53D CONGRESS, , SENATE. { Mis. Doc. REPORT COMMISSIONER OF FISH AND FISHERIES INVESTIGATIONS IN THE COLUMBIA RIVER BASIN IN REGARD TO THE SALMON FISHERIES. PRESENTED BY THE VICE-PRESIDENT, May 31, 1894.—lLaid on the table and ordered to be printed. WASHINGTON: GOVERNMENT PRINTING OFFICE. 1894. “IJYOW/LIVE'0I BNIOH VW HLIT Dew as “NISW4 YSAlY VIEIWNTO5S | | i — FHL NI Pons) | f+ NOWTVS 10 NOMA LSI 7 PE paxdae oS, | JO S\.LIND] IVHOLVN WH, )NIMOHS dv “IIILOTSSULULO!) PIEU0( SIN TILEY SIE, SoLloy sty pue ysty Jo UOTSSIULULOL*) LS “Ol T EEL wav GE) Ve | J ; | ey Cina 4 linia. } NY preuroZP a | —~/ cd | \ Verein S | 4 N= : i : Peete RO an sig 5 alas.) g ! Contrary IY - 4 3) t | sje as ee La = estas Giles THE SALMON FISHERIES OF THE COLUMBIA RIVER BASIN. By MARSHALL McDONALD, Onited States Commissioner of Fish and Fisheries. U.S. CoMMISSION OF FISH AND FISHERIES, Washington, D. C., May 31, 1894. Hon. ADLAI BH. STEVENSON, President of the Senate: Str: In compliance with instructions conveyed in the provisions of the Sundry Civil Bill, which became a law August 5, 1892, [ have the honor to submit a report of investigations in the Columbia River Basin. The first of the provisions above referred to authorized the expenditure from the appropriation for inquiry respecting food-fishes of $2,000, or so much thereof as may be necessary, ‘In examining the Clarke’s Fork of the Columbia River, with the view to ascertain the obstructions which prevent the ascent of salmon up said river to the Flathead Lake and adjacent waters.” The second provision directed an investigation and report respecting the advisa- bility of establishing a fish-hatching station at some suitable point in the State of Washington, and appropriated for the same “$1,000, or as much thereof as may be necessary.” It was not known whether the failure of the salmon to enter the Clarke Fork of the Columbia was due to natural obstructions preventing their ascent, or was to be attrib- uted to the extensive fishing operations prosecuted in the Lower Columbia, or possibly to other causes to be disclosed by the proposed investigation. Again, the location of the hatchery proposed for the State of Washington would be necessarily determined by our ability to secure an adequate supply of spawning salmon within convenient distance of the hatchery. It appearing probable that the methods of the large fisheries pursued in the Lower Columbia, if permitted to continue, would effectually intercept the run of salmon to the headwaters, and thus defeat the object for which the hatchery is proposed, it was thought proper and expedient to institute a general investigation covering the entire Columbia River Basin, and if conditions were disclosed threatening disaster to these valuable and productive fisheries, to bring the matter to the attention of Congress and the States interested in their prosperity. The direction of the field investigation was intrusted to Prof. B. W. Evermann, assistant in the Division of Inquiry Respecting Food-Fishes, whose report is appended to and constitutes an integral part of the report of the Commissioner of Fisheries. 3 4. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. A very complete statistical investigation into the history, methods, apparatus, present conditions, product, and annual value of the salmon fisheries of the Columbia has also been made by Mr. W. A. Wilcox, under the direction of Dr. H. M. Smith, assistant in charge of the Division of Statistics and Methods of the Fisheries, the results of which are embodied and discussed in the report which is herewith respect- fully submitted. CONDITIONS DETERMINING THE SALMON PRODUCTION OF A RIVER BASIN. There are fundamental conditions determining the salmon production of a river basin and the nature and extent of the fisheries which may be maintained without overtaxing the productive capacity of the river. Al the species of salmon which are the object of the fisheries are alike under the constraint of a natural law, which com- pels them to enter the fresh waters for the purpose of spawning. Some species ascend to a relatively short distance above tide water. Others, like the chinook, push their migrations to the remotest sources of the rivers and tributary streams when not pre- vented by natural or artificial obstructions. Where the area of distribution is con- tracted by the erection of barriers, dams, or other obstructions which the salmon can not surmount, the production of the river is diminished pro tanto, for the reason that the young salmon remain for some months in the waters in which they are hateched— they must here find their food—and consequently the extent of the feeding-grounds open to them will be the measure of nature’s ability to repair the waste occasioned by natural casualties and the fishing operations. If there be no contraction of the breeding area by artificial obstructions, but, on the other hand, the times, methods, and apparatus of the fisheries are such as to intercept or in a large measure prevent the run of salmon into and up the rivers, then a serious decline in the fisheries is inevitable. 7 It is possible by fish-cultural operations pursued on an adequate scale, by hatching and planting the fry in the head waters of the Columbia and its tributary streams, to realize the fuil productive capacity of the river, so long aS eggs can be obtained in sufficient numbers to furnish a basis for the extensive operations required. This would not be possible, however, if the fishing operations in the lower river practically excluded the salmon from the streams to which it would be necessary to have recourse to obtain a supply of eggs. It is evident, therefore, that fish-cultural operations can not be relied upon exclusively or chiefly to maintain the salmon supply in the Columbia. The regulation of the times, methods, and apparatus of the fisheries should be such as to assure the largest opportunity practicable for reproduction under natural condi- tions. Artificial propagation should be invoked as an aid and not as a substitute for reproduction under natural conditions. THE LIMITS OF MIGRATION OF SALMON. The limits of migration of salmon in the Columbia River basin, as determined by impassable falls in the larger tributaries of the Columbia and their affluents, is shown in the accompanying chart, there being no serious obstructions existing in the main river within the limits of the United States. The area of distribution is approximately 90,000 square miles. This immense tract is drained by innumerable streams of clear cold water, into which the salmon enter for the purpose of spawning and up which they ascend till their progress is stopped INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 9) by falls or other obstructions which they cannot surmount. These waters furnish the feeding grounds of the young salmon during their early life, which is spent in the fresh waters. Their migration seaward does not begin until they are at least a year old and have attained a length of from 8 to 10 inches. These streams are the nurseries of the great salmon fisheries of the lower Columbia. From each goes out every year a colony, more or less numerous, to swell the aggregate of young salmon necessary to repair the waste by natural casualty and by capture. The area of natural distribution has not as yet been very materially abridged. Certain streams, such as the Bruneau and the Boise, have been obstructed by dams near their mouths, but the vast extent of waters still accessible to salmon and affording suitable breeding and feeding grounds, indicates that we must look to other causes to explain any ascertained deterioration in the salmon fisheries of the Columbia. DECREASE OF SALMON IN THE HEAD WATERS OF THE COLUMBIA RIVER. The investigations made by Prof. Evermann and the parties under his direction establish conclusively the fact that there has been a very great reduction in the num- ber of salmon frequenting the head waters of the Columbia River and its tributaries. This decrease is more notable in the main river. In the early history of the fishery salmon were found in the head waters in marvelous abundance. According to the information obtained by Prof. Evermann: They were abundant in the Columbia River at Kettle Falls as late as 1878. Since then there has been a great decrease. They have been scarce since 1882. Since 1890 there have been scarcely any at Kettle Falls. The Meyers Brothers say that they have been almost unable to buy any salmon for their own table from the Indians for three years. Certain Indians with whom we talked at Kettle Falls said salmon were once very abundant there, but that very few are seen now. Other persons testified to the same effect. Essentially the same information was obtained regarding the decrease of salmon in other parts of the upper tributaries of the Columbia, viz: at Spokane, in both the Big and Little Spokane rivers, and in the Snake River and its various tributaries. Dr. O. P. Jenkins, an assistant of Prof. Evermann, makes the following reportin reference to the Yakima River, Washington: The Yakima is the main stream of the valley. It receives many tributaries, the main ones being Manistash and Wilson creeks. The river near the city (Ellensburg) is 160 feet wide, by an average of 10 feet deep, and flows with a velocity of 1 foot per second. Temperature at 9:15 a.m., August 24, 1893, 60° F.; water clear. Those acquainted with the facts state that formerly, up to about 1885, salmon of three or four kinds, including the quinnat, ran up the stream to this valley and spawned in the river in great numbers; at present very few make their appearance. There is no reason to doubt—indeed, the fact is beyond question—that the number of salmon now reaching the head waters of streams in the Columbia River basin is insignificant in comparison with the number which some years ago annually visited and spawned in these waters. It is further apparent that this decrease is not to be attributed either to the contraction of the area accessible to them or to changed con- ditions in the waters which would deter the salmon from entering them. We must look to the great commercial fisheries prosecuted in the lower river for an explanation of this decrease, which portends inevitable disaster to these fisheries if the conditions which have brought it about are permitted to contiiue. The relations of the decreased number of salmon in the head waters to the devel- opment of the commercial fisheries is brought out in a very instructive way by an analysis of the following table: 6 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. Summary of the salmon-canning industry of the Columbia River from its origin to the present time. Gross weight Number of Average Gross weight} Number of Average Year. of salmon cases Value. value Year. of salmon cases Value. value utilized. packed. per case. utilized. packed. per case. Pounds. Pounds. 260, 000 4, 000 $64, 000 $16. 00 35, 750, 000 550, 000 | $2, 475, 000 $4. 50 1, 170, 000 18, 000 288, 000 16. 00 35, 184, 500 541, 300 2, 600, 000 4.80 1, 820, 000 28, 000 392, 000 14. 00 40. 911, 000 629.400 | 3, 147, 000 5. 0U 6, 500, 000 100,000 | 1, 350, 000 13.50 40, 300, 000 620,000 | 2,915, 000 4.70 9, 750, 000 150,000 | 1,800, 000 12. 00 35, 997, 000 553, 800 } 2, 500, 000 4.51 18, 000, 000 200,000 | 2,100, 000 10. 50 29, 152, 000 448,500} 2,135, 000 4.76 16, 250, 000 250,000 | 2,325, 000 9.30 23, 140, 000 356,000 | 2, 124, 000 5.97 16, 250, 000 250,000 | 2, 250, 000 9. 00 24, 211, 005 872,477 | 2, 327, 981 6.25 22,750, 000 350,000 | 2, 625, 000 7.50 20, 685, 495 309, 885 | 1, 809, 820 5. 84 24, 375, 000 375, 000 | 2, 250, 000 6. 00 28, 781, 385 435,774 | 2, 407, 456 5. 52 29, 250, 000 450,000 | 2, 475, 000 5. 50 26, 450, 635 398,953 | 2, 240, 964 5. 62 24, 700, 000 380,000 | 2, 052, 000 5. 40 32, 185, 995 487,338 | 2,679, 069 5. 50 29, 900. 000 460,000 | 2, 300, 000 5.00 24, 050, 000 370,000 | 2, 107, 500 5.70 31, 200, 000 480,000 | 2, 640, 000 5. 50 i — —— |} —. —- 34, 450, 000 530,000 | 2, 650, 000 5. 00 Total .| 658,424,515 | 10, 098, 427 | 59, 029, 790 5. 85 Canning operations on the Columbia River began in 1866, when 4,000 cases were packed and sold at an average of $16 per case. As early as 1872 the total pack reached 250,000 cases, the price per case having declined to $9. Hach succeeding year operations were extended and reached their culmination in 1883 and 1884, when upwards of 600,000 cases were packed each season. From this time on the catch declined, having reached its lowest point in 1889, the number of cases packed that season being 309,885, or less than half the number of cases packed in 1883 and 1884. Up to 1888, practically the entire pack consisted of the king or chinook salmon, and the fishing season did not extend beyond the first of August. In 1889 the packers began canning bluebacks and steelheads to make up the deficiency in the supply, and extended their operations to the first of September. DETAILED STATISTICS OF THE SALMON INDUSTRY OF THE COLUMBIA RIVER, 1889-92. The following series of tables shows, in some detail, the extent of the salmon fishery and canning industry of the Columbia River during the years 1889 to 1892, inclusive, as determined by the inquiries conducted by this Commission. The number of fishermen and shore employes connected with the salmon industry in each of the years named is indicated in Table A: A.—Table showing the number of persons employed in the salmon industry of the Columbia River from 1889 to 1892. How engaged. 1889. 1890. 1891. 1892. Oregon: Hishermentset hs se tepu sce eee eee eee eae 1,606} 1,648} 1,929] 2,064 Shoresmen and cannery employes..-.-------- 870 | 1,028] 1,057} 1,100 Motal\emeomeccemmiccemesascccceas saseeiace 2,476 2,712 | 2,986) 3, 164 Washington: WMishermens- ss tse ease k eee ease nels sence mee 1,535 | 1,510} 1,575 | 4,677 594 602 654 704 2,129 | 2,112) 2,229) 2,381 3,141] 3,194] 3,504] 3,741 1,464 | 1,630] 1,711 | 1,804 4,605 | 4,824) 5,215 | 5,545 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 7 The number and value of boats and apparatus and the value of shore property and capital employed in the salmon fisheries of the Columbia River in 1889, 1890, 1891 and 1892 is given in Table B. B.—Number and value of boats and apparatus, and the value of shore property, and cash capital employed in the salmon industry of the Columbia River in 1889, 1890, 1891, and 1892, 1889. 1890. 1891. 1892. Apparatus and capital. ae No. Value. No. Value. No. Value. No. Value. i} Oregon: - WA sovodeobososaososaoocd 751 $99, 850 776 $104, 400 876 $120, 815 998 $131, 550 Pile-drivers and scows- 21 5, 900 23 6, 300 30 8, 300 29 7, 400 Pound nets - 102 72, 300 98 76, 500 140 98, 900 247 173, 400 Trap nets.....-. c 2 1, GOD bonesosdlocecnsoseass 2 1, 600 2 1, 600 Seines .......-. q 7 4, 800 6 2, 700 19 11, 150 12 5, 650 ill nets..-...- 757 152, 000 760 159, 450 790 181, 265 861 190, 100 Soe! 31 120, 052 29 107, 552 30 108, 152 40 132, 852 Dip nets and squaw nets -. 95 475 85 425 60 300 50 250 Shore property 502, 955 |..--.-.- 486, 355 |----.... 455, 205) |.-.--25- 507, 805 Cash capital............--- 395, 000 |.------- 581, 000 |....-... | 520, 000 |.--....- 614, 000 Total ..-.........-- coad|boboesec 1, 354, 932 |..-.---- 1, 524, 682 |...-.... 1,505, 687 |.-.---.. 1, 764, 607 Washington: Boatseees- seen a 475 60, 340 468 59, 780 534 67, 280 538 64, 895 Pile-drivers and scows..--- 39 9, 050 37 9, 950 42 10, 750 45 13, 550 Pound nets..-..---. 62 48, 200 70 55, 200 98 77, 000 131 103, 400 Trap nets... 2 1, 400 2 1, 400 2 1, 400 al 700 Seines -....-- 33 18, 700 29 16, 400 30 16, 900 26 10, 000 Gill nets.-.... 436 88, 775 432 89, 480 472 101, 780 453 98, 130 Wheels 9 25, 000 12 48, 500 14 45, 000 17 49, 100 Dip nets and squaw nets. - 15 75 18 90 23 115 25 125 Shore property-.--- 245, 950 }-.-..--. = LOE | eoeasood 321, 050 |.-....-- 282, 800 Cash capital.........-.... 304, 000 |.------- 331,000 |..-.---- 332, 000 |......-. 330, 000 TUR Salman Faharies of the Columsis River (To taco page 15.) PLATE Ill, _\sa" m & LOWER COLUMBIA RIVER © 9 \ SHOWING THE LOCATION OF } SALMON APPARATUS, FISHING GROUNDS (qeorara \ BCANNERIES & | AND CANNERIES 1892 PT ADAMS " 5 Loti Nom - seg = oad . Pound nets indicated (hus — ' RUNES S} Weirs we 9 ©) Scibe hauls mo mM ee 1 o S ciliast fe ‘anneries SN 3 = es ~ WU M B I C O. < tay fi me ' Ts L b ' oe \ \ \ ~ \ f \ \ 7za\o0! TEN S F Brookhield és % Pillar Rockey Heaveland Lis, FISHERS SS oT. U_S.Commission of Fish and Fisheries. Marshall MS Donald,Commissioner. CHART OF THE ae Monticello: Tet\oo™ SSS = = ND 7a \s0 TEs\a0 SMis 200 532 A REPORT UPON INVESTIGATIONS IN THE COLUMBIA RIVER BASIN, WITH DESCRIPTIONS OF FOUR NEW SPECIES OF FISHES. BY CHARLES H. GILBERT, Professor of Zoology, Leland Stanford Junior University, AND BARTON W. EVERMANN, Jchthyologist of the U. S. Fish Commission. INTRODUCTION. The investigations upon which this report is primarily based were provided for by two items in the sundry civil bill, approved August 5, 1892. The first of these items authorized the expenditure, from the appropriation for inquiry respecting food- fishes, of the sum of $2,000, or so much thereof as may be necessary, ‘‘in examining the Clarke’s Fork of the Columbia River, with the view to ascertain the obstructions which prevent the ascent of salmon in said river to the Flathead Lake and adjacent waters.” The second item provided “for investigation and report respecting the advisability of establishing a hatching station at some suitable point in the State of Washington, $1,000, or so much thereof as may be necessary.” The purposes of these two investigations were very intimately related. Any inquiry regarding obstructions which might interfere with the movements of salmon in any of the tributaries of the Columbia would have a bearing upon the advisa- bility of establishing a salmon-hatchery at any point in that river basin. These two inquiries were therefore conducted as one, and the results are presented in a single report. This work was begun in September, 1892, by Dr. Charles E. Gorham, engineer and architect of the Commission, assisted by Mr. Barton A. Bean, of the U. S. National Museum, and Mr. A. J. Woolman, teacher of science in the high school at South Bend, Ind. Dr. Gorham died before the completion of the investigation, and Prof. Evermann was instructed by the Commissioner to continue the work during the summer of 1893. While carrying on these investigations he had the assistance of Drs. Charles H. Gilbert, Oliver P. Jenkins, and W. W. Thoburn, and Mr. Cloud. Rutter, all of Leland Stanford Junior University. The work was taken up by us at Pocatello, Idaho, August 2, it having been determined to include an examination of the obstruc- tions in Snake River and a preliminary study of the natural-history features of the upper waters of the Columbia basin, with special reference to the present or former occurrence of salmon in those streams. iG) 20 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. To expedite matters as much as possible in the limited time which could be given to the work, the force was divided into two parties at Pocatello. Gilbert, Thoburn, and Rutter were instructed to go down Snake River, examine the various falls in that stream, make investigations as to the physical and natural-history characteristics of as many of the tributary streams as possible, and then carry on similar inquiries along the Columbia from Idaho to the Lower Columbia. Evermann and Jenkins went up the Snake River to Idaho Falls, where the rapids were examined, and then proceeded to Sand Point, Idaho, where was begun the examination of Clarke Fork or the Pend d’Oreille River, the latter being the name by which this river is generally known in that region. The Pend d’Oreille River was examined throughout the entire distance from Sand Point to within a few miles of the international boundary line. The two parties came together at Spokane. From this point Evermann returned east, and the work was continued by Gilbert, Jenkins, Thoburn, and Rutter. Investigations were made at various points in the Lower Columbia basin, chiefly for the purpose of selecting a site for a salmon hatchery and for gaining information respecting the occurrence and abundance of salmon in the various streams tributary to the Lower Columbia. While carrying on the investigations regarding the obstructions to the free move- ment of salmon in these rivers and the selection of a salmon-hatchery site, considerable opportunities occurred for a study of the natural history of the salmon and the general natural-history features of the waters of the Columbia basin. Considerable valuable information was obtained regarding the former as well as the present distribution of salmon in this region. : Large collections of fishes were made at the various places where collecting was possible, and their study has greatly increased our knowledge of the variations in and the geographic distribution of the fresh-water fishes of the northwestern United States. In this report we give (1) detailed descriptions of the various streams visited by the different members of the party; (2) a list of the species of fishes obtained in the Columbia River basin, together with a discussion of their relationships and distribu- tion; (3) notes on the breeding colors of the whitefish (Coregonus williamsoni), by~ Barton A. Bean; and (4) an annotated list of the reptiles and batrachians obtained. The time which has been given to the study of the various problems pertaining to the salmon question has been wholly inadequate to a satisfactory understanding of the matter, and any views which we venture to give in this paper must be regarded as tentative. An exhaustive study of the natural history of the various species of salmon and trout of the Columbia has never been made. The investigations now in progress will, it is confidently expected, lead to a much better understanding of the questions involved. or INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. Bil 7 LIST OF STREAMS EXAMINED. The following is a classified list of the streams examined, together with the dates upon which the various places were visited: Snake River: President Camp, Wyoming, August 14, 1891 (Evermann and Jenkins); Idaho Falls, August 4 and 5 (Evermann and Jenkins); American Falls, August 5 (Gilbert); Shoshone Falls, August 6 (Gilbert); Twin Falls, August 6 (Gilbert); Auger Falls, August 7 (Gilbert) ; Blue Lakes, August 7 (Gilbert); Upper and Lower Salmon Falls, August 7 (Gilbert); Mouth of Boise River, Caldwell, Idaho, August 8 (Gilbert, Thoburn, and Rutter); Payette, Idaho, August 10, and Lewiston, Idaho, August 15 (Gilbert, Thoburn, and Rutter). - Ross Fork of Snake River, near Pocatello, Idaho, August 4 (party). Port Neuf River, Pocatello, August 2 (Evermann and Rutter) and August 3 (party). Mink Creek near Pocatello, August 3 (party). Little Wood River near Shoshone, August 5 (Thoburn and Rutter). Boise River near Caldwell, August 8 (Gilbert, Thoburn, and Rutter). Payette River at Payette, August 9 (Gilbert, Thoburn, and