors favorable, spawning in 0. virginica begins when the high-tide temperature reaches the critical minimum of 20° C. It is important to know what conditions of temperature influence 0. lurida in this respect. (Hopldns, 1936.) A thermograph was placed on a frame in Oyster Bay so that the sensitive bulb was at the level of the oysters in the dike below. The records, therefore, represent with a high degree of accuracy the conditions of temperature of the water surrounding the SPAWNING AND SETTING OF OLYMPIA OYSTERS 461 oysters. The variation in water temperature on the oj^ster grounds during winter and summer and at different stages of tide has been described in table 1 and figures 5 and 6. Although he appeared to be uncertain of the specific factor involved in stimulating the beginning of spawning, Orton (1926) concluded that spawning of 0. edulis takes place primarily during the full-moon tidal period. Prytherch’s (1929) work indicated that the rise in water temperature, due to warming of the tide lands during the extreme low tides, was responsible for stimulating reproductive activity. He stated that in Milford Harbor, Conn., “the majority of the oysters spawned at the end of the July full-moon tidal period, when the water was brought to a favorable spawning temperature.” Nelson (1928 a, b) concluded that there is a definite relationship between the rapidity of the rise in temperature after the high-tide temperature reaches 20° C. and the time required for the initiation of spawning in 0. virginica. Pie found during several seasons that spawning started from 52 to 94 hours after the temperature of 20° was reached, depending upon the rapidity of the subsequent rise. His observation is in accord with Galtsoff’s (1930b, 1932) experimental finding that a sharp rise in temperature will induce discharge of germ cells. In the case of Olympia oysters, however, the grounds are diked and are all between low- and high-tide levels. The variation in salinity and pH at different stages of the tide is usually not very great (fig. 12) but the temperature is subject to wide fluctua- tions. When the tide is low the oysters are covered by only a few inches of water which quickly responds to weather conditions. During the day low tides in March, when the weather is favorable, the temperature may rise to 20° or 25° C., probably causing the maturation of eggs and sperms. Actual spawning does not usually begin until late in April or some time thereafter. During each season oysters were opened S37stematically on representative grounds in both Oyster Bay and Mud Bay to determine when spawning started and the number of adults bearing larvae throughout the season. Sampling was begun well in advance of the spawning period and consisted in the opening of 100 oysters on each of the test beds, 2 or 3 times a week. When a gravid individual was found the larvae were placed in a vial and preserved with formalin for examination in the laboratory. The method is described in greater detail with reference to the rate of larval development, and it is necessary here only to state that graphs of the results were made showing the percentage of oysters bearing larvae throughout each season. From these results and the thermograph records it is possible to correlate spawning activity and water temperature. Four graphs (figs. 15-18) are reproduced showing the percentage of gravid oysters on different days and the daily average and minimum temperature. Three of the figures refer to Oyster Bay, one to Mud Bay, but all agree with respect to the influence of temperature upon spawning. In 1932 (fig. 15) no specimens bearing larvae were found until May 17, when 12 out of 100 bore very young embryos. The temperature record shows a sudden rise at this time. For some days the average temperature had varied from 13° to 15° C., but the minimum, or high-tide temperature had been relatively stable. On the 16tli the minimum temperature rose from about 12° to over 13° C., and was followed by the sudden onset of spawning. For several weeks afterward, wlfile both minimum and average temperature steadily increased, spawn- ing was quite prolific. On the same graph the daily range of tide is plotted to indicate possible correlation with tidal periods, as described by Orton (1926). This is dis- cussed below. 149604—37 4 462 BULLETIN OF THE BUREAU OF FISHERIES Figure 15. — Daily average and minimum temperature in Oyster Bay during summer of 1932 as related to the frequency of spawning and range of tide. Open portions of columns refer to white larvae or embryos, shaded portions to conchiferous larvae. In Oyster Bay in 1933 (fig. 16) the graph has a somewhat different appearance because spawning was relatively light. On May 18, 2 percent of the oysters bore newly spawned eggs, though the minimum water temperature was only about 12° C. On the 24th or 25th, how- ever, spawning started at a considerable rate and con- tinued thereafter. At this time the minimum temper- ature was about 13° C. and the average about 14° C. It is remarkable that the few oysters which spawned early did not retain the larvae, for it may be noted on the graph that the first conchiferous larvae, of about 5 days development, were not found until the 31st. It is probable that a few oysters were able to discharge eggs but subse- quent low temperature caused abortions. Though this record is not as clear as could be desired, it in- dicates definitely that the first successful spawning took place when the minimum temperature reached approximately 13° C. During the same year, in Mud Bay, the beginning of spawning may be more closely correlated with a rise in minimum temperature from about 12° to over 13° C. (see fig. 17). The average temperature, at the same time, was about 14° C. The spring rise in tempera- ture in Mud Bay is charac- teristically late in compari- son with Oyster Bay and the entire breeding season is therefore later. In 1935, in Oyster Bay (fig. 18), spawn- ing started just after the minimum temperature reached a level of 13° C., although the average was between 15° and 16° C. These four series are selected for reproduction because they represent the most complete data at hand referring to specific localities. It is certain that it is not the low-tide temperature which initiates spawning in spring, for on many days preceding the time of beginning of spawning the water in the dikes would warm to 20° or 25° C. and remain so for several hours. It may be noted that, in the four cases presented, the average temperature varied over a wide range at the critical time. The minimum MAY JUNE JULY AUGUST Figure 16.— Daily average and minimum temperature in Oyster Bay during sum- mer of 1933 as related to frequency of spawning. Compare figures 15 and 18. SPAWNING AND SETTING OF OLYMPIA OYSTERS 463 temperature occurs commonly during the higher high tide and this shows a striking relationship to the onset of spawning. Prvtherch’s (1929) conclusion that the high- tide temperature must be adequate before spawning will begin appears to apply equally well to this species. Judg- ing from these data, the critical temperature for spawning may be placed at about 13° C., possibly from 12.5° to 13° C. It has frequently been noted that spawning is most likely to begin during or shortly after a period of neap tides. In spring and early summer, as shown in figure 15, at such times the mini- mum temperature is at a relatively high level. During a period of spring, or extreme tides, the tide flats warm in the sunshine and raise the temperature of the water coming in with the flood tides. The great range of tide in this place, 18 to 19 feet during a spring tide period, causes the colder water of the deep channels down the bays to reach the oyster grounds at high tide, while the warmer water is forced toward the head of the bay. During the neap tides a week later, however, the range may be only 11 or 12 feet, permitting the relatively warm water, resulting from the preceding low tide period, to remain over the oyster grounds. 30 10 Ct JUNE. JULY Figure 17.— Daily average and minimum temperature in Mud Bay dur- ing summer of 1933 as related to frequency of spawning. Compare figures 16 and 18. While the highest temperature is to be found in the dikes at low tide during a period of spring tides, the highest high-tide temperature may frequently occur in the neap-tide period, thereby inducing spawning. It is probable that this estimate of the critical temperature for spawning is not out of harmony with the results of Hori (1933) and Coe (1931a), who stated that the water temperature during spawn- ing was at least 14° and 16°, respectively. Their measurements appear to refer either to the average temperature or to that indicated by more or less frequent readings. As shown above, the average temperature at the time of the initial spawning is generally 14° to 16°. Figure 18.— Daily average and minimum temperature in Oyster Bay during summer of 1935 as related to frequency of spawning. 464 BULLETIN OF THE BUREAU OF FISHERIES SPAWNING SEASON The spawning period of the Olympia oyster appears to be quite different from that of 0. mrginica in Long Island Sound, as described by Prytherch (1929). In that place the oysters spawned prolifically on a particular day, when water conditions were favorable, so that the time of setting of the larvae could be exactly determined with reference to the time of discharge of the eggs. During several years, according to Prytherch, the first spawning was relatively light, and was followed about 2 weeks later, during the next period of spring tides when the water was warmer, by more general spawning. The native oyster, as described by Stafford (1914), spawns during a period of about 2 y2 months. According to this author, in 1913, in British Columbia waters, spawning occurred from about May 20 until the end of July, with the maximum spawning activity at the middle of June. On the coast of southern California Coe (1931a-1932b) found that spawning in this species continues for a period of about 7 months of the year. In the same manner, 0. virgivica on the Gulf coast spawns during many months (Hopkins, 1931b). Table 13.- — Percentage of oysters bearing larvae in Oyster Bay Date Number opened Number gravid Number not gravid Percent gravid Date Number opened Number gravid Number not gravid Percent gravid 11)31 ' 1931 May 25 _ 15 7 8 47 July 2 30 4 26 13 May 27 15 6 9 40 July 8 — 37 3 34 8 May 29__ 10 2 8 20 July 9 - 64 3 61 5 27 4 23 15 July 10. . . . 28 3 25 11 June 19 34 8 26 24 July 15. 53 0 53 0 June 21 20 3 17 15 July 24 30 0 30 0 11 1 10 9 July 27 21 3 18 14 June 27. . 29 2 27 7 Aug. 8 20 0 20 0 Table 14. — Percentage of adult oysters bearing embryos ( E ) and conchiferous larvae ( C ) in dikes 5 and S ( Oyster Bay) and dike B ( Mud Bay) during 1982 Date Dike 5 Dike S Dike B Date Dike 5 Dike S Dike B E c E c E c E c E c E C 12 July 1 13 1 4 6 19 12 2 5 1 21 20 3 4 13 4 8 4 24 32 3 5 5 27 25 11 6 6 10 30 26 13 8 3 6 12 31 15 25 9 9 2 15 40 11 10 8 3 14 3 13 12 6 1 4 15 28 13 11 8 6 12 16 14 10 8 1 25 15 16 6 10 7 30 16 5 11 5 15 18 17 4 8 13 11 12 4 1 12 19 7 14 7 19 20 18 4 15 3 5 2 8 21 5 2 16 2 11 22 2 11 4 17 5 4 4 25 3 8 4 6 18 4 10 26 6 20 5 12 4 27 4 6 21 5 28 8 22 5 30 6 23 5 4 2 24 4 2 4 2 2 10 27 5 12 4 1 28 5 5 1 4 29 2 5 8 2 31 i 1 No oysters opened before June 13. 1 No oysters opened before June 3. SPAWNING AND SETTING OF OLYMPIA OYSTERS 465 Table 15.— Percentage of adult oysters bearing embryos (E) and conchiferous larvae ( C) in dikes 5 and S (Oyster Bay) and dike B ( Mud Bay) Date Dike 5 Dike S Dike B Date Dike 5 Dike S Dike B E c E c E C E C E C E 0 1933 May 19 24 2 1933 June 27 1 22 7 6 28 11 9 2 12 26 4 6 July 3 4 3 14 12 29 10 10 2 9 14 31 10 4 10 8 6 6 4 June 5 6 10 ii 22 6 6 22 2 7 8 4 6 7 6 4 8 14 8 13 9 8 4 4 20 10 1 2 4 6 10 ii 15 6 12 14 10 4 22 17 10 4 13 12 3 18 3 16 11 13 19 6 17 11 8 20 10 19 9 21 32 21 8 12 20 14 22 2 21 15 4 8 20 24 2 7 6 22 17 15 26 2 1 2 23 12 i 20 31 1 24 15 17 Aug. 2 3 4 9 26 3 12 10 Table 16. — Percentage of adult oysters bearing embryos ( E ) and conchiferous larvae (C) in dikes 5 and T ( Oyster Bay) and dike B ( Mud Bay) during 1934 Date Dike 5 Dike T Dike B Date Dike 5 Dike T Dike B E 0 E c E O E C E c E c Apr. 19 27 7 June 1 6 6 6 2 6 5 7 3 8 28 3 2 1 8 6 4 May 1 2 3 7 9 8 14 1 20 19 6 11 1 2 5 9 18 3 12 3 6 10 2 17 9 17 13 3 5 1 11 16 29 14 3 8 12 10 3 15 9 1 7 14 9 18 13 18 16 6 17 14 15 22 2 4 18 10 23 5 18 25 1 6 3 19 11 24 29 4 4 2 25 1 10 1 20 July 6 9 1 26 18 1 4 3 28 6 7 9 10 3 29 1 17 13 1 1 2 30 7 4 6 21 1 31 15 25 1 1 No oysters opened until May 2. 2 No oysters opened until Apr. 28. Table 17.- — Percentage of adult oysters bearing embryos ( E ) and conchiferous larvae ( C) in dikes 5 and T (Oyster Bay) and dike A (Mud Bay) during 1935 1 No oysters opened until May 16. 2 Sampling discontinued. 466 BULLETIN OF THE BUREAU OF FISHERIES Table 18. — Number of spat caught on plane glass surfaces as determined by the angle of the surfaces: 0°, under horizontal; 90° , vertical; 180°, upper horizontal Average Angle of Area square Number of number of surface inches spat spat per 2,400 square inches 0° 2, 400 1, 195 1, 195 >45° 1, 200 42 } 181 2 45° 1, 200 139 >90° 2, 400 6 } * 2 qo° 2, 400 16 ‘135° 1,200 1 } 3 2 135° 1,200 2 180° 2,400 1 1 1 Perpendicular. 2 Parallel to general direction of current. A correct estimate of spawning activity of functional females was obtained by opening 100 adults two or three times weeldy throughout the season on selected typical beds. This mode of sampling has been carried on in two bays during 4 consecutive years. In 1931, the work was not begun in time to permit exact determination of the entire duration of the spawning season, but in table 13 the results of miscellaneous samples are given for comparison with later years. The table indicates, however, that after the end of May the proportion of adults bearing larvae became continuously less, until in July and August gravid specimens were only occasionally found. More complete data were obtained during the years 1932 to 1935 in both Oyster Bay and Mud Bay. Tables 14 to 17 summarize the results. In Oyster Bay two grounds were employed for sampling, one well up the bay, the other some distance below, or one high ground (dike 5) and one low ground (dike T). In Mud Bay samples were taken from dike B, a ground which is closely similar in all respects to most of those in the bay. The oysters were opened at the beds to eliminate the possibility of confusion due to the occurrence of spawning or abortion during transportation. In the tables the gravid specimens are divided into two groups, according to whether they bear unshelled embryos (E) or conchiferous larvae (C), in order to indicate more exactly the rate of spawning. While the complete data are given in the tables, a more significant picture of spawning activity may be obtained from figures 15 to 18, in which the percentage of gravid specimens on each day is shown as a column, the shaded portion of which represents conchiferous larvae, the open portion the embryos up through the trochophore stage. The tables and figures are not quite complete in that they do not include the very occasional gravid specimens that may be found as late as October, but these are too few to warrant any attention. In the tables it may be noted that the time when spawning begins varies over a period of about a month, depending upon climatic conditions which control the temperature of the water. While, in Oyster Bay, the first oysters bearing larvae were found at about the middle of May in 1932 and 1933, in 1934 spawning started a full month earlier. In Mud Bay the oysters generally start spawning some time later than in Oyster Bay, even early in June during some years. Spawning goes on at a significant rate for a period of about 6 weeks, although larvae may be found in some adults for as long as 5 or 6 months. The reproductory activity is best shown in figures 15 to 18, which indicate that after it once begins, the frequency of spawning slowdy increases to a maximum, then gradually diminishes until gravid individuals are found only occasionally. During most years, the rate of spawn- ing may be represented with fair accuracy by a simple symmetrical distribution curve, though sometimes (fig. 15) there is a later, secondary wave of spawning. In 1932, in Oyster Bay this later spawning was considerable, though not as important as the SPAWNING AND SETTING OF OLYMPIA OYSTERS 467 original activity early in the season. In other years it was not apparent that there was any definite renewed spawning activity. In his work on 0. edulis, Moebius (1883) said that as many as 20.6 percent of the adults bore larvae at once, and estimated from frequent observations, that at least 44 percent of the oysters produced broods during the season. Stafford's (1914) results indicate a comparable proportion bearing larvae at the same time, though he did not estimate the proportion of the population which produced broods. The data obtained during the sampling described above provide a means of estimating with some degree of certainty the total spawning activity throughout several seasons. Because of the frequency of the samples, it is possible to analyze the rate of development of the larvae, as is described below, and to detect the relative number of oysters bearing newly spawned eggs. From this one may reach an estimate of the total number of adults which bear broods. In 1932, as shown in table 14 and figure 15, the oysters spawned prolifically in Oyster Bay. At one time as many as 55 out of 100 carried broods. By referring each age group back to the date of spawning and then determining the total percentage of individuals spawning during the season it was possible to demonstrate that at least 1.5 broods per oyster were produced. That is, apparently all of the individuals bore one brood and at least half were gravid for the second time. In 1933, howe/ver, at the same place only about 75 percent of the individuals became gravid (fig. 16). During most years it appears that approximately 100 percent of the adults bear larvae, but only in 1932 was much greater spawning activity noted. Judging from all data available, it is probable that the variation in number of broods produced during different seasons is between about 75 and 150 per 100 adult oysters. In addition, the specimens would also spawn as males, as described by Coe (1931a, b). A source of error in such estimates is the possibility of abortions of young embryos which would, therefore, not be counted. It would be difficult to determine how frequently abortion of a brood occurs, but it is clear that it sometimes happens. DEVELOPMENT OF LARVAE Although Stafford (1914) reached the conclusion that the larvae develop normally for a period of 16% days within the maternal brood chamber it is probable that the method he emplojmd was unsatisfactory. He would periodically pry the valves of a gravid specimen partly open and take a sample of the larvae. Such handling of the specimen might readily result in a disturbance of normal function and interfere with larval development. A strictly biological method, therefore, would not appear to be adequate to solve the problems related to rate of development under natural conditions. It was necessary to use a system of sampling the oyster population and determin- ing at frequent intervals the stage of development of larvae in the various broods. On each of two typical grounds in Oyster Bay 100 adults were opened 3 times weekly. Larvae from gravid specimens were separately preserved in vials for later laboratory examination to determine their size or stage of development. By talcing samples at frequent intervals throughout the season it was possible to organize the results so that gravid specimens bearing broods of the same stage of development could be grouped and followed through the various stages. If on 1 day 10 percent of the oysters bore newly spawned eggs, 2 days later about the same number would be found with embryos of a certain stage. In subsequent samples the group would continue to recur until the larvae reached the size at which they are discharged. A single brood was found to consist of larvae of approximately the same stage, within relatively narrow limits. In no case were larvae of widely different stages found in a brood, and it may 468 BULLETIN OF THE BUREAU OF FISHERIES be concluded that an individual seldom, if ever, spawns as a female while carrying a brood of larvae. Records of the larvae taken in such collections were arranged in tabular form (fig. 19) to show the percentage of adults beeping larvae of different stages of develop- ment. By connecting the values from date to date and stage to stage, the age of each group is made clear. Division of the developmental process into 10 stages is largely arbitrary but at the same time convenient. While the em- bryonic stages, up through the trochopliore, are well defined, the only significant difference between straight-hinge larvae of different ages is in size. Measurements with an ocular micrometer were made of the larvae in each sample and while there is necessarily some uncertainty as to exact size, because of the variation among the larvae themselves, this is not great enough to be confusing. The example reproduced (fig. 19) is only one of many which are at hand but there is little difference between them. In the first place it will be noted that a total period of 9 to 1 1 days is required for development from eggs to the largest straight-hinge larvae. The general embryology has been well described by Stafford (1914) and is essentially the same as in 0. virginica, so that it is not necessary to describe it here. To illustrate some of the important stages there is reproduced in figure 20 a series of drawings from Hori (1933), an original copy of which Professor Hori kindly prepared for the writer. The figures are drawn accurately to scale so that they may be employed for the identification of larvae of the species. When they are discharged from the gonad the eggs are 100 to 105m in diameter, as stated by Stafford (1914) and Hori (1933). Development proceeds much more slowly than in the case of oviparous species. On the day after the eggs are discharged into the brood chamber they have be- come blastulae. At the age of 2 days they are usually in the gastrula stage, and 1 day later they have developed the swimming organ, or prototroch, and are actively swim- ming trochopore larvae. Usually on the fourth day the small valves may be seen developing on the dorso-lateral surfaces as a pair of clearly defined structures about 30 to 40^ long. This may be called the first conchiferous stage, and in figures 15 to 18 they are considered as such and in- cluded in the shaded portions of the columns. On the fifth day the valves have become complete and enclose the larvae entirely except when they are swimming with the velum protruding. In 0. virginica, as stated by Stafford (1913), “The age at which swimming begins may be considered to be about 5 hours, reckoned from fertilization * * As indi- cated above, the early embryology proceeds much more slowly in the viviparous 0. lurida, requiring between 3 and 4 days to reach the swimming stage. Larvae carried in the brood chamber are commonly spoken of as being either white or black. The expressions, “white-sick” and “black-sick” are frequently used Figure 19. — Graph showing percent- age of adult oysters bearing broods of larvae of each of 10 stages, as fol- lows: 0, eggs, or early segmenta- tion; 1, blastulae; 2, gastrulae; 3, trochophores; 4, first conchiferous larvae with incomplete valves; 5 to 10, straight n-hinge veliger larvae classified according to approximate length of valves; 5, 110-120^; 6, 120-130;i; 7, 130U40m; 8, 140-155m; 9, 155-170^; 10, 170-185ju- The per- centage values of larvae of definite size groups are connected to indicate rate of development. See text. SPAWNING AND SETTING OF OLYMPIA OYSTERS 469 in this respect. It is true, in general, that the young embryos show as pure white in the branchial chamber adjacent to the gills, and that older larvae in the well-advanced straight-hinge stage appear as a dark-gray or hluish-hlack mass. Yet one may not judge accurately the stage of development by estimating the depth of color. The Figure 20. — Developmental stages of larvae of Ostrea lurida. From Hori (1933). 1, sperm ball; 2, ovum; 3, ovum with polar body; 4, first cell division; 5, second cell division; 6, morula; 7, blastula; 8, gastrula; 9, early straight-hinge veliger; 10, veliger larva at time of discharge from brood chamber; 10-15, shells of free-swimming larvae of various sizes; 16, full grown larva of setting size; 17, young spat with growth of new shell. A, anus; A. M., anterior adductor muscle; A. P., posterior adductor muscle; C. S., crystalline style; D, digestive diverticula; F, foot; G, gill; Int., intestine; L. sh., larval shell; L. V., left valve; M, mouth; O, oesophagus; P. S., pigment spot; R, rectum; R. F., rudiment of foot; R. M., retractor muscle; R. V,, right valve; Sh., shell; 5. Sh., spat shell; St., stomach; T, teeth; V, velum. 149604—37 5 470 BULLETIN OF THE BUREAU OF FISHERIES pure whiteness of the eggs and young embryos slowly changes toward the gray after development of the valves. As the larvae grow older the mass becomes darker and darker, while the valves develop and pigment forms in the tissues. Frequently the largest larvae appear as black with a somewhat bluish tint, but it often happens that larvae of the largest size ever found in the brood chamber are only a medium gray. After 5 days of development in the brood chamber the larval valves become com- plete and the larvae are in the straight-hinge stage, so called because the dorsal border of the valves is straight, in contrast with the later pronounced umbo in the hinge region. The further developmental stages are arbitrarily arranged according to length of the larval valves, as measured under the microscope. These results appear not to be out of harmony with those of Stafford (1914) save that in the present case development is somewhat more rapid. His estimate of a period as long as 16J4 days may be correct for the locality in which he worked, or it may be due to his method of analysis. His data on water temperature, though in- complete, appear not to differ greatly from those taken in the present instance. Coe’s (1931a) estimate of a period of about 10 to 12 days required for development within the branchial chamber is more in accord with the records described above, which indicate a period of about 10 days, on the average. In his work on 0. edulis Orton (1926) estimated the gestation period, and stated (p. 219): An analysis of the spawning oysters into those with young embyros and those with mainly shelled larvae brings out the fact statistically that oyster larvae under natural conditions are retained in the mantle cavity a period of only 1 to 1)4 weeks from the date of their extrusion as fertilized eggs from the parent. In a later paper (1936) he wrote: The white-sick stage is thus normally of about 3 to 3J4 days duration, the grey-shelled stage about 1 34 to 2 days or less, and the black-sick stage of variable duration, probably 4 days or less. It seems probable that the oyster larva becomes fully developed normally in the sea in a period of 6 or 7 days and is expelled at an age between 7 and 10 days. Apparently his samples were not taken with sufficient frequency to permit analysis in the manner described above. Nevertheless it is probable that the period of larval development within the maternal brood chambers is not greatly different in the two species. Stafford used the word “swarming” to designate the final release of larvae from the maternal brood chamber, in contrast to the original spawning whereby the eggs are released from the gonad. It may be considered, in viviparous species, that swarm- ing is the delayed completion of the spawning process which in oviparous oysters, as described by Nelson (1928c) and Galtsoff (1930b-1932), is accomplished at once by means of rhythmic contractions of the adductor muscle. Whether discharge of the larvae is accomplished in the same manner is not known, but it has been observed that during abortions the embryos are forcibly ejected by means of shell movements. After discharge the larvae live and grow as free-swimming organisms for a period of approximately 1 month. The largest larvae found in the brood chambers are 180 to 185^ long> as described by Stafford (1914) and Hori (1933). The smallest straight-hinge larvae, 5 days after fertilization of the eggs, are about 110 to 120m long. Growth in length of the valves within the brood chamber proceeds at a rate of about 12m per day. At the time of setting the larvae are almost constant in size, of a length very close to 320 m- As is described in the following section, at least 30 days elapse between the time of discharge of the first larvae and the time when the SPAWNING AND SETTING OF OLYMPIA OYSTERS 471 first spat are found. During this period the larvae grow in length from 180/x to 320m, or an average rate of about 5m per day. Quantitative observations were not made on the abundance of larvae in the water, though plankton samples were frequently taken. In the case of this species it is not necessary to estimate the frequency and intensity of spawning from the age and abundance of larvae in the plankton, for this was more readily and accurately accom- plished by opening oysters periodically, as has already been described. Correlation between time of spawning and setting is considered below. Stafford (1914) wrote that the largest larvae found in plankton samples, and the smallest spat found on shells soon after attachment, had a length of 255m. Hori (1933), in a series of experiments on the artificial propagation of the species, stated that the length of newly-set spat is 320/z. He worked with oysters imported into Japan from Puget Sound, and was able to grow the larvae to maturity in dishes by feeding them ground sea lettuce ( Ulva ). Setting under the artificial conditions was successfully accomplished and his measurements of the spat are identical with those of the writer. A great many spat have been measured during these experiments and in no case has an attached spat been seen which was significantly less than 320m long. It is of interest also that Stafford’s (1913) measurements of larvae of 0. virginica at setting size are different from those of other observers. While he stated that newly-set spat are 380 m long, Prytherch (1934) measured them as 330m “across the shell at its greatest diameter.” Whether these differences are due to error in measurement or to actual differences in the size of the setting larvae may not be determined from Stafford’s works. It is certainly not impossible that under other environmental conditions the larvae of 0. lurida may be ready to attach at a smaller size. Stafford’s (1914) figures of older larvae and of newly-set spat do not show the umbo of the left valve to be as prominent as those of Hori (1933), which are in exact accord with observations of the writer. It is a possibility, also, that differentiation may proceed, under some circumstances, more rapidly than growth, with the result that the organs of the larvae reach the stage of maturity at a size smaller than that observed during the present work. There is, in addition, the possibility that the oysters studied by Stafford are of a distinct variety, having different characteristics from those of the Olympia oyster, though such is hardly likely. However, the fact remains that the measurements of Hori and the writer agree perfectly for both larvae and spat. Stafford’s measurements are exactly the same only up to the time of swarming. SETTING As was described in the preceding section, spawning is a long-continued process, and not as in some places an occurrence confined to a period of a day or two. It would therefore be expected that setting of larvae would also continue during a considerable period of time. It is important to know the frequency of attachment of larvae throughout the season and the relation of this to the rate of spawning. However, before describing the methods employed to analyze the results of setting during the different seasons it is necessary to consider the matter of cultch. Various questions have frequently arisen as to what constitutes the most favorable cultch and why oyster larvae attach as they do to certain surfaces. This phase of the investigation is of importance in throwing light upon the results described later. 472 BULLETIN OF THE BUREAU OF FISHERIES EFFECT OF ANGLE OF SURFACE Practical oyster growers have often noticed that most of the spat are to be found on the under surfaces of shells or other objects in the water. The question therefore arose as to whether this was due to some specific reaction of the mature larvae or merely to the fact that sedimentation and growth of algae and other organisms on the upper surfaces ordinarily prevent attachment of larvae. The opinion is frequently encountered that larvae seek the shadows and migrate to the relatively dark under surfaces. In order to determine, in the first place, whether larvae actually attach more abundantly to lower surfaces even when upper surfaces are equally clean an experiment was performed which provided various angles for comparison. Some of these results have already been published (Hopkins, 1935). Wire frames were made of galvanized hardware cloth of %-inch mesh, each frame holding three 8- by 10-inch panes of clear glass 1 inch apart and parallel. Some of the frames were designed to hold the panes in either a vertical or a horizontal position, others were so constructed that the panes were held at an angle of 45°. Thirty plates were used in each of the three posi- tions, horizontal, vertical, and at the 45° angle. These were placed in the water of a dike at low tide on one day and removed the following day when the dike was exposed. They were in the water only 24 % hours and may be considered as all equally clean since the time was too short for any considerable amount of fouling. Half of the vertical and 45° plates were placed perpendicular to the general direc- tion of tidal flow, the other half parallel, in order to indicate the effect of current. After removal from the water the plates were allowed to dry, removed from the frames, and the number of spat caught on all surfaces carefully counted with a binocu- lar microscope. In the analysis of the results the different angles are referred to as follows: 0°, under horizontal; 180°, upper horizontal; 45°, under, and 135°, upper, surfaces of the 45° panes; and 90°, vertical. In table 18 the results are presented in detail. The effect of the angle of the surface on its efficiency as a collector of spat is further illustrated in figure 21, in which the average values in table 18 are plotted graphically. The correlation between angle and number of spat caught is remarkably close, when it is considered that in each case, except for the vertical plates, the area of surface was only 2,400 square inches. It may be noted in the table that the panes which were parallel to the direction of flow of tide caught definitely more spat than those perpendicular to the current. This would appear to indicate that setting may be proportional to the rate of current. Prytherch (1929) noted that larvae of Ostrea virginica set most abundantly on the leeward side of objects in the water, where the current is reduced to a minimum, and concluded that the heaviest setting takes place when the current is least. The above results suggest that the larvae of the native oyster react differently. The values represented in the graph (fig. 21) are the totals, including plates which were both perpendicular and parallel to the current. It might possibly be a more 1200 1000 600 2 cr> EC £ coo £ Z 400 200 0 Figure 21.— Number of spat caught per 2,400 square inches of glass sur- face at different angles: 0°, under horizontal; 180°, upper horizontal; 90°, vertical. ANGLE OF SURFACE SPAWNING AND SETTING OF OLYMPIA OYSTERS 473 exact picture of setting behavior if only the results obtained with the parallel series were plotted. However, to do so would hardly alter the curve, since the difference in efficiency of the various angles is tremendously greater than that between plates of the same angle but in different positions with reference to tide. Also, in its present form, the curve is more typical of natural conditions, considered from an ecological point of view. The results of these experiments have an obvious practical application, for they point out the desirability of furnishing cultch which has a large amount of the ideal under horizontal surface. It is also of practical importance to know that the more freely the current flows along the surfaces the more spat will be caught, presumably because more water, bearing larvae, comes into contact with the surfaces. A com- parison of the efficiency of two types of manufactured spat collectors serves to illus- trate well the commercial application of the results described above. Prytherch, in 1929, used the egg crate filler, coated thinly with concrete, as a collector of seed oysters on the Atlantic coast. (See Bureau of Fisheries Document 1076, Improved Methods for the Collection of Seed Oysters.) These fillers, made of cardboard, provide a large amount of surface for the attachment of larvae, and the rough concrete surface is particularly favorable for setting. Such spat collectors are spread on the seed grounds at the correct time and have proven to be highly effective. The method is in use on some of the Olympia oyster beds where the fillers remain covered at low tide because of the dikes (see fig. 2). However, the experiments described above show that the best surface for the catching of spat of the Olympia oyster is the under-horizontal, and the egg crate filler lies on the bed with all partitions in the vertical position. Also, the filler, lying on the oyster ground, does not permit the free flow of water over the surfaces and the individual cells contain relatively still water, frequently resulting in their filling with silt. In order to develop an efficient collector for Olympia oysters it was, therefore, necessary to design a modification of the egg crate filler which would provide a large amount of under-horizontal surface and also permit the water to flow freely over all the surfaces. This special collector was made like the egg crate filler, but the individual parti- tions are twice as wide, and consists of two rows of six cells each. The cells are 2 inches square by 4 inches long (see fig. 22). It lies naturally on the ground in a posi- tion which is at an angle of 90° to that of the egg crate filler. Both vertical and hori- zontal surfaces are present, and the water flows freely through the cells. As they are now used they consist of two rows of seven cells, making the total area almost the same as that of the egg crate filler, but the counts of spat which were made for comparison refer to the original type. Table 19. — • Comparison of efficiency of 2 types of manufactured spat collectors Number of spat per 1,000 cm2 Total area (cm1) Total number of spat Average number of spat per 1,000 cm1 Under- horizontal Vertical Upper- horizontal Egg crate filler __ 427 1,066 4, 724 3, 464 2,064 4, 775 427 1, 378 3,235 28 Table 19 gives the results of counts of spat caught on egg crate fillers and special collectors which were put on the same grounds at the same time and removed together 3 months later. Each series is the average obtained by analysis of three collectors. 474 BULLETIN OF THE BUREAU OF FISHERIES The table shows that the average number of spat caught on 1,000 square centimeters of the egg crate filler is 427, while on the special collector the average is 1,378. This value refers to all surfaces, which in the egg crate filler are vertical, and in the special collector both vertical and horizontal. It was conclusively demonstrated that the special design is more than three times as effective as the standard filler. It is to be expected that the difference between the numbers of spat caught on horizontal and vertical surfaces would not be as great for concrete-coated paper as for plane glass, as described above, since the roughness of the concrete provides a large amount of surface which may be at all angles. The vertical concrete wall has a large horizontal component in the projecting grains of sand which are large in pro- portion to the oyster larvae. Therefore, the values in table 19 do not exactly fit the curve obtained with plane glass (fig. 21), but the points fall along a more gently sloping curve. The significance of the observation cited above, that flow of water along surfaces is necessary for most efficient setting, is well demonstrated by the difference between the numbers of spat caught per 1,000 square centimeters of vertical surface on the two types of collectors. The special collector is about two and one- half times as effective, considering only the vertical walls. On most grounds egg crate fillers collect large quantities of silt which fills the cells and kills the spat. Frequently only those caught on the upper edges survive. In the case of the special collector, however, the water is able to flow through the cells and prevent deposition of silt. Even on soft ground only the bottom layer suffers a loss as the collector settles into the surface. Such an example is shown in the upper photograph of figure 22, while the lower surface of a collector placed on firm ground shows little mortality. It would appear that the habit of attaching primarily to an under surface has the function of protecting the delicate young spat from various unfavorable conditions, such as hot sunshine, and deposition of silt. Immediately suggested by the results is the possibility that the larvae are photosensitive and react negatively, causing them to collect in the shadows where they set. Such a view would imply that setting takes place almost entirely during the day, and that at night the larvae would not concen- trate on under surfaces. To determine whether light is a factor in the setting behavior of larvae, two sets of wire frames, each containing 15 glass plates, were placed on an oyster ground so that the plates were horizontal and allowed to remain for about 24 hours. The plates of one set were painted black on the upper surfaces, the others left clear. Both surfaces of each pane of glass were carefully examined with a binocular miscroscope and all spat counted. On the lower surfaces of the black glass 435 spat had caught, while on similar surfaces of the clear glass 616 were counted. Not one was found on the upper surfaces of either group. It is not considered significant that the clear glass caught more spat than the black, but it is important that the shadow under the latter did not result in any increase in the catch. In other experiments, which are described in a later section, it was demonstrated that larvae set as well at night as during the day, and that in all cases the lower sur- faces of horizontal panes receive almost all of the spat. It is therefore obvious that light is not an orienting factor in the setting behavior of larvae of tins species. The pigment spots of mature larvae have often been looked upon as possibly sensitive to light, but Prytherch (1934) concluded that larvae are not photosensitive, and that the pigment spots have an entirely different function. The present results confirm his conclusion. U. S. Bureau of Fisheries, 1937 Bulletin No. 23 Figure 22. — Photographs of special spat collectors bearing oysters about 1 year old. Collector as a whole is shown above in upside-down position. Lack of spat on bottom due to soft ground on which it was placed. Lower photograph shows an under horizontal surface. Total length about 12 inches. SPAWNING AND SETTING OF OLYMPIA OYSTERS 475 The graph (fig. 21) suggests that the effect of angle of surface on the number of spat caught is purely mechanical and not due to any definitely biological reactions of the larvae. Hori (1933) stated that swimming larvae commonly are in an inverted position, with the velum uppermost. This has been observed also by the writer and is well illustrated in Prytherch’s (1934) work on larvae of Ostrea virginica. The velum projects through the valves as a flattened, ciliated swimming organ, while, the heavier shell hangs downward. The foot with which the larvae must adhere is beside the velum and projects more or less upward, although it is extensible in all directions. It is most likely that the swimming larva, as it comes into contact with a surface from below, is able to hold on with the foot, while on coming down upon a surface it is the hinge portion of the shell that touches. In this manner, as the angle of the surface departs more and more from the under horizontal there is constantly less chance of the foot touching. This interpretation, in effect, is that- the observed results are due to accidental contact of the foot with the surface as the larvae is swim ming and being washed about by tidal currents. Prytherch’s descriptions of setting, as directly observed with the microscope, suggest that larvae of 0. virginica may react differently in this respect. If the above-described interpretation is correct it would be expected that in places where the water is highly turbulent the larvae would frequently be turned over so that they might also catch on upper surfaces. It has frequently been observed on oyster grounds near Olympia, in places where the water flows over dikes, that the rocks and shells close to the dikes bear spat also on upper surfaces. METHOD OF DETERMINING FREQUENCY OF SETTING It is of considerable importance to know the duration of the setting season and the times when maxima are reached. For this purpose it was necessary to plant cultch periodically during the entire season. The system finally adopted, after the first year, was to plant a wire bag of shells on each experimental ground and allow it to remain for 7 days. It was then removed and brought to the laboratory. As one bag was taken from the ground a new one was planted and allowed to remain for the following 7 days. It was found necessary, however, to carry on at each place two such 7-day series so that one overlapped the other. In one series, for example, the bags would be in the water from Monday until the following Monday; in the other series, from Friday until the next Friday. A clean lot of shells was therefore put into the water every 3 or 4 days. After being brought from the grounds the shells were allowed to dry and counts made of the number of spat caught. Bags were made of 1-inch mesh galvanized wire netting and were about 30 inches long by about 8 inches in diameter. Each held something over tliree-quarters of a bushel of Japanese oyster shells. These shells were preferable because of their large size, generally 4 to 6 inches long, and the white color of the inside surfaces. In the bags the shells were held at all possible angles, eliminating any error that might be traceable to the angle of the surfaces. Counts were made only of the spat on the inside surfaces, because of their color and smoothness and because the outside surfaces are too rough, and often lamellate, so that all spat are not readily seen. This is not difficult to understand when it is realized that the shells were in the water only 7 days and the oldest spat a millimeter or less in diameter. Two bags of shells were left in the water for a month and a half to allow the spat to grow to a large enough size to permit accurate counts of the number on inside and outside surfaces. The results are summarized in table 20. In one case 33 percent were on inside surfaces, and in the other, 36.1 percent. The 476 BULLETIN OF THE BUREAU OF FISHERIES larger number on the outside surfaces is probably due to roughness as well as to greater area. The two series average 34.6 percent inside and 65.4 percent outside. In calculating the number of spat caught in a bag the counts made on inside surfaces are considered as 35 percent of the total. This proportion would not be correct if the shells were spread directly upon the grounds, for most of them, because of curva- ture, fall with the outer surfaces down. Table 20. — Number of spat caught on inside and outside surfaces of shells in wire bags Records were kept on the basis of a standard-size bag of shells, and at the time of counting the shells were carefully measured in a box of definite dimensions. Although some bags were fuller than others, the remeasurement eliminated any error from this source. It is true, of course, that it is impossible to obtain exactly uniform shells, and that at times they might be relatively smaller or larger than the average. It was noted, however, that this had little to do with the results. Generally, a standard bag contained 125 to 150 shells, though some held as many as 200 or as few as 100. When the shells are small they provide more surface per unit of volume, but impede the free circulation of water. When extra large, the water flows freely among them, though the area of surface is not so great. These factors appear to offset one another, in the case of bags of small diameter such as were used, and the resultant averages are relatively consistent. To standardize the system of counting and calculating the total number of spat caught by this method, 100 unselected shells from each bag were examined carefully, on their inside surfaces, with a binocular miscroscope, and every spat counted. To avoid error, circles were drawn around the spat as they were counted, for in some cases as many as 600 or 700 spat were found on the inside surface of a shell. The number of spat on the inside surfaces of 100 shells was used as a basis for calculating the number on both surfaces of all shells as described above. The variation between the different shells, with respect to the number of spat caught on the inside surfaces, was tremendous. In one typical case the extremes were 0 and 730. This was obvi- ously due to differences in the angles at which the shells are held as well as to their size and their position in the bag. On the shells in this particular bag, which was selected for SPAWNING AND SETTING OF OLYMPIA OYSTERS 477 description because it was in the water during a period of abundant setting, an average of 358 spat per shell were caught. During the first season the method employed was somewhat different. Bags of shells were left in the water for various lengths of time, the spat counted, and a great many of them measured for the purpose of determining from their size distribution the tune at which setting took place. The method of measurement was discarded because of the large error, due chiefly to differences in rate of growth during periods of spring and neap tides, which affect the low-tide temperature, and also probably to specific peculiarities of the different spat. The records for 1931 are therefore not as exact as those for the following years. Experimental series of this kind were carried on in the two principle oyster- producing bays near Olympia during five consecutive seasons. In addition, two other bays were studied for one season. The results of the analyses are given in detail in the following section. The time of beginning of setting was determined by daily examination of shells on the grounds until spat were found. SETTING SEASONS, OYSTER BAY SEASON OF 1931 -I.,? 3 s Figure 23. — Average number of spat caught daily on bags of shells left in dike 5 (Oyster Bay) for different periods during 1931. Daily range of tide is also shown. Observations were not begun in 1931 in tune to permit obtaining of complete data on spawning, though it is evident (table 13) that most of the larvae were dis- charged during May and early June, for after this time few gravid specimens were to be found. Bags of shells, however, were left in the water for various periods through- out the season. The first spat were found on June 12, and they were of the size of mature larvae, with no new growth, so that this date may be considered as the time when larvae began to attach. The results of counts of spat on the shells are given in table 21. Only two grounds were stud- ied in detail, and the samples from only one of these (dike 5) were thoroughly analyzed. The other series was checked sufficiently to show that the course of the setting season was iden- tical on the two grounds. Figure 23 gives a better picture of the abundance of larvae setting during different portions of the season. After setting had started on June 12, it continued for a period of about two weeks. Comparison of the graph (fig. 23) with table 21 is necessary for reaching an understanding of the season as a whole. The first bag of shells which caught spat was in the water for 15 days, from June 12 to 27, and caught a total of 6,065 spat, or an average of 404 per day. The bag which was in the water from June 30 to July 24 caught a negligible total of only 35 spat, and none at all was caught between July 10 and 18. The actual significant period of setting, there- fore, was between June 12 and 30, and during the last 10 days of this time the average daily catch was very slight. After this period, during which almost no larvae attached, a very profuse set began to take place. The table shows that the beginning of the second setting period was probably just after July 24, since the shells removed on that date bore a few spat. Between this date and August 3, when shells were brought in again, a great amount of setting occurred. The bags which were iu the water from July 30 to August 1 1 caught 478 BULLETIN OF THE BUREAU OF FISHERIES fewer spat than those in between , indicating that the peak of frequency of setting was probably between July 24 and 30. During the remainder of the season the rate of setting became gradually slower. Table 21. — Number of s-pat caught on hags of shells on two grounds in Oyster Bay, 1931 June 12_ June 18. June 20- Jane 30. July 10.. July 15_. July 18.. July 27.. July 30.. Aug. 8- Aug. 11. Aug. 15. Aug. 22. Aug. 27. Sept. 8_. Sept. 24. Oct. 7... Date planted June 27. June 30. July 10.. July 24_. July 18.. July 24_. Aug. 3.. do.. Aug. 11. Aug. 27. Aug. 22. Aug. 29. Aug. 27- Sept. 11. Sept. 25. Oct. 7.. Oct. 20. . Date removed Number of days 15 12 20 24 8 9 16 7 12 19 11 14 6 15 17 13 13 Dike 1 Total number of spat Number of spat daily 7,286 1, 900 485 158 22, 320 40, 520 35, 980 1, 395 5,789 2,998 14, 277 6, 957 1, 298 497 2,691 1,163 150 89 Dike 5 Total number of spat Number of spat daily 6, 065 404 2, 480 207 697 35 35 1 0 0 30 3 28, 071 1,754 39, 040 5, 577 31, 063 2,588 10, 863 67! 16,866 1, 533 7. 240 517 1, 340 68 7, 769 618 1,474 87 817 63 10 0.7 Table 22. — Number of spat caught on bags of shells on 3 grounds in Oyster Bay throughout the season of 1932 [Counts were made on only a few of the dike 5 series for comparison] Date planted Date removed Number of days Dike 1 Dike 5 Dike S Total num- ber of spat Number of spat daily Total num- ber of spat Number of spat daily Total num-‘ ber of spat Number of spat daily June 27 3 4, 286 1,429 527 178 July 1. 4 21, 694 5, 423 3, 913 978 July 4 5 30, 083 6, 016 3, 463 693 July 1 July 8 7 42' 037 6, 005 42, 040 6, 006 5, 342 763 July 4 July 11 7 38, 156 5, 451 16, 263 2, 323 July 8 July 15 7 36, 657 5, 237 3, 494 499 July 11 July 18 7 19, 981 2, 854 19, 963 2,852 4,756 679 July 22 7 9, 577 1,368 1,888 269 July 18 July 25__ 7 2, 654 379 2, 960 423 '281 54 July 22. __ July 29 7 2,614 373 1,220 174 July 25. _ Aug. 1 7 3' 352 479 1,474 210 July 29 Aug. 5 7 4, 981 712 4, 043 577 1,986 284 Aug. 1 Aug. 8 7 7, 146 1,021 4, 206 601 1,820 260 Aug. 5 Aug. 12 7 13, 132 1,876 10, 951 1,564 2, 665 381 7 11, 229 1,604 2, 214 216 Aug. 12 Aug. 19 7 25, 618 3, 659 18, 420 2, 631 2,086 298 Aug. 15. Aug. 22 7 31, 406 4, 486 31, 694 4, 528 3,315 473 Aug. 26 7 31, 824 4, 546 6,891 984 Aug. 22 Aug. 29 7 51,877 7,411 37, 023 5,289 4, 894 699 7 15, 591 2, 227 1, 521 217 7 6, 419 917 1,389 198 7 10, 694 1, 528 333 48 7 4, 441 ' 634 803 115 Sept. 16 7 2, 029 289 423 60 Sept. 12 Sept. 19 7 80 1, 374 196 843 120 Sept. 16 __ Sept. 23. __ 7 983 140 311 44 Sept. 19 .. Sept. 26 7 1, 468 209 1,802 257 651 93 Sept. 23 7 1, 466 209 655 93 Oct. 3 . 7 ' 211 30 307 44 Oct. 7 7 534 76 141 20 Oct. 3... Oct. 11_ _ 8 471 59 374 47 Oct. 7... Oct. 14 __ 7 160 23 360 61 Oct. 11 Oct. 17 6 69 11 Although not as complete as the results for later years, these records proved that two well-defined periods of setting occurred with their maxima about 6 weeks apart. The finding was particularly significant since the later setting was so much more profuse than the earlier, although the oyster growers had planted cultch only early in the SPAWNING AND SETTING OF OLYMPIA OYSTERS 479 season. It appears to have been a common idea that only by planting cultcb in time for the initial setting could a good catch be obtained. In 1931 this was certainly not the case. The larvae continued to set in small numbers up until about the middle of October. Either dike 5 or dike 1 (Olympia Oyster Co.) was used for experimentation during all five seasons. The former is at a level of about 1.5 feet above the zero tide, while the latter has a height of about 3.5 feet. Both grounds are considered as excellent seed-catching areas. According to the results obtained in these experiments the 2 grounds are almost equally favorable. As may be seen in table 21 the number of spat caught on bags of shells which were in the water at the same time is almost iden- tical for the two dikes. Considering only those bags which were in the dikes during the same periods, the total number of spat caught was 137,074 in dike 5, and 13 6,581 in dike 1. -Average number of spat caught daily per bag of shells on two grounds in Oyster Bay, 1932. Each bag was in the water for 7 days. Range of tide is also shown. SEASON OF 1932 < In 1932 the sys- s tern of carrying on two overlapping series of bags of shells for each ground was put into effect. Although involving the count- figure 24. ing of spat on a great many more shells the results well justify the effort because of the increased accuracy obtained, permitting more exact determination of the times of maximum and minimum frequency of setting. Complete counts were made of two series and on sample bags of another (table 22). No spat were found until June 26, although daily observations were made. After this date setting continued until the middle of October, a total period of over 3% months. Represented graphically (fig. 24) the more numerous samples make an excellent picture of setting activity. Dike S (Steele ground) is about 2 miles down the bay from dike 1 and is removed from the larger area in which most of the beds are located. The total number of larvae caught is therefore considerably less than on grounds up the bay. There are two distinct major periods of setting having their maxima during the first few days of July and near the end of August respectively. The graphs repre- senting the two grounds are substantially alike with respect to the times of the maxima and minima and differ only in the total number of spat caught at any time. As was found in the previous year the season consists of two distinct setting periods, with the maxima in the present case approximately 8 weeks apart. 480 BULLETIN OF THE BUREAU OF FISHERIES Table 23. — Number of spat caught on bags of shells on two grounds in Oyster Bay during the season 1933 of Date planted June 26_ June 30. July 3... July 7... July 10. July 14. July 17. July 21.. July 24.. July 28.. July 31. Aug. 4. . Aug. 7.. Aug. 11. Aug. 14. Aug. 18. Aug. 21. Aug. 25. Aug. 28. Sept. l._ Sept. 4_. Sept. 8.. Sept. 11. Date removed Number of days Dike 5 Dike S Total num- ber of spat Number of spat daily Total num- ber of spat Number of spat daily July 3 ‘7(1) 951 951 196 196 July 7 1 7(5) 16,200 3, 240 1,815 363 July 10 7 27, 566 3, 978 9, 416 1,345 July 14 7 23, 248 3, 321 5, 455 779 July 17 7 9, 900 1,414 2, 450 350 July 21 7 11,480 1, 640 2,225 318 July 24._ 7 14, 483 2, 069 1,060 151 July 28 ... _ 7 20, 082 2,809 3, 144 449 July 31. .. 7 16, 302 2, 329 4, 745 678 Aug. 4 7 8, 320 1, 189 4, 196 599 Aug. 7 7 9, 551 1,364 6, 619 945 Aug. 11. _ 7 8, 167 1, 167 1,297 185 Aug. 14. .. 7 7, 277 1,039 2, 838 405 Aug. 18. ... 7 23, 040 3,291 16, 872 2, 410 Aug. 21 ... _ ... ... 7 16, 766 2, 395 12, 954 1,850 Aug. 25 7 16, 778 2, 397 5, 928 847 Aug. 28 7 15, 784 2, 255 2, 040 291 Sept. 1 7 6, 003 858 823 117 Sept. 4 ... 7 3,602 514 343 49 Sept. 8 ... . _ ._ . 7 573 82 Sept. 11. . _. 7 467 67 Sept. 15 7 120 17 Sept. 18 7 100 14 ■ Bags in water for 7 days, but setting started some time after they were planted. Table 24. — Number of spat caught on bags of shells on grounds in Oakland Bay and Little Skookum in the season of 1933 Date planted Date removed Number of days Oakland Bay Little Skookum Total num- ber of spat Number of spat daily Total num- ber of spat Number of spat daily July 7 7 240 34 July 3 July 10. __ .. ... i 7 1,640 234 480 80 July 7 July 14 _ 7 982 140 577 82 July 17. 7 920 131 246 35 July 21 7 1,630 233 428 61 July 17 July 24 7 4,731 676 243 35 July 21 _ July 28 7 10, 620 1,517 791 113 July 24 _ _ __ July 31 7 5,650 807 440 63 July 28 Aug. 4_ _ _ 7 1, 340 191 403 58 Aug. 7 _ 7 680 97 157 22 Aug. 11 . _ 7 297 42 Aug. 18 14 1,560 112 Aug. 14_ 7 0 0 Aug. 21 14 3, 563 255 7 200 28 Aug. 21 _ _ __ _ 7 2,591 370 Aug. 25 7 2,614 373 709 101 Aug. 28 7 1, 131 161 Aug. 21 Sept. 21 31 86 3 Sept. 1. __ _ __ _ 7 0 0 Aug 28 8 185 23 1 In Little Skookum 0 days. SEASON OF 1933 In 1933 complete counts were made on shells planted in the same periodic manner in dikes 5 and S, and in addition two other adjacent bays were included (see tables 23 and 24). The three bays are treated together because of the similarity of the setting periods. Little Skookum (Skookum Inlet) is a small bay which branches off from Oyster Bay about half way between the mouth and the upper end, and well down from most of the oyster grounds (see chart, fig. 4). The mouth of Oakland Bay is very near to that of Oyster Bay and it would therefore not seem improbable that the water entering the two bodies of water is almost identical. As a result of pollution of the water in this region by waste liquor from a pulp mill several years before, the oyster beds in Oakland Bay and Little Skookum had been seriously depleted and the supply SPAWNING AND SETTING OF OLYMPIA OYSTERS 481 of spawning oysters reduced to a low level. For this reason the catch of seeds could not be expected to be as great as in the upper end of Oyster Bay, which retained more spawners. The results obtained on two grounds in Oyster Bay (dikes 5 and S) and on one typical ground each in Oakland Bay and Little Skookum are presented graphically in figure 25. On all of the grounds there were three periods during which setting was especially profuse. The graphs show the three max- ima as occurring early in July, at the end of July, and at the middle of August. The last maximum was somewhat later in Oakland Bay than in the other areas, but the first two were in all cases at almost the same time. The total length of the setting season was only about months, from the beginning of July until the middle of September, a full month shorter than in 1932. The first spat were found in Oyster Bay on July 3, and in Little Skookum the next day. In Oakland Bay the exact date was not noted, but the first bag of shells to bear spat was in the water from June 30 until July 7, indicating that setting be- gan at almost the same time as in Oyster Bay. In the season of 1933 setting started later and stopped earlier than during any of the other years. The reason for this short season may be seen in the records of water temperatures (see table 1, fig. 7). The water warmed to the spawning temperature much later in the spring, and did not reach as high a level by the end of the summer as during other years, thus reducing the length of the spawning season. Also, as was pointed out in the section on spawning, fewer adults bore broods of larvae. JULY AUGUST SEPT. Figure 25. — Average number of spat caught daily per bag of shells on two grounds in Oyster Bay and one each in Skookum Inlet and Oakland Bay, 1933. Range of tide is also indicated. SEASON OF 1934 In the summer of 1934 setting started on June 4, a month earlier than in the pre- ceding year, and continued until the end of September (see table 25, fig. 26). In both dikes (5 and S) there were again two distinct periods of setting, during June and early August, respectively. However, each period was of considerable duration and con- 482 BULLETIN OF THE BUREAU OF FISHERIES sisted of two minor maxima. In each instance the two submaxima are approximately 2 weeks apart, while the two peaks of the second period are about 6 weeks later than those of the first period. While in some seasons the second setting period is the greater, as in 1931, during 1934 the first was by far the more important. In 1932, however, the two were roughly equal. The results obtained in dike S are similar, but the sub- maxima do not show so clearly, possibly because the number of spat was too small to indicate such details. At the middle of July there was a period of about 2 weeks during which only an occasional larva attached. The figure shows that between July 9 and July 23 the number of spat caught was insignificant. A similar cessation of setting was also noted in 1931 for an even longer time in between the two major divisions of the setting season. As will be discussed later, this does not appear to be directly correlated with a similar variation in the frequency of spawning. Table 25. — Number of spat caught on bags of shells of 2 grounds in Oyster Bay during the season of 1984 Date planted May 28. June 1-. June 4. - June 8-- June 11- June 15- June 18. June 22. June 25- June 29 _ July 2— July 6--. July 9... July 13.. July 16- July 20.. Julv 23- July 27- July 30- Aug. 3_. Aug. 6.. Aug. 10. Aug. 13. Aug. 17. Aug. 20. Aug. 24. Aug. 27. Aug. 31. Sept. 3- Sept. 7._ Sept. 10- Sept. 14- Sept. 17. Sept. 21. Date removed Number of days Dike 5 Dike S Total number of spat Number of spat daily Total number of spat Number of spat daily June 4 1 7(1) 218 218 (2) June 8 1 7(5) 2, 267 453 (2) June 11. - 7 3| 569 610 549 78 June 15. . 7 47, 229 6, 761 3, 430 490 June 18 7 44, 825 6, 403 8, 510 1, 216 June 22. 7 39, 053 5, 579 8, 610 1, 230 June 25 7 34, 051 4, 864 10, 487 1,498 June 29 7 48, 309 6. 901 8, 000 1, 143 July 2 7 48, 600 6, 943 6, 855 979 July 6.. 7 28, 671 4,096 5, 407 772 July 9- 7 16, 448 2, 349 4, 307 615 July 13. - 7 4, 386 626 936 133 July 10. 7 398 57 79 11 July 20 ___ 7 65 8 (2) July 23 7 100 14 (2) July 27 '7(4) 2, 132 533 160 40 July 30 7 6, 193 885 892 127 Aug. 3 7 20, 785 2, 969 2, 204 315 Aug. 6 7 13, 728 1,961 1, 360 194 Aug. 10 .. 7 14, 694 2,099 2, 453 350 Aug. 13 7 24, 823 3, 546 1, 737 248 Aug. 17 - 7 8, 200 1, 171 526 75 Aug. 20. 7 6, 634 948 377 54 Aug. 24 7 2, 731 390 503 72 Aug. 27 7 1,047 149 417 59 Aug. 31 7 3, 528 504 31 4 Sept. 3. 7 2, 497 357 192 27 Sept. 7 — _ 7 1,480 211 69 10 Sept. 10-. - 7 1, 303 186 130 18 Sept. 14. 7 1,471 210 178 25 Sept. 17 7 1, 403 200 351 50 Sept. 21_ 7 123 17 34 5 Sept. 24 7 257 37 34 5 Sept. 28 7 62 9 1 Bags in water for 7 days, but setting started some time after they were planted. 2 Only occasional spat; too few to warrant counting. Table 26. — Number of spat caught on bags of shells in dike 5 in Oyster Bay during the season of 1935 Date planted Date re- moved Number of days Total number of spat Number of spat daily Date planted Date re- moved Number of days Total number of spat Number of spat daily June 14 June 21 1 7(3) 7, 991 2,664 Aug. 9 Aug. 16 7 22, 751 3,250 June 17 June 24 '7(6) 20, 451 3, 408 Aug. 12 Aug. 19 7 52, 810 7, 544 June 21 June 28 7 48, 848 6, 979 Aug. 16 Aug. 23 7 62, 110 8, 872 June 24 July 1 7 86, 034 12, 291 Aug. 19 Aug. 26 7 33, 628 4, 808 June 28 July 5 7 25,114 3, 587 Aug. 23 Aug. 30 7 19, 008 2,715 July 1 July 8 7 44,311 6, 330 Aug. 26 Sept. 2 7 18, 672 2,667 July 5 July 12 7 17, 890 2,555 Aug. 30 Sept. 6 7 15, 800 2, 257 July 8 July 15 7 9, 506 1, 358 Sept. 2 Sept. 9 7 5, 942 848 July 12 July 19 7 16, 505 2, 357 Sept. 6 Sept. 14 8 694 87 July 15 July 22 7 27, 751 3, 964 Sept. 9 Sept. 16 7 171 24 July 19 July 26 7 28, 097 4,013 Sept. 14 Sept. 20 6 380 76 July 22 July 29 7 16, 157 2, 308 Sept. 16 Sept. 27 11 514 46 July 26 Aug. 2 7 3, 791 540 Sept. 20 -__do 7 122 17 J ulv 29 Aug. 5 7 274 39 Sept. 27 Oct. 5 8 108 13 Aug. 2 Aug. 9 7 817 116 Oct. 5 Oct. 12 7 211 30 Aug. 5 Aug. 12 7 2,394 342 1 Bags in water for 7 days, but setting started some time after they were planted. SPAWNING AND SETTING OF OLYMPIA OYSTERS 483 SEASON OF 1935 Figure 26- Records for this year were analyzed completely only in dike 5, since it had already been well demonstrated that different portions of the bay differ only in the number of spat setting, while the time is the same. The results are given in table 26, and figure 27, and they show a marked similarity to those of previ- ous years in that there are two major divisions of the setting season. After start- ing on June 17 setting in- creased in frequency, reach- ing the first maximum at the end of the month. A second, smaller maximum centered about July 20. From July 29 until August 9 there was a time when almost no larvae attached, and then the second major setting period started, reaching a peak soon after the middle of August. Al- though the season was practically over early in September a few spat were caught up until mid-October, making a total setting season of nearly 4 months. Other bags of shells were planted in the same dikes at more frequent intervals, in order to give more exactly the times of the maxima. This series will be considered below with respect to the analysis of the significance of the results. SETTING SEASONS, MUD BAY Although parallel to Oyster Bay and separated from it by only a few miles, Mud Bay is somewhat different hydrograpliically, as was indicated above. Also, the spawn- ing and setting activities of the oysters are so different in the two bodies of water that they must be treated separately. It was shown in tables 14 to 17 that spawning does not begin until relatively late in Mud Bay, sometimes as much as 3 weeks later than in Oyster Bay. Setting is therefore similarly later. -Average number of spat caught daily per bag of shells on two grounds Oyster Bay, 1934. Range of tide is indicated also. Figure 27. — Average number of spat caught daily per bag of shells in dike 5 (Oyster Bay) in 1935. See also figure 33 for results with more fre- quently planted cultch. Table 27. — Number of spat caught per bag of shells on two grounds in Mud Bay, 1931 Date planted Date re- moved Num- ber of days Dike B Dike D Date planted Date re- moved Num- ber of days Dike B Dike D Total number of spat Num- ber of spat daily Total number of spat Num- ber of spat daily Total number of spat Num- ber of spat daily Total number of spat Num- ber of spat daily June 17 June 29 12 6,626 552 6, 240 520 June 29 Aug. 7 9 0 0 0 0 June 25 July 1 6 7, 390 1,231 3, 120 520 Aug. 7 Aug. 21 14 120 8 414 30 June 29 July 9 10 6, 949 695 10, 037 1,003 Aug. 21 Aug. 26 5 60 12 323 65 July 9 July 17 8 3, 021 377 3,709 463 Aug. 26 Sept. 10 15 331 22 628 42 July 13 July 25 12 331 27 228 19 Sept. 5 Sept. 26 21 871 41 1,988 95 July 20 Aug. 7 18 0 0 0 0 Sept. 26 Oct. 8 12 0 0 103 8 484 BULLETIN OF THE BUREAU OF FISHERIES SEASON OF 1931 Bags of shells were planted and removed periodically for spat counting on two grounds. Dike B is near the shore and dike D is next to the small channel which remains at low tide (see chart, fig. 4). As in Oyster Bay during this year a system of planting shells for regular in- 1 a DIKE 0 - mmmrn ■iii -L OIKE & - JUNE JULY AUGUST SEPTEMBER OCT. Figure 2S.— Average number of spat caught daily per bag of shells left in dikes B and D (Mud Bay) for different periods during 1931. tervals was not employed and the results are not as accurate as in later seasons. Setting began on the 16th of June and the maximum was reached at the end of the month, after which it diminished gradually in intensity (see table 27, fig. 28). From soon after the mid- dle of July until early in August no spat were caught, but after this time a few were found on every The total number of seeds caught was small The first setting period was bag of shells until the end of September as compared with Oyster Bay during the same season, by far the more important, as shown in the figure, while in Oyster Bay the later period of setting was more signifi- cant. All tests in Mud Bay were made in dikes A, B, C, D (J. J. Brenner Oyster Co.) and E (Charles Brenner). SEASON OF 1932 Figure 29.— Average number of spat caught daily per bag of shells in dike B, Mud Bay, 1932. During this year the improved method of sampling was employed and complete counts were made on the shells tested in dike B. For comparison, some counts were made on shells planted in dikes D and E. The latter is across the channel from dike D. The original results are given in table 28, and the dike B values are represented graphically in figure 29. The picture is in some respects different from that ob- tained during 1931, but the 2 years are alike in that there were two separate setting periods. In the graph the first period falls into two maxima. The late set- ting, though not as intense, was sufficient to be of com- mercial importance, although it continued only until early in September. SEASON OF 1933 The results for this year are given in table 29 and figure 30, and consist of complete counts on two series of bags of shells. The graphs are very similar save that the number of spat caught in dike E is only a small frac- tion of that obtained in dike B. It was necessary to dou- ble the scale in plotting the results in the former case in order to make the values distinct. It has generally been found that the catch of seeds Figure 30.— Average number of spat caught daily per bag of shells in dikes B and E, Mud Bay, 1933. in dike E and grounds nearby is much less than in other places a short distance away. SPAWNING AND SETTING OF OLYMPIA OYSTERS 485 Table 28. — Number of spat caught on bags of shells planted periodically in Mud Bay during 1932 [Counts were completed on the dike B series, and representative samples from dikes D and E were studied for comparison] Date planted Date removed Number of days Dike B Dike D Dike E Total num- ber of spat Number of spat daily Total num- ber of spat Number of spat daily Total num- ber of spat Number of spat daily July 2 July 9 7 441 63 466 66 July 12 7 935 134 609 87 July 9 7 2. 521 360 3, 663 623 July 12 July 19 7 8, 350 1, 193 7,626 1,089 3, 469 496 July 16 July 23 7 5, 153 736 4, 034 576 July 19 _ . July 26 7 2, 843 403 2,275 325 7 5, 590 798 July 26 Aug. 2 7 7, 692 1,099 5, 520 789 6, 047 864 July 30 . 7 1,705 243 2,126 304 Aug. 2 Aug. 9. 7 442 63 354 61 7 50 7 Aug. 9. Aug. 16 7 269 38 Aug. 20 7 1, 238 177 1,391 199 Aug. 16 Aug. 23 7 2, 992 427 4, 057 579 4,271 610 Aug. 20 Aug. 27 7 530 76 1, 395 140 Aug. 23. Aug. 30 7 2, 463 352 5, 177 739 Aug. 27 Sept. 3 7 220 31 Aug. 30 7 175 25 7 320 46 860 123 7 45 6 Table 29. — Number of spat caught on bags of shells on 2 grounds in Mud Bay, 1933 [Dike B is more favorable seed ground, although the time of most profuse setting is the same on both beds] Date planted Date re- moved Num- ber of days Dike B Dike E Date planted Date re- moved Num- ber of days Dike B Dike E Total number of spat Num- ber of spat daily Total number of spat Number of spat daily Total number of spat Num- ber of spat daily Total number of spat Num- ber of spat daily July 18 . July 25.. ■7(1) 1,494 1,494 Aug. 12. _ Aug. 19. 7 1,836 262 234 33 22.. 29.. >7(5) 2,318 464 157 31 " 15. _ 22. 7 913 130 163 23 25.. Aug. 1. . 7 9, 933 1,419 1,417 202 19.. 26. 7 530 76 55 8 29.. 5.. 7 27, 678 3, 954 3, 877 554 22.. 29. 7 226 32 (2) Aug. 1.. 8.. 7 10, 011 1,430 2, 951 421 26.. Sept. 2.. 7 183 26 (s) 5.. 12. 7 8, 154 1,165 1,892 270 29.. 5.. 7 61 9 M 8.. 15. 7 1,565 223 463 66 1 Bags in water for 7 days, but setting started some time after they were planted. 2 Only occasional spat; too few to warrant counting. Table 30. — Number of spat caught on bags of shells on 2 grounds in Mud Bay, 1934. Date planted Date removed Num- ber of days Dike B Dike E Date planted Date removed Num- ber of days Dike B Dike E Total number of spat Num- ber of spat daily Total number of spat Num- ber of spat daily Total number of spat Num- ber of spat daily Total number of spat Num- ber of spat daily June 9 June 16 1 7(1) 39 39 7 7 June 26 July 3 7 729 104 137 19 June 12 June 19 1 7(4) 377 94 123 31 June 30 July 7 7 1,360 194 355 50 June 16 June 23 7 2, 134 305 248 35 July 3 July 10 7 788 112 85 12 June 19 June 26 7 1,998 285 343 49 July 7 July 14 7 401 67 Juno 23 June 30 7 1, 562 223 266 38 July 10 July 17 2 7(4) 66 17 1 Bags in water for 7 days, but setting started some time after they were planted. 2 Setting stopped 4 days after planting. The figure shows only one period of setting, which started between July 24 and 25, reached a maximum about a week later, then gradually declined until early in September when it reached the zero level. The entire length of the setting season was onl^y about 1 % months. Nevertheless, oyster growers obtained a highly satisfactory catch of seeds, for during a short time the frequency of attachment was greater than was found in this bay during any of the other years. 486 BULLETIN OF THE BUREAU OF FISHERIES SEASON OF 1934 This was a relatively poor seed year in this bay, as may be seen in table 30 and figure 31. Even for the dike B results it was necessary to employ a scale 10 times as great as that used in most of the figures in order to obtain a satisfactory graph. After starting to set on about June 16 the larvae never attached in great numbers, so that at the time of the maximum the bags of shells caught an average of only about 300 spat per day in dike B and only about 50 per day in dike E. The entire setting period Figdre 31. — Average number of spat caught daily per bag of shells in dikes B and E, Mud Bay, 1934. Figure 32.— Average number of spat caught daily per bag of shells in dike E, Mud Bay, 1935. occupied only about 1 month, and no larvae became attached after the middle of July, although bags of clean shells were planted twice weekly until the end of September. Table 31. — Number of spat caught on bags of shells in dike E in Mud Bay during the season of 1935 Date planted Date removed Number of days Total num- ber of spat Number of spat daily Date planted Date removed Number of days Total num- ber of spat Number of spat daily June 19 June 26 7 74 10 July 13 July 20 7 428 61 June 22 June 29 7 128 18 July 17 July 24 7 94 13 June 26 July 3 7 54 7 July 20 July 27 7 68 9 June 29 July 6 7 102 14 July 24 July 31 7 22 3 July 3 Julv 10 7 122 17 July 27 Aug. 3 7 115 16 July 6 July 13 7 203 29 July 31 Aug. 7 7 0 0 July 10 July 17 7 445 63 SEASON OF 1935 Because of more concentrated investigations in Oyster Bay little attention was paid in 1935 to Mud Bay. One series of bags of shells ivas completed in dike E, which is not good seed ground. The catch in dike B, judging from records of previous years, was probably at least five times as great. The trend of the setting season is shown in table 31 and figure 32. Setting started at the last of June, reached a maximum at the middle of July, and stopped at the first of August. Even during the time of most SPAWNING AND SETTING OF OLYMPIA OYSTERS 487 abundant setting the bags of shells caught an average of only about 50 spat per day. There was some later set after the middle of August, but it was of small significance and samples taken are not complete enough to be included in the table and graph. Mud Bay is different from Oyster Bay in time of spawning, time of beginning of setting, time of maximum intensity of setting, and duration of the setting season. While Oyster Bay, Little Skookum, and Oakland Bay are closely similar with respect to setting seasons, Mud Bay is entirely different, and the results have to be presented separately. PERIODICITY OF SETTING In the foregoing account it was described, particularly with reference to Oyster Bay, that several periods of setting may occur during each season. Attempts to correlate these periods with conditions of salinity, pH, or temperature have resulted in no significant relationship. Local weather conditions appear to have little or no influence upon the setting of larvae, save in their effect upon water temperature which controls spawning and rate of larval development. A period of setting occurs generally as a matter of many days duration, seldom less than 2 weeks. In this locality it is not concentrated within a few days, as described by Prytherch (1929) for Long Island Sound. Ordinarily, only a few spat are found when a setting period begins, but during the following days the larvae attach more and more abundantly. The oyster growers have the problem of deciding when to plant cultch so that it will not be silted over or covered with organic growth before the larvae are able to attach. The system has always been in use to plant the cultch in advance of the time of setting, after it is known that spawning has started, so as to be certain that the cultch is in the water when setting begins. Naturally, it frequently occurred that cultch was planted far too early and the maximum catch of seeds was not obtained. During 1931, an opportunity was afforded to test roughly the depreciation in efficiency of cultch after being in the water for some time. Table 32. — Loss in 'percentage of efficiency of cultch after 9 days Bag number Date planted Date re- moved Number of spat Difference Dike 1 3578 July 18 July 27 July 18 July 27 Aug. 3 Aug. 3 Aug. 3 Aug. 3 22, 320 40, 520 28, 071 39, 040 Percent } 44. 91 } 28.09 3515 Dike 6 3577 3516 36. 50 During the middle of the summer the water is typically relatively free of silt and organic growth, as compared with spring and early summer. In 1931 the second major period of setting began between the 25th and the 28th of July. Two bags of shells had been planted on July 18 and two on the 27th. All were removed on August 3. The counts of spat on the two groups are given in table 32. It is assumed, for convenience, that the bags planted July 27 were placed in the water just at the correct time to obtain the maximum catch, though they may have been a little too late to get all of it. The other bags were planted 9 days earlier. When all bags were removed at the same time and the number of spat counted there was found to be a remarkable difference. One bag planted in dike 1 on July 18 caught a total of 22,320 spat, while the bag planted beside it on the 27 caught 40,520 spat. Similarly, in dike 5 the earlier bag caught 28,071 spat, and the later, 39,040. The average difference between 488 BULLETIN OF THE BUREAU OF FISHERIES the two groups is 36.5 percent, indicating that the earlier shells had lost one-third of their efficiency as spat collectors in 9 days, even during the time when the bay water was most free from fouling materials and organisms. The shells in wire bags are less subject to fouling than those thrown directly upon the grounds, for they are well supported above the silt of the bottom and tidal flow serves to keep them clean. The depreciation in effectiveness of shells thrown on the grounds is probably very rapid, particularly during early summer when there is still considerable silt. An understanding of the setting periods should serve to make it possible to eliminate much of the loss due to fouling of cultch. It was noted that setting periods appeared to be approximately 2, 4, or 6 weeks apart, rather definitely spaced, suggesting that tidal periodicity might be concerned. By plotting the daily range of tide throughout each setting season (fig. 23 to 27) this suggestion was shown to be well founded. In almost every case the time of maximum frequency of setting is near to tbe time of greatest tidal range; and in many cases it may be observed that during neap tides, when the range of tide is small, the frequency of setting is also low. In some of the series in which 7-day bags were used it is difficult to decide from tbe graphs the exact dates of maximum frequency of setting. To throw more light upon the nature of the periodicity and the possible relation to tidal conditions a series of bags of shells was tested in one dike in 1935 in such a manner that each bag generally remained in the water onty 2 or 3 days. Table 33. — -Number of spat caught on bags of shells planted at frequent intervals in Oyster Bay, dike 5, 1935 Date planted Date removed Number of days Total number of spat Number of spat daily Date planted Date removed Number of days Total number of spat Number of spat daily June 18 June 20 2 1,120 554 July 23 July 25 2 4,786 2, 393 June 19 June 21 2 1, S40 920 July 24 July 26 2 2, 788 1,394 June 20 June 22 2 4. 123 2, 061 July 25 July 27 2 2, 851 1,425 June 21 June 24 3 6, 909 2,303 July 26 July 29. 3 3, 231 1,077 June 24 June 25 1 5, 471 5, 471 July 27 -__do — 2 883 441 June 22 June 24 2 3, 226 1,613 Do— July 30 3 1, 220 407 June 25 June 26 1 1, 186 1,186 July 29 July 31 2 177 88 June 20 Juno 27 1 4, 131 4, 131 July 30 Aug. 1 2 88 44 June 27 June 28 1 8, 703 8, 703 July 31 Aug. 2 2 37 18 June 28 July 1 3 8,211 2, 737 Aug. 1 Aug. 3 2 83 41 June 29 --.do 2 37, 704 18, 852 Aug. 2 Aug. 5 3 122 41 Do... July 2 3 10, 157 3, 386 Aug. 3 Aug. 6 3 191 64 July 1 July 3 2 24. 120 12, 060 Do-.. Aug. 5 2 57 28 July 2 July 4 2 29, 149 14, 574 Aug. 5 Aug. 7 2 60 30 July 3 J uly 5 2 31,043 15, 021 Aug. 6 Aug. 8 2 34 17 July 4 July 6 2 20, 360 10, 180 Aug. 7 Aug. 9 2 57 28 July 5 July 8 3 16, 640 5, 547 Aug. 8 Aug. 10 2 191 95 July 0 ...do 2 7,500 3, 750 Aug. 9 Aug. 12 3 1, 200 400 Do... July 9 3 8, 805 2, 935 Aug. 10 -_-do 2 360 180 July 8 July 10 2 2,914 1, 457 Do... Aug. 17 7 45,314 6, 473 July 10 July 12 2 1, OSO 540 Aug. 12 __.do 5 47, 968 9, 593 July 11 July 13 2 622 311 Aug. 17 Aug. 19 2 17, 126 8, 563 July 12 July 15 3 3,162 1, 054 Do-.. Aug. 20 3 19, 571 6, 524 July 13 ___do 2 4, 520 2, 260 Aug. 19 Aug. 21 2 9,694 4, 847 Do.. July 16 3 6, 168 2,056 Aug. 22 Aug. 24 2 4, 351 2, 175 July 15 July 17 2 4, 617 2, 308 Aug. 23 Aug. 26 3 4,783 1,594 July 10 July 18 2 4, 445 2, 222 Aug. 24 Aug. 27 3 9, 434 3, 145 July 17 July 19 2 4,200 2, 100 Aug. 26 Aug. 28 2 4, 662 2, 331 July 18 July 20 2 4, 831 2, 415 Aug. 27 Aug. 29 2 6, 225 3,112 July 19 July 22 3 10, 460 3, 487 Aug. 28 Aug. 30 2 3, 663 1,831 July 20 July 23 3 5, 723 1, 908 Aug. 29 Aug. 31 2 2, 757 1, 378 Do-.. July 22 2 3, 448 1,724 Aug. 30 Sept. 2 3 5, 794 1,931 July 22 July 24 2 2, 843 1,421 Aug. 31 Sept. 3 3 6, 791 2, 263 Smaller units make the graph (fig. 33) more complete and permit a more certain statement of the correlation between frequency of setting and tidal periods. The results of the counts are given in detail in table 33. On the graph the daily range of tide is also plotted, and it may readily be seen that the maximum of the first setting period is centered almost exactly during a period of extreme, or spring, tides. Setting started on June 17, at the time of maximum range of the preceding tidal period, and SPAWNING AND SETTING OF OLYMPIA OYSTERS 489 reached a peak 2 weeks later. The first wave of setting may be considered as having ended about July 10, during neap tides, and a new set started immediately afterward, not reaching a definite sharp peak. It is to be noted that while this setting period continued until the end of the month, for a total time of about 3 weeks, the intervening neap tides had a high minimum range of 13 feet. The difference between spring and neap tides in this case was small, and little difference in rate of setting is to be observed. The next major setting cycle is concentrated during the extreme spring tides, having started at the time of the preceding neap tides. The following neap tides were not markedly different in range, and the effect is slight, though obvious. During the prominent neap tide period at the beginning of August almost no spat were caught. Figure 38.— Average number of spat caught daily per bag of shells left in water (dike 5, Oyster Bay) for periods of usually 2 or 3 days, 1935. Tidal range is also shown to illustrate correlation. Compare with figure 27. This series shows on a more exact basis that the previous interpretation of results obtained with weeldy bags is generally correct. An important source of possible error in reaching an understanding of the sig- nificance of counts of spat on the shells in bags is the fact that the shells are clean and thoroughly efficient as spat collectors only at the time they are put into the water. During the next 7 days they become increasing^ less efficient. The error is overcome to some degree by using overlapping series. It is to be noted, for example, in figures 27 and 33 that the average number of spat caught daily during any time is greater on the shells that were in the water 2 or 3 days than on those kept for 7 days. In inter- preting weekly series it is necessary to take the factor of fouling into consideration, for it would not be correct to say that the exact center of the highest column in any case represents the day of most abundant setting. STAGES OF TIDE AND SETTING Since periods of spring tides were shown above to pnmde most favorable condi- tions for the attachment of larvae it is of interest to determine the effect of different stages of tide. In Milford Harbor, Conn., Prytherch (1929) found for Ostrea vir- ginica that — “Heaviest setting occurs in the surface layer during the period of low slack water, which is the zone in which the oyster larvae were found to be most abundant. Setting continues as the tide 490 BULLETIN OF THE BUREAU OF FISHERIES begins to run flood, gradually becoming less intense as the velocity of the current increases, and finally ceasing altogether when the current attains a velocity of 10 centimeters, or one-third foot, per second.” Where this investigator worked the range of tide is less than half of that in south" ern Puget Sound and it is hardly to be expected that setting habits of larvae would be identical in the two places. When experiments on this subject were begun it was desired to determine at what stages of tide setting is most intense and the possible effect of such factors as daylight and darkness, sa- linity, temperature, and pH. Glass plates, sup- ported in wire frames as previously described, were arranged in units of fifteen 8- by 10-inch plates, mak- total area of under mg a total area surface of 1,200 square inches, almost 1 square yard. At low tide the frames were just covered by the water retained in Figure 34. — Number of spat caught hourly per unit of cultch with relation to stage of tide, temperature, and salinity. (Oyster Bay, dike 5, June 29 and 30, 1932.) the dikes. A set of frames was placed upon the ground and allowed to remain for a definite interval, then removed, allowed to dry and the number of spat counted. In the first series, plates were planted in the dike soon after it was exposed by the receding tide, at 6:30 a. m., and allowed to remain until just before the flood-tide water came over the dike (fig. 34). During this time the plates caught but 3 spat. The next set was in the water for a total time of 30 minutes, from near the end of the exposed period until the water was about 1 foot deep over the dike. Throughout the rest of the tidal cjmle it was arranged to have a set of plates in the water during each major tide except for about 1% or 2 hours at the times of high and low tides, which were separately tested. In figure 34 the results are given as number of spat caught hourly on each of the sets of plates. Shown also are depth of water throughout the period and values of temperature and salinity obtained each time samples were changed. The fewest spat were caught at low tide when the dike was exposed, the most on the two flood tides, although at, or near, the time of the higher low tide spat were also caught. During ebb tide few larvae set, in proportion to the activity at flood tide. It is of interest that salinity and temperature were quite uniform during the experimental period, save that the exposed dike permitted warming of the water at low tide. It is also clear that attachment of larvae is not markedly influenced by daylight or Figure 35. — Number of spat caught hourly per unit of cultch with relation to stage of tide, temperature, salinity, and pH. (Oyster Bay, dike S, July 6 and 7, 1932.) SPAWNING AND SETTING OF OLYMPIA OYSTERS 491 darkness. The two high tides in this case were almost identical and setting was almost equally heavy on the two floods, one day, the other night. In the next experiment, about 2 weeks later, made on a different ground, the first high tide was more than 4 feet higher than the second. Whether the different picture obtained (fig. 35) is due to the large difference between the two tides is un- certain though it would not seem unlikely that the greater flow of water in the former may account for the heavier set. On neither day was a single spat caught when the dike was exposed. Setting was about equally profuse on flood and ebb of the first tide, and it is suggested that the very low tide after midnight, which almost left the ground exposed, caused rate of current or other factors on the ebb and flood tides to be similar. The following secondary high tide produced few spat, most of which set during the flood. Except when the ground was exposed the temperature varied only slightly. It is of interest that the most prolific setting took place when both pH and salinity were quite high. The water in the dike at the end of the period of exposure had a pH of 7.8, while about an hour later it was 8.4. At the same time the salinity rose from a little over 26 to more than 29 parts per mille. In an attempt to reach a more definite conclusion a similar experiment was made in a later year with the additional help of an electric current meter. The boat was anchored in the channel near the oyster ground and the current meter suspended from a framework directly in the dike. The meter hung just above the oysters, at the level of the panes of glass, and the transmission line was led to the boat where counts of revolutions were made. The results are shown in figure 36. Spat were caught only at three well separated times: During the first flood and the following small ebb, and during the major ebb of the next day. The first and the last coincide definitely with the times of swift current. In the second case the current was not particularly rapid, but it is true that the heaviest setting took place about half way between high and low tide, when the current during this ebb was swiftest. Why no spat were caught on the second flood is not known, though it must be realized that in an experiment of this kind, necessarily carried on over a limited area and with a relatively small amount of cultch, chance is a large factor in determining whether the water in the particular place happens to contain larvae. For this reason the error involved in the tests is considerable. Nevertheless, the results are, within certain limits, of great interest. The final series to be described here is shown in figure 37. The experiment consisted of two parts. One set of plates was in the water for an entire tide, from high to low or low to high. In the other group two overlapping series of plates were used, each set being left in the water 3 hours. The lower portion of the graph represents the first group. Although spat were caught during all four tides, only -v 15 12 OQ Figure 36. — Number of spat caught hourly per unit of cultch with relation to stage of tide, rate of current, temperature, salinity, and pH. (Oyster Bay, Made Ground, June 26 and 27, 1934.) 492 BULLETIN OF THE BUREAU OF FISHERIES during the two floods did very many larvae set. By far the most were caught during the major flood tide following the extreme low of the first day. The sets of plates which were in the water for 3 hours each gave generally similar results. Those which caught most spat were exposed during the time when the flood tide was most rapid, from 1% hours after the water came over the dike, in the first case, until l}{ hours before high tide. During the first, or small, ebb fewer spat were caught than on the major ebb the next morning. The current records are in- complete in two places, because seaweed became entangled in the current meter. The pH was remarkably constant throughout the period, save that at low tide it dropped from 8.4 to 7.8, due to respiratory activity of the oysters in the shallow, warm water. Salinity and temperature also varied but slightly. Prytherch (1934) made the important finding that larvae of Ostrea virginica may vary over a wide range in the time required for completion of the setting process, and stated that: The most rapid setting was observed at salinities of 16 to 18.6 per mille and was complet- ed in from 12 to 19 minutes. He also determined that: In solutions that were above or below this salt concentration, the time for setting increased and reached a maximum of 140 and 144 minutes in salinities of 5.6 and 32.2 per mille respectively. It is of great significance that the time required for a larva to complete the process of setting may vary from 12 to 144 minutes, for it would therefore appear that a number of environ- mental conditions might become limiting factors. That the matter of rate of setting may have influenced the results of the experi- ments just described appears to be certain. In one case a set of plates was left in the water for only 15 minutes, yet it bore spat. In other cases the plates which were in the water for a longer period caught fewer spat than expected. In the last series described (fig. 37) the plates which were exposed during the major flood tide for a period of 6 hours caught a total of 103 spat. The three sets shown in the upper graph, covering the same period of time, caught a total of only 42 spat. It is suggested that those larvae which had not completed the setting process released their hold when the plates were withdrawn from the water, so that possibly only those that began to set soon after the plates were immersed were able to attach permanently. More clearly to illustrate the point it may be stated that a set of plates, left in the water during the entire 24 hours caught a total of 199 spat, while the four sets of the lower graph caught 151, and all of the double series of 3-hour plates caught but 129 spat. This information is of assistance in interpreting the results of the four experiments for it indicates that the first portion of the time that the plates are in the water must Figure 37. — Number of spat caught hourly per unit of cultch with relation to stage of tide, rate of current, temperature, salinity, and pH. (Oyster Bay, Gale Ground, June 27 and 28, 1935.) SPAWNING AND SETTING OF OLYMPIA OYSTERS 493 be given most weight. In such an instance as that shown in the upper graph of figure 37 the first 3-hour set of plates caught few spat; the next one, planted 1 % hours later caught many; the third, still 1% hours later caught few. It would seem safe to assume that most of this set occurred between 12 and 1:30 o’clock, or the third set would have caught a larger number. Summarizing these experiments it may be said that almost no spat are caught when the tide is low and the water in the dike is still. At this time the water is warmest, the salinity lowest, though only 1 to 2 per mille below that of the maximum, and the pH lowest, especially after the ground has been exposed for some time. As the tide comes in setting increases in intensity, most of it occurring when the water is 6 to 8 feet deep over the beds on which the tests were made. At this time the pH and salinity are relatively high and the temperature low. In some cases there appears to be a correlation between rate of current and frequency of setting. There seems to be no obvious parallel with conditions observed by Prythercli (1929) in Connecticut. DEPTH OF SETTING Most of the favorable grounds for the collection of spat are relatively high, at a level of 3 to 8 feet above the average lower low tide. While usually not so good as producers of the best oysters for market the higher grounds are used almost entirely for the collection of seeds. To the practical oystermen, who have always believed that most of the larvae set when the tide is low and the water on the grounds clear and warm, the highest grounds offered the warmest water at low tide and for this reason were especially favorable for catching seeds. However, as was shown in a preceding section, setting takes place at a lower frequency at low tide than at any other time. Only when the tide is as much as half high are some of the best seed beds com- pletely covered, except for the few inches of water held by the dikes. Because they are covered by deep water so much less of the time and are closer to the warmer surface water the seed beds are best for obtaining rapid growth of the spat, while on lower grounds, where growth is slower, the oysters fatten better but not so many spat are caught. It is probably true that cultch becomes fouled with organic growth more quickly on the lower grounds, thus preventing a heavy set of spat, but it appeared possible that some other factor might be concerned in determining that higher grounds are so much more effective. Table 34. — Number of spat caught on shells in four series of wire baskets suspended at fixed distances from surface of water [A basket covered a depth of 5 inches] Depth (inches) Series 1 number of spat Series 2 number of spat Series 3 Series 4 Depth (inches) Number of spat Depth (inches) Number of spat 0-5 402 228 0-5 338 0-5 182 6-11 1,029 517 6-11 612 8-13 696 14-19 1,393 640 14-19 662 16-21 604 24-29 779 392 24-29 723 24-29 654 39-44 757 417 42-47 624 32-37 544 64-09 718 447 71-76 591 99-104 530 495 Several series of experiments were performed for the purpose of finding at what depths the larvae set most profusely. Wire baskets were constructed 12 by 12 inches wide and 5 inches deep, filled with clam shells, suspended in series one above the other, 494 BULLETIN OF THE BUREAU OF FISHERIES and the entire series hung from a float which was anchored in the channel. The baskets were supported so that each was in a horizontal position, occupying 5 inches of depth. They were placed so as to be well separated in the series and to maintain a constant distance from the surface for periods of from 2 weeks to 2 months, and the spat then counted. The results of four series are given in table 34. In two series the baskets were at depths ranging from the surface down to 104 inches. The others reached only to 37 and 76 inches, respectively. In each basket 30 unselected shells were gone over carefully and every spat counted. The results are relatively uniform, considering the impossibility of measuring the exact area of individual shells, and when the results of the two longer series are averaged and plotted as number of spat per unit of culteh at different depths it becomes evident that the most spat were caught within the first 20 to 30 inches from the surface. In all cases the sample at the surface (0 to 5 inches) caught fewest spat, possibly because of the scouring action of waves, partly, perhaps, because larvae do not set as profusely within that area as they do a short distance below. On the bags suspended below this level of maximum setting fewer and fewer spat were caught. About twice as many spat were taken on the shells at a depth of 14 to 19 inches as on those at 99 to 104 inches. It was shown above that most of the larvae set when the water has a depth of about 6 to 10 feet or more The best seed grounds are feet, thereby placing them within or close to the area of maximum setting as shown in figure 38. The deeper the water on the oyster ground at the time setting takes place the fewer spat will be caught on cultch placed on the bottom. It is logical to conclude that this is one of the reasons why the higher grounds are best adapted to the catching of spat. Difficult to understand in view of these results is the fact that all natural beds in the region are located between low- and high-tide levels, or in shallow channels which are almost dry at low tide. The graph shows that the number of spat caught diminishes gradually with increasing distance from the surface. Tests were made only down to about 8.5 feet and the results suggest that larvae would set to some extent at much greater depths. It appeared likely that beds of oysters might be found in the deep channels well down the bay but extensive dredging in such places failed to disclose a single oyster. It may be that there is little clean cultch to which the larvae might attach, but clam shells were found abundantly in some places. The factor responsible for this localization of natural oyster beds is not clear, but in Yaquina Bay, Oreg., oysters of the same species occur almost exclusively in the deeper waters. These experiments have served a more immediately practical purpose. Follow- ing the original observations in 1931, which demonstrated that a very heavy set of spat could be obtained by employing floats, it was suggested to oyster growers that they try the method on a commercial scale. One of them tried it with a float made of two logs and a wire bottom, filed with Japanese oyster shells, in 1932. He was quite successful and during the following years others have started catching seeds in DEPTH IN INCHES Figuke 38. — Average number of spat caught on baskets of shells suspended from floats at different depths. See table 35. above the zero tide level, at which the tests were made, well above the zero level, frequently as high as 6 or 8 U, S. Bureau of Fisheries, 1937 Bulletin No. 23 Figure 39. — Photograph showing float with removable compartments filled with cultch for catching spat. SPAWNING AND SETTING OF OLYMPIA OYSTERS 495 this manner. In 1935 and 1936 the method has been put into practice on a large commercial scale, with floats filled with shells or manufactured collectors. Some of the growers have worked out a system of dividing the floats into a series of removable compartments, in which the cultch is placed, thus facilitating handling and minimizing possible storm damage. (See fig. 39.) Counts made on typical egg crate fillers or special-type collectors in 1935 showed that they caught an average of 5,000 to 10,000 spat each. These floats may be anchored in the channels or pot holes of Oyster Bay, where they get a swift current of water. The manufactured collectors are always placed so that water will flow through the cells bringing abundant larvae and washing out silt. Most of the growers using the method at the present time are those who have satis- factory growing ground but lack adequate seed beds. Formerly these growers pur- chased what seeds they were able to get, but in recent years, since almost complete destruction of oysters on the State-owned seed beds of Oakland Bay following the beginning of operations of a nearby pulp mill, almost no seeds have been purchasable. The float method now makes it possible for anyone with growing ground to obtain abundant seeds at a cost considerably less than would be required to maintain seed ground for the purpose. CORRELATION BETWEEN SPAWNING AND SETTING In the foregoing account detailed descriptions have been given of observations of spawning activities and on setting of larvae throughout the several seasons under different conditions. It is of interest to consider reproductory activities in their entirety in order to correlate the initial spawning with the somewhat later setting and metamorphosis of the larvae. Coe (1932 a, b) stated that spawning in this species continues during at least 7 months of the year on the coast of southern California, while in British Columbia waters, according to Stafford (1914), “The spawning season appeared to extend from about May 20 to about the last of July, and to have reached its maximum about the middle of June." This is a total spawning period of about 2% months, and he observed a setting period of about the same length, from early July until nearly the middle of September. The investigations described above indicate a total spawning period of 3 to 4 months, although the most intense spawning activity is confined to a much shorter time. Stafford estimated that at least a month is required between spawning and set- ting, while Coe (1932a) stated: Shortly after they have been spawned into the water these young bivalves attach themselves to almost any kind of solid objects. The free-swimming stage is thus very short and the opportunities for dispersal are limited. In a similar manner Galtsoff (1929) wrote: It is noteworthy that, although the whole development of the Pacific oyster is about twice as long as that of the eastern oyster, the duration of the free-swimming stage, when the organism is subjected to the vicissitudes of life in the open water and is not protected by the mother’s body, in both cases lasts for about a fortnight. Thus, the fact that Ostrea lurida spends half of the period of its development within the brood chamber of its mother is of no particular advantage, and the free swimming larvae of both species have an equal chance to become prey to plankton-feeding organisms or to be carried away by the tides. Because of the fact that the free swimming larval period lasts for a month or more, as was noted above, it is obvious that there is great opportunity for dispersal ; and in view of the fact that the maternal individuals protect their larvae until they have developed to an advanced bivalve stage at which they are presumably able to protect 496 BULLETIN OF THE BUREAU OF FISHERIES themselves from many of the unfavorable environmental factors it must be considered that the chances of survival are greater than in the case of oviparous species, which cast the unprotected eggs into the open water. Table 35. — Dates on which first spawning and first setting occurred Oyster Bay Mud Bay Date Spawn Spat Number of days Spawn Spat Number of days 1931 June 12 June 16 1932 May 17 June 26 40 May 25. July 7 43 1933 do July 3 47 June 4 July 25 51 1934 Apr. 17 June 4 48 Apr. 24 June 16 53 1935 May 5 June 17 44 May 13_ June 29 47 The only accurate estimate available of the duration of the free swimming period of larval life is that obtained by Hori (1933) who grew the larvae in the laboratory by feeding them macerated sea lettuce ( Ulva ). He removed black larvae from the brood chambers and kept them in dishes of seawater at a temperature of about 20° C. and found that they reached full size and attached after 22 days. The tem- perature of the water in Puget Sound is generally considerably lower and it is to be ex- pected that devel- opment o f larvae would proceed more slowly. Field observa- tions on the time when the first larvae and the first spat were found each season are summarized in table 36. In Oyster Bay the interval varied during 4 years from 40 to 48 days, wliile in Mud Bay the extremes were 43 and 53 days, or about 4 days longer each year. Avail- able data do not permit an exact statement of the total time from spawning until setting, for it is most probable that natural conditions may cause it to vary from year to year. Water temperature necessarily is concerned in determining rate of growth and it is probable that development may be affected by the abundance of food material. Hori (1933) was able to grow larvae of Ostrea gigas by feeding them Chlorella pacifica, but larvae of 0. lurida did not thrive on this alga. The experiments of Amemiya (1926) indicated that salinity, also, is an important factor in the development of larvae of several species. It has been described that during the first 10 days larval develop- ment takes place within the maternal brood chamber and the free swimming period in Oyster Bay is therefore some 30 or more days in length. Although tables and graphs of both spawning and setting activities have already been described, a complete picture of the season of propagation is better presented, as in figures 40 and 41, by including measurements of both spawning and setting on the same graph. In figure 40, referring to Oyster Bay in 1932, the frequency of occurrence of gravid adults is shown as a histogram while the time and abundance of larvae set- ting is indicated by a trend line derived from the results obtained by sampling with Figure 40. — Frequency of spawning and setting during season of 1932 in Oyster Bay. Setting is indicated by a trend line derived from values given in figure 24. Tidal range is also shown. SPAWNING AND SETTING OF OLYMPIA OYSTERS 497 hags of shells as previously described. Range of tide is also given. During this jrcar most of the spawning took place from the middle of May until the middle of June, and during July there was some further active spawning. In some respects the record of setting resembles that of spawning, though the break between the two setting periods may not readily be correlated with a comparable cessation of spawning. If the second setting period is traceable to larvae resulting from July spawning the mortality of larvae produced during the first spawning period was tremendously larger, for the later period of setting was very intense. In the graph referring to the year 1934 (fig. 41) the picture is somewhat similar, though one can hardly consider that there was a second distinct period of spawning. In this figure the correlation between tidal periods and setting is strikingly shown, while in figure 40 the second setting period appears to be correlated with the neap tides. However, in the latter case it is not quite correct to plot tides in this manner, for at the time of the second major set there were very low tides, but the high tides were not great, so that the total range shown is small. In both years, which were selected for presentation be- cause they represent marked differences in time of spawning and setting, the seasons of reproduction cover about 5 months. It may be noted that setting begins during the third period of tides following the beginning of spawning, and also that for 5 seasons, the second major setting period takes place during the third and fourth spring tide periods following that when the first set occurs. The time intervals seem to be predetermined, either by the spawning activity or by cyclic changes in the water which are correlated with tidal periodicity. DISCUSSION In the foregoing account various phases of the spawning and setting activities of the Olympia oyster have been described with particular reference to their application to commercial cultivation. Larvae of this viviparous species develop slowly within the maternal brood chamber, or that portion of the mantle chamber which contains the palps and the anterior ends of the gills, and require an average time of about 10 days before they reach the size at which they are normally discharged into the open water. While eggs spawned by the female of 0. virginica and other viviparous species develop to the trochophore, or earliest swimming stage, within a few hours, in the case of 0. lurida the same stage is not reached for about 4 days. Rapid early develop- ment is characteristic of those species which discharge the eggs directly into the open water, in contrast with the viviparous Olympia oyster which protects the embryos. Oyster culture in Puget Sound is somewhat different from that in other parts of the United States in that the range of tides is greater, with a maximum range of 20 feet. The oyster grounds are above the extreme low-tide level and are surrounded Figure 41.— Frequency of spawning and setting during season of 1934 in Oyster Bay. Setting is indicated by a trend line derived from values given in figure 26. Tidal range is also shown. 498 BULLETIN OF THE BUREAU OF FISHERIES by dikes which hold enough water to protect the oysters from freezing and drying, while at high tide they may be covered by as much as 16 or 18 feet of water. Such tides involve the movement of great quantities of water and swift currents, but most of the beds are located in the upper ends of the bays where currents are not so rapid. The interchange of cold water from the very deep portions of the Sound with that in the upper ends of the oyster-producing bays prevents the temperature from rising to a high level even during an exceptionally warm season. For this reason eastern oysters, transplanted to Puget Sound, were never able to spawn, since the high-tide temperature on the oyster beds does not reach the critical level of 20° C. The spawning season is of several months duration and, although in no case has a sudden burst of spawning been observed in which a great number of oysters were involved, as occurs frequently with oviparous species, it has been found that smaller numbers of individuals often bear embryos or larvae of the same age, indicating that favorable conditions may cause spawning to take place in a considerable portion of the population at the same time. Alternation of sexual phases (Coe, 1931a, b; 1932a) probably is responsible for the rather slowly developed wave of spawning, for different individuals are at any time in different stages of maturity. Sometimes as many as 12 to 15 percent of the adults bear larvae of the same age, so that a system of statistical sampling serves to show the rate of growth. Whether sperms or sperm extract will stimulate discharge of eggs b}7 functional females (Galtsoff, 1930b, 1932) has not been demonstrated in this species, but it is considered probable. It is hardly to be expected that the small native oyster would discharge as many eggs as the larger oviparous species (Galtsoff, 1930), not only because of difference in size but also because the eggs are held within the mantle chamber where they grow into larvae almost twice the diameter of the eggs, and space alone probably acts as a limit. Altlioiigh an individual produces in one brood onfy about 250,000 to 300,000 larvae, all individuals are capable of bearing at least one brood each season, while the eastern oyster is generally functional as only of one sex during a single season (Coe, 1932c, d). In some years as many as 150 broods are produced per 100 oysters, indicating that a large number bear second broods, while in other seasons as few as 75 broods per 100 individuals are produced. The degree of success of a spawning season depends upon the number of larvae per brood and the total number of broods produced. A problem which has never before been attacked is that of the relationship between angle of surface and frequency of attachment of larvae, although Prytherch (1934) observed that larvae in a dish set more abundantly on the vertical sides than on the bottom. It was shown in preceding pages that most larvae attach to under horizontal surfaces and that as the angle departs from this the larvae set in smaller numbers. It was demonstrated that this behavior of the larvae is not due to the effect of light, but the suggestion was put forward that in the normal swimming position of the larva the foot is projecting upward and therefore is able to take hold most readily to the under horizontal surface. Actual setting, according to Prytherch, is a specific process, and larvae may crawl for some time before definitely attaching themselves, but the foot must take hold before final attachment. It would seem probable, from considera- tions of structure, that the larvae of other species may also attach most abundantly to under horizontal surfaces. Various factors, however, may influence the reaction, and it would be of interest to determine the activities of other species in this respect. Incidentally, Prytherch’s (1934) observation that the pigment spots are not light- sensitive organs but have another function is in accord with the present results in that no evidence of a directive influence of light was noted. SPAWNING AND SETTING OF OLYMPIA OYSTERS 499 By planting cultch periodically throughout each season and allowing each unit to remain in the water for only a few days it was possible to obtain a picture of the fre- quency of setting at all times. The results, of course, represent the potential catch at any time, rather than the number of larvae setting on the commercial beds, for in the experimental work new shells were planted twice weekly to provide clean cultch at all times, while the older shells on the grounds are usually fouled and unfavorable. In this manner it was shown that in most oyster-producing bays the setting season is not limited to a short period early in the summer, as thought by many oyster growers, but is of several months duration. This information has resulted favorably for the growers for they no longer plant cultch well in advance of the beginning of setting, as was the previous practice. The fact that after setting starts there is still a week or more before the time of maximum setting gives them sufficient time for the planting of cultch. The setting season consists of several distinct periods which in certain bays are remarkably uniform from year to year. The first period of the season is followed by a second major setting period 6 to 8 weeks later. There is a marked parallel between tidal periodicity and periodicity in the setting of larvae. The peak of a setting period coincides generally with the maximum tidal range of a run of spring tides. Therefore, after setting begins, one may determine from the tide tables the time of the following extreme tides when the rate of setting will be at a maximum. It is probable that the total tidal range is not so much the important factor, but the incidence of extreme low tides without regard to the height of the following high tides. Of practical importance is the very prolific late setting period, which follows the first on the next third and fourth spring tide periods; for oyster growers are able to plant cultch at tins time, also, thereby improving their chance of obtaining a satisfactory catch of seeds. The exact reason for the control of setting by tidal periods is not now definitely known. The beginning of spawning, however, is associated with the tides, for the water warms more rapidly during spring tides. After the minimum water temperature reaches the critical level for spawning there appears to be no connection between further spawning and tides. Orton’s (1926) observation that a maximum of spawning occurred at about the time of fidl moon may in some instances apply also to the Olympia oyster, but analysis of data on spawning during several years indicates that maxima of spawning, as judged from the findings of newly spawned eggs or young embryos, occur during neap tides as well as during full-moon and new-moon tidal periods. The relation between setting of larvae and tidal periods appears not to be traceable to a similar correlation between spawning and tidal periods. It appears most likely that Prytherch’s (1934) work on the effect of copper brought into the bays with land drainage may be applicable to the Olympia oyster, also. He reached the conclusion, from both laboratory experimentation and field observation, that precipitation of copper from solution in the inflowing river water permits the mature larvae to absorb this substance which is required for setting and metamor- phosis. For this reason natural oyster beds are always found in relatively enclosed bodies of water which receive a considerable inflow of land drainage. A period of extreme low tides permits a more effective mixing of the fresh water with the sea water, providing the required mineral for the larvae. He found that most larvae attached during low and early flood tides in the surface layer of the water when the salinity was lowest and the rate of current very slow. In the present work it was found that the best set of spat was caught, on floating cultch, within about 2 feet of the surface of the water, and that with increasing depth the frequency of attachment became less and less, Although during summer there is 500 BULLETIN OF THE BUREAU OF FISHERIES very little salinity difference from surface to bottom it may be sufficient to account for the results on the basis of Prytherch’s conclusion. More difficult to understand, however, is the fact that on the oyster grounds most spat are caught at relatively high tide, when the water is deep and of the maximum salinity, while at low tide, when the salinity is lowest and the amount of mineral from land drainage presumably in highest concentration, almost no larvae set. At this time oilier factors, such as low pH, may inhibit setting. It is clear, also, that during a period of extreme tides the fresher water entering the upper end of a bay goes farther down the bay and is most thoroughly mixed with the sea water. These results appear to permit interpretation in the light of Prytherch’s conclu- sion, though the specific factor involved is not definitely known. Although copper may be the controlling factor in the bays studied it is not difficult to conceive that other substances may act in a similar manner. That is, copper may be only one of a number of factors which may control the setting process. As a result of field observa- tions near Galveston, Tex., (Hopkins, 1931b), it was concluded that setting occurred only when the salinity was relatively high, in the neighborhood of 20 p.p.m., for in that place the salinity was frequently very low. Prytherch (1934) disagreed with this conclusion, although he demonstrated experimentally that the setting process pro- ceeds most rapidly at a salinity of 15 to 25 p.p.m. Very slow completion of attach- ment may be of considerable disadvantage to the larvae and thereby constitute the reason for the writer’s observation that spat were caught chiefly when the salinity was high. In addition to salinity and copper there may be other factors which determine the time and frequency of setting under different conditions. It is not possible to give an exact statement of the number of days required for larvae to reach the setting stage, though it was demonstrated that they develop for about 10 days within the maternal branchial chamber before being discharged. The free-swimming period appears to be 30 to 40 or more days, depending largely, perhaps, on water temperature, so that the total larval life is at least 40 days. This is about tliree times as long as that of Ostrea virginica (Prytherch, 1929). The long larval life permits wide dispersal but also subjects the larvae to various plankton-feeding organisms as well as to the effects of tides and storms. Mortality of larvae is necessarily large in any species. It may be estimated that oyster growers catch and grow not more than about one out of a million larvae pro- duced, when it is considered that the 4-year-old oysters discharge about 300,000 eggs and all of the younger individuals also propagate on a smaller scale. Mortality of spat is also tremendous. It was shown that during a period of profuse setting as many as 12,000 spat per day might be caught on the shells in one bag. Since there were generally only about 125 shells in a bag, each shell caught several hundred spat within a few days. Yet, after 1 year it is remarkable to find a shell with as many as 50 spat. Most of the mortality appears to take place within the first few weeks after setting, and while some of it is due to overcrowding it cannot all be traced to this cause. SUMMARY 1. Grounds on which Olympia oysters are grown are surrounded by dikes to retain a few inches of water over the oysters at low tide. The maximum range of tide at this place is about 20 feet, the average about 14 feet, and most grounds are located between the minus 2 foot and plus 4 foot tide levels. 2. Average water temperature varies between a winter low of 6° to 9° C. and a summer high of 18° to 20° C. In summer the temperatui'e is highest when the tide is SPAWNING AND SETTING OF OLYMPIA OYSTERS 501 low, and the shallow water often reaches 30° C., while during winter low tides occur at night and a temperature as low as about — 2° C. has been recorded. 3. Salinity of the water on the oyster beds at high tide varies, in Oyster Bay, between about 26 p. p. m. in winter and about 29 p. p. m. in summer; in Mud Bay the range is about 27 to 29.5 p. p. m. Salinity of the surface water, however, is subject to greater variation. 4. Hydrogen-ion concentration varies throughout the year from a pH of 7.7 to 7.8 in midwinter to about 8.4 in late spring. It is probable that prolific growth of algae in spring, in the presence of fertilizing substances brought in by the winter rains, accounts for the high pH at tins time. 5. Market-size oysters bear broods of 250,000 to 300,000 larvae. The number of larvae per brood depends generally upon the size of the maternal oyster. 6. Generally each oyster produces one brood per season, but in some years as many as 50 percent bear second broods while in other seasons as few as 75 percent of the individuals spawn as females. Abortions of embryos frequently occur, however. 7. Spawning of functional females begins in the spring when the minimum, or high tide, temperature reaches 12.5° to 13° C. 8. Most broods of larvae are produced during a period of about 6 weeks at the beginning of the spawning season, though an occasional gravid individual may be found as late as October. 9. An average period of 10 days is required for development within the branchial chamber from the time the eggs (diameter, 100m to 105m) are extruded from the gonad until straight-hinge veliger larvae (length of valves, 180m) are discharged. 10. As compared with oviparous species, development of the larvae of 0. lurida is very slow, and the age of the various stages may be stated approximately as follows: 1 day, blastulae; 2 days, gastrulae; 3 days, trochopliores ; 4 days, first conchiferous larvae with incomplete valves; 5 days, straight-hinge veliger larvae completely en- closed by valves 110m-120m long; 10 day, veliger larvae with valves 180m~185m long. 11. The free-swimming period is 30 or more days in length and varies from year to year, probably according to water temperature. 12. Larvae set most frequently on an under horizontal surface, while fewest catch on upper horizontal surfaces. A definite relationship exists between angle of surface and number of spat caught. 13. This setting behavior of larvae is not due to a directive influence of light but to the swimming position whereby the larval foot projects upward. 14. A special type of manufactured spat collector, designed to take advantage of these habits, is now in 'use commercially. 15. In Oyster Bay the setting season consists of two distinct periods, 6 to 8 weeks apart. Secondary periods of setting may occur between these two or after the second. 16. Setting seasons in Oakland Bay and Skookum Inlet are similar to those in Oyster Bay. In Mud Bay seasons are shorter and maxima occur at different times. 17. Times of maximum frequency of setting fall within periods of spring tides when tidal range is greatest. 18. On cultch suspended from floats most spat are caught at a distance of 1 to 2 feet from the surface. This appears to be one reason why high grounds catch the most seeds. Floats filled with cultch are now being employed commercially to take advantage of these results. 19. Few spat are caught at low tide, most when the tide is about half high. Frequency of setting appears to be associated with swiftness of current. 502 BULLETIN OF THE BUREAU OF FISHERIES 20. Setting of larvae begins in the third tidal period following that during which spawning starts. Setting later in the season appears to depend upon larvae remaining in the water from earlier spawning as well as upon larvae resulting from late spawning. LITERATURE CITED Amemiya, Ikusaku. 1926. Notes on experiments on the early developmental stages of the Por- tuguese, American, and English native oyster, with special reference to the effect of varying salinity. Jour. Mar. Biol. Ass’n., vol. 14 (N. S.), pp. 161-175. Plymouth. Churchill, E. P. 1920. The oyster and the oyster industry of the Atlantic and Gulf coasts. Rept. U. S. Com. of Fish., 1919 (1920), appendix VIII, 51 pp. Washington. Coe, Wesley, R. 1931a. Sexual rhythm in the California oyster {Ostrea lurida). Science, vol. 74, pp. 247-249. Coe, Wesley R. 1931. Spermatogenesis in the California oyster (.Ostrea lurida). Biol. Bull., vol. 61, pp. 309-315. Coe, Wesley R. 1932a. Development of the gonads and the sequence of the sexual phases in the California oyster ( Ostrea lurida). Bull. Scripps Inst, of Oceanography, Univ. of Calif., Tech- nical Series, vol. 3, pp. 119-144. Berkeley. Coe, Wesley R. 1932b. Season of attachment and rate of growth of sedentary marine organisms at the pier of the Scripps Institution of Oceanography, La Jolla, California. Bull. Scripps Inst, of Oceanography, Univ. of Calif., Technical Series, vol. 3, pp. 37-86. Berkeley. Coe, Wesley R. 1932c. Sexual phases in the American oyster ( Ostrea virginica). Biol. Bull, vol. 63, pp. 419-441. Coe, Wesley R. 1932d. Histological basis of sex changes in the American oyster ( Ostrea virginica). Science, vol. 76, pp. 125-127. Dean, Bashford. 1890. The present methods of oyster-culture in France. Bull. U. S. Fish Com., vol. X, pp. 363-388. Washington. Elsey, C. R. 1933. The Japanese oyster in Canadian Pacific waters. Fifth Pacific Science Con- gress, section B8, pp. 4121-4127. Elsey, C. R. 1935. On the structure and function of the mantle and gill of Ostrea gigas (Thun- berg) and Ostrea lurida (Carpenter). Trans. R. Soc. Canada, section V, pp. 131-160. Galtsoff, Paul S. 1929. Oyster industry of the Pacific coast of the United States. Report U. S. Com. Fish., 1929, appendix VIII, pp. 367-400. Galtsoff, Paul S. 1930a. The fecundity of the oyster. Science, vol. LXXII, pp. 97-98. Galtsoff, Paul S. 1930b. The role of chemical stimulation in the spawning reactions of Ostrea virginica and Ostrea gigas. Proc. Nat. Acad. Sci., vol. 16, pp. 555-559. Galtsoff, Paul S. 1932. Spawning reactions of three species of oysters. Jour. Wash. Acad. Sci., vol. 22, pp. 65-69. Galtsoff, Paul S., and R. O. Smith. 1932. Stimulation of spawning and cross fertilization between American and Japanese oysters. Science, vol. 76, pp. 371-372. Gutsell, J. S. 1924. Oyster cultural problems of Connecticut. Report U. S. Com. Fish., 1923, appendix X, pp. 1-10. Gutsell, J. S. 1926. A hermaphroditic viviparous oyster of the Atlantic coast of North America. Science, vol. LXIV, p. 450. Hopkins, A. E. 1931a. Temperature and the shell movements of oysters. Bull. U. S. Bur. Fish., vol. XL VII, pp. 1-14. Hopkins, A. E. 1931b. Factors influencing the spawning and setting of oysters in Galveston Bay, Tex. Bull. U. S. Bur. Pish., vol. XLVII, pp. 57-83. Hopkins, A. E. 1935. Attachment of larvae of the Olympia oyster, Ostrea lurida, to plane surfaces. Ecology, vol. 16, pp. 82-87. Hopkins, A. E. 1936. Ecological observations on spawning and early larval development in the Olympia oyster ( Ostrea lurida). Ecology (in press). Hopkins, A. E., Paul S. Galtsoff, and H. C. McMillin. 1931. Effects of pulp mill pollution on oysters. Bull. U. S. Bur. Fish., vol. XLVII, pp. 125-186. Hori, Juzo. 1933. On the development of the Olympia oyster, Ostrea lurida Carpenter, trans- planted from United States to Japan. Bull. Jap. Soc. Sci. Fish., vol. 1, pp. 269-276. McGinitie, G. E. 1930. The natural history of the mud shrimp, Upogebia pugettensis (Dana). Annals and Magazine of Natural History, ser. 10, vol. 6, pp. 36-44. SPAWNING AND SETTING OF OLYMPIA OYSTERS 503 Moebius, Karl. 1883. The oyster and oyster-culture. Report U. S. Com. Fish., 1880, appendix XXVII, pp. 683-751. Nelson, T. C. 1922. Report, Dept, of Biol., New Jersey Agr. Exp. Sta., year ending June 30, 1921, pp. 287-299. Nelson, T. C. 1928a. Relations of spawning of the oyster to temperature. Ecology, vol. IX, pp. 145-154. Nelson, T. C. 1928b. On the distribution of critical temperatures for spawning and for ciliary activity in bivalve molluscs, Science, vol. LXVII, pp. 220-221. Nelson, T. C. 1928c. Report, Dept, of Biol., New Jersey Agr. Exp. Sta., year ending June 30, 1927, pp. 77-83. Nelson, T. C., and E. B. Perkins. 1931. Report, Dept, of Biol., New Jersey Agr. Exp. Sta., year ending June 30, 1930, pp. 1-47. Orton, J. II. 1920. Sea-temperature, breeding, and distribution in marine animals. Jour. Mar. Biol. Assoc., United Kingdom, vol. XII (N. S.), pp. 339-366. Plymouth. Orton, J. H. 1926. On lunar periodicity in spawning of normally grown Falmouth oysters (0. edulis) in 1925, with a comparison of the spawning capacity of normally grown and dumpy oysters. Jour. Mar. Biol. Assoc., United Kingdom, vol. XIV (N. S.), pp. 199-225. Plymouth. Orton, ,T. H. 1936. Observations and experiments on sex-change in the European oyster, Ostrea edulis L. Part 5. A simultaneous study of spawning in 1927 in two distinct geographical localities. M6moires du Mus6e Royal D’Histoire Naturelle de Belgique, Deuxieme S6rie, Fasc. 3, pp. 997-1056. Prytherch, H. F. 1929. Investigation of the physical conditions controlling spawning of oysters and the occurrence, distribution and setting of oyster larvae in Milford Harbor, Conn. Bull. U. S. Bur. Fish., vol. XLIV, pp. 429-503. Prytherch, H. F. 1934. The role of copper in the setting, metamorphosis, and distribution of the American oyster, Ostrea virginica. Ecological Monographs, vol. 4, pp. 47-107. Stafford, J. 1913. The Canadian oyster, its development, environment and culture. Commis- sion of Conservation, Canada. Committee on Fisheries, Game and Fur-bearing Animals. 159 pp. Ottawa. Stafford, J. 1914. The native oyster of British Columbia ( Ostrea lurida Carpenter). Province of British Columbia, Report, Com. of Fish., year ending December 31, 1913, pp. 79-102. Stafford, J. 1915. The native oyster of British Columbia (Ostrea lurida Carpenter). Province of British Columbia, Report, Com. of Fish., year ending December 31, 1914, pp. 100-119. Stafford, J. 1916. The native oyster of British Columbia ( Ostrea lurida Carpenter). Province of British Columbia, Report, Com. of Fish., year ending December 31, 1915, pp. 141-160. Stafford, J. 1917. The native oyster of British Columbia ( Ostrea lurida Carpenter). Province of British Columbia, Report, Com. of Fish., year ending December 31, 1916, pp. 88-120. Stafford, J. 1918. The native oyster of British Columbia ( Ostrea lurida Carpenter). Province of British Columbia, Report, Com. of Fish., year ending December 31, 1917, pp. 91-112. Townsend, C. II. 1893. Report of observations respecting the oyster resources and oyster fishery of the Pacific coast of the United States. Report, U. S. Com. of Fish, for 1889 to 1891, pp. 343-372. Washington. o 4 ) U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner FURTHER NOTES ON THE DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS AT BEAUFORT, N. C. By SAMUEL F. HILDEBRAND and LOUELLA E. CABLE From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 24 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1938 Price 30 cents For sale by the Superintendent of Documents. Washington, D. C. . ■ FURTHER NOTES ON THE DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS AT BEAUFORT, N. C.1 By Samuel F. Hildebrand and Louella E. Cable, United States Bureau of Fisheries CONTENTS Page Introduction.. 506 Scomberomorus maculatus (Mitchill). Spanish mackerel, with notes on related species 508 Characters of the adult 509 Spawning 510 Descriptions of the young 510 A discussion of the relationship of the species of Scomberomorus and the probable identity of the young 517 Lagodon rhomboides (Linnaeus). Pinfish.. 518 Characters of the adult 519 Spawning 520 Descriptions of the young 520 Distribution of the young 524 Growth 526 Archosargus probalocephalus (Walbaum). Sheepshead 526 Characters of the adult 526 Spawning 527 Descriptions of the young 528 Distribution of the young 532 Food 532 Growth 533 Chaetodiplerus faber (Broussonet). Spade- fish 534 Characters of the adult 535 Spawning 536 Descriptions of the young 536 Distribution of the young 543 Growth 543 Family Gobiidae. Gobies 543 Distinguishing characters of the young of the genera Gobiosoma, Microgo- bius, and Gobionellus 545 A comparison of the eggs and the young of some American and Euro- pean gobies 546 Page Family Gobiidae— Continued. Gobiosoma bosci (Lacepede) and Go- biosoma ginsburgi, Hildebrand and Schroeder. Naked gobies 548 Key to the adults of the local species 548 Spawning 550 Descriptions of the eggs and young 551 Distributions of the young 558 Growth 559 Microgobius holmesi Smith. Holmes goby 559 Spawning 560 Descriptions of the young 560 Distribution of the young 563 Growth 564 Local species of Gobionellus 564 Gobionellus boleosoma (Jordan and Gilbert). Scallop fish 565 Spawning 565 Descriptions of the eggs and young 566 Distribution of the young 571 Growth 571 Gobionellus oceanicus (Pallas). Ocean goby 571 Spawning 572 Descriptions of the young 572 Distribution of the young 573 Family Blenniidae. Theblennies 573 Key to the genera and species 574 The characters of the eggs and newly hatched young 574 Distinguishing characters 574 A comparison of the eggs and young of some American and European blennies 575 1 Bulletin No. 24. Approved for publication, June 19, 1937. 505 506 BULLETIN OF THE BUREAU OF FISHERIES Family Blenuiidae — Continued. Hypsoblennius hentz (LeSueur). Spot- ted seaweed fish Spawning Descriptions of the eggs and young Distribution of the young Growth Hypleurochilus geminatus (Wood) Blenny Spawning Descriptions of the eggs and young Distribution of the young Growth Chasmodes bosquianus (Lac6pede). Banded blenny Spawning Descriptions of the eggs and the newly hatched young Page 576 577 579 589 589 589 590 592 602 603 603 605 605 Page The hakes of the genus Urophycis 612 Key to the species 613 Spawning 613 Descriptions of the eggs and young 614 Distribution of the young 626 Growth 627 Archirus fasciatus Lac6pede. American sole 630 Characters of the adult 630 Methods of collecting 630 Spawning 631 Descriptions of the eggs and young. . 632 Growth 640 Bibliography 640 INTRODUCTION The following accounts of the development and life history of a miscellaneous group of teleostean fishes is a continuation of earlier studies by the same authors, published in the Bulletin of the United States Bureau of Fisheries, volume XL VI, 1930, pages 383 to 488, and volume XLVIII, 1934, pages 41 to 117 (see Bibliography). Most of the specimens and data used were collected at Beaufort, N. C. However, some of the specimens and data were secured elsewhere, principally by Dr. Lewis Radcliffe and the late William W. Welsh, working aboard the Fisheries vessels Albatross, Fish Hawk, and Grampus. The authors were very materially assisted in the field work, carried on from 1925 to 1932, by the various members of the staff of the United States Fisheries Biological Station at Beaufort, N. C., especially by Dr. James S. Gutsell and Capt. Charles Hatsel, who accompanied one or both of the writers on many trips, and also collected independently. The drawings presented herewith were prepared by the junior author, unless otherwise stated in the legends. The junior author, also, did much of the tedious work of sorting the young fishes from other forms and the general debris usually taken in towings, and made some of the preliminary identifications. The senior author, under whose direction the work was carried on, is responsible for final identifications, the interpretation of the data, and for the preparation of the manuscript. The principal collecting stations are indicated with small circles on the map (fig. 1). One-meter tow nets, one at the surface and one on the bottom, hauled simultane- ously, were the only nets used at the farthest offshore stations. At the stations near shore, and at those in the partly enclosed waters, otter trawls and beam trawls also were used. Furthermore, collecting seines, particularly small ones made of bobbinet, were employed along the shores both in the inside waters and along the outside beaches. An otter trawl having the cod-end surrounded by bobbinet, built as a modified 1 -meter tow net, with the collar laced to the meshes of the trawl, was found very useful for collecting young fish. The fish taken in the bobbinet generally were past the larval stage and too active to catch with an ordinary 1-meter tow net, yet small enough to pass through the one-fourth inch square mesh of the collecting trawl. This apparatus proved very satisfactory at Beaufort, where there is little or no rough or rocky bottom. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 507 Figure 1. — Map of Beaufort Harbor and neighboring waters. Numbers on the map show the depth of the water in feet at the principal collecting stations. 508 BULLETIN OF THE BUREAU OF FISHERIES It will be seen from the following accounts that generally many gaps remain in the series of developmental stages of the various species treated. However, in every case enough new information is presented to make publication seem quite worth while. It has been possible, at least for some of the species discussed, to determine from the time and place of collection of the eggs or early young or both, the approximate duration of the spawning season and also the place of spawning, even though ripe fish were not seen. The movements or migrations of the young, too, were determined for some of the species from the places of collection of immature fish. Considerable in- formation relative to the rate of growth during the first several months of life also was gained for several species, and is shown in tables and graphs presented. All the species discussed in this paper, exclusive of the pinfish and the hakes, spawn during the summer, and either are scarce or absent in the local shallow water during the winter. The pinfish and the hakes, however, spawn during autumn and winter, and the young sometimes were taken in large numbers during the winter in company with young spots and croakers, the last named species also being winter spawners, as shown in an earlier paper by the writers (1930, pp. 417 and 433). The drawings of the eggs and newly hatched fish are based on living material. All the rest of the illustrations were prepared from preserved specimens. SCOMBEROMORUS MACULATUS (MITCHILL). SPANISH MACKEREL, WITH NOTES ON RELATED SPECIES The development of the eggs and the early larvae, up to 6 days of age, of the Spanish mackerel was described and figured by John A. Ryder (1882, pp. 135-172). It is now possible to describe and figure some older stages of Scomberomorus. The eggs used in Professor Ryder’s study were secured directly from ripe fish at several different points in Chesapeake Bay. The eggs, according to Ryder, float in sea water and vary in size from “one-twenty-fifth to one-twentieth of an inch in diameter.” They generally hatched in 24 hours. Segmentation proceeded quite regularly, as usual in teleostean eggs. The newly hatched fish was scarcely 2 mm long. When 3 days old the larvae had absorbed the contents of the yolksac, and the mouth was wide open. On the sixth day after hatching (length not stated), according to Ryder’s figure 17, the mouth had grown very large and wide with a sharp angle at the joints of the lower jaw. Prominent teeth already were present. This is the most advanced larva described and figured (anterior part of body only) by Ryder, and it seems to be identifiable with the smallest larvae now at hand. The specimens upon which the present study is based were caught in nets, mostly on the coast of North Carolina in the vicinity of Beaufort. However, among the larger young, specimens from Massachusetts, South Carolina, Georgia, Florida, Louisiana, Cuba, St. Lucia, and Panama also have been studied. All the larvae under 14 mm in length were taken at sea and mostly several miles offshore along the coast of North Carolina. Neither the larvae nor the older young were found numerous during the extensive collecting done in the vicinity of Beaufort. Never- theless, adult Spanish mackerel occur there in season (spring and fall) in sufficient abundance to be of considerable commercial value. However, comparatively few seem to remain during the spawning season. The ceros (locally pronounced “zero”) or kingfish, S. cavalla and S. regalis, are too scarce (especially the last named one) on the coast of North Carolina to be of much commercial importance. They are sought, however, by sportsmen, who prefer them to Spanish mackerel because they run larger in size. S. cavalla sometimes DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 509 attains a weight of 50 to 75 pounds, and S. regalis 25 to 35 pounds, whereas S', macu- latus attains a maximum weight of only 20 to 25 pounds. The average run in weight of these species respectively in the order named, however, is only about 7, 5, and 2 pounds. The Spanish mackerel and the ceros or kingfishes are all of wide distribution. The Spanish mackerel (S. maculatus ) ranges from Cape Cod, Mass., (sometimes as far north as Maine) south to Brazil on the Atlantic coast and from San Diego, Calif., to the Galapagos Islands on the Pacific. The ceros ( S . cavalla and S. regalis) are about equally as widely distributed on the Atlantic, though they do not occur on the Pacific coast. S. cavalla and S. maculatus are recorded also from the Atlantic coast of Africa. All the species seem to be chiefly of southern distribution, large quantities being taken in southern Florida, on the Gulf coast, and southward. The Spanish mackerel, like the other species of the genus, is migratory. It appears on the coast off Beaufort, N. C., in the spring, generally arriving in April, and it returns again in the fall, comparatively few remaining during the summer. According to local fishermen and fish dealers the fish of the spring run are poor and contain green roe. This statement is affirmed by our limited observations. In the fall the fish are fat and without roe. Spawning takes place during the summer, as shown subsequently, at a time when the fish locally are scarce. Therefore, the vicinity of Beaufort evidently is not an important spawning area. The ceros occur off Beaufort chiefly in the fall, and are scarce or absent the rest of the 3Tear. All three species of Scomberomorus discussed are taken in large numbers in southern Florida and on the Gulf coast during the winter, when they are absent in North Carolina and northward. CHARACTERS OF THE ADULT Adult Scomberomorus are recognized by the elongate, little-compressed body; long pointed snout ; large mouth with strong teeth ; and by a keel of skin on the sides of the tail posteriorly. The dorsal fin is long, composed of 14 to 18 feeble spines, 14 to 18 soft rays, followed by 7 to 10 separate finlets. The anal fin similarly is followed by about an equal number of finlets and the caudal fin is deeply forked. The general color is silvery, generally with spots and markings that differ among the species. The three species of Scomberomorus herein considered are rather closely related. However, S. cavalla is distinguishable by the more slender body and the abruptly decurved lateral line under the second dorsal. Furthermore, the origin of the anal usually is under the middle of the dorsal, whereas in the related species it is a little farther forward. Large individuals of S. cavalla are plain bluish above and silvery below, without spots, though young ones are described as having bronze spots. S. regalis differs from the other species in having scales on the pectoral fins and in having one or two continuous black lines along the side. In addition to the black lines it retains elliptical bronze spots throughout life. /S', maculatus, as stated elsewhere, runs smaller in size than the related species. It has no scales on the pectoral fins, and no black line on the side, though it has bronze spots. The lateral line, as in S. regalis, is more gently decurved under the second dorsal than in S. cavalla. In S. maculatus the anterior part of the first dorsal, back to about the fifth spine, is wholly black, whereas in the limited number of specimens of S. regalis examined only the outer two-thirds of that part of the fin are black, the base being white. 