^Tnrrrr?- i! v~" \i - Ir^sir gyyv n m l i^UJLjirf ZOOLOGICA n SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME XXIV 1939 Numbers 1-31 PUBLISHED BY THE SOCIETY THE ZOOLOGICAL PARK, NEW YORK Jteto |9orfe Zoological Soctetp General Office: 90 Broad Street, New York City Officers President, W. Redmond Cross Vice-Presidents, Alfred Ely and Laurance S. Rockefeller Chairman, Executive Committee, W. Redmond Cross Treasurer, Cornelius R. Agnew Secretary, Fairfield Osborn Scientific Staff Zoological $3ark W. Reid Blair, Director Raymond L. Ditmars, Curator of Mammals and Reptiles Lee S. Crandall, Curator of Birds Leonard J. Goss, Veterinarian Claude W. Leister, Ass’t to the Director and Curator, Educational Activities H. C. Raven, Prosector Edward R. Osterndorff, Photographer William Bridges, Editor and Curator of Publications Aquarium Charles M. Breder, Jr., Acting Director Christopher W. Coates, Aquarist Ross F. Nigrelli, Pathologist G. M. Smith, Research Associate in Pathology Homer W. Smith, Research Associate in Physiology ©epartmrnt of tropical &escarcfj William Beebe, Director and Honorary Curator of Birds John Tee-Van, General Associate Gloria Hollister, Research Associate Jocelyn Crane, Technical Associate Cbitorial Committee Fairfield Osborn, Chairman Charles M. Breder, Jr. William Bridges W. Reid Blair William Beebe CONTENTS Part 1. April 27, 1939. PA 1. On Sphyrion lumpi (Kroyer), a Copepod Parasite on the Redfish, Sebastes marinus (Linnaeus), with Special Ref- erence to the Host-Parasite Relationships. By Ross F. Nigrelli & Frank E. Firth. (Plates I-IV; Text-figures 1-3) 1 2. Notes on the Functions of the Forebrain in Teleosts. By R. G. Meaoer. (Text-figures 1 & 2) 11 3. The Panama Canal as a Passageway for Fishes, with Lists and Remarks on the Fishes and Invertebrates Observed. By Samuel F. Hildebrand. (Plates I & II) 15 4. The Cytology of the Pituitary Gland of Two Varieties of Goldfish {Car as sins auratus L.), with Some Reference to Variable Factors in the Gland Which May Possibly Be Related to the Different Morphological Types. By Irving Levenstein. (Plate I) .. 47 5. Notes on Plumage Changes in the Quetzal. By Lee S. Cran- dall. (Plate I) 6l Part 2. July 31, 1939. 6. Deep-sea Fishes of the Bermuda Oceanographic Expedi- tions. Family Melanostomiatidae. By William Beebe & Jocelyn Crane. (Text-figures 1-77) 65 7. The Eye Structure of the Four-eyed Blenny, Dialommus f metis Gilbert. By C. M. Breder, Jr., & E. B. Gresser. (Plates I & II; Text-figures 1-3) 239 PAGE 8. Studies on Virus Diseases of Fish. III. Morphological and Experimental Observations on the Lymphocystis Disease of the Pike Perch, Stizostedion vitreum. By Richard Weissenberg. (Plate I) 245 9. Studies on Lymphocystis Disease in the Orange Filefish, Ceratacanthus schoeppi (Walbaum), from Sandy Hook Bay, N. J. By Ross F. Nigrelli & G. M. Smith. (Plates I-VIII) 255 Part 3. November 2, 1939. 10. Report of the Hospital and Laboratory of the New York Zoological Park, 1938. Mortality Statistics of the So- ciety’s Collection. By C. R. Schroeder 265 1 1 . Carcinoma of the Pancreas of Acinar Origin in a Bear. By Philip J. Kresky & Roy N. Barnett. (Plates I & II) _ 285 12. Congenital Macrofollicular Cystic Colloid Goiter in a Dromedary. By Leonard E. Finkelstein. (Plate I) 289 13. Pentatrichomonas macro pi Tanabe from Kangaroos. By Carlton M. Herman. (Text-figure 1 ) 293 14. A Report on the Dental Pathology Found in Animals that Died in the New York Zoological Park in 1938. By Theodore Kazimiroff. (Plates I-IX) 297 15. A Parasitological Survey of Wild Rats in the New York Zoological Park. By Carlton M. Herman 305 16. Serologic Relationships Among Bovidae and Cervidae. By Harold R. Wolfe 309 VI PAGE 17. Multi ceps serialis Infestation in a Baboon. Report of a Case Exhibiting Multiple Connective Tissue Cystic Masses. By Stephen R. Elek & Leonard E. Finkel- stein. (Plates I & II) 323 18. A Quantitative Study of the Testes of Certain Mammals. By Thomas H. Knepp. (Text-figures 1 & 2) 329 19. A Preliminary Report on the Os Opticus of the Bird’s Eye. By Otto W. Tiemeier. (Text-figures 1-24) 333 20. Treatment of Amoebic Dysentery in an Orang-utan. By Carlton M. Herman & Charles R. Schroeder 339 21. Deltokeras multilobatus, a New Species of Cestode (Parau- terininae : Dilepiididae) from the Twelve-wired Bird of Paradise ( Seleucides m. melanoleucus (Daudin) : Passeriformes). By O. Wilford Olsen. (Plate I) 341 22. The Urinary Nitrogen Distribution of Representative Members of the Carnivora. By Richard W. Jackson, Thomas J. Dring & Charles R. Schroeder. (Text- figures 1-5) 345 23. A New Nematode, Ascaris schroeden, from a Giant Panda, Ailuropoda melanoleuca. By Allen McIntosh. (Text- figures 1 & 2) 355 24. A Study of the Yellow-lipped Snake, Rhadinaea jlavilata (Cope). By Edmond Malnate. (Plate I; Text-fig- ure 1) 359 25. Variations in the Nesting Habits of Ameiurus nebulosus (Le Sueur) . By C. M. Breder, Jr. (Plates I & II; Text- figures 1-3) 367 vii PAGE 26. The Occurrence of Trematode Ova, Hapalotrema constric- tum (Leared), in Fibro-epithelial Tumors of the Marine Turtle, Chelonia mydas (Linnaeus). By G. M. Smith & C. W. Coates. (Plates I-V) 379 Notes: A Record Size (480 mm.) John Dory ( Zenopsis ocel- latus ) with Notes on Its Distribution in Our North and Middle Atlantic Waters. By Frank E. Firth 383 Part 4. December 22, 1939. 27. Studies on Fish Parasites of Lake Erie. Distribution Studies. By Ralph V. Bangham & George W. Hunter, III 385 28. Young Me galops cyprinoides from Batavia, Dutch East India, Including a Study of the Caudal Skeleton and a Comparison with the Atlantic Species, Tarpon atlanticus. By Gloria Hollister. (Text-figures 1-21) 449 29. The Feeding and Grooming Habits of the Galago. By Florence de L. Lowther. (Plate I) 477 30. Parasites Obtained from Animals in the Collection of the New York Zoological Park during 1938. By Carlton M. Herman 481 31. On the Life History and Development of the Sponge Blenny, Paraclinus marmoratus (Steindachner). By C. M. Breder, Jr. (Plates I-IV ; Text-figures 1-3 ) 487 viii Index to Volume XXIV 497 ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME XXIV Part 1 Numbers 1-5 PUBLISHED BY THE SOCIETY THE ZOOLOGICAL PARK, NEW YORK April 27, 1939 CONTENTS PAGE 1. On Sphyrion lumpi (Kroyer), a Copepod Parasite on the Redfish, Se bastes marinus (Linnaeus), with Special Ref- erence to the Host-Parasite Relationships. By Ross F. Nigrelli & Frank E. Firth. (Plates I-IV; Text-figures 1-3) 1 2. Notes on the Functions of the Forebrain in Teleosts. By R. G. Meader. (Text-figures 1 & 2) 11 3. The Panama Canal as a Passageway for Fishes, with Lists and Remarks on the Fishes and Invertebrates Observed. By Samuel F. Hildebrand. (Plates I & II) 15 4. The Cytology of the Pituitary Gland of Two Varieties of Goldfish ( Car as sins auratus L.), with Some Reference to Variable Factors in the Gland Which May Possibly Be Related to the Different Morphological Types. By Irving Levenstein. (Plate I) 47 5 . Notes on Plumage Changes in the Quetzal. By Lee S. Cran- dall. (Plate I) 61 ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY 1. On Sphyrion him pi (Krnyer), a Copepod Parasite on the Redfish, Se bastes marinus (Linnaeus), with Special Reference to the Host-Parasite Relationships. Ross F. Nigrelli N. Y. Aquarium & Frank E. Firth U. S. Bureau of Fisheries! (Plates I-IV; Text-figures 1-3). Introduction. Many parasites of fishes have little or no appreciable effects on the host tissues. A few, however, are known to be deleterious ; either they kill their hosts eventually or result in secondary infections of more virulent types of organisms. More often they weaken the fish so that they become easy victims for their predators. A large number of parasites, especially ex- ternal forms, render the flesh, in the case of commercially valuable fishes, unmarketable. Although many species of copepod parasites have been described from fishes, and although the general effects of these organisms on the hosts have been recognized, little is known about the histo-pathology of many such infections. It is the purpose of this paper, therefore, to describe the host- parasite relationships of a copepod infection occurring in the redfish, Sebastes marinus (Linnaeus). In recent years (since 1931), this species of fish has become important commercially, millions of pounds being caught annually and prepared for market. The copepod infecting the fish, referred to as Sphyrion lumpi, may be found attached to any part of the body, but usually it is buried in the region of the dorsal musculature. Description of the Parasite. The copepod found on the redfish belongs to the family Sphyriidae Wilson, 1919. It was first described by Kr0yer (1845) as Lestes and later mm 1939 [i] 2 Zoologica: New York Zoological Society [XXIV :1 (1863) as Lesteria. It was Bassett-Smith (1899) who gave it the correct name Sphyrion lumpi (Kr0yer). The parasite has been observed in limited numbers on the lump fish, Cyclopterus lumpi, from the Danish coast, by Kr0yer (1845, 1863) and Steenstrup (1869) ; from the British seas by Bassett-Smith (1899), T. and A. Scott (1913) and Leigh-Sharpe (1933). Scott (1905) and Leigh-Sharpe (1933) have also reported the copepod on the wolf fish, Anarrhichas lupus. Wilson (1901) recorded Sphyrion lumpi from Nematonurus goodei, Haloporhyrus viola and a salted hake. Again Wilson (1919, 1931) was the first to report it from redfish of the North Atlantic coast of the United States. The members of the genus Sphyrion are easily recognized by the presence of a cephalothorax, the anterior end of which, in females, is ex- panded transversely to form a process that has been referred to as the sphyra or “hammer” (Plate I, Figs. 1, 2). From the center of the anterior surface of this expanded structure projects the head with its modified appendages. The neck is elongated and smooth while the trunk broadens transversely and in most cases is depressed ventrally. These parasites have no abdomen, but there is a pair of knob-like caudal rami. The copepods are also characterized by the presence of multibranched processes extending from the trunk. The function of these structures is still uncertain. The ovisacs are long and smooth and extend beyond the trunk. In young females, the head appendages are discernible. There are two pairs of maxillae and a pair of maxillipeds which in the more matured parasites become transformed or replaced by simpler processes and in some cases may even be absent. The male Sphyrion is comparatively minute, measuring a little more than 2 mm. in length and 1 mm. in width. The appendages present are the same as those seen in young females. The head is separated from the trunk and there is a small carapace present. The trunk is folded and lacks the arborescent appendage described above. The present forms of Sphyrion lumpi were taken from the body of redfish caught by commercial fishermen off the coast of Maine. They agree in all essential details with the description given by Wilson (1919, 1931), although the measurements of the various parts of the body are slightly different. These are shown in Table I. Table I. Hammer Length Width Neck Length Width Trunk Length Width Thickness Post. Proc. Length Ovisacs Length Width Wilson1 13-16 mm. 10 35-15 2.5 16-12 12 6-5 8-16 20 2.5 Nigrelli Si Firth 8-18 2-11 10-30 2-3 9-12.5 10-15 3-8 6-20 20-46 2-3 In young females, the body is more or less transparent, a condition which disappears as the copepod becomes more matured and the body struc- tures thicken. During the growth process, the posterior arborescent struc- ture becomes more and more branched. From the measurements given above, it can be readily noticed that there is quite a variation in the size of the different parts of the body. There are other differences not dis- closed by these figures (Plate I, Fig. 2). Thus, the terminal parts of the “hammer” are usually swollen in bulb-like form. In some individuals the width of these ends is not much larger than the median parts of the “hammer,” while in others they may be considerably enlarged, more or less 1 Wilson’s measurements based on two matured specimens. Our measurements based on 20 specimens showing ovisacs. In these measurements length of hammer is transverse to body axis. 1939J Nigrelli & Firth: On Sphyrion lumpi 3 round, slightly bifurcated or even multilobed. In some forms, the long tubular neck may have a comparatively even diameter throughout its entire length, while in others it may be thicker towards the “hammer” or towards the trunk end; sometimes it is long and thin; at other times it is long and thick and often it is short and thick. The trunk may likewise differ. In a few individuals it appears as a transparent structure, so much so that the internal structures are visible to the naked eye. In others, the ventral region of the trunk is more or less depressed, while in still other individuals this ventral depression is lacking, being for the most part slightly raised above the surface much as it is on the dorsal side. Studies on the internal anatomy of Sphyrion lumpi from sectioned mate- rial agree with the description given by Wilson (1919). In the matured females the identity of the mouth appendages is entirely lost. The in- testine is narrow in the neck region, dilating considerably in the trunk to form a widened rectum. The peculiar processes described by Wilson (1919) for this part of the intestine were also observed in our material. However, besides these processes, the striking feature of the intestine is the presence of a comparatively large number of secretory gland cells found in the epithelium. The large amount of corpuscles and digested blood in the lumen of the intestine shows that these matured parasites feed almost ex- clusively on the blood of the host. The ovaries are two in number, more or less tubular, and situated in the anterior region of the trunk, one on each side of the intestine close to the body wall. The tubular cement glands are situated in the posterior half of the trunk, one below each ovary. The oviducts are coiled and suri-ounded by a thick layer of chitin. Each duct opens directly into the long chitinous ovisacs. The internal anatomy of the male was not studied. Although the complete life-history was not followed, due to a lack of living material, from studies of extremely young stages of female Sphyrion lumpi and from what we know of the life-cycle of closely related forms, the larvae (copepodid stage?) are the only forms in the cycle that can infect other parts of the body of the fish or other fish. Extremely small forms (2 mm.) found buried beneath the epidermis, beside being highly trans- parent, show well developed mouth parts, such as maxillae and maxillipeds. It is by means of these structures that the copepod moves about on the body of the host and burrows deeply into the skin and muscle. In such forms the sphyra is not well developed or only slightly so. The neck and trunk are distinct while what will be the arborescent appendage appears as two knob-like outgrowths. It is of interest to mention here that a large number of redfish examined for Sphyrion lumpi were also found infected with a gill species referred to as Chondrancanthopsis nodosus (see Wilson, 1931). Distribution of the Parasite on the Host. In the majority of the redfish examined, the parasites were found buried deep in the skin and musculature in the region dorsal to the lateral line, usually at the base of the dorsal fin. (Text-fig 1). As in the Lernaeidae (see Wilson, 1919), the parasites attempt to reach the dorsal blood vessel for their nourishment. Occasionally, however, parasites were found attached in the body wall in the region ventral to the lateral line. These invariably had penetrated the body cavity. Several specimens were found attached to the eye, anchoring in the anterior chamber. Quite a number of fish were found with the copepod attached to the bony plates of the pre-operculum and the operculum proper. Here, in the majority of instances, the soft bone was perforated so that the cephalothorax passed 4 Zoologica: New York Zoological Society [XXIV :1 Text-figure 1. Redfish, Sebastes marinus, showing typical infection with the copepod parasite, Sphyrion lumpi. into the branchial cavity and became attached to the gill arches, often de- stroying parts of the gill filaments. On a few occasions parasites were found pendant from the rim of the anal opening, anchoring between the rectum and the body wall of the host. The process involved in this penetration is not exactly known. It is assumed, since the young forms have well developed mouth parts, that they eat their way through the host tissues, for long before they reach any blood vessel there is quite a cavity formed. It is interesting to point out that these copepods have well developed digestive and secretory glands, and it is altogether possible that the secretion given off by these glands may aid the parasite in the penetration. The number of visible parasites found on each fish may vary somewhat. The largest number of matured parasites recovered from a single fish was 6. Actually, the damage done was considerable, for each copepod would eat and digest away tissue from 1 to 1.5 cm. in depth and about 1 cm. in width, some parasites even penetrating the entire width of the dorsal muscula- ture. However, the parasite population may be greater than that mentioned above, for any fish may harbor many minute copepods underneath the skin. Thus, one fish examined in the laboratory showed three external parasites and two sores. Microscopic examination of the skin revealed as many as twelve immature forms. Although the samplings for examination of parasites were taken at random and at various times of the year, there appears to be no definite correlation between temperature (time of year) and the number of visible parasites or the number of infected fish. It has been determined, however, that young parasites are present during the late summer and early fall. It is assumed that these grow and become visible by spring, reaching maturity by summer. It is during the latter periods that the parasites superficially appear to be more numerous. Geographical Distribution of Parasitized Redfish. In the North Atlantic coast of America the normal redfish population is distributed off the coast of Maine, Massachusetts and Nova Scotia. (See 1939] Nigrelli & Firth: On Sphyrion lumpi 5 71* 70° 69° 68° 67* 66° 65* 64.* Text-figure 2. Distribution of the redfish, Sebastes marinus. Distribution of the redfish, Sebastes marinus, showing parasitized areas. 6 Zoologica: New York Zoological Society [XXIV :1 Text-figs. 2 & 3). 2 Practically every area in 80 to 150 fathoms on or near the fishing banks supports a redfish population, with the exception of the southern rim and NE peak of George’s Bank. About 90% of the com- mercial catch ranges in size between 24 to 32 cm. from snout to fork of caudal fin in length, and it is among such fish that the copepods were re- covered. No redfish under 22-24 cm. were ever seen with parasites, while those under 16-20 mm. do not appear in the catches because most of them pass through the trawl meshes and the balance are culled out at the docks. Redfish are caught almost exclusively in the daytime, for fishing at night invariably yields only comparatively small catches. Why this should be is as yet not known. It was assumed that the fish might rise off the bottom and therefore not be caught by the deep set trawls of the fisher- men, or that they might scatter widely at night and congregate again during the day. The redfish appear to disperse also after spawning, and again in October and November. The spawning period seems to be concentrated between May 15 and July 15, but the season may extend from early May to August. It is probably during these periods and at the times when the fish are concentrated that the majority of infections occur. The incidence of infection should be higher at such times. That this is not demonstrable is not surprising for the reason that the presence of parasites is not recog- nized until they appear on the body surface or when sores are developed. One definite item that can be determined, however, is the fact that fishes from the banks off the southern tip of Nova Scotia appear free from the parasite, although the redfish population is nearly as large in this region, comparatively, as it is in the infective areas. Although a seasonal study of the distribution of the infections was made, there was no cor- relation. It definitely showed that fish from south of Nova Scotia, referred to as Brown’s Bank, are free of Sphyrion lumpi. Ecologically, this area appears to be characteristically different from the regions where parasitized fish occur, influenced chiefly by cold water currents. In this region the fish are on the average smaller than those found off the coasts of Maine and Massachusetts (Cape Cod). This is supported further by the fact that the growth rate of Brown’s Bank haddock is slower and these fish have a different scale pattei'n. Also, insofar as could be determined, there seems to be no mixing of redfish populations, for if they did, parasites would most certainly show up on fish from Brown’s area. The presence of a cold current from the NE is one factor that may possibly divide and separate the Canadian fish from those off our coast. The fish-parasite relationship is most interesting and is summarized as follows: from 4,971 individual redfish, or 52 samples from the same number of individual trips, examined for the copepod, the average infection for the inshore redfish grounds off Maine and Massachusetts, appears to be roughly 10%, while the average number of parasites per fish is approxi- mately 1%. The fish from the offshore or Channel grounds had an average of roughly 5% infection, with the number of parasites per fish the same as inshore. The redfish population off Nova Scotia was found to be consistently free of the infection. The striking feature of the data collected to date is the gradual reduction in the percentage infection as one leaves the coast of the United States. It must be realized, however, that these percentages refer only to those parasites visible to the naked eye. There is no doubt that the percentage would be greater under laboratory examination at any time, and very probably disproportionately greater in the late summer and fall, when more young are present than at other times. This study covers the first 11 months of 1938 and represents samples of 100 fish per trip in practically all cases. 2 Other areas, not shown on the map, where redfish are occasionally taken and from which no parasitized fish have been found, include Sable Island, Roseway, Liscombs and Banquero Banks. 1939] Nigrelli & Firth: On Sphyrion lumpi 7 The writers wish to thank Mr. Henry M. Bearse of the U. S. Bureau of Fisheries for some of the data presented in the above paragraph, and Mr. B. 0. Knake for the drawings. The Bureau of Fisheries is now carrying on research into the further details of the life history of the redfish and relationships between host and parasite. Pathological Effects of Sphyrion lumpi Infection. The majority of the parasites protrude from the body through a small opening in the skin of the host. The minute forms are indicated by small, dark, lumpy growths on the skin, the external openings of which are hardly visible. As the parasites increase in size, these openings become larger and larger until, in some cases, the skin and muscle are turned out, forming large flaps. (Plate II, Fig. 4). In such instances, the overlying epidermis and scales are lost. The striking feature of the histo-pathology of these infections is the tissue response of the host to the parasite. The buried part of the cepha- lothorax is encapsulated by a well defined wall composed of host connective tissue. (Plate III, Fig. 5; Plate IY, Fig. 8). In older infections, where the healing process is more or less completed, this connective tissue is fur- ther developed and entirely surrounds the cephalothorax of the parasite. (Plate II, Fig. 3). The general host response, however, is a typical in- flammatory reaction. The muscle bundles become interposed with blood elements and most conspicuous of all, by an extreme dilation of vessels. (Plate III, Fig. 6). In certain regions the tissue is oedematous, showing an exudate of leucocytes, erythrocytes, monocytes and plasma cells in a network of fibrin. (Plate IV, Fig. 7). The cellular elements between the muscle fibers are mostly leucocytes, chiefly small lymphocytes and neutro- philes. Occasionally, however, granulocytes, some distinctly eosinophiles, have been encountered. In certain regions many red blood cells are found free in the host tissue. The activity of the leucocytes is evident by foreign body inclusions in the cytoplasm of these cells. A similar type of host reaction is present wherever the parasite is attached. Those found around the anal opening cause an intense inflam- mation of the rectum and surrounding tissues. When the eye becomes the point of attachment, the anterior chamber is the place of anchorage. The cornea is split and stretched, resulting in a distortion of vision. A slight amount of fibrous connective tissue is developed and surrounds the anterior end of the parasite. Here again, a dilation of blood vessels has occurred and numerous leucocytes infiltrate the surrounding tissues. In two instances, small tumors were found in the region of the dorsal fins. These were more or less round in appearance and about .5 cm. in diameter. Although these fish are red in color, the tumors were black with some red pigmented cells intermingled. Microscopically, however, the melanophores and erythrophores are limited to the surface, while the re- mainder of the tumor was composed of dense fibrous tissues, infiltrated with all sorts of wandering cells, especially lymphocytes. The many sores encountered in these redfish are almost invariably the remains of what was previously the anchoring place for the copepods. A few are the sites in which immature parasites are implanted. The larger of these sores consistently contain a slightly viscid exudate. Micro- scopic examination of this fluid revealed the presence of many algae-like microorganisms, singly or in chains. These were round, ovate, elongated or rod-shaped. The elongated forms measured 18 x 12 /x. All such or- ganisms contained round greenish plastids, many of which were found free in the exudate. Beside these, many flagellates were present which also contained the greenish coccus-like bodies. Whether there is any 8 Zoologica: New York Zoological Society [XXIV :1 relation between the algae-like forms and the flagellates was not determined. Very little bacteria were present in these sores, although the usual host cellular elements were evident. These were leucocytes with ingested par- ticles in the cytoplasm, debris composed of dead cells and partly digested muscle fibers. Discussion. The pathological effects resulting from copepod infestations are not well known. Wilson (1917) reported that the female members of the family Lernaeidae become anchored in the flesh of the host by special organs of attachment. The irritation set up during the burrowing process results in the development of “a tough membrane or skin immediately around them, which adds considerably to the security of prehension. This tough membrane also envelops the free thorax or neck of the parasite and reaches as far as the real skin of the host.” In the case of fishes parasitized with Pennella these cysts may be found anywhere in the body, including the stomach, intestinal wall, the mesentery, liver and even the heart wall. “When the copepod dies the cyst simpiy shrivels and hardens still more, and such lumps, nearly as hard as bone, may be cut out of the liver or else- where alongside of the living ones!” Wilson further stated that the para- sites of the family Lernaeidae and even those of the Sphyriidae (Wilson, 1919) feed upon the blood of their host. He reported (1917), “The simple fact that they burrow through the flesh until their head and mouth are brought into close proximity with some large blood vessel, and sometimes penetrate into the heart itself, leaves us no choice but to conclude that the parasite is making sure of a copious food supply.” From our observations on Sphyrion lumpi, however, the parasites very seldom reach a large vessel, and in no cases did they reach as far as the dorsal aorta. On the other hand, what does happen in every instance, a strong irritation is set up, both mechanically and by means of secretions given off by the parasite, which result in an inflammatory reaction. There is a considerable dilation of blood vessels in this reaction, so much so that the fine capillaries, usually of microscopic size, become enlarged to such an extent that they are visible to the naked eye in killed and fixed material, and in many cases can be dissected out. Thus, the vessels widen and pour their fluid and cellular contents into the surrounding tissue until it be- comes oedematous and infiltrated with leucocytes and scattered erythrocytes. This, in the writers’ opinion, is the source of nutriment for these parasites and not any particular blood-vessel as it is supposed. The source of irri- tation is constant in so long as the parasite keeps growing and moving about. Once the growth process is completed, the parasite becomes an inert object which finally becomes surrounded by repaired host tissue, forming the so-called cysts observed in many cases. It is obvious, however, that in the early stages of growth, the young parasites must feed on muscle and other tissues beside blood. The allocation of parasitized redfish to a definite area off the coast of Maine and Massachusetts is the striking thing in the present studies. Why this should be we have as yet no definite explanation. That the fish do disperse at certain intervals has been more or less definitely established by investigators in the field. Yet nearby areas and especially those off the coast of Canada have yielded very few parasitized redfish and none in the samples investigated. The fish in the non-infective areas are apparently separated ecologically by a cold current, and on the average are smaller in size than those caught off the coast of Maine. This would indicate pos- sibly that the redfish population in both these areas are distinct and for some reason or other (temperature, current, etc.) the parasites are not present. 