Rutter). Clearwater River near Lewiston, August 15 and 16 (Gilbert, Thoburn, and Rutter). Potlatch Creek near Lewiston, August 16 (Gilbert, Thoburn, and Rutter). Palouse River near Colfax, Washington, August 17 (Gilbert, Thoburn, and Rutter). Grande Ronde River near La Grande, August 11 (Thoburn). Pataha River at Starbuck, August 14 (Gilbert, Thoburn, and Rutter). Pend @ Oreille River: Throughout the entire distance from Albany Falls, Idaho, to Big Eddy Cation, near the international boundary line, August 9 to 15 (Evermann-and Jenkins), and from its mouth to the international boundary, September 23 to 26, 1892 (Gorham and Bean). Deer Lodge, Little Blackfoot, Big Blackfoot, Hell Gate, Bitter Root, Missoula, and Flathead rivers, — together with many of their tributary streams, July and August, 1891 (Evermann and Jenkins). Flathead Lake, August 1 to 4, 1891 (Evermann and Jenkins), and September, 1892 (Gorham and Woolman). Thompson Falls, September, 1892 (Gorham, Bean, and Woolman). Lake Pend d’ Oreille at Sand Point, Idaho, August 7 (Evermann and Jenkins). Upper Columbia River: Kettle Falls, August 16 (Evermann and Jenkins), and at the mouth of Pend d’Oreille River, September 23, 1892 (Gorham and Bean). Colville River from Meyers Falls to its mouth, August 16 (Evermann and Jenkins). Spokane River in the vicinity of Spokane, September, 1892 (Gorham and Bean), and August 18 to 21 (Evermann and Jenkins). Little Spokane River below Dart’s Mill. September, 1892 (Bean), and near Dart’s Mill, August 18 (Evermann and Jenkins). Ceeur d’Alene River at Wardner, August 19, and Coeur d’Alene Lake at Cour d’Alene, August 21 (Gilbert, Thoburn, and Rutter). Hangman Creek near Spokane, September, 1892 (Bean), and at Tekoa, August 18 (Gilbert, Thoburn, and Rutter). Lower Columbia River: Pasco, Wallula, Umatilla, Dalles, Portland, and Astoria, August 11 to 27 (Rutter and Thoburn). Walla Walla River near Wallula, August 23 (Thoburn and Rutter). Mill Creek near Walla Walla, August 14 (Thoburn and Rutter). Umatilla River near Pendieton, Oregon, August 12 (Gilbert, Thoburn, and Rutter), and at Umatilla, August 11 and 23 (Thoburn and Rutter). Des Chutes River at its mouth, August 24 (Rutter). : Yakima River near North Yakima and Ellensburg, August 23 and 24 (Jenkins), Natchess River near North Yakima, August 24 (Gilbert and Jenkins). Cowlitz and Toutle rivers near Castle Rock, August 28 and 29 (Gilbert and Jenkins), Newaukum River near Chehalis, August 28 (Gilbert and Jenkins). Skookumchuck River near Centralia, August 27 (Gilbert and Jenkins). Lake Washington at Seattle, June 25, 1892 (Evermann). Snoqualmie River at Snoqualmie Falls, June 26 and 27, 1892 (Evermann), 22 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. INVESTIGATIONS WITH REFERENCE TO THE SELECTION OF A SITE FOR A SALMON HATCHERY IN THE STATE OF WASHINGTON. Every stream and every point visited was considered with regard to its fitness for salmon-hatching purposes. The majority of the places are, however, not suited at all to such ends, and only such locations as seem to possess most or all the required physical and biological conditions need be treated in detail in this report. LOWER COLUMBIA. There are Several reasons why a salmon hatchery would be better located on some tributary of the Lower Columbia rather than the Upper Columbia or the Snake. The supply of salmon would be more certain and the condition of the salmon better. So far as is known to us, salmon which enter the Columbia in the spring pass by the ‘mouths of the lower tributaries and press on higher up the stream. It is probably these fish which arrive in the Upper Snake in the vicinity of Glen’s Ferry and Salmon Falls in the latter part of August and in September. All observers on the Upper Snake agree that they arrive at this time and spawn from September 1 on to October or November. The fish of the fall run enter the Columbia a short time only before they are ready to spawn. So far as we now know, the most of these turn directly into streams near the mouth of the river and spawn a short time after their entrance into the Columbia. A second point in favor of such a location for a hatchery would be, perhaps, that the young fish when turned into the stream would stand a better chance of reaching salt water than they would if they had the whole course of the river to trayerse, during which time they are exposed to the attacks of all their fresh-water enemies. A third point in favor of such a location is the accessibility of various points in Washington along the lower course of the Columbia. Two streams were selected for examination, the Yakima River and the Cowlitz. Both of these rise in the high mountain region of southwestern Washington, and receive their waters largely from the snows of Mount Ranier, Mount Adams, and Mount St. Helen. They run through regions very different in their physical charac- teristics and in their climate. The Yakima lies to the east of the Cascade range and runs down through a dry valley covered with sagebrush and devoid of trees, except along the immediate vicinity of the stream itself. The summer season is very hot and the winter correspondingly cold. So far as the character of the stream itself is con- cerned, it seems admirably adapted for a hatchery. At North Yakima the stream is perfectly clear, flows rapidly in an open valley over gravel and sand, and had a tem- perature of 64° August 23. It receives an important tributary, the Natchess, 1 mile above the town. At its mouth this stream is about 75 feet wide with an average depth of 2 feet, and with a current of 14 feet per second. The temperature was 574° at 9:30 a.m. Were other conditions favorable, no better stream could be found for a hatchery than the Natchess. While salmon used to ascend the Yakima and its tributaries in large numbers, they have greatly fallen off of late years. It is now very doubtful whether a hatchery located at any point on this stream could depend for spawn on the fish which ascend ‘1354 081 LNOSY LNSOSAG WLOL ‘YSAIY SHVNS 'STIVS NIML Salmon Fisheries of the Columbia River. (To tace page 22.) PLATE INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 23 the stream itself. Ifit were considered desirable to ship spawn to such a hatchery, the Natchess might be favorably considered. In considering the possibility of establishing a hatchery on the Yakima or its tributaries, it should be borne in mind that the stream flows through a wide valley, only partially under cultivation. Extensive canals are now being constructed with a view to irrigating the entire valley. Recent litigation sees to show that more water has been claimed on behalf of these canals than the stream will be able to furnish. It seems probable, therefore, that the entire supply will be withdrawn from the river during the summer and fall. Toutle River.—The Toutle River is a tributary of the Cowlitz. Near its mouth, near Castle Rock, an excellent site for a salmon hatchery can be found. ‘This is a beautiful, clear, and cold stream, furnishing an abundance ot water, which is never likely to be required for other purposes. The temperature of the water at 11 a.m. August 27 was 59.59. The Toutle is a natural spawning-ground for the salmon, which still come into it in large numbers. They could be taken in the deeper pools in gill nets, and the character of the stream is such as to permit seining. The time at our disposal could not permit us to make a very thorough investigation of this stream and entirely prevented our visiting the Upper Cowlitz. From what we saw, however, we are inclined to recommend the Toutle River as being the best suited for hatchery purposes of any stream in Washington. THE UPPER COLUMBIA, Near Kettle Falls, Wash.—The Colville River flows into the Columbia at the town of Kettle Falls, about 2 miles below the Kettle Falls of the Columbia. An abundance of excellent water can be obtained from the Colville River, and plenty of suitable land can probably be had for nothing, as the people there are much interested in securing the hatchery. This site is about 2 miles from Meyers Falls, a station on the Spokane and Northern Railroad. The only objection to it is the uncertainty of getting a suffi- cient number of spawning salmon conveniently near. As already stated, salmon were abundant in the Columbia at Kettle Falls as late as 1878. Since then there has been a great decrease. They have been scarce since about 1882; since 1890 there have been scarcely any at Kettle Falls. The Meyers brothers say they haye been almost unable to buy any salmon for their own table from the Indians for three years. Certain Indians with whom we talked at Kettle Falls said salmon were once very abundant there, but that very few are seen now. Other persons testified to the same effect. Essentially the same information was obtained regarding the decrease of salmon in other parts of the upper tributaries of the Columbia, viz, at Spokane, in both the Big and Little Spokane rivers, and in the Snake River and its various tributaries. On the Little Spokane River, near Spokane, Wash.—This river, as elsewhere stated in this report, possesses all the natural conditions necessary for this purpose; and it has the advantage of having excellent shipping facilities in the numerous railroads centering at Spokane. The uncertainty of being able to obtain spawning salmon in sufficient numbers is, however, a fatal objection to this point, unless shipping the eggs from the Lower Columbia might be regarded as feasible. 24 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. SNAKE RIVER IN SOUTHWESTERN IDAHO. If the station does not necessarily have to be in Washington, a very good site can be found on Snake River in Idaho near Salmon Falls. Salmon seem still to ascend to that point in considerable numbers. For detailed description of this place see portion of this report pertaining to the Snake River. SUMMARY. In summing up the facts brought out by these investigations it may be said, first, that the absence of salmon from the Pend d’Oreille River is not necessarily due to the presence of falls in that stream, but to other causes, chief among which is the excessive catching of salmon in the Lower Columbia; second, that while it is true that the salmon are shut out by falls and dams from a large area, especially in the Upper Snake River basin, and that these limitations are increasing as the streams become useful for irrigation and mining purposes, it is nevertheless certain that the decrease in the salmon has been even greater and that the accessible waters suitable for spawning purposes are still more than ample to meet present needs ; and, third, that the desirability of establishing another salmon hatchery at some point in the Columbia basin will depend largely upon the nature of the fishery legislation in the States of Washingtoh, Oregon, and Idaho. It must be understood, however, that our knowledge regarding the present abun- dance of salmon, their relative abundance as compared with former years, and the location and extent of their spawning-grounds, is of the most general kind. While valuable so far as it goes, the information which we now possess upon these important questions is chiefly useful in indicating the nature of the investigations which must be carried on for several seasons before a thorough understanding of the salmon ques- tion can be reached. % DETAILED ACCOUNT OF THE VARIOUS WATERS EXAMINED. SNAKE RIVER. This river has been visited by us at the following points: President Camp, near the southern boundary of the Yellowstone Park; Idaho Falls; American Falls; Shoshone Falls; Twin Falls; Auger Falls; Blue Lakes; Upper and Lower Salmon Falls; at mouth of Boise River; at Caldwell, Payette, and Lewiston. The observations made at these various places are here given in order, beginning with the point farthest upstream which was visited. President Camp.—Snake River here flows through a wide meadow, grassy and open on the right side, but covered with a heavy growth of chapparal on the other. It is here a beautiful river with clear, cold water and gravelly bottom. The banks in the immediate vicinity of the camp are low, not exceeding 3 or 4 feet. In the main stream the current was pretty strong, but there are quiet nooks and coves where there was considerable water vegetation. The temperature of the water at 9 a. m., August 14, was 62.5°. Fishes were found to be abundant here, the red-horse sucker (Catostomus ardens), dace (Rhinichthys cataracte dulcis), chubs (Leuciscus hydrophlox and Leuciscus lineatus), white- fish (Coregonus williamsoni), cut-throat trout (Salmo mykiss), and the blob (Cottus bairdi punctulatus) being the species thus far known from the Snake River at that point.* Idaho Falls, Idaho, August 4 and 5, 1893.—At this point the river has cut its channel through the immense lava bed of that region. The banks of the stream are abrupt or vertical but broken and jagged walls of lava, reaching in some places as many as 15 to 20 feet or more above thesurface of the water. Large, detached masses of lava are frequent in the stream, and in the banks or bounding walls are many immense potholes, by far the largest and finest we have ever seen. The river is here con- fined to a relatively narrow channel, through which it rushes in a series of foaming rapids. There * See Evermann: Explorations in Montana and Wyoming, Bull. U. 8. Fish Comm. for 1891, 22. “L354 017 LNOGV LN3ADS3SG.1VLOL YSAIY SYVNS ‘S11¥V4 S3NOHSOHS Salmon Fisheries of the Columbia River (To face page 24.) PLATE V. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 25 are many comparatively quiet nooks, however, in the broken, irregular walls, and the water is very deep, perhaps 20 to 50 feet. While these rapids are quite turbulent, trout and even other species of fishes have no trouble in ascending them. Trout (Salmo mykiss) are common here, and ina large race which has been cut through the lava for milling purposes we obtained many specimens of dace, chubs, and suckers (Catostomus ardens). American Falls, Idaho. —At American Falls the Snake River is about 750 feet wide and flows but little below the general surface of the country. The shores have no abrupt banks, the northern shore only being followed by a low, rounded bluff 100 to 200 feethigh. This was apparently composed of gravel, as no lava could be seen projecting from it. The outline of the American Falls is very irreg- ular. Its position is determined by a basaltic ledge crossing the river. The position of the edge of this ledge is now marked by a series of islands, between which the river flows and below which it falls. This same ledge can be traced for some distance along the edge of the canon below the falls and is there seen to be underlaid by a layer of sandstone. At the falls, however, this seems not to be the case, the rock being lava from top to bottom. The western end of the falls is probably 200 to 300 yards farther upstream than the eastern end. The front of the falls is located, therefore, very obliquely to the course of the stream. They show nowhere any great vertical height, 15 feet being probably near the maximum. In several places the falls are so broken down as to present only a short stretch of steep rapids, with gentler rapids above and below. Below the falls the water becomes immediately deep, but the rapids above are, at the stage of water seen, extremely shallow. On the eastern side of the stream, especially, is a long stretch of these shallow rapids, in which the water averages not more than 6 inches deep, and it is here that the greatest obstacles to the ascent of fish would be found. When water is high in the spring, trout are seen to pass over the falls in large numbers, and it is probably true that even at a lower stage of water, as in the fall, fish can succeed “in passing this obstacle. A fishway could be made here at very little expense were it considered desirable. The stream here, as elsewhere in Idaho, flows throngh a country covered with sagebrush and the usual desert vegetation, bordered more or less thickly with willows. In the rapids at American Falls the rocks are thickly covered with green filamentous alge, and among the rocks are found very numerously crawfish, caddis worms, and other suitable food for fish. At American Falls the river descends about 70 feet and enters a cation, the surface of the country remaining about the same level, and from this point to below Shoshone Falls the stream descends deeper and deeper into its cation by a succession of falls and rapids. It flows here through what is known as the “ Lava Beds” of the Snake River, and the walls of its cation are composed of successive lava flows. But few streams find their way into the Snake River from the mountains of the north. As will be seen from the map, the greater number of these on flowing down from the mountains sink into the lava and are lost. Of this kind are Birch Creek, Little Lost River, and Big Lost River. There is thus a great stretch of country bordering the river on the north entirely without surface water. Towards the west the Malade or Wood River is the first stream. to find its way into the Snake from the north. The water which thus sinks near the base of the mountains apparently reappears inside the cation of the Snake, coming out as great springs at the base of the cliffs. The best-known of these lie between the Shoshone Falls and Glen’s Ferry. They emerge from the foot of the cliffs often as large streams and are used to irrigate the bottom lands which border the river on the north at that point. The water of these streams is beautifully clear and cold; trout abound in them, and the smaller minnows run up from the Snake into them. Crawfish (Astacus gambelii) also are very abundant. The temperature of the streams averages about 60°, and they would be admirably adapted for hatchery purposes. The salmon visit this part of the river in sufficient numbers to furnish roe for hatching, and this is probably the most available point where suitable water and an abundance of fish can be found for such a station in Idaho. Unnamed Falls.—The next falls in the course of the stream were not visited by any member of the party, as nothing was heard of them until we had passed that region, Théy seemed to be unnamed, According to Mr. J. L. Fuller, of Bliss, Idaho, the river has a vertical fall of about 40 feet a short distance above the mouth of Dry Creek, the latter a small stream coming in from the south, nearly midway between American and Shoshone Falls. Mr. Fuller worked a mining claim at the mouth of Dry Creek at one time, and is therefore well acquainted with the falls, which he states to be vertical and impassable to any kind of fish. Shoshone and Twin Falls.—The great obstacles to the passage of salmon up the Snake are found in Shoshone and Twin Falls, both of which are vertical and of great height. The erection of fish- 26 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. ways to permit the passage of salmon seems wholly impracticable. Both Shoshone and Twin Falls are formed by layers of more compact and lighter colored lava, which the stream wears away with great difficulty. Shoshone Falls can be reached by a stage ride of 28 miles from the town of Shoshone, on the line of the Union Pacific. The canon at this point is high and composed of black columnar basalt, which rises from the river’s edge as vertical cliffs, estimated to be about 800 feet high. The falls are said to be 210 feet high. The middle of the falls is higher upstream than either end, giving it a somewhat horseshoe-shaped appearance, and the front of the falls is about 1,200 feet wide. Twin Falls are 4 miles above Shoshone Falls and would be fully as serious an obstacle as the latter, even if fish were able to reach their foot. An island div ides the stream here into two portions, - both of which, however, fall nearly vertically a distance said to be 180 feet. On the north side of the stream the vertical portion of the falls is somewhat lower, the upper portion having worn back to form very strong rapids, through which no fish would be able to pass. We were not able to learn that salmon reached the foot of Shoshone Falls, although it is very probable that they do so. The stream immediately below the falls is deep and flows at the bottom of a very steep cation, and even if the salmon were there and spawned in the bed of the stream, it might be difficult to detect them. Four miles below Shoshone Falls is the first of the large springs already referred to. These rise near the northern shore of the river in what are known as Blue Lakes (see p. 27), and one of these springs forms a large river. The Snake River at this point has widened out and flows over a succes- sion of shallows, and has a considerable expanse of bottom lands, which can be cultivated whenever water can be put upon them. Auger Falls.—A gentleman living at Blue Lakes is of the opinion that salmon do not come above Auger Falls, which is found 4 miles below Blue Lakes. This was found to consist of a stretch of very strong rapids. At Auger Falls the river runs for a distance of at least 250 yards, hemmed in between basaltic walls, which vary in distance from 50 to 250 feet. As nearly as could be estimated, the stream falls in this distance about 50 feet, the last 20 feet of which is nearly vertical. In this entire stretch of 250 yards there is no resting-place for a fish, and the water dashes through it in whirls and eddies in such a way as to make it doubtful whether a salmon could sustain the long- continued effort necessary to pass the rapids. It is, however, certain that no single stretch of these so-called fallsis insurmountable. Both salmon and sturgeon are frequently taken below Auger Falls, but apparently not above them. At Auger Falls it was estimated that the current averaged 15 feet asecond. Marks on the rocks show that at high water the stream was at least 15 feet above the level seen at this time. ‘ Upper and Lower Salmon Falls.