510 BULLETIN OF THE BUREAU OF FISHERIES SPAWNING A fairly full report on the spawning season of the Spanish mackerel, S. maculatus, was given by R. Edward Earll (1882, pp. 395-426). This writer made a special investigation and stated (p. 404) that this fish begins to spawn in the Carolinas in April, in Chesapeake Bay in June, and in the vicinity of Long Island not until the last of August. Mr. Earll stated, furthermore (p. 405): * * * The spawning season on our coast continues throughout the summer, and, in an}' particular locality, it lasts from 6 to upward of 10 weeks. * * * Again, a single individual is a number of weeks depositing its eggs, as shown by the fact that when the first are excluded a large percentage are still small and immature. It seems to us from the evidence obtained during the investigation upon which this report is based that Mr. Earll set the beginning of the spawning season (“in April”) too early for North Carolina. It has been stated already (p. 509) that the Spanish mackerel arriving off Beaufort in April and May contain green roe. Further- more, no larvae were collected there prior to June 28 (1927). Other young of sizes stated were taken in the vicinity of Beaufort as follows: Larvae 4.0 mm and less in length, June 28 (1927), August 17 (1927), and August 26 (1929); larger larvae up to 8.0 mm in length, July 12 (1915), and September 1 and 2 (1914) ; young 14 to 20 mm long, July 7 (1913), July 9 (1915), 2 and September 2 (1914) ; and specimens up to 80 mm long, August 15 (1913), and October 7 (1930). The larger young, that is, fish 14 mm and upward in length are capable of swim- ming and may have traveled some distance from the spawning ground. In fact, some of these larger young were taken in inside waters, whereas the smaller ones were caught only in outside waters. Larvae 8 mm and less in length, as already shown, have no fins and are quite helpless. Except as wafted about by currents and tides, they no doubt remain where they were hatched. It may be concluded, then, that a limited amount of spawning (for the young, as stated elsewhere, are not numerous) takes place in the open waters off Beaufort Inlet, and apparently none in the inside waters. Furthermore, larvae under 8.0 mm in length quite certainly are only several days old. As these small larvae appeared in the collection from June 28 (1927) to September 2 (1914), it may be concluded also that spawning takes place off Beaufort at least from the latter part of June to near the end of August. It cannot be stated definitely that the earliest larvae of any one season were taken, yet the absence of young in our collections prior to the end of June does in a measure confirm the statement of local fishermen and fish dealers, as well as our observations, that the fish of the spring run (April and May) at Beaufort are not ripe, and that spawning very probably does not take place in the vicinity of Beaufort until sometime in June. So far as we are aware nothing is known definitely about the spawning habits of S. cavalla and S. regalis. The limited number of fish examined, taken in the fall of the year, contained no roe. No evidence indicating that they spawn on the coast of North Carolina has been found. DESCRIPTIONS OF THE YOUNG It cannot be stated positively that the young fish herein described are all Spanish mackerel, for even the adults of the species of Scomberomorus are rather closely > Some of the small specimens used in the preparation of this report were collected as early as 1913 to 1915 by Dr. Lewis Radciiffe, formerly of the Bureau of Fisheries, who already had identified some of them provisionally when they fell into our hands. Therefore, we wish to credit Dr. Radciiffe with laying the foundation that made this report on Scomberomorus possible. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 511 related, and the young may not be separable. Nevertheless, it seems highly probable that we are dealing with Spanish mackerel only, as shown subsequently. The descriptions that follow are all based on preserved specimens. Considerable shrinkage takes place during the hardening process. Consequently the smallest larvae herein described, though shorter than the largest ones described by Ryder (1882) in the fresh state, are more advanced in development. Specimens about 2.5 mm long. — The body is robust, but the tail is long and slender, being notably longer than the head and trunk. The greatest depth is contained in the total length about 3.25 times. The mouth is very large and broad and strongly oblique; the gape reaches under the eye; and the lower jaw projects slightly and is straight and broad. Teeth already are plainly visible. The myomeres are indistinct anteriorly and posteriorly and therefore cannot all be enumerated. They appear to be rather numerous. Slight indications of rays are present above and below the tip of the tail. Pectoral fin membranes are prominent, with indications of rays. The general color of the preserved specimens is brownish. A dark spot just behind the symphysis of the lower jaw is at least sometimes present, and another one appears on the abdominal wall a short distance in advance*of the vent (fig. 2). £ The chief distinguishing character, and the one that seems to “link” these larvae with) the smaller and larger ones, is the large broad oblique mouth with well developed teeth. Specimens 3.0 to 3.5 mm long. — The caudal portion of the body has grown pro- portionately much shorter and deeper, the vent now being situated near midbody length, and the greatest depth is equal to the head, and is contained about 3.0 times in the1; length. The mouth remains large and wide, and has become more strongly oblique. Two depressions, one over the snout and another at the nape, are present and rather more prominent than in the smaller fish already described. Several prominent spines are present on the preopercular margin (which disappear in the adult). Three slender spines are developed in the anterior part of the dorsal finfold, though no soft rays are developed, a sequence contrary to that found in other species studied, and apparently contrary to the general rule in spiny-raj^ed fishes, in winch the soft rays most usually are developed before the spines appear. A variation in the relative length of the dorsal spines seems to exist among individuals of about this size and larger ones, as in some specimens the first spine is longest and in others the second one. A few dark spots are present along the ventral surface of the caudal portion of the body, and generally some dark markings appear on the dorsal wall of the abdominal cavity (figs. 3 and 4). 512 BULLETIN OF THE BUREAU OF FISHERIES Figure 3. — Scomber omorvs maculatus. From a specimen 3 mm long. Figure 4 .—Scomberomorus maculatus. From a specimen 3.25 mm long. Figure 5.— Scomberomorus maculatus. From a specimen 4 mm long. Specimens 4-0 to 4-%5 mm long. — Two specimens of this range in size are at hand. They are very similar to specimens 3.5 mm long, differing principally in the develop- ment of two additional spines in the dorsal fin, five spines now being present (fig. 5). DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 513 Specimens about 6.0 mm long.-— The large mouth has become much less strongly oblique, the gape anteriorly being below the level of the middle of the eye, whereas in specimens about 4.0 mm long it is at or above the upper margin of the eye. The maxillary has become narrower, and now reaches well beyond the middle of the eye. Teeth are very prominent. The snout has become much more pointed, and there is a sharp demarcation and depression where the premaxillaries apparently articulate with the skull bones. In advance of this depression or groove there is a pronounced hump in some specimens, which has a tendency to form a backwardly directed hook over the groove. This depression is distinct in the smaller fish described, though the premaxillaries are not definitely outlined. A second depression present at the nape in smaller fish has now disappeared. Dorsal spines have increased to seven in num- ber, and are relatively high, the longest one being a little longer than the snout. There is variation in the relative length of the dorsal spines, the first spine particularly being shorter in some specimens than in others. Soft rays still are imperfectly developed. The notochord is directed upward sharply, as usual at this stage in fishes destined to have homocereal tails. Myomeres are numerous, but cannot be counted accurately. Four spines, though reduced in size, remain present on the preopercular margin as in younger fish. Figure 6 .—Scomberomoru) maculatus (?). From a specimen 5.75 mm long. The general color of the long-preserved specimens at hand is brownish. The only color marking is a broad black band on the dorsal spines (fig. 6). No specimens between a length of 4.0 and 5.0 mm are at hand. Unfortunately the 5.0-mm specimen is imperfect, especially in having the dorsal spines broken. The next smallest specimen in good condition is 5.25 mm long. Considerable ad- vancement in development took place, if the larvae actually are all of one species, while the fish grew in length from 4.0 to 5.0 mm. The chief connecting “links” between the smaller specimens and the present group are: The very large mouth with prominent teeth; the preopercular spines, four in number in each stage; the retention of the depression over the snout, marking the articulation of the premaxil- laries; and the prominent dorsal spines. The great increase in length of the dorsal spines is somewhat disturbing in the absence of intermediate specimens. There can be 514 BULLETIN OF THE BUREAU OF FISHERIES no doubt, however, that these fish, if not of the same species as the smaller and larger ones herein described, at least are of a related species. Specimens 7.0 to 8.0 mm long. — Three specimens of this size are at hand. They differ little from the somewhat smaller ones described in the preceding section. The upper jaw is now slightly arched as in larger fish; dorsal spines have increased to eight, and remain high as in the smaller fish; soft rays are fairly definite in all the fins, though no articulations are evident; the caudal fin shows a tendency to fork; and the color apparently remains unchanged. Specimen 14 Tam long. — Only one specimen of this size is at hand, and none intermediate of this one and those described in the immediately preceding section. Therefore, a considerable gap remains. However, several similar and identical char- acters “link” this fish with the smaller ones, showing that if not identical they at least are representatives of related species. The 14-mm specimen is much more elongate than the smaller ones described, the greatest depth being contained about four times in the standard length. The snout has become still longer and more pointed, being contained 2.1 times in the head, and it projects well beyond the lower jaw. The groove at the articulation of the premaxillaries remains prominent. Spines on the preopercular margin have increased to eight. The maxillary has become strongly arched, and the teeth are large. The dorsal spines have increased to 19 (the usual number present in adult Spanish mackerel being 18 or 19), and the anterior ones, which were developed in the smaller fish, are proportionately lower. Although the bases of the second dorsal and anal are well outlined, the development of soft rays in these fins, as well as in the pectorals, is still retarded, whereas those of the caudal and ventrals are rather better developed. The origin of the anal is somewhat in advance of the second dorsal, whereas in adults its origin generally is under or behind that of the second dorsal. Dorsal and anal finlets are not yet definitely developed but thickenings in the fin membranes that will constitute the bases of the finlets are evident. The primitive fin membrane, however, remains continuous in each fin. The caudal fin now is distinctly concave. The general color of the preserved specimen is brownish, with some black pigment at the posterior end of the maxillary, and scattered black specks on the head and snout. The black band on the spinous dorsal present in smaller specimens remains, but is somewhat broken up into spots in the 14-mm fish. Blackish specks also are visible along the base of the second dorsal and anal fins (fig. 7). Except for the comparatively great change in the height of the anterior dorsal spines, the 14-mm fish connects up well with the 8.0-mm ones. Additional specimens DEVELOPMENT AND LI EE HISTORY OF SOME TELEOSTS 515 will be required to determine positively the identity of this fish and the smaller specimens mentioned. Specimens about 17 mm long. — Five specimens of about this size are at hand. Development is much further advanced than in the 14-mm fish already described. The finlets of the dorsal and anal, eight or nine in number in each fin, are more distinctly outlined, yet remain connected by the primitive membrane. The caudal fin is now definitely forked. The origin of the anal remains slightly in advance of the origin of the second dorsal. Pigmentation has increased somewhat. Some specimens are partly silvery in color. The black band on the spinous dorsal, very pronounced in younger fish, is now broken up into spots (fig. 8). Figure 8. — Scomberomorus maculatus. From a specimen about i7 nun long. Figure d.—Scomberomorus maculatuz. From a specimen 22 mm long. Specimens 22 to 25 mm long. — Ten specimens of about this size have been studied. The body has become rather more slender, the depth now being contained about 4.5 times in the standard length, and it remains rather strongly compressed. The head is long, about 2.5 times in the standard length; the snout is sharply pointed and projects strongly beyond the lower jaw, as in younger fish, its length being contained about 2.2 times in head. The groove at the articulation of the premaxil- laries remains evident. The mouth is very large and the teeth are strong, a pair of large canines in the upper jaw being on the part projecting beyond the lower jaw. The maxillary reaches somewhat past the middle of the eye and is contained about 1.4 times in the head. Only two preopercular spines remain. The origin of the anal is now only slightly in advance of that of the second dorsal. The rays of these fins remain imperfectly developed, though those of the other fins are well formed. The finlets are all well developed and separate. The caudal fin is well forked, but not as broadly as in the adults. Pigmentation has progressed fairly rapidly. The general color is silvery, though the back has a brownish cast. The black markings are shown in figure 9. 516 BULLETIN OP THE BUREAU OF FISHERIES Specimens 85 to Jfi mm long. — Six specimens of about this size are at hand. The advancement over the 25-mm fish is not great. The upper jaw projects less promi- nently, and the articulation of the premaxillaries no longer is marked by a definite groove. The preopercular spines have been almost completely absorbed. In some specimens of this size a slight indication of a lateral line is present. The second dorsal and anal now have attained more nearly the relative position occupied in adult fish, as the origins are about opposite each other. (In adult S. maculatus and S. regalis the origin of the anal generally is slightly behind that of the second dorsal, whereas in S', cavalla it often is nearly under the middle of the second dorsal). The rays of these fins now are quite fully developed. The following counts are based on one specimen: D. XIX-17-VII; A. II, 17- VIII; vertebrae 22+31=53. Pigmentation remains about as in the 25-mm fish, except that black points on the middle of the side have become more numerous and more concentrated and tend to form a lateral band posteriorly. Much variation in the amount of black pigment on the spinous dorsal occurs among specimens. Generally the anterior part of the fin, to the fourth or fifth spine, is mostly black. Specimens 60 to 70 mm long. — The body remains quite strongly compressed. It does not differ in proportionate depth from the smaller fish described in the preceding section. The head is proportionately much shorter, however, being contained about 3.3 times in the standard length. The snout projects less prominently, the anterior pair of enlarged teeth no longer being far in advance of the mandible. The maxillary remains slightly arched, but much less so than in smaller specimens, having become gradually less bent since a length of about 14 mm was attained, reaching below posterior margin of pupil, and being contained about 1.7 times in the head. The lateral line is fairly well defined. It is curved downward rather gradually under the anterior rays of the soft dorsal, and posteriorly it is undulating. Gill rakers are very short, mere points, eight or nine can be seen. Vertebrae 23+30=53 in one specimen counted. The body is mostly silvery; more or less brownish on the back. The anterior part of the spinous dorsal, involving 3 to 5 spines, is wholly black, the rest of the fin has only a black margin, precisely as in adult Spanish mackerel. Specimens 85 to 100 mm long. — Fish of this size and for sometime afterwards remain more strongly compressed than adults. The upper jaw projects little at this size, the teeth remain strong, but less so than in younger fish; and the caudal fin is now broadly forked, about as in the adult. No dermal keel is as yet evident in the lateral line on the caudal peduncle. The following counts and proportions are based on a specimen 97 mm long: Head 3.8; depth 4.6 in standard length; snout 2.8, maxillary 1.8 in head; D. XVII-16-VIII; A. II, 14-IX; gill rakers minute, 9; vertebrae 22+31 = 53. The color is bright silvery, rather bluish silvery above. No spots or lines are dis- cernible in the preserved specimens at hand (fig. 10). Specimen 160 mm long. — A single fish of this length (and none intermediate of this one and one 115 mm long) is at hand. The 160-mm fish does not differ greatly from the smaller group described in the preceding section. The proportions given for the smaller fish have not changed. No change in color appears to have taken place. No indications of spots or other markings are present on the body in the old preserved specimen studied. Specimens 210 to 225 mm long. — One specimen of each length given is at hand. These fish were recently preserved. The larger one has dark spots (yellow in life) on the sides as in the adults. The smaller one has none, which seems to show that the DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 517 spots are not always developed at a length of 210 mm. In structure these fish do not differ essentially from the 160-mm fish already described. These specimens still remain rather more compressed than large fish. The maxillary remains gently arched, and the snout sharply pointed, projecting slightly, just as in the smaller fish described in the foregoing section. The lateral line remains unchanged, being rather gently decurved under the anterior part of the soft dorsal, precisely as in adult Spanish mackerel. The gill rakers have increased somewhat in length, though they do not yet exceed a fourth the length of the pupil. Scales now are present on the soft dorsal and on the anal, though none can be seen on the pectorals. The dermal keel on the caudal peduncle is quite evident. A DISCUSSION OF THE RELATIONSHIP OF THE SPECIES OF SCOMBEROMORUS AND THE PROBABLE IDENTITY OF THE YOUNG The relationship of the three species of Scomberomorus known from the Atlantic coast of the Americas, is rather close, as indicated elsewhere. We recognize only one species among the young studied, though the identity of the 6.0- to 8.0-mm specimens described is somewhat doubtful, owing to some missing stages. If the very young are separable into species it would be necessary to use characters different from those employed in recognizing the adults. We have not discovered any distinguishing “juvenile” characters. The first “adult” character that develops, which apparently is of some value in distinguishing the kingfish, S. cavalla, from the other local species of the genus, is the relative position of the soft dorsal and the anal. In the kingfish the origin of the anal is well behind the origin of the soft dorsal, often nearly under the middle of the soft dorsal, whereas in the other two species the origin of the anal is under or more usually slightly posterior to the origin of the soft dorsal. The soft dorsal and anal are not well developed until the fish reach a length of about 14 mm, and it is not until the fish reach a length of about 35 mm that the relative posi- tion occupied in adults is attained, as the origin of the anal is in advance of the second dorsal in smaller fish. The relative position of these fins remains unchanged in all the larger young (35 mm and upward in length) studied, the origin of the anal being slightly posterior to that of the second dorsal. Therefore, the specimens in our collection probably cannot be identified as S. cavalla. The next distinctive character that develops is the lateral line, which is abruptly decurved under the second dorsal in the kingfish, S. cavalla, and rather gradually in the other species. The lateral line sometimes is evident in specimens 70 mm long, but often not until later. Judging from the course of the lateral lines the kingfish again seems to be missing among the specimens that could be checked for this character. 518 BULLETIN OF THE BUREAU OF FISHERIES The Spanish mackerel, S. maculatus, and the spotted cero, S. regalis, are closely- related, apparently distinguishable by the presence of scales on the pectoral fins, and by the presence of one or two longitudinal black streaks along the side of the latter. It is not known at what size these distinguishing characters develop. It can only be stated now that no scales are present on the pectoral fins in any of the young at hand. Neither are dark stripes present. In two adult S. regalis examined the anterior part of the spinous dorsal is not wholly black, the lower third or so being white. This is shown also in an often reproduced drawing. Adult Spanish mackerel, S. maculatus, examined have the anterior part of the fin, involving from three to five spines, wholly black. This is true of young Scomberomorus of about 60 mm and upward in length that are at hand. The indications, therefore, are that at least the larger young studied and described are Spanish mackerel. Because of the scarcity or absence of ceros and kingfish during the probable spawning season (spring and summer), there is reasonable doubt that these fishes spawn on the coast of North Carolina. Therefore, the young Scomberomorus taken there very probably are all Spanish mackerel, even though their identity cannot be established positively by taxonomic characters. LAGODON RHOMBOIDES (LINNAEUS). PINFISH The name pinfish is most generally used for this species, though at Wilmington, N. C., it is sometimes called sand perch, and southward the name “sailor’s choice’’ is heard. The name pinfish suits the species well because of its extremely sharp spines. The pinfish is known from Cape Cod, Mass., to Texas, and is also reported from Bermuda and Cuba. On our coast it is common from Virginia southward. Its com- mercial value is not great, however, because of the small size attained. The maximum length reported is 13 inches, but the average length probably does not exceed 6 inches. In the statistical report of the Bureau of Fisheries for 1935, for example, it is listed only from North Carolina (180,000 pounds) and Florida (31,000 pounds). It is marketed in limited quantities in other States, mostly with other species as “mixed fish.” Therefore, the exact amount marketed is not obtainable. The pinfish is of good flavor, and no doubt the demand would be greater if the fish attained a larger size. Occasionally when large catches, running small in size, are made at Beaufort they are taken to the menhaden reduction plants and made into fish meal or fish scrap and oil. The pinfish is said to yield a very high grade of oil. The pinfish is one of the comparatively few species that is a year-round resident in the shallow water of the estuaries, bays, and sounds at Beaufort. It seems to withstand cold rather better than most of the other species that winter locally. For example, on January 7, 1926, and again January 4, 1928, during rather continued abnormally cold weather many individuals of such species as the speckled trout ( Cynoscion nebulosus ), the spot ( Leiostomus xanthurus ), and the croaker ( Micropogon undulatus), became numb and floated at the surface. No pinfish were seen among them. However, on December 28, 1925, a large number of this species (5 gallons), mostly rather large ones, were frozen to death in a “fish pool” from which they could not escape and which contained only about a foot of water at low tide. The tem- perature of the water at the time the fish perished is not known, but the air tem- perature dropped to 12° F., which is unusually low for Beaufort. It is of interest that some small mullets ( Mugil cephalus ) about 6 inches in length, that had been confined in the pool with the pinfish survived, indicating that this mullet can stand even more cold than the pinfish. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 519 The pinfish is a nuisance in some respects to anglers because of its ability to cut the bait off the hook without itself getting caught, and to net fishermen because when “gilled” it is hard to remove. The sharp spines make the fish difficult to grasp without injury to the hands. Furthermore, a small sharp spine, directed forward, precedes the dorsal fin. This procumbent spine prevents the fish from being forced through the mesh (the usual procedure in removing gilled fish) without lifting the thread over the spine. The fish taken in this way generally are too small to be of much value. Therefore, fishermen often have to labor long at the disagreeable task of removing the fish from the nets without receiving anything in return. The pinfish also was a most annoying pest to an investigator who had confined crabs in floating wire cages for the purpose of studying their life histor}7. The fish continually mutilated the crabs by biting off their legs and other appendages. The fish could be observed readily while “working” around the float and when underneath it they swam completely upside-down, that is, with the ventral side upward. Figure ll. — Growth curve based on length measurements of 3,348 f.agodon rhomboides of the 0-class. Solid line, average length for each month of all fish measured; dot and dash (upper) line, largest fish; dot and several dashes (lower) line, smallest fish. CHARACTERS OF THE ADULT The pinfish ( Lagodon rhomboides ) belongs to the family Sparidae with the sheeps- head ( Archosargus probatocephalus), discussed elsewhere in this publication. The adult pinfish is most readily recognized by its rather deep, compressed body, crossed by four to seven dark bars, and by its prominent deeply notched incisor teeth. The depth of the body is quite variable among specimens, being contained 2.2 to 2.9 times in the length to the base of the caudal. The head is rather long, and the snout is moderately pointed, being notably longer than the eye. The mouth is small and 154979 —38 — 2 520 BULLETIN OF THE BUREAU OF FISHERIES horizontal, with the maxillary reaching only to the eye. D. XI or XII, 10 to 12; A. Ill, 10 to 12; scales 62 to 68; vertebrae 9 + 15. SPAWNING Many adult fish were examined in the vicinity of Beaufort, from 1914 to 1917 and from 1925 to 1931, as to the state of development of the gonads. However, no ripe fish, nor even fish with developing roe, was found, notwithstanding that Smith (1907, p. 300) reported that a ripening female was seen at Beaufort on August 6 (1903), and a ripe male on November 20 (1903). The virtual absence of ripe, or ripening, fish in the local inshore waters suggests strongly that the fish go elsewhere to spawn. The collection of small young near the inlet and at offshore stations, onty, as shown subsequently, seems to indicate that spawning takes place at sea, probably a consider- able distance offshore. The presence of rather early young in the local waters over a long period of time, as shown in the following paragraph, indicates a long spawning season. Young, 10 mm and less in length, first appeared in the tow toward the end of October, and continued to be taken each succeeding month until toward the end of April. However, they were most numerous in December and January. The presence of such small fish over this long period of time seems to show that spawning begins in October and that it continues until the following March. All the smaller young, consisting of 242 fish of 10 mm and less in length, either were taken at offshore stations or in or near Beaufort Inlet, some of the stations being as much as 12 or 13 miles offshore, beyond which no collecting was done. However, as the eggs and early larvae, or fry under 5.0 mm in length, were not found, it seems prob- able that spawning takes place beyond the most distant stations made. Therefore, the young taken at sea presumably were migrating from the spawning grounds to the inshore waters where the larger young, of 12 to 15 mm and upward in length, and the adults are numerous. DESCRIPTIONS OF THE YOUNG The early larvae were not taken, and therefore remain unknown, as already stated. Specimens about 5.0 to 5.5 mm long. — Two specimens with damaged caudal fins are at hand. The body is decidedly elongate and compressed, the depth being con- tained 3.6 to 3.9 times in the length without the caudal fin. The dorsal outline is concave just in advance of the eyes and also at the nape, or just posterior to the brain, which is visible through the thin walls of the skull. The head is rather low, compressed 2.9 to 3.0 in length. The snout is moderately pointed, as long as the eye, 3.0 to 3.5 in head; the maxillary reaches nearly opposite the anterior margin of the pupil; and the gape anteriorly is very slightly below the level of the middle of the eye. Teeth are not evident. About 22 myomeres may be counted. The vent is situated slightly nearer the base of caudal than tip of snout. The primitive dorsal fin membrane in the 5.0-mm fish has suggestions of rays in the region of the anterior part of the soft dorsal of the older fish. These rays are considerably further developed in the 5.5-mm specimen. Rays are rather more definitely developed in the anal fin than in the dorsal. The notochord is bent upward posteriorly, and well-developed caudal fin rays appear below it, which are broken distally. Therefore, the exact shape of the fin cannot be determined. However, as somewhat larger specimens have a rounded caudal, it may be assumed that the fin also was more or less rounded in the small specimens. Pectoral fins are quite well developed and rather long, but the ventrals are minute. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 521 The general color is pale. On the median ventral line are three dark spots, one near the isthmus, another on the chest, and a third one just in advance of the vent. A row of black dots occurs along the ventral outline from the origin of the anal to the base of the caudal. A dark area, apparently internal, is visible on the side above and slightly posterior to the vent (fig. 12). Specimens 6.0 to 7.0 mm long.-— The advancement over the 5.0-mm fish, already described, is not great. The body apparently has become slightly more elongate, the depth in three specimens measured is contained 3.8 to 4.0 times in the length to the base of the caudal fin. The concavities in the dorsal outline (in advance of the eyes and at the nape) remain, but are less pronounced. No change, worthy of note, has taken place in the shape of the head, snout, eyes, or mouth. The principal advance- ment is the development of more definite soft rays in the dorsal and anal fins, of which about 12 can be counted in each fin. The spines, however, are not yet well differen- tiated. The caudal fin is quite long and round. The pectoral fins are long and reach to the vent, but the ventral fins are scarcely differentiated. The black dots, present in the smaller fish described, persist and are more definite. In addition a few to several black dots now are present on the base of the caudal, two or more on the upper surface of the caudal peduncle, one at the nape, and generally an elongate blackish one above the base of the pectoral (fig. 13). Figure 13.— Lagodon rhomboides. From a specimen 7 mm long. Specimens 8.0 to 10 mm long. — Development has progressed rather slowly. The body has become somewhat more slender, but it remains about equally com- pressed, the depth now being contained 4.3 to 4.6 times in the length to the base of the caudal. The dorsal outline remains as in the smaller fish, except that depressions in advance of the eyes and at the nape have disappeared, but the brain remains visible. The head now is contained 3.5 to 3.6 in head; eye 2.9 to 3.1 in head; and the snout 3.0 to 3.3. The mouth remains oblique, with the maxillary reaching nearly opposite anterior margin of pupil. Jaw teeth now are evident, but contrary to most spiny 522 BULLETIN OF THE BUREAU OF FISHERIES rayed fishes studied, no spines are visible at this size on the preopercular margin. The vent now is situated at midbody length, without caudal. The development of the fins has progressed rather rapidly. The spines in the dorsal and anal are well differentiated; the caudal fin is long and round, being nearly as long as the head; the pectoral fins, too, are long, reaching the vent; but the ventral fins are minute, being scarcely longer than the pupil. The only change in color, worthy of note, is the development of additional dark dots along the ventral outline of the chest and abdomen, which vary in number among individuals. Some specimens also have developed a few extra chromato- phores on the dorsal surface of the head. Specimens IS to 15 mm long. — No measureable changes in the proportions of the body have taken place. However, the snout has decreased in proportionate length and is definitely shorter than the eye, 3.6 to 4.0 in the head, whereas the eye is con- tained 2.8 to 3.0 times in the head. The mouth remains oblique, the gape anteriorly being only slightly below the level of the middle of the eye; the maxillary reaches only slightly beyond the anterior margin of the eye; and the teeth remain minute. The skull remains transparent, leaving the brain plainly visible from above. The rays in the dorsal and anal are all developed as the usual number present in adults may be counted, but the spines remain proportionately much shorter than in the adult. The caudal fin becomes square when the fish attains a length of about 12 mm and is definitely concave at a length of about 14 mm. The pectoral fins remain long, reaching nearly to origin of the anal; and the ventral fins have increased greatly in size, being nearly as long as the eye in 15-mm fish, but the spine is not yet well differentiated. No changes in color markings worthy of note have taken place since a length of about 10 mm was attained (fig. 14). Specimens 18 to SO mm long. — Specimens of this length are variable in shape and color. Some specimens up to 20 mm in length remain quite as slender as 15-mm fish, whereas others are notably deeper. The slender specimens of this size are as void of pigmentation as 15-mm fish, whereas the deeper bodied specimens are profusely pigmented and have dark cross bars as in the adult. A few specimens only 16 to 17 mm long already have increased considerably in depth and have evident cross bars, whereas others up to 20 mm in length remain slender and pale. It is evident, there- fore, that pigmentation and the deepening of the body take place simultaneously and that these changes occur at varying lengths. These changes apparently are associated with a change in habitat, as shown subsequently. The depth is contained in the length to the base of the caudal 4.3 to 4.5 times in three unpigmenfced specimens measured, whereas in three pigmented fish of the DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 523 same size it is contained 3.5 to 3.9 times. Other proportions do not differ measur- ably in the two groups, the head in six specimens (three of each group) measured being contained 3.3 to 3.6 times in the length without the caudal fin, eye 2.8 to 3.2 in the head, and the snout 3.3 to 3.8 in head. The teeth remain small and equally developed in the pigmented and unpigmented specimens. Pigmented specimens about 20 mm in length are at least partly covered with scales but smaller pigmented fish and unpigmented ones, up to 20 mm in length, have none. The ctenoid character of the scales is evident as soon as the scales are developed. The fins are longer and more fully developed in the pigmented fish, though the ventral spine is differentiated in each group. The pigmented specimens, however, have the first soft ray of the ventral produced into a short filament, which is not present in unpigmented fish. The caudal fin is deeply concave in all specimens. The specimens referred to in the foregoing paragraph as unpigmented retain a few dark markings, essentially as in much smaller fish. The pigmented ones already are more or less greenish in life. The preserved specimens, as seen under magnifica- tion, are profusely dotted with black; these dots being concentrated in certain places where they form cross bars. The dark spots extend more or less on the dorsal and anal fins (fig. 15). Specimens 25 to 30 mm long. — The body is strongly compressed and it has con- tinued to increase in depth, which now is contained 2.5 to 3.0 in the length without the caudal fin, proportions found also in adults. The dorsal profile is strongly ele- vated and round, being much more strongly curved than the ventral outline. The head is rather short and deep, being contained 2.8 to 3.1 times in the length without the caudal fin; the snout remains blunter and proportionately shorter than in the adult, 3.5 to 3.8 in head; eye 3.1 to 3.5. The mouth has become almost horizontal, the gape being wholly below the eye; the maxillary reaches slightly past the anterior margin of the eye; and the anterior teeth are somewhat enlarged. The exposed tips of the anterior teeth are pointed, and under magnification it is evident that these tips arise in pairs from a common base. The body is fully scaled; the pectoral and ventral fins remain shorter than in adults; the first soft ray of the ventral retains a short filament, which reaches the origin of the anal; and the second anal spine already is stronger than the third one, though not as much so as in the adult. In the general color pattern specimens 30 mm long do not differ greatly from larger fish (fig. 16). 524 BULLETIN OF THE BUREAU OF FISHERIES Specimens 40 mm and upward in length. — The body is quite variable in depth among individuals, and it may be deeper in rather smaller fish than in much larger ones. For example, the depth is contained 2.3 times in the length to the base of the caudal in a 40-mm specimen, whereas in a 125-mm one it is contained 2.6 times. The snout continues to become more pointed and proportionately longer with age, being contained 3.1 times in the head in a 40-mm fish and 2.5 times in 165-mm ones. The mouth has become horizontal and much below the eye, and the teeth already are broad and well notched in fish 40 mm long. The caudal fin becomes more deeply forked with age, and the lobes become more sharply pointed. The pectoral and ventral fins increase in length and become pointed with age. The pectoral fins reach the vertical from the vent in specimens about 40 mm long, whereas in large specimens they reach beyond the origin of the anal. Specimens up to about 100 mm in length retain the filament on the first soft ray of the ventral which thereafter decreases in length and is missing in large fish. The second anal spine, though already longer and stronger than the third in 40-mm fish, increases considerably in thickness in larger fish. The color is extremely variable. Dark cross bars are present in all specimens at hand, though variable in intensity. Furthermore, some specimens have prominent alternating bluish- and yellowish-green longitudinal lines which are indistinct or want- ing in others (fig. 17). DISTRIBUTION OF THE YOUNG The smallest young (5.0 to 10 mm in length) secured were all collected offshore or in or near Beaufort Inlet. These young presumably were en route from the spawn- ing grounds to inshore waters. Exactly where spawning takes place is not known. However, the indications are that it occurs a considerable distance offshore. The early larvae (under 5.0 mm in length), missing in the collections, which were made within 12 or 13 miles from the shore, would be expected to occur near the spawning grounds, as such small fish, lacking self-directive powers, could not have drifted far. The discovery of the habitat of these early young therefore remains for future in- vestigation, but apparently should be sought a considerable distance offshore. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 525 The smallest fry collected, ranging from 5.0 to 10 mm in length, were nearly all taken with 1-meter tow nets, and mostly at the surface. Young ranging upward of 10 mm in length enter the sounds and bays and estuaries freely, though some remain in the open outside waters until a length of about 20 mm or more is attained. After the fish enter the inshore waters they tend to settle more or less on the bottom. Many of the young after entering the inshore waters occupy areas overgrown with eelgrass or other bottom growths, a habitat also occupied by their relatives, the young sheepshead. It is interesting that the fish occupying the weedy areas become pigmented much earlier than those remaining in open water. In the description of young 18 to 20 mm in length it is shown that some specimens are well pigmented at a length of 18 mm and a few at even a smaller size, whereas others still are virtually unpig- men ted when 20 mm long. It is shown, also, that a pronounced deepening of the body occurs simultaneously. In every instance the specimens acquiring pigmenta- tion, as well as the deeper body early, were collected in weedy areas, whereas the large (up to 20 mm in length) unpigmented ones were taken in open water. During the winter (December to February) young ranging from about 12 to 16 mm in length are often numerous in the deeper channels, in the sounds and estuaries, in company with young croakers and spots of about the same size. At this season of the year large schools of young pinfish, and spots, were frequently seen in quiet coves of the breakwater and jetties along the eastern shore of Pivers Island. When winter is over young pinfish are not abundant in the deeper channels, as they then chiefly occupy the shallow weeded aieas; and they are seen also around piers, breakwaters, jetties, wrecks, etc., where presumably they find the food they need, which seems to be virtually identical with that of the sheepshead (see p. 532). 526 BULLETIN OF THE BUREAU OF FISHERIES GROWTH The rate of growth of the pinfish during the first several months of life, as shown in figure 11, is very slow, owing no doubt to the cold weather of winter. Other common fall- and winter-spawned fish, such as the spot and the croaker, also grow slowly at first (Hildebrand and Cable, 1930, pp. 426 and 443). However, in May when the water became warmer, and food probably became more plentiful, the fish began to grow rather rapidly, and this rate of growth apparently was maintained for several months, not slowing down again until after September. According to measurements taken the fish ranged from 50 to more than 100 mm in length when a year old. Since the average length of the usual catch of adult pinfish at Beaufort probably does not exceed 150 mm (6 inches) early maturity is probable. As some of the fast-growing fish already exceeded a length of 100 mm (4 inches) at one year of age, it seems highly probable that the faster-growing fish spawn in their second year. ARCHGSARGUS PROEATOCEPHALUS (WALBAUM). SHEEPSHEAD The eggs of the sheepshead are known only from a brief account by Rathbun (1892, p. L1X; republished in “A Manual of Fish Culture”, 1898, pp. 224-225; and revised edition, 1904, pp. 226-227). They are described as transparent, buoyant, about one thirty-second inch (about 0.8 mm) in diameter, requiring 1,600,000 eggs to fill a fluid quart. The eggs hatched in about 40 hours in a water temperature of 76° or 77° F. Unfortunately the development of the embryo was not studied, the work having been carried on aboard the Fisheries steamer Fish Hawk merely with the view of working out practical means of propagation. Neither were the young described beyond stating that they are “* * * very small, but active and strong and with- stand considerable rough handling.” It is not yet possible to add anything to the foregoing meager account of the eggs. Nor is it possible to give a complete picture of the development of the young. We can give an account only of the development of the young of about 6 mm and upward in length. Such fish are much smaller, however, than any previously described. The very small size at which the young acquire the characters of the adult, as shown sub- sequently, is quite remarkable. The salt water sheepshead is of wide distribution, ranging from Cape Cod, Mass, (rarely to the Bay of Fundy), south to Tampico, Mexico. At Beaufort, N. C., it is a year-round resident, though more numerous in the summer than winter. The sheepshead is sought not only by commercial fishermen, but also extensively by anglers, as it is one of the gamest of salt-water fishes. It is a food fish of excellent flavor and brings a good price in the market. It attains a maximum weight of 30 pounds according to published reports. However, the largest one which we saw at Beaufort, during 10 years of intermittent angling, collecting, and observation of fishermen’s catches, was a female weighing 12 pounds (length not recorded). This fish was taken in a seine on Shackleford Banks (inside) by commercial fishermen. Fish weighing from 1 to 2 pounds (11 to 15 inches long) make up the principal catch of the angler locally, though individuals up to 5 pounds (20 inches in length) are not rare. CHARACTERS OF THE ADULT Adult sheepshead are characterized by the oblong, deep, compressed body, crossed by about seven black bars on a greenish-yellow background. The mouth is rather small, nearly horizontal, and is provided in front with incisor teeth, which are entirely DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 527 or only slightly notched (not deeply notched as in some related species). The posterior teeth are composed of strong molars, which are used for crushing crustaceans and mollusks. The dorsal fin is long and continuous, being composed of 11 or 12 strong spines and 11 to 13 soft rays, and it is preceded by a short spine directed forward (more or less imbedded in large specimens). The caudal fin is forked; the anal con- sists of 3 sharp spines (the second one the largest) and 10 or 11 soft rays; and the pectoral fins are long, reaching to or beyond the origin of the anal. Vertebrae 9 + 15. SPAWNING No opportunity was found to examine a large number of sheepshead as to the development of the gonads. Of the comparatively few fish examined in the spring, when spawning evidentally takes place (many more specimens became available for dissection during the summer and fall) only one fish, taken June 16 (1926), contained fairly well developed roe. Therefore, little information was gained from that source. Neither were the eggs secured, or if so they were not recognized.3 Therefore, the period of time when comparatively small young appeared in the collections must serve as the chief indication of the time and duration of spawning. As very small young, under 6 mm in length, are not represented, the spawning areas in the vicinity of Beaufort cannot be determined from the collections. According to Rathbun (1892) the sheepshead spawns along sandy shores in southwestern Florida. A sandy shore is not the usual habitat of the sheepshead, which lives principally among rocks, piers, breakwaters, wrecks, sunken logs, and debris, and in Florida among mangroves. Therefore, it seems to leave its customary habitat to carry out its reproductive activities. Efforts were made repeatedly to catch ripe adults and the larvae on sandy shores in the vicinity of Beaufort, but with- out much success. The smallest specimen taken, however, was caught in Shackleford Channel, just off a sandy beach. All the other smaller young, ranging from 7.5 to 65 mm in length, were caught in “meadows” of seaweeds. Since the eggs are pelagic the larvae, also, no doubt, are pelagic. However, the young fish seem to abandon this habitat early in life, as indicated by the collections at hand. Rathbun (1892) stated, furthermore, that it was necessary to haul the nets after 4 o’clock in the afternoon to catch ripe females, the best time being about sunset. Late evening spawning seems to be quite general among marine fishes producing pelagic eggs. The smallest young secured at Beaufort was taken on May 20 (1930), in a tow net hauled at the surface in Shackleford Channel. This fish apparently was still living in its larval habitat, though already well past the larval stage. Why only this single specimen was taken in the pelagic stage cannot be explained, as great effort was made to secure others, many hauls with tow nets having been made in the same general vicinity from 1927 to 1931. Apparently the fish simply were not there. The next smallest young, ranging from 7.5 to 18 mm in length, were caught June 21 (1926), with a bobbinet seine hauled in eelgrass along the shores of Fivers Island. Small young, 11 to 21 mm in length, were taken as late as July 8 (1931). Thereafter they ran larger in size. However, a few specimens of 19 to 21 mm in length, and one 25 mm long, were taken as early as June 14 (1929). The range in length of the young collected each month, arranged in 5-mm groups, is shown in figure 23. * The several years of experience gained in endeavoring to identify marine-fish eggs taken in tow taught us that the task is very difficult. Aside from the many that obviously were unknown, we never could be quite certain that we reeogni/ed all the species included among the supposedly known ones. To gain an idea of the great similarity of the eggs of some of the common marine species the reader is referred to earlier papers (1930 and 1934) by the writers. 528 BULLETIN OF THE BUREAU OF FISHERIES Fish 6 to 8 mm long probably are not over 2 weeks old, but individuals 19 to 25 mm long, judging from the rate of growth of other species for which more data are at hand, may be 4 to 6 weeks old. It may be concluded, then, from the dates when young fish were caught, given in the preceding paragraph, supported by the single female with developing roe caught on June 16 (1926), that the sheepshead spawns from sometime in April to perhaps the latter part of June in the vicinity of Beaufort. DESCRIPTIONS OF THE YOUNG Specimens under 6 mm in length have not been taken. Therefore, the larvae, as already stated, remain unknown. The small size at which young sheepshead acquire the characters of the adults is remarkable. Specimen 6.0 mm long. — A single specimen with a damaged caudal fin of about 6.0 mm (5.2 mm to base of caudal) in length is at hand. This .fish already is well past the larval stage, as will be brought out in the following description. Body elongate, compressed, its depth 3.4 times in length to base of caudal; head rather short, compressed, 3.0 in length to base of caudal; snout short, blunt, with rounded profile, 4.2 in head; eye wholly lateral, rather small for such a young fish, 3.1 in head; mouth small, oblique, almost terminal; maxillary reaching about to pupil; preopercular spines present, but very short; vent a little behind midbody length; notochord bent upward posteriorly as usual in young teleosts having homocercal tails; myomeres about 27 (vertebrae in adults 9 + 15 = 24). The fins are remarkably well developed for such a small fish. Dorsal spines very short, about 7 discernible at this size (adults with 11 or 12); soft rays 12 (11 to 13 in adults); caudal fin with well devel- oped rays (broken); anal fin with 13 rays, the spines not well differentiated (adults with 3 spines and 10 or 11 soft rays); ventral fins not yet discernible; pectoral fins broad, damaged, apparently rather short. The general color of this preserved specimen is brownish, without very definite markings. Median ventral line with three obscure dark spots, one slightly behind isthmus, another below vertical from base of pectorals and the third one a very short distance in front of vent. A slight dark coloration is evident on side just posterior to vent, and two dark specks are present on the base of the anal fin (fig. 18). Specimen 7.5 mm long. — A single specimen with a damaged caudal fin of about 7.5 mm (6.25 mm to base of caudal) in length is at hand. A fairly complete description of this specimen when fresh, was prepared. It was then about 8 mm long, having shrunken somewhat during preservation. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 529 The differences between this specimen and the smaller one already described are not pronounced, except in color. The body has become somewhat more slender, the depth now being contained 3.9 times in the length to base of caudal, and the head remains short, compressed, 3.7 times in length to base of caudal. The snout also remains short, blunt, with a steep profile, only a little shorter than the eye, and is contained 3.0 times in the head. The mouth is slightly less oblique than in the smaller specimen. The dorsal spines remain short and feeble, though 13, the full number (11 to 13) present in the adult, may be counted. The three anal spines now are differentiated. However, the spines are more retarded in development than the soft rays. The caudal fin (broken now) was described as “slightly concave” in the field notes. The ventral fins are quite evident, though minute, being scarcely longer than pupil, and are inserted very slightly behind the base of the rather broad well-developed pectorals. The color of a fresh specimen is described in the field notes in part as, “pale, without cross bars, though small dark chromatophores are present along the side of body, but not yet forming cross bars.” However, the preserved specimen does show an arrangement of chromatophores which suggest a bar between the anterior part of the soft dorsal and anal, and another just posterior to these fins (see fig. 19), precisely where bars are present in larger fish. The upper margin of the eye in the fresh specimen was dark, and black chromatophores were present on the interorbital and also along the chest, abdomen, and base of anal. The three obscure dark spots along the median vental line in the smaller fish described persist in the larger one, the posterior one situated somewhat in advance of the vent having become rather more prominent (fig. 19). Specimens 10 to 12 mm long. — Several specimens ranging from about 10 to 12 mm in length were collected. These fish already resemble the adults so much that they are not difficult to identify. The body has become deeper, and it is much more robust, the ‘depth being con- tained about 3.0 to 3.2 times in the length to base of caudal. The head has become notably broader and is now contained 3.0 to 3.25 in the length. The snout remains short, rounded, with a steep profile, and much shorter than the eye, its length 4.3 to 4.7 in head; eye 3.0 in head. The mouth remains small, slightly oblique, and nearly terminal, the maxillary reaching about to pupil. The lateral line has made its appear- ance, being represented by a few pores anteriorly. The body is almost fully covered with scales at a length of 12 mm, though not so indicated in figure 20. The dorsal and anal spines are much better developed, but still proportionately much shorter than in adults. The caudal fin is distinctly concave, and the ventral fins have increased greatly in length, being longer than eye and reaching nearly to vent. 530 BULLETIN OF THE BUREAU OF FISHERIES Pigmentation has progressed rapidly, though it is not complete. Individual chromatophores are present everywhere and are concentrated in definite areas to form bars which are not developed equally early in all specimens. Usually, however, they are more or less definite in specimens 10 mm long, quite distinct in 12-mm fish, and generally 7 in number, as in the adult (fig. 20). Specimens 15 to 18 mm long.-—' The body has become somewhat deeper, the depth now being contained about 2.9 in the length to base of caudal. The head remains short and deep, about 3.0 in length. The snout is a little less blunt and slightly longer, about 4.1 in head; eye longer than snout, about 3.0 in head. The mouth remains small, slightly oblique, the gape anteriorly being nearly on the level with the lower margin of the eye; maxillary reaching slightly past anterior margin of eye. Preopercular spines, present in smaller fish, no longer are evident. The lateral line generally is rather fully developed in specimens 18 mm long, and the body is covered with scales. The dorsal spines have increased in proportionate length, but remain notably shorter than the soft rays. The anal spines are well developed, the second one already being the strongest as in the adult. The outer unbranched ray of the ventral is not yet fully developed as a spine; the second ray with a free filament distally, frequently reaching origin of anal. Specimens 18 mm long, in the fresh state, have the general color of the adult, including the characteristic black cross bars (fig. 21). DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 531 Specimens 25 to SO mm long. — The body has increased still further in depth, which is now contained about 2.3 to 2.5 in the length to the base of the caudal. The head has become longer, 2.8 to 2.9 in length. The snout is longer aud more pointed, 3.4 to 3.7 in head; eye longer than snout, 2.6 to 2.9 in head. The mouth is nearly hori- zontal, and wdiolly below the lower margin of the eye; maxillary scarcely reaching past anterior margin of eye. The lateral line is fully developed. The dorsal spines are strong and proportionately about as high as in adults; those of the anal also as in adults, the second one notably longer and stronger than the others. The caudal fin is slightly forked as in large individuals. The filament on next to the outermost ray of the ventral persists, reaching about to the origin of the anal, the outermost ray now being developed quite definitely as a spine. The pectoral fin has increased some- what in length, its distal margin, instead of being round as in the smaller fish is now oblique, the longest rays being in the upper half of the fin, reaching nearly to origin of anal. Fioube 22. — Archosargus probaiocephalws. From a specimen 30 mm long. The general color of the adults develops early, as already shown. Scarcely any changes of note have developed since a length of 18 mm or so was attained. Some specimens, though not all, have the ventrals quite black (fig. 22). Specimens 50 to 60 mm long. — The fish have continued to increase in depth, and are now proportionately as deep as large specimens, depth 2.0 to 2.1 in length to base of caudal. The snout has become proportionately longer and more pointed, though still notably shorter than in large individuals, equal to or a little longer than the eye, 2.8 to 3.1 in head; eye 2.8 to 3.3. The incisor teeth now are prominent, and the posterior molarial teeth are strong. The pectoral fin has acquired the shape it will retain, the fourth ray being longest with the rays below becoming shorter gradually, making the lower posterior margin straight and oblique. The ventral filament has become proportionately shorter, reaching vent in some specimens and to origin of anal in others. 532 BULLETIN OF THE BUREAU OF FISHERIES The color, though somewhat variable, does not differ from that of somewhat smaller and larger fish. Some specimens have two dark spots on the base of the caudal fin which are missing in others. Many specimens have the ventral fins mostly black, and most individuals have a definite dark shoulder spot near the beginning of the lateral line, partly in and partly in advance of the second cross bar. Specimens 75 mm and upward in length. — Although the proportionate depth of large specimens is attained when the fish reach a length of 50 to 60 mm, a pronounced change in the shape of the head and snout takes place as the fish continues to grow. The upper profile becomes notably more gently elevated, and the snout proportion- ately much longer and more pointed. In specimens about 75 mm long the snout (measured from anterior margin of eye to tip of upper jaw) is contained about 2.6 times in the head; in specimens about 100 mm in length, 2.3; and in specimens about 225 mm long, 2.1 times. The eye, as usual, becomes proportionately smaller as the fish grows, but the difference in the present species is unusually great. In fish about 75 mm long it is contained about 3.2 in head; in specimens 100 mm long, 3.5; and in specimens 225 mm long, about 4.5 times. The ventral filament continues to become shorter until it is scarcely longer than the longest rays in specimens around 100 mm in length, and soon disappears entirely. DISTRIBUTION OF THE YOUNG It has been pointed out that the early young (larvae) were not taken. Therefore, we cannot state where they live. However, Rathbun (1892) stated that spawning apparently takes place along sandy shores. The early young presumably are pelagic as they are hatched from floating eggs, and would be expected to occur not far from where the eggs are spawned. In an extensive search made in the vicinity of Beaufort for the eggs and larvae none were found. All the young at hand, except the smallest one, were seined from eelgrass and other growths in shallow water. The smallest specimen, which is about 6.0 mm long (caudal fin damaged), was taken in Shackleford Channel along a sandy beach. However, this specimen already had fins and no doubt was capable of self-directive swimming. Therefore, it may not have been very near the place where it was hatched, though it still seemed to be pelagic. It is evident, then, that no new information as to the abode of the larvae can be added at this time. When the young attain a length of about 7 to 8 mm they settle down in shallow water where an abundant growth of seaweeds is present. In 1926 and 1927, before the eelgrass began to disappear, young sheepshead, ranging in length upward of 8.0 mm, were common to numerous along the south shores of Pivers Island, Beaufort, N. C., where most of the specimens upon which the present study is based were taken by “cutting”, as far as possible, a bobbinet seine through dense growths of eelgrass. There the young remained until they reached a length of about 40 mm. Thereafter they seemed to become less abundant, though some stayed until they attained a length of 60 mm or so, when they left to occupy the habitat of the adults, which already has been described. When the young sheepshead has attained a length of about 40 to 50 mm the teeth are developed essentially as in the adult and thereafter they may be observed along stone jetties, breakwaters, around piers, and wrecks where larger fish also live. FOOD The chief food of young sheepshead, ranging from 9.0 to 55 mm in length, while dwelling in shallow water in weeded areas, according to the contents of 111 stomachs, is copepods. Those under 30 mm in length utilized ostracods, which were rarely eaten DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 533 by larger fish. Gammarus were sparingly eaten by the smaller fish, but abundantly by the larger ones. Small mollusks appeared early and continued to be eaten also by the larger fish. A few worms were eaten, and also some small decapod crustaceans, especially shrimp. In addition to the animal foods named filamentous alga was present in such abundance in the stomachs of fish ranging from 25 to about 40 mm in length that it quite certainly was not taken by accident. Some stomachs, in fact, contained almost nothing else. It seems definitely to constitute a part of the food of fish of the size range mentioned. Larger fish ate of it more sparingly if at all. The food of adults consists chiefly of mollusks and crustaceans which the fish find numerous along the breakwaters, piers, wrecks, etc-., where the adult fish live. A favorite bait at Beaufort is the fiddlar crab. The teeth of the sheepshead, described elsewhere, are well adapted for seizing and crushing these common foods. Figure 23.— Growth curve based on length measurements of 612 Archosargus probatocephalus of the 0-class. Solid line, average length for each month of all fish measured; dot and dash (upper) line, largest fish; dot and several dashes (lower) line, small- est fish. GROWTH Limited information relative to the rate of growth of the young during the first few months of life was obtained, and none for the older ones (fig. 23). After the fish leave their early habitat among seaweeds they no longer are obtainable in sufficient numbers, without much effort, to follow the rate of growth. It seems quite certain that as early as August some of the larger young of the season already had deserted their habitat among seaweeds. Therefore, the range at the upper limit and consequently the average length, of those taken in the habitat of the juveniles no longer give correct information as to the rate of growth. In September many young definitely had moved away from their earlier habitat, as the fish had become comparatively scarce, though a few remained there nearly all winter. The range in length of 46 young taken among seaweed in June from 1926 to 1931, is 7 to 25 mm. However, only one specimen, apparently a very fast growing one, exceeds a length of 21 mm. The average length of the 46 specimens is 12.8 mm. 534 BULLETIN OF THE BUREAU OF FISHERIES The range in length of 311 young taken during July under the same general conditions and over the same number of seasons is 11 to 42 mm, the average length being 21.8 mm. In August the range in length of 79 specimens, from the same localities and the same years, is 27 to 44 mm having an average length of 36.6 min. In September only 20 specimens, ranging in length from 38 to 47 mm, with an average length of 42.1 mm, were secured. It seems probable that the measurements for July alone show fairly accurately the range in length, as well as the average size, of the young fish for that month. In June the smallest young of the season had not arrived in the weeded areas. There- fore, the lower limit of the range, and consequently the average length, are not correct. By August some of the larger young had left the weeded areas, and therefore the upper end of the range, as well as the average, is incorrect. In September the fish had gotten so scarce that with the same fishing effort put forth each month in 1926 only 20 specimens were secured, whereas in August 69 were taken, in July 299, and in June 36. The data seem to justify the conclusion, however, that in June the largest young of the current season are around 20 to 25 mm long, the lower limit of the range and the average length being unknown. In July young range in length from about 11 to 42 mm, and the average length is close to 21.8 mm. In August the smallest young are about 27 mm long, and in September they are around 38 mm in length, the upper limit of the range and the average length being unknown for both months. A fairly slow rate of growth seems to be indicated if the evidence produced elsewhere, showing that spawning at Beaufort begins sometime in April and ends near the end of June, is correct. A slow rate of growth and late maturity would explain, in part at least, why the sheepshead has diminished rapidly under heavy fishing, whereas other, presumably faster-growing, species have withstood it without a serious decline. CHAETODIPTERUS FABER (BROUSSONET). SPADEFISH The development of the eggs and recently hatched larvae of the spadefish was described and figured by Ryder (1887, pp. 521-523). It is possible now to describe and figure some more advanced stages, though the series is not yet as complete as desirable. Professor Ryder did not state specifically that the eggs used in his study were taken directly from ripe fish caught in Chesapeake Bay, though this apparently may be assumed. He merely said, “This species spawns in the Chesapeake during the latter part of June and the early part of July. It is prodigiously fertile, the female probably discharging a million ova during a single season.” It must be further stated, however, that the number of eggs deposited probably depends on the size of the female, because large fish generally, if not always, produce more eggs than smaller ones of the same species. The eggs were not seen by us. Ryder states that they are pelagic, “somewhat over a millimeter in diameter”, and have a single oil globule. Cleavage took place rapidly, as only an hour intervened between the first cleavage and the morula stage. Hatching took place in about 24 hours in a water temperature of 80° F. The newly hatched fish were “about 2.5 mm in length.” In 63 hours the yolk was nearly all absorbed, young fish had increased greatly in depth, and were nearly 4.0 mm long. The snout was very blunt, the mouth (according to the figure) was DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 535 large and somewhat oblique, with the lower jaw projecting. An aggregation of pigment cells formed indications of a band above the base of the pectoral, and another at about midcaudal length. The specimens forming the basis of the present study were collected mostly at Beaufort, N. C. Other young were taken on the coast of Georgia and on the Gulf coast of the United States. Unfortunately, stages in development connecting directly with the largest larvae described by Ryder are not at hand. The smallest specimen in the collections, identified as a spadefish, is only about 2.5 mm long in the pre- served state. Although this larva is shorter than the oldest one described by Ryder, which was about 4.0 mm long when alive or fresh, it is much further advanced in development, showing apparently that much shrinkage took place during the harden- ing process. The characters connecting this larva with the younger ones described by Ryder are pointed out in the description of the specimen. The spadefish is known in some localities as angel fish and also as moonfish. At Beaufort it is called porgee (or pogee), a name also heard in the lower Chesapeake. This species ranges from Cape Cod, Mass., at least as far south as Rio de Janeiro, Brazil. It is not common north of Chesapeake Bay. On the Atlantic coast of Panama it is one of the common food fishes, and is seen in the market almost daily. It is not caught in large quantities on the coast of the United States. Indeed, the statistical report of the Bureau of Fisheries for 1935 lists a catch of only 6,000 pounds, which was made in North Carolina. However, the fish is taken commercially all along the Atlantic and Gulf coasts from Chesapeake Bay southward. It is a fish of good flavor and always in demand. Consequently, much of the catch is consumed locally and it often does not enter the markets. Therefore, it fails to get into the records, which do not show its full importance as a food fish. The spadefish is a summer resident at Beaufort, where it arrives in May and departs by about the beginning of October. At Key West, Fla., it is present the entire year, though most common during the summer. The species tends to congregate in small schools. It is caught chiefly with seines. However, it will take a hook baited with small bits of meat. Because of its small mouth, small hooks must be used. Furthermore, because of its tendency to nibble instead of swallowing the bait, consider- able patience and skill must be exercised by the angler. If he is successful in hooking one, a good fight follows. CHARACTERS OF THE ADULT The spadefish belongs to the family Ephippidae, of which it is the only repre- sentative on the Atlantic coast of America. It is readily recognized by the very short, deep, compressed body which is only a little longer than deep. The teeth in the jaws are in brushlike bands, the outer series being slightly enlarged. In large individuals the anterior rays of the second dorsal and anal are considerably produced, and the caudal fin is deeply lunate. Fish under a foot or so in length bear four to six broad black cross bars, which tend to fade in large individuals. The ground color varies from brown to silvery green. The maximum size of the spadefish is given as 20 pounds. However, fish weigh- ing as much as 12 pounds are comparatively rare. The average weight of the fish seen in the Beaufort market probably did not exceed 1% pounds, and those in the Colon, Panama, market were even smaller. 154979—38 3 536 BULLETIN OF THE BUREAU OF FISHERIES SPAWNING According to Ryder (1887, p. 521) the spadefish spawns in Chesapeake Bay during the latter part of June and the early part of July. Hildebrand and Schroeder (1928, p. 307) stated, “Fish with well-developed roe were taken at Crisfield, Md., on May 26, 1916,” and Smith (1907, p. 335) said “At Beaufort, ripe male and female fish have been found early in June.” We can add to these observations only that we saw some females with developing roe at Beaufort on May 25, 1916. The eggs either were not taken, or not recognized if taken, during the investigation at Beaufort. The scarcity of the young locally indicates that the vicinity of Beaufort is not an important spawning area. The smallest larva caught, which is about 2.5 mm long, was taken July 11 (1929). The next smallest one, which is about 4.25 mm long, was caught July 12 (1915); another small one, 9 mm long, was taken July 9 (1930); and still another one, 17 mm long, August 16 (1916). Larger young were taken locally in 1930 as follows: 10, ranging in length from 49 to 62 mm, August 23; 21, varying in length from 57 to 86 mm, from September 4 to 16; and 1 each on October 18 and 21, respectively, 72 and 74 mm long. These young no doubt are all in their first summer. Their size, espe- cially that of the smallest ones, suggests that at Beaufort spawning takes place at least during June, as indicated also by the few observations of ripe and ripening fish reported in a preceding paragraph. A definite determination of the duration of spawning, however, remains for future determination. The 3 smallest young were all taken at sea, suggesting that the fish may spawn offshore. Offshore spawning is indicated, furthermore, by the absence of small fish under about 15 mm in length, in the extensive and thorough collecting done in the inside waters in the vicinity of Beaufort. DESCRIPTIONS OF THE YOUNG Specimen about 2.5 mm long. — A single specimen of this size is at hand. The body is deep anteriorly, decreasing greatly in depth just posterior to the vent, the greatest depth being contained 1.9 times in the length to the end of the notochord. The head is very deep, with a steep profile, which is slightly concave just above upper jaw. The snout is short and blunt, and not quite as long as the eye. The mouth is strongly oblique, the gape anteriorly being on a level with the middle of the eye, and the maxillary reaches opposite the posterior margin of the pupil. The preopercular margin is provided with a few prominent spines, and a sharp transparent dermal crest is present on the occiput. The notochord remains straight. The primitive vertical fin membranes persist and contain only slight indications of rays where the soft dorsal, caudal, and anal develop later in life. The ventral fins are not evident, but the pectorals are short and broad. The general color of the preserved specimen is pale gray. Several dark chro- ma tophores are present on the chest and the abdomen, and also a few on the gill covers (fig. 24). This specimen resembles the largest larvae described by Ryder (1S87, p. 522) in having a deep body, which seems to have grown much deeper in the older fish herein described. It, also, resembles the younger fish in the steep anterior profile and oblique mouth. The younger larvae had dark chromatophores on the abdomen, which have been retained by the older one at hand. However, no indications of dark chromatophores, suggesting bands (one above the base of the pectoral and another at midcaudal length), described and illustrated for the younger larvae by Ryder, are evident in the fish before us. Ryder does not mention nor illustrate a dermal ridge DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 537 or crest at the occiput, present in the specimen herein described, and retained by some- what larger fish. It may be assumed, therefore, that the ridge is not present in the early larvae. This ridge is suggestive of the crest bearing a spine in the young fool fishes (Monacanthus). Specimen about 4.25 mm long. — Only one specimen of this length is at hand. The body remains deep anteriorly, but no longer decreases as abruptly in depth just posterior to the vent as in the smaller specimen already described. However, the body tapers sharply posteriorly. Its greatest depth is contained 1.7 times in the length to the end of the notochord. The head remains deep, with a somewhat more sloping anterior profile. Its length is contained 2.9 times in the length to the end of the notochord. The snout remains shorter than the eye, and the mouth is quite oblique, the gape anteriorly being about on a level with the lower margin of the pupil. The maxillary reaches below the middle of the eye. The crest or ridge on the head remains prominent, and slightly spinelike. The spines on the preopercular margin persist, but are less prominent. The notochord now is bent upward posteriorly, and Figube 2i.—-Chaetodiptems faber. From a specimen 2.5 mm long. below it are rather well developed rays, forming a moderately long caudal fin, the shape of which cannot be determined definitely because of the damaged condition of the fin, but it presumably had a round margin as in somewhat larger fish. The spines in the dorsal and anal fins are more retarded in development than the soft rays and cannot, be enumerated definitely. About 23 soft rays may be counted in the dorsal and 20 in the anal. The pectoral fins are broken, and the ventral fins appear as mere tufts of dermis. The preserved specimen at hand is very dark, apparently having become darkened by the action of a chemical in the denatured alcohol used. However, black chroma- tophores are visible along the side of the abdomen, on the head and back, and at midcaudal length. The last-mentioned ones are concentrated and slightly suggestive of a cross bar, shown by Ryder (1887, p. 522) for much younger fish, (fig. 25). Specimen ,9.0 mm long. — The body remains short and deep and has become some- what more robust, the greatest depth being contained 1.8 tunes in the length without the caudal fin. The dorsal profile remains quite steep anteriorly, and rather more 538 BULLETIN OF THE BUREAU OF FISHERIES strongly curved than the ventral outline. The head has become broader. It is scarcely as deep as in smaller fish, and its length is contained 2.5 times in the length without the caudal fin. The snout remains blunt and shorter than the eye, being contained 4.2 times in the head, whereas the eye is contained 3.1 times. The mouth has become less strongly oblique, the gape being wholly below the eye, and the maxil- lary scarcely reaches beyond the anterior margin of the eye. Small pointed teeth now are evident in the jaws. Preopercular spines remain quite prominent. The dermal crest, or ridge, at the occiput has become proportionately shorter, and ends in a blunt spinelike point. The body now is covered with blunt spinelike plates, scarcely resembling scales (not shown in fig. 26), and the upper surface of the head (where the plates are missing) bears short hairlike spines. The fins are all developed and have the usual number of spines and soft rays present in adults. However, the spines still are somewhat retarded in development and proportionately shorter than in the adult. The margins of the vertical fins are all rounded. The ventral fins are quite large, and exceed the length of the pectorals. The general color of the preserved specimen is dark brown. Dark chromato- phores present in younger fish and in large specimens at hand are not visible in the 9.0 mm specimen. A few dusky markings are present at the shoulder. The soft dorsal, caudal and anal are colorless, but the ventrals are dark brown (fig. 26). Specimen 11 mm long. — The body is proportionately a little deeper than in the 9.0-mm fish, the depth now being contained 1.5 times in the length without the caudal fin. The general shape of the head, and the proportions of the eye and snout have changed little. The preopercular spines have become rather shorter and blunter, and the occipital ridge of smaller fish now is represented by a small blunt projection. The lateral line is well developed. The scales no longer look like plates, and the spiny projections on them are smaller. The hairlike spines on the head, noted in the 9.0 mm specimen, have become minute. The dorsal spines have increased in length, but remain proportionately shorter than in larger fish. The ventral fins have con- tinued to increase in length, and now reach the vent. The general color is brownish, and the head and body nearly everywhere are dotted with black chromatopkores. An indefinite pale crossbar on the back of the head extends down on the preopercle. The ventral fins are black, and spinous parts DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 539 of the dorsal and anal are brownish and dotted with black, but the rest of these fins and the caudal and pectorals are entirely colorless (fig. 27). Specimens 15 to 18 mm long. — The differences between these larger specimens and the 11 mm one are not great. The depth remains proportionately about the same, being contained 1.4 to 1.6 times in the length without the caudal fin. The snout, Figure 26. — Chaetodipterus faber. From a specimen 9 mm long. Figure 27. — Chaetodipterus faber. From a specimen 11.6 mra long. which is shorter than the eye in younger fish, now is as long as the eye. The mouth is nearly horizontal and terminal, and the maxillary reaches scarcely to the eye. The preopercular spines have become quite small, and the occipital ridge is missing. Scales are well developed; they extend forward on the head to interorbital, and are 540 BULLETIN OF THE BUREAU OF FISHERIES strongly ctenoid. Very short hairlike spines remain visible on the head. The lobes of the soft dorsal and anal are high, and slightly pointed. The caudal fin, also, is somewhat pointed. The pectoral fins remain moderately short and rounded, whereas the ventral fins have increased still further in length, reaching fully to the origin of the anal. The general color is brownish, some specimens being much darker brown than others. Dark dots are visible on the lighter-colored specimens, much as in the 1 1-mm specimen, but none are discernible in the darker ones which have scales with blackish margins. A pale bar, extending across the nape and down on the preopercle, is very distinct, and the snout is about equally pale. The spinous part of the dorsal, the basal portion of the soft dorsal, as well as the basal parts of the anal and caudal are dark brown, and become abruptly entirely colorless. The pectoral fins remain wholly colorless, and the ventrals are black (fig. 28). Figure 2 S.—Cliaetodiplerus faber. From a specimen 17 mm long. (Drawn by Mrs. E. B. Docker.) Specimens about 20 mm long. — The proportions of the head and body have not changed measurably since a length of 18 mm was attained. The only change, worthy of note, is that of color. The pale bar at the nape, mentioned in the preceding descrip- tion, persists in some specimens, but is missing in others, and some specimens are blotched along the sides with the same pale color. Three dark cross bars now are present. The first and most distinct one crosses the interorbital, and extends through the eye to the chest. The second one crosses the nape, and extends down on and behind the margin of the opercle, through the base of the pectoral to the abdomen, just behind the ventral fin. The third one extends from the base of the spinous dorsal to the base of the anal spines. Numerous dark chromatophores still are visible on the head and body of the lighter-colored specimens. The fins remain unchanged in color, except that the brown color extends higher on the dorsal and anal, involving fully the basal half of the soft rays. Specimens 25 to SO mm long.— The depth in proportion to the length of the body has increased still further since a length of about 15 to 18 mm was attained, being contained 1.25 times in the length without the caudal fin. The general shape of the body now is very similar to that of the adult. The head is proportionately shorter, as DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 541 usual in larger fish, and is contained 2.75 in the length to the base of the caudal fin. The eye and snout are of equal length, being contained 3.2 to 3.5 in the head. The mouth is horizontal and terminal. The brushlike teeth are developed as in the adult. Preopercular spines remain present, but have become very small. Scales are fully developed, strongly ctenoid, and the hairlike spines present on the head in smaller fish have disappeared. The lobes of the dorsal, anal, and the caudal fins remain round. The ventral fins are long, the second ray (first soft ray) being produced and reaching about to the base of first soft ray of the anal. The general color of preserved specimens varies from light to dark brown, Some specimens retain a trace of the pale bar crossing occiput, and the pale blotches of smaller specimens. A fourth dark cross bar, extending from about the middle of the base of soft dorsal to the base of the anal, now is more or less distinct. The dark brownish color of the dorsal and anal extends farther on the fins, leaving only the margins translucent. The caudal and pectorals remain translucent, and the ventral fins are mostly black. Specimens 40 to 50 mm long.— The body has continued to increase in depth, and has attained about the proportions of full-grown fish, the greatest depth being con- tained 1.1 to 1.2 times in the length without the caudal fin. The snout is definitely longer than the eye, being contained 2.75 to 3.0 times in the head, whereas the eye is contained 3.3 to 3.7 times. Preopercular spines virtually have disappeared, only a few slight points remaining at and below the angle. In somewhat larger fish they no longer are evident. The dorsal spines are about as high as in full-grown fish, the third and largest one reaching the base of the last one if deflexed. Posterior to the longest spine is a black membrane, which reaches beyond the tip of the spine. The lobes of the soft dorsal and anal have become broadly rounded, and the caudal fin is broadly rounded to nearly square. The general color is not much different from that of fish 25 to 30 mm long. A fifth black bar, situated on the caudal peduncle, however, is present. The color of the fins remains unchanged, except that the lobes of the soft dorsal and anal are wholly dark brown and the interradial membranes of the spinous dorsal are partly black (fig. 29). Specimens 75 mm and upward in length.- — Specimens 40 to 45 mm long already have acquired essentially the shape and proportions of the body of much larger fish. However, a notable change in the shape of the soft dorsal, anal and caudal takes place in larger fish. At a length of 75 mm the anterior soft rays of the anal in some specimens already are produced, as in large fish, making the margin straight or even slightly concave. The dorsal fin apparently is a little more retarded in this same forthcoming development. The caudal fin margin is slightly rounded to nearly straight in fish around 75 mm long. It has been indicated already that development of the fins does not proceed equally in all spadefish. Thus, a specimen 90 mm long has the anterior rays of the soft dorsal and anal produced so as to form pointed lobes, and the outer rays of the caudal are sufficiently produced to make the margin of this fin slightly concave. Another specimen, 105 mm in length, by contrast, still has these fins shaped essen- tially as in specimens 40 to 45 mm long. The anterior rays of the dorsal and anal, as well as the outer rays of the caudal continue to increase in length, though unequally fast, as the fish grow, and become long and pointed in large individuals, often reaching beyond the midlength of the caudal fin. The membrane behind the third dorsal spine, already present in fish 40 to 45 mm long, which is at least somewhat longer than the spine, persists. This spine also develops, apparently reaching its maximum 542 BULLETIN OF THE BUREAU OF FISHERIES length in fish about 135 mm (5.5 inches) long, when it reaches to or sometimes beyond the base of the first soft ray of the dorsal, and thereafter it again becomes propor- tionately shorter. In a specimen 262 mm (10.5 inches) long it reaches only to the middle of the sixth spine if deflexed. The filament on the first soft ray of the ventral also persists in big fish, wherein it reaches to the origin of the anal. Figure 29. — Chaetodipterus faber. From a specimen 50 rnni long. The preserved specimens vary greatly in color, ranging from pale silvery to dark brown. Although the species is described as having four to six black bars, all the mature individuals at hand have five, except one old one (262 mm long) which has none. The caudal and pectorals are translucent except in fish upward of 110 mm in length wherein they are brownish. The ventrals remain black or at least dark throughout life. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 543 DISTRIBUTION OF THE YOUNG Too few early young were taken to cast much light on their distribution. The three smallest specimens at hand, having a length of 2.5, 4.25, and 9.0 mm were all taken at sea, the smallest one about 6 miles off Beaufort Inlet, the intermediate one somewhere off Southport (N. C.), and the largest a short distance west of Beaufort Inlet. It is not known whether the intermediate specimen was taken at the surface or bottom, but the other two were caught in bottom towings. The larger specimens were taken in inside waters in the immediate vicinity of Beaufort either with seines or with otter trawls. Other specimens, ranging upward of 11 mm in length, from the Gulf coast, principally from Louisiana and Texas, which have been studied, presumably were nearly all taken in shallow water with seines. GROWTH The literature contains little information on the rate of growth of the spade- fish, and insufficient specimens were measured during the recent investigation at Beaufort for definite determination. Smith (1907, p. 335) stated, “In the latter part of August fish about 3 inches (75 mm) long may be seined in Beaufort Harbor.” Hildebrand and Schroeder (1928, p. 307) reported the capture of small spadefish in Chesapeake Bay, which they believed were in their first summer, as follows: 1 fish, 55 mm long in September; and 35 fish, ranging in length from 65 to 100 mm in October. Young spadefish evidently in their first summer were taken on the coast of North Carolina during the recent investigation as follows: 3 fish, respectively 2.5, 4.5, and 9.0 mm long, in July; 12 fish, ranging in length from 49 to 80 mm (average length 56.6 mm) in August; 21 fish, ranging in length from 57 to 86 mm (average length 72.1 mm) in September; and 2 fish, 72 and 74 mm long in October. The data presented suggest that the young fish may reach a length of about 55 to 100 mm during their first summer on the coast of North Carolina and in Chesa- peake Bay. As a fully mature, ripening female 135 mm (5.4 inches) long was seen, it seems probable that at least some of the fish reach that length during their second summer and that they may spawn when 2 years old. FAMILY GOBIIDAE. GOBIES Nine species of gobies with united ventral fins are recorded from the coast of North Carolina in this paper. These gobies are assigned to four genera, namely, Gobiosoma (two species), Microgobius (two species), Gobionellus (four species), and Gobius (one species). Three species of the genus Gobionellus appear to be new to the fauna of North Carolina, as stated elsewhere (p. 564). The single species of Gobius, namely, glaucofraenum, is known from North Carolina (Cape Lookout) from one specimen (Gudger, 1913, p. 165), which has not been seen by us. This species will receive no further mention in this paper. It appears to differ from all the other local species in having larger scales, about 23 transverse series on the side, and in the shorter second dorsal and anal fins, each fin having 10 rays. In the present study we did not succeed in collecting eggs of gobies in their native environment. However, those of Gobiosoma bosci and Gobionellus boleosoma were secured through artificial means. The eggs of the other species dealt with in this paper remain unknown. Small larvae, usually under about 8 to 10 mm in length, were collected principally with 1-meter tow nets. The larger young, that is, fish from 8 to 10 mm and upward in length, were caught principally with especially adapted otter trawls, although some were taken with beam trawls and with bobbinet seines. 544 BULLETIN OF THE BUREAU OF FISHERIES The comparative abundance at Beaufort of the young gobies discussed in this paper indicates that the adults are more numerous than the number taken in net collections suggests. It is probable that the adults adhere to the bottom or to objects on the bottom by means of the ventral disk, permitting nets to pass over them. Gobio- soma was able to escape nets quite successfully when confined in a large tank, as explained elsewhere. (See. p. 549). This propensity of escaping nets probably is exercised in nature by most species of gobies. The original drawings of young and adult fish published in this paper are based upon preserved specimens. The illustrations showing the development of eggs are after Kuntz (1916) and were drawn from living material. The main differences among the young of the three genera, with which this paper deals principally, are shown in a parallel comparison of characters appearing herewith. The adults of the local species differ from each other rather markedly. In Gobiosoma the body is naked, or at most only two scales are present on the base of the caudal. In Microgobius and Gobionellus, on the other hand, the body is nearly or quite fully covered with scales. Microgobius is distinguished from Gobionellus in having a deeper and more compressed body. Furthermore, the mouth is large and strongly oblique, the maxillary reaching nearly opposite the middle of eye in Microgobius, while in Gobionellus the mouth is scarcely oblique and the maxillary reaches only below the anterior margin of the eye. The dorsal spines are 7 in number and are about equally spaced in Microgobius, whereas Gobionellus has 6 dorsal spines, with the last 2 notably farther apart than the others; and Microgobius has a larger second dorsal and anal fin, each fin being composed of 16 or 17 rays, while Gobionellus has only 11 to about 15 rays in each fin. The eggs of the gobies of North Carolina, as already indicated, have not been found in nature. Those of Gobiosoma bosci and Gobionellus boleosoma were secured by Kuntz (1916) by stripping the ripe fish. The eggs of the first-mentioned species also were secured recently by us. The eggs of all the other species remain unknown. Kuntz (loc. cit.) found that the eggs of Gobiosoma bosci and Ctenogobius stigmaticus {—Gobionellus boleosoma ) each had a bundle of adhesive threads attached to the egg membrane. These threads no doubt serve the purpose in nature of attaching the eggs to submerged objects. Ehrenbaum (1905), dealing with European species, stated that the eggs of Gobius niger are attached to plants, shells, ascidians, and stones ; G. Jlavescens are attached to plants; and those of G. minutus to molluskan shells. Hefford (1910) found the eggs, with an adult fish (sex not stated), of Gobius paganellus “on a stone between tide marks on the shore” at Plymouth, England. It is not stated that the adult fish guarded the eggs. Clark (1913) stated that the eggs, with males, of Crystallo- gobius nilssoni were found in abundance in the waters at Plymouth, England, attached to the inside of empty tubes of Chaetopterus. Petersen (1917) figured the eggs of four species occurring on the coast of Denmark, namely, Gobius niger, G. ruthensparri, G. minutus, and G. microps, showing that the eggs of each species possess an adhesive foot. Lebour (1919) said of Gobius paganellus: “From early spring to late summer the males may be seen guarding their eggs, which are attached to the under surface of stones in masses.” Lebour (1919) isolated several adults of Gobius ruthensparri in a tank and one deposited eggs on the inside of an empty oyster shell. It was not stated that the eggs were guarded by a male. Lebour (1920), who illustrated the eggs of Gobius minutus, G. microps, and G. pictus with adhesive organs, stated that those of the first-mentioned species were laid on an oyster shell, and those of the second one DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 545 on the valve of a Pectan. The eggs of the last-named species apparently were not found in nature, but were deposited in the aquarium on a molluskan shell. It would seem quite certain from the knowledge gained from the study of some American and European gobies, as shown in the preceding paragraphs, that these little fish generally, if indeed not always, attach their eggs to submerged objects, where they probably are guarded by the males. In these respects the spawning habits of the gobies agree in a large measure with the blennies, as shown in another part of this paper. Gobiosoma bosci and Gobionellus boleosoma are landlocked, except during extremely high storm tides, in the Mullet Pond on Shackleford Banks near Beaufort. The water in this pond ranges from brackish to nearly fresh. In fact, it becomes fresh enough, at times, to support such fresh water plants as Potamogeton and filamentous fresh water algae. Here the two species named evidently carry out their reproduc- tive processes, as ripe adults and young fish, less than 10 mm in length, have been collected. The bottom of the pond is quite muddy. In places luxurious growths of plants are present, and oyster shells, as well as live oysters, are found over some parts of the bottom. It would seem necessarv for the gobies to attach their eggs either to plants or to oyster shells in this pond, as few other submerged objects are present. DISTINGUISHING CHARACTERS OF THE YOUNG OF THE GENERA GOBIOSOMA, MICROGOB JUS, AND GOBIONELLUS. The young of the three genera, Gobiosoma, Microgobius, and Gobionellus, dis- cussed in the following pages are quite similar, especially when very small. The following comparison is offered with the hope that it may be found useful. Myomere and vertebrae counts in the three genera are almost identical, the range of vertebrae for the three genera (of Gobionellus only boleosoma was examined) being 11 or 12 + 15 to 17 and cannot be used in separating them and, therefore, are omitted. Length 2.0 to 3.5 mm GOBIOSOMA Body rather deep; depth just posterior to vent about equal to head. Vent typically well behind mid- body length. Eye small. Ventral outline of body usually with a few dark spots. Soft dorsal and anal bases short, the rays developed in some specimens, each fin with 11 to 14 rays. MICROGOBIUS Body as in Gobiosoma. Vent typically at midbody length. Eye slightly larger. Ventral outline usually with more numerous black spots, and a double row behind vent. Length 4-0 to 5.5 mm Soft dorsal and anal bases long, longer than in Gobiosoma, the rays partly undeveloped. GOBIONELLUS Body extremely slender; depth of body posterior to vent notably less than length of head. Vent notably behind midbody length. Eye small, bulging. Pigment spots as in Gobiosoma or more usually wanting. Soft dorsal and anal bases short, the rays partly undeveloped. The differences listed for smaller specimens, exclusive of the position of the vent which has changed, apply to fish 4.0 to 5.5 mm long. The two rows of dark spots behind the vent in Micro- gobius, now lying along the opposite sides of the anal base, have become very definite and serve as ready recognition marks. 546 BULLETIN OF THE BUREAU OF FISHERIES Soft dorsal and anal each with 11 to 14 rays. A dark sheath present above air bladder, not crescent-shaped. Length 6.0 to 7.5 mm Soft dorsal and anal each with 16 or 17 rays. A dark sheath over air bladder as in Gobiosoma. Soft dorsal and anal each with 11 to 13 rays. A distinct crescent-shaped dark area over air bladder. The differences listed in the shape of the body for smaller specimens apply to fish 6.0 to 7.5 mm long. Length 8.0 to 12 mm Body anteriorly rather robust somewhat rounded; head slightly depressed, as broad as deep. Mouth rather small; maxillary reaching about opposite an- terior margin of eye. Body and head compressed, rather deep. Mouth large, maxillary reaching opposite anterior margin of pupil. Body more or less round and very slender. Mouth small; maxillary scarcely reaching opposite anterior margin of eye. The differential color markings mentioned for smaller fish remain as described in specimens 6.0 to 7.5 mm long. Length 13 to 16 mm Body anteriorly quite robust; head notably depressed. Mouth only slightly oblique, terminal; maxillary reaching only slightly beyond anterior margin of eye. Dorsal spines all about equally spaced. Head and body compressed throughout. Mouth large, strongly oblique, terminal to slightly superior; maxillary reaching nearly op- posite middle of eye. Dorsal spines as in Gobiosoma. Body extremely slender, remain- ing round. Mouth moderately oblique, small, terminal; maxillary scarcely reaching opposite an- terior margin of eye. Last two dorsal spines much farther apart than the others. A COMPARISON OF THE EGGS AND THE YOUNG OF SOME AMERICAN AND EUROPEAN GOBIES It has been shown in the preceding paragraphs that the eggs of all the Ameri- can and European gobies that have been studied have adhesive threads or an adhesive foot by which the eggs become attached to submerged objects. A further comparison of the eggs, and the young, of the American and European species shows other simi- larities. The eggs of all the gobies, so far as they are known, are rather small. A dozen eggs of Gobiosoma bosci measured by us several hours after fertilization, when they had had time to expand, bad a length of 1.2 to 1.37 mm, and their greatest width was 0.52 to 0.59 mm. The eggs of Ctenogobius stigmaticus (— Gobionellus boleosoma ), according to Kuntz (1916), are extremely small, having an average diameter of only 0.3 mm. The eggs of European species, too, are rather small. Ehrenbaum (1905) gave the following lengths of the eggs for the European species named: Gobius ruthen- sparri, 0.7 mm; G. paganellus, 1.8 to 1.9 mm; G. pidus, 0.8 mm.; and G. niger, 1.5 mm. Lebour (1920) gave the length of the eggs of G. microps and G. minutus respectively as 0.85 to 1.0 and 1.08 to 1.4 mm. The eggs of Gobiosoma bosci when first spawned, as stated by Kuntz (loc. cit.) and verified by us, are nearly spherical. After fertilization they expand and become eliptical, the greater axis becoming nearly twice as long as the lesser one. The adhesive foot, consisting of a bundle of threads, remains attached to one pole of the longer axis. The eggs of Ctenogobius stigmaticus (= Gobionellus boleosoma ) according to Kuntz" (loc. cit.) are more or less irregular or variable in shape, and they remain so throughout the incubation period. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 547 All the eggs of the European gobies that have been described or figured, so far as the present writers are familiar with the literature, are more or less elongate. Accord- ing to Petersen’s figures (1917) the eggs of Gobius niger are very elongate, the greater axis being about four times as long as the lesser one, and one end of the egg is larger than the other one. The adhesive foot is attached to the smaller end. The eggs of Gobius ruthensparri, as figured by Petersen (loc. cit.) and by Lebour (1919), are more or less pear-shaped, and the greater axis is less than 2 times as long as the lesser one. The adhesive foot is attached to the larger end of the egg. The eggs of Gobius microps according to Petersen’s figures (loc. cit.) are very similar to those of G. ruthensparri. However, the upper end of the egg is rather less pointed and the lower end has a slight pedicel to which the adhesive foot is attached. Lebour (1920) stated that the eggs of Gobius pictus are very similar to those of G. microps. The eggs of Gobius minulus, as figured by Petersen (loc. cit.) and also by Lebour (1920) are similar in shape to those of G. microps, being more elongate, however, as the greater axis is nearly two times as long as the lesser one. The adhesive foot is attached to a slight pedicel. The eggs of Gobius paganellus, as figured by Lebour (1919), are eliptical and very similar in shape to those of the American species, Gobiosoma bosci. The greater axis is rather more than two times as long as the lesser one, and the adhesive strands are attached to one pole of the longer axis. A rather remarkable similarity exists among the young of the American and European gobies, that have been studied, as shown principally by the works of Kuntz (1916), Petersen (1917), and Lebour (1919 and 1920), and by the present paper. The newly hatched larvae are of a small to a moderate size. Kuntz (loc. cit.) gave the length of the newly hatched larvae of the American species, Gobiosoma bosci and Gobionellus boleosoma, respectively as 2.0 and 1.2 mm; Lebour (1919) stated that the newly hatched larvae of the European species, Gobius paganellus and G. ruthensparri, are respectively 4.0 and 3.1 mm long and (1920) that those of G. microps and G. pictus are respectively 3.0 and 2.7 to 3.0 millimeters in length; and Petersen (loc. cit.) gave the length of the newly hatched young of two other European species, Gobius niger and G. minutus as 2.6 millimeters each. In general the body of the larvae is quite elongate and slender, varying somewhat among the species. The caudal portion is especially slender and tapers gradually to a point. The vent is placed somewhere near midbody length ; generally, however, rather nearer the tip of the tail (without the finfold) than the end of the snout. At hatcliing the head is rather rounded and the mouth tends to be horizontal and inferior. Very soon the mouth becomes rather oblique. However, as the adult stage is reached the mouth, in at least several species, tends to assume again more nearly the position occupied at hatching. When the caudal fin is first developed it either has a straight or a slightly rounded margin. As development of the fish progresses this fin tends to become more or less concave. Even in Gobionellus boleosoma, which has a moderately long and more or less pointed caudal fin in the adult, this fin is concave in the young that are around 8 mm long. In some species as in those of Gobiosoma, reported upon in this paper, the caudal fin does not become round, as in mature fish, until virtually all the adult characters are developed at a length of about 15 mm. The spinous dorsal usually develops somewhat later than the other fins in teleosts. In the gobies this fin develops especially late, or not until all the other fins are quite fully formed. The body in larval gobies generally is quite transparent and often the notochord or vertebrae are visible in part. The plainly visible air bladder, commonly with a crescent-shaped black area over it, is characteristic. Usually a few pigment spots are present at hatching and others soon appear. General pigmentation, however, 548 BULLETIN OF THE BUREAU OF FISHERIES takes place at a rather advanced stage, that is, after virtually all the adult characters are developed. It is evident from the foregoing discussion that both American and European gobies generally have rather small eggs, which are variable in shape and somewhat in size among the species, and are equipped with an adhesive organ by which they become attached to submerged objects. The larvae as a rule are slender, quite transparent, and have at least a few pigment spots. Interesting phases in their development are the changes in the position of the mouth and in the shape of the caudal fin. In separating the species the myomere and vertebra counts sometimes are useful (although not in the American species discussed in this paper), the fin ray counts, as soon as they can be made, are extremely helpful, and the pigment spots are always important for identification purposes. GOBIOSOMA BOSCI (LACEPEDE) AND GOBSOSOMA GINSBURGI, HILDEBRAND AND SCHROEOER. NAKED GOBIES Two species of Gobiosoma, namely bosci and ginsburgi, occur in the waters at Beaufort. The last-mentioned species was described recently from Chesapeake Bay (Hildebrand and Schroeder, 1928, p. 324), Ginsburg (1933 p. 40) made a thorough study of the genus Gobiosoma, and found specimens of G. ginsburgi in collections from as far north as Cape Cod and southward to South Carolina, and of G. bosci from Long Island to Tampico, Mexico. Specimens of adult G. bosci are much more numerous in the collection from Beaufort than those of G. ginsburgi. However, the reverse is true with respect to the young. In Chesapeake Bay (Hildebrand and Schroeder, 1928, p. 325) bosci was taken in shallow water only, whereas ginsburgi was taken principally in rather deep water and rarely in shallow water. A similar vertical distribution of the species is indicated for Beaufort, since all adult bosci from this vicinity were taken in very shallow water along the shores, whereas all specimens of adult ginsburgi, except one, were taken in water from a few to several fathoms in depth. The young that are recognizable as to species (10 mm and upward in length) have a vertical distribution identical with that of the adults. KEY TO THE ADULTS OF THE LOCAL SPECIES a. Body rather robust, its depth 3.95 to 4.8 in its length without caudal fin; second dorsal normally with 13 rays, infrequently with 12 or 14; ventral disk short, reaching about half the distance from its base to the vent; no scales on base of caudal bosci. aa. Body more slender, its depth 6 to 7.15 in its length without caudal fin; second dorsal normally with 12 rays, infrequently with 11 or 13; ventral disk long, reaching two-thirds the distance from its base to the vent; two large scales on base of caudal, situated respectively on the base of the upper and lower rays of the fin ginsburgi. Although the characters mentioned in the foregoing key readily separate the adults when two or more of the characters are taken into consideration, they cannot be used successfully in separating the young under about 10 mm in length, because the characters either are entirely undeveloped at that size or so imperfectly developed that they are useless. Neither have we succeeded in finding other distinguishing characters. Therefore, the young (under about 10 mm in length) cannot be dis- cussed separately with respect to distribution, habitat, growth, etc. (figs. 30 and 31). Young Gobiosoma are rather generally distributed throughout the local waters and are very abundant, as shown elsewhere (p. 558). Their relative abundance, in fact, suggests that the adults are more common than indicated by the number DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 549 of grown fish generally secured in collecting nets. From the difficulty experienced in recapturing adult fish placed in a rather large tank table in the laboratory, it is evident that nets are not very efficient for capturing the fish, because they attach themselves, by means of the sucking disk, to the bottom or to objects in the water, or they hide under objects, thereby making it very easy for a net to slide over them. They no doubt escape the net in a similar way in nature. All of this is a further indication that the fish probably are more numerous than shown by the number of adults captured. No evidence indicating that any one individual produces an excessively large number of eggs was secured. A female 25 mm long with a greatly distended abdo- men, for example, contained only 249 eggs. Furthermore, the eggs in the ovaries in several specimens examined were all of uniform size, indicating that a single spawn- i cm. Figure 30. — Oobiosoma bosci. From adult 50 mm long. (Drawn by Louise Nash.) i cm Figure 31.— Gobiosoma ginsburgi. From adult 45 mm long. (Drawn by Louise Nash.) ing takes place during a season. The great abundance of the young — in several instances a hundred or more and in a few instances probably nearly a thousand specimens were taken in a single tow-net haul — then, would appear to be due to the presence of many adults. The local species of Gobiosoma, which rarely exceeded a length of 2 inches, obviously are too small to be of direct commercial value, yet they probably are of some importance as forage for commercial species. Notliing is known concerning the winter home of Gobiosoma. We have taken no specimens during the colder months of the year, or from early December to early May. It seems probable that these fish imbed themselves in mud during the winter. This probability is suggested by the abundance of gobies in seine hauls made in the Mullet Pond throughout the warmer months, in places where none were taken 550 BULLETIN OF THE BUREAU OF FISHERIES during the winter. This pond is connected with the adjacent sound only during exceptionally high tides, which occur generally only a few times during a year, and seldom near or in the winter months. It seems rather certain, therefore, that the gobies are present during cold weather but are not in open water where they can be caught with seines. Several species of Fundulus and Cyprinodon varigatus which also inhabit the Mullet Pond, are known to imbed themselves in mud during cold weather. Therefore, it seems probable that the gobies also enter the muddy bottom during the winter. No external structural characters by means of which the sexes may be distin- guished have been found. In general, the males range larger in size and are darker in color. However, specimens intermediate, both with respect to size and color, are nearly always present in collections, making a complete separation of the sexes from external characters impossible. In gravid fish the anal papilla, although present in both sexes, appears to be larger in the female than in the male. SPAWNING The information about spawning in Gobiosoma was derived mostly from the study of large collections of young. It has not been possible, however, as explained elsewhere, to separate definitely the young (less than about 10 mm in length) into two groups, representing the two local species, bosci and ginsburgi. Therefore, it is not known positively, although it is highly probable, that both species are repre- sented among the fry. Neither is it known definitely whether much of the infor- mation derived from the study of the collections is applicable to one or both forms. The data based on the study of the collections of young give no evidence of two predominating spawning periods. Therefore, if two species are represented among the fry, the spawning seasons probably occur simultaneously or overlap so fully that no distinction may be made either from the size of the young nor from their abundance. Collections of young Gobiosoma were made from 1927 to 1931. The larvae first appeared in the tow in May (the earliest date being May 11, 1929), and throughout the summer and fall until December (the latest date being Dec. 6, 1929). The larvae were common to abundant from June to September each year but most numerous during July and August. During October, November, and December only a few scattered ones were secured. Among the females in the relatively large collections of adults made in the Mullet Pond during August (1930), principally small individuals were in spawning condition, the larger ones evidently having spawned out. This condition suggests that the principal spawning season was past, and is in general agreement with the situation suggested by the data based on the collections of young. Although the fry were abundant in the towings during August it may be assumed, that many of those taken, particularly the larger ones, were hatched during July. Furthermore, in September there occurred a pronounced drop in the number of young. The evidence, therefore, is that spawning begins early in May or possibly during the latter part of April, that it occurs most abundantly during July and ends except for an occasional late spawner, in September. The larvae of Gobiosoma were taken over a wide variety of conditions and over a comparatively large area, as explained under that section of this paper dealing with distribution (p. 558). Since the eggs become attached and do not drift as already shown, spawning probably takes place over a large portion of this area and over a wide variety of conditions. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 551 Nothing definite can be said concerning the spawning places of the two species of Gobiosoma represented locally, first, because the young under 10 mm in length could not be separated, and second, because ripe G. ginsburgi were not taken. Ripe G. bosci were taken only in the Mullet Pond where they no doubt spawn, as the larvae were present in the same places where the adults were taken. It seems probable, therefore, that spawning takes place in the general habitat of the adults. If that were the case, then it would follow that G. bosci would spawn only in shallow water along the shores, whereas G. ginsburgi would spawn principally in somewhat deeper water (see p. 548). DESCRIPTIONS OF THE EGGS AND YOUNG Eggs. — The eggs stripped from ripe fish taken in the Mullet Pond were first described by Kuntz (1916, p. 423). Since the present investigators have found only G. bosci in that pond, it seems probable that the eggs described were of this species. The eggs were secured again by the present investigators in August 1930 from fish taken in the Mullet Pond. The description of the eggs given by Kuntz is essentially correct. The ova when expressed from the female adhere in clumps unless immediately separated. The eggs when seen in a mass with the unaided eye are yellowish in color and quite opaque. Under the microscope a “bundle” of gelatinous threads with small branches are seen to be joined to the egg membrane at a certain point. These threads cause the eggs to adhere. Their function no doubt is that of attaching the eggs to vegetation or other objects in the water. The eggs are slightly heavier than sea water and when placed in a dish of water they sink gradually. The mature unfertilized eggs, as observed by us, generally, are slightly elongate, but sometimes nearly spherical. The variation in the major axis of five selected eggs ranged from 0.637 to 0.675 mm and for the minor axis in the same eggs it was 0.52 to 0.6 mm. As soon as fertilization had taken place the eggs began to expand and became elliptical in shape. In the process of expansion the minor axis retained about its former length, for in 10 eggs measured during various cleavage stages it ranged from 0.573 to 0.592 mm. The major axis, however, becomes much longer, for its range in length in the same eggs ranged from 1.147 to 1.369 mm. Expansion appears to be fully completed by the time the first cleavage takes place and thereafter, ac- cording to our observations, the egg changes little or not at all in shape. According to Kuntz’s figures, eggs with large embryos are more pronouncedly elliptical than those in the early cleavage stages (fig. 32). When the egg is fully expanded a relatively large perivitelline space is present, for the yolk mass occupies somewhat less than half the space within the egg membrane. The position of the yolk varies greatly. Generally it lies toward one pole of the major axis of the egg and most frequently opposite the pole at which the gelatinous adhesive strands are attached. Occasionally, however, it is much nearer the opposite pole of the major axis, or it may occupy an intermediate position (fig. 33). The yolk mass of the egg when seen under magnification has a greenish-yellow cast and it contains many minute oil globules. Due to the opaqueness of the yolk, many of the phases in the development either cannot be seen at all or are obscure. The processes in the development of the egg are well described and accurately figured by Kuntz (1916, pp. 423 to 426, figs. 43 to 50). The cells in the early cleavage stages stand out very prominently as round elevations. As cleavage advances the fissures become less pronounced and gradually the blastoderm becomes circular in outline and sharply differentiated from the yolk (figs. 34 and 35). 154979—38 4 552 BULLETIN OF THE BUREAU OF FISHERIES The development is rapid at first. Eggs fertilized at 10:30 o’clock in the morning and held at a temperature close to 27° C., for example, reached the 2-cell stage in Figure 32.- Figure 33.- Figure 34.- Figure 35.- Figure 36.- Figure 37.- Figure 38.- Figure 39.- -Gobiosoma bosci. -Gobiosoma bosci. -Gobiosoma bosci. -Gobiosoma bosci. -Gobiosoma bosci. -Gobiosoma bosci. -Gobiosoma bosci. -Gobiosoma bosci. From mature unfertilized egg. (Drawn by Effie B. Decker. After Kuntz.) From egg with fully developed blastodisc, BD. (Drawn by Effie B. Decker. After Kuntz.) From egg with blastoderm of 2 cells. (Drawn by Effie B. Decker. After Kuntz.) From egg with blastoderm of 4 cells. (Drawn by Effie B. Decker. After Kuntz.) From egg with blastoderm of many cells. (Drawn by Effie B. Decker. After Kuntz.) From egg with recently differentiated embryo. (Drawn by Effie B. Decker. After Kuntz.) From egg with well-formed embryo. (Drawn by Effie B. Decker. After Kuntz.) From egg with large embryo. (Drawn by Effie B. Decker. After Kuntz.) 1% hours, the 4-cell stage in 1 % hours, and the 8- and 16-cell stages (for there were some eggs in each stage) were reached in about 2% hours. In about 22 hours, at a temperature varying from 26° to 27° C., the embryo had become well formed. There- DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 553 after development progressed less rapidly, and hatching took place in about 4 days at a water temperature varying from about 26° to 28° C. (figs. 36, 37, and 38). Although the yolk mass, after the expansion of the egg is completed, occupies less than half the area within the eggs, as stated elsewhere, nearly the entire space is utilized later by the advanced embryo which becomes bent back on itself, with the tail pointed in the general direction of the head. The position of the embryo within the egg is not always the same. In most instances the head is pointed toward the pole of the major axis at which the adhesive threads are inserted, but occasionally it is directed toward the opposite pole. This fact is not brought out by Kuntz (loc. cit.) (fig. 39). Figure iO.—Gobiosoma bosci. From a newly hatched fish about 2 mm long. (Drawn by Etfie B. Decker. After Kuntz.) Newly hatched young, 2.0 mm long. — The incubation period occupies about 5 days at the usual summer temperatures (around 24° to 27° C.) prevailing in the laboratory at Beaufort. The newly hatched fish is approximately 2.0 mm long and almost transparent. The air bladder is visible at the posteriodorsal aspect of the yolk mass. The vent is situated nearer the tip of the tail than the end of the snout. The mouth is inferior, although somewhat later it becomes strongly oblique, as shown subsequently. A few small pigment spots occur just over the vent and at the base of the ventral finfold posterior to the vent (fig. 40). Figure 41. — Gobiosoma bosci. From a fish hatched in the laboratory, a few days old, and about 3 mm long. (Drawn by Effie B. Decker. After Kuntz.) Kuntz (1916) was able to keep the fish hatched in the laboratory alive until a length of about 3.0 mm was attained. The mouth in the meantime, according to the illus- tration presented had moved forward and had become terminal and somewhat oblique. The line of pigment spots at the ventral outline of the tail had become somewhat more conspicuous than in the newly hatched fish (fig. 41). The eggs and young hatched in the laboratory, upon which the foregoing descrip- tions are based, are known definitely to belong to G. bosci. The young up to 10 mm in length, upon which the descriptions that follow are based, were taken in the tow and probably include both local species, as already explained. 554 BULLETIN OF THE BUREAU OF FISHERIES Kuntz did not have sufficient material for the preparation of descriptions and illustrations of all the stages in the development of the young. This information is supplied in the following pages.4 Specimens 1.8 mm long. — The smallest individuals in the collection of preserved specimens, which we assign with some doubt to this genus because of their similarity to Microgobius, are only about 1.8 mm in length. Such specimens are farther de- veloped than a fresh or live larval fish 3.0 mm in length (fig. 13), which indicates that considerable contraction probably has taken place during preservation. The body is rather slender and somewhat compressed. The yolk is completely absorbed and the abdominal mass is quite small. The air bladder is plainly visible through the abdominal wall, lying dorsally of the abdominal mass. The intestine is free or at most loosely attached posteriorly and the vent is far behind midbody length. The finfold remains continuous but has slight indications of rays where it surrounds the pointed tail. The eye is excessively large, being equal to about three-fourths the depth of the head. The mouth is almost vertical and the snout is turned up slightly at the tip. The color consists of a few dark clxromatophores on the median line of the abdomen and on the ventral surface of the tail, that is, at the base of the ventral finfold (fig. 42). The position of the mouth in our specimens differs sharply from that shown in Kuntz’s (loc. cit.) illustrations (figs. 40 and 41). Figure 40, based on a newly hatched fish, shows an inferior mouth, and figure 13, based on a fresh fish 3.0 mm long, represents the mouth as terminal and only slightly oblique, and shaped very much as in the adult. A sudden and a pronounced change in its position must take place since the specimens here described cannot be much older, as already indicated, than the 3 .0 mm fish shown in figure 41 . Illustrations, in various works, of the development of European gobies, too, show that the mouth in newly hatched larvae is inferior and that it tends to become oblique very early in life. Specimens 4.0 mm long. — The body is moderately slender and notably com- pressed. The caudal portion of the body which is much more slender than the trunk in smaller individuals has become much deeper and at this size the depth just posterior to the vent is nearly as great as it is in advance of it. The air bladder remains visible, microscopically, through the abdominal walls, but the intestine which is partly free posteriorly in smaller specimens is now quite fully invaginated. 4 Shropshire (1932, pp. 28 and 29, figs. 1 to 4) described four stages of young gobies under the name Gobiosoma molestrum, which is a synonym of G. bosci according to Ginsburgh (1933, p. 32). The two smaller stages are so different in the shape and position of the mouth from our series that they undoubtedly represent a different species. The two larger specimens figured could conceivably be identical with those of our series. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 555 The finfold no longer remains continuous, as the caudal fin is fairly well developed and the bases of the dorsal and anal fins have become indefinitely outlined and some of the rays have become differentiated. Pectoral fins, too, are evident, but the ventrals remain undeveloped. The notochord is curved upward posteriorly, giving the tail a heterocercal appearance. The mouth remains nearly vertical as in younger fish. The body is unpigmented, except for a few dark chromatophores on the median ventral outline of the body and tail, the last one of these dark spots, situated near the end of the anal base, being the largest (fig. 43). Specimens 5.0 mm long.— The shape of the body is essentially as in specimens 4.0 mm long. The mouth, however, is not quite as nearly vertical as in the smaller fish. The development of fins has progressed rather rapidly, the caudal, soft dorsal and anal being sufficiently developed to show the rays rather definitely and it is now possible to enumerate the rays of the second dorsal and the anal quite definitely which is a great help in identification. The first, or spinous dorsal, is partly developed in some specimens, but it is impossible to count the spines accurately. The pectoral fins are plainly evident but without distinct rays. The ventral fins, however, are still undeveloped. The air bladder remains slightly visible, microscopically, through the abdominal walls as an area which is slightly more transparent than the abdomen is elsewhere. The notochord is still turned upward posteriorly and pigmentation remains essentially as in 4.0 mm specimens (fig. 44). Specimens 7.5 mm long.— The body remains shaped essentially as in 5.0-mm specimens, that is, compressed and slender, the depth being contained in the length to base of caudal fin about 6.25 times. The fish has a somewhat different appearance, however, at this size, mainly because of the rather pointed snout; for the mouth, although still superior, has become rather oblique with a somewhat pointed projecting mandible, as seen in a lateral view. The muscular rings on the body remain rather dis- tinct, but the heterocercal character of the tail generally has disappeared. The air bladder remains visible, microscopically, through the body wall as a somewhat lighter area. The development of the fins has progressed slowly. The spinous dorsal is not yet evident, but the ventrals now appear as a short tuft of membrane 556 BULLETIN OF THE BUREAU OF FISHERIES without evident rays. The caudal fin seems to have a nearly straight to a some- what concave posterior margin. In life the body remains highly transparent and the fish are almost invisible when caught in a net, except for the dark eyes. Pig- mentation on the body consists of two short dark lines situated on the median ventral line, the anterior one being under the posterior part of the head and the second one on the chest; a dark chromatopliore appears just in advance of the vent; and usually a series of indefinite dark markings extends from the origin of the anal to the base of the caudal, the last spot of the series being on the base of the lower caudal rays and generally slightly vertically elongate. These markings are different from those in related species of gobies occurring locally and serve as a recognition mark (fig. 45). Specimens 10 mm long. — The two local species of Gobiosoma are rather defi- nitely separable at a length of 10 mm. In both species the body has become more robust since a length of about 7.5 mm was attained and generally it is somewhat rounded anteriorly. The greatest depth of the body in bosci is contained in the length to the base of the caudal about 5.3 times, whereas ginsburgi generally is more slender, the depth being contained in the length about 6 times. In bosci the head is nearly as broad as deep, the eyes have become slightly superior, the snout is comparatively round and blunt, and the mouth is moderately oblique and terminal. In ginsburgi the development in all these respects is rather less pronounced. The ventral fins are quite fully developed as a sucking disk in both species (reaching its greatest develop- ment in bosci at this size and becoming proportionately shorter later in life), and the first dorsal is present, although the spines are very weak and slender. The margin of the caudal fin remains straight to slightly concave. Variation in the progress of pig- mentation is evident among specimens of the same species. However, the develop- ment generally is further advanced in bosci than in ginsburgi. In the latter no general pigmentation has taken place and the markings remain virtually as described for the 7.5-mm fish. The posterior one of the two short dark lines on the median line of the chest is now situated at the base of the ventral disk. The dark markings along the base of the anal and on the ventral outline behind the anal clearly are short hyphen- shaped lines when viewed ventrally and are rather more distinct than in smaller fish. In the most profusely pigmented individuals of bosci indefinite cross bars are present on the upper part of the sides and back. Also, an oblique bar reaches from the eye to the mouth and another bar occupies the base of the caudal fin (fig. 46; based on a rather unusually well developed specimen for its size). A difference among specimens in the robustness of the body is present, as already shown. That the slender individuals generally are referable to ginsburgi is evident DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 557 from the presence of two scales on the base of the caudal fin, previously described (p. 548), which bosci does not possess (fig. 47). The development of adult characters at a larger size in ginsburgi than in bosci sug- gests that the first-mentioned species might reach a larger size. Judging from the adults taken this is not the case, for on the contrary the largest specimens in the collection are bosci. Specimens 15 mm long. — At this size bosci is robust anteriorly, the depth being contained about 5.3 times in the length to base of caudal fin; the head is depressed and quite as broad as deep; the snout is blunt; the eyes are directed slightly upward; the mouth is small, gently oblique, and terminal to slightly inferior; the maxillary reaches a little past anterior margin of eye; the fins are all fully developed, the margin of the caudal fin now being slightly rounded; and the body is fully pigmented. It is evident, therefore, that fish of this size have acquired nearly all the characters of the adult and are readily identifiable with the grown fish (fig. 48). Figure 48 .—Gobiosoma bosci. From a specimen 15 mm long. It is possible to identify the young at a length of about 15 mm as to species with a reasonable degree of certainty. G. ginsburgi, at this size, is notably more slender, the depth being contained in the total length about 6.1 times. The greater length of the ventral disk, winch (having attained its highest state of development at about this 558 BULLETIN OF THE BUREAU OF FISHERIES size, becoming proportionately shorter later in life) reaches nearly to or even past the vent in some specimens, also is evident. The interorbital space, too, is narrower, being equal to only about half the width of the pupil, whereas in bosci it is fully equal to the width of the pupil. Furthermore, ginsburgi has two scales on the base of the caudal fin which the other species does not possess. Pigmentation in ginsburgi in the specimens at hand has not progressed as far as in bosci of the same size. However, a considerable degree of variation in color development appears to exist among both species and the degree of pigmentation may be of no specific importance (fig. 49). General characteristics of the young. — In general young Gobiosoma, even before fin rays are developed, may be recognized by the rather deep body, by the vertical mouth, by the air bladder which is visible as a clear area through the body wall, and, perhaps most important of all, by the pigment spots present, which remain about the same throughout the larval stages, or until pigmentation becomes general. These spots are black and consist of a single row occupying the median ventral line of the body. Two elongate spots (short lines) are situated under the head and chest, one or two immediately in advance of the vent and several behind the vent, or along the base Figure 49. — Oobiosoma ginsburgi. From a specimen 15 mm long. of the anal when that member becomes developed, the last spot of the series being at the base of the caudal when that fin becomes differentiated. When the dorsal and anal fins become developed, at a length of about 5.0 mm, the rather low number (gen- erally 11 to 13) of rays in each fin is of much help in identification. At about this size the body becomes quite robust, the head rather broad, the caudal peduncle is short and deep, and the mouth is less nearly vertical than previously. Identification now is much simplified. DISTRIBUTION OF THE YOUNG Young Gobiosoma were taken in tow nets and seines in many places and over a wide variety of conditions, ranging from brackish water creeks and ponds, through salt and brackish estuaries, in Beaufort Harbor, along Bogue and Shackleford Banks (both shores) and at sea as far as approximately 15 miles offshore in 10 to 12 fathoms of water. The great majority of the hundreds of specimens collected were taken on the bottom, although occasionally a few appeared in the surface tow. The indica- tions are, then, that the young, like the adults, dwell principally on the bottom. Since the larvae (under about 10 mm in length) could not be separated as to species, the distribution cannot be given separately for each species. Among the young fish that are recognizable, bosci was taken only in shallow water and only once at an outside station, that is, about 1 mile off Bogue Banks in a few fathoms of water. G. ginsburgi, on the other hand, was taken in both shallow and rather deep water and frequently at offshore stations. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 559 As to time, the larvae are distributed over the entire summer, appearing first in May and a few stragglers as late as December, but they were common to abundant only from June to September. GROWTH Growth in Gobiosoma probably progresses moderately fast. It is not always an easy matter, however, to distinguish between those of the 0-class and the older ones, as the sizes intergrade during late summer. Specimens 18 mm long, taken in August, are recognizable as young of the current season and may be among the largest of their year class. Larger examples of the current year, if present, apparently could not be recognized, for they are as fully developed as the adults. Since the spawn- ing season may be said to end, except for an occasional late spawner, by the end of August, it seems unlikely that the largest young become mature during their first summer. However, sexual maturity is reached at a very small size, for we have seen a few gravid females only 23 mm long and many gravid ones from 25 to 30 mm in length. Therefore, the authors are not prepared to state positively that none of the young become mature during their first summer, but they regard it as quite unlikely. It seems highly probable, however, that sexual maturity is reached during their second summer. MICROGOBIUS HOLMESI SMITH. HOLMES GOBY Two very closely related species of Microgobius, namely eulepis and holmesi, are recognized from North Carolina. As understood by us eulepis has a somewhat more slender and less strongly compressed body than holmesi, the depth in the former in the two specimens at hand is contained respectively 6.5 and 7.0 in the standard length, whereas in eight specimens of the latter the depth goes into the length 4.7 to 5.75 times. The mouth in eulepis appears to be rather more nearly vertical and the ventral disk is shorter, failing to reach the vent, whereas in holmesi the disk usually reaches to or beyond the vent. It is possible that holmesi may grow some- what larger. However, it seems probable that a further study based on a larger number of specimens than is now available may show that the two nominal species intergrade, and in fact are identical. As now understood eulepis is very rare at Beaufort, whereas holmesi is moderately common. The known range of the two species is coterminous extending from Chesapeake Bay to North Carolina.5 The sexes are readily separable in M. holmesi, as the male has a row of prominent black spots on the interradial membranes of the anal just below the pale margin of the fin. These spots are entirely missing in the female, in which, as contrasted to the male, the membranes between the longest dorsal spines is jet black distally. In general, the males also have higher fins. The ventral disk, for example, usually reaches the origin of the anal in adult males, whereas it frequently reaches only to the vent in adult females. Furthermore, the females, at least during the breeding season, have a larger anal papilla. Whether similar sexual differences exist in eulepis cannot be stated at this time. The two specimens in the present collection in general agree in color with the females of holmesi .6 It is not surprising that the larval and young Microgobius do not appear to be separable into two species (if indeed more than one species is represented) since the adults of the local representatives are very closely related. Since M. holmesi is com- » Since the preparation of this manuscript Isaac Ginsburg, who has made a special study of the American gobies, concluded (Copeia, No. 1, 1934, p. 35) that M. holmesi and M. eulepis are identical and that both are synonyms of M. thalassinus. • Smith (1907, p. 367) presents a very satisfactory illustration of an adult male Microgobius holmesi. 