1939] Nigrelli & Firth: On Sphyrion lumpi 9 Summary. 1. Sphyrion lumpi (Kr0yer), a parasitic copepod found on the redfish ( Sebastes marinus ) is redescribed. 2. The host-parasite relationships are discussed and the pathological effects of the copepod infestation is described. 3. The geographical distribution of parasitized redfish is indicated. References. Bassett-Smith, P. W. 1899. A Systematic Description of Parasitic Copepoda found on Fishes, with an Enumeration of the Known Species. Proc. Zool. Soc. London. 438-507, pi. 26. Kr0yer, H. 1845. Danmarks Fishe. Kjpbenhavn. 1863. Bidrag til Kundskab om Snyltekrebsene. Naturhistorisk Tiidsshrift. 2: 75-426, pis. 1-18. Leigh-Sharpe, W. H. 1933. A List of British Fishes with their Characteristic Parasitic Copepoda. Parasitology, Suppl. to the Jour, of Hygiene , Chicago Press. 25: 109- 112. Scott, T. 1905. III. Observations on some Parasites of Fishes New or Rare in Scottish Waters. 23 rd Ann. Report Fish. Board, Scotland. Pt. 3: 108-119. 1900. Notes on Some Parasites of Fishes. 19t/i Ann. Report Fish. Board, Scotland. Pt. 3: 120-153, pis. 7-8. Scott, T. & A. 1913. The British Parasitic Copepoda. Vol. 1, Copepoda Parasitic on Fishes. Yol. 2, Plates. Steenstrup, J. 1869. Om Lesteira, Silenium og Pegesimallus, tre as Prof. Dr. H. Krpyer opstillede Slaegter as Snyltekrebs. Kongel. Danske Vidensk. Selsk. Forhandl. 179-202, 1 pi. Wilson, C. W. 1917. North American Parasitic Copepods belonging to the Lernaeidae with a revision of the entire Family. Proc. U. S. Nat. Mus. 53: 1-150, pis. 1-21. 1919. North American Parasitic Copepods belonging to the New Family Sphyriidae. Proc. U. S. Nat. Mus. 1920. 55: 549-604, pis. 50-59. 1931. The Copepods of the Woods Hole Region, Massachusetts. U. S. Nat. Mus. Bull. 158. pp. 1-623, 316 text-figs. 41 pis. 10 Zoologica: Neiv York Zoological Society EXPLANATION OF THE PLATES. Plate I. Fig. 1. Typical female Sphyrion lumpi, about 3X natural size showing collapsed ovisacs. All photographs by S. C. Dunton, N. Y. Aquarium. Fig. 2. Female copepods without ovisacs. Note variability in form and size of cephalothorax. The smaller and younger specimens are transparent. About 2X. Plate II. Fig. 3. An old female. One ovisac missing. The “hammer” part of the cephal- othorax is entirely surrounded with a thick fibrous covering. Fig. 4. Area just in front of the dorsal fin showing skin and muscle turned out. About 3X. Plate III. Fig. 5. Extreme low power photomicrograph of section through region of infec- tion, showing encapsulated parasite. Fig. 6. Low power photomicrograph showing inflammatory reaction to the copepod parasite infection. This reaction is indicated by the extreme dilation of blood vessels. Plate IV. Fig. 7. High power photomicrograph of one of the dilated vessels and surround- ing region filled with an exudate of lymphocytes, monocytes, plasma cells and enythrocytes. Fig. 8. Photomicrograph of same magnification as in Fig. 7, showing the develop- ment of fibrous connective tissue. NIGRELLI & FIRTH. PLATE I. FIG. 1. FIG. 2. ON SPHYRION LUMP1 (KROYER), A COPEPOD PARASITE ON THE REDFISH, SEBASTES MARINUS (LINNAEUS), WITH SPE- CIAL REFERENCE TO THE HOST-PARASITE RELATIONSHIPS. \ NIGRELLI & FIRTH. PLATE II. FIG. 3. FIG. 4. ON SPHYRION LUMPI (KROYER), A COPEPOD PARASITE ON THE REDFISH, SEBASTES MARINUS (LINNAEUS), WITH SPE- CIAL REFERENCE TO THE HOST-PARASITE RELATIONSHIPS. NIGRELLI & FIRTH. PLATE III FIG. 6. ON SPHYRION LUMPI (KROYER), A COPEPOD PARASITE ON THE REDFISH, SEBASTES MARINUS (LINNAEUS), WITH SPE- CIAL REFERENCE TO THE HOST-PARASITE RELATIONSHIPS. NIGRELLI & FIRTH. PLATE IV. V, »** • i * .• • * . ♦ '* ( -• 0 •4) ' *> % • . *> • fr ** • • »j * • • ' . • FIG. 7. FIG. 8. ON SPHYRION LUMPI (KROYER), A COPEPOD PARASITE ON THE REDFISH, SEBASTES MARINUS (LINNAEUS), WITH SPE- CIAL REFERENCE TO THE HOST-PARASITE RELATIONSHIPS. Meader: Functions of the Forebrain in Teleosts 11 2. Notes on the Functions of the Forebrain in Teleosts. R. G. Meader Section of N euro- Anatomy , Department of Anatomy, Yale University School of Medicine, and the Bermuda Biological Station for Research, Inc. (Text-figures 1 & 2). It has usually been assumed that the forebrain of fishes is used primarily for olfactory purposes. I wish to submit a report of some observations which, together with others recently published, suggest that the specialized fore- brains of some fishes may be of importance for other purposes as well. In the course of carrying out some experimental studies on the optic system of fishes at the Bermuda Biological Station it occurred to me to make lesions in the forebrains of a few species. The primary object of such lesions was to secure degeneration of fiber tracts arising in the forebrain. Ana- tomical investigations in one species, Holocentrus ascensionis (Osbeck), have revealed an elaborately organized telencephalon intimately connected by large and small bundles of fibers with other parts of the brain. In view of the complexity of this brain and in view of the current interest in the func- tions of the teleostean forebrain it seems worth while to summarize here the observations recorded in my protocols. The latter include notes on normal control animals, on individuals with one or both eyes enucleated, and on in- dividuals with forebrain lesions. The specimens of Holocentrus used ranged from 8 to 12 cm. in length. These squirrel-fish live very well in aquaria provided with running salt water and they withstand operative interference readily. For all operations they were anaesthetized by immersion in a solution of 1 part chloretone in 4,000 parts of sea water. They were then held in a damp cloth while a segment of the dorsum of the skull was removed to expose the forebrain. One or both hemispheres were extirpated or isolated lesions were produced. When the animals were replaced in sea water, they quickly recovered from the anaes- thesia. At first it was thought necessary to close the skull opening with some inert substance but such efforts were abandoned when the substances used failed to adhere and the animals showed no ill effects from the exposure of the brain. At no time was there evident any disturbance which could be attributed to the direct bathing of the brain by the sea water. Granula- tion tissue growing in from the periphery gradually filled the wound. After 54 to 61 days the fishes were decapitated and the heads were prepared for microscopic study to provide controls of the location and the extent of the lesions and to permit investigations of the degenerated fiber tracts. These preparations have shown that the lesions intended were made. In some cases the total forebrain, with the exception of the sessile olfactory bulbs, was removed. The figures illustrate such a case. A description (with figures) of the gross appearance of the brain and 12 Zoologica: New York Zoological Society [XXIV :2 Outline sketch of a lateral view of the brain of Holocentrus. The tissue removed in the most radical operations is indicated by broken lines. X8. of the anatomy of the optic system of this fish has previously been pre- sented by the author (1). A short account of the normal behavior of Holo- centrus was included, emphasizing the importance of vision in the life of this large-eyed, nocturnal teleost. In the aquarium, as in its natural habitat, it prefers the darker corners and recesses where during the daytime it lies more or less quiescent in the shadow of a rock. At dusk and after dark, and in dim light generally, it swims about more freely. For a few days after being placed in an aquarium most individuals remain in concealment and are with difficulty enticed into the open to feed. They notice food only when it is in motion and if anything prevents them from seeing it or securing it before it comes to rest, it will remain untouched. If food is dropped close to the place of concealment, it is usually seized in midwater. There is a notable absence of the investigative nibbling characteristic of so many fishes. In the course of a few days it is possible to induce them to leave their hiding places and come to the surface to take food from forceps or fingers. Inimical visual stimuli generally elicit the sudden erection of the rather large and spiny dorsal fin, a reaction which is usually followed by flight and conceal- ment. All of this behavior is modified in totally blinded individuals. They are less active and they are indifferent concerning the part of the aquarium they occupy. They tend to find some vertical surface against which they lean as they rest on the bottom. Food is found only with great difficulty even when it is placed close to the mouth or olfactory pit. Blinded fishes are excited by the proximity of food but are unable to localize it accurately enough to ob- tain it without aid. They react slowly and briefly to an object moving in the water close to the body but their dorsal fins are seldom erected except by direct tactile stimuli. Individuals blind on one side only illustrate even more strikingly the importance of vision to the squirrel-fish. Their activity is as great as that of intact animals and they respond to visual and tactile stimuli (including movements of objects in the water) presented on the intact side in the same way that normal fishes do. To tactile stimuli presented on the blind side, however, the reaction has a much higher threshold and the flight response is less both quantitatively and in temporal persistence. Such an observation 1939] Meader: Functions of the Forebrain in Teleosts 13 Outline sketch of a dorsal view of the brain of Holocentrus, showing the area extirpated (in broken lines). X8. leads one to suspect that the more lively response obtained from supposedly tactile stimuli presented on the normal side is really due to visual cues. The normal dorsal fin erection and flight response occur even when stimuli pre- sented on the intact side are separated from the body of the fish and from the water by the untouched glass walls of the aquarium. In no case did partial or complete removal of the forebrain of Holocen- trus have any apparent effect upon the elements of behavior noted above for the normal control animals. The operated individuals were just as active as the latter, reacted to feeding stimuli in the same way and could be trained to take food from my fingers. They reacted to inimical stimuli with a sim- ilar fin and flight response. When normal and operated individuals occu- pied the same aquarium, their only distinguishing characteristic was the head wound. It is surprising that Holocentrus, provided with such a specialized fore- brain, from the anatomical point of view, exhibits so little disturbance of normal behavior when the forebrain is removed. Other investigators (2) have found that decerebration of teleosts is followed by deficits in olfaction, in schooling reactions (3, 4), in breeding behavior (5), and by a rise in the stimulus threshold (6). A restriction to purely reflex types of response has also been noted (6). It may be that the difference between my observations and those of others can be explained in part by the differences in the habits of the species studied and in part by the adequacy of visual reflexes to carry out the solu- tion of all problems met by the squirrel-fish in uncomplicated aquarium life. Olfaction is relatively unimportant to it but an olfactory deficit would very probably be revealed by more refined methods of testing. Chemical cues alone were not sufficient for finding food, as they are for many other forms, whereas vision unaided by olfaction enabled the fishes to feed nor- mally. Very little is known of the social behavior of Holocentrus. Although it is a relatively individualistic fish, it does “school” on occasion. Mr. Louis Mowbray, Director of the Bermuda Aquarium, has told me that there is a seven-year cycle in the abundance of the squirrel-fishes in Bermuda waters. 14 Zoologica: New York Zoological Society At the times of great abundance these fishes swim along the shores in large schools. Under these conditions decerebrate forms might exhibit a deficit not otherwise evident. Inasmuch as the breeding habits of the squirrel-fishes are also unknown, it is obvious that this aspect of their social behavior cannot well be tested. It is possible that refined methods of investigation, such as those used by Hosch, would reveal some rise in the threshold of stimuli and some change in the patterns of response. The observations here reported, as well as many others, indicate that the problem of the function of the forebrain in fishes has many complica- tions. There is probably as great a variation in the cerebral physiology of the teleosts as in their cerebral anatomy. No other comparably limited group of vertebrates exhibits so wide a diversity of morphology in its nervous system. In this group, also, are to be found equally wide variations in behavioral habits, which are an expression of neural physiology. The ob- servations made on one form, therefore, cannot safely be generalized to apply to all forms. Careful studies must be made on many different species if we are to arrive at reliable conclusions. In the meantime, it should not be surprising that different species of fishes studied under different conditions (often those which are decidedly artificial) by different investigators appear to have varying deficits after forebrain extirpation. The work of previous investigators suggests that the forebrain may have many functions which vary in importance in the different fishes. For those species with relatively poor visual or tactile senses the olfactory function may be paramount. For those that normally swim chiefly in groups the forebrain may supply the necessary coordination. For those with other elaborate social behavior characteristics, such as special- ized breeding and brooding habits, it may provide the integration of sensory stimuli to produce the customary pattern of reproductive activity; and its massive fiber tract connections with the hypothalmus may possibly influence reproduction through the effect of stimuli on the hypophysis. In many cases it may prove to be true that the forebrain of teleosts, like that of higher vertebrates but to a lesser degree, is an integrative center capable of giving rise to non-stereotyped responses appropriate to the stimuli received. All of these functions, and many others, may be present to varying degrees in all fishes and yet not be evident in an experimental analysis because of in- adequate methods of testing or because their expression is masked by more prominent behavioral traits. If, however, one dares to predict physiology from anatomy, then the known anatomical variations in cellular distribution and in the fiber pathways relating the cells indicate a diversity of function of the forebrain in different groups of fishes. Literature Cited. 1. Meader, R. G. The optic system of the teleost. Holocentrus. I. J. comp. Neurol, 1934, 60, 361-407. 2. The literature in this field up to 1935 has been excellently summarized by J. ten Cate, Physiologie des Zentralnervensystems der Fische. Ergebn. Biol., 1935, 11, 325-409. 3. Noble, G. K. The function of the corpus striatum in the social behavior of fishes. Anat. Rec., 1935-36, 64, Suppl., 34. 4. Wiebalck, Ute. Untersuchungen zur Funktion des Vorderhirns bei Knochen- fischen. Zool. Anz., 1937, 117, 325-329. 5. Noble, G. K. Effect of lesions of the corpus striatum on the brooding behavior of cichlid fishes. Anat. Rec., 1937, 70, Suppl. 1, 58. 6. Hosch, Ludwig. Untersuchungen fiber Grosshirnfunktionen der Elritze (Phox- inus laevis) und des Grundlings ( Gobio fluviatilis) . Zool. Jb., Abt. Zool. Physiol.. 1936-37. 57. 57-98. Hildebrand: Panama Canal as a Passageway for Fishes 15 3. The Panama Canal as a Passageway for Fishes, with Lists and Remarks on the Fishes and Invertebrates Observed.1 Samuel F. Hildebrand United States Bureau of Fisheries. (Plates I & II). Contents Page Introduction 15 Remarks Concerning the Canal and Locks 16 The Locks as Physical Barriers 17 Difference in Salinity a Barrier 17 Fishes Using the Canal and Locks as Passageways 19 Invertebrates Observed and Collected 21 Gatun Locks 22 Pedro Miguel Locks 22 Miraflores Locks 23 Annotated List of Fishes from the Locks of the Panama Canal 24 Gatun Locks 24 Pedro Miguel Locks 30 Miraflores Locks 33 Introduction. Whether the Panama Canal serves as a passageway for fishes, per- mitting at least some of the species of the opposite oceans to cross the Isth- mus, has been a subject of conjecture ever since the Canal was built. It was questioned whether fish could successfully negotiate the locks, and if so whether any of them could endure the journey of about 40 miles through the fresh water between the locks at the opposite ends of the Canal. It is now possible to give limited information on these questions, and on the ani- mal life in the locks in general, as a result of observations and collections made in 1935 and 1937, together with subsequent study. This work was made possible largely through the financial aid given by Dr. Herbert C. Clark, Director of the Gorgas Memorial Laboratory, to whom the writer is greatly indebted also for numerous other courtesies. He is deeply appreciative also of the extensive help given by Dr. A. O. Foster of the Gorgas Memorial Laboratory, and in fact to the entire staff of that laboratory. Officers of the Panama Canal, as well as more than a few employees, too, i Published by permission of U. S. Commissioner of Fisheries. 16 Zoologica: New York Zoological Society [XXIV :3 gave valuable aid. Among them are Col. C. S. Ridley, Governor; Col. Glen E. Edgerton, Engineer of Maintenance; Maj. W. D. Styer, Assistant En- gineer of Maintenance; R. Z. Kirkpatrick, Chief of Surveys; E. D. Stillwell, Superintendent of Locks; H. M. Thomas and J. C. Myrick, Assistant Super- intendents of Locks; Fred Whaler, Carl G. Brown, S. A. Venable, R. A. Cauthors and many others who cannot be named for want of space. Special mention must be made of the extensive aid and many courtesies extended by W. H. W. Komp, medical entomologist, U. S. Public Health Service; and J. B. Shropshire, malariologist, U. S. Army. The writer is deeply grateful, also, to the several taxonomists, who identified various groups of animals collected. These specialists are named in the text in connection with the discussion of specimens identified by them. To these workers, who gave of their time and energy, the writer especially wishes to extend thanks. One or more specimens of the species herein named, whether fish or in- vertebrates, have been or will be deposited in the U. S. National Museum. Observations and collections were made in Gatun Locks, on the Atlantic side, during the early months of 1935, and in Pedro Miguel and Miraflores Locks, on the Pacific side, during the early part of 1937, when the writer was present to witness the dewatering of one side of each lock. The water is removed, partly by draining and partly by pumping, from the locks at intervals of about three years for the purpose of cleaning and generally overhauling them. Collections were made by others in the sides dewatered during the writer’s absence. A particularly fine collection was secured in the east side of Miraflores Locks under the supervision of Dr. A. 0. Foster of the Gorgas Memorial Laboratory. Representative specimens, as far as possible, were preserved from among the fishes and invertebrates stranded in the locks as the water was removed. Collections were made also in the fresh waters situated between the locks at the opposite ends of the Canal, as well as at several places on both shores of the Isthmus and some of the outlying islands.2 Remarks Concerning the Canal and Locks. For the convenience of the reader who has not seen the Panama Canal, it may be stated that the general direction of the Canal is somewhat west of north and east of south, though traffic is designated as “north and south bound.” The Atlantic terminus of the Canal is entered from Limon Bay. A ves- sel sailing from the Atlantic to the Pacific, that is, southbound, proceeds at sea level as far as Gatun, a distance of 6 or 7 miles from Cristobal, the Atlantic port. There the Gatun Locks are reached. These locks, like those at the other end of the Canal, are double, permitting two-way traffic, the two channels of the locks commonly being designated the east and west side. The Gatun Locks consist of three equal flights whereby a southbound ship is lifted approximately 85 feet, to the level of Gatun Lake, which lies beyond the locks. Gatun Lake is a large body of fresh water, having an area of about 196 square miles, created by damming the Chagres River. Though the deep water, that is, the course for vessels, is well marked with buoys and beacons in Gatun Lake, a definite channel or canal is not visible for a distance of somewhat more than 20 miles. Thereupon Culebra (or Gillard) Cut, is 2 The study of the fresh water fishes and data pertaining to the fresh waters has been com- pleted, and a report has been published, entitled, “A New Catalogue of the Fresh- Water Fishes of Panama” (See Field Mus. Nat. Hist. Pub., Zool. Ser., xxii, 1938, pp. 217-359). The results of a study of the data and specimens secured in the locks are set forth herein. Much of the rather large collection of marine species taken along the outside shores of the Isthmus and the islands, however, remains for future study. 1939] Hildebrand: Panama Canal as a Passageway for Fishes 17 reached. This cut, through the continental divide, is about 9 miles long, and leads to Pedro Miguel Locks. The Pedro Miguel Locks consist of a single flight, whereby a south- bound vessel is lowered from the approximately 85-foot level of Gatun Lake to a level of about 53 feet of Miraflores Lake, which lies below these locks. Miraflores Lake is a small body of fresh or brackish water (sometimes slightly brackish from lockage water when northbound traffic is heavy), scarcely 2 miles long in the direction of the Canal. A southbound vessel reaches the Miraflores Locks after crossing this small lake. The Miraflores Locks consist of two equal flights whereby a southbound ship is lowered to sea level. The vessel now follows a rather definite channel (canal) to Balboa, the Pacific port of the Canal, a distance of about 4 miles, and then enters Panama Bay. In passing a northbound ship through the Canal the processes described in the foregoing paragraphs are, of course, merely reversed. As already indicated the locks are double, making it possible to pass two vessels through them simultaneously either in the same or opposite directions. The Locks as Physical Barriers. The locks do not constitute serious physical barriers to fish, as ex- plained at some length by the writer in a paper entitled, “The Tarpon in the Panama Canal” ( Scientific Monthly, Vol. 44, Mar., 1937, pp. 245-246). Ob- viously fish may swim into the upper or lower chambers of the locks without meeting any obstruction whatever when the gates at the opposite ends of the locks are open. There they may remain more or less indefinitely, or they may follow the next ship through the locks. In the case of the Gatun Locks, with three flights, they could ascend from the lowest to the middle chamber with a southbound vessel, or descend to this chamber from the uppermost level with a northbound vessel. There they might remain for a time, or com- plete the transit through the locks with a single ship. It is understood, of course, that when the lock gates are open, as in passing a ship from one chamber to the next one, no physical obstruction remains to prevent the fish from following the vessel. Difference in Salinity a Barrier. The change in salinity from fresh to salt water or vice versa, depending upon the»direction a fish may be pursuing, in going through Gatun or Mira- flores Locks (this does not apply to Pedro Miguel Locks, as they are in fresh water), is a much more formidable barrier, to most fishes, than the locks themselves. That many marine fish enter the locks and go through a part of the way, at least, is evident from the large number present at each dewatering. Sev- eral marine and brackish water species appear to live in the locks indefinite periods of time, and a few probably are permanent residents. It is to be noted, however, that strictly fresh water species seem to avoid the locks, as very few individuals or species were present even in the fresh water of the Pedro Miguel Locks, and in the nearly fresh water of the upper chamber of Gatun Locks. The abundance of fish in the middle and lowest chambers of Gatun Locks and both Chambers of Miraflores Locks suggests that food probably is plentiful and that conditions otherwise are agreeable to a com- paratively large number of salt and brackish water species, as shown by the lists appended. The temperatures and particularly the salinities, as already pointed out, profoundly affect the animal life in different parts of the Canal. The tables 18 Zoologica: New York Zoological Society [XXIV :3 and some of the other data offered were very kindly furnished by R. Z. Kirk- patrick, Chief of Surveys of the Panama Canal. The temperatures, given in Table I, are a summary of records covering the period from 1908 to 1936 inclusive for Balboa and Colon, and from November, 1918, to December, 1936, inclusive for Gatun Lake. The period of time covered by the records of salinity given in Table II, was not furnished. The “Pacific Entrance (Inner Harbor)” and the “Atlantic Entrance (Inner Harbor)” temperatures, as well as the salinities, were taken respec- tively at the Balboa and Cristobal Docks. On the Pacific side a cold water period occurs during the dry season. Concerning this Mr. Kirkpatrick said : “Cold water period is from February to April ; it is caused by Antarctic colder water being welled-up over Panama Bay Bottom Shelf during these months.” As understood by the writer, the brisk trade winds blowing across the Isthmus from the Atlantic to the Pacific have something to do with the up-welling of cold water in Panama Bay, as they tend to drive the warm surface water off shore. Outside the inner harbor the water temperatures apparently drop con- siderably lower, as shown by some records kindly furnished by W. H. W. Komp, U. S. Public Health Seiwice. Mr. Komp took the water temperatures with a pocket thermometer at Amador Beach during the dry seasons of 1934, 1935, 1936 and 1937. Each season, exclusive of 1937, the temperatures ranged downward into the sixties, the lowest occurring in 1934 when the water along the beach, between 4 and 5 o’clock in the afternoon, ranged in temperature from 60.5° to 63° F. from Feb. 11 to 16. These low temperatures affect fishing profoundly in Panama Bay, as such important game fishes as the sailfish, marlin, and dolphin are missing during this season. On the other hand, the corbina (sea trout or weakfish) seem to become more numerous. The cold water probably causes the fish population to vary to some extent with the season also in Miraflores Locks. Tides and lockage water have a direct bearing on salinity at both ends of the Canal. Mr. Kirkpatrick stated : “Tidal ranges on the Pacific side vary between an elevation of +11.0 and — 10.5 feet, with a mean range of about 12.6 feet. . . . Tidal ranges on the Atlantic side do not exceed 24 inches.” He stated furthermore, “Inner harbor salinities and densities are affected, of course, by the down-lockage of fresh water from Miraflores and Gatun Lakes.” It is understood, of course, that the chambers of the locks in pass- ing ships through them are filled with water admitted from the lakes above them, which in each instance is fresh (except for the slight brackishness occurring at times in Miraflores Lake). Therefore, if traffic is heavy a large amount of fresh water reaches the sea level ends of the Canal, reducing the percentage of salinity. It is evident, then, that the water in the locks (exclusive of that in Pedro Miguel Locks, which are in fresh water) may vary from about the saline condition of the inner harbor, when the sea level gates are open, to a sort of half and half mixture of the harbor and lake water to almost fresh, as in the upper flight of Gatun Locks. Salinity records for the locks are not available, except for one day (June 10, 1935), and for Gatun Locks only. According to hydrometer readings furnished by Mr. Kirkpatrick (without making corrections for temperature) the chambers of the upper level were fresh. No appreciable amount of salt was indicated in the east chamber of the intermediate level immediately after the water had been lowered from a 71- to a 43-foot level, and only slight brackishness was evident in the west chamber of the same level after it had been filled from a 43- to a 71-foot level. The two lowest chambers, however, were decidedly brackish, the salinity varying from about 10,000 to 16,000 parts per million. At the Atlantic entrance (outside the locks) the water was about as salty as that shown for the “Atlantic Entrance (inner harbor)” in Table II. 1939] Hildebrand: Panama Canal as a Passageway for Fishes 19 Table I. Monthly water temperatures in degrees Fahrenheit, surface. (See text for periods of time covered). Locality Minimum Maximum Average Balboa Entrance (Inner Harbor) .... 69.3 84.9 80.0 Miraflores Lake 78.0 84.0 81.0 Gatun Lake 80.9 85.0 83.6 Atlantic Entrance (Inner Harbor) .... 77.6 85.3 82.1 Table II. Salinities, surface. Locality Parts per million Pacific entrance (Inner Harbor) .... 16,000 to 20,000 Miraflores Lake 100 to 3,000 Gatun Lake 5 to 20 Atlantic entrance (Inner Harbor) .... 