—From Auger Falls down to Salmon, Falls the valley of the Snake widens and the cliffs become broken down and more and more rounded, as though glaciated. On each side of the stream are found in. places extensive deposits of water-worn gravel, which are washed for gold. At the Upper Falls the stream flows over another lava ledge, the southern end of the fall being farthest down stream, and is there broken down into rapids, which present no serious obstacle to the ascent of the fish. This is also the case at various points along the front of the falls. The maximum vertical descent is about 20 to 25 feet. Salmon are known to go over these falls in large numbers. Indians encamp yearly on the island immediately below the falls, and spear the fish as they pass over the ripples. Well-known spawning-beds are said to be in the river about 2 miles above the falls, and salmon are known to ascend Salmon Creek, a tributary entering 2 or 3 miles higher up. A white man has been in the habit of catching salmon with a seine each year, and could obtain more than he could find market for. It seems evident, then, that a hatchery located near this point and drawing water from one of the many large spring-fed streams which enter here would have no diffi- culty in securing fish. The Lower Salmon Falls are about 6 miles below the Upper. We are informed that a man can descend this stretch of the stream in a small boat, although there are numerous shallow places and short rapids. The Lower Falls are very similar to the Upper. The river at this point falls over a lava shelf, for the most part vertically, and with a total descent of about 20 feet. The front of the falls is very wide, probably over a quarter of a mile, and runs obliquely, the northern end being farthest upstream. By far the greater part of the water falls over the southern half of the falls, so little coming over the northern part as to prevent the ascent of fish, except, perhaps, at one point. At the extreme southern end the falls are much lower. Here, and also near the center, the fish would apparently have no difficulty in ascending. To sum up what was learned about the salmon in this part of Snake River, it is certain that they visit Glen’s Ferry and the stretch of the stream between INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 27 there and a point 2 or 3 miles above Upper Salmon Falls in large numbers, and spawn mainly in the bed of the stream, some of them entering Salmon Creek, as before said. It is not known to us how far they ascend towards Auger Falls from the Upper Salmon Falls They appear late in August, and spawn in the bed of Snake River and the smaller tributaries from September on to November. Snake River below the various falls.—A short distance below Shoshone Falls, as already indicated, the valley of the Snake changes its character. The bluff recedes, leaving the valley several miles wide in places, and becomes, at the same time, less abrupt, and the iava walls are often entirely concealed by slopes of water-worn gravel and soil. The valley varies in width, but preserves this general character as far as the town of Huntington. It is along the upper part of this widened valley that the extensive springs already mentioned are found. The first of these are at Blue Lakes. Here they rise in the bottom of a lake at the base of the basaltic cliffs which forms the cafion wali. The outlet of this lake, after running ashort distance, widens into a second very deep lake, in which the water again sinks into the lava. This water, together with a much larger supply, reappears at a lower level as a very large spring, from which flows a small river of beautifully clear blue water. This finds its way among the lava bowlders down arather gentle incline to the Snake. Farther down the valley at intervals appear other similar springs. The streams that flow from these are used to irrigate the bottom lands, which are naturally covered with sagebrush and other desert vegetation. On the application of water they become very fertile, raising large crops of alfalfa and other hay, of garden vegetables, and fruit. If the supply of salmon were assured, these springs would offer model sites for a hatchery. They are located from 6 to 10 miles above Bliss, Idaho, and are reached by good roads. The most extensive of these springs empty into the lower course of the Malade or Wood River, which empties into the Snake a short distance from Bliss. (See p. 26). The long stretch of the Snake River which lies between Huntington and Lewiston was not visited by any member of the party. The stream was described to us as flowing for the greater part of this stretch through a deep cation in which were numerous rapids. A steamer once passed through this canon at high water, but arrived at Lewiston so battered and broken that none has dared attempt the passage since. No falls occur along this stretch of the stream, and there is nothing that can be considered an obstruction to salmon. But this part of the country is almost uninhabited and the river is difficult of approach. At Lewiston and below, the stream flows again through a comparatively open country, the canon walls being rounded and the slopes covered for the most part by deposits of water-worn gravel and soil. Mr. W. M. Stockton, of Glen’s Ferry, Idaho, who has resided there twenty-three years, says that the Snake River is usually highest in June, falls until the winter rains set in, and is lowest in October. Salmon caught in large numbers at Glen’s Ferry; speared. The run begins in September and lasts six weeks or two months. More numerous in former years than now, but plenty were caught last year, 1892. Indians spear them, salt and dry them for winter use. They spawn on the gravel beds in the river at and near Glen’s Ferry in water so shallow that the dorsal fins are outof the water. Knows of no obstructions in the river below Salmon Falls. Has heard that Salmon Falls is an obstruction; doesnot know so. The sturgeon are caught at all seasons of the year; more numerous in summer. Hasseen and caught salmon in Payette River and has seen them spawning there and in the Snake River on the ripples. The Boise is highest in June and lowest in October. Knows nothing definite about the redfish. Says they are a landlocked salmon. They are caught in Payette Lake and shipped to Caldwell and sold as food-fish during September. TRIBUTARIES OF SNAKE RIVER. Ross Fork of Snake River.—This is a small stream flowing into the Snake above Pocatello. It was examined on the Fort Hall Indian Reservation about 12 miles north of Pocatello. The stream there was about 15 feet wide, 10 inches deep, and had a very slow current—not over 6 inches per second, but somewhat swifter on the riffles. The water was somewhat muddy and the bottom of the stream was chiefly of mud, with gravel in some places. There was an abundance of Nostoc and other algoid vege- tation in the water, and the banks were well covered with willows and small cottonwood bushes, but no large bushes of any kind. Fishes, including trout, were abundant in this stream, and it was here that the types of a new sucker (Catostomus pocatello) were obtained. The temperature of the water at 1p. m., August 4, was 72.5°, when the air in the shade was 93°. Port Neuf River.—This stream has its rise in southeastern Idaho, on the low divide which now separates the Salt Lake Basin from that of the Upper Snake River, and flows into the Snake a few miles west of Pocatello. At Pocatello this stream averages about 30 feet wide, 6 inches deep, and 28 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. flows about 14 feet per second. There are many deep holes or pools with mud bottom, while in the shallower reaches the bottom is of gravel and the current is more swift. The banks are usually low and of clay, with occasional rocky places. The water is rather clear and cool, the temperature being 76° at noon, August 2, when that of the airin the shade was 90°. There appeared to be very little alge or other water vegetation in this stream. The banks were covered with a dense growth of willows, while back from the stream a short distance on either side are sagebrush plains. A few dead bivalves (Margaritana margaritifera) were found, but molluscous life seems to be rare in this stream. Crawfish (Astacus gambelii) were found in considerable abundance. Not many species of fish were found here. By far the most abundant species is Leuciscus hydrophlox, the next most common are the western dace (Rhinichthys cataracte dulcis) and the chub (Leuciscus lineatus). Suckers (Catostomus catostomus) and blobs (Cottus philonips) were also found, the latter in consider- able numbers. No trout were seen here, but we were informed that they are sometimes taken in the river near Pocatello, and that they are found rather plentifully further up the stream. The tempera- ture and other characters of the water are fairly suitable for trout, and no doubt plants of such fish would prove successful in this river. Mink Creek.—This is a small stream flowing into the Port Neuf about 6 miles above Pocatello. Near its mouth it averages about 6 to 8 feet wide, 2 feet deep, and has a 2-foot current. The water at the time of our visit was pretty clear and the temperature 59° at noon, August 3, when that of the air in the shade was 92°. The bed of the stream was of mud and sand in the more quiet portions and of gravel on the riffles. The banks were overhung by a heavy growth of willows. This is a typical trout stream, and we found the cut-throat trout to be quite common. About the same species of minnows and suckers which were found in the Port Neuf at Pocatello were also found here. Crawfish, toads, frogs, and mussels were also obtained here. The Port Neuf River at the mouth of Mink Creek is a clear, cooi stream with gravel and lime- deposit bottom in the shallower parts and mud and sand where deeper ani! more quiet. The same species of fishes were obtained here as elsewhere in this river. Salmon Creek.—The uppermost tributary of Snake River to which salmon have access is Salmon Creek, emptying into the river 3 or 4 miles above the Upper Salmon Falls. This was not visited by us, and little seems to be known about the general character of the stream. Mr. J. L. Fuller has seen salmon in the lower 2 or 3 miles of the stream, but does not know how far they ascend. Malade River.—The next stream is the Malade or Little Wood River, already mentioned. This was fished near Shoshone by Messrs. Thoburn and Rutter, August 5. Width, 25 feet; depth, 3 feet ; current, 2 feet; temperature at 7a. m.: air, 70°; water, 62.5°. During dry seasons the Malade becomes dry for the lower 40 or 50 miles of its course and is prevented from being a salmon stream by inaccessible falls near its mouth. As seen by us in its lower course, it runs on the surface of the country untila point about 4 miles above its month. Here it leaves the surface and enters a narrow cleft in the rocks by a succession of falls and rapids, two of which are designated the Upper and the Lower Falls. This cleft in the rock soon deepens and widens into an extensive canon, which seemed to be from 500 to 800 feet deep inits lower part. At the lower falls the stream descends vertically about 40 feet, shooting out of the canon, whichis here a mere cleft 20 to 30 feet wide, and falling into a deep pool at the bottom. As Mr. Fuller stated, it looks very much like the stream out of the spout of ateakettle. During high water the stream rises so as to obliterate these falls, and in the spring trout have no difficulty in ascending from the Snake into the Upper Malade. In autumn, however, these falls are an impassable obstacle to the salmon. It is below these falls that the large springs already referred to enter the Malade. These increase the size of the stream many times, so that even during the lowest stage otf water in autumn the Lower Malade flows full—this even at times when the upperstream is entirely dry. According to Mr. Fuller, who based his statement upon the reports of engineers, the Lower Malade at its lowest stage is a stream averaging 7 feet deep, 72 feet wide, having a current 15 miles an hour. It descends rapidly in its lower course and would offer a fine site for a hatchery. Salmon are seen as far as the base of the Lower Falls, i. e., 2 or 3 miles above its mouth. Bruneau River.—The next considerable tributary is the Bruneau, which enters from the south about opposite the town of Mountain Home. This was not visited by any member of the party. A large number of men were interviewed who were acquainted with the stream; these all agreed that it was a natural salmon stream. Mr. Fullei has seen the salmon spawning in the headwaters of the Bruneau, in October. Recently a dam has been placed in the lower course of the stream for irrigation purposes, The dam is without fishway, and salmon are now absolutely prevented from ascending. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 29 Owyhee River.—The Owyhee River is a large stream rising in the mountains of Nevada and flowing into the Snake at the boundary between Idaho and Oregon, south of Huntington, Oreg. The salmon are said to enter this in quantity. and are well known to the miners on the headwaters of the stream. This is a river of much importance, to which nearly all the streams of northern Nevada are tributary. Boise River.—Exatined August 8, one mile west of Caldwell, Idaho. Width, 200 to 400 feet; depth, 2 to 5 feet; current, 2 feet; temperature of water at 10 a. m., 66°. The bed of this stream is mostly sandy, with occasional patches of gravel. There is a riprap dam about 2 miles above Caldwell, belonging to the Howard Sebree Company. The lower slope of this dam is about 6 feet, and there is no fishway. Dr. J. B. Wright, of Caldwell, tells us that he caught salmon in the Boise, near Caldwell, in 1864, and that they were very numerous then. In 1865 placer mining began on the Upper Boise and but few salmon have been caught since. Occasionally he catches one in the upper waters of the Boise, but they are very rare. He further says that salmon trout enter this river in the spring, when the waters are high, and that he has caught them full of eggs in July in the Upper Boise. The dam already mentioned has been in five years, but he does nut think it has affected the run of fish. Dr. Wright says that the salmon run up Snake River in September, the run lasting until the middle of October. He has not noticed any decrease in recent years. We were also told that at Glen’s Ferry there is a run of salmon trout in April and May. Payette River.—Examined August 9, three-fourths of a mile southeast of Payette, Idaho, near its mouth. Average width 360 feet; depth, 3 feet; current, 14 feet; temperature of water, 63° at 5 p. m. Water clear; bottom sand and gravel. The Payette at this place is a rather shallow stream flowing rapidly over numerous shallows and much divided by gravellyislands. It flows over sand and coarse waterworn gravel. The river seems to be suitable for salmon, but no one in the vicinity seemed able to give’us any notes of value as to their occurrence. Salmon River.—This is, except the Snake, the largest and most important and is certainly the least known river of Idaho. It has its headwaters in the mountains forming the divide between Montana and Idaho, and enters the Snake where the latter is passing through its deep cation, near the northeast corner of Oregon. We were informed that the lower course of the Salmon River itself is through a deep narrow cation, which renders it difficult of access. It is claimed that salmon still ascend this stream in large numbers, and spawn in all the little creeks high in the mountains. Little, however, is known with certainty regarding the salmon or other fishes of this stream. Clearwater River.—Examined August 15 and 16 at various points from its mouth to 5 to 1% miles above Lewiston, Idaho, to the mouth of Potlatch Creek. It is there a clear, cold stream flowing over very large round bowlders. This kind of bottom makes it almost impossible to use a net, and salmon could not be obtained by this method if the stream were otherwise suitabte for a hatchery. The tem- perature of the water was 83.5° when the air was 83.5° at 4 p. m., and 63.5° when the air was 63° at 10 a.m. As in all of these larger, clear, cold streams, we found fishes very scarce. The smaller minnows and suckers could be obtained only at the rate of two or three toa haul. Fish may bemore abundant in the deeper parts of the stream, or the numbers may be kept down by the trout, which could éasily pursue the smaller fishes in the clear water. Potlatch Creek.—This is a small stream flowing into the Clearwater, near Lewiston. It was examined August 16 near its mouth. Palouse River.—This is a considerable stream rising in western Idaho and flowing westward through southeastern Washington to the Snake River north of Walla Walla about 45 miles. It was examined near Colfax, Wash., August 17. At this place the stream was quite low, being reduced to pools. Temperature of water 74°. Grande Ronde River.—This river rises in eastern Oregon, flows northeast, and joins Snake River near the forty-sixth parallel. It was examined near La Grande August 11. - According to Mr. J. B. Foley, of La Grande, salmon are very numerous in this river in September and October, coming as far as the dam 1 mile above La Grande. They try to jump this, but do not succeed. They are speared in large numbers by the Indians and boys, but are so worn and cut up by their trip up the river that they are of little value as food fish. The dam is of logs with two 4-foot steps on the lower side, and has no fishway. There are no dams below—that is, between La Grande and the Snake River. Plenty of salmon trout come in the spring in April and May. These can get over the dam in high water. Water lowest in August and September. There is placer mining in the upper parts of the river, and the water is milky. Trout are plentiful at Meacham, in the upper Grande Ronde River, and in the side streams. 30 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. Pataha River.—This stream was examined at Starbuck August 14. Temperature at 4 p. m., 68°; air, 66°. Width, 50 feet; depth, 1 foot; current, 2 feet. The Pataha is of some importance and is well supplied with the common fishes of the region. Pend d@ Oreille River.—There are two important lakes which are drained by this river—Flathead Lake, in Montana, and Lake Pend d’Oreille, which is in Idaho, near the Washington State line. Examinations were made at Flathead Lake* in 1891 by Evermann and Jenkins, and in 1892 by Gorham and Woolman, who found the ‘‘falls” in Flathead River near the outlet of Flathead Lake to consist simply of a series of rapids, which do not interfere in the least with the free movements of fish. From this point down Flathead River possesses no falls or obstructions of any kind, and there is none in Clarke Fork until near Lake Pend d’Oreille.t Not far above Lake Pend d’Oreille, in Clarke Fork proper, and near a station on the Northern Pacific called Thompson Falls, are some small rapids which are no more serious than are those in Flathead River. This is according to Dr. Gorham and Mr. Woolman. We did not deem it necessary to revisit these two places, as Dr. Gorham’s notes and the information which we gained through conversations with a number of persons who were familiar with that part of the river convinced us that there are no obstructions of anv importance above Lake Pend d’Oreille.. SSE SZ We examined this river pretty carefully from the outlet of Lake Pend d’Oreille to near its mouth, or where it joins the Columbia just across the British Columbia line. While that portion of the river above Lake Pend d’Oreille is still spoken of as Clarke Fork, the portion below Lake Pend d’Oreille is, in that region, known only as the Pend d’Oreille River. From Sand Point, Idaho, which is at the outlet of Lake Pend d’Oreille, to the Washington line is about 25 miles. In this portion of the river there is only one fall or rapid, and that is Albany Falls, sometimes known as Villard Falls or Seniaquoteen Falls. These falls are about 14 miles above the little town of Newport, Idaho. The falls are divided by a small, rocky island, npon which is built one of the piers of the railroad bridge which is used by the Great Northern in crossing the river at this place. The relative position of the bridge and the falls is shown in the above diagram. These falls are scarcely more than pretty steep rapids and would not interfere at all with the ascent of salmon. The part to the left of the islands (going down stream) is just above the bridge. “or information concerning the upper waters of this system see Evermann, in Bull. U.S. Fish Commission for 1891, pp. 1-90. tIn 1883 Mr. Livingston Stone, under the direction of the U. 8S. Commissioner of Fish and Fish- eries, made an extended exploration of Clarke Fork and the Columbia River with reference to the selection of a suitable site for a salmon-breeding station. In Mr. Stone’s interesting report (Report U.S. Fish Comm. for 1883, 237-255) is given much valuable information regarding the upper portion of Clarke Fork and the Big and Little Spokane rivers. He found, what our own inquiries confirm, that salmon never reach Lake Pend d’Oreille. but thought they were probably kept back by the falls at the mouth of the Pend d’Oreille. ; INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 31 At the time of our visit (August 9) the total descent was probably 10 feet, but as a rapid, not in a ver tical fall. During low water the descent would be somewhat greater. The fall on the right sideis of the same character and presents no greater difficulties. Just below Albany Falls the river is perhaps 1,000 feet wide and 20 to 30 feet deep in the channel. The stream was up, however, at this time, and would probably fall at least 10 feet before reaching low- water mark, according to the captain of the Dora, asmall steamer which makes irregular trips between Newport and the head of Box Canon. On August 9 we took this steamer and went down the river to Box Canon, a distance of about 60 miles, although the steamer people call it 80 miles. Throughout this distance the Pend d’Oreille is a beautiful, clear stream, with a good strong current, and varying in width from 500 to 1,000 feet. Box Canon is a narrow gorge about 14 miles long. The walls are quite close together and the river rushes through the narrow passage with a very strong current. There is, however, no fall in the canon and small boats have on several occasions been taken tarough withoutinjury. There is nothing here to stop the ascent of salmon. Metaline Falls.