560 BULLETIN OF THE BUREAU OF FISHERIES paratively common, whereas, M. eulepis is very rare, as previously stated, it seems logical to refer all the young, at least tentatively, to holmesi. The locally represented species of Microgobius reach a length of only about 2 inches and they probably are of only slight economic value even as forage fish, because of their comparative scarcity. The rather large number of young taken suggests, however, that the fish may be somewhat more common than is indicated by the scarcity of adults in the collections. Neither adults nor young were taken from December to February, inclusive. It seems probable, therefore, that these fish leave the local waters during the winter, or that they possibly seek shelter in the mud or sand like Fundulus and probably other minnows. SPAWNING The eggs of Microgobius have not been studied. M. holmesi with large roe were taken only during the first half of July. That the spawning period of this species is not limited to the month of July is evident, however, from the collection of larvae at hand, as shown subsequently. Smith (1907, p. 368) reported that a female M. eulepis distended with nearly ripe eggs was taken at Beaufort on May 18 (1905). No gravid fish of the last mentioned species were seen during the present investigation. A few young Microgobius, only about 3.0 to 4.0 mm long, were taken as early as March 11 (1929), and a few equally as small were taken as late as November 21 (1927). The larvae were numerous, however, only during July, August, and Sep- tember. The collection of young Microgobius indicates, therefore, that some spawn- ing takes place as early as March, that it continues throughout the summer, probably extending into the month of November, and that the principal spawning season occurs during July, August, and September. Larval Microgobius were taken over the entire area in which tow-net collections were made, including Beaufort Harbor, the adjacent sounds and estuaries, and off Beaufort Inlet to Cape Lookout and as far as 12 to 13 miles offshore. It seems reasonable to expect Microgobius to produce eggs which become attached like those of Gobiosoma and Gobionellus, and like those of the various European species that have been studied. If that be true the eggs do not drift and the recently hatched young should be expected to occur somewhere near the place where the eggs were spawned. It seems probable, therefore, that spawning takes place over much of the area in which the larvae were taken. BESCRIPTIONS OF THE YOUNG Specimens 1.6 to J+.O mm long. — Specimens of Microgobius 4.0 mm and less in length generally are difficult to separate from Gobiosoma (figs. 50 and 51). A careful study has revealed no outstanding structural differences in these small larvae, and color marking in preserved specimens, except in rather rare instances, are not of much help until a length of about 4.0 mm is attained. In general, the vent in Microgobius is slightly more anterior in position than it is in Gobiosoma. This difference is evident only when specimens of even size are compared, and it is not readily shown in a table of measurements, because few specimens are straight enough for an accurate measure- ment and, furthermore, the position of the vent evidently changes with growth. The quotients derived from dividing the caudal portion of the body into the total length to the tip of the notochord, even in specimens varying less than a millimeter in length, do not show a constant difference. An average difference, however, is evident, for in 12 specimens measured of each genus, the caudal portion of the body averages 2.15 DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 561 times in the length in Microgobius (with a range of 2.02 to 2.3) and 2.24 times in Gobiosoma (with a range of 2.13 to 2.33). Microgobius also has a slightly larger eye. However, the difference is so slight and the size of the eye changes so rapidly with growth that the difference cannot be shown easily in a table of measurements. Furthermore, the size of the eye in the delicate larvae of this size appears to have been affected by the strength of the preservative used which varied considerably with the different lots in the present collection. The first concrete difference observed between specimens of Microgobius and Gobiosoma are color markings which generally are fairly well developed at a length of 4.0 mm, frequently at 3.0 mm, and occasionally at a somewhat smaller size. Gobio- soma, as explained elsewhere (p. 555), has only a few dark markings on the ventral outline, consisting generally of about two elongate spots on the median ventral line under the head and on the chest, a larger spot or blotch at the vent and a few posterior to the vent, including one (situated at the last rays of the anal fin when that member becomes developed) that is notably larger than the others. Microgobius, on the other hand, has more numerous dark spots on the ventral outline and a double row of rather Figure 51. — Microgobius sp. From a specimen 4.5 mm long. well outlined dots extending from the vent to near the end of the tail, or nearly to the base of the caudal fin when that member becomes developed. The dark spot at the vent typically forms a short black line lying parallel with the upper margin of the loosely attached hind gut. Occasionally the larvae of Microgobius, only a few millimeters long, have a few dark spots at the nape and two more or less definite rows of minute black dots on the dorsal surface of the caudal portion of the body (fig. 51). No markings of any kind have been noticed on the dorsal outline in the larvae of Gobiosoma. Specimens 5.0 mm long. — At a length of about 5.0 mm the second dorsal and anal fins usually are sufficiently developed (although some of the posterior rays generally are undifferentiated) to admit of a count accurate enough to establish the fact that these fins are too long for Gobiosoma which has only 11 or 12 rays in each fin. The slightly larger eye in Microgobius remains equally as evident at this size as in the smaller individuals previously described. Differences in color markings 562 BULLETIN OF THE BUREAU OF FISHERIES between Microgobius and Gobiosoma which generally become established when the fish reach a length of about 4.0 mm, remain essentially as described in the preceding paragraph (fig. 51). Specimens 7.5 mm long. — The general shape of the body, as well as the head and mouth remain very similar in Microgobius and Gobiosoma. The slightly larger eye in the first-mentioned genus, noticed in the very smallest larvae at hand, remains evident. The second dorsal and anal fins are quite fully developed and a fairly accurate fin-ray count is obtainable. In addition to the higher fin-ray count (second dorsal and anal each with about 16 or 17 rays) in Microgobius, it is evident now that the caudal peduncle also is shorter. The color markings in Microgobius remain largely as when they first appeared, except that no dark markings are present on the back in any of the specimens examined. The two rows of dark spots along either side of the base of the anal fin are more distinct and each spot is now horizontally slightly elongate (fig. 52). Figure 52 —Microgobiut sp. From a specimen 7.6 mm long. Specimens 10 mm long. — A difference in the shape of the body between specimens of Microgobius and Gobiosoma is plainly evident in examples about 10 mm long. In Microgobius the body remains nearly as strongly compressed as in smaller fry, but in Gobiosoma it has become notably more robust anteriorly and the head is broader with a wider interorbital space. Although the mouth remains about equally oblique in the representatives of each genus, it is evident now that the gape in Microgobius is somewhat larger, the maxillary reaching about below the anterior margin of the pupil, whereas it ends slightly in advance of this point in Gobiosoma. All the fins, exclusive of the first dorsal, are now quite fully developed. The ventral disk is long and slender. The caudal fin is shorter than the head and its margin remains straight to slightly concave. The spinous dorsal usually consists of four very slender spines, with two or three of the posterior ones still missing. Pigmentation has advanced only slightly. In addition to the pigment spots described in smaller fish, at least some specimens now have a few dark markings about the mouth, a few on the side of the head and an indication of a slight dark bar at the base of the caudal. Some specimens also have some black dots along the bases of the dorsal fins (fig. 53). DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 563 Specimens 15.0 mm long. — The body has become more robust but remains com- pressed throughout. The eyes are lateral in position; the mouth is large, terminal to slightly superior and nearly or quite as oblique as in 10-inm fish; and the maxil- lary reaches nearly opposite middle of eye. The fins, including the spinous dorsal, are all well developed. The caudal is strongly rounded to pointed and fully as long as the head; the long ventral disk usually reaches nearly to or even beyond the origin of the anal; the dorsal spines are rather long and slender, the longest ones being equal to or slightly longer than the eye and snout; and the posterior rays of the second dorsal and anal frequently reach the base of the caudal when deflected. Pigmenta- tion has progressed considerably, for nearly the entire body is covered with minute dark points. The color markings along the ventral and dorsal outlines, described for smaller fish, in most specimens have become somewhat less pronounced (fig. 54). It is evident from the foregoing description that nearly all the structural char- acters of the adult, exclusive of scales, are developed at a length of 15.0 mm and identification is comparatively easy. The scales first appear when the fish is about 18 mm long and they become evident on the caudal peduncle first, and from there squa- ma tion proceeds forward until it is virtually completed when the fish reaches a length of about 23 mm. Figure 54. — Microgobius sp. From a specimen 15 mm long. It is interesting that in specimens from about 16 to 25 mm long the ventral disk is proportionately longer than in larger fish, generally reaching to or a little beyond the origin of the anal at this size. Microgobius and Gobiosoma, at a length of about 15 mm, differ strongly. Microgobius now has a more compressed body ; a much narrower and deeper head with the eyes fully lateral; the mouth is much larger, more strongly oblique, and terminal to slightly, superior; and the fins are much higher than those in Gobiosoma. DISTRIBUTION OF THE YOUNG It already has been stated under the section of this paper dealing with spawning that the larvae of Microgobius were taken virtually over the entire area in which tow net collections were made. This area includes Beaufort Harbor, the adjacent sounds and estuaries, and at sea, 13 to 15 miles offshore. The variation in depth over this area ranges from a few feet to 12 fathoms. The larvae were taken in inside waters 37 times and offshore 20 times. The inside catches, besides being more numerous, generally contained a larger number of fish per catch. A total of about 90 larvae was taken outside, whereas a total of about 760 larvae wTas taken inside. Although a somewhat greater number of hauls probably was made in the inside waters, that difference would not offset the great difference between the number of larvae taken in inside and outside waters. It may be con- 564 BULLETIN OF THE BUREAU OF FISHERIES eluded, therefore, that young Microgobius are decidedly more numerous in Beaufort Harbor and adjacent sounds and estuaries than they are off Beaufort Inlet. In about an equal number of surface and bottom hauls Microgobius was taken at the surface 13 times and on the bottom 45 times. Young Microgobius, like the adults, therefore, are primarily bottom dwelling. With respect to distribution over time, a few young appeared in the catches in March, April, and May, they become rather numerous in June, and abundant in July, August, and September; only a few were taken in October and November, and none from December to February. GROWTH 1 The largest young of the season were caught in September and had reached a length of 32 mm. It seems quite certain that such individuals will reach an adult size of 40 to 50 mm, and sexual maturity, during their second summer. However, some of the young taken with the large ones mentioned are only about 5.0 mm long. It seems doubtful that such individuals will attain an adult size and sexual maturity before their third summer. It may be concluded, therefore, that some of the larger and faster- growing individuals certainly reach an adult size and sexual maturity during their second summer when about a year old. Others almost certainly are not full grown, nor sexually mature, before their third summer. Figure 55. — Gobiontllus boleosoma. From an adult 34 mm long. LOCAL SPECIES OF GOBIONELLUS A single species of Gobionellus, namely, boleosoma, heretofore has been recorded from the coast of North Carolina. This fish was assigned to Ctenogobius stigmaticus by Smith in “The Fishes of North Carolina” (1907, p. 365). However, Ginsburg (1932, p. 23) in an exhaustive study of extensive collections in the Bureau of Fisheries and the National Museum failed to find stigmaticus north of Florida This investi- gator assigned the common Atlantic coast species of scaled goby (which ranges from North Carolina at least as far south as Panama) to boleosoma and places it in the genus Gobionellus. Mr. Ginsburg’s nomenclature has been adopted by the present writers. Attention is called to the fact that Smith’s illustration (loc. cit., fig. 167) is not correct for boleosoma. The figure probably represents an entirely different species. Accordingly, a drawing of an adult based on a specimen from Beaufort has been prepared (fig. 55). A few representatives of a second species of Gobionellus, namely, shufeldti, recently were taken in fresh water in Newport River. 6. shufeldti according to Ginsburg (1932, p. 14) differs from boleosoma: (a) In having one or more rows of scales on the median line of the back in advance of the first dorsal (in one specimen from Beaufort, assigned to this species, however, the median line of the back is naked as in boleosoma) ; ( b ) in having a slightly higher average number of rays in the dorsal and anal DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 565 fins, the typical number being dorsal 12, anal 13, as compared with dorsal 11, anal 12 in boleosoma; (c) in having no definite dark shoulder spot, nor V-shaped da.rk markings along the sides whereas these color markings generally are quite evident in boleosoma; (d) in reaching a larger size (according to the few specimens at hand from North Carolina the maximum size attained by shujeldti is about 70 mm, while the largest specimen of boleosoma, of which many specimens have been collected, is only 55 mm long); ( e ) in inhabiting fresh and slightly bracldsh water, whereas boleosoma ranges from salt to bracldsh water. If the young of shujeldti are present in the collection, they are confused with boleosoma. Two other species of gobies which Mr. Ginsburg (loc. cit.) also places in the genus Gobionellus, not as yet recorded from Beaufort, have been taken there recently, namely, oceanicus and hastatus. These species both differ from the other local Gobionellus in the more numerous rays in the dorsal and anal fins, the usual number in the dorsal being 14 and in the anal 14 or 15. The two species differ from each other largely in the number of scales in a lateral series, hastatus, according to 10 specimens from Beaufort, has 81 to 89, and oceanicus, according to 3 specimens from Beaufort, has 61 to 68 scales. The counts agree with those made by Ginsburg (1932, p. 39) based on specimens from Panama, Puerto Rico, Cuba, Florida, and Louisiana. The young of at least one of these species apparently also are included in the present collection. GOBIONELLUS BOLEOSOMA (JORDAN AND GILBERT). SCALLOP FISH The young of Gobionellus are much less numerous in the collection than those of Gobiosoma and Microgobius, but they are not rare, as 177 larvae are at hand. Adults of Gobionellus boleosoma have been taken more frequently locally than those of Micro- gobius, but much less often than Gobiosoma. The local distribution of adult G. boleosoma seems to be rather general. They were taken most frequently with seines in shallow water and on muddy bottom, both in salt and slightly bracldsh water, including in one instance a small drainage ditch. In somewhat deeper water they were secured only twice, once in the channel of Newport River, a few miles north of (he laboratory, and again at sea off Bogue Banks. Many variations or differences have been noticed among adults from Beaufort. In some specimens the body is more slender than in others. Also, the size of the mouth, the teeth, and the eyes varies. The median portion of the abdomen is variously scaled or naked and in some specimens the caudal fin is much longer than in others. Most of these differences certainly are associated with sex. In general, large males are more slender than females, and they have larger eyes, more prominent teeth, and a longer and more pointed caudal fin. The maximum size attained by this goby is about 55 mm. It no doubt is preyed upon by larger predatory fishes, but it evidently is not abundant enough to be of much importance as a forage fish. Therefore, its economic value must be very slight locally. Gobionellus boleosoma is present in the vicinity of Beaufort throughout the year, as both adults and young have been taken occasionally during the winter months as well as during the summer. Therefore, it evidently does not migrate, and no evidence indicating that it seeks protection from the cold by burying itself in mud or sand has been secured. SPAWNING Ripe or nearly ripe fish have been taken locally during July and August. Very small larvae (2.5 to 5.0 mm long), however, were taken as early as May 15 (1929), 566 BULLETIN OF THE BUREAU OF FISHERIES and as late as November 3 (1928). The young were not abundant at any time. The largest number of specimens was secured during July and August, which may represent the principal spawning period. It is not definitely known where this goby spawns. Tbe eggs are demersal and bear adhesive threads, which suggests that spawning probably takes place where there are sufficient objects in the water for the attachment of the eggs. Small larvae, only a few to several millimeters long, were taken virtually over the entire area in which towings were made. This area includes Beaufort Harbor, and the neighboring sounds and estuaries, as well as the waters off Beaufort Inlet, extending 12 to 15 miles offshore. It seems probable, therefore, that spawning takes place both in the inside protected waters and along the outside shores. DESCRIPTIONS OF THE EGGS AND YOUNG Eggs. — The eggs were described by Kuntz (1916, pp. 426-428) on the basis of samples stripped from fish taken and identified by the present senior author as Cteno- gobius stigmaticus, following Smith (1907, p. 365). The following descriptions of the eggs and their development is a condensed account, based on Kuntz ’s paper. The illustrations of the development of the egg and the figure of the newly hatched fish are also from Kuntz. The eggs are yellow in color, highly translucent, somewhat irregular in shape and have a diameter of about 0.3 mm. Their specific gravity is only slightly greater than sea water. The egg membrane is thin and delicate and usually drawn out into a blunt apex at the insertion of the “peduncle”, that is, at the insertion of the gelati- nous threads. The egg contains a relatively enormous amount of protoplasm and very little yolk (fig. 56). The fully developed blastodisc covers about half the area of the surface of the yolk. The first cleavage act, at ordinary summer laboratory temperature, takes place in about 30 minutes and the successive cleavages occur in rapid succession. The first cleavage plane cuts deep into the blastodisc and the first cells usually, although not always, are quite symmetrical. Until the 16-cell stage is reached the cells are in a single row. Thereafter they become heaped up on one side of the yolk (figs. 57, 58, and 59). As cleavage advances the blastoderm becomes more distinctly dome-shaped and it soon becomes thickest at the periphery. The peripheral growth of the blastoderm advances and the yolk becomes entirely engulfed, the blastopore closing within 6 hours after fertilization (figs. 60 and 61). Soon after the closing of the blastopore a distinct linear thickening of the blasto- derm, representing the axis of the future embryo, grows anteriorly from the blastopore. As the differentiation of the embryonic axis advances the anterior region of the differen- tiated area of the blastoderm becomes distinctly broader than the posterior region. That is, the differentiation of the embryo begins in the anterior or head region and advances posteriorly (fig. 62). The subsequent growth of the embryo advances rapidly. Within 11 hours after fertilization the embryo is well formed and it shows 10 to 12 somites. An hour later the embryo almost completely encircles the egg and the posterior region of the body is already free from the greatly reduced yolk mass. The embryo, although highly transparent, is marked by small areas of delicate pigment. It now more than encircles the periphery of the egg membrane, and the entire period of incubation at laboratory temperature occupies not over 18 hours (figs. 63 and 64). DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 567 — B3> Figure 56. — Gobionellus boleosoma. From egg with undivided blastodise, BD; yolk, Y. (Drawn by Effie B. Decker. After Kuntz.) Figure 57. — Gobionellus boleo- soma. From egg with 2-cell blastoderm. (Drawn by Effie B. Decker. After Kuntz.) Figure 59. — Gobionellus boleosoma. From egg with a 16-cell blastoderm. (Drawn by Effie B. Decker. After Kuntz.) Figure 60. — Gobionellus boleosoma. From egg with blastoderm of many cells. (Drawn by Effie B. Decker. After Kuntz.) Figure 63. — Gobionellus boleosoma. From egg with well-differentiated embryo. (Drawn by Effie B. Decker. After Kuntz.) Figure 62. — Gobionellus boleosoma. From egg showing an early stage in the differentiation of the em- bryo. (Drawn by Effie B. Decker. After Kuntz.) Figure 58. — Gobionellus boleosoma. From egg with a 4-cell blasto- derm. (Drawn by Effie B. Decker. After Kuntz.) Figure 61. — Gobionellus boleosoma. From egg with blastoderm growing around yolk shortly before closing. (Drawn by Effie B. Decker. After Kuntz.) Figure 64 .—Gobionellus boleosoma. From egg with large embryo, just before hatching. (Drawn by Effie B. Decker. After Kuntz.) 154979-38- 568 BULLETIN OP THE BUREAU OF FISHERIES The newly hatched fish, 1 .2 mm long— The following account of the newly hatched fish has been compiled from Kuntz’s description (loc. cit.) based on a fresh specimen: It is exceedingly delicate, and only about 1.2 mm long. It is highly transparent and marked by small areas of delicate yellow pigment on the dorsal surface of the head, over the vent and with a vertical band about half way from the vent to the tip of the tail. The vent is located slightly in advance of midbody length. The dorsal and ventral finfolds are continuous and the depth of each fold is equal to or greater than the depth of the body posterior to the vent (fig. 65). Figure 65 .—Gobionellus boleosoma. From newly hatched fish, length 1.2 mm. (Drawn by Effie B. Decker. After Kuntz.) Kuntz was able to keep the delicate larvae alive in the laboratory only a few hours, and he did not have any advanced larval stages. Descriptions and illustra- tions of the subsequent development of the young are based on preserved specimens contained in the collection studied by the present writers. Specimens 2.5 mm long. — The larvae of this genus are extremely slender and this character generally distinguishes them from those of Gobiosoma and Microgobius. The caudal portion of the body is especially slender and at this size notably longer than the rest of the body. The head is rather broad, its width being nearly as great as its depth; the mouth is almost vertical and very close in front of the moderately large protruding eyes. The air bladder is plainly visible as a round or slightly elon- gate clear area within the abdominal cavity; dorsally of the air bladder the dark peritoneum is visible (and at a slightly larger size the black peritoneum at this point becomes very pronounced and forms a recognition mark). Fins are undeveloped, except for a slight indication of rays in the fin fold around the tip of the notochord (fig. 66). Specimens 3.5 mm long. — The fish has become only slightly more robust than it was at a length of 2.5 mm. The mouth remains very close to the eyes, but has become slightly less vertical. Pectoral fins have appeared as tufts of membrane without rays. The notochord remains straight, with indications of rays around its tip. The soft dorsal and anal bases are in part evident, but no definite rays have developed. On some specimens a few very small dark spots are present along the ventral outline of the body and tail. The black peritoneum over the air bladder is moderately distinct and has acquired a crescent shape which is characteristic of the young of this species (fig. 67). Specimens 5.0 mm long. — Little change has taken place in the shape of the head and body since a length of 3.5 mm was attained. The mouth has become slightly DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 569 less strongly oblique, but remains superior, and the gape is proportionately further removed from the eye. The notochord is bent upward sharply at the tip, giving the tail a heterocercal appearance. The caudal fin is fully formed, with definite rays, and has a nearly straight posterior margin. The soft dorsal and anal contain some well-developed rays, but a definite fin ray count is not yet obtainable. The pectorals remain as tufts of membrane, and the ventral fins (disk) are not yet evident. Pig- mentation has made little progress. It now consists of two or three short, narrow, dark lines on the chest, a very small dark spot at the vent and a slightly larger one at or near the end of the anal base. The crescent-shaped dark area over the air bladder, visible through the abdominal wall, is prominent and serves as a ready recognition mark (fig. 68). Figure 69. — Gobionellus boleosoma. From a specimen 7.5 mm long. Specimens 7.5 mm long. — The body has increased somewhat in depth but remains comparatively very slender, the depth being contained in the length to the base of the caudal about 7.5 to 8.0 times. Some progress in the development of the fins has been made. It is now possible, in at least some specimens, to make a fairly accurate count of the soft dorsal and anal rays, each fin having 11 to 13 rays. The caudal fin is well developed and its margin is straight to slightly concave. The pectoral fins have indications of rays and the base of the ventral disk is just becoming evident. The spinous dorsal is undeveloped, or in some specimens just becoming evident. The notochord, sharply bent upward at its tip, remains visible. A crescent-shaped dark area over the air bladder is quite distinct. This dark area is visible with the unaided eye in somewhat larger specimens and is a definite aid in identification. Pigmenta- tion on the body remains virtually as in 5.0 mm fish (fig. 69). 570 BULLETIN OF THE BUREAU OF FISHERIES Specimens 10 mm long. — The body remains much more slender than in other local genera of gobies, and fully as slender as in 7.5-mm fish, the depth being contained in the length to base of caudal about 6.5 to 8.0 times. The mouth is still quite oblique, nearly terminal and small, the maxillary scarcely reaching opposite anterior margin of eye. The fins are all developed. However, the spines of the first dorsal usually remain short and slender. The ventral disk is fully developed and long, reaching about three-fourths of the distance from its base to the vent. The pectoral fins, too, are rather long but do not reach quite as far back as the ventral disk. The margin of the caudal fin is straight to rounded. The crescent-shaped dark area over the air bladder has become quite pronounced and in some specimens is clearly evident with the unaided eye. Pigmentation has made no definite advancement (fig. 70). Specimens 13 mm long. — The body remains extremely slender, the depth being contained in the length to base of caudal about 10 times. The head has become slightly broader and somewhat depressed. The mouth is small, oblique, and terminal, and the maxillary scarcely reaches the vertical from the anterior margin of eye. The air bladder is visible microscopically, but the crescent-shaped black area above it (that is, the dark peritoneum), very evident in somewhat smaller fish, is quite indistinct and somewhat changed in shape. The spinous dorsal now is fully developed and it is plain that the last two spines are much further apart than the others, which appears to be characteristic of the local species of the genus. The pectorals and ventral disk are long, but do not extend as far back on the body as in somewhat smaller fish. However, there appears to be some variation in this respect among individuals. The caudal fin is about as long as the head and its margin is slightly convex. Progress in pigmentation varies greatly. In a 13-mm specimen it has progressed little further than in the 10-mm fish, described in the foregoing paragraph. However, there is at hand one specimen 11 and another 12 mm long which have some dark markings on the head, including indications of a dark oblique bar between the eye and the mouth (characteristic of the adult); scattered dark dots on the back and along the ventral edge of the abdomen, a more definite series of black spots on the base of the anal, and with indications of wavy dusky bars on the caudal fin, as in the adult. No perfect specimens from Beaufort suitable for drawing are at hand. Further- more, the differences between fish 10 mm and 13 mm long are slight, as pointed out in the description. For these reasons no illustration of the size described in the fore- going paragraph is offered. Unfortunately specimens between 13 and 22 mm in length (the latter being adults) are not at hand. The fish described in the foregoing paragraph are quite immature, yet sufficient adult characters are developed to make identification fairly easy and certain. The characters that are especially helpful in the identification of specimens of the size described in the foregoing paragraph are: (a) The fin-ray counts (dorsal and anal each having 11 to 13 rays), which may be made accurately, ( b ) the DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 571 greater distance between the last two dorsal spines than between the others, (c) the long pectoral fins and ventral disk (reaching about two-thirds the distance from their bases to vent), ( d ) the general shape of the head and the small mouth (the maxillary scarcely reaching the eye). The head is still somewhat less strongly depressed than in the adult, the snout more pointed and the mouth more oblique, (e) the elongate body, winch remains more slender than in the adult (depth in the length to base of caudal in the adult about 5.0 to 6.25 times, in 13 mm young about 10.0 times), and {f) the characteristic color developed on some specimens, especially the oblique bar between the eye and the mouth and the wavy dusky bars on the caudal fin. However, a series of about five slightly elongate dark spots along the middle of the side, the last one being situated on the base of the caudal, which are quite characteristic of the adult, are undeveloped in the 13 mm specimens at hand. Gobionellus boleosoma, at about 13 mm in length, differs from Gobiosoma bosci and gmsburgi, and Microgobius holmesi, of similar length, very prominently in the much more slender body, as well as in the smaller mouth, and in the long space between the last two dorsal spines. DISTRIBUTION OF THE YOUNG The young were taken in tow nets over nearly the entire area in which collections were made, including Beaufort Harbor, the adjoining sounds and estuaries, and off Beaufort Inlet to Cape Lookout and at stations as far as 13 to 15 miles offshore. They were taken 22 times at offshore stations and 17 times in the inside waters. The young appeared in surface hauls only 6 times and in bottom hauls 33 times, indicating that the young, like the adults, dwell cluefly on the bottom. We also have many specimens taken by the United States Fisheries schooner Grampus in 1917 from Florida to Texas. The distribution of the young as to time differs from that of the other common gobies at Beaufort in that some individuals, 13 mm and less in length, were taken throughout the year, while the other species are not present in collections made during the winter months. Very small larvae, under 5.0 mm in length, were taken from May to November. GROWTH The scarcity of the species and the long spawning season resulted in capturing comparatively few young, which vary widely in size. It is consequently impossible from the few specimens taken to determine definitely the rate of growth. The presence in the tow of larvae only about 8.0 mm in length during March and April, which evidently were hatched the previous summer or fall, suggests a slow rate of growth, at least during the winter months. Sexual maturity apparently is reached at a length of about 25 to 30 mm, but it is not known how old a fish is when it attains that length. GOBIONELLUS OCEANICUS (PALLAS). OCEAN GOBY In addition to the young of G. boleosoma described in the forgoing pages, at least one other species is represented. Most of the specimens of the second group are large enough to permit a fairly accurate count of the dorsal and anal rays which is about the same for each fin, namely, 14 or 15 (rarely 13). This number of rays suggests that the specimens either are oceaniciLs or hastatus. The adults of these species are separable by the difference in the number of scales in a lateral series (see p. 565). However, no scales are developed in the young at hand. Therefore, the specimens cannot be definitely identified at this time and are only tentatively referred to oceanicus. The specimens of the second group of Gobionellus differ from the first one, furthermore, in having a more slender body and in the somewhat more retarded 572 BULLETIN OF THE BUREAU OF FISHERIES development. For example, in specimens of boleosoma 10 mm long the spinous dorsal is equally as well or even better developed than in specimens of oceanicus (?) 15 mm long. Specimens of this group of Gobionellus, with the very slender body, range in length from 9.0 to 18 mm. If smaller ones are contained in the collection they were not recognized as different from boleosoma. Only 15 specimens were collected in the vicinity of Beaufort. In addition 23 specimens, taken by the United States Fisheries schooner Grampus off the eastern coast of Florida and in the Gulf of Mexico, are at hand. SPAWNING The time and place of spawning remain largely undetermined, as the eggs and very small larvae have not been taken. Since two of the common local species of gobies ( Gobiosoma bosci and Gobionellus boleosoma) are known to spawn in the usual habitat occupied by the adults, it seems reasonable to expect this species to do like- wise. However, the number of adults taken locally is too small to admit of a definite statement in regard to their habitat. The examples at hand were taken on rather muddy bottom, two in Newport River and one along the shores of Fivers Island. The young were collected at Beaufort over such a long period of time that it is impossi- ble to judge definitely when spawning takes plane, the specimens having been taken in February, April, August, September, October, November, and December. Since these young generally were taken with those of G. boleosoma, a species known to spawn throughout the summer, it seems probable that the spawning period of G. oceanicus may extend over the same period of time. DESCRIPTIONS OF THE YOUNG Specimen 9.0 mm long— Only one specimen of this, the smallest size recognized, is at hand. It was taken by the Grampus at latitude 27°39', longitude 83°36' on Janu- ary 24, 1917. This fish differs from boleosoma of the same size in the extremely slender body, the depth being contained in the standard length about 9.7 times (as compared with 7.3 times in boleosoma). The bases of the dorsal and anal rays are visible in part, but the rays themselves are almost wholly undeveloped. Twelve or 13 fulcra can be counted in each fin, the posterior ones being feebly and very probably in part undeveloped, whereas in boleosoma the rays are quite well formed at this size. Pec- toral fins are evident, although without definitely formed rays. The ventral disk, already well formed in boleosoma of this size, is not evident (fig. 71). Specimens 14 mm long. — The body remains very slender, the depth being contained in the standard length about 9.0 times. The soft rays of the dorsal and anal are now well developed and easily enumerated, each fin having 14 or 15 rays. However, generally only two or three spines have become visible in the first dorsal, this fin being scarcely as well developed at this size as it is in specimens of boleosoma only 10 mm long. The pectoral fins are well developed and have definite rays, but the ventral disk is rudimentary, whereas it is long and prominent in specimens of boleosoma when only 10 mm long (fig. 72). DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 573 Specimen 18 mm long. — Only one specimen of this size is at hand, which is the largest post larva in the collection. Development has progressed rather slowly and is about in the same stage as a boleosoma of a length of 10 to 12 mm. The body has become rather more robust, the depth being contained in the length to base of caudal about 8.0 times. The spinous dorsal is partly developed, five slender spines being present, whereas in the fully developed individuals the normal number is six. The ventral disk is quite well formed and reaches about half the distance from its base to the vent. This specimen, like the smaller ones, is void of color. The transparent air bladder with the dark membrane over it remains visible through the abdominal wall (fig. 73). DISTRIBUTION OF THE YOUNG The number of specimens taken is entirely too small to permit drawing a con- clusion in regard to the distribution of the young of this species. Eight of the 14 specimens from Beaufort were collected off Beaufort Inlet, while the others were taken in the harbor and a neighboring estuary. The 23 specimens collected by the Grampus were all taken offshore. The young, therefore, may be expected along the outer shores, as well as in inside waters. Only two specimens were taken in surface towings at Beaufort, all the others appearing in bottom hauls. Of the 23 specimens collected by the Grampus 21 were taken on the bottom. This information is missing for the other specimens. It seems probable, therefore, that the young may occur at any depth in the water inhabited, but that they are most commonly on the bottom. Nothing can be reported at this time concerning the rate of growth. FAMILY BLENNIIDAE. THE BLENNIES Three species of blenny, namely Hypsoblennius hentz, Hypleurochilus geminatus, and Chasmodes bosquianus are common on the coast of North Carolina. The develop- ment of the eggs of all these species has been studied, and also the development of the young of the first two named. The young of C. bosquianus, a species less common 574 BULLETIN OF THE BUREAU OF FISHERIES at Beaufort than the other two, remain unknown. A fourth species, Blennius stearnsi, was recorded from Beaufort by Radcliffe (1914) without comment. This species was not seen by us. The adults are not especially difficult to identify, yet care is required as the species superficially are not strikingly different. This is true especially of young adults. Accordingly the following key, embodying characters thought to be readily usable and dependable, is offered. KEY TO THE GENERA AND SPECIES a. Head short, deep; forehead very steep (nearly vertical) ; snout scarcely projecting. Mouth small; maxillary scarcely reaching middle of eye. Canine teeth wanting. P. 14, rarely 13 or 15; D. XII, 14 or 15; A. II, 16 Hypsoblennius hentz aa. Head somewhat longer, not quite as deep; forehead not very steep, strongly convex; snout projecting moderately. Mouth small; maxillary reaching only slightly past anterior margin of eye. Each jaw with a strong canine tooth posteriorly, near angle of mouth. P. 14; D. XI to XIII, 14 or 15; A. II, 16 or 17 Hypleurochilus geminatus. aaa. Head notably longer and not as deep; forehead not steep, rather gently convex; snout strongly projecting, very pointed. Mouth large; maxillary reaching to or past posterior margin of eye. Canine teeth wanting. P. 12, rarely 11; D. XI or XII, 18; A. II, 17 or 18 Chasmodes bosquianus. THE CHARACTERS OF THE EGGS AND NEWLY HATCHED YOUNG The eggs of the three species of blennies discussed in this report are not difficult to recognize. That is not true for the young, however, which are very similar in appearance. The distinguishing characters of the eggs and newly hatched larvae are shown in the parallel comparison which follows. The distinguishing characters of young Hypsoblennius hentz and Hypleurochilus geminatus taken in the tow are described in the text. Since young Chasmodes bosquianus were not taken in collections made in nature, our present knowledge of its development ends with the newly hatched larvae. DISTINGUISHING CHARACTERS EGGS Hypsoblennius hentz Moderately small, about 0.77 mm in diameter. Eggs with violet or old rose colored bod- ies (disappearing in advanced stage of development) and yellow oil globules in yolk. Larvae moderately small, average length about 2.7 mm. Myo- meres behind vent about 23. Black markings on abdomen (yolksac) generally scattered, usually not especially concen- trated at upper edge of ab- dominal mass. Lower two-thirds or so of inner surface of pectoral fin mem- branes with black ehromato- phores. An elongate branching black spot under auditory vesicle. Hypleurochilus geminatus Small, about 0.69 mm in diam- eter. Eggs with purple spots (disappearing in advanced stage of development) and bright golden yellow to orange oil globules in yolk. NEWLY HATCHED YOUNG Larvae small, average length about 2.4 mm. Myomeres behind vent about 24. Black markings mostly concen- trated at upper margin of ab- dominal mass. Pectoral fin membranes at most with only a few black chroma- tophores at base. An elongate branching black spot under auditory vesicle. Chasmodes bosquianus Large, about 1.04 mm in diam- eter. Eggs with pale yellow oil globules, never with violet or purple bodies in yolk. Larvae large, average length about 3.66 mm. Myomeres behind vent about 28. Black markings mostly concen- trated at upper margin of ab- dominal mass. Lower two-thirds or so of pec- toral fin membranes with black chromatophores. No black under auditory vesicle. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 575 A COMPARISON OF THE EGGS AND YOUNG OF SOME AMERICAN AND EUROPEAN BLENNIES The eggs of the three species of blennies from North Carolina, forming the basis for the present report, all have an adhesive disk or foot by which they become firmly attached to objects in the water at the time they are laid, remaining attached through- out the period of incubation. The eggs of the European blennies, Blennius pholis, B. ocellaris, and B. gattorugine, all have similar organs of attachment (Lebour, 1927). In all these species, both American and European, the eggs are laid in a single layer. The adhesive organs generally, if not always, have a greater diameter than the eggs and keep them from touching each other. In another European blenny, B. montagui, the eggs are described (Guitel, 1893) as having a number of glutinous threads which attach them to the under side of stones. The eggs are said to press against each other, although presumably laid in a single layer. Such distantly related forms as Pholis gunnellus (Ehrenbaum, 1909), Anoplarchus purpurescens (Schultz and De Lacy, 1932) and Heterostichus rostratus (Barnhart, 1932), formerly assigned to the family Blenniidae, but now referred to separate and distinct families, also have eggs which adhere. The adhesive organ, if any, for the first mentioned species are not described, it merely being stated that the eggs adhere in clusters. The egg of the third species has a number of adhesive threads like the one of Blennius galerita. The eggs of the two last-mentioned species, therefore, resemble those of the silversides (Menidia) in the structure of their adhesive organs. The eggs may be attached to mollusk shells, particularly to the inner surface of the valves of empty oyster shells, as in Hypsoblennius hentz, Chasmocles bosquianus, and Blennius ocellaris. Or they may be attached to rocks in crevices or to the under side of overhanging rocks, as in B. pholis, B. gattorugine, and B. galerita. Again, they may be laid in such places as the hollow of an ox bone or bottle, as in B. ocellaris, or in clusters loosely attached to stones, as in the distantly related Anoplarchus purpurescens of the Pacific coast. The eggs in one “nest”, if laid in a single layer, often cover several square inches of surface. The male was observed guarding the nest in most of the species studied by various investigators, as in II. hentz, C. bosquianus, B. pholis, B. ocellaris, B. sphinx, B. gattorugine, B. montagui, and Clinus argentus. Gudger (1927) reported that both sexes guard the eggs of Pholis gunnellus. Finally Shultz and De Lacy (1932) reported that the female guards the eggs of the Pacific coast blenny A. purpurescens. The eggs of the three species of blenny from North Carolina, constituting the subject of a part of this paper, are all slightly flattened at the place of attachment, as stated in the descriptions of the eggs in the text. The eggs of two European species, namely, Blennius pholis and B. gattorugine, are described as decidedly flattened, the egg of the first-mentioned species being a little more than three-fourths of a sphere, and that of the other one only slightly more than half a sphere. The eggs of another European species, B. ocellaris, having a smaller adhesive organ, is described as nearly spherical. It is pointed out in the text (pp. 579 and 592) that the eggs of two species of blennies from North Carolina have yolk containing brightly colored bodies. The eggs of Hypleurochilus geminatus have yolk with purple bodies, and those of Hypsoblennius hentz violet to old-rose colored ones when first spawned. The colored bodies gradu- ally lose their outline as development of the egg progresses, and the color becomes diffuse, generally disappearing before hatching. Bright colors in the yolk seem to be usual in the eggs of European blennies also. The egg of Blennius ocellaris, B. gattorugine, and B. pholis are all said to have pink, red, or purple yolk, though no 576 BULLETIN OF THE BUREAU OF FISHERIES definite spots or bodies are mentioned. Chasmodes bosquiannus seems to be the only- true blenny (family Blenniidae according to Jordan 1923) studied to date which has eggs containing yolk without pink, red, or purple color. The eggs of the North Carolina blennies all contain many oil globules. The oil spheres are yellow, being especially bright golden yellow in Hypleurochilus geminatus. In the European species oil globules are mentioned only in the eggs of Blennius ocellaris. The newly hatched fish of the European species Blennius ocellaris, B. gattorugine, and B. pholis are respectively 4.4, 4.9, and 5.4 mm long, and therefore larger than those of the American species, Hypsoblennius hentz, Hypleurochilus geminatus, and Chasmodes bosquianus, which are respectively 2.7, 2.4, and 3.6 mm long. Somewhat larger young would be expected as the eggs of the European species are larger than those of the American ones. The greater axis of the eggs of the European species, in the order named, are 1.2, 1.6, and 2.0 mm, whereas those of the American species, respectively, are only 0.77, 0.69, and 1.04 mm. Furthermore, the European species grow larger than the American ones. The former, in the order named, reach a length of about 175, 225 and 150 mm, whereas the latter attain a length, respectively, of only about 100, 75, and 90 mm. It is understood, of course, that the size of a fish is no criterion relative to the size of the egg it produces. However, in this instance the larger European species evidently do produce larger eggs than the smaller American ones. Although the newly batched young of the European species are larger than those of the American ones, as pointed out in the preceding paragraph, they are all strik- ingly similar in general appearance throughout the larval stages. The newly hatched larvae are fairly stocky anteriorly and have rather long slender tails, the vent being situated far in advance of midbody length. The pectoral fin membranes are com- paratively large and generally more or less spotted with black. Usually black is present also on the abdomen which most often is concentrated on the side of the fish along the upper edge of the abdominal mass. Short black cross lines on the ventral edge of the tail may be present on only a few to several myomeres or on all the caudal segments. In the older larvae the tail becomes proportionately shorter and heavier and the black on the sides and on the pectorals tends to become more prominent. In the postlarval stages the pectoral fins generally are proportionately much longer than in adults, and the caudal fin, which is round in the adult, tends to be slightly concave. Such a development of the caudal fin must be considered rather unusual, though a similar evolution has been observed in the gobies. (For descriptions and figures of the eggs and the young of the European blennies, Blennius ocellaris, B. gattorugine, and B. pholis, see Cunningham, 1889; Ford, 1922; and Lebour, 1927.) HYPSOBLENNIUS HENTZ (LeSUEUR). SPOTTED SEAWEED FISH Hypsoblennius hentz is common, but not abundant at Beaufort, N. C., and is known to range from Chesapeake Bay to Florida. It is recognized chiefly by its very steep forehead; small, horizontal mouth, the maxillary scarcely reaching under the middle of the eye; by the absence of canine teeth; the small gill opening; the broad pectoral, with 14, rarely 13 or 15 rays; and the moderately long and low dorsal and anal fins, the former consisting of 12 spines and 14 or 15 soft rays and the latter of 2 spines and 16 soft rays. The males appear to grow larger than the females (largest male at hand 104 and the largest female 84 mm long) and the males have a much longer tentacle over DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 577 the eye. Although the tentacle is variable in length among individuals of the same sex, it is rarely as long as the eye in adult females, whereas it always exceeds the length of the eye in adult males. Males differ from the females in external structure, furthermore, in having fleshy expansions or hoods, opening forward, attached to the two anal spines, and the fin itself is preceded by a low elliptical membranous hood, opening backward. Females have a distinct genital papilla, at least during the breeding season, which is not evident in the males (fig. 74). The shallow water areas with rather hard, often somewhat shelly, bottom sup- porting growths of plants, sponges, ascidians, hy droids, etc., are the common summer habitat of the adults. The shallow areas are deserted during the winter when speci- mens occasionally are taken in the deeper channels and in holes, where the species also occurs sparingly during the summer. It is a game little fish and like its relative, Chasmodes bosguianus, it fights when handled. It will seize the skin (its mouth being too small to catch more than the skin) of a man’s hand and hold on bulldog fashion, allowing itself to be lifted by its grasp. However, its jaws are not strong enough to inflict a wound. Figure 74. — Hypsoblennius hentz. Adult male 96 mm long. Note membranous expansions attached to the anal spines. Color assimilation is well developed in this blenny. It is hardy, stands handling, and endures confinement in small aquaria very well and, therefore, it constitutes a fairly favorable subject for the study of its reactions to various color stimuli. The species no doubt is preyed upon to a limited extent by various predatory fishes. It is not abundant enough locally to be of much importance even as a forage fish, and of course it is too small to be of direct commercial use, as 100 mm (4 inches) is near the maximum size attained. The figures of the developing egg and the newly hatched larva are based on living material. The other illustrations were prepared from preserved specimens. SPAWNING Eggs of several sizes are present in the ovary at one time, just as in the other local species of blenny, suggesting a long spawning season, as well as repeated spawn- ing. This expectation is substantiated by the presence of fry less than 5.0 mm in length in the tow from May 13 (1930) to September 13 (1927). The young of this species were never as abundant in collections as those of HypLeurochilus geminatus. They were taken in fair numbers, however, from about the middle of May to the end 578 BULLETIN OF THE BUREAU OF FISHERIES of August. It may be concluded, therefore, that the spawning period extends from May to August. The eggs were seen first on August 25, 1927, when a female held in a battery jar spawned. Since no male was at hand the eggs could not be fertilized. Several “nests”, each containing many eggs, were taken from May 31 to June 27, 1932. From this material it was possible to study the embryology in detail. This blenny does not make a nest in the true sense of the word. However, it uses empty oyster shells (possibly also clam and scallop shells) with the hinge in- tact, from which the oysters probably have not been removed very long and which are still clean and white within. Therein the eggs are deposited, and they become firmly attached by means of an adhesive disk. In several instances, nearly the entire inner surface of both valves of the oyster shell was covered with eggs. Occa- sionally only a part of each valve was occupied by the eggs. It seems probable that in these last mentioned instances the nests were not completed, and that more eggs would have been deposited. Nests were found only on a natural oyster reef at the west end of Pivers Island, at or near the usual low tide line. It is evident from the difference in the development of the eggs in a nest that they are not all spawned at the same time. The difference in development may range from an early cleavage stage to an advanced embryonic stage, suggesting that the eggs are laid over a period of several days. In general, the eggs near the hinge of the oyster shell are furtherest advanced, whereas those most distant from the hinge of each valve show the least development. The eggs are in a single layer in the nest, not always in definite rows, and are well separated by the adhesive disks which have a greater diameter than the eggs. It is estimated that some of the larger nests found contained as many as 3,750 eggs. The eggs are so firmly attached that they can be removed without injury only by cutting the adhesive disk close to the oyster shell with a sharp instrument. It is not known whether all the eggs in one nest are the product of one female, although this seems quite possible since all the eggs within an ovary evidentally do not mature at one time, as already stated. It is possible, therefore, that a female may go to the same nest several days in succession to spawn.7 The eggs probably always are guarded by a male. The foregoing statement is made notwithstanding the fact that a few nests were found with which no males were seen. On the other hand, a fish was seen to escape from a nest in a few instances, and several nests were taken with the male within the valves of the oyster shell con- stituting the nest. It is assumed in those instances when no males were seen that they escaped unnoticed. The male stays within the oyster shell in taking care of the eggs. In case the shell is shorter than the fish it bends the tail forward to get within the shell, for it allows only the snout and eyes to protrude. A decided difference in the temperament of different males was noticed. It already has been shown that some males fled when someone approached. Others stayed with the nest and allowed themselves to be picked up (by hand) with the oyster shell containing the eggs. Only one male, upon being transferred from its native habitat to the aquarium, reoccnpied his nest almost immediately, although others, after being in confinement for some time would even occupy empty shells. In a few instances males failed to return to their nests in 7 Guitel (1893) has reported that among specimens of Blennius montagui=B. galerita, from the coasts of France, kept in a tank made of a small boat, in which their natural habitat was reproduced as nearly as possible, several females laid eggs in one nest situ- ated on the under side of an overhanging stone. The eggs were all fertilized and guarded by the same male. It cannot be stated of course, that this procedure would obtain in nature. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 579 nature after being disturbed. One male transferred to the aquarium devoured the eggs, instead of taking care of them. The single male which reoccupied his nest in the aquarium not only stayed in the oyster shell until the eggs were hatched (not- withstanding frequent disturbances during the incubation period for removing eggs for study), but also for several weeks afterwards or until removed from the tank. During the incubation period, as well as afterwards, he came out only to feed on bits of oyster and fish that were supplied, and even retreated between bites. The male undoubtedly drives away intruders, for it was noticed that the eggs in nests deserted by the parent fish were destroyed very soon. The chief enemy noticed was the flat mud crab, Eurypanopeus depressus (Smith).8 Specimens of this crab were taken with three deserted nests. The male, however, appears to have another function, namely that of keeping the eggs clean and in healthy condition. Just how this is accomplished is not evident. It was not noticed that he fanned the eggs especially, for he seemed to lie quietly within the oyster shell.9 Yet, the eggs in a nest cared for by a male, and held in a tank with running water almost all hatched, whereas the eggs treated identically, but without a male attendant all died in ad- vanced embyronic stages. A very small number of eggs in one deserted nest was hatched by providing special treatment. That is, in addition to keeping the nest in a tank with running water, the eggs were washed vigorously once or twice each day by playing a jet of water directly on the eggs, and by rushing the nest through water rapidly. A few eggs removed from a nest were hatched in standing water (changed twice daily) in a glass bov/1. Eggs in deserted nests in tanks became infested with hydroids and copepods, which caused death before hatching. No infestations were noticed in eggs guarded by the male fish. Spawning apparently takes place early in the morning. This conclusion is arrived at from the fact that eggs in early cell division stages were present only in nests taken before 9 o’clock in the morning. All eggs collected even as late as 11 o’clock in the morning already had passed the early cell division stages and those taken during the afternoon had progressed correspondingly further in development. DESCRIPTIONS OF THE EGGS AND YOUNG Description of the eggs. — The eggs of Hypsoblennius hentz are slightly flattened next to the adhesive disk or foot which attaches them to oyster shells, as already explained. The foot and the slight depression in the contour of the egg at the place of attachment are shown in only two of the accompanying drawings (figs. 77 and 83). The greater axis in 11 eggs measured varied from 0.72 to 0.8 mm, the average being 0.769 mm. The smaller axis which is difficult to measure accurately because the opaque foot obscures the outline of the egg, varied in four specimens from about 0.64 to 0.68 mm. The eggs, as seen with the unaided eye, if still in rather early developmental stages, are pinkish in color. Under magnification it becomes evident that the color is within the yolk and in the form of spherical or more or less elongate bodies. The longest diameter of the latter apparently is always perpendicular to the plane to which the egg adheres. These bodies, as seen under magnification, are violet to old rose in color. They are variable in size within the same egg, as well as in shape and number in dif- ferent eggs. They lie at various depths within the yolk, and therefore it is necessary to refocus the microscope to see all of them. The variation in number in different 8 The writer is indebted to Dr. Mary J. Rathbun of the U. S. National Museum for the identification. • Guitel (1893) stated that the male of Blennius montagui=B. galerita, a European species, does fan the eggs and that he will remove with his mouth any foreign object which may enter the nest. 580 BULLETIN OP THE BUREAU OF FISHERIES eggs apparently ranges from about 12 to 24. During development these colored bodies become less and less definite in outline, and in advanced embryonic stages the color becomes paler and diffuse, often disappearing entirely several days before the egg hatches. Golden yellow oil globules are also present. These spheres are equally as variable in size within an egg and in number in different eggs as the old- rose colored bodies. In general they are somewhat concentrated near the blastoderm. The oil globules persist in part at least until the egg hatches or even in the small yolksac attached to the newly hatched fish. In the accompanying drawings the old rose colored bodies are shaded, while the oil globules are unshaded. Some of the variations in the shape of the colored bodies are shown in the illustrations. The total number of colored bodies and oil globules is not shown. Those indicated are the ones which came into focus under the microscope at one level. The egg, furthermore, has a large central body, apparently denser in texture than the rest of the egg, grayish in color like the adhesive disk, and quite opaque. This body disappears in the advanced embryonic stage. A similar central opaque body is present also in the eggs of the other two species of blenny discussed elsewhere in this paper. The entire egg is moderately opaque, becoming more so as development proceeds. Fair perception is obtainable, however, in recently spawned eggs if viewed in a plane parallel with the surface to which they are attached. In the opposite direction the dirty -gray opaque adhesive foot, which cannot be detached without injury to the egg, and the opaque central body obscure vision. The yolk is granular in appearance. The egg membrane has deep lines and elevations, suggesting rugged eroded land. This sculpture on the egg case is not shown in the accompanying illustrations. Segmentation and the development oj the embryo— The eggs forming the bases for the present account were taken in nature. The exact time of fertilization is not known. Therefore, the length of the period intervening between fertilization and the beginning of cell cleavage cannot be stated definitely. The earliest cell division stage found, namely four cells, occurred in two nests taken at 8:30 o’clock in the morning. These eggs probably had been laid 2 hours or so before the nests were found. This tentative conclusion is based on the results obtained with Hypleurochilus geminatus (p. 593). In that species segmentation started about 2 hours after fertilization. It seems reason- able to expect that the intervening time in these related species would be about equal at the nearly identical temperatures which prevailed (28° to 28° C.). The blastodisc is apparently always situated next to the adhesive disk. This position of the blastodisc makes it difficult to observe cell division, as the opaque adhesive disk below and the opaque central body above it obscure vision. Fortunately, eggs in the early stages are more transparent than those in the more advanced stages. Consequently, it was possible to see the cells, even though dimly, through the mass of the egg (fig. 75). In a lateral view the cells could be seen more definitely. The first blastomeres apparently are about equal in size and the second cleavage cuts the blas- todisc at right angles to the first. The perivitelline space is comparatively large at the positive pole and very small or wanting at the negative one (fig. 76). Segmentation proceeds rapidly, the 8- and the 16-cell stages (figs. 76 and 77) fol- lowing the 4-cell one at intervals of about 30 minutes each at a water temperature of about 26° C. As development proceeds the egg becomes more granular, and it becomes more and more difficult to see exactly what is taking place. While the blastoderm no doubt is dome-shaped, as usual in teleosts, it cannot be seen because of the opaqueness of the yolk (fig. 78). An advanced cleavage stage is reached in about DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 581 8 hours at a water temperature of 25° to 27° C. (fig. 79). In about 24 hours the germ ring becomes evident (fig. 80) and in 48 hours the embryo already is well differentiated. Generally only a part of it is visible from one viewpoint, as any part lying underneath Figtjbe 75. — Hypsoblennius hentz. From egg in 4-cell stage, about 2 to 3 hours after fertilization. (Drawn by Nell Henry.) Figure 77. — Hypsoblennius hentz. From egg in about 16-cell stage, lateral view; about 3)4 hours after fertilization. Adhe- sive disk shown at upper margin of egg. Note round form of shaded bodies in yolk. (Drawn by Nell Henry.) Figure 76. — Hypsoblennius hentz. From egg in 8-cell stage, lateral view about 3 hours after fertilization. A slight de- pression in the egg at place of attachment opposite the blas- todisc is not shown. Note elongate shape of the shaded bodies. (Drawn by Nell Henry.) Figure 78. — Hyposblennius hentz. From egg in a moderately advanced cleavage stage; probably about 6 hours after fertil- ization. Only that part of the blastoderm projecting above the yolk is shown as the rest is obscured by the opaqueness of the egg. (Drawn by Nell Henry.) the now very dense yolk cannot be seen (fig. 81). The tail was curved underneath the yolk where it could not be seen and for that reason was not shown in figure 81. The development proceeds slowly for such a small egg after the embryo once is well formed. The embryo may be expected to extend three-fourths of the distance around the egg in about 3 days at a water temperature of 25° to 27° C. Somites are evident in at least a part of the body, the eyes are well formed and punctuated with 582 BULLETIN OF THE BUREAU OF FISHERIES dark dots, and circulation is established, although the blood flows slowly. The heart is situated under the anterior tip of the head. A large artery courses through the ventral part of the embryo. It recurves rather sharply in the caudal portion where it leaves the embryo. This vessel then divides and several branches course over the yolk, reuniting just before reaching the heart. No return circulation is established in the embryo. Large dark blotches with irregular outlines, sometimes merely Figure 79. — Hypsoblennius hentz. From egg in an advanced cleavage stage; probably about 8 hours after fertilization. (Drawn by Nell Henry.) Figure 81.— Hypsoblennius hentz. From egg with an early embryo; tail underneath the opaque yolk; about 2 days after fertilization. (Drawn by Nell Henry.) Figure 80. — Hypsoblennius hentz. From egg showing blasto- derm growing around egg; about 1 day after fertilization. (Drawn by Nell Henry.) Figure 82. — Hypsoblennius hentz. From egg with well- formed embryo, showing blood vessels. Arrows indicate direction of flow of blood. About 3 days after fertilization. (Drawn by Nell Henry.) branching blotches resembling crows feet, are now present on the surface of the yolk (fig. 82). In about 6 days the embryo encircles the egg, the tip of the tail reaching to or past the head. The eyes are very large and black with a greenish sheen. Heart action is very brisk, the beats following each other so rapidly that it is difficult to enumerate them accurately. The number of beats probably is close to 200 per minute. A return circulation is now established in the embryo. Large vessels still course over the yolk DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 583 and these together with the large vein in the embryo all pour their contents into the heart which has somewhat the appearance of a pit. Corpuscles are plainly evident, and from their rapid progress it is obvious that circulation is brisk. The old-rose colored bodies previously present have disappeared. However, some eggs retain a diffuse pinkish cast in the yolk near the heart of the embryo. The embryo is capable of much movement. The tail is free and is switched frequently. The embryo is able to turn within the egg membrane, carrying the yolk with it in its movements, (fig. 83). Progress in the development after about the sixth day of incubation seems particu- larly slow. The embryo increases little in length and the yolk is absorbed very slowly. Dark markings on the yolk, which tend to decrease in size after about the fourth or fifth day of incubation, generally disappear entirely a day or two before hatching. In the meantime black markings become evident on the embryo. A rather large branching blotch is present on the head between the eyes, numerous black chroma- tophores also appear on the comparatively large pectorals, and short branching cross lines mark the myomeres along the ventral surface of the caudal region of the embryo. Just before hatching the egg becomes some- what distorted, the egg membrane being pushed out somewhat at the head of the embryo. Eggs taken on May 31, 1932, which were in several different stages of develop- ment, ranging at the time of collection from a rather advanced cleavage stage to a stage in winch the embryo already was well differ- entiated, hatched from June 8 to 12. The temperature of the water during this tune varied from about 25° to 27° C. The eggs in a nest taken June 16, 1932, which ranged in development about equally as much as those taken on May 31, hatched from June 24 to 26. The temperature of the wrater during this period varied from about 24.5° to 27° C. Assuming that the last eggs hatched in each nest were those which were in an advanced cleavage stage when taken, and that these eggs were laid on the day of collection (concerning which there can be little or no doubt), the incubation period has a duration, at the tempera- tures stated, of about 10 to 12 days. The incubation period in this species, therefore, is longer than in Hypleurochilus geminatus (see pp. 596 and 610), and about the same as in Chasmodes bosquianus. Hatching, like spawning, apparently takes place early in the morning. At the time of hatching, the yolk was almost wholly absorbed and the young fish generally died by the evening of the day on which they were hatched. However, for 4 days in succession a new lot was present each morning. Several efforts were made to keep the fish alive and to induce them to feed and to grow. Some were kept in a tank with running water, others were transferred to shallow glass bowls with standing sea water. The lots in running water were not fed, those in standing water in part were offered towings and in part very finely minced oyster. However, none lived more than 2 154970— 3S 6 Figure 83 .—I-Iypsoblennius hentz. From egg with large embryo: about 6 days after fertilization. H, heart. Arrows show direction of blood flow in the larger vessels. (Drawn by Nell Henry.) 584 BULLETIN OF THE BUREAU OF FISHERIES days. The fish presumably did not feed, but it is quite unlikely that they died of starvation as a bit of the yolksac remained even in those individuals that lived longest. Newly hatched fish. — The newly hatched fish range in length from about 2.6 to 2.8 mm. The yolk is nearly all absorbed at hatching. The fish are robust anteriorly, with a broad depressed head. The tail is long and slender, with the vent situated much in advance of midbody length; distance from snout to vent about 1.0 mm, from vent to tip of tail, without finfold, about 1.5 mm. The snout is short and very blunt; the eye is large, having a diameter of about a quarter of a millimeter; the mouth is large, the gape reaching to or past the middle of eye. Large pectorals with suggestions of rays are present. The body is fairly transparent. Consequently the outline of the brain and the circulation can be seen rather definitely. The aorta may be seen close to the notochord, turning upon itself about an eye’s diameter from the tip of the tail to form the caudal vein. Heart action is too rapid to permit accurate enu- meration of the beats. However, the number of beats probably is close to 230 per minute. About 28 to 30 myomeres may be enumerated, being indefinite in advance of the vent (only 3 or 4 visible) and again toward the tip of the tail. The vertebrae count in the adult is 33, there being 9 body and 24 caudal ones. The number of more or less definitely outlined myomeres in the newly hatched fish, therefore, is not far below the number of vertebrae in the adult. Figure 84. — Ilypsoblennius hentz. From newly hatched fish. Length of live specimen 2.6 mm. AV, auditory vesicle. (Drawn by Nell Henry.) Several dark markings are present on the newly hatched fish which correspond for the most part with those already present in the advanced embryo as described else- where. The eye is very dark with a greenish sheen above the pupil; an irregularly outlined dark spot is present on the head between the anterior part of the eyes, or in some specimens several black chromatophores are distributed over the snout to the interorbital ; generally a blackish blotch with branches is present at the auditory vesicle ; many black chromatophores (or in some specimens in part solid black) are present on the abdominal region; and the ventral side of the tail is marked by short black branching cross fines. The large pectorals are marked on the inner surface, with dark chromatophores. The dark markings often are present only on the basal two- thirds, although sometimes they may cover nearly the entire inner surface and extend to the margin of the fin (fig. 84). Comparatively few of the numerous fish hatched lived as long as 2 days. At 2 days after hatching the fish apparently had become more slender and had increased in length only slightly, being 1 .8 to 1 .9 mm long. The yolksac was almost all absorbed. Only minor changes in color had taken place. The black chromatophores on the abdominal region had become more concentrated along the side near the upper boundary of the abdomen in the form of an indefinitely outlined oblique band extend- ing from near the eye to the vent. Only a few separate black branched markings remained on the ventral surface of the abdomen (yolksac) where a shade of yellow DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 585 had appeared. Other color markings remained essentially as in the newly hatched larvae. The black color along the upper margin of the abdomen, described in the foregoing lines; the short black lines on the ventral surface of the tail, which remain as in the newly hatched fish; and the black chromatophores on the pectoral fins are very useful characters in identifying the larvae hatched in the laboratory with others taken in towing and described in the following pages. Specimens 1.5 to 2.0 mm long. — Although the newly hatched fish when alive, or before preservation, were around 2.7 mm in length, the smallest specimens taken in the tow are only 1.5 to about 2.0 mm long. These small larvae quite surely belong to the species under discussion. Their size does not exclude them, as young tender fish generally shrink greatly when preserved in formalin and alcohol. Young hatched in the laboratory, preserved when less than a day old in 65 percent alcohol, for example, decreased in length from about 2.7 to 2.0 mm. The specimens taken in the tow were killed in formalin and later transferred to about 75 percent alcohol. Some of these specimens, although 2.0 mm and less in length, evidently are several days old, as showm by the more advanced development. The older fish are less robust anteriorly, the pectoral fins are more elongate and less broadly rounded, and suggestions of rays are present. Indications of rays also are appearing in the vertical finfold around the tip of the tail. The color on the abdomen has become more concentrated along the upper margin of the abdominal mass, the black spots on the inner surface of the pec- Fiouee 85. — Ilypaobtennius hentz. From a preserved specimen 3.0 mm long. torals are prominent, and the cross lines on the ventral edge of the caudal region remain as in younger fish. Specimens 2.5 to S.O mm long. — The head and trunk are robust, the head being about two-thirds as broad as deep. The head and trunk have become longer in proportion to the tail, the distance from tip of snout to vent being contained about 2.4 times in the total length without the caudal finfold. The snout is very short and blunt, scarcely longer than the pupil, the forehead is very steep, and the mouth is slightly inferior, moderately oblique, with the tip of the lower jaw a little below the level of the middle of eye. Three minute preopercular spines are evident in some specimens. The vertical finfold remains continuous, with indications of rays pos- teriorly. The pectoral fins are long and rather narrow, with definite rays, and about three-fourths as long as the head. The most prominent color marking is an oblique black bar extending from the axile of the pectoral to the ventral outline just in front of the somewhat protruding hindgut. Several dark dots are present on the ventral surface in advance of the vent, a distinct dark bar crosses the forehead between the eyes, and generally several chromatophores are present on the upper surface of the head and nape. A row of small, vertically elongate, dark spots is situated on the ven- tral outline of the tail. The most important color markings for the purpose of identi- fication are dark dots, covering most of the pectoral fins, which extend to the tips of at least some of the rays (fig. 85). 586 BULLETIN OF THE BUREAU OF FISHERIES Specimens 4-0 to 4-5 mm long. — The head and trunk remain rather robust, although less so than in somewhat smaller specimens. The caudal portion of the body is moderately deep, strongly compressed, and scarcely longer than the head and trunk, the vent being situated at about midbody length, exclusive of the caudal fin. The head is deep and rather broad, the interorbital space being scarcely narrower than the eye. The snout is very short and round, projecting scarcely half the diameter of the orbit in front of the eye. The mouth is placed low, slightly inferior, oblique, the tip of the lower jaw being only a little above the level of the lower margin of the eye. The eye is placed low, that is, nearer to the ventral than the dorsal outline of the head. Fin rays are only partly developed in the dorsal and anal fins, but more fully in the caudal fin which is round in outline. The notochord is bent upward at the base of the fin, as usual in larval teleosts at about this stage of development. Ventral fins are not evident. The pectorals, however, are long and rather narrow, and scarcely shorter than the head. A few obscure dark markings generally are present on the ventral surface of the chest and abdomen ; a dark band extends across the fore- head between the eyes; the occipital surface of the head has one to several dark dots and a large median black spot is present at the nape. An oblique black band extends from the axile of the pectoral nearly to the vent; the long pectoral fin, exclusive of two or three of the upper rays, is densely dotted with black; and a row of very small black points begins a short distance behind the vent and extends to the base of the caudal fin (fig. 86). Specimens 5.0 to 6.0 mm long. — The body has continued to grow deeper and somewhat more compressed since a length of about 4.0 to 4.5 mm was attained. The head especially is deep and short; the snout remains very short and blunt, being scarcely more than half as long as the eye. A rather definite bony ridge is evident now over and in front of the orbit, making the interorbital space quite flat and fully as broad as the eye. The position of the mouth remains low and is slightly inferior, the tip of the lower jaw being only a little above the level of the lower margin of the eye. Five preopercular spines are now visible. Advancement in the development of rays in the dorsal and anal fins is not pronounced. The caudal fin, however, has grown proportionately longer and remains round. Ventral fins now are evident as mere tufts of membrane. The pectoral fins are fully equal to the length of the head and, exclusive of the upper rays, are dotted with black as in the younger fish, and the oblique, dark bar behind them remains as described in smaller specimens. A few indefinite dark spots occur on the chest and sides of the head, and several rather definite dark chromatophores usually are present on the occipital surface of the head and nape. A row of very small dark dots on the ventral outline of the tail, or base of the anal, described in smaller specimens, remains (fig. 87). DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 587 Specimens 8.0 to 10 mm long. — The body is rather deep and strongly compressed, the depth being contained about 3.4 to 4.0 times in the length without the caudal fin. The head is deep, the snout remains excessively short, as in smaller specimens, and the forehead is very steep. The snout projects in front of the orbit a distance scarcely equal to half the diameter of the eye. The mouth is small, placed very low, almost horizontal, and terminal to slightly inferior. The tip of the lower jaw is on, or a little below, the level of the lower margin of the eye, and the maxillary reaches to, or slightly past, the anterior margin of the pupil. The interorbital remains quite flat, Figure 87.— I-Iypsoblennius hentz. From a young fish 6.2 mm long. with a prominent bony ridge over and in front of the eye. The preopercular spines, while well developed in specimens 8.0 mm long, are proportionately longer in specimens 10 mm in length. The dorsal and anal fins are quite fully developed and a fairly accurate count of the rays can be made; the caudal fin has an almost straight margin and is about as long as the head; the ventral fins are well developed and are long and slender, being nearly as long as the head without the snout; and the pectoral fins are large, being nearly or quite as long as the head. The lower surface of the head and chest is variously dotted with black, generally with a few definite dark spots slightly behind the articulation of the lower jaw, also with a pair of dark spots a short distance in advance of the ventral fins, and another pair in the axiles of the ventrals. The side of the head has a few indefinite spots or blotches and the upper surface of the head, that is from the interorbital backward, bears brownish spots with dark center and dark outline. The pectoral fin is almost wholly black in some specimens; in others two to four of the upper rays are pale, while the rest of the fin is black, and the oblique dark bar behind the pectoral, very prominent in smaller fish, has become quite obscure in 10-mm fish. A row of fine dark points along the base of the anal is present in some specimens, but not evident in others (fig. 88). 588 BULLETIN OF THE BUREAU OF FISHERIES Specimens about 12 mm long. — The differences between fish 10 and 12 mm long are not pronounced. However, the body in the larger fish is considerably more robust, especially anteriorly, the depth as in the smaller specimens being contained 3.4 to 4.0 times in the length without the caudal fin. The bony ridge over and in advance of the eye is quite as prominent as in the smaller fish. The forehead re- mains very steep to vertical and projects slightly beyond the low, almost horizontal mouth. The gape of the mouth is now wholly below the level of the lower margin of the mouth as in the adult. The preopercular spines have continued to increase in proportionate length, the one situated at the lower posterior angle having become much larger than the other, being equal to the length of the eye in one specimen, but somewhat shorter in others. It is probable that the preopercular spines, which are not present in the adult, reach their greatest development at this stage and that they gradually decrease in size in larger fish. Specimens of the proper sizes for a study of the recession of these spines, however, are not at hand. A small fleshy tentacle is now visible over the eye for the first time. Although nearly or quite as long as the eye in the adult it is scarcely as long as the pupil in fish 12 mm long. A definite notch between the spinous and soft portions of the dorsal fin is present, as in the adult. No pronounced development in pigmentation has taken place since a length of 10 mm was attained. However, dark areas and spots about the head have increased somewhat in size and number in at least some specimens. A row of very small black dots still persists on the base of the anal. The dots are not evenly spaced and not definitely on each interradial membrane, although some variation in this respect exists (fig. 89). Figure 89 .—Hypsoblennius hentz. From a young fish 12 mm long. Unfortunately no specimens ranging from about 13 to 24 mm in length are at hand, and therefore, a complete picture of the development of the late juvenile stages cannot be given at this time. Specimens 25 mm long are “young adults” and have virtually all the characters of mature fish. In such specimens the prominent bony ridge over each eye, very characteristic of the young, has disappeared entirely; the fleshy tentacle over the eye is nearly or quite as long as the orbit, the preopercular spines no longer are evident; the caudal fin is round; and pigmentation is complete and similar to that of fully matured fish. Although the series is not complete, the largest young (12 mm long) at hand have developed sufficient adult characters to make identification certain. The extremely steep forehead in the young and the size, shape, and position of the mouth are quite characteristic of the adults and unlike the other local species of blennies. Furthermore, the fins are rather fully developed and the shape and number of their rays agree with those of the adult. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 589 DISTRIBUTION OF THE YOUNG The fry were taken in to wings in outside waters 39 times and in inside waters 28 times. No definite record of the number of hauls made was kept. It is probable, however, that nearly an equal number of towings was made in the inside and outside waters. Therefore, the indications are that the fry are somewhat more common off Beaufort Inlet than in Beaufort Harbor and adjacent sounds and estuaries. The fry were taken in surface towings 48 times and in bottom ones only 16 times. Although a considerably larger number of surface than bottom hauls was made, the discrepancy in the number of hauls certainly is not great enough to account entirely for the difference. Furthermore in 1927 and 1928 the number of surface and bottom hauls made was about equal and during those years the fry appeared 21 times in the surface nets, and only 1 1 times in the bottom ones. Since the bottom nets were not closed while they were lifted, it is entirely possible that some of the fry occurring in these nets were not actually taken on the bottom. The larger larvae, 5.0 to about 10 mm in length (after which they seldom appeared in the tow), occurred in the surface nets quite as frequently as the smaller ones. It may be concluded, therefore, that the young of this species, until they reach a length of at least 10 mm, occur in the open waters and are chiefly surface dwelling. The size at which the young cease to be chiefly pelagic and begin to occupy the habitat of their parents (mainly shallow “grassy” areas) is not definitely known, as no specimens ranging from about 13 to 24 mm in length were taken. Fish 24 mm in length are “young adults” and may be taken with collecting seines in shallow water in the usual summer habitat of the adult, while the fish of the smaller size (13 mm) are still pelagic. GROWTH Insufficient specimens were taken to determine the rate of growth. Specimens 5.0 to 6.0 mm in length first occurred in towings early in June and several specimens 10 mm and one 12 mm long were taken in July. Therefore, the indications are that the larval stages are passed rather quickly, or within 2 or 3 months, and the earliest young of the season probably become “young adults” during their first summer. Specimens ranging from about 13 to 24 mm in length are lacking in the collection. HYPLEUROCHILUS GEMINATUS (WOOD). BLENNY The genus Hypleurochilus contains a single species, which is common at Beaufort, N. C., and from there it ranges southward to the coast of Texas. This blenny, in general, is recognized by its rather deep, compressed, naked body; short, blunt head; low horizontal mouth, with a strong canine tooth on the posterior part of each jaw; broad pectoral, with 14 rays; long, low, continuous dorsal fin, with 11 to 13 spines and 14 or 15 soft rays; and by a tentacle over the eye; which is much larger in the male than in the female. Adult males differ from the females, furthermore, in having fleshy bulbs, covered with folded or creased skin, on the two anal spines, and an eliptical membranous hood in advance of the anal fin and in the presence of folds of skin around the vent. Females have a very distinct genital papilla, at least during the breeding season, which is not evident in the other sex (fig. 90). Males reach a larger size than the females, for the largest fish in the numerous catches invariably were males. Furthermore, the largest male seen was 72 mm long and many others nearly as large were present in the collections, whereas the largest female was only 58 mm in length. 590 BULLETIN OF THE BUREAU OF FISHERIES The common habitat of this blenny locally is among marine growths attached to wharf and bridge piling and to rocks of breakwaters. The ripe fish used in the present investigations were secured from the marine growths (principally asc.idians) attached to the wooden piling of a railroad bridge near the laboratory. The fish were caught with a scrape net, that is, a dip net with a flattened side, provided with a blunt cutting edge. With such a net the marine growths, in part, may be scraped from the piling and frequently a blenny is contained among them. Collecting is most conveniently and efficiently done on low tide. During July and August 1930, one man could catch from two to four dozen fish with a scrape net on the low stages of a single tide. Hypleurochilus geminatus is very hardy and it lives well in an aquarium. There- fore, if captured specimens were not quite ripe, they could be retained several days until their sexual products matured. This species occupies a habitat which is almost identical with that of the adult sheepshead ( Archosargus probatocephalus) , the gamest of the locally represented salt water fishes. Both species feed on attached marine growths and on free swimming forms (principally crustaceans) which also frequent these marine growths. However, Figure 90.— Hypleurochilus geminatus. From adult male 55 mm long. Note plicate membranous bulbs attached to anal spines, covering the anterior one almost completely. the competition probably is not great, as the blenny requires much smaller bits of food than the sheepshead. And young sheepsheads do not enter into the competition, because they have an entirely different habitat. (See p. 532.) Hypleurochilus probably is preyed upon to a limited extent by predatory fishes, but its habitat is very restricted, as already shown, and of such a nature that it is not visited by many species. Neither is this blenny abundant enough to be of much im- portance as a forage fish. Therefore, its economic value locally must be very slight. SPAWNING The presence of eggs of several different sizes within the ovary at one time, as will be pointed out subsequently, suggests a long spawning season. That the period of reproduction is a long one is substantiated by the presence of small fry, under 5.0 mm in length, in the tow from spring to autumn, or to be exact, from May 11 (1929) to October 5 (1927). Such small young were common from the last half of May, through June, July, and August. In September they became less numerous. It seems evident, therefore, that the spawning period extends from May to September, DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 591 inclusive, or possibly into the early part of October, with the principal spawning activities taking place during May, June, July, and August. Adult fish were examined for their spawning condition only during July and August when the egg development reported upon in the following pages was studied. Nearly all the adults taken during July contained ripe or nearly ripe roe. However, during August the percentage of spawned-out fish increased steadily, indicating that the end of the spawning season was approaching. "Nests” containing the eggs of this blenny have been found from time to time for several years and were first reported by Dr. R. E. Coker (in Smith, 1907, p. 377), who found them attached to "rocks, ascidians, shells, etc.” During the present investigation they have been taken on ascidians only. The ova are neatly arranged in rather regular series and in a single layer. They are placed close together, but do not touch, and one nest may cover an area of 2 to 3 square inches. Although naturally spawned eggs were taken several times, the development coidd not be observed satisfactorily within the nests, and it was found impracticable to remove the eggs from their place of attachment without injury. Therefore, other means and methods for their study had to be devised. On one occasion ripe fish were secured and artificially spawned in a glass dish. The eggs adhered equally as tightly to the glass as to the ascidians. Ripe or nearly ripe fish wTere confined in 1930 in a small aquarium, the bottom of which had been covered with microscope slides. It was hoped that the slides woidd receive the eggs when cast and, if so, they could be placed under the microscope for the study of the development. Eggs apparently were cast, as shown by "marks” on the slides where they had been attached. However, they apparently had been eaten by the fish. Thereupon, ripe fish were secured and the eggs were expressed directly on microscope slides where they were fertilized and then placed into sediment dishes, containing sea water, for development. The slides with the eggs attached were removed from the water from time to time for study under the microscope. The eggs did not suffer injury by being exposed to the air for several minutes at a time. By adding water at frequent intervals in sufficient amounts to keep them moist, the observation could be carried on as long as desired. The presence of several different sizes of eggs within the ovary and the com- paratively small number that ripens at one time suggest that this blenny spawns sev- eral times during a breeding season. The eggs in the "nests” observed also were in several different stages of development, ranging from apparently recently laid eggs to others with large embryos, showing that they were not all deposited at the same time. It is not known, however, whether a nest is the product of a single female or whether it receives eggs from two or more individuals. Since all the ovaries examined contained eggs of several sizes, it seems possible that a nest may be the product of a single female and that it returns from time to time to deposit additional eggs as they become mature. (See footnote 7, p. 578). No males were found from which milt could be expressed. To obtain fertilization, males were killed, the testes removed, placed on a slide or in a small dish, with several drops of sea water, and cut and mashed into a pulp with a scalpel. Then the liquid was drawn off with a pipette, distributed over the freshly expressed eggs, and allowed to remain there for about 5 minutes before the eggs were transferred to sea water. Fertilization resulted readily. It was not observed that the nests of Hyplcurochilus geminatus are protected by a. parent fish like those of Hypsoblennius hentz and Chasmodes bosquianus, as stated elsewhere in this paper. Since the eggs of nearly all species of blennies, as far as 592 BULLETIN OF THE BUREAU OF FISHERIES known, are guarded by a parent fish, it seems probable that they are similarly protected in the present species. DESCRIPTIONS OF THE EGG AND YOUNG Description oj the eggs. — Mature eggs within the ovary, according to preserved material examined, are slightly flat on one side to which a sort of disk, or “foot”, is attached. Immediately after spawning the eggs adhere firmly, by means of this disk, to objects (probably principally ascidians in nature) with which they come in contact. The disk as seen under the microscope in newly spawned eggs is granular, slightly irregular in shape, and a little greater in diameter than the egg. The eggs are relatively small, as the diameter of 25 ova secured from several different females (measured in the same plane as the surface to which they were attached) ranged from 0.6 to 0.75 mm, with an average of 0.694 mm. The mature unfertilized egg is so opaque that its structure cannot be seen definitely. The center of the egg as seen with the microscope, using transmitted light, is somewhat paler in color and more densely opaque than the rest of the egg. The pale center is surrounded by purple and orange spots, or spheres, which vary among themselves Figure 91.— Hypleurochilus geminatus. From egg before fer- tilization. The adhesive foot is shown extending beyond the outline of the egg. Figure 92. — Hypleurochilus geminatus. From egg with blasto- disc, shortly before the first cleavage; about 25 minutes after fertilization. in the intensity of their color. By changing the focus of the microscope a slight network of cellular structures, too, is evident on the surface of the egg. The peri- vitelline space is very small and the yolk is slightly granular (fig. 91 ).10 Segmentation and the development oj the embryo. — Fertilization does not cause a change in the size and shape of the egg. The blastodisc becomes evident about 25 to 35 minutes after fertilization at a water temperature of about 28° C. It is not a perfect disc, however, as it is somewhat irregular in shape and generally slightly elongate. Neither does it occupy the center of the upper surface of the egg. Owing to the density of the egg its outline usually cannot be seen definitely (fig. 92). The first cleavage occurs about 1% to 2 hours after fertilization at a water tem- perature of 27° to 28° C. The cells, wdiile notplainly visible throughout, appear unequal in size. Some variation in this respect, however, is evident. Upon completion of io Figures 89 to 102 were drawn from live material; those from 103 to 107 from preserved specimens. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 593 the first cleavage some influence is exerted on the color markings within the egg (previously described as clustered rather closely around the opaque center), which suddenly, as through an explosion, become rather widely scattered within the egg (fig. 93). The second cleavage plane generally follows the first very quickly. In fact some eggs reach the four-cell stage as quickly as the two-cell stage is attained in others. An irregularity in size and shape of the cells remains evident. It is obvious, also, that the purple and orange markings have increased somewhat in size (fig. 94). The eight-cell stage may be expected about 2/ to 3 hours after fertilization, when the water temperature ranges from 27° to 28° C. The cells, although not plainly evi- dent throughout, appear even more irregular and unequal in size than in the earlier stages. The purple and orange spots, or spheres, have continued to increase in size, Figure Q3.—Hypleurochilus geminatus. From egg in 2-cell stage; 154 hours after fertilization. Figure 94. — Hypleurochilus geminatus. From egg in 4-cell stage; 154 hours after fertilization. some of them being nearly twice as large as in the four-cell stage, and they cover most of the germinal disc, further obscuring vision of the segmental processes (fig. 95). The eggs that failed to adhere to the slides by means of the foot, because of crowd- ing or other causes, did not develop. It is important, therefore, that they become attached in the proper position. Judging from the neat and even arrangement of the eggs in the “nests”, it would seem highly improbable that a loss from a similar source would occur in nature. The 16-cell stage follows the 8-cell stage rather quickly and may be expected within about 2 / to 3% hours after fertilization at a water temperature of 27° to 28° C. The cells remain irregular in shape and unequal in size. The germinal disc now spreads over nearly the entire upper surface of the yolk (fig. 96). Cell division continues to progress rapidly, the 32-cell stage following the 16-cell one very quickly. Owing to the opaqueness of the egg and the large color markings, segmentation is very obscure and it generally cannot be followed after the 32-cell stage is reached. The germinal disc now appears to cover the entire upper surface of the yolk. The large opaque center of the egg remains unchanged. In the advanced cell stages the purple spots, varying among themselves in intensity, have all become somewhat less brilliant in color and are irregularly distributed in the yolk. The 594 BULLETIN OF THE BUREAU OF FISHERIES orange and yellow spots, too, are much scattered, but appear to remain unchanged in the intensity of color. It has been stated that the development cannot be followed for some time after the early cleavage stages, owing to the opaqueness of the egg. As a result, the next phase in the process that is clearly evident is the appearance of a notch in the edge of the yolk which no doubt is occupied by the head of the newly formed embryo. This stage is reached in about 20 to 40 hours at a water temperature of 27° to 28° C. (fig. 97). Figure 95.— Hypleurochilus geminatus. From egg in 8-cell stage; about 2}4 hours after fertilization. Figure 98 .—Hypleurochilus geminatus. From egg in abou I 16-cell stage; 254 hours after fertilization. Figure 97. — Hypleurochilus geminatus. From egg showing early stage of differentiation of embryo, 21 hours after fertilization. Figure 98.— Hypleurochilus geminatus. From egg with well- differentiated embryo; 26 hours after fertilization. The outline of at least the head (with large eyes) and tail, which extend beyond the periphery of the opaque yolk, becomes distinctly visible about 25 to 27 hours after fertilization at a water temperature of 27° to 28° C. It is evident now that the embryos are not all in the same position. Some of them lie above the yolk, others curve underneath it, and still others occupy positions intermediate between the ones mentioned (fig. 98). DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 595 Very soon after the embryo becomes well differentiated the purple spots become quite diffuse and shortly disappear. The pale, densely opaque center and the yellow spheres persist somewhat longer. At about this stage two black wavy lines or bars become evident on the yolk. These bars usually meet at one end in the form of a U or a V, although occasionally they are separate. The cellular structures on the yolk of the egg, mentioned previously, have become smaller and much less definite. About 48 hours (2 days) after fertilization at a water temperature of 26° to 28° C. the embryo has nearly encircled the egg. Several somites are visible, but a definite count cannot be made. Circulation is evident. The heart is located under the anterior part of the head, from which a large blood vessel rises and courses through the length of the embryo to near the tip of the tail and then runs across the yolk back to the heart. If other blood vessels are present, they cannot be seen. The embryo already is capable of some movement. The black bars on the yolk, previously mentioned, are now broad and distinct and show indications of breaking up into spots in some specimens. The opaque center of the egg has disappeared, and very rarely a few spots, previously purple but now changed to pink, remain. In some eggs yellow spots are still definitely present. In others they seem to have become diffuse, giving the yolk a yellowish tinge (fig. 99). Figure 99. — Hypleurochilus geminatus. From egg with well- developed embryo; 2 days after fertilization. AV, auditory vesicle; H, heart; BV, blood vessels. Arrows indicate direc- tion of flow of blood. Figure 100. — Hypleurochilus geminatus. From egg with mod- erately large embryo, encircling the egg; 3 days after fertilization. Note pigment on the eye and compare pigment on the yolk with that shown in figures 98, 99, and 101. About 72 hours (3 days) after fertilization, at a water temperature of 26° to 28° C., the embryo slightly more than encircles the egg, although its entire outline generally cannot be seen. The eyes now are completely pigmented with black and overcast with green, particularly above the pupil. Circulation is brisk. The violet colors have entirely disappeared from the yolk, but a few yellow spheres still persist in some specimens. The two dark bars on the yolk usually are broken up into dark spots at this stage. Considerable variation in this respect, however, was noticed. The yolk has been reduced greatly and now occupies only about half the space within the egg case (fig. 100). Development progresses rather slowly in the advanced embryonic stages. About 96 hours (4 days) after fertilization, at a water temperature of 26° to 28° C., the embryo has curved somewhat further round the periphery of the egg than in the last- 596 BULLETIN OF THE BUREAU OF FISHERIES described stage (72 hours), and shows greater activity. The tail appears to be quite free and the eyes, which are very prominent, frequently are “rolled” within their sockets. Circulation is very brisk, corpuscles now being plainly evident in the blood. The aorta can be seen to turn on itself in the tail of the embryo in those specimens that happen to be in such a position that a lateral view is obtainable. Several blood vessels are now visible in the yolk. Some golden color markings without definite outlines have appeared on the head of the embryo in some of the eggs and mixed with the golden color are two dark chromatophores. The yolk which has become greatly reduced and somewhat half-moon shaped usually is marked with several large irreg- ularly shaped dark spots or blotches. Much variation in the size, shape, and number of these spots exists among specimens. One specimen, for example, had a single large elongate black blotch, whereas others had many smaller spots. A few specimens were seen in which a few yellow spheres remained in the yolk. The egg remains round, as seen from above in its attached position, and according to measurements made of four eggs no measurable change in the diameter has occurred (fig. 101). Hatching begins during the sixth or seventh day after fertilization when a tem- perature of about 26° to 28° C. prevails, and it may extend over a period of at least 24 hours. That is, among a batch of eggs all fertilized at the same time, some of the eggs may hatch fully a day earlier than others. An incubation period of 6 to 8 days at the comparatively high temperature which pervaded during the present study is regarded as a very long one for such a small egg. Many other marine fish eggs of simdar size which have been studied hatched in 2 to 3 days. Just before hatching, the eggs become somewhat distorted, that is, the portion of the egg case at the head of the embryo protrudes, causing the egg to become elon- gate and to have a somewhat uneven out- line. The egg and embryo are more opaque than previously and the structures are even more difficult to see. On the head of the embryo is a network of yellow and black markings, and at midbody length are dark, more or less branched, cross lines. The embryo is capable of much movement and appears to struggle, probably in an effort to break the egg case. Newly hatched fish.- — The newly hatched fish is close to 2.4 mm long. It emerges with an extremely small yolksac. Although the fish is very stocky anteriorly its tail is long and rather slender; preanal length is contained 2.45, and postanal length 1.7 times in total length without the caudal fin membrane. The head is blunt, the mouth large and slightly inferior, and the pectoral finfold is prominent. The eye is relatively large, much longer than the snout, and nearly half the length of the head. About 26 myomeres are present. The body is quite transparent and the outline of the brain, the heart, and the circulation can be seen rather plainly. The head and trunk are largely overcast with a yellowish tinge; two irregular dark spots (or simply a blotch in some specimens) are present below the auditory vesicle; a large dark area is present on the upper part of the abdominal mass; and Figure 101. — Hypleurochilus geminatus. From egg with large embryo: 4 days after fertilization. H, heart; BV, blood vessels. Arrows indieato direction of flow of blood. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 597 generally several dark chromatophores occupy the ventral edge of the abdomen, besides a few to several dots which are variously distributed. Dark bars, appearing as spots in a lateral view, are present on the ventral edge of several of the caudal myomeres in some specimens, and on most of them in others. The large eyes are black with a greenish sheen over the pupil. The newly hatched fish swims or floats on its back and is very active (fig. 102). The fish hatched in the laboratory lived only about 2 days. No change worthy of note, except that the color on the abdomen became more diffuse, took place in the meantime. Newly hatched larvae of this species are a little smaller (length 2.4 mm) than those of Hypsoblennius hentz (length 2.7 mm). The larvae of the latter species are also rather more stocky anteriorly. Furthermore, the black marks on the abdomen appear as separate branching chromatophores and are quite generally distributed, whereas in H. geminatus the black is concentrated mostly on the side near the upper margin of the Figure 102.— Hypleurochilus geminatus. From a newly hatched fish 2.4 mm long when alive. A V, auditory vesicle. Figure 103. — Hypleurochilus geminatus. From a preserved specimen 1.6 mm long. This larva is smaller than the newly hatched live fish (Fig. 102) quite certainly because of shrinkage in preservative. abdominal mass into almost solid black with only a few scattered chromatophores elsewhere. In H. hentz most of the inner surface of the pectoral fin membrane is dotted with black branching chromatophores, whereas in H. geminatus only a few black dots at most are present at the base of this fin. Specimens 1 .5 mm long. — The head and trunk in preserved specimens of this size are short and rather robust, while the tail is long, rather slender and compressed, the head and trunk being contained about 2.4 to 2.9 times in the total length without the caudal finfold. The snout is very short and round, scarcely extending beyond anterior margin of eye. The mouth is small, oblique, and terminal, with the tip of the lower jaw slightly below the level of the middle of the eye when the mouth is closed. The vertical finfold is continuous and without rays. The pectoral fins appear as mere tufts of membrane, scarcely longer than the pupil and the ventral fins are not evident. An oblique dark bar extends from the axileof the pectoral to the ventral outline just above the vent; the ventral surface of the chest and abdomen generally is marked with a few to several dark points ; and a distinct dark bar crosses the forehead between the eyes. A rather close-set row of fine, vertically elongate, dark spots is present on the ventral outline of the tail, and the base of the rudimentary pectoral is mostly black (fig. 103). 598 BULLETIN OF THE BUREAU OF FISHERIES Blennies of this size are difficult to identify. However, the specimens assigned to this species are a little less robust than those referred to Hypsoblennius hentz. Further- more, those of the last-mentioned species have rather larger pectoral fin membranes with the basal two-thirds or three-fourths spotted with black, whereas specimens of H. geminatus have black only on the fleshy base of that fin. It will be seen from the two foregoing descriptions that the live fish at hatching are longer than the larvae just described, although the latter are somewhat more advanced in development. The difference, no doubt, is the result of shrinkage during preservation in the last-mentioned group. Specimens 2.0 to 3.0 mm long. — The body is rather strongly compressed through- out, the head being about half as broad as deep. The head and trunk are short in proportion to the tail, the distance from the tip of the snout to the vent being con- tained about 2.9 times in the total length without the caudal finfold. The snout, although rather blunt, especially in 2.0-mm fish, is more pointed than in Hypsoblennius hentz and about one-third (in 2.0-mm fish) to two-thirds (in 3.0-mm fish) the length of the eye. The mouth is terminal and strongly oblique, with the tip of the lower jaw a little above the level of the middle of the eye when the mouth is closed. The vertical finfold is continuous with indications of rays posteriorly in 3.0-mm fish, but not in smaller ones. The pectoral fins are short and broad and scarcely longer than the eye. A broad black oblique band extends from the axile of the pectoral to ventral outline just above the protruding hindgut. The ventral outline of the chest and abdomen usually bear a few dark dots, the upper surface of the head and nape generally has one or more dark cliromatophores, and usually a faint dark bar across the forehead between the eyes. A row of small, vertical, slightly elongate dark spots is situated on the ventral outline of the tail. Dark dots also are present on the base of the inner surface of the short pectorals (fig. 104). This species is distinguished from Hypsoblennius hentz at this size chiefly by the longer and more pointed snout, less strongly elevated forehead, the more strongly oblique mouth, and the much shorter and broader pectorals which bear dark dots only at the base on the inner surface, whereas in H. hentz the dark specks extend to the tips of the fins. These differences are evident in specimens as small as 2.0 mm in length. Smaller larvae, as already stated, are difficult to identify. Specimens J+.O to 4-5 mm long.— The head and body are compressed, the caudal portion of the body being longer than the head and trunk, with the vent situated well in advance of midbody length. The head is deep and rather narrow, the interorbital space being only about half the width of eye. The snout is moderately pointed and about three-fourths as long as eye. The mouth is terminal, rather strongly oblique, the tip of the lower jaw being scarcely below the level of the middle of the eye. The eye is placed moderately high and is about equidistant from the dorsal and ventral DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 599 outlines of the head. Fin rays are only partly developed in the dorsal and anal fins, but much better in the caudal fin which has a round outline. Ventral fins are not evident and the pectoral fins are short and broad, and scarcely longer than the eye. Several more or less distinct dark dots are present on the ventral surface of the chest and abdomen, and on the head and nape; an obscure dark bar crosses the forehead between the eyes. An oblique black bar extends from the axile of the pectoral nearly to the vent; and the inner surface, only, of the base of the pectoral is black. A row of small black dots begins a short distance behind the vent and extends to the base of the caudal fin (fig. 105). Specimens of this species at this size differ from those of the same size of Hypso- blennius hentz principally in the longer and more pointed snout, more strongly oblique, terminal mouth, in the much shorter pectoral fins, which are black at the base on the inner surface only, and in the somewhat more anterior position of the vent, the caudal portion of the body being longer than the head and trunk. Specimens 5.0 to 6.0 mm long. — The body is moderately deep and rather strongly compressed, having made no pronounced change in shape since a length of 4.0 mm was attained. The snout is slightly more rounded than in smaller specimens and usually only a little shorter than the eye. The interorbital is strongly convex and somewhat narrower than the eye. The mouth is terminal and the tip of the lower jaw is on or somewhat below the level of the center of the eye. The preopercle in some specimens shows indications of minute spines, but it appears to be smooth in others. Advance- ment in the development of rays in the dorsal and anal fins, since a length of 4.0 mm was attained, is not pronounced and a definite enumeration of the rays cannot be made. The caudal fin, however, has grown proportionate!}7 longer, is broadly rounded, and similar to the adult. Very minute ventrals are evident in only a few of the rather numerous specimens of this size examined. The pectorals have increased in propor- tionate length and frequently are about as long as the eye and snout. The black, confined to the inner base of the pectoral fin in smaller specimens, now extends some- what on the lower rays of the fin, and the prominent oblique dark bar originating in the axile remains as described in smaller specimens. The ventral surface of the chest and abdomen usually bears several indefinite dark spots; the sides of the head generally have a few very small dark points; and the occipital surface of the head and nape has several more definite ones. A row of somewhat obliquely elongate dark spots begins a short distance behind the vent and extends to the base of the caudal fin, It is evi- dent now that the spots are situated between the bases of the anal rays (fig. 106). The principal characters distinguishing this species from Hypsoblennius hentz at this size do not differ greatly from the ones mentioned for specimens 4.0 to 4.5 mm in length. The snout in H. geminatus remains longer, although scarcely as pointed; the mouth is terminal and more strongly oblique; the pectoral fins, although they have 154979—38 7 600 BULLETIN OF THE BUREAU OF FISHERIES increased in proportionate length, remain shorter and the black on their bases is much less extensive. H. hentz, in the meantime, has developed a bony ridge over and in advance of the eye making the interorbital quite flat. This bony ridge is entirely missing in H. geminatus and the interorbital is strongly convex. Specimens 8.0 to 10 mm long. — The body is moderately elongate and rather strongly compressed, the depth being contained about 5.3 to 5.7 times in the length without the caudal fin. The head is moderately deep ; the snout tapers and is about three-fourths as long as the eye; and the forehead is not very steep, being rather evenly and fairly strongly convex. The mouth remains rather strongly oblique and terminal, as in smaller specimens. The tip of the lower jaw is slightly below the level of the middle of the eye, and the maxillary reaches only a little past the anterior margin of the orbit. The interobital remains strongly convex, with only slight indications of a bony ridge over and in advance of the eye. Very small preopercular spines are present, the longest scarcely exceeding the length of the pupil. The dorsal and anal fins are quite fully developed and a fairly accurate count of the rays can be made. The caudal fin has a straight to a round margin and is nearly as long as the head without the snout. The ventral fins remain very small in specimens 8.0 mm long, but have increased considerably in length in fish 10 mm long, when they are about equal to the eye. The pectoral fins are broad at the base, the middle rays being somewhat produced and about as long as the head without the snout. The ventral surface of the chest and abdomen usually bears a few to several dark dots, sometimes a few dark markings also are present on the sides of the head, and the occipital portion of the head is marked either with small dark dots or with somewhat larger, less well-defined dark or brownish spots. The pectoral fin has a few to several dark dots at the base on its inner surface, and the oblique dark bar behind the pectoral, prominent in smaller specimens, has become quite indistinct in some specimens. Small elongate dark dots situated between the bases of the anal rays, described in smaller specimens, have become more elongate. Each one bends back abruptly and reaches the ray situated immediately behind it a short distance above the base of that ray (fig. 107). The characters distinguishing the young of this species, when 5.0 to 6.0 mm long, from Hypsoblennius hentz of the same size, in general, also separate young 8 to DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 601 10 mm long. The difference in the depth of the body is somewhat more evident than in smaller fish, H. hentz being notably deeper and also somewhat stockier. The much longer preopercular spines, particularly those at the lower posterior angle in H. hentz, too, are useful in separating the species. Another helpful difference is found in the shape of the dark dots at the base of the anal, which have become elongate and form lines in II. geminatus, whereas they are round in II. hentz. Specimens about 12 mm long. — The body has increased somewhat in robustness since a length of 10 mm was reached, but it remains much more slender than in speci- mens of Hypsoblennius hentz, of this size, the depth being contained in the length without the caudal fin 4.4 to 4.75 times. The forehead is strongly convex, but not vertical; the snout projects quite prominently, being fully three-fourths as long as the e}7e; and the mouth is terminal and only slightly oblique, with the tip of the lower jaw scarcely above the level of the lower margin of the eye. A low bony ridge, evident over the eye in somewhat smaller specimens, is scarcely visible now and the inter- orbital remains strongly convex. Minute preopercular spines remain present and do not seem to have either increased or decreased in proportionate length since the fish attained a size of 10 mm. A small fleshy tentacle, notably shorter than the pupil, is visible over the eye for the first time. The dorsal spines, as in the adult, are shorter than the soft rays of that fin; the caudal fin has a nearly straight margin; and the ventral fins have increased in proportionate length, being almost twice as long as the eye. Pigmentation remains virtually as in the smaller fish described in the preceding paragraph (fig. 108). Hypsoblennius hentz is distinguished from the present species at this size (as in smaller fish) by the vertical forehead, the slightly inferior horizontal mouth, the prominent bony ridge over and in advance of the eye, by the much larger preopercular spines, and by the greater amount of black color on the pectoral fins. Specimens 15 to 16 mm long. — Specimens of this size are very similar in shape to the adult, and they have the appearance of being much older fish than 12- or even 14-mm specimens. The body has become deeper and more robust, the depth being contained in the length, without the caudal fin, 3.3 to 3.6 times. The snout projects rather prominently in advance of the forehead and it is nearly or quite equal to the length of the eye. The small mouth is now wholly below the level of the lower margin of the eye. It is almost horizontal, as in the adult. The lower jaw is slightly shorter than the upper one and the maxillary scarcely reaches beyond the anterior margin of the pupil. Preopercular spines, present in somewhat smaller specimens, are not evident. Three or four fleshy tentacles, placed close together and in a transverse row and rising from a common base, are present over each eye, the longest one being about as long as the pupil. Another fleshy tentacle is present behind the nostril. The fins are all shaped virtually as in the adult. The color cannot be fully described, 602 BULLETIN OF THE BUREAU OF FISHERIES as only old alcoholic specimens (collected in 1912 and 1913) of this size are at hand. They are rather pale in color and have only a few dark points in advance of, as well as behind, the ventral and pectoral fins. Similar dots occur on the occipital portion of the head, and a row of elongate dark spots are present on the base of the anal fin (fig. 109). Specimens 20 to 22 mm long. — A canine tooth on the posterior part of each jaw, constituting a generic character, has become evident at about this size. Pigmenta- tion is general and complete, and similar to that of the adult. Recently preserved specimens (which do not differ greatly in color from live fish) are brownish. Some are plain dark brown and others, somewhat lighter in color, have indications of dark bars on the upper part of the sides. Indefinitely outlined dark spots are present along the middle of the sides and also on the base of the anal fin. The dorsal and anal fins are profusely dotted with brown, similar to the body, as seen under mag- nification. A dark spot is present on the membrane between the first two dorsal spines, the margin of the anal is pale, and the caudal fin has dark cross bars. The ventral and pectoral fins are finely dotted like the dorsal and anal, and the pectorals in addition bear larger dark spots. Specimens 20 to 22 mm long virtually are “young adults” with posterior canines, and with the color almost identical with that of the adult. The size of the fish at which general pigmentation takes place, however, has not been determined, as only greatly bleached alcoholic specimens ranging from 14 to 18 mm in length are at hand. It can be stated at this time only that general pigmentation lias not begun at a length of 14 mm, whereas it is complete at 20 mm. DISTRIBUTION OF THE YOUNG The fry were taken in to wings in outside waters 76 times and in inside waters 12 times. No record of the numbers of towings taken was kept, but it is probable that nearly as many hauls were made in inside waters as in the outside ones. The collections show, therefore, that the young are more numerous off Beaufort Inlet than they are in Beaufort Harbor and adjacent sounds and estuaries. The fry were taken in surface towings 65 times and in bottom hauls only 16 times. Although a considerably larger number of surface than bottom hauls was made, the discrepancy in the numbers of hauls certainly was not great enough to equalize the difference. Furthermore, in 1927 and 1928 the surface and bottom hauls were nearly equal in number and during those years the fry occurred in surface towings 51 times and in bottom drags 9 times. Since the bottom nets were not DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 603 closed while they were hauled in, it is possible that some of the fry caught may have been taken somewhere between the bottom and the surface. The larger fry (5.0 to 10 mm) were taken no more frequently in bottom drags than the smaller ones. Therefore, the evidence is that the larvae of this species, until a length of about 10 mm is reached, live in the open waters and are chiefly surface dwelling. Fish of all sorts, after reaching a length of about 10 to 15 mm, are taken sparingly in 1-meter tow nets. Many species at this size may be taken in an otter trawl, having the cod end covered with bobbinet. That mode of collecting failed, however, for the present species, very probably because the fish no longer occurred in the open waters. The usual habitat of the adults, as stated elsewhere, is among marine growths attached to wharf and bridge piling, rocks, shells, etc., and specimens as small at 16 mm in length have been taken in such an environment. However, no special effort to collect small fish in the favorite haunts of the adults has been made. It seems probable that the young fish, after abandoning the open waters, take up their abode with the adults and that they will be found there when collections are made with suitable apparatus. GROWTH The data on the rate of growth are meager, owing to the scarcity in the collections of specimens ranging from about 10 to 15 mm in length. A change in habitat apparently takes place at about this time in the life of the fish, as already shown. The new habitat is not well known and requires further exploration. Examples around 8.0 mm long occurred in the tow as early as June 2 (1928), and are rather common thereafter throughout the summer. Also, several specimens about 10 mm long were caught in the tow during the summer, the first one of this size having been taken on July 3 (1928). However, a larger one (12 mm) was caught as early as June 28 (1927). These data indicate, therefore, that the larval stages are passed rather quickly and that a length of 8.0 to 10 or even 12 mm may be attained within 1 to 2 months after hatching. Specimens 16 to 22 mm long were dredged on shelly bottom and caught on wharf piling in July and August. Such fish are “young adults” and may or may not belong to an older year class. They, at least, look much older and more mature than the single 14-mm specimen secured in the tow. The indications are, therefore, that no great increase in length takes place at the time (between about 14.0 and 16.0 mm) when the fish acquires nearly all the characters of the adult. It is during this time and probably somewhat earlier, as already pointed out, that the fish leaves the surface waters and begins to live with the adults among marine growths attached to rocks, shells, submerged timbers, wharf piling, and other objects. CHASMODES BOSQUIANUS (LACEPEDE). BANDED BLENNY This blenny is not very common at Beaufort, and the least numerous of the three local species. It is reported from New York to Florida, apparently being more numerous in Chesapeake Bay than elsewhere. It may be distinguished from the other blennies occurring locally by the more pointed snout, by the larger mouth (the maxillary reaching to or past posterior margin of the eye), the absence of canine teeth (present in Hypleurochilus geminatus only), by the rather longer dorsal and anal fins (the dorsal formula being XII, 18, and that of the anal II, 17 or 18), and the fewer rays in the pectoral fin (12, rarely only 11). Insufficient specimens are at hand to determine the relative sizes attained by the sexes. No males or females exceeding a length of 70 mm were seen at Beaufort. 604 BULLETIN OF THE BUREAU OF FISHERIES The secondary sex characters do not differ noticeably from, those of the other local species. Each anal spine bears a fleshy expansion at the tip in adult males, and a membranous expansion is present immediately in advance of the anal fin. Also the female has a more or less distinct anal papilla (fig. 110). This blenny apparently inhabits shelly bottom only at Beaufort, though in Chesapeake Bay it was taken on clay, mud, and sand (Hildebrand and Schroeder, 1928, p. 333). A few specimens were taken at Beaufort in nets hauled over shelly grounds. A somewhat larger number of fish, however, was taken by hand in oyster, clam, and scallop shells. The shells were occupied not only during the spawning season, but at other times also and by both sexes. The banded blenny is hardy. It lives well in confinement, and during cool weather at least it can live out of water a long time. For example, on the afternoon of November 24, 1927, an individual occupying a scallop shell containing some sand and mud was picked up by hand. It was placed in a dry container with the shell and left over night. The next morning the fish was still in a lively condition. Upon being placed in an aquarium it at first deserted the shell, but soon afterwards reoc- cupied it. It lived in the aquarium for several weeks, and allowed itself to be lifted from the water with the shell numerous times. Some of the nests found were so near the usual low-tide line that they must become exposed when rather exceptionally low tides occur. At such times if the male fish guarding the nest does not desert it, he may have to live for a while either without water, or at most only from water brought by the wash of waves. Observations indicate, however, that if the nest is deserted the eggs most probably will be destroyed soon by enemies, as explained subsequently. This little fish is game and when handled fights vigorously. It will grasp the skin and flesh of the hand and hold on bulldog fashion. However, its jaws and teeth are too weak to inflict a wound. The illustrations of the development of the egg and of the newly hatched fish, presented herewith, are all based on live specimens. The young were not taken in collections made in nature, and those hatched in the laboratory died within a day or so after hatching. Consequently, no material for the study of their develop- ment is available. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 605 SPAWNING This blenny, like the others reported upon in this paper, evidently does not spawn all of its eggs at one time, as ova of several different sizes are present in the ovary during the spawning period. The length of the spawning season has not been deter- mined fully. The young have not been taken, and insufficient adults have been collected to make a full determination from the study of the gonads. However, a nest was found as early as May 15 (1920), and as late as August 14 (1930). Other nests were taken in June and July. Hildebrand and Schrocder (1928) report a nest taken May 22 (1922), at Cherrystone Island, Va. From these data it may be con- cluded that the spawning season extends at least from May to August. The nesting habits of this blenny, so far as known, are identical with those of Hypsoblennius hentz. The eggs have been found only in oyster shells, although clam and scallop shells probably also are used. A full nest covers the entire inside of both valves of an oyster shell. The eggs are firmly attached in a single layer, though not always in definite rows, and are well separated by the adhesive disk which has a greater diameter than the egg itself. For study, the eggs with the disk may be removed from the shell with a sharp instrument, but the disk could not be separated from the egg. In this species, as noted for Hypsoblennius hentz and Hypleurochilus geminatus, the eggs in a nest are not all in the same stage of development, a range from an early cleaveage stage to an advanced embryonic stage having been observed. The remarks made under the discussion of Hypsoblennius hentz (p. 578) as to whether all the eggs in one nest are the product of one female apply equally as well to Chasmodes bosquianus. Presumably the nests are always guarded by the male, as already indicated. The care of the male is evidently necessary to prevent the destruction of the eggs by enemies and to keep them clean and healthy. The eggs in a deserted nest in nature were destroyed quickly by the small flat mud crab, Eurypanopeus depressus (Smith), that also attacked the eggs of Hypsoblennius hentz (p. 579). Those in two other deserted nests, placed in tanks with running water, all died in an advanced embryonic stage, having become infested with liydroids and a copepod, Tisbe jurcata (Baird).11 A small percentage of several dozen eggs removed from a nest when in rather early developmental stages and placed in glass bowls, in which the water was changed twice daily, hatched successfully. Spawning in this species, as in Hypsoblennius hentz, apparently takes place early in the morning, as only those nests taken before 10 o’clock contained eggs in the early cleavage stages. DESCRIPTIONS OF THE EGGS AND THE NEWLY HATCHED YOUNG Description oj the egg. — The eggs of Chasmodes bosquianus are slightly flattened next to the adhesive disk or “foot” which attaches them to the inside of oyster shells and possibly to other bivalve mollusks also, as already explained. The eggs of the present species are larger than those of the other blennies discussed in the preceding pages. The greater axis has a length of 0.93 to 1.1 mm in 27 eggs measured and an average length of 1.04 mm. The lesser axis which cannot be measured accurately, because the grayish opaque adhesive disk obscures the outline of the egg, has a length of about 0.8 to 0.9 mm. The slightly flattened contour of the egg at the place of attachment is not shown in the drawings portraying lateral views, as the degree of depression could not be determined definitely. ii The writers are indebted to Dr. C. B. Wilson, State Teachers College, Westfield, Mass., for this identification. 606 BULLETIN OF THE BUREAU OF FISHERIES The eggs as seen on an oyster shell with the unaided eye are very pale yellow. Under magnification just a tinge of yellow is evident. Numerous yellowish oil globules, mostly in that half of the yolk nearest the adhesive disc, are present. The eggs have a dense opaque central body, as in the other blennies studied. No bluish or reddish spots are present in the yolk and therein the eggs of this species differ conspicuously from those of the other local forms. The eggs of this species apparently are even more opaque than those of the other blennies studied. The yolk is quite granular, becoming more so as develop- ment progresses. The egg membrane is cellular in appearance. When the microscope is refocused the lines have the appearance of deep ravines with elevations between them. This sculpture of the egg membrane is not shown in the accompanying illustrations. The adhesive disk described above, is shown in only one drawing (fig. 118), although of course it is always present. Figure 112. — Chasmodes bosquianus. From egg in 2-cell stage; about 2 hours after fertilization. (Drawn by Nell Henry.) Figure 111. — Chasmodes bosquianus. From egg with blast- odisc; shortly before the first cleavage; probably about an hour after fertilization. (Drawn by Nell Henry.) Segmentation and the development of the embryo. — The following account is based entirely upon eggs collected in nature. The exact time of spawning and fertilization is not known. Therefore, the time intervening between fertilization and the begin- ning of cleavage cannot be stated definitely. In a nest taken at 9:30 o’clock in the morning eggs were present in which the first cleavage took place about an hour after collection. It seems probable that these eggs were laid early on the morning the nest was brought to the laboratory, as already explained (p. 605). In Hypleurochilus geminatus about 2 hours intervened between fertilization and cleavage at a temperature of 26° to 28° C. It probably may be assumed that in Chasmodes about an equal length of time elapses between fertilization and segmentation, at nearly identical temperatures. The blastodisc in all the eggs examined lay next to the adhesive foot by which it was largely obscured when viewed in the normal position. However, when the egg was turned so that the adhesive surface of the disk was at right angles to the side upon which it rested, a fair lateral view of the blastodisc and segmentation was obtainable. Accordingly, the illustrations showing different stages of cleavage are all lateral views. In general, only that part of the disc extending beyond the yolk is shown, as the opaqueness of the egg obscured the rest. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 607 The blastodisc is large and projects prominently beyond the yolk. The pervi- telline space is wide at the positive pole of the egg and very narrow or wanting at the negative pole (fig. 111). The first blastomeres are large and about equal in size, as usual in teleosts (fig. 112). The second cleavage plane is approximately at right angles to the first, and it followed the first, at a water temperature of about 26° C., in about 20 minutes (fig. 113). The third and fourth cleavages followed equally as rapidly. Figure 113. — Chasmodes bosquianus. From egg in 4-cell stage; about 2H hours after fertilization. Owing to opaqueness of egg all the cells could not be seen from one viewpoint. (Drawn by Nell Henry.) Figure 115.— Chasmodes bosquianus. From egg probably in the 64-cell stage; about 3H hours after fertilization. Owing to opaqueness of the egg the cells could not be counted accu- rately. (Drawn by Nell Henry.) Figure 114. — Chasmodes bosquianus. From egg in 16-cell stage; about 3 hours after fertilization. (Drawn by Nell Henry.) Figure 116. — Chasmodes bosquianus. From egg in rather ad- vanced cleavage stage; about 6 or 7 hours after fertilization. Owing to opaqueness of the egg only that part of the blast- oderm projecting above the yolk is visible from one view point. (Drawn by Nell Henry.) When the eight-cell stage is reached, all the cells are no longer visible in a lateral view, and they cannot be seen in a surface view, as already explained. Therefore, further divisions cannot be clearly observed. The blastomeres are large and prominent until about the 16-cell stage is reached (fig. 114). Thereafter they get smaller and flatter rather rapidly (fig. 115). 608 BULLETIN OF THE BUREAU OF FISHERIES The eggs in which cleavage started at about 10:30 in the morning reached a fairly- advanced cleavage stage by the evening of the same day (fig. 116). The temperature of the water had remained near 26° C. throughout the period. No pronounced changes had taken place in the egg in the meantime, except that the yolk apparently had become more granular and rather more opaque. Twenty-four hours after cleavage started a very early embryonic stage was reached, that is, the embryo was just becoming differentiated, though it was not yet possible to distinguish between the head and tail. The temperature of the water had dropped to 24.5° C. Little obvious headway was made during the next 12 hours. However, a fairly well-formed embryo, with the eyes partly developed, was present, at about 48 hours after fertilization. The temperature of the water had advanced to 26° C. The embryos evidently were not all in the same position in relation to the adhesive foot. In some eggs the embryo lay underneath the yolk, next to the foot, with only the head and tail visible if viewed from the side opposite the foot. Other embryos lay mostly above the yolk and therefore were entirely visible in eggs seen from the same angle. Positions intermediate of these also were observed. A few grayish blotches, variable in size, were noticed on the yolk for the first time (fig. 117). About 60 hours (2 % days) after fertilization, the temperature of the water remain- ing near 26° C., the embryo was well formed, with a large head and partly pigmented eyes. It curved about two-thirds the distance around the egg. Indications of somites were present at midbody length and the heart beat slowly and rather feebly (about 90 beats per minute). Circulation was evident only near the heart, no definite blood vessels apparently having been formed. Black blotches, with irregular outlines, variable in size and shape in any one egg and variable in number in different eggs were present on the surface of the yolk (fig. 118). On the fourth day of incubation, with the temperature of the water remaining quite constant at 26° C., the body segments had become plainly marked in the anterior caudal region, although the embryo had gained little in length. The eyes had many black pigment dots, most numerous along the upper margin ; the yolk appeared very granular and had been cut into deeply by the embryo, and the dark spots on its surface Figure 117 —Chasmodes bosquianus. From egg with moder- ately well-differentiated embryo; 2 days after fertilization. (Drawn by Nell Henry.) Figure 118.— Chasmodes bosquianus. From egg with well- formed embryo; 2Yi days after fertilization. Tail of embryo curved under the opaque yolk. (Drawn by Nell Henry.) DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 609 in general bad become smaller and more numerous. Tbe heart was beating rapidly, sending the blood to the body through a large vessel situated near the ventral outline. This artery left the embryo somewhat more than an eye’s diameter from the tip of the tail where it entered the yolk and divided into several branches. These branches all coursed over the yolk and in a general way ran toward the snout of the embryo, underneath which the heart is situated. There they united just before pouring their contents into the heart (fig. 119). On the fifth day of incubation, with the temperature of the water still remaining near 26° C., the embryo had encircled the egg. The tail reached opposite the head. It was free and moved frequently. The eyes were very prominent, being fully pig- mented, and visible without magnification. The yolk was reduced to about two-thirds its original size and more or less crescent-shaped, having been cut into very deeply in the head region of the embryo. Oil globules of various sizes remained distributed throughout the yolk. The central opaque body, previously described, and still visible a day or so earlier, had disappeared. Dark markings on the yolk had become more Figtjre 119. — Chasmodes bosquianus. From egg with devel- oping embryo; 4 days after fertilization. Arrows indicate the direction of the flow of blood in the larger vessels. (Drawn by Nell Henry.) Figure 120. — Chasmodes bosquianus. From egg with ad- vanced embryo; 6 days after fertilization. H, heart. Arrows indicate direction of flow of blood in the larger vessels. (Drawn by Nell Henry.) numerous and in general smaller. They now consisted mostly of lines branching more or less from a central point, and many of them were shaped somewhat like crow’s feet. A concentration of dark markings was taking place in the trunk region of the embryo. Circulation was brisk, and the blood returned within the embryo, the caudal vein being quite fully developed. Corpuscles were distinct, and the heart and large vessels near it had a pinkish tinge (fig. 120). Development progressed slowly after about the fifth day of incubation. By the seventh day, with a drop in temperature to 24.5° C. between the sixth and seventh day, the tail of the embryo reached a little past the head. The embryo was capable of considerable movement, carrying the yolk with it as it turned in the egg case. The yolk had been cut into more deeply and was definitely crescent-shaped. The black markings on the yolk, described in the foregoing paragraph, although variable in number in different eggs, had become less numerous, and a further concentration of black had taken place in the trunk region of the embryo. Also an irregular black blotch was present at each auditory vesicle. Almost innumerable blood vessels were 610 BULLETIN OF THE BUREAU OF FISHERIES visible in the vicinity of the head of the embryo and all poured their contents into the heart, which had the appearance of a pit (fig. 121). No important changes in the embryo itself appeared after about the seventh day of incubation. The temperature remained near 24.5° C. from the fifth to the ninth day when it advanced to 26° C. On about the ninth day it was evident that the black color concentrated in the abdominal region of the embryo, first noticed on the fifth day, was on the embryo, whereas it at first appeared to be on the yolk. On some eggs a few “crow’s feet” remained on the yolk, whereas in others they had all dis- appeared. A dark blotch was present between the anterior part of the eyes, and in some specimens short, branching, cross lines were evident on the ventral margin of some of the caudal myomeres. Also distinct black spots were present on the pectoral fin membranes which could be seen clearly through the egg case. The eggs in two unguarded nests (the writer was not successful in inducing a male of this species, to stay with his nest in the aquarium) all died between the fifth and ninth days of incu- bation, having become infested with hydroids and protozoa. Eggs removed from the nests while in early cleavage stages and placed in glass bowls, in which the water was changed twice daily, also nearly all became infested and only four hatched. The rest of this account is based on the few remaining eggs and the four larvae that emerged successfully. On the tenth day of the incubation period the temperature of the water advanced to 27° C. The embryos were very active. The dark color markings on the embryo, already evident on the ninth day or earlier, had become more distinct. The number of blood vessels had increased and the blood, when viewed under moderately low magnification, was seen pouring over the head and eyes in minute vessels as if in a sheet. Heart action was extremely rapid, the beats following each other in such close succession that they could not be enumerated accurately. The heart had a distinct reddish tinge, the red probably being in the blood. Only four eggs survived, as previously stated, and these all hatched on the eleventh day of incubation. The temperature dropped from 27° to 25° C. between the tenth and the eleventh day. The incubation period of this blenny, therefore, is around 11 days when the temperature of the water in which the eggs are incubated ranges between 24.5° and 27° C., with a mean temperature around 26° C. No attempt was made to keep the larvae alive. After measurements and a sketch had been made and a description prepared they were preserved. Newly hatched fish.