18,000 to 20,000 It is claimed by employees of the Canal that when the Gatun Locks first were operated dead fish were seen in the locks from time to time, which presumably died from the change in salinity caused by filling the locks with fresh water from Gatun Lake. Dead fish no longer are seen. The em- ployees believe the fish have become “educated” to the necessity of avoiding fresh water. The extent to which marine fishes have invaded fresh water, neverthe- less, is remarkable, as shown by the large number of salt water species listed from fresh or nearly fresh water subsequently. This is true especially of Miraflores Lake where fresh and salt water species seemingly intermingle freely. Fishes Using the Canal and Locks as Passageways. The species that most logically would be expected to pass through the locks and possibly complete the transit from ocean to ocean, are those in- habiting more or less indiscriminately salt, brackish and fresh water. To this group of fishes in Panama belong some of the guavinas (Gobiidae), several species of snook or robalos (Centropomidae), some of the so-called marine mojarras (Gerridae), a few species of rancons or burros ( Poma - dasys) , and the tarpon ( Tarpon atlanticus) . Among the fishes named the tarpon definitely has completed the transit from the Atlantic to the Pacific, as 4 individuals were present in the lower chamber of Miraflores Locks (east side) when dewatered in 1937. When the gates to the lower flight of the locks are open, as they often are, when 20 Zoologica: Neiv York Zoological Society [XXIV :3 vessels are not actually in transit, nothing remains to prevent the fish from swimming into the sea level end of the Canal and out into Panama Bay. Tarpons, indeed, have been reliably reported from the Pacific sea level terminus of the Canal, though to date this fish does not seem to have been caught in Panama Bay. While tarpons are present in Gatun Lake at all times, there is as yet no evidence that this fish breeds there, as pointed out by the writer ( Scientific Monthly, Vol. 44, Mar., 1937, p. 242). Therefore, it may be assumed, for the present at least, that the fish came from the At- lantic (or Caribbean Sea) ; that they use Gatun Locks as a passageway to Gatun Lake whence they pass on through Culebra Cut, and the Pedro Miguel Locks into Miraflores Lake (where they are seen frequently), and then on through the Miraflores Locks. Among the guavinas, or fresh water gobies, which inhabit principally fresh and brackish water, Dormitator maculatus of the Atlantic slope and shores was taken in the lower chamber, that is, at sea level, of Miraflores Locks. Leptophilypinus fluviatilis, another species of the Atlantic side, also was taken in Miraflores Locks, though not in the lower chambers. This species was numerous in the Pedro Miguel Locks and one specimen was secured in Miraflores Lake. On the other hand, Gobiomorus maculatus, a species of the Pacific side, was secured in Gatun Lake in company with its near relative Gobiomorus dormitor, of the Atlantic side. Gobiomorus macu- latus is very common in Miraflores Lake, though it was not taken in Pedro Miguel Locks. It conceivably could have reached Gatun Lake without pass- ing through the locks, as a few small Pacific slope streams empty into the Canal above Pedro Miguel Locks. Finally, it seems probable that Eleotris pisonis, of the Atlantic side, has crossed over to the Pacific, as shown by some specimens taken in the lower chamber of Miraflores Locks, which appear to be hybrids, that is, a cross between Eleotris pisonis and E. picta. No typical examples of E. pisonis were taken, however, on the Pacific side. Among the snooks or robalos, some of which range from the shores far up fresh water streams, Centropomus parallelus, an Atlantic side species, was taken in Miraflores Lake. To reach this lake the fish had to pass through Pedro Miguel Locks. Centropomus pectinatus, which occurs on both coasts of Panama, also was taken in Miraflores Lake. Because of its natural dis- tribution this species may have come from either coast. No positive proof has been found, so far as the writer is aware, that any of the snooks breed in fresh water. Two species of mojarra (Gerridae), namely, Eucinostomus calif ornien- siss and Gerres cinereus, were taken in the locks and the latter also in Mira- flores Lake. However, as these species are common to both coasts of Pan- ama it is not known that they have traversed the Isthmus, though they probably pass through the locks freely. Specimens of rancon or burro ( Pomadasys ) were taken in Gatun Lake, and in Pedro Miguel Locks and Miraflores Lake. These specimens all appear to be one species, but it is not possible at this time to state whether they are Pomadasys crocro from the Atlantic or P. bayanus from the Pacific, two nominal species which may not be distinct. Though these fish seem to pass through the locks, it has not been determined whether they have crossed the Isthmus. Neither is it known that they breed in fresh water, though they frequent it. The small anchovy, Anchovia parva, present in large numbers in all three chambers of Gatun Locks in 1935, was common in both chambers of Miraflores Locks in 1937. As this species has not been recorded from the Pacific, it seems possible that a migration has taken place, though it was not taken in Gatun Lake, Pedro Miguel Locks nor Miraflores Lake. The silverside, Menidia ( Thyrina ) chagresi, though belonging to a 3 Though I have not been able to date to separate the Atlantic and Pacific coast specimens as to species, other investigators at least have attempted to do so. 1939] Hildebrand : Panama Canal as a Passageivay for Fishes 21 family whose members are mostly marine, lives in fresh and brackish water. It was found common in the Chagres Basin during our investigations in 1911 and 1912, before the opening of the Canal. In 1935 it was found in the middle and uppermost chambers in Gatun Locks, and numerous in Gatun Lake. In 1937 several specimens were secured in Pedro Miguel Locks and the lower chamber of Miraflores Locks, presumably a result of a migration from the Atlantic to the Pacific slope through Culebra Cut. Among the more or less strictly fresh water species the chogorro, Cichlasoma maculicauda, an Atlantic slope fish sometimes descending to slightly brackish water, seems to have crossed to the Pacific side, as it was taken in Pedro Miguel Locks, Miraflores Lake and the upper chamber of Miraflores Locks, as well as in the Gatun Locks. Two Atlantic slope species of “sabalo pipon,” Brycon chagrensis and B. petrosus, were taken on the Pacific side, the former in Pedro Miguel Locks and Miraflores Lake, and the latter in the Rio Cocoli, a short distance above Miraflores Lake. Only one of the numerous species and generally abundant “sardinas,” namely Astyanax fasciatus, a Pacific slope species, seems definitely to have crossed the divide through Culebra Cut, as it was taken in Gatun Lake. Other species of characins may have crossed through Culebra Cut. How- ever, as several species of this family are common to both slopes, crossing over cannot be determined from specimens. It was particularly surprising that the abundant “sardina,” Astyanax ruberrimus, of both slopes, which literally swarms everywhere in Gatun and Miraflores Lakes, did not occur in the locks. Not one specimen even was found in Pedro Miguel Locks, which are in fresh water, though it occurs in abundance above and below them. It was rather surprising also that Roeboides, another “sardina,” avoids the locks, wherein no specimen was secured. Furthermore, no crossing over nor intermingling of the two easily distinguishable species, guatemalensis of the Atlantic slope and occidentalis of the Pacific, through Culebra Cut, seems to have taken place, as shown by numerous specimens collected in Gatun and Miraflores Lakes. The pipefish, Doryrhamphus ( Oostethus ) lineatus, has been reported as having been caught “in transit through the Panama Canal” by Chickering ( Copeia , No. 173, 1930, p. 85). However, this fish probably inhabits chiefly fresh and brackish water as shown by many specimens taken during our investigation in 1911 and 1912 (Meek and Hildebrand, Field Mus. Nat. Hist. Pub., Zool. Ser., XV, Pt. I, 1923, p. 262), before the opening of the Canal, when none was secured in salt water. In March, 1935, the writer collected 9 specimens in a few hours seining along the shores of Barro Colorado Island, in Gatun Lake. Four of the specimens are males with abdominal pouches filled with eggs, showing that this fish breeds in the Lake, where it probably is a permanent residence. It was not seen in any of the locks, nor in Miraflores Lake. Therefore, it is not known that it frequents the locks, nor that it has crossed the divide through Culebra Cut. Invertebrates Observed and Collected. It is stated in the foregoing pages that some fishes are so numerous in parts of the locks at each dewatering (see appended lists for the relative abundance of the various species of fishes observed in the chambers of the locks) that they must find conditions agreeable. In this connection a brief account of the condition of each lock, together with remarks on the inverte- brates observed, seems desirable. Many of these lower forms of course are eaten by fish. The collections of invertebrates of necessity are incomplete, as the water drops rapidly in the dewatering process. After the floors of the dif- ferent chambers become exposed water remains only in the “manholes” in the floors, and in the sumps, at the gates. Some animals are stranded, but 22 Zoological New York Zoological Society [XXIV :3 many more of the free swimming forms manage to reach either the bottom holes or the sumps. In any event, collections must be made quickly. As the writer and his helpers were interested chiefly in securing a representa- tive collection of fishes, to which they gave most of their attention, more than a few invertebrates, even of the larger forms, no doubt escaped notice. Therefore, the collections of these lower forms must be considered far from complete. Gatun Locks: The walls and floors of Gatun Locks (east side), when de- watered in 1935, had not accumulated much sediment or rubbish since the previous overhaul in 1932, though sufficient slush was present on the bottom that the collectors were well covered with mud splattered by stranded fish. Neither were the growths on the walls, gates and floors especially luxuriant. The growth in the uppermost and middle chambers consisted mostly of a hydroid, identified by Prof. Charles McLean Fraser as Cordylophora lacus- tris Allman. This hydroid was most abundant in the nearly fresh water of the uppermost chamber, and little in evidence in the much saltier lowest chamber, wherein that growth was largely replaced by oysters and barnacles. Clusters of mussel-like bivalves, examples of which were identified as Brachidontes exustus Linnaeus by Dr. Paul Bartsch, who furnished identifi- cation also for the other molluscs mentioned herein, were present in the mid- dle and lowest chambers of Gatun Locks. The gastropod, N eritina meleagris Lam., was numerous in the uppermost chamber, less so in the middle one, and was not observed in the lowest chamber. Small crabs were present in each chamber of Gatun Locks. However, examples were preserved only from the uppermost one. These were identi- fied by Dr. Mary J. Rathbun, who identified the other crabs mentioned herein also, as Callinectes sapidus acutidens Rathbun, and who supplied the follow- ing note, “A marine species ranging from each coast of Florida to Rio de Janeiro, Brazil.” This crab was inhabiting the nearly fresh water in the uppermost chamber of Gatun Locks. Small shrimps, too, were present in the Gatun Locks. One species, identified as Macrobrachium acanthurus (Wiegmann) by Dr. Waldo L. Schmitt, who identified the other shrimps mentioned herein also, was taken only in the lowest chamber, though it may have been present in the others. Dr. Schmitt remarked, “A fresh water species, ranging from Florida to Brazil and Uruguay, and West Coast of Mexico to Ecuador.” Juveniles of another species, provisionally identified as Macrobrachium olfersii (Wieg- mann), were common in the uppermost chamber, and a third species, Cran- gon armillatus (H. Milne-Edwards) , was taken in the lowest chamber. Con- cerning the latter Dr. Schmitt remarked, “A marine species distributed from North Carolina and Bermuda to Brazil; and through the West Indies.” One specimen of Macrobrachium olfersii was infested with a bopyrid isopod in the right branchial cavity. This isopod was identified by J. O. Maloney as Palaegyge meeki Richardson. Pedro Miguel Locks : The Pedro Miguel Locks, which are in fresh water, though possibly at times very slightly brackish from lockage water from Miraflores Locks, contained mud several inches to a few feet deep (east side). The concrete walls and the iron gates, as high as the permanent water level, were almost entirely covered with a mussel-like bivalve, iden- tified by Dr. Bartsch as Congeria ( Mytilopsis ) sallei Recluz. On the floors of the locks, where objects for attachment were present, clumps of this bi- valve also occurred, and under and around the clumps amphipods, probably of the same species as the one from the upper chamber of Miraflores Locks, identified as Grandidierella megnae (Giles) by C. R. Shoemaker, were nu- merous. Unfortunately the specimens collected were lost. No other molluscs or amphipods were noticed, though they may have been present. Shrimps were rather common, and no doubt are fed upon by some of the fish that inhabit these locks. Dr. Schmitt identified 4 species, namely, 1939 | Hildebrand: Panama Canal as a Passageway for Fishes 23 Penaeus stylirostris Stimpson, Macrobrachium jamaicense (Herbst), M. acanthurus (Wiegmann), and Palaemonetes sp. Dr. Schmitt referred to the one first named as a marine species, the next two as well known fresh water shrimps, and the last one as most likely a fresh water form. No crabs were seen in these locks. Miraflores Locks : The fauna of the Miraflores Locks was much more diversified than that of the Pedro Miguel Locks, no doubt because the water ranges from salt to nearly fresh. The upper chamber contained fully as much sediment as the Pedro Miguel Locks, but the lower one contained much less. The walls of the upper chamber were almost as fully over-grown as Pedro Miguel Locks with the same species of bivalve, but in the lower chamber this mollusc was missing, presumably because of the higher salinity. This bivalve occurred also in clusters on the floor of the upper chamber wherever there were objects reaching above the bottom slush to which it could attach itself. Among these clusters were numerous amphipods, pre- sumably all of the species identified as Grandidierella megnae (Giles) by Mr. Shoemaker from the single specimen, of many collected, not lost in shipment. In addition to the abundant bivalve mollusc a scant growth of a hydroid, somewhat doubtfully identified (because of the unsatisfactory condition of the specimens) as Bimeria gracilia Clark by Prof. Fraser, was present in the upper chamber. In the lower chamber only a few clumps of the hydroid were seen attached to objects on the floor. The abundant growths of the bivalve and hydroid mentioned, of the upper chamber, were replaced in large part by barnacles in the lower one. A few other attached forms of which only scattered examples were seen, was the sponge, identified by M. W. de Laubenfels as a cosmopolitan form, Haliclona permollis (Bowerbank), and the alcyonarian, Leptogorgia alba Duchassaing & Michelotti, as identified by Miss Elisabeth Deichmann, who referred to it as “one of the most common forms in the lava pools off Panama.” In the lower chamber of Miraflores Locks the wooden beams against which the bottoms of the iron gates close were badly infested with teredo, of which no examples were secured. Only one mollusc, the gastropod, Thais kiosquiformis Duclos, in addition to the numerous bivalve already mentioned, was taken in the upper chamber. This gastropod was found also in the lower chamber with four others, identified by Dr. Bartsch as Phyllonotus radix Gmelin, Pustularia pustutala Lamark, Triumphis distorta Wood, and Cyma- tium ( Linatella ) wiegmanni Anton. Limpids were fairly common in the lower chamber. Examples of two species, Crepidula aculeata Gmelin and C. incurva Broderip, were preserved. Small squids were common in the sump at the lowest gates of the locks. The examples transmitted to the National Museum were identified as Loligo sp. Crabs and shrimps were in evidence in both chambers of Miraflores Locks and were numerous in the bottom holes after the water had been pumped somewhat below floor level. The crabs in particular were difficult to catch in these holes, as they clung closely to the walls from which they were not readily removed, and generally dived out of reach of a dipnet after some agitation. Because of the difficulty of catching crabs and shrimps, and more particularly because of lack of time, more than a few species surely were missed. Only two species of crab, identified by Dr. Rathbun as Panopeus rugosus A.M.E., and Callinectes arcuatus Ordway, were collected in the upper cham- bers. These two were taken also in the lower chamber with Panopeus chil- ensis M. Edw. & Lucas and Petrolisthes armatus (Gibbes). The occurrence of the blue crab, Callinectes sapidus Rathbun in fresh water recently was discussed by Gordon Gunter ( Science , Vol. 87, Jan. 28, 1938, p. 87). It is not surprising, therefore, that other species of the genus also enter brackish and fresh water. 24 Zoologica: Neiv York Zoological Society [XXIV :3 Seven species of shrimp were collected in the upper chambers of Mira- flores Locks, which Dr. Schmitt identified as Penaeus brevirostris Kingsley, Macrobrachium jamaicense (Herbst), M. acanthurus (Wiegmann), Palae- monetes sp., and two species of snapping shrimp, Crangon, unidentifiable as to species because of the “meager and incomplete material.” and because of “our too limited knowledge of the west American species.” Four species of shrimp collected in the upper chambers, namely Macro- brachium acanthurus (Wiegmann), Palaemonetes sp., and the two unidenti- fiables species of Crangon, were taken also in the lower chamber, with the three following species not secured in the upper one : Penaeus stylirostris Stimpson, P. occidentalis Streets and a third unidentifiable Crangon. Besides the shrimps, a stomatopod, Chloridella aculeata (Bigelow), was taken. Dr. Schmitt referred to Macrobrachium jamaicense (taken in Pedro Miguel Locks and the upper chamber of Miraflores Locks), and M. acan- thurus (found in all three levels) as “well known fresh water shrimps,” and he regarded it likely that Palaemonetes sp. (taken in all three levels) also is a fresh water form. Though these species live chiefly in fresh water they appear to enter brackish water, or even at times salt water. The rest of the shrimps apparently may be regarded as salt and brackish water forms. Annotated List of Fishes from the Locks of the Panama Canal. The nomenclature and sequence of families of the earlier works by Meek & Hildebrand (a. “The Fishes of the Fresh Waters of Panama,” Field Mus. Nat. Hist. Pub., Zool. Ser., X, No. 15, 1916, pp. 217-374; b. “The Marine Fishes of Panama,” Field Mus. Nat. Hist. Pub., Zool. Ser., XV, No. 215, Pt. I, 1923; No. 226, Pt. II, 1925; and No. 249, Pt. Ill, 1928, 1045 pages) have been followed as far as it seemed permissible to do so. In order to retain as far as possible this nomenclature and sequence, the rearrangement and splitting of some of the old families (as for example Siluridae, Serranidae, Sciaenidae, and Gobiidae), as well as some of the genera, by some recent writers have not been adopted. This course was followed chiefly for the convenience of those perhaps not entirely familiar with the nomenclature who may wish to check the lists against the descriptions and accounts in the earlier publications. The large number of marine species present in the locks and their great tolerance for fresh or nearly fresh water are interesting facts shown in these lists. GATUN LOCKS. The level or levels at which the specimens collected in the east side of the locks were taken are shown, without stating that they are from the east side. Those from the west side, collected by workmen, were not kept separate, and therefore are listed simply as from the “west side.” In other words, all specimens listed are from the east chambers of the locks unless otherwise stated. The collections in the east chambers were made by me and helpers from February 20 to 24, 1935. Those from the west side were made by workmen, as already stated, in January, 1935, before my arrival on the Canal Zone. The marine gobies listed in this paper were kindly identified by Isaac Ginsburg, who is a specialist on that group of fishes. Family Myridae. Worm Eels. Myrophis punctatus Liitken. Lowest Chamber: 1 specimen, 204 mm. long, picked up in bottom slush. 1939 J Hildebrand: Panama Canal as a Passageway for Fishes 25 Family Elopidae. Tarpon attentions (Cuvier & Valenciennes). Tarpon; “Sabalo real.” Uppermost Chamber: A half-dozen or so large ones, all removed by laborers, except one, before the writer could reach them. One female, about 150 cm. long, with small roe, examined. Middle Chamber: 2 males, ripe or nearly so, 100 and 106.5 cm. long; 6 females, respectively 97, 108, 109, 153, 167 and 199 cm. long, the largest one with well developed gonads, the others entirely undeveloped. The tarpon, locally known as sabalo real, i.e., king shad, is highly prized as a food fish by the natives and the West Indian negro immigrants. This fish occurs regularly in Gatun Lake and it has passed on to the Pacific locks (see lists). It apparently does not spawn in the fresh waters of the Canal. Therefore, it evidently passes through the locks, which it may use as a feed- ing ground also. (For a more extended account see “The Tarpon in the Panama Canal,” by the author, Scientific Monthly, Vol. 44, Mar., 1937, pp. 239-248, 4 figs.). The tarpon has a modified air bladder, which is developed somewhat as a lung, enabling it to breathe air at least in part. This adapta- tion presumably aids it in tolerating fresh water. Slops sanrus Linnaeus. “Bonyfish” ; Big-eyed herring. Middle Chamber: Hundreds of fish present, probably 1,000 pounds or more; 20 individuals (selected at random) consisted of 7 males, 56.5 to 62 cm. long; and 13 females, 63.5 to 73 cm. long; all with large roe, indicating that spawning time was near at hand. Lowest Chamber: About one-sixth as many fish as in middle chamber. West Side: Several, according to workmen; 2 preserved. The gillraker counts for 18 specimens are 13 and 14, showing that no crossing over had taken place, as the Pacific Coast bonyfish ( Elops affinis ) has 18 to 24 gillrakers on the lower limb of the first arch. There is, in fact, no evidence that this species enters the fresh water of the Canal. It seems probable, therefore, in view of the large number present, that this species uses the locks as a feeding ground. Though the closely related Pacific coast species is not uncommon in Panama Bay, not a single individual was seen in the locks at the Pacific end of the Canal. This fish was rejected by the natives and negroes as unfit for food, and the large quantity stranded had to be disposed of by burial. This species, nevertheless, is seen in the local markets from time to time. Family Engraulidae. Anchovies. Anchovia parva Meek & Hildebrand. “Sardina.” Numerous in all chambers, many in bottom holes of locks, mostly small, almost transparent; many specimens preserved. This anchovy was not taken in Gatun Lake, but specimens were secured in Miraflores Locks. This species heretofore was not known from the Pacific. Anchovia spinifer (Cuvier & Valenciennes) . “Sardina.” A single specimen, 88 mm. long, was taken in the locks, shortly after they were refilled, by Felipi Torris, and preserved by Dr. J. R. Martin. It was not seen when the locks were dewatered. This species, reported from both coasts of Panama, apparently is rare on the Atlantic coast. It was numerous in the locks on the Pacific end of the Canal. 26 [XXIV :3 Zoologica: New York Zoological Society Family Belonidae. Salt Water Garfish. Tyhsurus raphidoma (Ranzani). West Side: 1 mutilated specimen 62.5 cm. long. Family Atherinidae. Silversides. Menidia IThyrinal chagresi Meek & Hildebrand. Uppermost Chamber: 4 juveniles, 13 to 18 mm. long. Middle Chamber: 1 specimen 50 mm. long. This silverside is very common in Gatun Lake. So far as known, it does not enter salt water. Several specimens were taken in the locks at the Pacific end of the Canal, probably the result of a migration through the Canal. These fish were thought to be young tarpons by some of the native tarpon fishermen. Family Mugilidae. Mullets; “Liza.” Mugil brasiliensis Agassiz. Middle Chamber: 5 large fish, ranging in length from 56.5 to 86 cm.; 2 large females, respectively 82.5 and 86 cm. long, contained large roe. Although this mullet enters brackish water freely in Panama, it is not known to pass through the locks into Gatun Lake. Family Sphyraenidae. Barracudas. Sphyraena sp. Barracuda. None seen in 1935; reliably reported from earlier dewaterings, gener- ally one or two large ones present. Family Carangidae. Cavallas. Caranx hippos (Linnaeus). “Jack.” Middle Chamber: Many (500 or so) all large; 19 individuals (selected at random) consisted of 11 males, ranging in length from 69 to 88 cm., and 8 females, varying from 67 to 98 cm. in length; all except 4 males with large or developing roe. Lowest Chamber: About half as many as in middle chamber; 11 indi- viduals (selected at random) consisted of 4 males, 67.5 to 83 cm. long, and 7 females, 76 to 100 cm. long; all of these fish and a dozen or so others, cut open but not measured, contained large or developing roe, showing that spawning time was near at hand. West Side: 1 specimen 50 cm. long. This fish quite surely does not pass through the locks into fresh water. It, indeed, was not present in the uppermost chamber, though numerous in the middle one. The locks apparently are a feeding ground for the jack, which is valued locally as a food fish. The largest individual noticed was 100 cm. (40 in.) long, which may be near the maximum size attained. Though it is common also on the Pacific coast of Panama, it was not numer- ous in the locks at that end of the Canal. Selene vomer (Linnaeus). Moonfish. West Side: A single specimen 320 mm. long. 1939] Hildebrand : Panama Canal as a Passageway for Fishes 27 Family Centropomidae. Robalos; Snook. Cenfropomns undscimalis (Bloch). Robalo ; Snook. Middle Chamber: 1 specimen 275 mm. long. Most of the species of the genus ascend brackish and fresh water streams in Panama. However, this species was not taken in fresh water during the recent investigation, nor in earlier (1911 and 1912) investiga- tions. Ceniropamus paraHelus Poey. This species was not seen in the locks. However, three fish, respectively 250, 270 and 300 mm. long, were taken in Gatun Lake near Gamboa. It was reported from Barro Colorado Island by Breder ( Zoologica , IX, 1933, p. 568). Since there is as yet no evidence that this Atlantic coast fish breeds in fresh water, it apparently may be assumed that the specimens collected had passed through the Gatun Locks. Specimens were secured in Miraflores Lake, indicating a migration through Pedro Miguel Locks. Family Serranidae. Sea Basses. Promicrops itaiara (Lichtenstein). Spotted jewfish. Middle Chamber: 1 specimen 208 mm. long. Lowest Chamber: 1 specimen, juvenile, 24 mm. long; identification un- certain because of extreme youth. The jewfish, which is common to both coasts of Panama, occurred also in the Miraflores Locks. Ryptieus sapomaceus (Bloch & Schneider) . Soapfish. Lowest Chamber: 3 specimens, respectively 133, 140 and 150 mm. long; picked up from bottom slush. Many more were present. West side: 6 specimens, 123 to 145 mm. long. Family Lutianidae. Snappers. Lutianus g risen s (Linnaeus). Gray or mangrove snapper. Lowest Chamber: 3 specimens, 100 to 110 mm. long, preserved. Exam- ined six othei’s ranging in length from 370 to 540 mm. for spawning condi- tion, all of which were undeveloped. This species was scarcely as numerous as the dog snapper. West Side: 2 specimens, 150 and 225 mm. long. Lutianus jo cm (Bloch & Schneider). Dog snapper. Middle Chamber: 3 specimens, 60, 235 and 260 mm. long. Lowest Chamber: 16 specimens preserved, ranging in length from 34 to 310 mm. Many others were present; examined six ranging from 270 to 335 mm. in length for spawning condition, which were all undeveloped. West Side: 8 specimens, varying in length from 62 to 300 mm. Lutianus apodus (Walbaum). Schoolmaster. Middle Chamber: 2 specimens, respectively 205 and 250 mm. long. Lowest Chamber: 4 specimens preserved, ranging in length from 245 to 290 mm. Several others were present, but the species was less numerous than the dog snapper. Lutianus synagris (Linnaeus). Lane snapper. Lowest Chamber: Only 4 small ones, ranging in length from 106 to 152 mm. West Side: 1 specimen 105 mm. long. 28 Zoologica: New York Zoological Society [XXIV :3 Family Haemulidae. Grunts. Pomadasys crocro (Cuvier & Valenciennes). Ranco. Uppermost Chamber: 14 small specimens, 36 to 57 mm. long. This fish was taken also in Gatun Lake, near Gamboa. In earlier investi- gations (1911 and 1912) specimens were collected all the way from strictly salt water at Porto Bello to fresh water, above many rapids, in the upper Chagres River. It seems probable that this fish passes through the locks, as no evidence has been secured indicating that it reproduces in fresh water. Specimens of this species, or the closely related P. bayanus, were taken in Pedro Miguel Locks and Miraflores Lake. Family Sparidae. Porgies. Archosargus aries (Cuvier & Valenciennes). Sheepshead. West side: 1 specimen 325 mm. long. This fish is new to the fauna of Panama. The species originally was described from Rio de Janeiro and Maracaibo, and later recorded from Belize, Honduras. Family Gerridae. Marine Mojarras. Eucinostomus calif orniensis (Gill). Uppermost Chamber: 11 juveniles, 24 to 32 mm. long. This species, as here understood, is common to both coasts of Panama. Specimens were taken also in Miraflores Locks and Miraflores Lake, but none at intermediate points. Though it ranges into brackish water streams, it is not known to enter strictly fresh water in Panama. Diapterus plumieri (Cuvier & Valenciennes). Uppermost Chamber: 1 individual, quite surely of this species, was seen in a bottom hole but not captured. Middle Chamber: 3 specimens, respectively, 270, 270 and 315 mm. long preserved. Others were seen. West Side: 2 specimens, 330 and 340 mm. long. Family Sciaendae. Sea Trout, Croakers, etc. Bairdiella sp. Lowest Chamber: 12 postlarvae, 8 to 10 mm. long, probably belong to this genus. Bairdiella ronclnis (Cuvier & Valenciennes). Lowest Chamber: 8 specimens, ranging in length from 33 to 95 mm. Family Pomacentridae. Pomacentrus fuse us Cuvier & Valenciennes. Lowest Chamber: 1 specimen 75 mm. long. Family Cichlidae. Mojarras de Rio. Cichlasoma maculicauda Regan. Uppermost Chamber: 6 specimens, varying in length from 82 to 210 mm. This Atlantic slope species is not confined to strictly fresh water, as it 1939] Hildebrand: Panama Canal as a Passageway for Fishes 29 frequently was taken in brackish water in Panama during earlier investiga- tions (1911 and 1912). It apparently has crossed the divide through Culebra Cut, as specimens were secured in Pedro Miguel and Miraflores Locks as well as in Miraflores Lake. Family Gobiidae. Gobies; Guavinas. Gobiomorus dormitor Lacepede. Uppermost Chamber: 7 specimens, 35 to 75 mm. long. Lowest Chamber: 1 specimen 73 mm. long. This fish was not taken in strictly salt water in Panama during the recent investigation nor the earlier ones (1911 and 1912). It most commonly was found in quiet brackish water, though occasionally far upstream. It, also, was secured in Gatun Lake in 1935. Eleotris pisonis (Gmelin). Lowest Chamber: 7 specimens, 15 to 75 mm. long, preserved. Others were seen. This species ranges from brackish to fresh water. It apparently has crossed the divide to the Pacific side locks where it seems to have hybridized with its closely related congener, picta. (See Hildebrand, Field Mus. Nat. Hist. Pub., XXII, 1928, pp. 344-347). Eleotris isthmensis Meek & Hildebrand. Uppermost Chamber: 12 specimens, 20 to 60 mm. long, preserved. This species was common in the bottom holes after the flight had been drained, but it was difficult to catch. This fish ranges from salt to fresh water, but it appears to be most numerous in brackish water. No evidence of crossing over was found, though nothing would appear to prevent it. Leptophilypmis tluviatilis Meek & Hildebrand. Uppermost Chamber: 28 specimens, 15 to 45 mm. long, preserved. It is very common, like the preceding species, in the bottom “manholes” where many adhering to the walls could be seen. Middle Chamber: 2 specimens, 49 and 52 mm. long. Lowest Chamber: 1 specimen 35 mm. long. Three specimens of this species were taken in Gatun Lake at Barro Colorado Island. It apparently ranges from brackish to fresh water. This species seems to have crossed the divide to the Pedro Miguel and Miraflores Locks. Lophogobius cyprinaides (Pallas). Lowest Chamber: 25 specimens, 30 to 81 mm. long, preserved; many present. This goby previously was known from Panama from only one specimen, taken at Porto Bello. Gobionellus boleosoma (Jordan & Gilbert). Uppermost Chamber: 1 specimen 17 mm. long. Lowest Chamber : 2 specimens, 27 and 30 mm. long. Bathygobius separator (Cuvier & Valenciennes). Middle Chamber: 1 specimen 38 mm. long. Lowest Chamber: 7 specimens, 25 to 59 mm. long, preserved; many present. 