—These falls are just below the Metaline mining camp, or 7 miles below the foot of Box Canon. The river between Box Cafion and Metaline Falls has a good strong current, but no falls orrapids. The falls are over a ledge of limestone, through which the river has cut, and are the largest and most important of any found in this river. The total fall is perhaps as much as 30 feet, but it is in a series of rapids, there being no vertical drop at all. The stream is here inclosed between high rocky walls and is very turbulent for some distance. Salmon could probably ascend these falls with- out much difficulty. A little blasting near the left (west) wall would make it still easier for fish to getup. Just above Metaline Falls, Sullivan Creek flows into the Pend d’Oreille from the right bank. From Metaline we walked down the river about 14 miles farther, on August 10, to the head of what is known as the Big Eddy Cafion. This canon is about 3 miles long and is quite narrow, the limestone walls being so close together that in one place a fallen tree lies across from one wall to the other. The river rushes through this canon with great fury, but there are no falls, and we do not believe that the ascent of salmon would be seriously interfered with. If it should be shown that salmon can not swim against such a strong current for so great a distance, we see no easy way by which it could be made less difficult. There are some relatively quiet nooks or eddies here and there, however, in which salmon would be able to rest and we therefore do not consider Big Eddy Canon a serious obstacle to the ascentof fish. Lime Creek, a small but fine trout stream, flows into theriver at the head of this cafion. The river between Metaline Falls and Big Eddy Caton is quite swift, but contains no falls or rapids worth mentioning. The lower end of Big Eddy Canon is but a short distance from the British Columbia line, just north of which the Pend d’Oreille turns abruptly westward and runs approxi- mately parallel with the international boundary until it flows into the Columbia, a distance of about 27 miles from where it leaves the United States. We did not visit this part of the river for two reasons: (1) Dr. Gorham’s notes and Mr. Bean’s report upon the obstructions were sufficiently full to enable us to judge of its character; and (2) several persons familiar with it, and with whom we talked, _all agreed that there are no obstructions below Big Eddy Canon which are nearly as serious as Big Eddy Canon or Metaline Falls. All agree that Metaline Falls is the most serious obstruction found anywhere in the Pend d’Oreille. From Mr. Bean’s report and from our conversations with prospectors and others living along the Pend d’Oreille, it appears that there is a series of rapids near the mouth of the river and another just above the mouth of Salmon River, which empties into the Pend d’Oreille just above the Washington line. These are all said to be rapids rather than falls and probably would not interfere with the ascent of salmon in the least. From the foregoing it therefore appears that there are no serious obstcuctions in Clarke Fork of the Columbia which would prevent salmon from reaching Lake Pend d’Oreille and Flathead Lake, or other parts of that river basin. The Pend d’Oreille River is one of the most beautiful and picturesque in Agneta. It is a magnifi- cent river, probably averaging over 1,000 feet in width and being very deep throughout most of its course. In most places there is a good, strong current, becoming dangerous rapids in the narrower places. The water is clear and pure and cold—an ideal trout stream. The depth varies greatly, high water occurring in July from melting snows. Late in August or September the water is many feet lower than in July. High mountain slopes ascend abruptly from the river’s banks throughout most of its course, and these are covered with a heavy evergreen forest and a dense growth of underbrush. 32 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. In other places, as at Usk, La Claires, and Metaline, the river bottom widens out and there are many acres of excellent farming land. During high water large areas of this level land are covered by water, but when the waters subside these tracts become valuable meadow lands. Trout are abundant in this river; salmon trout are also quite abundant, and both bite readily. We know of no stream which offers finer opportunities for sport with the rod than the lower Pend @Oreille. Deer, wild geese, and ducks were also seen in considerable numbers. From the Big Canon below Metaline we were compelled to walk back to Newport, a distance of about 75 miles. As there was no trail for the greater part of this distance, except a cattle trail, which was used by cattle only later in the summer and which was now under water, we found the trip a very difficult one, attended by many hardships. We reached Newport early in the morning of August 15, where we took the train for Colville, Washington. THE UPPER COLUMBIA RIVER. The Upper Columbia River was visited only at Kettle Falls, Washington, but several of its tribu- taries were examined, notes upon which are given in the following pages. Kettle Falls, about 9 miles from Colville, Washington, are the only falls in the Upper Columbia that need mentioning in this connection. At this place the river cuts through a ledge of highly crystalline rock, the strata of which have a gentle dip upstream. A large island divides the river into two parts, as shown in the following diagram: i MIM SS Pes At the present stage of water we judged these falls to have a vertical fall of 12 to 15 feet each, but they are not of equal height throughout their entire width. The upper falls (a) was at least 14 feet vertical near the island and in the middle, but toward the right bank it seemed to be lower and less vertical. The lower fall (6) is probably 15 feet high in its highest places, but at the right shore it, too, is not so high nor so nearly vertical. At ¢ is a seething whirlpool, the water coming around the left side of the island, having to make an abrupt turnin order to get out. The upper fall is probably not of great importance in this connection, for, when salmon have once gotten above the lower falls they can go around to the right (going upstream) of the island where there are no serious obstruc- tions, but they are seen to swim up over the upper falls. George E. and Jacob A. Meyers are two intelligent and well-informed men who have lived at Kettle Falls for 23 years, and are quite familiar with the falls and their relation to the salmon. From them we obtained the following information: Up to 1878 salmon were very abundant in this part of the Columbia; “millions were seen ascending the falls every season.” The run would begin in June and continue until October, the biggest run being in the last half of August. The run toward the end of June was also large, but while there was PLATE VI. (To face page 32.) Salmon Fisheries of the Columbia River. LOWER KETTLE FALLS, COLUMBIA RIVER. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 33 a decrease in the number from then until late in the summer, some salmon were to be seen all along; so that there were not two distinct runs, but one continuous run from June to November with two periods of great numbers—June and August. The salmon caught early in the season are regarded as the best. The salmon have no trouble getting up Kettle Falls; indeed, they usually swim right up the current, seldom haying to jump out of the water. The time when it is hardest for them to get up is during a medium stage of water; it is easy at high water, asthe fall is then wiped out to some extent; it is also easy at low water, as there are eddies and pools then in which the fish can rest. Salmon formerly spawned in great numbers just below Kettle Falls. The spawning beds were toward the right side of the river on gravel bottom, usually just above a riffle. A great many spawned in the Colville River just below Meyers Falls. The Colville flows into the Columbia from the east just below Kettle Falls a short distance. Meyers Falls is in the Colville 2 or 3 miles above its mouth. The height of the lower Meyers Falls is 80 feet, that of the upper about 26; the total descent, including rapids, being about 125 feet. The width of the falls is about 150 feet. Salmon still enter Colville River and spawn on the gravel beds below Meyers Falls, but they are very rare. A fishway could be placed here which would enable salmon to ascend the Colville, which is, so far as the other features are concerned, an excellent stream for salmon and trout. The temperature of the water at the falls, August 16, was 62°. While we think the evidence shows that salmon are able to ascend the Lower Kettle Falls, the evidence that they have ever gone much, if any, farther, is not conclusive. Indeed, one of the earliest accounts of these falls which we have seen, states positively that no salmon are taken above these falls. In volume tv of the Narrative of the United States Exploring Expedition under Captain Wilkes we find the following: “‘The Kettle Falls are one of the greatest curiosities in this part ofthe country. They are formed oy a tabular bed of quartz that crosses the river, and which, being harder than the rocks, either above or below, has of course suffered less by abrasion, and thus formed a basin that renders the name appro- priate. The total descent of the water is 50 feet, though the perpendicular fall in no place exceeds 15 feet, which is, however, more than sufficient to prevent the passage of boats. At the foot of the falls the breadth of the river is 2,330 feet, and the rate of the current is4 miles an hour. This breadth is somewhat narrowed by an island, about midway of which is the first fall, which is almost entirely unbroken. Thence the river forces its way over a rocky bed until it reaches the main fall, where the water is thrown into every variety of shape and form, resembling the boiling of a kettle, from which the falls derive their name. “There is an Indian village on the banks of the great falls, inhabited by a few families, who are . ealled *‘ Quiarlpi” (Basket People), from the circumstance of their using baskets to catch their fish (salmon). The season for the salmon fishery had not yet [in June ?] arrived, so that our gentlemen did not see the manner of taking the fish; but as described to them, the fishing apparatus consists of a large wicker basket supported by long poles inserted into it and fixed in the rocks. The lower part, which is of the basket form, is joined to a broad frame, spreading above, against which the fish, in attempting to jump the falls, strike, and are thrown back into the basket. This basket, during the fishing season, is raised three times in the day (twenty-four hours), and at each haul, not untrequently, contains 300 fine fish. A division of these takes place at sunset each day, under the direction of one * of the chief men of the village, and to each family is allotted the number it may be entitled to; not only the resident Indians, but all who may be there fishing, or by accident, are equally included in the distribution. “At the lower end of the falls are large masses of quartz rock, on which the Indians dry their fish. Few of the salmon, even if able to pass the lower fall, ever get by the upper one, being generally caught between the two falls ; consequently, above this place no salmon are taken, A short distance below the Kettle Falls are the Thompson Rapids, which begin at the mouth of Mill River, and extend for some distance below that point.” This visit to Kettle Falls and eastern Washington was made by Captain Wilkes in 1841.* Spokane River.—The Spokane River has its source in Ceeur d’Alene Lake, in Kootenai County, Idaho. From the northern end of the lake the river flows approximately due west about 30 miles to * Narrative of the United States Exploring Expedition, during the years 1838, 1839, 1840, 1841, and 1842 ; by Charles Wilkes, U.S. Navy, commander of the expedition, In five volumes. Vol. Iv, pp. 444 and 445. Philadelphia, Lea & Blanchard, 1845. S. Mis. 200 3 34 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. the city of Spokane, where it turns to the northwest and flows into the Columbia, about 45 or 50 miles distant. The total length of the Spokane River probably exceeds 125 miles, as its course is extremely winding. For a considerable part of its course it Hows through vast fields of lava, into which it has cut a deep and picturesque channel. The stream is large and in most places quite deep; the bed in many places is strewn with large granite bowlders or large irregular masses of lava, which render seining next to impossible. About 6 miles below Coeur d’Alene Lake are Post Falls, which probably do not interfere with the free movement of fish. In the city of Spokane, where the river breaks through a lava flow, there are several very beautiful falls and rapids, which have been modified in various ways in utilizing the water power for milling purposes. These obstructions, natural and artificial, are impassable to fish. Salmon find no obstructions in the Lower Spokane and ascend as far as these falls. Formerly salmon were said to be abundant as far up as the falls,* but now they are seldom seen farther up than the mouth of the Little Spokane. This stream was examined in and near the city of Spokane by Mr. B. A. Bean in October, 1892, and by Profs. Evermann and Jenkins August 19 and 20, 1893. The water is clear, cold, andpure. The only contamination is that from the city of Spokane, and that does not seem to be at all serious as yet. An abundance of fish food, such as insects and their larve, small mollusks, and crawfish, was noticed in this river. Cour @ Alene Lake.—This is one of the largest and most picturesque lakesinIdaho. Itis very irreg- ular in shape, occupying, as it does, a narrow mountain valley together with its lateral ramifications. Its greatest length from north to south is probably not less than 22 miles, while its average width is less than 3 miles. The Cour d’Alene, St. Joseph, and other mountain streams are tributary to this lake, the outlet of which is Spokane River. Trout are abundant and of excellent quality in the lake, but salmon are not known to occur in it. The numerous falls in the Spokane River in the city of Spokane undoubtedly prevent the ascent of salmon to Coeur d’Alene Lake.t This lake was fished August 21 near the outlet, 14 miles west of Ceurd’Alene. Temperature of water at 4 p.m., 75°; air, 86°. Hangman Creek.—Vhis is an unimportant stream, tributary to the Spokane. It was examined in the vicinity of Tekoa, Wash., where it was found to bea small, rather filthy stream, not suitable for trout or other food-fishes, but well supplied with minnows and suckers of several species. * Regarding the salmon fishing at these falls in 1841, Capt. Wilkes has the following: “The number of Indians actually resident about the falls is 150; but during the height of the fishing season there are often nearly 1,000, consisting of all the Spokane tribe, who are generally included under the name of the Flatheads. They subsist for the most part on roots, fish, berries, and game. At the opening of the spring, in March and April, or as soon as the snow disappears, they begin to search for a root resembling the cammass, which they call pox-pox. This lasts them till the beginning of May, when it gives place to a bitter root, termed spatylou. This is a slender and white root, not unlike vermicelli in appearance, and when boiled it dissolves into a white jelly, like arrow- root. It has a bitter but not disagreeable flavor, and is remarkable for growing in gravelly soils where nothing else will thrive. In June the itzwa, or cammass, comes in season, and is found in greater quantities than the others all over the country, particularly in the meadow grounds. This root was thought by many of us to have the taste of boiled chestnuts. Before this fails the salmon make their appearance, and during the summer months the Indians enjoy a very plentiful supply of food. While the men are employed fishing, the women are busy digging the cammass, which may be termed the principal occupation of the two sexes. They devote a portion of their time to the collection of berries, a work which is principaily the duty of the younger part of the tribes. “Tn September and October the salmon still claim their attention, although they are, after having deposited their roes, quite exhausted and about to perish, yet these are dried for their winter consump- tion, and unless they had recourse to these much want would ensue, which is always the case if the ~ salmon should be searce.” +The Indian legend given in Wilkes’ Narrative, vol. Iv, p. 449, is interesting, in that it shows that the falls at Spokane have always been regarded by the Indians as a barrier to the ascent of salmon to Ceeur d’Alene Lake. “They have, in common with the other tribes, many traditions connected with the rivers and remarkable features of their country. In these the prairie wolf bears always aconspicuous part. This wolf was not an object of worship, but was supposed to be endowed with supernatural powers, and to exert them in many ways. On one occasion it is related that the wolf was desirous of having a wife, and visited the tribes on the Spokane for that purpose, demanding a young woman in marriage. This request being granted, he promised that the salmon should be abundant, and for this purpose he raised the rapids, that they might be caught with facility. After he had been gratified in this first instance he made the same request of the others, among them of the Sketsui (Cur d’Alene) tribe, who were the only ones to refuse. He thereupon formed the great falls of the Spokane, which have ever since prevented the fish from ascending to their territory.” Salmon Fisheries of the Columbia River. (To face page 34.) PiaTte VII. CANON OF SPOKANE RIVER, THREE MILES BELOW SPOKANE, WASHINGTON. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 35 Little Spokane River.—This was visited by Mr. Barton A. Bean in 1892, and by us August 18, 1893. The Little Spokane rises among the low hills in a system of small streams and lakes in Stevens County, Wash., only 4 or 5 miles from the Pend d’Oreille River. Fed, as it is, by numerous springs, its water is very clear and cold. It flows through a narrow, fertile valley, the low meadows bordering it haying a black loamy soil. The immediate banks are for the most part covered with a network of brushes. High hills rise on either side of the valley and they are sparsely covered with pines. Such trees as cottonwood, maples, and alders are common along the banks. At Dart’s mill, where the Little Spokane was examined by us, it averages about 40 feet wide, 20 inches deep (on the ripples), and had a current of 24+ feet per second. The temperature at 2 p.m., August 18, was 63°. The bottom there was of coarse gravel in most places. Just above the dam the water was, of course, deeper and the bottom is of sand and soft mud. Here we found such water vegetation as Ranunculus aquatilus trichophyllus, and Myriophyllum in abundance. A single species of Unionidx, Margaritana margaritifera was not uncommon at this place. Fishes were also rather abun- dant, some 8 or 10 species being obtained. The Little Spokane is an excellent salmon and trout stream, as is fully evidenced by the great abundance of salmonoid fishes which we found. The cut-throat trout was abundant, as were also young whitefish. Large whitefish (Coregonus williamsoni) were seen at the dam at the mill, where Indians were spearing them with fair success. Salmon are said to enter the Little Spokane in considerable numbers even yet, but much less abundantly than formerly. The dam at Dart’s mill interferes with their farther ascent and a fishway should be putin. Salmon were quite abundant in this stream in 1882, as reported by Mr. Lane C. Gilliam, of Spokane, to Mr. Livingston Stone. * Mr. Gilliam says: “JT have just completed my second trip to the Little Spokane, and as yet no salmon to speak of are running. The Indians, who are encamped here in great numbers, anticipating a large run, are uneasy and fear the fish are not coming. Yesterday morning they caught eight, which was the largest number taken at any one time as yet. A white man living in the neighborhood told me that last year he made a rough estimate of the salmon taken by the Indians. He thinks they had between 40,000 and 50,000 drying at one time, about October 1. I will make another investigation about Oetober 1.” In the same letter reporting this information to Prof. Baird, Mr. Stone says: “The result of my researches on the Snake River are that no salmon ascend as high asthe crossing of the Utah and Northern Railroad, and that there are no salmon as high as the foot of the American Falls on the Oregon Short Line. Thesalmon probably can not get overShoshone Falls. Inthespawn- ing season there are a great many salmon at the foot of these falls, 27 miles from the Oregon Short Line Railroad.” It should be added that the character of this streamis being materially changed by the advent of civilization, a fact which is, or has been, true of most streams of this country. The cutting away of the timber and brush on the immediate banks and the cultivation of the land within the drainage area of the stream have greatly increased the surface erosion and, in consequence, the impurities of the stream. LOWER COLUMBIA RIVER. Very little work was done by us on the Lower Columbia. Some fishing was done August 22 at Pasco, near the railroad bridge 1 mile east of town, where we made twelve hauls on sand and gravel bars on both sides of the river in water from 1 to 5 feet deep. Took very few fish. Water very clear and cold. Rocks nearly free from alge. Mr. John E. Gantenheim, an educated and intelligent fish- erman of Pasco, says that he fishesevery year at the mouth of the Snake and Yakima rivers. The salmon bite readily at a spoon and are in good condition for eating. They are caught by trelling only, and bite greedily, even when full of eggs. Theirstomachs are alwaysempty. ‘hey spawn on the rip- ples near the mouths of the Snake and the Yakima rivers. Mr. Gantenheim caught his first salmon for this season on August 20. It was the first he knew of as being caught this year. It was a silver- side (O. kiswtch?), and he took it from the Columbia River near the mouth of the Yakima. We saw three salmon while at the river. Mr.Gantenheim says that the salmon begin their run about the 20th of August, are at their best during September, and last until the high water in October. The last ones are spent and not good eating. He calls the ones he catches silversides and chinooks. Does not know of other forms. Though many fish are caught, none are shipped to the canneries because of railroad charges. It is probable that some of the salmon which are caught by trolling are steelheads, *Bull. U. S. Fish Com. for 1883. 36 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. Walla Walla River.—This is a river of some importance flowing into the Columbia at the town of Wallula, about 30 miles west of Walla Walla. It was examined August 23, at Wallula, below the railroad bridge. It is here a good-sized stream, 3 to 8 feet deep in the channel, and has a velocity of about one-half foot per second. Temperature at noon, 70°; air, 80°. The bed of the stream was of soft mud, with an abundance of Chara and other vegetation in places, and the water was rather muddy. At this place Messrs. Thoburn and Rutter obtained the only specimens of Columbia trans- montana that were secured by any of us. Mill Creek.—This is a small stream, tributary to the Walla Walla River near Walla Walla. It was examined August 14 south of Walla Walla one-half mile. Width, 12 feet; depth, 10 inches; cur- rent, 1} feet. Temperature at 8:30 a. m., 56°; air, 73°. The bottom here is of coarse gravel. We could not learn that salmon are ever taken in this stream. Umatilla River.—The Umatilla River was examined August 23 near its mouth, and on August 12 near Pendleton, Oreg. At Pendleton it had an average width of 25 feet, depth of 14 inches, and a velocity of 1 foot. Temperature at 1la.m.,70°. The bottom was of coarse gravel covered with alge, and the water wasclear. Mr. Smith, of the Commercial Stables at Pendleton, says that no salmon come as far up the river as Pendleton. He has never known any salmon to occur there. They probably occur in the lower part of the stream, but we could get no reliable information upon the matter. Des Chutes River.