- — The newly hatched fish range in length from 3.56 to 3.78 mm. The yolk is small at hatching. The head and trunk are short and robust, and the tail is long and slender. The vent is situated far in advance of midbody length ; distance from snout to vent being 1.25 to 1.3 mm, from vent to tip of tail without finfold 2.1 to 2.3 mm. The snout is short and blunt, its length being less than half the diameter of the eye. The eye is large, its diameter (0.36 mm) being a little greater than the depth of the body just behind the vent (0.32 mm). The mouth is placed rather low, Figure 121. — Chasmodes bosquianus. From egg with large embryo; about 7 days after fertilization. H, heart. Arrows indicate direction of flow of blood in the larger vessels. (Drawn by Nell Henry.) DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 611 anteriorly scarcely above tlie lower margin of the eye. The gape reaches to or a little behind the vertical from the anterior margin of eye. The vertical finfold is rather broad; originating above the auditory vesicle, it is continuous and extends to the vent. Large pectoral fin membranes, somewhat longer than the diameter of the eye, are also present. The body is fairly transparent. As a consequence, the outline of the brain and the circulation of the blood can be seen rather clearly. The aorta and the caudal vein remain rather close together, as in the embryo, both being located ventrally in the long tail. About 8 or 9 partly indefinitely outlined myomeres may be counted in advance of the vent and from 28 to 30 behind it. The vertebra count in two adults examined was 9 + 24 and 10 + 23. These counts seem to indicate that the total number of myomeres in the newly hatched fish is greater than the number of vertebrae in the adult. Since the demarcations between the myomeres both ante- riorly and posteriorly are indistinct, it is barely possible that the count is excessive. Nearly all the color markings on the newly hatched fish already were evident on the embryo several days before hatching. The black on the snout is in elongate “twin” blotches, situated over the anterior part of and somewhat in advance of the eye. The abdomen is largely black along its upper margin, the black reaching from the upper edge of the base of the pectoral to the vent. A few small black chromato- phores remain on the ventral surface of the abdomen (yolksac). The tail has short branching cross lines on the ventral edge, which sometimes are wanting anteriorly and also posteriorly. The basal three-fourths of the inner surface of the pectoral has black cliromatophores, the lowermost spot being very large, while those more distant from the base of the fin are smaller (fig. 122). Figure 122. — Chasmodes bosquianus. From a newly hatched fish, 3.6 mm long. AV, auditory vesicle. Arrows indicate position and direction of the flow of blood in the aorta and caudal vein. (Drawn by Nell Uenry.) Chasmodes bosquianus is about 3.6 mm long at hatching, whereas Hypsoblennius hentz is about 2.7 mm long, and Hypleurochilus geminatus is only about 2.4 mm. Chasmodes apparently has a larger number of myomeres behind the vent, having about 28 to 30, whereas the other species have only about 23 or 24 at hatching. No black color markings were noticed under the auditory vesicle in Chasmodes, whereas more or less black is present in the other species. Chasmodes and Hypsoblennius agree in having most of the inner surface of the pectoral fin membrane dotted with black chromato- phores, whereas Hypleurochilus at most has only a few black dots on the base of that fin. On the other hand, Chasmodes and Hypleurochilus agree in having a concen- tration of black points, forming almost solid black, along the upper margin of the abdomen, extending from above the base of the pectoral to the vent, while Hypsoblen- nius more usually has scattered branching chromatophores quite generally distributed over the abdomen.' Considerable variation in the distribution of the black markings on the abdomen, however, has been noticed in all three species. 612 BULLETIN OF THE BUREAU OF FISHERIES THE HAKES OF THE GENUS UROPHYCIS The development and other life history data of four species of Urophycis, namely chuss, regius, floridanus, and earlli,12 are discussed in the following pages. Some of the hakes are rather widely distributed. U. chuss has been recorded from the Gulf of St. Lawrence southward to Cape Henry, Va. The known range is now extended southward to the coast of North Carolina, on the basis of some small speci- mens at hand, taken by the Albatross at sea off Kitty Hawk. U. regius is known to range from Nova Scotia southward to South Carolina; U . floridanus from Beaufort, N. C., to Pensacola, Fla., and U. earlli from Beaufort, N. C., to Charleston, S. C. U. floridanus was first recorded from Beaufort by Hildebrand (1916, p. 306). Since that time, the young of this species, have been found to be common locally in shallow water during the winter and early spring, but the species apparently is absent there during the summer. U. regius is more common, the young being numerous during their first winter, but adults were rather rarely taken. It seems possible that these hakes, after spending the first several months in shallow water, live chiefly in deep water offshore where very little collecting has been done. The habitat of both the young and adults is discussed under the heading, “The distribution of the young.” According to our field records only four specimens of U. earlli were taken during the senior author’s connection with the biological station at Beaufort from 1914 to 1917 and 1925 to 1931, notwithstanding that Smith (1907, p. 384) stated, “Tins hake * * * is not uncommon in the Beaufort and Cape Lookout regions. * * * On the adjacent shores the fish is common enough to have received a local name, ‘Dickie,’ although it has no economic value as yet.” In view of the later, much more intensive collecting, one wonders if there was not confusion with one of the other more common species. The southern species of hake do not grow large. U. regius is reported to attain a maximum length of 16 inches. The largest individual seen at Beaufort was 13)( inches long. The largest specimen oi floridanus taken was only 8)£ inches long, and the largest one of earlli 15K inches, though one 18 inches long has been recorded. U. chuss is reported to reach a length of about 30 inches, or even 42 inches, if tenuis is not dis- tinguishable from that species, as suggested by Vladykov and McKenzie (1935, p. 71). The hakes as yet are of no commercial value in North Carolina. Small catches are made in Chesapeake Bay and off Cape Henry, Va. The catch for Virginia (not separated by species) for 1934 is given as 21,000 pounds in the statistical report of the Bureau of Fisheries. Northward the hakes increase in importance, the catch for New Jersey for 1934 being 22,171 pounds, and for New York 139,954 pounds. Large catches are made in Massachusetts and Maine, and smaller ones in the other New England States, the total catch for those states for 1934 being 15,319,692 pounds. The meat of the hakes is soft, but it is of good flavor, and generally sells readily. Adult hakes of the genus Urophycis, as here understood, are recognized by the elongate somewhat compressed body; subconical head; rather large, nearly horizontal mouth, with the maxillary generally reaching to or beyond the posterior margin of the eye; with unequal teeth on the jaws and vomer, none on the palatines; a small barbel i* Jordan, Evermann, and Clark (1930, pp. 212-213) place earlli in the older genus Phycis, leaving the other three species herein discussed in Urophycis. This classification does not seem justifiable, as earlli is closely related to floridanus, differing only in the smaller scales, rather longer dorsal and anal fins, and in color. In turn, floridanus differs in the same characters and in about the same degree from regius, the type of Urophycis. In addition, floridanus has a longer chin barbel, wherein it agrees with earlli. Cer- tainly earlli is more closely related to regius than to chuss with its low broad head, large eyes, and produced dorsal ray. Evidently a further study of the group is necessary to determine the status of Urophycis. Perhaps all the species herein discussed should be assigned to the genus Phycis. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 613 at the chin; and two dorsal fins, composed of soft rays only, the first one short, the second one long, and similar to the anal. The ventral fins are described in current works as consisting of three slender rays, closely joined, and appearing as a bifid filament. In the young three separate rays are plainly evident, and sometimes a fourth one may be discernible. Upon removing the skin, it was found that two articulated rays are enclosed in each filament of the adult and a short “remanent” of a fifth ray (unarticulated) is present at the inner side of the base of the fin. These hakes, therefore, actually have four-rayed, or five- rayed ventral fins if the unarticulated remanent is counted. The distinguishing characters of the adults of the species herein discussed are most readily shown in a key. KEY TO THE SPECIES a. Chin barbel very short, not exceeding the pupil of the eye in length. b. Head depressed, notably broader than deep; eye large, equal to or wider than interorbital; scales small, 104 to 112 or more oblique series above the lateral line; first dorsal with a long filamentous ray; dorsal rays 9 to 11 — 56 to 61; anal rays 52 to 56; lateral line not in a black streak and with- out white spots; no white on first dorsal chuss. bb. Head scarcely depressed, deeper, its depth about equal to its width; eye smaller, not as wide as interorbital; scales larger, 89 to 97 oblique series above lateral line; first dorsal without a produced ray; dorsal rays 8 or 9 — 46 to 51; anal rays 43 to 49; lateral line in a black streak, interrupted by pale spots; first dorsal largely black, margined with white regius. aa. Chin barbel notably longer, always longer than the pupil, frequently nearly or fully as long as eye. c. Scales moderately small, 110 to 130 oblique series above lateral line; dorsal rays 12 or 13 — 54 to 59; anal rays 40 to 49; color bluish or brownish above, silvery below; lateral line in a black streak, interrupted by pale spots; vertical fins mostly pale brownish, often with dusky margins, the first dorsal largely black, not margined with white floridanus. cc. Scales very small, 153 to 175 oblique series above lateral line; dorsal rays 8 or 9 — 54 to 63; anal rays 50 to 56; color dark brown to nearly black, sometimes with pale blotches; lateral line not in a black streak and without pale spots; the vertical fins frequently nearly black, no white on first dorsal earlli. SPAWNING The eggs of Uropliycis, as already stated, were not secured at Beaufort. Neither were ripe adult fish seen by us. However, the capture of spawning fish by the Albatross on the coast of the Carolinas in December, 1919 is reported in the field notes by the late W. W. Welsh. Small larvae, that is, young under 5.0 mm in length, were taken only a few times, as follows: One, 3.0 mm long, November 12, 1927, 13 miles west southwest of Cape Lookout; 13, ranging in length from 2.75 to 4.5 mm, December 6, 1927, at the same station; and 1, 3.0 mm long, December 6, 1927, 6 miles west south- west of Cape Lookout. These larvae, as stated elsewhere, apparently represent about equally regius and floridanus. Larger young were taken frequently and sometimes in abundance, during December and the following several months, as shown by tables 1 and 3. The very small larvae taken are very probably only several days old, which seems to show that both regius and floridanus spawn in the general latitude of Beaufort at least during November and December. The small size of some of the young, though beyond the larval stage, taken during the several succeeding months suggests, however, that the spawning season extends over a longer period of time. A few specimens of floridanus and several of regius, 30 to 40 mm long, were collected as late as March, and a few of regius 38 and 39 mm long as late as April 15 (1931). Judging from the 614 BULLETIN OF THE BUREAU OF FISHERIES growth data contained in tables 1-4, it apparently may be assumed that these species in the general latitude of Beaufort spawn from about November to February. U. earlli is so scarce at Beaufort that very little material was obtainable. In fact only three young, 37, 75, and 103 mm long, were secured. Therefore, virtually nothing was learned concerning its life history. However, the two larger young were taken March 24 (1931), and the smallest one April 15 (1931), when regius of about the same size also were taken. It is possible, therefore, that earlli, like the other local species of hake spawns during the winter on the coast of North Carolina. It may be stated with some assurance that the hakes do not spawn in the bays and estuaries at Beaufort, as the eggs and larvae were not taken in these waters during several years of intensive collecting. All the larvae of regius and floridanus at hand were taken at sea from 6 to 13 miles offshore, beyond which no collecting was done. It apparently may be assumed, therefore, that these hakes spawn only at sea in the vicinity of Beaufort. The abundance of young floridanus, and especially of regius, during the winter and early spring indicates rather extensive spawning in the Beaufort region. We have included in the present discussion TJ. chuss for reasons already stated, though this species is not recorded from Beaufort. In regard to the spawning Bigelow and Welsh (1925, p. 452) stated that the height of the spawning season of this species falls in early summer in the Massachusetts Bay region and at least as early as June south of Cape Cod. Also, that the extreme limits of the spawning season were not known, but that the evidence collected indicated that it spawns in the Gulf of Maine from late spring until early autumn. We have at hand specimens 2.75 to 15 mm in length collected by the Albatross off Cape Henry, Va., October 30, 1919, and off Kitty Hawk, N. C., October 31, 1919. We, also, have specimens of similar size collected on the coast of New Jersey by the Grampus, July 19, 1912. It seems, therefore, that the spawning season of this species is a very long one. DESCRIPTIONS OF THE EGGS AND YOUNG The eggs of Urophycis were not recognized in collections made at Beaufort, and the larvae were not taken often. Those collected apparently are separable into two species, namely, regius and floridanus, as shown subsequently. The smallest specimen of earlli taken, the only other species of Urophycis known from Beaufort, is 38 mm long. Various additional collections of young hakes from both north and south of Beau- fort, made principally by the Albatross, the Grampus and the Fish Hawk, are at hand for study. These include almost a complete series of the northern hake, U. chuss, which is not known from Beaufort, though specimens taken off Kitty Hawk, N. C., are at hand. We also have the notes and some rough camera lucida drawing of the development of the eggs, and newly hatched young of chuss, made by the late W. W. Welsh. Some of this information, together with two of the drawings, already has been published by Bigelow and Welsh (1925, p. 454). It seems desirable to bring to light more of the information gathered long ago (1916) by Mr. Welsh, and to include as full an ac- count of the development of this species as the data and specimens at hand permit. This seems especially desirable because of the close similarity of the young to the species occurring at Beaufort. The development of the shape of the body is most peculiar, as may be seen from the descriptions and illustrations of the stages of development. The early larvae are slender; next, at a length of 4.0 mm or so, they become considerably deeper and more DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 615 compressed. Then, at a length of about 10 mm, they have become slender again, and they remain so until they are fairly large fish, ranging upward of 100 mm in length, when, at least regius and floridanus again become deeper, and especially more robust, that is, less strongly compressed. It is interesting also that contrary to most of the other species discussed in this paper, the hakes, at no time have spines on the pre- opercular margin. The eggs of U. chuss and their development. — The eggs were obtained by Mr. Welsh at Gloucester, Mass., evidently directly from ripe fish. It may be assumed that the eggs of all the species of Uropliycis are similar. Therefore, the descriptions and draw- ings of those of chuss, offered herewith, may be useful in identifying those of the other species when they are taken. The eggs are small and vary little in size. The diameter of 10 eggs ranged from 0.72 to 0.76 mm, the average being 0.74 mm. They are clear and buoyant, and con- tain many (54 counted in one egg) oil globules when first spawned. During incuba- tion the oil globules decreased rapidly in number, until most eggs retained a single large one, much larger than any originally present, 6 hours after fertilization. Oc- Fiqure 123. — Urophycis chuss. From egg in 2-cell stage; 1H hours after fertilization. (From a camera lucida drawing by W. W. Welsh.) Figure 124. — Urophycis chuss. From egg with early embryo; 50 hours after fertilization. (From a camera lucida drawing by W. W. Welsh.) casionally, however, a few minute scattered ones, in addition to the large one, were retained 26 hours after feritlization. The first cleavage took place 1 K hours after fertilization at a temperature of about 60° F. The number of oil globules already had decreased (fig. 123). Segmentation progressed rather rapidly, as the morula stage was attained about 26 hours after fertilization. The embryo was well formed 50 hours after fertilization. It extended fully half the distance around the periphery of the egg, and the eyes were evident. Black chromatophores dotted the embryo (fig. 124). During the next 24 hours, that is, 74 hours after fertilization, no important changes took place, except that the embryo grew larger, and the amount of yolk was reduced. Pigmentation of the embryo re- mained unchanged. Some convulsive movements now were noticed (fig. 125). At 90 hours of incubation, with a more or less constant temperature of 60° F., the eggs were ready to hatch. The pigment spots on the embryo had become notably 154979—38 8 616 BULLETIN OF THE BUREAU OF FISHERIES larger and bad branched. The eyes, too, were now slightly pigmented. The single remaining oil globule lay under the abdomen (fig. 126). Figure 125. — Urophycis chuss. From egg with well- Figure 126. — Urophycis chuss. From egg with formed embryo; 74 hours after fertilization. (From large embryo; 90 hours after fertilization. (From a camera lucida drawing by W. W. Welsh.) a camera lucida drawing by W. W. Welsh.) Newly hatched U. chuss. — The newly batched larvae ranged from 1.83 to 1.98 mm in length. The oil globule lay in the posterior part of the yolksac, or at midlengtb of the larva. Large pigment spots were present, principally along the dorsal and ventral outline, and also on top of the bead. A few small dots were present on the eye, a few larger ones on the yolksac, and about three large branched ones on the oil globule (fig. 127). Figure 127.— Urophycis chuss. From newly hatched larva. (From a camera lucida drawing by W. W. Welsh.) Specimens oj U. floridanus (?) 2.75 to 3.0 mm long. — The body is rather deep, robust, the greatest depth being contained 3.0 to 3.3 times in the length to the end of the notochord. The caudal portion of the body is relatively short and deep, notably shorter (without caudal fin membrane) than head and trunk, its depth just posterior to vent being contained about 2.2 times in its length. Myomeres are indistinct posteriorly. Upward of 40 may be counted. (Number of vertebrae in regius, 14+31; in floridanus, 16+34; one specimen of each species examined.) The head is large, compressed, and is contained about 3.0 times in the length to the end of the notochord. The mouth is almost vertical, the tip of the lower jaw being about at a level with the upper margin of the eye. The eye is large, fully twice as long as the snout, being contained about 2.0 times in the head. The vertical fin membranes remain continuous around the tail where there are rather distinct indications of the formation of rays. The pectorals are represented by broad short membranes, and DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 017 the ventrals appear as three hairlike rays, which do not nearly reach the vent, and which at this stage are inserted laterally below the base of the pectorals. The color markings consist of some dark chromatophores on the head, some more on the back at the base of the dorsal finfold, one to several on the ventral edge of the body above base of the anal, and generally a few to several on the middle of the side, sometimes forming a more or less continuous black line. A few dark dots frequently are present around the mouth, and on the side along the upper margin of the abdomen. The distal part of the ventrals already are slightly dusky (fig. 128). The specimens described in the foregoing paragraphs are from Beaufort, N. C., and may be U. floridanus. In the same lot are specimens that apparently differ only in the absence of dusky color on the ventral fins and generally in having no color mark- ings above the base of the anal. As larger easily recognizable specimens of U. regius have no black on the ventral fins, it seems probable that small specimens destitute of this color also are U. regius. Figure 128. — Urophycis floridanus (?). From a specimen 3 mm long. Figure 129. — Urophycis chuss. From a specimen 2.75 mm long. In addition to the specimens already described, there are at hand specimens of the same size, taken off Cape Henry, Va., which apparently are representatives of a third species. The larvae differ rather markedly in having a proportionately much longer and more slender tail, the caudal portion of the body (without the finfold) being about equal in length to the head and trunk, and its depth just posterior to the vent is contained about 4.0 times in its length. The development of the ventral fins is some- what more retarded in these specimens, no rays being present. However, in slightly larger ones in the same lot they are developed, and are distinctly black distally. Other color markings agree with the specimens already described. Larger and easily re- cognizable specimens of U. chuss have the distal parts of the ventrals black. As U. floridanus, which also has black ventrals, is not known to occur as far north as Cape Henry, it seems probable that the last described larvae are U. chuss (fig. 129). 618 BULLETIN OF THE BUREAU OF FISHERIES Specimens about 4-0 mm long. — The advancement in development is not great. In specimens probably of U. jloridanus the body has become rather more robust, the depth being contained in the length to the end of the notochord about 3.0 times. The caudal portion of the body has become proportionately rather longer, yet it remains decidedly shorter than the rest of the body. The mouth is less strongly vertical, the tip of the lower jaw now being slightly below the level of the middle of the eye. The ventral fins have increased in length and reach to or a little beyond the vent. No change in color apparently has taken place. Most larvae have more dark dots above the base of the anal than the specimen drawn (fig. 130). The difference between U. jloridanus and U. regius remains one of color only, as in the smaller specimens, if both species actually are represented among the young at hand. The distal part of the membranes of the ventrals being black in jloridanus, and pale in regius. Furthermore, in regius of this size, there generally are no black chromatophores above the base of the anal, though a few exceptions have been noticed. U. chuss continues to differ from both regius and jloridanus in having a longer and more slender tail, though it has become proportionately shorter. Yet it is fully equal (without the caudal finfold) in length to the rest of the body, and its depth just posterior to the vent is contained 3.2 times in its length. The mouth is less nearly vertical than in the other species, and the dorsal profile is rounder. In color this species differs very little from jloridanus, the black markings being similarly placed, though rather more numerous. Specimens about 5.0 mm long. — U. jloridanus apparently is missing among the specimens of this size. In fact, no specimens between a length of about 4.0 and 21 mm appear to be at hand. In U. regius the body has continued to increase in robustness, the depth now being contained in the length to the end of the notochord about 2.8 times. The caudal portion of the body has increased further in proportionate length, and is con- tained about 1.6 times in the length to the end of the notochord. It is deep and compressed, its depth just posterior to the vent being contained about 2.0 times in its length. The head is rather deep, compressed, and is contained 2.75 times in the length to the end of the notochord. The eye is nearly twice as long as the snout and is contained 2.75 in the head. The mouth is strongly oblique (not vertical), the tip of the lower jaw being slightly below the middle of the eye, and the maxillary reaches about under the middle of the eye. The notochord is bent upward very slightly distally. The vertical fin membranes remain continuous. The rounded caudal contains fairly well developed rays, but the dorsal and anal are more retarded DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 619 in development. The pectoral fin membranes remain short and broad, and without definite rays. The ventral hairlike rays (apparently three in number) have increased in length and reach well beyond the vent, the fins remaining inserted laterally below the base of the pectorals. Black chromatophores remain present on the upper surface of the head, on the back below the anterior half to two-thirds of the dorsal, and occasionally one or more black dots are present at the base of anal. A dark lateral stripe, variable in length, is generally situated above the anterior half of the anal. A dusky area extends upward and forward from the vent, and often a dusky area is present at the upper angle of the gill opening, sometimes extending downward just posterior to the opercle. The fins remain without color, the ventrals being pale throughout. U. chuss seems to differ from U. regius principally in the rather more slender body and in having a proportionately longer and more slender tail, the depth of the body being contained 3.5 times in the length to the end of the notochord; and the tail, from the vent to the tip of the notochord, is contained 2.2 times in the length. The depth just posterior to the vent is contained 2.3 times in the length of the caudal portion of the body to the tip of the notochord. In color U. clmss differs principally in that the interradial membranes of the ventrals are black distally (fig. 131). Figtjee 131. — Urophycis chuss. From a specimen 5 mm long. Specimens about 7.0 mm long. — The most important advancement is the develop- ment of rays, or at least the fulcra, of most of the dorsal, and to a somewhat lesser extent of the anal rays. The number of rays in the second dorsal, as pointed out else- where, is diagnostic, as thereby regius (with 46 to 51 rays) is distinguished from the other local species, which have a greater number of rays. It is possible now, with transmitted light and fairly high magnification, to count about 43 fulcra in the second dorsal and about 45 in the anal in specimens with deep short tails, winch have no black on the ventral fins. The specimens with short deep tails and without black on the ventrals among the younger stages, as indicated, were suspected of being regius. At a length of about 7.0 mm they may be so designated quite definitely, as shown subsequently. The specimens with the rather longer and more slender tails, and with the ventrals distally black have about 50 to 52 fulcra developed in the second dorsal. The anal is somewhat more retarded in development than the dorsal and the rays and fulcra are not nearly all developed. Although the dorsal fulcra evidently, too, are not quite all developed, it is evident that the number that will be developed is greater than in adult regius. The specimens with the higher number of fulcra, developed at a length of about 7.0 mm, are from the vicinity of Cape Henry, Va., farther north than flori- danus is known to occur. The only species recorded from the coast of Virginia are 620 BULLETIN OF THE BUREAU OF FISHERIES regius and chuss. The specimens with the larger number of rays or fulcra in the dorsal certainly are not regius, and therefore apparently must be chuss. It has been pointed out that in smaller specimens the caudal portion of the body was proportionately longer and more slender in chuss than in regius. This difference persists, but it is no longer pronounced. The distance from the vent to the tip of the notochord, in specimens about 7.0 mm long, is contained in the length of the fish, without the caudal fin, about 2.25 times in regius, and 2.1 times in chuss, and the depth just posterior to the vent is contained in the length of the tail 2.1 times in regius, and 2.5 times in chuss. The body in both species has become more elongate, the depth in regius being contained 3.4 times in the length without the caudal fin, and 3.6 times in chuss. The color is variable among specimens of both species, though not essentially different from smaller ones already described (figs. 132 and 133). Specimens 9.0 to 11 mm long. — The body in regius, as well as in chuss, has con- tinued to grow proportionately more elongate, though it remains decidedly com- pressed. The depth in regius is contained 3.8 to 3.9 times in the standard length, and in chuss 4.0 to 4.3 times. The caudal portion of the body (without the caudal fin) is almost exactly equal in length to the head and trunk in regius, whereas in chuss it is noticeably longer. It is also deeper in regius, the depth just behind the vent being contained 2.9 times in the distance from the vent to the base of the caudal, whereas it is contained 3.1 to 3.3 times in that distance in chuss. The mouth has become much less strongly oblique. However, it remains a little more strongly oblique in regius (wherein the tip of the lower jaw is about at the level of the lower margin of the eye) than in chuss, in which it is well below the eye. The first dorsal is partly formed in both species under discussion, and is situated over the base of the pectoral. The second dorsal is well enough developed to permit a fairly accurate count of the rays, and especially of the fulcra. In regius 47 and 50 fulcra were counted, and in chuss 55 and 56, in two specimens of each species examined. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 621 The anal rays are more retarded in development, and a full count is not yet obtainable. The pectoral fins remain short and rounded in both species. The ventral fins, though still lateral, are lower on the side and rather farther forward than in smaller fish, being inserted somewhat in advance of the base of pectorals, and the rays, of which three distinct ones of about equal length are present, reach well beyond the origin of the anal. In color the two species do not seem to differ, except for the black on the ventrals in chuss, which is missing in regius. The number of dark dots have increased some- what, though there is much variation among specimens. All specimens at hand of both species have black chromatophores on the head and back. Some specimens have black dots on the cheeks and opercles, some have a dark lateral stripe variable in length, and in others these markings are missing, apparently without regard to species. In the larger specimens of this group, dusky specks have begun to appear on the first dorsal (fig. 134). Specimens about 15 mm long.-— No pronounced changes in development have taken place since a length of 9.0 to 11 mm was attained in either species. However, the first dorsal is considerably higher and better developed, and the pectoral fins have become much longer, as shown in the accompanying illustrations. The caudal fin is quite variable in shape, for it may be rounded, straight, or slightly concave. The chin barbel first becomes evident in specimens of about this size. Figuke 13 i.—Urophycis chuss. From a specimen 9.5 mm long. The color is variable among specimens, some being more profusely spotted than others. In general, dark pigment has increased in both species. However, the only distinguishing feature in color noticed is the black on the distal part of the ventral in chuss, which is missing in regius, just as in much smaller specimens (figs. 135 and 136). The proportionate length and depth of the caudal portion of the body, which aided in separating smaller specimens of regius and chuss, are now so nearly the same that the distinction has vanished. The rays in the dorsal and anal, at least in some specimens, are not quite all formed. The development in 15-mm specimens, as in smaller ones, is rather more retarded in chuss than in regius. Some of the rays and fulcra remain difficult to see. However, with the use of comparatively high magnification and transmitted light, 9-54 rays were counted in the dorsal and 54 or 55 in the anal in three specimens of chuss. In two specimens of regius 7-46 and 7-47 rays were counted in the dorsal, and 44 and 47 in the anal. The counts, as shown in the key to the species, for adult chuss are — dorsal 9 to 11-56 to 61, anal 52 to 56; and for adult regius, dorsal 8 or 9-46 to 51, anal 39 to 49. Therefore, the difference in the counts between the two species in 15-mm specimens is quite evident. 622 BULLETIN OF THE BUREAU OF FISHERIES The anal ray counts definitely separate chuss from floridanus , as the adults of the latter have only 40 to 49 anal rays. The specimens herein described as chuss were taken off Ivitty Hawk, N. C., and northward, where floridanus is not known to occur. The smaller specimens were so identified largely by “locality”, as in the absence of specimens of floridanus of similar size it was not possible to know how the two species differed. The anal fin ray counts in 15-mm fish, however, aid in establishing the identification on a morphological basis. Specimens about 25 mm long. — At this length three species; namely, regius, floridanus, and chuss, are recognizable among the specimens studied, principally by the number of rays in the dorsal and anal fins, and by the length of the chin barbel, as shown subsequently. The species are not distinguishable by the shape of the body, the shape and length of the head, the eye, the snout, nor the mouth. The body has become quite slender, and remains compressed, the depth in any one of the three species named being contained about 4.0 to 4.6 times in the standard length, and the head 3.3 to Figure 135.— Urophycis chuss. From a specimen 15 mm long. Figure 136. — Urophycis regius. From a specimen 15 mm long. 4.0 times. The snout is gradually increasing in length, being contained in the head about 4.0 to 4.4 times, and the eye 3.3 to 3.6 times. The mouth remains only slightly oblique, and it has become somewhat inferior, with the upper jaw a little in advance of the lower one, and the snout projecting slightly beyond the upper jaw. The maxillary reaches to or a little beyond the posterior margin of the pupil. The barbel at the symphysis of the lower jaw, which first made its appearance in regius and chuss when about 15 mm long, remains minute, being scarcely a fourth the length of the pupil in those species. No specimens of floridanus around 15 mm in length are at hand. In specimens of this species, about 25 mm long, it is much longer than in the other species, being fully equal to the length of the pupil. The greater length of the chin barbel is a readily available morphological character at this size, as well as among larger fish, for separating floridanus from both regius and chuss. Scales are present at a length of 25 mm in all three species, though not shown in the accompanying illustration. The series cannot be definitely enumerated, but it is evident already that the scales are larger in regius than in the other species. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 623 The difference in the number of dorsal rays between regius and chuss is pointed out in a preceding section, as well as in the key to the species. However, chuss and Jloridanus have so nearly the same number of rays in the dorsal that they cannot be separated readily, if at all, by that character. Nevertheless, the last-mentioned species differ in the number of anal rays, chuss having 52 to 56, whereas jloridanus has only 40 to 49 (the same number as in regius), as shown in the key to the species. The anal rays are well enough developed when the fish reach a length of about 25 mm to permit the use of this distinguishing character. Much variation in color exists among specimens, some individuals being much more profusely spotted than others. The two specimens of chuss of this size at hand are much more profusely dotted than any others. However, insufficient specimens are available to determine whether it is of specific significance. The ventral fins of jloridanus are black distally. However, no black is evident on these fins in the old preserved specimens of chuss, though of course the smaller ones have it. The first dorsal is partly dusky, there being as yet no distinction among the species in this respect (fig. 137). Specimens 35 to 50 mm long. — The three species; namely, regius, jloridanus, and chuss, discussed in the preceding section, are readily recognizable among specimens 40 to 55 mm in length. They are distinguishable by the characters pointed out in the preceding section and some additional ones, as shown subsequently. A fourth species, namely, earlli, also is present. This species is discussed separately. The body has continued to grow more slender and less strongly compressed, especially anteriorly, in regius, jloridanus, and chuss. No measurable difference in the range in depth seems to exist among these species. In nine specimens, including three of each species, the range of the depth in the standard length is 5.0 to 5.75. In the same specimens, the head is contained 3.7 to 4.1 times in the standard length. The snout now is equal to, or only slightly shorter than the eye, being contained 3.75 to 4.5 in the head. The mouth is slightly oblique, and is definitely inferior, being situated essentially as in adults. The maxillary is broad posteriorly, and reaches nearly or quite opposite the posterior margin of the eye, being contained 1.8 to 2.2 times in the head. The maxillary barbel remains minute in regius and chuss, wherein it is scarcely half as long as the pupil. In jloridanus it is much longer, being equal to fully half the diameter of the eye. The scales are quite fully developed and the series can be counted fairly accurately. (The number present in the different species is shown in the key to the species.) It is plainly evident, under magnification, without counting, that regius has notably larger scales than the other species. The fins are all developed essentially as in adults. The pectoral fins are longer in regius, wherein they reach beyond the origin of the anal, than in the other species 624 BULLETIN OF THE BUREAU OF FISHERIES in which they generally fail to reach opposite the origin of the anal. The ventrals are scarcely lateral; they are inserted well in advance of the pectorals, or nearly under the margin of the preopercle. The rays now generally appear as two in number, though a third short one sometimes remains evident. The two long rays (filaments) are free from each other distally, and they are rather variable in length, within any one species, the upper or outer one, which is the longer, generally reaches to or beyond the origin of the anal. The caudal fin varies in shape in all species, as its margin may be round, straight, or concave. The differences in the counts of the rays in the dorsal and anal fins are shown in the key to the species. Pigmentation has become quite general, though variable among specimens of any one species. The color of preserved specimens is pale to rather dark brown above, and generally silvery below. In life regius and j doridanus (no fresh specimens of chuss seen by us) may be bright green to bluish above, and the sides and lower parts bright silvery. The extent to which the body is covered with brownish dots varies even among specimens of the same species caught in one haul. The black on the distal part of the ventrals, present in smaller specimens of jloridanus and chuss, rarely remains visible in specimens 40 mm long, and was not seen in any fish 50 mm and upward in length. The black on the first dorsal is now quite definitely surrounded by white, at least distally, in regius and distinguishes that species from jloridanus and chuss in which the black extends nearly or quite to the margin of the fin and is not surrounded by white (fig. 138). Figure 138,—Urophycis floridanus. From a specimen 40 mm long. The fourth species, namely, earlli, is represented by a single specimen 37 mm long. This fish does not differ from the other species in the proportions usually calculated. However, it has much smaller scales, and the dorsal and anal rays are more numerous. (The counts are given in the key to the species.) The mouth is nearly horizontal and inferior, as in the other species, and the maxillary reaches almost below the pos- terior margin of the eye. The chin barbel is long and slender, even longer than in jloridanus, as it exceeds half the length of eye. The general color is dark brown, much darker than the darkest specimens of the other species, and this color extends on the dorsal and anal fins, only the margins posteriorly being pale. In fact, these fins are darker than the body. Only the chest and abdomen are silvery. The first dorsal is no darker than most of the second one. The caudal fin is dark brown at the base, and the rest of the fin is plain translucent. The pectorals and ventrals are brown at the base and colorless elsewhere (fig. 139). Specimens 60 to 75 mm long. — The four species discussed in this work; namely, chuss, regius, fioridans, and earlli, are all represented among the specimens of this size. In general, the characters of the adults, pointed out in the key to the species, can be used in separating the species. Therefore, only changes in development are pointed out in the following paragraphs. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 625 The body has become slightly more elongate, and the head proportionately lower and broader. The interorbital now is equal to or wider than the eye in all four species. The snout has increased in proportionate length, and is definitely longer than the eye. If specimens of equal length are compared it is evident that regius, and especially chuss, have somewhat larger eyes than the other species. The longer pectoral of regius is quite distinctive at this size, falling short of the length of the head by only half the snout, whereas in the other species this fin does not exceed the length of the head without the snout. The color continues to vary markedly among the specimens of any one species. However, earlli is notably darker, and has almost black fins. The dark stripe, broken by roundish pale spots at intervals, rather shorter than the diameter of the eye, in which the lateral line lies in adults of regius smd jloridanus, only, sometimes is present when the fish have reached a length of 60 mm, but often not until much later. Some- times the pale spots appear in advance of the black stripe. In regius and jloridanus four black dots in a vertical row sometimes are present behind the eye in specimens 60 mm long, but often not until much later, the upper one of which is in line with another spot over the eye and a third one over the posterior nostril. In addition about three dark dots are situated on the opercle. A pore is present in the center of each black spot. The pores, apparently are present in chuss and earlli, too, but in those species they are not surrounded with black. Figure 139. — Urophycis earlli. From a specimen 37 mm long. Specimens 100 mm and upward in length. — The body in regius, jloridanus, and earlli becomes more robust with age, and also rather deeper, the proportionate depth being about equal in all these species and contained 3.9 to 5.0 times in the standard length. Adults of chuss are notably more elongate, especially large fish, than those of the other species, the depth being contained 5.1 to 5.5 times in the length. The head becomes broader and more depressed with age. This change is especially pronounced in chuss, for in large specimens of this species it is notably wider than deep, at the middle of the eyes, whereas in the other species the width of the head at the same point is about equal to its depth. The larger eye in chuss becomes more noticeable with age. It is quite as wide as the interorbital in fish about 200 mm long, and is contained 4.1 to 4.6 times in the head. In specimens of about the same size of regius, which have a rather larger eye than those of jloridanus and earlli, it is much narrower than the interorbital, and is contained 5.1 to 6.5 times in the head. The snout increases in proportionate length as the eye decreases, and in all species in specimens 100 mm and upward in length the snout is noticeably longer than the eye. It is a little broader in chuss than in 626 BULLETIN OF THE BUREAU OF FISHERIES the other species, and, perhaps because of the larger eye, the maxillary reaches only to the posterior margin of the eye, whereas it reaches well beyond this point in large specimens of the other species. The snout projects more prominently beyond the mouth with age in all the species, and it becomes quite conical, though a little de- pressed in chuss. The chin barbel remains short throughout life in regius and chuss, in which it never exceeds the pupil of the eye in length. In large specimens of Jioridanus and earlli it is nearly or quite equal to the diameter of the eye. The third ray of the first dorsal is greatly produced in adults of chuss, reaching about to the end of the first third of the second dorsal. It is not evident at what size the ray becomes produced from the specimens at hand, as fish ranging from 70 to 185 mm in length are missing. In the smaller fish it is not produced, but in the larger ones it is long. In all the other species the first dorsal becomes rather pointed, but none of the rays are especially produced. The differences in the length and shape of the other fins remain about the same as in smaller specimens already discussed. The differences in color remain virtually the same as in the smaller specimens described in the immediately preceding section. The general color of earlli remains much darker than in any of the other species. A specimen about 100 mm long is uniform dark brown, with the vertical fins almost black. Larger fish sometimes are more or less blotched with pale color. No deep black is evident on the first dorsal, as in the other species herein considered. Of these, regius differs strikingly in having the deep black color of this fin margined with snow white (figs. 140 and 141). DISTRIBUTION OF THE YOUNG It is shown under the heading, “Spawning,” that the early larvae, under 5.0 mm in length, apparently consisting of both regius and jioridanus, the only common species of Urophycis at Beaufort, were taken only at stations 6 to 13 miles offshore, DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 627 beyond which no collecting was done. Larger young, too, were taken offshore as well as near the shore. However, no young measuring less than 40 mm long were found in the bays and estuaries, and not many until a length of 50 to 60 mm had been attained. Thereafter, for a period of about 3 months, February, March, and April, they were common to numerous on muddy bottom in the estuary of Newport River, but no more so than on the very muddy bottom in the vicinity of the “sea buoy” at the entrance of Beaufort Inlet. However, those taken in Newport River averaged larger in size. U. regius nearly always was greatly in the majority, though in its habitat it did not seem to differ from jloridanus. The young, up to 40 mm in length, sometimes were taken at the surface, though more frequently at the bottom. All the larvae, except one specimen measuring under 10 mm in length were taken in surface nets. Presuming that the eggs of regius and doridanus are pelagic, like those of chuss, the young would be expected to remain at the surface, where they are hatched, at least until the yolksac, with its oil globule, is absorbed. That they not only stay there until the yolk is absorbed but for some time afterward, is indicated by the catches made. However, the hakes, with inferior mouths and ventral fins developed as feelers (presumably for prowling around on the bottom), are typical bottom-dwelling fishes. According to our data that is the common, if not the exclusive, habitat after a length of 40 mm or so is attained. Futhermore, they seem to prefer muddy bottom, as already stated. Although young regius and jloridanus are common to numerous, in the areas named, during winter and early spring, they disappear almost entirely from shallow water by June 1, and the adults either are scarce or missing at all times. Definite information as to where the young go was not obtained. It seems probable, however, that they merely withdraw to deeper water. It at least seems rather certain that the adults are common in the offshore waters in the vicinity of Beaufort during the re- productive season as considerable spawning must take place locally, for the young at times were taken in great abundance, outnumbering all other species. In the absence of information to the contrary, it may perhaps be assumed that the deeper offshore water is the regular habitat of the local species of hakes. The abundance of the young in the shallow water during their first winter suggests an abundant population. It might even be possible to develop a hake fishery if their habitat could be located, which apparently should be sought in rather deep water with muddy bottom. GROWTH The measurements of fish tabulated in the accompanying tables are based wholly on young fish believed to be 6 or 7 months and less of age. Certainly there is no break in the growth curve. Very few larger fish were taken. Assuming that the fish for which measurements are given are all under 6 or 7 months old a rapid rate of growth is indicated, for a few individuals of both regius and jloridanus apparently reached a length slightly upward of 8 inches at an age of about 6 months. If regius attains a length of only about 16 inches and jloridanus is even smaller, as shown by the data available, early maturity surely would result if such a rapid rate of growth were maintained. 628 BULLETIN OF THE BUREAU OF FISHERIES Table 1. — Length frequencies of 2,054 hakes ( Urophycis regius ), all less than a year old [Measurements to nearest mm; in 5-mm groups] Millimeters November December January February March April May June 0-4.. 1 6 5-9. 2 10-14 1 15-19 4 20-24 6 1 25-29 30-34 1 1 35-39. 1 13 2 40-44 4 71 4 4 45-49.. 22 89 10 2 50-54 25 53 20 12 2 55-59 10 47 29 24 2 60-64 . 4 57 31 10 65-69 i 46 33 8 70-74 3 56 32 7 75-79 1 35 41 1 1 80-84 24 62 1 1 85-89 1 28 52 4 90-94. 23 75 7 1 95-99. 20 59 9 100-104 2 18 66 ii 105-109 9 63 17 110-114 12 46 31 1 115-119. 3 50 26 1 120-124. 7 46 18 2 125-129 7 33 18 2 130-134 5 37 24 135-139 4 22 16 140-144 2 28 19 2 145-149.. 3 23 20 1 150-154 11 17 1 155-159 13 16 160-164 8 12 1 165-169 3 15 170-174 14 175-179 1 7 180-184 1 7 2 185-189 1 2 10 190-194 i 3 195-199 - 4 200-204 3 205-209 2 210-214 215-219 i 1 Table 2. — Monthly summaries of length measurements of 2,054 hakes (Urophycis regius ) during the first several months of life [Measurements based on the same fish as in table 1] Month Fish measured Smallest Largest Average 1 Mm 3.0 Mm Mm December 18 2.75 23 18.8 January 74 38.00 101 53.9 February 634 24.00 147 67.0 Month Fish measured Smallest Largest Average March 902 Mm 30. 00 Mm 192 Mm 102.3 April 402 38.00 215 123.9 May 21 33. 00 219 115.3 June 1 152. 00 DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 629 Table 3. — Length frequencies of 218 hakes {Urophycis floridanus) , all less than a year old [Measurements to nearest mm; in 5-mm groups] Millimeters 0-4 5-9 10-14 15-19 20-24 25-29.... 30-34.... 35-39.... 40-44... 45-49... 50-54.... 55-59—. 00-04.... 05-09.... 70-74.... 75-79.... 80-84.... 85-89— 90-94.... 95-99.... 100-104.. 105-109., 110-114.. 115-119. 120-124. 125-129. 130-134. 135-139. 140-144- 145-149. 150-154.. 155-159.. 100-164.. 165-169. 170-174. . 175-179- 180-184.. 185-189.. 190-194.. 195-199- 200-204., 205-209.. 210-214- December January February March April May Table 4. — Monthly summaries of length measurements of 218 hakes {Urophycis floridanus) , during the first several months of life [Measurements based on the same fish as in table 3] Month Fish measured Smallest Largest Average Month Fish measured Smallest Largest Average 11 Mm 3 Mm 40 Mm 14.8 March 51 Mm 37 Mm 135 Mm 79.6 5 25 62 45.0 April 66 83 212 152.8 February 47 24 91 51.7 May 38 80 202 139.5 630 BULLETIN OF THE BUREAU OF FISHERIES ACHIRUS FASCIATUS LACEPEDE. AMERICAN SOLE The American sole, Achirus fasciatus ,n as understood here, ranges from Massa- chusetts to Texas, and is also recorded from the Atlantic coast of Panama. This species is very common on the coast of North Carolina, where it is often found in abundance in estuaries, and the mouths of fresh-water streams, on muddy bottom. It generally may be secured in numbers in the estuary of the Newport River at Beau- fort, and the young especially range in abundance up the river into fresh water. Small examples of this sole sometimes are taken in fresh water far from the sea. For example, it is a more or less permanent resident, at least during the summer, of the Potomac River as far up as Washington. The senior author also has seen a small specimen taken in the Savannah River at Augusta, Ga., slightly more than 200 miles from the sea, following the course of the river. He also has a specimen from the Pas- cagoula River, taken at Merrill, Miss., probably fully 75 miles by the course of the river from the Gulf, where we were informed by a local game warden the fish, though considered a curiosity, is taken from time to time. It may be said, therefore, that this sole ranges from salt, through brackish water, and sometimes far into fresh water. However, in the vicinity of Beaufort, N. C., at least, it is most numerous in water that is more or less brackish. It is, of course, a bottom-dwelling fish, like other flat- fishes. The usual book name of this species is American sole. In the field the names, sole, flounder, and hogchoker, are heard. In North Carolina hogchoker is almost universally used. In bygone times, and to a limited extent to the present day, hogs have fed on waste fish, cast on the shore by fishermen. Among them, of course, was the sole, for it has no commercial value. It is related that occasionally this sole act- ually became a hogchoker. In case the hog masticated poorly and tried to swallow the fish tail foremost, the fish sometimes lodged in the hog’s throat, because of its extremely rough (ctenoid) scales. The hog, apparently being unable either to swallow or regurgitate the fish, eventually was choked to death. CHARACTERS OF THE ADULT The hogchoker is characterized chiefly by the short deep body, the depth generally being contained in the length to the base of the caudal fin considerably less than two (1.6 to 1.81) times. The eyes, which are very small, and the color, are on the right side of the body. The color is variable ; generally it is brownish with darker blotches, and with about seven or eight dark cross lines. The mouth is very small, terminal, and the jaws are twisted; the maxillary reaches under the lower eye. The dorsal and anal fins are long, the former having 50 to 56 rays and the latter 36 to 42. The caudal fin is round, and the pectoral fins are missing. METHODS OF COLLECTING Adult fish, as well as young ones ranging upward of 18 mm in length, were collected mostly with otter trawls, though larger ones frequently were taken with seines also. 13 Considerable discussion relative to the correct scientific name of the American sole has taken place during the past several years The reader interested is referred t-o Chabanaud (1928 and 1935), Myers (1929),and Hubbs (1932.) If the set-up of genera proposed by Chabanand, who recognizes more genera than most authors, is accepted the name, Achirus, is not available for the American sole, and is replaced by Trinectes. According to Hubbs and Chabanaud, fasciatus should be replaced by maculatus. Notwithstanding that this name was assigned to a fish of the Indian fauna by the original describes, it is now claimed that the designation of that locality was an error. Therefore, the last mentioned authors arrive at the conclusion that the correct name of the American sole is Trinectes maculatus (Bloch and Schneider). The present writers, nevertheless, prefer to retain the long familiar name, Achirus fasciatus, for the reason that the extensive splitting of genera does not seem to us to be advantageous, and because there seems to be insufficient evidence that fasciatus actually is a synonym of maculatus. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 631 The eggs were often secured in abundance with meter tow nets made of number 20 bolting silk. They w'ere also obtained several times from the overflow of tanks supplied with running water in which gravid fish were confined. Unfortunately the eggs secured from fish in captivity never appeared to be fertilized, presumably be- cause no ripe males were present. However, these eggs served a very useful purpose, as they aided us in positively identifying hogchoker eggs taken in the tow\ Hogchoker eggs, indeed, had been taken in the tow by us over a period of several summers before their identity became definitely known on June 8, 1929, from a com- parison with eggs secured from fish held in confinement.14 The more advanced stages of the larvae shown in the accompanying illustrations were drawn from fish reared in the laboratory by the junior author. In the rearing experiments comparatively large numbers of young were placed in glass evaporating dishes having a depth of about 3 inches and a diameter of 8 to 10 inches. Only about an inch of water was used in each dish, thus exposing a large surface, in comparison with the small amount of water, to the air for absorption of oxygen. No artificial aeration was used. To keep the water at a fairly uniform temperature the dishes were partly submerged in the large laboratory tanks supplied with running sea water fed from a 12,000 gallon tank by gravity. While the larvae were very small they were fed daily with towings strained through number 20 bolting silk. After the fish had gained some growth, towings were introduced without straining. Ample time for feeding, that is, an hour or so, was given after introducing the towings, and then the fish were removed with a pipette to clean dishes supplied with water brought to the laboratory in a clean container directly from the laboratory pier. SPAWNING The spawning season of the hogchoker seems to be a long one, the eggs having been obtained from spawning fish held in tanks, as early as May 18 (1931), and as late as August 14 (1930). In the tow the eggs were noticed as early as May 20 (1931), and as late as August 5 (1928). It is evident, then, that at Beaufort the spawning season extends, at least from midspring to midsummer. Ripe or nearly ripe fish were taken only in the estuary of Newport River, where the eggs also were secured. However, eggs also were taken in several other localities within the harbor, as well as at sea as far out as 6 miles off Bogue Banks. Hogchoker eggs often were collected in abundance, being among the most numerous fish eggs in season. Spawning evidently takes place only in the evening, principally from about 6 to 8 o’clock. It was during that time when the eggs were spawned in the laboratory tanks, and it was only in the early evening, as shown by many towings, that eggs in early cleavage stages were secured. In addition to the very recently spawned eggs, older ones with rather w'ell-developed embryos, extending about two-thirds the distance around the periphery of the eggs, were present in the early evening towings. The older eggs evidently had been spawned the previous evening, and wrere about 24 hours old. According to other studies made at Beaufort, partly published b}^ the writers (1930) and partly still unpublished, it would seem that early evening spawning is quite usual among marine fishes. 14 Dr. Albert Kuntz, working for the Bureau of Fisheries at Beaufort, N. C., temporarily, secured the eggs, drew up descriptions and had sketches of the development of the eggs and early young prepared (unpublished) as early as 1913. However, the eggs were not identified. In 1916 Dr. Lewis Radcliffe secured the eggs in Chesapeake Bay, drew up descriptions and some sketches (also un- published), which he labeled “hogchoker.” How he arrived at the tentative conclusion is not evident from his notes. 154979—38 9 632 BULLETIN OF THE BUREAU OF FISHERIES Although the eggs were very numerous in the tow at times during the spawning season, the larvae were not found, notwithstanding that an extensive search was made for them. The smallest young taken in nature was 18 mm long. Therefore, nothing can be reported at this time on the habitat and distribution of the larvae. DESCRIPTIONS OF THE EGGS AND YOUNG The eggs are spherical, richly supplied with oil globules, and float at the surface. According to 200 unfertilized eggs, spawned in a tank on two different dates (the product of two or more fish), the diameter varies from 0.66 to 0.84 mm, the average being 0.73 mm. Eggs especially selected for range in size from several hundred taken in the tow, and in an advanced cleavage stage when measured, ranged in diameter from 0.67 to 0.71 mm. The oil globules are variable in number, as few as 15 and as many as 34 having been counted. They also are variable in distribution, sometimes lying close together, giving the egg a beaded appearance, and sometimes more or less uniformity distributed over the surface of the egg. They also vary in size from very minute dots to about 0.06 mm in diameter. The variation in position, number, and size is shown, at least in part, in the accompanying drawings. The large number of oil gloubles give the egg buoyancy. No perivitalline space is noticeable (fig. 142). Figure 142— A.chirus fascialm. From egg with fully formed Figure 143.— Achirm fasciatus. From egg in 2-cell stage; blastodisc. about half hour after fertilization. The eggs, though quite transparent, have a slight greenish tinge (described as yellowish by Albert Kuntz, MS.) This color seems to be contained in minute yolk granules, discernible under rather high magnification (fig. 143). Cell division proceeds rapidly after fertilization. Eggs collected between 7:30 and 8 p. m. (May 20, 1931), quite surely spawned after 6 p. m. of the same day, ranged from four-cell to many-cell stages when examined in the laboratory at 8:30 p. m. on the same evening (figs. 144 to 147). An hour later, or from about 2 to 3 hours after fertilization, all had reached advanced cleavage stages (fig. 147). On the morning of the following day, or about 13 to 14 hours after fertilization, the eggs contained well-outlined embryos, with eyes just becoming visible, the stage showing the embryonic streak having been passed in the meantime (figs. 148 and 149). About 20 hours after fertilization the embryo extended;almost two-thirds the distance DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 633 around the periphery of the egg; and at 26 hours after fertilization it extended fully three-fourths the distance around the egg. Its tail was sharply recurved, its heart was beating slowly, and it was capable of considerable movement (fig. 150). When next observed, about 36 hours after fertilization, the eggs had hatched; that is, hatching had taken place sometime between 26 and 36 hours after fertilization. The temperature of the water in the dishes in which the eggs were hatched had varied only from about 74° to 76° F. Figure 144. — AcUirus fasciatus. From egg in 4-cell stage, following the 2-cell stage in about 15 minutes. Figure HS.—Achlrus fasciatus. From egg in 8-cell stage, following the 4-cell stage in about 15 minutes. Figure 146.— -Achirus fasciatus. From egg in 16-cell stage, showing irregularity of cells. (Drawn by Effie B. Decker.) Figure H7.— Achirus fasciatus. From egg in advanced cleavage stage; about 3 to 4 hours after fertilization. The older eggs taken at the same time as those for which the development is described in the foregoing paragraphs, which contained advanced embryos, hatched within 12 hours after collection. As these eggs quite certainly were spawned a day earlier than the others, the period of incubation also fell between 26 and 36 hours. It may be stated rather definitely, therefore, that the incubation period almost surely does not exceed 36 hours, at temperatures usually prevailing at Beaufort during the spawning period of the hogchoker. 634 BULLETIN OF THE BUREAU OF FISHERIES The development of the egg of the hogchoker is quite usual for a teleost and is well shown by the drawings presented herewith. Therefore, extended descriptions of the different stages in the development are not necessary. In the series of illustrations prepared during our investigation the cells are all shown as of fairly uniform size and shape. Some eggs were observed, however, in which the cells were more or less unequal in size and somewhat different in shape. Figure 146, prepared by Mrs. Effie B. Decker under the supervision of Dr. Albert Kuntz, is introduced to show a rather extreme case of unsymmetrical cleavage. Too many eggs with more or less unequal cells were seen to permit us to consider variation an abnormality. It may be assumed, therefore, that the cells in the early cleavage stages are apt to vary somewhat. During the early cleavage stages the blastoderm appears as a rather flat mass of cells. However, in the more advanced cleavage stages it is very distinctly dome- shaped, with a cavity beneath it, as shown in a side view, in figure 147. This stage apparently is reached within 2 or 3 hours after fertilization. Figure 148. Achirus fasciatus. From egg showing an early Figure 149. — Achirus fasciatus. Egg showing later stage in stage in the differentiation of the embryo (the shaded differentiation of embryo; about 12 hours after fertilization, streak to the right). The development shown in figure 148, a rather early stage in the differentiation of the embryonic axis (the dark streak to the right), is attained about 6 to 8 hours after fertilization. Many greenish granules are present within the egg. Note the concentration of the oil globules in figure 148 in contrast with their scattered positions in figure 149, as well as in some of the other illustrations. This concentration is not characteristic of this nor any other particular stage of development, but varies in individual eggs. The rather early embryonic stage shown in figure 149 was attained in about 13 to 14 hours. Note that the eyes with lenses are just becoming differentiated. No somites are visible, probably because the embryo is too opaque. Many greenish specks are present on the embryo and some scattered ones on the yolk. A few more or less definitely formed chromatophores, too, are becoming visible. The moderately advanced embryonic stage shown in figure 150 was attained in about 26 hours. The embryo evidentally is too opaque to allow the somites to be seen. It is capable of considerable movement, and slow heart action may be seen. Green specks still are numerous, and comparatively many chromatophores are present, DE VELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 635 both on tlie embryo and the yolk. Note that few oil globules, most of them being small, are present. This again is not characteristic of this stage, but only of the particular egg drawn. Newly hatched larvae. — The newly hatched fish is only about 1.7 to 1.9 mm long. The dorsal and ventral finfolds are very wide, making the larva seem short and deep. The head is slightly deflected, and on its dorsal surface is a pronounced hump. The tail is straight and pointed. The yolksac is comparatively large and some or all the oil globules present in the egg are retained. The number of globules in the yolksac is equally as variable as in the egg. The yolksac also retains the green specks and cliromatophores described for the eggs in advanced embryonic stages. Green specks also are present on the body of the larva, except on the distal part of the tail. On the vertical finfolds green specks are concentrated to form blotches, which are somewhat variable among indi- viduals in size, intensity, and position. There is one on the dorsal finfold above the yolksac; another on the fold above the vent; generally a more or less definite corresponding one on the ventral finfold just behind the vent; and another pair on the fold at about mideaudal length. In some individuals the concentration of color is continued more or less on the body of the larvae, forming indications of cross bars, which become more distinct as the fish grow. In addition to the green specks more or less definite dark chromatopliores are variously distributed over the body and finfolds. Heart action is visible, but due to the opaqueness of the fish the circulation cannot be seen. The vent is located at about midbody length. The newly hatched fish swims, or floats on its back, presumably being held in that position by the bouyancy of the many oil globules in the yolksac (fig. 151). Figure 150. — Achirus fasciatus. From egg with rather advanced embryo. 636 BULLETIN OF THE BUREAU OF FISHERIES Larvae 16 hours old, 2.2 to 2.4- mm long. — In about 16 hours the yolksac is nearly all absorbed. The very small amount remaining is crowded with oil globules. The fish is now about 2.2 to 2.4 mm long. The head no longer is deflexed, but is rather elevated, with the hump even more prominent than in the newly hatched fish. The mouth is open, and the pectoral finfold on each side is plainly visible. It is interesting that, although pectoral finfolds are present in the larvae, pectoral fins are not developed. At least, not a rudiment of a pectoral fin was found in 186 adults, collected at various places along the Atlantic and Gulf coasts of the United States, especially examined for this character (fig. 152). Larvae 24 hours old.— In the aquarium the fish did not increase much, if any, in length for several days after the yolk was absorbed. The fish shown in figure 153 was only 2.18 mm long at 2 days of age, and therefore a little shorter than the younger fish shown in figure 152. Development, nevertheless, progressed somewhat. The yolksac with its oil globules had entirely disappeared. In its place there was a body wall through which the internal organs in part were visible. The hump on the head had become rather lower, and the mouth had moved forward somewhat with the lower jaw projecting slightly. The general color remained about the same as in the younger larvae. However, the pigment areas in the finfolds had become smaller and the pigment dots more concentrated (fig. 153). Larvae 4 days old. — Four days after hatching the larvae still did not exceed a length of 2.5 mm. The head now had become more elongate, and the mouth more prominent with the lower jaw projecting rather strongly. The body had remained slender, as in the younger larvae. The critical stage in the fife of the larvae seems to be reached about the fourth day, and few survived until the fifth in the glass dishes in which they were kept in the laboratory. However, among those fed with towings, of which mention already has DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 637 been made, a few survived much longer and those specimens furnish the clue to the further development of the larvae. Larvae 7 days old. — The fish shown in figure 154 is 7 days old and was 3.0 mm long. This fish differs little from the 4-day old one, already described. Very definite dark chromatophores are now present on the head and abdomen, and rather definite cross bars generally are present. Anteriorly the body has increased considerably in depth. Larvae 14 days old. — Figure 155 is based on a preserved specimen 14 days old, measuring only 2.0 mm in length after preservation. It may be assumed that consider- able shrinkage had taken place. The fish at this stage is deep and strongly compressed. The pigment spots on the dorsal finfold, tending to form bars on the body, are present Figure 165. — Achirus fanciatiLS. From a young fish 14 days old. about as in the early larvae. The pigment on the ventral finfold, however, is not concentrated in blotches in the single specimen studied. A great increase in dark dots on the body has taken place. Young fish 17 days old. — It is evident from figure 156, based on a specimen 17 days old, 3.8 mm long after preservation, that development progressed rapidly. The fish has become much more shapely. The distal part of the tail, instead of being curved downward, is now bent upward, giving it the heterocercal appearance character- istic at this stage of development of telosts with homocercal tails. Indications of rays are present in the finfolds. Note that the pectoral finfold remains prominent. The eyes are quite symmetrically placed on the opposite sides of the head, and there is as 638 BULLETIN OF THE BUEEAU OF FISHERIES yet no indication that one of them (the left one) will “migrate.” Pigmentation has increased greatly, and is equally developed on both sides. Young fish 26 days old. — Growth among the five fish alive at the age of 26 days was quite unequal. In the larger ones, 3.8 mm long when alive, rays are quite fully Figure 156. — Achirus fasciatus. From a young fish 17 days old. Figure 157.— Achirus fasciatus. From a young fish 32 days old. Figure 158.— Achirus fasciatus. From a young fish 34 days old. Though only a few days older than the one illustrated in fig. 157, development is much further advanced. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 639 developed in the vertical fins. A slight depression is evident above the right eye through which the left eye is destined to migrate, as illustrated in the older fish shown in figure 158. However, at this stage the eyes remain symmetrically placed. The fish still swam upright in deep water, but on a hollowed microscope slide containing insuffi- cient water to “float” them vertically, they invaribly swam or rested on the left side. Young fish 82 days old. — The specimen illustrated in figure 157, which is 32 days old and 3.0 mm long, after preservation, has the depression over the right eye, described for the fish 26 days old, more pronounced. The second line shown in the drawing is the actual outline of the interorbital, for above it is only a transparent membrane. However, the eyes remain symmetrically placed on the opposite sides of the head. Note in the illustration that the pigment blotches on the dorsal and anal fins remain placed essentially as in the early larvae. Young fish 34 days old. — The fish illustrated in figure 158 is 34 days old. There- fore, only 2 days older than the one shown in figure 157. However, this fish is larger, being 5.0 mm long after preservation, and it is much advanced in development. Note in the illustration that the left eye has just entered the depression in the forehead. At this size the mouth already is much twisted, essentially as in the adult. Ventral fins now are developed. Note also that the pectoral fin membranes persist. It seems significant, however, that in contrast with the other fins no definite rays have devel- oped. It is, furthermore, noteworthy that the pigment spots on the dorsal and anal fins persist essentially as in the early larvae. No loss of pigmentation on the blind side has taken place. Although one fish lived in the aquarium to reach the age of 41 days, development had not proceeded as far as in the specimen 34 days old, just described. As the smallest specimen taken in nature is 18 mm long, and essentially a young adult, an unfilled gap in the development remains between this specimen and the 5.0-mm one, described in the foregoing paragraph. 154979—38 10 640 BULLETIN OF THE BUREAU OF FISHERIES Young fish 18 mm long. — The 18-mm specimen illustrated in figure 159 already is fully scaled. Fleshy tentacles are developed on the head, though not in the profusion of fully grown fish. The left eye has completed the migration, both eyes appearing close together on the right side of the head, the upper one (originally the left one) being a little in advance of the lower one just as in the fully grown fish. The pectoral fin- fold remains in part, having the appearance of a fleshy tentacle as shown in the illus- tration. Most of the fish examined had lost this rudiment of a pectoral at a length of 25 to 30 mm. However, one specimen 43 mm long retained it. It would not be sur- prising, therefore, if occasionally it were retained longer, or even throughout life. The ventral and anal fins have approached each other as in the adult. In the process the vent and the anal fin have “migrated” forward. (See figures 158 and 159.) The 18-mm specimen is fully pigmented on the right or eyed side. It is inter- esting that the dark blotches on the dorsal and ventral fin folds of the newly hatched larvae have been retained on the right side of this fish in essentially the same position. These juvenile spots are lost, however, when the fish reach a length of about 25 mm. As the fish grows, spots generally become much more numerous than in the 18-mm one illustrated. Frequently about seven or eight more or less definite blackish cross lines are also developed in adults. Much variation in color among individuals exists. The small specimen drawn is destitute of pigment on the blind side, which is quite usual among adults. Yet, many exceptions have been noticed. In fact, vari- ous degrees of pigmentation have been seen, ranging from a few obscure dusky spots or a dusky shade here and there to a general dusky to blackish coloration with definite blackish spots. GROWTH An insufficient number of hogchokers has been measured to determine the rate of growth with any degree of certainty. According to length measurements of 440 specimens, regarded as belonging to the 0-class, taken during April (1931), this class ranges in length from 18 to about 100 mm, the mode being at about 55 mm. A considerable range in size would be expected because of the very long spawning season, which extends at least from May to August. The specimens measured, therefore, may have varied in age from about 7 to nearly 12 months. Subsequent growth remains almost entirely under termined, though among a limited number of larger specimens, measured in April (1931), there is another slight mode around 140 mm, indicating that fish approaching an age of 2 years probably are fully grown. The largest hogchocker taken at Beaufort was 184 mm (J){ inches) long. The largest one ever reported (Hildebrand & Schroeder, 1928, p. 176), so far as known to the writers, was 200 mm (8 inches) long. The average size of mature fish apparently is around 125 mm (5 inches). It is obvious, therefore, that the hogchocker is too small to be of commercial value. So far as the writers are aware it is never eaten. BIBLIOGRAPHY A Manual of Fish Culture. 1898 and 1900. Report U. S. Com. Fish and Fisheries for 1897 (1898), pp. 1-340, pis. 1-62 and I-XVII, illus.: Revised ed. (separately published by U. S. Bur. Fish.) 1904, X; 340 pages, 64 and XVII pis., illus. Washington. Barnhart, Percy Spencer. 1932. Notes on the habits, eggs, and young of some fishes of south- ern California. Bull., Scripps Instit. Ocean., vol. 3, no. 4, 1932, pp. 87-99, 11 figs. Berkeley. Chabanaud, Paul. 1928. Revision des poissons h£t6rosomes de la sous-famille des Achirinae, d’apres les types de Kaup, de Gunther et de Steindachner. Bull., l’lnstitut ocdanographi- que, no. 523. 1928, 53 pages. Monaco. DEVELOPMENT AND LIFE HISTORY OF SOME TELEOSTS 641 Chabanaud, Paul. 1935. Achiridae nec Trinedidae caractdres et synonymie de deux g6notypes systematiques certains: Achirus achirus Linn6 1758 et Trinedes maculatus (Bloch MS) Schneider 1801. Bull., l’Institut ociianographique, no. 661, 1935, 24 pages, 11 figs. Monaco. Clark, R. S. 1913. General report on the larval and post-larval teleosteans in Plymouth waters. Jour. Mar. Biol. Asso., United Kingdom, vol. 10 (N. S.), 1913-15, pp. 327-394, 11 figs. Plymouth. Cunningham, J. T. 1889. Studies of the reproduction and development of teleostean fishes occurring in the neighbourhood of Plymouth. Jour. Mar. Biol. Asso., United Kingdon, vol. 1 (N. S.), 1889-90, pp. 10-54, pi. 1-4, with 39 figs. Plymouth. Earll, R. Edward. 1882. The Spanish mackerel, Cybium maculatum (Mitch.) Ag.; its natural history and artificial propagation, with an account of the origin and development of the fishery. Report, U. S. Com. Fish and Fisheries, 1880 (1883), pp. 395-426, pi. 1-3. Washington. Ehrenbaum, E. 1905-1909. Eier und Larven von Fischen des Nordischen Planktons. Pt. 1, 1905, pp. 1-216, figs. 1-82; pt. 2, 1909, pp. I-IV, 217-413, figs. 83-148. Kiel and Leipzig. Ford, E. 1922. On the young stages of Blennius ocellaris L., Blennius pholis L., and Blennius gattorugine L. Jour., Mar. Biol. Asso., United Kingdom, vol. 12 (N. S.), 1919-22, pp. 688-692, 12 figs. Plymouth. Ginsburg, Isaac. 1932. A revision of the genus Gobionellus (family Gobiidae). Bull., Bingham Ocean. Col., vol. 4, art. 2, 1932, 51 pp., 7 figs. New Haven. Ginsburg, Isaac. 1933. A revision of the genus Gobiosoma (family Gobiidae). Bull., Bingham Ocean. Col., vol. 4, art. 5, 1933, 59 pp., 3 figs. New Haven. Gudger, E. W. 1913. Natural history notes on some Beaufort, N. C. fishes, 1910-11. No. 3. Fishes new or little known on the coast of North Carolina. Collected by Russell J. Coles. Jour., Elisha Mitchell Scien. Soc., vol. 28, 1912, pp. 157-172. Chapel Hill. Gudger, E. W. 1927. The nest and the nesting habits of the butterfish or gunnel, Pholis gunnellus. Nat. Hist., Amer. Museum, vol. 27, 1927, pp. 65-71, 4 figs. New York. Guitel, Frederic. 1893. Observations sur les moeurs de trois blenniid^s Clinus argentatus, Blennius montagui et Blennius sphynx. Arch, zoologie exp6r. et g<5n., ser. 3, vol. 1, 1893, pp. 325-384. Paris. Hefford, A. E. 1910. Notes on teleostean ova and larvae observed at Plymouth in spring and summer, 1909. Jour. Mar. Biol. Assoc., United Kingdom, vol. 9 (N. S.), 1910-13, pp. 1-58, 2 pis. Plymouth. Hildebrand, Samuel F. 1916. The United States Fisheries Biological Station at Beaufort, N. C., during 1914 and 1915. Science, N. S., vol. 43, pp. 303-307. New York. Hildebrand, Samuel F., and Louella E. Cable. 1930. Development and life history of fourteen teleostean fishes at Beaufort, N. C. Bull., U. S. Bur. Fish., vol. 46, 1930 (1931), pp. 383-488, 101 figs. Washington. Hildebrand, Samuel F., and Louella E. Cable. 1934. Reproduction and development of whitings or kingfishes, drums, spot, croaker, and weakfishes or sea trouts, family Sciaenidae, of the Atlantic Coast of the United States. Bull., U. S. Bur. Fish. Bull. no. 16, vol. 48, 1934, pp. 41-117, 44 figs. Washington. Hildebrand, Samuel F., and William C. Schroeder. 1928. Fishes of Chesapeake Bay. Bull., U. S. Bur. Fish., vol. 43, pt. 1, 1927 (1928), 388 pages, 211 figs. Washington. Hubbs, Carl L. 1932. The scientific name of the common sole of the Atlantic Coast of the United States. Proc., Biol. Soc. Washington, vol. 45, 1932, pp. 19-22. Washington. Jordan, David Starr. 1923. A classification of fishes. Stanford Univ. Pub. University series, Biol. Sciences, vol. 3, no. 2, 1923, 243 pp. Stanford University. Kuntz, Albert. 1914. The embryology and larval development of Bairdiella chrysura and Anchovia mitchilli. Bull., U. S. Bur. Fish., vol. 33, 1913 (1915), pp. 3-19, 46 figs. Washington. Kuntz, Albert. 1916. Notes on the embryology and larval development of five species of tele- ostean fishes. Bull., U. S. Bur. Fish., vol. 34, 1914 (1916), pp. 407-429, 68 figs. Washington. Lebour, Marie V. 1919. The young of the Gobiidae from the neighbourhood of Plymouth. Jour., Mar. Biol. Asso., United Kingdom, vol. 12 (N. S.), 1919-22 (1919), pp. 48-80, 3 text figs., pi. 1-4. Plymouth. Lebour, Marie V. 1920. The eggs of Gobius minutus, pidus and microps. Jour., Mar. Biol. Asso., United Kingdom, vol. 12 (N. S.), 1919-22 (1920), pp. 253-260, 3 pis. Plymouth. Lebour, Marie V. 1927. The eggs and newly hatched young of the common blennies from the Plymouth neighbourhood. Jour., Mar. Biol. Asso., United Kingdom, vol. 14 (N. S.), 1926-27, pp. 647-650, 1 fig. Plymouth. 642 BULLETIN OF THE BUREAU OF FISHERIES Myers, George S. 1929. Notes on soles related to Achirus. Copeia, no. 171, 1929, pp. 36-38. Northampton, Mass. Petersen, Joh. C. G. 1917. On the development of our common gobies (Gobius) from the egg to the adult stages, etc. Report, Danish Biol. Stat., 24, 1916 (1917), pp. 3-16, 3 figs., 1 pi. Copenhagen. Radcliffe, Lewis. 1914. The work of the U. S. Fisheries Marine Biological Station at Beaufort, N. C., during 1913. Science, (N. S.), vol. 40, 1914, pp. 413-417. New York. Rathbun, Richard. 1892. Successful hatching of sheepshead eggs on Fish Hawk. Report, U. S. Com., Fish and Fisheries, 1888-89 (1892), p. LIX. Washington. Ryder, John A. 1882. Development of the Spanish mackerel ( Cybium maculatum). Bull., U. S. Fish Com., 1881 (1882), vol. 1, pp. 135-172, pi. 1-4. Washington. Ryder, John A. 1887. On the development of osseous fishes, including marine and freshwater forms. Report U. S. Com., Fish and Fisheries, 1885 (1887), pp. 489-604, 30 pis., 174 figs. Washington. Shropshire, Ralph F. 1932. A contribution to the life history of Gobiosoma molestum. Copeia, no. 1, 1932, pp. 28 and 29, figs. 1-4. New York. Schultz, Leonard P., and Allan C. DeLacy. 1932. The eggs and nesting habits of the crested blenny, Anoplarchus. Copeia, no. 3, 1932, pp. 143-147. New York. Smith, Hugh M. 1907. The fishes of North Carolina. North Carolina Geol. and Econ. Sux., vol. 2, 1907, XI, 453 pages, 21 pis., 188 figs. Raleigh. Vladykov, V. D., and R. A. McKenzie. 1935. The marine fishes of Nova Scotia. Proc., Nova Scotian Insti. Sci., vol. 19, pt. 1, 1935, pp. 17-113, 130 figs. Halifax. U. S. DEPARTMENT OF COMMERCE DANIEL C. ROPER, Secretary BUREAU OF FISHERIES FRANK T. BELL, Commissioner THE MIGRATIONS OF PINK SALMON (Oncorhynchus Gorbuscha ) IN THE CLARENCE AND SUMNER STRAITS REGIONS OF SOUTHEASTERN ALASKA By FREDERICK A. DAVIDSON and LEROY S. CHRISTEY From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 25 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1938 For sale by the Superintendent of Documents, Washington, D. C, Price 10 cents * ■ . ; . . . THE MIGRATIONS OF PINK SALMON (ONCORHYNCHUS GORBUSCHA) IN THE CLARENCE AND SUMNER STRAITS REGIONS OF SOUTHEASTERN ALASKA 1 By Frederick A. Davidson, Ph. D., Aquatic Biologist, and Leroy S. Christey, Field Assistant, United States Bureau of Fisheries J> CONTENTS Page Introduction 643 Channels of migration 644 Tagging methods 645 Pink-6almon tagging experiments in Clarence Strait and adjacent waters, 1924-32 646 Tagging experiments in the vicinity of Cape Fox 648 Tagging experiments in the vicinity of Cape Chacon 649 Tagging experiments on Gravina Island and in the vicinity of Kasaan Bay 651 Tagging experiments in the vicinity of Cape Muzon 654 Pink-salmon tagging experiments in Clarence Strait in 1935 and 1936 654 Summary of Cape Chacon experiments 660 Pink-salmon tagging experiments in Sumner Strait, 1924-36 661 Tagging experiments at Ruins Point and Cape Decision 662 Tagging experiments at Point Colpoys 664 Summary of Point Colpoys experiments 665 Conclusions 665 Literature cited 666 INTRODUCTION Southeastern Alaska lies on the Pacific shore of North America between latitudes 54° and 60° N. It is composed of a narrow coastal strip and a broken chain of moun- tainous islands known as the Alexander Archipelago. The geography of that part which includes the Alexander Archipelago is illustrated in figure 1. Owing to the temperate climate and heavy rainfall of this region most of the islands and mainland shore are covered with dense growths of timber and are drained by hundreds of streams that range in size from brooks to small rivers. These streams support one of south- eastern Alaska’s most valuable natural resources in that they form the breeding grounds for millions of salmon that migrate into them each year to spawn. Through the utilization of these salmon a large and flourishing food-packing industry has developed which is providing nutritious food for the nation as well as a substantial income to the citizens engaged in it. The Pacific salmon of the genus Oncoribynchus spend part of their lives in the sea where they mature and part in the streams where they spawn and then die. The period during which the adults migrate from the sea extends from early summer to 1 Bulletin No. 25. Approved for publication August 20. 1937. 643 644 BULLETIN OF THE BUREAU OF FISFIERIES late fall. The eggs are deposited in the gravel beds of the streams to incubate during the winter months. They hatch out in the following spring and early summer. The young of some of the species remain in fresh water for a few years but they all eventually migrate to the sea where they mature. Since no definite knowledge has yet been obtained as to the location of most species of these salmon during their sojourn in the sea, their spawning migration is of primary importance to the fishing industry for it is only at this time that they are captured in large numbers. The responsibility of protecting this natural resource of Alaska from overexploita- tion, so that it may be preserved for future generations, is vested in the Secretary of Commerce who is advised by the United States Bureau of Fisheries. Since the demands of the industry for salmon are usually greater than the supply, the Bureau has found it necessary to regulate the fishery. This regulation aims to provide for an adequate escapement of the adult salmon to the streams so that they may reproduce and maintain their bounteous numbers. In the life-history studies of the Pacific salmon it was found that they have a high degree of homing instinct; i. e., the majority of the adults return to spawn in the streams of their origin. A discussion of the results of these studies may be found in the following references: Gilbert (1913), Snyder (1921 to 1924), Rich and Holmes (1928), Foerster (1929), Pritchard (1933 and 1934), and Davidson (1934). Owing to this peculiar characteristic of the salmon the population in each stream is self per- petuating and if once destroyed it will not be readily restocked through the straying of salmon native to other streams. Hence, in order to insure the maintenance of the salmon populations in the streams, it was necessary for the Bureau to provide for the protection of the spawning fish each season. In general it imposes definite limitations on the length of time fishing may be carried on in each locality, and prohibits fishing in and near the mouths of the streams and in the small bays that form resting areas for the migrating salmon.1 2 CHANNELS OF MIGRATION Figure 1 shows six main channels through which salmon may enter the inside waters among the islands on their way to the streams. These are, namely: Icy Strait-Lynn Canal, Chatham Strait-Frederick Sound-Stephens Passage, Sumner Strait, Cordova Bay, Clarence Strait-Ernest Sound, and Revillagigedo Channel- Behm Canal. Most salmon spawning in interior localities migrate through one or more of these channels to reach their destinations. In considering the conservation of the salmon populations in each locality the Bureau realized that some provision had to be made for the protection of these populations during their migration through the channels as well as in the streams. Hence, in order to expedite the patrol and regulation of the commercial fishery, the first and most logical step was to set apart the various spawning localities into fishing districts according to the main channels of entry through which their populations migrated from the sea. Before this could be accomplished satisfactorily, however, it was necessary to make a study of the migratory routes and destinations of the salmon populations passing through each main channel. 1 For detailed information concerning the regulation of the salmon industry in Alaska see Laws and Regulations for the Pro- tection of the Fisheries of Alaska. Department Circular 251. Department of Commerce, Bureau of Fisheries, Washington, D. C. MIGRATIONS OF PINK SALMON 645 TAGGING METHODS In 1924 the Bureau started a series of salmon-tagging experiments in each of the main channels of entry for the purpose of securing this information. The work of the experiments was carried on as follows: The trap was first closed and the web of the spiller lifted to bring the salmon near the surface. They were then caught with a dip net and slid, one by one, onto the tagging table from which they were guided head first into a small box held on the outside of the table. This box was short and from 3 to 4 inches of the salm- on’s tail projected beyond the open end. The tag- ging operator was thus enabled to grasp the tail and clamp a tag, about V/2 inches long and %-inch wide, on it. After tagging, the operator holding the small box tossed the salmon back into the water beyond the trap. The entire operation re- quired but a fraction of a minute and, if conditions were favorable, from 150 to 200 fish could be tagged in an hour (see figs. 2 and 3). It is assumed that the great majority of the tagged individuals, when released, continue to follow their original course of migration. The recovery of a few tagged fish in areas far distant from the point of liberation does not necessarily indicate that the tagging operation affected them, for it is not improb- able that salmon occasionally stray from their nor- mal course of migration. Most of those recovered are picked up by the commercial fishery at various points along the migratory routes. Small numbers have also been recovered on the spawning grounds in streams. It has never been possible to recover all the salmon that are tagged. Many of them escape the commercial fishery and spawn in the streams un- noticed and even some of those that are caught by the fishery lose their tags in shipment and remain undiscovered. The Bureau has for several years offered a small reward, from 25 to 50 cents, for the recovery of tagged individuals, and in this way has encouraged the search for them in areas where they are being tagged. No tags, however, are accepted by the Bureau unless they are accompanied by in- formation as to the date and place of recapture of the fish. Without this information the tags are worthless as a means of tracing the migratory routes of the salmon. The first tagging experiments that were carried on in southeastern Alaska attempted to locate, as quickly as possible, the general migratory routes and destinations of the salmon migrating through each of the six main channels of entry This necessitated covering the entire region in a comparatively short time and consequently it was not possible to tag more than two or three times at each Figure 1. — The Alexander Archipelago in southeastern Alaska. The dots indicate the locations of the early pink-salmon tagging experiments from 1924 to 1932. The triangles indicate the locations of the 1935 and 1936 pink-salmon tagging experiments. The boun- dary lines and included numbers in the Clar- ence and Sumner Straits region show the geographic areas used in classifying the local- ities in which the tagged salmon were re- covered. 646 BULLETIN OF THE BUREAU OF FISHERIES location during a season. Tlie locations of these experiments are given in figure 1. The results from them were most valuable and greatly assisted in the formation of the present fishing districts in the region.3 Plans were made to continue these tagging experiments in a more detailed manner so that the destinations of the salmon passing through each main channel of entry at different times during the migratory season could be determined. Information of this nature is very important, for it provides a basis upon which to regulate the fishery in order to protect the salmon during their migration in the entrance channels as well as in the streams where they spawn. The grouping of the waters of southeastern Alaska into fishing districts, for the purpose of regulating the commercial fishery, is based upon two separate studies of the life history and habits of the salmon. The first study deals with the determination of the migratory routes of the salmon. This gives a general picture of the channels of entry frequented by the salmon during their migrations to the various spawning localities. The second study — beyond the scope of this paper — deals with the time in the fishing season during wliich the salmon migrate through each district, and is based upon a study of the daily catch records of the fishing gear in the districts. Information from this study was used to set the opening and closing dates for fishing in the districts so as to provide for the escapement of an adequate proportion of each run of salmon to the streams for reproduction. Since the pink salmon are by far the most abundant species of the Pacific salmon in southeastern Alaska, the grouping of the various waters in this region into fishing districts and the regulations imposed therein have, to a great extent, been directed toward the conservation of this species. Hence, the more detailed tagging experiments have been limited to the pink salmon. The results from a study of the salmon catches, as well as from the early tagging experiments in Clarence and Sumner Straits, indicated that each of these large bodies of water formed the migratory channel of distinct runs of pink salmon. Although these waters were originally included in one fishing district, the Bureau realized that each should constitute a separate district and made the change at the first oppor- tunity, which occurred in 1934. After making this change further tagging work was carried on in the summers of 1935 and 1936 to determine the destinations of the pink salmon passing through each of these main channels at different times in the season. In order to give as complete a picture as possible of the pink-salmon migrations in these waters, the results from the earlier tagging experiments will be summarized and compared to those of this later work. PINK-SALMON TAGGING EXPERIMENTS IN CLARENCE STRAIT AND ADJACENT WATERS, 1924-32 An inspection of the map in figure 1 will show that Clarence Strait and its ad- jacent waters, Revillagigedo Channel and Cordova Bay, do not open directly into the ocean but into a large body of water known as Dixon Entrance. Early in the develop- ment of the salmon fishery in these waters it was found that the first pink salmon to ap- pear each season usually migrated easterly through Dixon Entrance and turned north- ward and eastward along the mainland shores in the vicinity of Cape Fox. As the season progressed the numbers of salmon following this course of migration became more and more abundant and built up a very definite run into Revillagigedo Channel, Portland Canal, and their contiguous waters. A week to 10 days after the beginning of the 3 For the results of the early tagging experiments see Rich (1926), Rich and Suomela (1927), Rich and Morton (1929), and Rich (1932). The locations of these experiments are given in figure 1. U. S. Bureau of Fisheries, 1938 Bulletin 23 Figure 2.— Tagged pink salmon showing relative size of tag and location of attachment. Figure 3.— Tagging pink salmon at a fish trap in southeastern Alaska. MIGRATIONS OF PINK SALMON 647 runs in the region of Cape Fox, large numbers of the pink salmon migrating through Dixon Entrance began to turn northward into Clarence Strait along the east shore of Prince of Wales Island in the vicinity of Cape Chacon. Here again, the salmon fol- lowing this course of migration increased in numbers as the season progressed and built up a separate and definite run into Clarence Strait. From a week to 10 days later a third run of pink salmon began to leave Dixon Entrance and migrate north- ward; this time into Cordova Bay. This run, like the others, increased in abundance as the season progressed, thus forming three definite runs of pink salmon originating in Dixon Entrance. This information was secured from a study of the daily catches of the fishing gear operating in these waters. Studies of similar recent records by the Bureau indicate that three definite runs of pink salmon continue to appear in these waters. In order to provide a uniform method for classifying and reporting the recoveries of tagged salmon, the spawning localities of Clarence and Sumner Straits were grouped into definite areas. Although these areas include waters which are similar to those of the fishing districts, they cannot be construed as being identical. They were not formed with any idea of using them to replace the present districts as factors other than the locations of the migratory channels of the salmon enter into their forma- tion. They will be referred to hereafter as geograplnc areas of recovery which are shown in figure 1 and may be described as follows: Area 1 includes all the waters of Revillagigedo Channel, Portland and Behm Canals, and the waters surrounding Duke, Annette, and Gravina Islands. Area 2 includes all the waters of Clarence Strait along the east shore of Prince of Wales Island from Cape Chacon to Approach Point. Area 3 includes all the waters of Clarence Strait above a line from Approach Point to Caamano Point and below a line from Narrow Point to Ernest Point. This area also includes the waters of Ernest Sound, Zimovia Strait, Bradfield Canal, and Blake Channel. Area 4 includes the remainder of the waters of Clarence Strait above a line from Narrow Point to Ernest Point, and the waters of Snow Passage and Stikine Strait. Area 5 includes the waters of Sumner Strait, Keku Strait, Wrangell Narrows, and the lower extremity of Frederick Sound below latitude 56°30' N. Area 6 includes the waters of Cordova Bay and contiguous channels. Since no information was available concerning the localities in which the salmon comprising these migration waves spawned, the Bureau, in 1924, began three series of tagging experiments in the region. The first was in the vicinity of Cape Fox, the entrance to Revillagigedo Channel, and Portland Canal; the second in the vicinity of Cape Chacon, the entrance to Clarence Strait; and the third in the vicinity of Cape Muzon, the entrance to Cordova Bay. The locations at which these tagging experiments were carried on are shown by the black dots in figure 1. In classifying the results from the tagging experiments summarized in tables 1 to 8, all tagged salmon recovered at the location of tagging were considered as not having migrated from that location and were not included either in the number of individuals tagged, or in the number recovered. The recovery of a tagged salmon cannot be considered as indicating a route of migration unless capture has been made at some distance from the point of tagging. Accordingly, no record is given of the recoveries at the point of tagging nor those where the locality of recapture is doubt- ful. Hence, the total percentage of tagged fish recovered in the experiments does not represent the entire influence of the intensity of the fishery. 648 BULLETIN OF THE BUREAU OF FISHERIES TAGGING EXPERIMENTS IN THE VICINITY OF CAPE FOX The tagging experiments carried on in the vicinity of Cape Fox in 1924 were continued intermittently through 1932, and a total of 17 experiments were made during this period. Since it was not possible to tag frequently throughout the season at any one location, an effort was made to change the seasonal time of tagging each year. In this way the results from these experiments, as a whole, approximately indicate the final destinations of the salmon migrating into this region at different times during the season. A summary of the results from these experiments is given in table 1. Table 1. — Pink salmon tagged in the vicinity of Cape Fox, 1924.-32, and number recovered 1 [Column headings indicate locality and date of tagging] Foggy Point, Kanag- unut Island, June 24- July 1, 1926 Sitklan and Kanag- unut Islands, July 13, 1930 Cape Fox, Sitklan, and Kanag- unut Islands, July 25-26, 1930 Foggy Point, July 30-31, 1925 Kanag- unut Island, Aug. 7, 1924 Cape Fox, Sitklan, and Kanag- unut Islands, Aug. 7-8, 1930 Tree Point, Aug. 8, 1924 Duke Point, Kelp Island, Aug. 6, 1932 1 2 Duke Point, Aug. 8, 1924 Point White, Aug. 8, 1924 Total 137 312 685 1,000 21 667 203 467 245 194 3,931 LOCALITY OF RECOVERY Lower Clarence Strait. — Area 1 East of Cape Fox-_ . 13 87 39 3 130 18 10 18 15 333 British Columbia 3 4 10 44 6 13 79 Revillagigedo Channel and Ton- gass Narrows 17 46 46 241 2 66 35 32 27 21 533 Behm Canal, south arm 2 12 33 33 4 5 4 7 3 103 1 3 1 8 1 14 West shores of Gravina, Annette, and Duke Islands 3 2 10 5 8 4 3 4 39 26 80 187 362 5 221 62 59 56 43 1,101 28.0 19.0 25.6 27.3 36.2 23.8 33.1 30.5 12.6 22.9 22.2 Lower Clarence Strait.— Area 2 East shore Prince of Wales Island from Cape Chacon to Approach Point: Total recovered _ _ _ 2 2 1 4 1 22 3 35 Percent recovered 0.0 0.0 0.3 0.2 4.8 0.6 0.5 4.7 0.0 1.5 0.9 Middle Clarence Strait- Area 3 Approach Point, Caamano Point to Narrow Point, and Ernest Point . 1 2 4 6 1 14 Ernest Sound 5 1 6 1 7 4 7 1 20 Percent recovered. 0.0 0.0 0.2 0.7 0.0 0.6 0.0 1.5 0.4 0.0 0.5 Upper Clarence Strait- Area 4 Total recovered ___ 3 3 Percent recovered- _ 0.0 0.0 0.0 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.1 SUMMARY: Total recovered _ 26 80 190 374 6 229 63 88 57 46 1,159 Percent recovered- 19.0 25.6 27.7 37.4 28.6 34.3 31.0 18.8 23.3 23.7 29.5 1 These data do not include recoveries reported from the point of tagging nor those doubtful as to location of capture. See text, p. 647. The original records of the tagging experiments from 1924 to 1930 are given in the reports of Rich (1926), Rich and Suomela (1927), Rich and Morton (1929), and Rich (1932). 2 From unpublished data, U. S. Bureau of Fisheries. MIGRATIONS OF PINK SALMON 649 From an inspection of table 1 it will be seen that, regardless of the time in the sea- son in which the salmon were tagged, practically all those recovered were captured in area 1. A total of 3,931 pink salmon were tagged in all of the experiments, of which 1,159, or 29.5 percent, were recovered. Of the total number recovered 1,101 were captured in area 1, 35 in area 2, 20 in area 3, and 3 in area 4. The tagging experiments conducted on the east, south, and west shores of Duke Island were included with those on and near the mainland shore in the vicinity of Cape Fox because the percentage of recoveries from these experiments shows a greater rela- tion to those carried on near Cape Fox than to those carried on elsewhere in this general region. It will be noted in table 1 that the total number of tagged individuals re- covered in areas 2 and 3 are composed largely of the recoveries reported from the Kelp Island experiment carried on in 1932. A further analysis of the recoveries reported from area 1 shows that of those captured east of Cape Fox 333 were taken in Alaskan waters and only 79 in Canadian waters. Of those captured north of Cape Fox 533 were taken in the waters of Revillagigedo Channel and Tongass Narrows, 103 in the south arm of Behm Canal, 14 in the north arm of Behm Canal, and 39 off the west shores of Duke, Annette, and Gravina Islands. Although far greater numbers were captured in the south arm of Behm Canal than in the north arm, it cannot be definitely assumed that these results represent the exact ratio of distribution, for many of the salmon bound for the north arm of Behm Canal may have been picked up enroute in Revillagigedo Channel and Tongass Narrows. However, it may be definitely as- sumed that most of the pink salmon migrating through Dixon Entrance to the main- land and island shores in the vicinity of Cape Fox are bound for the localities in some part of area 1 ; i. e., those of Portland Canal and other waters east of Cape Fox, and those of Revillagigedo Channel and Behm Canal. TAGGING EXPERIMENTS IN THE VICINITY OF CAPE CHACON The early tagging experiments carried on in the vicinity of Cape Chacon were not as numerous, nor as varied in the time of the season they were conducted, as those in the vicinity of Cape Fox. A total of six experiments, five in 1925 and 1 in 1926, were made during the second week of August of each year, which is beyond the middle of the migratory season for pink salmon in this region. However, the results are sufficient to show that there is a distinct difference between the final destinations of the pink salmon migrating northward from Dixon Entrance into Clarence Strait in the vicinity of Cape Chacon, and those migrating northward from Dixon Entrance along the main- land shores in the vicinity of Cape Fox. 41860—38- ■2 650 BULLETIN OF THE BUREAU OF FISHERIES Table 2. — Pink salmon tagged in the vicinity of Cape Chacon, 1925 and 1926, and number recovered [Column headings indicate locality and date of tagging] Cape Cha- con, Aug. 8, 1926 Stone Rock Bay, Aug. 9, 1925 Stone Rock Bay, Aug. 9, 1926 Cape Cha- con, Aug. 11, 1925 Stone Rock Bay, Aug. 12, 1925 Cape Cha- con, Aug. 13, 1925 Total, Cape Chacon and Stone Rock Bay Kaigani Strait, Aug. 10-11, 1926 Cape Muzon, Aug. 15, 1925 Kaigani Point, Aug. 16-21, 1925 Total, Kaigani Strait and Cape Muzon Number tagged.. 36 504 479 546 609 455 2, 629 1, 354 579 1,876 3, 809 LOCALITY OF RECOVERY Lower Clarence Strait- Area 1 East of Cape Fox.__ 2 3 3 8 British Columbia 1 1 1 1 1 3 Revillagigedo Channel and Ton- 1 6 7 11 17 10 52 13 2 1 16 1 1 2 2 3 9 Bebm Canal, north arm ... 2 4 9 4 6 25 3 3 West shores of Gravina, Annette, and Duke Islands 1 23 25 42 39 21 151 18 ii 6 35 Total recovered.. 3 34 37 67 65 40 246 35 14 8 57 Percent recovered 8.3 6.7 7.8 12.3 10. 7 8.8 9.4 2.6 2.4 0.4 1.5 Lower Clarence Strait. — Area 2 East shore Prince of Wales Island, from Cape Chacon to Approach Point (all points): Total recovered _ 4 63 38 55 50 51 261 64 26 32 122 in 12.5 7.9 10.1 8.2 11.2 9.9 4. 7 4.5 1.7 3.2 Middle Clarence Strait. — Area 3 Approach Point, Caamano Point, to Narrow Point, and Ernest 3 11 10 34 23 16 97 23 3 26 Ernest Sound 7 1 11 10 2 31 1 1 2 Total recovered. 3 18 11 45 33 18 128 23 4 1 28 Percent recovered. 8.3 3.6 2.3 8.2 5.4 4.0 4.9 1.7 0.7 0.1 0.7 Upper Clarence Strait.— Area 4 Narrow Point, Ernest Point to Point Harrington, and East Is- 4 1 8 7 7 27 1 3 4 2 2 4 8 Total recovered 6 1 8 9 ii 35 1 3 4 Percent recovered 0.0 1.2 0.2 1.5 1.5 2.4 1.3 0.0 0.2 0.2 0. 1 West Coast Prince of Wales Island (All Points).— Area 6 Total recovered - 51 45 18 101 26 241 249 115 391 755 Percent recovered 0.0 10. 1 9.4 3.3 16. 5 5.7 9.2 18.4 19.8 20.8 19.8 OUTLYING AREAS Chatham Strait and Frederick Sound (all points): Total recovered 1 1 2 2 0.0 0.2 0.0 0.0 0.0 0.0 0.0 0.2 0.0 0.0 0.1 SUMMARY: 10 173 132 193 258 146 912 373 160 435 968 27.7 34.3 27.6 35.3 42.4 32.1 34.7 27.6 27.6 23.2 25.4 1 These data do not include recoveries reported from the point of tagging nor those doubtful as to location of capture. See text, p. 647. The original records of the tagging experiments in 1925 and 1926 are given in the reports of Rich (1926), and Rich and Suomela (1927). MIGRATIONS OF PINK SALMON 651 A total of 2,629 pink salmon were tagged, of which 912, or 34.7 percent, were recovered. Of the total number recovered 246 were captured in area 1, 261 in area 2, 128 in area 3, 35 in area 4, 241 in area 6, and 1 in Chatham Strait, an outlying dis- trict. Of the 246 salmon captured in area 1, 151 were taken by the fishery on the west shores of Annette and Gravina Islands, 25 in the north arm of Belim Canal, 9 in the south arm of Belim Canal, and the remaining 61 in Revillagigedo Channel and waters east of Cape Fox. It will be recalled that the results from the Cape Fox taggings showed more recoveries from the south arm of Behm Canal than from the north arm, whereas the results from the Cape Chacon taggings show just the reverse. That larger numbers of the salmon migrating through Clarence Strait are bound for the north arm of Behm Canal, than for the south arm, will be definitely demonstrated later in the discussion of the 1935 and 1936 taggings near Cape Chacon. Although only 35 tagged salmon were captured in area 4 it cannot be assumed that this repre- sents the exact proportion bound for this area. Since the salmon bound for area 4, at the extreme upper end of Clarence Strait, must run the gauntlet of all the fishing gear along the shores of the strait from Cape Chacon northward, it is not unlikely that some of the tagged salmon captured in areas 1, 2, and 3 were destined for area 4 but were intercepted en route. This is one of the difficulties that make it impossible to determine the exact distribution of the salmon migrating through a main channel of entry by tagging experiments conducted near the entrance to the channel. By tagging at different points along the channel part of the difficulty encountered from interception of the tagged individuals may be overcome. However, in using this system of tagging in a channel such as Clarence Strait, where the fish are migrating in both directions, further difficulties arise that far offset the advantages gained. From the results of these experiments it appears that the pink salmon migrating into Clarence Strait during the latter part of the season are bound mainly for localities in areas 1, 2, and 6. u TAGGING EXPERIMENTS ON GRAVINA ISLAND AND IN THE VICINITY OF KASAAN BAY The results from the early tagging experiments carried on at points in the vicin- ity of Kasaan Bay and along the west shore of Gravina Island will be considered next. Three experiments were carried on in the latter region; one at Nelson Cove early in July 1926, and one each at Nelson Cove and Dali Head early in August 1927. The exact locations of these points are shown by dots on Gravina Island in figure 1. Nelson Cove was reported by Rich (1927-29) as located at the north end and Dali Head at the south end of the island, and the results from the experiments are given in table 3. 652 BULLETIN OF THE BUREAU OF FISHERIES Table 3. — Pink salmon tagged in the vicinity of Kasaan Bay, 1926-80, and number recovered 1 [Column headings indicate locality and date of tagging] Number tagged - LOCALITY OF RECOVERY Lower Clarence Strait. — Area 1 Revillagigedo Channel and Tongass Narrows. Behm Canal, south arm Behm Canal, north arm West shores of Qravina, Annette, and Duke Islands. . Total recovered Percent recovered. Lower Clarence Strait. — Area 2 East shore Prince of Wales Island, from Cape Chacon to Approach Point: Total recovered... - Percent recovered - — Middle Clarence Strait.— Area 3 Approach Point, Caamano Point to Narrow Point, and Ernest Point - Ernest Sound - Total recovered- .. Percent recovered. Upper Clarence Strait. — Area 4 Narrow Point, Ernest Point to Point Harrington, and East Island — Snow Passage and Stikine Strait Total recovered — Percent recovered. West Coast Prince Of Wales Island (All Points).— Area 6 Total recovered — Percent recovered. Outlying Areas Chatham Strait... Stephens Passage. Total recovered — Percent recovered. SUMMARY: Total recovered — Percent recovered. South Entrance to Kasaan Bay, July 29, 1930 146 16 10.9 9 6.2 0.0 1 0.7 0.0 36 24.6 South Entrance to Kasaan Bay, Aug. 14, 1930 125 5 3.7 49 36.2 9 6.7 0.0 0.0 0.0 63 46.6 Windfall Harbor, Aug. 3, 1930 423 27 6.4 5 1.2 20 20 4.7 10 2.4 1 0.2 1 0.2 64 15.1 Windfall Harbor, Aug. 14, 1930 425 15 3.5 40 9.5 93 93 21. 1 0.2 0.0 0.0 149 35.1 Nelson Cove, July 6-7, 1926 Dali Head and Nelson Cove, Aug. 5-6, 1927 284 8 1 24 6 39 13.7 1 0.4 13 4 17 6.0 1 0.4 0.0 2 0.7 60 21.1 321 8 1 14 13 36 11.2 4 1.2 19 11 30 9.4 2 0.6 0.0 0.0 72 22.4 i These data do not include recoveries reported from the point of tagging nor those doubtful as to location of capture. See text, p. 647. The original records of the tagging experiments from 1926 to 1930 are given in the reports of Rich and Suomela (1927), Rich and Morton (1929), and Rich (1932). From an inspection of column 5 in table 3 it will be seen that most of the pink salmon migrating along the west shore of Gravina Island early in July are bound for the northern localities in area 1 and the general localities in areas 3 and 4, there being only a small percentage of them taken in area 2. From an inspection of column 6 in table 3 it will be seen that most of the salmon migrating along this shore in the early part of August are also bound for the northern localities in area 1 and the general localities in areas 3 and 4. However, there is the appearance of a tendency for the fish to turn back from this shore to localities in area 2, which becomes quite marked later on in August. Evidence of this is afforded by the slight increase in the number of tagged individuals recovered from area 2 as reported from the August taggings. Further reference will be made to this tendency in the discussion immediately follow- MIGRATIONS OF PINK SALMON 653 ing. In all three of these experiments the small number of recoveries in area 4 may have been due partially to interception in area 3. Four lots of salmon were tagged in the vicinity of Kasaan Bay in 1930, 2 just south of the entrance to the bay near Island Point, 1 on July 29, and 1 on August 14, 2 just north of the bay at Windfall Harbor on August 3 and 14. The results are given in columns 1-4 in table 3. The tagged salmon recovered from the July 29 tagging just south of the bay were captured in areas 1, 2, and 3, in almost equal proportions. The largest number, however, came from area 2. Those recovered from the August 14 tagging at the same location were captured also in areas 1, 2, and 3, but this time by far the greater part were captured in area 2. In both of these experiments the localities in which the salmon were recovered in area 2 extended as far south as Cape Chacon. It is evident, then, that the majority of the pink salmon reaching the eastern shore of Prince of Wales Island, in the vicinity of Kasaan Bay, during the second week in August are migrating southward rather than northward. The individuals recovered from the August 3 tagging at Windfall Harbor, north of Kasaan Bay, were recaptured mainly in areas 1,3, and 4, with only a relatively small number coming from area 2. Of the 64 tagged individuals recovered, 33 came from areas south of Kasaan Bay, 27 from area 1, 5 from area 2, and 1 from area 6, while 30 came from areas north of the bay, 20 from area 3, and 10 from area 4. Thus, it may be assumed that the pink salmon at this time of the season are migrating in equal numbers in both directions from their point of tagging at Windfall Harbor. The recoveries from the August 14 tagging at this point, however, show a much dif- ferent picture. Of the 149 tagged salmon recovered from this experiment, 15 were captured in area 1, 40 in area 2, 93 in area 3, and 1 in area 4. Here we find 94 of the recoveries made in areas 3 and 4, north of the point of tagging, and only 55 made in areas 1 and 2, south of the point of tagging. Hence, it may be assumed that in the latter part of the season the pink salmon reaching the shores of Clarence Strait, in the vicinity of Windfall Harbor, are migrating northward in greater numbers than south- ward. It is also interesting to note that the recoveries from area 2 in this experiment were more numerous than from area 1, 40 as compared to 15, whereas the reverse was true in the August 3 tagging, 5 as compared to 27. This, together with the fact that the majority of the recoveries from the tagging south of Kasaan Bay on August 14 were recovered in area 2, leads to the conclusion that most of the salmon reaching the shores of Clarence Strait below Approach Point, during the latter part of the season, are bound for localities along the east shore of Prince of Wales Island, in area 2, as far south as Cape Chacon. Where do these salmon come from? Are they part of a population migrating southward from Sumner Strait, or do they come from the popu- lations migrating northward in Clarence Strait that have turned near Kasaan Bay and move southward? A review of the tagging experiments carried on in Sumner Strait at Point Colpoys (see table 7) will show that most of the salmon migrating south- ward from Sumner Strait at this time of the season are bound for localities in area 4, with only a few migrating as far south as area 2. Therefore, considering the large volume of salmon caught in area 2 from Kasaan Bay south at this time of the season, there is only one probable origin of these southward-bound salmon. They must come from populations migrating northward in Clarence Strait that have turned westward and southward from the west shore of Gravina Island and the lower shore of Cleve- land Peninsula just above Caamano Point. This turning back of the salmon from the west shore of Gravina Island during the latter part of the season was indicated in the discussion of the taggings carried on along this shore in 1926 and 1927. Further 654 BULLETIN OF THE BUREAU OF FISHERIES evidence in support of this conclusion will be presented in the discussion of the 1935 and 1936 taggings near Cape Chacon. TAGGING EXPERIMENTS IN THE VICINITY OF CAPE MUZON In discussing the tagging experiments in the immediate vicinity of Cape Chacon it was pointed out that a considerable number of the recoveries were reported from area 6 and the localities of Cordova Bay and contiguous channels. Of the 924 salmon recovered from the experiments near Cape Chacon, 241 were captured in area 6. It would appear that considerable numbers of the pink salmon migrating into Clarence Strait during the latter part of the season are bound for the localities of Cordova Bay and are milling about in an attempt to find them. A review of the results from the tagging experiments carried on late in the season at Cape Muzon, Kaigani Point, and Kaigani Strait points (see table 2), located at the entrance to Cordova Bay, will show that part of the pink salmon migrating into Cordova Bay are bound for areas 1, 2, 3, and 4. The results from these experiments are given in the last four columns of table 2. A total of 3,809 pink salmon were tagged at these three locations, of which 968, or 25.4 percent, were recovered. Of the total number recovered 57 were captured in area 1, 122 in area 2, 28 in area 3, 4 in area 4, 755 in area 6, and 2 in an outlying area. Since these taggings were carried on at approximately the same time in the fishing season as those near Cape Chacon, it is not surprising to find considerable numbers of the tagged individuals captured in area 2. In summarizing the results from the tagging experiments carried on in the vicinity of Cape Chacon and Cape Muzon, it is evident that during the latter part of the season the pink salmon move in from Dixon Entrance to the southern shores of Prince of Wales Island, and after milling about for some time separate into populations that are bound for the localities in area 6, and populations that are bound for the localities in areas 1, 2, 3, and 4. PINK-SALMON TAGGING EXPERIMENTS IN CLARENCE STRAIT IN 1935 AND 1936 In the summers of 1935 and 1936 series of consecutive weekly taggings of pink salmon were carried on in Clarence Strait in the vicinity of Cape Chacon. These taggings were made from a trap at McLean Point, located on the east shore of Prince of Wales Island, approximately 7 miles north of Cape Chacon. The taggings at McLean Point were all made from the trap operated at this location by the Alaska Pacific Salmon Co., and the Bureau of Fisheries wishes to express its appreciation for the cooperation of this company in furnishing these facilities, and the salmon that were tagged. In Alaska, commercial fishing for salmon is prohibited by law from 6 p. m. Satur- day to 6 a. m. Monday during the entire fishing season. All of the taggings in both series, with the exception of the first and fourth in 1935, were carried on at the begin- ning of the weekly closed periods. This provided an opportunity for the tagged in- dividuals to migrate from the location at which they were tagged for at least 34 hours before they were again subject to capture. By following this procedure only a few were recaptured in the trap from which they had been tagged. The dates of the weekly experiments, and the number of pink salmon tagged in each, are given in table 4 for the 1935 series and in table 5 for the 1936 series of experiments. MIGRATIONS OF PINK SALMON 655 Table 4. — Pink salmon tagged at McLean Point in 1935 1 [Column headings indicate date and number tagged] Salmon recovered and average time en route Locality of recovery July 22, 381 July 27, 395 Aug. 3, 395 Aug. 12, 389 Aug. 17, 297 Total, 1,857 Lower Clarence Strait. — Area 1 East of Cape Fox: Num- ber Days Num- ber 1 Days 12 Num- ber Days Num- ber Days Num- ber Days Num- ber 1 Day 8 12 1 4 1 4 Revillagigedo Channel and Tongass Narrows: 2 14 1 9 3 12 1 3 1 5 2 4 2 6 3 10 6 6 2 2 13 1 6 1 3 3 6 4 6 4 8 4 3 3 11 4 1 0 i a i o 9 6 1 9 1 1 6 Behm Canal, south arm: 1 7 7 1 ? 1 ? 1 ? 1 ? 2 5 2 5 1 6 1 Behm Canal, north arm: 1 4 1 4 1 8 2 3 3 7 6 6 1 8 1 3 2 6 8 7 9 7 3 6 20 7 1 ? 1 ? 1 26 1 26 4 6 4 3 1 4 9 4 6 7 3 3 3 4 i 7 13 5 2 4 2 4 3 4 3 4 3 12 3 12 1 ? 1 ? 2 8 2 3 4 6 West and south shores of Gravina, Annette, and Duke Islands: 1 7 7 5 1 4 9 5 2 5 2 5 12 4 3 6 6 6 3 3 24 6 1 5 3 3 1 4 5 4 2 19 1 0 1 ? 3 5 2 4 9 8 27 5 23 5 19 7 n 5 18 4 98 5 61 95 44 37 32 269 16.0 24.0 ii. i 9.5 10.8 14.5 Lower Clarence Strait.— Area 2 Prince of Wales Island from Capo Chacon to Ap- proach Point: 4 2 5 7 6 11 18 4 8 3 41 5 Polk Island 3 14 22 5 10 4 35 5 2 15 4 5 5 4 11 6 Point Adams. _. 5 3 2 7 i 10 19 6 6 4 33 5 Wedge Island 1 4 2 4 3 4 i 9 3 13 2 9 1 4 7 10 1 13 i 13 1 3 1 9 2 6 1 ? 2 8 1 4 4 5 8 5 1 7 1 7 2 5 4 6 Total recoveries 12 13 16 67 37 145 3.1 3.3 4. 1 17.2 12.5 7.8 Middle Clarence Strait.— Area 3 Approach Point, Caamano Point, to Narrow Point, and Ernest Point: 3 9 3 6 5 5 1 13 12 7 Grindall Island and Niblack Point 7 6 10 6 4 4 3^ 7 6 3 30 5 Niblack Point - 3 5 3 5 3 8 5 10 8 6 1 14 8 7 25 8 14 9 9 5 17 5 4 4 8 5 52 6 Ernest Sound: Union Point. i 13 i 8 2 10 Eaton Point.. . 4 9 4 4 8 6 1 25 1 25 Santa Anna 1 4 1 4 Point Warde 3 27 3 27 i 6 1 6 Total recoveries 38 28 39 10 23 138 Percent recoveries 10.0 7.0 9.9 2.6 7.7 7.4 1 These data do not include recoveries reported from the point of tagging nor those doubtful as to location of capture. See text, p. 647. 656 BULLETIN OF THE BUREAU OF FISHERIES Table 4. — Pink salmon tagged at McLean Point in 1935 — Continued [Column headings indicate'date and number'tagged] Locality of recovery Salmon recovered and average time en route July 22, 381 July 27, 395 Aug. 3, 395 Aug. 12, 389 Aug. 17, 297 Total, 1,857 Upper Clarence Strait.— Area 4 Narrow Point, Ernest Point to Point Harrington, East Island: Num- ber 2 3 Days 5 34 Num- ber Days Num- ber 1 Days 4 Num- ber Days Num- ber Days Num- ber 3 6 2 4 1 1 Days 5 23 3 6 3 3 2 15 1 8 2 3 1 11 1 4 2 1 4 3 Snow Passage and Stikine Straits: 1 3 5 1.3 1 0.3 3 0.7 4 1.0 4 1.3 1,7 0.9 Sumner Strait.— Area 5 1 1 0.3 6 1 1 0.1 8 West Coast Prince of Wales Island.— Area 6 3 6 2 11 1 1 9 7 6 3 8 4 2 2 3 0.8 2 0.5 2 0.5 2 0.7 9 0.5 119 31.2 138 34.9 104 26.3 120 30.8 98 33.0 579 31.2 In 1935, five weekly taggings were carried on from July 22 to August 17. In 1936, be- cause of a heavy storm on the week end of August 8, only four weekly taggings were carried on over a sim- ilar period of time; July 18 to August 15. It was for this reason that twice the usual number of salmon were tagged on August 15. Tables 4 and 5 also give the number and percent of tagged salmon recovered from each of the experiments and the localities in which they were captured. These localities, as stated in tables 1, 2, and 3, are listed under the geographic areas in which they occur. The average number of days the tagged salmon were en route before recapture are also reported but, owing to the ex- treme variability in the rate of travel, as indicated by the dates the salmon were recaptured, the authors do not feel justified in drawing definite conclusions about this phase of the salmon migrations. However, since this information is of general interest, it is reported in the tables. The seasonal trends in the percentages of salmon recovered in the geographic areas, as Figure 4. — Trends in the percentage recoveries of tagged pink salmon from the weekly tagging experiments in 1935 and 1936 at McLean Point. MIGRATIONS OF PINK SALMON 657 indicated by the results from the weekly taggings, are shown in figure 4 for both the 1935 and 1936 series of experiments. Table 5. — Pink salmon tagged at McLean Point in 1936 1 [Column headings indicate date and number tagged] Salmon recovered and average time en route Locality of recovery July 18, 339 July 25, 625 Aug. 1, 493 Aug. 15, 984 Total, 2,441 Lower Clarence Strait.— Area 1 East of Cape Fox: Nakat Bay . - Revillagigedo Channel and Tongass Nar- rows: Cape Fox Tree Point ... Foggy Bay Buke Point Slate Island--. Black Island Point Alava Lucky Cove Thorne Arm Carrol Point Annette Island, north and east shores.. General Behm Canal, south arm: Point Sykes Smeaton Bay Chickamin River Behm Canal, north arm: Point Higgins Bond Bay— Clover Passage Betton Island Traitors Cove Belle Island General West and south shores of Gravina, An- nette, and Duke Islands: Percy Islands - Annette Island, south shore.. Annette Island, northwest shore Point McCartey Warburton Island Seal Cove Blank Point Gravina Island, west shore Total recoveries Percent recoveries.. 118 34.8 Lower Clarence Strait. — Area 2 Prince of Wales Island from Cape Chacon to Approach Point: Kendrick Bay Polk Island, Scott Point, Rip Point, and Hidden Point Moira Sound Point Adams Wedge Island Chasina Point Island Point Patterson Island Skowl Arm. Twelve Mile Arm Kasaan Bay... Total recoveries Percent recoveries. Middle Clarence Strait.— Area 3 Approach Point, Caamano Point, to Nar- row Point, and Ernest Point: Caamano Point and Grindall Island. . Niblack Point Ship Island and Street Island Meyers Chuck.: False Island Ernest Sound: Vixen Point Union Point and Watkins Point Total recoveries Percent recoveries. 11 3.2 62 148 23.7 20 3.2 12 60 9.6 106 21.5 37 7.5 19 3.9 72 133 13.5 212 21.6 24 2.4 2 2 77 5 8 15 2 218 505 20.7 115 76 2 13 10 21 10 1 12 3 6 269 11.0 114 4.7 4 4 14 2 7 7 4 11 7 10 7 6 5 5 4 4 21 8 18 7 1 These data do not include recoveries reported from the point of tagging nor those doubtful as to location of capture. See text, p. 647. 658 BULLETIN OF THE BUREAU OF FISHERIES Table 5. — Pink salmon tagged at McLean Point in 1936 — Continued [Column headings indicate date and number tagged] Salmon recovered and average time en route Locality of recovery July 18, 339 July 25, 625 Aug. 1, 493 Aug. 15, 984 Total, 2,441 Upper Clarence Strait.— Area 4 Narrow Point, Ernest Point to Point Har- rington, and East Island: Ernest Point Narrow Point Olson Cove Eagle Creek Lincoln Rock Marsh Island Steamer Rock Snow Passage and Stikine Strait: Snow Passage Total recoveries Percent recoveries. 3 0.9 Sumner Strait.— Area 5 Affleck Canal. Total recoveries Percent recoveries. West Coast Prince of Wales Island.- Area 6 Nichols Bay Brownson Bay Hunter Bay Point Webster Kdakas Inlet Natkwa Inlet Cordova Bay Cape Muzon Total recoveries Percent recoveries. 4 1.2 Outlying Areas Kingsmill Point Total recoveries Percent recoveries. 1 1 0.3 Summary of all recoveries. Percent of all recoveries... 137 40.4 25 15 2.4 1 0.2 36 14 2.2 10 3 3 0.5 261 41.8 9 1.8 1 1 0.2 176 35.7 15 1.5 40 4. 1 424 43. 1 10 37 1.5 1 0. 1 67 2.7 5 5 0.2 998 40.9 10 18 16 11 16 11 22 The percentage of recoveries in area 1, from the weekly taggings at McLean Point, tended to be highest in the early part of the season and lowest in the latter part, thus indicating a seasonal decrease in the portion of the run passing McLean Point that was bound for area 1. In reviewing the number of tagged salmon recov- ered in area 1 (see tables 4 and 5), it was found that very few were captured in the localities east of Cape Fox and in the south arm of Behm Canal. Considerable numbers were captured in Kevillagigedo Channel and the north arm of Behm Canal. The largest numbers, however, were recovered consistently along the west shores of Annette and Gravina Islands. The percentage of recoveries in area 2 from the weekly taggings tended to be lowest in the early part of the season and highest in the latter part. This is just the reverse of the seasonal trends in area 1. In other words, there was a very definite seasonal increase in the percentages of the pink salmon tagged at McLean Point that were recovered in area 2. The localities in which these recoveries were made were scattered along the east shore of Prince of Wales Island from Kendrick Bay to Kasaan Bay. During the first part of the season a large proportion of the salmon were recap- tured just north of the point of tagging but, as the season progressed, more and more were reported from localities in the vicinity of Kasaan Bay. MIGRATIONS OF PINK SALMON 659 The percentage of recoveries in area 3 from the weekly taggings was much smaller than those in areas 1 and 2. No doubt many of the tagged salmon bound for this area were intercepted in the latter regions. Nevertheless, there were sufficient recoveries in area 3 to give some indication of the time when these salmon migrated through Clarence Strait. Although the seasonal trends in area 3 are not very marked, there is indication that more of the recoveries were made in the early part of the season than in the latter part. In other words, more pink salmon bound for area 3, and the areas above it, migrate through Clarence Strait in the early part of the season than in the latter part. The percentage of recoveries in areas 4 and 5 were so small that no definite trends occur in them. Many of the salmon bound for these areas were, no doubt, intercepted in areas 1,2, and 3. The percentage of recoveries in area 6 for the 1935 series of experiments was too small to indicate a definite seasonal trend. However, the percentage recovered from the 1936 series of experiments, although only slightly greater, indicates a possible upward trend in the latter part of the season. Since, in 1935, there was a misunder- standing on the part of the cannerymen in this area as to the method of reporting recoveries of tagged salmon, there is reason to believe the 1935 data are incomplete. This tendency toward an increase in the percentage of recoveries as the season pro- gresses is in accord with the results from the experiments carried on near Cape Chacon during the second and third weeks of August in 1925 and 1926. In fact the whole distribution of the salmon tagged during the second and third weeks of August in the 1935 and 1936 experiments is in accord with the distribution of the salmon tagged in the 1925 and 1926 experiments. In discussing the distribution of the recoveries from the early tagging experiments in Clarence Strait, evidence was pointed out winch indicated that many of the pink salmon migrating along the west shores of Gravina and Annette Islands during the latter part of the season were bound for localities along the east shore of Prince of Wales Island in area 2. Further evidence in regard to this peculiar migration of the pink salmon in Clarence Strait may be found in the distribution of the recoveries from the 1935 and 1936 experiments. In discussing the localities in area 1 in which the tagged salmon were recovered (see tables 4 and 5), it was pointed out that the majority were recovered from the west shores of Gravina and Annette Islands. The total numbers, by weekly taggings, recovered in area 1 in the localities east of Cape Fox, Revillagigedo Channel, and Behm Canal, as a group, and the recoveries from the west shores of Gravina and Annette and Duke Islands, as a group, are given in table 6. Table 6. — Pink salmon recovered in area 1 from tagging experiments at McLean Point in 1935 and 1936 Recoveries in 1935 Recoveries in 1936 Date of tagging experi- ment East of Cape Fox, Revilla- gigedo Channel, and Behm Canal West shores of Gravina, An- nette, and Duke Islands Total East of Cape Fox, Revilla- gigedo Channel, and Behm Canal West shores of Gravina, An- nette, and Duke Islands Total July 18 Number Percent Number Percent Number Percent Number 51 Percent 43 Number 67 Percent 57 Number 118 Percent 100 July 22 29 48 32 52 61 • 100 July 25 52 35 96 65 148 100 July 27 51 54 44 40 95 100 Aug. 1 38 37 68 63 106 100 Aug. 3. 20 45 24 55 44 100 Aug. 12. 14 38 23 62 37 100 Aug. 15 35 26 98 74 133 100 Aug. 17 .. 8 25 24 75 32 100 660 BULLETIN OF THE BUREAU OF FISHERIES The relation between these recoveries and the total recoveries in area 1 , expressed in percentage, is also given in table 6. It will be noted that the weekly percentage of recoveries in area 1, from the localities east of Cape Fox, Revillagigedo Channel, and Behm Canal, are highest in the beginning of the season and lowest toward the end. The weekly percentage of recoveries along the west shores of Gravina, Annette, and Duke Islands is, on the other hand, lowest at the first part of the season and highest toward the last. In other words, there is a seasonal increase in the percentage of tagged salmon caught in area 1 along the west shores of Gravina and Annette Islands, whereas there is a seasonal decrease in the percentage of the tagged salmon caught in other localities of the area. Since there are very few streams along the west shores of Annette and Gravina Islands in which salmon can spawn, most of the tagged salmon caught in this region were evidently bound either for other localities in area 1 , or in one of the other areas. During the first part of the season considerable numbers of tagged salmon were recovered in the north arm of Behm Canal and in area 3. No doubt many of those recovered from the west shores of Gravina and Annette Islands during the first part of the season were bound for those regions. In the latter part of the season, however, very few were recovered from the north arm of Behm Canal and area 3, but there were still large recoveries from the west shores of Gravina and Annette Islands. Where were these tagged salmon going at this time of the season? In discussing the percentage of recoveries of tagged salmon in area 2 from the weekly taggings it was pointed out that there was a seasonal increase in these recov- eries and that larger numbers were reported from the localities in the vicinity of Kasaan in the last part of the season than in the first part. Hence, it is not at all improbable that many of the tagged salmon recovered from the west shores of Gravina and Annette Islands during the latter part of the season were bound for the localities in area 2 in the vicinity of Kasaan Bay and southward. These salmon no doubt migrated across Clarence Strait to the west shores of Gravina and Annette Islands and then turned back to the localities on the east shore of Prince of Wales Island in area 2. Whether or not the salmon make this journey may depend to some extent upon the prevailing winds. If a southeast wind is blowing there seems to be a greater tendency for the salmon to migrate northward along the east shore of Prince of Wales Island rather than to cross over to the west shores of Gravina and Annette Islands. If a southwest wind is blowing the opposite course is more apt to be taken. No definite conclusions may be drawn because of the lack of sufficient information, at the present time, concerning the influence of wind direction on migration. SUMMARY OF CAPE CHACON EXPERIMENTS 1. The results from the tagging experiments in Clarence Strait, both past and present, indicate that the pink salmon migrating into the strait at different times of the season follow rather definite migratory routes to the localities for which 'they are bound. 2. Most of the first pink salmon to appear each season are bound for localities in Revillagigedo Channel, the north arm of Behm Canal, Ernest Sound, or the northernmost regions in Clarence Strait. These salmon enter Clarence Strait by way of Cape Chacon and, after migrating for a short distance northward along the east shore of Prince of Wales Island, leave this shore and either follow directly up the middle of the strait or turn eastward until they reach the west shores of Annette and Gravina Islands, From here they either continue northward into the north MIGRATIONS OF PINK SALMON 661 arm of Bekm Canal and the northern regions of Clarence Strait, including Ernest Sound, or they continue eastward and southward into Revillagigedo Channel. 3. This peculiar migration of the salmon along the west shores of Annette and Gravina Islands results in many of the salmon, bound for the north arm of Bekm Canal and the northern regions of Clarence Strait, being intercepted by the traps and purse seines operated along these shores. 4. As the season progresses, more and more of the pink salmon entering Clarence Strait by way of Cape Chacon arc destined for the localities along the east shore of Prince of Wales Island below Approach Point. Most of these salmon con- tinue northward along the east shore of Prince of Wales Island instead of crossing the strait to the west shores of Annette and Gravina Islands. Many of those that do cross the strait turn back from the shores of these islands to the localities on the east shore of Prince of Wales Island. 5. At the very close of the season many of the salmon migrating into Clarence Strait are bound for localities in Cordova Bay on the extreme southwest shore of Prince of Wales Island. The migration of the pink salmon along the west shores of Annette and Gravina Islands, especially the latter, makes these shores one of the most productive fishing areas in Clarence Strait. The fishing gear operated in this area intercepts the runs of pink salmon that are bound for practically all of the localities in Clarence Strait and its adjacent waters to the east. It is for this reason that so many of the pink salmon tagged in Clarence Strait were recovered along the west shores of these islands. PINK-SALMON TAGGING EXPERIMENTS IN SUMNER STRAIT, 1924-36 It has long been known that the run of pink salmon in Sumner Strait appears earlier in the summer than the run in Clarence Strait. The time the salmon ap- peared in the commercial catches and the location of these catches in the strait indicated that most of these early migrants were bound for localities in the extreme eastern section and the adjoining waters of Zimovia Strait, Eastern Passage, Blake Channel, and Bradfield Canal. However, the extent to which this run penetrated the waters of Clarence Strait and other adjoining channels was not known. In order to determine more completely the distribution of the localities in which these pink salmon spawned, the Bureau laid plans for a number of tagging experiments to be carried on in various parts of Sumner Strait. This work began in 1924, was continued each year through 1927, and taken up again in 1935 and 1936. Since only one experiment was carried on in each of the latter seasons, their results will be discussed with those from the early experiments. The locations where the salmon were tagged in Sumner Strait from 1924 to 1936 are shown in figure 1. The dots indicate the locations of early taggings and the triangle is that of the later ones. The early tagging was done at Cape Decision, Ruins Point, and Point Colpoys, and during 1935 and 1936 only at Point Colpoys. Although none of these experiments continued throughout the entire season of any year, they varied sufficiently in point of time so that, taken together, they give a picture of the movements of fish in this region over an entire season. Thus, experi- ments were carried on at Cape Decision and Rums Point in 1924, 1925, and 1927, between July 12 and August 3; those at Point Colpoys on July 10, 1926, July 26 to 30, 1927, August 13, 1935, and August 16, 1936. A summary by geographic areas of the results from all experiments from 1924 to 1936 is given in table 7. The indi- vidual localities in the geographic areas, in which the tagged salmon were recovered 662 BULLETIN OF THE BUREAU OF FISHERIES from the 1935 and 1936 experiments at Point Colpoys, are given in table 8. The taggings at Point Colpoys were made from the trap operated at this location by the Pacific American Fisheries, Inc., and the Bureau of Fisheries wishes to express its appreciation for the cooperation of this company in furnishing these facilities, and the salmon that were tagged. Table 7.- — Pink salmon tagged in Sumner Strait, 1924-86, and number recovered 1 [Column headings indicate locality and date of tagging] Point Col- poys, July 10, 1926 Point Col- poys, July 26- 30, 1927 Point Col- poys, Aug. 13, 1935 Point Col- poys, Aug. 16, 1936 Total, Point Col- poys Ruins Point, July 18- 25, 1925 Ruins Point, July 12- Aug. 3, 1924 Cape De- cision, July 30, 1927 Total, Ruins Point and Cape De- cision Number tagged. 259 569 386 498 1,712 1, 217 250 162 1,629 LOCALITY OF RECOVERY Sumner Strait (all points).— Area 5 Total recovered 2 6 9 9 26 295 20 8 323 Percent recovered 0.8 1. 1 2.3 1.8 1.5 24. 2 8.0 4.9 19.8 Upper Clarence Strait.— Area 4 Point Colpoys 32 6 5 43 Snow Passage and Stikine Strait 3 13 19 84 119 36 2 38 Narrow Point, Ernest Point to Point Harrington, and East Island.. _ 8 34 47 66 155 30 3 2 35 Total recovered 11 47 66 150 274 98 9 9 116 Percent recovered 4.2 8.3 17. 1 30. 1 16.0 8.2 3.6 5.6 7. 1 Middle Clarence Strait.— Area 3 Ernest Sound — - - 28 88 3 6 125 29 2 2 33 Approach Point, Caamano Point to Narrow Point and Ernest Point 11 39 20 7 77 15 2 2 19 Total recovered 39 127 23 13 202 44 4 4 52 Percent recovered _ 15. 1 22.2 6.0 2.6 11.8 3.6 1.6 2.5 3.2 Lower Clarence Strait.— Area 2 East shore of Prince of Wales Island from Approach Point to Brownson Bay: 1 4 2 6 13 6 3 9 Percent recovered- 0.4 0.7 0.5 1.2 0.8 0.5 1.2 0.0 0.6 Lower Clarence Strait.— Area 1 West shores of Gravina, Annette, and Duke Islands. 2 21 4 9 36 13 1 14 4 6 2 12 1 1 1 1 1 8 2 3 14 7 2 9 1 1 4 4 2 2 1 1 2 Total recovered.. 8 38 6 14 66 26 3 1 30 Percent recovered 3. 1 6.7 1.6 2.8 3.8 2. 1 1.2 0.6 1.8 Outlying Areas 27 7 34 Chatham Strait and Frederick Sound (all points) 2 1 3 7 3 16 26 2 1 3 34 10 16 60 Percent recovered..- 0.8 0.2 0.0 0.0 0.2 2.8 4.0 9.9 3.7 SUMMARY: Total recovered. . 63 223 106 192 584 503 49 38 590 Percent recovered 24.3 39.2 27.5 38.5 34. 1 41.3 19.6 23.5 36.2 1 These data do not include recoveries reported from the point of tagging nor those doubtful as to location of capture. See text, p. 647. The original records of the tagging experiments from 1924 to 1927 are given in the reports of Rich 1926, Rich and Suomela 1927 , and Rich and Morton 1929. TAGGING EXPERIMENTS AT RUINS POINT AND CAPE DECISION Ruins Point is located on the east side and Cape Decision on the west side of the entrance to Sumner Strait. Point Colpoys is located in the central part of the strait where, through Snow Passage, Sumner Strait connects with the northern part of MIGRATIONS OF PINK SALMON 663 Clarence Strait. Because of the rather wide geographic separation of these points, separate summaries were made of the Ruins Point-Cape Decision experiments and the Point Colpoys experiments. Table 8. — Pink salmon tagged at Point Colpoys in 1935 and 1936 1 Locality of recovery Salmon recovered and average time enroute Aug. 13. 1935, 386 Aug. 16, 1936, 498 Total, 884 Kah Sheets Bay. Red Bay Point Baker Rocky Pass Port Beauclerc... Calder Bay Warren Channel. Sumner Strait.— Area 5 Nuviber Total recoveries Percent recoveries. Upper Clarence Strait.— Area 4 Snow Passage and Stikine Strait: Snow Passage Steamer Point Point Harrington, East Island to Ernest Point, and Narrow Point: Steamer Rock East Island Whale Passage Marsh Island Screen Islands Lincoln Rock. Lake Bay Eagle Creek.. Ratz Harbor Olson Cove.. Total recoveries Percent reeoveries. Middle Clarence Strait.— Area 3 Ernest Sound: Ernest Point Union Point and Watkins Point Emerald Bay Point Ward Narrow Point, Ernest Point to Approach Point, and Caamano Point: Meyers Chuck... Wolf Creek Ship Island Niblack Point Grindall Island Total recoveries Percent recoveries. Lower Clarence Strait.— Area 2 East shore Prince of Wales Island from Approach Point to Brown- son Bay: Chasina Point Moria Sound Ripp and Scott Points Cape Chacon Brownson Bay... Total recoveries Percent recoveries. Lower Clarence Strait.— Area 1 West shores of Gravina, Annette, and Duke Islands: West Shore of Gravina Island West Shore of Annette Island.. Seal Cove Behm Canal and Revillagigedo Channel: Cape Caamano Point Sykes Foggy Bay Total recoveries Percent recoveries. Summary of all recoveries. Percent of all recoveries ... 2 1 2 1 1 2 9 2.3 66 17. 1 23 6.0 2 0.5 106 27.5 Days Number 2 65 2 84 19 3 19 2 3 2 3 9 12 16 11 3 11 3 26 6 26 3 12 9 41 2 9 4 4 7 4 150 30. 1 6 1. 2 14 2.8 192 38.5 Days Number 2 2 1 2 8 1 18 2.0 216 24.4 36 4. 