30 Zoologica: New York Zoological Society [XXIV :3 West Side: 1 specimen 58 mm. long. This species apparently does not enter fresh water. Garmaswia hildebrandi Ginsburg. Uppermost Chamber: 26 specimens, 13 to 37 mm. long, type material. This goby was taken also in the Pedro Miguel Locks. Family Batrachoididae. 'Toadfishes. Amphichthys crypfocenfrus (Cuvier & Valenciennes). Lowest Chamber: 2 specimens, 72 and 75 mm. long. Family Blenniidae. Blennies. Hypleurochilus sp. Lowest Chamber: 1 specimen 46 mm. long. This probably is a new species. Blennius sp. Lowest Chamber: 4 specimens, 25 to 37 mm. long. This may be a new species. Family Antennariidae. Frogfishes. Anfennarius scaber (Cuvier). Lowest Chamber: 1 specimen 45 mm. long. PEDRO MIGUEL LOCKS. All specimens noted in the following list were taken in the east side of the locks, no collection having been made in the west side. The Pedro Miguel Locks are situated in fresh water, as stated elsewhere, and as the fresh water species of the Canal Zone in large part seem to shun the locks, comparatively few species, or even individuals, were present when the east side of these locks was dewatered February 20, 1937. It will be noticed that most of the species recorded are more or less regular inhabitants of brackish water. Several marine forms are included in the list, which in this instance have invaded fresh water. Family Elopidae. Tarpon atlanticus (Cuvier & Valenciennes). Tarpon. This species was not present in the east side of the locks, but Mr. Myrick, the superintendent of the locks, stated that several were stranded in the west side when dewatered in January, 1937. Several tarpons were pres- ent in Miraflores Locks. It was seen also in Miraflores Lake. Family Engraulidae. Anchovies. Anchovia lucid® (Jordan & Gilbert). Six specimens, 93 to 106 mm. long. This species was found in Miraflores Locks also. Anchovia spinifer (Cuvier & Valenciennes). Many specimens of this anchovy were taken, ranging upward to 160 1939] Hildebrand: Panama Canal as a Passageway for Fishes 31 mm. in length. It was numerous also in Miraflores Locks. No evidence was secured, indicating that this anchovy, which inhabits both coasts of Panama, invades the fresh water between Gatun and Pedro Miguel Locks. Family Characinidae. Astyanax fasciatus (Cuvier). “Sardina.” This Pacific slope species was taken in Gatun Lake. Though no speci- men was seen in Pedro Miguel Locks when dewatered, it may have passed through them to reach Gatun Lake. Brycon ch agrensis (Kner). “Sabalo pipon.” A single specimen of this common Atlantic slope fish, 295 mm. long, was found in Pedro Miguel Locks. Many more were taken in Miraflores Lake, where the species apparently is now well established. To reach Miraflores Lake, the fish originally very probably descended from Gatun Lake and Culebra Cut, through the Pedro Miguel Locks. Brycon petrosas Meek & Hildebrand. “Sabalo pipon.” Eight small specimens, 49 to 76 mm. long, were taken in the Rio Cocoli, just above Miraflores Lake. This Atlantic slope fish presumably reached the Rio Cocoli from Miraflores Lake by descending from Gatun Lake and Culebra Cut through Pedro Miguel Locks. Family Atherinidae. Silversides. Menidia IThyrina J chagresi Meek & Hildebrand. Four specimens, 45 to 70 mm. long, of this Atlantic slope fresh water silverside were secured. It was taken also in Miraflores Locks. This small species, which usually does not exceed a length of about 120 mm., has been mistaxen by some Canal Zone residents for young tarpon. Family Mugilidae. Mullets. Chaencimigi! pr&boscideus (Gunther). A single small specimen 37 mm. long. Mug H curema Cuvier & Valenciennes. Three small specimens, 41 42 and 44 mm. long. This species was taken also in Miraflores Locks and Miraflores Lake. It occurs on both coasts of Panama. No evidence indicating that it has invaded the fresh water between Pedro Miguel and Gatun Locks was secured. Family Centropomidae. Robalos; Snook. Centropoemis parallehis Poey. Although this Atlantic coast fish was not taken in Pedro Miguel Locks, it was secured in Miraflores Lake, where it seems to be common. To reach this lake it apparently had to pass through Pedro Miguel Locks. Cenfropomus robalifo Jordan & Gilbert. Four specimens 80 to 125 mm. long. Many more were seen and taken in Miraflores Lake, though none was seen in Miraflores Locks, through which they presumably had to pass to reach Miraflores Lake and Pedro Miguel Locks. 32 Zoologica: New York Zoological Society [XXIV :3 Centropomus aramatus Gill. Three specimens, 220, 223 and 240 mm. long. It was not seen in Mira- flores Locks through which it presumably had to pass to reach the upper locks. Family Lutianidae. Snappers. Lutianus novemfasciatus Gill. One specimen 335 mm. long was preserved. It is quite certain that others were present, but disappeared by the “route of the fish-hungry.” It also occurred in Miraflores Locks. This snapper evidently has great tolerance for fresh water, but -to date it is not known to have advanced into the fresh water of the Canal beyond the Pedro Miguel Locks. Lutianus Colorado Jordan & Gilbert. Three specimens, 330, 430 and 580 mm. in length, were preserved, and others were seen. It was secured also in Miraflores Locks and Miraflores Lake. Family Haemulidae. Grunts. Pomadasys bayanus Jordan & Evermann. Two small specimens, 27 and 50 mm. long. Two considerably larger specimens, 250 and 310 mm. long, were taken in Miraflores Lake. It was not seen in Miraflores Locks. This fish probably is not distinct from P. crocro of the Atlantic, which was secured in Gatun Locks and Gatun Lake. Family Sciaenidae. Croakers, Sea Trout, etc. Cynoscion albus (Gunther). “Yellow corbina.” Three specimens, 153, 225 and 305 mm. long. This species was taken also in the Miraflores Locks. Family Cichlidae. Mojarras de Rio. Cichlasoma maculicauda Regan. “Chogorro.” Three specimens, 260, 293 and 300 mm. long. This Atlantic slope fish seems well established and common in Miraflores Lake where numerous individuals were taken, ranging upward to 320 mm. in length. It was taken also in Miraflores Locks and Gatun Locks and Gatun Lake. Family Gobiidae. Gobies; Guavinas. Gobiomorus maculatus (Gunther). “Guavina.” This Pacific slope fish was found in Gatun Lake, where 2 specimens, 127 and 210 mm. long, were taken not far below Madden Dam. Though no speci- mens were secured in Pedro Miguel Locks the species presumably reached Gatun Lake by passing through these locks. This fish is common in Mii’a- flores Lake, below Pedro Miguel Locks. Eleotris picta Kner & Steindachner. “Guavina.” Twelve specimens, 35 to 480 mm. long, preserved. This species was numerous in the locks, where specimens up to 495 mm. in length were 33 1939] Hildebrand: Panama Canal as a Passageway for Fishes measured. It is common in Miraflores Lake, and it was found also in Miia- flores Locks. No evidence indicating that this fish has passed through Culebra Cut to Gatun Lake was obtained. However, signs of the hybridization of this species and its closely related Atlantic slope congener, pis onus, was found. This apparent crossbreeding is discussed in another paper ( Field Mus. Nat. Hist. Pub., Zool. Ser., XXII, 1928, pp. 344 to 347). Leptophilypnus fluviatilis Meek & Hildebrand. This Atlantic slope species was numerous in the “manholes” in the floor of the locks, where it seemed so much at home that it may be a permanent resident. Here it clings, very goby-like, to the walls of the holes. Thirty- five specimens 20 to 33 mm. long were preserved. This fish was taken also in Miraflores Locks and Miraflores Lake. G armannia hildebrandi Ginsburg. This species was obtained also in the uppermost chamber of Gatun Locks. Four specimens, 19 to 35 mm. long, were secured in the Pedro Miguel Locks, type material. G armannia homochroma Ginsburg. Ten specimens, 12 to 33 mm. long were preserved, type material. Family Soleidae. Soles. Achirus fluviatilis Meek & Hildebrand. This species was fairly common in the “manholes” in the floor of the locks, where it was rather difficult to catch. Five specimens, 19 to 34 mm. long, were preserved. It was taken also in the Miraflores Locks. MIRAFLORES LOCKS. Collections were made by the writer and assistants in the two chambers of the west side of the Miraflores Locks, when dewatered from March 24 to 29, 1937, and in the east side by Dr. A. 0. Foster of the Gorgas Memorial Laboratory, Panama City, on April 28 and 29, 1937. The specimens secured in the upper and lower chambers of the west side were all kept separate and most of those from the east side similarly were labeled as to the level in which they were taken. When the level is known it is given in the following list. Family Carcharhinidae. Gray Sharks. Carcharhinus sp. Lower Chamber, East Side: 1 specimen, a partial skin, about 950 mm. long; a young male. Dr. Foster reported that 3 other sharks, similar to the one preserved, were present. To date the writer has not succeeded in identifying the specimen. Family Siluridae. Catfish (Marine). Sciadeichthys troschelii (Gill). Upper Chamber, West Side: 3 specimens, 440, 470 and 490 mm. long, were preserved. Many others, supposedly of this species, were seen. Many males of this and other species carried eggs and young in the mouth, which sometimes were dropped when the fish became distressed as 34 Zoologica: New York Zoological Society [XXIV :3 the water receded. In places in both chambers (west side) the floor of the locks was fairly covered with eggs and young catfish with large yolksacs. A male of this' species examined in the laboratory retained a single young, 70 mm. long, with a large yolksac, in his mouth. One might judge that this young could easily have reached a length of 100 mm. (4 inches) on the large amount of yolk remaining. It would seem probable, then, that the young of this species are retained and cared for in the mouth of the male parent until they reach the comparatively large size of around 100 mm. Galeichthys seemanni (Gunther). Lower Chamber, West Side: 1 specimen, 360 mm. long, was preserved. According to my field notes this species was very common. However, as this and the next mentioned closely related species were not distinguished in the field, it seems probable that both were present in some abundance. Galeichthys jordani (Eigenmann & Eigenmann). East Side: 1 specimen, 375 mm. long, was preserved. This species is closely related to the preceding one from which it is distinguished with difficulty. This species and the foregoing ones very prob- ably were present in the east side, though no specimens were preserved. Galeichthys dasycephalus (Gunther). Upper Chamber, West Side: 4 specimens, each about 65 mm. long, re- taining a large yolksac, evidently dropped by the parent, were preserved. Lower Chamber, West Side: 1 adult female 353 mm. long; 1 young 80 mm. long with yolk fully absorbed; 4 young removed from mouth of male parent, respectively 52, 54, 55 and 56 mm. long. These young retained a very large yolksac, measuring about 17 mm. in diameter. According to my field notes this species was very numerous in the locks. It no doubt was present also in the east chambers, though no specimens were preserved. Arius multiradiatus Gunther. Upper Chamber, West Side: 1 specimen 250 mm. long. Lower Chamber, West Side: 1 specimen, a male 287 mm. long, with eggs 6 mm. in diameter in the mouth. Family Muraenesocidae. Eels. Neoconger vermiformis Gilbert. This eel was common in the bottom silt of all four chambers. Thirty-six specimens, ranging in length from 42 to 152 mm., were preserved. Many more could have been taken. Specimens under about 75 mm. in length are glassy; the larger ones are pinkish. This eel previously had not been taken in shallow water at Panama. Hoplunnis sp. Lower Chamber, West Side: 1 specimen 272 mm. long. I have not yet been able to identify this eel, and in fact am not certain that it belongs to the genus Hoplunnis. Upon further study it may prove to be new. Family Myridae. Worm Eels. Myrophis voter Jordan & Gilbert. Lower Chamber, West Side: 1 specimen 68 mm. long. 1939] Hildebrand: Panama Canal as a Passageway for Fishes 35 Family Elopidae. Tarpon atlanticus (Cuvier & Valenciennes). Tarpon; “Sabalo real.” Upper Chamber, West Side: 1 female, with undeveloped gonads, 131.25 cm. (52.5 inches) long, became stranded. There was none present in the lower chamber of the west side. Upper Chamber, East Side: 6 tarpons, ranging in length from 118 cm. (47 inches) to 162.5 cm. (65 inches), present at this level. A female, 141 cm. (561/2 inches) long, contained well developed roe. Lower Chamber, East Side: Four individuals, 2 males and 2 females, were present, ranging from 122.5 cm. (49 inches) to 150 cm. (60 inches) in length. The smallest one, a female, weighed 32 pounds, and the largest one, a male, weighed 57 pounds. No tarpon were present in the east side of the Pedro Miguel Locks when dewatered in 1937, though reported from the west side, as stated in the preceding list. This species was seen in Miraflores Lake and it has been reliably reported from the sea level end of the Canal below Miraflores Locks. The tarpon, an Atlantic species, then, has completed the transit through the Canal. Family Clupeidae. Herrings. Sardinella stolifera (Jordan & Gilbert). Upper Chamber, West Side: This species was very numerous; 15 speci- mens, ranging from 95 to 130 mm. in length, were preserved. Lower Chamber, West Side: This small herring was somewhat less numerous at this level than in the higher one. Eleven specimens, ranging from 51 to 132 mm. in length, were retained for the collection. It was taken also in Miraflores Lake. It is a rather conspicuous fish because of its very bright, broad, silvery, lateral band. Misha fiirthii (Steindachner) . Upper Chamber, West Side: A single specimen 150 mm. long. Odontognafhus sp. Lower Chamber, West Side: 1 specimen 77 mm. long. Lower Chamber, East Side: 3 specimens, 52, 53 and 62 mm. long. Because of the immaturity of the specimens it has not been possible to identify them with any degree of certainty even as to the genus. Family Engraulidae. Anchovies. Anchovia balboae (Jordan & Seale). Lower Chamber, West Side: 4 specimens, each close to 90 mm. in length. This is the species listed as A. brevirostris in our earlier work (1923, p. 198). The name, brevirostris, however, is regarded as preoccupied by a Brazilian species of this genus. Anchovia naso (Gilbert & Pierson). Lower Chamber, West Side : 3 specimens, 60, 62 and 63 mm. long. Lower Chamber, East Side: 1 specimen 52 mm. long. This species was not numerous in the locks. Anchovia parva Meek & Hildebrand. Lower Chamber, West Side: 15 specimens, 43 to 65 mm. long. 36 Zoologica: New York Zoological Society [XXIV :3 Upper Chamber, East Side: 62 specimens, 35 to 62 mm. long. Lower Chamber, East Side: 12 specimens, 33 to 50 mm. long. This small anchovy heretofore was recorded only from the Atlantic coast of Panama and Trinidad. The specimens have been carefully compared with specimens from the Atlantic coast of Panama (the type locality), without detecting any differences. This species was numerous in the Gatun Locks, even in the uppermost chamber. It was not taken, however, in Gatun Lake, the Pedro Miguel Locks, nor in Miraflores Lake. Therefore, evidence indi- cating that it has crossed the Isthmus through the Canal is lacking. Its presence in abundance in the uppermost flight of the Gatun Locks, never- theless, shows that it has great tolerance for fresh water, and the possibility that it has crossed definitely exists. Anchovia ischana (Jordan & Gilbert). This anchovy was numerous in all four chambers of the locks. Many specimens, 47 to 93 mm. in length, were preserved. The great abundance of this species, as well as other anchovies, prob- ably attracts some of the larger fish to the locks. Anchovia lucida (Jordan & Gilbert) . Upper Chamber, West Side: 1 specimen 85 mm. long. Lower Chamber, West Side: 18 specimens, 80 to 97 mm. long. Lower Chamber, East Side: 1 specimen 92 mm. long. This species was found also in the Pedro Miguel Locks, showing that it can endure fresh water. Anchovia curta (Jordan & Gilbert). Although this anchovy was not taken in the locks it seems to belong to this list as 10 specimens, 22 to 54 mm. in length, were taken in Miraflores Lake above the locks. To reach this lake it presumably had to pass through Miraflores Locks. Its tolerance for fresh water suggests that this anchovy could traverse the Isthmus, but to date no evidence indicating that it has done so has been found. Anchovia spinifer (Cuvier & Valenciennes). This species was very numerous in every chamber of the locks. Many specimens, ranging from 33 to 165 mm. in length, were preserved. It was common also in Pedro Miguel Locks, and a single specimen was secured in Gatun Locks, though not at intermediate points. As this anchovy was recorded from both coasts of tropical America before the completion of the Canal, it would be impossible to determine whether crossing over has taken place from the study of specimens. Its tolerance for fresh water apparently would enable it to complete the transit. Its great abundance in the locks was surprising because it was not secured on either coast of Panama during our earlier extensive collecting. Many of the larger specimens were very conspicuous because of their bright orange color, though others were plain silvery. The great abundance of this fish in the locks very probably helps to attract many large predatory species. Anchovia panamensis (Steindachner) . Upper Chamber, West Side: 4 specimens, 58 to 72 mm. long. Lower Chamber, West Side: 1 specimen 88 mm. long. Two specimens, 46 and 55 mm. long, were taken in Miraflores Lake. It was not seen in the Pedro Miguel Locks. 1939] Hildebrand: Panama Canal as a Passageivay for Fishes 37 Ancho via rastralis (Gilbert & Pierson). Upper Chamber, West Side: 5 specimens, 53 to 55 mm. long. Lower Chamber, West Side : Many present ; 40 specimens, 57 to 105 mm. long, were preserved. Lower Chamber, East Side: A single specimen 63 mm. long. Lycengraulis poeyi (Kner & Steindachner) . Upper Chamber, West Side: 3 specimens, 140, 160 and 185 mm. long. Cetengraulis mysticetus (Gunther). Upper Chamber, West Side: 1 specimen 53 mm. long. Lower Chamber, West Side: Numerous; 41 specimens, 53 to 90 mm. long, were preserved. Family Poeciliidae. Top Minnows. Poeeiliopsis elongafus (Gunther). Upper Chamber, West Side: 11 specimens, 19 to 38 mm. long. Lower Chamber, West Side: 54 specimens, 26 to 47 mm. long. Upper Chamber, East Side: 2 specimens, 56 and 60 mm. long. This brackish water minnow was numerous in both chambers of the west side, where many either were stranded or occupied the “manholes” in the floor of the locks. Family Hemirhamphidae. Halfbeaks. Hyporhamphus snyderi Meek & Hildebrand. Lower Chamber, East Side: 2 specimens, 143 and 175 mm. long. Family Atherinidae. Silversides. Menidia I Thyrina I chagresi Meek & Hildebrand. Lower Chamber, West Side: 3 specimens, 25, 27 and 88 mm. long. This Atlantic slope fish, confined to fresh and brackish water, was taken also in the Pedro Miguel Locks. Kirflandia gilberti (Jordan & Bollman). Lower Chamber, East Side: 2 specimens, 38 and 117 mm. long. East side: 3 specimens, 36, 48 and 120 mm. long. Family Mugilidae. Mullets. Mugil curema Cuvier & Valenciennes. Lower Chamber, West Side: 9 juveniles of one school, 46 to 56 mm. long, and 14 of another school, 24 to 33 mm. long, were preserved. Upper Chamber, East Side: 3 specimens, 30, 38 and 45 mm. long. East side: 1 adult 235 mm. long. Juveniles were taken, also, in Miraflores Lake and Pedro Miguel Locks. Family Polynaemidae. Threadfins. Polynemus approximans Lay & Bennett. Lower Chamber, East Side: 1 juvenile 29 mm. long. East Side: 2 adults, 195 and 205 mm. long. 38 Zoologica: Neiv York Zoological Society [XXIV :3 Family Carangidae. Crevalle, Jacks, Pompanos, etc. Caranx hippos (Linnaeus). “Jack.” Upper Chamber, West Side: 1 specimen 135 mm. long. Lower Chamber, West Side: 8 specimens, 105 to 125 mm. long. East Side: 1 large individual 610 mm. long. This species as here understood occurs on both coasts of Panama. Though some writers have attempted to separate the specimens from the opposite coasts, giving the name C. caninus to the Pacific coast one, I am not convinced that the supposed differences are well founded. The scarcity of this species in the Miraflores Locks contrasts sharply with its abundance in the Gatun Locks. It is not known to invade strictly fresh water. Caranx marginatus (Gill). Upper Chamber, West Side: A single specimen 140 mm. long. Oligoplites mundus Jordan & Starks. Leather jack. East Side: 1 specimen 365 mm. long. Oligoplites saurus (Bloch & Schneider). Leather jack. Lower Chamber, West Side: 3 specimens, 29, 38 and 62 mm. long. East Side: 2 specimens, 190 and 205 mm. long. This species, which is common to both coasts of Panama, was not seen in the Gatun Locks, and is not known to enter fresh water. Trachinotus kennedyi Steindachner. Pompano. East Side: 1 large individual 645 mm. long. Chloroscombrus orqueta Jordan & Gilbert. Lower Chamber, West Side: 1 specimen, a juvenile 54 mm. long. East Side: 8 specimens; one juvenile 40 mm. long, and 7 adults 185 to 205 mm. long. Selene brevoortii (Gill). Moonfish. Lower Chamber, East Side: 1 juvenile 70 mm. long, and an adult 265 mm. long. Vomer declivifrons Meek & Hildebrand. Moonfish; horsefish. Lowest Chamber, West Side: 1 specimen 335 mm. long. East Side: 1 individual 250 mm. long. Family Apogonidae. Cardinal-fishes. Apogon dovii Gunther. Lower Chamber, West Side: 1 specimen 100 mm. long. Family Centropomidae. Robalos; Snook. Centropomus peetinatus Poey. This species, which is known from both coasts of tropical America, was not seen in Miraflores Locks. However, 2 specimens, 95 and 330 mm. long, were taken in Miraflores Lake. The fish presumably had to pass through the Miraflores Locks to reach this lake. It was not taken in the Gatun Locks. 1939] Hildebrand: Panama Canal as a Passageway for Fishes 39 Cenfropomus unionensis Bocourt. Upper Chamber, West Side: 3 specimens, 110, 160 and 200 mm. long. Lower Chamber, West Side: 2 specimens, 115 and 170 mm. long. Cenfropomus nigrescens Gunther. Lower Chamber, West Side: 1 individual about 450 mm. long, quite certainly of this species, was laid aside, but disappeared by the “route of the fish-hungry.” In addition 2 small specimens, 55 and 95 mm. long, were taken in Mira- flores Lake, presumably having passed through Miraflores Locks to reach that lake. Family Serranidae. Sea Basses, Groupers, Jewfish, etc. Epinephelus labriformis (Jenyns). Grouper. Upper Chamber, West Side: 5 specimens, 38 to 180 mm. long. Promicrops itaiara (Lichtenstein). Jewfish. Upper Chamber, West Side: 1 specimen about a foot long was seen, but disappeared, having been taken by someone who was hungry for fish. Lower Chamber, East Side: 1 individual 975 mm. (39 inches) long, weighing 47 pounds, was present. Only the head and tail were preserved. Rypticus saponaceus bicolor (Valenciennes). Soapfish. Lower Chamber, West Side: 19 specimens, 61 to 155 mm. long, were preserved. Many more were present in the sumps and bottom slush. East Side: 2 specimens, 86 and 121 mm. long. The identification is by Dr. L. P. Schultz of the U. S. National Museum, who regards the Pacific Coast specimens as only subspecifically distinct from the Atlantic ones. The two nominal species listed in our earlier work (1925, pp. 481 and 482), R. xanti Gill and R. nigripinnis Gill, are regarded as synonyms of the above name. Family Lutianidae. Snappers. Lutianus novemfasciatus Gill. Lower Chamber, West Side: 1 specimen, 395 mm. long, was preserved. Only a few others were seen. This species was taken also in the Pedro Miguel Locks. Lutianus Colorado Jordan & Gilbert. Lower Chamber, West Side: 1 specimen, 410 mm. long, was measured and critically examined. Many more were present, mostly quite large. Prob- ably about 50 pounds of this species and L. argentiventris were removed from this flight by Canal employees. East Side: 1 specimen, 150 mm. long, was preserved. According to Dr. Foster this fish was numerous in the east side of the locks. Although this snapper was not taken in the upper chamber (west side) of the locks, specimens were secured in Pedro Miguel Locks and Miraflores Lake, where it appears to be common. Though this snapper tolerates virtu- ally fresh water, it is not known to have advanced beyond the Pedro Miguel Locks. Lutianus argentiventris (Peters). Upper Chamber, West Side: 2 specimens, 65 and 353 mm. long, were preserved. Several others were seen. 40 Zoologica: New York Zoological Society [XXIV :3 Lower Chamber, West Side: 13 specimens, 90 to 230 mm. long, were preserved. Many present. Probably about 50 pounds of this species and of L. Colorado were removed from this flight by Canal employees. East Side: 3 specimens, 90, 160 and 270 mm. long, were preserved. It was common in the east side of the locks, according to Dr. Foster. Family Haemulidae. Grunts. Orthopristis ehalceus (Gunther). East Side: 5 specimens, 170 to 225 mm. long. Pomadasys leueiseus (Gunther). East Side: 2 specimens, 370 and 385 mm. long. This is Br achy deuterus nitidus of Jordan & Evermann {Bull. U. S. Nat. Mus. XLVII, 1898, p. 1326) and others, which merges with P. leueiseus according to our earlier studies (1925, p. 451). Anisotremus pacifici (Gunther). Upper Chamber, West Side: 3 specimens, 200, 280 and 320 mm. long, were preserved. Many more were present. Ten were dissected and were found to be 8 females and 2 males, all nearly ripe. Lower Chamber, West Side: 1 specimen, 315 mm. long, was preserved. This species w’as nearly as numerous in this flight as in the upper one. Six specimens were examined for spawning condition; only 2 had large roe. .Anisofrentus dovii (Gunther). Lower Chamber, West Side: A single small specimen, 55 mm. long, was doubtfully identified as this species. East Side: 1 specimen 200 mm. long. Family Gerridae. Mojarra (Marine). Eucinostomus californiensis (Gill). Upper Chamber, West Side: 2 specimens, 42 and 60 mm. long. Lower Chamber, West Side: 13 specimens, 47 to 82 mm. long. East Side: Two specimens, 62 and 78 mm. long. One specimen, 145 mm. long, was taken in Miraflores Lake. To reach this lake the fish presumably had to pass through Miraflores Locks. Gerres c inereus (Walbaum). Upper Chamber, West Side: 3 specimens, 56, 59 and 68 mm. long. Lower Chamber, West Side: 1 specimen 55 mm. long. One large specimen, 370, mm. long, was taken in Miraflores Lake, pre- sumably having passed through the locks to reach this lake. This species inhabits both coasts of tropical America. Although