—This is a southern tributary of the Columbia, into which it flows at a distance of 10 or 15 miles above The Dalles. It was visited August 24. The falls of Des Chutes River, located near its mouth, are about 30 feet high in low water; in high water a series of rapids. In the Des Chutes the difference between high and low water marks varies from 40 to 90 feet, according to the width of the river. The highest water is about June 20, the lowest during the coldest part of the winter. Salmon usually find the falls no obstruction. Numerous salmon are said to run up the John Day River. They are caught in large numbers by the Indians, but we find no authentic information concerning their spawning. Yakima River.—The Yakima is a good-sized stream, rising in numerous lakes near Snoqualmie Pass, southeast of Seattle about 50 miles, and flowing southeast about 150 miles to Pasco, where it joins the Columbia. At Ellensburg and North Yakima, where this river was visited by Dr. Jenkins, it runs through a broad, fertile valley, and its waters are extensively used for irrigation purposes. At Ellensburg the stream is about 160 feet wide and 10 feet deep, and flows about 1 foot per second. The water is clear and cold; its temperature at 9 a. m., August 24, was 60°. At North Yakima the stream is very clear and flows with a rapid current through an open valley, over gravel and sand, and had a temperature of 64°. The Yakima has many important tributaries, in all of which trout are said to abound. Wilson Creek near Ellensburg had an average width of about 18 feet, depth of 18 inches, and a current of 2 feet per second. Manistash Creek empties in on the right bank of the Yakima near Ellensburg. For a few miles above its mouth nearly all the water is taken out for irrigation purposes. Six miles from Ellensburg it comes through a cation into the valley. At this point it is a fine stream, abounding in trout. Itis here 25 feet wide, with a velocity of about 3 feet per second. The water is clear and excellent. The temperature at 11 a.m. was 55°. Below this point, about 2 miles from its mouth, where most of the water is taken ou¢ for irrigation, the stream was about 6 feet wide, with an average depth of 6 inches and a velocity of one-half foot per second. Temperature, 58° at 9:45 a. m. The Yakima was visited also at Prossen, at which point there is a low fall of some 3 or 4 feet, with a long gentle ripple above it. The fall would form no obstacle to the ascent of salmon unless at time of very low water. The temperature was 70° at 10a.m. At North Yakima the Yakima receives one of its principal affluents from the west. This is the Natchess River, which takes its rise among the snowfields of Mount Ranier and Cowlitz Pass. This is a clear, cold stream, admirably suited to trout. In its lower course such common species as the chisel-mouth (Acrocheilus alutaceus), Agosia nubila, and Pantosteus jordani were found. Those acquainted with the facts state that formerly, up to about 1885, salmon of three or four kinds, including the quinnat, ran up the Yakima River to this valley and spawned in the river in great numbers. At present very few make their appearance. Cowlitz River.—The Cowlitz River, made famous in Dr. Jordan’s delightfully interesting ‘‘ Story of a Salmon,” has its sources in the snowfields on the west slopes of Mount Ranier, and flows through the densely wooded country west of the Cascades for more than 100 miles before it joins the Columbia. This region is very moist and is little suited to agriculture, and the stream will never be needed for irrigation. The Cowlitz was visited by us at Castle Rock. It is there a very deep, sluggish stream, INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 37 extensively used for rafting lumber. It had a temperature of about 60°, August 25. Salmon ascend the stream in large numbers to and above Castle Rock. They make their appearance in the fall about the first of September, and are caught by the ton at Castle Rock and at numerous points below. We were informed that two kinds of salmon are taken, quinnat and the silversalmon. The quinnat makes its appearance first, and is, according to reports, obtained in great numbers. We do not know how safely one may rely upon the reports of the fishermen, however. Toutle River.—Toutle River is a fine, clear stream entering the Cowlitz from the east, about 4 miles above Castle Rock. It was visited by us 2 miles above its mouth. At that point it was about 100 feet wide, averaging perhaps 1 to 14 feet deep, and was flowing rapidly over rounded bowlders and stones of small size. Its current was perhaps 1} or 2 feet per second. At 11 a.m. the temperature was 594°. The stream flows through a very sparsely inhabited country. A few miners and a larger number of lumbermen live on its upper course. It flows everywhere through a dense fir forest, in which are some deciduous trees. All agree that the salmon ascend this stream yearly in large numbers. In addition to the investigations which were inade in the Columbia River basin, some little work was done on streams tributary to Puget Sound, or which flow directly into the Pacific. Drs. Gilbert and Jenkins examined Newaukum and Skookumchuck rivers, and in June, 1892, Prof. Evermann spent parts of two days examining Lake Washington at Seattle, and the Snoqualmie River in the vicinity of Snoqualmie Falls. NEWAUKUM RIVER. This stream is a small tributary of Chehalis River, into which it flows near the town of Chehalis. It was visited near its mouth August 27. SKOOKUMCHUCK RIVER. This river rises on the divide near the headwaters of the Newaukum, and, flowing to the north- west, empties into an arm of Puget Sound near old Fort Steilacoom. The Newaukum and the Skookumchuck are both interesting as having furnished us many specimens of young dog salmon. They were found in both of these streams in abundance and were evidently the young of the preceding year. LAKE WASHINGTON. This lake is a magnificent body of fresh water, extending for more than 20 miles north and south, just east of Seattle. Some collecting was done here on June 25, 1892. Nothing was found, however, except two or three species of Cyprinid@ and a number of blobs. SNOQUALMIE RIVER. This riverrises near Yakima and Snoqualmie passes and, flowing westward, joins the Snohomish, which in turn flows into the Sound. The Snoqualmie was visited June 26 and 27, 1892, and a small collection of fishes obtained. At the falls this river was 150 to 200 feet wide and about 6 feet deep, entirely too deep for seining, only occasional shallow places being found where the seine could be drawn. At Snoqualmie Falls the river descends 268 feet in asingle plunge. Trout, however, are abundant both above and below the falls. The only other species obtained were a few minnows and suckers. We were unable to secure any reliable information as to the occurrence of salmon in Sno- qualmie River or in Lake Washington. 38 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. NOTES ON THE FISHES OF THE COLUMBIA RIVER BASIN, WITH DESCRIPTIONS OF FOUR NEW SPECIES. In the following notes on the fishes of the Columbia River basin we have included not only those collected by us, but also the small collections made by Messrs. Bean and Woolman in 1892, and the few species obtained in Newaukum and Skookum- chuck rivers by Drs. Gilbert and Jenkins, and in Lake Washington and Snoqualmie River in 1892. The exact status of several of the species of Salmonide, as well as some of the minnows and suckers and all the Cottide of this region, is a matter which will require much additional investigation to determine. Most of the forms which have been regarded as good species are but poorly differentiated. The range of variation seems to be very great, and characters which are of undoubted specific value when applied to Atlantic-drainage species, do not possess any such value for classification of Pacific coast fishes. Each so-called species seems to be in a very unstable state of equilib- rium, and not to have yet assumed or been able to retain with any degree of permanence any set of specific characters. Thisis particularly true of the species of Agosia, Catos- tomus, Salmo, and, possibly, Oncorhynchus. In sequence of species in this paper we follow Jordan’s Catalogue of Fishes of North America, 1885. 1. Entosphenus tridentatus (Gairdner). Three-toothed Lamprey. Petromyzon tridentatus Gairdner ms., Richardson, Fauna Boreali-Americana, 293, 1836. Type locality: Falls of Walamet (Willamette) River. Petromyzon lividus Girard, P. R. R. Survey, 379, 1858. Type locality: Wahlahmath (Willamette) River, Oregon. Petronyzon astori Girard, loc. cit., 380. Type locality: Astoria, Oregon. This lamprey was first seen by us at Lower Salmon Falls, on Snake River, on August 8. Over 40 specimens were here found dead on a sand bar below the falls. They had probably died the night before, and had been deposited on the spit, where buzzards were busily feasting on them when we arrived. We were informed that the lampreys in their upward migration reach this point in the river sometime during July, after the water has begun to go down. They are said to make good sturgeon bait, and can be best caught in the evening or in the early morning, when they are found clinging to the rocks at the falls. On August 11, a large number of decayed specimens was found on the banks of the Umatilla River at its mouth. They were high up on the banks, and had apparently died and drifted ashore several weeks before, atatime when the river was higher. They ascend the Umatilla, and are caught by the Indians for food. One dead specimen wasseenat Pendleton. A numberof larve, 14 to2 inches long, were taken from débris in the bottom ofa pool in the Natchess River at North Yakima. The lampreys are well known to the owners of salmon-wheels on the Lower Snake and the Columbia, and are universally called eels. At Lewiston, we learned that the lampreys begin their run very early, being already in the stream when the salmon-wheel is first put in place in the spring. They are occasionally caught by these wheels in such numbers as to fill the beat, and are said to be valuable for the oil they contain. This lamprey was seen also by Dr. Eigenmann, at La Grande and Caldwell, in 1892. 2. Acipenser transmontanus Richardson. Oolwmbia River Sturgeon. Acipenser transmontanus Richardson, Fauna Boreali-Americana, 111, 278, 1836. Type locality: Columbia River at Fort Vancouver. The sturgeon ascends the Snake River to above the Upper Salmon Falls, between which and Auger Falls it is frequently taken. We were unable to learn that they passed the Auger Falls, which apparently serve as a barrier to both sturgeon and salmon. We are informed by numerous fishermen that the sturgeon are in the river throughout the year, and can be taken at any season. They are found at Glen’s Ferry throughout the year, and we were told of individuals taken there weighing as much as 600 to 800 pounds. No definite information as to their spawning season could be secured. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 39 3. Pantosteus jordani Evermann. Pantosteus jordani Evermann, Bull. U. 8. Fish. Comm. for 1892, January 27, 1893,51. Type locality: Red Rock River, Red Rock, Montana. Pantosteus columbianus Eigenmann & HKigenmann, American Naturalist, February 4, 1893, 151. Type locality: Boise River, Caldwell, Idaho. Recent explorations of the Fish Commission have shown this sucker to be an abundant species in the region about the Black Hills in South Dakota and Wyoming. Dr. Eigenmann was the first to obtainit in the Columbia Basin, he having found it at Caldwell, Idaho, in 1892. During our investigations we found this to be an abundant and widely distributed species in the Columbia Basin. Specimens were obtained by us at the following places: Snake River at Idaho Falls, 1; Ross Fork near Pocatello, 49; Boise River at Caldwell, 4; Payette River at Payette, 13; Umatilla River at Pendleton, 3; Columbia River at Umatilla, 1; Natchess River near North Yakima, 9. A comparison of these specimens with a large series from various places in the Missouri River Basin shows them to be specifically identical. Young examples from Payette, Caldwell, and elsewhere, agree perfectly with Dr. Eigenmann’s description of P. columbianus. The dorsal rays vary from 10 to 13; the scales from 82 to 107. 4. Catostomus catostomus (Forster). Cyprinus catostomus Forster, Philos. Trans., 1773, 155. Type locality: Streams about Hudson Bay. Specimens from Little Wood River, Shoshone, Idaho, 25; Ross Fork near Pocatello, Idaho, 10; Payette River, Payette, Idaho, 2; Cour d’Alene Lake, Cceur d’Alene, Idaho, 7; Umatilla River, Pendleton, Oreg., 4; Columbia River, Umatilla, Oreg., 1; Pataha River, Starbuck, Wash., 3; Mill Creek, Walla Walla, 81; Creek at Sand Point, Idaho, 38. D. 11 or 12; scales, 90 to 104. This species differs from latipinnis, griseus, and catostomus (Evermann; EKigenmann) in its thin and rather narrow lower lip, which is incised for but little over half its depth. Two well- separated series of large papillew cross the lip between base of incision and sheath. 5. Catostomus pocatello sp. nov. Moo-gad-ee of the Fort HallIndians. (Pl. 1x.) Type locality: Ross Fork of Snake River near Pocatello, Idaho, where 18 specimens were collected August 4, 1893. Type, No. 45385, U. S. Nat. Mus. Co-types, No. 45386, U. S. Nat. Mus., and Nos. 1135 to 1141, Museum Leland Stanford Junior University. Related to Catostomus catostomus (Forster). Description: Head, 4; depth, 5; eye, 44; snout, 24; interorbital width, 22; D. 10; A. 7; scales, 19-95-14, about 50 before the dorsal. Body moderately stout; head heavy; snout not very pointed; eye rather large—larger than in any related species, its diameter 24 in snout or 24 in interorbital width; eye placed high; middle of pupil a little nearer posterior edge o opercle than to tip of snout. Mouth narrow; upper lip rather thick, but not pendent, with three definite rows of papillz ; lower lip incised nearly to base, a single series of small papille between sheath and base of incision; lobes of lower lip short and rounded; cartilaginous sheath of lips rather strongly developed. Scales small, crowded, and very mech reduced in size on anterior part of body; lateral line imperfect. Origin of dorsal fin midway between tip of snout and base of caudal rays; greatest height of dorsal fin 14 in head, its free edge very slightly concave. Heightof anal alittle greater than that of dorsal, 1; in head; pointed, reaching base of caudal fin. Pectoral about equal to anal; ventral 1g in head. Peritoneum silvery, with dark punctulations. Air-bladder large. Color in alcohol, dark olivaceous above, and on sides to below lateral line somewhat mottled with darker; under parts pale. Length, 150 millimeters. An examination of the series of eighteen specimens shows some variation. Head, 3% to 4; depth, 5 to 54 3 eye, 44 to 5-4 in young; snout, 2} to 23-23 in young. The numberof dorsal rays is usually 10, but in one example there are but 9. There is considerable variationin number of scales in the lateral line, the number in eleven examples counted being 90, 93, 93, 95, 96, 96, 100, 101, 105, 106, 107, and 108, respectively; the lateral line is frequently irregular and imper- fectly developed. From Catostomus catostomus, which this species resembles, it differs in its larger eye, fuller lower lip, and somewhat larger head. These characters may all prove unreliable, however. From C. griseus and (. latipinnis of the same size it differs in its narrower upper lip and larger eye, a3 well as in other minor characters. 40 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. This species was found only in Ross ork just above the Fort Hall Indian Agency. It does not seem to be very common, asa day’s collecting in this stream resulted in taking only 18 specimens of the species. It apparently does not attain a length of more than 6 to 8 inches, The Indian name Moo-gad-ee means sucker, or that which sucks. 6. Catostomus macrocheilus Girard. Catostomus macrocheilus Girard, Proc. Acad. Nat. Sci. Phila. 1856, 175. Type locality: Astoria, Oregon. Specimens obtained from Payette River at Payette, 5; Boise River at Caldwell, 17; Clearwater Creek at Lewiston, 2; Hangman Creek at Tekoa, 6; Hangman Creek at Spo- kane, 1; Pataha River at Starbuck, 7; Walla Walla River at Wallula, 5; Colville River near Colville, 10; Umatilla River at Pendleton, 2; Snake River, at Payette, 2; Columbia River at Umatilla, 1; Skookumehuck River near Centralia, 7; Post Creek, St. Ignatius Mission, Mont., 1; Pend d’Oreille River, Newport, Idaho, 19. This is the common sucker of the Columbia and Lower Snake rivers, and large numbers were frequently seen feeding in the shallow waters along shore. In Y5 specimens the dorsal rays were as follows: Thirteen rays in 1 specimen, 14 in 15, 15 in 8, 16in 1. Scales 67 to 70. The four specimens reported by Eigenmann from Idaho Falls are more likely referable to C. ardens; C. macrocheilus probably does not occur in the Upper Snake. 7. Catostomus ardens Jordan & Gilbert. Catostomus ardens Jordan & Gilbert, Proc. U. S. Nat. Mus, 1880, 464. Type locality: Utah Lake, Proyo, Utah. Six specimens from Mink Creek, near Pocatello, are identified with this species. No adults of C. ardens were obtained, and the status of ardens and macrocheilus in the Columbia can not be determined until a larger series is available for comparison. In all suckers of this type thus far taken from Snake River above the falls, including those from President Camp and from Heart Lake, the dorsal fin is small, containing but 11, 12, or 13 rays; and the caudal peduncle is thicker than in specimens of macrocheilus of equal size. Measurements of our specimens are given in the following table: Length in inches. 8. Acrocheilus alutaceus Agassiz & Pickering. ‘‘ Chisel-mouth.” Acrocheilus alutaceus Agassiz & Pickering, Amer. Jour. Sci. and Arts, 1855, 99. Type localities: Falls of the Willamette and in Walla Walla River. : Specimens obtained from Payette River at Payette, 53; Pataha Creek at Starbuck, 1; Umatilla River at Pendleton, 15; Natchess River at North Yakima, 2; Walla Walla River at Wallula, Wash., 1; Columbia River at Umatilla, 26; Potlatch Creek, 2 miles above mouth, 19; Snake River at Payette, 17; Boise River at Caldwell, 5. So far as known this species is confined to the Columbia River basin, where it is one of the most abundant and most widely distributed of the minnows. It has not yet been found in Snake River above the falls, nor is it known from the Pend d’Oreille basin. 9. Rhinichthys cataracte dulcis (Girard). Western Dace; Mot-to-nwt-se of the Fort Hall Indians. Argyreus dulcis Girard, Proc. Acad. Nat. Sci. Phila. 1856, 185. Type locality: Sweetwater River, Nebraska. This widely distributed species is represented in the collection by the following: Mouth of Colville River, 1; Snake River at Idaho Falls, 2; Ross Fork near Pocatello, 64; Little Wood River near Shoshone, 9; C@ur d’Alene Lake, 14; Columbia River at Pasco, 3; Natchess River at North Yakima, 11; Post Creek, St. Ignatius Mission, Mont., 6; Clarke Fork at Thompson Falls, Montana, 2. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 41 This species has hitherto been reported from the Columbia River basin from but one place—Snake River, at President Camp;* it seems, however, to be a pretty common fish throughout that basin. It was obtained by Woolman and Bean in Post Creek and at Thomp- son Falls, the only places in the Pend d’Oreille system where it has yet been found. The Indian name refers to the motion of the nose in eating. 10. Agosia nubila (Girard). Argyreus nubilus Girard, Proc. Acad. Nat. Sci. Phila.1856, 186. Type locality: Fort Steilacoom, Washington. Since the original description of this species no specimens have until now been taken from near the type locality. ‘The name has been recently used by Jordan and others for the Agosia inhabiting the Upper Snake River and the Great Basin in Utah, being thus considered Synonymous with the numerous nominal species (carringtonii, vulnerata, rhinichthyoides, hen- shavii, and novemradiata) described by Cope from streams tributary to Great Salt Lake. As this identification has been based upon a comparison with the imperfectly preserved types of nubila, the present collection is of great interest, containing, as it does, material from 15 localities, distributed between the Newaukum River in western Washington and the tribu- taries of the Upper Snake River in southeastern Idaho. A study of this material has shown the desirability of recognizing as a distinct subspecies Agosia nubila carringtonit, the form found in the Great Basin and the Upper Snake River. Examination of the annexed tabular statement will show the astonishing amount of variation which this species exhibits. Thus, the crosswise series of scales varies from 47 to 70 in number; the barbel is present or absent; the pharyngeal teeth vary from 1, 4-4, 0 to 2, 4-4, 2; and the dorsal fin varies much in position and somewhat in size. These characters occur in various combinations, and with some of these are often correlated peculiarities of physiognomy and general appearance, all of which may serve to put a certain stamp upon the individuals from a single stream, or even from one locality ina stream. Disregarding such local variations, we find that our material, exclusive of the specimens of 4. nubila carringtonii, falls more or less clearly into three groups, distributed around certain geographical centers. Whether we are here dealing with subspecies seems doubtful, and can be determined only by more extensive and detailed exploration. The first of these forms, typical nubila, is represented in our collection by a large number of specimens from the Newaukum and Skookumchuck rivers in western Washington, very near the type locality of the species. These are all very dark in coloration, and have a jet-black lateral band which extends along sides of head and encircles the snout. This band is absent in our second and third groups, found east of the Cascades, or it is at most only faintly indicated. The darker coloration of the coastwise form may be due to its inhabiting a densely forested area, possessing different climatic conditions from those characterizing the dry semidesert of eastern Washington and western Idaho. Both the typical nubila and the lighter interior form which centers about Umatilla are characterized by their coarse scales (averaging 54 along the lateral line) and their peculiar markings. The latter are due to the fact that numerous scattered scales along the back and sides are of a dark slate color, contrasting sharply with the lighter ground. The third group centers in the Spokane region, and is characterized by smaller scales, the less-marked peculiarities of coloration, and the almost uniform absence of the maxillary barbel. The inconstancy of this important generic character within the limits of the species has been heretofore noticed only by Cope, who in notes on Apocope vulnerata t calls attention to its occasional absence. in our specimens from other than the Spokane district the barbel is very rarely lacking. The significance of the groups above outlined can be determined satisfactorily only by the study of a much more extensive series than that on which this paper is based. An open waterway exists between them, and it is useless to attempt to indicate their value while so large a part of the Columbia and adjacent basins remain unexplored. 