1 8 0.9 20 2.3 298 33.7 Days 1 These data do not include recoveries reported from the point of tagging nor those doubtful as to location of capture. p. 647. See text, 664 BULLETIN OF THE BUREAU OF FISHERIES Seven tagging experiments were carried on at Ruins Point in 1924 and 1925, and one at Cape Decision in 1927. Since these experiments were carried on at different times in the season, from July 12 to August 3, their results should indicate the desti- nations of the pink salmon that migrate through Sumner Strait at all times of the season. A total of 1,629 pink salmon were tagged, of which 590, or 36.2 percent, were recovered. Of these, 393, or 19.8 percent, were captured in Sumner Strait, area 5; 116, or 7.1 percent, in upper Clarence Strait, area 3; 9, or 0.6 percent, in lower Clarence Strait, area 2 ; 30, or 1 .8 percent, in lower Clarence Strait, area 1 ; and 60, or 3.7 percent, along the northwest shore of Prince of Wales Island, Chatham Strait, and Frederick Sound. Hence it may be assumed that most of the pink salmon migrating into Sumner Strait are bound for localities in the strait and the northern regions of Clarence Strait, with only a small percentage migrating to localities in the lower regions of Clarence Strait. It is also important to note that of the 52 tagged salmon recovered in middle Clarence Strait, area 3, 33 were captured in Ernest Sound, indicating that many of the Sumner Strait pink salmon use the northern region of Clarence Strait only as a means of reaching Ernest Sound and its adjoining channels. TAGGING EXPERIMENTS AT POINT COLPOYS Further and more exact proof of this distribution of the Sumner Strait pink salmon in Clarence Strait and Ernest Sound may be found in an analysis of the results from the Point Colpoys taggings. A total of five tagging experiments were carried on at Point Colpoys; one on July 10, 1926, one each on July 26 and 30, 1927, one on August 13, 1935, and one on August 16, 1936. Thus, these experiments cover a period from July 10 to August 16 and rep- resent both the odd- and even-year runs in equal proportion. The results were compiled in a manner similar to the procedure followed with the Clarence Strait data. In other words, they were used to show the seasonal differences in the distribution of the pink salmon to the spawning localities in the eastern section of Sumner Strait and its adjoining channels, and in Clarence Strait and Ernest Sound. The percentages of the tagged pink salmon recovered in the various geographic areas from each of the experiments are given in figure 5. From an inspection of this figure it will be noticed that very few of the salmon tagged at Point Colpoys were recovered in area 5 (Sumner Strait). The seasonal trend in the percentages of the recoveries in this area was not very definite and cannot be considered as indicative of any sea- sonal change in the number of pink salmon bound for the area. The seasonal trends in the percentages of recoveries in areas 3 and 4, on the other hand, were very marked and are certainly indicative of a seasonal increase in the number of pink salmon bound for the localities in area 4, and a definite seasonal decrease in the numbers bound for the localities in area 3. The percentage of recoveries in area 2, like those in area 5, area 5 4 3 2 J DATE or TAGGING & AREA °r RECOVERY Figure 5.— Trends in the percentage recoveries of tagged pink salmon from the 1926, 1927, 1935, and 1936 tagging experiments at Point Col- poys. MIGRATIONS OF PINK SALMON 665 was very small. The percentage of recoveries in area 1, although by no means as great as those in areas 3 and 4, did show some indication of a seasonal decrease. It is not improbable that the majority of the Sumner Strait pink salmon migrating as far south as area 1 in Clarence Strait come from the early, rather than the late, part of the run. SUMMARY OF POINT COLPOYS EXPERIMENTS In reviewing the distribution of the pink salmon tagged at Point Colpoys (see tables 7 and 8), it will be noted that the majority of the tagged salmon recovered during the early part of the season were captured in the Ernest Sound region of area 3. This region supports a large number of excellent spawning streams whose pink- salmon populations enter them during the early part of the season. The majority of the tagged salmon recovered during the latter part of the season, on the other hand, were captured in area 4. This region also supports a large number of excellent spawn- ing streams whose pink-salmon populations are known to migrate into them during the latter part of the season. Hence it may be assumed that most of the pink salmon migrating through Sumner Strait as far as Point Colpoys are bound for the localities in Ernest Sound and its contiguous channels, and the localities in the northern region of Clarence Strait. Furthermore, the pink salmon destined for localities in Ernest Sound and its adjoining channels, which are the farthest from the sea, pass through Sumner Strait early in the season, whereas those bound for the localities in Snow Pas- sage and the northern region of Clarence Strait, which are closer to the sea, migrate later in the season. The results from both the Clarence and Sumner Straits experiments indicate that most of the pink salmon migrating through these channels in the early part of the season are bound for the inside localities farthest from the sea, and as the season progresses they tend more and more to migrate into the localities which are closer to the sea. In view of these results the authors do not feel that the contention of the salmon packers, that their catches of pink salmon in area 2 during the latter part of the season are made from the runs migrating through Sumner Strait, is well founded. CONCLUSIONS Cape Fox region. — The pink salmon migrating through Dixon Entrance to the shores near Cape Fox are either bound for localities east of the cape or those in Revillagigedo Channel and the south arm of Behm Canal. Cape Chacon region. — Most of the pink salmon entering Clarence Strait by way of Cape Chacon, during the early part of the fishing season, are destined for localities in Revillagigedo Channel, the north arm of Behm Canal, and the more distant localities in the northern region of Clarence Strait. Most of those migrating by the same route later in the season are bound for localities on the east shore of Prince of Wales Island south of Approach Point. Thus, the early migrants are native to the streams farthest distant from the sea, whereas the later migrants are native to those in the more proximate localities. Cape Muzon region.- — The tagging experiments carried on near the entrance to Cordova Bay were all made during the latter part of the season. The results indicate that at this time the incoming pink salmon are bound for localities in Cordova Bay, with a small percentage continuing around Cape Chacon to the southeast shore of Prince of Wales Island. 666 BULLETIN OF THE BUREAU OF FISHERIES Point Colpoys region. — Those pink salmon migrating through Sumner Strait which pass Point Colpoys early in the season are destined for localities in Ernest Sound and the central region of Clarence Strait with a few continuing as far south as Revillagigedo Channel. Those passing the point later in the season are destined almost exclusively for localities in Snow Passage and the northern region of Clarence Strait. Thus, again, we find a seasonal difference in distribution. The early migrants are destined for the localities remote from the sea, the later migrants for the more proximate localities. LITERATURE CITED Davidson, F. A. 1934. The homing instinct and age at maturity of pink salmon ( Oncorhynchus gorbuscha). Bull. U. S. Bur. Fish., vol. XL VIII, No. 15. Washington. Foerster, R. E. 1929. An investigation of the life history and propagation of the sockeye salmon 0 Oncorhynchus nerka) at Cultus Lake, British Columbia. No. 3. The downstream migration of the young in 1926 and 1927. Contr. Can. Biol, and Fish., N. S., vol. 5, No. 3. Toronto. Gilbert, C. H. 1913. Age at maturity of the Pacific coast salmon of the genus Oncorhynchus. Bull. U. S. Bur. Fish., vol. XXXII, 1912 (1913). Washington. Pritchard, A. L. 1933. Natural run of pink salmon ( Oncorhynchus gorbuscha ) in Masset Inlet. Annual Report of Biol. Board of Canada, 1933, p. 92. Pritchard, A. L. 1934. Propagation of pink salmon. Annual Report of Biol. Board of Canada, 1934, p. 26. Rich, W. H. 1926. Salmon-tagging experiments in Alaska, 1924 and 1925. Bull. U. S. Bur. Fish., vol. XLII, Doc. No. 1005. Washington. Rich, W. H. 1932. Salmon-tagging experiments in Alaska, 1930. Bull. U. S. Bur. Fish., vol. XLVII, No. 11. Washington. Rich, W. H., and H. B. Holmes. 1928. Experiments in marking young chinook salmon on the Columbia River, 1916 to 1927. Bull. U. S. Bur. Fish., vol. XLIV, 1928 (1929). Washington. Rich, W. H., and F. G. Morton. 1929. Salmon-tagging experiments in Alaska, 1927 and 1928. Bull. U. S. Bur. Fish., vol. XLV, Doc. No. 1057. Washington. Rich, W. H., and A. J. Suomela. 1927. Salmon-tagging experiments in Alaska, 1926. Bull. U. S. Bur. Fish., vol. XLIII, Doc. No. 1022. Washington. Snyder, J. O. 1921. Three California marked salmon recovered. Calif. Fish and Game, vol. 7, No. 1, Jan. 1921, pp. 1-6, figs. 1-4. Sacramento. Snyder, J. O. 1922. The return of marked king salmon grilse. Calif. Fish and Game, vol. 8, No. 2, Apr. 1922, pp. 102-107, figs. 40-50. Sacramento. Snyder, J. O. 1923. A second report on the return of king salmon marked in 1919 in Klamath River. Calif. Fish and Game, vol. 9, No. 1, Jan. 1923, pp. 1-9, figs. 1-5. Sacramento. Snyder, J. O. 1924. A third report on the return of king salmon marked in 1919 in Klamath River. Calif. Fish and Game, vol. 10, No. 3, July 1924, pp. 110-114, pis. 1-2. Sacramento. U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner THE GEOGRAPHIC DISTRIBUTION AND ENVIRONMENTAL LIMITATIONS OF THE PACIFIC SALMON (GENUS ONCORHYNCHUS) By FREDERICK A. DAVIDSON and SAMUEL J. HUTCHINSON From BULLETIN OF THE BUREAU OF FISHERIES Volume XLVIII Bulletin No. 26 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1938 For sale by the Superintendent of Documents, Washington, D, C. Price 10 cents THE GEOGRAPHIC DISTRIBUTION AND ENVIRONMENTAL LIMITATIONS OF THE PACIFIC SALMON (GENUS ONCORHYNCHUS)1 By Frederick A. Davidson, Ph. D., Aquatic Biologist, and Samuel J. Hutchinson, B. S., Junior Aquatic Biologist, United States Bureau of Fisheries J* CONTENTS Page Introduction 667 Geographic distribution 668 Native 668 Foreign 669 Environmental limitations to occurrence 673 North Pacific region 673 South Pacific region 680 North Atlantic region ' 685 South Atlantic region 687 Summary 687 Literature cited 688 INTRODUCTION There are five principal species of Pacific salmon, all of which are classified in the genus Oncorhynchus. They are cliinook, or quinnat ( 0 . tschawytscha) ; sockeye, or red ( 0 . nerka)-, coho, or silver ( 0 . kisutch ); pink, or humpback ( 0 . gorbuscha) ; and chum, or dog (0. keta). These fish are anadromous; they spend part of their lives in the sea and part in the streams. The eggs are deposited in gravel beds in the streams and lakes during the summer and fall and hatch out during the following spring months. The fry remain in fresh water for varying lengths of time, depending upon the species, but all eventually migrate to the sea where they make over 95 percent of their growth. Upon attaining maturity in the sea the adults return to the streams where they spawn and die. The studies of Gilbert (1913), Snyder (1921-24), Rich and Holmes (1928), Pritchard (1933), Davidson (1934), and Foerster (1936), on the life histories of the Pacific salmon show that they have a pronounced homing instinct and in general return to their parent streams to spawn. The locations and depths at which the salmon feed while in the sea have not been definitely determined. Catches of cliinook and coho salmon are made by the troll fishery as far as 100 miles offshore and at depths as great as 90 fathoms. Com- mercial and Government vessels operating in Alaskan waters have reported the presence of the salmon even farther out to sea. The continental shelf along the Pacific coast of North America averages less than 40 miles in width, thus it is evident 1 Bulletin No. 26. Approved for publication January 14, 1938. 667 668 BULLETIN OF THE BUREAU OF FISHERIES that the feeding salmon frequent the waters of the open sea as well as those of the immediate coast. Bigelow and Welsh (1924), in discussing the habits of the pink salmon transplanted in the coastal streams of Maine, state that: During their first months in salt water the fry linger near the mouths of the home streams, where they feed chiefly on copepods and other small crustaceans, or pteropods, and on insects that drift down stream with the current, and occasionally on fish fry. After they are 5 or 6 inches long they move out into deep water, and very little is known of their habits and wanderings there- after until they reappear on the coast as adults to breed. Since the Pacific salmon live alternately in two distinctly different environments, fresh-water and marine, their geographic distribution is influenced by the limiting factors in each environment. This study was made for the purpose of determining the geographic distribution of the salmon and gaining knowledge of the environmental limitations to their occurrence. GEOGRAPHIC DISTRIBUTION NATIVE The native distribution of the Pacific salmon is confined almost entirely to the temperate and arctic waters of the North Pacific. They are found in the streams along both the North American and Asiatic coast lines within similar geographic limits. On the North American continent O’Malley (1920), Cobb (1930), and Ever- mann and Clark (1931), give Monterey Bay, 70 miles south of San Francisco, Calif., as the southernmost limit of their common occurrence, although a few specimens have been taken at odd times as far south as the Ventura River, Calif. From here O’Malley (1920), Gilbert (1922), and Cobb (1930), report them inhabiting the coastal streams, in varying degrees of abundance, northward along the continent to Kotzebue Sound in Bering Strait. Dymond and Vladykov (1933) give the probable occurrence of chum salmon in the Mackenzie River of northern Canada and the definite occurrence of this species in the Lena River of northern Siberia. These rivers flow into the Arctic Ocean. From the Lena River, the northernmost point of occurrence on the Asiatic continent, they are found to a limited extent southeastward along the Arctic shores to the Chukchee Peninsula in Bering Strait. From the Anadir River just south of the Chukchee Peninsula all species, according to Caldwell (1916), Lebedev (1920), Russian Economic Monthly (1920), Baievsky (1926), and Pravdin (1932), are present in varying degrees of abundance in the coastal streams southward along the continent to the Amur River. The range of the pink and chum salmon extends farther south- ward to the Tumen River in northern Korea which is given by Mori (1933) as the southernmost occurrence of these salmon. All species other than the chinook salmon, according to Jordan, Tanaka, and Snyder (1913), Tanaka (1931), Handa (1933), Oshima (1933), and Tokuhisa and Ito (1933), are found in the coastal streams of Sakhalin, Hokkaido, and Kurile Islands and the northern shore of Honshu Island. The range of the pink and chum salmon extends farther southward on Honshu Island to the Tonegawa River near Cape Inuboye on the eastern shore, and to the Omonogawa River near Akita on the west- ern shore. A report has also been received through correspondence from Dr. Fujita of the Hokkaido Imperial University, of the limited occurrence of the chum salmon GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 669 along the west shore of Honshu Island as far south as the Joganjigawa River. The geographic distribution of the Pacific salmon on both the North American and Asiatic continents is shown in figures 1, 2, 3, and 4. FOREIGN In 1872 the United States Commission of Fish and Fisheries, later the United States Bureau of Fisheries, established an egg-taking station (Baird Station, Calif.) on a tributary of the Sacramento River for the sole purpose of collecting chinook or quinnat salmon eggs for transplantation in foreign waters.2 This station formed the source of supply of millions of these eggs which were shipped to the Atlantic Coast States and to countries in many parts of the world. With the development of the Pacific coast, additional egg-taking stations were established in Oregon and Washington by the Bureau of Fisheries. Following 1900, these stations furnished eggs and young of the other species of the Pacific salmon that were likewise shipped to various parts of the world for transplantation. The introduction of salmon into foreign waters was continued actively through 1930. The number, by species of eggs and young shipped, and the States and foreign countries receiving them, are given by 10-year periods in tables 1 and 2. 3 Only those eggs and young that were transplanted in coastal streams for the purpose of developing natural sea-run popu- lations are included in these tables. All transplantations in inland waters for the establishment of landlocked populations have been omitted. Table 1. — Foreign distribution of Pacific salmon eggs and young [By 10-year periods, 1872-1930] CHINOOK SALMON (0. tschawytscha) Localities of distribution Periods and number distributed Total 1872-1880 1881-1890 1891-1900 1901-1910 1911-1920 1921-1930 STATES Connecticut 1. 410. 000 31. 400 79. 000 43. 400 215, 000 4, 445, 000 640, 000 43, 400 550, 000 2, S00, 000 975. 000 1. 150. 000 2, 545, 000 340. 000 200. 000 40, 000 1.270.000 1, 410, 000 31, 400 119.000 43, 400 3. 765. 000 5, 117, 200 768. 000 43, 400 2, 122, 670 3, 350, 000 9, 172, 190 1, 150, 000 2, 805. 000 340, 000 500, 000 516,820 1, 322, 000 1. 058. 000 150, 000 1.415.000 200, 000 150, 000 1, 053, 000 955. 000 Delaware.- Georgia 40, 000 Louisiana Maine 3, 450, 000 10, 000 400 100, 000 Maryland. .. 500, 000 22, 500 117, 500 139, 700 Massachusetts. 10, 100 Mississippi New Hampshire 50, 000 550, 000 50, 000 567, 960 184, 710 720, 000 New Jersey New York 7. 097, 400 114, 240 985, 550 North Carolina Pennsylvania 150, 000 100, 000 10, 000 Rhode Island South Carolina .. 300, 000 304, 070 7,000 50, 750 45, 000 1, 058, 000 122, 000 Virginia COUNTRIES Argentina Australia 100, 000 915, 000 50, 000 500, 000 Canada Chile 200, 000 England __ . 150, 000 358. 000 830. 000 y developmt 300, 000 395. 000 125. 000 tion see Stor Germany 2 For history of establishment and ear ,nt of this sta ie (1878). 3 For more detailed information on the data reported in these tables see United States Bureau of Fisheries reports on the prop- agation and distribution of food fishes (1871 to 1935). 670 BULLETIN OF THE BUREAU OF FISHERIES Table 1. — Foreign distribution of Pacific salmon eggs and young — Continued CHINOOK SALMON (0. tschawytscha) — Continued Localities of distribution Periods and number distributed Total 1872-1880 1881-1890 1891-1900 1901-1910 1911-1920 1921-1930 countries — continued 30, 000 99, 000 129. 000 50, 000 50. 000 100. 000 900. 000 ' 3,550,000 20. 000 25, 000 494. 000 50, 000 50, 000 50, 000 50, 000 500, 000 1, 175, 000 400, 000 775, 000 20, 000 1, 600, 000 « . 25, 000 494, 000 20, 835, 200 2, 475, 000 12, 523, 870 4, 040, 050 1, 442, 260 1, 558, 700 42, 875, 080 1 No information was secured on the disposition of these shipments. 2 This shipment was refused in Norway and sent to one of the northern Europe countries (see Aagaard, 1930). Table 2. — Foreign distribution of Pacific salmon eggs and young [By 10-year periods, 1901-30] SOCKEYE SALMON (O. nerka) Localities of distribution Periods and number distributed Total 1901-1910 1911-1920 1921-1930 STATE 17, 500 17, 500 397, 500 30, 700, 000 314, 000 COUNTRIES 397, 500 30, 700, 000 Chile 314, 000 Total-- 397, 500 30, 717, 500 314, 000 31, 429, 000 PINK SALMON (O. gorbuscha ) STATES 991, 141 27, 482. 826 15, 000 18, 000 1, 722, 340 6,350 30, 196, 307 21, 350 18, 000 2,000 2. 000 Total . _ - -. _ 993, 141 27, 515, 826 1, 728, 690 30, 237, 657 COHO SALMON (O. kisutch) Maine Maryland New Hampshire. New York Pennsylvania Vermont STATES 1,317,387 09, 800 12, 000 315, 000 5, 000 350. 000 5,800 8, 500 ~41,875~ 5, 600 1, 387, 187 12, 000 315, 000 13, 500 355, 600 47, 675 COUNTRIES Argentina _ Chile.. Total 377, 180 225, 040 377, 180 225, 040 2, 595, 407 120, 175 17, 600 2, 733, 182 GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 671 In this study a transplantation lias been considered successful only when it survived to the extent of producing subsequent sea-run populations with migratory and spawning habits characteristic of the species in the native distribution. In other words, the mere hatching of the eggs or rearing of the young under landlocked conditions has not been considered as indicating the successful introduction of the species. In order to secure complete and authentic information concerning the ultimate success or failure of the attempts to introduce these salmon into foreign waters, letters requesting this information were sent to the fish commissions and scientific fishery societies of the States and countries listed in tables 1 and 2. These letters, together with the replies that were received, are on file in the office of the Bureau of Fisheries, Washington, D. C. The history to date of the attempts to introduce the Pacific salmon in foreign waters is not very encouraging. Only 1 State and 3 countries reported the develop- ment of natural sea-run populations in their coastal streams. The others reported that negative results or, to the best of their knowledge, no natural populations had developed from the transplantations. Maine is the only State in which natural runs of salmon have been definitely established. According to Bigelow and Welsh (1924), pink-salmon fry planted in the Dennys, Medomak, St. Georges, St. Croix, Pembroke, and Penobscot Rivers survived and developed populations having characteristics similar to those in their native distribution. However, adverse sentiment of the residents in this region has greatly contributed to their present lack of abundance. Dymond, Hart, and Pritchard (1929) report the establishment of sea-run popu- lations of chinook salmon in the St. John River, New Brunswick, and the Port Credit River, Ontario. These salmon have been taken in the St. John River by the hundreds and vary in weight up to 8 pounds. They are also quite plentiful in the Port Credit River where fish weighing up to 30 pounds have been taken. It is assumed that the Port Credit salmon migrate to and from the sea by way of Lake Ontario and the St. Lawrence River. Other streams tributary to these waters may maintain small runs of chinooks which to date have not been identified. The streams and coastal regions of Maine, New Brunswick, and Ontario are the only foreign waters on the North American continent in which natural populations of the Pacific salmon have been established. Transplantations of Pacific salmon have been made in both Chile and Argentina in South America. Chile reports the presence of either coho or sockeye salmon running in the San Pedro River in the southern part of the country. Legislation has been promulgated by the Chilean Government which prohibits commercial fishing for these salmon until 1940. In Argentina none of the transplantations to date, according to Marini (1936), have been successful. However, final information in Chile as well as Argentina is not available owing to the lack of adequate scientific surveys throughout the sparsely inhabited regions in which the salmon have been introduced. All European waters stocked with Pacific salmon, according to Bottemanne (1879), Behr (1882), Aagaard (1930), and correspondence received, have been unfavorable to the survival of the species. The countries acknowledged receiving the eggs but none could cite a single instance in which adult salmon had returned from the sea 672 BULLETIN OF THE BUREAU OF FISHERIES to spawn. Various methods of propagation were used but none proved successful. From the many unsuccessful attempts at introducing the Pacific salmon into European waters it may be concluded that the establishment of sea-run populations in them is very improbable. However, Finland is at present importing chinook eggs with the hope of establishing natural runs of this species. The possibility of successfully introducing the Pacific salmon into the coastal waters of Norway may never be determined since the Norwegian Government has always, with thanks, declined offers to plant these salmon in their waters. Attempts to establish natural runs of Pacific salmon in the waters of Hawaii have been unsuccessful. The eggs were hatched successfully and the young reared to the migrant stage before planting but no adults have ever returned to the streams. Although recent shipments of eggs have been made to Hawaii it is not deemed ad- visable to continue this practice. The waters of Australia and Tasmania, according to McCulloch (1927), Tasmania Fisheries Commission (1933 and 1935), and correspondence received, have all been unfavorable to the introduction of Pacific salmon. Many attempts have been made to establish natural runs in the coastal streams but all have been unsuccessful. No particular difficulty was encountered in hatching the eggs and rearing the young to the migrant stage, see Baird (1878), but no adults ever returned from the numerous plantings in the streams. The Tasmania Fisheries Commission (1933) states that in the confined waters of the Great Lake, chinook salmon thrive and grow rapidly to support a flourishing sport fishery. Other than to maintain landlocked populations for sport fishing it is considered that attempts to stock the streams for the establish- ment of sea-run populations would not justify the necessary expenditure of eggs and effort. The introduction of the Pacific salmon into the waters of New Zealand has been successful only on South Island and even there, only within definite limits. The streams in which sea-run populations have been established, and those which have been stocked consistently with salmon but which have never developed sea-run populations, are shown in figure 5. The well-defined distributional range of the salmon on South Island will be explained in the discussion on environmental limitations to their occurrence. The many attempts to establish runs of chinook salmon in New Zealand prior to 1900 were all unsuccessful. During this period the eggs and young were divided into small consignments and distributed to many rivers throughout the colony. Following the year 1900 this practice was discontinued and only one river system, the Waitaki, was stocked. In 1905 many adult salmon returned to this river to spawn, thus establishing the first natural run of chinooks on South Island. This run survived and through natural and artificial propagation has been spread to other rivers on the island. Although the Bureau of Fisheries records show only the shipment of chinook salmon stock to New Zealand, shipments of sockeye salmon stock were also received from another source according to W. L. Calderwood (Fishery Board for Scotland, Salmon Fisheries, 1924, No. 2) who states: In operating with Sockeye, some curious results have appeared. Eggs were imported in 1902, and adult specimens of this fish began to appear in 1907. Dead examples were first noticed, and GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 673 these were found to have spawned and died in the usual way. A brief note in the last official year- book states that “a number exist in Lake Ohau, having acquired a landlocked habit. These fish run up creeks at the head of the lake and spawn there every season [year] in March and April.” Hatcheries have been constructed, and each year there is an abundant egg take. The eggs are hatched and the fry used to restock the parent stream or are planted in other streams on the island. Eggs collected and eyed in New Zealand have been sent to Tasmania for transplantation. Of the four successful foreign regions in the world to develop sea-run populations, New Zealand has been the most outstanding to date. The authors wish to acknowledge the kind cooperation of A. E. Hefford, Chief In- spector of Fisheries, New Zealand, in furnishing them with complete information concerning the history of the transplantation and development of the chinook salmon in the waters of New Zealand. Pictures of the chinook salmon and the streams in which they spawn, also scenes of the sport and commercial fisheries, are shown in figures 6-9. The sea-run populations of Pacific salmon that have been established in both Chile and New Zealand have adjusted their life cycle to the change in occurrence of the seasons in the Southern Hemisphere. The spawning migrations of these salmon occur in January, February and March which are the seasonal equivalents of July, August, and September in the Northern Hemisphere. The foreign regions into which the Pacific salmon have been introduced are shown in figures 1-4. The solid black areas indicate the regions in which the salmon have been transplanted successfully, and the dots show the regions in which transplantations failed. ENVIRONMENTAL LIMITATIONS TO OCCURRENCE During the first years of the introduction of the Pacific salmon into foreign waters very little was known concerning the proper methods for shipping or propagating these species. The failure of many transplantations to survive during this early period may have been due to excessive mortality in the eggs or unsuccessful rearing and planting of the young. However, with the improvement in fish-cultural methods the mortality during shipment and early propagation declined in importance so that following the year 1900 a high percentage of eggs shipped survived and the young were reared and planted successfully. This information was secured mainly through correspondence received from the various States and countries participating in this work. In view of this fact it is believed that the ultimate success or failure of these latter transplantations was dependent, to a high degree, upon the favorable or unfavor- able influences existing in the foreign waters in which they were made. NORTH PACIFIC REGION The environmental components of the fresh-water habitats in the native distri- bution of the salmon, which appear to be definite limiting factors, are temperature of water and character of stream bed. The degree of tolerance to temperature is much greater for the adults than for the eggs and young. The temperatures of the streams in which the salmon have been found spawning ranged from slightly above 0° to a maximum of 21°C. This range of temperature has been determined from recording thermographs operated yearly in Alaskan streams by the Bureau of Fisheries, and from stream surveys made by the Bureau’s biologists, in both the Pacific Coast States and 50352—38 2 674 BULLETIN OF THE BUREAU OF FISHERIES Alaska. Records taken at the Bureau’s hatcheries on the Pacific coast also show the range of temperature tolerated by the adult salmon. Adult salmon have also been found migrating through estuaries, and streams fed by hot springs, whose tempera- tures were as high as 27° C. The California Department of Public Works (1931) reports a temperature range of 16° to 26° C. in the lower reaches of the Sacramento River during the months the salmon are migrating into the river. In July and August it is not uncommon for the temperature of the lower estuary of the river to hover for days around 24° C. Although the temperature of the streams in the native range may fluctuate at a high level during the spawning period it rapidly decreases with the onset of the winter season during which time the eggs pass through the incubation period. The studies of Donaldson (1936) have shown that the eggs can withstand temperature below 4° and above 11° C. for short periods of time but that the optimum lies between these limits. The mortality was extreme in eggs maintained constantly at tempera- tures below 4° or above 11° C. After hatching, the optimum range of temperature in fresh water, which controls the rate of growth and survival of the young, shifts to a level of 13° to 17° C. Constant temperatures above 17° C. retarded growth and increased the mortality of the young and at 20° C. the mortality was excessive. Con- stant temperatures below i3° C. retarded growth and at 3° mortality was excessive. In view of the results from these studies it may be assumed that temperature in the fresh-water habitats becomes a limiting factor in the early developmental period of the salmon. The eggs of the Pacific salmon are spawned free and, being of a higher specific gravity than water, sink to the bottom. Eggs of this type require a medium that will hold and cover them for protection and at the same time permit the free flow of well- aerated water for incubation. Such a medium is found in clean gravel beds, or in pockets among rocks, but not in mud or sand. The former conditions are invariably found in all of the native fresh-water habitats of the salmon. Spawning in the streams is usually confined to the comparatively shallow areas where the current is swift, and in the lakes to areas provided with flowing water from seepage or surface drainage. In the Pacific Coast States deforestation, agricultural developments, and mining operations have, in some areas, produced excessive erosion of the watersheds. This has always resulted in the silting of the streams and the subsequent destruction of their salmon populations. An excessive amount of silt in the water influences the normal respiration of the salmon and destroys the eggs by suffocating them with a blanket of mud. The character of the stream bed, therefore, becomes a very definite limiting factor in the distribution of the Pacific salmon. The environmental components of the marine habitats in the native distribution, which appear to be limiting factors, are ocean currents and associated temperatures and salinities (salt content). The mean directional drifts in the North Pacific from June through September are shown in figure 1. This period was selected because it is during these months that the salmon are known to be definitely migrating in the ocean. The adults are migrating from the open ocean to the streams to spawn and the young are migrating seaward from the streams. The currents in figure 1 were determined from the limits of the directional drifts during this period as given by Dali (1880), Schulz (1911), McEwen (1912), Mariner (1926), Hatai and Kokubo (1928), Uda and Okamoto (1930), Uda (1931 and 1933), Schumacher (1932), Zeusler GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 675 (1934), Schott (1935), Thompson and Van Cleve (1936), and Wiist (1936). Since it is known that the salmon mature in the offshore waters and begin their spawning migration there, more emphasis was laid upon the general movements in these waters than upon the local and complex shiftings along the immediate shores. The Japan, or Kurosliio Current; Bering Sea, or Oyashio Current; Okhotsk Sea Current, and other less perceptible currents form an intricate maze of water move- Figure I. —The geographic distribution of the Pacific salmon and the directional drifts of ocean currents during the spawning migration period of the salmon. The bars indicate the native distribution of the salmon, the solid black areas indicate the regions in which the salmon have been transplanted successfully, and the dots indicate the regions in which the transplantations were unsuccessful. The directional drifts of the ocean currents in the Northern Hemisphere are means calculated from the monthly averages of June, July, August, and September, and in the Southern Hemisphere are means calculated from the monthly averages of January, February, and March. A, Japan or Kuroshio Current and West Wind Drift; B, North Equatorial Current; C, Okhotsk Sea Current; D, Bering Sea, Oyashio or East Kamchatka Current; E, West Australian Current; F, East Australian Current; H, South Equatorial Current; J, South Pacific Current and Antarctic Drift; K, Peruvian or Humboldt Current; N, East Greenland Current; P, Labrador Current; E, Irminger Current; S, Gulf Stream or Florida Current; T, Canaries Current; U, North Equatorial Current; V, South Equatorial Current; W, Brazil Current; X, Benguela Current; Y, Falkland Current; Z, South Atlantic Current and Antarctic Drift. ments iu the North Pacific. The Japan Current has the most outstanding circulation. It is a tropical drift originating from the North Equatorial Current off the east shore of the Philippine Islands. From here it flows northward and strikes the shores of the Islands of Japan. Part is forced to the west of the islands and enters the Japan Sea. The bulk, however, closely follows the east shore of Honshu Island to Cape Inuboye just east of Tokyo. Here it is met by the Bering Sea Current of arctic origin which 676 BULLETIN OF THE BUREAU OF FISHERIES flows southward along the island. The southern limit of occurrence of the Pacific salmon on this shore of the island is at Cape Inuboye near the point of confluence of these currents. A similar relation exists between the distributional limits of the salmon along the shores of the Japan Sea and the confluence of the Japan and Okhotsk Sea Current. The Okhotsk Sea Current, which is of arctic nature, flows southward along the continent and is dissipated in the waters off the northern shore of Korea. The Tumen River, which marks the distributional boundary of the salmon in this region, flows into these waters. After entering the Japan Sea the Okhotsk Sea Cur- rent influences the Japan Current flowing northward along the west coast of Honshu Island. The salmon are found on this shore of the island as far south as the Jogan- jigawa River which is near the southern point of confluence of these currents. There is a definite correlation between the distribution of the salmon and the influence of ocean currents in these regions. The Japan Current, after encountering the Bering Sea Current off the east coast of Honshu Island, takes a northeasterly course across the Pacific to form a fan-shaped divergence commonly known as the West Wind Drift. During this course it is greatly tempered. Upon nearing the coast of North America, off Vancouver Island, it divides into two branches; the northern branch forming the Alaska Current and the southern branch the California Current. According to McEwen (1912), an upwell- ing of cold waters along the coasts of Oregon and California influences the California Current as it flows southward. The southern distribution of the salmon on the North American continent appears to be correlated with the region dominated by this cold upwelling current. The adult salmon are subjected to the influence of surface temperatures in the ocean during their spawning migration to the streams, for it is definitely known that they frequent the surface waters at this time. Accordingly a study was made of the mean surface temperatures from June through September in relation to the distribution of the salmon. The mean temperatures rather than the limits of temperature during this period were used owing to the continuous character of the spawning migration which in some areas extends over the entire period from June through September. The mean surface temperatures in the North Pacific during this period are given by the isotherms in figure 2. These isotherms were determined from the seasonal and monthly surface temperatures given by Dali (1880), Rosse (1881), McEwen (1912), Uda and Okamoto (1930), Uda (1931), Kolcubo (1932), Zeusler (1934), and Schott (1935). Along the coasts of Japan and Korea the mean 20° C. isotherm touches the shores near the southern limits of the native range of the salmon. Accord- ing to Uda and Okamoto (1930), and Uda (1931), the surface temperatures of these coastal waters range from 15° C. in June to approximately 24° in September. On the North American continent the mean 15° C. isotherm touches the shores of California below the southern limits in the distribution of the salmon. According to Schott (1935), the surface waters along the coast of California have an annual range of only 3° C. The northern distribution of the salmon on both continents is bounded by the mean 5° C. isotherm. However, the salmon migrating to and from the Mackenzie, Lena, and other streams tributary to the Arctic Ocean may be subjected to surface temperatures only a few degrees above freezing. It is, therefore, possible that the GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 677 surface temperatures tolerated by the salmon during their spawning and seaward migrations approximate 0° C. at the minimum and are in the vicinity of 20° C. at the maximum. Information thus far available indicates that the Pacific salmon, during their sojourn in the sea, frequent the subsurface waters to depths of 200 meters. Hence, the mean annual temperatures at 200 meters were studied in relation to the distribu- tion of the salmon. These mean temperatures for the North Pacific are shown by the Figure 2.— The geographic distribution of the Pacific salmon and the mean surface ocean temperatures during the spawning migra- tion period of the salmon. The bars indicate the native distribution of the salmon, the solid black areas indicate the regions in which the salmon have been transplanted successfully, and the dots indicate the regions in which the transplantations were unsuccessful. The isotherms in the North Pacific Ocean give the mean surface temperatures for the period of June, July, August, and September. Those in the North Atlantic Ocean give the mean surface temperatures for the period of July, August, and September, and those in the South Atlantic and South Pacific Oceans give the mean surface temperatures for the period of January, February, and March. isotherms in figure 3. Schott (1935) gives a review of the oceanographic studies carried on in the North Pacific and describes the subsurface temperatures in this region. The isotherms in figure 3 were taken primarily from Schott’s review. Along the coasts of Japan and Korea the mean annual 10° and 5° C. isotherms, respectively, at 200 meters, describe the subsurface temperatures of the waters at the southern distributional boundaries of the salmon. The mean annual temperatures at 200 meters on the east coast of Honshu Island, Japan, decrease from 10° to 3° C. within the comparatively short distance from Cape Inuboye to the north end of the island. 60352—38 3 678 BULLETIN OF THE BUREAU OF FISHERIES Uda and Okamoto (1930), and Uda (1931), give the monthly temperatures at 100 meters along these shores. The temperatures at 100 meters along the Korean coast do not differ appreciably from those at 200 meters and vary no more than 3° C. throughout the year. The temperatures at 100 meters along the east shore of Honshu Island average from 3° to 4° C. higher than at 200 meters and also vary no more than 3° throughout the season. In general the subsurface temperatures in this region do not fluctuate widely throughout the year. Figure 3.— The geographic distribution of the Pacific salmon and the mean annual subsurface ocean temperature at 200 meters depth. The bars indicate the native distribution of the salmon, the solid black areas indicate the regions in which the salmon have been transplanted successfully, and the dots indicate the regions in which the transplantations were unsuccessful. The isotherms give the mean annual subsurface temperatures at 200 meters depth. The southern distribution of the salmon along the North American continent falls well within the region bounded by the mean 8° C. subsurface isotherm. Here again the subsurface waters vary only 2° to 3° C. throughout the year. In the Bering Sea the mean annual temperature at 200 meters is less than 3°, and in the Arctic Ocean it is less than 0° C. Hence, if the salmon frequent the subsurface waters to depths of 200 meters, they must be tolerant to temperatures ranging from slightly below 0° at the minimum to the vicinity of 10° C. at the maximum. Donaldson (1936) has shown that the optimum range of temperature for growth of the young salmon in fresh water is between 13° and 17° C. Furthermore, he found that mortality was excessive at constant temperatures of 20° and 3° C. In view of the GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 679 temperature data given in figures 2 and 3 it is obvious that the optimum range for growth in the sea must shift to a lower level. The mean annual temperatures of the ocean waters at 200 meters throughout almost the entire distribution of the salmon do not exceed 8° and in the northern half of the range they fluctuate close to 3° C. Even if the salmon frequented only the surface waters during their sojourn in the sea they would encounter temperatures of less than 15° C. throughout the greater part of the year, for it is only during the summer that the surface waters reach this temperature Figure 4. — The geographic distribution of the Pacific salmon and the mean annual surface salinities in parts per 1,000. The bars indicate the native distribution of the salmon, the solid black areas indicate the regions in which the salmon have been trans- planted successfully, and the dots indicate the regions in which the transplantations were unsuccessful. The isohalines give the mean annual surface salinities in parts per 1,000. and then only in the southern part of the range. In the northern part of the range the temperature of the surface waters at no time exceeds 10° C. Hence, the salmon must become acclimated to the colder waters of the ocean during their sojourn in them, and are able to grow and survive at lower temperatures than in fresh water. Since the salinity of the water forms an important environmental component of marine habitats, it was likewise studied in relation to the distribution of the salmon. Although salinity data for the subsurface waters would have been more desirable it was found that only surface data were available for all of the oceanographic regions. The mean annual surface salinities in parts per thousand for the North Pacific are shown by the isohalines in figure 4. The data from which these isohalines were 680 BULLETIN OF THE BUREAU OF FISHERIES determined were taken from the works of Schott (1928 and 1935), Uda and Okamoto (1930), and Uda (1931). The isohalines of 33 and 34 parts per thousand describe the mean annual salinities of the surface waters at the southern boundaries in the distribution of the salmon along the coasts of Korea and Japan. Owing to the direct contact of warm currents of high salinity with cold currents of low salinity, and the continuous shifting of these currents off the eastern coast of Honshu Island, the isohaline of 34 parts per thousand is not confined to any one district but shifts about over a broad area. Hence, the sal- mon frequenting the waters in this area may at times be subjected to surface salinities as high as 35 and as low as 33 parts per thousand. The southern distributional limits of the salmon on the North American continent fall within an area whose coastal waters are characterized by mean surface salinities from 33 to 34 parts per thousand. In the northern range of the salmon the mean salinities of the surface waters do not exceed 30 parts per thousand. It is, therefore, quite possible that the salmon orient themselves in the open ocean to surface waters of salinities ranging from 30 to 35 parts per thousand. The analysis of the marine habitats thus far has been confined mainly to the deter- mination of the ranges in certain physical and chemical properties of the waters in the North Pacific Ocean within the limits of the native distribution of the salmon. Briefly, it was found that the occurrence of the salmon is associated with the presence of ocean currents bearing waters of low temperature and salinity. The mean surface tempera- tures during the spawning migration period of the salmon ranged from 0° to 20° C. The mean annual temperatures at 200 meters ranged from slightly below 0° to 10° C. and the mean annual surface salinities varied from 30 to 35 parts per thousand. Since the salmon frequent the ocean waters of these temperatures and salinities, it may be assumed that they are tolerant to them. In this analysis, however, it has not been possible to definitely determine if temperatures and salinities outside these ranges are also tolerated by the salmon or form definite limiting factors governing their survival. The further analysis of this relationship may be found in a similar study of the marine waters in the foreign regions where the salmon have been transplanted. In other words, if the foreign marine waters in which the transplantations have survived have physical and chemical properties similar to those in the native distribution of the salmon and if the foreign waters where the transplantations have failed have properties unlike those in the native distribution, fresh-water conditions being favor- able to survival, then it is logical to assume that temperature and salinity values beyond the ranges of the native distribution may form limiting factors to the marine survival of the species. SOUTH PACIFIC REGION In discussing the foreign distribution of the Pacific salmon in this region it was pointed out that transplantations were made in Hawaii, Chile, New Zealand, Tas- mania, and Australia. Natural sea-run populations developed from the transplanta- tions in New Zealand and Chile but failed to develop from those planted in Hawaii, Tasmania, and Australia. New Zealand is composed of two large islands, known as North Island and South Island. Some of the streams on each island were stocked with chinook salmon from the Sacramento River, Calif., but only those on South Island have developed natural GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 681 sea -run populations and even here only within certain limits. The streams on both islands are shown in figure 5. The streams on South Island which support natural runs of chinooks are indicated by a solid circle, and those which have been stocked frequently from 1910 to 1929, but which have never developed natural runs, are indicated by a solid triangle. Does the explanation of this failure of the chinooks to develop natural runs in certain streams on South Island and in none of the streams on North Island lie in unfavorable conditions in the fresh- water or the marine environments? The streams on both North Island and South Island are quite similar in origin and type. Most of the larger ones originate in mountain lakes and flow rapidly to the sea over gravelly and rocky beds, see figures 6-8. Percival (1932), in describing the streams of New Zealand, states: The geological youthfulness of the present land-surface of New Zealand accounts for the rela- tive absence of slowly flowing rivers such as, in other countries, give shelter to a great variety of free-swimming organisms and allow of the growth of much vegetation on the bed. dfe H: ije 5|e sH # Generally speaking the rivers of New Zealand are comparable with the portions of the Euro- pean rivers called by Thienemann (28) “Aschenbach” (Grayling stream), where the bed is stony and liable to flooding through the accumulation of surface water. The streams on North Island, owing to the milder climate, are somewhat warmer than those on South Island. At Rotorua, North Island, the mean air temperature for January is 18° C. and for July is 7.5° C., while at Queenstown, South Island, the mean air temperature for January is 15.5° and for July is 3° C. Phillips (1929) reports stream temperatures on North Island, during the winter and spring, as low as 8° C., and during the fall from 12.5° to 15.5° C. Hobbs (1937) reports the mean monthly temperatures of salmon-bearing streams on South Island as ranging from 3° C. in midwinter to 16.5° C. in midsummer. Percival (1932), in discussing the presence of fish food in the streams on both islands, states that it is sufficiently abundant in most of the streams to support trout and other fresh-water fishes. In view of these facts it may be assumed that the streams on both islands provide favorable environ- mental conditions for the survival of the salmon during their fresh-water existence. An examination of the environmental conditions found in the coastal waters of the islands, however, gives an altogether different picture, for North Island is almost wholly bathed by a tropical current and South Island by an Antarctic current, The directional drifts of these currents are shown in figures 1 and 5. Hefford (1929), in discussing the reasons why runs of chinooks have not been established in the Wairau and Hokitika Rivers on South Island, and in all of the rivers on North Island, makes the following statement: It is known that off the south-eastern coasts of South Island the water in the sea is of Antarctic origin. There is a general set or drift in a north-easterly direction of cold water from the south, and this water produces the prevailing conditions in the sea off the Otago and Canterbury coasts where the quinnat have been established for some years. The South Equatorial Drift, which sets from the eastward and impinges upon the east coast of North Island, may be said to dominate the conditions to the northward of East Cape; while between that point and Cook Strait there is a mix- ture of this subtropical water with water from the south. For a long time navigators have been familiar with these “sets” or surface movements of the sea, but it was not until the hydrographer of the Danish research steamer Dana had applied physical and chemical tests to the water sampled at intervals between the east coast of Auckland and the coast of Otago, in January 1929, that the significant differences in the character of the water along this line were ascertained. It seems clear from the Dana's observations that the present distribution of quinnat salmon off the New Zealand coasts coincides with the occurrence of practically unmixed Antarctic water, with its characteristic physical and chemical qualities. Not a single individual of the quinnat species has ever been 682 BULLETIN OF THE BUREAU OF FISHERIES Figure 5. — The distribution of Chinook salmon in New Zealand. The dots indicate the streams on South Island in which natural sea-run populations of these fish have been established. The triangles indicate the streams which have been stocked frequently from 1910 to 1929 with young Chinooks but which have never developed natural runs. All of the streams on North Island have at some time or other been stocked with young Chinooks but have also never developed natural runs. The solid arrows indicate the directional drifts of cold currents. The broken arrows indicate the directional drifts of warm currents. U. S. Bureau of Fisheries, 1938 Bulletin No. 26 Figure 0. — A Chinook salmon caught at mouth of the Rangitata River, Canterbury, New Zealand. The general form and mark- ings, such as the spots on the back and fins, are typical Chinook salmon color markings. Gashes on the body are due to attacks by predators, probably barracuda. New Zealand Government Publicity Photo. Figure 7. — Fresh-water stream conditions as found in the Rakaia River, Canterbury, New Zealand. The clean gravel beds in the foreground and the snow-covered mountains in the background are prime factors in the fresh-water life history of the Chinook salmon. New Zealand Government Publicity Photo. U. S. Bureau of Fisheries, 1938 Bulletin No. 26 Figure 8. — Sport*fishery at mouth of the Rangitata River, Canterbury, New Zealand. Anglers in foreground landing a Chinook salmon, while others fish the surf at the mouth of the river. New Zealand Government Publicity Photo. Figure 9. — Commercial fishery as carried on along the Waimakariri River, Canterbury, New Zealand. Note range in size of the chinook salmon that were netted with a small seine carried in skiff at left. New Zealand Government Publicity Photo. GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 683 planted in a Canterbury stream, yet the Canterbury rivers now provide the best quinnat fishing in the Dominion, the species having migrated to their mouths from the Waitaki, where the original fry were planted. The Wairau has been fairly generously stocked and yet shows no appreciable run of fish. The inference is that it is probably too far north — outside the influence of the purely Antarctic water which attracts the bulk of the species — though an odd few are known to run into the Wairau, and, in fact, into some of the southern rivers of the North Island. This season an indubitable quinnat was caught in the Tukituki River, Hawke’s Bay. It does not follow that these parts are suitable for the permanent establishment of the species in abundance. The limit to which the influence of hydrographical factors pertaining to Antarctic waters extends will doubtless vary at different times, and it may be that in odd years the Cook Strait neighborhood, or even farther north, may provide suitable and congenial conditions for the quinnat salmon. But the indications afforded both by experience and by theoretical considerations seem to emphasize the probability of the fundamental relationship between the nature of the sea-water and the distribution of these salmon. There is also the case of the attempted acclimatization of the quinnat in the Hokitika River, on the west coast of the South Island. Our departmental reports show that between 1910 and 1924 the fry from over three million ova were planted in the head-waters of this river. The only apparent outcome has been a stock of lake-dwelling quinnat which has established itself in Lake Kanieri. As is well known, the west coast of the Dominion is washed by a warm current which has eddied across the Tasman Sea from the coast of eastern Australia, and which was origin- ally a branch of a westerly-trending subequatorial current. Again it seems to be a case of the wrong sort of sea-water for a salmon species. Where the quinnat smolts, which have presumably entered the Tasman Sea to the number of thousands or hundreds of thousands, have disappeared to is a mystery which may never be solved. This discussion is admittedly somewhat speculative, but it seems necessary to ventilate these considerations in view of the frequent recommendations, based rather on what is desirable than on what is probably feasible, to stock this or that river with salmon. The analysis of the physical and chemical properties of the coastal waters of New Zealand bears out Mr. Hefford’s assumptions as to the unfavorable character of the marine waters off the north and west coasts of South Island and the entire coast of North Island. The directional drifts of the ocean currents, the mean seasonal surface temperatures, the mean annual subsurface temperatures at 200 meters, and the mean annual surface salinities are given in figures 1-4, respectively. These data were calculated in the same manner as those for the North Pacific and were taken from the studies of Buchan (1894) and Schott (1928 and 1935). Schott (1935) gives a complete summary of all the hydrographic data collected in this region. The directional drifts of the ocean currents shown in figures 1 and 5 were cal- culated for the months of January, February, and March, which cover the spawning migration period of the Pacific salmon in this region. The South Pacific Current and the Antarctic Drift, which are so closely related that they may be considered as one current, carry waters of low temperatures and salinity. South of New Zealand a portion of this combined current divides. Part flows northward along the west coast of South Island and merges with a branch of the warm East Australian Current near the central coast of the island. Natural runs of California chinooks have been estab- lished only in the streams along this coast south of the point of confluence of these currents. The remainder flows northward along the east shores of South Island to Cook Strait where it is met by counter drifts from the warm South Equatorial Cur- rent. Natural runs of California chinooks have also been established only in the streams along this coast south of the point of confluence of the cold and warm currents. This shows a relationship between ocean drifts and the occurrence of natural runs of salmon similar to that found along the coasts of Korea and Japan. The mean surface and subsurface temperatures, and mean surface salinities, given in figures 2, 3, and 4, all show that the coastal waters of North Island are warmer and more saline than those of South Island. The mean 15° C. surface and 10° C. subsurface isotherms and mean 35 parts per thousand surface isohaline all touch the shores of South Island near the upper limits of the range in which the sal- 684 BULLETIN OF THE BUREAU OF FISHERIES mon have established natural runs. Similar mean temperatures are also found in the coastal waters near the native limits of occurrence of these salmon in California. Hence, from these observations in New Zealand, it may be assumed that the Cali- fornia chinooks react unfavorably to temperatures beyond the ranges found in their native habitats, but that they tolerate surface salinities of higher values up to 35 parts per thousand which is the maximum found throughout the entire native range of the Pacific salmon. Further demonstration of the unfavorable influence of coastal waters of high temperature and salinity on the marine survival of these salmon may be found in the failure of the attempts to introduce them into the streams of Hawaii, Australia, and Tasmania. The upper reaches of the streams in these countries have been fav- orable to the introduction of trout, whose fresh-water requirements are similar to those of the salmon. In fact, the salmon eggs shipped to these countries, according to correspondence received, Baird (1878), McCulloch (1927), and Tasmania Fish- eries Commission (1933 and 1935), were hatched without considerable loss and the young reared successfully to the stage of seaward migration. Landlocked populations of chinooks have been established in the Great Lake of Tasmania but no adults have ever returned from the plantings in the rivers, although chinook eggs were also im- ported from New Zealand for stocking them. No adults have ever returned from the many plantings of salmon made in the rivers of Australia and Hawaii. The mean directional drifts in figure 1 show that Hawaii, Australia, and Tas- mania are completely surrounded by currents of tropical origin during the spawning and seaward migration periods of the salmon. Schott (1935) shows that this same condition also prevails throughout most of the year. The mean isotherms and iso- halines given in figures 2, 3, and 4, show that in general the surface and subsurface temperatures and surface salinities of the coastal waters of these countries exceed the values found in the native marine habitats of the salmon. The mean surface temper- ature during the spawning migration period appears to be an exception in the case of Tasmania. This may indicate that all marine conditions must be favorable before survival of the salmon is possible. The attempts to introduce sockeye and coho salmon from Washington and Oregon into the waters of southern Chile have been successful. The coastal streams of southern Chile are similar in origin and character to the streams of southeastern Alaska. The climates of the two regions are also quite similar, being characterized by heavy rainfall and comparatively mild temperature. The hydrographic conditions of the waters along the southern coast of Chile (see figs. 1-4) are also similar to those in the native marine habitats of these salmon. The returns, thus far, of adult salmon have been reported only in the most southern streams in which transplantations were made. This does not mean, however, that other streams in the region are not suit- able for the establishment of natural runs, but merely that no returns have as yet been reported in them. It is for this reason that areas of both success and failure have been indicated on the distributional charts in figures 1-4. The successful trans- plantation of sockeye or coho salmon in Chile supports the conclusion that environ- mental conditions in both the marine and fresh waters of a foreign region must be similar to those in the native habitats of the salmon before successful introduction may be expected. GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 685 NORTH ATLANTIC REGION Many attempts have been made to introduce the Pacific salmon into the streams along the eastern coast of North America and the countries of northern Europe (see tables 1 and 2). Of these many transplantations all but those in the streams of Maine, New Brunswick, and Ontario were failures. The origin and character of the streams and lakes along the North American coast, north of the State of Maryland, indicate that they originally provided the physical requirements essential to the fresh- water survival of these salmon. Kendall (1935) states that the original range of the Atlantic salmon ( Salmo salar), which has fresh-water requirements similar to those of the Pacific salmon, probably extended from Delaware to Labrador. The establish- ment of natural runs of Pacific salmon in Maine, New Brunswick, and Ontario, as well as the development of landlocked populations in lakes throughout this region, gives further evidence of the suitability of these fresh waters for the introduction of these salmon. Many of the streams in this region have been gradually altered, through the intro- duction of power dams and pollution, so that at present they may not provide the essential requirements for the fresh-water survival of the salmon. However, these hazards were not so serious from 1872 to 1900, during which time the majority of the transplantations of Pacific salmon were made (see tables 1 and 2). Mather (1887) states that natural runs failed to develop from the transplantations of chinook salmon in the Hudson River but that the runs of Atlantic salmon in the river could be greatly improved through the introduction of eggs and young from other coastal streams. Since the Pacific and Atlantic salmon have similar fresh-water requirements, the indications are that the Hudson River, in 1887, provided the essential fresh-water conditions for both species. The failure of these salmon to develop natural runs in the coastal streams from Maryland to the Gulf of Maine cannot be wholly attributed to the presence of unfavorable conditions in them. In fact, the streams of Maine would still support natural runs of Alaska pink salmon had they not been destroyed through adverse sentiment. The consistent lack of returns from plantings made in the coastal streams of Virginia, North Carolina, South Carolina, Georgia, and espe- cially in the warm and muddy streams of Louisiana and Mississippi, may be in part attributed to their unsuitability for the fresh-water existence of the salmon. The millions of California chinook eggs sent to the countries of northern Europe, according to Baird (1878), Bottemanne (1879), Behr (1882), Aagaard (1930), and correspondence received, were all hatched with little loss and the young reared success- fully to the stage of seaward migration. Many of the young were also reared to the adult stage hi natural or artificial ponds in France, Germany, and Holland. In France these landlocked fish were spawned artifically for propagation in inland waters. The rivers and lakes of northern Europe in which the Chinooks were reared and liberated have in the past supported large populations of trout and Atlantic salmon, see Kendall (1935), all of which have fresh-water requirements similar to those of the chinooks. In fact, many of these rivers and lakes still support populations of Atlantic species. With the exception of artificial barriers and hazards introduced in these rivers through the progress of civilization, they all provide the essential conditions necessary for the fresh-water survival of the Pacific salmon. 686 BULLETIN OF THE BUREAU OF FISHERIES The warm silt-bearing streams of southern Europe, in which efforts were made to establish natural runs of chinooks, do not provide the conditions essential to the sur- vival of these fish. It is not surprising that the transplantations in these streams were unsuccessful. However, the failure of the chinooks to develop natural runs in the rivers of northern Europe cannot be logically attributed to this cause. Figures 1-4, inclusive, give hydrographic data for the North Atlantic Ocean similar to those given for the other oceanographic regions. These data were calculated in the same manner as in the other regions and were taken from the works of Rathbun (1882), Townsend (1901), Nansen (1913), Bigelow (1917 and 1933), Sandstrom (1918), Bjerkan (1919), Huntsman (1921), Dawson (1922), Schott (1926), Zeusler (1926), Smith (1928), Church (1932, 1934, and 1936), Helland-Hansen (1933), and Parr (1933). The mean directional drifts in figure 1 show the general movements of the North Atlantic waters from June through September, the period during which the spawning migration of the salmon occurs in this zone. Three major currents dominate the waters of the North Atlantic; namely, the Gulf Stream or Florida Current, the Labra- dor Current, and the East Greenland Current. The North Equatorial Current, banking up the waters in the Caribbean Sea and the Gulf of Mexico, gives rise to a strong current, the Gulf Stream, which flows out of the gulf through the straits between Florida, Cuba, and the Bahamas. This current follows the coast line of Florida and Southeastern United States until it reaches Cape Hatteras. Here it turns more to the eastward toward the banks of Newfound- land, thus allowing a cold current from the north to bathe the shores of Canada and the United States as far south as Cape Hatteras. However, the influence of this cold current is not appreciably effective south of Cape Cod. South of the banks of New- foundland the Labrador Current meets the Gulf Stream. This cold current has only a minor influence on the Gulf Stream as it continues eastward toward the coast of Europe. The cold current which bathes the Northeastern shore of the United States is not a continuation of the Labrador Current, but originates in the Gulf of St. Lawrence. As it leaves the gulf it turns southward and effectively carries waters of low tempera- ture and salinity to the shores as far south as Cape Cod. It is interesting to note at this point that the Pacific salmon have not developed natural runs in the coastal streams south of Cape Cod. Here again, as in the North and South Pacific regions, the occurrence of natural runs of these salmon is associated with the presence of ocean drifts bearing waters of low temperature and salinity. As the Gulf Stream follows its eastward drift toward the coast of Europe it branches into a number of lesser currents whose warm waters greatly temper the areas influenced by them (see fig. 1). The major branch passes to the northward of the Faeroes and flows toward and along the coast of Norway, where it divides and sends branches to Spitsbergen and the Barent Sea. A portion of the Gulf Stream also flows around Scotland and enters the North Sea. Other branches penetrate the English Channel and bathe the shores of France, Spain, and Portugal. The North Sea, which averages considerably less than 200 meters in depth, is readily influenced by the warm and saline waters of the Gulf Stream. The Baltic Sea, being likewise very shallow, warms rapidly in the summer months, and the waters flowing from it during this period fluctuate around 17° C. Furthermore, all of the fish in the Baltic Sea area migrating to and from the open ocean must pass through GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 687 the warm and saline waters of the North Sea. The failure of the California chinook salmon to develop sea-run populations in the streams tributary to these seas is con- sistent with their failure to develop natural populations under similar conditions in the South Pacific region. The mean temperatures and salinities of the coastal waters in the areas of the North Atlantic (see figs. 2, 3, and 4), where the Alaska and California salmon have failed to establish natural nms, are beyond the ranges of temperature and salinity found in the native marine habitats of these salmon. The areas in which they have been successfully established all have coastal waters with temperatures and salinities similar to those in the native marine habitats of the salmon. In other words, the reactions of the Pacific salmon to both fresh-water and marine environmental condi- tions in the North Atlantic are consistent with their reaction to similar conditions in the South Atlantic and South Pacific regions. SOUTH ATLANTIC REGION The only attempts to introduce the Pacific salmon into the South Atlantic region have been those made in the waters of southern Argentina, which have apparently failed. The streams of Argentina, with the possible exception of those in southern Patagonia, receive a great deal of drainage from plateaus and are essentially alluvium- bearing streams with sandy and mud bottoms. Since it is known that the Pacific salmon do not spawn in sandy- or mud-bottomed streams, nor coidd the eggs survive under such conditions even if so spawned, it is not surprising that sea-run populations have failed to develop from the transplantations in these streams. Marini (1936) also reports unfavorable high temperatures in some streams in which the salmon were transplanted. Complete surveys have not as yet been made of all the streams in Patagonia in which the salmon have been introduced. There may still be streams in the southern extremity of this province that will support natural runs which are at present unknown. The mean directional drifts of the currents, the mean seasonal surface tempera- tures, the mean annual subsurface temperatures (200 meters), and the mean annual surface salinities for the South Atlantic, are given in figures 1-4, inclusive. These data were taken from the oceanographical studies of Buchan (1894), Schott (1926), and Church (1934). The hydrographic conditions of the coastal waters of Patagonia, as given in these figures, are in every case similar to those found in the native marine habitats of the salmon. Hence, it appears that the failure of the attempts to intro- duce these salmon in Argentina lies in the unfavorable environmental conditions in its fresh waters. SUMMARY The native distribution of the Pacific salmon (genus Oncorhynchus ) is confined almost entirely to the temperate waters of the North Pacific. They are found in varying degrees of abundance along the North American coast from Monterey Bay, Calif., to Kotzebue Sound, Bering Sea, and along the Asiatic coast from the Anadir River, Siberia, to the Tumen River, Korea, and Cape Inuboye, Honshu Island, Japan. They also occur in isolated streams along the Arctic coast. 688 BULLETIN OF THE BUREAU OF FISHERIES From 1872 to 1930, millions of eggs and young of Pacific salmon from California, Oregon, Washington, and Alaska were shipped to the Atlantic Coast States and foreign countries for the purpose of establishing natural runs in their coastal streams. Transplantations were made in Hawaii, Australia, Tasmania, New Zealand, Chile, Argentina, Eastern United States, Eastern Canada, England, Ireland, France, Hol- land, Germany, Finland, and Italy. Of these many transplantations only those in New Zealand, Chile, the State of Maine, and the provinces of New Brunswick and Ontario have developed natural populations of these salmon with characteristics similar to those in their native distribution. The environmental components, as considered in this study of the foreign streams and lakes and coastal waters in which these salmon have developed natural runs, have in every case been similar to the components of the waters frequented by the salmon in their native range. On the other hand the environmental components of the foreign waters in which these salmon have failed to develop natural runs have differed from those of the waters native to the salmon. The failures of the transplantations in some areas have been due to the lack of suitable fresh-water conditions; in others, to the lack of suitable marine conditions, while some areas provided neither fresh water nor marine conditions favorable to the introduction of the salmon. Owing to the dependence of the Pacific salmon on particular environmental conditions, as shown in this study, there are no oceanographic regions in the world that can support populations of these fish as great as those supported by the North Pacific region. LITERATURE CITED Aagaard, Birger. 1930. B0r vi innf0re Stillehavslaksen? Fiskeri-inspekt0rens innberetning om ferskvannsfiskeriene for &ret 1930 (1931), pp. 32-33. Oslo. Baievsky, Boris. 1926. Fisheries of Siberia. Appendix II, Report, U. S. Comm. Fish., 1926 (1927), pp. 37-64, 2 figs. Washington. Baird, S. F. 1878. Correspondence connected with the transmission of eggs of the quinnat salmon and whitefish to Australia and New Zealand, 1877, 1878, and prior years. Appendix XL, Report, U. S. Comm. Fish., part. VI, 1878 (1880), pp. 825-905. Washington. Behr, Herr Von. 1882. Five American Salmonidae in Germany. Bull. U. S. Bur. Fish., vol. II, 1882 (1883), pp. 237-246. Washington. Bigelow, H. B. 1917. Explorations of the coast water between Cape Cod and Halifax in 1914 and 1915, by the U. S. Fisheries Schooner Grampus. Oceanography and Plankton. Bulletin of the Museum of Comparative Zoology, vol. 61, No. 8, July 1917, pp. 161-359. Cambridge, Mass. Bigelow, H. B. 1933. Studies of the waters on the continental shelf, Cape Cod to Chesapeake Bay. 1, The Cycle of Temperature. Papers in Physical Oceanography and Meteorology, vol. II, No. 4, contribution No. 34, Dec. 1933. Mass. Inst, of Technology and Woods Hole Oceanographic Inst., Cambridge, Mass. Bigelow, H. B., and W. W. Welsh. 1924. Fishes of the Gulf of Maine. Bull. U. S. Bur. Fish., vol. XL, part 1, 1924 (1925). Washington. Bjerkan, Paul. 1919. Results of the hydrographical observations made by Dr. Johan Hjort in the Canadian Atlantic waters during the year 1915. Canadian Fisheries Expedition, 1914-15. Department of the Naval Service, 1919, pp. 349-403. Ottawa. Bottemanne, C. J. 1879. California salmon in the Netherlands. Appendix XXVI, Report, U. S. Comm. Fish., part VII, 1879 (1882), pp. 709-713. Washington. GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 689 Buchan, Alexander. 1894. The voyage of H. M. S. Challenger. A summary of the scientific results. Appendix (Physics and Chemistry, Part VIII). Report on oceanic circulation. Challenger Reports, Physics and Chemistry — vol. 1, 1872-76 (1895). Neill and Co., Edin- burgh. Caldwell, John K. 1916. Fishing in the Priamur district of Siberia. Appendix VI, Report, U. S. Comm. Fish., 1916 (1917). Washington. California Department of Public Works. 1931. Economic aspects of a salt water barrier below confluence of Sacramento and San Joaquin Rivers. State of Calif. Dept, of Public Works, publications of the Division of Water Resources, Bull. No. 28, 1931. Calif. State Printing Office, Sacramento. Church, Phil E. 1932. Progress in the investigation of surface-temperatures of the western North Atlantic. Trans. American Geophysical Union, 1932, pp. 244-249. Mass. Church, Phil E. 1934. Surface-temperatures on the New York to Cape Sao Roque steamship- route. Trans American Geophysical Union, 1934, pp. 205-208. Mass. Church, Phil E. 1936. Temperatures of the western North Atlantic from thermograph records Association D’Oceanographie Physique, Union Geodesique et Geophysique Internationale. Publication Scientifique No. 4, 1937. Liverpool. Clark, G. H. 1929. Sacramento-San Joaquin salmon ( Oncorhynchus tschawytscha) fishery of California. Division of Fish and Game of California, Fish Bull. No. 17, 1929, pp. 1-73, 32 figs. Calif. State Printing Office, Sacramento. Cobb, John N. 1930. Pacific salmon fisheries. Appendix XIII, Report, U. S. Comm. Fish., 1930 (1931), pp. 409-704, 48 figs. Washington. Dall, Wm. H. 1880. Report on the currents and temperatures of Bering Sea and the adjacent waters. Appendix No. 16, Report, U. S. Coast and Geodetic Survey. Fiscal year ending June 1880 (1882), pp. 297-340. Washington. Davidson, F. A. 1934. The homing instinct and age at maturity of pink salmon ( Oncorhynchus gorbuscha) . Bull. U. S. Bur. Fish., vol. XLVIII, 1934, No. 15, pp. 27-39, 10 figs. Washington. Dawson, W. Bell. 1922. Temperatures and densities of the waters of eastern Canada. Depart- ment of the Naval Service, 1922, pp. 1-102. Ottawa. Donaldson, L. R. 1936. Unpublished data; School of Fisheries, U. of W., 1936, Seattle, Wash. Dymond, J. R., J. L. Hart, and A. L. Pritchard. 1929. The fishes of the Canadian waters of Lake Ontario. LTniversity of Toronto Studies, Publications of the Ontario Fisheries Research Laboratory, No. 37, 1929. Toronto. Dymond, J. R., and V. D. Vladykov. 1933. The distribution and relationship of the salmonoid fishes of North America and North Asia. Proceedings of the Fifth Pacific Science Congress, vol. V, 1933 (1934), pp. 3741-3750. The University of Toronto Press, Canada. Evermann, B. W., and H. W. Clark. 1931. A distributional list of the species of freshwater fishes known to occur in California. Division of Fish and Game of California, Fish Bull. No. 35, 1931, pp. 1-67. Calif. State Printing Office, Sacramento. Foerster, R. E. 1936. The return from the sea of sockeye salmon ( Oncorhynchus nerka) with special reference to percentage survival, sex proportions, and progress of migration. Jour. Biol. Board Canada, vol. Ill, No. 1, 1936, pp. 26-42. Toronto. Gilbert, C. H. 1913. Age at maturity of the Pacific coast salmon of the genus Oncorhnychus. Bull. U. S. Bur. Fish., vol. XXXII, 1912 (1914), pp. 1-22, 29 figs. Washington. Gilbert, C. H. 1922. The salmon of the Yukon River. Bull. U. S. Bur. Fish., vol. XXXVIII, 1921-22 (1923), pp. 317-332. Washington. Handa, Yoshio. 1933. Salmon propagation in Hokkaido. Proceedings of the Fifth Pacific Science Congress, vol. V, 1933 (1934), pp. 3601-3605. The University of Toronto Press, Canada. Hatai, S., and S. Kokubo. 1928. The marine biological station of Asamushi. Its history, equip- ment, and activities. Records of Oceanographic Works in Japan, vol. I, No. 1, March 1928, pp. 26-38. Tokyo. 690 BULLETIN OF THE BUREAU OF FISHERIES Hefford, A. E. 1929. Fisheries Report. New Zealand. Marine Department. Report on fisheries for the year ended 31st March, 1929, pp. 10-11. Wellington. Helland-Hansen, B. 1933. Oceanographic observations in the northernmost part of the North Sea and the southern part of the Norwegian Sea. James Johnstone Memorial Volume, Lan- cashire Sea-Fisheries Laboratory, 1933 (1934), pp. 257-274. University Press, Liverpool. Hobbs, D. F. 1937. Natural reproduction of quinnat salmon, brown and rainbow trout in certain New Zealand waters. New Zealand. Marine Department. Fish. Bull. No. 6, 1937. Welling- ton. Huntsman, A. G. 1921. Climates of our Atlantic waters. Trans. American Fish. Soc., vol. L, 1920-21 (1921), pp. 326-333. Washington. Jenkins, J. T. 1921. A Textbook of Oceanography. Constable and Co. Ltd., 10-12 Orange Street, W. C. 2, London. Jordan, D. S., S. Tanaka, and J. O. Snyder. 1913. A catalogue of the fishes of Japan. Journal of the College of Science, Imperial University of Tokyo, vol. XXXIII, art. 1, March 31, 1913. Tokyo. Kendall, W. C. 1935. The fishes of New England. The salmon family. Part 2. — The salmons. Memoirs of the Boston society of natural history, vol. 9, No. 1, 1935. Boston. Kokubo, S. 1932. Quantitative studies on microplankton in Aomori Bay during 1929-30. Rec- ords of Oceanographic Works in Japan, vol. IV, No. 1, June 1932. Tokyo. Lebedev, E. T. 1920. Japan’s trade in Russian humpbacked salmon. The Russian Far East Economic Monthly, No. 1, October 1920, pp. 12-16. K. Lavrov, editor, 4 Hikava-Cho, Akasaka-Ku, Tokyo, Japan. Marini, Tomas L. 1936. Los salmonidos en nuestro parque nacional de Nahuel Huapi. Comen- tarios a los trabajos de piscicultura realizados en el pais. 1936. Est. Grafico “Tomas Pal- umbo,” 321 - La Madrid - 325, Buenos Aires, Argentina. Marmer, H. A. 1926. Coastal currents along the Pacific coast of the United States. U. S. Coast and Geodetic Survey, special publication No. 121, Serial No. 330, pp. 1-80. Washington. Mather, F. 1887. Report upon the results of salmon planting in the Hudson River. Bull. U. S. Bur. Fish., vol. VII, 1887 (1889), pp. 409-424. Washington. McCulloch, Allan R. 1927. The fishes and fish-like animals of New South Wales. Royal Zoological Society of New South Wales, 2d edition, 1927. Box 2399, General Post Office, Sydney. McEwen, George F. 1912. The distribution of ocean temperatures along the west coast of North America deduced from Ekman’s theory of the upwelling of cold water from the adjacent ocean depths. Internationale Revue der Gesamten Hydrobiologie und Hydrographie, vol. V, 1912, pp. 243-286. Dr. Werner Klinkhardt, Leipzig. Mori, T. 1933. On the geographical distribution of Korean Salmonidae. Proceedings of the Fifth Pacific Science Congress, vol. V, 1933 (1934), pp. 3775-3776. The University of Toronto Press, Canada. Nansen, Fridtjof. 1913. The waters of the north-eastern North Atlantic. Cruise of the Frithjof; Norwegian Royal Navy; July 1910 (1913). Dr. Werner Klinkhardt, Leipzig. New Zealand, Marine Department. 1928. Report on fisheries for the year ended 31st March, 1928. By authority: W. A. G. Skinner, Govt. Printer, Wellington. New Zealand, Marine Department. 1929. Report on fisheries for the year ended 31st March, 1929. By authority: W. A. G. Skinner, Govt. Printer, Wellington. New Zealand, Marine Department. 1930. Report on fisheries for the year ended 31st March, 1930 (1931). By authority: W. A. G. Skinner, Govt. Printer, Wellington. New Zealand, Marine Department. 1931. Report on fisheries for the year ended 31st March, 1931. W. A. G. Skinner, Govt. Printer, Wellington. New Zealand, Marine Department. 1932. Report on fisheries for the year ended 31st March, 1932. W. A. G. Skinner, Govt. Printer, Wellington. New Zealand, Marine Department. 1933. Report on fisheries for the year ended 31st March, 1933 (1934). G. H. Loney, Govt. Printer, Wellington. New Zealand, Marine Department. 1934. Report on fisheries for the year ended 31st March, 1934. G. H. Loney, Govt. Printer, Wellington. GEOGRAPHIC DISTRIBUTION OF THE PACIFIC SALMON 691 New Zealand, Marine Department. 1935. Report on fisheries for the year ended 31st March, 1935. G. H. Loney, Govt. Printer, Wellington. O’Malley, Henry. 1920. Artificial propagation of the salmons of the Pacific coast. Appendix II, Report, U. S. Comm. Fish., 1919 (1921), pp. 1-32, 9 pis., 11 figs. Washington. Oshima, M. 1933. Life-history and distribution of the freshwater salmons found in the waters of Japan. Proceedings of the Fifth Pacific Science Congress, vol. V, 1933 (1934), pp. 3751-3773. The University of Toronto Press, Canada. Parr, A. E. 1933. A geographic-ecological analysis of the seasonal changes in temperature con- ditions in shallow water along the Atlantic coast of the United States. Bull, of the Bingham Oceanographic Laboratory, vol. IV, art. 3. Yale University, New Haven, Conn. Percival, E. 1932. On the depreciation of trout-fishing in the Oreti (or new river), Southland, with remarks on conditions in other parts of New Zealand. New Zealand. Marine Depart- ment. Fish. Bull. No. 5, 1932. Wellington. Phillips, J. S. 1929. A report on the food of trout and other conditions affecting their well-being in the Wellington district. New Zealand. Marine Department. Fish. Bull. No. 2, 1929. Wellington. Pravdin, I. F. 1932. The humpback-salmon from the Amour River. U. S. S. R. Institute of Fresh Water Fisheries. Bull. vol. 14, 1932, pp. 53-98, English summary, pp. 94-98. Leningrad. Pritchard, A. L. 1933. Natural run of pink salmon ( Oncorhynchus gorbuscha (Walbaum)), in Massett Inlet. Ann. Rept. Biol. Board, Canada, 1933 (1934), pp. 92-95. Ottawa. Rathbun, Richard. 1882. Ocean temperatures of the eastern coast of the United States, from observations made at twenty-four light-houses and light-ships. Section III, Fishery Industries of the United States, 1882 (1887), pp. 157-238, 32 charts. Washington. Rich, Willis H., and Harlan B. Holmes. 1928. Experiments in marking young chinook salmon on the Columbia River, 1916-27. Bull. U. S. Bur. Fish., vol. XLIV, 1928 (1929), pp. 215-264, 85 figs. Washington. Rosse, Irving C. 1881. Cruise of the Revenue-Steamer Corwin in Alaska and the N. W. Arctic Ocean in 1881. (1883). Washington. Russian Economic Monthly. 1920. Fisheries of Siberia and the Far East (written in English). The Russian Far East Economic Monthly, No. 2, Nov. 1920. K. Lavrov, editor, 4 Hikava- Cho, Akasaka-Ivu, Tokyo, Japan. Sandstrom, W. J. 1918. The hydrodynamics of Canadian Atlantic waters. Canadian Fisheries Expedition, 1914-15. Department of the Naval Service, 1919, pp. 221-345. Ottawa. Schott, G. 1926. Geographie des Atlantischen Ozeans. 1926, 2d edition. Verlag von C. Boysen, Hamburg, Germany. Schott, G. 1928. Die Verteilung des Salzgehaltes im Oberflachenwasser der Ozeane. Ann. d. Hydr. usw., LVI. Jahrg. (1928), Heft V, 1928, pp. 145-166. Hamburg, Germany. Schott, G. 1935. Geographie des Indischen und Stiffen Ozeans. 1935. Verlag von C. Boysen, Hamburg, Germany. Schulz, Bruno. 1911. Die Stromungen und die Temperaturverhaltnisse des Stiffen Ozeans nordlich von 40 N-Br. einschliesslich des Bering-Meeres. Annalen der hvdrographie und maritimen Meterorologie, bd. 39, Apr. und May, 1911, pp. 177-190, 242-264. E. S. Mittler & Sohn, Berlin. Schumacher, A. 1932. Movements of sea water. A survey of present knowledge of oceanic circulation based upon modern physical and chemical observations. Physics of the Earth — V, Oceanography. Bull, of the National Research Council, No. 85, June 1932, pp. 358-383. The National Academy of Sciences, Washington. Smith, Edward H. 1928. The Marion expedition to Davis Strait and Baffin Bay. U. S. Coast Guard, Bull. No. 19, Scientific Results, part 3, 1928 (1931). Washington. Snyder, J. O. 1921. Three California marked salmon recovered. Calif. Fish and Game, vol. 7, No. 1, Jan. 1921, pp. 1-6, figs. 1-4. Sacramento. Snyder, J. O. 1922. The return of marked king salmon grilse. Calif. Fish and Game, vol. 8, No. 2, Apr. 1922, pp. 102-107, figs. 40-50. Sacramento. Snyder, J. O. 1923. A second report on the return of king salmon marked in 1919 in Klamath River. Calif. Fish and Game, vol. 9, No. 1, Jan. 1923, pp. 1-9, figs. 1-5. Sacramento. 692 BULLETIN OF THE BUREAU OF FISHERIES Snyder, J. O. 1924. A third report on the return of king salmon marked in 1919 in Klamath River. Calif. Fish and Game, vol. 10, No. 3, July 1924, pp. 110-114, pis. 1-2. Sacramento. Stone, L. 1878. Report of operations at the United States salmon-hatching station on the McCloud River, Calif., in 1878. Appendix XXXIII, Report, U. S. Comm. Fish., part VI, 1878 (1880), pp. 741-770. Washington. Tanaka, S. 1931. On the distribution of fishes in Japanese waters. Journal of the Faculty of Science, sec. IV, Zoology, vol. Ill, part 1, 1931. Imperial University of Tokyo, Japan. Tasmania Fisheries Commission. 1933. Salmon and freshwater fisheries commission. Report for year ending 30th June, 1933. (No. 11.) Walter E. Shimmins, Govt. Printer, Hobart, Tasmania. Tasmania Fisheries Commission. 1935. Salmon and freshwater fisheries commission. Report for years ending 30th June, 1934, and 30th June, 1935. (No. 7.) Walter E. Shimmins, Govt. Printer, Hobart, Tasmania. Thompson, T. G., B. D. Thomas, and C. A. Barnes. 1934. Distribution of dissolved oxygen in the North Pacific Ocean. James Johnstone Memorial Volume, Lancashire Sea-Fisheries Laboratory, 1934, pp. 203-234. University Press, Liverpool. Thompson, W. F., and Richard Van Cleve. 1936. Life history of the Pacific halibut, (2) Dis- tribution and early life history. Report of the International Fisheries Commission, No. 9, 1936. Wrigley Printing Co. Ltd., Vancouver, B. C. Tokuhisa, M. and T. Ito. 1933. On the artificial propagation of salmon, trout, and other kinds of fish in Japan. Proceedings of the Fifth Pacific Science Congress, vol. V, 1933 (1934), pp. 3599-3600. The University of Toronto Press, Canada. Townsend, C. H. 1901. Dredging and other records of the United States Fish Commission steamer Albatross, with bibliography relative to the work of the vessel. Appendix IV, Report, U. S. Comm. Fish., 1900 (1901), pp. 387-562, 7 pis. Washington. Uda, M. 1931. Of the monthly oceanographical charts of the adjacent seas of Japan based on the averages for the thirteen years, 1918-30, with a discussion of the current-system inferred from these charts. Journal of the Imperial Fisheries Experimental Station, No. 2 (papers No. 13-23), Sept. 1931, pp. 80-82, 12 pis. Tokyo, Japan. Uda, M. 1933. The results of simultaneous oceanographical investigations in the North Pacific Ocean adjacent to Japan made in August 1933. Journal of the Imperial Fisheries Experi- mental Station, No. 6 (papers No. 43-51), March 1935, pp. 126-130. Tokyo, Japan. Uda, M., and G. Okamoto. 1930. Of the monthly oceanographical charts of the adjacent seas of Japan based on the averages for the eleven years, 1918-29, with a discussion of the current system inferred from these charts. (Part 1: From July to December.) Journal of the Im- perial Fisheries Experimental Station, No. 1 (vol. 1, pt. 1), Nov. 1930, pp. 54-56, 12 pis. Tokyo, Japan. United States Bureau of Fisheries. 1871-1935. Propagation and distribution of food fishes. (For each year, 1871 to 1935, inch) Reports, U. S. Comm. Fish., parts I to XXIX and years 1904 to the fiscal year 1935, inch Washington. Wilmot, S. 1881. Introduction of California salmon into Ontario, with remarks on the disap- pearance of Maine salmon from that province. Bull. U. S. Bur. Fish., vol. I, 1881 (1882), pp. 347-349. Washington. Wust, Georg. 1936. Kuroshio und Golfstrom. Eine Vergleichende Hydrodynamische Unter- suchung. Veroffentlichungen des Instituts fur Meereskunde an der Universitat Berlin. Neue Folge, Heft 29. Geographisch-naturwissenschaftliche Reihe. Germany. Zeusler, F. A. 1926. International ice observation and ice patrol service in the North Atlantic Ocean. Season of 1925. U. S. Coast Guard Bull. No. 13. Washington. Zeusler, F. A. 1934. Report of oceanographic cruise United States Coast Guard Cutter Chelan, Bering Sea and Bering Strait, 1934, and other related data. U. S. Coast Guard, 1934 (1936). Washington. U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner THE SALMON AND SALMON FISHERIES OF SWIFTSURE BANK, PUGET SOUND, AND THE FRASER RIVER By GEORGE A. ROUNSEFELL and GEORGE B. KELEZ From BULLETIN OF THE BUREAU OF FISHERIES Volume XLIX Bulletin No. 27 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1938 For sale by the Superintendent of Documents. Washington, D. C. Price 30 cents THE SALMON AND SALMON FISHERIES OF SWIFT- SURE BANK, PUGET SOUND, AND THE FRASER RIVER 1 By George A. Rounsefeel, Ph. D., and George B. Kelez, M. A. & CONTENTS Page Introduction. By George A. Rounsefell and George B. Kelez 694 The Pacific salmons 695 Fishing districts 695 Development of the fisheries 697 Production and value 699 Need for investigation 700 Acknowledgments 701 Gill net fishery. Bv George A. Rounse- fell 701 Fraser River 701 Early commercial development. _ 701 Relative importance of different species 702 Number of canneries . 704 Evaluation of fishing intensity.. 704 Company licensing system.. 704 Nationality of the fishermen. 705 Number of licenses 705 Units of fishing effort 705 Changes in gill-net boats 707 Changes in the gill net 708 Fishing seasons 709 Changes in location of the canneries 710 Seasonal occurrence of each species 711 Puget sound 712 Localities fished 712 Relative importance of various species 713 Trap fishery. By George A. Rounsefell.. 713 Reef nets 713 Construction of the traps 714 Number in operation 716 Locations fished 717 Cannery expansion from the trap fishery 719 Season 719 Seasonal occurrence of each species.. 721 Relative importance of each species and of each district 723 1 Bulletin No. 27. Approved for publication May 28, 1938. Page The purse-seine fishery. By George B. Kelez 725 Drag seines 725 Development of the purse seine 726 Early seines 726 Scow seines 726 Development of the modern purse-seine vessel 728 Introduction of power 728 Improvements in vessel de- sign 729 Increase in vessel size 730 Evaluation of fishing intensity 730 Seasonal fluctuations in fleet size. 730 Factors affecting seasonal in- tensity 730 Size of summer and fall fleets on Puget Sound 731 Size of cape seine fleet 733 Changes in composition of the fleet 734 Relation of vessel size to effi- ciency 736 Seasonal occurrence of each species. . 740 Puget Sound fishery 740 Cape fishery 741 Fishing seasons in different districts. _ 742 Puget Sound 742 Cape Flattery 744 Relation of fishing intensity to sea- sonal occurrence 745 Relative importance of each species.. 747 Puget Sound 747 Cape Flattery 748 The troll fishery. By George B. Kelez 749 Development of the fishery 749 Importance 750 Seasonal occurrence of cohos and kings 751 Sport fishing 753 Sockeye Salmon. By George A. Rounse- fell 754 693 694 BULLETIN OF THE BUREAU OF FISHERIES Sockeye Salmon — Continued. Page Introduction 754 General life history 754 Spawning 754 Age at maturity 755 Sockeye rivers of the region 756 Outer coast streams 756 Puget Sound streams 757 Gulf of Georgia streams 757 Migration in salt water 758 Total pack of the Fraser River sys- tem 758 Method and locality of capture 759 Indian fishing in the Fraser 759 Extent of the Indian fishery 760 Catch by commercial gear 760 Locality of trap catches 762 Locality of purse-seine catches.. 762 Changes in abundance of different portions of the run 763 Changes in abundance 765 Average catch per unit of effort with gill nets 766 Index of abundance from traps... 767 Purse seines 769 Combined index of abundance.. 772 Explanation of changes in abun- dance 772 Abundance of cycle ending in 1934 774 Abundance of cycle ending in 1933 775 Abundance of cycle ending in 1932 777 Abundance of cycle ending in 1931 778 Coho Salmon. By George B. Kelez 781 Introduction 781 Life history 781 Spawning 781 Growth 782 Age at maturity 783 Individuality of populations 783 Locality of capture by different types of gear 784 Catches in various districts 784 Locality of trap catches 785 Seasonal occurrence in various areas. 786 Page Coho Salmon. Changes in abundance 789 Calculation of trap indices 789 Calculation of purse-seine index. 792 Trends of abundance 794 King Salmon. By George B. Kelez 795 Introduction 795 Life history 795 Locality of capture by different types of gear 797 Catches in various districts 797 Locality of trap catches 798 Seasonal occurrence in various areas. 799 Seasonal occurrence of red and white king salmon 800 Changes in abundance 803 Pink Salmon. By George A. Rounsefell.. 804 General life history 804 Migration 805 Method and locality of capture 805 Seasonal occurrence in northern and southern districts 808 Changes in abundance between early and late years 808 Indices of abundance from traps 811 Abundance from purse-seine catches. 812 Comparison of purse-seine and trap indices 812 Chum Salmon. By George A. Rounsefell. 813 General life history 813 Method and locality of capture 814 Seasonal occurrence in northern and southern districts 814 Abundance from Admiralty Inlet traps 815 Abundance from purse seines 815 Summary. By George A. Rounsefell and George B. Kelez 817 The gill-net fishery 817 The trap fishery 817 The purse-seine fishery 818 The troll fishery 818 Sockeye salmon 819 Coho salmon 819 King salmon 819 Pink salmon 820 Chum salmon 820 Bibliography 820 INTRODUCTION By George A. Rounsefell and George B. Kelez The decrease in abundance of sockeye salmon in the waters of Swiftsure Bank, Puget Sound, and the Gulf of Georgia has been readily apparent, but no previous attempt has been made to measure accurately this change, nor has the decline of other species been previously demonstrated. The studies included in this report on the seasonal occurrence of each species, and the history and development of each form of gear, were necessary in arriving at logical conclusions as to the causes and extent of the changes in abundance that have occurred. The interrelations of the various species of salmon and the different types of gear in this region are such that the problem cannot be understood unless all of these factors are considered. Not since the general report in 1899, entitled “A Review of the Fisheries in the Contiguous SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 695 Waters of the State of Washington and British Columbia,” by Richard Rathbun, has this region been considered as an entity. The region is of considerable extent, including that portion of the high seas in the vicinity of Swiftsure Bank, the Strait of Juan de Fuca, and the narrow inland sea, over 200 miles in length, formed by Puget Sound and the Gulf of Georgia (see fig. 1). Of the numerous tributary streams, only the Fraser River penetrates the Coast Range into the interior. Many shorter rivers, however, such as the Skagit, Snohomish, and Squamish on the mainland, and the Cowichan and Nanaimo Rivers on Vancouver Island, together with a host of smaller streams, also furnish spawning grounds for the salmon of these waters. THE PACIFIC SALMONS The Pacific salmons (genus Oncorhynchus ) inhabiting this region, like the At- lantic salmon ( Salmo solar ) and the steelhead trout ( Salmo gairdneri), spend varying lengths of time in fresh water after hatching, before descending to the sea where most of their growth is attained. They differ from the Atlantic salmon and the steelhead in that all of the adults, upon returning to fresh water, die shortly after spawning. The adult salmon, returning from the ocean to spawn in the streams from whence they came, form the object of intensive fisheries on Swiftsure Bank, among the inlets and islands of Puget Sound, the Gulf of Georgia, and in the estuary and lower reaches of the Fraser River. This region has five species of Pacific salmon: The sockeye ( Oncorhynchus nerlca), known as the red salmon in Alaska and as the blueback on the Skagit, Quinault, and Columbia Rivers; the coho or silver salmon ( 0 . kisutch), also known as the silverside; the king or spring salmon ( 0 . tschawytscha) , known as the chinook on the Columbia River and the quinnat on the Sacramento River; the pink or humpback salmon ( 0 . gorbuscha) ; and the chum or dog salmon ( 0 . keta), also called keta or fall salmon. In addition to the confusing array of names given above, the immature king salmon are often called blackmouth, a term which is also sometimes applied to immature cohos. In the Gulf of Georgia the immature cohos taken early in their third summer are termed bluebacks. In size the pinks are the smallest, averaging around 4 pounds. The sockeyes average under 6 pounds, the cohos about 7-8 pounds, and the chums about 9 pounds. The kings are by far the largest, averaging about 22 pounds, with occasional indi- viduals of 60 pounds and upwards. The pink salmon are unique in that they appear in abundance over the greater part of this region during the odd-numbered years, whereas only a few thousand are taken in the even-numbered years. FISHING DISTRICTS The region may be roughly divided into fishing districts, not only geographically, but also in accordance with the types of gear used and the abundance of the various species. Swiftsure Bank is unique in that the vast majority of the cohos and kings caught by trolling are taken there. Here the purse seiners meet the incoming schools of pinks, cohos, and sockeyes that are bound for the Strait of Juan de Fuca, and 696 BULLETIN OF THE BUREAU OF FISHERIES Figure 1. — General map of the region. — ~ - -I »-r„30/n« . UAH3MI jJ3t> .mm suiijgft #At Miwo'i': .aoeBW j0lo? o) ivrIB’z-'vt I •»(<; a Jnoi • uJ: 71941—38 (Face p.697) SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 697 thence to their spawning grounds in the myriad streams of the region. Here also, during the early summer, immature cohos and kings, actively feeding on this ocean bank, are taken in large quantities. The waters inside of the strait, our so-called “inland sea,” also fall into natural categories. The waters of Puget Sound east of Whidbey Island (see fig. 2), and south of Point Wilson (see fig. 3), are traversed almost entirely by salmon bound for local streams; the dominant species being the coho, chum, and pink. The only sockeyes taken are a few headed for the Skagit River. Traps, purse seines, and gill nets are employed. The remainder of Puget Sound, north of Point Wilson and west of Whidbey Island, is often spoken of as the “outside” waters. In this district, which should include also the southern tip of Vancouver Island, the sockeye and pink salmon greatly outnumber the other species in the catches. The trap and purse seine are both employed to advantage and a few gill nets are used in Bellingham and Boundary Bays. The last district is the Fraser River itself, from Mission Bridge to the mouth, and the adjoining waters of the Gulf of Georgia. Here the sockeye is the paramount species, although pinks are taken in abundance and fair catches of kings, cohos, and chums are made. The only gear permitted is the drift gill net, except late in the fall when portions of the district are opened to purse seining. The remainder of the Gulf of Georgia is fished by purse seines for cohos, chums, and pinks. A few sockeyes are taken near Quathiaski. DEVELOPMENT OF THE FISHERIES Exploitation of the salmon fisheries on a commercial scale began with the build- ing of the first sockeye cannery at New Westminster in 1866 (see fig. 2). Since sockeye were plentiful and the fishing, conducted with gill nets, was easy, the indus- try flourished (see table 1). Some changes have occurred in the gear, the skiffs used at first were replaced by roundbottomed boats in the 1890’s, and engines were in- stalled in practically all of the gill-net boats between 1911 and 1913. Since 1914 the gear has not undergone any significant changes in this Fraser River district. The second of the aforementioned districts to be commercially exploited was the inside waters of Puget Sound. Here the first cannery was built at Mukilteo (see fig. 3) in 1877, followed soon by canneries at Seattle and Tacoma. In these waters the early forms of gear were the gill net, set net, drag seine, and a primitive type of purse seine. Traps were used near Seattle as early as 1885-87, but were not successful in this portion of the district until about 1899, although east of Whidbey Island they were successful by the early 1890’s. In later years the gill nets, set nets, and drag seines became of minor importance, while the power-driven purse seiners became a major factor in the fishery. The northern or “outside” waters of Puget Sound were lightly fished until the erection of the first cannery in this district at Semiahmoo in 1891 (see fig. 2.) Can- neries were built at Point Roberts (see fig. 7) in 1893 and at Friday Harbor in 1894. By 1900, 15 canneries were operating in the district, out of a total of 19 in Puget Sound (see table 1). The sudden expansion of the fishery here was due to the success- 698 BULLETIN OF THE BUREAU OF FISHERIES Figure 3.— Map of Puget Sound from Point Wilson to Olympia, showing the fishing areas. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 699 ful use of traps in the capture of sockeye Purse seines did not become of great importance in this district until 1907 when power-driven vessels had come into general use. In the Gulf of Georgia the fishery developed slowly, except for the area near the mouth of the Fraser River. The first cannery in this district was built at Quathiaski in 1904 and canned chiefly cohos, caught by troll in the northern end of the Gulf of Georgia, as well as small quantities of sock- eye. Later pinks and chums were also utilized. Except for a small cannery 5 at Pender Harbor in 1906 g and 1907, this was the only cannery in this district for several years. « Swiftsure Bank was the 5 last district to be exploited, z as the development of this fishery in the open ocean depended upon the in- creased mobility of power- , driven vessels. About 1908 trolling vessels were „ fishing in the Strait of §, Juan de Fuca as far as the £ open sea, and by 1912 the greater part of the fleet was fishing at the cape. Purse- z seine vessels also began ^ to fish here by 1911 and, since 1912, a fair share of the fleet has spent a por- tion of the summer there. PRODUCTION AND VALUE Because of variations in economic conditions, and in the abundance of the various species, it is difficult to appraise the value of these fisheries. During the 8-year period, 1927-34, the average annual production was 113,450,000 pounds of raw salmon which had a wholesale market value of $10,400,000. If the 2 worst depression years, 1931 and 1932, are omitted, the averages are raised to 116,660,000 pounds and $11,720,000 (see fig. 4). However, this region is capable of producing a great deal more wealth than it does at present. By way of illustration one need only refer to the reduced catches of sockeye. From 1898-1913, a 16-year period when the sockeye fishery was flourishing, the average pack of sockeye was 790,000 cases per year, worth on the average $4,930,000 (average price of just over $6.00 per case). During the 8-year period, 1927-34, the sockeye 4. — Salmon production of the region in pounds of raw salmon, and wholesale value of the products. 700 BULLETIN OF THE BUREAU OF FISHERIES pack has averaged 229,147 cases, valued at $3, ISO, 000 per year (average price just under $14.00 per case). At present prices the former sockeye pack would be worth $10,960,000 per year — as much as the present fishery for all five species combined — and yet the present sockeye catch only averages about 15,000,000 pounds, or 13 per- cent of a regional total of 1 13,000,000 pounds. NEED FOR INVESTIGATION Although the entire region should be considered in general as a biological unit, the fact that the salmon are taken on the high seas, and in both Canadian and Ameri- can waters, has caused each governmental agency to keep only records of the catches landed under their own jurisdiction. Furthermore, during the period covered by this report, these agencies have usually collected onty such records as have been necessary for purposes of taxation or general production statistics. Hence, only a few of the existing catch records were of any biological value. In order to determine such relative factors as the seasonal progression of the runs, or changes in abundance of the various species, it was imperative that catch data be obtained which included the daily landings of individual units of fishing gear. Many valuable records of this type still exist in private hands, although, with the passage of time, a large part of various individual company records have been destroyed or lost when certain companies changed ownership or ceased opera- tion. Accordingly, the authors gathered a vast quantity of these records from both American and Canadian companies which, together with total catch records from the publications of various agencies, have been analyzed in this report. Such analyses were complicated by the many changes which have occurred during the long period of development of these fisheries. Not only were new fishing areas pioneered, and new types or radical improvements of the old forms of gear developed, but there has been a considerable shift in intensity of the fisheries for some of the species, both for economic reasons and because of changes in abundance. Because these changes directly influenced the exploitation of the resource, the his- tory and development of the major forms of gear have been carefully traced. Differ- ences in fishing locality, seasonal operation, and effectiveness in the capture of the various species of salmon have necessitated the separate consideration of each of the more important forms of gear. The different species of salmon enter the fishery in varying abundance at certain parts of the season, hence it has been necessary to determine the curves of seasonal occurrence for each species. The changes in abundance that have occurred during the course of the fishery have in the past been measured largely from the total annual production of canned fish, a measure which is especially inaccurate in view of the influence of changing economic conditions, changes in fishing effort, and the obscuring of the decline in certain species by the increase in intensity of the fishery for others. The authors have endeavored to present, for each species, the best measure of abundance possible from the available data. The varying importance of the species in certain districts and in different types of gear, and the differences in production of the major spawning areas have also been treated. The complexity of these problems and the differences in their life histories have made it necessary to consider them, like the major types of gear, in separate sections of the report. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 701 It has been the desire of the authors not only to make the above material avail- able, but to present it in such a way as to provide a thorough understanding of the fisheries of the region and to establish a background which will form the basis for future conservation efforts in the region. ACKNOWLEDGMENTS The authors wish to express their appreciation for the splendid cooperation in the furnishing of information and statistics by the following companies: Anglo- British Columbia Packing Co.; The British Columbia Packers; The Canadian Fish- ing Co.; Francis Millerd; Greatwest Packing Co.; J. H. Todd & Sons; Johnston Fishing & Packing Co.; Kingcome Packers; Nelson Fisheries; Quathiaski Canning Co.; Queen Charlotte Fisheries; Sooke Harbour Fishing & Packing Co.; Alaska Packers Association; American Packing Co.; Anacortes Canning Co.; Astoria & Puget Sound Canning Co.; Beach Packing Co.; Bellingham Canning Co.; Booth Fisheries Corporation; Carlisle Packing Co. (S. P. Kelly) ; Everett Fish Co.; Far- west Fisheries; Fidalgo Island Packing Co.; Fishermen’s Packing Corporation; Friday Harbor Canning Co.; W. A. Lowman; New England Fish Co.; Northwestern Fisheries Co.; Pacific American Fisheries; Puget Fisheries; San Juan Fishing & Packing Co.; Sebastian-Stuart Fish Co.; Icy Straits Packing Co.; Western Fisheries; Western Sea Foods Co. For valuable information and statistics of early fishing on the Fraser River the authors are indebted to Mr. Henry Doyle, of Vancouver. Capt. T. E. Eggers, of Seattle, supplied information of the early fishing on Puget Sound. The officials of the Fisheries Departments of the Dominion of Canada, the Province of British Columbia, and the State of Washington have extended numerous courtesies, in addition to giving the authors access to their files and records. GILL NET FISHERY By George A. Rounsefell FRASER RIVER EARLY COMMERCIAL DEVELOPMENT Gill nets were the first to be developed of the four main types of gear used com- mercially in this region. Since 1873 they have captured 46 percent of all of the sockeyes taken, as well as large quantities of the other species. The gill net fishery is so inextricably bound up with the Fraser River that its story is largely that of the Fraser itself. The salting of salmon was begun soon after 1800 by the Northwest Company, later the Hudson Bay Company (Rathbun 1899), which exercised a monopoly of the fishing (Howay 1914), and by 1835 was shipping 3 to 4 thousand barrels of salt salmon each year to the Hawaiian Islands. These early trading companies depended very largely upon salmon for their food supply. Thus, in 1836, the supplies gathered for the upper Fraser River trading posts included 67,510 salmon, 11,941 smaller fishes, 781 sturgeon, and 346 trout (Morice 1904). In 1858 the Hudson Bay Com- pany’s license was revoked and its claim of monopoly fell. 702 BULLETIN OF THE BUREAU OF FISHERIES The first salmon were canned on the Fraser River in 1863, when Mr. Annandale canned a limited quantity for local use (Doyle 1920). This pre-dates by 1 year the establishment of the first salmon cannery on the Pacific coast by Hapgood, Hume & Company, in 1864, on the Sacramento River. The first real cannery on the Fraser River was built in 1866 at New Westminster. The first cannery on the Columbia River was built the same year at Eagle Cliff. Thus, salmon canning on the Pacific coast started almost simultaneously on three of the largest salmon streams. The first recorded pack on the Fraser River, in 1873 (Rathbun 1899), was 8,125 cases. Howay (1914), mentions the unsuccessful use of Scotch trap nets in 1864 by the Annandale saltery, and the change to drift gill nets. The gill netting during the earlier years was done by Indian fishermen from canoes and flat-bottomed skiffs. The packs were restricted because of the crudeness and inefficiency of the canning equipment, and because the necessary tinplate had to be shipped around Cape Horn in sailing vessels in advance of the season. Thus, in 1882, because of an unexpectedly large run of salmon, the supply of tinplate became exhausted in the middle of the season and the packers were forced to close down. RELATIVE IMPORTANCE OF THE VARIOUS SPECIES In the early development of the Fraser River fishery the sockeye was by far the most important species. The deep color and firmness of its flesh was most important for producing an attractive product with the crude canning methods then in use. Also, sockeyes were tremendously abundant, the run reaching its peak during the summer months when fishing conditions were at their best. So important were they to the canning industry that, for the period before 1900, when accurate records of the number of cases of each species canned were not always available, the total canned pack has often been used to represent the sockeye pack. In seasons when sockeye were not abundant the canners would often, even during the earlier years, supplement their pack with coho and king salmon. However, when the packers were unable to handle all of the sockeye that the fishermen delivered they could not afford to waste time, effort, or their sometimes inadequate supply of tin- plate, to put up a cheaper product. Thus, 1905 was the first of the “big” years of the quadrennial sockeye run to the Fraser River in which as many as 30,000 cases were canned of the other four species combined. Meanwhile the fishery for king salmon began to attain importance after freezers were built on the Fraser River. The first of these appeared in 1886 and two others in 1887. In early years the canning of king salmon usually began before the sockeye runs made their appearance. Thus, one cannery, in the period from 1887-91, usually started canning king salmon during the latter half of April, more than 2 months before the sockeyes were due to appear. Gradually they commenced operations later in the season until, from 1900-1902, they did not start until after the sockeyes had arrived. There was much variation between individual canneries, however, as to their season of operation. Since the 1880’s a few canneries have remained open, after poor sockeye runs, for the fall fishing. For many years this fishing was confined largely to cohos, and the fall run of king salmon, which are inferior to those running in the spring. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 703 The pink salmon were for a long time considered inferior in value for canning because of their light-colored, soft flesh. However, as the sockeyes became scarcer and a demand for cheaper grades of salmon increased, the pinks eventually became important. The first pack of any consequence on the Fraser River was in 1907 when 63.000 cases of pinks and chums were canned. In 1909, a big sockeye year, only 2.000 cases of pinks were canned, but in 1911, the next pink-salmon cycle, 142,000 cases were packed and the pink salmon had definitely become an important factor in the fishery. Table 1. — Number of canneries operated in the region Year Fraser River 1 Victoria and Gulf of Georgia ! Puget Sound and Neah Bay 3 Total Year Fraser River 1 Victoria and Gulf of Georgia 2 Puget Sound and Neah Bay 3 Total 1876 3 3 1906 23 4 17 44 1877 5 1 6 1907.. _ 18 3 13 34 1878... 8 1 9 1908... 10 2 11 23 1879 7 2 9 1909 34 3 23 60 1880 7 2 9 1910... 21 2 15 38 1881 8 3 11 1911 15 2 20 37 1882... 11 3 14 1912 15 2 20 37 1883 13 3 16 1913 35 4 31 70 188-1 6 3 9 1914 20 3 22 45 1885 6 3 9 1915 22 3 41 66 1886 ii 3 14 1916 21 5 32 58 1887 12 3 15 1917 29 5 47 81 1888 12 4 16 1918 18 5 37 60 1889... 15 2 17 1919 14 3 36 53 1890... 17 1 18 1920 11 3 11 25 1891... 22 2 24 1921 13 3 23 39 1892. _ 22 2 24 1922. 10 4 15 29 1893 26 3 29 1923 11 4 18 33 1894 28 4 32 1924 9 4 12 25 1895 33 6 39 1925 10 4 23 37 1896 35 u 46 1926 10 3 14 27 1897... 43 12 55 1927 10 3 21 34 1898 49 18 67 1928... 8 3 14 25 1899 41 17 58 1929 9 3 21 33 1900 45 19 64 1930 8 3 13 24 1901 49 16 65 1931 7 3 18 28 1902 42 20 62 1932 8 3 10 21 1903 36 19 65 1933. 10 3 19 32 1904 25 1 12 38 1934 11 3 21 35 1905 38 2 24 64 > Includes canneries in Vancouver and environs. 2 Extending north to and including Quathiaski. 3 Neah Bay is just inside of Cape Flattery. Number estimated from 1878 to 1887, inclusive, except for 1881, which is from Hitt ell (1882). Chum salmon were long regarded as a nuisance by the fisherman, although the Indians always used them to some extent, especially in years of poor sockeye runs. In 1897 the Japanese commenced drysalting chum salmon on the Fraser River for the Japanese market, and for use in the Yukon for dog feed. The Report of the Department of Marine and Fisheries for 1899 (1900) says: A new feature in the fishing industry this season was the salting for shipment to Japan of 4,000,000 pounds of dog salmon (O. keta) by Japanese fishermen. The fish were mostly caught by fishermen when fishing for cohos for the canners, and bought by the Japanese. Formerly this class of fish, when caught, were allowed to go to waste. In 1900 the canners commenced using chum salmon. The sockeye run was very small and a good price was being offered for lower grades of salmon, so 105,000 cases were canned. Difficulty was experienced in marketing, however, on account 704 BULLETIN OF THE BUREAU OF FISHERIES of a large production in other areas, and the chum-salmon pack remained small until 1910, when 52,000 cases were packed. The pack did not again exceed 100,000 cases until 1923. NUMBER OF CANNERIES Judging from the number of canneries in operation on the Fraser River or near its mouth each season since 1876 (see table 1), exploitation of salmon increased almost continuously from 1876-98. The great majority of the canneries were built during this 23-year period and the peak was reached when nine new canneries were built in 1897. The decline in the number of canneries in 1884 was possibly due to unfavorable economic conditions at that time. The Annual Report of the Department of Fisheries for 1884 says: There is estimated to be over in Great Britain now — 1st January, 1885 — in an unsalable con- dition, . . . , over two hundred thousand (200,000) cases of fall salmon, that will not bring much more than freight, insurance and charges. In 1901, the large packs both on the Fraser River and in Puget Sound again brought about an oversupply of salmon. The British Columbia Packers Association, which was formed at this time, included 29 of the 49 canneries on the river. The number of canneries in operation was considerably curtailed through this and other combines, especially during the “off” years when a few canneries were sufficient to handle all the catch. During the war years the number of canneries increased some- what, but at the end of the war it dropped sharply, and there have been less than a dozen since 1921. EVALUATION OF FISHING INTENSITY Company Licensing System In the early years of the fishery the majority of the fishing licenses were taken out by the canneries, who then hired men to fish them on whatever arrangement the company wished to make. At first they usually hired men to fish by the day or month, but later this custom was largely supplanted by the share system in which a certain percentage of the price of the fish, usually one-third, was deducted by the company, which supplied the net and rented a boat for a nominal charge. The independent fisherman was required to fish under his own license. The canneries often hired 2 gangs (2 men in each gang) for each of their boats. Thus, by working in shifts, the license and boat might be used day and night. For instance, Hittell (1882) says of the cannery of Laidlaw and Co. in 1881, “It has 25 boats, which run day and night, with 4 men to each boat.” Of a total of 1,174 gill-net licenses issued in 1893 the companies obtained 909, varying from 27 to 40 licenses per company. Apparently the companies were re- stricted as to the total number of licenses they might have for 1 company had 27, 7 had 30, 4 had 35, 7 had 36, and 7 had 40. In 1894 the number of company licenses was reduced by law to a maximum of 20 each for canneries, and 7 for dealers in fresh, frozen, salted, cured or smoked salmon. By 1898 this limit was further reduced to 10, and after 1907 company licenses were abolished. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 705 Nationality of the Fishermen Because of differences in fishing ability it lias been important to a study of the gill netting to note the changes in the nationalities of the fishermen. According to Henry Doyle the fishermen were practically all Indians as late as 1882. The first Japanese fishermen were engaged by English and Company at their Steveston cannery in 1888. Only a few were employed at first, however, and up to 1892 they were not given independent licenses. Doyle estimated that they formed at least one-third of the fishermen by 1895. The statement by Doyle that in 1882 most of the fishermen, if not all of them, were Indians, is borne out by Hittell (1882) who says that the Delta Packing Com- pany in 1881 had 36 boats and employed 200 Chinese, 150 Indians, and 30 white men. The Chinese, of course, were used as cannery labor, the white men were prob- ably nearly all clerks and mechanics, and the 150 Indians would be about the number required to furnish 2 crews of fishermen (4 men) to each of the 36 boats. From 1900 to date the license registers for indi- vidual fishermen have been available at the New West- minster office of the Do- minion Fisheries Depart- ment. Since 1915 these registers have given the nationality of each fisher- man. For previous years we have divided them into three groups: Japanese, Indian, and white, being guided both by the name and residence of each fisherman. Figtjke 5.— Fraser River gill nets, showing for each year the total number of gill-net licenses issued, the number issued to Japanese fishermen, and the total units of fishing effort. For an explanation of units of fishing effort see text. Number of Licenses The number of licenses issued to each of these three groups of fishermen, plus company licenses — which we have not attempted to segregate before 1900 — and special licenses issued since 1908 permitting bona fide residents along the banks of the Fraser River between the New Westminster and Mission bridges to fish only in that area are given in table 2. The figures for the Fraser River, except the totals, for years previous to 1900 were empirically determined from available information. Units of Fishing Effort Having made an estimate of the number of each type of fishermen, it has been necessary, in order to obtain the best measure of the intensity of the gill-net fishery 706 BULLETIN OF THE BUREAU OF FISHERIES during each season, to determine the relative efficiency of each type. For one com- pany we have records from 1905-16, inclusive, giving the catches of their individual fishermen. During this 12-year period the average annual catch of their Japanese fishermen was 1,782 sockeyes, their white fishermen 1,057 sockeyes, and their Indian fishermen 768 sockeyes (see table 3). Table 2. — Gill net licenses of the Puget Sound-Fraser River region, 1877-1934 Year Fraser River Puget Sound Grand total Type of license 1 Total Type of gill net Total Com- pany Individual Between- bridges license Drift Set Japanese Indian White 1877 285 285 1878 449 449 1879__ ... 304 304 1880._ - 274 274 1881_ 396 396 1882 666 666 1883 __ 715 49 764 1884 645 57 702 1885 611 44 655 1886 625 109 734 1887 615 320 935 1888_ 10 323 167 500 1889 25 308 167 500 1890 25 308 167 500 1891 50 283 167 500 1892 108 373 240 721 1883 235 558 381 1, 174 1894 417 549 701 1,667 1895 * 434 539 731 30 1,734 1896 926 530 1, 130 60 2, 646 1897 928 520 780 90 2,318 422 668 1,090 3, 408 1898 1,321 511 690 120 2, 642 281 460 ' 741 3, 383 1899 1, 361 501 710 150 2, 722 322 344 666 3, 388 1900. 393 1, 659 555 1,076 3, 683 380 330 710 4, 393 1901 416 1, S05 396 '909 3,526 414 369 783 4, 309 1902 381 929 583 781 2, 674 353 361 714 3, 383 1903 343 1,416 477 860 3, 096 334 470 804 3; 900 1904 232 795 446 742 2, 215 438 540 978 3, 193 1905 339 1,056 464 915 2, 774 348 574 922 3, 696 1906 200 494 392 660 1.746 310 618 928 2, 674 1907 193 769 270 494 1,726 329 755 1,084 2,810 1908. 3 717 175 273 195 1,363 362 836 1, 198 2, 561 1909 1, 263 584 638 243 2, 728 366 686 1,052 3,780 1910.. 766 236 426 148 1, 576 403 666 1, 069 2, 645 1911 607 232 411 146 1,396 459 813 1,272 2,668 1912 655 217 486 72 1. 430 377 829 1,206 2,636 1913.... 1, 132 476 843 109 2,560 427 807 1,234 3, 794 1914... 1,250 333 842 231 2, 656 544 458 1,002 3, 658 1915 1,332 317 768 199 2,616 512 559 1,071 3, 687 1916- 1,435 211 437 157 2,240 449 541 990 3,230 1917 1, 520 300 570 237 2, 627 537 658 1, 195 3, 822 1918 l’ 025 106 303 149 1, 583 417 646 1, 063 2, 646 1919 874 56 294 113 1, 337 540 686 1, 226 2, 563 1920 875 36 275 102 1,288 364 439 803 2, 091 1921 857 68 359 153 1,437 346 318 664 2, 101 1922 871 32 277 116 1, 296 119 37 156 1,452 1923 523 26 304 111 964 136 14 150 1, 114 1924 523 40 289 117 969 181 10 191 1, 160 1925 444 36 357 132 969 391 17 408 1,377 1926 444 53 429 137 1, 063 361 11 372 1,435 1927 400 68 619 172 1,249 397 18 415 1,664 1928 400 57 695 151 1,303 353 22 375 1,678 1929 400 73 830 170 1,473 368 23 391 1,864 1930... 400 60 863 200 1, 523 398 20 418 1, 941 1931 400 35 739 184 1, 358 319 19 338 1,696 1932 400 26 840 180 1, 446 254 8 262 1,708 1933 400 25 1,026 234 1, 685 302 9 311 1,998 1934. 400 31 1, 105 267 1, 803 318 12 330 2,133 1 From 1877 to 1899 the nationalities have been estimated from various notes. The company licenses before 1900 are not separated from the total, and so are allocated amongst the other types. There were no special “between bridges” licenses prior to 1908, so the figures from 1895 to 1899 merely represent a rough estimate of the number of this type of resident up-river fishermen before 1900. From 1900-1907, inclusive, no estimate of these fishermen was made as it was impossible to segregate the nationalities accurately. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 707 Table 3. — Annual catches of sockeyes by white, Indian, and J apanese fishermen at a Steveslon cannery, 1905-16, inclusive Year 1905.. 1906. . 1907.. 1908 «. 1909.. 1910.. 1911.. 1912.. 1913 1914.. 1915.. 1916.. Total... Unweighted average. Japanese Whites Indians Cannery license > Number 114 46 132 132 122 94 69 62 / 85 l 21 138 141 168 Average 4, 064 1,537 425 788 2,393 1,270 824 1,283 3, 588 3, 546 1,053 435 178 Number 72 77 46 42 34 28 58 56 62 14 92 106 30 Average 2,872 860 234 645 1, 437 852 328 611 1,832 2, 365 517 164 63 Number 29 60 19 27 31 10 11 15 13 10 29 27 20 Average 2,414 550 183 370 1, 102 527 412 660 1,204 1,142 476 122 53 Number 9 8 9 Average 3,154 717 249 1, 324 21, 384 717 12, 680 291 9,215 110.3 1,782 59.3 1,057 24. 25 768 i From 1904-7, inclusive, out of 40 company licencees, 38 were white, 2 Japanese during the summer fishery, and a few Indians were employed for fall fishing. * Includes a very few cohos and some kings. 5 Two canneries. From the averages shown in table 3, and the variations in the number of each type of fishermen, it is obvious that in order to obtain a true picture of the intensity of fishing the total number of licenses must be broken into component groups and each group weighted according to an estimate of its efficiency. This has been done by as- signing to Indian licenses and “between bridges” licenses a weight of 1.00, to white and company licenses a weight of 1.375, and to Japanese licenses a weight of 2.32. From 1900-1907, inclusive, we have estimated that 150 of the fishermen not falling into other classifications, grouped as whites in table 2, were up-river resident fisher- men of the same type that later used the special between bridges license. These are given the same efficiency weighting as the Indian licenses. The total units of effort for each year, estimated on the above basis, have been used in the sockeye section of this report to determine the average annual catch per unit of fishing effort on the Fraser River (see fig. 5 and table 33). Changes in Gill-net Boats In addition to differences in the efficiency of each license holder, according to his nationality, there have been changes in the form of the unit of gear itself. The first of these to be considered is the change in type of boat used. According to Greenwood (1917) the fishermen still used a two-oared skiff in 1896, 20 years after salmon canning began. Rathbun (1899, p. 307) says: The boats are mostly small skiffs, about 20 feet long, generally manned by two, occasionally by three persons. In recent years the Columbia River boat has been introduced and is now used to a considerable extent in the lower part of the river and outside. Its breadth and centerboard make it much safer for the more exposed places. Greenwood also says the round-bottomed 30-foot sail boats were introduced “a score of years ago”, when 20 were built for the Alliance cannery. This would place their introduction about 1897. However, Rathbun establishes their introduction in the early 1890’s. 71941—38 2 708 BULLETIN OF THE BUREAU OF FISHERIES In 1903 the records for one cannery show that their 25 white fishermen all used round-bottomed boats while their 66 Indian fishermen used 36 round-bottomed boats and 30 skiffs. Since the Japanese all fished on contract no record was kept of their gear, but it is safe to assume that all of their boats were round-bottomed, as they were very progressive fishermen. Among 3,096 licenses issued in 1903 only 477 were for Indians,2 and it is therefore evident that the transition from skiffs to Columbia River boats was almost complete. After 1905 the records of this company show no skiffs in use. The introduction of motorpower in gill-net boats, to replace oars and sails, took place soon after the turn of the century. According to old-timers on the river, gasoline engines were used as early as 1902, although only a few were in use until a decade later. Thus records of one of the largest canneries on the river, located at Steveston, show very few gasoline boats in 1909 and 1910. From then on, however, the number increased rapidly and large numbers of engines were installed in 1911-13. By 1914 the change appears to have been almost complete. The data have been insufficient to measure the increase in efficiency brought about by the adoption of engines, but such an increase existed and should be remembered when comparing the catches of the earlier years with those made during and after the World War. Changes in the Gill Net The gill-net fishery on the Fraser River is remarkable for the few changes that have taken place in the net itself over a long period of years. There has been no change of any consequence in the length of the net, and the deep nets, used for only a few years, were confined to a small percentage of the fishermen. In 1882, when the Richmond cannery was built on the North Arm, the nets used in that section of the river were 27 and 30 meshes in depth, 150 fathoms in length, and of 5%-inch mesh, according to Charles F. Todd. The Government regulations that went into effect May 1, 1894, provided for a maximum length of 150 fathoms. Rathbun (1899) says that although there was no restriction upon their depth, custom fixed it at 50 to 55 meshes, though some were shallower. In the years 1903 and 1905, the men fishing on shares for the Imperial cannery used a total of 8 nets of 40-mesh depth, 101 of 45 meshes, 37 of 50 meshes and 1 each of 55 and 60 meshes, placing the average at less than 50 meshes. The records for these years do not give any indication of the depth of the nets used by the Japanese, who formed over 40 percent of the fishermen on the river. Testimony as to the depth of gill nets is given in the Interim Report of the British Columbia Fisheries Commission (Report of the Fisheries Commission for B. C., 1906, pp. C18-C40), in which one witness, a canneryman, stated: This summer I had over 20 boats of Japanese fishing in the river, and there was not one of them with a net of less than 80 meshes. The same witness says later: It is only 8 or 10 years ago that the fishermen commenced to use these extra deep nets * * * I think it is only 4 or 5 years ago since 80-mesh nets were common. i This figure does not include Indians that may have fished on the 343 company licenses. Figure 8. — Brailing crew lifting the spiller of a salmon trap preparatory to brailing. In this operation one side of the spiller is lowered sufficiently to permit a small pot scow to enter the spiller. The side is then raised. Starting at one side of the spiller the crew overhauls the web until the salmon are crowded enough for brailing. Figure 9.— Brailing a salmon trap. The lower end of the heavy net, or brail, is attached to the side of the scow. The upper end is attached to a heavy pipe so that when the brail is lowered over the side of the spiller it sinks quickly. As soon as the brail sinks it is hauled under the densely schooled salmon by the men on the pot scow (in the background). The brail is then hoisted with a winch and the salmon are dumped into the large transporting scow. SALMON AND SALMON FISHERIES OF SWIFISURE BANK 709 From the foregoing it would appear that the depth of the gill nets commenced to increase somewhat after 1899, the last year for which Ratlibun gives any informa- tion. In 1906 our records for the Imperial cannery give 4 nets of 40 meshes in depth, 52 of 45, 42 of 50, 4 of 72, 4 of 75, and 3 of 80 meshes, so that out of 109 nets only 11 were over 50 meshes in depth. The 1906 records included both share and con- tract white fishermen, and unless the Japanese fishermen were using radically different gear, our records do not support the viewpoint of the witnesses as to the preponderance of deep nets. The British Columbia Fisheries Commission also stated: We favour the limitation of the length of salmon gill nets to 150 fathoms (300 yards). This was formerly the length of net universally used in the sockeye fishery, but for some years nets double the length, viz., 300 fathoms (600 yards) have been permitted outside the mouth of the Fraser River. To prevent all risk of abuse arising from the alleged use of long nets inside the Fraser River, a length of 150 fathoms is recommended as a maximum limit. Their statement is at variance with a statement by Inspector C. B. Sword in the Dominion Fisheries Report for 1904, p. 214, in which he says the canners suggest that a gill net longer than the prescribed 150 fathoms should be allowed in the Gulf of Georgia, as the shallower nets in use there would permit handling of 300 fathoms. That the longer nets were not used in the Gulf of Georgia is also the opinion of the cannerymen. From 1908-30 the size of gill nets in the whole area was restricted to a maximum length of 150 fathoms and a maximum depth of 60 meshes. Since 1930 a maximum length of 200 fathons has been permitted in the Gulf of Georgia. The size of the meshes in the sockeye nets were restricted as early as 18S2, and probably earlier, to a minimum of 5% inches. In 1916 the minimum size of mesh was lowered to 5% inches and in 1928 the minimum was abolished. Fishing Seasons In studying changes in fishing intensity one must know not only the relative effectiveness of the gear used in different years, but also the length of time during which it was employed and the proportion of the run that occurred during that period. On the Fraser River the closed seasons had little effect on sockeye fishing, especially during the earlier years. At one lower-river cannery the earliest sockeye canning date was July 5, 1887 and 1890, and the latest was August 30, 1888. The shortest season was 26 days in 1885, and the longest was 50 days in 1888; averaging 39 days. The closing date of August 25, effective in most years, had little influence on the pack. At another lower-river cannery, over the period 1887-1902, the sockeye pack was put up, on the average, in 52 days — from July 5 to August 25. The earliest start was made on June 27, 1896, and the latest on July 13, 1901. The season ended on August 12 in 1887 and on September 6 in 1902. The sockeye fishing seasons, as far as we have been able to determine from avail- able data, are given in table 4. 710 BULLETIN OF THE BUREAU OF FISHERIES Table 4. — Fraser River sockeye fishing regulations Fall season Opening Closing (?) (?) Sept. 25 Sept. 25 Sept. 25 Sept. 25 Sept. 25 Sept. 16 Sept. 16 Sept. 16 Oct. 31 Oct. 31 Oet. 31 Oct. 31 Oct. 31 (?) Oct. 7 Sept. 31 Sept. 16 Sept. 31 Sept. 16 Oct. 31 Oct. 1 Oct. 31 Year Closing summer season 1 Week end closed season General Between bridges 1 Remarks Before 1878. 1878-81 1882-87..., 1888 1889-92.... 1893 1894-1900. 1901 1902 1903 1904-07... 1908 1909 1910 1911 1912-14... 1915 1916 1917-20... 1921 1922 1923-24... 1925 1926-27... 1928 1929 1930 1931. 1932. 1933. 1934. (?) Aug. 25 Aug. 31 Aug. 25 Aug. 31 Aug. 25 Aug. 31 Sept. 6 Aug. 31 Aug. 25 Aug. 25 Aug. 25 Aug. 25 Sept. 31 Aug. 25 Sept. 31 Aug. 25 Sept. 31 Sept. 6 Sept. 22 Sept. 30 Nov. 21 Sept. 30 Sept. 30 Nov. 30 Sept. 20 Sept. 29 Sept. 30 Sept. 30 Sept. 15 Hours " 40" 40 40 40 40 36 36 36 36 36 42 42 42 42 42 42 3 42 42 42 42 42 42 42 48 48 48 Hours No regulations. No gill netting above tide water— must not obstruct over one-tbird of chan- nel. Nets 5£6-inch mesh, minimum. Nets 150 fathoms maximum length. Nets 60 meshes maximum depth. Nets 5%-inch mesh. Mesh limitation abolished. Nets 200 fathoms permitted outside of river. Purse seining Aug. 25-Sept. 30.4 Purse seining Sept. 1-8 and Oct. 1 -27. 1 Closing dates of summer season 1882 to 1903 partly from cannery pack records, opening date July 1 at least as early as 1894. 8 Fraser River between New Westminster and Mission bridges. • 54 hours weekly closed season during fall of 1916. 4 Purse seining in area 17, see map. Changes in Location of the Canneries At first the gill-net fishing was conducted inside the river, chiefly from New Westminster to Sumas and beyond, a distance of over 50 miles from the river mouth. At times the canneries received shipments of sockeye that were caught by the Indians with dip nets in Yale Canyon, near Hope, a distance of nearly 100 miles from the river mouth. The first canneries, as a consequence, were located at New Westminster. Meanwhile the fishermen had discovered that it was possible to make large catches in the lower river and the canneries found it advantageous to be closer to these fishing grounds. Consequently the first down-river cannery was built on Deas Island in 1876, followed by a second in 1878, and a third in 1880. In 1882 two more were built in this area, as well as one each at Steves ton and in the North Arm. The Indian fishermen did not have good boats for fishing outside the river, although they -went out at least as far as the sandheads. In 1885 we find the Do- minion Report suggesting that the distance between gill nets, while drifting over the sandheads outside the river, should be increased from 250 to 400 yards. That they did not, as yet, venture far from the river mouth is attested by the Dominion Report SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 711 for 1887 which states that the fishermen go out only as far as the lightship, 4 or 5 miles from land. Table 1 gives the number of canneries operated annually from 1876 to 1934. For nearly 20 years the proportion of the canneries located at New Westminster declined, while the proportion near Steveston and Ladner continued to rise. The few remaining canneries were either at the river mouth, in the North Arm, or entirely outside the river proper. The canneries at Ladner reached their peak in 1885, when half the total number operating were located there, and have since declined steadily to a point of little con- sequence. Many ascribe much of this decline to the fact that the fish have entered the river through Canoe Pass in decreasing numbers since the driving of traps at Point Roberts. The decline may possibly be further ascribed to the silting up of Canoe Pass and the change in the main channel at Woodwards Slough, effected dur- ing the flood of 1 894, which made it difficult to reach most of the canneries with large boats. SEASONAL OCCURRENCE OF EACH SPECIES Seasonal occurrence is of prime importance in any fishery wherein more than one species is taken, as the intensity of fishing for a species is not governed by its abundance alone, but by a combination of factors, such as the relative abundance of the several species at any time during the season, as well as the relative prices. In determining the seasonal occurrence for sockeyes, data for 1,982,735 fish taken in 30,706 gill-net deliveries were used, covering 3 complete 4-year cycles, 1898-1909, inclusive. The occurrence shown in these early years was considerably different than that shown in the last three cycles, 1923-34. This difference is treated in the sock- eye section of this report (see page 754). The king salmon curve is derived from 102,123 fish taken in 26,193 deliveries over a 5-year period, 1929-33. For pink salmon 8 years are represented, all of the odd-numbered years from 1915-33, except 1917 and 1921; the data totaling 597,774 fish in 15,581 deliveries. The coho curve is also based on 8 years’ data, 1904, 1905, and 1929-34, and repre- sent 155,957 fish in 22,117 deliveries. The chum-salmon curve represents only 3 years, 1932-34, but is quite repre- sentative of those particular years, comprising 263,703 fish from 10,608 deliveries. In analyzing these data the average catch per delivery for each 7-day period was computed for each year and then given equal weight in determining the average curve for all years (see table 5). Table 5 shows that the period over which one or more species can be taken in some measure of abundance extends from June 24 (week ending June 30) to November 17; 21 weeks, or 147 days. As mentioned above, in earlier years the season was very much shorter, corresponding largely to the more abundant portion of the sockeye run. The sockeye and pink-salmon runs, which overlap to a slight extent, are both of short duration. Approximately 79 percent of the pinks are caught in 4 weeks, September 2-29, and 83 percent of the sockeyes are taken in the 5 weeks from July 22-August 25. 712 BULLETIN OF THE BUREAU OF FISHERIES Table 5. — Seasonal occurrence in Fraser River gill nets Week ending — Percentage occurrence Sockeye King Pink Coho Chum June 30 __ 3. 73 July 7 5. 17 .083 0. 42 July 14. 4. 15 5. 21 0. 37 .92 .42 July 21 ' 5.68 6. 53 .46 .84 .42 July 28 9. 90 5.24 .43 .87 .42 Aug. 4 18. 24 5. 36 .98 .84 .42 Aug. 11. 24. 95 7. 45 1.42 .99 .42 Aug. 18 20. 21 7. 11 1.28 .87 .51 Aug. 25. 10. 12 8. 06 2. 66 1. 11 .47 Sept. 1 6. 75 10. 04 5.65 2. 65 .56 Sept. 8 7. 35 14. 73 5. 96 .58 Sept- 15 8. 40 31. 40 14. 64 .75 Sept. 22 5. 19 18. 25 12. 52 1. 18 3. 90 15. 84 14. 57 2. 86 Oct. 6 3. 77 3. 37 13.69 6. 69 Oct. 13 3. 38 2. 27 9. 43 11.68 Oct. 20 1.87 .54 6. 72 13. 01 Oct. 27 2.24 .37 7. 38 13. 37 Nov. 3. 2. 29 28. 46 Nov. 10 1.34 9. 58 Nov. 17 1.53 7.81 Number in sample... 1, 982, 735 102, 123 597, 774 155, 957 263, 703 Number of catches... 30, 706 26, 193 15, 581 22, 117 10, 608 Week ending — Cumulative percentage occurrence Sockeye King Pink Coho Chum June 30 3. 73 July 7 8. 90 0. 83 0. 42 July 14 4. 15 14. 11 0. 37 1. 75 .84 July 21 9. 83 20. 64 .83 2. 59 1. 26 July 28 19. 73 25. 88 1. 26 3. 46 1. 68 37. 97 31. 24 2.24 4. 30 2. 10 Aug. 11. 62. 92 38.69 3. 66 5. 29 2. 52 Aug. 18 83. 13 45. 80 4. 94 6. 16 3. 03 Aug. 25 93. 25 53. 86 7. 60 7. 27 3.50 100. 00 63. 90 13. 25 9. 92 4. 06 71. 25 27. 98 15. 88 4.64 Sept. 15 79. 65 59. 38 30. 52 5. 39 Sept. 22 84. 84 77. 63 43. 04 6.57 Sept. 29 88. 74 93. 47 57.61 9. 43 Oct. 6 92. 51 96. 84 71. 30 16.12 Oct. 13. 95. 89 99. 11 80. 73 27.78 Oct. 20 97. 76 99. 65 87.45 40.79 Oct. 27 100.00 100. 02 94. 83 54. 16 Nov. 3 97. 12 82. 62 Nov. 10 98. 46 92. 20 Nov. 17 99. 99 100. 