it evidently has much tolerance for fresh water it does not seem to have ad- vanced in the Canal beyond Miraflores Lake on the Pacific side. It was not seen in Gatun Locks. Diapterus peruvianas (Cuvier & Valenciennes). Lower Chamber, West Side: 15 specimens, 23 to 110 mm. long (mostly juveniles). Lower Chamber, East Side: 6 specimens, 70 to 145 mm. long. Eight specimens of this species, ranging in length from 50 to 285 mm., were taken in Miraflores Lake. It is probable that the fish passed through Miraflores Locks to reach this lake. 1939] Hildebrand: Panama Canal as a Passageway for Fishes 41 Diapterus axillaris (Gunther). Upper Chamber, West Side: A single specimen, 247 mm. long, seems to belong to this species, which previously was not reported from Panama. Family Kyphosidae. Rudder-fishes. Kyphosus elegans (Peters). East Side: 1 specimen 410 mm. long. Family Sciaenidae. Croakers, Sea Trout, Drums, etc. Micropogon altipinnis Gunther. Upper Chamber, West Side: 3 specimens, 72, 103 and 180 mm. long. Lower Chamber, West Side: 13 specimens, 52 to 152 mm. long; and 69 juveniles, 12 to 40 mm. long, which probably also belong to this species. East Side: 2 specimens, 135 and 182 mm. long. The young of this species seem to be common in Miraflores Lake where 18 specimens, 40 to 125 mm. long, were taken. Stellifer oscitans (Jordan & Gilbert). Lower Chamber, West Side: 1 specimen 103 mm. long. Stellifer illecebrosus Gilbert. Lower Chamber, West Side: 4 small specimens, 67 to 80 mm. long; and 3 juveniles, 22, 33 and 35 mm. long, which may belong to this species. Bairdiella ensifera (Jordan & Gilbert). Upper Chamber, West Side: 2 specimens, 78 and 130 mm. long. Lower Chamber, West Side: 3 specimens, 85, 120 and 150 mm. long. Ophioscion typieus Gill. Upper Chamber, West Side: 3 specimens, 175, 180 and 185 mm. long. Lower Chamber, West Side: 1 specimen 168 mm. long. East Side: 2 specimens, 183 and 200 mm. long. Ophioscion scierus (Jordan & Gilbert). Lower Chamber, West Side: 1 small specimen, 89 mm. long, doubtfully identified as this species. Cynoscion albus (Gunther). “Yellow corbina.” Upper Chamber, West Side: 5 specimens, 235 to 265 mm. long. Lower Chamber, West Side: 2 specimens, 240 and 310 mm. long, were preserved. Some fairly large ones, measuring up to 30 inches in length, were present. Although no specimens were preserved, this species was numerous in the east side of the locks, according to Dr. Foster. Small individuals were taken also in Pedro Miguel Locks, showing that this species has much tolerance for fresh water. The “yellow corbina” is a highly prized foodfish in Panama. A length of about 40 inches and a weight of 25 pounds are not unusual. It is sought extensively by sportsmen. Larimus effulgens Gilbert. East Side: 1 specimen 280 mm. long. 42 [XXIV :3 Zoologica: New York Zoological Society Family Ephippidae. Spade-fishes. Chaetodepterus zonatus (Girard). East Side: 3 specimens, 220, 220, and 235 mm. in length. Family Cichlidae. Mojarras de Rio. Cichlasoma maculicauda Regan. Upper Chamber, West Side: 3 specimens, 32, 70 and 85 mm. long. This Atlantic slope fish was taken also in Miraflores Lake and in Pedro Miguel Locks. Family Tetraodontidae. Puffers. Sphoeroides annulatus (Jenyns). Lower Chamber, West Side: 1 specimen 225 mm. long. East Side: 1 specimen 195 mm. long. Sphoeroides fiirthii (Steindachner) . Lower Chamber, West Side: 1 juvenile 28 mm. long. Guentheridia form osa (Gunther). East Side : 1 specimen 225 mm. long. Family Gobiidae. Gobies; Guavinas. Small gobies, principally naked ones, were very common in both cham- bers of the west side in the slush, where an inch or two of water remained on the floor after pumping had ceased. Although the different species were not recognized in the field, their relative abundance is believed to be shown more or less by the number of specimens collected of each species. Dormitator maculatus (Bloch). Lower Chamber, East Side: 15 small specimens, 20 to 85 mm. long, seem to be of this Atlantic slope species. Dormitator latifrons (Richardson). This species was not seen in the locks when they were dewatered, but later, after the locks had been refilled (Sept. 30, 1937), Mr. C. B. Lear took several fine specimens which were in spawning condition. Eieotris picta Kner & Steindachner. Upper Chamber, West Side: 5 specimens, 43 to 80 mm. long. Lower Chamber, East Side : 2 specimens, 20 and 245 mm. long. This fish is common in Miraflores Lake, and it was numerous in Pedro Miguel Locks. Some evidence of cross breeding with its near relative, pisonus, of the Atlantic slope was found (see under this species in the list for Pedro Miguel Locks). Leptophiiypinus Huviatilis Meek & Hildebrand. Upper Chamber, West Side: 16 specimens, 28 to 48 mm. long, were cap- tured. Others were seen in the “manholes” in the floor of this flight. It was not found in the lower chamber. This Atlantic slope fish was taken also in Pedro Miguel Locks and in the Gatun Locks. 1939] Hildebrand: Panama Canal as a Passageway for Fishes 43 Erotelis ciarki (Hildebrand). Lower Chamber, West Side: 5 specimens, 77 to 150 mm. long, including the type of this species. Upper Chamber, East Side: 2 specimens, 70 and 115 mm. long. East Side (level unknown) : 4 specimens, 80 to 150 mm. long. This species recently described by me ( Field Mus. Nat. Hist. Pub., Zool. Ser., XXII, 1938, p. 352) has not been taken elsewhere. Bathygobius separator (Cuvier & Valenciennes). Lower Chamber, West Side: 4 specimens, 69 to 86 mm. long. Others were present, but the species was not numerous. Upper Chamber, West Side: 2 specimens, each 84 mm. long. Enypinas as eras Ginsburg. Lower Chamber, West Side: 8 specimens, 24 to 46 mm. long, type material. Enypinas seminudus (Gunther). Upper Chamber, West Side: 6 specimens, 23 to 45 mm. long. Lower Chamber, West Side: 95 specimens, 16 to 51 mm. long. Upper Chamber, East Side: 12 specimens, 36 to 57 mm. long. Lower Chamber, East Side: 6 specimens, 36 to 53 mm. long. Gobionellus lioiepis (Meek & Hildebrand). Upper Chamber, East Side: 1 specimen 31 mm. long. Gobionellus manglicola (Jordan & Starks). Upper Chamber, West Side: 6 specimens, 27 to 32 mm. long. Lower Chamber, West Side: 28 specimens, 18 to 32 mm. long. Upper Chamber, East Side: 16 specimens, 21 to 40 mm. long. Gobionellus sagittula (Gunther). Lower Chamber, West Side: 1 specimen 38 mm. long. Germania paradoxa (Gunther). Lower Chamber, West Side: 33 specimens, 19 to 35 mm. long. Lower Chamber, East Side: 2 specimens, 25 to 35 mm. long. Gobioides peruanus (Steindachner) . Lower Chamber, West Side: 1 specimen 248 mm. long. This goby was seined on deep mud in the sump next to the “seagates” of the locks. It was thought to be a snake by my helper, who warned me to keep hands off. A new record for Panama. Gobiosoma nudum (Meek & Hildebrand). Lower Chamber, West Side: 1 specimen 27 mm. long. Microgobius tabogensis Meek & Hildebrand. Upper Chamber, West Side: 1 specimen 30 mm. long. Lower Chamber, West Side: 28 specimens, 22 to 50 mm. long. Microgobius emblematicus (Jordan & Gilbert). Lower Chamber, West Side: 1 specimen 44 mm. long. Parrella spilopteryx Ginsburg. Upper Chamber, East Side: 1 specimen, 74 mm. long, the type. 44 Zoologica: New York Zoological Society [XXIV :3 Parrella fusca Ginsburg. Lower Chamber, West Side: 1 specimen, 42 mm. long, the type. Family Batrachoidae. Toadfishes. Batrachoides pacifici (Gunther). Lower Chamber, West Side: 5 specimens, 32 to 290 mm. long. Porichthys greenei Gilbert & Starks. Lower Chamber, West Side: 2 specimens, each 87 mm. long. Porichthys margaritatus (Richardson). East Side: 1 specimen 177 mm. long. This species previously was not taken in shallow water in the vicinity of the Canal Zone. The determination is by Dr. L. P. Schultz. Thalassophryne reticulata Gunther. Lower Chamber, West Side: Saw one badly decomposed specimen, about 150 mm. long, floating in a sump a few days after the main body of water had been removed. The specimen certainly was of this genus, though there is doubt as to whether it was reticulata or dowii. Family Blenniidae. Blennies. Hypsoblennius sp. Lower Chamber, West Side: 17 specimens, 16 to 70 mm. long. The specimens are near H. lignus, but differ in minor structures, re- quiring further study. Family Soleidae. Soles. Small soles were fairly common in the bottom “manholes,” but were rather difficult to catch. Achirus Bmbriatus (Gunther). Upper Chamber, West Side: 17 specimens, 15 to 40 mm. long. Lower Chamber, West Side: 5 specimens, 24 to 50 mm. long. Lower Chamber, East Side: 2 specimens, 35 to 40 mm. long. Achirus Huviatilis Meek & Hildebrand. Lower Chamber, West Side: 1 specimen 50 mm. long. This sole was taken also in the Pedro Miguel Locks. Achirus klunzingeri (Steindachner) . Upper Chamber, West Side: 1 specimen 71 mm. long. Lower Chamber, West Side: 3 specimens, 58, 62 and 63 mm. long. Upper Chamber, East Side: 2 specimens, 70 and 95 mm. long. Symphurus elongafus (Gunther). Tongue fish. Upper Chamber, West Side: 14 specimens, 20 to 64 mm. long. Lower Chamber, West Side: 28 specimens, 17 to 80 mm. long. Upper Chamber, East Side: 2 specimens, 43 and 65 mm. long. Lower Chamber, East Side: 3 specimens, 51, 58 and 75 mm. long. 1939] Hildebrand: Panama Canal as a Passageway for Fishes 45 EXPLANATION OF THE PLATES Plate I. Fig. 1. Gatun Locks, Panama Canal, looking toward Gatun Lake, showing vessels in transit. Courtesy Panama Canal. Fig. 2. Scene in base of dewatered Gatun Locks, showing stranded fish, consisting mostly of “jacks,” Caranx hippos. Plate II. Fig. 3. Wall of upper chamber of dewatered lock, showing growth of bivalve mollusc, Congria ( Mytilopsis ) sallei; also the “skiff” operated by a crane, lowering collectors into the lock. Fig. 4. Lifting collecting seine after a haul in the sump of Pedro Miguel Locks. HILDEBRAND PLATE I FIG. 1. FIG. 2. THE PANAMA CANAL AS A PASSAGEWAY FOR FISHES. WITH LISTS AND REMARKS ON THE FISHES AND INVERTEBRATES OBSERVED. HILDEBRAND. PLATE II. FIG. 3. FIG. 4. THE PANAMA CANAL AS A PASSAGEWAY FOR FISHES, WITH LISTS AND REMARKS ON THE FISHES AND INVERTEBRATES OBSERVED. Levenstein: Cytology of Goldfish Pituitary 47 4. The Cytology of the Pituitary Gland of Two Varieties of Gold- fish ( Carassius auratus L.), with Some Reference to Variable Factors in the Gland Which May Possibly Be Related to the Different Morphological Types. Irving Levenstein1 Department of Biology, Washington Square College, New York University (Plate I). In recent years the pituitary gland, as compared to other glands of internal secretion, has received a greater amount of concentrated atten- tion and experimental study. However, most of this work has centered around the pituitary of the mammal and only a few investigators have attempted to give any sustained attention to that structure in lower forms. This seems rather unfortunate for the greater part of the con- fusion, regarding the pituitary, concerns itself with its supposed mani- fold hormonal activities and inter-relationships. In the higher forms, for example, such functions as corpus luteum formation, uterine growth, lac- tation and proportional growth in general, have all been shown to be related to the pituitary gland. In lower forms most of these physiological activities are absent. This alone should result in a comparative simplifica- tion of the functional activity of the gland. Therefore, it is reasonable to believe that a study of the pituitary gland in the lower forms may lead more readily to a better understanding of the function of the gland. However, before the function of any gland may be adequately studied, it is always valuable to have as complete a knowledge of its microscopic structure as possible. This, of course, becomes especially true if any attempt is made to correlate its function with its structure. In addition, there is still a great need for a cytological study of the pituitary in these lower forms by the use of modern techniques (Charipper, 1937). The morphology and microscopic anatomy of the pituitary gland of only a few species of Cyclostomes, Pisces and Amphibia have been worked out with varying degrees of completeness. This work has been covered for the most part by Tilney (1911), Stendell (1914) and Bock (1928). More recently, Bell (1938) has reviewed the previous work on the pituitary of the lower vertebrates and has added a morphological description of the pituitary gland of the common goldfish ( Carassius auratus). None of these workers has attempted a complete cytological description of the cells form- ing the gland in the species which they considered. 1 Accepted in partial fulfillment of the requirements for the degree of Doctor of Philosophy, New York University. 48 Zoologica: New York Zoological Society [XXIV :4 Aside from the work on amphibian metamorphosis, there has been very little, if any, investigation of the relationship of the pituitary in lower forms to deviations from the normal body structure known to occur in such animals. Cushing (1932, a, b) and his co-workers have described cellular changes in the anterior lobe occurring in acromegalic individuals and those suffering from “Basophilia.” Keith (1922), Crew (1923) and Stockard (1936) have also related body abnormalities to changes in the endocrine glands. These workers believe that these structural abnormalities are car- ried on from generation to generation. The present work has two objectives: (1) to describe the histology and cytology of the pituitary of two varieties of goldfish, and (2) to investigate whether or not easily apparent differences in body structure occur along with variations in the pituitary. I wish to acknowledge gratefully the interest and helpful suggestions of Dr. Harry A. Charipper under whose guidance this work was carried out. I wish also to thank Dr. C. M. Breder, Jr., of the New York Aquarium for his invaluable assistance and co-operation. Materials and Methods. The pituitary glands of two varieties of goldfish ( Carassius auratus L.) were examined both histologically and cytologically. They were removed from animals, both male and female, purchased at local supply houses during the months from October to March. The varieties of goldfish studied were the “common” and the “telescope moor.” The former has an elongate body, a scaleless head, yellow to gold pigmentation and two sets of paired fins, the ventral and pectoral, as well as three single fins including the tail. The “telescope moor” has a short blunt body which is rounded and egg-shaped and possesses a black pigmentation. All the ventral fins are paired and its tail is divided to its base. In this variety the eyes are set into stalks which project from the sides of the head. To follow cytological procedures, it was found best to remove the pitui- tary gland from the animal before placing it into the fixing solutions. In order to obtain the gland, the animal first was quieted in an ice water bath. The head then was severed from the body, the ventral surface of the brain case exposed and the portion of bone below the gland carefully cut away. The exposed gland was removed easily and placed into the fixing solution. The majority of the pituitary glands were fixed, either by the method of Champy or by Nassonov’s modification of the method of Kolatchev. In addition, a few of the glands were fixed according to the Mann-Kopsch method as described by Gatenby. Pituitary glands were also treated accord- ing to the Champy-Kull method in order to demonstrate mitochondria. After fixation and impregnation, all the glands were dehydrated and cleared in dioxan. Dioxan was advantageous in keeping the brittle osmicated tissue from fragmenting. Sections were cut in hard paraffin (60° C.) at two, three and four micra at low temperatures. Glands were sectioned seri- ally in frontal, sagittal and horizontal planes, and stained by either the Masson method or the Dawson modification of the Heidenhain azan proce- dure. It is important to note that throughout the work the pituitary glands of the two varieties of goldfish being studied were fixed in sets and run through all the procedures side by side thereby assuring the validity of cytological comparison. Further, the glands were removed from animals of approximately the same weight. 1939] Levenstein: Cytology of Goldfish Pituitary 49 Observations and Description. The pituitary gland of the goldfish ( Carassius auratus ) lies below the brain and is connected with it by a short heavy stalk. A tough membrane separates the ventral surface of the brain from the pituitary gland. The infundibular stalk leaves the brain case, along with the optic nerves, through a foramen in the parasphenoid bone. After emerging from the foramen, the stalk continues posteriorly as the pars nervosa and forms a center about which the rest of the gland is arranged. Part of the pituitary gland also extends anteriorly and, as the gland is applied closely to the ventral surface of the brain case, it acts to close the large optic-pituitary foramen. These morphological relationships confirm the findings of Bell (1938). The pituitary glands of two varieties of goldfish, the common and the moor, were examined cytologically. No easily apparent differences were found in cell types, cell structures or Golgi configurations. Therefore, the following description of the gland may be applied to the pituitary of either the common or black moor varieties. The pituitary gland is divided, as Bell (1938) has shown, into four easily distinguishable areas. The infundibular stalk enters the gland and is continuous with the pars nervosa which ramifies throughout all parts of the gland. The blood supply of the gland is intimately connected with these rami- fications (Plate I, Fig. 9). Blood capillaries are only present between the fibers of the nervosa tissue. The cells found among the pars nervosa fibers are few in number. Pars intermedia. The entire posterior end of the goldfish pituitary gland consists of a compact intermediate lobe. The pars nervosa ends here, after having passed through the ubergangsteil or transitional lobe, as a heavy clump of fibers which spread out in all directions. The cells of the pars intermedia are arranged along the invading pars nervosa and are large, measuring from ten to fifteen micra in diameter. The cells composing the entire free surface of the intermediate lobe contain many osmiophilic granules while in the interior of the lobe and along the attached border there is a mixture of granular and non-granular cells (Plate I, Fig. 1). These are oriented along the strands of invading nervous tissue. Only granular cells, however, are found in the spaces between the strands of nervosa tissue. The non-granular cells are vesicular, elongate, and contain large nuclei which are clear after Champy fixation. The granular cells are larger and contain nuclei which, after the same fixation, stain a solid red with acid fuchsin. Even after ordin- ary fixation, the nuclei of the two types of cells can be distinguished by a difference in intensity of staining. They both contain fine strands of nuclear material in a clear nucleoplasm. These nuclei are round or oval after most types of fixation, but take on many bizarre shapes and stain deeply red after exposure to an osmo-sublimate solution. The cytoplasm of the intermedia cells does not stain in a similar manner following the use of different fixing solutions. After Bouin fixation, a blue granulation is present, while following osmic acid fixing solutions the cytoplasm stains light blue or not at all. After exposing the intermediate lobe cells to osmic acid fumes, a Golgi apparatus appears in these cells (Plate I, Fig. 2). It is of the same type and occupies the same position in both the granular and non-granular cells. The osmiophilic substance forms a circular cap which is applied closely to the nucleus and. never extends over more than one-half of its surface. On cross section it is usually granular although in some cells several strandlike 50 Zoologica: New York Zoological Society [XXIV :4 layers may be found. Sometimes, tiny vesicles are seen between the strands. After the same length of exposure to osmic acid the granular cells have a much blacker and more extensive Golgi apparatus than the non-granular cells. There is no apparent orientation of the Golgi material in relationship to the position of the cell. In one cell the network may be found applied to the nucleus on the side nearest the invading nervosa tissue while the next cell may have its Golgi network on the opposite side of the nucleus. After the use of the Champy-Kull technique and staining with acid fuchsin, the mitochondria, which are in the cytoplasm on one side of the nucleus, appear as fine granules. In some cells the granules spread out to fill up more than one-half of the cell cytoplasm. Pars anterior. Each of the cells along the outer surface of the anterior lobe contains a small nucleus, placed at one end of a clear cytoplasm. These cells are from six to ten micra in diameter and are grouped closely together (Plate I, Fig. 8). Their nuclei, after Champy fixation, are clear, while after non-osmic acid fixing solutions they contain a great many chromatin strands and granules. Nucleoli which stain heavily are present at one edge or in the center of the nuclei. The cytoplasm surrounding the nuclei is filled with a flocculated non-stainable material. A few osmiophilic granules are found near the nuclei and a Golgi apparatus is present. This Golgi figure is small, granular, and applied closely to the surface of the nucleus (Plate I, Fig. 3). It covers an area about one-fourth that of the nucleus, and is situated on the side which contains the most cytoplasm. No radiating strands or gran- ules leave the loosely collected central portion of the apparatus to pass into the surrounding cytoplasm. The osmiophilic granules, even after prolonged exposure to osmic acid fumes, do not become greatly blackened. There is no orientation of the Golgi material in respect to the pars nervosa which spreads throughout the lobe. In addition to the cells which have been described above, there are others which are smaller in size and fewer in number. These cells contain small nuclei which stain similarly to those found in the non-granular cells. The cytoplasm of these cells stains a reddish brown. It contains two types of granules, one of which stains with acid fuchsin and the other which becomes blackened after exposure to osmic acid. These granules fill the entire cell and completely surround the nucleus. A Golgi network is next to the nucleus and is of the same size and extent as that found in the non-granular cell. However, the apparatus is strand-like and contains an increased amount of osmiophilic material which results in more intense blackening. Along the anterior surface of the lobe these cells are few in number and lie be- tween the non-granular cells while in the deeper parts of the lobe large groups are found. They are most numerous at the border between the anterior and transitional lobe. A third variety of cell is pi’esent in the anterior lobe. It is as large as the non-granular cell and has a cytoplasm which contains a few granules. The nucleus is similar to that present in the other types of cells found in this lobe, and is at the edge of the cell. In some of these cells the cytoplasm, immediately surrounding the nucleus, is clear while that at the outer rim is highly granular. In others, the clear area of cytoplasm has increased and only a few granules remain. Cells are present which show different degrees of granulation, ranging from a completely granulated cytoplasm to one con- taining no granules at all. Thus, these cells may be intermediate cell stages between the clear cells found along the periphery of the lobe and the highly granular cells found in the interior of the lobe. These cells are found mainly 1939] Levenstein: Cytology of Goldfish Pituitary 51 at a point midway between the inner and outer edges of the anterior lobe, where the lobe extends deeply into the transitional lobe. The Golgi apparatus in this third variety of cell is similar to that found in both the granular and agranular cells. It is not as strand-like and heavily staining as the one in the granular cells but is less granular and stains more deeply than does the apparatus in the non-granular cells. The appearance of Golgi apparatus and the changes in granulation from a granular to a non- granular cell seem to indicate a morphological relationship between all the cells described. Occasionally a basophile is found among the acidophiles of the anterior lobe. This cell resembles in form and structure the basophilic cells which are present in the transitional lobe. It is larger than the acidophiles, has a smaller nucleus and a cytoplasm which may be highly granular or have a homogeneous structure. The blue-staining granules, when present, are very large and few in number, and are found in a background of more finely granulated material which take a light blue stain. The basophiles generally appear singly, although a group of four or five is seen occasionally. In each cell a heavy strand Golgi network radiates from the nucleus. It is not as small or compact as that present in the acidophiles. Transitional lobe. A general survey of the transitional lobe of the goldfish pituitary gland is somewhat confusing in that there does not appear to be any definite architectural pattern, such as a nesting or cordlike arrangement, present anywhere in this region of the gland (Plate I, Fig. 4). The fibers of the nervosa ramify, apparently at random, passing as readily between groups of basophilic staining cells as between acidophilic staining cells. Occasion- ally, bands of nervosa fibers completely surround a group of cells, tending to separate them from their neighbors. The cells in such groups do not appear oriented in any definite way with respect to the investing fibers. It seems that either chromophobes, acidophiles, or basophiles may be opposed to the nervosa tissue. Histological examination of carefully prepared sections of the gland, in which the transitional area has been treated according to the method of Severinghaus, discloses many different cellular entities. These different cell types can be distinguished, one from the other, by their variation in size, granular or agranular cytoplasmic nature, and chromophilic reactions. The following three major types of cells are found in this portion of the pituitary : (A) The granular basophilic cells which are characterized by the presence of many large, spheroidal deep blue staining cytoplasmic bodies set in a gray matrix, and crowding on to a large vesicular nucleus. (B) The granular acidophilic cells which are somewhat smaller than the preceding type and diagnostically contain a fine eosinophilic granulation set in a light clear cytoplasm and surrounding a large vesicular nucleus. (C) The chromophobic cells which are smaller than any of the other cell types and contain large clear vesicular nuclei set in a scant, non-staining, agranular cytoplasm. It is interesting to note that both the granular basophilic cell type and the granular acidophilic cell type may be found in an agranular form. Ac- companying the degranulation of these types, the nucleus undergoes a change from the large clear vesicular karyosome to a compact almost pycnotic one. A careful study of the transitional region, in frontal and sagittal serial sections, discloses a rather characteristic distribution of the various cell 52 Zoologica: New York Zoological Society [XXIV :4 types in relation to one another. The acidophiles and basophiles are present throughout the lobe as aggregates, among which may be found a few chro- mophobes (Plate I, Fig. 4). The acidophiles vary in number in different parts of the lobe. They are relatively more numerous in a mid-sagittal sec- tion than in a section near the periphery. The basophiles are found in greater numbers along the dorsal border of the lobe, posterior to the over- lying pars anterior, and also appear to be more compactly grouped at the antero-ventral border, posterior to the stalk as it leaves the body of the gland. In addition, a large group of basophiles border the transitional lobe in the area of its contact with the pars intermedia. The chromophobes are the least numerous of the cell types found in the transitional lobe. These cells occur scattered in groups of two and three, throughout the lobe, among both the basophiles and acidophiles. Small groups of chromophobes are present along the nervosa tissue which extends between the transitional and intermediate lobes. Much larger groups are to be found associated with the nervosa tissue passing through the substance of the transitional lobe. In fact, in frontal section, the chromophobes are seen to surround completely the larger branches of the nervosa tissue. A close study of the individuality of these various cell types under conditions of special techniques presents some interesting configurations of the Golgi material and mitochondrial substance. Both Golgi nets and granu- lar mitochondria are demonstrable in all the cell types to be found in the transitional lobe. The Golgi material, however, presents various differences in configuration which are characteristic of the different cell types. In the acidophiles this osmiophilic substance is compact and closely applied, in the form of a cap, to the side of the nucleus (Plate I, Fig. 5). It consists of several heavy strands which form short loops in the cytoplasm. These loops do not extend very far from the nucleus and free ends from the network are rarely seen. In some instances, these cells contain a larger Golgi net which extends a considerable way out into the cytoplasm. The mitochondria in these cells are present throughout the cytoplasm, in the form of a fine granulation. Areas of condensation may be present on one side of the nucleus in the region of the Golgi net. The mitochondrial granules can easily be differentiated from the secretory granules, after proper destag- ing, by their lighter pink coloration. The Golgi configuration present in the basophile is more strand-like and more extensive than that found in the acidophilic cells (Plate I, Fig. 7). Here again the Golgi material is limited to one side of the nucleus, but does not form a cap. Instead it appears to be free in the cytoplasm only occasion- ally touching the nucleus and then only at a point or two. The loops are large and made up of fine strands of osmiophilic material. The cytoplasm between the strands is finely granular and the large basophilic granules or spheroids are found completely outside the region of the net and along the borders of the cell. The mitochondria here are also granular and are dis- persed throughout the cytoplasm. The finer granular chondriosomes are found nearest the nucleus while the coarser ones are toward the periphery and in between the basophilic secretory granulations. There does not appear to be any points of increased condensations. The chromophobes present a series of extremely interesting Golgi con- figurations (Plate I, Fig. 6). All of these cells have a characteristic com- pact cap of osmiophilic substance which is limited to one side of the nucleus. This Golgi material forms either a high or low triangle. The high triangle Golgi network consists of fine strands of osmiophilic substance and appears to tend more toward fine loop formation while the low triangular form is made up of coarser strands, compactly arranged with only an occasional visible opening. Finely granular mitochondria are present in both these varieties of cells and are found to be unevenly dispersed in the narrow rim of cytoplasm surrounding the nucleus. 