11. Agosia nubila carringtonii (Cope). Mo-sha-pog-gee. Apocope carringtonit Cope, Hayden’s Fifth Annual Report, 1871 (1872), 472. Type locality: Warm Springs, Utah. *Evermann, Bull. U. S. Fish Comm. for 1891 (1892), 42. + Cope, Zool. Wheeler’s Survey W. 100th Merid., 647, 1876. 42 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. We include under this name the Agosia of the Great Salt Lake basin (exclusive of the Sevier River) and of the Upper Snake River. Our collection contains specimens from Port Neuf River, Mink Creek, and Ross Fork near Pocatello, and from Little Wood River at Shoshone, It differs from nubila in its finer scales (average about 65) and in the absence of the sharply marked blackish scales on the sides. The Fort Hall Indians, to whom we showed specimens of this minnow, called it Mo-sha-pog-gee, a word which they say describes its eating habits. 12. Agosia umatilla sp. noy. (Pl. 1x.) Type locality: Columbia River at Umatilla, Gleeen where 15 specimens were collected, August 11, 1893, by Messrs. Gilbertand Rutter. Type, No. 45390, U.S. Nat. Mus. Co-types, No. 45391, U. S. Nat. Mus., and Nos. 1142 to 1147 (Umatilla) and 1148 to 1150 (Payette) Museum Leland Stanford Junior University. Associate type locality: Payette River at Payette, Idaho, where 3 specimens were secured, August 9, 1893, by Messrs. Gilbert, Thoburn, and Rutter. Related to Agosia falcata and Agosia nubila carringtonii. Description: Head, 33; depth, 4%; eye, 4; snout, 3. D.1,9; A.1,7; scales, 14-68-8, about 30 before the dorsal. Teeth, 1,4-4,1 hooked. The body is rather slender, the back somewhat elevated; head pointed, narrow; mouth inferior, nearly horizontal, narrow; caudal peduncle compressed, slender, its least depth 2 in head. Origin of dorsal fin slightly behind insertion of ventrals and about midway between base of middle caudal rays and nostril; dorsal fin faleate, the anterior rays nearly as long as head, their tips reaching well behind posterior rays when deflexed; rudimentary ray not enlarged nor spinelike; anal strongly falcate, the anterior rays much produced, about as long as head and more than twice the length of the posterior rays; pectoral not quite reaching ventrals, 1} in head; ventrals reaching well beyond front of anal, 14 in head; caudal deeply forked; no ventral stays. Barbel minute; upper lip without frenum. Color in alcohol, olivaceous above, covered with obscure patches of darker; sides with a distinct plumbeous band following the course of the lateral line and extending forward through the eye and around snout; side with a number of dark blotches, usually imperfectly defined, partly covering the plumbeous band; a large dark blotch on base of caudal peduncle, and one or two smaller ones on base of caudal rays. Length, 65 mm. The range of variation, as shown by the co-types, is not great. Depth, 44 to 5; eye, 34 to 4; scales, 13 or 14, 60 to 70-7 or 8: D. 80r¥. ‘There are slight but unimportant color differ- ences. (For variation in measurements, see table.) This species is somewhat intermediate between Agosia falcata and A. nubila carringtonii. From the former it may be distinguished by its notably smaller scales, absence of ventral stays, and smaller eye; from the latter it differs in its longer, more slender snout and larger, strongly faleate fins. As in A. falcata, the top of head and anterior portion of trunk are often covered with minute nuptial tubercles. From Agosia adobe it differs in the larger eye, which is con- tained 14 to 14 times in snout, while in 4. adobe it is contained 2 to 2} times; the fins are higher and more faleate, and the scales below lateral line are larger. The 3 specimens obtained from Payette River do not differ materially from those found at Umatilla. 13. Agosia falcata Eigenmann & Higenmann. Ayosia falcata Kigenmann & Higenmann, American Naturalist for February, 1893, 153. Type locality: Boise River, Caldwell, Idaho. This interesting species was obtained by us at the following places: Boise River at Caldwell, 97; Payette River at Payette, 27; Snake River at Payette, 1; Mill Creek near Walla Walla, 1; Columbia River at Umatilla, 55; Columbia River at Pasco, 5. In the following table we give measurements of a number of specimens: | Boise River at Caldwell. Payette River at Payette. | : [ae ae lon | | Head. |4 [4 | 88] 88) 4 | 4 | 38 | 32 | 32 | 38 | 88 | 36 | 38 | 4 |\4 | 38 | 38 Depth |44)4¢/5 |52/5 15 | 4¢]4¢)5 | 4 | 3h] ay] 5 [5 | 4g 15 |5 Ry eueeeee .-| 34 | 4 | 34 | 34 | 36] 34 | 84] 33] 4 | 34] Bh] BR | 42 | 3g | Ba | 2g | 8g Smo cist Ges. seat eee 22 25 | 98/3 13 2s) 25 13 SEN I alee 38 | 8 Dorsal. eee 9 |9 |9 /9 fo |9 | 9/9 |2 3 9 | ‘Amal! -- = cape NT NO ON eee We Scales soe) 375-0 manatees 56 |56 (53 |53 [55 |55 |55 54 be ie \59 | | INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 43 One of the most characteristic marks of this species is the presence of two or three mem- branous stays connecting the inner ventral rays with the skin of the body, thus forming pockets under the ventral fins and holding them down quite firmly. Adults show an extraordinary development of the nuptial tubercles, which are present on top of head, and on back and sides of body. On the body a single tubercle is located on the middle of the free edge of each scale, being formed by a thickening of the integument. On the belly, this thickening involves the entire surface of the scales, giving to this region a mosaic-like appearance. Beneath this thickened epidermis, the scales are often partially absorbed, especially on the breast. Tubercles are also present on the inner (superior) surfaces of the pectoral fins, where they follow the fin rays in single series, branching to correspond with the forking of the ray. We find the origin of the dorsal fin in this species constantly behind the frontinsertion of the ventrals. It varies from midway between base of median caudal rays and nostrils (its usual position) to a point midway between caudal and posterior margin of pupil. The variation includes the position of the dorsal assigned as a distinguishing feature in Agosia shuswap Eigenmann, this being the only character assigned as distinguishing shuswap from falcata.* Table showing variation in species of Agosia. oa Hs : S oe C) D a g| 4 ole Te ed "Hea |S Sis dS ae E o= | 28 | 2 A 2 | Ss ep Be ; a. ae 1/35 | ss eceuty S| Soe | ge beam & aa ¢ og | = | Sa A or dat |) oe é a 3 : oo aca) a | She] = |Stol. 2 E = al & |e Sei Se a = >) ° mH es) = = a = Om a) -o © 2 p> a | 3 S 5) 3 al § ba] o 5 ss A i) nm | a A im qq) n |42/4 4 Agosia nubila. Colville River, Meyers | 33-4 | 33-4 4 3/0 1,44,1 8 | Toeye........|-.- 52-57 55 10 6 Falls. Little Spokane River, 4/4 43) 44-5 3/0 2,4-4,2 | 8or9]| To preopercle.|-..| 52-63 62 17 9 Dart’'s Mill. Hangman Cr.,Tekoa -.| 33-4 | 4 -44 (ey locsoas +inl| 1,44,2} 8o0r9 | Beyond pre- |---| 51-65 57 50 10 2, 4-4, 2 opercle Coeur d’ Alene Lake ..-.|.-.-.-|.--...]-.----|------ Oin34| 1,44,2] 8o0r9} Middle of |..-| 64-70 66 44 9 +in 10 pupil. Clearwater R., Lewis- |.-----|.--.--|------|------|.---.-..|.--..-.--- GORE) |becasecseccescsu|eecllestoocdloacoes 9 9 ton. Boise R., Caldwell -.-. 4 43 33) 32) + 1,44,1 8 | To eye.---..-. 7 56 56 1 1 Potlatch Cr.,Lewiston_| 33-41! 33-43) 4 -44| 22-3 | O 1,4-4,0! 8or9 | To preopercle-| 7 | 60-66 62 10 10 1,4-4,1 2,4-4,2 Pataha R., Starbuck -.| 4 -44] 4-43) 44-5 | 28-3] + =‘ |...-.-.-.. 8or9 | To pupil.---.-]-- -| 60-64 62 13 9 Ret Walla R., Wal- pacosdbad kacsnaad To preopercle.}..-| 47-55 49 39 15 ula. Mill Cr., Walla Walla. 55 20 20 Umatilla R., Pendleton. GY) lloscacs 3 Columbia R., Umatilla. 53 7 r/ Natchess River, North AAD |esescelisccoca Yakima. Newaukum River, Che- |....-.|.-----]------|------ + BS) |lsscocc||zcbose! halis. Skookumchuck River, 4 | 41-42) 33-4 | 24-33) + 55 6 6 Chehalis. A. nubdila carringtonii. Port Weut: River, Po- 4 43 33 33) O To eye.------- if 69 69 al 1 catello. | Ross Fork, Pocatello..}......!..-...|.-----|-.---- 4 Tofrontofeye.| 7 | 53-72 64 15 13 Mink Creek, Pocatello. 44) 4h 4 3) + To eye.----.-.| 7 65 65 1 1 Port Neuf River, Po- |.....-|.-.---|....--|------ + To pupil. --.-- --| 61-67 64 4 4 catello. | Idaho Falls ..-...--.-- 43 41) 4 30+ Tofrontofeye.| 7 | 62-63 624 2 2 Little Wood River .... 33) 42 4/3 -34 4 To eye..-.-.-- --| 63-79 71 48 10 42 5 43 Payette R., Payette... 38 4h a5. sesso db Ibsocosseed| BGP |bsssecisssssaccaq| 7) 60-65 64 6 6 A. falcata. Boise River, Caldwell.| 33-4 | 44-54) 34-4 | 23:3 | + 97 11 Eaves R., Payette...) 3-4 | 44-5 | 34-41 3) + 27 6 ColumbiaR., Umatilla.| 4 —43) 43-5 | 34-4 | 23-3 | + 55 8 A. umatilla. | Columbia R., Umatilla.) 34-4 | 44-5 4 2h + 1,44,1) 8or9} To se nt of | 7 ; 63-70 66 10 10 | pupil. } | Payette R., Payette...) 33) 44-5 | 43-5 | 23-3 | 4 WEA | BOeO |bs-sesocscosssce | 7 | 60-65} 63| 6 6 t | *Higenmann, American Naturalist, February, 1893, 154, 44 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 14. Couesius greeni Jordan. Couesius greeni Jordan, Proc. U.S. Nat. Mus. 1893, 313. Type locality: Stuart Lake near Fort St. James, British Columbia. (Type, No. 44454, U. S. Nat. Mus.) In the collection made by Messrs. Bean and Woolman, September 20, 1892, in a small creek at Sand Point, Idaho, we find three examples of this species. In length they measure 23, 3, and 34 inches, respectively. Head in length of body, 4+, 44,44; depth, 4%, 5, 5; eye, 4, 34, 34; snout, 34, 34, 384; interorbital width, 3}, 34,3; D. 8; A. 8; scales, 10-55-6, 11-60-5, 11-60-6; 34 before the dorsal. The origin of the dorsal fin is at a point midway between base of caudal fin and the preorbital (not “‘preopercle,” as given in the original description of C. grecni, evidently a misprint for ‘‘preorbital”). These Sand Point specimens agree well with the type of C. greent with which we have compared them. This species seems to differ from the Couesius of the Upper Missouri basin (Cowesius dissimilis) in the somewhat larger scales and in having the scales Jess crowded on anterior part of body. 15. Cyprinus carpio Linnzus. Carp. The carp has been introduced into a number of ponds ané small lakes of the Columbia basin and from these has escaped into the streams. We saw it in Payette River at Payette and heard of it elsewhere. 16. Mylocheilus caurinus (Richardson). ‘‘ Chub”; ‘ Whitefish.” Cyprinus (Leuciscus) caurinus Richardson, Fauna Boreali-Americana, 111, 304, 1836. Type locality: Columbia River, at Fort Vancouver. Specimens obtained from Pend d’Oreille River at Newport, Idaho, 1; Boise River at Cald- well, Idaho, 19; Payette River at Payette, Idaho, 7; Snake River at Payette, Idaho, 8; Columbia River at Umatilla, Oreg., 16; Walla Walla River at Wallula, Oreg., 18; Blue Lakes, Idaho, 8; Umatilla River at Pendleton, Oreg., 1. An abundant and widely distributed fish in the Lower Columbia basin; notknown from Snake River above the falls, and probably does not occur there. Observed to be very abundant in the Pend d’Oreille below Newport. 17. Ptychocheilus oregonensis (Richardson). ‘‘ Squawfish.” Cyprinus (Leuciscus) oregonensis Richardson, Fauna Boreali-Americana, 11, 305, 1836. Type locality: Columbia River, at Fort Vancouver. Specimens obtained from Payette River at Payette, Idaho, 27; Hangman Creek at Tekoa, Wash., 9; Clearwater Creek at Lewiston, Idaho, 2; Potlatch Creek near Lewiston, Idaho, 3; Snake River at Payette, Idaho, 2; Boise River at Caldwell, Idaho, 28; Walla Walla River at Wallula, Wash., 1; Columbia River at Pasco, Wash., 6; Columbia River at Umatilla, Oreg., 4; Umatilla River at Pendleton, Oreg., 6; Natchess River at North Yakima, Wash., 4; Skookumchuck River near Centralia, Wash., 28; Newaukum River near Chehalis, Wash., 8; Mouth of Colville River, Colville, Wash., 6; Spokane River below Spokane, Wash., 10; Lake Pend d’Oreille at Sand Point, Idaho, 1; Pend d’Oreille River at Newport, Idaho, 7; Flathead Lake, Mont., 28. In the Pend d’Oreille River the squawfish is even more abundant than M. caurinus. No differences could be discovered between the above-mentioned specimens and others from the Sacramento River basin in California. 18. Leuciscus hydrophlox (Cope). Po-he-wa. Clinostomus hydrophlox Cope, Hayden’s Fifth Annual Report, 1871 (1872), 475. Type locality: Blackfoot Creek, Idaho. Clinostomus montanus Cope, 1. ¢., 475. Type locality: Grass Creek, Idaho. Clinostomus tenia Cope, Trans. Amer. Philos. Soc. Phila. 1874, 133. Type locality: Utah Lake. Numerous specimens of this species were collected in the Snake River at Idaho Falls, in Ross Fork of Snake River (on the Fort Hall Indian Reservation), and in Mink Creek and Port Neuf River near Pocatello, Idaho. The Indian name Po-he-wa means striped. The following tabular statement shows the variationin the number of anal fin rays among the examples col- lected at these places. Number of | 9 an. Locality. specimens | as Tays. | rays. | rays. | rays. | ofrays collected. | in anal Ndahowh alllskeeeeer esse eeeeee 18} || ssc6ao8 Ross Fork. -- se 52 1 Mink Creek... 5 Port Neuf River........---- ag Fi 2h | eee 19. 20. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 45 Numerous specimens (83) obtained in 1891 in the Snaxe River at President Camp and in a small creek at the head of Jackson Lake, Wyoming, * have 11 or 12 anal rays, 12 being the number in most of the examples counted. In 1892, Dr. Eigenmann obtained specimens at Idaho Falls, 2 of which have 12, 14 have 13, and 4 have 14 analrays each. Putting these with the 13 col- lected by us, gives an average of 124 anal rays for that locality. This species was obtained also by Jordan & Gilbert in 1889 in Heart Lake and Witch Creek, in Yellowstone Park.t The specimens examined by them had 13 anal rays. In specimens from Idaho Falls the scales of lateral line range as follows: 52, 53, 53, 53, 55, 55, 57, 57, 58, 61; from Port Neuf River, 51, 51, 51, 52, 58, 538, 55,57. Head from 4 to 44 in length, eye 33 to 3{in head. The maxillary scarcely reaches front of eye, and is 3 to 3} times in head. Leuciscus lineatus (Girard). Tigoma lineata Girard, Proc. Acad. Nat. Sci. Phila. 1856, 206. Type locality: Not definitely known, but probably somewhere in the Utah basin. Tigoma atraria Girard, 1. c., 208. Type locality: ‘‘ A spring in the Utah district, near the desert.” Tigoma obesa Girard, Proc. Acad. Nat. Sci. Phila. 1856, 206. Type locality: Salt Lake Valley. Tigoma squamata Gill, Proc. Bost. Soc. Nat. Hist. 1861, 42. Type lecality: Salt Lake Basin. Squalius cruoreus Jordan & Gilbert, Proc. U. S. Nat. Mus. 1880, 400. Type locality: Utah Lake; young specimens. Squalius rhomaleus Jordan & Gilbert, 1. c., 461. Type locality: Utah Lake; large specimens. Siboma atraria longiceps Cope. Zool. Wheeler Surv.,v, 667, 1876. Type locality: Snake Creek, Nev. This species was obtained at the following places: Idaho Falls; Port Neut River at Pocatello; Payette River at Payette. Others were seen at American Falls, in Snake River. Besides the localities given in the above synonymy, this fish has been reported from the fol- lowing additional points in the Snake River basin: Heart Lake and Witch Creek, in Yellow- stone Park (Jordan & Gilbert, 1889); Snake River at President Camp; Jackson Lake and a small creek at upper end of Jackson Lake, Wyoming, (Evermann & Jenkins, 1891); Suake River at Idaho Falls, Idaho (Eigenmann). In the Columbia basin it seems to be entirely confined to the Snake River, and in that stream its occurrence below the falls is exceptional. It is an excessively abundant fish in the Great Salt Lake basin, particularly in Utah Lake. Inthe canal at Idaho Falls we easily caught this and the preceding species by placing dough inside asmall dip net and allowing it to rest a short time on the bottom; great numbers of the two species, particularly of L. hydrophlox, would soon begin feeding on the dough, when they could be easily secured by lifting the net. In all recent papers this species has been listed under the name Leuciscus atrarius, but it seems quite certain that Tigoma lineata Girard is the same fish; and this, being the oldername, must take the place of atrarius. Leuciscus aliciz Jouy. Tigoma gracilis Girard, Proc. Acad. Nat. Sci. Phila. 1856, 206; not Cyprinus (Leuciscus) gracilis Richardson. Squalius copei Jordan & Gilbert, Proc. U. S. Nat. Mus. 1880, 461; not Leuciscus copei Giinther. Leuciscus alicie Jouy, Proc. U. 8. Nat. Mus. 1881, 10. Type locality: Utah Lake. This species was found very abundant in Little Wood River at Shoshone, where 86 speci- "mens were obtained. Slight differences are found on comparison of these specimens with others from the Sevier River, Utah, but these differences are not likely to prove constant and are not more extensive than are frequently found on comparing specimens from adjacent tributaries of the same stream. The eye is a trifle larger, the caudal peduncle rather more slender, the pectoral and ventral fins seem to average a little shorter, and the dorsal may be a little more anterior in position. This species has been previously known only from Sevier River, Beaver River, and Provo River in Utah. Its occurrence in the valley of the Upper Snake River is one more evidence of the identity of the two faunas. At least 8 of the species of our collection from the Upper Snake River are also found in the Great Salt Lake Basin of Utah. They are Catostomus ardens, Rhinichthys dulcis, Agosia nubila carringtonii, Leuciscus hydrophlox, Leuciscus aliciw, Leuciscus lineatus, Coregonus williamsoni, and Cottus punctulatus. The following is a detailed description of our specimens from Little Wood River: Head, 4 (32 to 44); depth, 4 to 44; eye, 33 to 4; snout, 32 to 4. D.8; A. 8, in 52 specimens, 9 in 34. Scales, 18-79 to 83-13. Body elongate, moderately compressed; head short and heavy, *Evermann, Bull. U. S. Fish Comm. for 1891, 44. + Jordan, Bull. U. 8. Fish Comm. for 1889, 48. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. interorbital width a little greater than snout; snout decurved; mouth wide, oblique, lower jaw very slightly projecting; maxillary reaching to within front of orbit; caudal peduncle long, the distance between anal fin and base of caudal about 1} in head, least depth of caudal peduncle 23 in head. Fins small; height of dorsal 14 in head, its free edge slightly convex; origin of dorsal somewhat behind ventrals, a little nearer base of caudal than tip of snout; anal about size of dorsal; length of pectoral equal to height of dorsal, their tips rarely reaching base of ventrals; ventrals short, equal to snout and eye. Lateral line somewhat interrupted, little decurved; scales crowded anteriorly. 21. Leuciscus balteatus (Richardson). Cyprinus (Abramis) balteatus Richardson, Fauna Boreali-Americana, 111, 301, 1836. Type locality: Columbia River, presumably at Fort Vancouver, Washington. Richardsonius lateralis Girard, Proc. Acad. Nat. Sci. Phila. 1856, 202. Type locality: Fort Steil- acoom, Puget Sound. This is one of the most abundant species of the Columbia and Lower Snake River, but appears not to occur in the Upper Snake, where itis replaced by LZ. hydrophlox. The accom- panying table shows the stations at which this species was taken, together with the number of individuals from each locality, and the number of rays in the anal fin. It will be seen that the latter vary, in our specimens, from 11 to 22 in number, the averages in the counts from the different localities ranging from 13 to 18. In a recent paper,* Dr. Higenmann announces the discovery that in this species the number of anal rays decreases with increasing altitude or that ‘“‘the higher the altitude the fewer the number of rays and the narrower the limits of variation.” To successfully establish such a generalization a very large amount of evidence would be necessary. As a contribution to this question we append the following table, which can not, we think, be interpreted as showing the truth of the theory. In this table we give, (1) the localities from which specimens were examined; (2) the range in variation in anal rays among the individuals; (3) the total number examined from each locality; (4) the average number of anal rays for each locality, and (5) the approximate altitude of each place. Fractions of 4 or more are included in the next higher number: avet 5 Approx- Number of anal rays r Total avers ee Locality. = No. ex- |"8°%°:| eleva- j | | lamined.| 828! | tion i 11 | 12 | 13 | 14 | 15 | 16 | 17| 18 | 19 | 20 | 21 | 22 | 23 | 24 | | rays. | “poo | | | é I | H | Little Spokane River, Dart's Mill)| 1) 14) 30)] 19} 6 |-_--)----) ee || fae} |e | 70 13 1, 850 Colville River, Meyers Falls-.--. ecodiocse| i) O| Gi Biko 0 eae asaleace 21 14 1, 200 Spokane River, Spokane --------- seoa ah ik 5 I} 8 11 14 1, 910 Revelstoke, B. Cia -.-..---.----.- sdca|fooccllecedlosas|| IW Jose 1 14 1,475 Browns Gulch, Silver Bow-.-- bee) 2 UO he ih i {ss 14 14 5, 344. Umatilla River, Pendleton .-.-.. LG ee On LOM | el! aalls 29 14 1, 070 Lake Washington, Seattle -... of ab |) ak) ala} as} eb al al 47 14 1 Mlathead) Wallies sa eee een soodloosa| Uy iti GB feco-|. 11 15 3, 100 Griffin Lake, B.C. a@.-------.-..-- pacdlossolese-) 8 i Bi We 14 15 1, 990 Small Creek at Sand Point, Idahoj.---}.--.| 2]19 | 36) 3} 6; 1 95 15 2, 100 Pend d’Oreille River, Newport -.|---.|----|----| 5] 1) 2 |-.--).-. 8} 15 2, 000 Hangman Creek, Spokane --..--- ssadlocos) 2] Bi BI Alf alg i ais) 1,910 Natchess River, North Yakima ..|----. zose| Bi 2] Bile alle 8 15 1, 078 Newaukum River, Chehalis -.-.- cal By ab loos 3 15 264 Post Creek, Flathead Lake - Asal ee ih esas 6 15} 3, 100 Golden sBS Cranes -eeseeene Saaleoel ons ous | By ee aE Ne 18 16 2, 550 Boise River, Caldwell.-.---...... fs 2 10) il aly a 65 16 2,372 Skookumehuck River, Chehalis..}....|.-.-| 2 }....] 1 7 2 Wllsssells 13 16 204 Payette River, Payette ...--..... seedleses| 3 || 25) || 2) |} Bp|| 2s || ay |) 12 154 16 2, 15u La Grande, Oregon a@.-----.------ a pero 6 it OS eallascul| 23 16 2, 786 Potlatch Creek, Lewiston ...-.... Bo BOE) leeT 1 OF esate 4 17 1, 200 | Silesimnome, 18), Che ccces cansscoacs : 3] 13 28} 8} 5 |-- 58 17 1, 300 Walla Walla River, Wallula : soa) 2] B ileeacl] 2 Il 6 17 326 @aldwell; Tdahoja)s2- 22-25-25. = oe BO Wy Bl 4 34 18 2, 372 Clearwater River, Lewiston Bealls Sod) Gap Bal) ee een. 16 18 750 Columbia River, Pasco..--.-..._. Be asiees Seelloace) forge lard hie slate 3 18 375 Umatilla River, Umatilla......._. Lees ab} al PRN 73 i 3) 12 18 300 | Snake River, Payette ........._.. peal sees sofa) yp aie B ----| 10 18 2,150 | Umatilla River, Umatillaa.....