01 The chum season is of almost as short a duration, 76 percent being taken in the 5 weeks from October 7-November 10. The coho season is somewhat more protracted, only 65 percent being taken in the 5-week period from September 9-October 13, and 7 weeks being required, September 9-October 27, to take 79 percent of the catch. The king salmon run rather steadily over a long period, 11 weeks, from July 1 -Sep- tember 15, being required to cover 76 percent of the run. Fifty percent of the sockeye catch has been made by about August 7 (see table 5). The pinks do not reach the 50 percent mark until about September 12, a difference of 36 days. This is followed about 2 weeks later by the cohos, which reach the 50- percent mark on September 26. Another month usually elapses before 50 percent of the chum run has passed. The king salmon run slowly but steadily and reach the halfway point about August 22. PUGET SOUND LOCALITIES FISHED Gill nets have been employed in Puget Sound since the earliest days of the fishery, but have never attained the importance that they have on the Fraser River. There are two reasons for this: First, in the clear waters of Puget Sound gill nets can be used only at night, as the fish avoid them in daylight; and second, it is difficult to compete with other forms of gear. The gill nets employed were of two kinds, drift and set, and, as their name implies, one was used adrift and the other anchored. They were Used chiefly in a few localities such as Skagit Bay and Skagit River, the estuary of the Snohomish River, and off the mouths of the Nooksack and Samish Rivers. A few were used in other localities, especially south of Point Wilson, among the San Juan Islands and in Boundary Bay. The addresses of the drift net licensees in 1899, from the State of Washington Fisheries Department files, showed that of 322 licenses issued, 154 were taken out in SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 713 areas south of Point Wilson, 78 from Seattle, 38 from Tacoma, 26 from Hood Canal, and 12 from scattered localities. More than one fourth, or 86, were from Skagit Bay and the Snohomish River. Of the remainder 1 was from Port Angeles, 5 from the San Juan Islands, and 76 from Bellingham and Boundary Bays. A second check was made, for the year 1901, of both drift and set gill nets, and it was found that out of 414 drift gill net licenses, only 63 were from Boundary Bay and the San Juan Islands. Out of 369 set net licenses 15 were from the San Juan Islands and none from Boundary Bay. It is evident that gill nets played a very minor role in the sockeye fishery in Puget Sound. The set nets were employed chiefly in river mouths, and especially in the Skagit, Snohomish, Duwamish, and Puyallup. A few were used away from the river mouths at such places as Open Bay on Henry Island, Andrews Bay on San Juan Island, and along the northwest shore of Orcas Island. There is some confusion as to the number of set nets operated, and as to their location during the earlier years. This is because a set net license was sometimes bought merely to hold a trap location during a year when it was not desired to drive the trap. The license fee for a trap was from 4 to 10 times as much as for a set net. No accurate estimate of the numbers of the different species taken by the gill-net fishery is available for early years, but the fishery was essentially the same then as today, except for the areas around Seattle and Tacoma, and the head of Puget Sound, where the salmon runs declined several years ago. RELATIVE IMPORTANCE OF VARIOUS SPECIES The set nets, fishing chiefly in the river mouths, caught mostly cohos and kings. In the 4 years from 1917-20, inclusive, they caught, on the average, 5.8 percent of the cohos and 3 percent of the kings taken in Puget Sound. They took but 1.3 percent of the chums and negligible quantities of pinks and sockeyes. After the formation of the Washington State Fishery Board in 1921, set nets ceased to be a factor in the fishing because of their subsequent strict seasonal regulation and their removal, by law, from the rivers. The drift gill nets, fishing in the more open waters, caught a greater variety of salmon than the set nets. During the 18-year period 1917-34, inclusive, they took, on the average, 12.1 percent of the kings, 8.9 percent of the cohos, 4.9 percent of the chums, 1.1 percent of the sockeyes, and 1 percent of the pink salmon caught in Puget Sound. TRAP FISHERY By George A. Rounsefell REEF NETS Reef nets, being the forerunners of the traps, will be considered first. They were used almost exclusively by the Indians, deriving their name from the kelp-covered reefs on which they were fished. Originally made from the fiber of cedar bark or roots, they were changed to cotton twine when it became available. According to Rathbun (1899) a reef net consisted of a piece of webbing, varying more or less in size, but averaging perhaps 36-40 feet long by 25-30 feet across, the mesh being about 3% inches. 714 BULLETIN OF THE BUREAU OF FISHERIES To fish a reef net a channel was cut through the kelp. The net was suspended between two canoes, anchored at both the sides and bows, with the forward end of the net sloping downward and the rear end curving back upward to the surface. In deep locations strands of rope were sometimes strung across in front of the net and below it, to lead the salmon closer to the surface. The nets were fished when the tide was running strongly, but a tide of over 5-6 knots per hour was considered too fast for fishing. Reef net crews often had two locations and fished them at different stages of the tide. A lookout was stationed in the bow of each canoe and when a school of salmon passed over the net they signaled for it to be lifted. The net crews immed- iately let go the side anchor lines and, since the bow anchors were placed close together, the canoes were swung toward each other by the current. At the same time the forward edge of the net was swiftly lifted, enclosing the salmon in a bag from which they were dumped into the canoes. Because of the manner in which these nets were operated, only a few localities were well suited to this type of fishing. One of the principal reef-netting grounds was off the southeastern point of Point Roberts, before that region was disturbed by the introduction of traps. Another excellent ground was along the western shore of Lummi Island, but the introduction of traps here diverted the salmon from these reefs. Other grounds, of lesser importance, were along the south shore of Lopez Island, the west shore of San Juan Island, the east and west shores of Stuart Island, and at Point Doughty on Orcas Island. The number of these nets in the earlier years of the fishery must have been con- siderable, as Rathbun says that 15 to 20 nets were formerly fished at Point Roberts, 16 operating there in 1889. By 1894 the string of traps had destroyed the advantage of this reef for nets. Wilcox (1898) lists 25 reef nets in Whatcom County and 14 in San Juan County in 1894. As late as 1901 there were 27 reef nets licensed, 15 to Lummi Island Indians and 12 to residents of the San Juan Islands. Because of the amount of labor involved, and the scarcity of favorable fishing locations, this gear was gradually supplanted, and only about a dozen have been used each year for the past 20 years. According to Rathbun the reef-net fishermen confined their attention almost exclusively to sockeyes, taking only a few king salmon. However, in late years they have taken more of the other species, especially pinks and cohos. A day’s catch has declined until, in recent years, it has rarely amounted to more than a few hundred salmon, but this decrease has been due largely to the fact that the more favorable locations have been rendered useless by traps. CONSTRUCTION OF THE TRAPS The trap fishery, which was abolished after 1934 in Puget Sound by the passage of an initiative measure in the State of Washington, was the second of the four main types of gear to attain prominence. From 1873-1934 they have taken 37 percent of the sockeyes caught in the region, as well as enormous numbers of the other species. Trap nets were tried at Point Roberts some years earlier than at other places, the first trap being built in 1880 by John Waller at Cannery Point, Rathbun (1899), (see fig. 7). Several years elapsed before the fishermen discovered the most desirable Figure 6. — Modern Fraser River gill-net boat. These round-bottomed motorized boats are very seaworthy, being relatively independent of the vagaries of the weather. They are more efficient than the skiffs in use during the earlier years, or the round- bottomed boats in use before motors were installed. Figure 7.— View from Cannery Point, on Point Roberts, showing the cannery established there in 1893. Note the 10 traps in the background. The trap nearest the cannery is approximately on the original trap location in Puget Sound, first established bv John Waller in 1880. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 715 locations for intercepting the salmon runs, and before they learned to build their traps sufficiently strong to withstand the storms that occasionally swept all exposed locations. The first traps consisted essentially of a barrier, or “lead” of webbing hung from a row of driven piling, which diverted the passing fish into a pen, or “crib,” similarly constructed. Although patterned after the pound nets of the Great Lakes, with a crib, heart, tunnel, and lead, they were built with much heavier piling which was usu- ally strengthened by having the pilings bound together with a capping of timbers, lashed on with cables. At first the heart was merely two rows of piling that formed a V with the lead pointed toward the bottom of the V. The fish followed the lead, which usually extended out from shore, until they found themselves between the lead and one of the outstretched arms of the heart. Continuing farther they swam through a narrow opening, or tunnel, into the crib. By 1895 the traps were much improved. The heart was often partially closed at its base, so that if the fish failed to enter the tunnel into the crib, they would, on circling back, find themselves in a semienclosure pointing toward the tunnel. A few traps had double hearts to minimize the chances of escape, and some had a leadlike extension, the forerunner of the “jigger” often employed on later traps. The jigger was essentially a supplementary lead consisting of a row of pilings connecting at about a right angle with the arm of either side of the heart, depending on the direction from which the fish usually approached the trap, and extending out toward deeper water, with the pilings driven to form a hook on the far end. The purpose of the jigger was to direct back to the lead such fish as passed the opening into the heart. The cribs in several traps measured by Rathbun were rectangular but not always square in shape, ranging from 35-80 feet on a side, and were driven in water from 3-9 fathoms in depth. The catches were sometimes much larger than could be han- dled by the canneries at once and, while a large catch might be held in the crib for several days, such accumulation prevented continuous fishing during a period when the salmon might be running best. To meet this contingencjq an adjunct to the crib, called a “spiller,” was devised and appeared to be coming into general use. It was, in fact, an additional crib, square in shape, and connected with the first by means of a tunnel, through which the surplus fish of any catch could be driven. The netting on the earlier traps was cotton twine, usually of 3-inch mesh in the crib and heart and from 3%-4 inches in the lead. Galvanized wire netting, in place of cotton, was experimentally used for the hearts and leads at Point Roberts in the late 1890’s, Rathbun (1899). The modern fish trap differs from the majority of those described by Rathbun in several respects. All of the trap, except the lead, is now customarily capped. If no capping is used the piles are tightly connected with a heavy wire cable to which the netting is attached to prevent sagging. All netting, except the spiller, is of gal- vanized wire which is cheaper and much more easily kept clean of seaweed and floating debris. All traps use a spiller of tarred cotton web. As a general rule the spiller is 40 feet square, and the pot is usually the same. If a trap fishes very well a second spiller is sometimes driven on the opposite side of the pot to take care of the surplus fish. 716 BULLETIN OF THE BUREAU OF FISHERIES A spiller is so placed that the fish, which enter the trap with the tide and then turn and swim against it, are led into the spiller through a narrow web tunnel which can easily be closed when the current is running in the opposite direction. Two spillers thus have a big advantage over one in that each one can be filled in turn, unless the trap is in an eddy where the current does not reverse itself with the tide. The pot aids in the fishing as the fish would not readily pass from as large a chamber as the heart directly through the narrow tunnel leading to the spiller, but the salmon are removed only from the spiller. The construction of the earlier traps was modified to some extent when certain regulations were put into effect. In 1897, the length of a trap lead was restricted to 2,500 feet, and it was further provided by law that there should be an end passageway of at least 600 feet, and a minimum lateral passageway of 2,400 feet, between all traps. These regulations had the effect of preventing a complete blockade of a whole area. For instance, in 1895 a string of three traps, each one connected with its neigh- bor, extended in a southeasterly direction off Cannery Point, the southeast tip of Point Roberts, for a mile. Two other connected traps near the international boundary extended for four-fifths of a mile. Such long strings of traps were not uncommon, and the law advanced conservation by breaking them up. Another law, passed in 1897, prohibited traps from operating in water over 65 feet in depth. However, this law was not observed for several years. In 1913, soundings by the State Fish Commissioner (Washington State, 24th and 25th reports, 1916, p. 36) revealed 11 traps operating in water exceeding the legal depth by l%-27 feet, The owners admitted having driven these traps in the same locations for 12 years, but changed them to conform with the law. NUMBER IN OPERATION The total number of traps operated each year in Puget Sound has been rather difficult to obtain owing to the fact that a trap need be driven only once in 4 years in order to hold a location. Furthermore, where the driving of one trap would tend to lead fish away from another it has been the general practice among companies to drive the one location for fishing and to hold the other by driving a “dummy” trap there at least once every 4 years. A dummy trap was very poorly constructed, and hung chiefly with old, worn-out gear. The object was merely to comply with the law, the dummy not being expected to catch more than a few dozen fish. In addition to these dummy traps there have always been some traps of an experimental nature, especially in years of abundant runs and good prices. Many of these locations have been driven but once, others have been tried from time to time. The efficiency of the traps has not varied as much as the number in operation from year to year might seem to indicate, since the best locations are practically always fished, and many of the extra traps, added during years of abundant runs or high prices, are driven in inferior locations. The number of traps in operation, exclusive of dummies, is given in table 6. Between 1895 and 1900 the traps doubled in number three times, reaching a peak of 163 in 1900. During this first great expansion many inferior locations were tried and later abandoned, as shown by the lessened number in all years except for those of the SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 717 big sockeye runs. During the World War the number of traps remained high even during years of poor runs owing to the high prices of salmon, but immediately there- after the number fell off sharply and never fully recovered. The number of traps has been reduced to a slight extent by regulations closing certain areas to fishing. In 1921 the State fishery board set aside certain areas as salmon preserves, but they were areas that had been without regular traps for several years. The San Juan Island preserve had had a few traps at times, especially on Shaw Island, but none of them had been successful. In 1924 the Hood Canal preserve, which was created in 1921 to protect the lower end of Hood Canal, was extended to take in nearly all of the canal. Two or three traps that had been operating in the fall, chiefly for chum salmon, were thus removed. In the same year traps wTere prohibited in the Hope Island area, thus removing about a dozen traps catching chiefly Skagit River salmon. However, this prohibition was modified the following year. Table 6. — Number of salmon traps operated from. 1893-1934, exclusive of dummy traps Year Traps operated in Puget Sound 1 Traps with data before 1915 British Colum- bia traps Total Year Traps operated in Puget Sound 1 Traps with data before 1915 British Colum- bia traps Total Regu- lar Exper- imen- tal Total Regu- lar Exper- imen- tal Total 1893 2 13 3 13 1914 116 71 8 10 126 1894 3 19 1 20 1915 121 27 148 8 10 158 1895 * 21 11 2 23 1916 96 14 110 4 114 1896 26 2 1917 121 32 153 7 160 1897. ___ 5 71 35 4 75 1918 98 11 109 8 11 120 1898 39 6 32 3 48 1919 101 13 114 8 8 122 1899 98 14 112 76 3 115 1920 71 8 79 8 8 87 1900 130 33 163 74 3 166 1921-._ 91 5 96 8 104 1901 140 9 149 88 3 152 1922 62 1 63 4 67 1902 105 37 142 82 3 145 1923 90 6 96 6 102 1903 104 2 106 67 3 109 1924. 68 3 71 4 75 1904 75 4 79 44 1 80 1925 104 13 117 5 122 1905 137 1 138 70 7 17 155 1926 86 5 91 6 97 1906 88 8 96 61 8 10 106 1927. 97 3 100 5 105 1907. « 98 60 7 12 110 1928 - 86 2 88 5 93 1908 80 49 7 12 92 1929 116 14 130 6 136 1909 152 76 8 15 167 1930. 102 9 111 6 117 1910 93 8 10 103 1931 93 5 98 4 102 1911 111 68 8 10 121 1932. 47 1 48 4 52 1912 110 66 8 10 120 1933 - 80 3 83 5 88 1913 168 84 8 176 1934 84 8 92 5 97 1 1898-1906 partly from State license flies at Auburn. 2 At Point Roberts only, Rathbun (1899). > Partly estimated from Rathbun (1899). 4 Rathbun (1899). * Fidalgo Island Packing Co. records. 6 1907-14 estimated. Number for which we had data estimated as 61 percent of traps operated, as from 1901-06 (except 1905), when it varied from 56-64 percent. In 1905 twice as many operated and this was used for 1909 and 1913. : Partly from Pacific Fisherman. * Number licensed. •Estimated. LOCATIONS FISHED Because of the sketchy nature of the available data no attempt has been made to give accurately the number of traps operating in each area prior to 1898. Traps were first tried at Point Roberts in 1880, but could hardly be considered a success until 1891. In the few years from 1891-97 traps were driven in numerous localities through- out Puget Sound, but mostly without much success. The locations that proved suc- cessful were continued, and for the others only a few records are available. 718 BULLETIN OF THE BUREAU OF FISHERIES The number of traps fishing in each locality since 1898 is shown in table 7. 3 It is apparent that while the trap fishery was widespread its use was emphasized only in those few localities where trap sites could be favorably situated to intercept the salmon runs, and where there was a depth and a bottom suitable for driving. Where these conditions were well satisfied, as in Boundary Bay, the number of traps was large. In some areas, like the Salmon Banks or Rosario Strait, the fish were present, but suitable places for driving were scarce, and few traps were constructed. Table 7. — Number of traps fishing in various localities, 1898-1934 Area 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 Point Roberts 5 7 7 5 5 5 4 5 5 4 3 6 5 4 4 6 4 6 Boundary Bay (U. S. traps) 18 30 26 35 35 30 21 29 23 23 21 29 20 21 19 31 21 24 Birch Bay ... . 5 12 14 15 16 15 8 10 9 8 6 14 9 10 8 13 9 16 I.ummi Island 3__ _ _ __ 1 1 3 3 3 2 5 6 4 1 8 5 5 5 10 5 9 Rosario Strait 1 i 5 1 i 2 2 i 2 2 2 1 3 i 3 4 5 5 12 South Lonez 2 3 4 Salmon Bank _ 5 6 6 7 5 4 4 4 4 3 5 5 4 4 4 4 5 Haro Strait 1 3 2 3 2 1 1 1 1 1 3 2 6 Waldron Island ___ _ __ _ 1 1 1 1 1 1 1 2 West Beach __ . .. ... 3 4 2 i 2 2 i 4 4 6 6 4 4 6 6 6 6 13 Ebevs Landiner __ 1 1 1 1 1 Middle Point 1 2 1 1 2 4 Strait of Juan de Fuca 5.. _ 1 2 Admiralty Bay and Bush Point 1 4 3 4 2 5 4 4 3 3 3 3 3 2 3 8 Oak Bay and Point No Point - 1 1 1 1 1 2 1 1 1 1 1 3 1 1 2 4 4 4 4 Useless Bav and Possession Sound _ 1 1 Meadow Point and south ... _ _ __ 1 3 East of Whidbey Island 8 5 4 3 i 3 3 3 3 3 3 2 2 10 Total 32 70 74 80 83 65 44 70 61 60 49 76 58 70 63 85 72 132 Area 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 Point Roberts 5 7 5 6 5 6 2 4 4 6 7 8 5 8 8 5 3 3 5 Boundary Bay (U. S. traps) 19 29 19 18 14 21 11 17 12 19 14 19 16 28 18 21 10 22 22 Birch Bay 2 - 9 14 10 9 3 8 3 8 4 10 7 9 7 10 8 7 3 7 6 Lummi Island 3 6 9 5 5 4 5 3 4 3 4 4 4 4 5 5 6 4 4 3 Rosario Strait 4 7 13 8 5 6 7 1 2 5 9 5 3 4 10 6 7 1 5 8 South Lopez 2 5 4 6 3 5 4 4 5 4 4 4 3 5 3 3 3 3 4 Salmon Bank _ 5 7 6 8 3 6 3 8 5 7 5 8 6 9 8 5 3 5 8 Haro Strait 4 7 5 6 2 6 2 6 4 9 4 6 4 9 5 5 3 5 Waldron Island 2 1 1 1 1 1 1 1 1 i 1 1 West Beach . ... 11 13 12 13 12 11 9 13 12 12 11 11 13 12 15 8 4 5 7 Ebevs Landing... 1 2 2 2 1 1 1 1 2 1 3 2 2 2 1 1 1 Middle Point 2 2 2 3 1 1 2 3 1 2 2 2 2 2 3 2 1 1 Strait of Juan de Fuca 4 2 1 1 2 1 1 1 1 1 1 2 2 2 2 1 1 i 1 Admiralty Bav and Bush Point.. ....... 9 7 9 7 6 4 3 7 7 8 8 8 9 9 11 11 3 8 7 Oak Bay and Point No Point 3 4 4 3 3 2 3 3 3 3 3 2 2 3 3 2 1 Hood Canal __ ... 2 4 4 4 2 3 4 1 1 1 1 3 2 Useless Bay and Possession Sound . . 1 3 1 1 1 1 1 1 1 1 Meadow Point and south . 2 i i 2 2 1 i i 3 5 2 2 2 4 3 2 3 3 2 East of Whidbey Island. 7 8 7 9 8 8 a 10 .... 14 11 11 5 9 8 8 8 9 10 Total 97 138 106 110 78 94 63 96 71 117 91 100 88 129 111 98 48 82 92 ' Incomplete before 1915. 2 Including Alden Bank. 3 Including Bellingham Bay. * Including Padilla Bay and Guemes Island. 4 South side. During the period from 1915-34, 33 percent of all the traps have been located north of Sandy Point — Point Roberts, Boundary Bay and Birch Bay areas; 27 percent south of Sandy Point and north of Deception Pass — Rosario Strait, Salmon Banks, 3 These trap locations have been determined from charts made by the U. S. Army Engineer’s Office in Seattle, 1919-34, from the files of the State of Washington Department of Fisheries, and from numerous records and maps obtained from various operators. This table is not complete for years before 1915, and a few minor traps have not been identified as to location since that date. Since table 7 is based only on traps for which locality data are available, the numbers of traps do not check with table 6 giving the total number operated. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 719 Haro Strait, Lummi Island, etc.; 16 percent along West Beach, Ebeys Landing, and the south side of the Strait of Juan de Fuca; 9 percent east of Whidbey Island — chiefly Hope Island area; and 15 percent south of Point Wilson — Admiralty Bay, Hood Canal, etc. CANNERY EXPANSION FROM THE TRAP FISHERY After more than a decade of cannery operation in the southern portion of Puget Sound, 1877-90, during which time 3 or 4 small canneries were annually engaged in the industry, business had fallen off to such an extent that only 1 cannery operated in 1890. The successful use of salmon traps at Point Roberts resulted in the building of a salmon cannery at Semiahmoo in 1891, one at Point Roberts in 1893, and another at Friday Harbor in 1894, the number quickly increasing to 19 by 1900. In 1901, a big sockeye year, the number dropped to 16, owing to overproduction the previous year, especially of the cheaper grades. In 1902, however, the number rose again to 20 (see table 1). In 1902, in addition to the original sockeye cannery at Semiahmoo, there were 2 at Point Roberts, 3 at Blaine, 3 at Fairhaven (now South Bellingham), 1 at Chuckanut, 1 on Lummi Island, 6 at Anacortes, and 1 each at Friday Harbor, Port Angeles and Seattle. The successful use of salmon traps near Sooke, on Van- couver Island (see fig. 2) caused the building of a cannery at Victoria in 1905. SEASON One very striking instance of the increased intensity of fishing in later years is furnished by changes in the season when the fish traps were operated. The season has been measured by the dates of the first and last lift of a trap. Since the traps usually fish from about two days to as much as a week before the first lift, all seasons mentioned are slightly less than the actual time fished. In Boundary Bay, the most important sockeye area, the date by which half the traps had been lifted for the first time was July 9 in the period 1897-1902, in the next 8 years, it advanced to July 7, in the following 16 years it averaged July 4, and in the last period, 1927-34, it had advanced to June 25, a total for the whole period of 14 days. (See fig. 10.) The change at the end of the season is more striking. From August 23 the closing date became later and later until, in the last 8-year period, it was September 27. A 46-day season had changed to one of 95 days. The reasons for the change are best explained by comparing trap seasons with the curves for seasonal occurrence of each species. It is evident that the late spring fishing is to increase the catch of kings. In the early days the traps usually stopped fishing in the odd-numbered years before the sockeye run was quite over in order to avoid bothering with the tremendous pink runs which were of little value. In recent years the traps have usually fished until the pink runs are over. A somewhat similar story is told of the traps in the area between Point Wilson and Point No Point (Admiralty Inlet). Admiralty Inlet was a fall fishing area for many years. The opening date for the period 1900-1910 averaged August 27, and for the next 8 years August 23. From 1919-26 it had advanced to June 14 and in the last 8-year period, 1927-34 it was May 30, a change in the opening date of 85 days. 720 BULLETIN OF THE BUREAU OF FISHERIES During the earlier years this southern area was fished chiefly for cohos and chums and the pink run was usually in full swing before fishing commenced. Later the fishing was advanced to take in all of the pink run, and more recently a large proportion of the traps made their first lift about May 4 ; evidently fishing immediately after the cessation of the April closed season to catch the early run of king salmon. In the areas east of Whidbey Island the season was always very long. The traps opened in late April and early May to take kings and steel heads, and to be in time to fish the June run of sockeye to the Skagit River. They then remained open for the coho run in the fall. The season fished by the traps has been modified somewhat by regulation. The first of these was a law, enacted in 1905, imposing a weekly closed season of 36 hours. Our data do not in- dicate any observance of this law prior to 1908. This weekly closed season was modified in 1915 to apply only during July and August. june july august sept oct nov Commencing in 1921 it was TRAP FISHING SEASON applied during the balance Figure 10. — Length of the trap-fishing season in Boundary Bay by groups of years from of the year to the districts 1897-1934. The length of the season is gaged by the percentage of the traps that were gagj. "Whidbey Island and actually in operation during the various parts of the season. Note the progressive •C , increase in the length of the season during which they fished. SOUtll of a line from Pomt Wilson to Partridge Point. A closed season was introduced in 1915 from January 18-April 15, inclusive. This affected some forms of gear but had almost no effect on trap fishing. In Hood Canal an additional closed season from November 16 to January 1, inclusive, probably had some effect on the few traps in that area. The area near Tacoma, including Poverty Bay, was also closed from November 16-30, inclusive. The closed periods from 1921-34 are given in the following table. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 721 Table 8. — Puget Sound closed seasons from 1921-34 1 Year All districts Southern district 1 Middle district Northern district 3 From — To— From — To— From — To— From — To— From — To— 1921. 1922 Oct. 26 Nov. 6 do Apr. 30 do Sept. 6 Sept. 15 Sept. 6 Sept. 15 Sept. 6 Sept. 16 1923 ...do Sept. 6 Aug. 25 ...do Sept. 15 Sept. 3 ...do Sept. 6 Sept. 15 Sept. 6 Sept. 15 1924 __ do ...do 1925 ...do -..do do Sept. 6 Sept. 15 Sept. 6 Sept. 15 1926 do 1927 ...do ...do ...do Aug. 25 Sept. 3 Sept. 6 Sept. 15 Sept. 6 Sept. 16 1928 ...do 1929 . do do Aug. 25 Aug. 25 ...do ...do ...do Sept. 3 Sept. 3 --.do ...do ...do Sept. 6 ...do ...do .. .do Sept. 11 Sept. 2 Sept. 15 do -..do ...do Sept. 20 Sept. 11 Sept. 6 ...do ...do ...do Sept. 11 Sept. 2 Sept. 15 ...do ...do ...do Sept. 30 Oct. 1 1930 1931 1932 1933 1934 ...do Nov. 11 ...do ...do ...do ...do ...do __.do ...do ...do Sept. 21 Sept. 30 1 All dates are closed days. 1 East of Whidbey Island and south of line Point Wilson to Point Partridge. * North of line Sand Point to Patos Island (Birch Bay, Boundary Bay and Point Roberts areas). The closed periods were introduced largely for the protection of the pink salmon and so at first were confined to the odd-numbered years, except in 1924, when it was hoped that there might be a fair run of pinks from the fry liberated by the hatcheries from eggs taken in Alaska. Since 1930 this closed period has been extended to the even-numbered jmars for the protection of the sockeye. The fall closing date was inaugurated in 1921 and applied to all districts. This closing protects a considerable portion of the chum salmon runs, and a small percentage of the cohoes. SEASONAL OCCURRENCE OF EACH SPECIES The seasons during which each species migrates through the salt water toward the spawning grounds is of the utmost importance from a standpoint of conservation as it determines, to a great extent, the possibilities of so regulating the fishery as to allow the talcing of the more abundant species, while protecting the less abundant. There is, of course, considerable variation from season to season in the time of run, although a general average may be obtained. The traps furnish the best measure of seasonal occurrence since a trap does not fluctuate from day to day in its fishing effort, but con- tinuously samples the runs that are passing by. For sockeyes data were used for 12 traps, all located north of Deception Pass. They fished in various years from 1896-1934, catching a total of 13,129,869 sockeyes. In making a seasonal curve (fig. 11), the total catch of each 7-day period was divided by the number of trap-fishing days. However, for sockeyes the trap-fishing days for each trap were weighted by the fishing efficiency of that trap. (Cf. page 768.) For species other than sockeye the traps were not weighted. For king salmon the catches of 17 traps were employed; 7 were north of Deception Pass, 4 at West Beach, 2 at Middle Point, 2 in the Hope Island Area, and 2 in Ad- miralty Inlet. They caught a total of 580,698 fish from 1900-1934. The pink-salmon curve was derived from 4,467,115 fish caught in 16 traps; 9 located north of Deception Pass, 1 at Ebeys Landing and 6 in Admiralty Inlet. Since little effort was made to take pinks during the earlier years of the fishery, the material used is from odd-numbered years from 1919-33. As 1919 is the only year in 722 BULLETIN OF THE BUREAU OF FISHERIES which a fall closed season was not in effect it was necessary to determine a small portion of the curve by empirical methods. The curves for the 9 northern and the 7 southern traps were each calculated separately and combined with equal weighting to obtain the final curve. (See fig. 11.) To obtain the seasonal occurrence for coho salmon 26 traps were used; 15 located north of Deception Pass, 2 in the Hope Island Area, 2 in Middle Point Area, 1 at Dungeness Spit, and 6 in Admiralty Inlet. They fished from 1900-1934, taking 5,652,592 fish. For the chum salmon, as for the pinks, a northern and a southern curve were each calculated and then combined. However, in the case of the chums, the southern curve was given double weight, as more chums are always caught in the southern areas. A total of 13 traps were used; 7 north of Deception Pass and 6 in Admiralty Inlet, catching 946,094 fish. The curves for all species are given in table 9 and shown in fig. 11. Table 9. — Seasonal occurrence in Puget Sound traps Percentage occurrence weeK ending— Sockeye King Pink Coho Chum Sockeye King Pink Coho Chum Apr. 21 0. 425 0. 425 1.353 1.778 0. 391 2.259 0.018 0. 391 4. 037 0. 018 May 12 .351 3. 212 .035 0. 001 .742 7.249 .053 0. 001 .328 3.649 .059 .001 1.070 10. 898 .112 .002 May 26 .149 3.780 0. 002 .054 .002 1.219 14. 678 0. 002 .166 .004 June 2 .061 4. 166 .002 .084 .006 1.280 18. 844 .004 .250 .010 June 9 .018 4. 770 .005 .080 .011 1.298 23. 614 .009 .330 .021 June 16. .015 5. 145 .006 .103 .008 1.313 28. 759 .015 .433 .029 Juno 23_— .087 5. 921 .007 .175 .007 1.400 34. 680 .022 .608 .036 June 30 .468 6. 330 .010 .174 .013 1.868 41.010 .032 .782 .049 July 7 2. 206 7.292 .017 .351 .026 4.074 48. 302 .049 1. 133 .075 July 14 4. 495 6. 696 .027 .393 .032 8. 569 54. 998 .076 1.526 .107 July 21 8.408 6. 252 .170 .466 .063 16. 977 61.250 .246 1.992 .170 July 28 16. 098 6.188 1.463 .532 . 172 33. 075 67. 438 1.709 2. 524 .342 Aug. 4 26. 344 6. 072 3. 660 .709 .450 59.419 73. 510 5. 369 3. 233 .792 Aug. 11 20.911 6. 149 6. 875 .962 .816 80. 330 79. 659 12. 224 4.195 1.608 Aug. 18. . _ 11.224 5. 565 10.117 1.413 1.234 91. 554 85. 224 22. 361 5. 608 2. 842 Aug. 25.. 5. 542 4.456 21. 120 2.717 1.863 97. 096 89. 680 43. 481 8. 325 4. 705 Sept. 1 1.530 3. 406 23. 837 3.911 1.835 98. 626 93. 086 67. 318 12. 236 6. 540 Sept. 8 .493 2. 875 19. 591 6. 953 1.977 99. 119 95. 961 86. 909 19. 189 8. 517 Sept. 16... .071 2.074 8. 660 10. 795 2. 298 99. 190 98. 035 95. 569 29. 984 10.815 Sept. 22. . 121 1.105 3. 452 12. 652 3. 246 99. 311 99. 140 99. 021 42. 636 14. 061 Sept. 29 . 132 .451 .830 12. 129 7. 376 99. 443 99. 591 99. 851 54. 765 21. 437 Oct. 6 .048 .167 .107 12. 628 9. 244 99, 491 99, 758 99, 958 67. 393 30. 681 Oct. 13... .018 .069 .041 11.978 11.803 99. 509 99. 827 99. 999 79. 371 42. 484 Oct. 20 .036 .041 .004 8. 357 12. 656 99. 545 99. 868 100. 003 87. 728 55. 140 Oct. 27 .096 .038 .003 5.313 13. 643 99. 641 99. 906 100. 006 93. 041 68. 783 .099 .064 3. 108 10. 335 99. 740 99. 970 96. 149 79. 118 .258 .030 1.628 7. 131 99. 998 100. 000 97. 777 86. 249 1.502 5. 747 99. 279 91. 996 .667 3.282 99. 946 95. 278 .051 1.865 99. 997 97. 143 Dec. 8 ... .005 1.097 100. 002 98. 240 .431 98. 671 .669 99. 340 .553 99. 893 .066 99. 959 Jan. 12 .015 99. 974 .006 99. 980 Cumulative percentage occurrence The seasonal occurrence of each species is quite distinct from any of the others and the modes of the five curves are about a month apart; kings, sockeyes, pinks, cohos and chums following in that order. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 723 The king-salmon run covers a long period of time, but averages much earlier than those of the other species. Thus 40 percent of the run is over by June 30, whereas no other species has reached 2 percent of its run by that date. The next species to appear in abundance is the sockeye, overlapping the latter portion of the king-salmon run. On the average, over a long period of years, the sockeye runs have been practically over by August 25. By that date only 5 percent of the chums, and less than 10 percent of the cohos, have passed the traps. However, over 40 percent of the pink salmon run is complete. The pink salmon run lasts for such a short period that it is practically over before the cohos appear in abundance, 85 percent hav- ing passed by the time 20 percent of the cohos are taken. The coho and chum sal- mon are the backbone of the fall fishery. Neither species presents a well-defined mode, but the centers of the two distributions are between three weeks and a month apart. Since both species run for a considerable length of time there is a considerable degree of overlapping in their time of run. During the five 7-day periods, from September 23- October 27, inclusive, 54.7 percent of the chum and 50.6 percent of the coho runs occur. RELATIVE IMPORTANCE OF EACH SPECIES AND DISTRICT The relative importance of each species of salmon to the trap fishery is shown in figure 12 which illustrates the number of each species of salmon caught by traps in the 5 major areas during the past four decades. The areas shown are (1) North of Sandy Point, (2) Sandy Point to Deception Pass, (3) West Beach and Ebeys Land- ing, (4) the Strait of Juan de Fuca, and (5) the waters east of Whidbey Island and south of Point Wilson. For the past two decades the Puget Sound data are com- plete. Before that they represent only that portion of the trap catches for which original records could be secured. For sockeye this portion was about 80 percent of the trap catches in Puget Sound and practically all of the Canadian trap catches. For the other species the proportion represented is even higher than is the case for the sockeyes, as the data are more complete in the latter part of the period when more of the other species were used. For Canadian traps the other species are not included, as the data were not available. From figure 12 it is to be noted that 53 percent of the entire catch came from the district north of Bellingham — Point Roberts, Boundary Bay, and Birch Bay Areas. The next largest district, from the standpoint of catch, was that south of Bellingham 71941—38 3 Figure 11.— Seasonal occurrence of all species of salmon as shown by Puget Sound trap catches. Each ordinate shows the percentage of the run occurring during the indi- cated 7-day period. 724 BULLETIN OF THE BUREAU OF FISHERIES (Sandy Point) and north of Deception Pass, which includes the San Juan Islands. The second district accounted for an additional 27 percent. In other words, 80 percent of the trap catches during the past 40 years have been from the areas north of Deception Pass. Of the remaining 20 percent, less than 11 percent came from the inside waters of Puget Sound — east of Whidbey Island and south of Point Wilson. 12 3 4 ECADtS 12 3 4 2 3 4 12 3 4 2 - _l _ia Xeb 2 - 0 I 0 TRAP CATCHES BY DECADES 1895-1934 SOCKEYE Fifty percent of the trap catch were sockeyes, 34.9 percent pinks, 9.3 per- cent cohos, 3.6 percent kings, and only 2.2 percent chums. These figures, however, give only a general picture. If the catches are considered by districts it is found that the two districts north of Deception Pass caught 56.8 percent of the sockeyes, 36 percent of the pinks, 4.3 percent of the cohos, 2.6 per- cent of the kings, and 0.4 percent of the chums. That is, all but 7.3 percent of the catch consisted of but two species, sockeye and pink. In the West Beach and Ebeys Landing district the catch was 32.2 percent pinks, 25.7 percent sockeyes, 20.8 percent cohos, 16.9 percent kings, and 4.4 percent chums; the sockeye and pink, the two dominating species north of Deception Pass, thus accounting for but 58 percent of the catch. East of Whidbey Island and south of Point Wilson, except for the pinks, the catches are very different, being 43.5 percent cohos, 35.9 percent pinks, 13.9 percent chums, 6 percent kings, and only 0.8 percent sockeyes. The changes in the catch by decades in each of the 5 districts are apparent. The catches of pinks, for example, after being subjected to exploitation in the second decade, 1905-14, fell off tremendously in the third and fourth decades in the two northern areas. In district 5, however, they have continued to rise. PINK COHO CHUM j=d □ : AREA NUMBERS Figure 12. — Showing the number of each species of salmon caught by traps in the major areas during each of the past four decades. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 725 THE PURSE-SEINE FISHERY By George B. Kelez The importance of the purse seines has varied considerably during the history of the salmon fisheries. Shortly after their introduction they surpassed the drag seine, their forerunner, and were in turn superseded by the traps. They again be- came an important factor when motor-driven vessels were employed. Since the use of traps has recently been prohibited in Puget Sound waters they are the only important gear operating in that district, and a knowledge of their effectiveness, the species taken, and the seasonal nature of operations in various areas, is of extreme importance to the administration of the fishery. DRAG SEINES One of the earliest forms of fishing gear to be used on Puget Sound was the drag seine. This was a long shallow net provided with cork floats on the upper edge and lead weights on the bottom, and was pulled by lines attached to each end. In use the net was loaded into a skiff and one of the hauling lines passed to a man on shore. The skiff was pulled directly away from the beach until all the line was payed out, then turned parallel to shore and the net run out, after which the skiff returned to the beach with the second line. The lines were rapidly hauled in until the wings of the net were ashore and the fish concentrated in the center or “bunt” of the net, whereupon the remaining web was quickly hauled onto the beach, landing the catch. Since a beach free of large rocks or other obstructions was nec- essary for landing the catch, the drag seiners worked in unobstructed places where the fish were concentrated by favorable currents, or where their migration routes led them close inshore. The mouths of streams where the mature fish schooled before ascending to spawn were favorite locations prior to the passage of legislation protecting these areas. The number of drag seine-licenses from 1897-1934 is shown in table 10. The greatest number of licenses was issued during the period from 1908-14, and that number steadily decreased thereafter. Drag seines were commonly used in early years along the east shore of Van- couver Island and in Puget Sound near the cities of Seattle, Tacoma, and Olympia. They later appeared on the sands at the mouth of the Skagit River, the Nooksack River, and Lummi Slough, as well as at Point Roberts. They were also used exten- sively in the inlets and passages of the west shore of Puget Sound and in Hood Canal. In early years the catch of this gear consisted chiefly of coho and pink salmon. Later, chum salmon became of considerable importance, and in some years large numbers of king salmon were caught. Subsequent to 1924 the total catch of the drag seines has been only a few thousand fish per year, consisting chiefly of pink salmon. Sockeye, which were caught only occasionally in former years, are now second in importance. These changes in the proportion of various species in the catch have been due in part to the competition of other forms of gear, but have re- 726 BULLETIN OF THE BUREAU OF FISHERIES suited chiefly from the closure, by legislation, of many districts which were frequented by the drag seines. This gear is still used in the region, but it is now of very little importance. Table 10. — Puget Sound drag seine licenses, 1897-1984 Year Number Year Number Year Number 1897 59 1911 307 1923-... 111 1898 59 1912 __ 243 1924 109 1899 125 1913 238 1925 144 1900 - 114 1914 354 1926 130 1901 74 1915 187 1927 135 1902 74 1916. 189 1928 120 1903 171 1917 218 1929 123 1904 95 1918 185 1930 123 1905 69 1919 187 1931 104 1906 - 123 1920 144 1932 84 1907 176 1921 116 1933 109 1908 283 1922 108 1934 90 1909 242 1910 247 DEVELOPMENT OF THE PURSE SEINE EARLY SEINES The purse seine is a net not unlike the drag seine in shape, but much longer and deeper. Its chief characteristic is the purse line, a stout rope or cable, rove through metal rings attached to its lower edge. This net is used in deep water. When a school of fish have been observed the net is set around them, the two ends are brought together, and the purse line hauled in. This closes the bottom of the net, trapping the fish within it. Although the purse seine is inseparably associated at the present time with the highly specialized vessel from which it is fished, the seine itself has undergone but little change, except in size, whereas the vessel is the product of long years of development and experience. The date this gear was originally introduced on Puget Sound is a matter of con- jecture. Hittell (1882) reported it to be an important form of gear in 1882. He stated that the fishery was prosecuted almost entirely by Indians and that the nets were from 50-80 fathoms in length, and 4-8 fathoms in depth. These seines were set from large canoes from which they were also pursed when the set was complete. Other canoes cruised around the net, the crews beating the water with their paddles to keep the fish schooled. Coho, pink, chum, and king are listed as the species caught, and from two to five thousand fish might be taken at a single haul. Hittell offers no information as to the date of introduction or as to the number of years that these nets had been used. SCOW SEINES This type of fishing must have undergone a considerable development in a brief space of time. Collins (1892) reports purse seines to be “the most effective form of apparatus yet used in the salmon fishery,” and states that they were introduced in 1886. They are described as being approximately 200 fathoms long and 25 fathoms deep. They were set from a four-oared skiff, the after 8-foot portion of which was decked to form a platform for stowing the seine. A scow 20 feet long and 8 feet wide, equipped with a hand winch, was used for pursing the net and carrying the SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 727 catch. One end of the net was attached to the scow and the bulk of the seine was carried by the skiff, from which it was set around the school of fish. The free end was brought back to the scow where the two ends of the purse line were then hauled in by the means of the winch. A “plunger,” consisting of a stout pole with a wooden box shaped like a truncated pyramid and attached to the lower end, was thrust re- peatedly into the water at the opening between the purse lines to keep the fish from escaping there. This was necessary, since pursing the net was a very slow proce- dure. As high as 6,690 fish were taken in a single haul. At this time the principal fisheries on the Sound were at Seattle, which then had three canneries, at Tacoma, and at Port Townsend (see figs. 2 and 3). Rathbun (1899) describes the purse seines in use about 1895 as essentially similar to those of 1888. He also dates their introduction to these waters as 1886, doubtless based on Collins’ report, and gives their size as ranging from 150-250 fathoms in length, from 14-25 fathoms in depth, and being of 2%-3-inch mesh. Rathbun states that in 1893 and 1894 several seines fished regularly at Point Roberts, some were employed at Port Angeles, and some in the San Juan Islands. The principal purse-seine fishery remained at Seattle, however, where the catches were sold to the fresh-fish markets as well as to the canneries. Eleven seines fished out of Seattle in 1895, and at least 20 in 1896. Individual hauls of from 1,500 to 2,500 fish were not uncommon, and one Seattle cannery received from 6 seines an average of 12,000 cohos a day during the height of the 1895 run. Although traps had become the chief source of salmon in other districts by 1895, the seines still supplied the greater part of all fish used in the Seattle aiea. Purse-seine fishing in the San Juan Islands received considerable impetus from the location of a cannery at Friday Harboi in 1894, and three at Anacortes in 1896. Large shore camps, established at points close to the best fishing grounds, provided living quarters for the crews. The seine scows and skiffs were towed to these camps at the beginning of the run and remained there during the season. The individual seine outfits also had to be towed to various parts of the fishing grounds, for their own movements were limited to the distance that the boat-pullers could row the heavy skiff and attendant scow, and at the close of the day’s fishing the whole apparatus had to be returned to the camp ground. Because of these limitations, fishing by purse seines was confined to a radius of a few miles from the base camps. The first purse seines had been employed during the fall season in the southern districts of the Sound where the bulk of the catches consisted of coho salmon. Although large quantities of chum and pink salmon were available, the lack of a ready market curtailed the fishing for these species. A considerable increase in the number of canneries after 1895 furnished a better market for species other than coho, and the fishing season of the seines was considerably lengthened. The license records of the Washington State Department of Fisheries show that, during 1897, 22 licenses were issued during the month of June, 11 during July, 1 in August, and 13 in September, In 1898 approximately 31 licenses were issued up to and including duly 6, and none thereafter until September 10. Nine licenses were issued after the latter date. It will be noted that the larger number of licenses were issued during the early summer, that few or none were issued during a slack period of several weeks, and that an addi- 728 BULLETIN OF THE BUREAU OF FISHERIES tional number were issued in the later summer or fall. For convenience, the first group of licenses will hereafter be designated as “summer licenses,” and the second group as “fall licenses” (see table 11). Although somewhat obscured by a general increase, the odd-numbered years show a larger number of licenses than do the even- numbered years. This is largely due to the greater availability in those years of the pink salmon, which by this time could be marketed in sufficient quantity to encourage their pursuit by the seine fleet. Table 11. — Puget Sound purse-seine licenses, 1897-1915 Year Summer Fall Total Year Total 1897 34 13 47 1907 64 1S98 31 9 40 1908 69 1899 58 14 72 1909. 95 1900 41 16 57 1910 120 1901 45 22 67 1911 133 1902 - 59 19 78 1912 169 1903 _ 79 8 87 1913 252 1904 53 19 72 1914 288 1905 73 18 91 1915 308 1906 73 6 78 DEVELOPMENT OF THE MODERN PURSE-SEINE VESSEL Introduction of Power Perhaps the most important single factor which influenced the development of the purse-seine fishery was tbe introduction of the internal-combustion engine for fishing vessels. The Pacific Fisherman Yearbook for 1919 states that the first gaso- line-powered boat on Puget Sound, exclusively engaged in the fishing industry, was a 32-foot fish carrier, the Silverside, built in Tacoma about 1898 for T. E. Eggers, a pioneer operator of that city. In a few years the success of power in other fishing vessels encouraged the purse seiners to take advantage of this new development. The complete change of the purse-seine fleet from oars to power was accomplished n a very few years. The Pacific Fisherman Annual Review for 1910 states: Skansie Brothers of Gig Harbor, pioneers in the use of gas engines, have ordered two new boats. They started six years ago (1904) with one boat powered with a 7 hp. “Frisco Standard”. They have since bought 15 more. The same publication, in the issue of 1907, includes in the caption of a picture of a power seiner the statement: Gasoline power is now universally used in seine boats. From these statements we may conclude that the change to power in the seine fleet was completed in but little more than 3 years. This change to power necessitated a revision of purse-seine fishing methods. The scow was replaced with a small open power boat and, although the skiff was retained, its function was reversed. The seine was now carried in the after part of the power boat. In setting, one end was made fast to the skiff while the seine boat circled the school of fish and payed out the net. The end of the net which had been made fast to the skiff was now brought aboard the seine boat and the purse line Figure 13. — An early Puget Sound purse-seine vessel, of 12 net tons, built in 1909. Note the large house on deck containing the crew’s quarters, and the outside steering wheel in front of the house. Figure 14. — A gasoline-powered purse-seine vessel built in 1920, of 27 net tons. The ports forward indicate that the crew’s quarters are in the forecastle. The "flying bridge” with steering wheel and controls, atop the wheel house, are visible. A seine skiff is towed astern. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 729 hauled in by means of a winch. The time necessary to reach and surround a school of fish was thus greatly decreased, with a corresponding increase in the efficiency of the seine. It has already been noted that purse seines became the most important type of gear in use on Puget Sound shortly after their introduction, and that by about 1895 the successful development of the salmon traps had relegated them to a position of much less importance. The adoption of power by the purse-seine fleet, which was consummated by 1907, now altered this position of minor influence in the fishery to one of considerable consequence, lor what had been a relatively fixed type of gear became an extremely mobile one when the seine scows were superseded by power boats. This newly acquired mobility, allowing rapid shifting of operations during the season to any district in which salmon were abundant, has remained the outstand- ing characteristic of the purse-seine fishery. Improvements in Vessel Design The introduction of power was followed by a gradual but positive change in the type of vessels used. As the fishermen moved farther afield, the unsuitability of the open boat under adverse weather conditions soon became apparent, and seaworthi- ness became the major consideration when the seiners began fishing far out in the Strait of Juan de Fuca. The first improvement in design, a compromise hull partially decked forward, appeared shortly after power was introduced. Later vessels were built with a full deck, and, at the same time, their depth was increased considerably, providing greater carrying capacity and increasing their seaworthiness. By 1912 most vessels were full-decked. This roving type of fishery was greatly impeded by the necessity of the crew sleeping ashore, and crew’s quarters were soon placed on board. At first a long superstructure was built, but the quarters were later arranged in a forecastle under a slightly raised forward deck. The wheel house and galley were brought forward partially over the raised deck, which afforded more deck space and increased the seaworthiness of the vessel. The speed and maneuverability of the vessels was increased considerably as engine efficiency improved. These developments, together with the use of larger seines, brought about the introduction of the “turntable” upon which the seine was stowed. This was a free-turning platform mounted above the gunwales of the vessel at the stern, and still retaining the roller at the after edge, which had been used for many years. The seine could be payed out freely and rapidly from this turntable and also stowed thereon with far greater ease than before. At about the same time engine power was further utilized to operate the pursing winch. This reduced the labor and increased the speed of pursing the nets, thus effecting an increase in their efficiency. Figure 13, which was taken before 1913, shows that the outside wheel had been adopted by that date. The fishing captain was thus enabled to steer the vessel while standing on the forward deck where he was better able to observe the fish and set the net. Some 10 years later this outer wheel was moved to the top of the wheel house, allowing still greater range of observation (see fig. 14). At about the same time a power drive was applied to the turntable roller, allowing the net to be gotten on board for stowing far more rapidly and easily than before. 730 BULLETIN OF THE BUREAU OF FISHERIES Although the first Diesel-powered vessel on Puget Sound, the cannery tender Warrior, which was built in 1914 at Seattle by Nilson and Kelez (Pacific Fisherman Yearbook for 1919), was successful in operation and very economical, the original cost of these engines was too great to encourage their ready acceptance. However, during the years of expansion of the fleet following 1925, the many advantages of Diesel engines encouraged their installation in a majority of the new vessels. In recent years there have been no further radical changes in type or design of purse-seine vessels. Increase in Vessel Size Improvements in vessel design were accompanied by a parallel increase in vessel size. It is impossible to determine the exact size of all vessels in the fleets of early years, since most of them were of less than 5 net tons and were not required to be officially registered. We may obtain some indication of the increase in vessel size, however, from records of the vessels large enough to be registered. The average size of vessels of this class, built in 1906, was only 6 net tons. That of 1907 was 7.5 net tons, that of 1908 was 8.92 net tons, that of 1909 was 9.43 net tons, and that of 1910 was 9.97 net tons. This tendency to build larger vessels received great impetus with the beginning of the high-seas fishery at Cape Flattery and on Swiftsure Bank, where there were frequent storms, few harbors, and no protection. Practically no seiners had fished there prior to 1911, but the development of this fishery was very rapid. Several vessels were laid down during 1911 of more than 10 net tons, and in 1912 nearly 50 vessels of 15-25 net tons were constructed. The size of vessels has continued to increase since that time. EVALUATION OF FISHING INTENSITY SEASONAL FLUCTUATIONS IN FLEET SIZE Factors Affecting Seasonal Intensity Variations in number of licenses in odd- and even-numbered years, and the licensing of an additional amount of gear in the fall of the year, have been noted in the discussion of scow seines. The operation of these factors was intensified by the conversion of the purse-seine fleet to power vessels and by the increase in vessel size which followed. The larger seine vessels were now able to run from their home ports on Puget Sound to southeastern Alaska with little difficulty, and some even voyaged as far as Bristol Bay. The termination of the fishing season in Alaska usually occurred early enough to allow them to return to Puget Sound and fish during the coho and chum runs in the fall. Since about 1925 the development of Alaskan herring-reduction plants attracted a fleet of large, able seine boats which fished from June to August or September, and many of which then returned to the Sound to further swell the fall fleet. Other large seiners, which fished in the California sardine fleets during the winter months, often fished in this region later in the year. During seasons when heavy runs of salmon were anticipated, certain vessels from the halibut fleet, which were constructed with a low stern suitable for seining, also engaged in the purse-seine fishery. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 731 In even-numbered years, when the pink salmon did not appear, the departure of the larger vessels to other fisheries was especially common, and when the decreas- ing abundance of sockeye rendered summer fishing even less profitable many smaller vessels followed suit. Other factors have further intensified the annual change in the number of vessels. Prior to 1921, when regulations in waters of the State of Washington were undertaken by the State Fisheries Board, a considerable fishery for immature coho salmon was carried on in lower Puget Sound, especially off the southern end of Whidbey Island, in Possession Sound, and in Port Susan (see fig. 3). This fishery was pursued by a number of very small boats which fished during April and May of each year. When the regular seining season began, in June or July, most of these small boats transferred their licenses to larger vessels and engaged in gill netting during the remainder of the season. Closure to early fishing of a large part of these waters discouraged seining by the smaller boats. These various factors have caused considerable fluctuations in the size of the Puget Sound seine fleets, but have not obscured the striking difference in the number of seiners operating in the summer fleets of alternate years, or the distinct difference between the total fleets of odd and even years. Size of Summer and Fall Fleets on Puget Sound During the period from 1909-15, the number of seine licenses issued increased from 95 to 308 (see table 11). However, the dates on which fishing licenses were issued are available for only a few of those years, and the number of vessels fishing during different parts of the seasons cannot be determined for this early period. Beginning with 1916, the vessels fishing on Puget Sound in each year have been classified as summer or fall seiners; all those obtaining early licenses were tabulated as the summer fleet, and all vessels fishing after September 6 as the fall fleet. In most years there was a period of from one to four weeks preceding this date during which no licenses were obtained. A more detailed discussion of the time of change from summer to fall fishing will be presented later under a discussion of the fishing seasons of the fleets. The number of vessels in the summer fleets of each year from 1916-34 are presented as column totals in the bottom line of table 12; those of the fall fleet of each year are similarly presented in table 13. Table 12. — Summer purse-seine fleets on Puget Sound, 1916-34 Registered net tonnage 1 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 Below 5 4 30 23 21 25 2 1 2 3 1 1 1 1 1 5-9 32 40 11 14 5 7 2 5 3 5 3 3 3 6 5 7 3 5 5 10-14 78 103 37 46 16 45 12 24 9 27 13 22 9 20 13 22 17 23 22 15-19 81 121 52 56 36 69 15 30 13 41 19 43 34 46 35 55 42 56 54 20-24 44 82 43 53 30 51 13 24 6 21 13 22 20 39 35 45 41 42 43 25-29 3 25 16 18 31 47 8 22 11 24 12 26 21 41 37 48 42 46 41 30-34 I 23 9 20 10 21 6 14 8 11 5 19 17 24 21 40 33 32 34 35-39 1 1 3 1 1 1 2 1 1 1 2 1 6 2 20 15 16 14 40-44 1 2 2 5 4 fi 4 45-49 1 1 3 4 1 2 1 Total 243 425 192 231 154 243 58 121 51 133 66 138 106 194 154 247 199 228 219 1 Vessels of 5 net tons and larger from official registers; boats below 5 net tons from State license applications. 732 BULLETIN OF THE BUREAU OF FISHERIES Table 13. — Fall purse-seine fleets on Puget Sound, 1916-34 Registered net tonnage 1 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 Below 5 6 31 24 23 26 3 1 1 1 3 1 2 4 4 1 1 4 1 5-9 36 40 11 15 7 7 3 4 3 6 4 3 3 6 5 8 4 5 5 10-14 90 106 45 54 23 46 27 26 20 29 20 26 22 26 18 22 18 25 22 15-19 88 125 58 63 43 69 30 35 26 47 36 44 44 49 47 56 45 56 57 20-24 51 83 50 58 34 53 27 28 16 27 19 27 25 39 38 45 42 44 43 25-29 4 26 19 19 33 49 22 23 16 29 28 33 34 41 40 48 42 47 43 30-34 - 1 23 15 20 13 22 12 14 9 14 20 32 31 27 30 40 34 34 36 35-39 1 1 3 1 1 2 2 2 2 2 8 6 10 8 20 15 17 15 40-44 1 1 3 2 5 4 8 4 4.5-49 1 1 1 4 3 1 2 Total-. 276 435 223 255 180 250 130 133 93 158 130 175 168 211 196 248 206 242 226 1 Vessels of 5 net tons and larger from official registers; boats below 5 net tons from State license applications. The data given in tables 12 and 13 are presented in graphical form in the top section of figure 15. The dotted line represents the size of the unallocated fleets from 1909-15. The size of the summer fleets from 1916-34 is represented by the solid line, and that of the fall fleets of the same years by the broken line. A general consideration of the number of licenses indicates a continuous increase in numbers from 1909-15, an extremely high point in 1917, very small numbers during the years of post-war depression, and a considerable increase thereafter. The year 1917 stands apart as the peak in number of vessels during the entire history of purse- seining in this region; 425 vessels fished during the summer season and 435 during the fall. Pink salmon were abundant, the appearance of a big run of sockeye was antici- pated, and a war-created demand for food had caused the price of raw fish to increase enormously. As a result, 122 new vessels were built that year, and almost every vessel on the Sound large enough to carry a net, including tow boats and pleasure craft, was engaged in purse seining. Although the regular seiners enjoyed a successful season, the sockeye run did not reach expectations, nor was the fall fishing especially profitable. Newcomers to the fleet found that purse seining was a most arduous vocation and that successful fishing was largely dependent upon the ability and expe- rience of the vessel captain. These factors, coupled with the fact that 1918 was an off year for the summer fishery, caused the fleet of that year to shrink to more normal proportions, even in the face of a continued demand for fish. Except for the alternate rise and fall in odd-numbered years, the fleets remained approximately constant in number from 1918 to 1921. The abundance of most species of salmon had diminished considerably and this, together with the financial depression of 1921, resulted in a marked decrease in the number of vessels fishing in 1922. Only three more vessels fished in the fall fleet of 1923 than in that of the previous year. This was the first year since the period of early development that the odd year showed so small a rise in number. The year 1924, when only 51 vessels fished during the summer season and 93 in the fall, was the first since 1909 in which less than a hundred vessels were licensed on the sound. However, beginning in 1925, the fleets again began to increase steadily in number. Although expansion ceased during the depression years following 1929, there followed no such decline as appeared in the period from 1922-24. The fleets of the 1930’s, were of approximately the same size as were those immediately following 1917. SALMON AND SALMON FISHERIES OF SWIFTSIJRE BANK 733 Size of Cape Seine Fleet The purse-seine fishery in the waters ofF Cape Flattery and in the vicinity of Swift- sure Bank, which has long been called the “cape” fishery in this region, experienced a development similar to that of Puget Sound. For years the cape fleet has consisted of the larger vessels of the Puget Sound fleet, which fished there before the salmon runs began in inside waters, together with a few large vessels which have proceeded to other fisheries when the season was over. During the years im- mediately following its de- velopment, tremendous catches encouraged many seiners to engage in this fishery. Most of the catch, however, consisted of im- mature fish which spoiled quickly, and the refusal of the canneries to accept them reduced the size of the cape fleet. This situ- ation was met temporarily by butchering the fish at Neah Bay, and by icing the catches. Somewhat later the canners employed a fleet of fast tenders or “buy- ing-boats”, to which the seiners transferred their catches, and which then returned immediately to the canneries. This not only enabled the seine boats to remain at sea for longer periods of time, but insured the delivery of the fish ashore soon after they were caught. This development again encouraged the increase of the fleet. Since this fishery was conducted in waters outside the jurisdiction of the State of Washington, the vessels were not licensed and no record is available of the size of the annual fleets. Gilbert (1913) reported 22 vessels fishing at the cape in 1911, and more than 100 in 1912. Data furnished by the major part of the fishing companies in the region, which include the greater part of the landings from the cape, are quite complete Figure 15. — Changes in numbers and efficiency of the Puget Sound purse-seine fleets. The early increase in size of the fleets, the decrease following the World War, and the increase during recent years may be seen, together with the general rise in efficiency throughout. 734 BULLETIN OF THE BUREAU OF FISHERIES for the period from 1927-34. These figures indicate that the numbers of vessels fishing there during these years were 64, 88, 122, 75, 163, 1 17, 104, and 142, respectively. CHANGES IN COMPOSITION OF THE FLEET The size-composition of the annual purse-seine fleets was essential to a determination of fishing intensity, for vessel size is an important aid in the calculation of vessel efficiency.4 The changes in size that have taken place during the history of the purse-seine fishery are partially indicated in figure 16, which shows the size distribution of vessels fishing on Puget Sound during the years 1911, 1917, 1925, and 1933. Of the entire fleet fishing during 1911, there 1 — t’ 1911 Ui 1 1 t J SIZE DISTRIBUTION PUGET SOUND SEINE FLEETS were only 6 vessels of 15 or more net tons. By 1917 ves- sels of this larger size consti- tuted the major portion of the fleet, although a consid- erable number of smaller vessels were still fishing. A number of vessels of 24 or more net tons fished for the first time that year. By 1925 vessels of less than 9 net tons had become scarce and the remaining fleet showed almost a bi- modal size distribution, somewhat obscured by the presence of several vessels of 22 net tons built in 1915, and several of 24 net tons built in 1917; there is a mode at about 16 net tons, and an- other some 12 tons greater. Three vessels of more than 35 net tons fished in 1925. In 1933 the small vessels had become even less nu- merous, and the remainder of the vessels, although similar in distribution to the fleet of 1925, show a considera- ble increase in the number of large vessels. Because we are especially concerned with the fleets of the past 18 years, the year of building the vessels fishing during that time, and their size, are shown in table 14. The persistence of old vessels in the fleet is noteworthy, even though most of the smaller ones of early years have disappeared. 4 Id order to establish the size of vessels composing the fleets of various years, it was necessary to identify as many as possible of the individual vessels which had engaged in the purse-seine fishery of the region. By means of the license applications in the files of the Washington State Fisheries Department, the Fireman’s Fund register of vessels documented on the Pacific coast, and the official Merchant Vessels Register of the United States Bureau of Navigation, the identity of 924 vessels was established, and the net tonnage, horsepower, and the year of building of each was recorded. 20 25 30 NET TONNAGE Figure 16. — Size distribution of vessels in the Puget Sound seine fleets at various in- tervals of development. The first histogram pictures the fleet shortly after the intro- duction of power; the second that of the exceptional year, 1917; the third the resum- ption of building after the post-war depression; and the fourth that of a recent year. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 735 The increase in larger vessels in 1912, which resulted from the development of the cape fishery, is very apparent. These larger craft had been underpowered and not particularly successful, and smaller vessels were more popular during the next few years. The two large vessels, built in 1909 and 1911, were not built as purse seiners but were converted in later years. The second abrupt size increase, beginning in 1916, was terminated by the depres- sion in 1921. Building was resumed in 1924, but construction never reached the pro- portions of earlier years, for the declining abundance of salmon discouraged sustained building. It was at this time, however, that Diesel-engined vessels began to appear in the fleet. The depression following 1929 sharply curtailed the number of vessels under construction, and a recession in size similar to that in the years following 1921 is evident. Table 14. — Relation of size and year of building for vessels in the Puget Sound purse-seine fleet from 1915 to 1934 ‘ i All vessels powered with gasoline engines prior to 1925, gasoline and Diesel (“oil”) powered vessels listed separately thereafter. 736 BULLETIN OF THE BUBEAU OF FISHERIES Table 14. — Relation of size and year of building for vessels in the Puget Sound purse-seine fleet from RELATION OF VESSEL SIZE TO EFFICIENCY Any comparison of the number of vessels fishing in recent years to the number in any early year is of little significance unless consideration is given to the effi- ciency of the individual vessels of these respective periods. Many reasons may be offered for variation in vessel efficiency, but the greater number of these may be either directly or indirectly ascribed to the size of the vessel itself. With the exception of two brief periods of unfavorable economic conditions, the size of the new vessels added to the fleet each year has been gradually increasing. The SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 737 newer vessels have been fitted out with better engines and equipment, and in recent years Diesel engines have been used almost exclusively by the larger vessels. These engines, allowing a far greater range of operation and greater economy than had been possible with gasoline engines, contributed much to the efficiency of the larger vessels. The average horsepower of engines has also gradually increased. For example, the average power of vessels in the 10-14 net-ton class has increased from 22.4 hp. in 1915 to 30.9 hp. in 1934. Larger vessels show a lesser increase in the case of gasoline engines, but the many Diesel engines are of much greater power. The maximum power of the largest vessels prior to 1918 was 55 hp., whereas vessels above 45 net tons now average 132.5 hp. The present averages for the 7 size-classes of vessels between 10 and 40 net tons are 36.5, 46.0, 56.6, 68.1, 88.1, 97.0, and 109.8 hp., respectively. The relatively greater power of the larger vessels undoubtedly adds to their efficiency. An important difference in earlier years existed in the size of the seine carried. In general, the larger seines were more efficient than the smaller ones and, since the size of the seine was necessarily limited by the space available for handling and stowing, it was generally proportional to the size of the vessel. Throughout the years the human factor, although difficult of measurement, has always been of great importance. The most successful fishermen have constantly built larger and better vessels, while the older, smaller craft ha ve usually been manned by less active men or by newcomers to the fishery. For these reasons the present analysis of vessel efficiency has been confined to a study of the relation of vessel size to size of catch. In order to facilitate vessel-catch comparisons, the fleets of all years from 1916 to 1934 have been divided into size classes of 5 net tons each. The annual numbers of vessels in each class, for the summer and fall fleets, are given in tables 12 and 13. Theoretically, any difference in efficiency between vessels of varying size should be reflected in a proportional difference in the average size of their catches. In order to determine such differences and to measure their degree, the average catches, over a considerable period of time, of vessels of different size classes were compared. Catch data used were from the years 1916-19, 1922-25, and 1928-34, in order to include the various building periods of the vessels and the fluctuations in fleet size. The size class of vessels from 10 to 14 net tons was selected as the unit of relationship since this class was well represented throughout the period of years covered. Direct comparisons of annual average catches could not be used because of the seasonal fluctuations in abundance of the various species of salmon, with the resultant influence that the presence of one species might have on the size of the catch of another. Therefore, data for different species were used for comparison during different parts of the fishing season. Sockeye catches were used for determining averages for the summer fishery of even years, pink-salmon catches for that of odd years, and coho and chum catches for the fall fishery in all years. Data for individual species were limited to the part of the season when they were sufficiently abundant to warrant fishing, and when other species were less numerous than the one sampled. Pink-salmon catches for most years were those from a period between July 29 and September 15. This period was shifted one week earlier in 1929 and one week later in 1933 in accordance with the time of appearance 738 BULLETIN OF THE BUREAU OF FISHERIES of the runs. Catches of coho salmon used were those taken during a period between September 16 and October 27; this period was decreased by one week in both 1929 and 1930. The period used for chum salmon was from October 13 to November 6, except in years when the season was extended beyond the latter date. The periods for sockeye salmon were necessarily more varied than those for other species because of greater fluctuations in the time of run. Catches used were generally from the period between July 15 and August 15, although these dates were shifted when necessary, for example, to the period from July 29-September 8, in 1930, when the run was very late. For each species the average delivery by vessels of each size class was deter- mined by dividing the total number of fish caught, during the period selected, by the total number of deliveries made. No class was used in which less than 5 vessels fished with a minimum of 10 catches. For years in which the 6 size classes between 10 and 39 net tons were represented, the average catches of the individual classes were determined as percentages of the average catch of all classes. For early years, when data were available for only the smaller classes, the average catches were determined as percentages of the average catch of the total class range represented. In order to make the data for early years comparable with those for later ones, the percentages of the individual size classes were proportionately reduced so that their sum was equal to the average sum of the percentages of an equal class range in the years when all classes were represented. The sums of the percentages in all years were divided by the number of years to determine the average percentage for each class, and the ratio of these averages to that of the class from 10 to 14 net tons was calculated for each species. These relative-efficiency ratios for each species, and for the average of all species, are presented in table 15. The sockeye salmon show the smallest and least con- sistent differences between vessel classes. Large catches of this species have fre- quently been made by vessels of all sizes fishing in certain limited areas on the Salmon Banks, near Lummi Island and at Point Roberts (fig. 2). Here peculiarities of winds and tides, or advantages of geographical location in relation to migration channels, have caused dense schooling for brief periods of time, and disproportional catches have been made by many vessels. Table 15. — Relative efficiency of Puget Sound purse-seine vessels 1 Vessel size in five-ton classes 1 Species 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40 and larger Sorkeve 3 _ 0. 66 1.00 0.99 1.46 1. 56 1. 43 1.55 1. 59 Pink i .92 1.00 1.27 1.64 1.85 2. 02 2. 33 2. 25 Coho_ .83 1. 00 1. 15 1.69 2. 19 2. 27 2. 37 2. 12 Chum .79 1. 00 1. 21 1.43 1.70 1.78 1.91 1.98 All species .80 1. 00 1. 16 1. 56 1.82 1.88 2. 08 1. 99 1 Proportion of the average annual catch of each species taken by each size class, calculated on basis of 10-14 class as unity. 2 Size in net tons, official register. 2 For even years only. • For odd years only. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 739 The ratios of the other species show a consistent increase with vessel size except in the group of vessels of 40 or more net tons. In this particular class, two species show increases and two decreases as compared with the next smaller class. The average ratios of all species were used as the final measure of relative vessel efficiency. The efficiency of boats of less than 5 net tons was arbitrarily set at 0.5, since sufficient data were not available from which a ratio for this class might be calculated. The average vessel efficiencies of the Puget Sound fleets from 1909-34, based upon these ratios, are shown graphically in the center section of fig. 15. The abrupt increase in the efficiency of the 1912 fleet is due to the construction of large vessels in that year. Efficiencies of the summer and fall fleets are quite similar, with the exception of the period after 1923. The divergence shown here is due to considerable variations in the number of small boats fishing. The general trend of the average efficiency is upward, with a slight decline in 1933 and 1934. It is evident that the fleets of recent years are, boat for boat, about twice as efficient as were those of 1909 and 1910. The total efficiency figures for the fleets from 1909-34 are presented in table 16. The same data are shown graphically in the bottom section of fig. 15. The great increase in efficiency in early years, as well as the considerable rise during recent years, is obvious. Judging from the actual number of licenses issued, as shown in the top section of the figure, there were 7 years between 1913 and 1921 in which the number of vessels fishing was greater than the average number fishing between 1931 and 1934. However, it is apparent from the figures of total vessel efficiency that the average of the last 4 years has been exceeded only once, in 1917, and approached closely in only 2 other years, 1915 and 1921. It is thus evident that, with the exception of the abnormal year 1917, the intensity of the purse-seine fishery on Puget Sound has been potentially greater during recent years than at any previous time in the history of the fishery. Table 16. — Relative efficiencies of Puget Sound purse-seine fleets, 1909-84 1 Year Summer fleet Fall fleet Unallo- cated Year Summer fleet Fall fleet 1809 74. 58 1922 79. 70 180. 54 1910 96.68 1923 172. 84 186. 32 1911 112. 06 113. 86 1924 72. 98 128. 22 1912 174. 22 1925 181. 75 217. 19 1913 263. 10 1926 91. 04 189. 31 1914 304. 30 1927 197. 79 260.91 1915... 343. 48 1928 154. 30 250. 07 1916 275. 54 312. 60 1929 284. 69 304. 62 1917 .. 509. 10 520. 62 1930 233.63 295. 58 1918 . 232.82 275. 94 1931 384. 17 384. 14 1919 _ 294. 02 319. 48 1932 312. 21 320. 93 1920 198. 36 231. 10 1933. 351.06 369. 32 1921 338. 30 348. 44 1934 333.83 344.80 1 For years 1909, 1912, 1913, and 1914, actual sizes of all boats unknown; efficiencies calculated from proportionate sizes of identified boats, which were 84, 70, 42, and 45 percent of the respective fleets of those years. 71941-38- 4 740 BULLETIN OF THE BUREAU OF FISHERIES SEASONAL OCCURRENCE OF EACH SPECIES PUGET SOUND FISHERY In certain areas several species of salmon may be present in considerable numbers at the same time, and during parts of the season a single purse-seine haul usually contains all five species. The seiners are able to make a certain amount of selection as to the species they wish to catch, however, by operating in different localities. In order to determine the seasonal progression of the various species in the fishery, the average daily delivery, by 7-day periods for each year from 1911— 34, was calculated for each of them ; data from all ves- sels of 10-39 net tons being used. The 7-d ay averages over the 24-year period were then calculated, and determined as percentages of their sum (see table 17 and fig. 17). The curves do not show the relative week ending abundance between spe- Figure 17. — Seasonal occurrence of the various species in the catch of Puget Sound cies but indicate for each purse seines. . . species the average pro- portion appearing in the catches of successive weeks during the fishing season. The pink-salmon curve represents occurrence only in odd-numbered years. Although there is considerable overlapping in the time when the various species appear, a distinct progression throughout the season is apparent, and the peaks of the runs of all species, except king salmon, occur at intervals of 3-4 weeks. These curves correspond closely to those from the trap fishery. The more prolonged periods of abundance of the various species indicated by these data may be attributed to the ability of the seiners to move with the fish, making their catches in whatever region that affords the greatest abundance at any particular part of the season. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 741 Table 17. — Seasonal occurrence in Puget Sound purse seines Week ending — Percentage occurrence Cumulative percentage occurrence Sockeye King Pink Coho Chum Sockeye King Pink Coho Chum 6.314 6.314 Julv 7 3. 685 3. 960 4. 181 3. 685 10. 280 4. 181 2. 534 5. 804 1. 709 6.219 16. 144 5. 890 July 21. 4. 629 6. 466 0. 805 3. 561 10. 848 22. 610 0. 805 9. 451 July 28 9. 070 5. 542 .847 2. 690 19.918 28. 152 1.652 12. 141 16. 352 5. 136 1. 189 2. 055 36. 270 33.288 2. 841 14. 196 Aug. 11..... 20. 914 6. 110 4.393 1.020 0. 102 57. 184 39. 398 7.234 15. 216 0. 102 Aug. 18 16. 481 4. 847 13. 066 1.738 1.437 73. 665 44. 245 20. 300 16. 954 1.539 Aug. 25 9. 199 6. 119 20. 484 3. 162 1. 536 82. 864 50. 364 40. 784 20. 116 3. 075 Sept. 1. 6. 142 6. 398 26. 069 o. 602 1.297 89. 006 56. 762 66. 853 25. 718 4.372 Sept. 8 7. 437 8. 737 15. 539 6. 396 1.391 96. 443 65. 499 82. 392 32. 114 5. 763 Sept. 16. .488 6. 432 10. 895 8. 282 2.017 96. 931 71.931 93. 287 40. 396 7. 780 Sept. 22 .816 5. 576 4. 843 9. 043 1.602 97. 847 77. 507 98. 130 49. 439 9.382 Sept. 29 .994 2. 449 .999 10. 834 2. 324 98. 841 79. 956 99. 129 60. 273 11. 706 Oct. C 1. 158 3. 212 .204 10. 286 4. 076 99. 999 83. 168 99. 333 70. 559 15. 782 Oct. 13 3. 873 .291 8. 564 8. 138 87. 041 99. 624 79. 123 23. 920 Oct. 20. __ 3. 449 .128 7. 147 14. 272 90. 490 99. 752 86. 270 38. 192 Oct. 27. . 2. 169 . 142 4. 522 18. 593 92. 659 99. 894 90. 792 56. 785 Nov. 3... 2. 508 . 106 3. 948 19. 487 95. 167 100. 000 94. 740 76. 272 Nov. 10 .. 2. 288 2. 279 12. 866 97. 455 97. 019 89. 138 2. 542 1.672 10. 860 99. 997 98. 691 99. 998 Nov. 24 . 1.311 100. 002 CAPE FISHERY The seasonal occurrence of the various species in the cape fishery has been de- termined in the same manner as that for Puget Sound. Adequate data, however, were not available prior to 1927. These data are presented in table 18. The sockeye and pink-salmon runs at the cape reach their seasonal peaks at about the same time as in the inside fishery (see fig. 17), but the former species is more concentrated at the tune of the peak of the run. The king salmon run is gen- erally similar to that of the inside fishery. The coho season at the cape differs con- siderably from that on Puget Sound. Large numbers of fish are taken during the first part of the season and the early cessation of fishing obscures what undoubtedly would be a fall run similar to that in Puget Sound waters. Occurrence of chum salmon has not been calculated because they form only a very minor part of the cape catches. Table 18. — Seasonal occurrence in cape purse seines Week ending— Percentage occurrence Cumulative percentage occurrence Sockeye King Pink Coho Sockeye King Pink Coho 0. 192 2. 526 13. 959 0.192 2. 526 13. 959 June 30— .635 7. 629 0. 805 7.319 .827 10. 155 0. 805 21. 278 July 7 .947 6. 134 1.206 8. S31 1.774 16. 289 2.011 30. 109 July 14 .939 8.296 2.511 5. 275 2.713 24. 585 4. 522 35. 384 July 21 2. 021 6. 114 4. 101 5. 828 4. 734 30. 699 8. 623 41.212 July 28 6. 048 9. 176 6. 003 6. 536 10. 782 39. 875 14. 626 47. 748 Aug. 4. 12. 496 7. 528 6. 540 5. 531 23. 278 47. 403 21. 166 53. 279 Aug. 11 30. 768 10. 075 12. 880 3. 799 54. 046 67. 478 34. 046 67. 078 Aug. 18 30. 368 12. 096 9.886 3. 526 84. 414 69. 574 43. 932 60.604 Aug. 25 8. 561 11.449 18. 723 4.988 92. 975 81.023 62. 655 65. 592 Sept. 1 3. 151 10. 570 22. 554 4. 906 96. 126 91.593 85. 209 70. 497 Sept.. 8 3.341 6. 265 14. 145 8. 884 99. 467 97. 858 99. 354 79. 381 Sept. 15 2. 142 .484 10. 930 100. 000 99. 838 90.311 .531 . 163 3. 038 99. 998 100. 001 93. 349 Sept. 29... 1.863 95. 212 Oct. 6 . 4. 788 100. 000 742 BULLETIN OP THE BUREAU OP FISHERIES FISHING SEASONS IN DIFFERENT DISTRICTS PUGET SOUND Purse seining on Puget Sound usually begins in the early summer in the region of the San Juan Islands, the greater number of vessels fishing on or near the Salmon Banks (see fig. 2). As the season progresses the vessels work farther inside to Rosario Strait, Lummi Island, and Point Roberts, and, especially in years when pink salmon are abundant, in Haro Strait. In even years there is a slack period between the summer and fall seasons in winch little fishing is done. In the odd years the pink-salmon run extends to the late summer closed period (see table 8). Fall fishing begins shortly after this slack period. In odd-numbered years some vessels may remain in the northern districts for the last of the pink-salmon run, but the remainder of the fleet will shift to the eastern part of the Strait of Juan de Fuca from Ediz Hook to Middle Point, and the southern shores of the San Juan Islands. A short time later most of the vessels will move to Admiralty Inlet. Much of the late fall fishing is in the inlets and passages of lower Puget Sound. In even years the fall fishery is similar, except that such vessels as fish during the slack period between the summer and fall fisheries usually operate in the lower part of the strait at an earlier date. Seining is carried on by Canadian vessels along the eastern shore of Vancouver Island and in seining areas 17-20 (see fig. 2), except that the portion of area 17 which is adjacent to the mouth of the Fraser River has, until recent years, been open to fishing only during the time of the pink and chum runs. The intensity of the seine fishery during different parts of the season is dependent largely upon the abundance of fish. However, the number of fish caught does not truly represent the effort expended by the fleet for fishing intensity may be very high, even though only moderate catches are made. The best measure of effort which may be determined from present records is the average number of deliveries made in a uniform period of time. During the greater part of the season buyers pick up fish at fairly regular intervals, and the number of deliveries made to them should closely approximate the intensity of the fishery. The number of deliveries in each week of odd- and even-numbered years from 1916-34, except 1920 and 1930, were calculated as percentages of the total number of deliveries made in each year. The year 1930 was omitted because of unusual differences in the time when the run of certain species occurred, and because of the curtailment of fishing in certain areas by administrative orders; 1920 was omitted because of inadequate data. The average percentage of the season’s deliveries, of the Puget Sound fleet, made in each week for both odd and even years were then determined, and are shown in the first two columns of table 19. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 743 Table 19. — Average proportion in each 7-day period of the total annual deliveries of the Puget Sound and cape seine fleets Week ending — Puget Sound fleet, 1916 to 1934 • Cape fleet, 1927 to 1934 Odd years Even years Even years weighted J Odd years Even years Even years weighted * 0.032 .021 0. 051 0. 041 June 23 . .615 2. 933 2.361 June 30 0. 020 0.005 0. 003 3. 705 10. 765 8. 665 July 7.... — .324 .143 .091 5. 934 9. 858 7. 935 July 14 - .906 .728 .463 7. 570 10. 432 8. 397 July 21 2. 462 2.882 1.833 13. 474 15. 043 12. 108 July 28. 3.836 6. 586 3. 553 9.193 11.546 9. 293 Aug. 4 4. 841 5.945 3. 781 16. 971 10. 795 8.689 Aug. 11.. 5. 470 5.583 3. 551 18. 926 4.942 3. 978 Aug. 18 6. 768 4. 873 3. 099 11. 812 9.011 7. 253 Aug. 25 10. 303 4. 153 2.641 4. 870 3. 084 2. 482 Sept. 1. 10. 461 2. 402 1. 528 4.616 2. 189 1.762 Sept. 8. 7. 425 1.504 .957 .473 4.328 3. 484 2. 431 1. 546 .235 3. 552 2.859 Sept. 22 4.324 3. 926 2.497 .667 .664 .534 Sept. 29 4.771 5.956 3. 788 .541 .247 . 199 Oct. 6.. - 5. 753 7. 70S 4. 902 .028 . 180 .145 Oct. 13. 6.767 8.846 5. 626 . 266 .038 .071 Oct. 20... 7. 367 10. 401 6.615 .050 .138 . Ill Oct. 27 7. 432 11.095 7. 056 .064 . 052 Nov. 3 _ 7. 077 10. 458 6. 651 .076 .061 3. 178 4. 449 2. 829 .013 .010 .416 .875 .099 .050 .032 Total 100. 000 99. 999 63. 598 99. 999 99. 999 80. 490 i 1920 and 1930 omitted. ‘ Percentages in even years weighted by ratio of average number of deliveries in even years to average number of deliveries in odd years. The week ending September 15 has been omitted from the odd years, since in ail years except 2, 1917 and 1919, a closed period has been enforced. The catches made during this week in these 2 years were not included when the percentages for these years were calculated. The last 3 days of the preceding week were also included in the closed period. The catches for this week have been estimated on the basis of the daily average for the 4 days of the week during which fishing was permitted, and the percentages calculated from the estimated figures. Because a similar closed period has been enforced in the last 2 even years, the percentages for the closed weeks during these years were estimated on the average of the same weeks of the remaining 6 even years in which this closure was not operative. Because the fleets in odd years have been larger than those in even years it was necessary, in order to show the proportionate intensity of the fishery, to reduce the even-year percentages in the same proportion that the average number of even-year deliveries bore to the average number of odd-year deliveries. From these weighted figures, appearing in the third column of table 19 and shown in the lower section of figure 18, it is immediately apparent that the increased intensity in odd years is con- fined largely to the summer fishery, and that the relative size of the summer and fall fisheries is reversed in odd and even years. In both odd and even years deliveries start shortly before July 1. In even years they increase rapidly to a peak during the last part of July and the early part of August, begin to decline about the middle of August, and by the first week in September have almost ceased. Shortly after this the fall fishery begins, with a gradual increase 744 BULLETIN OF THE BUREAU OF FISHERIES each week until a peak is reached in the last week of October. From this point the fishery declines abruptly. Fishing in odd years also increases during July, but, whereas the even years show a decline in August, the odd-year fishery continues to increase during that month to the highest point in the season. The slack season between summer and fall fishing follows, but is not so ac- centuated as in even years, even though the closed period terminates fishing entirely for a short time. After September 15 the fall fishing begins, increases to a peak about the middle of October, and then declines rapidly; the mode is more protracted than in even years. CAPE FLATTERY This fishery is generally carried on during the early summer, after which most of the vessels move to the inside waters where better protection from adverse weather is afforded, and where there is a greater concentration of fall-run- ning fish. The average proportion of deliveries made during each week of the season was. calculated for odd and even years in a manner similar to that for the Puget Sound fishery. These data are presented in the last three columns of table 18. The even-year percentages have again been weighted by the proportion of the average total numbers of even- and odd-year deliveries. These data are less smooth than those of the Puget Sound fishery because of the small number of years, 4 odd and 4 even, for which records are available. The curves of proportional intensity in odd and even years are presented in the upper section of figure 18. It will be noted that in both cases fishing begins during the latter part of June and is generally concluded early in September. In even years more than 60 percent of the deliveries have been made before the first of August, the catches being largely taken from the coho populations feeding on the banks. 9 23 JUNE 7 21 JULY 29 13 OCT. 27 10 NOV. 18 I 15 AUG. SEPT. WEEK ENDING Figure 18.— Fishing seasons of the cape and Puget Sound fleets in odd and even years. The early season at the cape, the influence of the large runs of pink salmon in odd years in both districts, and the summer and fall fisheries on Puget Sound are indi- cated. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 745 In the odd years a considerable number of catches are made throughout July, but the peak of the season is reached during the pink run in the month of August. Fishing terminates rather abruptly thereafter, the bulk of the vessels moving to the inside waters. RELATION OF FISHING INTENSITY TO SEASONAL OCCURRENCE Both seasonal occurrence and fishing intensity determine the proportion of the annual catch made at different intervals in the season. In order to portray the seasonal distribution of the catch, , , , , , . ODD YEARS the percentage taken in each 7-day period was cal- culated, for vessels of 10-39 net tons, for each year from 1916-34. The years 1920 and 1930 were omitted for reasons previously ex- plained. The average per- centages, by 7-day periods, were calculated for both odd and even years. These weekly percent- ages differ from the pre- viously calculated figures for fishing intensity in that they show the relative number of fish caught dur- ing uniform parts of the season, whereas the inten- sity figures represent the fishing effort during similar periods. Since it is also impor- tant to know the contribu- tion of the individual species, their proportionate representation in the week- ly catches of each year from 1916-34 were calcu- lated and the average week- Figure 19.— Seasonal distribution of the catch of Puget Sound purse seines and the pro- portional contribution of the various species. The striking difference of the impor- tance of the summer and fall fisheries in odd and even years is readily apparent. 8 22 SEPT. WEEK ENDING ly proportions for odd and even years determined. Corrections were made for closed periods in a manner similar to that described in the calculation of seasonal fishing intensity. The average proportion of the catch taken by weeks, and the average representation of the individual species are presented, for both odd and even years, in table 20 and in figure 19. 746 BULLETIN OF THE BUREAU OF FISHERIES Table 20. — Average proportion of each species in the weekly catch of Puget Sound purse seines and percentage occurrence of total catch, 191 6-3 4 July 7 July 14— July 21.— July 28— Aug. 4 Aug. 11— . Aug. 18— Aug. 25... Sept. 1— . Sept. 8... Sept. 15 '. Sept. 22- Sept. 29.. Oct. 6— Oct. 13... Oct. 20— Oct. 27... Nov. 3... Nov. 10- No v. 17- Nov. 24. . Week ending— Odd years Percentage total catch Sockeye Pink Coho Chum King 71.766 19. 500 56. 546 34. 241 57. 669 37. 819 58. 108 35. 618 59. 657 44. 138 36. 438 58. 393 19. 450 77. 595 6. 588 92. 098 2.790 95. 650 2.013 95. 096 7. 018 0. 577 5. 767 1. 761 3. 069 .250 3. 165 1.833 4. 082 . 163 3. 767 .450 2. 132 .149 .965 .091 1.116 .119 1.769 .650 1.140 0. 091 1. 685 . 196 1.203 .674 1.275 1. 390 .960 1.918 .953 3. 187 .675 8. 007 .258 20. 568 .324 27. 791 .472 17. 446 1.673 58. 246 1. 142 14. 147 .371 2.811 . 138 .882 .002 .076 . 054 .006 1. 179 .721 35. 109 4. 62.718 21. 57. 527 39. 39. 625 59. 24. 294 75. 15.315 84. 16. 236 82. 11.909 87. 6. 210 93. 11. 584 88. 540 .432 447 .546 046 .245 077 .278 298 .329 447 .184 408 .172 219 .151 440 .349 331 .085 2. 548 1. 034 1.580 2. 631 3.608 3.081 2. 931 1. 191 . 102 .026 July 7— July 14.. Juiy 21.. July 28.. Aug. 4.. Aug. 11. Aug. 18. Aug. 25. Sept. 1— Sept. 8.. Sept. 15. Sept. 22- Sept. 29. Oct. 6... Oct. 13- Oct. 20.. Oct. 27.. Nov. 3.. Nov. 10. Nov. 17. Nov. 24. Even years » 78. 415 1. 453 19. 094 2. 038 .011 77. 125 1. 734 15. 074 .328 5.739 .057 75. 982 1.863 16. 738 .288 5. 388 .682 85. 061 3. 669 6. 949 .028 4. 264 2.289 86. 776 5. 072 4. 746 .082 3. 324 3. 070 87. 745 3. 527 4. 426 .181 4. 121 3. 197 71. 489 3. 964 11.527 4.215 8.805 4. 174 37. 173 13.813 26. 204 17. 536 5. 274 5. 753 31. 253 2. 252 42. 864 18.214 5.416 2. 566 13. 785 1. 251 56. 441 22. 021 6. 503 1.316 .217 1.410 75.811 20. 243 2.319 1.035 .296 .392 78. 907 18. 978 1.428 2.081 .397 .072 67. 506 31. 576 .449 4. 920 1.086 .155 55. 412 43. 137 .209 6. 114 .002 .015 36. 479 63. 319 .184 9. 529 .033 .035 23. 647 76. 066 .219 15. 052 .011 .028 12.412 87. 285 .265 18. 367 .010 9. 996 89. 816 .178 13. 625 .012 7. 475 91.544 . 970 5. 423 12. 802 87. 149 .049 . 680 2. 807 97. 126 .067 .058 Omitted because of closed period. > 1920 and 1930 omitted. The curves for even years are presented in the lower section of the figure, and those for odd years in the upper section. The curves for kings were omitted, since the highest point in any one week in even years was less than 0.4 percent, and in odd years was less than 0.1 percent. The scale for odd years was increased so that the proportion of the run afforded by all species other than pinks should be equal in odd- and even-numbered years. Because of the extreme peak the odd-year curve was truncated, hence the percentage occurrence of the total catch and the proportion represented by pink salmon are not shown for the weeks from August 18 to September 8. These curves vary most from those of fishing effort in the more extreme differ- ences between the summer and fall fisheries. It is evident that the number of fish per delivery is much greater during the height of the run of pink salmon in odd years and during that of chum salmon in even years. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 747 It is apparent that the late summer fishing for pink salmon in odd years in the northern districts of the sound has caused some extension of sockeye catches, and this is further demonstrated by the absence of chum salmon in the catches. In even years, although the summer fishery begins to decline much earlier, such vessels as are fishing are operating in districts where the early chum runs are found, and increased catches of chums appear more than a month earlier than in odd years. The predominance of chum salmon in the fall fishery of even years indicates a greater effort to take this species when the lack of pink salmon has resulted in a poor season for the seiners. The peak of the fall fishery is reached during the week ending October 27. In odd years the peak of the total catch is reached a week earlier, shortly after the coho run has reached its maximum, and the curve begins to decline while chums are still abundant. RELATIVE IMPORTANCE OF EACH SPECIES PUGET SOUND The sum of the Puget Sound purse-seine catches from 1917-34 was 64,978,888 salmon, of which 37,559,326 were pink salmon, 12,653,382 were chum salmon, 9,121,238 were sockeye, 5,383,438 were coho, and 261,504 were king salmon. Large and small runs of pink salmon appear in alternate years. In years of abundance, odd years, they have averaged over 4 million fish a year and have provided approxi- mately 75 percent of the catch, in the even years they have averaged little more than 6.000 a year and have furnished less than 1 percent of the catch. Their average for all years is 37.44 percent (see table 21). The average chum-salmon catch over 18 years has been approximately 700,000 fish per year. Seven of the 9 even-year totals are considerably above tins figure, reflecting the more intense even-year fishery. During this period the average pro- portion of chums in the annual catches was 32.07 percent. Although over 9 million sockeyes have been taken during this period, nearly 6 million were caught during ODly 3 years; almost 2 million in 1917, nearly 2){ million in 1930, and over 1% million in 1934. The remaining 15 years averaged approximately 226.000 fish. The annual average for sockeyes was 15.63 percent over the 18-year period. The catches of coho salmon show smaller fluctuations than do those of the above species; then- average has been approximately 300,000 fish per year during this period. They averaged 14.16 percent of the annual catches. King salmon are a negligible factor in the purse-seine catches, averaging less than 15.000 fish per year. This species has provided an average of only 0.7 percent of the total catches during the 18-year period. 748 BULLETIN OF THE BUREAU OF FISHERIES Table 21. — Pro-portion of various species in total annual catches of Puget Sound purse seines, 1917-34 Year Sockeye King Pink Coho Chum Total eaten > 1917. 14.31 0. 29 62. 99 3.71 18.70 11, 804, 026 1918 2. 35 2. 13 .26 32.35 62.91 1, 376, 757 1919 4. 10 .93 47. 25 10. 64 37. 08 4, 349, 421 1920 3. 05 .66 .03 22. 82 73. 44 775, 421 192l_ - 5.06 .39 78. 18 11.86 4.51 3, 079, 015 1922 - 9.88 .79 .43 45.51 43. 39 875, 233 1923__ 4. 39 .12 82. 10 4.80 8. 59 4, 042, 288 1924_ 10. 35 .52 1. 11 17. 91 70. 10 1, 127, 020 1925... 3. 32 .21 83. 85 5. 38 7.25 5, 656, 515 1926... 13.69 .47 .36 28. 33 57. 15 1,158, 848 1927 11.12 .43 78.64 5.04 4. 76 4, 549, 007 1928 6. 65 2.08 1.53 27. 06 62. 68 1, 164, 682 1929 6. 79 .21 72. 72 4. 19 16.09 6, 359, 144 1930... 81. 24 .61 .80 4.25 13. 10 3, 567, 442 1931... 5.28 .39 81. 44 4. 23 8. 66 5, 468, 739 1932 24. 43 1.32 .36 17.40 56. 49 1,716,772 1933 9. 93 .34 81. 64 2. 77 5. 32 5, 531, 318 1934. 65. 43 .61 .30 6.61 27. 05 2, 367, 240 Average 15. 63 .70 37. 44 14. 16 32. 07 1 Total catch of all species in numbers of fish. Although approximately 58 percent of the total number of fish caught during this period have been pink salmon, they have been abundant only in odd-numbered years. In the alternate years chums have provided approximately half the catch, with cohos and sockeye next in importance. CAPE FLATTERY The contributions of various species to the purse-seine catch at the cape differ considerably from those on Puget Sound. Records are not available for the numbers of seine-caught fish taken at the cape before the period from 1927-34, during which 14,166,769 salmon were caught. Of these, 10,395,194 were pink salmon, 2,305,290 were cohoes, 1,348,553 were sockeye, 69,433 were kings, and 48,299 were chums. Pink salmon have averaged 84.56 percent of the catch in odd years and 8.53 percent in even years. Their average for all years is 46.54 percent. During the period for which accurate figures are available, more than 73 percent of the total number of fish landed at the cape have been pink salmon (see table 22). Table 22.- — Proportion of each species in the total annual catches of Cape Flattery purse seines, 1927-34 Year Sockeye Pink Coho Chum King Total catch 1 1927. 2. 10 89. 66 7. 92 0. 03 0.29 2, 382. 838 1928 6.81 23.73 67.48 .57 1.40 290, 222 1929 2. 60 85. 96 11.01 .07 .35 3, 924, 375 1930- 23.49 6. 85 66. 33 1. 89 1. 43 614, 170 1931 4.97 89. 37 5.16 . 17 .32 4, 367, 412 1932. 5.44 1.66 87. 60 3.38 1.92 359, 900 1933. 10. 50 73. 25 15. 15 .71 .40 1, 153, 429 1934. 62. 80 1. 87 34. 01 .35 .98 1, 074, 423 14. 84 46. 54 36. 83 .90 .89 1 Total catch of all species in number of fish. Coho salmon are next in importance, furnishing the greater part of the early- season catch in all years. During the even years they averaged 63.86 percent of the catch, and during the odd years, 9.81 percent. Their all-year average is 36.83 percent. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 749 The sockeyes show the same heavy catches in 1930 and 1934 noted in the Puget Sound fishery, providing 23.49 percent and 62.80 percent of the catch, respectively. Their average for the even years is 24.64 percent, for the odd years 5.04 percent, and over the 8-year period 14.84 percent. King salmon, although taken throughout the season, provide only a very small proportion of the cape landings. The catch figuies are somewhat reduced, however, by the practice of bujfing small kings as pink salmon, and occasionally as cohoes. The average in the even years is 1.43 percent, in odd years 0.34 percent, and over the 8-year period was 0.89 percent. Chum salmon are caught infrequently in the offshore waters. Their average is 1.55 percent in even years, 0.25 percent in odd years, and was 0.90 percent over the 8-year period. THE TROLL FISHERY By George B. Kelez Fishing with hook and line was engaged in by natives of the region long before commercial salmon fishing began, but this gear never became of significance until the introduction of power boats. As was true of the purse seine, little change has taken place in the gear itself, whereas a considerable improvement has been made in the boats from which it is fished. Although individuals of all five species of salmon are landed occasionally, only the coho and king salmon are readily taken by trolling. The early Indian gear consisted of lines twisted from bark or animal sinews, a stone weight, and a hook of bone or of wood with a bone point. Although “spoons” (lures) of shell were in use, the principal Indian fishery was with baited hooks, herring being chiefly used for this purpose. According to Rathbun (1899) the fishing season at Neah Bay was during the months of June, July, and August. Another interesting but little-known type of native gear, which developed from the trolling line, is shown in fig. 20. It consists essentially of a bladder float to which is attached a line of twisted sinew suspending a stone weight. A second line is fastened to the weight, and the free end is attached to a shank of whalebone bearing a double hook of bone lashed with bark. As many as thirty of these units were attached together, each hook was baited with a whole herring, and the string was drifted from a canoe. Both types of gear were fished close to the surface, and the principal catch was coho salmon, preferred by the natives because of its suitability for drying. DEVELOPMENT OF THE FISHERY For many years commercial trolling was of little importance. Collins (1892) did not include it among the commercial fishing methods listed for the region, but stated : The Indians employ trolling hooks and spears in the Sound and small streams tributary thereto, and parties fishing for pleasure also use spoon hooks and trolling lines. Also, the Indians at Neah Bay use trolling lines, and in 1888 took 7,000 pounds of salmon valued at $140. A much larger catch could, no doubt, be made at this place. . . . Rathbun (1899) included trolling gear among commercial methods, but stated that its use was restricted both as to locality and number of men employed, and that 750 BULLETIN OF THE BUREAU OF FISHERIES it was still chiefly fished by Indians. The principal catch was king and coho salmon. Kings were fished from November to February, and sometimes to April, in the Gulf of Georgia, both in the region of Nanaimo and off the mouth of the Fraser River (see fig. 2). They were also taken in the vicinity of Victoria, in the San Juan Islands, off Port Townsend, in the upper part of Admiralty Inlet, and in Hood Canal. Cohos were taken in smaller numbers, although good catches were made in Boundary Bay and in the waters of lower Puget Sound. Rathbun also stated that the catch of trolling gear was much less than that of the gill nets in the region. Fishing was conducted from canoes or skiffs, and by one or not more than two men to a boat. Spoons and hooks baited with herring were in general use. The introduction of power, which had almost as great an effect on trolling as it did on fishing with purse seines, eliminated the rowing or paddling of the skiff or canoe, and thus greatly reduced the labor of fishing. The fishermen were now able to cover greater distances, were less subject to the force of the tides, and could attend to more lines. Larger, more able boats soon came into use, and the fishing area was extended over the entire inner waters of the region, while the size of the catch of the boats was increased remarkably. By 1908 the trollers were fishing well out into the Strait of Juan de Fuca, and by 1911 they were operating on the open ocean in the vicinity of Cape Flattery. With the development of the offshore fishery, still larger boats appeared in the trolling fleet. These carried a small cabin which housed the engine and provided cramped quarters for the crew when at anchor. Although the greater part of the trolling boats remained at some base, such as Neah Bay, and fished during the early hours of the day, the larger boats, which were of 30-35 feet in length, made trips of 2 or 3 days duration. These were designated as “overnight” boats, in contrast to the majoiity, which were “day” boats. The gear fished by these boats now consisted of as many as six lines, often carrying from two to three spoons and hooks each. The lines were suspended from poles of varying lengths hung outboard over the sides of the boat, one pair usually at the bow and one amidships. Metal spoons were almost universally used, but herring bait was still favored by a few single-liners. The power gurdy, which was introduced in 1918, was a multiple reel, driven off the motor, by means of which the lines could be hauled in whenever a fish was hooked. This greatly increased the speed of handling the lines. Figure 21 illustrates the mounting of this device, together with the lead-in blocks by means of which the lines are brought from the poles to the gurdy reels. The fish hatch is forward of the gurdy, and the cockpit, from which the boat is steered and the lines handled while fishing, is immediately aft of it. With the exception of the adoption of the Diesel engine, giving greater cruising radius and more economical operation, there has been little further change to the present time. IMPORTANCE It is difficult to obtain accurate records as to the number of trollers operating in the region during most of the past years. Some of these boats fished entirely on the high seas and were not licensed by the State of Washington, while others roved from Mon- terey Bay in California to Southeastern Alaska, fishing for varying periods along the coast according to the abundance of the fish. Figure 20. — Modified floating liook-and-line gear used for coho salmon by the natives at Neali Bay before white fishermen operated in that district. The bone hook was baited with a whole herring. From the collection of Captain T. E. Eggers. Figure 21. — Stern view of trolling boat. Note the two hand-operated gurdies and the lead-in blocks directly over them and on both gunwales. The two main poles may be seen at, the sides of the mast.. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 751 Records of licenses issued between 1917 and 1934 by the Department of Fisheries of the State of Washington for the Puget Sound district, which also embraces the territorial waters in the vicinity of Neah Bay, are presented in table 23. Boats fishing exclusively offshore did not have licenses prior to 1917 as none were issued. Gilbert (1913) reported 250 trailers in the cape region in 1911 and stated that this was an “unprecedented number.” He estimated more than 400 there the following season. Smith and Kincaid (1920) reported more than 500 boats fishing at the cape in 1918. We may assume that the fishery was of little importance prior to about 1910, and that the number of boats increased thereafter to a maximum in 1919, the last 3 years of tins period being included in table 25. There was a marked decrease in licenses during the period of economic depression following 1921, and again in the similar period after 1931. Table 23. — Puget Sound trolling licenses, 1917-84. Year Number Year Number Year Number 1917 782 1923 221 1929 656 1918.... 982 1924... 374 1930 784 1919.... 1,032 611 1925 438 1931 599 1920.. 1926 684 1932 259 1921 415 1927 820 1933.. 220 1922 165 1928 672 1934 478 During recent years practically all the boats have fished in the region of the cape, some as far as Forty-mile (La Perouse) Bank. A few of those fishing in Puget Sound operate hi the San Juan Islands, but most of them fish the waters south and east of Point Wilson. A large fleet of Canadian trailers operates off the west coast of Van- couver Island, and a small fleet fishes in the upper part of the Gulf of Georgia for coho salmon. Some boats work off the southeastern part of Vancouver Island for kings. The catches of the cape and Puget Sound fleets for recent years may be found in the sections on coho and king salmon. For the 8-year period from 1927-34, Puget Sound trailers took 104,692 cohos and 18,285 kings. During the same period, the cape fleet took 2,411,312 cohos and 1,545,178 kings. In addition, a few thousand pink salmon are taken at the cape in years of abundance, and occasional catches of the other species are made. SEASONAL OCCURRENCE OF COHOS AND KINGS Species other than coho or king appear so infrequently in trailers’ catches that their occurrence may be disregarded. In the early part of the season kings are taken almost exclusively, but after the first of May both species appear in most of the catches. Seasonal occurrence is not so well defined in the troll catches as in other gear, for land- ings at any station, such as Neah Bay, may contain fish caught at a considerable distance from the landing point. In the early season the trailers fish longer, more heavily weighted fines, thus increasing their chance of taking the deeper-swimming king salmon. In the latter part of the season they fish closer to the surface in order to take cohos. Many fishermen shift during the fall from the plain metal spoons used in early summer for kings to ones which ha ve been painted red on one side and which 752 BULLETIN OF THE BUREAU OF FISHERIES seem to be more efficacious for cohos. For these reasons the occurrence of the species in the troll catches do not reflect their relative runs as accurately as do those from less selective gear. Catches were available for from 174-261 trolling vessels landing at Neah Bay during the years from 1922-28. Because of the extreme difficulty in identifying the boats, no attempt was made to treat their catches individually. For both kings and cohos the average daily delivery per boat during each week of the season was calcu- lated for the individual years, and from these data the averages over the 7-year period were calculated. These were then determined as percentages throughout the season. The percentage occurrence of both species by weeks is presented in table 24. Table 24. — Seasonal occurrence in cape trolling gear Week ending— Percentage occur- rence Cumulative per- centage occurrence Week ending — Percentage occur- rence Cumulative per- centage occurrence King Coho King Coho King Coho King Coho 1. 523 1.523 July 28 4. 960 6. 255 61. 046 45. 604 3. eoo 5. 123 Aug. 4 5. 036 4.547 66. 082 50. 151 4. 787 1. 846 9. 910 1.846 Aug. 11 4. 494 5. 212 70. 576 55. 363 May 12 3. 320 .761 13. 230 2. 607 Aug. 18 4.846 6. 585 75. 422 61.948 2. 905 2.867 16. 135 5. 474 Aug. 25 4. 477 5.840 79. 899 67. 788 4. 543 2.072 20. 678 7. 546 Sept. 1 2. 932 4. 608 82. 831 72. 396 6. 519 3. 139 27. 197 10. 685 Sept. 8- 2.448 5. 515 85. 279 77.911 4. 152 2. 182 31. 349 12. 867 Sept. 15 4. 304 6. 227 89. 583 84. 138 3.669 3. 186 35. 018 16. 053 Sept. 22 3. 493 6.713 93. 076 90. 851 4. 114 3.861 39.132 19. 914 Sept. 29 3.882 3.427 96. 958 94. 278 3.347 4. 717 42. 479 24. 631 Oct. 6 2. 046 2. 450 99. 004 96. 728 July 7 4. 184 3.813 46. 663 28. 444 Oct. 13. .949 2. 038 99. 953 98. 766 4. 859 5. 520 51. 522 33. 964 Oct. 20 .047 .740 100. 000 99. 506 4. 564 6.385 56. 086 40. 349 Oct. 27 .494 100. 000 There is a short period of heavy catches of king salmon in early May, followed by the main period of occurrence lasting from June to the latter part of August. There is a third small run in the latter part of September which decreases immediately after the first week in October. The coho catches build up slowly during a period of about two months prior to the middle of July, remain, with some fluctuations, at that level until the third week in September, and decrease thereafter to the last week in October. A comparison of the cumulative percentage occurrence figures from trolling gear (see table 24) with those for cohos and kings in cape purse seines (see table 18) indicates some of the differences in these two fisheries. The first troll-caught kings are taken in the week of April 15-21, 25 percent of the catch is made by the end of May, 50 percent by July 14, 75 percent by August 18, and 100 percent by October 20. Seine-caught kings do not appear before the middle of June, 25 percent are taken by July 14, 50 percent by August 6, 75 percent by August 22, and 100 percent by September 15. The trollers will have been operating for about two months before the seiners begin, and slightly more than 50 percent of the troll catch has been made by the time that 25 percent of the seine fish are landed. The two curves cross during the latter part of August, and the seine season is over before 90 percent of the troll-caught fish are landed. Trollers begin landing cohos about the first of May, 25 percent of the catch is taken by the first of July, 50 percent during the first week in August, 75 percent by the first week in September, and the season ends during the latter part of October. The seiners begin fishing cohos about the middle of June, and 25 percent of the catch is SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 753 made by July 4, 50 percent by August 3, 75 percent by September 4, and 100 percent by the first week in October. It will be noted that the differences in time of the king catches are due mainly to the length of season fished, and that there is little similarit}1- in the time of the 25th percentiles. In the case of cohos, however, the 25th, 50th, and 75th percentiles of both types of gear coincide. The heavy catches of immature cohos by the seiners allow them to take the first quarter of their catch in some two weeks; the trollers require approximately two months to take 25 percent of their catch, since they are fishing primarily for king salmon during the early part of the season. During the remainder of the coho rim, the curves of both types of gear are very similar. SPORT FISHING King and coho salmon have provided popular sport fishing in the region for many years. With the exception of fly-fishing in a few restricted localities, this fishery has been carried on entirely by trolling, or by modifications of this gear, hence catches of other species of salmon are rare. Collins (1892) referred to trolling for salmon as a recreation, saying: In autumn, when salmon are most numerous in the Sound, Seattle Bay is literally covered with pleasure boats for days in succession. Rathbun (1899) mentions sport trolling for king and coho, either with spoons or bait, and also refers to good fishing in the spring for king salmon in the pools of such rivers as the Nanaimo and Cowichan. At the present time the Campbell River is best known for fly-fishing for kings, and many cohos are taken by this method at the mouth of the Cowichan River. Throughout the southern part of the region the greater part of the spring and summer king-salmon catches, and a considerable number of coho catches, are made with “spinning” gear. This is a highly specialized development of trolling, and consists of fishing from an anchored boat with a rod, light line, and small hook. The bait is a spinner which is usually cut from fresh herring. In use, the line is cast from the boat, allowed to sink almost to the bottom, and then recovered by drawing it in with successive pulls, allowing the recovered line to coil in the bottom of the boat. The largest longs are landed in a few favorable places by7 trolling with “plugs” some- what similar to those used in bass fishing. The bulk of the sport catches on the sound consist of coho salmon, and these are most frequently taken by trolling with spoons, although many fishermen use cut herring or candlefish. Mature cohos are taken in the fall on copper spoons which are nickelplated on one side. Although sportsmen fish in nearly all the inner waters of the region and as far out in the Strait of Juan de Fuca as Port Angeles and Victoria, the most heavily fished waters are in the region of Whidbey Island and the lower part of Puget Sound. Many resorts located in this region have 50 or more boats available for rental, and several thousand sportsmen fish from early spring to fall. Fishing is conducted in places such as Elliot Bay at Seattle throughout the entire year. This sport has become increasingly popular in recent years, and the outfitting of fishermen, together with the rental of boats and sleeping quarters, may now be ranked as one of the fishing industries of the region. 754 BULLETIN OF THE BUREAU OF FISHERIES SOCKEYE SALMON By Geoege A. Rotjnsefell INTRODUCTION The Fraser River, with its numerous tributary streams and chains of lakes, is potentially the best sockeye river in the world. Over a period of 24 years, six genera- tions, from 1894-1917, it produced 195,740,000 sockeyes; an annual average of 8,160,- 000. The Kvichak River, flowing into Bristol Bay, ranked next, producing, during the same period, 155,330,000 sockeyes, an annual average of 6,470,000. The pro- duction of the Nushagak River, also flowing into Bristol Bay, was 78,010,000, with an annual average of 3,250,000. The river ranking fourth in North America was the Karluk, on Kodiak Island, with a production of 47,700,000 fish and an annual average of 1,990,000. This comparison cannot be made over a longer period of time because in the earlier years none of these rivers were fished with sufficient intensity for the catch to be any measure of the size of the run, and in later years the Fraser River runs were so depleted by the blocking of the river at Hell’s Gate in 1913 and 1914, and the intense fishing of the War years, that the catches have no longer given any measure of the productive capacity of the river. From an annual average catch of 8,160,000 sockeyes for the 24-year period from 1894-1917, the production of the Fraser River, for the 17-year period from 1918-34, has fallen to an annual average of 1,830,000. The consequent annual loss to the fishermen of several millions of sockeye, through the failure of sufficient adult salmon to reach the spawning grounds, is a waste of the potential capacity of this great river. Such a waste of a natural resource, although less obvious, is just as real as the needless burning of thousands of acres of forest. GENERAL LIFE HISTORY SPAWNING The sockeye, unlike the other four species of Pacific salmon in this region, rarely spawns elsewhere than in a tributary of a lake, or in gravel provided with spring seepage within a lake. Sockeyes spawn in one or another of the vast Fraser River lake systems from August until December, spawning, in general, being earlier in the Nechako River and Stuart-Trembleur-Takla lake systems and later below Hell’s Gate and in the tributaries of the Thompson River, although a lake system may have both an early and a late run of sockeyes during the same season, forming two spawning peaks. The fry, after absorption of the yolksac, wriggle free from the gravel, usually during the spring and summer months. Those that are hatched in the tributaries of the lakes find their way downstream into the lakes. In some localities a considerable portion of the adult run may occasionally spawn in the sluggish outlet stream of a lake. Whether or not the fry, upon hatching, ascend the slow-moving stream into the lake is not known, but it would appear probable that such may be the case. Young sockeyes spend varying lengths of time in lakes before descending to the sea. In the Fraser River the majority of the young migrate in their second year. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 755 From scale reading (Clemens, 1934) it appears that approximately 91 percent of the returning adults had left the lakes in their second year, 5 percent in their third year, and 4 percent in their first year. Foerster (1929b, 1934) shows that from the 1925 spawning at Cultus Lake, 6.2 percent of the migrants were in their first year (fry), 92.9 percent were in their second year (yearlings) and 0.9 percent were in their third year. AGE AT MATURITY The majority of the sockeyes of this region reach maturity and return from the ocean to their spawning grounds in their fourth summer. From 1920-33, inclusive, a period of 14 years, the ages of the sockeyes taken by the traps near Sooke, on Van- couver Island, (Clemens, 1934) have averaged as follows: 3-year-olds, 3.2 percent; 4-year-olds, 76.4 percent; 5-year-olds, 19.6 percent; and 6-year-olds, 0.6 percent. Since the proportion of the fish at each age varied considerably in different parts of the season, these figures are only an approximation of the number of fish at each age composing the catch, but they show the preponderance of 4-year-olds. The cycle, or generation of sockeyes occurring quadrennially in the year following leap year (1909, 1913, 1917, etc.) was, as is shown below, tremendously abundant up to 1913, and fairly abundant in 1917, but much less abundant in 1921 and later years. Cilbert (1914) showed that the sockeyes running in 1913 were 99.5 percent 4-year-olds. In 1917 they were 94 percent 4-year-olds. In the past 4 years of this cycle (Clemens 1934) they have averaged but 77.4 percent 4-year-olds. There is reason to believe that the change in the proportion of sockeyes 4 and 5 years of age is caused, at least in part, by changes in the proportion of the runs coming from different lake systems. This was pointed out by Gilbert (1917), who said that the runs to the various tributaries did not show the same proportions of 4- and 5-year- olds as did the samples of the run as a whole, the 5-year-olds being especially prominent in many localities below Hell’s Gate. During May and early June a run of sockeyes occurs that contains a large propor- tion of 5-year-old fish. This run is too small to be of any importance, as can readily be seen from the trap curve of seasonal occurrence (fig. 11), and is distinguished by the small size of the individuals, the lack of oil, and light-colored flesh. Since these fish lose most of their color in the canning process, they are usually sold as cohos. Some of these very early sockeyes may be Skagit River fish, which are taken in late June along West Beach, but the larger part are probably bound for the Fraser River, as the traps in Rosario Strait, Lummi Island, Boundary Bay, and Point Roberts Areas all take them, and in about the same amount as the traps in the Salmon Banks and South Lopez Areas. A third group of sockeyes that merit attention are the “grilse.” These fish, usually males, have migrated to the ocean at the usual time, in the second year, but have matured precociously, returning after only 1 year in the sea, instead of the customary 2 years. On the years that preceded the former big years grilse were always numerous. Gilbert (1913, 1916) estimated them at 21.5 percent of the run in 1912, and 10 percent in 1916. The presence of these small sockeyes on such years was well- known to the cannerymen. On years preceding the off years the percentage of grilse in the run was quite small ; very often negligible. 71941— 3S 5 756 BULLETIN OF THE BUREAU OF FISHERIES SOCKEYE RIVERS OF THE REGION OUTER COAST STREAMS In order to determine whether or not one is justified in regarding practically all of the sockeyes caught on Swiftsure Bank, in Puget Sound, and in the Gulf of Georgia as originating in the Fraser River, it has seemed advisable to show the extent of the runs to other sockeye streams in the region and to discuss the probability of any of these sockeyes being included in the records as Fraser River fish. The largest run of sockeyes on the outer coast, immediately south of Puget Sound, is that of the Quinault River, which enters the ocean 65 miles south of Cape Flattery. The runs appear to fluctuate from about 50,000 to 500,000 sockeyes, as shown in table 25. The Indians commence catching a few sockeyes at the mouth of the river as early as January, the bulk of the run reaching Quinault Lake between May 20 and July 7, and the mode occurring in the week ending June 9. In the 1922-24 runs, for which accurate weir counts by the Bureau of Fisheries are available, 77 percent had entered the lake by June 30. Of the remaining 23 percent there is reason to suppose that most of them were already in the river by this date, as fishing at the mouth of the river is usually practically over by July 1. The sockeyes run considerably later, however, on Swiftsure Bank, the seiners taking almost none before July 1 and the season not reaching its height until early in August. Table 25. — Quinault River sockeye ( blueback ) run, 1908-34 Year Pack in cases i Actual catch Escape- ment Year Pack in cases 1 Actual catch Escape- ment 1908 » 75, 000 1921 2,590 19, 213 10, 454 8, 473 3,313 1,729 5,260 2, 000 4, 449 21, 536 8, 476 14, 263 6, 754 4, 960 * 20, 000 248, 935 174, 602 136, 774 19, 395 1909 1922- 265, 649 138, 148 104, 571 54, 000 1910. 4, 350 2,031 4, 700 712 12, 274 24, 484 10, 315 4, 608 2, 490 1,244 235 1923 1911 1924 3912- 1925 1913 1926 1914 1927 1915 3 355, 007 1928— 1936 1929 1917 1930 1918 1931 1919 14, 947 1932 1920 1933 1934 > 1910-28 from Cobb (1930, pp. 559-500), 1929-35 from Pacific Fisherman. > New York Sun, July 19, 1908. It also states: “This is 27,000 more fish than have ever been caught in any previous season.” s From Cobb (1930, p. 426). < Only 11,786 counted, balance estimated. The Ozette River (fig. 1) empties into the ocean 12 miles south of Cape Flattery. The Bureau of Fisheries placed a weir across this river in 1926, discovering that the run, which is nearly over by July 1, amounted to only a few thousand fish. The Hobarton River empties into Nitinat Inlet, which reaches the ocean just north of the entrance to the Strait of Juan de Fuca. The Nitinat Inlet sockeye catch is given in the Fisheries Reports of the Dominion of Canada as follows: 12,000 in 1928, 20,130 in 1930, 16,487 in 1931, and 56,000 in 1932. Barclay Sound, (fig. 1) a little farther to the north, has two runs of sockeyes, one ascending the Anderson River, which is 18 miles from Cape Beale, and the other the Somass River at the head of Alberni Canal, a northeasterly extension of Barclay Sound that cuts deeply into Vancouver Island. SALMON AND SALMON FISHERIES OF SWIFTSUKE DANK 757 The Anderson River spawning escapement has been estimated from 1925-34 in the Dominion Reports. The lowest escapement was 7,500 in 1933, the highest 135,000 in 1929, with an average for the 9 years of 55,000 sockeyes. In the only 2 years for which figures are given, 1928 and 1932, the catch was 15,000 and 28,000 respectively. The total annual run may therefore be considered as approximately 75,000. The run to the Somass River appears to be larger. The Stamp River falls were formerly difficult for sockeye to ascend, most of the run to the Somass River spawning in Sproat Lake. In 1927, a permanent fishway was constructed, so that the run now spawns in Sproat Lake, Great Central Lake and Ash Lakes ; all of considerable extent. The Reports of the Dominion give the catch of Somass River sockeyes as 24,000 in 1928, 47,860 in 1930, 77,000 in 1932, 60,000 in 1933, and 75,000 in 1934. The escape- ment is unknown but, if we assume it was 50 percent, the run since 1932 has been close to 150,000. The annual run then to Barclay Sound appears to total in the neighborhood of 225,000 sockeyes. That a few of these fish may be captured on Swiftsure Bank is not impossible and it is unlikely that this can be adequately determined until such time as sockeyes are tagged on the bank. PUGET SOUND STREAMS The Skagit River, the only sockeye stream in the Puget Sound area, is no longer an important producer of sockeye salmon although it once supported a fair run. The Baker Lake sockeye hatchery, built in 1896 by the State of Washington on the Baker River, tributary to the Skagit, was bought by the Bureau of Fisheries in 1899 and has operated continuously since. The records of tliis station previous to 1916 were burned, but the remainder have been available. The annual escapement to Baker River from 1898-1901 was estimated at 20,000 sockeyes. Within a few years the run had become somewhat reduced, and by 1916 the escapement was about 5,000 sockeyes per year. The escapement of 14,558 in 1924 w^as due to the closing of the salmon traps in the waters east of Wliidbey Island during that season. The building of the Baker River dam destroyed all but 40 fish of the 1925 run, but since then the greater portion of those reaching the dam has been caught and hoisted over. This small run of sockeyes is distinguished from that of the Fraser River by the season of its migration. The traps east of Whidbey Island, which catch only Skagit River sockeye, commence taking them by the first of June. The run, which reaches its peak during the last week in June or occasionally the first week in July, and is practically over by July 20, averages about a month earlier than that to the Fraser River. The traps on West Beach usually show two modes in their sockeye catches; a small early mode due to Skagit River fish and a later one when the bulk of the Fraser River sockeyes are migrating. GULF OF GEORGIA STREAMS The only sockeye stream in the Gulf of Georgia proper is Saginaw Creek (see fig. 1). The 1926 catch, mentioned as being very small, was reported as 3,000 sock- eyes, while the escapement was estimated as between 18,000 and 19,000 fish. 758 BULLETIN OF THE BUREAU OF FISHERIES Just north of the Gulf of Georgia proper, there are small runs of sockeyes to sev- eral streams, the chief being the run to Phillips Arm, which is practically over before the run of Fraser River fish makes its appearance. MIGRATION IN SALT WATER Tagging experiments (O’Malley and Rich, 1919) have shown that the sockeyes entering through the Strait of Juan de Fuca strike the Salmon Banks and pass along the southern shore of San Juan and Lopez Islands, and, to a slight extent, the western shore of Whidbey Island, thence past Lummi Island, Whitehorn Point, Boundary Bay and Point Roberts to the mouth of the Fraser River. A few migrate north through Haro Strait Another tagging experiment (Dominion Report for 1929-30, p. 155; 1930), indi- cates that the run of sockeyes which enters the northern end of the Gulf of Georgia through Discovery Passage is bound chiefly for the Fraser River. Out of 519 sock- eyes tagged at Deepwater Bay in Discovery Passage, 107 were recaptured. The 17 recaptured at the point of tagging must be disregarded. Out of the remaining 90 a total of 82 fish, or 91 percent, were recaptured either in the Fraser River or at Point Grey (7 fish) just at the mouth of the river. TOTAL PACK OF THE FRASER RIVER SYSTEM The first real sockeye cannery was built at New Westminster in 1866 but no pack records are available for the first 7 years of the industry. The pack of 1873 was 8,125 cases (Rathbun 1899). The packs of 1874 and 1875 are unknown, but figures are available since 1876. The annual sockeye packs of the Fraser River system are given in table 26. 5 The Canadian fishery is much older than the American, reaching 100,000 cases by 1878 and 300,000 cases by the big sockeye year of 1889. By 1896 the Canadians had packed a total of 3,209,000 cases against 254,000 cases by the American operators. However, the introduction of traps in the early 1890’s gave a great impetus to the industry in Puget Sound. From 1898-1934, a 37-year period, the Canadian pack was larger than the American in only 6 years: 1903, 1905, 1906, 1915, 1922, and 1926. Up to the end of 1934 the packs of both countries aggregated the amazing sum of 21% million cases of sockeye, of which the Canadians had packed 10,773,000 cases, the Americans, 10,721,000 cases. * In compiling these data several sources have been used: The Dominion of Canada reports (1882-1834), the reports of the British Columbia Commissioner of Fisheries (1801-34), the Washington State reports (1890-1934), the Pacific Fisherman annual numbers (1903-34) and reports by the U. S. Bureau of Fisheries in various years from 1883 to 1834; as well as much unpublished material including printed tabulations of the pack by companies, prepared by R. P. Rithet & Co., Ltd., Victoria, B. C. for 1900; Fraser River Canner’s Association (1904-8); British Columbia Salmon Canners Association, and since 1923 by the canned salmon section of the Canadian Manufacturers’ Association. Material for recent years has been supplied by the Office of the Chief Supervisor of Fisher- ies for British Columbia and by the State of Washington Fisheries Department. In the earlier years the published reports of the packs are not segregated according to species and for these years we have made use of very extensive and careful notes kept by Henry Doyle of Vancouver, B. C. In addition, original records of various operators have been available. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 759 Table 26. — Sockeye pack of Fraser River system, in 4 8-pound cases Year 1873. 1876. 1877. 1878. 1879. 1880. 1881. 1882. 1883. 1884. 1885. 1886. 1887. 1888. 1889. 1890. 1891. 1892. 1893. 1894. 1895. 1896. 1897. 1898. 1899. 1900. 1901. 1902. 1903. 1904. 1905. Cases canned Ye^r Cases canned Fraser River 1 Puget Sound 2 Total Fraser River 1 Puget Sound 2 Total 8, 125 8, 125 1906. 185, 440 182, 241 367, 681 9, 847 9,847 1907 65, 061 96, 974 162, 035 64’ 387 64, 387 1908 79,211 170, 951 250', 162 100, 000 100, 000 1909 585, 935 1, 102, 399 1, 688, 334 50; 000 50, 000 1910 151, 595 248, 041 399, 636 25, 000 25, 000 1911... 64, 470 127, 761 192, 231 142, 516 142, 516 1912 124, 967 184, 680 309, 647 175, 000 175, 000 1913 739, 601 1, 673, 099 2, 412, 700 100, 000 100, 000 1914. 201, 498 335, 230 ' 536, 728 25, 000 25, 000 1915 95, 407 64, 584 159, 991 89, 617 89, 617 1916... 35, 070 84, 637 119, 707 36', 000 36', 000 1917.. 154, 415 411, 538 565', 953 125, 0C0 125, 000 1918... 21, 598 50, 723 72, 321 40' 000 40, 000 1919 38, 854 64; 346 103, 200 303, 875 1920 49, 184 62, 654 111.838 225', 000 225, 000 1921 41,731 102, 967 144, 698 131, 000 12, 000 143, 000 1922 54, 829 48, 566 103, 395 59, 000 15, 000 74, 000 1923 34, 574 47, 402 81, 976 455, 000 47, 852 502, 852 1924 39, 732 69, 369 109, 101 360, 000 41, 300 401, 300 1925. 36, 954 112, 023 148, 977 360, 000 65, 143 425, 143 1926.. 86, 765 44, 673 131, 438 325, 000 72, 979 397, 979 1927 65, 154 97, 594 162, 748 850, 000 312, 048 1, 152, 048 1928 30, 128 61,044 91, 172 216, 000 252, 000 468, 000 1929... 60, 823 111,898 172, 721 486, 409 512, 500 998, 909 1930 103, 662 352, 194 455, 856 172,617 229, 800 402, 417 1931 40, 947 87, 211 128, 158 974, 911 1, 106, 643 2, 081, 554 1932. 69, 792 81, 188 150,980 295, 079 372, 301 667, 980 1933. 54, 146 128, 518 182, 664 204, 848 167, 211 372, 059 1934... 139, 276 349, 602 488, 878 73, 175 123,419 196, 594 838, 813 837, 122 1, 675, 935 Grand total 10, 772, 638 10, 721, 425 21, 494, 063 ' Includes packs at Victoria, Quathiaski, and points in the Gulf of Georgia. Quathiaski packs not available for 1931 and 1934. 2 Includes 4,496 cases packed at Grays Harbor and the Columbia River in 1909 (see Cobb, 1930). Some idea of the former abundance of the sockeyes can be gained by noting that in 4 years of the former big-year cycle the pack was in excess of 1,675,000 cases, and, in 1901 and 1913, it was over 2,000,000 cases. METHOD AND LOCALITY OF CAPTURE INDIAN FISHING IN THE FRASER The Indians fishing in the Fraser River, except commercially, have depended largely on dip nets, gaffs, set nets, and spears. Dip nets are used chiefly in the larger rivers at points where the salmon have difficulty in ascending, such as Hell’s Gate canyon; the canyon of the Fraser just above the mouth of Bridge River; Fish Canyon, Hanceville and Indian Bridge on the Chilcotin River, and at Fort George on the Nechako River above its confluence with the Fraser River (fig. 25). The fishing at both Hell’s Gate and Bridge River canyons is much more successful during seasons of low water when the salmon have greater difficulty in passing. Set nets are used but slightly, not being practical in swift water. Spears are for use in the smaller tribu- taries, especially on the spawning grounds. Gaffs are mentioned in the 1917 report of the B. C. Commissioner of Fisheries as being used, along with dip nets, at Bridge River canyon. At one time the salmon were also taken by barricading the streams. The fishing in the streams near Stuart Lake in 1830 is thus described by John McLean (Wallace, 1932) who says that the natives built weirs of stakes and brush and caught the salmon in wicker baskets as they swam through openings in the weirs. 760 BULLETIN OF THE BUREAU OF FISHERIES In addition to catching the adult salmon the Indians formerly caught large quantities of the young sockeyes on their migration from the lakes to the sea. John P. Babcock (Report of the Fisheries Commissioner for British Columbia for the year 1903) describes how the Indians had built a dam of rocks and brush across a stream in the form of a great funnel with a basket trap at the lower end. Besides those caught in the trap many thousands were destroyed by becoming entangled in the brush. EXTENT OF THE INDIAN FISHERY Salmon fishing on the Fraser River was always carried on by the Indians, who consumed large quantities of fresh salmon and dried larger quantities for their own use and for barter with the tribes of the hinterland. Those living near the mouth of the river obtained some of all species of salmon, but the Indians dwelling nearer the headwaters depended chiefly on sockeye, and a few king salmon. The extent of this fishing is rather difficult to determine. At some points, such as Bridge River, Kam- loops, Stuart Lake, Hell’s Gate, Pemberton, and the Chilcotin River, large catches were made in good years (see fig. 25). Fishery officials have made many estimates of the Indian catch at the chief fishing camps by counting the numbers of salmon on the drying racks. According to their reports the sockeye catch at Bridge River in big years averaged 40,000. For the Chilcotin River system the catches of 1905 and 1909 were also estimated at 40,000, the catch of 1908 at over 20,000, and that of 1913 at 25,000. Of the Lillooet River, Crawford (13th Annual Report of the State Fish Commissioner (Washington) 1902) says: Every year the Indians gather here to secure their salmon for the winter and thousands of sock- eyes are taken and dried every season. One Indian speared seventy sockeyes in two hours, the first day I was there. A toll of between 400,000 and 500,000 sockeyes in the former big years is a con- servative estimate of the Indian catch. Even as late as 1929, with a greatly reduced abundance, as well as a much smaller Indian population, an accurate estimate showed that they caught 48,000 sockeyes, 20,000 kings, 25,000 cohos, 4,500 pinks, and 6,500 chums (Dominion Report, 1930). During years of poor sockeye runs the Indians living on tributaries where the runs failed were often on the verge of starvation, so complete was their dependence on the salmon for their livelihood. This was the case at Stuart Lake in 1841 and at Alexandria, on the Fraser River between the mouths of the Chilcotin and the Quesnel Rivers, in 1855 (Morice, 1904). CATCH BY COMMERCIAL GEAR In determining the number of sockeyes captured by the various methods in the different localities, the records of the actual number of sockeyes taken have been used wherever possible, and where these have not been available the number of cases canned has been converted into number of fish.6 » The number of sockeyes required to fill a 48-pound case of cans varies considerably from year to year, so that the use of the same conversion factor year after year would not give the best results. From two Canadian and two United States canneries we have obtained records covering 23 years, of the number of sockeyes required to fill a case. This varies from about 10 to 13 fish per case, tending to be higher in the earlier years, especially on the years of the big run. For years in which no conversion data were available we have used the average conversion factor of the other years of the same 4-year cycle, as the size tends to be the same from one cycle to the next. This is probably on account of the differences in size of the sockeyes spawning in the different lake systems, as the various lakes do not contribute equally to the runs of each cycle. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 761 Table 27 shows the annual catch by the principal forms of gear. The total com- mercial take of sockeye from 1873-1934 comes to 253% million, of which 116% million, or 46 percent, have been caught by gill nets in, or off the mouth of the Fraser River. The traps, both Canadian and American, account for 94 million, or 37 percent, and of the remaining 17 percent, 14 percent were taken by purse seines and 3 percent by miscellaneous gear. The miscellaneous included most of the fish caught at Qua- thiaski, as well as fish taken by minor Puget Sound gear such as gill nets, set nets, drag seines, and reef nets. Approximately 5 million of the trap fish and one-half million of the purse seine fish were taken by Canadian gear, so that, if the miscellaneous gear is ignored, the catches total 122 million by Canadian gear and 124 million by United States gear. The slight difference in pack in favor of the Canadians was due largely to ship- ments of fresh sockeye from Puget Sound waters to the canneries on the Fraser River, outweighing shipments in the other direction. In the early days the canning facilities on Puget Sound were too limited to handle the catch, and the Fraser River canneries were much closer to the sockeye fishing grounds. In 1894 the Canadians placed an embargo on the shipment of fresh sockeye out of the Province. This embargo, however, was not always in effect. In 1905, for instance, over 2 million pounds of late-run sockeyes were shipped from the Fraser River to Puget Sound canneries. Table 27. — Sockeye catch of the Fraser River system by various types of gear Year Fraser River gill nets Purse seines Traps Miscellaneous gear Total Territorial waters High seas 1 1873 100, 839 (?) 