1939] Levenstein: Cytology of Goldfish Pituitary 53 Cell Counts. In the absence of any obvious difference between the two varieties of animals studied, it was deemed advisable to resort to the method of differen- tial cell counts. This method should bring to light such differences in the numerical distribution of cell types which may be present and yet not be apparent under routine microscopical examination. Initial samplings of cell counts of the transitional lobes of many glands, chosen at random, were made first. Such studies yielded results which warranted a further and more careful examination of the percentages of chromophilic cells in the glands of both varieties. The transitional lobe was chosen for examination because it contained the different varieties of chromophiles in greater number. Differential cell counts were made on the chromophilic cells of the transitional lobe of the pituitary glands of both varieties of goldfish. The animals compared were similar in weight and since they were purchased from local supply houses their ages could only be approximated. Pituitary glands removed from both sexes of each variety were studied. Care was taken to compare the glands which had been fixed and stained simultaneously, thereby eliminating any possible errors due to differences in techniques. The results of these cell counts showed that the transitional lobe of the pituitary gland in both male and female animals of the same variety, contained similar percentages of basophilic and acidophilic cells, thus indicating that the sex of the animal could be only, at best, a slightly modifying factor. The pitui- tary glands were removed from all the animals during a normally occurring sexually inactive period. With the aid of an ocular counting chamber comparable areas were selected and the chromophilic cells in one hundred squares of the chamber counted. The following precautions were taken to insure a normal sampling and to minimize the possibility of selecting areas subjectively: Firstly, using low power magnification, the anterior end of the gland was always placed against the highest horizontal line of the counting chamber and the ventral surface of the gland oriented against the most lateral vertical line. Secondly, after the section to be studied had been oriented under low power, the oil immersion objective was placed in position and the cells falling within the one hundred squares were counted. Depending upon the thick- ness of the serial section, every third or fourth section was examined. The cell counts throughout the lobe were made at intervals of twelve micra. An average of 1,000 cells were counted from each gland. Table I gives the results of such counts. A statistical method was used to evaluate the results obtained. Al- though many pituitaries were studied, only those in which every section of the gland was properly cut, mounted, and stained were used for calculation and analysis. The validity of the statistical results rests, in part, upon the perfectness and completeness of the sets of serial sections. In view of the extremely small coefficient of variation (see Table I) obtained in this work and the fact that the statistical method employed in this analysis permits the use of few determinations, seven different animals of each variety were employed. In addition, the rigid criteria of technique applied in the choice of the series used, made it expedient, without detracting from the validity of the analysis, not to increase the number of animals. Discussion. The pituitary glands of the two varieties of goldfish studied in the present investigation are similar in their anatomical relationships and gross morphological subdivisions. These findings confirm the description which Bell (1938) gives for the common variety. The four different lobes described 54 Zoologica: New York Zoological Society [XXIV :4 for the goldfish are easily distinguished one from the other by their ana- tomical position and by their reactions to polychromatic stains. The presence of a pars nervosa, pars intermedia, pars anterior and a special portion re- ferred to as the iibergangsteil or transitional region is typical of the pi- tuitary in all forms of fish thus far studied (Stendell, 1914; Bock, 1928; Matthews, 1936; and Bell, 1938). The blood supply of the pituitary gland is intimately connected with the pars nervosa and no sizable vessels are found entering the gland from its surface. A similar condition has been reported by Stendell (1914) and Bock (1928) for the many form which they studied. Blood capillaries are present in the goldfish only where the nervosa fibers penetrate the substance of the gland. TABLE I. Differential cell counts on the transitional lobe of the pituitary gland of the goldfish. Mean S.E. Diff. of Means S.E. of Diff. Coeff. of variation Number of pituitaries Acidophiles in Common 45 .54 3.1 7 Acidophiles in Moor 56 1.1 9.2 5.0 7 Basophiles in Common 55 .54 2.6 7 Basophiles in Moor 45 1.2 8.1 6.7 7 In relation to the presence of the blood vessels in this part of the gland in the goldfish, it is interesting to note the occurrence of masses of colloid material. Herring (1908) was the first to describe colloid in the gland and he believed it to be a secretory product from one of the lobes. More recently, Matthews (1936) reported large masses of this material among the fibers of the stalk in Fundulus. The presence of large blood vessels in the pars nervosa of the goldfish together with Florentin’s (1934) concept of a hypophyseal portal system which drains the epithelial portions of the gland, through these vessels in the pars nervosa, indicates that if secretions are poured directly into the blood stream they probably find their way into the pars nervosa. Thus, the presence of colloid between the fibers of the pars nervosa may be the result of a concentration of some secretory product and appears to support some of the earlier concepts of pituitary secretion. Another point worthy of note in regard to the pars nervosa is its rela- tion to the pars intermedia. In most forms, thus far reported, the nervosa is connected directly to the intermedia without coming into contact with any other portion of the pituitary gland. In the two forms of the family Cypri- nidae (carp and goldfish) which have been studied, the pars nervosa must first pass through the transitional lobe before it reaches the pars intermedia. In the cod an interesting condition exists (Herring, 1908) which may be interpreted as being intermediate between the more common arrange- ment and the apparently special situation occurring in the carp and its related form, the goldfish. In the cod the intermediate lobe is divided into two parts by a chromophilic band of cells and one part is invaded directly by strands of the pars nervosa while, to reach the other portion of the inter- 1939] Levenstein: Cytology of Goldfish Pituitary 55 media, the strands of the nervosa must first go through the chromophilic band of tissue. It may very well be that this chromophilic band of tissue rep- resents the transitional or iibergangsteil portion of this gland. It should be noted in passing that recognition of that part of the pituitary did not occur until sometime after Herring’s investigation (Stendell, 1913). The cells of the intermediate lobe are described as small and taking a basic or acidic sti'ain depending upon previous fixation. It was seen that if an osmic acid fixing solution is used the cells do not stain with acid fuchsin. Occasionally they retain some aniline blue which gives them a basophilic appearance. After the use of a non-osmic acid fixing solution these same cells stain red with acid fuchsin or eosin treatment. Gentes (1907), Herring (1908) and Matthews (1936) found that the cells in the intermediate lobe did not take any stain after the methods which they employed. Other workers, however, have reported acidophiles or basophiles and sometimes both to be present at one time. In the eel, Tilney (1911) and Stendell (1914) described small basophilic cells with clear cytoplasm. Stendell also found a few weakly staining eosinophiles present while Bock (1928) mentioned that the basophilic pars intermedia cells contained acidophilic granules in their cytoplasm. Matthews (1936) found not only acidophiles and basophiles in the intermediate lobe, but cells which took no stain and others which stained a pale lavender after azur carmine. As a result of these investigations, it can be seen that all types of cells have been reported in the intermediate lobe of the fish pituitary gland. The cells present may be of only one type or a mixture of several types. The fact that the affinity of the same cells for different stains can be modified by the type of fixing solution used has been overlooked by most workers. Uniformity in the types of fixing solu- tions employed is a necessary precaution before the results of different in- vestigations can be compared. In general, the Golgi network present in the cells of the pituitary gland of the goldfish appears to be characteristic for each epithelial lobe. In the intermediate lobe the Golgi network is present as a fine, heavily granulated structure capping the nucleus. Although the cells of the intermediate lobe are divided into granular and agranular types, the Golgi structure is the same in both forms. This may indicate that the cells are of the same kind and that the difference in granulation is due merely to phases of physiologi- cal activity. The same picture is seen in the acidophiles of the anterior lobe. The Golgi apparatus of the granular and non-granular cells, as well as that of the intermediate cell types, is exactly alike. There is a difference in the extent of blackening between the Golgi apparatus of the granular and agranular cells. On the basis of our present concept of the activity of the Golgi apparatus as related to cellular secretion, this change in stainability may be indicative of cellular activity. The cells of the transitional lobe differ markedly among themselves in staining reaction, granulation, and in the form and position of the Golgi apparatus. The basophiles contain a Golgi network which is large, loosely arranged, free in the cytoplasm, and usually not in contact with the nucleus. The Golgi network of the acidophiles is compact, smaller, and caps the nu- cleus. The presence of two types of Golgi configurations in the chromophilic cells of the transitional lobe of the goldfish pituitary is a condition similar to that reported by Addison (1916), Atwell (1929) and Severinghaus (1933) for the anterior lobe of the mammal. Severinghaus (1933) has pointed out that the different Golgi structures present in the chromophilic cells can also be found in the chromophobic cells. This led him to state that the chromo- phobes could be divided into acidophilic chromophobes and basophilic chro- mophobes, suggesting that the acidophilic cells arise from the acidophilic and basophilic cells from the basophilic chromophobes. Kirkman (1937), in the guinea pig, also described two types of Golgi configurations in the chro- mophobes. Similarly in the present work, two types of Golgi configurations can be distinguished in the chromophobic cells of the transitional lobe. The 56 Zoologica: Neiv York Zoological Society [XXIV :4 presence of two types of Golgi networks in the chromophobes suggests (on the basis of the work of Severinghaus) , that a relationship exists between these cells and the chromophilic cells of the transitional lobe. The compact type of Golgi network in the chromophobes resembles the same structure in the acidophiles while the loose, more strandlike type can be associated with the loosely arranged, ramifying Golgi network of the basophilic cells. From the above discussion it can be seen that the Golgi apparatus of the anterior lobe cells in the goldfish pituitary is different from that found in the cells of the transitional lobe. This may be used as added evidence in support of the contention that the cells of the transitional lobe are not re- lated to the cells of the anterior lobe in the goldfish. The great similarity between the Golgi configurations present in the transitional lobe cells of the goldfish and the anterior lobe cells of higher forms gives a firmer basis for postulating a homology between the two lobes. Another item of interest in the present investigation is the comparison of the pituitary glands from the two different varieties of goldfish studied on the basis of the differential counts of the chromophilic cells in the transi- tional lobe. A statistical analysis of these cell counts shows that a significant difference does exist in the percentages of chromophilic cells present in the transitional lobes of these two varieties. Table I shows that in the forms investigated, there are more basophilic cells present in this lobe of the com- mon variety than are present in the same lobe of the black moor variety. This difference in cell counts may possibly be associated with the morpho- logical differences existing in the two forms of goldfish under observation. The black moor has been shown to contain a greater number of acidophiles. These cells in the anterior lobe of the mammal are believed to be associated with growth, and changes in their number from the normal are related to abnormal changes in their morphology (Crew, 1923; Smith & MacDowell, 1930; Stockard, 1931; Cushing, 1932, a,b; and Vicari, 1937). All these workers found, after investigating a form which showed marked differences, anatomically, from others of the same species, that these differences were associated with a change in the pituitary gland. They concluded that dif- ferences in body structure may be reflections of cellular differences in the pituitary. The abnormal growth of the black moor as well as the presence of telescopic eyes and black pigmentation may also, in some way, reflect the differences in cell counts occurring in their pituitary glands. In the goldfish, as well as in other forms examined, in which morpho- logical differences occur, these differences in structure are not necessarily restricted to the individual animal but may be transmitted to the offspring. That is, of course, if the form being studied can survive to sexual maturity. Stockard’s investigations have indicated that the inheritance of these ab- normal forms, which may owe their abnormality to some changed activity of a gland in the organism, follows a definite genetic ratio. Goldfish breeders have found that when two of the so-called fancy types of goldfish are mated, the offspring may range in form from that of the common variety to forms even more divergent, morphologically, than the parents. Many of the young are so abnormal that they die during the first few days of life. Others appear to be the common variety for several weeks and then begin to show atypical form. If this phenomenon is depen- dent on the glands of internal secretion, it is at this time that the glands causing the change in body types would be most apt to show the greatest variation from that found in the typical variety which undergoes no radical morphological change. If the pituitary is really the gland which is modifying body type, then an examination of that gland in the two varieties at that possible “critical” period should yield greater divergence of cell type, dis- tribution, or number, than reported as the result of the present investiga- tion. It is planned to obtain the necessary materials to carry out an examina- 57 1939] Levenstein: Cytology of Goldfish Pituitary tion of pituitaries in animals sacrificed at the “critical” periods of develop- ment. Summary and Conclusions. 1. Confirmation is offered for the anatomical position and relations of the goldfish pituitary gland and its morphological division into four lobes, the pars anterior, pars nervosa, pars intermedia and the transitional lobe. 2. The pars anterior contains acidophilic cells with a granular Golgi net- work applied to the nucleus. 3. The pars intermedia consists of both granular and agranular chromo- phobic cells each containing a similar granular Golgi apparatus which caps the nucleus. 4. The transitional lobe contains the following three types of cells: (a) A basophilic cell which contains large granules and an extensive loosely arranged Golgi network made up of fine strands and lying free in the cytoplasm. (b) An acidophilic cell which has a finely granular cytoplasm and contains a heavy strand Golgi network capping the nucleus. (c) Two varieties of chromophobes are present. Both have a scanty cytoplasm but can be differentiated by their Golgi configuration. One variety has a compact heavy Golgi apparatus while the Golgi apparatus of the other is looser and more strand-like. These cells are described as being related to the different chromophilic cells found in the transitional lobe of the gland in a way similar to that described by Severinghaus for the pars anterior of the mammal. 5. The pituitary glands of the two varieties of goldfish studied (the com- mon and the black moor) are anatomically and histologically similar. 6. Statistical evaluation of cell counts made on the transitional lobe of both forms shows a significant difference in the proportion of chromophilic cells. The black moor contains a greater number of eosinophiles than does the common variety. 7. The morphological differences between the black moor and the common variety of goldfish may possibly be related to the differences in the pro- portional distribution of the different chromophilic cells in their pituitary glands. Literature Cited. Addison, W. H. F. 1916. The Golgi apparatus in the cells of the distal glandular portion of the hypophysis. Anat. Rec., vol. 11, Suppl., pp. 317-318. Atwell, W. J. 1929. Characteristics of the Golgi apparatus in the different types of cells in the anterior lobe of the Cat’s hypophysis. Anat. Rec., vol. 42, Suppl., p. 44. Bell, W. R. 1938. Morphology of the hypophysis of the common goldfish (Carassius auratus L.). Zoologica, vol. 23, pp. 219-234. Bock, F. 1928. Die Hypophyse des Strichlings (Gasterosteus aculeatus L.) unter besonderer Beruchsiehtigung der jahrescyklischen Veranderungen. Zeit. f. TFfss. Zool., Bd. 131, S. 643-710. 58 Zoologica: New York Zoological Society [XXIV :4 Charipper, H. A. 1937. The morphology of the hypophysis in lower vertebrates, particularly fish and amphibia, with some notes on the cytology of the pituitary in Carassius auratus (the goldfish) and Necturus maculosus (the mud- puppy). Cold Spring Harbor Symposia on Quantitative Biology, vol. 5, pp. 151-164. Crew, F. A. E. 1923. The significance of an achondroplasia-like condition met with in cattle. Proc. Roy. Soc. B., vol. 95, pp. 228-255. Cushing, H. 1932a. The basophile adenomas of the pituitary body and their clinical mani- festations (pituitary basophilia). Johns Hopkins Bull., vol. 50, pp. 137-195. 1932b. Papers relating to the pituitary body, hypothalamus and parasym- pathetic nervous system. Charles C. Thomas, Baltimore. Florentin, F. 1934 Histophysiologie comparee de l’hypophyse. L’excretion de la colloide hypophysaire chez les Teleosteens. Ann. de Physiol, et Physico. Biol., T. 10, pp. 963-965. Gentes, L. 1907. L’hypophyse des vertebres. Compt. Rend. Soc. de Biol., T. 63, pp. 120- 122. Herring, P. T. 1908. A contribution to the comparative physiology of the pituitary body. Quart. Jour. Exp. Physiol., vol. 1, pp. 261-280. Keith, A. 1922. The evolution of human races in the light of the hormone theory. Johns Hopkins Hosp. Bull., vol. 33, pp. 155-159. Kirkman, H. 1937. A cytological study of the anterior hypophysis of the guinea pig and a statistical analysis of its cell types. Amer. Jour. Anat., vol. 61, pp. 233-287. Matthews, S. A. 1936. The pituitary gland of Fundulus. Anat. Rec., vol. 65, pp. 357-367. Severinghaus, A. E. 1933. A cytological study of the anterior pituitary of the rat, with special reference to the Golgi apparatus and to cell relationship. Anat. Rec., vol. 57, pp. 149-176. Smith, P. E. & E. C. MacDowell. 1930. An hereditary anterior pituitary deficiency in the mouse. Anat. Rec., vol. 46, pp. 249-257. Stendell, W. 1913. Zur vergleichenden Anatomie und Histologie der Hypophysis cerebri. Arch. f. Mikr. Anat., Bd. 82, S. 289-333. 1914. Die Hypophysis Cerebri. Oppel, Lehrbuch der vergl. mikr. Anat., 8 Teil, Jena. Stockard, C. R. 1931. The Physical Basis of Heredity. Norton and Co., New York. 1936. Defective endocrine glands associated with structural disharmonies and lethal reactions. Anat. Rec., vol. 64 Suppl., pp. 47-48. 1939] Levenstein: Cytology of Goldfish Pituitary 59 Tilney, F. 1911. Contributions to the study of the hypophysis cerebri with especial reference to its comparative anatomy. Memoirs of the Wistar Insti- tute of Anatomy and Biology, No. 2. Vicari, E. M. 1937. Pituitary acidophilia and basophilia in the German shepherd dog. Anat. Rec., vol. 67, Suppl., p. 51. 60 Zoologica: New York Zoological Society EXPLANATION OF THE PLATE. Plate I. Fig. 1. Section of the pars intermedia showing the relationships of the granular and agranular cells to each other and to the invading strands of the pars nervosa. X 240 Champy. Masson stain. Fig. 2. A group of cells from the pars intermedia showing the compact Golgi ap- paratus capping their nuclei. X 1350 Nassonov-Kolatchev. Masson stain. Fig. 3. A group of cells from the pars anterior showing the granular Golgi ap- paratus next to the nuclei. X 1350 Nassonov-Kolatchev. Masson stain. Fig. 4. A group of transitional lobe cells showing the large granules in the basophilic cells and the highly granular, heavily staining acidophilic cells. X 450 Champy. Masson stain. Fig. 5. A section showing the heavy strand, compact Golgi apparatus capping the acidophilic cells of the transitional lobe. X 1350 Nassonov-Kolatchev. Masson stain. Fig. 6. A group of chromophobic cells from the transitional lobe showing the types of Golgi apparatus present in these cells. One type is more loosely arranged and made up of thinner strands than the other. X 1350. Nas- sonov-Kolatchev. Masson stain. Fig. 7. A group of basophilic cells from the transitional lobe containing a loosely arranged strand-like Golgi net in the cytoplasm of the cells. X 1350 Nassonov-Kolatchev. Masson stain (Compare with Fig. 5). Fig. 8. A group of cells from the pars anterior showing the complete lack of cellular arrangement present here. X 600 Champy. Masson stain. Fig. 9. A section of the gland showing the ramifications of the pars nervosa and its intimate connection with the large blood vessels of the gland. X 170 Champy. Masson stain. LEVENSTEIN. PLATE I. FIG. 1. FIG. 4. FIG. 7. FIG. 8. FIG. 3. FIG. 6. FIG. 9. THE CYTOLOGY OF THE PITUITARY GLAND OF TWO VARIETIES OF GOLDFISH (CARASSIUS AURATUS L.). Crandall: Plumage Changes in the Quetzal 61 5. Notes on Plumage Changes in the Quetzal. Lee S. Crandall Curator of Birds, New York Zoological Park (Plate I). On October 29, 1937, nine young Quetzals ( Pharomachrus mocinno mocinno De la Llave) were received at the New York Zoological Park. Six were forwarded a few weeks later to the Zoological Society of London. Of the remaining three, one died on December 14, 1937, from a bacterial in- vasion of the lung. The second, which had suffered a wing injury before arrival, survived until June 16, 1938. The third bird, a male, lived until March 3, 1939, and it was on this specimen that the following observations were made. Before the arrival of these birds, there appear to be no records of living specimens having reached either Europe or America. In 1914, when in San Jose, Costa Rica, I called on the late Senor Jose C. Zeledon who was closely connected with the development of Costa Rican ornithology. The discussion turned to Quetzals and I was informed that Senora Zeledon had possessed two birds of the Costa Rican form ( Pharomachrus mocinno costari- censis (Cabanis)), at different times. Each had been kept without difficulty for six months but had been liberated in turn “because they smelled so bad.” I know of no other instance of Quetzals having been kept in captivity. The birds received at the Zoological Park had been taken fi-om the nest in the mountains of Honduras and were hand-reared by Dr. Wolfgang von Hagen. Collections were made in July and August, the birds being still unable to fly at that time. On arrival, they checked well against the nestling plumage described by Ridgway1 except that all showed some infusion of green in the scapulars and upper wing coverts. Since Ridgway’s description was taken from a nestling of the Costa Rican form, this diffei’ence may be accounted for. In all, the bill was black, iris dark brown, feet gray-blue. About January 1, 1938, a molt was begun by the bird under obsei'vation. By January 24, the upper plumage had become bright gi’een, except for the head, which remained brown. The throat and upper breast wei'e gi’ay, scaled with green, the lower chest was clear gray, the abdomen and under tail coverts were rosy salmon. The elongated upper wing coverts reached a length of about two inches. Remiges and rectrices wex-e not molted. The bill was black. On March 15, it was noted that both upper and lower mandible wei’e becoming suffused with yellow. The two bright green middle upper tail coverts, which had been growing slowly, now reached their maximum length, which was just short of the tips of the rectrices. The bi’own feathei’s of the head began a slow process of x’eplacement about April 15. By June 9, the change to bright green was complete, l Ridgway, Robert. Birds of North and Middle America, part V, p. 737. 62 Zoologica: Neiv York Zoological Society [XXIV :5 although the crest was but slightly developed. Yellow areas of the bill had increased, leaving only a dark spot over each nostril and at the tip of each mandible. At this stage, the bird checked closely with Ridgway’s “immature male.” On July 10, a tail feather was dropped and on the 28th, another. Begin- ning on July 20, there was a general molting of body feathers. This con- tinued for several days but then checked and was not resumed until about September 10. On the 16th of this month, a middle upper tail covert was dropped and was found to measure 185 mm. in total length. By mid-October, the molt was in full swing, with body feathers, remiges and rectrices being rapidly replaced. The plumage change was complete by November 15, when the following description was taken: Upper parts golden green, crest well developed. Upper breast green, gray of lower chest interspersed with deep crimson. Vent and under tail coverts geranium red. The two middle upper tail coverts projected about three-quarters of an inch beyond the rectrices. Elongated greater wing coverts green for most of their visible length, the black bases barely discernible. Remiges black, primaries and outer secondaries edged with buff. Three inner pairs of rectrices black, three outer pairs white, the bases barred with black, white areas increasing outward. Bill clear yellow, feet pale gray-blue, iris dark brown. This plumage appears to represent a stage intermediate between Ridg- way’s “immature male” and “adult male.” It is quite possible that there may be further stages, since it seems unlikely that the superb beauty of the adult could be attained except by degrees. The molt of this bird is note- worthy for its almost continual progression, since there was hardly a time when change was not occurring in some part of its plumage. On the other hand, once the stage described above had been reached, on November 15, 1938, there were no changes up to the time of the bird’s death on March 3, 1939. Immediately after death, the following notes were made, measurements being in millimeters : length, to end of rectrices, 357 ; wing, 188 ; tail, 190 ; right middle upper tail covert, 203; left, 213; crest, 26; culmen, 16; tarsus, 16. Weight, 108 grams. Sex $. 