-. Saal ee Beall 5 1 RSet 10 19 300 jeMEISsion BAC seme eee eee ee eee B At 2h) Ye |) ale} |) Ps 2 | 79 19 al | Kamloops, B: @.a@---.22 202222222). Se Meseel (Sel otl [asea emed aces iseaa Seca Beas ---| 2 204 1, 158 | ‘ t @ These are the localities from which specimens were examined by Dr. Eigenmann. The data are those given by him, * Results of Explorations in Western Canada and the Northwestern United States. Bull. U.S. F. C. 1894, 131. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 47 In the above table we have arranged the data beginning with the lowest average number of anal rays (13), and proceeded from that to the highest (204). In the table which follows we have arranged the localities in order of elevation, beginning with the lowest. Table showing the relation of altitude to number of anal fin rays in Leuciscus balteatus. No. of Approxi-| Average | SPCC! Localities. mate alti-| No. of | ents | tude. anal rays:| Beaasen | | |. re based. Lake Washington -....---------.----------- | 1 14 47 NIGESGING sesoasocoscoocoase Se| 1 19 719 Newaukum River. .--..---- 25] 204 id 3 Skookumchuck River | 204 16 13 Umatillaa -- ae a0 300 19 10 (Uma till aeseeree eae co 300 18 12 AWN oo oe ob Shes be cocsSe Scapa EneasOEGoaTd | 326 7 6 LENCO a maudosconcabascdese S04 375 18 3 JEG ARLIOIN oh scocouoscounc5 Sa 750 18 16 Pendleton. - seisl| 1, 070 14 29 Kamloops @-.-.---- eo Baedl 1,158 204 2 Potlatch Creek at Lewiston. --- oal| 1, 200 7 4 Meyers Falls -......---.---- Bose| ie 20) 14 21 Sicamous @....--.-.-------- eae 1, 300 17 58 Revelstokea- Saas 1,475 14 1 DartisoMills\ee sees cobs 1, 850 13 70 Griffin Lakea ---.....--.-- od05 1,900 15 14 Spokane River, Spokane. -- oan 1,910 14 11 Hangman Creek, Spokane - 6558 1,910 15 ape WOW ANDOU saescagsccspepeaces Seco 2, 000 15 8 SandtPointeree eee ee oBce) 2, 100 15, 95 Snake River, Payette ..-..- | 2,150 18 10 Payette River, Payette ...-..-- <|| 2,150 16 154 Caldwella o 2,372 18 34 Caldwell -| 2, 372 16 65 Goldena 52 -| 2, 550. 16 18 MalGrandeapeeceseecee eee .| 2, 786 16 23 Flathead Lake ........-..---.-- os 3, 100 15 11 PostiCneekeeseseceeee eee dl 3,100 154) 6 Silver! BoOWwasere ee oocee cee ee eeeeee y 5, 344 14 14 a These data are from Dr. Higenmann’s paper. Comparing these with Eigenmann’s results, it will be seen that the average number of rays from our lowest elevation (14 at Lake Washington) is fewer by 2 than the average from his highest elevation (16 at La Grande) and that our average for Newaukum River (204 feet elevation) is fewer than any average found by him except at Revelstoke (1,475 feet), where his average is 14, and at Griffin Lake (1,900 feet), where it is 15. The average found by us at sea level (Lake Washington) is identical with that found at Silver Bow,* whose elevation is 5,344 feet, the greatest elevation from which specimens have been examined. When there is absolutely no difference between the averages for the lowest and the highest elevations it is not possible for us to see any reason for the generalization, ‘‘the higher the altitude the fewer the number of rays.” If the figures show anything, they show that the number of anal fin rays does not detrease with increasing altitude. It is also stated that the greatest range of variation is at the lowest altitudes, but further on it is stated that the ‘‘ greatest variation in this [the Columbia] system was not at the lowest altitude, but at an elevation of 2,372 feet.” The range here was found by him to be through 9, or from 12 to 20. We find the same range of variation (from 12 to 20) in our specimens from Payette (2,150 feet.) The range found at the point nearest sea level in the Columbia basin (at Umatilla, 300 feet) was only through 6 (15 to 20), and the average for that place is nearly identical with that at Caldwell. The vari- ation found by Eigenmann among his Mission specimens is through 9 (16 to 24), while that of our Lake Washington specimens is through 7 (12 to 18). The averages for these two places, both of which are at sea level, are 19 and 14, respectively. We consider Leuciscus lateralis (Girard) a simple synonym of JL. balteatus. Our material covers well the type localities of both (Columbia River at Fort Vancouver and Fort Steila- coom on Puget Sound) and indicates but one form. We are certainly not prepared to consider lateralis a subspecies of balteatus, occupying the same brook with its parent form, as indicated by Eigenmann. * Ti. gilli, probably a good species, 48 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 22. Coregonus williamsoni Girard. (Plate rx). Coregonus williamsoni Girard, Proc. Acad. Nat. Sci. Phila. 1856, 136. Type locality: Des Chutes River, Oregon. Specimens obtained from the Payette River at Payette, Idaho, 24; Clearwater River at Lewiston, Idaho, 7; Columbia River at Umatilla, Oreg., 1; Natchess River at North Yakima, Wash., 8; Newaukum River at Chehalis, Washington, 4; Little Spokane River at Dart’s Mill, Wash., 6; Flathead Lake, Montana, 2; Post Creek, St. Ignatius Mission, Montana, 9; Clarke Fork at Thompson Falls, Mont.,1; Creek at Sand Point, Idaho, 5; Spokane River, Spokane, Wash., 1. Abundant and widely distributed throughout the Columbia and Upper Colorado and Missouri basins. Specimens of this fish were obtained by Mr. Bean which are of unusual interest in that they show the breeding colors. Mr. Bean’s report upon these specimens is given at the end of this paper. 23. Oncorhynchus gorbuscha(Walbaum). Humpback Salmon. The humpback salmon was ruuning in great abundance at the date of our visit to Puget Sound (August 26) and was the only species then being handled at the canneries. We were informed that the humpbacks did not run last year, and in corroboration of the fact that this species runs on the Sound in alternate years only, it is recalled that it did not appear in 1880, when one of us visited this region. As is well known, the female humpbacks are plump, symmetrical, silvery fishes of attractive appear- ance and good flavor. Aside from the traditional requirements as to color of flesh, the species is well adapted for canning purposes. No young of this species were found in the streams. 24. Oncorhynchus keta (Walbaum). Dog Salmon. The young of the dog salmon were abundant in the Newaukum and Skookumehuck rivers at Chehalis, Wash. They average slightly smaller than young quinnats, and are readily distinguished by their larger eyes. They have 12 or 13 branchiostegal rays, 14 rays in the anal fin, 130 to 135 scales in lateral lines, and 7 ++ 13 gill-rakers. Like the young quinnats, these were all nearly uniform in size, and were evidently young of the preceding year. 25. Oncorhynchus tschawytscha (Walbaum). Quinnat Salmon. Only the young of the Columbia River salmon were seen by us during our short investigation of the Columbia and the Snake. It had not yet reached its spawning beds on the Snake at the time of our visit, and the ‘‘close season” prevented its capture later in the lower river. Such faets as we weie able to ascertain concerning its run and spawning-grounds are therefore based on interviews with fishermen and others, a kind of testimony which must, in this case, be scrutinized with more than usual care. On the upper river it was repeatedly found that no distinction was seemingly made between the salmon and the steelhead, and of the two species of salmon that almost certainly spawn in the upper course of these streams, no distinctive accounts could be had. A ‘silver salmon” was, indeed, frequently mentioned, but we were unable to ascer- tain whether the fish thus distinguished was the female quinnat or the blueback (0. nerka). As already indicated in our notes on the streams, salmon ascend the Snake River to and above the Lower and Upper Salmon Falls, and have important spawning-beds in the main stream, both above and below these falls. They are not known above the Auger Falls, and probably do not even reach the foot of the Great Shoshone. They appear first in this portion of the river early in September, or occasionally in the latter part of August, at a time when the streams are so low that falls or dams which would form no barrier earlier in the season now keep them out from otherwise favorable spawning-beds. This fish is not much used for food in the upper waters. The remnants of the various Indian tribes make yearly visits to the spawning-beds and occasionally white men have tried to put them on the market, but without success. Great numbers, are, however, annually killed through mere Jove of destruc- tion. The advent of the salmon brings out from every town men and boys with pitchforks or other weapons, curious to see how many of these fish they can destroy. It is to be held in mind that these localities in Idaho and in the eastern portions of Oregon and Washington are 80 remote from the canneries that the people have no interest whatever in the preservation of the salmon. We can not, therefore, depend upon public sentiment to enforce protective legislation. We give below such information as we possess concerning the distribution of salmon in the Snake and Upper Columbia rivers. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 49 The prineipal tributaries of that portion of Snake River which is accessible to salmon are the following: Salmon Falls River or Salmon Creek, Malade River, Bruneau River, Owyhee River, Boise River, Payette River, Salmon River, Grande Ronde River, Clearwater River, and Palouse River. : Beyond the fact that the fish enter Salmon Creek and ascend it for a mile or more, we were able to ascertain nothing concerning the value of this stream. Malade River is effectually shut off by high falls near its mouth. Trout are said to ascend the stream during high water in the spring, but it becomes impassable later. The stream often goes dry for a distance of many miles along the lower part of its course, and has also other impassable falls in its upper course. Bruneau River was formerly an important stream for spawning salmon, which reached its head waters in October, according to the statement of Mr. J. L. Fuller, of Bliss, Idaho. We are informed that a dam recently constructed in the Lower Bruneau now wholly prevents the ascent of fish. The Owvhee River is still open to salmon, so far as could be learned from reports. Mr. J. L. Fuller has seen them in the extreme head waters of the Owyhee in Nevada. Boise River, like the Bruneau, was formerly a salmon stream, but is now partly or entirely closed by a dam near Caldwell, and is unsuitable by placer mining in the upper part of the stream. The Payette, Salmon, and Clearwater rivers are all available spawning-grounds for the salmon, and we learned of no obstructions in these streams. The Grande Ronde River is ascended as far as La Grande, where a dam obstructs further progress. A high falls at the mouth of the Palouse River prevents the ascent of salmon. The principal streams and lakes tributary to the Columbia River above the mouth of the Snake are: Yakima River, Wenatchee River, Chelan Lake, Okanagan River, Spokane River, Little Spokane River, Cour d’ Alene Lake, Colville River, Kettle River, and Pend d’Oreille River and its numerous tributary streams and lakes. Up to 1885 the Yakima River was visited by three or four species of salmon, including the quinnat, in great numbers. Important spawning-beds were located in the bed of the streain; but in recent years but few salmon have made their appearance in thisriver. No artificial obstructions have been interposed, and the decline can hardly be due to any change iti the character of the stream. We were not able to visit Chelan Lake and Wenatchee and Okanagan rivers, and do not know to what extent salmon frequent these waters. Spokane River, below the falls, was formerly an important salmon stream containing large spawning-beds, but salmon are rarely seen there now. The steelhead still occurs in considerable numbers in the Spokane. The Little Spokane, as already stated elsewhere, was visited by salmon in great numbers in 1882 and previous years, but since 1882 the number has been very few. Salmon have never been able to reach Ceur d’Alene Lake, Spokane Fallsapparently having proved an effective barrier to their ascent, Only the few miles of Colville River below Meyers Falls can be reached by salmon, and it is certain that this was formerly an important spawning-ground. The portion of Colville River above the falls would prove excellent for salmon, but Meyers Falls form an absolute barrier. Kettle River flows into the Columbia from the west, just above Kettle Falls. It was not visited by us, and we were not able to get any reliable information regarding it. The Pend d’Oreille River and the numerous important streams and lakes tributary to it have been discussed elsewhere in this paper. The occurrence of salmon in Kettle River and the lower part of the Pend d’Oreille is dependent upon their ability to ascend Kettle Falls. That salmon formerly reached and ascended the Lower Kettle Falls seems pretty well estab- lished; but whether they have ever passed the Upper Kettle Falls is not so certain. The Upper Falls, upon examination, donot appear to be as formidable as the Lower. The fact that so little evidence could be secured regarding the occurrence of salmon in any of these waters renders it highly probable that at no time have they ever ascended in any considerable numbers above the Upper Keitle Falls, S. Mis. 200——4 50 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. While it is true that the salmon are shut out by falls and dams from a large area of the Columbia and especially the Snake River basins, and while it is also true that the limitations are increasing as streams become useful for irrigation purposes and for mining, it is never- theless certain that the decrease in the numbers of salmon, due to ill-regulated fishing in the lower Columbia, has so far outstripped the decrease in area of spawning-beds that the latter are now more than ample for all the fish that appear. We do not, therefore, believe that increasing the spawning-grounds through the removal of obstructions would materially benefit the salmon industry. In our judgment, the streams can be repopulated only by regu- lating the fishing in the lower Columbia and at the same time increasing the output from the hatcheries. The young of this species were taken in abundance at the mouth of the Natchess River, near North Yakima, Wash., and in the Pataha River at Starbuck, Wash. The largest indi- vidual seen was 88 mm. long, the smallest 55 mm.; the average is about 70mm. These are evidently the young of the previous year, and their uniformity in size indicates clearly that all pass out of the brooks to the sea, or at least to the deeper river channels, during the first and second years. We count in these young specimens 133 to 140 scales in the lateral line, 16 or 17 branchiostegal rays, 7 + 12 gill-rakers, and 15 or 16 rays in the anal fin. 26. Salmo gairdneri Richardson. Steelhead. ‘The steelhead is an abundant fish in the larger streams of the Columbia basin, especially about Spokane and the mouth of the Pend d’Oreille. Sev- eral fine examples of this fish were taken with the spoon by Mr. B. A. Bean in September, 1892, near Spokane. These were called “salmon” by the residents, and Mr. Bean is of the opinion that most, if not all, the ‘‘salmon” which they take by trolling are really not salmon, but the steelhead. Mr. Bean was also told that the ‘‘salmon” about the mouth of the Pend d’Oreille and Salmon rivers are taken by trolling, and it is quite likely that these also are — steelheads, 27. Salmo mykiss (Walbaum). Rocky Mountain Trout; ‘‘Sa-pen-que” of Fort Hall Indians. A very large series of trout from the Columbia basin has been examined and from widely separated localities. Specimens were not preserved in every case, but an examination was made of trout from the following places: Snake River at Idaho Falls; Ross Fork and Mink Creek near Pocatello; Little Wood River at Shoshone; Pataha River at Starbuck; Little Spokane River at Dart’s Mill: Lake Ceurd’Alene; Lake Pend d’Oreille at Sand Point; Pend Oreille River at various places between Newport and the mouth of Salmon River; mouth of Colville River; Newaukum River at Chehalis; Green River at Hot Springs. With every additional collection of black-spotted trout it becomes increasingly difficult to recognize any of the distinctions, specific or subspecific, which have been set up. The present collection adds not a little to the difficulty. Weare now convinced that the greater number of the “subspecies” of S. mykiss have no sufficient foundation. We find our speci- mens from the Upper Snake River (Ross Fork and Mink Creek at Pocatello) to be typical mykiss, having small seales, in 176 to 180 transverse rows, and a deep red dash on inner side of mandible. The spots are most abundant posteriorly, and the specimens are scarcely to be distinguished from the so-called Salmo mykiss pleuriticus of the Colorado River. When taken in the larger river channels the fish is lighter colored, with finer spots and fainter red mark on lower jaw. Between such typical myliss and the form represented in our collection from such coastwise streams as the Newaukum River at Chehalis, Wash., there seems to be a wide difference. The latter has conspicuously larger scales (in 120 to 130 cross rows) and no red streak on lower jaw. The sea-run individuals of this latter kind we believe to be the steelhead (S. gairdneri), and between it and the mykiss we are now unable to draw any sharp line. Thus the Wood River specimens have fine scales (150 to 163 transverse rows) and usually no red dash under the jaw. Some specimens show traces of the latter, and in such it is usually faint and irregular. From the Umatilla River at Pendleton, the Natchess River at North Yakima, and the Pataha River at Starbuck the scales are intermediate in size, ranging from 142 to 163 in number, averaging perhaps 148. In these the lower jaw shows nored. Specimens from the Ceur d'Alene region have the red dashes usually very distinct, but vary greatly in size of scales. Wardner examples look much like typical mykiss, with 165 to 170 scales. From Cur d’Alene Lake we find 130 to 166, with the average about 145, while from the Little Spokane River at Dart’s Mill specimens with conspicuous red dash on mandibles haye the INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 51 ‘scales averaging 125 in number. Trout from the Green River at Hot Springs, Wash., and from the Newankum River at Chehalis have also 123 to 130 scales. We think it not unlikely that the coastwise form should be recognized as Salmo mykiss gairdneri, though the question is sadly in need of systematic and thorough investigation. The Fort Hall Indians call the trout Sa-pen-que, which, they say, means good fish. 28. Salvelinus malma (Walbaum). Charr; “Salmon Trout ;” “Bull Trout.” Salmo malma Walbaum, Artedi Piscium, 66, 1792. Type locality: Kamchatka. Abundant in the Pend d’Oreille River. At La Claires we saw in the possession of an Indian several fine specimens, the largest of which was 26 inches long, 11 inches in greatest circumference, and weighed 5 pounds and 1 ounce; the length of the head was 6 inches. The people along the river know this fish as the ‘‘charr,” while in Montana, from Flathead Lake to Missoula, it is called ‘‘salmon trout” or ‘‘bull trout.” One example was obtained by Bean and Woolman from Lake Pend d’Oreille. One specimen of 3 pounds weight was seen at North Yakima, which had been caught in the Natchess River. Salvelinus malma has not yet been found in the Snake River, and it is doubtful if it occurs in that basin above the falls. The only Salvelinus yet known in that basin is from Henry Lake, and was identified by Dr. Bean as S. namaycush. 29. Columbia transmontana Higenmann & Higenmann. Columbia transmontana Eigenmann & Higenmann, Science, October 21, 1892. Type locality: Umatilla River, at Umatilla, Oregon. Seventy-three specimens of this very interesting species were obtained in the Walla Walla River at Wallula, Wash., August 23, 1893, by Messrs. Thoburn and Rutter. The largest specimen measures 3}; inches in total length. Diligent search was made for this fish at the locality where the types were obtained, but none was found. The specimens taken at Wallula were found in only one place, about 250 feet below the railroad bridge, on the edge of a large patch of Chara and in water about 3 feet deep, where the bottom was somewhat muddy. The temperature of the water here was 70° at 11:30 a. m., when the air was 80°. Measurements of the larger specimens give the following results: Head, 3} to 34; depth, 3} to 34; eye, 33; snout, 3. D. 1,9; A. u, 6; V.8; P.10 to 12. Scales 9-43-8. Ground color pale straw-color, profusely covered with fine dark-brown specks which form irregular blotches along the course of the lateral line and on median line of back, there being 1 at anterior base of dorsal and 3 or 40n the caudal peduncle; head with fine dark spots on sides; dorsal, anal, and caudal barred with dark. Dorsal and anal spines strong; the first dorsal spine short, less than length of eye, the second much longer, as long as from tip of snout to middle of pupil; first anal spine scarcely as long as first dorsal; second anal spine about one-third length of head; longest dorsal rays about 14 in head; those of anal shorter. 30. Gasterosteus microcephalus Girard. Abundant in the Walla Walla River at Wallula, in Lake Washington. and in Skookumchuck River near Chehalis, Wash.; not seen elsewhere by us. Specimens taken were found to be extremely variable in the extent to which lateral shields are developed. Ina considerable proportion the plates are developed along the entire length of sides of body and the caudal peduncle is sharply keeled. Others have but 4 or 5 plates developed, the caudal peduncle being then without trace of carina. Between these extremes are all possible intermediate conditions, thus establishing a series comparable with that recently reported on by Dr. G. A. Boulenger as occurring in the common European species. 