107, 332 799, 107 1,077,000 571, 350 272, 500 1, 768, 766 1,884, 750 1. 142, 700 272, 500 1, 112, 257 387, 720 1,428,375 433, 000 3,651,393 2, 263, 250 1,296,937 543, 100 5, 397, 005 3, 737, 200 4, 033, 720 3,120, 523 9, 959, 350 2,293, 715 4, 514, 385 1, 873, 981 11, 792, 692 3. 142, 814 2, 338, 987 742, 081 10, 143, 517 1, 983, 698 584, 033 707,011 100, 839 (?) (?) 107, 332 799, 107 1. 077. 000 571, 350 272, 500 1, 768, 766 1,884,750 I, 142, 700 272, 500 1, 112,257 387, 720 1, 428,375 436, 000 3,771,393 2, 423, 250 1, 840, 937 943, 100 6, 240, 896 4. 282. 001 5, 150, 757 4, 297, 971 14, 422, 178 5, 040, 360 II, 368, 243 4, 386, 345 25, 760, 031 7, 179, 255 4,252,619 2, 399, 071 20, 681, 236 4, 097, 154 1, 721, 569 2, 749, 880 1874 1875 1876 1877 1878 1879... 1880 1881... 1882 1883 1884.. 1885- 1886... 1887 1888 3,000 120, 000 160, 000 a 200,000 a 100, 000 a 372, 535 • 194, 801 207, 183 • 283, 134 a 734, 342 a 216, 502 438, 759 • 389, 856 a 509, 382 a 499, 784 a 261, 620 163, 264 a 500, 000 107, 602 » 33, 729 •60, 000 1889 1890 1891 . >344,666 300, 000 371,356 > 200, 000 903, 852 • 694,314 •3,128,486 2, 230, 143 6, 610,418 • 1, 722, 508 • 12, 457, 957 a 2, 736, 657 • 1,252,012 1,239,069 a 8, 662, 974 1, 505, 854 • 903, 807 a 1, 667, 295 1892 1893 » 100,000 a 150, 000 6,002 a 200, 000 a 600, 000 > 300, 000 804, 681 a 400, 000 3 1, 000, 000 1894 1895 — 1896 1897... 1898 1899 1900 1901 1902 a 800, 000 a 400,000 254, 657 > 1, 374, 745 a 500, 000 a 200, 000 > 325, 574 1903 1904 1905 1906 1907 1908 , 1 High seas catch 1925-1934 from TJ. S. Fishery Industry reports, before that from our data, plus sockeye canned at Neah Bay. Some taken before our records. • Estimated: From 1900 to 1912 the U. S. trap catch equals our data plus 20 percent, from 1896 to 1898 plus 50 percent, 1894 purely an estimate, and 1891 equals our data times 2. 762 BULLETIN OF THE BUREAU OF FISHERIES Table 27. — Sockeye catch of the Fraser River system by various types of gear — Continued Year Fraser River gill nets Purse seines Traps Miscellaneous gear Total Territorial waters High seas 1909 4, 869, 134 3, 484, 799 12, 026, 263 546, 278 20,92 6,474 1910 1, 459, 297 2 1,060,558 J 1, 905, 962 2 30, 000 4, 455, 817 1911 659, 496 2 392, 300 3 1, 101, 837 3 25, 000 2, 178, 633 1912 1, 185, 746 2 269, 603 3 1, 877, 945 3 30" 000 3, 363, 294 1913 - 8', 761 , 249 10, 049, 295 12, 493,’ 687 38, 808 31,343,039 1914 2, 035, 630 lj 344, 004 2, 276, 554 36, 879 5, 693 i 067 1915 1, 050" 672 ’ 244, 693 456, 542 73, 556 1, 825, 463 1916 311, 196 150, 446 768’ 369 56, 305 1, 286,316 1917 1, 402’ 327 1, 989, 191 3, 292', 193 199, 690 6, 883, 401 1918 ' 19"! 352 45,073 2, 495 538, 903 27, 546 811,369 1919 368,395 286, 365 25, 366 539, 618 29, 125 1, 248, 868 1920.. 486,118 53, 083 828 666,917 12, 783 1, 209, 729 1921 - 433, 852 221, 152 35, 820 915,313 80, 104 1, 686, 241 1922 514, 249 88, 277 5, 167 436, 848 49,461 1, 093, 992 1923.. 300, 115 142, 355 5,717 370, S74 37, 892 856, 953 1924 372, 333 99, 098 25,931 680, 554 36, 390 1, 214, 306 1925.. 397, 386 287, 329 142, 224 975, 252 26, 525 1, 828, 716 1926. 891, 0,45 90, 523 14, 286 355, 848 30, 764 1, 382, 466 1927.. 643, 254 435, 693 50, 000 586, 944 62, 596 1,783,487 192S 267, 457 61,716 19, 770 566, 280 26, 460 941,683 1929... 605, 170 368, 155 102, 134 926, 939 56, 780 2, 059, 178 1930 964, 987 2, 504, 978 144, 278 908, 066 65, 723 4, 588,032 1931 450, 532 316, 141 217,015 444, 366 5,585 1, 433, 639 1932 657, 222 353, 849 19, 579 510, 113 46, 378 1,587, 141 1933 546, 026 641, 505 121,061 1, 198, 887 42, 957 2, 450, 436 1934 1, 230, 986 1,716,055 674, 716 1, 391, 104 7,497 5, 020, 358 Total 116, 543, 814 34, Oil, 888 1,606,376 94, 132, 880 7, 226, 575 253, 521, 533 Percent 46 13 1 37 3 100 2 Estimated: From 1900 to 1912 the U. S. trap catch equals our data plus 20 percent, from 1896 to 1898 plus SO percent, 1894 purely an estimate, and 1891 equals our data times 2. LOCALITY OF TRAP CATCHES In addition to the locality segregation given in the foregoing table, the following detailed analysis of the locality of capture of the trap fish shows the relative importance of each fishing district in Puget Sound. Since records were obtained for about 82 percent of all of the trap-caught sockeyes, 100 percent from 1915 to 1934, inclusive, the figures given in table 28 may be considered representative of all of the 94 million taken by this method. LOCALITY OF PURSE-SEINE CATCHES Of the 35K million taken in purse seines, 1% million are definitely assigned to extraterritorial waters off the mouth of the Strait of Juan de Fuca. The locality of capture of the remainder cannot be as easily established as in the case of those caught by traps. The principal sockeye seining grounds are the Salmon Banks and Point Roberts Areas, with lesser amounts from Rosario Strait, Haro Strait, Lummi Island, Birch Bay and Boundary Bay Areas, and a very few from West Beach. Data from companies buying purse seine fish show that during the 4-year period covering a year of each sockeye cycle, from 1931-1934, about two-thirds of the seine- caught sockeyes were taken on the Salmon Banks. This includes the Salmon Bank and South Lopez Areas. Of the remainder the larger share were caught at Point Roberts, with lesser amounts from Rosario Strait, Lummi Island, and Haro Strait Areas. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 763 Table 28. — Sockeye catch by traps in different areas, 1893-19$/+ 1 Areas in which caught Year 1893. 1894. 1895. 1896. 1897. 1898. 1899. 1900. 1901. 1902. 1903. 1904. 1905. 1906. 1907. 1908. 1909. 1910. 1911. 1912. 1913. 1914. 1915. 1916. 1917. 1918. 1919. 1920. 1921. 1922. 1923. 1924. 1925. 1926. 1927. 1928. 1929 1930. 1931. 1932. 1933. 1934. Total. North of Sandy Point 1 Sandy Point to Deception Pass West Beach and Ebeys Landing Strait of Juan de Fuca 1 East of Whidbey Is- land and south of Point Wilson Undeter- mined Total 185, 678 185, 678 600, 957 GOO, 957 454,831 8, 045 462, 876 1, 904, 593 208, 191 2, 873 2, 1 15' 657 1, 482, 549 15, 081 1, 497, 630 3’ 247’ 248 832, 680 4, 079’, 928 1, 098’ 886 324, 505 7, 148 6, 142 1, 436, 681 7, 931 ; 801 1, 864, 905 21,925 6,411 595, 388 lo! 420’, 430 1, 573, 961 687, 050 19, 907 4,089 2, 285, 007 775, 692 245, 923 22, 214 1, 043, 829 728, 780 205, 141 5, 1 19 50, 000 989, 040 5, 039, 241 1, 517, 202 238, 906 524, 535 2, 897 7, 322, 841 832, 147 338, 049 20, 348 72, 357 4,037 1, 266, 938 455, 356 208, 059 27, 749 70, 822 2,990 764, 976 1, 004, 494 253, 066 22, 485 128, 218 2, 519 1,410,782 5, 789, 782 2, 306, 825 212, 033 725, 736 1, 377 9, 035, 753 991,026 391, 156 21,819 218, 461 2, 250 1,624,712 572,511 287, 167 6,541 59, 212 2, 635 928, 066 911,978 488, 636 24, 118 164, 530 3, 109 1,592, 377 6,011,680 3, 080, 543 100, 027 881, 123 1,350 10, 074, 723 968, 885 683, 530 153, 991 171,078 1, 213 1,978, 697 240, 670 149, 264 27, 572 26, 506 11,046 1,484 456, 542 386, 446 278, 566 39, 765 55, 550 6,581 1,461 768, 369 1, 584, 230 1,091, 186 164, 683 437, 175 4, 197 10, 722 3, 292, 193 220, 785 233, 426 33, 382 48, 312 2, 938 60 538, 903 284, 714 142, 805 17, 136 86, 608 8, 331 24 539, 618 307, 707 258, 877 34, 496 45,416 9. 441 980 656, 917 476, 128 347, 135 38, 713 46, 508 6,208 621 915,313 220, 710 152, 200 2-4, 203 38, 393 1, 342 436, 848 168, 851 161,238 9,115 28, 365 3, 305 370, 874 382, 755 232, 610 17,410 45, 933 1, 846 680, 554 543, 310 338, 279 34, 279 52, 897 6, 594 975, 309 192,818 129, 592 6, 389 25,324 1,720 5 355, 848 322, 282 203, 828 7, 853 51, 383 1, 598 580, 944 308, 092 204, 315 18, 156 33,812 1,905 566, 280 488, 018 328.918 .54, 851 46, 564 4, 062 4, 526 926, 939 488, 386 323, 461 35, 503 58,184 2, 532 908, 066 206, 338 184, 492 18, 332 31, 150 4, 054 444, 366 236, 248 202, 470 19,716 48, 843 2,836 510, 113 510, 053 539, 848 25, 137 122, 349 1, 500 1, 198, 887 821, 737 469, 463 27, 507 69, 751 2, 646 1, 391, 104 50, 952, 364 19, 917, 787 1, 561, 401 4, 465, 101 125, 701 615, 271 77, 637, 615 1 North of Sandy Point includes Canadian traps in Boundary Bay; the Strait of Juan de Fuca includes Canadian traps near Sooke and American traps west of Point Wilson. From 1915-34 our data include all trap-caught sockeye. All but portions of the Sooke data are actual numbers of fish, not converted figures. During the late sockeye run of 1934, seining was permitted from September 1-8 in the portion of seining area 17 directly off of the mouth of the Fraser River, and 328,000 fish were taken. Small amounts of sockeyes are sometimes seined around Pender Island in seining area 18. In 1930 this area produced 31,000 sockeyes, in 1931, 3,000, and in 1934, 45,000. CHANGES IN ABUNDANCE OF DIFFERENT PORTIONS OF THE RUN The gill nets have been used as giving the best measure of the change in the time of the run. The average gill net delivery for each 7-day period was derived by combining the averages for each year and dividing by the number of years with data (see table 29). 764 BULLETIN OF THE BUREAU OF FISHERIES The curves for the 12 early years, 3 sockeye cycles, and for the 12 late years are shown in figure 22. For the 12 early years sockeye fishing usually terminated on August 25, although consid- erable fishing was carried on during the heavy fall runs of 1905 and 1909. No data are available for the fall of 1905, but those for 1909 are shown in figure 22. Because of the lack of fall fishing during most of the earlier years it is often thought that there were no abundant late runs in those years, but the figure shows plainly that the late run of 1909 was many times as abundant as that of 1930, the most abundant of the late rims during the last 12 years. That some sockeye were ordinarily present in the river after the usual cessation of fishing on August 25, during the years before we have accurate records, is indicated by Rathbun (1899, p. 270) who says: Figure 22. — Occurrence of sockeye as shown by Fraser River gill-net catches. Note the peak in the week ending August 11 in the three early cycles (1898-1909), which is entirely missing in the three late cycles. The late runs of 1909 and 1930 are also shown. The big years of 1901, 1905, 1909, and 1913 were characterized by a second heavy run coming late in the fall. . . . the average fishing season ends somewhere about the 20th to the 25th of August, and years are recalled when nothing could be done after the first week of that month. Small numbers usually continue present during more or less of the early part of September, but with the near approach of the spawning period the fish rapidly deteriorate in appearance and condition and lose their com- mercial value. Table 29. — Change in seasonal occurrence ofsockeyes between early and late years in Fraser River gill nets Week ending 1898 to 1909 1923 to 1934 Week ending 1898 to 1909 1923 to 1934 Num- ber of years with data Average catch per gill net delivery Num- ber of years with data Average catch per gill net delivery Num- ber of years with data Average catch per gill net delivery Num- ber of years with data Average catch per gill net delivery July 7 3 33. 34 5 10. 83 Sept. 22 1 202. 90 9.03 9 14. 40 6 10. 41 Sept. 29 1 107. 32 4 3.09 July 21 10 19. 73 12 8. 44 Oct. 6 1 80.93 3 15. 64 12 34. 37 12 13.09 Oct. 13... 3 2. 77 12 63.31 12 15. 15 Oct. 20 2 1.63 12 86. 58 12 16. 98 Oct. 27 2 1.44 12 35! 14 12 14. 54 Number of fish 1, 982, 735 1, 469, 746 23. 41 12 11. 10 Number catches... 30,700 87, 514 Sept. 15 i 101. 52 12 16. 60 Wbat has happened to the early runs is clearly shown by table 30, giving the average catches during the period from July 15-August 25, which embraces almost all SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 765 of the period usually fished during the earlier years. The decrease in abundance is astounding, the average of 14.85 sockeye per delivery during the later years being but 24 percent of the earlier average. Even if the former big-year cycle is omitted from both periods, the deliveries in the later period are only 32 percent of the earlier. Table 30. — Average catch per gill net delivery of sockeye on the Fraser River Years Number caught July 15 to Aug. 25, inclusive Number of de- liveries Average delivery Years Number caught July 15 to Aug. 25, inclusive Number of de- liveries Average delivery 1898. 38, 636 76, 910 38, 208 186, 797 45, 736 164, 058 64, 867 724, 000 128, 484 71. 292 129, 662 201, 467 1,240 1,201 1,172 1,345 607 3,640 2,845 4, 901 2,237 3,062 3, 872 2, 598 31.16 64.04 32.60 138. 88 75. 35 45. 07 22.80 147. 72 57.44 23.28 33. 49 77. 55 1923 8, 823 IS, 266 17, 005 22, 134 18, 600 37, 873 85, 811 81, 557 57, 064 152, 847 104, 944 141, 932 783 1,018 1, 183 1, 172 2,386 3, 282 3,512 6, 181 6, 101 7, 389 7, 677 9, 427 11.27 17.94 14. 37 18.89 7.80 11.54 24. 43 13. 19 9. 35 20.69 13. 67 15. 06 1899 1924 1900 1925.... 1901 1926 1902 1927 1903 1928 1904 1929.... 1905 1930... 1906... 1931 1907 1932 1908 1933.... 1909 1934... Sum. Sum 749. 38 178.20 Average 62. 45 14.85 Average of “off” years Average of “off” years 42.83 13. 86 The most unfortunate feature in the depletion of the earlier-running sockeyes is the accompanying fall in the quality of the pack as a whole. Not only have the sockeyes been depleted, but worse, the depletion has been much heavier during the early run when the quality is of the best. The late-running sockeyes have been encouraged by several circumstances; first, during the earlier years the late run was seldom fished on account of its inferior quality; second, the Fraser River closed season, which began on August 25 during most years, was a protection; third, the 10-day fall closed season in odd-numbered years from 1921-29, and in all years since 1930 in Puget Sound waters, has enlarged the escapement of the late-running fish. This serves to emphasize the fact, common to nearly all fisheries, that the most valuable portion of a population is usually the first to be destroyed. CHANGES IN ABUNDANCE Because the sockeye has always been the chief object of the gill net and trap fisheries, its abundance may be more accurately measured than that of the other species. The abundance of a salmon run cannot be measured in the same manner as that of a marine species for which each unit of gear may fish throughout the season upon the same general population. The salmon are running a gauntlet, each school avoiding capture as it approaches closer to its goal. Therefore, because variations in temperatures, currents, winds and tides cause changes in the rate and exact route of migration, the productivity of the different fishing areas may exhibit annual variations independent of those produced by the actual numbers of migrating sockeyes. Conditions often favor one form of gear more than another, so that the availability of the schools to one method of fishing must not be accepted as the final criterion of 766 BULLETIN OF THE BUREAU OF FISHERIES abundance without comparing it with the availability to other forms of gear. Also, the number of sockeyes caught on Swift sure and Salmon banks is bound to influence the catches in Boundary Bay, and they aid in influencing the catches in the Fraser River. The gill nets in the Fraser River, covering a restricted area, undoubtedly sample the portion of the run that escapes thus far more thoroughly than the traps and seines can hope to do. If the number and efficiency of the gill nets remained constant they might then give an adequate picture of the escapement, but, unfortunately, their number varies considerably. To work out these complexities so as to allow for the difference in seasonal avail- ability to different gear, the effect of one form of gear on the catch of another, the amount of competition between units of gear according to their numbers, and, finally, the changes in abundance of some races due to the difference in fishing intensity at different parts of the season, is beyond the scope of this report. General indices of abundance are presented for the major forms of gear and such general conclusions drawn as appear justified. AVERAGE CATCH PER UNIT OF EFFORT WITH GILL NETS The number of sockeyes actually captured by gill nets in the Fraser River, taking into consideration, whenever possible, fish shipped to and from the Fraser River, is given in table 31. This has been divided by the number of units of fishing effort and the results shown in figure 23. In the earlier years the catch was often limited by the capacity of the canneries, and this continued in the big-year cycle up to 1913. Under these conditions the curve does not give a true picture of the actual early abundance which was undoubtedly somewhat higher. Year 1877— 1878,-. 1879— 1880— 1881— 1882... 1883— 1884.. . 1885.. . 1886— 1887— 1888.. . 1889— 1890.. . 1891.. . 1892.. . 1893— 1894.. . 1895— 1896— 1897.. . 1898.. . 1899.. . 1900.. . 1901— 1902.. 1903.. . 1904.. . 1905— 1906- Table 31. — Catch per unit of effort by gill nets, 1877-1934 Number gill-netted Total units of effort Catch per unit of effort Year Number gill-netted Total units of effort Catch per unit of effort 799, 107 285 2,804 1907 584, 033 2,942 199 1, 077, 000 449 2, 399 1908 707,011 2,410 293 571,350 304 1,879 1909 4, 869, 134 4, 634 1,051 272, 500 274 995 1910 1, 459, 297 2,745 532 1, 768, 766 396 4, 467 1911 659. 496 2, 350 281 1, 884, 750 666 2, 830 1912 1, 185, 746 2, 476 479 1, 142, 700 782 1, 461 1913 8, 761, 249 4, 369 2, 005 272, 500 723 377 1914... 2, 035, 630 4, 621 441 1, 112,257 672 1, 655 1915... 1, 050, 672 4, 683 225 387, 720 775 500 1916 311, 196 4, 299 72 1, 428, 375 1, 055 1, 354 1917 1, 402, 327 4, 849 289 433, 000 576 752 1918 197, 352 3, 049 65 3, 651, 393 596 6, 126 1919.... 368, 395 2, 600 142 2, 263, 250 596 3, 797 1920 486, 118 2, 545 191 1, 296, 937 629 2, 062 1921 433, 852 2,702 181 543, 100 954 569 1922 514, 249 2,548 202 5, 397, 005 1,626 3,319 1923. 300, 115 1,768 170 3, 737, 200 2,481 1, 506 1924 372, 333 1,768 211 4, 033, 720 2, 580 1, 563 1925. 397, 386 1, 689 235 3, 120, 523 4, 291 727 1926— 891,045 1,810 492 9, 959, 350 3,832 2,599 494 1927 648, 254 2, 010 323 2, 293, 715 4, 642 1928 267, 457 2, 092 128 4, 514, 385 4,785 943 1929 605, 170 2,312 262 1,873, 981 6,369 294 1930 964, 987 2, 375 408 11,792, 692 6, 350 1,857 1931 4.50, 532 2, 163 208 3, 142, 814 4, 278 735 1932 657, 222 546, 026 2, 2S9 287 2, 338, 987 5, 362 436 1933 2, 598 210 742, 081 3,571 208 1934.... 1, 230, 986 2,745 448 10, 143, 517 1, 983, 698 4,582 3,178 2, 214 624 116, 335, 643 767 SALMON AND SALMON FISHERIES OF SWIFTSURE BANK On account of economic conditions only six canneries operated in 1884 and 1885; but the number of licenses issued was as great as in years when double the number of plants were busy. Therefore, the low points of 1884 and 1885 should be regarded with suspicion, as the catch per net was obviously lowered by the inability of the canneries to utilize their full catching capacity. Eliminating these doubtful years, 1886 appears to be the low point of the early period. Since about 1897 the whole curve is lower than would be the case were the whole sockeye population to have reached the river, as it did before the expansion of fishing in Puget Sound. Regard- less, however, of all the fac- tors that presumably affect the level of the curve to some extent the fall is far too pronounced to mean anything but depletion. INDEX OF ABUNDANCE FROM TRAPS The salmon traps form a very reliable means of deter- mining the abundance of the sockeye, inasmuch as they were driven year after year in the same location; and, although the fishing ability of the individual trap may have varied somewhat from year to year, on ac- count of weather or tides, yet the decrease in the catch of one trap is apt to be compen- sated for by the increase in another if a sufficiently large sample is utilized. In making this index traps were selected from various localities so as to discount the effect of any slight changes in migration routes or any diminution of the numbers migrating past any one locality, which might be caused by hydrographic conditions or by sockeyes of different lake systems using different migration routes through the salt water channels leading to the mouth of the river. Of the 43 traps selected, 3 were from the Point Roberts Area, 12 from Boundary Bay, 5 from Birch Bay, 4 from Lummi Island, 6 from Rosario Strait, 3 from the South Lopez Area, 4 from Salmon Banks, 1 from Waldron Island Area, and 5 from Haro Strait. No trap selected fished less than 10 years and 5 of them fished from 1898 to 1934, or 37 years, without a single break. They averaged 27 fishing years each between 1896 and 1934. The use of more traps would have given too much weight to the Boundary Bay Area which was already well represented. In most of the other areas all available traps were used to aid in compensating for changes in the route followed. No traps were used from West Beach as they also catch sockeyes bound for the Skagit River, but, as this area is a small producer of sockeyes, its omission can be of no consequence in determining the trend. Figure 23. — Annual catch per unit of fishing effort of Fraser River gill nets for the 68-year period from 1877-1934. Note the decrease in the catch in each of the four cycles, These cycles are caused by the sockeye maturing predominately at 4 years of age. 768 BULLETIN OP THE BUREAU OF FISHERIES As not all of these traps fished every year during the period under consideration, it was necessary to determine the relative efficiency of each trap, especially since no two traps are exactly alike in their potential capacity to catch fish. In determining these efficiencies it was, of course, necessary to use a base. The use of any one year as a base could not give a very accurate picture of their relative efficiencies, so a 28-year period was employed, from 1902-31; with the excep- tion of 1908 and 1922. Fifteen traps were found that had fished every year during this period, of which 1 was from the Point Roberts Area, 10 from Boundary Bay, 2 from Birch Bay and 1 each from the Lummi Island and Salmon Bank Areas. For these traps an average annual catch per trap was computed. Using these average annual catches as a standard, or base, the proportion that the total annual catches of each of the 43 traps formed of the same annual catches of the standard was found. In- stead of using these proportions as weights, each trap was assigned an efficiency weight- ing which was the calculated average annual catch it theoretically would have caught had it fished for the whole 28 years represented by the standard, or base, curve. This was done for each trap by merely multiplying the average annual catch of the standard curve for the 28 years by the above-mentioned proportion. Having determined the relative efficiency of each of the 43 traps, the index was made by dividing for each year the total catch of such of the 43 traps as were driven by the total efficiency weightings of the same traps. The index figures are not actual numbers of fish but, as with most other indices, are to be considered in relation to one another. However, they give roughly the percentage that each year’s catches are of the average of the 28 years represented in the standard curve. Even though the trend of the base curve for the 15 traps rose or fell at a different rate than did the trend of the traps as a whole, this method of determining the effi- ciencies would prevent this difference in the trend from having any effect on the final index unless a large share of the traps selected fished for only a short number of years at one end of the period of time. Since this condition does not obtain, the index is believed to be a reliable measure of the changes that have occurred in the trap catches. Table 32. — Sockeye index of abundance from traps, 1896-1934 Year Catches Efficiency weights Num- ber of traps Index of abun- dance Index from stand- ard curve Year Catches Efficiency weights Num- ber of traps Index of abun- dance Index from stand- ard curve 1896 259, 512 157, 152 6 165. 134 1917 1, 777, 158 1, 361, 590 43 130. 521 139. 225 1897 843, 303 349, 089 10 241. 572 1918 350, 451 1, 316, 830 39 26. 613 23. 548 1898 821* 677 381, 254 11 215, 520 1919 306, 114 1, 161,984 35 26. 344 29. 226 1899 2, 663, 376 755, 475 21 352. 543 1920 499, 406 932, 553 27 53. 553 52. 306 1900 ' 942* 721 868', 394 26 108. 559 1921 621, 190 1, 310, 431 42 47. 403 47. 050 1901 5, 095, 464 833, 241 26 611. 523 1922 328, 554 802, 564 20 40. 938 1902 1, 403' 869 983, 037 30 142. 809 132. 035 1923 276, 658 1, 180, 625 36 23. 433 24.950 1903 ' 703' 336 983', 037 30 71. 547 70. 728 1924 555, 636 972, 745 27 57. 120 65. 769 1904 609, 681 912, 297 25 66. 829 68. 660 1925 679, 459 1, 302, 442 38 52. 168 59. 422 1905 4, 273, 212 1, 033, 479 31 413. 478 424. 986 1926 272, 170 1, 171, 431 33 23. 234 24. 999 1906 ' 875, 782 990, 361 29 88. 431 90. 637 1927 392, 468 1, 263, 574 39 31. 060 36. 742 1907 512, 369 976, 475 28 52. 471 53. 652 1928 418, 199 1, 121, 823 32 37. 279 35. 085 1908 _ 907', 670 824' 243 23 110. 122 1929 552, 836 1,310,431 42 42. 187 47. 686 19C9 4, 621, 094 1, 095, 853 31 421. 689 406. 717 1930 629, 889 1, 195, 611 36 52. 683 48. 720 1910 1, 058, 917 1, 042, 569 28 101. 568 98. 828 1931 298, 260 1, 169, 332 36 25, 507 23. 957 1911 657, 770 1, 232, 865 33 53. 353 56. 610 1932 338, 576 743, 919 21 45. 513 1912 1. 082', 917 1, 198', 760 32 90. 336 86. 775 1933 753, 311 1, 115, 144 32 67. 553 1913 5, 790, 820 1, 226, 629 35 472. 092 492. 264 1934 921, 829 1, 076, 915 33 85. 599 1915 244, 628 1, 342, 578 40 IS! 221 20. 320 Total. 45, 110, 330 1916.. 487, 271 1, 106, 225 33 44.048 42. 167 SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 769 The index would appear to be extremely reliable for trap-caught sockeyes, as, during the period from 1896-1934 the 43 traps caught 45 million sockeyes while the total trap catch since the beginning of the fishery totals but 94 million. During the past 20 years, when complete figures for trap catches are available, our sample comprised as high as 82 percent of the trap catch in 1924, and fell only as low as 54 percent, in 1915 and 1917. The index (table 32 and fig. 24) shows a marked decline in abundance in all four age cycles, comparing favorably with the average catch per unit of gill net effort, except in a few years. In 1897 the abundance shown by the trap index is decid- edly lower than that shown by the gill net averages, but a large part of this discrepancy may be due to the fact that a great many of the traps were driven for the first time in 1897 and so had not yet been efficiently located. The de- tails of the levels of abun- dance shown will be dis- cussed under the various cycles. PURSE SEINES 1897 1901 1905 1909 1913 1917 (921 1925 1929 1933 Figube 24— Sockeye index of abundance calculated from the catches of Puget Sound traps for the 39-year period from 1896-1934. A decrease in abundance has occurred in all cycles. In the 26 years since 1909 when purse seines be- came an important factor in the sockeye fishery, their catch has exceeded that of the traps in only 3 years: 1930, 1931, and 1934. Their success in 1930 was due to the heavy schooling, especially at Point Roberts, of the abundant late run which, massed in the shallows off the river mouth, were easily seined. The 1931 catch exceeded that of the traps because the seines had their second most successful season on Swiftsure Bank. In 1934 the purse seiners were prepared for a repetition of the abundant late run of 1930 and, although they did not do as well in the inside waters, they caught over three times as many sockeyes on Swiftsure Bank as in any previous season. In 7 of the 26 years their catch in both inside and offshore waters totaled less than 150,000 sockeyes per year. Six of these were even-numbered years when no pink salmon were running. The seiners fished during the early season for cohos in the offshore waters, and during the late season for both cohos and chums in the inside waters. 770 BULLETIN OF THE BUREAU OF FISHERIES In the odd-numbered years, which have abundant pink salmon runs, usually three or four times as many sockeyes are seined as in even years, because there are more seine boats fishing, and they are largely concentrated during the late summer in the areas where the pink salmon are migrating on their way to the Fraser River and other streams in the Gulf of Georgia. The average size of the purse-seine delivery is not a good measure of sockeye abundance. In the even-numbered years it tends to be high, as the boats fish only during the height of the run. In the odd-numbered years it tends to be low, as the boats often made a large number of catches, containing few sockeye per catch, while fishing primarily for pink salmon. The purse seine catches are thus not as reliable as a measure as either the trap or gill-net catches, but they do show how the purse seines have fared under varying conditions of abundance. In making this index the number of sockeye taken each year during each 7-day period was divided by the weighted number of deliveries. The weights were given according to the size of the boats making the catches in accordance with the efficiency weighting for all species described in the purse seine section of this report. Data were available for every year, except 1920, from 1911-34. Of the 23 years remaining, the data for 1918 cover such a short period of time that they were not used in comput- ing a normal curve for each week From the other 22 years a normal average daily delivery was made for each of the 6 weeks between July 15 and August 25, by merely dividing the sum of the averages for all years by the number of years. No week had less than 19 years data. For each year the sum of all the average daily deliveries for the six 7-day periods between July 15 and August 25, or as many of these six periods as there were data for, was divided by the sum of the average daily deliveries for the same periods for the noimal. The resulting index then is a measure of the annual abundance ex- pressed as a peicentage of the normal. The purse-seine index of abundance differs from the trap index in a number of years, but before deciding on the meaning of these differences several factors must be considered. Thus the actual catch of sockeye in 1918, 1922, 1924, 1926, and 1928 by purse seines in Puget Sound was less than 100,000 fish. In 1918 it was only 45,000 and in 1928 it was but 62,000. In such years the total quantities caught by purse seines were very low in relation to the actual abundances. In ceitain other years the purse-seine index is very high in relation to that for traps, as the purse seines may make catches out of all proportion to the abundance when the fish are heavily concentrated, as they were at Point Roberts in 1930. Although it has seemed unwise to lay any stress on the purse-seine index as an accu- rate measure of abundance, yet, considered in relation to the trap and gill-net indices, it portrays the fluctuations in availability of sockeyes to the purse seines, and is thus necessary to an understanding of the fishery. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 771 Table 33. — Sockeye index of abundance from Puget Sound purse seines, 1911-84 Year 1911. 1912- 1913- 1914- 1915. 1916. 1917. 1918. 1919. 1920. 1921. 1922. 1923. 1924- 1925. 1926- 1927. 1928- 1929. 1930. 1931 - 1932. 1933. 1934. Number of fish 25, 634 80, 955 1, 312, 188 405, 937 60, 644 36, 645 180, 477 7, 70S 22, 229 Number of catches 262 442 799 1, 176 1,815 513 890 153 212 Weighted number of catches 238. 52 452. 16 904. 26 1, 308. 00 2, 184. 32 590. 52 1,181.70 192. 20 311.22 Average size of delivery for week ending- July 14 29.24 39. 48 35. 81 15.99 22. 36 July 21 65. 92 53.64 196. 42 87.31 9. 12 55. 15 88.20 47. 63 July 28 138. 72 180. 21 555. 81 230. 62 30. 27 71. 77 243.28 36. 31 Aug. 4 178. 85 201.96 1, 999. 39 555. 60 40. 35 65. 22 460. 00 40. 15 72.20 Aug. 11 134. 93 286. 11 2, 938. 88 261. 75 40. 13 55. 73 120. 46 42. 20 138. 27 Sum 1 Number of years. Normal average.. 59, 340 19, 954 80, 279 26, 253 61, 764 74, 904 257, 741 54,603 349, 740 1, 503, 020 235, 864 364, 018 495, 126 1, 227, 634 1,174 149 1, 253 184 837 504 1, 714 1,212 4,031 2, 451 4, 206 3,711 7, 368 2, 758 1, 668. 70 137. 54 1, 780. 62 247. 76 1, 220. 70 666. 34 2, 483. 58 1, 755. 96 5, 958. 18 3, 921. 71 6, 516. 87 5, 832. 54 11,375.44 4, 222. 36 81. 19 17.24 33. 88 10. 15 24. 15 34. 39 18. 00 52. 23 99. 59 18. 09 12. 30 45. 21 78. 01 10. 55 16. 82 26. 50 49. 88 21.74 72. 67 46. 00 12. 83 160. 78 115. 38 62. 57 16. 74 54. 44 89.51 31.52 35. 72 55. 70 58. 09 33. 36 53.17 69. 78 32. 36 131.59 87.20 115. 05 21.47 45. 58 73. 98 82. 94 38. 12 102. 86 72. 40 50. 02 6, 942, 663 37, 814 1, 057. 77 2, 296. 05 4, 550. 09 38. 89 113.51 32. 87 113. 68 40.41 194. 10 53. 06 34. 65 102. 54 76. 52 74.45 74.97 97.41 164. 51 5, 187. 83 19 21 22 22 55. 67 109. 34 206. 82 235. 81 Year Average size of delivery for week ending —Continued Sum of weekly aver- Sum of normal aver- Index Aug. 18 Aug. 25 Sept. 1 Sept. 8 ages July 15 to Aug. 25 ages for same weeks 1911.. 73. 55 53.83 635. 80 70. 19 1912 97. 43 819. 25 787. 13 104. 08 1913 _ 2,119. 56 24. 04 4C8. 27 274. 85 128. 86 8, 278. 33 1, 270. 62 172. 36 905. 79 913. 93 1914 111.30 905. 79 140. 28 1915... 21.98 30.51 16.20 6. 56 905. 79 19. 03 1916 36. 82 284. 69 787. 13 36. 17 1917 93.31 40. 29 28.05 9.50 1, 045. 54 905. 79 115. 43 1918 166. 29 607. 64 27.37 1919 106. 18 73. 33 35. 25 29. 34 389.98 740. 78 52.64 1920 — 1921... 32. 88 23.03 8.70 6. 36 301. 83 905. 79 33. 32 1922... 144. 11 133. 17 506. 63 850. 12 59. 60 1923 67. 44 65.02 42. 79 13. 96 210. 52 850. 12 24. 76 1924 48. 07 506. 35 787. 13 64. 33 1925 44. 22 32. 16 21. 22 18. 37 418. 96 905. 79 46. 25 1926 114. 32 51.09 555. 22 905. 79 61.30 1927 54. 07 125. 09 147. 26 152. 07 282. 73 905. 79 31.21 1928 24.73 10.26 4.56 1.37 214. 87 905. 79 23. 72 1929 90. 01 50.41 27. 74 9. 01 484. 46 905. 79 53.48 1930. 147. 00 413. 27 603. 17 673. 61 761. 80 905. 79 84. 10 1931 56.95 33. 90 26. 77 13. 78 255. 96 905. 79 28. 26 1932 65.59 24.41 19. 31 21.72 350. 03 905. 79 38. 64 1933 56. 09 27. 75 17. 94 9. 15 361. 62 905. 79 39. 92 1934. 430. 36 487. 49 296. 52 690. 08 1, 187. 48 905. 79 131. 10 Sum 1 3, 948. 71 2, 254. 58 Number of years 22 19 Normal average 179. 49 118. 66 1 Excluding 1918. 71941-38- -6 772 BULLETIN OF THE BUREAU OF FISHERIES COMBINED INDEX OF ABUNDANCE In years when fishing conditions favored the traps the gill net measure of abun- dance was usually lower owing to the toll exacted by the traps, but when conditions were reversed, as in 1915 and in 1926, the gill net index was the higher. Since the two measures are thus somewhat interdependent, neither one gives as clear a picture of the actual abundance as the two considered together. Therefore, the two have been combined. In making the combination each index was, from 1896 to 1934, expressed each year as a percentage of its average over the whole 39-year period. In each year each percentage was then weighted in accordance with the percentage of the com- bined trap and gill net-caught sockeyes that had been taken by that form of gear. The weighted percentages were then combined to form the final index, which is given by 4 -year cycles in table 34. EXPLANATION OF CHANGES IN ABUNDANCE Having reviewed briefly some of the causes of changes in the sockeye fishery, the question arises as to the present state of the fishery and the present state of abun- dance. In order to arrive at any reasonable conclusions account must be taken of the changes that have occurred within each cycle of 4 years — four years, as men- tioned above, is the age at which the majority of the Fraser River sockeye mature — in regard to the size of the spawning escapements, and the extent of the areas seeded Table 34. — Abundance by cycles of Fraser River sockeyes Year Combined index of abundance Year Combined index of abundance 1896.... 134.1 1899 236.0 1900 69.4 1903___ 72.6 1904. 48.3 1907 40.2 1908 78.8 1911 46.5 1912 79. 1 1915 33.4 1916. 29.3 1919 23.3 1920. 39.9 1923 24.5 1924 43.7 1927. 43.2 1928 28.1 1931 29.6 1932 45.9 Year Combined index of abundance Year Combined index of abundance 1897 411.7 1898 132.4 1901.... 421.4 1902.. 126.4 1905 374.7 1906.... 96.4 1909. 299.3 1910. 89.2 1913. 377.5 1914.. 80.8 1917 90.1 1918.... 19.0 1921 35.6 1922 35.4 1925 42.6 1926 70.3 1929.... 39.8 1930... 69.3 1933 49.8 1934. 75.6 The providing of a large number of spawners, while of importance, cannot achieve permanent rehabilitation unless these spawners are members of several different “races” or “colonies” of sockeye, so that they will migrate to many different lake systems. Such a distribution of spawners will insure ample spawning gravel for the adults, will guard the fishery against failure when on occasion unfavorable conditions of weather or enemies destroy the spawning of any single lake system, and will give a greater stability to the fishery as it is far better to have successive waves of migrating adults passing through the gear, than to have the whole season’s migration occur in a very few weeks, as may easily happen when the total migration is to one lake system. A clearer conception of these waves of migration may be gained by thinking of the main river merely as an extension of the salt water channels up which different races of fish migrate to their spawning grounds on several independent lake systems. The principal lake systems of the Fraser River, the tributaries of which are sockeye spawning grounds, are shown in figure 25. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 773 Figure 25. — Sockeye spawning ground of the Fraser River. All of the lakes shown are mentioned as sockeye lakes either in the reports of the British Columbia Commissioner of Fisheries or in the reports of the Department of Fisheries of Canada. Other lakes in this river system have been omitted. The sockeye spawn chiefly in the streams tributary to these lakes. Tue sockeye fry descend into these lakes and spend some time there, usually about a year and a half, before migrating to the sea. 774 BULLETIN OF THE BUREAU OF FISHERIES Abundance of Cycle Ending in 1934 The cycle of years — 1934, 1930, 1926, etc. — immediately following the big years showed a decline from 1898-1914 amounting to 39 percent from the 1898 level. The catch of 5,000,000 sockeyes in 1898 did not appear to be unduly heavy at the then existing level of abundance, only a 4 percent drop, which may not be statistically significant, occurring between 1898 and 1902. In 1902, however, the catch was in- creased to over 7,000,000 fish, resulting in a drop of 23 percent in the abundance of the 1908 run. Catches of over 4,000,000 in 1906 and 1910 were both too heavy for these lower levels of abundance and the catch continued to decline. In 1914, the lowest level of abundance the cycle had thus far experienced, the fishing was very intense. One hundred traps fished in the sockeye areas, the most in any off year since 1903, and the gill net effort was exceeded only by 1900, 1901, and 1903, resulting in a catch of 5,700,000 sockeyes. The spawning ground reports for 1914 indicated the poorest escapement on record, which was amply borne out by the run of 1918, the next year of this cycle, which was the poorest in the whole history of the Fraser River fishery. The intensive fishery of 1914 was doubtless instrumental in causing this remark- ably low escapement, but there is little doubt that at least a small portion of the blame must be laid on the blockade of Hell’s Gate in 1914. The report on this blockade stated that no salmon were able to ascend through the canyon from August 10 to 25, and that the fish had great difficulty in passing at other times, some 20,000 being put over the rapids with dipnets. Although a fair amount of gear was employed in 1918 the catch of just over 800,000 was relatively much less than that of 1914, considering the very low level of abundance. However, the remarkable increase in abundance between 1918 and 1922 cannot be explained in terms of catch or escapement. The survival rate of the sockeyes being influenced to a great extent by conditions in the lakes, and probably, to a lesser extent, by conditions in the ocean, is subject to occasional violent fluctuations. In this case the result was a doubling in the level of abundance between 1918 and 1922. In 1922, with the sockeyes much more numerous than in 1918, the catch was only slightly over 1,000,000 fish. The number of sockeye traps was the lowest since 1898 and the gill net effort had fallen considerably since the war years, permitting the best off year escapement for several years, possibly since 1912. One feature of the 1922 run was a fair escapement to the Shuswap-Adams Lake system. The relatively good escapement of 1922 was reflected in an improved run in 1926. The run was exceptionally late, and, in addition, appeared not to have followed its usual migration routes through the salt-water channels leading to the mouth of the Fraser River. As a result, neither the traps nor the purse seines in Puget Sound caught many sockeyes, and the gill net operators on the Fraser River received the full benefit of the run, catching more per unit of fishing effort than in any year since 1913. However, the number of gill nets was so small that the escapement was relatively very high in proportion to the catch, which was slightly under 1,400,000. The results of the 1926 escapement are shown in the catches of 4,600,000 and 5,000,000 in 1930 and 1934, respectively. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 775 Abundance of Cycle Ending in 1933 The big year cycle, ending in 1933, 1929, 1925, etc., was tremendously abundant from the earliest records of the commercial fishery in 1877 up until the cycle following the Hell’s Gate disaster of 1913. In earlier years the catch was so strictly limited by the capacity of the canneries that the index of abundance was always too low. All one can say is that the cycle was far more abundant than the others. In 1897 the trap index is considerably lower than in 1901, due largely to the fact that many of the traps were driven for the first time in 1897. That the big-year cycle was somewhat higher, as indicated by the combined index of abundance, in 1897 and in 1901 than in any of the succeeding years is undoubtedly true. In 1901, for instance, one trap in Boundary Bay caught 680,000 sockeyes between July 10 and August 29, which is as much as the entire trap catch of sockeyes for 11 out of the 21 years since 1913. The 1901 catch of 25,800,000 sockeyes was, next to 1913, the largest in the history of the fishery. The trap catch in 1901, with less gear, was equal to that of 1913, and the gill net catch of 11,800,000 was 3,000,000 higher than that of 1913. In 1913 the power purse-seine fleet, which was nonexistent in 1901 (only hand-propelled seine boats were then in use), took 10,000,000 fish. However, the difference in the catches of 1901 and 1913 was not due in any measure to a difference in the amount of gear, but rather to the great increase, by 1913, in the canning capacity of the plants. The number of sockeye wasted in 1913 was as nothing compared to the squandering of a natural resource that took place in 1S97 and 1901. Rathbun (1899) says: The run of 1897 was one of the largest, if not the largest, in the history of the region. Prepara- tions had been made in anticipation of a good year, both on the Fraser River and in Washington. The great body of sockeye first made its appearance about the middle of July and continued until about the end of the first week in August, a relatively short season, but during this period the cannery pack was completed and in addition an immense amount of fish was thrown away, the daily catch being often much larger than could be disposed of. It has, in fact, been claimed, though this is probably an exaggeration, that more fish were caught and wasted than were utilized. Concerning the waste of sockeyes in 1901 the Report of the British Columbia Commissioner of Fisheries for 1909, page I 11, says: The catch that year (1901) was so great that every one of the canneries on both sides of the inter- national line filled every can they had or could obtain; and in addition to the millions of fish which they packed that year, many millions more were captured, from both the Canadian and American waters of the Fraser River District, which could not be used, and were thrown back dead into the water. The waste of sockeye of our own catch and of that of the Americans in 1901 is believed to have been greater than the number caught and packed by all the canners on the waters mentioned in any year since, with the exception of 1905 and this year. Despite catches averaging 24,700,000 sockeyes per year in the big years from 1901 to 1913, huge numbers escaped to the spawning grounds. The spawning ground surveys made by the Provincial Fisheries Department estimated millions in 1901 and 1905. In 1909 estimates made by counting, for a portion of each day, the number of sockeyes ascending the fishway at Quesnel Dam showed that over 4,000,000 fish entered the lake. The sockeyes were thicker in the Chilco River than the observer had ever seen them in any unobstructed stream. Fully 1,000,000 were estimated to have entered Seton and Anderson Lakes. Shuswap and Adams Lakes were better 776 BULLETIN OF THE BUREAU OF FISHERIES seeded than in 1905, when most of the very heavy late run went to that lake system. The runs to Lillooet and Harrison Lakes, below Hell’s Gate, were practically a failure. The fact that tremendous numbers of sockeyes escaped to the spawning grounds on the big years, despite the huge catches, may have occurred because of the presence in all of the big-year cycles from 1901-13 of very abundant late runs, appearing after most of the fishing had ceased. The extent of this late run on the big years is indi- cated in the following quotation from the British Columbia Commissioner of Fisheries Report for 1909: On September 16, 1905, there appeared in the channels at the mouth of the Fraser a run of sockeye so numerous as to lead many competent observers to state that it equalled that which ap- peared during the first two weeks in August. This late run continued until the first week in October. None of these fish were observed in Juan de Fuca Strait, or in the American channels leading to the Gulf of Georgia and the Fraser River. During the first week of this movement several of our canners packed the fish, and a considerable number of them were purchased for and shipped to American canneries . . . Nothwithstanding the fact that there had been a similar run in the Fraser in Septem- ber and October of 1901, the claim was made that the late run of 1905 was most unusual. The same claim was again advanced as to the late run this year (1909). It appears evident, however, from the numbers of sockeye which ran in the lower Fraser in September and October of 1901 and 1905, and again this year, that a late run is characteristic of the big years. Whether the huge catch of 1913 had enough effect on the spawning escapement to have affected the abundance of the 1917 run will never be definitely known, as a portion of the sockeye ascending the Fraser River in 1913 were prevented from reach- ing the spawning grounds on accoimt of rock slides, incidental to the construction of a railway at Hell’s Gate in the canyon near Yale. The spawning-ground estimates of 1913 show 552,000 entering Quesnel Lake, contrasted to 4,000,000 in 1909, the pre- vious year of the cycle. Chilco Lake was likewise estimated to have had about one- eighth as many as in 1909. Anderson and Seton Lakes had an estimated escapement of 30,000 against 1,000,000 in 1909. Lillooet and Harrison Lakes, below Hell’s Gate, had poor runs. However, large numbers were seen in Adams River; and in Little River, connecting the outlet of Shuswap Lake with Little Shuswap Lake, the spawn- ing sockeyes appeared as thick as in 1905 or 1909. The run at Stuart Lake was re- ported to be one-twentieth as large as on most big years, and that at Fraser Lake about 50 percent as large. From the foregoing it is evident that, whether due chiefly to the obstruction at Hell’s Gate, or to the tremendous catch, the spawning escapement of 1913 was con- siderably curtailed. In spite of this curtailment, the run of 1917 was of such size that, had the fishing effort been sufficiently reduced to allow an escapement even comparable to that of 1913, the big-year cycle might have continued to dominate. However, the total fishing effort was probably as great as in any of the preceding big years, a relatively large portion of the run being taken before it even reached the river, as is shown by the small gill-net catches. Spawning-ground surveys in 1917 showed 26,000 spawners arriving at Quesnel Lake as against 552,000 in 1913. The Chilcotin Indians caught but 15,000 in the Chilcotin River compared with 25,000 in 1913. Seton Lake had not to exceed 200 fish caught by actual weir count. Shuswap and Adams Lakes had much less than in 1913. Harrison and Lillooet Lakes had the poorest spawning escapement that they bad known. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 777 The returns from this spawning brought a run in 1921 only two-fifths as abundant as that of the parent year. The catch of 1,700,000 in 1921 was relatively a great deal less, for the abundance level, than that of 6,800,000 in 1917. Since 1921 this cycle has been very slowly recuperating, increasing about 25 percent in abundance by 1933, according to the combined index. Besides producing the best pack of the last 4 years of this cycle, 1933 also had the best spawning escapement since 1917. Especially worthy of note was the good escapement to the headwater lakes as compared to other recent years. For instance, over 100,000 are estimated to have reached Chilco Lake. A fair number reached the lakes of the Stuart system. The escape- ment to the Fraser-Francois Lake system was twice that of 1929 and for the first time in years numbers of sockeyes reached Burns Lake. Abundance of Cycle Ending in 1932 The cycle of years, 1932, 1928, 1924, etc., immediately preceding the big years was the poorest of the 4 throughout the early years of the fishery, and in common with the other off years, this cycle commenced to decline before the beginning of the century. In 1900, while still at a fair level of abundance, this cycle was fished with extreme intensity, the gill-net effort being the highest in the whole history of the fishery and the number of sockeye traps as great as in the big year of 1901. The resulting catch of 4,400,000 was too great a proportion of the run, the abundance declining over 30 percent by 1904. In 1904 the fishing intensity was greatly reduced, only 2,400,000 sockeyes being taken, and the cycle recuperated. In 1908 the fishing intensity was again dropped, yet a larger catch of 2,700,000 was made. The abundance in 1912 was apparently as great as in 1908, as is shown both by the combined index and by the catch of 3,400,000 which was made with slightly more traps and about the same gill-net effort as the catch of 2,700,000 in 1908. Further- more, the proportion taken by the gill nets was much greater in 1912 than in 1908 which might indicate a better escapement. This is confirmed by spawning-ground estimates that would certainly place 1912 ahead of 1908. The index of abundance fell 63 percent between 1912 and 1916. In 1916, although the number of traps was fairly low, the gill-net fishery was very intense, yet only 1,300,- 000 fish were taken, and the unusually small proportion taken by the large number of gill nets would indicate a small escapement. The estimates show that it was probably the smallest escapement in the history of the fishery. Because the spawning of 1912 produced a run so very far below the average expectation for such a relatively good escapement, we are forced to conclude that the failure in 1916 was not caused by overfishing, but by some natural condition, pos- sibly connected with spawning, that greatly reduced the rate of survival. It is impossible, at this date, to know what all of the spawning-ground conditions were, but we have noted that the early months of 1913, when the eggs would have been incu- bating in the gravels of the spawning beds, were extremely cold. Average monthly temperatures from 1888-1930 at Barkerville and from 1891-1930 at Kamloops were studied. These two points were chosen for having long series of observations and for being close to the spawning grounds. For each locality the 778 BULLETIN OF THE BUREAU OF FISHERIES average monthly temperatures for January, February, and March were added for each year, and the sum subtracted from the mean average of the sum of these 3 months for the whole series of years. The two series of temperature deviations were added for each year and divided by two (see table 35). It will be noted that in both series the winter of 1913 was the second coldest in 42 years. That this long protracted cold spell might well have had a deleterious effect on the success of the 1912 spawning is obvious, but the point cannot be pressed until information on the effect of severe cold upon spawning has been collected. Although the escapement was reported as very poor in 1916, the abundance was somewhat higher in 1920, a much less intense fishery producing about the same catch as in 1916. The abundance was at practically the same level in 1924 as in 1920. The cycle fell off slightly in 1928 but recovered in 1932 owing probably to the very small catch that was made in 1928 in proportion to the abundance. Table 35.- — Winter temperatures of the upper Fraser River valley, 1888-1930 Year Barkerville Kamloops Average devia- tion in degrees Sum of average temper- atures, Jan., Feb., and Mar. Devia- tion from average in degrees Sum of average temper- atures, Jan., Feb., and Mar. Devia- tion from average in degrees 1888- 70.0 +8.9 1889 72.4 +11.3 1890 64. 1 +3.0 1891 53.8 -7.3 86.9 -0.7 -4.00 1892 65.0 +3.9 94.3 +8.7 +5. 30 1893. 55. 1 -6.0 73.2 -14.4 -10. 20 1894 60. 7 -. 4 1895_ 66.8 +5.7 1896 58.9 -2.2 95.0 +7.4 +2. 60 1897 56.4 -4.7 83.5 -4. 1 -4. 40 1898 68.8 +7.7 91.7 +4.1 +5. 90 1899. 53. 1 -8.0 80.8 -6.8 -7. 40 1900 74.5 + 13.4 103.5 + 15.9 +14. 65 1901 63.4 +2.3 91. 1 +3.5 +2. 90 1902 68.8 +7.7 100.8 +13.2 +10. 45 1903 60.6 -.5 84.7 -2.9 -1.70 1904 49.6 -11.5 81.0 —6. 6 -9.05 1905 68.8 +7.7 97.7 +10. 1 +8. 90 1906 70. 1 +9.0 101.2 +13.6 +11. 30 1907 1908 61.7 +0.6 8S.9 +1.3 +. 95 1909 49. 1 -12.0 82.2 -5.4 -8. 70 1910 60.6 -.6 94.7 +7.1 +3. 30 1911 50.2 -10.9 73.2 -14.4 -12.65 Barkerville Kamloops Y’ear Sum of average temper- atures, Jan., Feb., and Mar. Devia- tion from average in degrees Sum of average temper- atures, Jan., Feb., and Mar. Devia- tion from average in degrees Average devia- tion in degrees 1912 63. 1 +2.0 — 17.4 82.0 -5.6 -1.80 1913. 43. 7 63.8 -23.8 -20. 60 1914. 63. 7 +2.6 +16.1 -25.7 94.9 +7.3 + 17 7 -22. 1 +4. 95 +16. 90 -23. 90 1915 77.2 105.3 1916 35.4 65.5 1917 49.0 -12. 1 70. 1 -17.5 -14. 80 1918 63.4 +2.3 —1.8 89.2 +1.6 +3.5 +2.3 +8.7 -18.3 +1.95 +.85 +2.30 +5. 90 -16. 45 1919-._ 59.3 91. 1 1920. 63.4 +2.3 +3.1 -14.6 89. 9 1921 64 2 96.3 1922__ 46. 5 69.3 1923- 59. 1 -2.0 84. 4 -3.2 -2. 60 1924 71.8 +10.7 +4.7 +23.9 -3. 1 99.4 +11.8 +5.4 +23.4 +1.4 +8.4 —16.6 +11. 25 +5. 05 +23. 65 -.85 1925 65.8 93.0 1926. 85.0 111.0 1927 58.0 89.0 1928 73 0 +11.9 -9. 1 96.0 +10. 15 -12. 85 1929 52.0 71.0 1930 49.0 -12. 1 76.0 -11.6 -11. 85 Sum 2, 565. 1 3, 241. 6 Number of years Average. 42 37 61.1 87.6 Abundance of Cycle Ending in 1931 The cycle of years containing 1931 — 1931, 1927, 1923, etc. — has been the least abundant since 1899. The gill-net index shows that for six consecutive cycles, up to and including 1899, it was more abundant than the cycle following it. In 3 of the 6 years, 1887, 1895, and 1899, it was also more abundant than the cycle preceding it. Between 1899 and 1903 this cycle fell 69 percent according to the combined index of abundance — the largest drop in abundance in recent years with the exception of that of the big-year cycle after 1913. SALMON AND SALMON FISHERIES OF SWIFTSTJRE BANK 779 In 1899 both the trap and gill-net fisheries, especially the latter, were quite intense, resulting in a catch of 11,400,000 sockeyes. This catch does not appear to be excessive in relation to the index of abundance when compared to the catches of the big years. On the other hand, there is a possibility that the escapement in 1899 (no surveys were made of the spawning grounds) was much less than the mere comparison of the catch with the level of abundance would indicate, as neither the trap nor the gill-net data point to any late run in 1899, although the evidence is not conclusive. This same cycle had a late run in 1887, mentioned in the Dominion Report for that year, which states that many sockeyes were caught as late as October, which was very unusual. In all of the big-year cycles, from 1901-1913, very abundant late runs appeared after most of the fishing had ceased and provided heavy escapements. Since there is no evidence of a late run in 1899, it is quite possible that the catch was too heavy to allow a sufficient escapement. Some have ascribed this fall in abundance to the blocking of the Quesnel River by a dam at the outlet of Quesnel Lake, built in 1898, which caused the majority of the sockeyes reaching the dam to die below it without spawning, until after the con- struction of a fishway in 1904. That some of the sockeyes could not ascend the race is quite possible but that the majority did not enter the lake would seem to be refuted by the run of several millions that passed into the lake in 1905. If none spawned there in 1901, the run of 1905 cannot reasonably be accounted for. The darn and fishway are thoroughly described in the British Columbia Com- missioner’s Report for 1904. The dam was 18 feet high and the race was 124 feet wide and 382 feet long, with a drop of only 6 inches. At the head of the race there were 9 gates, each 12 feet wide. At the time of the sockeye run the water in the race was said to average 4 or 5 feet in depth, with a velocity of 12-14 feet per second. The fishway was merely a walled-in section along one side of the race. It was 26 feet wide and every 25 feet timbers 2 feet high were placed on the bottom to form an inverted V pointing upstream. The fishway led to two of the gates, one of which was kept open during the sockeye run. The dam was constructed for the purpose of shutting off the waters of Quesnel Lake in the fall of the year in order that mining operations could be carried on in the bed of the Quesnel River. Obviously the lake was permitted to become as low as possible during the summer so that the gates were merely openings through which the lake water flowed into the race. In 1905 the wall separating the fishway proper from the race was washed out, but the fish continued to ascend, and a low wall was substituted for the former high one. It is obvious that the problem was not passage through the gates but merely that of getting the sockeyes through the race. There would appear to be little doubt but that the majority of the sockeyes passed this obstruction. That a matter of some thousands could not, should be regarded as of no greater moment than the residue that fail to negotiate any fall or rapid of any consequence in a natural stream. Since the first great decline in this cycle, between 1899 and 1903, there has been a further decrease. From 72.6 in 1903 the combined index fell to 40.2 in 1907, due, as before, to overfishing. Remembering the good pack of 1899, large preparations were made in 1903, resulting in a catch of 4,300,000. The traps were numerous and the number of gill nets was exceeded only in 1900 and 1901. It is not surprising 780 BULLETIN OF THE BUREAU OF FISHERIES therefore that so large a catch was made at so low a level of abundance, or that the abundance had declined an additional 47 percent by 1907. In 1907 only three-quarters as many traps and one-half as many gill nets were employed as in 1903. The catch of 1,700,000 doubtless permitted a larger escape- ment than in 1903. This is reflected by a slightly increased abundance in 1911. In 1911 the number of traps remained about the same as in 1907 and the gill-net intensity was slightly lower, yet the yield was larger, being 2,200,000. According to the combined index of abundance there was a fall of 39 percent between 1911 and 1915, but this figure is undoubtedly too large. The trap index for 1915 was the lowest of the whole 39 years, but that it was so low chiefly on account of the failure of the run to pass by the traps is shown by the gill-net catch. This was nearly twice that of 1911, or about what one would have expected if the number of sockeyes reaching the gill nets in 1915 had been somewhat comparable to the number reaching them in 1911, as the gill-net fishing effort was about twice that in 1911. Since the number removed by the traps before reaching the gill nets was much greater in 1911 than in 1915 it is probably true that the 1915 level of abundance was slightly lower than that of 1911. Between 1915 and 1919 the abundance declined another 30 percent, according to the combined index, and probably more if the 1915 level were higher than shown. The spawning ground reports claim that in 1915 fewer sockeyes passed through Hell’s Gate to the spawning grounds of the upper Fraser than in any year since observations were started in 1901. On the other hand, the number spawning in the tributaries below the canyon, Lillooet Lake, Harrison Lake, Cultus Lake, Pitt Lake, etc., was estimated as being the largest for some years, even including 1913. Because of the failure of the traps to take many sockeyes, the total catch of 1915 was but 1,800,000. Considering the catch of 1915 in relation to the abundance, it does not appear to have been sufficiently large to have been the sole cause of the drop in 1919. Rather, it would appear that the extremely cold weather early in 1916, when the eggs deposited in 1915 were incubating (see table 35), had some part in it. The temperatures pre- vailing early in 1916 were even colder than in 1913. The reason for tnis second instance not showing as great a fall in abundance as in the first instance, when the temperatures were not quite as low, probably lies in the fact that in 1912 by far the larger portion of the spawning escapement went to the lakes above Hell’s Gate, in 1915 most of the spawning was below Hell’s Gate where it would not be affected by the cold temperatures of the upper Fraser. This is borne out by the 1919 escapement estimates, which for the region below the canyon were as high as in 1915, whereas practically none were found above the canyon. The survey was more thorough than usual and the dearth of up-river fish was very marked. In 1923 the abundance level was about on a par with 1919. There were only two-thirds as many traps and slightly fewer gill nets than in 1919, resulting in a eaten of 850,000 compared to 1,250,000 in 1919. Since 1919 was able to bring back a comparable run in 1923 witn a larger catch it is not surprising that 1927 showed a much improved condition. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 781 In 1927 both the trap- and gill-net fisheries were slightly more intense than in 1923. The purse seine boats were also more numerous. The net result was a catch of 1,800,000 in 1927 against 850,000 in 1923, and, as might be expected, the level of abundance fell off somewhat in 1931. COHO SALMON By George B. Kelez INTRODUCTION Ascending almost every stream and river of the region on their spawning migra- tions, cohos are the most widely distributed salmon present in these waters. Although suffering a severe decrease in numbers in recent years, they have formed a considerable portion of the catch throughout the history of the salmon fishery. This species provided the bulk of the pack of the first Puget Sound cannery and of the establishments which immediately succeeded it in that district. They formed the major portion of the catch of the natives resident at Neah Bay when fishery operators first visited that region in quest of new supplies of salmon. The catches of the early type of purse seines were composed almost entirely of cohos, and they have provided the chief source of the seiner’s income in off years up to the present time. This species is also the principal salt-water catch of summer vacationists and recreational fishermen throughout the region. The first coho catciies of the season are made during the early summer by the troll and purse-seine fleets operating in the waters off Cape Flattery, and on Swiftsure Bank. Great schools of immature fish feed there at that time, and large catches are common for a period of several weeks. In late summer the adult cohos begin their migration through the inner waters of the region to the tributary rivers where they will spawn, and the major part of the commercial catch is made during the period of this migration by traps, seines, and gill nets. LIFE HISTORY SPAWNING The majority of the mature fish enter fresh water during the months of October and November, although some may run as early as September, and a few individuals may tarry in salt water until the latter part of January. Actual spawning usually begins a week or two after the fish first enter the streams, and often extends through- out the winter months. Some of the salmon hatcheries in the region have con- tinued to strip eggs up to the middle of March, but most of the natural spawning has terminated before that date. In general, late spawning is confined to the smaller, shorter streams. Active and highly adaptive to different conditions, coho salmon may spawn on suitable gravel beds only a few miles from salt water, or may ascend the larger rivers to tributary streams in the mountains which surround the region. Such varia- tions in time and locality of spawning cause considerable differences in the time of hatching of the eggs and in the growth of the fry. 782 BULLETIN OF THE BUREAU OF FISHERIES GROWTH The time of hatching of the eggs depends on temperature conditions, but usually occurs during the early spring. The greater part of the young fish remain in the streams throughout the summer and the following winter, and usually migrate to salt water early in their second year. Growth in fresh water is quite rapid, especially in the streams of southern Puget Sound where temperatures are favorable and food is plentiful. In these streams the fry usually have attained a length of approximately 30 mm by early March, whereas those in the more northerly part of the region may not reach this size until the latter part of May. By September the size range of the southern fingerlings is from 60-70 mm. Collections of fish in their second year, taken in early March, show a size range of from 80-95 mm. By early May these fingerlings measure from 100-130 mm. During spring and early summer the fingerlings migrate from the upper reaches of the rivers to the estuaries, and finally into salt water. Scale collections from these populations indicate that the majority of the fingerlings migrate to salt water during the early spring freshets, but that many remain in the streams for a much longer period of time. After reaching the inner waters of the region, young cohoes may be found in large schools for a period of several weeks. At this time they have reached a size of from 14-20 cm. The greater part of these young fish gradually migrate to the waters of the Pacific Ocean. Clemens (1935) states that tagging experiments have indicated that some of the cohoes never leave the Strait of Georgia. Sport-fishing catches in the lower sound confirm the presence of cohoes there throughout all stages of their life in salt water. These fish remain in salt water during the second winter of their life, and through- out the following summer, during which time they experience a remarkable increase in size. Gilbert (1913) reported the cohoes at the cape to average 13.35 fish per case on July 23 and 7.56 fish per case on September 2. Smith (1921) stated that the average weight of cohoes taken by troll ers in the same region increased from 5.63 pounds on July 8 to 9.75 pounds on September 2. Recent samples from the com- mercial catches taken in the inside waters of Puget Sound during October indicate a size range from 5.13-14.90 pounds, and an average weight of 9.47 pounds at this time. Individual fish of more than 20 pounds in weight have been taken by sport fishermen in this region. Some indications of the migrations of cohoes in inside waters are given by tag- ging experiments reviewed by Clemens (1930). Recoveries were made of forty- seven immature cohoes tagged in 1927 at Deep Bay, in the northern part of the Gulf of Georgia. Of these, 29 were recovered north of the point of tagging, or on the lower coast of Vancouver Island, 3 were recovered in the Fraser River, and 1 in the nearby Capilano River. Approximately 30 percent were recovered in Puget Sound, some being taken as far south as Whidbey Island. From a similar experiment at Nanaimo in 1928, 163 recoveries were made. Of these, 34 were taken north of Nanaimo and 34 in the general vicinity of the tagging, 43 were taken in the Fraser River and vicinity, 8 were taken in the Strait of Juan de Fuca, or west of it, while 44, approximately 27 percent, were taken in Puget Sound. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 783 Of these latter recoveries, 15 were in the vicinity of Whidbey Island and in the Skagit River. These results would indicate that some individuals of the southern runs must either remain in the Gulf of Georgia during their life in salt water, or migrate inside of Vancouver Island on their return from the sea to the streams where they will spawn. AGE AT MATURITY Pritchard (1936) reported that commercially caught fish, secured for tagging experiments along the British Columbia coast during the years 1927-31, ranged in age from 2 to 4 years, but that 97.89 percent of these fish were in their third year. A small number of grilse, almost entirely precocious males, returned to the streams in the fall of their second year. Fraser (1920) reported that, of 2,000 cohoes examined from the Gulf of Georgia in 1916, all but 28 were in their third year, and that these 28 fish were all males in their second year. Gilbert (1912) reported a very few “sea-type” scales, from fish which have descended to salt water during then1 first summer, in his collections from Puget Sound. Pritchard reported 0.35 percent of this type of scale in his collections. It is reasonable to expect considerable fluctuations in the size of the runs of any species of which a high proportion of the individual fish mature at the same age. For those salmon which descend to salt water shortly after hatching, a considerable spawning escapement, combined with favorable conditions on the spawning grounds, often results in an extremely high return at maturity of that particular brood. That coho salmon, which mature almost entirely at 3 years of age, have not ex- perienced any sudden increase in numbers may be largely due to the fact that they have a long stream residence during their early life history. Because the carrying capacity of streams is physically limited, and there exists a considerable competition between the young stream-dwelling salmon and resident trout or other species, the numbers of fingerlings surviving until they begin their seaward migration cannot be increased beyond a certain point, even in very favorable years. Although this factor has doubtless had considerable influence in preventing large increases in numbers of coho salmon, the existence of so many populations in various streams has conversely aided in averting any sudden decrease in abundance, hence fluctuations in the numbers of this species have never been violent. INDIVIDUALITY OF POPULATIONS That the populations of different streams tend to be individual in nature is sup- ported by some experimental evidence. Gilbert (1913) reported the return in the fall of 1911 at Scotts Creek, California, of several coho salmon grilse from fingerlings marked there during the preceeding winter; no data as to returns of mature fish from this experiment were published. Fraser (1921) reported the recovery in Cowichan Bay, on October 11, 1917, of 1 coho salmon from 1,000 fry marked at the Cowichan Lake hatchery in March 1915. Pritchard (1936) reported the recovery in 1927 of 19 adult cohos in Cultus Lake, B. C., from 72 fish marked there during the spring of the same year. These fish were in the early part of their third year when marked, and returned as adults after having remained only a few months in the sea. 784 BULLETIN OF THE BUREAU OF FISHERIES During the spring of 1934, 26,000 coho fingerlings, averaging 47.4 mm in length, which were made available through the cooperation of the Washington State De- partment of Fisheries, were marked by the author at Friday Creek, a tributary of the Samish River. During the same month, 9,800 coho fingerlings, averaging 49.2 mm in length, were transferred from the Skykomish River and were marked and liberated in Friday Creek. In November of that year an additional 26,000 finger- lings from the same brood as the fish used in the first experiment were also marked and liberated at Friday Creek. This lot averaged 101.6 mm in length at the time of marking. Complete data on returns to the Samish River of six grilse from the third marking experiment were obtained during the spawning run of 1935, and the capture of two additional marked grilse was reported from a reliable source. The run of normally maturing three-year-olds appeared during the winter of 1936-37, and 480 marked fish were recovered from the Samish River, 7 from the first experiment, 11 from the second, and 462 from the third. No recoveries have been made from nearby streams or from the Skykomish River. From these results it would appear that mortality is much higher for the smaller fish, and that there is a definite tendency for mature cohos to return to spawn in the stream from which they migrated to the sea. LOCALITY OF CAPTURE BY DIFFERENT TYPES OF GEAR CATCHES IN VARIOUS DISTRICTS Cohos have been second in demand only to kings for consumption as fresh fish, and large quantities have always been used in local markets. Because of their suit- ability for freezing they have surpassed all other species as a supply for the consider- able demand of cold-storage units which have maintained an active market since the earliest years of the present century. For these reasons the canned-pack figures for this species are an unreliable measure of the commercial catch in past years. Al- though they have been the mainstay of the cape purse-seine fishery throughout its history, Gilbert (1913) reporting over 850,000 cohos taken there as early as 1911, and have formed the major part of the offshore catch of trailers, no records of the high-seas catches have been kept for other than very recent years. It is impossible without thorough tagging experiments to determine the propor- tion of the cape catch provided by the populations of the Puget Sound-Fraser River region. Because of their widespread range of operation, part of the trailer’s catch landed in Washington may well be drawn from other sources. The purse seiners, however, are usually concentrated in the area off the entrance to the Strait of Juan de Fuca, and their catch doubtless consists mainly of the populations from the region. We may infer, from the far greater size of the runs entering the Strait of Juan de Fuca than of those conceivably passing the Banks en route for any other nearby district, that the major portion of the catch there is drawn from the regional popula- tions. In Puget Sound the trap fishery usually suspended operations in early years before the coho run had begun, except in the inside waters where the catch consisted SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 785 mainly of this species. However, from the time that the fishing season of the north- ern traps was increased to include the fall runs up to the last decade, traps took the major part of the cohos caught in Puget Sound waters. In late years purse seines have become the chief source of this species. The major part of gill-net catches in the estuaries of such rivers as the Skagit and the Snohomish have been coho salmon. Although considerable catches of coho salmon have been made on the Fraser River, especially in years when sockeye were not abundant, fall fishing has never been prosecuted as strenuously in that district as in the Puget Sound region. Except for recent years data are not available for catches other than in a portion of the region, hence it is not possible to present complete figures for coho salmon production prioi to 1926. During this latter period the catch has been considerably smaller than in previous years. The total catch of coho salmon for Swiftsure Bank, Puget Sound, and the Fraser River, by various types of gear from 1926-34, is pre- sented in table 36. LOCALITY OF TRAP CATCHES Because of the mobile nature of the purse-seine fleet, the determination of the particular district of the region in which their catches were made is not possible from past records. The best indication of the coho production of specific localities may be had from a consideration of the catches of the traps located therein. The total catches of traps in restricted areas are presented for the period from 1896-1934 in table 37. Most of the areas in this table may be readily located from figures 2 and 3. “Lower sound” includes the water south of Useless Bay on Whidbey Island. “Miscellaneous” includes such inner bays as Bellingham, Padilla, and Samish, as well as Possession Sound and Hood Canal, but four-fifths of these fish were from the waters south of Point Wilson. Table 36. — Catch of coho salmon, 1926-84 Year Fraser River catch 1 Puget Sound traps Purse seines Trollers Puget Sound gill nets Minor Puget Sound gear Total, all gear Puget Sound High seas Puget Sound High seas 1926.. 120, 663 384, 600 232, 721 375, 000 22, 269 325. 000 57, 436 6, 266 1, 523, 955 1927 226, 710 536, 937 354, 976 188, 750 23, 491 400, 000 108, 360 5, 051 1, 844, 275 1928 203, 580 436, 819 236, 085 195, 844 18, 538 339, 311 65, 092 4, 163 1,499, 432 1929. 334, 467 397, 381 319, 847 432, 095 19, 331 329, 026 61, 757 8, 655 1, 902, 559 1930 71,280 285, 310 204, 692 407, 405 15, 589 355, 040 65, 228 4, 125 1,408, 669 1931. 79, 2.54 241, 873 449, 081 225, 798 6, 655 267, 916 40, 527 1,099 1,312,203 1932.. 160, 452 102, 727 331, 565 315, 290 3. 457 281. 686 22, 240 1,262 1, 218, 679 1933 125,883 244, 755 248, 686 174, 728 4,922 176, 529 35, 421 2, 194 1,013,118 1934 113, 382 164, 504 233, 418 365, 380 12, 709 261, 804 40, 038 507 1, 191,742 Total... 1, 435, 671 2, 794, 906 2, 611,071 2, 680, 290 126, 961 2, 736, 312 496, 099 33, 322 12, 914, 632 ■ Converted from cases at 9 fish per case, does not include cohos caught elsewhere in the Gulf of Georgia and canned on the Fraser River, or Fraser River cohos used for purposes other than canning. 786 BULLETIN OF THE BUREAU OF FISHERIES Table 37. — Annual catch of coho salmon by traps, in different areas, 1896-1934 1 Year 1896. 1897. 1898. 1899. 1900. 1901. 1902. 1903. 1904. 1905. 1906. 1907. 1908. 1909. 1910. 1911. 1912. 1913. 1914. 1915. 1916. 1917. 1918. 1919. 1920. 1921. 1922. 1923. 1924. 1925 1926. 1927. 1928. 1929 1930. 1931 1932. 1933. 1934. Total... i Incomplete before 1915. Out of a total catch in all areas of nearly 14 % million fish, approximately 4% nhllion were taken in the northern part of the region, 2}i million in areas through which the populations of both northern and southern districts migrate, and more than 7 million in the southern areas of the region. Of the latter total, more than 5 / million fish were taken from Admiralty Inlet alone. SEASONAL OCCURRENCE IN VARIOUS AREAS The general seasonal occurrence of coho salmon from different types of fishing gear has already been presented. However, as might be anticipated in the case of migrations of populations from widely scattered streams, occurrence vaiies consider- ably in different districts. These variations do not seem to be correlated with the dis- tances which the fish must travel along their migration routes, but appear to depend largely upon the characteristics of the individual populations. Because the traps sample individual runs in exact localities, their catches were used as the best measure of seasonal occurrence in various portions of the inner waters of the region. Area Point Roberts and Bound- ary Bay Sandy Point to Bound- ary Bay Lummi Island and Rosario Strait Haro Strait and Wal- dron Island Salmon Banks and South Lopez Hope Island and Skagit Bay West Beach West of Point Wilson Admi- ralty Inlet Lower Sound Miscel- laneous Uni- denti- fied Total 1,632 2, 170 61, 753 1,270 20. 095 3, 181 1,020 5, 540 751 7,708 9, 296 33, 559 18. 096 226 62, 782 49, 647 64, 851 7,808 18, 355 67, 608 36, 093 80, 524 85, 164 86, 545 31, 581 31, 550 53, 589 48, 621 63, 090 60, 440 55, 973 76, 058 75, 948 66, 880 19, 047 28, 247 2, 693 15. 845 15, 170 27, 938 22, 375 19,913 2, 185 6, 570 655 8, 457 16, 093 15, 135 12, 835 51, 504 65, 523 107, 906 52, 223 35, 299 21, 169 45, 389 27, 046 38, 972 64, 159 52, 551 5, 214 11, 213 22, 400 24, 171 27, 432 25, 616 16, 808 34, 756 43, 134 32, 631 9, 549 31, 146 3, 199 14, 874 3, 184 31,000 42, 107 6,893 22, 171 19, 576 3,841 409 8, 622 10, 421 14,919 18, 496 61, 876 82, 728 105, 140 54, 671 42, 682 49, 846 64, 972 38, 741 37, 671 100, 167 62, 028 20, 693 25, 821 69, 311 50, 049 46, 974 32, 738 33, 617 47, 234 48, 130 56, 387 20, 581 23, 354 12, 374 17, 421 22, 162 1,607 20, 909 33, 588 24,811 5, 476 49, 113 10, 696 66, 102 49, 020 155 690 12, 529 1, 134 6, 611 60, 516 30, 035 40, 062 40, 382 42, 307 7, 759 24, 308 16, 857 22, 478 18, 290 17, 231 17, 993 17, 772 9, 270 10, 574 6,708 6, 165 2,015 6, 411 3, 710 32, 832 46, 204 52, 919 52, 756 27, 486 16 29, 087 23, 065 38, 533 14, 446 70,061 64, 552 46, 124 35, 583 3, 871 9, 711 16, 642 21, 317 19, 278 66, 632 46, 442 14, 970 9,548 9,407 10, 502 20, 142 15, 301 12, 172 19, 334 22, 203 11,814 5, 372 4, 594 3,717 14, 908 6,508 38, 378 41, 600 39, 710 28, 181 27,818 33, 952 27, 065 42, 785 35, 329 54, 815 7, 768 1, 659 34, 963 38, 730 52, 355 34, 634 46, 614 61, 023 22, 457 44, 639 99, 831 49, 884 41, 222 30, 144 59, 197 19, 470 28, 431 49, 278 32, 895 22, 652 16, 292 16, 380 13, 062 54, 361 26, 628 25, 587 13, 499 24, 405 18, 566 15, 272 9,544 27, 432 45, 012 51, 321 43, 073 52, 952 37, 215 64, 693 47, 249 18, 802 31, 881 51, 921 39, 015 31, 106 46, 467 50, 982 20, 845 14, 553 17, 377 21, 680 36, 097 13, 452 16, 232 31, 764 23. 469 21, 19S 27, 784 17, 883 2, 869 14, 590 8,442 15, 459 31, 894 83, 830 55, 274 36, 590 13, 398 27, 289 9, 163 12, 761 10, 190 15, 796 1, 667 3, 295 9,612 4, 059 6, 171 13, 825 37, 683 21, 895 17, 382 12, 536 11,896 10, 256 6, 390 23, 829 8, 354 43, 329 152, 757 211,079 142, 348 9, 621 127, 130 221,547 266, 710 234, 528 165, 302 258, 668 148, 860 254, 102 190, 532 143, 723 64, 539 205, 632 118, 047 117, 263 173, 963 184, 763 99, 898 103, 691 105, 081 186, 609 203, 304 186, 083 134, 455 211,748 130, 361 129, 908 102, 317 72, 879 36, 107 108, 676 47, 019 10, 500 14, 914 5, 310 22, 752 15, 517 16, 465 22, 741 27, 083 11,400 64, 422 6, 041 19, 619 16, 981 12, 527 13, 579 21, 844 6, 948 6, 992 11, 677 13, 163 "29, "846 878 16, 047 11, 467 20, 148 38, 064 10, 584 24, 629 32,061 31, 332 40, 132 51, 516 23, 617 17, 534 7, 338 35, 498 16, 169 19, 215 11, 865 11, 672 6, 034 9,289 8, 464 14, 939 3, 490 12, 837 8, 729 265 1,034 4, 048 13, 340 65 32, 052 60 11, 080 20, 621 6, 792 4, 190 1,635 443 1, 500 192 524 14, 959 87. 531 88. 381 213, 947 325, 158 452, 042 300, 972 64, 250 187, 618 341, 730 465, 395 437, 994 348, 333 538, 227 572, 402 787, 155 558, 744 313, 232 275, 102 641,881 424, 850 461, 903 709, 641 642, 014 260, 583 298, 889 466, 046 445, 622 505, 137 423, 814 386, 389 536, 739 407, 928 393, 751 282, 840 249, 306 102, 498 245, 688 151,335 1, 370, 406 999, 224 1, 405, 823 683, 284 898, 049 1, 180, 151 1, 128, 280 500, 494 5, 282, 958 336, 933 526, 557 97, 867, 14, 410, 026 SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 787 Data were available from 26 traps which fished throughout the duration of the coho run in most of the years from 1911-34. Districts selected (see figs. 2 and 3) and number of traps were as follows: Point Roberts 2, Boundary Bay 4, Birch Bay 5, Rosario Strait 4, Dungeness Spit and Middle Point 3, Admiralty Bay 3, Bush Point 3, and Hope Island 2. The total number of fish included in the catches of these traps was 5,652,592. From these data the average proportions of the season’s catch taken in each 7-day period by the traps in various districts were calculated. These figures are presented in table 38. Because of the essential similarity in occurrence, Point Roberts and Boundary Bay have been grouped, as have Admiralty Bay and Bush Point. Table 38. — Seasonal occurrence of coho salmon from traps; proportion of total catch taken in each 7-day period Area Week ending — Point Roberts Bound- ary Bay Point Roberts and Bound- ary Bay Birch Bay Rosario Strait Dunge- ness Spit and Middle Point Admi- ralty Bay Bush Point Admi- ralty Bay and Bush Point Hope Island All dis- tricts May 5 0.062 0.002 0.056 0.044 0.018 May 12 0.090 .038 .202 .024 .056 .035 May 19__ 0.822 0.871 0.046 .124 . 105 .216 .034 .071 .059 May 26 .554 .588 .024 . 100 .077 .138 .041 .073 .054 June 2 .496 .526 .327 .077 .068 .157 .072 .107 .084 June 9 .265 .281 .348 .073 .070 . 131 .069 .094 .080 June 16 .148 0. 162 . 154 .297 .121 .072 .224 . 103 . 151 . 103 June 23 .143 3.703 1. 207 .396 .114 .086 .494 .158 .298 . 175 June 30. .461 1.058 .627 .343 .132 .127 .494 . 132 .302 0.001 .174 July 7 .277 .649 .414 .695 .174 .261 1.273 .227 .757 .003 .351 July 14 .207 1. 944 .937 .676 .370 .451 .867 .447 .667 .003 .393 July 21 .526 1. 460 .946 .940 .636 1.214 .730 .405 .601 .057 .466 July 28.. .451 1.610 1.001 1.210 .784 1.781 .599 .565 .583 .081 .532 Aug. 4 1.053 1.358 1. 171 1.367 1. 127 2. 600 .613 . 745 .681 .421 .709 Aug. 11 .518 1.464 .962 1. 810 1.484 3.684 .790 .817 .808 1.825 .902 Aug. 18 .517 1. 512 .985 1.880 1.915 3. 563 1.269 1.540 1.412 3. 178 1.413 Aug. 25 2.018 2.373 2. 132 2.997 3.998 4. 938 2. 853 2. 809 2. 829 4.457 2.717 Sept. 1 5.068 5. 117 4. 950 6.393 6.091 7. 840 3.732 4. 026 3.891 5.366 3.911 Sept. 8. 11.926 6.476 9.020 7. 805 10. 671 8. 335 6.955 6. 191 6.495 7. 853 6. 953 Sept. 15 12. 540 14. 586 13. 173 13.341 13. 462 7. 360 10. 041 9. 459 9.689 10. 329 10. 795 Sept. 22 15. 238 16. 145 15. 262 10. 382 13.845 11.857 13. 724 13. 382 13. 485 14.373 12. 652 Sept. 29.. 15. 194 11.935 13.513 9. 271 11.967 12. 869 13. 004 13.411 13. 200 12. 789 12. 129 Oct. 6... 9. 737 8. 276 9. 127 9. 126 9. 950 11.686 14. 033 14.038 13. 991 14. 159 12.628 Oct. 13 8. 501 9. 454 8.819 10. 385 10.211 10. 076 12. 091 12. 479 12. 279 12.351 11.978 Oct. 20.. 4. 209 7. 579 5. 221 9. 648 6. 339 4.894 6. 894 9.025 8.091 6. 629 8.357 Oct. 27.. 6. 042 2. 681 5.315 7. 727 4. 149 4. 041 3.897 4. 862 4. 441 2. 993 5.313 Nov. 3_ 2. C31 .459 2.314 1.756 1. 553 1.845 2. 256 2. 290 2. 270 1.362 3.108 Nov. 10 .457 .484 .810 .441 1. 262 .939 1.061 .914 1.628 Nov. 17 1.053 1. 000 1.000 .546 1.502 Nov. 24 .547 .528 .301 .667 Dec. 1 .044 .043 .008 .051 Dec. 8 .003 .003 .005 Total.. 99.999 100.001 100. 000 100. 000 99.998 100. 000 100. 000 100. 000 100. 001 99. 999 100. 002 Comparing the data for the Point Roberts-Boundary Bay area with those for Birch Bay, occurrence in both areas is slight until the latter part of August, when the runs increase abruptly in size. The run at Point Roberts and Boundary Bay increases steadily to a peak in the week ending September 22, and abundance decreases mate- rially thereafter, with minor fluctuations, to the end of the season. Birch Bay clearly shows the presence of the same run as that of the former area, but occurrence is distinctly bimodal. The peak of the main run occurs in the week ending September 15, after which there is a definite decrease in numbers. A second run of smaller pro- portions follows, reaching its peak between October 6 and October 20, the run falling 71941 — 38 7 788 BULLETIN OF THE BUREAU OF FISHERIES off abruptly thereafter. The main portions of the runs of both areas are probably contributed by the Fraser and other northern rivers, but the second peak in the Birch Bay area may be composed largely of populations of such rivers as the Nicomekl and the Serpentine. A comparison of the data for the Point Roberts-Boundary Bay area with those from Rosario Strait indicate that the run in the latter area corresponds closely with that of the more northern dis- tricts; the somewhat earlier appearance here is probably due to the lesser distance of migration from the sea. There is a strong indication that a large part of the runs passing through Rosario Strait continues to Boundary Bay without entering Birch Bay. The Rosario Strait data are shown graphically in the lower section of figure 26. A comparison of seasonal occurrence of the Dungeness Spit-Middle Point area with that of the Admiralty Bay- Bush Point area indicates that the early appearance of cohos in the former area is consonant with its more sea- ward location. In this area a relatively heavy run follows the first group of fish, appear- ing during late August and early September, after which the intensity slackens. This is followed by the main run, Figure 26.— Seasonal occurrence of coho salmon in trap catches from the southern which reaches its peak in the part of Puget Sound. In the lower section of the figure occurrence in the principal w0ek in September and southern area is compared with that of one of the northern areas. R drops abruptly after the second week in October. It is evident that traps in this area fish a mixed popula- tion, but that the main run consists of fish bound for the southern areas. The Admiralty Bay-Bush Point run increases steadily from the last week in August to the third week in September, remains at a high level for three more weeks, and decreases steadily thereafter to the middle of November. Unlike the other areas, the run does not terminate here, but continues at a low level until early December. The data for seasonal occurrence in these areas is presented graphically for comparison in the upper section of figure 26. JUNE II 25 6 22 AUG. SEPT. WEEK ENDING SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 789 In the lower section of figure 26, occurrence in the Rosario Strait area is com- pared with that of the Admiralty Bay-Bush Point area. Although both of these areas are immediately adjacent to waters in which their individual runs mingle, seasonal occurrence of cohos in Rosario Strait is considerably earlier than in the southern area, in fact the run in the former area has begun to decrease almost at the time that the southern run has first reached its peak. The Admiralty Bay-Bush Point runs also show a more prolonged peak of occurrence, and continue much later in the season. From table 38 it is evident that the Hope Island run is almost identical in occur- rence with that of the Admiralty Bay-Bush Point area, although the small catches in the early and late portions of the season appear only in the latter area. It is appar- ent that a large proportion of the Skagit River runs must pass around the southern end of Whidbey Island in the course of their migration. CHANGES IN ABUNDANCE CALCULATION OF TRAP INDICES It is apparent, from both trap and purse-seine catches, that there has been a considerable diminution in abundance of cohos in recent years. Inasmuch as traps and purse seines have been the principal types of gear catching coho salmon in this region, trends of abundance were determined from catches by both types of gear, and are presented together for comparison. In measuring abundance from traps, several difficulties are encountered which arise from the lateness of the coho runs in relation to those of other species taken by this type of gear. During early years, in certain areas where other species formed the principal catch, the traps were often removed from the water after fishing during only part of the coho season. Closed periods, which were imposed through legislation in many of the years after 1920, also prevented the traps in certain areas from fishing during the entire coho run. Years in which these closures were enforced cannot be compared directly to those in which fishing was not restricted unless some provision is made to offset the shorter fishing period. In most early years the traps were per- mitted to fish well into the winter months, while in later years legislation has often terminated the season before the entire run has appeared. For these reasons it was impossible to use the catches of any trap unless the opening and closing dates of its annual fishing seasons were known. This requirement sharply curtailed the available amount of data. In order to make the annual catch data comparable for both early and late years, they were weighted according to the length of the period fished. Inasmuch as the coho runs in the various districts are quite uniform in time from year to year, the average seasonal occurrences already presented were used as a basis for determining the time period of the runs in their respective districts. November 10 was arbitrarily selected as the end of the fishing season. The catches of traps which fished later in the year were reduced in proportion to the per- centage occurrence of the run after that date, and catches of traps which ceased 790 BULLETIN OF THE BUREAU OF FISHERIES fishing before that date were similarly increased. For 1921, and for other late years in which closed periods have been in force, trap catches were increased by the average percentage occurrence during the closed periods in their respective areas. Catches of traps which fished for a lesser period of time than that in which 75 percent of the run for their district normally occurred were not included in the analysis. A certain amount of error is unavoidably introduced by empirically increasing or decreasing catches for particular years to compensate for irregular length of fishing season. However, catches which were decreased in size were confined almost entirely to early years when fishing was less restricted, and nearly all increases in catches were made in later years when closed periods were imposed and fishing seasons were short- ened by legislative action. Such error as may have accompanied these necessary corrections would tend to reduce the apparent level of abundance in early years and to increase it in later years. Any decline shown in the trend of abundance would thus be given added validity. Three particular districts were selected for analysis. The first was that extending from Sandy Point to the international boundary, and included the Birch Bay, Bound- ary Bay, and Point Roberts areas (see fig. 2). Because of the size of the district and the large number of traps situated therein, catches were used from all traps for which suitable data were available. Prior to 1910 the data were meager, for sockeyes were of such importance in this region that catches of other species were often not recorded. After the tremendous sockeye run of 1913 most of the traps were removed before the coho run. In 1932, unfavorable economic conditions sharply reduced the number of traps fishing. During the remaining years, suitable data for from 7-12 traps were available. These traps, although but a small part of the total number fishing in the area, represent a considerable portion of those which were fished late enough in the season to intercept the coho migration. The number of traps available in this area, and their total catch for each year, are tabulated in the first two columns of table 39. The second area selected was Rosario Strait. Although the runs in this district are largely composed of the same populations which pass through the northern areas, fishing conditions differ considerably, for the area of water through which the runs must pass is much more restricted and the number of traps is very small. Three of these traps, located in strategic positions, have taken the bulk of the catch in this area. Data for the 16-year period, from 1919-34, when at least two of these traps fished every year, all three of them for fifteen years, are tabulated in the third and fourth columns of table 39. It is evident that the efficiency of Rosario Strait traps is greater than that of those in the northern area, and their index of abundance should provide a useful check on that calculated from the larger group. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 791 Table 39. — Indices of abundance of coho salmon from traps Year Data by areas Index figures North of Sandy Point Rosario Strait • Admiralty Inlet North of Sandy Point Rosario Strait Admiral- ty Inlet Number of fish Number of traps Number of fish Number of traps Number of fish Number of traps 1900 122, 723 2 2. 166 1901 201, 962 2 3. 544 1902 113; 063 2 2. 071 1903 1904 140, 004 2 2. 558 1905 26, 353 2 195, 469 4 4. 409 1. soe 1906 33, 066 3 248 189 4 3. 923 2. 03e 1907 50, 087 6 180, 145 4 3. 828 1. 466 1908 47, 856 5 165, 054 3 5. 431 1. 644 1909 31, 174 i 56, 871 2 294', 398 3 9.888 8. 288 3. 195 1910 . 99, 579 10 64, 779 2 163, 185 4 4. 009 9. 238 1. 548 1911 _ 54, 261 10 77, 585 3 277, 837 4 2. 278 7. 813 2. 472 1912 113, 122 12 56, 977 3 205, 888 4 4. 083 5. 798 1. 701 1913 36, 503 5 47, 753 3 159, 062 3 2. 598 4. 787 1.449 1914 58, 640 10 32, 398 2 68, 166 3 2.328 4. 551 .718 1915 56, 881 12 30,114 3 131, 229 4 1.835 2. 988 .992 1916 50, 375 10 25, 440 3 70, 722 4 2. 245 2. 507 .517 1917 53, 562 12 23, 418 3 94, 769 4 1.838 2. 403 .754 1918 108, 710 12 69, 690 3 116, 083 5 3. 433 6. 812 .811 1919 73, 488 . 12 44, 131 3 122, 723 5 2. 461 4. 473 .862 1920 44,597 10 21, 485 3 69, 010 4 1.852 2. 199 .731 1921 59, 820 12 14,844 2 78, 967 3 2. 143 2. 330 1.297 1922 86, 589 7 46, 791 2 61, 253 2 6. 332 6. 740 1.715 1923 70, 529 12 32, 619 2 166, 608 6 2. 672 4. 950 1.017 1924 92, 276 10 25, 802 2 166, 520 6 3. 978 3. 748 1.044 1925 57, 501 12 21, 950 3 165, 477 5 1.852 2. 191 1.078 1926 49, 163 11 20, 396 3 99, 946 5 1.898 2. 066 .690 1927 56, 853 12 26, 584 2 186, 117 6 1.879 3. 907 .996 1928 51,430 10 23, 784 2 104, 412 6 2. 009 3. 510 .561 1929 63, 629 12 29, 402 2 115, 487 7 2.291 4. 122 .608 1930... 26, 169 u 10, 108 2 81, 478 7 .933 1.434 .443 1931 28, 659 n 13, 470 3 58, 381 4 1. 182 1.381 .722 1932 1, 184 3 19,919 3 33, 155 2 . 175 1. 951 .710 1933 23, 828 7 14, 026 3 106, 083 7 1.342 1. 350 .578 1934 21, 781 8 19, 317 3 54,639 5 1. 151 1. 847 .388 Total .. 1, 633, 665 869, 653 4, 618, 204 The third area selected was Admiralty Inlet (see fig. 3). Here, as in Rosario Strait, the runs are concentrated while migrating through a restricted passage, and the effectiveness of most of the traps is correspondingly great. Catches of from two to seven traps were available each year from 1900-1934, with the exception of 1903. The number of traps, and their corresponding total catches by years, are tabulated in the fifth and sixth columns of table 39. The calculation of index figures from these data is complicated by the fact that only a few traps in each district have fished continuously throughout this period of years. There is a wide variation in the fishing effectiveness of different traps, and any determination, such as the average annual catch per trap, must be affected considerably by the proportions of efficient and inefficient traps fishing each year. In order to minimize this variation it was necessary to weight the catches in such a manner that the relative annual change in the average catch of each trap might be measured irrespective of the actual sizes of the catches from which the average was derived. Such weighting was accomplished for each district by selecting a group of traps which had fished in the same years over a long period of time, and from which the relative effectiveness of all traps in that district might be measured. From the sums of the annual catches of these traps the average annual catches were calculated ; 792 BULLETIN OF THE BUREAU OF FISHERIES each of these was then determined as a proportion of the average annual catch of the base group. The average annual catch of any trap which fished for a lesser period of years than did the standard traps was determined as a proportion of the average of the total catches of the base group for the same years as those in which that par- ticular trap fished. The annual catches of the traps were then divided by the proportional weights of the same traps, and the average of the resultant figures for any 1 year is the index figure for that year. The index figures for the three areas, tabulated in the last three columns of table 39, are not directly comparable as they now appear, but measure only the degree of change from year to year in the individual areas. The relative changes in the different areas will be considered in conjunction with the index de- rived from purse-seine catches. CALCULATION OF PURSE-SEINE INDEX Inasmuch as a considerable portion of the coho salmon taken in Puget Sound waters have been caught by means of purse seines, a determination of changes in abundance of the species based on purse-seine data provides a valuable comparison with the indices from trap catches. The purse-seine index is similar to those derived from traps in that it is a measure of relative variation, from year to year, in the average catch of a unit of fishing effort. However, its construction is materially different in that the total seasonal catches of individual vessels are unknown, hence the size of the average delivery was used as the unit of measurement instead of the annual catch. In order to eliminate the influence of deliveries made by the vessels fishing for other species of salmon than coho, only such deliveries as were made between Sep- tember 2 and October 20 were included. Data were also limited to vessels of more than 9 net tons and less than 40 net tons. This restriction excluded both the very small vessels, which were not regular purse seiners, and the largest vessels, which fished on Puget Sound only occasionally in the fall. Since the average delivery of the small vessels operating in early years could not be compared directly to that of the large-sized, modern vessels, the catches neces- sarily were weighted to compensate for the changes in efficiency. In determining the weighted average delivery of the fleet, the vessels of 10-14 net tons were considered as unity, and the weighted number of deliveries of vessels in larger size-classes were the product of their actual number of deliveries and the vessel efficiency of that par- ticular size-class, taken from table 15. For each year from 1911-34 the sum of the number of fish in the catches of all vessels in the fleet was divided by the weighted number of deliveries. The weighted average delivery figures represent the average catch in terms of one size-class of vessels, hence they are directly comparable through- out the series of years. These figures are presented in the last column in table 40. The other columns in the table show the same data broken down according to group- ings of vessels of various sizes, SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 793 Table 40. — Index of abundance of coho salmon from Puget Sound purse seines Indices from individual size classes Year 10-14 net tons 15-24 net tons Number of fish Number of deliveries 1 Average delivery Number of fish Number of deliveries Average delivery Weighted number of deliveries Weighted average delivery 1911 ... 5, 760 214 65 88. 62 1912 8 26. 75 11, 169 93 120. 10 no 101. 54 1913 - 4, 231 18, 784 12, 732 8,940 60 70. 52 8, 261 29,015 45, 495 105 78. 68 124 66. 62 1914 146 128. 66 255 113. 78 303 84. 16 1915 243 52. 40 606 75.07 788 57. 73 1916 266 33.61 26, 218 410 63. 95 533 49. 19 1917 5, 689 387 14. 70 18, 050 58, 685 754 23. 94 1,010 620 17.87 1918 27, 552 13, 846 2, 188 12, 637 266 103. 58 459 127. 85 94. 65 1919 247 56.06 28, 481 430 66.23 581 49.02 1920 66 33. 15 13, 808 45, 580 146 94. 58 196 70. 45 1921 154 82.06 477 95. 56 634 71.89 1922 7,999 6, 049 132 60.60 35, 216 31, 433 326 108. 02 434 81. 14 1923 128 47.26 435 72. 26 583 53. 92 1924 4,643 80 68. 04 23, 038 305 75. 53 400 57. 60 1925 11, 921 222 53.70 38, 578 553 69. 76 724 53. 28 1926 - 3,406 6, 962 122 27. 92 34, 477 498 69. 23 647 53. 29 1927 178 39. 11 35, 280 655 53. 86 858 41. 12 1928 8,174 14, 327 242 33. 78 70, 146 1,046 67.06 1, 360 51.58 1929 440 32. 56 74, 835 1, 489 50. 26 1, 995 37. 51 1930 6, 275 135 46. 48 25, 781 542 47. 57 726 35.51 1931 7,360 8, 255 253 29.09 54, 040 1,277 42.79 1,711 31.93 1932 157 52.58 58, 166 1, 072 54.26 1, 447 40.20 1933 - - 5, 545 334 16. 60 35, 987 1,600 22. 49 2, 144 16. 78 1934 4,486 240 18.69 36, 357 1,069 34. 01 1, 421 25. 59 207, 975 4, 571 838, 696 14, 602 19, 349 Indices from individual size classes Index from grouped size classes Year 25-39 net tons Number of weighted deliveries Number of fish Number of deliveries Average delivery Weighted number of deliveries Weighted average delivery Number of fish Weighted average 1911 1,024 10 102. 40 18 56.89 6, 784 83 81.73 1912 11, 383 13, 559 48, 329 58, 872 118 96. 47 1913 1, 067 530 9 118. 56 16 66.69 200 67. 80 1914 - 8 66. 25 15 35. 33 464 104. 16 1915 645 16 40. 31 29 22.24 1,060 821 55. 54 1916 - 1,345 12 112. 08 22 61.14 36, 503 35, 629 44. 46 1917 11, 890 14, 677 11, 833 211 56. 35 390 30. 49 1, 787 1,090 1,062 406 19. 94 1918 - - 110 133. 43 204 71.95 100, 914 64, 160 27, 308 92. 58 1919 - 125 94. 66 234 50. 57 51.00 1920 11,312 78 145. 03 144 78. 56 67. 26 1921 58, 882 40, 542 34 578 424 138. 87 780 75. 49 117, 099 1, 568 74. 68 1922 176 230. 35 326 124. 36 83, 757 892 93. SO 1923 295 117. 21 546 63. 33 72, 060 43, 889 82, 208 98, 465 97, 939 175, 706 180, 655 85, 653 1, 257 672 57. 33 1924 16, 208 103 157. 36 192 84. 42 65.31 1925 31, 709 60, 582 303 104. 65 661 56.52 1, 507 1,753 54. 55 1926 532 113. 88 9S4 61.57 56. 17 1927 55, 697 624 89. 26 1,167 47. 73 2,203 3,907 5,214 2, 321 5,290 4,664 6, 209 44.46 1928 97, 386 1, 239 78. 60 2, 305 42.25 44. 97 1929 91, 493 53, 597 1, 486 61.57 2, 779 32. 92 34. 65 1930 781 68. 63 1,460 36.71 36. 90 1931 95, 752 153, 377 1, 760 54.40 3, 326 28.79 157, 752 219, 798 107, 796 106, 012 29. 82 1932 1,619 1 974 94. 74 3,060 3,731 50.12 47. 13 1933 66, 264 33. 57 17. 77 17. 36 1934 65, 169 1,047 62. 24 1, 968 33.11 3, 629 29. 21 975, 559 12, 942 24,257 2, 022, 230 48, 177 1 Number of deliveries and weighted number of deliveries identical for this group, as efficiency weighting is unity. 794 BULLETIN OF THE BUREAU OF FISHERIES TRENDS OF ABUNDANCE The indices from traps and seines represent the relative availability of coho salmon to the particular type of gear and for the particular area in which that gear operated. The trap indices are for three individual areas, whereas the purse-seine index is necessarily based on catches made throughout the entire Puget Sound region. In order to show more clearly the trend of abun- dance of the species, the in- dices were smoothed once by threes. Since some 98 percent of these fish ma- ture at 3 years of age, such smoothing also minimizes the effect of any predomi- nate age-cycle. To facili- tate comparison of the curves, it was necessary to reduce the indices to the same general range of vari- ation, therefore each curve was proportionally reduced or increased so that the sum of the index figures for the years 1912-33 equalled 100.00. These smoothed indices are shown graph- ically in figure 27. In the three sections of the figure the indices from traps north of Sandy Point, from Rosario Strait, and from Admiralty Inlet are compared with the seine index. The same general trend is apparent in all the indices. A general high level of abundance is indicated prior to 1910, but the continued increase thereafter in numbers of all types of gear was accompanied by a decrease in abundance of the species. During the post-war depression, the considerable decrease in fishing effort throughout the region resulted in a general rise in abundance; this increase was quickly terminated, however, when fishing once more became profitable and the trend of abundance has declined from that point. The indices correspond fairly well throughout, and since the post-war years the seine index is very similar to that for Admiralty Inlet traps. This is a direct corollary of the more intense purse-seine fishery in the southern district during these years, ms 06 oq 15 18 YEAR 24 27 30 33 Figure 27.—Trends of abundance of coho salmon. Indices calculated from trap catches in three different areas are compared to the index calculated from purse-seine catches taken from the entire Puget Sound region. A considerable decrease in abundance has taken place since the early years of the fishery. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 795 throughout which this area has been the heaviest producer of coho salmon. It is further evident from these curves that the general level of abundance throughout the region has been lower in recent years than at any previous time in the history of the fishery. That changes in the intensity of the fishery have exerted a considerable influence on the abundance of cohos has been indicated. However, abundance has been fur- ther affected by changing conditions in the streams where the adult fish spawn and the young are reared. Lumbering has been, or is now being carried on in the drainage basins of almost every river in the southern part of the region, and most of the cut-over lands have been cleared for agricultural purposes. Rapidity of run-offs and resultant flood conditions have become increasingly prevalent on these streams, many former spawning grounds have been rendered useless, and the carrying capacity of the streams for young fish during their stream residence has been reduced. Utilization of streams for water power or for industrial purposes has had a similar effect. There is a further possibility of the withdrawal of spawning grounds due to the impounding of waters in the upper reaches of these rivers for the purpose of controlling floods and erosion. It is difficult to determine how far the level of abundance can decline before the populations of some areas pass the point at which they are able to rehabilitate them- selves, even under the most stringent protection that legislation might offer. In view of these conditions it appears highly probable that the decline in numbers of this species will be continued unless there is a drastic change in the factors influencing their abundance. KING SALMON By George B. Kelez INTRODUCTION Populations of king salmon are found in most of the important salmon streams in the region, the heaviest runs usually appearing in the larger rivers. Averaging more than 20 pounds in weight, the kings are the largest of the 5 species of Pacific salmon. Their large size and high quality have always commanded the highest indi- vidual price of any of the species, and the greater portion of the catch has been absorbed by the fresh-fish markets or used for mild curing. Kings from the troll fishery of Cape Flattery appear in the city markets in early spring and they are taken in gill nets throughout the fall months, but the bulk of the commercial landings are made in late spring and summer. Except in the gill-net catches in the rivers, both immature and mature fish appear together in the landings during the greater part of the fishing season. Sport fishing for kings, which has been popular with residents of the region for nearly 50 years, is carried on from April to September. LIFE HISTORY Possibly because of their greater size and strength, kings usually spawn in deeper, faster water than do the other species of salmon. Although the spring runs may ascend to small head-water streams, the later runs often spawn in the larger tributaries or even in the main channels of the rivers. There is a recognizable difference in the 796 BULLETIN OF THE BUREAU OF FISHERIES time at which these runs enter the rivers; the races which spawn far upstream usually appear during the spring months, whereas the lower-spawning races do not appear until later in the summer or in early fall. Gilbert (1913) and Fraser (1917 et seq.) both found that the greater part of the fry descend to salt water shortly after hatching, and a lesser proportion remain in the stream throughout the first winter and migrate seaward during the following spring. These findings were based on scale readings. Scales from fish which migrated to the sea as fry showed a typical rapid growth in the nucleus, those which migrated as yearlings showed a distinctly different nucleus, due to the less rapid growth in the stream. Fraser reported the proportions of these types in lower Gulf of Georgia fish to be 65.4 percent sea-type and 34.6 percent stream-type. His collections from the upper part of the Gulf of Georgia contained 78.2 percent of the former type and 21.8 percent of the latter. Rich (1925) stated that in the Columbia River rims the stream-type nuclei indi- cated spring-running fish which spawned in the headwater streams, whereas sea-type nuclei predominated later in the season when the lower-spawning races of fish were entering the river. After migrating to salt water, the young kings are frequently caught in the inner waters of the region before reaching the ocean. At this time they are called “black- mouth” by the fishermen. Tagging experiments reported by Canadian investigators, Williamson (1925, 1926), Mottley (1929), Williamson and Clemens (1932), Clemens (1932), and Pritchard (1934), indicate that a considerable proportion of the young kings migrate northward and return along the coast of Southeastern Alaska and British Columbia on their migration to the streams where they will spawn. These experiments have also indi- cated the presence of large numbers of kings from the populations of other coastal rivers, both north and south of the region, in the same localities along the British Columbia coast. It is evident that a considerable mixture of populations occurs in the waters of the Pacific, and that catches of gear operating in the offshore waters may well contain large numbers of fish from streams other than those of the region. Gilbert (1913) stated that kings taken in the commercial fishery of the region ranged in age from 3-7 years, and that the fish in their third year were grilse. Fraser (1921) reported that the commercial catch from the upper part of the Gulf of Georgia contained fish from 2-6 years of age, only part of which were mature. Of those indi- viduals which had entered the sea shortly after hatching, nearly 50 percent were in their third, and approximately 35 percent were in their fourth year. The remainder were 2 and 5 years of age. Of those which had entered the sea after a considerable time in fresh water, some 30 percent were in their third year, 44 percent in their fourth year, 23 percent in their fifth year, and the remainder in their sixth year. The bulk of the mature fish were in their fourth and fifth years. An important characteristic of the king salmon, unique to that species, is the considerable variation in the color of the flesh. Rathbun (1899) stated: While in some of the fish the flesh has its ordinary deep pink color, in others the flesh is white or only slightly tinged with pink. All intermediate gradations of colorations, as well as intermixtures of the two, occur, and no degree of this variation is distinguishable from the outside. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 797 Cobb (1911) stated: In most places the flesh is of a deep salmon red, but in certain places, notably Southeast Alaska, Bristol Bay, Puget Sound, and British Columbia, many of the fish, the proportion being sometimes as much as one-third of the catch, have white flesh. No reasonable explanation of this phenomenon has yet been given. Aside from color, the flesh of white and red kings taken at the same time in the fishery is of the same quality. This, together with the definite difference in proportion of white kings in various districts of the region throughout the season, which will be discussed later, indicates a strong possibility of a hereditary color-characteristic. The Fraser River king pack is canned as red, pink, and white kings. It is possible that a part of the late-season pack may consist of red kings whose color has faded with approaching sexual maturity. However, heavy catches of white kings are made by trailers off the west coast of Vancouver Island in late July and August, at which time the color cannot be ascribed to changes accompanying sexual development. Since these fish are not caught below Destruction Island, southwest of Cape Flattery, it is highly probable that they are part of the run which appears in the northern part of Puget Sound in September. LOCALITY OF CAPTURE BY DIFFERENT TYPES OF GEAR CATCHES IN VARIOUS DISTRICTS The demands of the fresh-fish markets, and methods of processing other than canning, have absorbed the greater part of the catches of king salmon, hence the canned packs are of little use in determining the annual catch of the species. The catch on Puget Sound alone has averaged 264,000 fish a year during the 20-year period from 1915-34. The catch by 5-year intervals during this time was 1,597,246 fish from 1915-19; 1,219,492 fish from 1920-24; 1,380,225 fish from 1925-29; and 1,087,693 fish from 1930-34. It is exceedingly difficult to obtain catch records for all districts of the region, but data are available for the period from 1927-34 for all districts except the Fraser River. For this district the canned pack has been converted to number of fish and it represents only a part of the early run on the river, but includes the greater part of the fall run. A small number of kings landed by trailers in the northern portion of the Gulf of Georgia have not been included. These data are presented in table 41. Table 41. — Catch of king salmon, 1927-34 Year Puget Sound traps Purse-seines Trailers Puget Sound gill nets Minor Puget Sound gear Fraser River catch i Total, all gear Puget Sound High seas Puget Sound High seas 1927 - 227, 909 18, 370 6, 818 1, 870 235, 866 37, 580 2, 033 53, 770 584, 216 1928 198, 443 11,025 4,067 1,651 213, 784 31, 195 900 11,629 472, 694 1929 - 249, 353 14, 181 13,817 1, 366 206, 073 44, 485 2, 257 23, 533 555, 065 1930 208, 872 17, 136 8, 791 2,645 235, 425 49, 934 1, 558 51, 084 675, 445 1931 - - - 156, 207 21, 497 13, 957 1,156 245, 611 28, 522 516 28,712 496, 178 1932 137, 770 20, 670 6, 897 192 169,530 20,910 24 84, 722 440, 716 1933 162, 991 23, 916 4, 596 68 113,512 22, 960 667 16, 483 345, 193 1934 165, 013 15, 606 10, 490 9, 337 125, 377 19, 250 276 46, 227 391, 676 Total 1, 506, 658 142, 401 69, 433 18, 285 1, 545, 178 254, 836 8,231 316, 160 3, 861, 082 Converted from cases packed from fish caught on the Fraser River; does not include kings used for purposes other than canning. 798 BULLETIN OF THE BUREAU OF FISHERIES The total catch by all gear in the region during these 8 years was 3,861,082 fish. Trollers on the high seas landed 40.02 percent of the total catch, and inside trollers 0.47 percent, a total of 40.40 percent. Puget Sound traps took 39.02 percent of the king catch during these years, Fraser River gill nets 8.19 percent, and Puget Sound gill nets 6.60 percent. Purse seiners on the high seas took 1.80 percent and those on Puget Sound 3.69 percent, a total of 5.49 percent. Landings from miscellaneous gear amounted to 0.21 percent of the total. The catch of trollers in the region of Swiftsure Bank differs from the “inside” gear in that it must, in view of the migrations indicated by tagging experiments, contain a considerable proportion of fish from populations other than those of the Puget Sound-Fraser River region. LOCALITY OF TRAP CATCHES A consideration of catch data from traps shows the general proportions of the king-salmon catches in the different parts of the Puget Sound district. These data, from 1895-1934, are presented in table 42. The districts used are similar to those discussed under trap catches of coho salmon in the preceding section. The total catch of king salmon includes 5,659,793 fish. Of this total, 2,644,524 were taken in traps north of Deception Pass, the greater part of these being from the populations of rivers in the northern part of the region. There were 1,741,479 fish from districts wherein a considerable mixture of populations migrating to both north- ern and southern streams must be present; 1,128,835 fish were taken in the southern portions of the region, and 144,955 fish were taken in miscellaneous and unidentified areas. These data indicate that the greater portion of the catch of king salmon on Puget Sound is supplied by the populations of the northern rivers, and the size of the catch in the northernmost districts would further indicate that a considerable portion of these populations are migrating to the Fraser River and to the smaller streams entering the Gulf of Georgia. Table 42. — Annual catch of king salmon in different areas, 1895-1934 1 Year Area Total Point Roberts and Bound- ary Bay Sandy Point to Bound- ary Bay Lummi Island and Ro- sario Strait Haro Strait and Wald- ron Island Salm- on Banks and South Lopez Hope Island and Skagit Bay West Beach West of Middle Point Admi- ralty Inlet Lower Sound Mis- cella- neous Un- identi- fied 1895 912 10, 192 1,449 30, 255 19, 980 30, 979 7,881 5,312 6, 005 15, 695 14, 105 8,731 14, 952 8, 843 7, 374 14, 542 28, 054 22, 442 912 11,077 8,788 33, 394 40, 019 98, 397 41, 103 53, 550 46, 965 61. 194 56, 545 72, 264 79, 955 98, 980 58, 675 107, 229 136, 362 109, 544 1896 __ 97 3, 164 41 12, 286 364 3, 568 5, 497 5, 563 5,807 4,883 8, 048 6, 543 8, 475 9, 877 18, 144 25, 996 22,785 788 1897 __ 720 3, 000 4,635 9, 215 4,442 7,681 15, 427 17, 478 5, 065 7, 9S1 8,829 8, 922 5,666 21, 478 26, 590 19, 461 71 3, 384 94 79 5, 758 5, 047 9, 077 1898 4 1899 96 2,412 2, 933 5, 378 2, 489 2, 343 2,080 9,081 2, 814 20, 365 5, 963 14, 891 17, 360 6, 960 7, 472 3,627 5,011 5, 892 6,911 8, 307 9, 647 4, 374 129 1900 14, 755 5, 180 4, 829 8, 922 461 885 850 218 4, 777 5,410 1901 1902 1903 121 1904 12,911 10, 469 19, 103 34, 977 47, 986 19, 648 24, 414 31,090 28,886 1905 _ 9, 787 5, 609 8,903 9,640 8,945 8, 583 4, 752 4, 374 2, 684 10, 084 740 3, 461 254 11,505 6,770 3, 343 1906__. _ 1907 190S 5, 761 1909 .. 1910 256 83 1911 380 629 1912 3, 250 1 Incomplete before 1915. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 799 Table 42. — Annual catch of king salmon in different areas, 1895-1934 — Continued Area Year 1913. 1914. 1915. 1916. 1917. 1918. 1919. 1920. 1921. 1922. 1923. 1924. 1925. 1926. 1927. 1928. 1929. 1930. 1931. 1932. 1933. 1934. Total Point Roberts and Bound- ary Bay Sandy Point to Bound- ary Bay Lumrai Island and Ro- sario Strait Haro Strait and Wald- ron Island Salm- on Banks and South Lopez Hope Island and Skagit 'Bay West Beach West of Middle Point Admi- ralty Inlet Lower Sound Mis- cella- neous Un- identi- fied Total 22, 349 17, 984 20, 967 2, 126 9, 114 4,038 18, 306 4, 55S 14 99, 456 30, 134 25, 585 25, 619 3, 323 11, 838 4, 040 15, 330 6, 479 449 128| 797 43, 359 15,867 31,878 15', 386 20, 632 13', 903 42, 502 13,310 19,712 3, 780 4, 302 12, 553 237, 184 40,819 26,219 29, 319 14, 567 21, 222 8, 607 46, 473 23,917 24, 901 1,332 6, 327 7, 930 251, 633 45, 172 21, 257 26, 209 24, 795 41, 582 16, 971 42, 975 17, 332 26, 559 3,321 9, 572 9, 486 285, 231 42, 160 34,916 35, 773 16, 772 35, 574 21, 057 59, 606 25, 921 40, 478 1,992 10, 013 724 324, 986 47, 869 17, 493 19, 937 12, 382 23, 707 22, 475 44, 627 14, 831 32, 913 2,639 10, 085 448 249, 406 58, 172 19, 053 33. 490 4, 577 19, 781 10, 583 35, 012 6, 104 22, 304 3, 548 6, 109 1, 069 218, 802 41, 573 19, 627 17, 335 16, 268 27, 994 13,313 31, 681 7, 140 23,812 1, 678 3, 049 1,832 205, 302 41,419 16, 430 21,333 4, 525 17,311 20, 234 29, 452 9, 749 20, 630 2, 121 3, 697 186, 901 37, 375 17, 406 16, 860 8, 119 22, 278 21,409 32, 739 6, 801 28, 246 1, 434 16, 157 208, 824 51,005 16, 511 21, 775 9, 861 16, 773 39, 123 12, 903 35, 127 8, 117 3, 827 215, 082 48, 491 18, 223 24,231 8, 746 25, 022 21, 745 33, 844 13; 087 33; 142 3, 526 4,310 234; 367 43, 639 15, 497 18, 066 10, 899 12, 757 17, 165 29, 641 13, 855 31, 689 5, 330 2, 560 367 201, 765 63, 690 26, 463 18, 568 8, 471 21, 230 21,429 38, 585 13, 897 26, 614 3,453 231, 400 39, 856 16, 638 14, 583 4, 925 16, 318 10, 779 27, 607 11,710 31. 689 6, 699 385 181, 189 48, 708 18, 557 30, 105 9,117 10; 389 24, 873 35,611 8,937 32, 298 10, 121 1,504 7, 506 243; 726 45, 431 13, 940 22, 370 7, 507 16, 853 25, 741 28, 201 9, 086 31,314 5, 153 2, 186 207, 782 36^ 912 8, 569 17, 586 5, 573 9, 550 20, 951 21, 630 5, 279 23, 939 2,015 ' 916 152; 920 34, 493 4, 292 20, 943 2, 007 13, 243 17, 626 13, 875 7, 832 18, 818 3, 524 136, 653 32, 407 16, 087 23, 855 5,965 17, 606 12, 673 15, 072 2, 300 40, 088 1,305 1,307 2, 476 171, 141 50, 187 10, 792 21, 143 7, 728 15, 990 12, 104 14, 764 12,351 22, 606 3,289 1,279 172, 293 1, 188, 983 562, 990 653, 083 239, 462 552, 429 427, 433 952, 579 236, 471 627, 025 74, 377 89, 111 55, 844 5, 659, 793 In only one tagging experiment of those reviewed by Pritchard (1934) were the recoveries in southern Puget Sound greater than those in the northern part of the region. The major part of the recoveries of fish tagged in this experiment, on the northeast coast of Vancouver Island, were taken in the vicinity of the mouth of the Skagit River or in the river itself. This stream, which has supplied the greater portion of the kings gill netted in the Puget Sound region, supports the largest run in the southern area. The other tagging experiments confirm the inferences drawn from the trap-catch data as to the importance of the northern streams, since the greater proportion of recoveries made in the inner waters of the region have been taken in the northern districts of Puget Sound or in the Fraser River itself. SEASONAL OCCURRENCE IN VARIOUS AREAS The general seasonal occurrence of king salmon in different types of gear has already been presented, but a further consideration of occurrence in traps indicates certain specific differences in the runs in various parts of the region. The occurrence of kings from traps in several restricted areas was calculated in a manner similar to that used for the entire region, see section on trap fishery. The average proportions of the annual catch taken in each week of the season in these different areas are pre- sented in table 43. For location of areas, see figures 2 and 3. 800 BULLETIN OP THE BUREAU OF FISHERIES Table 43. — Seasonal occurrence of king salmon from traps; proportion of total catch taken in each 7-day period Area Week ending— North of Sandy Point Rosario Strait West Beach Middle Point Admiralty Inlet Hope Island All dis- tricts Apr. 21 0. 556 0. 425 Apr. 28 2. 724 1.365 1.353 May 5 _ 3. 298 2. 334 3.595 2. 392 3.513 1. 587 2. 259 May 12 3.089 4. 232 3. 779 3. 799 4. 742 1.785 3.212 May 19 3. 195 3.808 3. 735 4. 384 5.875 2. 551 3. 649 May 26 3. 137 3. 539 4. 469 5. 142 4. 859 3. 265 3. 780 June 2 4. 056 3. 539 4. 573 4.713 5. 782 3. 578 4. 166 June 9 4. 774 4.354 4. 994 4. 943 5. 481 4. 625 4.770 June 16 - 4.910 4. 845 5. 450 5. 865 5. 349 5. 151 5. 145 June 23 5. 421 5.582 7.049 5. 835 5.397 7. 400 5. 921 June 30 __ 5. 547 6. 140 7. 398 5. 507 5. 742 8. 829 6. 330 July 7 6.317 7. 433 7. 736 6. 571 5. 854 10. 796 7. 292 5. 679 6. 925 6. 950 7.028 5.867 8. 701 6. 696 July 21 5.516 6.267 7. 184 6. 705 5.618 7.258 6. 252 July 28 5. 296 6.666 6. 552 7.179 6.013 5. 925 6. 188 Aug. 4 __ 5. 069 6, 153 6. 224 7. 191 6. 590 6. 385 6. 072 Aug. 11 5. 379 5. 981 4. 897 8. 040 7.418 6. 407 6. 149 Aug. 18. 5.796 5. 654 4.430 6.151 5. 132 5. 776 5. 565 Aug. 25 4. 768 4. 510 3. 621 4. 272 4.742 4. 234 4.456 Sept. 1 5.037 3. 693 2.022 2. 050 2. 554 2. 133 3. 406 Sept. 8 4. 912 3.703 1. 277 1.276 1.602 1.042 2. 875 Sept. 15 5.011 2. 544 .611 .584 .900 .371 2. 074 Sept. 22 2. 452 1.248 .300 .197 .385 .229 1. 105 Sept. 29 . .956 .554 .303 .076 .168 .024 .451 Oct. 6 .222 .198 . 106 .044 .119 .020 .167 Oet. 13 . 145 .063 .014 .039 .074 .007 .069 Oct. 20.. .015 .029 .019 .062 .041 Oct. 27. .004 .015 .059 .038 Nov. 3. .011 .073 .064 .030 .030 Total 100. 001 100. 000 99. 999 100. 002 100. 000 100. 000 100. 000 In all areas the run is much more prolonged than that of the other species, and there are no extreme peaks of occurrence. The highest percentages for any single week in the district north of Sandy Point or in Rosario Strait occur in the first week of July. There is an additional run in these areas in late August and September, especially in the more northern one. West Beach, where the catches probably con- tain a considerable mixture of populations, shows a similar peak in that week, but there is no indication of the late-season run. The southern areas show proportion- ately higher percentages in the early part of the season, a peak early in August, and an abrupt decrease thereafter. There is also no indication of a late run in these areas. SEASONAL OCCURRENCE OF RED AND WHITE KING SALMON Thus far the runs of king salmon have been treated as entities, but some of the distinct differences between their occurrence in northern and southern areas may be explained by a consideration of the proportionate runs of red and white kings in these districts. The catches of king salmon from certain traps in the region have been segregated as to red and white kings by the operators, especially where the fish were sold for mar- ket purposes. Such a segregation into only two classes undoubtedly introduces some errors in the determination of the proper classification of the individuals which inter- grade between the color extremes of red and white. Grading has been purely on the SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 801 basis of market demand, and the general practice has been to classify the vari-colored fish with the whites, since the reds bring a higher price. The following determina- tions are necessarily confined to the two main classes, but the presence of intergrading colors must not be overlooked. Data were available for some early years, and for most of the years between 1923 and 1934, for 3 traps in Haro Strait, 3 in Birch Bay, 1 on Lopez Island, and 2 in Admiralty Inlet. The average proportionate occurrence of red and white king salmon throughout the season was calculated for these four areas. These data are presented in table 44. Table 44. — Seasonal occurrence of red and white king salmon in different areas; proportion of total catch taken in each 7-day period Week ending — Red king White king Haro Strait Birch Bay Haro Strait and Birch Bay South Lopez Admir- alty Inlet Haro Strait Birch Bay Haro Strait and Birch Bay South Lopez Admi- ralty Inlet 3. 696 3. 569 0. 380 3.054 3. 400 4. 658 6. 181 4.986 0. 163 0. 266 2. 296 2. 883 May 19 3. 758 9. 053 5.199 7. 398 6. 001 .713 3. 209 1.391 8. 406 4. 760 May 26 5. 645 4. 799 5. 533 6.887 4.947 .909 1.399 1. 140 2. 546 3.625 June 2 4. 101 3.410 4.000 6. 193 5. 442 .769 1.567 1. 051 2. 133 5. 398 June 9 5.614 8. 932 6.437 5. 555 5. 347 .979 2. 775 1. 505 2.438 6. 247 June 16 5. 297 6. 881 5. 624 6. 118 5. 209 .952 1.791 1.231 3. 003 5.415 June 23 6. 709 6. 543 6. 493 6. 267 5. 209 1.343 2. 059 1.619 3.613 5. 637 June 30 5. 052 6. 422 5. 323 6. 095 5. 665 1. 161 1.522 1. 333 3. 352 6. 034 July 7 4. 856 5.718 4. 985 7. 325 6. 023 .865 1. 985 1. 210 5. 663 6.496 July 14 5. 727 5. 689 5. 573 6. 374 5. 746 1.094 3. 448 1.769 6. 323 6. 350 July 21 5. 953 3. 352 5. 050 6. 180 5. 483 . 849 2. 089 1.223 7. 643 5. 604 July 28_ 5. 989 5.718 5. 761 6. 374 5. 981 1.583 3. 396 2. 154 5.713 5. 993 Aug. 4 5. 357 6.281 5. 491 5.812 6. 681 1.926 5. 120 2. 872 9. 090 6. 560 Aug. 11 10. 988 6. 957 9. 544 4. 774 7. 514 3. 608 4. 179 4. 003 9. 953 7. 565 Aug. 18 4. 904 6. 534 5. 254 4.018 5. 361 3. 232 4. 806 3. 856 7. 669 4.714 Aug. 25 5.787 3. 746 5.050 3. 060 4. 740 3. 526 3. 134 3. 670 6. 475 6. 170 Sept. 1 4.183 4. 021 4. 115 1. 066 2.587 6. 398 13. 671 8. 777 5. 472 2. 457 Sept. 8. . . _ 3. 365 2. 298 2. 562 .360 1. 61 1 35. 849 12. 662 25. 931 3. 383 2. 403 Sept. 15 2. 267 .739 1. 724 . 158 .926 27. 491 10. 186 24. 348 1. 676 1.203 Sept. 22 .603 .845 .701 .041 .360 5. 443 10. 521 7. 551 1. 490 .834 Sept. 29 . 150 2. 060 .542 .051 .161 .948 10. 480 2.835 .702 .332 Oct. 6 .057 .058 .015 . 125 . 137 . 164 .057 .084 Oct. 13 . . 134 .184 .078 .061 . 100 . 129 .043 Oct. 20__ . 102 . 140 .070 .335 .016 Oct. 27.. .067 .016 Nov. 3 .075 .109 Nov. 10. .035 Total 99. 998 99. 998 100. 001 99. 998 99. 999 99. 999 99. 999 99. 999 100. 000 100. 002 Occurrence of red kings does not differ materially in the various areas, although there is a greater early run in the northern districts and a heavier run in the Admiralty Inlet aiea. White kings differ considerably, however, with heavy fall concentrations in the northern areas. More than 75 percent of the season’s catch in Haro Strait is made during the month of September, as is approximately 60 percent of the catch in Birch Bay. Occurrence of white kings in the southern portion of Rosario Strait (South Lopez) is more even throughout the season, the peak of the run appearing dur- ing the month of August, while in Admiralty Inlet no definite peak of occurrence is shown. The average proportion of white kings in the total catch of red and white kings combined was then calculated for each week in the season and for the total season from the trap catches of the various areas. These data are presented in table 45. 802 BULLETIN OF THE BUREAU OF FISHERIES In order to determine the proportionate occurrence of red and white kings in both northern and southern runs, the weekly percentages of kings from table 43 were divided as to proportion of red and white kings on the basis of the data presented in table 45. Since percentages were not available for the entire area north of Sandy Point, a combination of the Haro Strait and Birch Bay proportions were used for this northern district. Proportionate seasonal occurrence in the area north of Sandy Point is shown graphically in the upper section of figure 28, that of Admiralty Inlet in the lower section of the same figure. Table 45. — Proportion in each 7 -day period of white king salmon in total king salmon catches in different areas Area Week ending — Haro Strait Birch Bay Haro Strait and Birch Bay South Lopez Admiralty Inlet May 5 2. 646 7. 358 May 12 3. 390 3. 390 8. 921 5. 101 May 19__ . . .. 12. 167 16. 045 14. 124 23. 182 6. 857 May 26 10. 535 13. 587 11. 236 8. 884 6. 367 June 2 12. 061 19. 858 13. 903 8. 326 8. 431 June 9 11. 309 14. 352 12. 559 10. 370 9. 785 June 16 11. 604 12. 308 11.855 11. 462 8. 800 June 23 12. 766 14. 511 13. 283 13. 196 9. 128 June 30_ 14. 385 11.333 13. 341 12. 664 8. 995 July 7 11.522 15. 768 12. 981 16. 932 9. 099 July 14 . 12. 249 24. 627 16. 319 20. 733 9. 302 July 21 9. 437 25. 157 12. 958 24. 592 8. 665 July 28 16. 194 24. 254 18. 685 19. 116 8. 510 Aug. 4___ 20.811 30. 530 24. 324 29. 201 8. 352 Aug. 11 19. 352 24. 465 20. 490 35. 475 8. 545 Aug. 18 32. 512 28. 395 31. 083 33. 481 7. 545 Aug. 25 30. 813 31.088 30. 872 35.815 10. 779 Sept. 1 . _ 52. 778 64. 706 56. 720 57. 513 8. 102 Sept. 8___ _ 88. 620 74. 815 86. 150 71. 242 12. 165 Sept. 15 89. 861 88. 136 89. 668 73. 684 10. 759 Sept. 22 86. 830 87. 037 86. 885 90. 566 17. 699 Sept. 29 82. 193 73. 288 76. 256 78. 261 16. 107 Oct. 6 63. 636 63. 636 50. 000 5. 825 Oct. 13.. 25. 000 25. 000 4. 839 Oct. 20. 1.961 Oct. 27_ 2. 128 Nov. 3 11. 765 34. 161 31. 503 33. 390 20. 295 8. 665 The two peaks of occurrence of red kings, in early July and in late August, in the northern area may be compared to the sustained run in the southern area, where the highest percentages occur in early August. The run in both areas diminishes uniformly during late August and early September. The run of white kings in the northern area is in striking contrast to that of Admiralty Inlet. In the latter area the white kings form a very small proportion of the run and are distributed throughout the season. In the northern area, however, they form a considerable portion of the run in the early part of the season, become increasingly important in midsummer, and form the major part of the run from early September to the end of the season. It thus appears that the more prolonged occurrence of king salmon in the northern areas is due to the presence of a considerable run of white kings, and that the major portion of white kings caught in the region must have been contributed by the populations of the Gulf of Georgia streams. SALMON AND SALMON FISHERIES OF SWIFTSUEE BANK 803 RED AND WHITE KING SALMON SEASONAL OCCURRENCE “i r NORTH OF SANDY POINT AJ. CHANGES IN ABUNDANCE Traps are the only major gear taking king salmon for which sufficient data are available for any determination of changes in abundance of this species. The records of this gear are inadequate prior to 1910, but the fol- lowing calculations are pre- sented as the best measure which can be determined from present data. The data were necessa- rily restricted, because of fishing seasons of varying lengths, to include only those traps for which the opening fishing dates for each season were known . The catches of all traps were then weighted according to the length of the season fished in a man- ner similar to that discussed under the trap index for coho salmon. Suitable data from the area north of Sandy Point were available for the period from 1910-34. During these years, from 6 to 11 traps fished in each year except 1932. Four of these traps were in Birch Bay, 4 in Bounday Bay, and 3 at Point Roberts. During the same period of time, data were available from 2 traps in Rosario Strait for every year except 1910. In Admiralty Inlet, data were available between 1916 and 1934 from 4 traps, at least 2 of which fished in every year except 1916 and 1932. The number of traps fishing in each area, and their combined catches, are presented in table 46. Figure 28.— Seasonal occurrence of red and white king salmon in trap catches of the northern and southern districts of Puget Sound. The greater abundance of white kings and the heavy late-season run in the northern district are apparent. 71941—31 -8 S04 BULLETIN OF THE BUREAU OF FISHERIES Table 46. — Indices of abundance of king salmon from traps Year Data by areas Index figures North of Sandy Point Rosario Strait Admiralty Inlet North of Sandy Point Rosario Strait Admir- alty Inlet Number of fish Number of traps Number of fish Number of traps Number of fish Number of traps 1910. 20, 821 6 12, 053 1 1.611 1. 741 1911 28, 475 6 19, 265 2 2. no 1. 780 1912 31, 841 9 13, 904 2 1. 631 1. 382 1913 . 22, 806 6 9, 844 2 1. 747 1. 007 1914 31,317 7 13, 385 2 1. 956 1.344 1915 _ 23, no 8 11, 106 2 1. 145 1. 168 1916. 29, 577 8 7,808 2 4, 869 1 1. 661 .920 2. 394 1917 27, 640 8 9, 367 2 11,420 2 1. 627 .974 3. 362 1918 35, 562 8 13, 432 2 17, 244 2 2. 290 1.416 5. 105 1919 32, 308 9 10, 166 2 9, 728 2 1. 744 1. 017 2. 975 1920 44, 266 7 19, 595 2 16, 279 4 2. 859 1.848 1.548 1921 38, 922 9 10, 227 2 14, 359 2 2. 088 1. 053 2. 216 1922 39, 717 7 IS, 333 2 7, 217 2 2. 676 1. 675 1. 937 1923 28, 078 7 11, 464 2 21, 996 4 1. 962 1. 096 2. 160 1924 46, 926 8 11,650 2 28, 693 4 2.887 1. 061 2.645 1925... 48, 234 10 10, 500 2 18. 407 4 1.915 1.020 1.974 1926 36, 890 10 10, 749 2 16, 220 4 1. 534 1. 071 1. 640 1927 51, 161 11 10, 060 2 11,638 3 2. 057 1. 056 1.494 1928 26, 972 9 7, 129 2 7,915 3 1.590 .775 1.034 1929 44, 423 10 11,287 2 12, 862 4 1.868 1. 113 1. 368 1930. 27, 112 11 12. 656 2 9. 428 4 1. 274 1. 292 1.054 1931... 24,914 10 7, 060 2 3, 927 2 1.204 .940 .697 1932 2, 764 2 19, 703 2 5, 223 1 1. 428 1. 964 1.927 1933 16, 559 7 13. 431 2 11, 046 3 1.529 1. 325 1. 426 1934. 38, 405 9 11, 952 2 4,407 2 2. 094 1.116 .947 Total 798, S03 306, 126 232, 878 Indices of abundance for these three areas, calculated in the same manner as were those from traps for coho salmon, are presented in the last three columns of table 46. The indices are high for the northern areas during the post-war period prior to 1925. Increased catches during this period may be due in part to the lesser competition of trolling gear, which fishes the rims before they reach the traps. The number of trolling licenses issued by the State of Washington for the Puget Sound district decreased from 1,032 in 1919 to 165 in 1922, and then increased considerably in number to 820 in 1927 (see table 23). There is little difference in levels of abundance in early and late years in the two northern areas. In Admiralty Inlet, however, abundance is highest before 1920 and reaches a lower level by 1924; a decrease in the size of the runs in recent years is strongly indicated. PINK SALMON By George A. Rounsefell GENERAL LIFE HISTORY Because pink salmon invariably mature in their second year, as has been well established, there is no overlapping of generations as in the sockeye, king, and chum salmon, and, in some regions, in the coho. In this region there is an abundant run of pink salmon every second year, in the odd-numbered years. They spawn in scores of small streams, as well as the lower tributaries of the main rivers. In the Fraser River they even spawned above Hell’s Gate in Seton and Anderson lakes and the Nicola and Thompson rivers until the blockade at Hell’s Gate in 1913, which, coming in an odd- SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 805 numbered year, destroyed this up-river run. In the even-numbered years no pink salmon spawn in Puget Sound streams or in the Fraser River, although a few thousand are usually caught north of Deception Pass. Most of these pinks are probably bound to the streams in the northern end of the Gulf of Georgia, which have pink runs in both odd- and even-numbered years. The pink-salmon fry, upon emerging from the gravel, migrate at once to the sea, which permits great numbers to propagate in streams that might be unsuitable for the support of large numbers of young fish. Recently evidence has been gathered on the homing instinct in pink sahnon. Pritchard (1934) in an experiment at McClinton Creek, Masset Inlet, in which 108,000 fry were marked by clipping of fins before being liberated, recovered 3,285 when they returned from the sea as adults. Of this total, over 3 percent of the number marked, only 7 fish were taken outside of the Queen Charlotte Islands, and 2,950 were recap- tured in the same creek. Davidson (1934) in an earlier experiment marked 50,000 pink fry at Olive Cove, Alaska. Twenty-three marks were recovered there from 7,944 adult salmon dipped over the counting weir. Since 10,640 of the run were not examined for scars the total number of marked fish in the run was calculated as 54. MIGRATION Information is scarce on the migrations of pink salmon in the region. Pritchard (1930) tagged 205 pinks in Johnstone Strait in 1928. All of the recoveries were made in local streams. In 1929 the experiment was repeated (Pritchard, 1932) and out of 468 tagged in the same area 37 were recovered, 20 in the Fraser River, and 1 at West Beach, WThidbey Island. None were recaptured farther to the north than the point of tagging. The difference between the 1928 and 1929 results was quite as expected, since Puget Sound and the Fraser River support a tremendous run of pinks in the odd-numbered years, but almost none in the even-numbered years. The recoveries show that a fair share of the run to this region may ordinarily come around the north end of Vancouver Island. Pink salmon were also tagged in 1929 from the traps at Sooke. Out of 185 released there were 14 recoveries, 1 at the point of tagging, 6 in Puget Sound waters (3 from north of Deception Pass), and 7 in the Fraser River. METHOD AND LOCALITY OF CAPTURE The Swiftsure Bank-Puget Sound-Fraser River pink salmon catch from 1925-34 amounts to 52,240,000 fish, excluding Vancouver Island and the Gulf of Georgia for which sufficient data are not at hand (see table 47). Previous to 1925 data are lacking on the Swiftsure Bank catch or of the amounts canned on the Fraser River that were not shipped in from other districts. 806 BULLETIN OF THE BUREAU OF FISHERIES Table 47. — Catch of pink salmon, 1925-34 Year Fraser River catch 1 Puget Sound traps Purse seines Miscella- neous Puget Sound gear Total Puget Sound High seas 1925 1, 355, 592 19, 236 1, 378, 762 938 1, 957, 760 13, 118 186, 298 1, 950, 468 21,669 3, 062, 604 5, 882 2, 945, 720 7, 057 3, 688, 006 3, 678 1, 729, 775 2. 964 4, 602, 188 1, 764 3, 341,419 3,445 4, 365, 513 9, 520 4, 346, 600 5,130 4, 298, 591 10, 044 729, 702 1,529 2, 136, 570 68, 877 3, 373, 529 42, 058 3, 903, 188 5, 981 844, 895 20, 096 108, 386 1,052 125, 142 114 152, 962 738 52, no 21 58, 384 117 8, 746, 336 45, 250 10, 044, 497 79, 256 12, 795, 484 72, 491 12, 176, 202 14, 810 8,230, 411 38, 009 1926 1927 1928 1929__ 1930 1931 1932 1933. 1, 298, 766 4, 788 1934 . 6, 215, 258 13, 417, 823 20, 984, 214 11, 126, 625 499, 026 52, 242, 746 1 Converted from cases at 14 per case, does not include pinks caught elsewhere in the Gulf of Georgia and canned on the Fraser River. The purse seines are the most important factor, accounting for 32 million fish, or about 60 percent of the total catch, during the past 10 years. Purse seines do better, compared to the traps, in taking pinks than they do in the capture of sockeyes. The pink salmon swim in dense schools, frequently jumping or “finning,” so that the schools are much easier to locate. Also, a much larger proportion of the pinks may use Haro Strait than is the case with the sockeyes, as the pinks that are bound northward spawn not only in the Fraser River, but in a number of smaller rivers and streams entering the Gulf of Georgia from both the mainland and Vancouver Island shores, and, since only a few traps are favorably located to capture fish using Haro Strait they would catch relatively less. Accurate data on the locality of capture is available for the trap-caught pinks. Traps north of Deception Pass have taken over 45 million, whereas the southern traps have caught but 9 million, or a proportion of 5 to 1. During the past 10 years the proportion has been 2 to 1 ; 9 million northward and 4% million to the south. The records of one large company over a 7-year period show that the bulk of the seine-caught pinks are from the Salmon Bank area, with large numbers from around Stuart Island and Mitchell Bay in Haro Strait, and also from Lummi Island, Birch Bay, Boundary Bay and Point Roberts areas, only minor quantities being captured south of Deception Pass. It would thus appear that a large proportion of the pink salmon captured in Puget Sound waters, probably well over half, are bound toward Canadian spawning grounds. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 807 Table 48. — Pink salmon caught by traps north of Deception Pass 1 Year 1895. 1896. 1897. 1898. 1899. 1900. 1901. 1902. 1903. 1904. 1905. 1900. 1907. 1908. 1909. 1910. 1911. 1912. 1913. 1914. 1915. 1916. 1917. 1918. 1919. 1920. 1921. 1922. 1923. 1924. 1925. 1926. 1927. 1928. 1929. 1930. 1931. 1932. 1933. 1934. Point Roberts and Boundary Bay 28, 660 Birch Bay 1 Rosario Strait and Lummi Island Haro Strait and Waldron Island Salmon Bank South Lopez Undeter- mined Total 28,660 "47, "663 38,637 "1, 198, 461 9,026 56^861 4, 655 24, 493 28, 139 353, 640 6, 634 1, 644, 644 ’"'239, "973 ~L 494,707 69, 564 94, 246 19, 068 38, 956 ’ 198," 195" "~8L 522* 959, 905 6, 062 "79^520" 66, 988 175, 270 ’"75," 138" ,287 238, 504 55, 825 528, 516 1, 885, 463 97ft 1, 992, 165 457 3, 049, 686 3, 309 2,211,470 1,308 744, 701 129 1, 777, 330 8, 822 932,419 3,753 723, 232 7, 706 974, 883 25, 714 834, 226 7, 515 1, 124., 516 1,692 805, 697 1,485 648, 250 355 490, 513 1, 264 708, 077 449 809, 846 22 1, 227, 578 2,076 1,471,812 3,279 166, 854 161 584, 472 12,838 272, 762 2,470 200.577 3, 547 173, 289 8, 096 66, 300 4,427 223, 648 1,011 113,454 1,438 183, 059 558 132, 857 566 472,717 280 1, 472, 042 1, 180, 549 381 1, 443, 287 2, 448 519,415 96 536, 769 13, 143 248, 654 3, 180 167, 678 4,440 252, 037 7,764 158, 589 2, 626 333, 276 867 409, 851 1,040 598, 836 148 134, 507 380 761 50,260 83, 396 371 715, 922 1,200 160, 924 206 386, 945 3,203 201,849 434 170, 399 2,108 235, 262 17,054 276, 117 2, 674 248, 316 549 247,587 477 267, 078 77 131,061 336 397, 916 838 387, 316 335 650, 980 774 685, 701 5, 654 117, 114 232 306, 233 6, 905 203, 004 832 157, 478 3,234 153, 731 12,683 201, 707 1,011 162, 896 615 169, 910 243 132, 664 251 161, 709 126 168, 520 19 83, 287 36 161, 085 3,084 79, 801 538 87, 867 2,888 67,812 8, 831 93. 