1939] Crandall: Plumage Changes in the Quetzal 63 EXPLANATION OF THE PLATE. Fig. 1. Nestling plumage, photographed November 18, 1937. Fig. 2. First immature plumage, photographed July 5, 1938. Fig. 3. Second immature plumage, front view, photographed February 20, 1939. Fig. 4. Second immature plumage, rear view, photographed February 20, 1939. CRANDALL. PLATE I. ro 6 c\j 0 Ll. NOTES ON PLUMAGE CHANGES IN THE QUETZAL. JJcto gork Zoological Society General Office: 90 Broad Street, New York City Officers President, W. Redmond Cross Vice-Presidents, Alfred Ely and Laurance S. Rockefeller Chairman, Executive Committee, W. Redmond Cross Treasurer, Cornelius R. Agnew Secretary, Fairfield Osborn Scientific Staff Zoological Hark W. Reid Blair, Director Raymond L. Ditmars, Curator of Mammals and Reptiles Claude W. Leister, Ass’t to the Director and Curator, Educational Activities H. C. Raven, Prosector Edward R. Osterndorff, Photographer William Bridges, Editor and Curator of Publications Squarium Charles M. Breder, Jr., Acting Director Christopher W. Coates, Aquarist Ross F. Nigrelli, Pathologist G. M. Smith, Research Associate in Pathology Homer W. Smith, Research Associate in Physiology Departmrnt of tropical &tstarcf) William Beebe, Director and Honorary Curator of Birds Lee S. Crandall, Curator of Birds Charles R. Schroeder, Veterinarian John Tee-Van, General Associate Gloria Hollister, Research Associate Jocelyn Crane, Technical Associate Cbitorial Committee Fairfield Osborn, Chairman W. Reid Blair William Beebe Charles M. Breder, Jr. William Bridges ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME XXIV Part 2 Numbers 6-9 PUBLISHED BY THE THE ZOOLOGICAL PARK, July 31, 1939 SOCIETY NEW YORK CONTENTS PAGE 6. Deep-sea Fishes of the Bermuda Oceanographic Expedi- tions. Family Melanostomiatidae. By William Beebe & Jocelyn Crane. (Text-figures 1-77) 65 7. The Eye Structure of the Four-eyed Blenny, Dialommus fuscus Gilbert. By C. M. Breder, Jr., & E. B. Gresser. (Plates I & II; Text-figures 1-3) 239 8. Studies on Virus Diseases of Fish. III. Morphological and Experimental Observations on the Lymphocystis Disease of the Pike Perch, Stizostedion vitreum. By Richard Weissenberg. (Plate I) 245 9. Studies on Lymphocystis Disease in the Orange Filefish, Ceratacanthus schoepfi (Walbaum), from Sandy Hook Bay, N. J. By Ross F. Nigrelli & G. M. Smith. (Plates I- VIII) 255 Beebe & Crane: Family Melanostomiatidae 65 6. Deep-sea Fishes of the Bermuda Oceanographic Expeditions. Family Melanostomiatidae.1 William Beebe & Jocelyn Crane Department of Tropical Research, New York Zoological Society. (Text-figures 1-77). Contents. Page I. Introduction 66 II. Summary of Important Points 66 III. Suborder Stomiatoidea i 69 Superfamily Gymnophotodermi 69 Family Melanostomiatidae 70 A. Taxonomic Discussion 70 B. Family Characters in Brief 73 C. Development 74 History and Taxonomy ••• 74 Present Material 75 Division of Developmental Period into Stages 76 Diagnostic Characteristics of Stomiatoid Growth Stages 76 Identification of Melanostomiatid Larvae 79 D. Form and Development of Individual Characters 80 1. Color and Luminescence 80 2. Body Form and Proportions 85 3. Barbel 88 4. Light Organs 89 5. Teeth and Gill Teeth 91 6. Branchiostegal Rays 96 7. Fins 96 8. Epidermal Grooves 97 9. Osteology 97 10. Coelomic Organs 99 11. Shrinking 102 E. Ecology 102 1. Horizontal Distribution 102 2. Vertical Distribution 102 3. Abundance 103 4. Food and Enemies 103 5. Activity 104 F. Phylogeny 104 G. Synopsis of the Genera 109 H. Report on the Collection of the Bermuda Oceanographic Expeditions, includ- ing Revisions of Genera and Species Ill Genus Chirostomias Regan & Trawavas Ill C. pliopterus Regan & Trewavas 116 Genus Pachystomias Gunther 117 P. atlanticns Regan & Trewavas 119 Genus Leptostomias Gilbert 121 L. bermudensis Beebe 125 L. gladiator Regan & Trewavas 127 Genus Echiostoma Lowe 130 E. tanneri (Gill) 134 Contribution, Bermuda Biological Station for Research, Inc. AUG < 3 1939 66 Zoologica: New York Zoological Society [XXIV :6 Genus Melanostomias Brauer 142 M. spilorhynchus Regan & Trewavas 148 M. biseriatus Regan & Trewavas 152 Genus Photonectes Gunther 154 P. dinema Regan & Trewavas 162 P. leucospilus Regan & Trewavas 164 P. braueri (Zugmayer) 165 P. mirabilis Parr 167 P. cornutus Beebe 169 P. parvimanus Regan & Trewavas 170 P. bifilifer Beebe 173 P. margarita (Goode & Bean) 175 Genus Flagellostomias Parr 179 F. boureei (Zugmayer) 184 Genus Grammatostornias Goode & Bean 185 G. dentatus Goode & Bean 190 G. flagellibarba Holt & Byrne 192 Genus Bathophilus Giglioli 196 B. brevis Regan & Trewavas 199 B. altipinnis Beebe 203 B. metallicus (Welsh) 205 B. longipinnis (Pappenheim) 208 B. chironema Regan & Trewavas 209 Genus Eustomias Vaillant 210 E. obscurus Vaillant 217 E. bibulbosus Parr 219 E. simplex Regan & Trewavas 221 E. dubius Parr 222 E. lipochirus Regan & Trewavas 224 E. bigelowi Welsh 225 E. silvescens Regan & Trewavas 228 E. schmidti Regan & Trewavas 230 E. fissibarbis Pappenheim 232 IV. Bibliography 233 Introduction. For detailed data in regard to nets, locality, dates, etc., concerning the capture of the deep-sea fishes treated in this monograph, refer to Zoologica, Vol. XIII, Nos. 1, 2 and 3 and Vol. XX, No. 1, pp. 1-2. For physical data, methods of measurement and general definitions of growth stages, see Zoo- logica, Vol. XVI, No. 1. For the related family Idiacanthidae, see Zoologica, Vol. XX, No. 4. The majority of the drawings in the present paper are the work of Harriet Bennett; Figs. 8 and 12 are by George Swanson. For the dyeing and clearing of many specimens and for the tail drawings in Text-fig. 9 we are indebted to Gloria Hollister. We wish to express our appreciation to the following persons for their generous cooperation in lending specimens and enabling us to examine and sex type material deposited in various museums: Dr. J. R. Norman of the British Museum; Dr. A. Vedel Taning of the Carlsberg Foundation’s Marine Biological Laboratories, Copenhagen; Dr. Clinton V. MacCoy of the Boston Society of Natural History; Dr. A. E. Parr of the Peabody Museum; Dr. William C. Schroeder of the Museum of Comparative Zoology, and Dr. Leonard P. Schultz of the U. S. National Museum. Summary of Important Points. Material. The Bermuda collection of Melanostomiatidae, taken in a cylinder of water 8 miles in diameter and a mile deep, consists of 250 speci- mens belonging to 10 genera and 32 species. Previously known melanostomia- tids number about 1,450 specimens belonging to 16 genera and, at a generous estimate, slightly less than 100 valid species. The advantages, therefore, of continued, concentrated collecting in a single, definite area of ocean are again evident, since from the Bermuda 8-mile circle have come more than 62% of all known genera and at least a third of the species taken in all seas. In addition to the study of our own collection, we have examined (a) examples of all genera except Opostomias and Pareicstomias, (b) all of the AUS s 1939 1939] Beebe & Crane: Family Melanostomiatidae 67 melanostomiatids deposited on this side of the Atlantic and (c) a number of specimens on loan from abroad. Taxonomy. 1. The sub-division of the Stomiatoidea into Gymnophoto- dermi, Lepidophotodermi and Heterophotodermi, suggested by Parr in 1927, is adopted, except that the three groups are given the status of superfamilies instead of suborders. 2. Parr’s family Melanostomiatidae, exclusive of the malacosteids, is maintained, the family Stomiatidae being limited to Stomias, Macrostomias and Stomioides. 3. The following genera are synonymized: Lamprotoxus Holt & Byrne, 1913 = Grammatostomias Goode & Bean, 1895. Haplostomias Regan & Trewavas, 1930 = Melanostomias Brauer, 1902. Stomiatella Roule & Angel, 1930 {part.) = Bathophilus Giglioli, 1884, and ? Flag ellostomias Parr, 1927. Parastomias Roule & Angel, 1931 = Eustomias Vaillant, 1888. Microdontostomias Fowler, 1934 = Stomias Cuvier, 1817 (Family Stomiatidae) . Pseudeustomias Fowler, 1934 = Stomias Cuvier, 1817 (Family Stomi- atidae) . Photonectops Chapman, 1939 = Tactostoma Bolin, 1939. 4. The following species are synonymized : Chirostomias lucidimanus Beebe, 1932 = C. pliopterus Regan & Tre- wavas, 1930. Leptostomias problematicus (Parr, 1927) = L. gladiator (Zugmayer, 1911). Leptostomias ramosus Regan & Trewavas, 1930 = L. gladiator. Echiostoma ctenobarba Parr, 1927 = E. tanneri Gill, 1883. Echiostoma guentheri Regan & Trewavas, 1930 = E. barbatum Lowe, 1843. Echiostoma calliobarba Parr, 1934 = E. tanneri Gill, 1883. Echiostoma ctenobarba ramifera Parr, 1934 = E. tanneri Gill, 1883. Melanostomias bulbosus Beebe, 1933 = M. spilorhynchus Regan & Tre- wavas, 1930. Melanostomias heteropogon Regan & Trewavas, 1930 = ?M. valdiviae Brauer, 1902. Melanostomias melanocaulus Regan & Trewavas, 1930 = ?M. valdiviae Brauer, 1902. Melanostomias albibarba Regan & Trewavas, 1930 = M. melanops Brauer, 1902. Melanostomias stewarti Fowler, 1934 = M. valdiviae Brauer, 1902. Melanostomias vierecki Fowler, 1934 = M. valdiviae Brauer, 1902. Photonectes richardi (Zugmayer, 1913) = P. margarita (Goode & Bean, 1895). Photonectes flagellatus Parr, 1927 = P. margarita (Goode & Bean, 1895). Photonectes intermedins Parr, 1927 = P. margarita (Goode & Bean, 1895). Photonectes ovibarba Regan & Trewavas, 1930 = P. braueri (Zugmayer, 1913). Photonectes caerulescens Regan & Trewavas, 1930 = P. achirus Regan & Trewavas, 1930. Photonectes monodactylus Regan & Trewavas, 1930 = P. margarita (Goode & Bean, 1895). Lamprotoxus phanobrochus Regan & Trewavas, 1930 = Grammatosto- mias flag ellibarba Holt & Byrne, 1910. Lamprotoxus paucifilis Regan & Trewavas, 1930 •— Grammatostomias flagellibarba Holt & Byrne, 1910. 68 Zoologica: New York Zoological Society [XXIV :6 Lamprotoxus angulifer Beebe, 1932 = Grammatostomias dentatus Goode & Bean, 1895. Bathophilus alberti (Roule & Angel, 1931) = B. metallicus (Welsh, 1923). Eustomias bibulbosus arborifer Parr, 1927 = E. bibulbosus Parr, 1927. Eustomias bituberatus Regan & Trewavas, 1930 — ?E. micraster Parr, 1927. Eustomias schiffi Beebe, 1932 = E. dubius Parr, 1927. Eustomias bigelowi paucifilis Parr, 1927 = E. bigelowi Welsh, 1923. Eustomias bigelowi parvibulbus Parr, 1927 = E. bigelowi Welsh, 1923. Eustomias dendriticus Regan & Trewavas, 1930 = ?E. fissibarbis Pap- penheim, 1914. Eustomias frondosus Regan & Trewavas, 1930 = E. binghami Parr, 1927. Eustomias satterleei Beebe, 1933 = E. silvescens Regan & Trewavas, 1930. Eustomias triramis Regan & Trewavas, 1930 = ?E. bigelowi Welsh, 1923. Other species, especially in the genera Leptostomias and Eustomias, will doubtless prove also to be synonymous. Color and Luminescence. Color notes and sketches were made from more than 100 freshly caught specimens belonging to 28 species; 10 indi- viduals, representing 5 species, were living and their luminescence and behavior noted. This work was supplemented by observations made from the Bathysphere. Up to the present, only eight specimens in the entire family had been studied when freshly caught or recently preserved. Sexual Dimorphism. In the majority of genera the postorbital light organ is almost or completely atrophied in adult females. Striking sexual differences are found in the barbels of Eustomias ; similar differences are suspected in other genera. The importance of sexing type specimens in this family is obvious. Development. Larvae and post-larvae of the following species have been identified for the first time: Flagellostomias boureei, Leptostomias gladiator, Melanostomias spilorhynchus , Melanostomias biseriatus, Photo- nectes parvimanus, Grammatostomias flagellibarba, Bathophilus brevis, Bathophilus metallicus, Bathophilus near longipinnis, Bathophilus sp., Eus- tomias bibulbosus, Eustomias dubius, Eustomias spp. In addition, our collec- tion includes adolescents of most other species taken by the Bermuda Expe- ditions. This material has been sufficient for the tabulations of family and generic juvenile characters. A number of recorded species prove synonymous with others due to their being based on juvenile characters, such as partly developed barbels. Dissection alone can determine whether or not a specimen is wholly adult, with well developed gonads, fully pigmented stomach of relative length typical of the genus, and hour-glass-shaped centra. Special Characteristics. In the comparison of the form and develop- ment of various parts of the body in the different genera, certain structures have proved to be of unexpected phylogenetic or taxonomic importance. Among the most interesting are the form and distribution of gill-teeth in adults, the presence of spiny gill-rakers in larvae and post-larvae, the varia- tion of larval pigment patterns and the development of the eye. Probable relationships of the genera to each other and to adjacent families are dis- cussed. Vertical Distribution. As with most families of Bermuda deep-sea fish, the depths at which these fish are taken are greater than the average in other areas, practically none except very young melanostomiatids having been taken above 500 fathoms, although they were seen above this level from the Bathysphere. 1939] Beebe & Crane: Family Melanostomiatidae 69 Suborder Stomiatoidea. Characteristics : Oceanic isospondyls differing from the Clupeoidea and Salmonoidea in the presence of photophores, which are arranged typically in a double series along the abdomen and in a single series above the anal fin. Discussion : It should now be generally agreed that the suborder Stomia- toidea be divided into eight families, namely, the Gonostomatidae, Sterno- ptychidae, Chauliodontidae, Stomiatidae, Astronesthidae, Melanostomiatidae, Malacosteidae and Idiacanthidae. In 1927 (p. 1) and 1930 (p. 136), Parr proposed separating the smooth- skinned members of the old family Stomiatidae (including the malacosteids) from the scaly Stomias and placing the former in a family of their own, Melanostomiatidae. This new family he grouped with the other smooth- skinned stomiatoids, the Astronesthidae and Idiacanthidae, in a new sub- order, Gymnophotodermi. A second suborder, Lepidophotodermi, was pro- posed to include the Stomiatidae proper ( Stomias and Macrostomias ) and, provisionally the Chauliodontidae. A third suborder, Heterophotodermi, was suggested to embrace the Gonostomatidae and Sternoptychidae. Regan and Trewavas, on the other hand, in the Dana “Fishes of the Families Stomiatidae and Malacosteidae” (1930) employed the old classifica- tion, including Stomias (as an aberrant genus), Idiacantlius, and all the smooth-skinned fishes with posterior vertical fins and complete floors to their mouths in the single family Stomiatidae. The malacosteids were treated as a separate family. From our own studies we draw the following conclusions: 1. Parr’s three suborders, the Gymnophotodermi, Lepidophotodermi and Heterophotodermi, are valid and useful subdivisions of the stomiatoid isospondyls. However, since the apparently quite natural group of Stomia- toidea as defined on the preceding page is generally regarded as a suborder of the order Isospondyli, we propose that each of Parr’s three divisions be given the rank of superfamily instead of suborder. 2. Parr’s family Melanostomiatidae (excluding Malacosteus, Aristosto- mias and Photostomias) , should unquestionably be maintained. 3. The three genera just mentioned, along with the more recent genus Ultimostomias described by Beebe in 1933, should form the family Mala- costeidae, as suggested by Regan and Trewavas (1930). 4. The family Idiacanthidae should be maintained, due chiefly to its exceptional life-history (Beebe, 1934.1). 5. The family Stomiatidae should be limited to Stomias, Macrostomias, and Stomioides Parr, 1933. Superfamily Gymnophotodermi. Characteristics : Naked Stomiatoidea with black skin, large mouths, barbel (except in Malacosteus) , postorbital luminous organ present at least in males, serial photophores developed on branchiostegal membranes and isthmus as well as in the usual lateral and ventral rows on each side (vesti- gial in Malacosteus and Bathophilus brevis ) ; these organs lacking lumen or duct; smaller organs usually scattered on skin. Parietals small and well separated or absent ; orbitosphenoid absent ; opisthotic absent ; entopterygoid membranous or very thinly ossified; preoperculum slender; vertebral centra thin cylinders of bone enclosing notochord; parapophyses and at least anterior neural arches not ankylosed with centra; parapophyses with pleural ribs; epipleurals present or absent; epineurals present. Long caecal stomach present, giving off a short arm anteriorly which opens into the usually straight intestine. 70 Zoologica: Neiv York Zoological Society [XXIV :6 Key to the families : A. Dorsal fin not confined to caudal peduncle. B. Dorsal fin short, ending before anal origin Astronesthidae. BB. Dorsal fin very long, extending almost to caudal base Idiacanthidae. AA. Dorsal fin confined to caudal peduncle. C. Lower jaw and hyoid arch joined by a membrane, forming a floor to the mouth Melanostomiatidae. CC. Lower jaw and hyoid arch not joined by a membrane, the sym- physis and hyoid being connected only by a muscular cord Malacosteidae. Family Melanostomiatidae. A. Taxonomic Discussion. Thirty-two generic names have been proposed for fishes referred to this family, or to naked fishes of the old family Stomiatidae. Of these we recognize 16 as valid. In chronological order, the 32 names, along with their present standing and the type species, are as follows : 1. Echiostoma Lowe, 1843. Valid. Type: E. barbatum Lowe, 1843. 2. Opostomias Gunther, 1878. Valid. Type: O. micripnus Gunther, 1878. 3. Pachystomias Gunther, 1878. Valid. Type: P. microdon Gunther, 1878. 4. Lucifer Doderlein, 1882. = Photonectes. Name given by Gunther, 1887, because Lucifer preoccupied. Type : Lucifer albipinnis Doderlein, 1882. 5. Hyperchoristius Gill, 1883. = Echiostoma. Synonymized by Parr, 1927. Type: H. tanneri Gill, 1883. 6. Bathophilus Giglioli, 1884. Valid. Type: B. nigerrimus Giglioli, 1884. 7. Photonectes Gunther, 1887. Valid. Type: P. albipinnis (Doderlein, 1882). 8. Eustomias Vaillant, 1888. Valid. Type: E. obscurus Vaillant, 1888. 9. Grammatostomias Goode & Bean, 1895. Valid. Type: G. dentatus Goode & Bean, 1895. 10. Dactylostomias Garman, 1899. = Bathophihis. Synonymized by Parr, 1927. Type : B. filifer Garman, 1899. 11. Melanostomias Brauer, 1902. Valid. Type: M. valdiviae Brauer, 1902. 12. Leptostomias Gilbert, 1905. Valid. Type: L. macronema Gilbert, 1905. 13. Neostomias Gilchrist, 1908. = Eustomias. Synonymized by Parr, 1927. Type: E. filiferum Gilchrist, 1908. 14. N ematostomias Zugmayer, 1911. = Leptostomias. Synonymized by Parr, 1927. Type: N. gladiator Zugmayer, 1911. 15. Trichostomias Zugmayer, 1911. = Bathophilus. Synonymized oy Parr, 1927. Type: T. vaillanti Zugmayer, 1911. 16. Gnathostomias Pappenheim, 1911. = Bathophilus. Synonymized by Parr, 1927. Type: G. longifilis Pappenheim, 1911. 17. Lamprotoxus Holt & Byrne, 1913. = Grammatostomias. Synony- mized in present paper; see below. Type: L. flagellibarba (Holt & Byrne, 1910). 18. Flagellostomias Parr, 1927. Valid. Type: F. tyrannus Parr, 1927. = F. boureei (Zugmayer, 1911). 1939] Beebe & Crane: Family M elanostomiatidae 71 19. Chirostomias Regan & Trewavas, 1930. Valid. Type: C. pliopterus Regan & Trewavas, 1930. 20. Trigonolampa Regan & Trewavas, 1930. Valid. Type: T. miriceps Regan & Trewavas, 1930. 21. Thysanactis Regan & Trewavas, 1930. Valid. Type: T. dentex Regan & Trewavas, 1930. 22. Haplostomias Regan & Trewavas, 1930. = Melanostomias. Synony- mized in present paper ; see below. Type : H. tentaculatus Regan & Trewavas, 1930. 23. Odontostomias Norman, 1930. Valid. Type: O. micropogon Norman, 1930. 24. Stomiatella Roule & Angel, 1930 {part.). = Bathophilus (larva) and Flagellostomias (larva). Synonymized in present paper (pp. 74, 75). 25. Stylophthalmella Roule & Angel, 1930 {part.). = Eustomias (larva). Synonymized in present paper (p. 75). 26. Pareustomias Bailly, 1930. Probably valid, although apparently very close to Eustomias. Type: P. chabanaudi Bailly, 1930. 27. Parastomias Roule & Angel, 1931. — Eustomias. Synonymized in present paper. Type. P. tetranema (Zugmayer, 1911). 28. Elapterostomias Fowler, 1934. — Borostomias, family Astrones- thidae. Synonymized by Myers, 1935 (in footnote, p. 2). 29. Microdontostomias Fowler, 1934. = Stomias, family Stomiatidae. Synonymized in present paper; see below. 30. Pseudeustomias Fowler, 1934. = Stomias, family Stomiatidae. Syn- onymized in present paper; see below. 31. Tactostoma Bolin, 1939. Valid. Type: T. macropus Bolin, 1939. 32. Photonectops Chapman, 1939. = Tactostoma Bolin, 1939, the latter genus having priority by a few weeks. Type: P. midtipunctata Chapman, 1939, probably synonymous with T. macropus. The systematic position of the six-foot fish Bathysphaera Beebe, 1932, described from a specimen observed from the Bathysphere off Bennuda, is of course still uncertain, since no specimen has yet been taken. That it is a stomiatoid, near the Melanostomiatidae, seems certain, but in all proba- bility it does not belong in this family, and is therefore omitted from sys- tematic treatment in the present paper. For reference, however, we include herewith a copy of the type description : “On the twentieth dive in the Bathysphere, at a depth of 2100 feet, we saw two large, elongate, barracuda-shaped fish, which twice passed within eight feet of the windows, once partly through the beam of our electric light. These were at least six feet in length. Bathysphaera Intacta. Actual length about 6 feet. Numbers of photophores and finrays approximate. 72 Zoologica: New York Zoological Society [XXIV :6 “No direct lights were visible on the head, yet the rather large eye and the faint outline were distinct. There was a single row of strong, pale blue lights along the side, large and not far from twenty in number. The mouth, with strongly undershot jaw, and numerous fangs was illumined either by mucous or indirect internal lights along the branchiostegals. “The fish reminded me in general of barracudas, with deeper jaws open all the time. Posteriorly placed vertical fins were seen when they passed through the electric beam. There were two ventral tentacles, each tipped with a pair of separate, luminous bodies, the superior reddish, the lower one blue. These twitched and jerked along beneath the fish, one undoubtedly arising from a mental base, the other so far back that its origin must have been at the anal fin. Neither the stem of the tentacles nor paired fins were distinguishable. “I assume from the position of the vertical fins and the general facies, that the position of the fish must be somewhere near the Melanostomiatidae, but the single line of large, lateral photophores and the two ventral tentacles set it apart from any known species or genus. “The depth was 2100 feet, the date September 22nd, 1932, the position 32°17' No. Lat., 64°36' West Long., 5 miles southeast of Nonsuch Island, Bermuda. “Relying on this recognizable diagnosis I propose for it the name of Bathysphaera intacta, the Untouchable Bathysphere Fish.” (Beebe, 1932.2 pp. 175-177). We have compared specimens of Lamprotoxus with the type of Gram- mat os tomias dentatus at the United States National Museum, and found the two genera to be unquestionably synonymous. For a detailed discussion, see p. 185. Through the kindness of Dr. Norman, we have been able to examine a specimen of Haplostomias Regan & Trewavas from the British Museum. In view of the very slight differences between this genus and Melanostomias, compared with the large differences between these genera and their nearest relatives, we have little hesitation in synonymizing these two groups. (See p. 143.) At the United States National Museum we have also examined the type specimens of Microdontostomias and Pseudeustomias. In his descriptions (1934) Fowler does not mention that both of these fish have the hexagonal scales of S tomias. They are, in fact, both clearly members of that genus. Microdontostomias orientalis, the type and unique species, is close to or identical with Stomias nebulosus Alcock. It is likely that a third barbel fila- ment, which would make the synonymy certain, has been broken off. Simi- larly, Pseudeustomias myersi, the type and unique specimen, belongs unques- tionably in the elongatus-valdiviae-affinis group of the genus Stomias (see Parr, 1934). The barbel bulb, instead of terminating in a single filament, as is stated in the description, ends in three of about equal length, as is usual in the genus. The figures (Fowler, loc. cit. figs. 21 and 22) of both the proposed genera are inexact in regard to the general appearance of the fish, since both specimens have the pronounced slenderness and strongly curved jaws characteristic of Stomias. In regard to the advisability of dividing Eustomias into two or more genera, we agree with Regan and Trewavas that cleancut divisions cannot be made; also the groups of species are so much closer to one another than they are to other genera that any division seems wholly inadvisable. There- fore, we do not accept the proposal of Roule and Angel (1931) that the name Parastomias be given to the species which have branched barbels. (See p. 210.) In addition to the 10 genera of Melanostomiatidae taken by the Ber- muda Expeditions, we have examined specimens of Trigonolampa, Odontos- tomias, Thysanactis, Haplostomias (which we synonymize with Melanosto- 1939] Beebe & Crane: Family Melanostomiatidae 73 mias ) and Tactostoma (the type of Photonectops; examination superficial). This leaves Opostomias and Pareustomias as the only valid genera which we have not examined; both are known from unique types. A key to the genera of Melanostomiatidae as now understood will be found on page 109. B. Family Characters in Brief. Gymnophotodermi, usually elongate and little compressed, with very ' short caudal peduncle, to which the vertical fins are entirely confined ; lower jaw and hyoid arch joined by a membrane, forming a floor to the mouth; teeth in jaws highly developed, often depressible; premaxillary almost always with a small, anterior ascending process; maxillary usually much longer than premaxillary, forming posterior part of upper margin of mouth, and usually furnished with normal, erect teeth anteriorly and oblique denticles posteri- orly ; vomer with or without teeth ; palatine usually toothed ; gill-arch teeth usually present, often in pairs or groups; a series of teeth, usually strong, on third and fourth pharyngobranchials (= upper pharyngeals) ; branchi- ostegals 8 to 22; pectorals well developed, reduced or absent; pelvics typi- cally of seven rays, their insertion usually at or behind middle of body, rarely in front of it; caudal fin very short, forked, the ventral lobe the longer; adipose fin absent except in Chirostomias; no pseudobranchiae; special grooves for barbel and pectoral fins often present. Skeleton moderately well developed, the jaws always more strongly ossified than any other portion of the body. Mesethmoid with or without lateral process; frontals united by suture; parietals present or absent; hyomandibular and quadrate forming with the jaw an angle of 45 degrees or less; one supramaxillary ; opercular apparatus weak, reduced or rudi- mentary; hyoid and branchial apparatus well developed; pectoral girdle moderately or feebly developed; post-temporal present or absent; upper and lower coracoids always present, mesocoracoid sometimes absent, all 3 ele- ments often reduced ; actinosts often reduced ; caudal fin alone strongly sup- ported; vertebrae moderately numerous, 35 to 82, an average number being around 60; anterior vertebrae usually more or less modified, permitting free movement of head. Usually two pyloric caeca ; gonads dorsal. Sexual dimorphism usually apparent in development of postorbital photophore, sometimes in form of barbel. Size: The largest known melanostomiatid is the unique specimen of Opostomias, measuring 380 mm. in length. The size records in the remaining genera are as follows: Echiostoma, 355 mm., (375 mm. when fresh) ; Photo- nectes, 340 mm. ; Odontostomias, 290 mm. ; Tactostoma, 280 mm. ; Leptosto- mias, 270 mm. (285 mm. when fresh); Melanostomias, 242 mm.; Trigono- lampa, 223 mm.; Flag ellostomias , 222 mm.; Grammatostomias, 206 mm.; Chirostomias, 205 mm. ; Eustomias, 204 mm. ; Pachystomias, 165 mm. ; Batho- philus, 140 mm. ; Tliysanactis 139 mm. ; Pareustomias, 62 mm. The specimens listed above of Echiostoma, Leptostomias , Melanosto- mias, Grammatostomias and Chirostomias vtere taken by the Bermuda Expe- ditions. In Leptostomias, Flag ellostomias and Thysanactis, at least, and doubt- less in other genera as well, no fully adult specimens have been taken. Judging from this fact and from the number of melanostomiatids more than a foot long which were seen from the Bathysphere, it is probable that larger specimens of these swiftly swimming fishes escape the net. The largest Echiostoma, Photonectes and Melanostomias, however, were definitely in breeding condition, while others near the lengths given in the above paragraph had well developed gonads. Larva, as far as known, moderately elongate, translucent, with the 74 Zoologica: New York Zoological Society [XXIV :6 posterior, unpaired fins of the adult; no yolk sac but a gut hanging below myomeral body and extending as a free tube beyond the anal fin; pigment spots usually present in a longitudinal series just below the dorsal mid-line, sometimes in additional rows above and/or below the lateral mid-line; temporary, small teeth present in jaws, and temporary gill-rakers, often bristling with minute spines, usually present; larval pectoral pad with raylets always present, even when pectoral is much reduced or absent in adult. The family characters given above will be discussed in detail, after a genera] account of development. C. Development. History and Taxonomy: Thanks to Lo Bianco, Jespersen & Taning, Ege, Sanzo, Regan & Trewavas, Roule & Angel, and Beebe, developmental stages of a number of representative stomiatoids have been recognized and described. In some cases, complete series have been obtained; in others, only one or two stages, known from single specimens, have so far been identified. Exclusive of the Melanostomiatidae, the stomiatoid genera of which one or more juvenile stages have been identified include the fol- lowing: the gonostomids Gonostoma, Cyclothone, Maurolicus, Ichthyococcus and Vinciguerria; the sternoptychids Argyropelecus and Sternoptyx; Stomias; Chauliodus; one or two astronesthids; a questionable Malacosteus ; and Idiacanthus. Representative references are included in the bibliography. Sanzo’s (1931) account of Chauliodus is especially complete, since he succeeded in raising a larva from the last three days in the egg through the thirteenth day after hatching, at which stage it was apparent that the little fish was