31. Cottus asper (Richardson). A single specimen, 145 mm. long, from the Walla Walla River at Wallula, Wash., and two small specimens from Lake Washington. The dorsals contain 9 spines and 21 rays, and are slightly connected at base. Soft dorsal high, the longest ray 2} in length of head, the posterior rays reaching base of caudal. Anal fin with 17 rays; pec- torals, 16. Interorbital space very wide, equaling diameter of eye, not concave. Prickles coarser and less closely crowded than in the Sacramento River specimens, distinctly visible without the aid of a lens, and directed upwards and backwards. The lateral line is complete, without abrupt angle under last dorsal rays, and contains 38 pores. Anus separated from first anal ray by a space equaling two-thirds diameter of orbit. Band of palatine teeth weak. Anterior nostril in a short tube. The coarser prickles, more anterior position of vent, and flatter interorbital space probably indicate that C. asper is separable, at least subspecitically, from the Sacramento River form. Material is not now at hand to settle this point, 52 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 32. Cottus rhotheus (Rosa Smith). This strongly-marked species is abundant in the Spokane region, and was taken at the following stations: Little Spokane River at Dart’s Mill near Spokane, and at Chattaroy, Wash.; Columbia River at Colville, Wash; Cour d’Alene Lake near Cour d’Alene, Idaho; Clearwater River near Lewiston, Idaho; Walla Walla River at Wallula, Oreg.; Hangman Creek at Tekoa, Wash.; Natchess River at North Yakima, Wash.; Newaukum River near Chehalis, Wash.; Snoqualmie River at Snoqualmie Falls, Wash. The salient features of this species are: (1) The pointed wedge-shaped profile of head, as viewed from above; this contrasting strongly with the usual broadly-rounded contour of other species. (2) The wide horizontal mouth, quite at lower profile of head. (3) The noticeably concave interorbital and occipital regions. (4) The very broad and long palatine band of teeth. (5) The rather slender body and the extremely slender caudal peduncle, the latter expanding fan-like at base of caudal fin. (6) The dorsals usually separate; when united, at extreme base only. (7) Laterallinecomplete. (8) Sides usually well invested with prickles, which are triangular and coarse, and less closely placed than in asper. They are arranged more or less definitely in oblique series. There is considerable variation in the completeness of the investment, and in one specimen from Chehalis an axillary patch only is present. The specimens from Chattaroy and from Snoqualmie Falls are referred to this species with doubts as to their identity. The following table will give an idea of the amount of variation in number of fin-rays, found in this species: Spinous dorsal. | Soft dorsal. Anal. Pectoral. Locality. | 2 vit. | vir. | 16 | 17 Titers Little Spokane River ---.---- 4 6 6 4 82 to 34 WNatchess River----..-.------.|.-..--.- 2 OP etary Newaukum River ---- soc\acanoeod 2 1 1 Walla Walla River ..---...--- 1 2 1 2 Colville River 1 Ieee! Tekoa)--------.-.-.--- 5 5 1 Cceur d’ Alene 4 3 2 Lewiston ..0c..22.---=---ncnee 1 1 |bescsalestace 33. Cottus punctulatus (Gill). Ah-we. Cottopsis semiscaber Cope. Hayden’s 5th Report, 1871, 476. Type locality: Fort Hall, Idaho. Numerous specimens from Green River, Wyoming. Comparing these with typical semis- caber from the vicinity of Pocatello, Idaho, we can appreciate no difference whatever, except in the matter of armature. None of our Green River specimens show any prickles whatever. About half the Pocatello specimens are also naked and are indistinguishable from typical punctulatus; in the others more or less prickles are developed, varying from a few in axil of pectorals to a band covering more than half of the sides. As this is not an unusual amount of variation, we do not consider C. semiscaber worthy of recognition. The relations of C. punctulatus with the eastern species have not been carefully worked out, and it seems best to recognize it for the present as distinct. Specimens from Green River, Wyoming, and from Mink Creek, Ross Fork, and Port Neuf River, Pocatello, Idaho. The collection contains also two specimens from Thompson Falls and three from Flathead Lake, which seem to be this species. By the Fort Hall Indians this fish is called ah-we, a word meaning horns. 34 Cottus perplexus, sp. nov. (Plate vill.) Type locality: Skookumchuck River, near Chehalis, Wash., where 13 specimens were col- lected, August 28, 1893, by Drs. Gilbert and Jenkins. Associate type locality: Newaukum River near Chehalis, Wash.; 26 specimens, collected by Drs. Gilbert and Jenkins, August 28, 1893, Type, No. 45387, U. S. Nat. Mus. Co-types, No. 45386 (Newaukum River), U. S. Nat. Mus.; and Nos. 1324 to 1343 (Skookumchuck River), Museum Leland Stanford Junior Uni- versity. Related to Cotius punctulatus. Salmon Fisheries of the Columbia River. (To face page 52.) PLaTe VIII. Fig. 1. COTTUS PERPLEXUS, sp. nov. (Type.) Skookumchuck River, Chehalis, Washington. Fig, 2. COTTUS LEIOPOMUS, sp. nov, (Type.) Little Wood River, Shoshone, Idaho. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 53 Description: Head, 3}; depth, 44; eye, 4; snout, 4; interorbital width,54. D. vir, 21; A. 15; P. 16; V. 1, 4. Least depth of caudal peduncle greater than snout, 3} in head; interorbital space rather broad, about 14 in eye. The body is deeper and more compressed than in any other species known to us, this being especially noticeable posteriorly; the caudal peduncle is very short and deep, and is entirely overlapped by the posterior dorsal rays which extend beyond base of caudal fin. Length of caudal peduncle from base of last dorsalray about } depthof same. Depth of body at origin of anal fin } length of head. Interorbital space slightly concave; occiput flat or transversely convex. Mouth oblique, the maxillary reaching vertical from posterior margin of pupil, 24 in head. Teeth in a very narrow crescentic band on vomer, none on palatines. Upper preoper- cular spine short and broad, curved or simply directed upward; below this two stout, blunt spines directed downward. Body, in the type specimen, entirely naked; lateral line incom- plete, not reaching end of soft dorsal. Spinous dorsal low, the longest spines not greater than length of snout; soft fins all high, the 15th dorsal ray equal to snout and eye; a broad membrane always connecting the two dorsals, the notch inconspicuous. Last rays of anal as well as dorsal extend beyond base of caudal; first anal ray under third ray of soft dorsal; ventral spine and rays slender and weak. Anus midway between base of caudal fin and front of eye Color in alcohol, back and sides with vermiculations of light and dark, the back with 5 oré ill-defined black crossbars, which usually reach the lateral line; the usual black bar at base of caudal, emarginate posteriorly; below the lateral line a number of small, quadrate, dark blotches, arranged in two irregular series; lower parts unmarked except with fine dark punctn- lations; dorsal, pectoral, and caudal fins crossbarred with dark; anal and ventrals with numerous small dark specks. Length, 91 mm. The co-types show that this species is subject to some variations which should be men- tioned. Head, 3to3}; depth, 44to44. D. vitor vil, 18 to 21; A. 14t016; P.14t016. While the body is usually entirely naked, there is occasionally an axillary band of prickles, sometimes supplemented by a single irregular series of prickles along base of the dorsal fin. The notch in the membrane connecting the dorsals is usually inconspicuous. In some examples the black crossbars on the back do not reach the lateral line. The small, quadrate, dark blotches below the lateral line are sometimes arranged in a single series parallel with base of anal, sometimes in two irregular series. The following table gives the fin formula in a number of individuals of this species: Dorsal spines. Dorsal rays. Anal rays. No. of Locality. 7 specimens VI. | VIII. | 18 19 20 21 14 | 15 16 | examined. Skookumchuck River ----... 3 10 4 5 2 2 6 6 1 13 Newaukum River ---..-..--- 3 2 2 74 lloeceso 1 1 3 1 26 Natchess River-.-......---.-- 4 1 1 3 I boodes 2 2 | 1 5 From Cottus punctulatus, which it most closely resembles, this species may be distinguished by its deeper body, more elongate anal fin, the broad union between the dorsals, the absence of palatine teeth, and the different coloration. 35. Cottus leiopomus sp. nov. (Acioc, smooth; zaua, cover, opercle.) (Plate vil.) Typeloc >: Upper Little Wood River, Shoshone, Idaho, where 2 specimens were col- lected, Septewver, 1893, by Mr. H. H. Kinsey. Type, tu. 45389, U. S. Nat. Mus. Co-type, No. 1151, Museum Leland Stanford Junior University. Related to Cottus philonips. Head, 34; depth, 54; eye, 44; snout, 34; interorbital width, 74; D. vi-17; A. 12; P.13; V. 1,4. Least depth of caudal peduncle about equal to length of snout; interorbitalspace very narrow, much less than diameter of eye; mouth small, maxillary reaching vertical from middle of eye, a wide strip visible laterally in the closed mouth. Preopercular spines entirely absent, the preopercular margin evenly rounded throughout, without prominence, and without the least trace of a spine. Ot INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. Vomer with a very narrow band of teeth; palatines naked; skin wholly naked, lateral line complete. Dorsal fins not joined unless at extreme base; fins all low, the pectorals barely reaching front of anal; front of anal under third ray of second dorsal, its last ray under fourth from last ray of latter. Free portion of caudal peduncle (behind last anal ray) contained 13 in head; portion behind baseof last dorsal ray 3 in head; neither dorsal nor anal reaching base of caudal when depressed. Color in alcohol: head on sides rather finely vermiculated with light and dark; plain whitish below; not coarsely spotted or blotched as in philonips; dorsal bars indistinct; two narrow black lines downward and backward from the eye; an evenly convex dark bar at base of caudal; dorsals, pectorals, and caudal faintly crossbarred. Length 31 mm. : The second specimen, which is 71 mm. long, agrees closely in every respect with the type. This species is very closely related to Cottus philonips, from which it differs only in the total absence of any preopercular spine. In both specimens, and on each side, the preoper- cular margin is entirely rounded throughout, without any prominence and without the least trace of a spine, It seems very improbable that the two should agree in being merely abnor- mal in this respect, and we are forced to conclude that a form exists which is peculiar to the Malade River, a stream otherwise remarkable in its ichthyologic features, 36. Cottus philonips Eigenmann. This name was proposed as a substitute for Cottus minutus Pallas, supposed to be pre- oecupied, and Cottus microstomus (Lockington), not of Heckel. The first mentioned is perfectly available, but was applied to a specimen from the island of Talek, near Tanisk, in the Okhotsk Sea. It is very doubtful, therefore, whether C. minutus should be used for any American species in advance of comparison with the Siberian form. From the Aleutian Island species (C. microstomus Lockington), C, philonips differs in many important respects, and is undoubtedly distinct. Thus the Alaskan form has the posterior nostrils in short but conspicuous tubes, the preorbital produced into a lobe which conceals all of the maxillary except the extreme tip, and the dorsal fin with 8 or 9 spines and 18 to 20 soft rays. Cottus philonips is a small-headed form, typically with perfectly smooth skin and unarmed palatines. Like most other species of the genus it occasionally develops a band of postaxillary prickles, which are often accompanied in the same specimens, by a small patch of teeth on the palatine bones. The head is less strongly armed than usual, the single preopercular spine being short, the preopercular margin otherwise wholly unarmed. In this respect C. philonips differs from all other western species of Cottus, except the Alaskan form above mentioned. The dorsal varies from vil or vil, 16 to 18; the anal from 12 to 14. The nostrils are without tubes, and the preorbital little produced, exposing the greater part of maxillary in closed mouth. Specimens were obtained in the Port Neuf River near Pocatello, at Snoqualmie Falls, and in a spring branch emptying into the South Fork of the Coeur d’Alene River, near Ward- ner, Idaho. We have also seen specimens taken from Birch Creek, in western Idaho, by Merriam and Bailey. 37. Cottus marginatus Bean. Six small specimens from Mill Creek at Walla Walla (the type locality of marginatus) agree with Bean’s description and differ from all other western specimens of Cottus which we have seen in having but three soft rays in the ventral fins. So far as can be ascertained from our very immature specimens, marginatus strongly resembles perplexus, with which it agrees in fin rays, naked skin, the incomplete lateral line, and the absence of palatine teeth. C. perplecus has constantly 4 soft rays in the ventral fins, and other differences may appear when compared with adult specimens. In our specimens of marginatus, the anus varies in position, being sometimes nearer base of candal fin than snout, sometimes nearer snout. Twenty-two small specimens, collected by Bean and Woolman at Sand Point, Idaho, are for the present referred to this species, though we are not certain that this identification is correct. The ventrals seem to be I, 3, but the body is more or less covered with prickles. oe (To face page 54.) Salmon Fisheries of the Columbia River. Pocatello, Idaho Ross Fork of Snake River, (Type.) CATOSTOMUS POCATELLO, sp. nov. Fig. 1. (Type.) Columbia River, Umatilla, Oregon Fig. 2, AGOSIA UMATILLA, sp. nov. Little Spokane River, Washington, Breeding male. Fig. 3. COREGONUS WILLIAMSON| Girard. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 55 NOTES ON WILLIAMSON’S WHITEFISH IN BREEDING COLORS, FROM LITTLE SPOKANE RIVER, WASHINGTON, AND REMARKS ON THE DISTRIBUTION OF THE SPECIES. [By Barton A. Bean, assistant curator, Department of Fishes, U. S. National Museum. ] In the fall of 1892, while on an investigation of the streams in eastern Washington with a view of locating a site for a fish-hatchery for the U.S. Fish Commission, the writer had the good fortune to secure a very fine specimen of Coregonus williamsoni approaching the breeding condition. As will be seen in the accompanying illustration (Plate 1x) the tubercles on the seales at this time are very prominent, situated on the middle of the scales, milk-white in color, and forming horizontal lines along the body from head to tail. About six- teen of these lines can be counted between the back and ventral edge of the body. The tubercles show on the abdomen, but the color of that portion of the body and of the tubercles being similar, they are indistinct. Color: Dark on back, sides a lighter steel-gray, and under parts white; all fins tipped with black; caudal and adipose fins steel-blue. D. 14; A.13; scales, 9-83-10; pores in lateral line, 80; head, 5; depth, 4; eye, 44; snout,3. Gill-rakers short, about 12 below the angle. Mouth very small, the maxilla barely reaching vertical through front of eye. Dorsal fin highest in front, gradually graduated to last ray, which is less than one-half length of anterior rays. The greatest height of the dorsal slightly exceeds its length of base, which equals the length of the ventral fins, being considerably less than the length of the pectorals, but exactly that of the anal fin. Caudal fin deeply forked. The specimen here described and illustrated measures not quite 11 inches; it is a male and was taken in Little Spokane River near Clark Springs, October 5, 1892. Another example, a little larger (124 inches), was obtained by Mr. A. J. Woolman in the Little Spokane, in September, 1892. The tubercles on this specimen are well preserved, those on the former having disappeared owing to exposure to the air and handling while being drawn. : Mr. Woolman’s example has the following characters: D.13; A.13. Scales 10-80-10. Head, 5; depth, 4. Adipose fin very long, the length of its base being contained 24 times in the head’s length. Color as in the preceding. So far as we are aware the breeding whitefish, of any species, has not been here- tofore described in America. In Faune des Vertebres de la Suisse, volume vy, Hist. Nat. des Poissons, Géneve, 1890, Dr. Victor Fatio presents the following note on the breeding colors of Coregonus exiguus of Switzerland: Adult males differ from the females by a more slender body, larger head, stouter muzzle of snout, the greater development of the paired fins, especially the pectorals, and during the breeding season by a more intense coloration, also by the stronger and more numerous epidermic buttous, and some- times by the more strongly arched scales on the lateral line. Tt is unfortunate that we failed to secure female fish. The whitefish were very abundant in the Little Spokane; large numbers were observed. They were, however, exceedingly shy and difficult of capture, and our efforts to net them were entirely ineffectual. 56 INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. In most of the streams seined by Mr. Woolman and the writer in western Montana, Idaho, and Washington the young, parr-marked whitefish was taken. These young fish were obtained in several quiet streams, almost sluggish, so weak was the current. In Spokane River at the city of Spokane large numbers of adult whitefish could be seen from the city bridges. They would lie or swim close to the bottom, keeping in the shade of the bridges, and would bite at grasshoppers. The artificial fly seemed to have no attraction for the fish. This whitefish grows to a length of 15 inches, has excellent flesh, and is by many confused with the grayling, Thymallus. The Columbia River chub, Mylocheilus cau- rinus is often called ‘“‘whitefish.” The type of Coregonus williamsoni was obtained in the Des Chutes River, Oregon, and described by Dr. Girard in the Proceedings of the Philadelphia Academy of Sciences in 1856, and again in 1858 in the Pacific Railroad Survey reports. Later recorded localities are: Willamette and Columbia rivers, Oregon; the Columbia and its tributaries in Washington; Lake Tahoe, Trout Creek and Truckee River, California; lakes and streams of Idaho, Montana, Wyoming, Colorado, and Utah. The species ranges north into British America and eastward to the moun- tain tributaries of the Upper Missouri, several of the localities given by Prof. Evermann being east of the Continental Divide. ANNOTATED LIST OF REPTILES AND BATRACHIANS. [By John Van Denburgh, student, Leland Stanford Junior University.] But few reptiles and batrachians were observed by members of the expedition, partly because no special effort was made to collect them, and partly becatise they were not abundant in most of the region traversed. Both lizards and snakes seemed to be rare on the barren lava plains of the Snake River, where Sceloporus graciosus, Phrynosoma douglassi, and Phrynosoma platyrhinus were the only species seen. The case was different in the bottom lands of the Snake River below Shoshone Falls, Idaho. Here lizards were extremely abundant and in considerable variety. Species of Uta, Crotaphytus, and Onemidophorus were as numerously represented as on the hot deserts of southern California. Two species of Sceloporus were also seen, but of these no specimens were secured. The richness of reptilian life on the sandy floor of the valley contrasted strongly with the reverse condition on the lava plains immediately bordering the valley and but a few hundred feet above it. 1. Crotaphytus wislizenii Baird & Girard. Two full-grown specimens collected in the bottom lands of the Snake River near Bliss, Idaho, August 8. With the exception of Cnemidophorus tigris, this was the most abundant species observed. 2. Uta stansburiana Baird & Girard. A single badly mutilated specimen from Snake River bottoms, near Bliss, taken August 8. The species was not rare, and was usually observed on rocks. 3. Sceloporus graciosus Baird & Girard. Six specimens from near Pocatello, Idaho, August3; two specimens from near Idaho Falls, Idaho, August 5; three specimens from near Umatilla, Oreg., August 11. This species was always observed on the ground, and was seen both among the coarser basalt and on the finer sands and gravels. 4 Phrynosoma douglassii (Bell). This pygmy horned toad seemed by no means abundant. The largest specimen, 68 min. long, was taken at Pocatello, Idaho, August 4. Two other specimens were captured: one on the Clearwater River, 7 miles above Lewiston, August 15; the other at North Yakima, Wash., August 23. INVESTIGATIONS IN THE COLUMBIA RIVER BASIN. 57 5. Phrynosoma platyrhinos Girard. One specimen from Bliss, Idaho, August 8; two—a male and a female—from the lava plains between Shoshone and the Snake River, August 7. In the latter locality the species was abundant. The two specimens there secured have the series of enlarged gulars almost obsolete, it being represented on each side by two or three scales slightly more pointed than the other gulars, The number of femoral pores is 8 in the male, 10 in the female. The tympanum is fully scaled in one, only partially so in the other. 6. Cnemidophorus tigris Baird & Girard. The ‘‘sand lizard” was the most abundant species in Snake River bottoms. A single specimen was collected near Bliss, Idaho, August 8. This seems not to differ from Owen’s valley specimens, and we therefore follow Dr. Stejneger in the use of the above name. 7. Thamnophis vagrans (Baird & Girard). Although the six garter snakes brought in by the expedition show considerable color variation, they unquestionably represent a single species. They were collected at the following localities: 1 specimen, Sand Point, Idaho, August 7; 2 specimens, Clearwater River near Lewiston, Idaho, August 15; 1 specimen, Potlatch Creek near Juliaetta, Idaho, August 16; 1 specimen, Wardner, Idaho, August 18; 1 specimen, Umatilla, Oregon, August 23. 8. Thamnophis sirtalis parietalis (Say). One specimen of this variety was obtained on the Little Spokane River near Spokane, in September, 1892, by Mr. A. J. Woolman. 9. Crotalus lucifer Baird & Girard. A single specimen taken on the Snake River between Twin Falls and Shoshone Falls, Idaho. The rattler is said to be not abundant in that vicinity. 10. Diemyctylus torosus (Esch). Two specimens found in Skookumchuck River near Chehalis, Wash., August 28. 11. Bufo columbiensis Baird & Girard. One young specimen, Umatilla, Oregon, August 23,-and one adult, Pocatello, Idaho, August 3. 12. Hyla regilla Baird & Girard. One specimen, Chehalis, Wash., August 28. 13. Rana aurora Baird & Girard. Four specimens from Skookumchuck River near Chehalis, Wash., August 28. 14. Rana pretiosa Baird & Girard. Seven specimens, Sand Point, Idaho, August 7; 1 specimen, Pendleton, Oreg., August 12; 2 specimens, Colfax, Wash., August 17; 1 specimen, Little Spokane River, Dart’s Mill, Washington, August 18; 4 specimens, Caur d’Alene, Idaho, August 21; 1specimen, Post Creek, Montana, September 18, 1892; 1 specimen, Thompson Falls, Montana, September 19, 1892; 5 specimens, Sand Point, Idaho, September 20, 1892; 7 speci- mens, Little Spokane River, Spokane, Wash., September, 1892; 3 specimens, Clark Spring, Spokane, Wash., October 5, 1892. 15. Rana pipiens brachycephala (Cope). One specimen of this handsome frog was secured in the Boise River, at Caldwell, Idaho, August 8. The collection made by Messrs. Bean and Wool- man contains three specimens from Post Creek, Montana (September 18), and one from Sand Point, Idaho (September 20). i) wuss PLATE X. (To face page 58.) Salmon Fisheries of the Columbia River. PART OF THE UPPER FALLS. SPOKANE FALLS, SPOKANE, WASHINGTON. PLATE XI. (To face page 58.) Salmon Fisheries of the Columbia River. SPOKANE FALLS, SPOKANE, WASHINGTON. FROM MONROE STREET BRIDGE. I ME phn he PLaTE XIl. (To face page 58.) Salmon Fisheries of the Columbia River. UPPER SPOKANE FALLS, SPOKANE, WASHINGTON. FROM POST STREET BRIDGE, Salmon Fisheries of the Columbia River. (To face page 58.) PLATE XIII. LOWER KETTLE FALLS, COLUMBIA RIVER. Pa Chih a Y