034 474 221, 063 329 89, 695 309 95. 035 730 93, 877 98 35 3, 464, 173 2,606 4, 711, 629 953 6, 360, 744 6, 911 6, 528, 192 13. 908 1, 792, 295 860 3, 762, 834 47, 998 1, 938, 489 11,207 1, 507, 231 23, 923 1, 857, 014 80, 142 1, 628, 973 18, 727 2, 313, 615 5,063 1, 836, 094 4,990 1, 924, 922 2,119 1, 144, 524 2,770 Total. 20, 780, 069 6, 620, 142 8, 111, 678 3, 507, 606 4, 594, 405 1, 258, 312 94, 956 44, 967, 068 1 Incomplete before 1915. 1 Including Alden Bank. Table 49. — Pink salmon caught by traps south of Deception Pass 1 Year West Beach Hope Island Middle Point and Ebeys Landing Admi- ralty Bay and Bush Point Oak Bay and Hood Canal Useless Bay and Point No Point Meadow Point and south South side Strait of Juan de Puea Total » 1897 125 125 1898 1899 5,050 5, 050 1900 1901_ 400 14,383 429 9, 455 24,238 429 15,816 1902 1903 15, 816 1904 1905 18, 498 19, 613 10, 859 48, 970 1906 1907...; 123, 140 333 90, 506 59 64,288 302 285, 221 782 145, 771 522 61, 044 2 553 417, 812 63, 356 655, 687 335 91, 059 69 381, 398 311 570, 195 905 1, 079, 382 741 1908. 1909.. 1910.. 1911 10, 745 9 44, 721 160, 303 146, 062 1912 1913 128, 757 113 160, 507 92 110,772 10 558, 897 22 1914... 1915... 109,971 6 15, 052 32, 470 7 32,000 24, 714 92 1916.... > Incomplete before 1915. 1 Total for 1913 includes 724 with locality undetermined. 808 BULLETIN OF THE BUREAU OF FISHERIES Table 49. — Pink salmon caught by traps south of Deception Pass — Continued Year West Beach Hope Island Middle Point and Ebeys Landing Admi- ralty Bay and Bush Point Oak Bay and Hood Canal Useless Bay and Point No Point Meadow Point and south South side Strait of Juan de Fuca Total 1917 1918. 1919.. . 1920.. . 1921 1922 1923 1924 1925 1926 1927.. . 1928 1929 1930.. 1931.. 1932 1933 1934 Total 107, 709 5, 989 12, 791 1,412 76, 416 1,158 62, 562 3, 188 45, 710 1, 181 130, 280 228 224, 968 590 249. 637 59 90, 892 36 1, 765, 619 56, 783 751 11,601 10 72, 308 147 50, 790 33, 497 80 74, 019 150, 968 105 166, 426 135 42, 847 5 121, 128 1,886 11, 990 197 36, 034 313 32, 545 724 33, 068 486 56, 265 109 118, 409 649 210, 816 164 100, 532 25 273, 722 898 56, 553 1, 164 201, 839 434 422, 933 1,062 161, 496 304 414, 171 323 461, 574 326 987, 625 312, 500 56 10, 815 55 6,869 1 75 29, 625 137 2, 552 21 29, 909 106 6, 033 13 32, 274 14, 936 "lsoo' 10, 357 10 13, 859 1 6, 311 46 6, 180 37 4,509 73 18, 669 9, 351 16 15, 409 15, 420 8 12, 999 "32," 578' 18, 979 20, 673 14 3,955 37, 998 838 1,736 170 33, 176 258 10, 453 921 26, 804 796 49, 274 159 109, 957 473 99, 499 596 1,947 72 653, 000 10, 523 109, 373 2, 967 452, 047 2,431 615, 445 6, 079 321, 495 2,942 748, 989 819 1, 109, 626 2,067 1, 763, 084 1, 559 585, 251 194 921, 948 1,175, 012 4, 551, 460 89, 978 245, 711 121, 871 399, 933 9, 262, 256 SEASONAL OCCURRENCE IN NORTHERN AND SOUTHERN DISTRICTS The southern pink salmon runs are earlier than the northern. The southern run, south of Admiralty Head, Ebeys Landing, Admiralty Bay, and Bush Point areas, reaches its peak about August 22, the northern run, areas north of Deception Pass, about September 1, making a difference of about 10 days in the modes. By August 11 about 22 percent of the southern run has passed, but only about 2% percent of the northern. By September 8, over 95 percent of the southern run has appeared, as against 78 percent of the northern. This difference in time of rim of trap-caught pinks in the two districts is good evidence of the existence of different populations or groups of populations. It is therefore necessary to allow a sufficient number of spawners in each district, as either one can doubtless be depleted regardless of the size of the escapement to the other. CHANGE IN ABUNDANCE BETWEEN EARLY AND LATE YEARS In the earlier years pink salmon were evidently tremendously abundant. Rath- bun (1899) says that in 1891 four drag seines operating for the Seattle cannery caught 275,000 pinks, but this number represented only a small part of the fishery in progress that year. At that time, and for a few years thereafter, pinks were canned only in Seattle, the output finding a ready sale at a low price in the southern part of the United States. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 809 Table 50. — Seasonal occurrence in traps of odd-year pink salmon in northern and southern districts , 1919-38 Week ending North of Deception Pass 1 South of Admiralty Head * Percentage Cumulative percentage Percentage Cumulative percentage 0. 003 .004 .010 .011 .014 .019 .034 .051 .302 2. 759 6.704 11.994 15. 579 27. 969 17. 887 12. 136 3.464 .395 .659 .006 .001 0. 003 .007 .017 .028 .042 .061 .095 . 146 .448 3.207 9.911 21. 905 37. 484 65. 453 83. 340 95. 476 98. 940 99. 335 99. 994 100. 000 100. 001 June 2 June 16. June 23 J u ne 30 July 7_ 0.002 0.002 .039 .206 .822 2. 578 7. 232 21. 502 51.289 78. 334 92. 189 98. 697 99. 698 99. 906 99. 986 99. 993 99. 998 Julv 21 . .037 . 167 .616 1.756 4. 654 14. 270 29. 7S7 27. 045 13. 855 6. 508 1.001 .208 .080 .007 .005 Julv 28 Aug. 18 Sept. 8 Sept. 15 Sept. 22 Sept. 29 Oct. 6 Oct. 13 Oct. 20 Oct. 27 Number of fish 2, 537, 611 1, 929, 504 Number of traps 9 7 i Week ending Sept. 15, empirically determined. } Week ending Sept. 1, empirically determined. Speaking of the trap fishery Kathbun says: The trap nets would appear, however, to afford the best means for the capture of the humpback in the salt water, and they are sometimes so taken in immense quantities during the sockeye run. In fact, they often compose by far the larger part of the catch, and as it is generally impracticable to do the sorting in the water at the net, the entire catch may be emptied into scows and the over- hauling take place at the wharves. Here the humpbacks are culled out and discarded, causing a wholesale destruction of the species. In addition to discarding pink salmon, the traps were often closed in odd-numbered years while some sockeyes were still available, in order to avoid capturing the later- running pink salmon for which they had no use. Owing to these factors during the early years of the fishery, the total catch figures are entirely unreliable for measur- ing abundance. Since the total catches of the individual traps do not give us an adequate measure of abundance in these years the problem has first been attacked by plotting the frequency distributions of the pink-salmon catches of all regularly operated traps north of Deception Pass in the odd-numbered years from 1899-1933 (see table 51). From 1899-1905 there was practically no demand for pink salmon, and only small quantities were used; the remainder was discarded. This is especially obvious in 1901 and 1905, both of which were big years for sockeye. 810 BULLETIN OF THE BUREAU OF FISHERIES Table 51. — Pink salmon catch 'per trap north of Deception Pass Catch in thousands 1899 1901 1903 1905 1907 1909 1911 1913 1915 1917 1919 1921 1923 1925 1927 1929 1931 1933 0 0-10 10-20 20-30 30-40 40-50.-. 50-60.... 60-70.... 70-80.... 80-90... 90-100— 100-110. 110-120. 120-130. 130-140. 140-150. 150-160. 160-170. 170-180. 180-190. 190-200. 200-210. 210-220. 220-230. 230-240. 240-250. 250-200. 260-270. 270-280.. 280-290. 290-300. 300-310. 310-320. 320-330. 330-340. 340-350. 350-360. 360-370. 370-380. 380-390. 390-400. 400-410. 410-420. 420-430. 430-440. 440-450. 450-160. In 1907 there was some demand for pinks and the medium take per trap was over 60,000. In 1909, a big sockeye year, only 50,000 per trap were utilized. In 1911, with a small sockeye run and an increasing demand for pinks, the median catch per trap was over 100,000. The median catch per trap was only 60,000 in 1913, again a big sockeye year, but on comparing it with 1911 and 1909 it is obvious that in the big years, either no pinks, or very few, were used from many of tlie traps. Eliminating those traps taking less than 20,000 pinks from the 1913 distribution, and they are not part of the distribution, as shown by 1911, the median catch is over 110,000. Since 1913 the demand for pink salmon has been good, and yet the highest median catch, in 1917, has only been over 30,000 per trap. If this evidence of a tremendous decline in abundance is not sufficiently convincing, one needs but note the size of the maximum trap catches. In the past 10 cycles, 1915-33, only 8 trap catches have exceeded 120,000 pink salmon, yet in the 8 earlier years this was exceeded 64 times. Considering only the earlier years when there was some demand, 1907-13, it was exceeded 62 times. In the same 4 cycles 29 catches were made of over 190,000 — larger than any single SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 811 catch in the past 10 cycles. Therefore, we must conclude that a tremendous decline in the abundance of pink salmon took place between 1913 and 1915. INDICES OF ABUNDANCE FROM TRAPS Because of the great dilforence in the time of the run between the northern and southern pinks, separate indices were made for the two districts. For the district north of Deception Pass 31 traps were selected fishing in the 14 odd years between 1907 and 1933, and taking 21,051,873 pinks, up to and including September 8 of each year. To use a longer season was impractical as the traps did not fish late during the early years and were subjected to a 10-day closed period from September 6-15 in the later years. The 31 traps selected were distributed as follows: Point Roberts 3, Boundary Bay 9, Birch Bay 6, Lummi Island 4, Salmon Bank 4, South Lopez 2, Rosario, Waldron Island, and Haro Strait areas 1 each. The index was calculated in the same manner as described for sockeye. For a standard curve 12 traps were used, 3 each from Boundary Bay and Birch Bay areas, 2 each from Lummi Island and Salmon Bank areas, and 1 each from Point Roberts and Rosario Strait areas. The standard covered the years from 1911-31. For the southern district only 7 traps were available, 2 from Middle Point area, 2 from Admiralty Bay, and 3 from Bush Point. For a standard curve all 7 traps were used for the 4 odd years from 1923-29. The northern index (table 52) shows a tremendous fall in abundance after 1913. In 1911 and 1913 the index was 284, in the following 20 years, 10 odd years, it has averaged 67.7 or about 24 percent of the former level. The reason for this sudden drop in abundance can best be explained by the following quotation from the Report of the British Columbia Commissioner of Fisheries for 1915: . . . That there would be a great decrease in the run of pink salmon to the Fraser River District this year was clearly indicated in the Department’s report from the spawning grounds in 1913. Owing to the blockade in the canyon of the Fraser at Hell’s Gate in 1913, no pink salmon were able to reach the spawning-beds in the waters above that point. Up to that year countless millions spawned in the Thompson and Nicola Rivers and in the vicinity of Seton Lake. As is shown in our report for the spawning-beds this year, no pinks reached those waters. Since, as pointed out above, the pinks invariably mature at two years of age, the very abundant odd-year run of pinks spawning in the Fraser River above Hell’s Gate Canyon was completely wiped out. Table 52 .-—Pink salmon index of abundance from traps north of Deception Pass, 1907-33 Year Catches Efficiency weights Number of traps Index of abundance Year Catches Efficiency weights Number of traps Index of abundance 1907 1, 103,010 1, 220, 370 4, 136, 212 689, 171 343, 969 1, 453, 493 1, 225, 884 1, 833, 634 1, 713, .587 1, 557, 144 10 203. 579 1921. 967, 059 1, 354, 003 937, 627 1, 731, 927 1, 556, 160 1,581,422 1, 556, 160 1, 500, 928 1,520,336 1,394, 611 30 65. 837 1909 5 354. 791 1923.. 26 87. 009 1911 24 284. 570 1925 27 59. 290 1913 i 487, 853 909, 462 1, 517, 903 988, 092 20 284.517 1927 1, 395', 948 947, 559 1, 262, 263 524, 512 26 89. 705 1915 31 49. 599 1929 27 63. 132 1917. 29 88. 580 1931 24 83. 025 1919 25 63. 455 1933 23 37. 610 812 BULLETIN OF THE BUREAU OF FISHERIES The southern pink-salmon index is very different from the northern (see table 53). There was no fall after 1913 because the Hells Gate slide, which so seriously affected the northern run, had, of course, no effect on the spawning grounds of the southern run. From 1915-33 the two indices differ at many points, the northern index not showing the extreme fluctuations of the southern. In 1919 the southern abundance was extremely low, possibly due to the intense fishery of 1917. The highest point reached was in 1931. In this southern district our data show no depletion within a recent year. Table 53. — Pink salmon index of abundance from traps south of Deception Pass, 1907-33 Year Catches Efficiency weights Number of traps Index of abundance Year Catches Efficiency weights Number of traps Index of abundance 1907 1909 400, 054 185, 762 2 215. 358 1921 1923 223, 143 495. 933 134, 863 432, 280 3 7 165. 459 114. 725 1911 314, 603 290, 426 134, 375 338, 059 397, 919 3 108. 325 1925.. 254, 732 492, 875 432, 280 432. 280 7 58. 928 1913 1.54, 210 531, 439 432, 541 49, 891 1 114. 761 1927 7 114. 018 1915 5 157. 203 1929... 485, 619 813, 810 334, 525 432, 280 239, 527 290, 914 7 112. 339 339. 757 1917-.. 6 108. 701 1931 4 1919 397, 919 6 12. 538 1933 _... 5 114.991 ABUNDANCE FROM PURSE-SEINE CATCHES The purse-seine catches have been a fairly reliable guide to the abundance of pink salmon in Puget Sound since 1911, except in 1913 and to some extent in 1917, as they were usually the chief object of the summer seine fishery. To measure the abundance the average catch per seine boat delivery has been employed, using all of the catches made from August 5-September 8, inclusive, these 5 weeks taking in all of the important part of the season. Because of the difference in efficiency between purse-seine vessels of different size, the number of deliveries made by vessels of each 5-net-ton class was tabulated separately, and then weighted according to the efficiency scale for all species (see p. 738). The weighted numbers of deliveries for all sizes of purse-seine vessels were pooled, as were the catches, and the average catch per weighted delivery calculated (see table 54). Table 54. — Pink salmon abundance from Puget Sound purse seines Year Number of fish Number of catches Weighted number of catches Average catch 1911 441,920 194 175.6 2, 516. 63 1913 471,627 272 301. 3 1, 565. 31 1915... 1, 059, 304 1,558 1, 866. 2 567. 63 1917... 763, 626 705 898.3 850. 08 1919 251,337 272 391.0 642. 81 1921 699, 099 982 1, 408. 0 496. 52 Year N umber of fish Number of catches Weighted number of catches Average catch 1923. 1, 493, 749 1, 514, 755 1, 136 1, 621. 7 921. 10 1925 . 745 1, 067. 4 2, 181. 5 4, 546. 0 1, 419. 11 1927 1, 800, 778 3, 686, 797 3, 399, 825 3, 677, 705 1, 497 825. 48 1929 3, 019 3,678 5,003 811. 00 1931. 5' 765. 5 589. 68 1933 7, 719. 3 476. 43 COMPARISON OF PURSE SEINE AND TRAP INDICES The indices of abundance from Puget Sound purse seines and northern traps are compared in figure 29. The similarity between the indices is striking, as in only 2 out of 12 years do they show any degree of divergence, namely 1913 and 1925. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 813 In 1913 the purse seiners were fishing primarily for sockeyes. Consequently, when the sockeye run was over the seiners quit; only 4 out of 272 catches being made in the last week of the 5-week period covered, and 89 catches being made in the first week; before the pinks wore really abundant. For this reason the difference in level of the curves in 1913 cannot be considered significant. In 1925 the purse-seine curve is considerably higher than the northern trap curve, but the data do not suggest any reason for this difference. The purse seines take large quantities of pink salmon from the areas north of Deception Pass, and the close correspondence with the northern trap index would seem to indicate that the southern run does not con- tribute much to their catch. Correlating the northern trap index with the average purse-seine delivery gives a coefficient of correlation of .8468 with a probability of less than .01. Such a high correlation certainly indi- cates that they are drawing largely upon the same gen- eral population. CHUM SALMON By George A. Rounsefell GENERAL LIFE HISTORY Chum salmon spawn in the lower tributaries of the. main rivers of the region as well as in a great many of the smaller streams. They are the latest running of the Pacific salmons; although there a're runs that reach some streams as early as Septem- ber, the bulk of the run is much later. In earlier years chums were often seined in salt water as late as January. As with the pink salmon, the chum-salmon fry, upon emerging from the gravel of the spawning beds, migrate to salt water. Because less is known of the life history of the chums than of the other species of Pacific salmon, data were collected during the 1935 fishing season on several hundred adults. Out of 890 individuals taken in Admiralty Inlet between October 1 0 and November 11, the scales could be read for age on 875. Of these there were 334 three-year-olds, 463 four-year-olds, and 78 five-year-olds, or percentages of 38, 53, and 9. However, none of these percentages are more than an indication of the true proportion, since the percentage of 3-year-olds increases, and that of 5-year-olds de- creases, as the season progresses.7 These ages compare favorably with those reported by Pritchard (1932) in Johnstone Strait, except that we had fewer in their fourth year. Figure 29.— Showing two measures of the abundance of pink salmon. One measure is an index calculated from the catches of Puget Sound traps located north of De- ception Pass. The other measure of abundance is the average weighted purse-seine delivery for the period from August 5 to September 8, inclusive. The average purse-seine delivery has been plotted to one-tenth scale to facilitate comparison between the two measures. Note their close correspondence. 7 These chum salmon ages were read by Milton Lobell. 814 BULLETIN OF THE BUREAU OF FISHERIES METHOD AND LOCALITY OF CAPTURE Chum salmon are taken chiefly by purse seines in Puget Sound and the Gulf of Georgia and by gill nets in the Fraser River. Chums run so late in the fall that most of the traps close before they are abundant, and very few are taken on Swiftsure Bank, as the weather is not conducive to ocean fishing at that season The chum- salmon catches have depended as much upon economic conditions as upon abundance, usually being larger on the even-numbered years, due to the absence of pink salmon, which furnish the cheaper grades on the odd-numbered years. The actual number of chum salmon caught in Puget Sound is shown in table 55. These figures cannot be correlated with the canned pack as large quantities of chums were sometimes bought in British Columbia. The numbers taken in adjacent Canadian waters cannot be estimated from material on hand as chums were used for canning, freezing, smoking, dry-salting, and for export in a raw state. SEASONAL OCCURRENCE IN NORTHERN AND SOUTHERN DISTRICTS With the chums, as with the pinks, there is a considerable difference in time of run between the northern and southern districts of Puget Sound. However, the southern pink salmon run earlier than the northern, whereas for chums the situation is reversed. Table 55. — Puget Sound chum salmon catch, 1918-84 Year Purse seines 1 Traps Other gear Year Purse seines 1 Traps Other gear 1913 445, 384 1, 431, 983 1, 280, 931 1, 852, 859 832, 922 799, 833 1,112,404 641, 213 211, 198 405, 905 528, 542 159, 473 254, 154 177, 764 191,492 131, 804 173, 782 185, 292 111,433 32,414 89, 427 74, 465 1924 713, 258 438, 408 838, 371 398, 549 852,411 1, 291, 448 903, 081 581, 781 1, 009, 605 418, 620 777,833 84, 200 67, 204 125, 164 99, 472 142, 708 128, 214 78, 688 85, 576 60, 017 67, 445 51, 893 62, 525 31, 200 100, 160 28, 847 48, 982 66, 772 29, 591 15, 136 32, 687 18, 074 37, 103 1914 ... 1925 1915 127, 383 146, 757 130, 289 182, 956 177, 395 30, 424 26, 581 6,898 34, 875 1926 1916 1927 1917 ... 1928... 1918 . 1929 1919 1930 1920 1931 1921 . 1932 1922 . 1933 1923 1934 __ 1 Includes other gear in 1913 and 1914. Table 56. — Seasonal occurrence in traps of chum salmon in northern and southern districts, 1900-34 Week ending North of Deception Pass South of Admiralty Head Week ending North of Deception Pass South of Admiralty Head Percent- age Cumulative percentage Percent- age Cumulative percentage Percent- age Cumulative percentage Percent- age Cumulative percentage 0. 001 0. 001 Sept. 22 5.880 14. 878 1.929 13. 650 .001 .002 Sept. 29 10. 520 25. 398 5.804 19. 454 .003 .005 Oct. 6 15. 704 41. 102 6. 014 25. 468 0. 007 0. 007 .005 .010 Oct. 13 17. 387 58. 489 9. 011 34. 479 .010 .017 .012 .022 Oct. 20_ 15. 033 73. 522 11. 467 45. 916 .002 .019 .011 .033 Oct. 27 14.115 87. 637 13. 407 59. 353 .000 .019 .010 .043 Nov. 3 9. 120 96. 757 10. 943 70. 296 .015 .034 . 012 .055 Nov. 10- _ 3. 244 100. 001 9. 075 79. 371 July 7 . 003 . 037 .037 .092 8. 621 87. 992 .003 .040 .046 . 138 4. 923 92. 915 July 21 . 010 . 050 . 089 .227 2. 798 95. 713 July 28 . 035 .085 .241 .468 1.645 97. 358 . 079 . 164 .635 1. 103 .646 98. 004 .332 . 496 1.058 2. 161 Dec. 22 1.004 99. 008 . 691 1. 187 1. 506 3. 667 Dec. 29 ___ .829 99. 837 Aug. 25 1. 146 2. 333 2. 221 5. 888 .099 99. 936 1.442 3. 775 2. 031 7. 919 Jan. 12_„__ .022 99. 958 1.880 5. 655 2. 026 9. 945 .009 99. 967 Sept. 15 3. 343 8.998 1. 776 11. 721 Feb. 23 .031 99. 998 SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 815 For the district north of Deception Pass, data were analyzed for seven traps catching 124,831 fish from 1902-34. For the district south of Admiralty Head, the six tiaps used caught 821,263 chums from 1900-1934. In the northern district the run really commences about the middle of September and reaches its peak by October 10. In the southern district there is a small early run in late August and early September, but the main run does not really start until nearly the end of September, and the peak is not reached until October 24, just 2 weeks later than the northern run. Because of the difference in time of run in the 2 districts, only a small fraction of the northern chums are protected by the closed season commencing November 11. This same closing date, however, protects about 20 percent of the southern run. ABUNDANCE FROM ADMIRALTY INLET TRAPS For the chum-trap index, 8 Admiralty Inlet traps were employed, 3 each from the Admiralty Bay and Bush Point areas, and 1 each from the Oak Bay and Point No Point areas. The total catch of each trap up to and including November 3 was used, as this period normally includes 70 percent of the southern run and it was not feasible to use a longer period as many of the traps ceased fishing by that date. In 1934, 1921, and 1920 they all closed too early to be usable. The index was calculated in the same manner as that described for sockeyes. Three traps, over a 19-year period, were used for the standard curve. Because a small number of traps were used, and only a portion of the run occurred during the period they fished, the index is not especially reliable for any particular year. However, it does show that the chums of the southern district were very abundant at one time. In the last 12 years they were less than half as abundant as during the period just previous to the war (see table 57). ABUNDANCE FROM PURSE SEINES An estimate of the abundance of chums was made from the Puget Sound seine catches. The average catch per weighted delivery, each delivery was weighted by the efficiency weight given in the purse-seine section of this report, was first obtained for a 6-week period from September 23-November 3. From 1910-34 data were available for 25,838 deliveries containing 5,322,546 chums. The first 2 weeks of the 6-week period chosen represented a large number of catches but only a few chums, the run having not yet attained any proportions. The efforts of the fleet up to this time had been almost wholly directed toward the capture of cohos. For this reason the average delivery was also obtained for a 4-week period from October 7-November 3, which, over the 25 years, represented 19,584 catches and 4,973,971 fisb (see table 58). The average catch per delivery obtained from the purse seine data appears to reflect economic factors as well as abundance. Thus 11 out of 12 of the even-num- bered years are higher than the year preceding them, whereas 8 out of 12 of the odd- numbered years are lower than the preceding year. Since the chums vary from 3-5 years in age at maturity, there is no apparent biological reason for a higher level of abundance in the even years. 816 BULLETIN OF THE BUREAU OF FISHERIES Table 57. — Chum, index of abundance for Admiralty Inlet traps, 1902-38 Year Catches Efficiency weights Number of traps Index of abundance Index from si andard curve 1902 21,952 15, 324 2 143. 252 1903 1904... 1905 36, 589 34, 921 4 104. 776 118.043 1906 34,911 15, 324 2 227. 819 1907 19, 068 33, 988 2 56. 102 1908 26, 221 24, 586 3 106. 650 106. 650 1909 86, 368 24, 586 3 351. 289 351. 289 1910 94, 885 41, 906 4 226. 423 207. 435 1911 67, 474 41,906 4 161.013 155. 483 1912 48, 357 41,906 4 115. 394 155. 829 1913 24, 777 18, 503 2 133. 908 1914 31, 730 24, 586 3 129.057 129.057 1915 32, 629 41,906 4 77. 862 83. 238 1916. 26, 690 24, 586 3 108. 558 108. 558 1917 35, 209 48, 336 5 72. 842 68. 592 1918... 31,578 48, 336 5 65. 330 65. 907 1919 28, 815 29,833 3 96. 588 1920 1921 1922 8,871 24, 992 2 35. 495 1923 22, 897 56, 660 6 40.411 36. 094 1924 29. 604 50, 577 5 58. 533 1925 13, 809 56,660 6 24. 372 17. 929 1926 34, 565 56 660 6 61. 004 70. 630 1927 26, 439 50, 327 6 52. 534 55. 349 1928. 24, 539 50, 327 6 48. 759 62. 637 1929 45, 843 67, 647 7 67. 768 61. 962 1930 20, 883 67, 647 7 30. 871 19. 556 1931 21,346 34; 737 3 61.450 1932 9, 626 18, 659 2 51. 589 1933 24, 275 69, 323 6 40. 920 38. 062 Total 929, 950 Table 58. — Chum-salmon index of abundance from Puget Sound purse seines Year From September 23-November 3 From October 7-November 3 Number of fish Number of catches Weighted number of catches Average catch Number of fish Number of catches Weighted number of catches Average catch 1910 7, 211 20 18.40 391.90 7, 211 20 183. 40 391. 90 1911 42. 190 111 103. 24 408. 66 42, 190 111 103. 24 408. 66 1912. 88, 268 163 155. 44 567. 86 86, 156 124 117. 56 732. 87 1913 37,612 174 199. 78 188. 27 36, 851 163 187. 34 196. 71 1914 169, 628 360 405. 10 418. 73 154, 475 261 295. 26 523. 18 1915 129, 855 620 779. 38 166. 61 121,178 461 576. 84 210. 07 1916 157, 217 665 786. 90 199. 79 151,755 520 614. 76 246. 85 1917. 190, 120 1,471 1, 838. 92 103. 39 186, 042 1, 330 1, 659. 78 112. 09 1918 149, 824 749 973. 54 153. 90 140, 178 598 764. 04 183. 47 1919 174,512 753 963. 04 181.21 154, 926 551 709. 50 218. 36 1920 76, 038 298 394. 04 192. 97 70, 043 217 283. 18 247. 34 1921 48, 546 688 1, 004. 46 48. 33 40, 606 383 551. 74 73. 60 1922 79.111 552 761. 30 103. 92 71, 675 412 541.50 132. 36 1923.. 146, 388 671 971. 16 150. 74 136, 875 526 759. 82 180. 14 1924 176, 332 490 640. 26 275.41 157, 939 376 486. 74 324. 48 1925 117, 305 817 1, 102. 88 106. 36 111,886 672 901.16 124. 16 1926 285, 644 1, 116 1, 698. 14 168. 21 258, 216 758 1, 149. 84 224. 57 1927 95, 651 1,061 1,581.94 60. 46 88, 328 766 1, 122. 44 78.69 1928. 462, 882 2,004 3,134.97 147. 65 441,033 1,511 2, 378. 50 185. 42 1929 725, 733 2, 527 3, 766. 80 192. 67 674, 788 1,858 2, 721.71 247. 93 1930 342,117 1.167 1, 904. 25 179. 66 327, 533 695 1, 146. 29 285. 73 1931 335, 268 2,061 3,314. 75 101. 14 319, 927 1, 596 2, 538. 51 126. 03 1932 693, 046 2,528 4, 140. 78 167. 37 648, 097 1,931 3,113. 67 208. 15 1933 230, 945 2,519 4.013. 42 57.54 215, 860 1,988 3, 168. 02 68.14 1934 361, 103 2,253 3, 425. 82 105. 41 330, 203 1,756 2,651.12 124.55 There is usually a greater demand for chums in the even-numbered years, owing to the lack of pinks, and the deliveries are raised by increased effort on the part of the SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 817 fishermen. Another factor may be lessened competition between gear on the even years, as usually there is a smaller fall fleet than on the odd years. All that can safely be said is that the purse-seine data seem to indicate that the general trend has remained about the same since 1915. Before that the data are scant but seem to indicate a higher level of abundance. SUMMARY By George A. Rottnsefell and George B. Kelez THE GILL-NET FISHERY On the Fraser River sock eye salmon was at first used to the practical exclusion of other species, but in later years the fishery was extended to include the others. Drift gill nets, introduced in 1864, have been the only gear used there. The fishery developed rapidly and the number of canneries increased steadily, reaching maxima of 49 plants in 1898 and in 1901 ; mergers and decreasing runs caused many of the plants to be closed thereafter. Less than a dozen have operated in any year since 1921. The Fraser River gill nets were at first fished mainly by Indians, later more white fishermen were engaged, and Japanese fishermen were introduced on the river in 1888. The early flat-bottomed skiffs were replaced in the 1890’s by round-bottomed Columbia River boats, which were generally equipped with engines by about 1914. Each of these changes increased the efficiency of the individual units of gear. The number of gill nets licensed on the river reached a peak of more than 3,600 in 1900, but decreased considerably within a few years, until at the present time about half that number are employed. Regulations, some in effect since 1878, have limited the size and the mesh of gill nets, and have provided for a week-end closed season intended to permit escapement of salmon up the Fraser River. The sockeye, pink, and chum salmon overlap but slightly, in their seasonal occurrence on the Fraser River, but the runs of coho and king salmon are more extended. The bulk of the sockeye catches have been made between July 22 and August 25, those of the pinks, which are abundant only in odd-numbered years, between September 2 and September 29, and of the chums between October 7 and November 10. The major catch of cohoes is made between September 9 and October 13, that of the kings between July 1 and September 15. Gill nets are of minor importance on Puget Sound, where they are used chiefly in or adjacent to the estuaries of the larger Puget Sound rivers, catching mainly coho and king salmon. THE TRAP FISHERY Salmon traps were driven in Puget Sound as early as 1880, but were not developed to a point of success until about 1891, at which time the first sockeye cannery was built on Puget Sound. This success caused a great expansion of the American fishery, and 163 traps were driven by 1900. The peak year for traps was 1913, when 168 were driven on Puget Sound, 2 in the Canadian waters of Boundary Bay, and 6 near Sooke on Vancouver Island. Available data show that between 1895 and 1934, over 156,- 818 BULLETIN OF THE BUREAU OF FISHERIES 000,000 salmon were taken by traps, 53 percent of which were caught in the waters north of Sandy Point, 27 percent in the region of the San Juan Islands, 4 percent on the west shore of Whidbey Island, north of Point Wilson, 5 percent west of Point Wilson, and 1 1 percent in areas south and east of Point Wilson. In the period from about 1900-1934 the average number of days of operation of each trap has increased from 46-95 days in Boundary Bay, and in Admiralty Inlet the time at which they commence operations has advanced 85 days. The average seasonal occurrence of each species of salmon is quite distinct in the trap catches. Kings run very early, 40 percent of the catch being made by June 30. They are followed by the sockeyes, whose run is practically over by August 25, at which date only 40 percent of the pinks have been taken. The latter species reaches a peak about August 29, the cohos about October 1 and the chums about October 23. THE PURSE-SEINE FISHERY Purse seines were used in this region before 1882, and within a decade had become the most important type of gear on Puget Sound. Later they were surpassed by the traps, but the introduction of the gasoline engine, completed by 1907, returned them to a place of considerable consequence in the fishery. The purse-seine vessels have improved steadily in design and equipment, and have increased in size throughout the history of this fishery. The average efficiency of the fleets has correspondingly increased so that, although the modern fleet is smaller in numbers than were those of many earlier years, the total fishing efficiency of today is greater than in all but 1 previous year. Both fishing season and the size of the fleets vary considerably in odd- and even- numbered years. The summer fishery is most important in the odd-numbered years, when pink salmon are abundant, while the fall fishery for cohoes and chums is considerably greater in even years. The number of vessels fishing is usually greater in odd than in even years. The larger vessels fish on the high seas in spring and early summer, moving into Puget Sound later in the season. Seasonal occurrence of the various species in purse-seine catches is similar to that in trap catches, but the periods of abundance are more prolonged. From 1917-34, pink salmon have averaged 75 percent of the catch in odd years, but less than 1 percent in even years. Over this 18-year period their average was 37.44 percent of the catch, chums were 32.07 percent, sockeyes 15.63 percent, cohoes 14.16 percent, and kings 0.70 percent. The proportion of pink salmon in odd and even years at the cape is similar to that on Puget Sound. During the period from 1927-34, pinks averaged 46.54 percent of the cape catches, cohoes were 36.83 percent, and sockeyes 14.84 percent. Chums and kings both averaged less than 1.0 percent. THE TROLL FISHERY Coho and king salmon provide almost the entire catch of the troll fishery, which was of slight consequence until the introduction of engines increased the efficiency of the boats. During recent years almost the entire troll fleet has fished at the cape, SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 819 the season extending from April to October. Over the 8-year period from 1927-34, Puget Sound trollers took 104,692 cohos and 18,285 kings, while the cape fleet took 2,411,312 cohos and 1,545,178 kings. SOCKEYE SALMON The Fraser River produces the only sockeye run of consequence in the region. From 1873-1934, over 250 million sockeyes have been canned, of which 46 percent were taken by Fraser River gill nets, 37 percent by traps, 14 percent by purse seines, and 3 percent by miscellaneous gear. An analysis of seasonal occurrence from gill-net catches indicates that the heavy, early-season run of superior quality sockeyes has suffered the greatest decrease in abundance. Indices of abundance from gill-net and trap catches both show a tremendous decline in all cycles. The cycle of years ending in 1934 fell about 39 percent in abundance between 1898 and 1914, reached a very low point in 1918, and has been increasing considerably in each cycle after that date. The big year cycle, 1933, etc., tremendously abundant in early years, was severely reduced by over fishing and the Hell’s Gate slide, but has recuperated slightly in recent years. The cycle of years containing 1932 was the least abundant in the early years of the fishery, and declined still further in 1904. The run of 1932 was the best since that of 1912. The cycle of years containing 1931 has been the least abundant since 1899, although it was second in abundance for several years preceding that date. COHO SALMON Cohoes are the most widely distributed species of salmon found in the region. Approximately 98 percent mature at 3 years of age, and the migration to the spawn- ing beds occurs during the fall months, at which period the greater part of the catch is made. During the 9 years from 1926-34, approximately 5% million cohoes were taken on the high seas, a slightly greater number in Puget Sound waters, and about one-half million in the Fraser River. The greater part of the Puget Sound catches are taken in the southern part of that district. Seasonal occurrence is generally earlier in the northern than in the southern districts. Indices of abundance from both trap and purse seine catches show a high level of abundance in early years and a present level that is lower than at any previous time in the history of the fishery. KING SALMON King salmon are caught from early spring to fall, the bulk of the catches being made during early summer. During the 8 years from 1927-34, nearly 4 million were landed in the region, of which trollers landed approximately 40 percent, traps 39 percent, gill nets 15 percent, and purse seines and miscellaneous gear 6 percent. Indices of abundance from trap catches do not show any definite trends in the north- ern areas, but do indicate a decrease in the runs of recent years in the southern part of Puget Sound. 71941—38 9 820 BULLETIN OF THE BUREAU OF FISHERIES PINK SALMON In the 10-year period from 1925-34, the pink salmon catch in the region was more than 50,000,000 fish, of which 60 percent were taken by purse seines, 27 percent by traps, 12 percent by Fraser River gill nets, and 2 percent by minor gear. Of the trap-caught fish, taken between 1895 and 1934, about 5 times as great a catch was made north of Deception Pass as south of that point. The peak of the seasonal runs in the southern part of Puget Sound is about 10 days earlier than in the northern part. Indices of abundance from purse seines and traps indicate that, following the obstruction at Hell’s Gate in 1913, which prevented them from reaching their spawn- ing grounds in the upper Fraser River, the pinks declined to about one-quarter of their former abundance. CHUM SALMON The runs of chum salmon occur during the last part of the fishing season, and have been taken chiefly by purse seines in the Puget Sound district, as most of the traps have ceased fishing by the time that the runs appear in any quantity. The chums of Admiralty Inlet were found to be approximately 38 percent 3-year-olds, 53 percent 4-year-olds, and 9 percent 5-year-olds at maturity. The peak of the runs in the northern part of Puget Sound occurs about 2 weeks earlier than in the southern part. An index of abundance from Admiralty Inlet traps shows abundance in recent years to be less than half that of the period previous to the war. The average size of delivery by purse seines also indicates a higher level of abundance previous to 1915. BIBLIOGRAPHY Babcock, John P. 1902-1932. The spawning-beds of the Fraser River. In annual reports of the Commissioner of Fisheries (British Columbia) for the years 1901-1931. Victoria. Babcock, John P. 1918. Salmon-fishery of the Fraser River district. Appendix to the report of the Commissioner of Fisheries (British Columbia) for the year ending December 31, 1917, pp. Q 116-Q 123. Victoria. Babcock, John P. 1920. The Fraser River salmon situation. Canada’s position. Canadian Fisherman, vol. VII, No. 6, pp. 101-104. British Columbia Commissioner of Fisheries. 1901-1934. Annual reports of the Commis- sioner of Fisheries (British Columbia) and Appendices. Victoria. Clemens, Wilbert A. 1930. Pacific salmon migration: The tagging of the Coho salmon on the east coast of Vancouver Island in 1927 and 1928. Biological Board of Canada, Bulletin 15, pp. 1-19, 1930. Toronto. Clemens, Wilbert A. 1932. Pacific salmon migration: The tagging of the spring salmon on the east coast of Vancouver Island in 1927 and 1928 with notes on incidental tagging of other fish. Biological Board of Canada, Bulletin 27. Ottawa. Clemens, Wilbert A. 1935. The Pacific salmon in British Columbia waters. Report of the Commissioner of Fisheries (British Columbia) for 1934, pp. K 103-K 105. Victoria. Clemens, Wilbert A. and Lucy S. 1926-1934. Contributions to the life history of the sockeye salmon. Appendices to the annual reports of the Commissioner of Fisheries (British Colum- bia). Victoria. Cobb, John N. 1911. The salmon fisheries of the Pacific coast. U. S. Bureau of Fisheries Docu- ment No. 751. Appendix to the report of the U. S. Commissioner of Fisheries for 1910. 180 pages. Washington. Cobb, John N. 1930. Pacific salmon fisheries. Bureau of Fisheries Document 1092. Appendix XIII to the report of the U. S. Commissioner of Fisheries for 1930. 4th edition, pp. 409-704. Washington, SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 821 Collins, J. W. 1892. Report on the fisheries of the Pacific coast of the United States. Report U. S. Fish Commission for 1888, pp. 3-269, Fisheries of Puget Sound, pp. 243-269. Washington. Crawford, John M. 1902. (Letter to the State Fish Commissioner.) 13th annual report of the State Fish Commissioner (Washington), pp. 10-15. Seattle. Davidson, Frederick A. 1934. The homing instinct and age at maturity of pink salmon (On- corhynchus gorbuscha) . Bulletin, United States Bureau of Fisheries, vol. XLVIII, 1933 (1934) pp. 27-39. Washington. Dominion of Canada 1882-1891. Annual reports of the Dept, of Fisheries of Canada. Dominion of Canada 1892-1914. Annual reports of the Dept, of Marine and Fisheries of Canada. Dominion of Canada 1915-1919. Annual reports of the Naval Service (Canada). Dominion of Canada 1920-1930. Annual reports of the Dept, of Marine and Fisheries of Canada, Fisheries Branch. Dominion of Canada 1931-1934. Annual reports of the Dept, of Fisheries of Canada. Dominion of Canada 1927-1934. Fisheries Statistics of Canada. Fisheries Statistics Branch. Dominion Bureau of Statistics. Doyle, Henry. 1920. History of the Pacific coast salmon industry. Canadian Fisherman, vol. VII, No. 6, pp. 72-74. Foerster, R. Earle. 1929. An investigation of the life history and propagation of the sockeye salmon ( Oncorhynchus nerka) at Cultus Lake, British Columbia. No. 1. Introduction and the run of 1925. Contributions to Canadian Biology and Fisheries, New Series, vol. V, No. 1, 1929. Toronto. Foerster, R. Earle. 1929. An investigation of the life history and propagation of the sockeye salmon ( Oncorhynchus nerka ) at Cultus Lake, British Columbia. No. 3. The down-stream migration of the young in 1926 and 1927. Contributions to Canadian Biology and Fisheries, New Series, vol. V. No. 3, 1929. Toronto. Foerster, R. Earle. 1934. An investigation of the life history and propagation of the sockeye salmon ( Oncorhynchus nerka ) at Cultus Lake, British Columbia. No. 4. The life history cycle of the 1925 year class with natural propagation. Contributions to Canadian Biology and Fish- eries, vol. VIII, No. 27 (Series A, General, No. 42). Toronto. Fraser, C. McLean. 1917a. On the scales of the spring salmon. Contributions to Canadian Biology for 1915-1916, pp. 21-38. Ottawa. Fraser, C. McLean. 1917a. On the life-history of the coho. Contributions to Canadian Biology for 1915-16, pp. 39-52. Ottawa. Fraser, C. McLean. 1920. Growth rate in the Pacific salmon. Transactions Royal Society of Canada, Series III, vol. XIII, for 1919, Sect. V, pp. 163-226. Ottawa. Fraser, C. McLean. 1921. Further studies on the growth rate in Pacific salmon. Contri- bution to Canadian Biology, 1918-1920, pp. 7-27. Ottawa. Gilbert, Charles H. 1912. Age at maturity of the Pacific Coast salmon of the genus Oncorhyn- chus. Bulletin, United States Bureau of Fisheries, vol. XXXII, pp. 1-22. Washington. Gilbert, Charles H. 1913. The salmon of Swiftsure Bank. Appendix to the report of the Com- missioner of Fisheries (British Columbia) for the year ending December 31, 1912, pp. I 14r-I 18. Victoria. Gilbert, Charles H. 1913-1924. Contribution to the life history of the sockeye salmon. (1-10.) Reports of the British Columbia Commissioner of Fisheries for the years 1913-15, 1917-19, 1921-24. Victoria. Gilbert, Charles H. 1923. The salmon of the Yukon River. Bulletin, U. S. Bureau of Fisheries for 1921-22, vol. XXXVIII, pp. 317-332. Washington. Greenwood, W. H. 1917. The salmon fishermen. Canadian Fishermen, July 1917, pp. 288-290 292-294. Hittell, John S. 1882. Commerce and Industries of the Pacific Coast. Bancroft & Co. San Francisco. Howay, F. W. 1914. British Columbia, from the earliest times to the present. 4 vols., illus. Vancouver. &22 BULLETIN OF THE BUREAU OF FISHERIES Morice, A. G. 1904. The history of the northern interior of British Columbia, formerly New Caledonia (1660-1880), 349 pages. Toronto. Mottley, Chas. McC. 1929. Pacific salmon migration. Report on the study of the scales of the spring salmon tagged in 1926 and 1927 off the west coast of Vancouver Island. Contri- butions to Canadian Biology and Fisheries, N. S., vol. 4, No. 30, 1929. Myers, George T. 1905. Early canning days on Puget Sound. Pacific Fisherman Annual, January 1905. pp. 25-6. Seattle. O’Malley, Henry and Willis H. Rich. 1919. Migration of adult sockeye salmon in Puget Sound and Fraser River. U. S. Bureau of Fisheries Document 873. Appendix VIII to the report of the United States Commissioner of Fisheries for 1918. pp. 1-38. Washington. Pacific Fisherman. 19Q4U1933. Pacific Fisherman Annuals. Seattle, Wash. Pritchard. Andrew L. 1930. Pacific salmon migration: The tagging of the pink salmon and the churn salmon in British Columbia in 1928. Biological Board of Canada, Bulletin 14. Toronto. Pritchard, Andrew L. 1932. Pacific salmon migration: The tagging of the pink salmon and the chum salmon in British Columbia in 1929 and 1930. Biological Board of Canada, Bulletin 31, 1932. Ottawa. Pritchard, Andrew L. 1934. Pacific salmon migration: The tagging of the coho salmon in British Columbia in 1929 and 1930. Biological Board of Canada, Bulletin 40, 1934. Ottawa. Pritchard, Andrew L. 1934. Pacific salmon migration: The tagging of the spring salmon in British Columbia in 1929 and 1930. Biological Board of Canada, Bulletin 41, 1934. Ottawa. Pritchard, Andrew L. 1934. The recovery of marked pink salmon in 1934. Biological Board of Canada. Progress reports of Pacific Biological Station, Nanaimo, B. C. and Pacific Fisheries Experimental Station, Prince Rupert, B. C., No. 22, pp. 17-18. Prince Rupert. Pritchard, Andrew L. 1936. Facts concerning the coho salmon ( Oncorhynchus kisutch ) in the commercial catches of British Columbia as determined from their scales. Progress reports of Pacific Biological Station, Nanaimo, B. C. and Pacific Fisheries Experimental Station, Prince Rupert, B. C., No. 29, pp. 16-20. Prince Rupert. Rathbtjn, Richard. 1899. A review of the fisheries in the contiguous waters of the State of Washington and British Columbia. Report U. S. Fish Commission for 1899, pp. 251-350, plates. Washington. Rich, Willis H. 1925. Growth and degree of maturity of Chinook salmon in the ocean. Bulletin, United States Bureau of Fisheries, vol. XLI, pp. 15-90. Washington. Rich, Willis H., and Edward M. Ball. 1928. Statistical review of the Alaska salmon fisheries. Part I: Bristol Bay and the Alaska Peninsula. Bulletin, United States Bureau of Fisheries, vol. XLIV, pp. 41-95. Washington. Rich, Willis H., and Edward M. Ball. 1931. Statistical review of the Alaska salmon fisheries. Part II: ChigDik to Resurrection Bay. Bulletin, United States Bureau of Fisheries, vol. XLIV, 1930, pp. 643-712. Washington. Rich, Willis H., and H. B. Holmes. 1928. Experiments in marking young cliinook salmon on the Columbia River, 1916-1927. Bulletin, United States Bureau of Fisheries, vol. XLIV, pp. 215-264. Washington. Smith, E. Victor. 1921. The taking of immature salmon in the waters of the State of Washington. The taking of immature salmon in the waters of the State of Washington during the 1920 fishing season. The thirtieth and thirty -first annual reports of the State Fish Commission (Washington). Olympia. United States Bureau of Fisheries 1925-1934. Fishery industries of the United States. Ap- pendices to the reports of the U. S. Commissioner of Fisheries for 1925-1934. Washington. McLean, John (Wallace, W. S. ed.). 1932. Notes of a twenty-five year’s service in the Hudson’s Bay Territory. The Champlain Societjr, Toronto. 1932. x Washington State 1890-1920. Annual reports of the State Fish Commissioner. Washington State 1921-1926. Annual reports of the State Supervisor of Fisheries. Washington State 1927-1930. Annual reports of the Division of Fisheries, State Department of Fisheries and Game. Washington State 1931-1934. Annual reports of the State Department of Fisheries. SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 823 Wilcox, William A. 1895. Fisheries of the Pacific Coast. Appendix to the report of the Com- missioner, U. S. Fish Commission for 1893, pp. 139-304. Wilcox, William A. 1898. Notes on the fisheries of the Pacific Coast in 1895. Appendix to the report of the Commissioner, U. S. Fish Commission for 1896, pp. 575-659. Wilcox, William A. 1902. Notes on the fisheries of the Pacific coast in 1899. Appendix to the report of the Commissioner, U. S. Fish Commission for 1901, pp. 501-574. Wilcox, William A. 1907. The commercial fisheries of the Pacific Coast States in 1904. U. S. Bureau of Fisheries Document No. 612, 74 p. Washington. Report of Com. for 1905. Special paper. Williamson, H. Chas. 1927. Pacific salmon migration: Report on the tagging operations in 1925. Contr. Canad. Biol. Fish, N. S., vol. 3, No. 9. Williamson, H. Chas. 1929. Pacific salmon migration: Report on the tagging operations in 1926, with additional returns from the operations of 1925. Contributions to Canadian Biology and Fisheries, N. S. vol. 4, No. 29. Williamson, H. Charles and Wilbert A. Clemens. 1932. Pacific salmon migration: The tagging operations at Quatsino and Kyuquot in 1927, with additional returns from the opera- tions of 1925 and 1926. Biological Board of Canada, Bulletin 26. Ottawa. o U. S. DEPARTMENT OF COMMERCE Daniel C. Roper, Secretary BUREAU OF FISHERIES Frank T. Bell, Commissioner THE LIFE HISTORY OF THE STRIPED BASS, OR ROCKFISH, Roccus saxatilis (WALBAUM) By John C. Pearson From BULLETIN OF THE BUREAU OF FISHERIES Volume XLIX Bulletin No. 28 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1938 For sale by the Superintendent of Documents, Washington, D. C. Price 10 cents THE LIFE HISTORY OF THE STRIPED BASS, OR ROCKFISH, Roccus saxatilis (WALBAUM) 1 By John C. Pearson, Assistant Aquatic Biologist, United States Bureau of Fisheries CONTENTS Page Introduction 825 Distribution 826 Abundance 827 Spawning grounds 829 Spawning season 830 Size and age at maturity 830 Eggs and young 831 Growth 837 Food habits 839 Movements 840 Fishery 845 Summary 848 Bibliography 849 INTRODUCTION The purpose of this Bulletin is to review a considerable amount of scattered information on the life history of the striped bass, or rockfish, and to present data collected by the author during the course of a study of the spawning habits and migrations of the fish in Chesapeake Bay during 1930-31, and in the Roanoke River, N. C., during May 1937. The striped bass ranks close to the immortal codfish in the vital part which our fishery resources played in early American history. In the year 1623 the Plymouth colonists had but one boat left, and that none of the best, which then was the principal support of their lives, for that year it helped them for to improve a net where- with they took a multitude of bass, which was their livelihood all that summer — Hubbard (1815). The striped bass astonished the early settlers in New England by its abundance and choice food qualities. The Basse is an excellent Fish, both fresh and Salte. They are so large, the head of one will give a good eater a dinner, and for daintiness of diet, they excell the Marybones of Beefe. There are such multitudes, that I have seene stopped into the river close adjoining to my house with a sande (seine) at one tide, so many as will loade a ship of 100 tonnes — Morton (1637). The striped bass and the codfish were probably the first natural resources in colonial America that were subject to conservation measures enacted by statute. 1 Bulletin No. 28. Approved for publication July 28, 1937. 825 826 BULLETIN OF THE BUREAU OF FISHERIES The following act, passed by the General Court of Massachusetts Bay Colony in 1639, ordered that neither cod nor bass should be used as fertilizer for farm crops: At the Generali Courte, holden at Boston, the 22nd of the 3rd M., called May, 1639 — “And it is forbidden to all men, after the 20th of the next month, to imploy any codd or basse fish for manuring the ground, upon paine that every pson, being a fisherman, that shall sell or imploy any such fish for that end, shall loose the said priviledge of exemption from public charges, & that both fishermen, or others who shall use any of the said fish for that purpose, shall forfeit for every hundred of such fish so imployed for manuring ground twenty shillings & so pportionally for a lesser or greater num- ber; pvided, that it shall bee lawful to use the heads & offal of such fish for corne, this order not- withstanding.” The value and probably the limited supply of striped bass seemed to be realized by the colonists within 19 years after the landing of the Pilgrims at Plymouth. Another distinction shared by the striped bass was an act of the Plymouth Colony in 1670 that granted all income that should accrue annually to the Colony from the fisheries at Cape Cod for bass, mackerel, or herring, be employed for and toward a free school in some town of this jurisdiction. As a result of this act the first public (free) school of the New World was made possible through moneys derived in part from the sale of striped bass. A portion of this fund was also expended in helping the widows and orphans of men formerly engaged in the service of the Colony. Appreciation of the striped bass as a superb game fish and a source of unexcelled recreation came during the last century when Herbert (1849) noted that with the sole exception of salmon fishing, striped bass fishing was the finest of the “seaboard vari- eties of piscatorial sport” and that the striped bass was the “boldest, bravest, strong- est, and most active fish that visits the waters of the midland States.” Today the striped bass is esteemed far more by sportsmen than by epicures and its value to the Nation is far greater from a recreational than from a food standpoint. DISTRIBUTION The striped bass, or rockfish, Roccus saxatilis (Walbaum)1 2 ranges along the Eastern coast of North America from the St. Lawrence River, Canada, to the Tche- functa River, La. Introduced on the Pacific coast in San Francisco Bay, in 1879 and 1882, the species now occurs from the Columbia River, Wash., south to Los Angeles County, Calif. The striped bass has probably the most extended geographic range of any American food and game fish. Its ability to exist in fresh, brackish, or salt waters throughout the year and from the cold rivers of Eastern Canada to the sub- tropical bayous of Louisiana, provides a unique record of successful adaptation to environment. (See fig. 1.) The most distant inland fresh-water range on the Atlantic coast from which striped bass have been recorded is Quebec, on the St. Lawrence River. Most coastal rivers from New Brunswick to Georgia contained striped bass in abundance in early colonial times according to various writers.3 Inland coastal ranges for the species have included the Hudson River at Albany, the Delaware River at Lambertville, the Susquehanna River to Luzerne County, Pa., the Potomac River to Great Falls, the Roanoke River at Roanoke Rapids, the Alabama River at Montgomery, and 250 miles up the Sacramento River in California. Few records exist to define the exact range of the striped bass in tributaries of the Gulf of Mexico. The Escambia River, at Pensacola, Fla., the Alabama River, at 1 The scientific name of the striped bass has been corrected from Roccus lineatus (Bloch) to Roccus saxatilis (Walbaum). a Early records of striped bass distribution and abundance are noted by Perley (1850), Atkins (1889), Wood (1634), Mease (1815), Schoepf (1788), and Burns (1886). LIFE HISTORY OF THE STRIPED BASS 827 Montgomery, Ala., and the Tangipahoe River, at Osyka, Miss., have been recorded as localities where the species has been taken previous to 1884. In recent years addi- tional distributions of the fish have been secured from the Tchefuncta River, La., from the Jordan and Wolf Rivers, Miss., and from various coastal streams along the west coast of Florida, from St. Marks to Pensacola.4 On the Gulf coast the striped bass appears to be confined to fresh or brackish coastal rivers and is unknown in salt water. The introduction of the striped bass into California has provided a classic example of successful fish transplantation. From an initial stocking of 435 small fish, brought from New Jersey and liberated in San Francisco Bay in 1879, the species has Figube 1. — Geographic distribution of striped bass, or roekfisb, Roccus saxalilis (Walbaum), within the United States. Circle numbers represent centers of commercial abundance. gradually extended its range to about 850 miles along the Pacific coast. It has become a favorite game fish among many sport fishermen and was reported as a commercial food fish in San Francisco within 10 years after its introduction. The favorite habitat of this species appears to be the fresh and brackish rivers and coastal estuaries. They range freely along the coast line but captures at sea are practically unknown. The record of a 6-pound striped bass, taken on Cod Ledge, 4 miles off Cape Elizabeth, Maine, on October 15, 1931, provides the most distant offshore record. ABUNDANCE The notes of early writers indicate that the striped bass formerly occurred in considerable abundance in areas now recognized as completely depleted of this fish. In addition to the striped bass conservation law, enacted in 1639 by the colonists of * The occurrence of considerable numbers of striped bass in various coastal streams of Louisiana, Mississippi, and Florida, bas been reliably reported to the author by Whitaker Riggs, of Covington, La., U. A. Cuevas, of Cuevas, Miss., and Robert O. Lincoln, of Minneapolis, Minn. 828 BULLETIN OF THE BUREAU OF FISHERIES Massachusetts Bay, New York, in 1758, passed a law to prohibit the sale of bass during the winter months on account of the great decrease of that kind of fish. In 1762 the inhabitants of Marshfield, Mass., also sought to regulate the fishery for bass by passing favorably on a petition to the General Court to enact a bill for the preserva- tion of the fish and to prevent its capture in the winter season. Abundant as the supply of striped bass may have seemed to many early his- torians, its ease of capture, because of its large size and habit of dwelling close ipshore about coastal streams throughout the year, made possible the depletion of the species over the greater part of its northerly range along the Atlantic coast. North of Cape Cod only one localized population of striped bass, at Parker River, Mass., appears to have maintained itself in appreciable quantities (Pearson, 1933 b). The gradual decline of striped bass in southern New England waters was indicated by Bean (1905) who reported a decrease in the annual catch at various angling clubs during the last century. Although overfishing was probably the original cause of depletion in the northern rivers, the industrial uses to which nearly all rivers along the North Atlantic seaboard INCHES Figure 2.— Length-frequency distributions of mature striped bass taken by commercial fishing gear near Havre de Grace, Md., in April and May 1932. Solid line indicates male, and dotted line indicates female fish. have been devoted for many years has had an effective part in the diminishment of the species and in the retardation of reestablishment of the fish in depleted areas. The construction of impassible dams in the lower reaches of the Merrimack, Connecti- cut, and Susquehanna Rivers, shutting off probable spawning grounds; the pollution of the Hudson and Delaware Rivers; and unregulated commercial fishing in all sections have been exceedingly detrimental to the striped bass as well as to other anadromous food fishes such as the shad, smelt, salmon, and sturgeon. Restoration of these fishes to their original abundance clearly involves a restoration of coastal streams, where possible, to their primeval conditions of purity and accessibility, together with adequate restrictions against overfishing. A reduction in the natural supply of striped bass at centers of greatest abundance in Chesapeake Bay and North Carolina has not been so marked as in more northerly areas. This condition has been influenced by the relative absence of industrial devel- opment and the limited population in these localities. Nevertheless, a diminishing annual catch of striped bass is noted in many sections of the Southern States. (See fig. 26.) In California, particularly in the San Francisco Bay region, the striped bass has increased many fold since its introduction. The commercial fishery, prior to 1930, yielded over 17 million pounds despite many years of restricted fishing. In 1931 it became unlawful to take the species by nets and no estimation of their present abun- dance is possible. LIFE HISTORY OF THE STRIPED BASS 829 SPAWNING GROUNDS The first mention of the spawning grounds of the striped bass was probably by Josselyn (1672), who stated: “The Basse is a salt-water fish too but most an end taken in Rivers where they spawn.” 6 A more definite spawning habitat was sug- gested by Schoepf (1788), who, describing the vast number of fishes that came up to the falls on the upper Roanoke River, N. C., every spring, stated that the rockfish (striped bass) especially came up the river in millions to spawn and that being checked at the falls “sprang” and “tumbled” so that the water foamed with the fish. This spawning area in the Roanoke River, 100 miles above tidewater, was so well defined that it was possible for Holton (1874) to artificially fertilize and hatch the eggs of the striped bass at Weldon, N. C. It appears probable that the most important spawning ground for the species, at least along the Atlantic coast, is in the upper Roanoke River where there occurs a fall of 50 feet in about 6 miles and that in the rapids, where the current is exceedingly strong and rendered erratic by islands, boulders, and rocks, the striped bass prefers to spawn. Collections of eggs from ripe fish for artificial propagation have occurred at irregular intervals during the past 64 years at Weldon. It has also been noted by observers that ripe striped bass are found during May at the head of Chesapeake Bay. Past efforts made by fish culturists at Havre de Grace, Md., to obtain eggs suitable for artificial fertilization and hatching proved unsuccessful because of the difficulty experienced in obtaining ripe male and female fish simultaneously. The most important spawning grounds were believed to be located along a rocky swift-running stretch of the Susquehanna River extending from Port Deposit to Octoraro, Md. Eggs of the striped bass were secured by the author in river plankton at night during various times in May and June 1932, in the Susquehanna River at Garrett Island. The occurrence of these eggs, brought down the river by the strong current, definitely establishes a spawning ground for the fish at a point between the locality of capture and the impassable Conowingo Dam, 12 miles upstream. The eggs taken in 1932 would normally have been carried into the head of Chesapeake Bay near the Susquehanna flats. There occurs only one record of a spawning striped bass from the Gulf coast of the United States. A female with eggs was taken on April 7, 1883, in the Alabama River, near Montgomery. The deltas of the Sacramento and San Joaquin Rivers are believed to be the principal spawning grounds of the striped bass in California. The fish appear to spawn, according to Scofield (1931), in fresh -water sloughs and creeks. Free eggs have not been taken in California. The definite records of striped bass eggs in the lower Susquehanna River, and of spawning adults in the upper Roanoke River, indicate that spawning occurs in rock- strewn coastal rivers characterized by rapids and strong currents. Rivers, such as the James, Potomac, and Hudson, offer a similar environment for the spawning fish, and are known to contain either young or ripe adult striped bass. While some writers have stated that the striped bass spawns in brackish water, there is no conclusive evidence to justify this belief. Ripe striped bass, presumably taken at the entrance to the Hudson River off Governors Island, were noted by Rice (1883). A ripe female fish was caught by Corson (1926) near Bamegat Inlet, » Other early notes on the spawning grounds are in the works of Belknap (1792), Mease (1815), and Mitchell (1815), 830 BULLETIN OF THE BUREAU OF FISHERIES N. J. It is well to remember, however, that many anadromous fishes often appear to be near spawning on entering tidal estuaries from the sea but that actual spawning probably does not occur until fresh water is reached. Many larvae of the striped bass were taken by Leim (1924) during the summers of 1922 and 1923 in the Shubenacadie River, Nova Scotia. The young were taken in plankton near the head of the tidal zone. SPAWNING SEASON The spawning season of the striped bass has generally been recognized to occur in the spring and early summer months. Ripe fish have been noted in the rivers of New Brunswick about the middle of June ; in Delaware and New York Bays about the middle of May; in the Roanoke River during May; in the Alabama River in April; and in upper San Francisco Bay principally during May.6 Ripe fish were taken by Worth (1884 b) at Weldon, N. C., from April 19 to May 17, 1883, at a water temperature rising from 58° to 71° F. This observer also recorded the spawning period at Weldon during 1904, as extending from May 2 to May 24. During hatchery operations at Weldon, in 1931, the first ripe female fish was secured on May 5 and the last on May 21. Water temperatures at the hatchery, supplied by filtered and underground piped city water from the Roanoke River, gradually increased from 61° to 71° F. During hatchery operations at the same point, in 1937, ripe females were taken from about May 7 to May 22. The eggs of the striped bass were taken in river plankton in the lower Susque- hanna River, Md., from May 16 to June 8, 1931. These eggs were secured during the early half of the night and were probably only a few hours spawned. Water temper- atures in the river increased from 60° to 70° F. during the period of egg collection. SIZE AND AGE AT MATURITY The weights of 19 female striped bass, taken and stripped for eggs at Weldon, N. C., were recorded by Worth (1904). Three females ranged between 3 and 7 pounds, seven from 10 to 18 pounds, four from 23 to 35 pounds, and five from 40 to 70 pounds. The approximate lengths of these ripe fish would have ranged from 20 in. (50.8 cm) to over 4 ft., according to a length-weight correlation given by Scofield (1932). It appears, on the basis of considerable data collected by Scofield (1931) for California striped bass, that 35 percent of the female fish mature and spawn by their fourth year at an average length of 50 cm (19.7 in.), 87 percent by their fifth year at an average length of 54 cm (21.2 in.), 98 percent by their sixth year at an average length of 61 cm (24 in.), and 100 percent by their seventh year. It was observed that many male striped bass mature and spawn in their third year while all are mature by their fifth year. Ripe male striped bass, 12-18 in. (30.5-45.7 cm) in length, were taken in the Potomac River, Md., late in April 1875, by Milner (1876). Length measurements were obtained of 70 mature male and 29 egg-bearing female striped bass taken by commercial fishing gear near the entrance to the Susquehanna River, Chesapeake Bay, during April and May 1932. The lengths of the male fish ranged from 33-78 cm (13-30.7 in.) with an average length of 40-45 cm (15.7-17.7 in.). Most males were approximately 3 years old. The lengths of the female fish « Adams (1873); Mease (1815); Holton (1874); Bean (1884); Smith (1895); and Scofield (1931). LIFE HISTORY OF THE STRIPED BASS 831 ranged from 50-78 cm (19.7-30.7 in.). No females under 4 years of age were obtained with eggs. The smallest fish taken in these collections probably represent the mini- mum size of spawning fish in both sexes. The largest fish taken do not represent the maximum size which the species attains because the samples were limited to 15 pounds, about 32 inches in length, by legal-size restrictions. The length-frequency distri- butions for these striped bass are given in figure 2. There appears to be many more mature male than female fish on the spawning grounds and the average size of the males is much smaller than that of the females. Both numerical superiority and smaller size of the males may be due to their earlier age at maturity. It was observed by Worth (1903), at Weldon, that where the female fish are in spawning condition the males gather around them in great numbers and there will be 1 large female, weighing from 5-50 pounds, surrounded by 20-50 small males weighing not more than 2 pounds each. A somewhat similar predominance of small males was also noted at Weldon by the writer in May 1937. EGGS AND YOUNG The number of eggs spawned by the striped bass was calculated by Worth (1904), who found a total of 14,000 eggs in a 3-pound fish and 3,220,000 eggs in a 50-pound fish. The Manual of Fish Culture (1900) estimated 1,280,000 eggs from a 12-pound striped bass taken in the Susquehanna River in 1897. This estimate is closely approx- imated by volumetric measurement of the eggs taken from a 1 3-pound fish, measuring 70 cm (27.5 in.) in length, on May 14, 1932, at Havre de Grace, Md. The count totaled 1,337,000 eggs. No complete description of the eggs and young of the striped bass has been avail- able, despite frequent artificial propagation of the species. Various writers have offered partial descriptions of the eggs and fry, however, based on fish-cultural oper- ations.7 A series of eggs and larvae of the striped bass was obtained during May 1931, at Weldon, N. C., through the artificial fertilization and hatching of the eggs at this point on the Roanoke River by the Bureau of Fisheries and the State of North Caro- lina. Samples of eggs and larvae were preserved in a weak formalin solution at 12 hour intervals after the fertilization of the eggs. The eggs were stripped from a ripe female at the fishing grounds, fertilized by the usual dry-pan method, and placed in McDonald hatching jars supplied with filtered river water within 30 minutes after fertilization. The eggs were taken and fertilized about 10 p. m. on May 5 and hatched in approx- imately 48 hours at a water temperature averaging 64.2° F. during the incubation period. No effort was made to rear the larvae through the introduction of food and con- sequently all young fish had perished by 312 hours after fertilization of the eggs. Successful attempts were made during May and June 1937, at Weldon and Edenton, N. C., to rear the larval striped bass in aquaria and outdoor ponds through the intro- duction of natural foods such as Daphnia. These rearing experiments provided addi- tional specimens of larval and post-larval fish.8 The eggs of the striped bass immediately after fertilization are spherical, nonad- hesive, and measure 1.28-1.36 mm in diameter after preservation. The eggs are 7 These writers include Ferguson and Hugblett (1880), Worth (1882) (1883), Ryder (1885), Scofield and Coleman (1907), Bigelow and Welsh (1924), and Scofield (1931). s The author expresses appreciation to W. C. Bunch for the time and care spent in the preservation of the series of eggs and larvae and Louella E. Cable for the painstaking and accurate drawings contained in this report. 73430—38 2 832 BULLETIN OF THE BUREAU OF FISHERIES slightly heavier than fresh water and sink to the bottom of unagitated water. How- ever, a slight movement of the water serves to float the eggs and keep them in suspen- sion. The egg membrane, or chorion, appears heavily corrugated and nearly opaque after preservation. Living eggs show a transparent chorion at all times. The yolk sphere is heavily granulated, about 1.10 mm in diameter, of a rather intense green color in live eggs, and usually of a pale amber color in preserved eggs. It contains an amber-colored oil globule that measures 0.56 mm in diameter. Several much smaller oil globules may also be present. (See fig. 3.) The egg at 15 minutes after fertilization increases by the rapid absorption of water to about 1.84 mm in diameter. The size of the yolk sphere and the oil globule remains the same during the early developmental stages. The blastodisc appears differentiated at one end. of the yolk sphere. The chorion, becoming stretched, is less corrugated. It was noted by Scofield and Coleman (1907) that the first cleavage of the germinal disc takes place about 2 hours after fertilization. (See fig. 4.) The egg at 12 hours after fertilization shows a considerable increase in size. The egg diameter may range from 3. 2-3. 8 mm and water absorption appears complete. The chorion is thin, transparent, and fragile. The blastoderm is in late cleavage and the periblast appears clearly differentiated about the yolk sphere and becomes a paler green with age. (See fig. 5.) The egg at 24 hours after fertilization shows no further expansion of the chorion. The embryo becomes differentiated and extends about half way around the circumfer- ence of the yolk. A moderately intense pigmentation of the embryo occurs and consists of small black dots distributed over the dorsal aspect of the body and over a part of the adjacent blastoderm. (See fig. 6.) The egg at 36 hours after fertilization has an embryo about 1.6 mm in length. Eyes become differentiated but lack pigmentation. The posterior part of the embryo body is free from the yolksac. (See fig. 7.) The egg at 48 hours after fertilization, kept at a temperature averaging 64.2° F., is about to hatch. The embryo is approximately 2.5 mm in length upon leaving the egg. (See fig. 8.) The larva at 60 hours after fertilization of the egg measures about 3.2 mm in length. The oil globule, embedded in the anterior end of the yolksac, projects beyond the head of the larva. The newly hatched fish tends to settle to the bottom of a still aquarium despite swimming efforts to remain near the surface. A strong current of water, however, enables the fish to keep suspended and in more or less constant motion. Hatchery fish are usually liberated soon after this stage. (See fig. 9.) The larval fish at 84 hours after fertilization of the egg increases to about 4.4 mm in length. The head projects beyond the oil globule. A series of small chroma- tophores appears along the ventral surface of the body posterior to the vent but the eyes continue to lack pigmentation. (See fig. 10.) The larva at 120 hours after fertilization of the egg measures 5.2 mm in length. The eyes now possess pigmentation and the jaws are somewhat developed. The oil globule and yolksac are considerably reduced as the rudiments of the digestive tract appear. The pectoral fins become differentiated. The ventral chromatophores become somewhat stronger and several of them now lie along the edges of the gut. (See fig. 11.) LIFE HISTORY OF THE STRIPED BASS 833 3 4 5 6 7 Figure 3. — Striped bass egg at fertilization; diameter 1.3 millimeters. Figure 4.— Striped bass egg 15 minutes after fertilization; diameter 1.8 millimeters. Figure 5. — Striped bass egg 12 hours after fertilization; diameter 3.7 millimeters. Figure 6. — Striped bass egg 24 hours after fertilization. Figure 7. — Striped bass egg 36 hours after fertilization. 834 BULLETIN OF THE BUREAU OF FISHERIES Figure 8.— Striped bass egg 48 hours after fertilization. Figure 9.— Striped bass larva GO hours after fertilization of egg; length 3.2 millimeters. Figure 10.— Striped bass larva 84 hours after fertilization of egg; length 4.4 millimeters. I-::/ LIFE HISTORY OF THE STRIPED BASS 835 13 Figure 12. — Striped bass larva 144 hours after fertilization of egg; length 5.8 millimeters. Figure 13. — Striped bass larva 192 hours after fertilization of egg; length 6 millimeters. Figure 14.— Striped bass larva 288 hours after fertilization of egg; length 6 millimeters; no food available to fish. 836 BULLETIN OF THE BUREAU OF FISHERIES The larval striped bass at 144 hours after fertilization of the egg reaches about 5.8 mm in length. The mouth parts and digestive tract become better developed preparatory to feeding. A series of small chromatophores now extends along the ventral edge of the entire yolksac. (See fig. 12.) The larva at 192 hours after fertilization of the egg has the oil globule and yolksac nearly absorbed. The length of the fish increases only slightly, to about 6 mm. Pigmentation on the ventral surface of the body becomes stronger. (See fig. 13.) Figure 15.— Striped bass young 240 hours after fertilization of egg; length 9 millimeters; food available to fish. The larva at about 288 hours after fertilization of the egg, about 6 mm in length, commences to die rapidly in an aquarium not supplied with food. No fins have developed on the fish except the pectorals. The finfold still extends from the region of the head around the body to the abdomen, becoming interrupted at the vent. A more or less continuous line of pigmentation extends along the ventral portion of the body from the opercle to a point about midway between the vent and the tail. A large chromatophore lies on the upper surface of the swim bladder. (See fig. 14.) Figure 16. — Striped bass young 18 days old; length 13 millimeters; reared in aquarium at Edenton, N. C., May 1937. The young striped bass reach a post-larval stage at 240 hours after fertilization of the egg provided food is made available. At 9 mm in length the postlarva has lost most of the larval finfold. The second dorsal and anal finrays become slightly differ- entiated although the first dorsal and ventral fins are wanting. Large chromato- phores are scattered profusely on the top of the head and a series of branching chro- matophores run along the ventral edge of the body from the head to the tail, becoming interrupted along the intestine and the vent. A regular but broken line of pigmenta- tion extends medially along the side from the pectoral fin to the base of the tail. Mouth parts are well developed. (See fig. 15.) The postlarva at 13 mm (one-half inch) in length, 18 days after fertilization of the egg, has the dorsal and anal finrays well differentiated. The spinous and soft dorsal fins are still connected by the finfold and the spines are still quite rudimentary. The LIFE HISTORY OF THE STRIPED BASS 837 body has become much more robust and all traces of the larval finfold are gone. The ventral fins are now present. Pigmentation is heavy and consists of a medial line of chromatophores that extend along the side from the pectoral fin to the tail, a large number of heavy chromatophores on the head, and various scattered markings on the body, expecially in the ventral region. (See fig. 16.) The young striped bass at a length of 36 mm (1.4 in.) and from 3 to 4 weeks old has the general shape of the adult fish, is well scaled, and has fully developed fins and Figure 17.— Striped bass young, 36 millimeters (1.4 inches) in length. Taken August 28, 1929, at Back River, Va. rays. Pigmentation consists of minute black dots scattered over the entire body. Larger chromatophores are present on the top of the head. A series of about nine oblique V -shaped lines appear along the lateral line of the fish and probably represent blood vessels. (See fig. 17.) The young fish at a length of 130 mm (5.1 in.) and approximately 1 year old pos- sesses the characteristic lateral black stripes ranging from six to eight in number and extending from the edge of the opercle to the base of the tail. There appears also about Figure 18.— Striped bass young, 130 millimeters (5 inches) in length. Taken June 1, 1932, at Sassafras River, Md. seven fainter vertical bars extending from the base of the dorsal fins to somewhat below the lateral line. The dorsal and caudal fins are quite heavily marked with fine dots. (See fig. 18.) GROWTH A thorough study of the growth of the striped bass in California was made by Scofield (1931). This investigator found that on the basis of length -frequency dis- tributions the average length of the fish at the end of the first year of life (April) 838 BULLETIN OF THE BUREAU OF FISHERIES was approximately 10 cm (4 in.), at the end of the second year 25 cm (9.8 in.), at the end of the third year 34 cm (13.4 in.), and at the end of the fourth year about 47 cm (18.5 in.). It was found impossible to determine the age or growth of the species beyond the fourth year by the length-frequency groupings. Calculations of growth for the first 4 years by scale examinations of winter annuli were approximately the same as indicated by the length-frequency distributions. The scales revealed further that at the end of the fifth year the average length of the female striped bass is 54.2 cm (21.3 in.) and of male fish 51.6 cm (20.4 in.), at the end of the sixth year 61.3 cm (24 in.) for female fish and 56.3 cm (22 in.) for males, and at the end of the seventh year 68 cm (26.8 in.) for female fish and 61.2 cm (24 in.) for males. It was noted that both sexes grow at about the same rate during the first year. From then until the fourth Figure 19.— Striped bass adult, 21.4 inches (54.7 centimeters) in length. Taken in April 1880, at Washington, D. C. year the males are larger, but beyond this point the females continue their rapid growth while the males show a retarded growth. At the end of the tenth year the males are about 7 cm (2.7 in.) shorter than the females. Male striped bass older than 10 years were found to be rare, as were females beyond 16 years. Various length-frequency distributions of striped bass were secured during the summer months, chiefly in 1931 and 1932, from Chesapeake Bay. Although the num- ber of fish represented are few in most age groups, the annual growth for the first 3 years of life appears approximately the same as for the species resident in California waters. The O age group, or fish in their first year (spawned about May), attain a length of about 4 cm (1.6 in.) by July, and about 9 cm (3.5 in.) by September. The I age group, or 1-year-old fish, attain a length of 20-27 cm (7.9-10.8 in.) with a mode at 25 cm, by August. The II age group, or 2-year-old fish, reach a length of 26-38 cm (10.4-15 in.), with an average length of 31 cm, by July. The III age group, or 3-year- old fish, may reach a length of 34-50 cm (13.5-19.7 in.), with an average length of 40-43 cm by May, or at the approximate third birthday (see fig. 20). Selectivity of the fishing gear, and the nature of the environment, affected the length-frequency distributions considerably. Likewise, the limited sampling occa- sioned an unknown error in the determination of the average growth rate. The study of the scales verified the age as indicated by the length-frequency distributions. No attempt was made to determine age or growth after the third year, as material was inadequate, LIFE HISTORY OF THE STRIPED BASS 839 The early growth of striped bass appears quite rapid, for larvae hatched at Weldon, N. C., on May 14-16, 1937, and planted in a pond at Edenton, N. C., several days after hatching, attained a total length of 30-33 mm (1% in.) by June 10, less than a month after hatching. The maximum growth of the striped bass is indicated by the capture of several fish at Edenton in April 1891, weighing about 125 pounds each. It is of interest to note that Worth (1882) recorded a seine catch of striped bass at Avoca, N. C., on May 6, 1876, composed of 840 fish, totaling over 35,000 pounds. Three hundred and fifty fish are said to have averaged 65 pounds each. A female striped bass kept on exhibition at the New York Aquarium for over 19 years reached only 20 pounds. It is assumed that this fish did not reach its full CMS. O 10 ao 30 40 30 INS. 0 3B 7.9 11.8 15.7 19.7 Figure 20.— Length-frequency distributions of striped bass taken principally during summers of 1931 and 1932 in Chesapeake Bay. The smaller distribution of O age-group taken in early July; larger distribution iu late August. The distribution of 1 age-group taken during July and August. The smaller distribution of II age-group taken in July; larger distribution in August. The III age-group taken from April to June and composed of mature male fish. development in captivity where variety of food and freedom of movement were re- stricted. Length-weight and length-age correlations for California fish (Scofield, 1932) are given in figure 21. FOOD HABITS The striped bass is carnivorous, predacious, and an active feeder. The species is known to consume all kinds of fishes and crustaceans. Shad, river herring, and men- haden are favorite prey in fresh and brackish waters, while crabs and lobsters are eaten along rocky coast lines. Shrimps, squids, clams, and other crustaceans have been noted in the stomachs of striped bass taken along the Atlantic seaboard. Young striped bass reared in aquaria were fed live Daphnia. Young fish taken in the Hudson River were found to feed largely on the shrimp, Gammarus (Curran, 1937). 840 BULLETIN OF THE BUREAU OF FISHERIES Investigators on the Pacific coast have found that the species feed on every marine form common to the San Francisco Bay region.9 The food included the Pacific herring, silver salmon, steelhead trout, shad, carp, and perch; such crustaceans as Gammarus and Neomysis, and even Velella, the Portuguese man-of-war. It has 0 4 8 12 16 20 24 28 32 36 40 4 4 48 52 WEIGHT" IN POUNDS Figure 21.— Length-weight and length-age correlation of striped bass in California. After Scofield (1832). been observed that striped bass feed heaviest in the warm months of the year and in salt water. Nearly every type of angling bait can be successfully used to hook striped bass. Eel tails, and silvery trolling lures resembling fishes, are particularly effective. MOVEMENTS The movements of striped bass can be broadly classed as coastal, seasonal, and spawning. The exact nature and intensity of these movements are probably deter- mined largely by the character of the environment. That coastal movements occur is clearly indicated by the geographic range of the species along the Pacific and Atlantic coasts of the United States. Along the Pacific coast the striped bass has spread from the initial stocking in San Francisco Bay over a coastal range of about 850 miles. Self-sustaining colonies of striped bass are known to exist in San Francisco Bay and « Various writers on the food of striped bass are Ayres (1842), Verrill (1873), Rice (1883), Scofield (1931), Shapovalov (1936), and Merriman (1937). LIFE HISTORY OF THE STRIPED BASS 841 its tributaries, and in Coos Bay, Oreg., about 400 miles north of San Francisco. Marking experiments in California waters have indicated, however, that no regular or definite coastal movement of striped bass occurs, and that the fish appear to diffuse at random to all points from the locality of release. In a marking experiment in San Francisco Bay, Clark (1936) found that the time elapsing between release and recap- ture ranged from 4 to 477 days with an average of 111 days of freedom. Yet the distances traveled by the marked fish varied from 0 to 46 miles. Such a restricted dispersion indicates limited coastal movement. Along the Atlantic seaboard Merriman (1937) has recently shown that seasonal coastal movements of striped bass occur in southern New England with an apparent incursion of fish from southern waters in early summer and a return movement to the south in late fall. Local seasonal movements of striped bass are quite pronounced. In November and December, as noted by Mease (1815), the fish leave the sea and run into the rivers along the New Jersey coast to pass the winter, where they remain, unless disturbed, until the following spring. In colonial times a winter fishery for striped bass along the North Atlantic coast was possible because the fish moved into the deep river channels during cold weather and lay semidormant near the bottom, from whence they could be easily captured by large dip nets operated under the river ice. In the tidal Parker River, Mass., the fishery now depends entirely on the formation of firm river ice. It is believed by fishermen that the ebb-tide movement of the river water also tends to force the striped bass off the shallow tidal flats into the deeper channel holes where dip nets can be operated to best advantage. In Chesapeake Bay the striped bass are known to winter in the deeper channels of the bay and river mouths. A concentration of fish is known to occur in a deep channel near Kent Island where fishermen find it profitable to sink gill nets for the sluggish fish. A movement of bass takes place in the fall of the year in California. The fish come out of the bays, run into sloughs, and for some distances up the rivers. When cold weather sets in the fish leave the flats and seek the depths of the channels and sloughs. In the summer, following spawning, the striped bass leave the rivers and creek and move out into more open areas in the sea or estuaries. This summer movemen of fish appears to be induced by food requirements. As observed by Wood (1634): These (striped bass) are at one time when the Alewives passe up the Rivers to be catched in Rivers, in Lobster time at the Rocks, in Macrill time in the Bayes, at Michelmas in the Seas. In southern waters the species prefers to dwell in fresh or brackish water at all times and relatively few fish are found near ocean inlets or in the open sea. North of Chesapeake Bay a more pronounced movement of bass occurs along the open sea- coast during the summer months The annual summer appearance of large striped bass along the sandy beaches of New Jersey and off the rocky headlands of southern New England has provided angling sport for many years. In California during the summer striped bass from the region of San Francisco Bay move down along the coast of southern California and from upper Suisan Bay down into San Francisco Bay after spawning. These movements, according to Scofield (1931), appear induced by the more abundant food supply in the salt water than in the fresh water of the delta country. Spawning movements of the species consist essentially of the migration of adult fish from salt, brackish, or fresh waters, up suitable rivers where spawning occurs. 842 BULLETIN OF THE BUREAU OF FISHERIES A spawning movement of striped bass was definitely noted in May 1932, at the entrance to the Susquehanna River where captures of ripe fish indicated a nocturnal spawning migration up the river from Chesapeake Bay. On the upper Roanoke River, at Weldon, N. C., a pronounced spawi ing move- ment occurs during the latter part of April and throughout May. This movement provides an opportunity for fishermen to enjoy the sport of capturing the spawning fisii with large skim nets. This fishing operation is carried on during the early evening in and just below the rapids. In tbe region of San Francisco Bay the spawning migration is made up of fish that come from the deeper holes in the lower rivers and bays, and from the ocean, to run up the Sacramento and San Joaquin Rivers and some of the smaller tributaries. It was thought by Scofield (1931) that a spawning movement of fish also occurred Figure 22.— Length-frequency distribution of striped bass marked for migration studies in Chesapeake Bay. Solid line indicates number of fish marked and released; dotted line indicates number of fish recaptured. from the coast of southern California back to common spawning grounds in the San Francisco Bay area. An early suggestion of a spawning migration of striped bass involving a parent- stream theory, and the feasibility of stocking depleted streams with fish, was advanced by Belknap in 1792. This historian wrote: It is said by some, that fish which are spawned in rivers, and descend to the sea, return to those rivers, only where they are spawned. If this principle be true, the breed might be renewed by bring- ing some of the bass, which are caught in the Merrimac River, alive, over the land, to the nearest part of the waters of the Piscataqua, a distance of not more than 12 miles. This must be done before the spawning season, and might easily be accomplished. The first attempt to determine the migrations of the striped bass through marking experiments was made by the author in July and August 1931, in upper Chesapeake Bay (Pearson, 1933 a). A total of 305 fish, ranging in length from 26-40 cm (10.2-15.7 in.) were caught, marked, and released. Most fish were in their third summer (2 years of age) and were immature so far as could be determined. The fish were taken on hook and line and were released immediately at or near the place of capture, about 1 mile east of Hacketts Point, off Annapolis, Md. (See fig. 22.) The tags were the modified Nesbit disk consisting of two 12-mm (one-half inch) circular celluloid disks connected by a length of nickel wire. The wire, sharpened at one end and headed at the other, was run through one disk and then through the back of the fish slightly below the second dorsal fin and another disk was placed against LIFE HISTORY OF THE STRIPED BASS 843 the other side of the fish and secured by twisting the end of the wire. One disk was red and bore a serial number to identify the individual fish; the other was white and bore the words: “Bureau of Fisheries, Washington, D. C. Return Both Disks. Reward.” A nominal fee was paid for the return of the disks together with informa- tion as to date and place of capture. Soon after marking operations were commenced, on July 7, 1931, disks were returned from various localities in upper Chesapeake Bay and they continued to be returned over a 2-year period. From July 1931, to September 1933, a total of 89 marked fish were recaptured either by sportsmen or commercial fishermen. The JULY SEPT NOV. JAN. MAR. MAY JULY SEPT. NOV. JAN. MAR. MAY JULY SEPT. AUG. OCT. DEC. FEB. APR. JUNE AUG. OCT. DEC. FEB. APR. JUNE AUG. OCT. 1931 1932 1933 Figure 23.— Bimonthly recapture of marked striped bass in upper Chesapeake Bay from July 1931 to September 1933. recaptured fish totaled 29.1 percent of the number released, or about 1 out of every 3 fish marked. Twenty percent of these fish were retaken within the first 6 months after release. (See fig. 23.) None of the marked striped bass were recaptured at the immediate point of release. Only 9 fish out of 89 were recaptured south of the point of release off Annapolis, Md. The majority of fish were taken at various points along the shores of upper Chesapeake Bay from the Magothy River and Love Point north to the Susquehanna and Elk Rivers. The point of greatest concentration of marked fish was in the vicinity of Rock Hall near the entrance to the Chester River. (See fig. 24.) Six out of nine marked fish taken in Chesapeake Bay, or tributaries below the point of release off Annapolis, were recaptured the following spring after marking. One striped bass was recaptured off Maryland Point in the Potomac River on March 17, 1932, while another was secured in the Wicomico River, near Salisbury, on March 23, 1932. These localities were the most distant points to which the marked fish dispersed over a 2-year period. The steady decrease in the number of recaptured fish after the first 2 months (see fig. 23) was probably caused by the ultimate detachment of the disks from the back of the fish and by the continually reduced number of marked fish available for capture. 844 BULLETIN OF THE BUREAU OF FISHERIES Figure 24. — Movements of striped bass in upper Chesapeake Bay, 1931-33. The solid square off Annapolis indicates point of release; dots represent individual recaptures of fish. LIFE HISTORY OF THE STRIP! D BASS 845 It is obvious that a clear-cut movement of fish occurred to the north of the point of release which indicates that the bass preferred the fresh or slightly brackish water at the head of the bay to the more saline water down the bay. South of the point of release the four most distant recaptures were at considerable distances up rivers in brackish water. Although the absence of marked fish south of the Potomac River might indicate a local stock of striped bass in upper Chesapeake Bay, a recent increase in the stock of fish within the entire bay, together with a simultaneous increase in the number of fish annually visiting southern New England waters, suggests that the limited distributions of the marked fish during 1931-33 were perhaps caused by a low population density of striped bass in the upper bay and by an abundance of food for the fish with little incentive for widespread movements in or out of the bay. FISHERY The fishery methods employed to capture the striped bass afford ample evidence of the severity of the struggle that this food and game fish has undergone in order to survive. These methods are applicable in most instances to other anadromous fishes, such as the shad and salmon, which have suffered alarming decreases in abun- dance along our Atlantic coast. The early settlers in New England laid efficient traps for the striped bass during the summer months as they used to “tide it in and out to the Rivers and Creekes by stretching long seins and weirs across coastal streams at high tide.” As the water ebbed from the creeks the stranded fish were often obtained in far greater quantities than the fishermen could haul to land. The fish were consumed either fresh, salted, pickled, or smoked. Pickled bass furnished a medium of trade in the West Indies along with salted codfish. The earliest colonial records of the smoking of striped bass as a means of preservation contain the following statement of Wood (1634): They drie them to keepe for Winter, erecting scaffolds in the hot sunshine, making fires likewise underneath them, by whose smoake the dies are expelled till the substance remaine hard and drie. In this manner they dry Basse and other fishes without salt, cutting them very thin to dry suddenly, before the flies spoyle them, or the raine moist them having speciall care to hang them in their smoaky houses, in the night or dankish weather. In the St. Johns River, New Brunswick, according to Adams (1873), the Indians captured the bass at spawning time. A few canoes would drop down the river, each with an Indian in the bow, spear in hand, and another in the stern paddling gently. A sudden spash close by would indicate a bass and like an arrow the birchbark skiff was shot toward the spot while the man in front, resting on his knees, with much force and dexterity sent his three-pronged harpoon into the fish. The winter months proved the most destructive to the striped bass in northern waters. The fish normally sought the shelter of river channels during cold weather, lying more or less dormant along the bottom until spring. Fishermen soon learned to capture them under the ice by means of large dip nets (Pearson, 1935 b). The havoc of this type of fishery on the resident stock of striped bass was noted by various early writers.10 Various methods have been developed to capture the striped bass in southern rivers. It has been the practice for many years on the Roanoke River near Weldon, N. C., to secure the spawning fish in and below the rapids each spring. The adult fish move up the river in late April and May and, if there is sufficient water in the 10 Tenney (1795), Mease (1815), and Perley (1850). 846 BULLETIN OF THE BUREAU OF FISHERIES river, they distribute themselves about the falls where the strong current renders them inaccessible to fishermen. The fish work upstream into numerous channels between various islands lying amid the rapids. During summer low water fishermen at one time prepared traps to capture the fish in these channels by constructing wooden slides at favorable points in the rapids. The fish, forced to descend the rapids through lowered river level, were guided onto the slides and were forced to remain against slats by the pressure of the current and could be easily removed by the fishermen. As many as 300 fish of 30 pounds each have been removed from a slide in a single day. This efficient fishing device has been recently outlawed by the State of North Carolina. (See fig. 25.) The striped bass congregate at the foot of the rapids at Weldon and are taken in large quantities during the spawning season by skim nets. A skim net consists of a large bow frame of hickory, about 6 feet long and 4 feet wide, to which is hung a linen net about 6 feet deep and 1%-inch square mesh. The bow frame is fastened to a stout wooden pole at least 20 feet long. Two such nets may be fished from a small power boat simultaneously but a man must sit in the stem of the boat and keep it broadside to the river current as it drifts downstream. A fisherman usually stands amidship holding the net in a rigid vertical position against the gunwale with the bow frame lifted a few inches from the river bottom. The touch of a fish against the net signals the fisherman who quickly lifts the net vertically out of the water and deposits the fish in the boat. The catch consists chiefly of ripe fish from which eggs and milt are taken for artificial propagation. Most commercial fishery methods for the capture of striped bass are confined, through legal restrictions, to more open areas than narrow river channels and rapids. Pound nets, haul seines, and gill nets effectively take the fish from Rhode Island to North Carolina. Salt-water areas provide the most abundant catches in northern waters while brackish and fresh-water estuaries and rivers afford the best fishing from Chesapeake Bay south. Sunken gill nets are used in winter and drift gill nets in summer. Pound nets or trap nets are most advantageous along the open coast line and off river mouths. Haul seines are favored in large estuaries and purse seines, now outlawed, were formerly employed to capture schooling striped bass in Chesa- peake Bay. No commercial fishery for the striped bass now exists in California; the species is reserved for hook-and-line sportsmen. Commercial catch statistics for striped bass from the waters of Maryland, Virginia, North Carolina, and the Middle Atlantic States are given in figure 26. The annual catch records show a decreasing supply of fish despite the more efficient gear employed. The catch in Maryland decreased from a peak of 1,413,000 pounds in 1925 to 314,000 pounds in 1933. The striped bass in various Middle Atlantic States has provided an annual catch of less than 207,000 pounds (40,000 pounds in 1933) since the early part of the present century although this area in 1889 produced over a million pounds of the fish. North Carolina, relatively free from coastal river obstructions and wide- spread industrial water pollution, shows a decrease in the catch from 1,175,000 pounds in 1902 to 362,000 pounds in 1934. Virginia, however, shows a steady catch at about one-half million pounds annually. The intensity of the fishery for striped bass in upper Chesapeake Bay may be estimated by the high return of released fish in a marking experiment conducted in 1931. A total of 29 percent of the 305 released fish were recaptured by fishermen within 2 years, and about 20 percent of these fish were retaken within 6 months after Figure 25.— Rockfish slide in a channel of the Roanoke River at Weldon, N. C. The striped bass spawn in these channels characterized by swift currents. LIFE HISTORY OF THE STRIPED BASS 847 release. The high rate of recapture is indicative of a severe strain on the local stock of fish. It is surprising to note that after an extended period of lean years the catch of striped bass in Maryland waters increased from 332,000 pounds in 1934 to 928,000 pounds in 1935. This increase of nearly threefold cannot be definitely explained in the absence of field observations but a likely cause for the greater abundance of fish is suggested. In 1932 the use of the purse seine was forbidden in Maryland. This type of net had accounted for about 25 percent of the annual catch for several years prior to 1931. Although the catch remained low from 1932 to 1934, it is significant that the striped bass do not generally attain commercial size until their third summer. Hence, fish which were spawned in 1933 did not appear in the catch until 1935. It might be assumed that enough adult striped bass 3 years old or older were spared by the abolition of the purse-seine fishery in 1932 to aid greatly in spawning production Figure 26. — Commercial catch of striped bass in the Middle Atlantic States, Maryland, Virginia, and North Carolina, compiled by Bureau of Fisheries for various years since 1887. in the spring of 1933. Many fish spawned in 1933 undoubtedly reached the com- mercial catch during 1935. If such a condition actually occurred then a heavy production of young also occurred in 1934, making possible a large commercial catch in 1936. Field reports again indicate that the striped bass was as abundant in 1936 as in 1935, and that most catches were composed of small fish. Another indication of the recent increase in the stock of striped bass along the Atlantic seaboard is shown by the incursion of many 2-year-old fish to the coast of southern New England in 1936 as noted by Merriman (1937). This movement of fish into southern New England is perhaps definitely correlated with the increase of striped bass shown by the commercial catch in 1935 in upper Chesapeake Bay. A movement of fish out of Chesapeake Bay and into northern areas may therefore occur at times when the local stock of fish becomes so abundant as to seriously reduce its food supply. Whether depleted northern waters will be permanently restocked as a result of this recent influx of striped bass from apparently overstocked southern areas is unknown. 848 BULLETIN OF THE BUREAU OF FISHERIES The striped bass has shown, by its remarkable reproduction in California and its recent increase in Chesapeake Bay, that it has the ability to establish itself as an important aquatic resource in favorable environments within a short period of time. However, unless the fishery strain on the stock of fish in most eastern waters is eased appreciably by adequately restricted fishing, it is feared that only the past glories of this superb food and game fish will remain for future generations to contemplate. Nevertheless, with vigorous and wTell-considered conservation measures adopted, the striped bass can be expected to increase to some degree of its former abundance and assure the future of the “boldest, bravest, strongest, and most active fish that occurs the year round in our American coastal waters.” SUMMARY The striped bass, or rockfish, occurs over an extended coastal range along the Atlantic and Pacific coasts of the United States. This fish is also found in small numbers in streams tributary to the Gulf of Mexico from St. Marks, Fla., to Lake Pontchartrain, La. The species is taken most abundantly at the present time in the fresh and brackish waters of Chesapeake Bay, Albemarle Sound, and San Francisco Bay. The spawning grounds are located in coastal rivers, apparently characterized by strong rapids and rock-strewn bottoms, and the spawning season extends from late April to early June in most areas. Sexual maturity, accompanied by spawning, is attained by most male fish at the end of the third year and at a minimum length of about 10 inches. Female fish mature at the end of the fourth year and at a minimum length of about 19 inches. The eggs of the striped bass are semibuoyant, spherical, and measure about 1.3 millimeters in diameter at fertilization, increasing to about 3.5 millimeters within 12 hours. The eggs hatch in 48 hours at about 65° and in about 36 hours at 71° F. The yolk is absorbed and the young begin feeding by 240 hours after fertilization of the egg. The average length of the striped bass is 4 inches at the end of the first year; 10 inches at the end of the second year; 15 inches at the end of the third year; and 18.5 inches at the end of the fourth year. The food of the species is largely fishes and crustaceans. The striped bass show coastal, seasonal, and spawning movements. Coastal movements are widespread but are probably regulated by the population density of fish in natural centers of abundance. Seasonal movements consist of a summer movement of fish into more open water with better feeding grounds and a winter movement into deep river channels for a semihibernation period. Spawning migration occurs in the spring of the year when the striped bass move up favorable rivers from the sea or estuarine areas. A marking experiment in upper Chesapeake Bay in 1931 showed purely local movements within the upper bay over a 2-year period with a 29.1 percent recapture of marked fish. The fishery for striped bass has shown a general decline over the greater part of its range despite a more intensive fishing effort. Restrictive fishing regulations appear to offer suitable means for increasing the stock of fish appreciably. LIFE HISTORY OF THE STRIPED BASS 849 BIBLIOGRAPHY Abbott, Charles Conrad. 1878. Notes on some fishes of the Delaware River. Report, U. S. Commissioner of Fish and Fisheries, 1875-76 (1878), pp. 825-845. Washington. Adams, A. Leith. 1873. Field and forest rambles, with notes and observations on the natural history of eastern Canada, xvi, 333 pp., illus. London. Atkins, Charles G. 1887. The river fisheries of Maine. In the fisheries and fishery industries of the United States, by George Brown Goode and associates, sec. V, vol. I, pp. 673-728, and sec. V, vol. II, Pis. 164—168. Washington. Ayres, William O. 1842. Enumeration of the fishes of Brookhaven, Long Island, with remarks upon the species observed. The Boston Jour, of Nat. Hist., vol. IV., No. 1. Boston. Bean, Tarleton H. 1884. On the occurrence of the striped bass in the lower Mississippi Valley. Proc. U. S. Nat. Mus., vol. 7, 1884 (1885), pp. 242-244. Washington. Bean, Tarleton H. 1903. Catalogue of the fishes of New York. Bull. 60, Zoology 9, N. Y. State Mus., 784 pp. Albany. Bean, Tarleton H. 1905. The striped bass. In the basses, fresh-water and marine, by William C. Harris and Tarleton H. Bean. Edited by Louis Rhead, pp. 167-211. New York. Belknap, Jeremy. 1792. History of New Hampshire, vol. Ill, 480 pp. Boston. Bigelow, Henry B. and William W. Welsh. 1925. Fishes of the Gulf of Maine. Bull. U. S. Bur. Fish., vol. XL, pt. I, 1924 (1925), 567 pp., 278 figs. Washington. Bigelow, Henry B. and William C. Schroeder. 1936. Supplemental notes on the fishes of the Gulf of Maine. Bull. U. S. Bur. Fish., vol. XLVIII, pp. 319-343. Washington. Burns, Frank. 1887. Rockfish in South Carolina. Bull. U. S. Fish Commission, vol. VI, 1886 (1887), pp. 124—125. Washington. Cole, Charles E. 1930. Angling for striped bass. Calif. Fish and Game, vol. 16, No. 4, pp. 286-293, 5 figs. Sacramento. Clark, G. H. 1932. The striped bass supply, past and present. Calif. Fish and Game, vol. 18, No. 4, pp. 297-298, 1 fig. Sacramento. Clark, G. H. 1934. Tagging of striped bass. Calif. Fish and Game, vol. 20, No. 1, pp. 14-19, 1 fig. Sacramento. Clark, G. H. 1936. A second report on striped bass tagging. Calif. Fish and Game. vol. 22, No. 4, pp. 272-2S3. Sacramento. Corson, R. H. 1926. Striped bass ( R . lineatus ), a study of weights and lengths as compared with ages. Bull. N. Y. Zool. Soc., vol. XXIX, No. 6, pp. 210-211, 2 figs. New York. Craig, J. A. 1928. The striped bass supply of California. Calif. Fish and Game., vol. 14, No. 4, pp. 265-272, 3 figs. Sacramento. Craig, J. A. 1930. An analysis of the catch statistics of the striped bass ( Roccus lineatus) fishery of California. Division of Fish and Game. Fish Bull. No. 24, 41 pp., 22 figs. Sacramento. Curran, H. W. 1937. Fisheries investigations in the lower Hudson River. In A Biological Survey of the Lower Hudson Watershed. Biol, survey (1936), No. XI, N. Y. State Conserva- tion Dept. Albany. DeVries, David Pietersz. 1655. Korte historial ende journaels aenteyckeninge van verschey- den voyagiens in de vier deelendes wereldtsronde, also Europa, Africa, Asia, ende Amerika. Hoorn. Transactions by Henry C. Murphy in Vo37ages from Holland to America, 1632-1644, coll. N. Y. Hist. Soc., 2nd ser., vol. Ill, pt. I, pp. 11-129, 1857. New York. Dunn, Horace D. 1889. Fish-culture on the Pacific coast. Bull. U. S. Fish Commission, 1887 (1889), pp. 49-50. Washington. Endicott, Francis. 1883. Striped bass. In Sport with gun and rod, in American woods and waters, edited by Alfred M. Mayer, pp. 449-471. New York. Fearing, D. B. 1903. Some early notes on striped bass. Trans. American Fish Soc., pp. 90-98. Appleton. Ferguson, F. B. and Thomas Hughlett. 1880. Report, Commissioner of Fish, of Maryland, pp. XXIII-XXVI. Annapolis. Goode, George Brown. 1888. American Fishes. A popular treatise upon the game and food fishes of North Carolina, with especial reference to habits and methods of capture, xv, 496 pp., illus. New York. Gowanloch, James Nelson. 1933. Fishes and fishing in Louisiana. La. Dept. Conservation, Bull. 23, 638 pp., illus. New Orleans. 850 BULLETIN OF THE BUREAU OF FISHERIES Herbert, Henry William. 1849. Frank Forester’s fish and fishing in the United States and British provinces of North America, xvi, 455 pp., illus. London. Higginson, Francis. 1630. New England’s plantation, or a short and true description of the commodities and discommodities of that countrey. London. Reprint New England Soc. of City of New York. 1930. New York. Hildebrand, Samuel F., and William C. Schroeder. 1928. Fishes of Chesapeake Bay. Bull. U. S. Bur. Fish., vol. XLIII, pt. I, 1927 (1928), 388 pp., 211 figs. Washington. Holton, Marcellus G. 1874. Letters to Spencer F. Baird. In the progress of fish-culture in the United States by James W. Milner. Report, Commissioner of Fish and Fisheries, pt. II, 1872-1873 (1874), pp. 553-554. Washington. Hubbard, William. 1815. A general history of New England from the discovery to MDCLXXX. 676 pp. Cambridge. Josselyn, John. 1672. An account of two voyages to New-England. London. Reprint by William Veazie, 1865, vii, 211 pp. Boston. LeCompte, E. Lee. 1926. The striped bass (roekfish) of the Atlantic coast. Trans. American Fish. Soc., pp. 203-206. Hartford. Leim, A. H. 1924. The life history of the shad ( Alosa sapidissima, Wilson) with special reference to the factors limiting its abundance. Cont. Can. Biol. N. S., vol. II, No. 11. Toronto. McFarland, Raymond. 1911. A history of the New England fisheries. 457 pp. New York. Mease, James. 1815. Facts respecting the roekfish or streaked bass of the United States. Trans. Lit. and Philos. Soc. N. Y., vol. I, pp. 502-504. New York. Merriman, Daniel. 1937. Notes on the life history of the striped bass ( Roccus lineatus). Copeia, No. 1, pp. 15-36. Ann Arbor. Milner, J. W. 1876. Report of the Triana trip. Report U. S. Commissioner of Fish and Fish- eries, 1873-74, 1874-75 (1876), pp. 351-362. Washington. Mitchell, S. L. 1815. The fisheries of New York, described and arranged. Trans. Lit. and Philos. Soc. N. Y., vol. I, pp. 355-492, pis. I- VI. New York. Morton, Thomas. 1637. New English Canaan or New Canaan; containing an abstract of New England. Amsterdam. Reprinted by Prince Society, 1883, 349 pp. Boston. Mosher, Gideon. 1883. Do striped bass ( Roccus lineatus ) feed on menhaden. Bull. U. S. Fish. Comm., vol. Ill, p. 410. Washington. New York Zoological Society. 1913. A long-lived fish. Bull. N. Y. Zool. Soc., vol. XVI, No. 60, p. 1049. New York. Norny, E. R. 1882. On the propagation of the striped bass. Bull. U. S. Fish Commission, vol. I, 1881 (1882), pp. 67-68. Washington. Pearson, John C. 1933a. Movements of striped bass in Chesapeake Bay. Maryland Fisheries, no. 22, pp. 15-17. Baltimore. Pearson, John C. 1933b. A unique fishery for the striped bass or roekfish in Massachusetts. Maryland Fisheries, No. 24, pp. 16-18. Baltimore. Perley, H. M. 1850. Report on the sea and river fisheries of New Brunswick, within the Gulf of St. Lawrence and Bay of Chaleur. 137 pp. Fredericton. Rice, H. J. 1883. Letter to E. G. Blackford, in a few facts in relation to the food and spawning seasons of fishes on the Atlantic coast by E. G. Blackford. Trans. Amer. Fish Cultural Assoc., pp. 6-8. New York. Roosevelt, Robert B. 1865. Superior fishing, or the striped bass, trout, and black bass of the northern States. 304 pp. New York. Ryder, John A. 1887. On the development of osseous fishes, including marine and fresh-water forms. Report, U. S. Fish Commission, Appendix D, 1885 (1887), pp. 489-604, pis. I-XXX, 2 figs. Washington. Schoepf, Johann David. 1788. Reise durch einige der mittlern und sudlichen vereinigten Nord Americanischer Staaten. 2 vols. Erlangen. Trans, by A. J. Morrison in Travels in the Con- federation, 1783-1784. 1911. Philadelphia. Scofield, Eugene C. 1928. Striped bass studies. Calif. Fish and Game., vol. 14, No. 1, pp. 29-37, 4 figs. Sacramento. Scofield, Eugene C. 1928. Preliminary studies on the California striped bass. Trans. Amer. Fish. Soc., pp. 139-144. Hartford. Scofield, Eugene C. 1931. The striped bass of California ( Roccus lineatus). Calif. Div. Fish and Game, Fish Bull. 29, 84 pp., 47 figs. Sacramento. LIFE HISTORY OF THE STRIPED BASS 851 Scofield, Eugene C. 1932. A simple method of age determination of striped bass. Calif. Fish and Game, vol. 18, No. 2, pp. 1(38-170, 1 fig. Sacramento. Scofield, N. B. and G. A. Coleman. 1910. Notes on spawning and hatching striped bass eggs at Bouldin Island hatchery. Report, Bd. Fish and Game Comm, of Calif., 1907-1908 (Appen- dix to report for 1909-1910), pp. 109-117, 3 pis. Sacramento. Scofield, N. B. 1910. Notes on the striped bass in California. Report, Board of Fish and Game Comm, of Calif., 1907-1908 (Appendix to report for 1909-1910), pp. 104-109. Sacramento. Scofield, N. B. and H. C. Bryant. 1926. The striped bass in California. Calif. Fish and Game, vol. 12, No. 2, pp. 55-74, 4 figs. Sacramento. Shapavalov, Leo. 1936. Food of striped bass. Calif. Fish and Game, vol. 22, No. 4, pp. 261- 271, 2 figs. Sacramento. Smith, Jerome V. C. 1833. Natural history of the fishes of Massachusetts, embracing a practical essay on angling, vii, 400 pp. Boston. Smith, Hugh M. 1896. A review of the history and results of the attempts to acclimatize fish and other water animals in the Pacific States. Bull. U. S. Fish Commission, vol. XV, 1895 (1896), pp. 397-472, pis. 73-83. Washington. Smith, Hugh M. 1907. The fishes of North Carolina. N. C. Geol. and Econ. Survey, vol. II, 1907, xi, 453 pp., 21 pis., 187 figs. Raleigh. Snyder, J. P. 1914. Notes on striped bass. Trans. Amer. Fish. Soc., pp. 93-96. New York. Snyder, J. P. 1918. Report on fish hatcheries, 1918. Official Bull. Conservation Commission of Md., No. 6, pp. 19-20. Baltimore. Snyder, J. P. 1919. Report on fish hatcheries, 1919. Official Bull. Conservation Commission of Md., No. 8, pp. 19-20. Baltimore. Tenney, Samuel. 1795. Topographical description of Exeter in New Hampshire. In Mass. Hist. Soc. Coll., vol. 4. Boston. Throckmorton, S. R. 1882. The introduction of striped bass into California. Bull. U. S. Fish. Commission, vol. I, 1881 (1882), pp. 61-62. Washington. (J. S. Commission of Fish and Fisheries. 1898. A manual of fish-culture. Report, U. S. Com- missioner of Fish and Fisheries, 340 pp., 62 pis. Washington. Verrill, A. E. 1873. Report upon the invertebrate animals of Vineyard Sound and adjacent waters, with an account of the physical characters of the region. Report, U. S. Commissioner of Fish and Fisheries, 1871-72 (1873), pp. 295-778. Washington. Wood, William. 1634. New England’s Prospect. London. Reprinted 1898, 103 pp. Boston. Worth, S. G. 1882. The artificial propagation of the striped bass ( Roccus lineatus ) on Albemarle Sound. Bull. U. S. Fish Commission, vol. I, 1881 (1882), pp. 174r-177. Washington. Worth, S. G. 1883. Letter to Col. M. McDonald. Trans. Amer. Fish-Cultural Assoc., pp. 9-10. New York. Worth, S. G. 1884. The propagation of the striped bass. Trans. Amer. Fish-Cultural Assoc., pp. 209-212. New York. Worth, S. G. 1884. Report upon the propagation of striped bass at Weldon, N. C., in the spring of 1884. Bull. U. S. Fish Commission, vol. IV, pp. 225-230, 1 pi. Washington. Worth, S. G. 1889. The striped bass or rockfish industry of Roanoke Island, North Carolina, and vicinity. Bull. U. S. Fish Commission, vol. VII, 1887 (1889), pp. 193-197. Washington. Worth, S. G. 1903. Striped bass hatching in North Carolina. Trans. Amer. Fish. Soc., pp. 98-102. Appleton. Worth, S. G. 1904. The recent hatching of striped bass and possibilities with other commercial species. Trans. Amer. Fish. Soc., pp. 223-228. Appleton. Worth, S. G. 1910. Progress in hatching striped bass. Trans. Amer. Fish. Soc., pp. 155-159. Washington. Worth, S. G. 1912. Fresh-water angling grounds for the striped bass. Trans. Amer. Fish. Soc., pp. 115-123. Washington. o ; t > ; - ' ■ ' ' GENERAL INDEX & Page Acipenser brevirostris 326 sturio.. 326 Adaptation of the feeding mechanism of the oyster ( Ostrea gigas) to changes in salinity 345-364 Age and growth of the cisco, Leucichthys artedi (Le Sueur), in the lakes of the northeastern highlands, Wisconsin- 211-317 Alaska, southeastern, pink salmon tagging experiments in 643-666 Alepisauris ferox 328 alewife 327 Alopias vulpinas. 322 Alutera scripta 334 Anarrhichas minor 337 anchovy __ 327 striped 328 Ancho viella epsetus 328 mitchilli 327 Anderson, W. W., F. W. Weymouth and Milton J. Lind- ner: Preliminary report on the life history of the common shrimp, Penaeus setiferus (Linn.) 1-26 Anguilla rostrata 326 Argentine 328 Argentina silus 328 Artediellus uneinatus 335 Balistes carolinensis 333 banded blenny 603 barracuda, northern 329 Bass, striped, or rockfish, Roccus saxatilis (Walbaum), life history of 825-851 bass, rock 213 sea 333 striped 333 Bigelow, Henry B. and William C. Schroeder: Supple- mental notes on fishes of the Gulf of Maine 319-343 black bass, largemouth 213 smallmouth 213 Blennies (family Blenniidae) 573-611 Blenny, banded 603 snake 336 blueback 327 salmon 695,754-778,819 bluefish 332 bluegill 213 bonito, common 330 brasiliensis, Penaeus 2, 4, 6, 8, 21 British Columbia, pink salmon marking experiments in.. 36 Brosme brosme 339 burbot 213 butterfish 332 Cable, Louella E. and Samuel F. Hildebrand: Further notes on the development and life history of some tele- osts at Beaufort, N. C 505-642 Cable, Louella E. and Samuel F, Hildebrand: Reproduc- tion and development of whitings or kingfishes, drums, spot, croaker, and weakfishes or sea trouts, family Scia- enidae, of the Atlantic coast of the United States 41-117 Page Caranx crysos . 331 hippos 331 Carcharhinus obscurus 312 Centropristes striatus _ 333 Cestracion Tiburo - 322 zygaena ... 322 Charcharodon charcharias. 322 Chauloidus sloanei 328 Chimaera affinis... 326 Chinook salmon 667, 695, 795-803, 819 Christey, Leroy S. and Frederick A. Davidson: The migra- tions of pink salmon (Oncorhynchus gorbuscha ) in the Clarence and Sumner Straits regions of southeastern Alaska 643-666 chum salmon 667,695,813-815,820 Cisco, Leucichthys artedi (Le Sueur), age and growth of in the lakes of the northeastern highlands of Wisconsin .. 211-317 cisco, lake herring 213 or chub 213 Citharichthys arctifrons _. 341 Clupea harengus 327 pallsii, races of in southeastern Alaska 119-141 cod 338 coho salmon 667, 695, 781-794, 819 common shrimp 1 Conger oceanica 326 constrictus, Trachypenaeus 2, 8, 21 Coryphaena hippurus 332 Cottunculus microps 335 Crago vulgaris 6, 12 Crude oil pollution, effects of on oysters in Louisiana waters 143-210 Cryptacanthodes maculatus 337 cusk 339 Cynoscion nebulosus 91 no thus 91 regalis 91 dabs, rusty 340 Dahlgren, Edwin H. and George A. Rounsefell: Races of herring, Clupea pallasii, in southeastern Alaska 119-141 Davidson, Frederick A. and Samuel T. Hutchinson: The geographic distribution and environmental limitations of the Pacific salmon (genus Oncorhynchus) 667-692 Geographic distribution 668 Native.. 668 Foreign 669 Environmental limitations to occurrence 673 North Pacific region. 673 South Pacific region 680 North Atlantic region 685 South Atlantic region 687 Davidson, Frederick A. and Leroy S. Christey: The mi- grations of pink salmon ( Oncorhynchus gorbuscha ) in the Clarence and Sumner Straits regions of southeastern Alaska 643-666 Channels of migration 644 Tagging methods - 645 177988—40 853 854 GENERAL INDEX Page Davidson, Frederick A. and Leroy S. Christey — Con. Pink-salmon tagging experiments in Clarence Strait and adjacent waters, 1924-32 646 Tagging experiments in the vicinity of Cape Fox. . 648 Tagging experiments in the vicinity of Cape Chacon.. 649 Tagging experiments on Oravina Island and in the vicinity of Kasaan Bay ... 651 Tagging experiments in the vicinity of Cape Muzon. 654 Pink-salmon tagging experiments in Clarence Strait in 1935 and 1936 654 Summary of Cape Chacon experiments 660 Pink-salmon tagging experiments in Sumner Strait, 1924-36 661 Tagging experiments at Ruins Point and Cape Decision 062 Tagging experiments at Point Colpoys 664 Summary of Point Colpoys experiments 665 Davidson, Frederick A.: The homing instinct and age at maturity of the pink salmon (Oncorhynchus gorbuscha) ... 27-39 Marking the pink salmon fry 28 Interpretation of results from marking experiments 29 Homing instinct 29 Age of pink salmon at maturity 34 Pink salmon marking experiments in British Columbia 36 Decapterus macarellus 331 deep-sea angler 342 Detection and measurement of stream pollution . ... 365-437 dog salmon 667,695,813-815,820 dogfish, smooth 321 spiny 322 dolphin, common 332 drum, banded ..... 84 star 75 eelpout 337 eel, American conger 326 Anguilla rostrat a 326 rock 336 snake 326 wolf 338 Ellis, M. M,: Detection and measurement of stream pollu- tion 365-437 Stream pollutants and aquatic environment 366 Physical and chemical characteristics of waters suitable for fresh-water stream fishes 366 General field methods _ 368 Equipment 370 Dissolved oxygen 370 Hydrogen-ion (pH) limits. 379 Ionizable salts 383 Specific conductance 383 Carbon dioxide 386 Iron — 390 Ammonia 391 Suspensoids 394 Depths 396 Bottom conditions as affected by stream pollution. 398 Action of pollutants on fishes 400 Injuries to gills and external structures 400 Pollutants entering the body of the fish and exert- ing true toxic action 402 Lethality of specific substances occurring in stream pollutants 403 General consideration 403 Test animals 404 Water types.. 406 Specific lethality tables 408 Lethal limits of 114 substances which may be found in stream pollutants 417 Enchelyopus chimbrius 339 Etrumeus sadina.. 327 Page filefish.... ._ 334 Fishes of the Gulf of Maine, supplemental notes on 319-343 Fist.ularia tobacaria.. 328 fish: alewife 327 anchovy 327 striped 328 Argentine 328 banded blenny ... 603 barracuda, northern 329 bass, rock 213 sea 333 striped 333,825-851 black bass, largemouth 213 smallmouth 213 blenny, banded 603 snake. 336 blennies (family Blenniidae) 573-611 blueback 327 bluefish. 332 bluegill 213 bonito, common 330 burbot 213 butterfish . . 332 chimaera 326 cisco 211-317 lake herring 213 or chubs 213 cod 338 cusk 339 dab, rusty.. . . 340 deep-sea angler 342 dogfish, smooth ... 321 spiny... 322 dolphin, common 332 drum, banded 84 star 75 eel, American conger 326 Anguilla rostrata 326 rock 336 snake.. 326 wolf - 338 eelpout 337 filefish 334 flounder, four-spotted .... 340 Georges Bank... 341 Gulf Stream 341 winter 340 witch ... 341 gobies (family Gobiidae) 543-573 goosefish, American. . . . 341 grenadier, common... 340 haddock 339 hagfish 321 hake - 612 long-finned 339 silver.. ... 338 spotted 339 hardtail 331 harvestfish 333 herring, lake 213 round 327 thread 327 kingfishes, or whitings 51 lake herring 213 lamprey, sea 321 lancetfish 328 leatherjacket 332 lookdown 332 mackerel — 329 Spanish. 508 marlin, spearflsh 330 needlefish - 328 GENERAL INDEX 855 fish— continued. Page opah 332 pearlsides... 328 perch, yellow 213 pike perch 213 pilotflsh 331 pinfish... 618 pink salmon. 27-39, 643-666, 667, 695, 804-812, 820 pipefish, common 329 pelagic... 329 plaice, American 340 pollock 338 puffer 334 remora 336 rock bass 213 rockling, four-bearded 339 rosefish 334 black-bellied.. 334 rudderfish 331 salmon: blueback 695, 754-778, 819 Chinook. 667,695,795-803,819 coho 667,695,781-794,819 chum 667,695,813-815,820 dog 667,695,813-815,820 humpback 667, 695, 804-812, 820 king 695,795-803,819 Pacific 667-692,695 pink 27-39, 643-666, 667, 695, 804-812, 820 quinnat.. 667,695,754-778,819 red 667, 695, 754-778, 819 silver 667, 695, 781-794, 819 silverside... 667, 695, 781-794, 819 sockeye. 667,695,754-778,819 spring 695,795-803,819 sauger 213 sargassum fish. 341 scad, mackerel 331 big-eyed 331 scul pin, deep-sea 335 hook-eared 335 longhorn 335 mailed 335 sea bream 332 lamprey 321 raven 335 robin, red-winged 336 snail 336 striped. 336 trouts (weakfishes) 91 seaweed fish, spotted 576 shanny. 337 Arctic 337 radiated. 337 shark, dusky 321 great blue 321 greenback 323 hammerhead 322 mackerel. 322 sharp-nosed 322 shovel-head 322 thresher 322 white 322 sheepshead 526 skate, barn -door 325 big 324 brier 325 little 324 prickly 324 smooth 325 sole, American 630 spadeflsh 534 fish— continued. Page Spanish mackerel 508 spearfish, marlin 330 sturgeon, common 326 short-nosod 326 sucker 213 swordfish 330 torpedo 325 triggerfish... 333 trout, bastard 110 gray 102 lake.. 213 speckled 92 spotted 92 trumpetfish 328 tuna 329 unicornfish 334 viperfish 328 weakfishes (sea trouts) 91 gray 102 spotted 92 whitefish 213 whitings, or kingfishes 51 wolffish, spotted 337 wrymouth.. 337 yellow perch 213 gaduscallarias.- 338 Galtsoff, Paul S., Herbert F. Prytherch, Robert O. Smith, and Vera Koehring: Effects of crude oil pollution on oysters in Louisiana waters 143-210 Preliminary field investigations, 1933 146 Survey of oyster bottoms in areas affected by oil-well pollution, 1934 150 Methods 150 General conditions 150 LakeBarre 152 Lake Felicity and Lake Chein 152 Terrebonne Bay 163 Timbalier Bay 153 Lake Baccourci 154 Lake Pelto and Pelican Lake 154 Examination of oyster beds at mouth of Bayou Grey and Little Lake 155 Experimental studies of the effect of oil on oysters 158 Survival of oysters in oil-polluted water 159 Experiments with surface film of oil 160 Survival of oysters in sea water passed through oil 161 Immersion of oysters in oil 162 Effect of oil on glycogen content of oysters 163 Experiments with brine 164 Effect of brine on glycogen content of oysters. . . 167 Effect of oil on feeding of oysters 167 Effect on the adductor muscle 167 Effect of oil and oil-well bleed-water on the rate of feeding of oysters. 170 Results obtained with the cone method 174 Results obtained with the drop-counting method 183 Effect of bleed-water on the rate of feeding 188 Effect of consecutive treatments. 191 Effect of crude oil on diatoms 193 Method. 194 Effect of heavy surface layer of oil on Nitzsehia culture 196 Effect of oil held on the bottom 197 Effect of water-soluble fraction of oil 199 Effect of oil held in collodion bags... 200 Effect of brine on Nitzsehia 200 Glyptocephalus cynoglossus 341 gorbuscha, Oncorhynchus 27-39, 643-666, 667, 695, 804-812, 820 856 GENERAL INDEX Page goosefish, American 341 gobies (family Gobiidae) 534-573 grenadier, common 340 haddock - 339 hagfish 321 hake 012 long-finned 339 silver 338 spotted.. 339 hardtail 331 harvestfish 333 Heiicolenus dactylopterus 334 Hemitripterus americanus 335 herring 119,327 round 327 thread 327 Hildebrand, Samuel F. and Louella E. Cable: Further notes on the development and life history of some teleosts at Beaufort, N. C 505-642 Spanish mackerel, Scomberomorus maculatus, with notes on related species 508 Characters of the adult 509 Spawning 510 Descriptions of the young 510 A discussion of the relationship of the species Scomberomorus and the probable identity of the young 517 Pinfish, Lagodon rhomboides 518 Characters of the adult 519 Spawning 520 Descriptions of the young 520 Distribution of the young 524 Growth 526 Sheepshead Archosargus probatocephalus 526 Characters of the adult 526 Spawning. 527 Descriptions of the young 528 Distribution of the young 532 Food 532 Growth 533 Spadefish Chaetodipterus faber 534 Characters of the adult 535 Spawning 536 Descriptions of the young 536 Distribution of the young 543 Growth 543 Family Gobiidae (gobies) 543 Distinguishing characters of the young of the genera Oobiosoma, Microgobius and Gobionellus . 645 A comparison of the eggs and the young of some American and European gobies 546 Naked gobies, Oobiosoma bosci and Oobiosoma ginsburgi _ 548 Key to the adults of the local species 548 Spawning. 550 Descriptions of the eggs and young 551 Distribution of the young 558 Growth 559 Holmes goby, Microbogius holmesi 559 Spawning 560 Descriptions of the young 560 Distribution of the young 563 Growth 564 Local species of the Gobionellus 564 Scallop fish Oobionellus bole soma 565 Spawning 565 Descriptions of the eggs and young 566 Distribution of the young 571 Growth 571 Local species of the Gobionellus— Continued. Page Ocean goby Gobionellus oceanicus 571 Spawning 572 Descriptions of the young 572 Distribution of the young 573 Family Blenniidae (blennies). 573 Key to the genera and species 574 The characters of the eggs and newly hatched young 574 Distinguishing characters 574 A comparison of the eggs and young of some American and European blennies 575 Seaweed fish, spotted, Hypsoblennius hentz 576 Spawning 577 Descriptions of the eggs and young 579 Distribution of the young 589 Growth 589 Blenny, Hypleurochilus geminatus 589 Spawning 590 Descriptions of the eggs and young 592 Distribution of the young 602 Growth 603 . Banded blenny, Chasmodes bosquianus 603 Spawning 605 Descriptions of the eggs and the newly hatched young 605 Hakes (genus Urophysis). 612 Key to the species 613 Spawning 613 Descriptions of the eggs and young 614 Distribution of the young 626 Growth 627 Sole, American , Archirus fasciatus 630 Characters of the adult 630 Methods of collecting 630 Spawning 631 Descriptions of the eggs and young 632 Growth 640 Hildebrand, Samuel F. and Louella E. Cable: Reproduc- tion and development of the whitings or kingfishes, drums, spot, croaker, and weakfishes or sea trouts, family Sciaenidae, of the Atlantic coast of the United States 41-117 Artificial keys to the eggs and young as far as known __ 42 The whitings or kingfishes ( Menticirrhus americanus, M. saxatilis, and M. littoralis) 51 Characters of the adults 52 Key to the species 52 Menticirrhus americanus 53 Spawning 53 Descriptions of the young 54 Distribution of the young 62 Growth 62 Menticirrhus saxatilis 64 Spawning 65 Descriptions of the eggs and young 65 Distribution of the young 70 Growth 70 Menticirrhus littoralis 70 Spawning 71 Descriptions of the young 71 Distribution of the young 75 Growth 75 Star drum Slellifer lanceolatus 75 Spawning 76 Descriptions of the young 76 Distribution of the young 83 Growth 83 Banded drum Larimus fasciatus 84 Spawning 84 Descriptions of the young 85 Distribution of the young 91 Growth 91 GENERAL INDEX 857 Page Hildebrand, Samuel F. and Louella E. Cable — Con. Weakfishes or sea trouts (Cynoscion nebulosus, C. regalis, and C. nolhus) 91 Key to the species 91 C ynoseion nebulosis _ 92 Spawning 93 Descriptions of the young 94 Distribution of the young 101 Growth 102 Cynoscion regalis. 102 Spawning... 103 Descriptions of the eggs and young 104 Distribution of the young 107 Growth 108 Cynoscion nothus. 110 Spawning.. 111 Descriptions of the young 112 Distribution of the young 115 Growth 116 Hile, Ralph: Age and growth of the cisco, Leucuchthys artedi, in the lakes of the northeastern highlands of Wisconsin. 211-317 Materials 214 Methods 215 Gear used in collecting 215 Methods of fishing 215 Field data recorded for individual specimens 216 Treatment of preserved specimens and resulting shrinkage 217 Preparation and examination of scale material 217 Miscellaneous considerations 218 The scale method.. 218 Assessment of age and calculation of growth 218 Lee’s phenomenon in the Silver Lake cisco 222 Possible causes of Lee’s phenomenon in the Silver Lake cisco. 224 General growth curves for the Trout Lake, Muskel- lunge Lake, Silver Lake and Clear Lake cisco pop- ulations 226 Growth in length 226 Growth in weight 230 Comparison of the growth of the Trout Lake, Muskel- lunge Lake, Silver Lake and Clear Lake cisco popu- lations with that of cisco populations in other regions. 232 Range of length in individual age groups; maximum length and weight 234 Condition and the relationship between length and weight. 237 Length of growing season 249 Relationship between density of population and rate of growth 253 Age composition of the samples and the relative abun- dance of year classes. 263 Age at maturity and sex ratio 267 Annual increments of growth 271 Variation in the amount of growth in different calendar years 271 Bimodality in the calculated growth for the first year of life. 280 Growth compensation 282 Growth relationships in the Trout Lake, Muskellunge Lake, Silver Lake and Clear Lake cisco populations. 286 Physical-chemical factors 287 Density of population 288 Length of the growing season 288 Parasitization 289 Condition 289 Sex ratio 292 Fishes associated with the cisco 293 Selective action of gill nets 294 Hippoglossoides platessoides 340 Histrio histrio. 341 Page Hopkins, A. E.: Experimental observations on spawning, larval development and setting in the Olympia oyster, Ostrea lurida 439-503 Method of cultivation 441 Enemies of the oyster 441 Aims of investigation 443 Hydrographical observations 443 General description of region 443 Temperature 444 Salinity and pH 448 Spawning 456 Size of broods 458 Relation of temperature to spawning 460 Spawning season 464 Development of larvae 467 Setting 471 Effect of angle of surface. 472 Method of determining frequency of setting 475 Setting seasons, Oyster Bay 477 Setting seasons, Mud Bay 483 Periodicity of setting 487 Stages of tide and setting 489 Depth of setting 493 Correlation between spawning and setting 495 Hopkins, A. E.: Adaptation of the feeding mechanism of the oyster, Ostrea gigas, to changes in salinity 345-364 Homing instinct and age at maturity of the pink salmon, Oncorhynchus gorbuscha 27-39 Hutchinson, Samuel J. and Frederick A. Davidson: The geographic distribution and environmental limitations of the Pacific salmon (genus Oncorhynchus) 667-692 humpback salmon 667, 695, 804-812, 820 Isurus nasus 322 tigris 322 Kelez, George B. and George A. Rounsefell: The salmon and salmon fisheries of Swiftsure Bank, Puget Sound and the Fraser River 693-823 kingfishes, or whitings 61 king salmon 695, 795-803, 819 Koehring, Vera, Paul S. Galtsoff, Herbert F. Prytherch, and Robert O. Smith: Effects of crude oil pollution on oysters in Louisiana waters 143-210 kroyeri, Xiphopenaeus 2,8,13,21 lake shrimp 1 Lamprisregius 332 lamprey, sea.. 321 lancetfish 328 Larimus fasciatus 84 leatherjacket 332 Leptoclinus maculatus 337 Leucichthys artedi, cisco, age and growth of in the lakes of the northeastern highlands of Wisconsin 211-317 Limanda ferruginea 340 Lindner, Milton J., W. W. Anderson and F. W. Wey- mouth: Preliminary report on the life history of the common shrimp, Penaeus setiferus 1-26 Liparis liparis 336 lookdown 332 Lophius americanus 341 Louisiana waters, effects of crude oil pollution on oysters in.. 143-210 Lumpenus lampetraeformis 336 Lycenchelys verrillii 338 mackerel 329 Spanish 508 Macrobrachium sp 1 Macrourus bairdii 340 Makaira albida 330 858 GENERAL INDEX Page Mancalias uranoscopus 342 Marking experiment, pink salmon, in British Columbia.. 36 marlin, spearfisb 330 Maurolious pennanti 328 Melanogrammus aeglefinus 339 Menticirrhus amerieanus. 53 littoralis 70 saxatilis 64 Merlnceius bilinearis 338 Migrations of pink salmon ( Oncorhynchus gorbuscha) in the Clarence and Sumner Straits regions of southeastern Alaska 643-666 Monacanthus ciliatus 334 hispidus 334 Mustelus mustelus 321 Myoxocephalus octodecimspinosus 335 Myxine glutinosa 321 Narcacion nobilianus 325 Naucrates ductor 331 needlefish.. 328 Neoliparis atlanticus.. 336 Observations on spawning, larval development, and setting in the Olympia oyster 439-503 Oligoplites saurus 332 Oncorhynchus gorbuscha 27-39, 643-666, 667, 695, 804-812, 820 opah 332 Opisthonema oglinum 327 Ostrea gigas 345-364 lurida 439-503 virginica 345 Oyster, effects of crude oil pollution on, in Louisiana waters 143-210 Oyster, Ostrea gigas, adaptation of the feeding mechanism of to changes in salinity 345-364 Oyster, Olympia, observations on spawning, larval devel- opment, and setting 439-503 Pacific salmon (genus Oncorhynchus), geographic distribu- tion and environmental limitations of 667-692 Paralichthy s oblongus 340 pearlsides 328 Pearson, John C.: The life history of the striped bass, or rockfish, Roccus saxatilis 825-851 Distribution 826 Abundance 827 Spawning grounds 829 Spawning season 830 Size and age at maturity 830 Eggs and young 831 Growth 837 Food habits 839 Movements 840 Fishery 846 Penaeus brasiliensis 2, 4, 6, 8, 21 Penaeus setiferus, common shrimp, preliminary report on the life history of 1-26 Peprilus alepidotus. 333 perch, yellow 213 Petromyzon marinus 321 Pholis gunnellus 336 pike-perch 213 pilotfish 331 pinfish.. 518 Pink salmon, Oncorhynchus gorbuscha, the homing instinct and age at maturity of 27-39 Pink salmon, Oncorhynchus gorbuscha. Migrations of in the Clarence and Sumner Straits regions of southeastern Alaska 643-666 pink salmon, Oncorhynchus gorbuscha 27-39, 643-666, 667, 695, 804-812, 820 Page pipefish, common 329 pelagic 329 Pisodonophis cruentifer.. 326 plaice, American 340 Pollachius virens 338 pollock 338 Pomatomus saltatrix 332 Pomolobus aestivalis 327 Pomolobus pseudoharengus 327 Preliminary report on the life history of the common shrimp, Penaeus setiferus .. 1-26 Prionaee glauca 321 Prionotus strigatus 336 Pronotus triacanthus 332 Prytherch, Herbert F., Robert O. Smith, Vera Koehring and Paul S. Galtsoff: Effects of crude oil pollution on oysters in Louisiana waters 143-210 Pseudoplcuronectes amerieanus 340 dignabilis 341 puffer 334 Quinnat salmon . _ 667, 695, 754-778, 819 Raja diaphanes 324 eglanteria 325 erinacea... 324 scabrata 324 senta 325 stabuliforis 325 red salmon 667, 695, 754-778, 819 Remora remora... 336 Roccus lineatus 333 rock bass.. 213 rockling, four-bearded 339 rosefish 334 black-bellied 334 Rounsefell, George A. and Edwin H. Dahlgren: Races of herring, Clupea pallasii, in southeastern Alaska 119-141 Spawning and feeding localities 120 Analysis of vertebral counts 123 Discussion of factors influencing vertebral count distribution within a population 123 Existence of races proven by heterogeneity of samples from all localities 124 Homogeneity of material from individual locali- ties 129 Segregation of races 129 Analysis of growth rates 133 Analysis of year classes 138 Tagging 138 Rounsefell, George A. and George B. Kelez: The salmon and salmon fisheries of Swiftsure Bank, Puget Sound, and the Fraser River 693-825 The Pacific salmons 695 Fishing districts 695 Development of the fisheries 697 Production and value 699 Need for investigation 700 Gill-net fishery .... 701 Fraser River 701 Puget Sound 712 Trap fishery 713 Reef nets 713 Construction of the traps. 714 Number in operation. 716 Locations fished 717 Cannery expansion from the trap fishery 719 Season 719 Seasonal occurrence of each species. 721 Relative importance of each species and each dis- trict 723 GENERAL INDEX 859 Rounsefell, George A. and George B. Kelez— Contd. Page The purse-seine fishery.. 725 Drag seines. . 725 Development of the purse seine 726 Development of the modern purse-seine vessel 728 Evaluation of fishing intensity 730 Seasonal fluctuations in fleet size 730 Changes in composition of the fleet 734 Relation of vessel size to efficiency 736 Seasonal occurrence of each species 740 Pugot sound fishery. 740 Cape fishery 741 Fishing seasons in different districts 742 Puget sound 742 Cape Flattery - 744 Relation of fishing intensity to seasonal occurrence. . 745 Relative importance of each species.. 747 Puget Sound 747 Cape Flattery 748 The troll fishery 749 Development of the fishery 749 Importance 750 Seasonal occurrence of cohos and kings 751 Sport fishing 753 Sockeye salmon .. 754 Genera! life history 754 Sockeye rivers of the region 756 Migration in salt water 758 Total pack of the Fraser River system 758 Method and locality of capture 759 Changes in abundance of different portions of the run 763 Coho salmon 781 Life history 781 Locality of capture by different types of gear 784 Seasonal occurrence in various areas 786 King salmon 795 Life history 795 Locality of capture by different types of gear 797 Seasonal occurrence in different areas 799 Seasonal occurrence of red and white king salmon. 800 Changes in abundance 803 Pink salmon __ 804 General life history 804 Migration 805 Method and locality of capture 805 Seasonal occurrence in northern and southern districts. .. 808 Changes in abundance between early and late years 808 Indices of abundance from traps 811 Abundance from purse-seine catches 812 Comparison of purse-seine and trap indices 812 Chum salmon 813 General life history 813 Method and locality of capture 814 Seasonal occurrence in northern and southern districts 814 Abundance from Admiralty Inlet traps 815 Abundance from purse seines 815 rudderfish 331 Salmon, pink, Oncorhynchus gorbuscha, the homing instinct and age at maturity of 27-39 salmon, pink 27-39, 643-666, 667, 695, 804-812, 820 Salmon and salmon fisheries of Swiftsure Bank, Puget Sound, and the Fraser River 693-823 seabrem 332 lamprey 321 raven 335 robin, red-winged. 336 snail 336 striped 336 Page Sarda sarda 330 sargassum fish 341 sauger - 213 scad, mackerel 331 big-eyed 331 Schroeder, William C. and Henry B. Bigelow: Supple- mental notes on fishes of the Gulf of Maine 319-343 Scomber scombrus 329 Scomberesox saurus 328 sculpin, deep-sea.. 335 hook-eared 335 longhorn — 335 mailed 335 sea trouts (weakfishes). 91 seaweed fish, spotted — 576 Sebastes marinus.. — 334 Selene vomer 332 Seriola zonata - 331 shanny, Arctic - ... 337 radiated. . 337 shark, dusky 321 great blue 321 greenback. 323 hammerhead 322 mackerel 322 sharp-nosed 322 shovel-head. 322 thresher.. 322 white 322 sbeepshead 526 Shrimp, common, Penaeus setiferus, preliminary report on the life history of 1-26 shrimp, lake.. 1 Sicyonia sp 2 silver salmon 667, 695, 781-794, 819 silverside salmon 667, 695, 781-794, 819 skate, barn-door.. 325 big 324 brier. 325 little.... — 324 prickly... 324 smooth 325 Smith, Robert O., Vera Koehring, Paul S. Galtsoff and Herbert F. Prytherch: Effects of crude oil pollution on oysters in Louisiana waters 143-210 sockeye salmon. 667,695,754-778,819 sole, A.:merican 630 Somniosus microcephalus. 323 Southeastern Alaska: Races of herring ( Clupea pallasii) in 119-141 The migrations of pink salmon ( Oncorhynchus gor- buscha) in the Clarence and Sumner Straits regions of.. 643-666 spadefish 534 Spanish mackerel 508 spearfish (marlin).. 330 Spheroides maculatus— 334 Sphyraena borealis 329 spring salmon 695, 795-803, 819 Squalus acanthias — 322 Stellifer lanceolatus 75 Stichaeus punctatus 337 Stream pollution, detection and measurement of— 365-437 sturgeon, common 326 short-nosed 326 sucker 213 Supplemental notes on fishes of the Gulf of Maine 319-343 swordfish 330 Syngnathus fuscus 329 pelagicus 329 860 GENERAL INDEX Page Page Taractes prineeps 332 Teleosts, further notes on the development and life history of at Beaufort, N. C 505-642 Thunnus thynnus 329 torpedo 325 Trachurops crumenophthalma 331 Traohurus trachurus 331 Traehypenaeus constrictus 2,8,21 triggerfish 333 Triglops omnatistius 335 trout, bastard _ 110 gray --- 102 lake... 213 speckled - 92 spotted.. 92 trumpetfish - 328 tuna - 329 Ulvaria subbifurcata 337 unicornflsh 334 Urophycis chesteri 339 regius 339 viperflsh - - 328 vulgaris, Crago 6,12 Weymouth, F. W., Milton J. Lindner and W. W. Ander- son: Preliminary report on the life history of the com- mon shrimp , Penaeus setiferus 1-26 Production and value of 2 Previous work on life history 6 Life history 8 Nature of data 8 Interpretation of data 9 Recognition of age groups _ - 9 Spawning 11 Sex-ratio 14 Larvae- 15 Young 15 Growth 18 Fate of adults - 19 Habits 21 Depletion and protection - 22 Whitings or kingfishes, drums, spot, croaker, and weak- fishes or sea trouts, family Sciaenidae, of the Atlantic coast of the United States, reproduction and develop- ment of 41-117 whitefish — 213 wolffish, spotted 337 wrymouth 337 Xiphias gladius 330 Xiphopenaeus kroyeri - 2, 8, 13, 21 weakfishes (sea trouts) 91 spotted- 92 gray 102 yellow perch 213 Zoarces anguillaris- 337