identical with a free-swimming larva taken at the surface. The latter specimen, in turn, unquestionably formed a link with older Chauliodus in which generic characters were well established. Sanzo’s evidence for the continuity of the series is convincing, and of great interest to us in our study of the Melanostomiatidae, since Chauliodus is the most closely related genus of which the egg and pre-larva have been identified. On the other hand, very little previous work has been done on the larvae of the Melanostomiatidae. In 1914 Sanzo (pp. 1-12) described the first known larval melanostomiatid, Bathophilus nigerrimus. In 1930 Regan and Trewavas (p. 73) stated that Regan’s larva de- scribed in 1916 (p. 136) as Stylophthalmus macrenteron was in all prob- ability a larval Eustomias. They described briefly a number of similar young Eustomias from the Dana collection, including some metamorphosing and juvenile forms which could be subgenerically and sometimes specifically identified. Parr, in 1927, described as new species two immature Eustomias (adolescents) and Regan and Trewavas in the same Dana collection found a number of specimens belonging to other genera which could be specifically identified, or described, although they had juvenile characters remaining. It may also be remarked here that a majority of the Dana melanostomiatids, as well as those of other collections including our own, are immature — that is, in the transitional adolescent stage, in which most or all of the external characteristics of adults are present, but with immaturity apparent internally. These advanced specimens, however, obviously shed little light on the characteristics of early stages in the family. In 1930 Roule and Angel described and figured a number of stomiatoid larvae under the general names of Stomiatella and Stylophthalmella, sug- gesting, where possible, their systematic positions. Thanks to our addi- tional material, we are able to contribute further suggestions in regard to the identity of these larvae. Stomiatella A Roule & Angel (1930, p. 14; pi. I, fig. 6) : We agree with the authors and with Sanzo (1930, p. 89), that this larva should be re- ferred to Bathophilus. We do not, however, agree with Sanzo that it is 1939] Beebe & Crane: Family Melanostomiatidae 75 B. nigerrimus, providing that the pelvic fin is shown correctly in the figure, since it is much too far forward and too high. Comparison with Sanzo’s own figures of B. nigerrimus larvae (1931, pi. vii, figs. 7, 8) will show the difference. If the fin of Roule and Angel’s specimen is accurately shown, it is very likely that this fish is B. brevis Regan & Trewavas, 1930. Stomiatella D Roule & Angel (1930, p. 17; pi. I, figs. 10, 11): Roule and Angel suggest that both these larvae may be young Malacosteus niger Ayres. We agree that the specimen shown in their Fig. 11 should in all probability be referred to this species. We are certain, however, that their Fig. 10 represents a quite different form and belongs to some genus of Melanostomiatidae. Myomere counts are not given, but from the number shown in the figure (about 70 to the end of the anal), the distribution of the pigment, and the general facies, it appears very likely that the speci- men should be referred to Flag ellostomias or a closely related genus (cf. our Text-fig. 47). Stylophthalmella B Roule & Angel (1930, p. 52; pi. Ill, figs. 62, 63) : This larva almost certainly belongs to the genus Eustomias. Stylophthalmella D Roule & Angel (1930, p. 53; pi. Ill, figs. 66, 67, 68) : This larva remains a puzzle, and should be remembered by subsequent workers on the Melanostomiatidae and Malacosteidae, since in all proba- bility it belongs to one or the other of those families. The other Stylophthalmella larvae, as has already been pointed out (Beebe, 1934, p. 155) include argentinids (bathylagids), but no Idiacanthus. Stylophthalmella C ( loc . cit., p. 52, pi. Ill, figs. 64, 65) may be Chauliodus (cf. Sanzo, 1931, p. 82, pi. VI). The most recent study of melanostomiatid larvae is by Sanzo who re- describes and figures early stages of Bathophilus nigerrimus, the only species of the family known to occur in the Mediterranean (1931, p. 89; pi. VII). In the summary, up to the present time the only larvae and post- larvae of the Melanostomiatidae (as defined in the present paper) have been Bathophilus nigerrimus, B. brevis?, Flag ellostomias? and Eustomias ssp. Present Material: To this list we add the following forms, of which larvae, post-larvae or both have been taken by the Bermuda Oceanographic Expeditions : Flag ellostomias boureei : larvae and post-larvae. Leptostomias gladiator: larvae and post-larvae. Melanostomias spilorhynchus : larva and post-larva. Photonectes parvimanus: larva and post-larva. Grammatostomias fiagellibarba: post-larva. Bathophilus brevis : post-larva. Bathophilus, near longipinnis: larva. Bathophilus metallicus: post-larva. Bathophilus sp. : larva. Eustomias bibulbosus: post-larva. Eustomias dubius: post-larva. Eustomias ( N ominostomias ) spp. : post-larvae. Eustomias ( Dinematochirus ) spp.: post-larvae. Eustomias spp.: larvae. Only three of the ten genera of Melanostomiatidae taken by the Ber- muda Expeditions are not represented in the collection by either larvae or post-larvae, namely, Chirostomias, Pachystomias and Echiostoma. We have, however, adolescent specimens of Chirostomias and young transitional adoles- cent specimens of every genus. Thanks to borrowed specimens, we have found under the skin remains of larval pigment spots in the genera Odontostomias and Echiostoma which will aid in the identification of the larvae of these genera in the future. (Text-fig. 2). 76 Zoologica: New York Zoological Society [XXIV :6 Division of Developmental Period into Growth Stages: The growth stages of Stomiatoidea in general and Melanostomiatidae in particular fall easily into our accepted classification of larvae, post-larvae, adoles- cents and adults (see Beebe, 1933.3, p. 7; 1934.1, p. 158 ff. ; Beebe & Crane, 1936, p. 80; 1937, p. 357). In addition, the prelarval stage is more clearly defined in the Stomia- toidea than in the other groups we have previously studied, although no pre-larval melanostomiatids have been found in our collection. Roule & Angel (1930, p. 6) define the stage succinctly as the one immediately fol- lowing hatching, often characterized by temporary traits which rapidly disappear. Their conception of larvae, post-larvae (hemi-larves) , and adol- escents (alevins) , drawn from the work of their predecessors as well as from their own experience, corresponds well with our own, which will be redefined relative to the Stomiatoidea in succeeding pages. It seems that at last the unfortunate confusion in the nomenclature of the growth stages of fishes, and the definition of the boundaries of these stages, is becoming a thing of the past. In spite of the number of stomiatoid young which have been previously described, there has been as yet no effort to characterize the larvae of the group as a whole, and this we propose to do, showing at the same time likenesses and differences of the young of the superfamily Gymnophotodermi, including the family Melanostomiatidae, to those of other stomiatoids. In this study of the young the importance of the recognition of the group Gymnophotodermi has become especially apparent. Diagnostic Characteristics of Stomiatoid Growth Stages : From a study of the published records discussed above and from our own material we find that the known early stages of all Stomiatoidea have in common the following characters: (1) transparency, or at least, translucence when freshly caught; (2) a somewhat compressed but non-leptocephalic body; (3) moderate to extreme slenderness; (4) well-developed finfolds; (5) a lack of pigment except for a few evanescent spots which, varying with the group are of great taxonomic importance; (6) with few exceptions, a large number of myomeres; (7) the early disappearance of the yolk sac, during the pre-larval stage; (8) the occurrence of a period, usually absent in the pre-larva and confined wholly to the larva, when the eyes are elongate and rotated more or less forward. This combination of characters does not seem to occur in the development of any other group of fishes. The diagnostic characteristics of the various growth stages of Stomia- toidea in general and of Gymnophotodermi and Melanostomiatidae in pai’- ticular are as follows: 1. Pre-larva. Stomiatoid Characters: Yolk sac present throughout most of stage; special pre-larval pigment often present; teeth lacking; eye round; pectoral pad present; dorsal, anal and pelvic completely lacking. Gymnophotodermid Characters: (Not known). 2. Larva. Stomiatoid Characters: True yolk sac absent, but intestine not yet en- closed by myomeres ; typical larval pigment spots usually present ; temporary larval teeth usually present ; eye small, elongate, rotated forward ; dorsal and ana! appearing; pelvic rudiment usually appearing toward end of stage. A period of growth. Gymnophotodermid Characters (Present in known larvae of Astrones- thidae, Melanostomiatidae, Malacosteidae and Idiacanthidae) : End of gut prolonged beyond anal origin, sometimes more than a third the total length 1939] Beebe & Crane: Family Melanostomiatidae 77 of the body; pigment in longitudinal series of spots, along full length of body, rarely absent; head strongly inclined; barbel absent; rudiments of photophores appearing at end of stage; temporary, small teeth present in jaws; temporary gill-rakers present or absent; these teeth and gill-rakers both reach their maximum development late in the stage; larval pectoral pad with a continuous frill of undifferentiated rays always well developed, even when pectoral is much reduced or absent in adult; dorsal and anal clearly visible in their normal positions, although some of the rays are often not developed (especially the anterior anal rays in fishes where this fin originates before the dorsal; and the anterior dorsal rays of I diacanthus) ; finfolds usually very high in early part of stage, but dwindling as post-larval stage is reached; stomach absent. The larvae of the family Melanostomiatidae may be distinguished from those of other gymnophotodermids by use of the following key : A. Dorsal fin ending before anal origin Astronesthidae. AA. Dorsal and anal opposed, confined to caudal peduncle. B. Eyes stalked; larval gill-rakers absent Idiacanthidae. BB. Eyes not stalked; larval gill-rakers present. C. Pigment sparse or absent Malacosteidae. CC. Pigment spots almost always present in longitudinal series immediately below dorsal mid-line and sometimes in additional row or rows below the lateral mid-line Melanostomiatidae. It will be observed that the distinction between malacosteid and mel- anostomiatid larvae is not satisfactory; this is because, save for the speci- men of Malacosteus? described by Roule and Angel (1930, pi. I, fig. 11) and for several transitional larvae and early post-larvae in the present col- lection of Photostomias and Aristostomias (an account of which will be published at a future date), the young of the family are unknown. 3. Post-larva. Stomiatoid Characters: Intestine partly enclosed by myomeres; larval pigment spots remaining; general body pigment appearing; light organs appearing; eye once more becoming round, directed laterally; all fins present, but rays not completely differentiated, nor full relative length at- tained; traces of finfolds remaining. A period of metamorphosis, often accom- panied by shrinking in length. Gymnophotodermid Characters: End of gut protruding only as a papilla; larval pigment spots remaining under the developing, lightly pig- mented epidermis, sometimes more distinct and numerous than in larvae; depth greatest toward middle of length instead of at shoulder; barbel a stump; serial photophores well developed; larval teeth absent, but larval gill-rakers remaining throughout all or most of stage; permanent teeth and gill-arch teeth not yet apparent, or very rudimentary; caudal fin relatively larger than in larva or adult; stomach a papilla on wall of intestine. Thanks to the presence of photophores and relatively well-developed fins in this and succeeding stages, melanostomiatid larvae can be easily distinguished from those of related families and referred to their proper genus. 4. Adolescent. Stomiatoid Characters: Intestine completely enclosed in body cavity; larval pigment spots disappear during this stage; pigment, light organs, proportions and internal organs all gradually approaching adult conditions; fin rays fully developed, except, sometimes, in the case of highly specialized n M Text-figure 2. 0 Pigment patterns of young Melanostomiatidae. Each diagram represents a typical series of myomeres from near the middle of the body, the upper and lower boundaries being, respectively, the dorsal mid-line and the upper level of the lateral serial photophores. A, Leptostomias gladiator, larva, standard length 14 mm. ; B, same, post-larva, standard length 42 mm. ; C, Odontostomias micropogon, transitional adolescent, standard length 42 mm. (pattern subdermal) ; D, Melanostomias spilorhynchus, post-larva, standard length 24 mm. ; E, Photonectes parvimanus, larva, standard length 26 mm. ; F, same, post-larva, standard length 26 mm. ; G, Photonectes braueri, late adoles- cent, standard length 25 mm. (pattern subdermal) ; H, Echiostoma tanneri, adolescent, standard length 25 mm. (pattern subdermal) ; I, Flagellostomias boureei, post-larva, standard length 34 mm.; J, Grammatostomias flagellibarba, adolescent, standard length 30 mm. ; K, Bathophilus sp., larva, standard length 7 mm.; L, Bathophilus sp., near longipinnis, larva, standard length 11 mm.; M, Bathophilus metallicus, post-larva, standard length 29 mm. ; N, ? Eustomias sp., larva, standard length 12 mm. ; O, Eustomias sp., larva, standard length 13 mm. C, G and J, from specimens in the Dana collection of the British Museum ; remainder from present collection. 1939] Beebe & Crane: Family Melanostomiatidae 79 fins, the details of which are gradually differentiated during this stage; caudal fin still relatively larger than in adult; finfold absent. The latter part of adolescence, which as in many fishes apparently lasts a long time, is conveniently referred to as “transitional adolescence” ; it is characterized by the fish’s being externally almost or completely adult in appearance, its immaturity being attested only by its size and undeveloped gonads, and by the incomplete development of any or all of the following characters: minute details of light organ and tooth development, skeleton (especially in form of vertebrae) and digestive organs. The entire period of adolescence is marked by growth. Gymnophotodermid Characters: End of gut completely enclosed; sub- dermal larval pigment spots become reduced and vanish; barbel short and roughly formed; permanent teeth on both jaws and gill-arches growing slowly; stomach short, its pigment lacking or incomplete. The transitional adolescent is especially well marked in this group, having completely developed external pigment, proportions and light organs, and the barbel more or less perfectly formed. This stage is sometimes reached when the fish is very small — at 18 mm., for example, in certain species of Photonectes. Immaturity, however, in the form of slightly de- veloped gonads; few maxillary teeth; short, partially pigmented stomach; and often a different barbel, is shown as usual. Sometimes, too, imma- turity is shown in the slow development of the specialization of an organ, such as a luminous or elongate pectoral fin. These transitional adolescents form by far the greater part of our collection, and it would seem that these fish are designed to win through to the safety of an adult external appear- ance as soon as possible. Identification of Melanostomiatid Larvae: We have found the fol- lowing characters to be of the most diagnostic value in the identification of melanostomiatid larvae: 1. Number of Myomeres : These are counted only to the end of the anal fin, since in most cases the few segments occurring behind this point are too ill-defined to count. The myomere count thus gained is be- tween one and four more than the vertebral count of the grown fish, and usually three to six more than the number of ventral serial photo- phores between pectoral origin and caudal base. Unlike the myomeres of some fish, such as nemichthyid eels, the full number found in the adult is present even in young larvae. 2. Myomeres from Nape to Pelvic Rudiment : The pelvic bud can often be detected by careful manipulation of lighting while it is still sub- dermal. The number of myomeres between the nape and this point, is of course, roughly equivalent to the number of vertebrae and to the future number of photophores in the P-V series ; usually the count is two to four more than the photophore count. 3. Myomeres from Pelvic Rudiment to Anal Origin : The count will roughly equal the future number of photophores in the V-A series, minus, of course the number of organs in the series above the anal fin, and to the number of corresponding vertebrae. 4. Pigment : Although, next to myomere counts, the pigment pattern is the most persistent of larval characters, it must always be used only as a partial guide, since almost identical pigment patterns are sometimes found in closely related genera (e.g., Melanostomias, Echiostoma and Photonectes) , while in others the arrangement of spots differs between species of the same genus (e.g., Eustomias) and in still other cases certain chromatophore rows are found in post-larvae which are absent in larvae of the same species (e.g., Leptostomias) . 5. Larval Teeth: The larval teeth, since their appearance, growth and disappearance are all confined to this single growth stage, are not a good 80 Zoologica: New York Zoological Society [XXIV :6 identification character. We can see apparent differences between genera, but the material at present is far too scanty to make up a table of frequencies. 6. Larval Gill-rakers: Their number and form are a useful secondary check in larvae of doubtful position. 7. Dorsal and Anal Fin Counts : Except in the youngest larvae, approxi- mate counts can be made, which either fall within, or slightly below, the generic or specific limits. 8. Anal Origin with Reference to that of the Dorsal: This chai’acter is very useful; it must always be remembered in using it that in larvae the anal origin usually is slightly behind its position in the adult; that is, in genera where the anal originates well in front of the dorsal, it stai'ts only barely in front in the larva; similarly, when the adult origin is immediately below the dorsal, in the larva it is usually under the second to fourth dorsal ray. 9. Length: Metamorphosis begins at slightly different lengths in dif- ferent genera. In most melanostomiatids in which the young stages are known, the larval period ends when the fish is around 20 mm. long, while post-larvae and adolescents (excluding transitional adolescents) measure between 20 and 30 mm. In Flag ellostomias , Leptostomias and Eustomias, however, post-larvae and adolescents measure between 30 and 50 mm. or more. Absolute specific identification always, of course, waits upon the study of intermediate forms, which retain characteristic larval pigmentation subdermally, along with recognizable adult characters. With the single exception of the youngest larva referred to Eustomias, we have series com- plete enough so that we are confident of generic identification in every case, and of specific in most of these. Because of the evanescence of a number of characters — the change in pigment spots at different phases, of relative gut lengths and fin develop- ment, and the relatively few genera of which true larvae are known — we cannot yet give an adequate key. The accompanying table, however, used in conjunction with Text-fig. 2 and the full-length figures of larvae scattered throughout the text, should form a basis for future work. D. Form and Development of Separate Characters. 1. Color and Luminescence: Up to the present time the colors of the light organs in living and freshly dead Melanostomiatidae have been almost completely unknown. Lowe in 1843 (p. 88), in describing Echiostoma bar- batum, the first known melanostomiatid, remarked that the postorbital was rose-colored. Murray, writing pioneer notes on the Challenger Expedition, observed that in Opostomias “the end of the barbel, which was thickened, was flesh colour with a rose tint; there was also a rose tint on the dorsal and anal fins. The rest of the animal was of a dark colour with a perceptible slate- coloured tint. The phosphorescent spots along the belly and lateral line were red, as was also that below the eye.” (Tizard, etc., 1885, p. 412). In regard to Pachystomias microdon, which was living, he wrote (ibid., p. 521), “It had one club-shaped spot of a rose colour directly below the eye, and another, about half the size, directly in front of this, of the same colour. . . . The two rows of probably phosphorescent dots along the body were red sur- rounded by a circle of pale violet.” Only one of the Monaco melanostomiatids seems to have been described while still fairly fresh. Zugmayer (1911.2, p. 78) reported in his descrip- tion of Trichostomias vaillanti that the fish had the organs of the lateral series yellowish-white, and the suborbital (postorbital) pale red. The colored plates of the family in both the Monaco reports and in Brauer’s Valdivia Tief-see Fische (1906) were obviously not made from field sketches, since in none do the organs have the brilliant colors which we have found, by repeated observations in Bermuda, to be characteristic of these fishes. 1939] Beebe & Crane: Family Melanostomiatidae 81 Table I. Characteristics of larvae and post-larvae in the Bermuda Collection2. 1 Myomeres (nape to anal end) Myomeres (nape to pelvic) Myomeres (pelvic to anal origin) Larval teeth, premaxillary Larval teeth, maxillary Larval teeth, each half mandible Larval gill-rakers (Numbers refer to arches) Length of larvae (mm.) Length of post-larvae (mm.) Leptostomias gladiator 75-78 43-44 16-18 7 18 12 Long on 1st, 2nd, 3rd. Mounds on 4th, 5th. 12-30 38-45 M elanostomias spilorhynchus 49-52 29-30 10-12 7 15 7 Long on 1st, 2nd, 3rd. Rudiments on 4th, 5th. 17 21-32 M elanostomias biseriatus 55-56 33-34 10-12 - — — Long on 1st, 2nd, 3rd. Rudiments on 4th, 5th. — 23,25 Photonedes pammanus 64-67 39-41 12-13 5 10 7 Long on 1st, 2nd, 3rd- Mounds on 4th, 5th. 14, 26 25 Flagellostomias boureei 67-68 32-33 14-16 4 14 5 Long on 1st, 2nd, 3rd. Mounds on 4th, 5th. 20-21 34, 39 Grammatoslomias flagellibarba 55 — — — — — Short on 1st, 2nd. Mounds on 3rd. Absent on 4th, 5th. — 29 Bathophilus sp. 45-46 — - — 15 10 None. 7 — Bathophilus, near longipinnis 42-44 19-21 11-12 5 12 6-7 Short on 1st, 2nd, 3rd. Mounds on 4th, 5th. 11,12 — Bathophilus metallicus 45 19 17 — — — Moderate on 1st, 2nd. Absent on 3rd, 4th, 5th. — 25, 29 t Eustomias sp. 55 — — 6 18-19 9 Short on 1st, 2nd, 3rd. Mounds on 4th. Absent on 5th. 12 — Eustomias sp. 77 — — 8 10 10 Rudimentary mounds on all 5. 13, 15J4 — Eustomias bibulbosus ca. 72 37 14 — — Moderate on 1st. Short on 2nd. Mounds on 3rd, 4th. Absent on 5th. — 42,52 Eustomias sp. (Dinematochirus) 78 32 14 — — — Long on 1st. Short on 2nd, 3rd. Mounds on 4th. Absent on 5th. — 43 2 Fins and photophore counts of post-larvae, typical of species: diagrams of pigmentation in Text-fig. 2; typical teeth and gill-rakers in Text-figs. 5, 7 ; full length drawings in Text-figs. 21, 30, 42, 47, 57, 58, 61, 66, 67, 68. 82 Zoologica: Neiv York Zoological Society [XXIV :6 The first notes on the actual luminescence of a living melanostomiatid were taken on the Arcturus by Beebe in 1926. Of Echiostoma tanneri (then identified as E. barbatum) the following observations were made: “It was alive and stayed so for several hours while we got movies. The most notice- able character of this otherwise brownish-black fish was a wedge-or-pear- shaped light organ of rich rose color below the eye. In the dark this gave forth a warm reddish glow. The lateral light organs were all tinged with rose.” (Beebe, 1926, p. 422). Finally, Borodin (1931, pp. 65 and 67) noted that the postorbital was red in Echiostoma, rose in Photonectes, while Bolin (1939, p. 41) describes this same organ as “pale luminous green” in Tactostoma. This handful of observations apparently includes recently preserved as well as actually living fish, and all told represents only six out of the sixteen valid genera in the family. Except for the single Arcturus observation, the actual luminescence of the fish has never before been recorded. The Bermuda observations which follow, while naturally incomplete, at least lay a sound foundation for future research, which will emphasize the functions of these organs. Our notes give hints of the relationship between the colors of the photophores in daylight and their luminescence in the dark, as well as of their respective uses, and on their appearance under natural conditions, as observed from the Bathysphere. The field notes and color sketches were all made while the fish were actually alive or else when they were freshly dead, still in water, and within two hours of the nets’ having reached the surface. We have made these observations upon all 10 genera taken by the Expeditions and on 29 of the 32 species, or about one-third of the known forms. From one to more than a dozen individuals of each species were painted and described. In the case of the more common forms, such as Melanostomias spilorhynchus , the colors of the barbel and postorbital were so consistent and unvarying that after having made a number of sketches and notes, we simply checked the various organs mentally in order to devote more time to rarer forms taken in the same nets. The colors usually faded with extreme rapidity, the organs often being almost or quite white after only a few minutes at the surface, so that the taking of notes was useless by the time they reached the laboratory. Occa- sionally, however, the serial organs retained traces of violet even after two years in alcohol, and the postorbitals of Echiostoma usually remained pink for several months. Individuals of five of the species, Chirostomias pliopterus, Pachystomias atlanticus, Echiostoma tanneri, Photonectes margarita and Eustomias bibul- bosus, were still alive when they reached the laboratory. In the case of Echiostoma, five specimens of both sexes came up alive, and two of them lived overnight. Thanks to these living examples, we could compare the colors in daylight with those soon after death. There was found to be no difference at all except, sometimes, in relative brilliancy, so that we may consider our notes on fresh fish very accurate. No dead fish, no matter how fresh, showed any luminescence in the dark-room. Since the functioning of the organs in the dark-room was exactly similar on a small scale to their functioning when seen from the Bathysphere, objections that our observa- tions were made upon dying specimens are not as valid as they would be in the case of freshly caught shallow water fish, where the superficial colora- tion is known often to go through unusual or abnormal phases. In the dark-room an ultra-violet lamp was sometimes used to help in the observation of luminescent areas. A table, summarizing the observations on each species, will be found at the end of this section ; detailed color descriptions are included under the headings of the various genera and species further on. We will now sum- 1939] Beebe & Crane: Family Melanostomiatidae 83 marize what we have learned of the general body color, and of the colors, luminescence and uses of the barbel and light organs. General Color : When alive or recently dead, these fishes appear velvety, jet black in the shade; in direct sunlight, however, the color is dark brown. The presence and arrangement of the evanescent pigment spots of the trans- lucent larvae has already been discussed above. The pigment of the adult begins to appear at the end of the post-larval or beginning of the adolescent stage; by early transitional adolescence it is nearly or quite as well developed as in adults; advanced transitional adolescents are invariably as deeply pigmented as mature fish. Pigmented skin, forming on top pf the larval spots, appears first along the sides, last around the snout and end of the caudal peduncle. Iridescence has been observed in the following fishes: Chirostomias pliopterus, green bronze on shoulder; Bathophilus longipinnis and B. metal- licus, head and body iridescent in both males and females. In the latter species this iridescence is often completely lacking; it is obviously very easily damaged, however, and we consider it a normal, rather than a variable characteristic. It appears in transitional adolescence. The fins of melanostomiatids are usually translucent, or whitish, because of luminous mucous (of which more will be said later). Often, however, the skin of the body, as in certain species of Grammatostomias and Photonectes, extends almost to the tips of the dorsal and anal, and often all of the fins are tinged with pink because of the coursing blood. Barbel : This highly specialized and variable structure cannot be listed definitely as either a luminous or a tactile organ; in all probability it is sometimes one, sometimes the other, sometimes both, and perhaps always connected with one or more senses of which we have no knowledge whatever. The Bathysphere observations on this subject were necessarily disap- pointing. “In the great majority of cases, it was quite impossible to make accurate generic identifications. By the time I had satisfied myself that I was looking at a member of this family, the Bathysphere or fish would move. So I invariably lost the chance of seeing the barbel and its light. In Bathy- sphaera I thought, on the occasion of their first passing, that a parti-colored jelly or small fish was swimming beneath. Only on their return did I sud- denly realize that the bobbing red and blue lights terminated a dangling, invisible barbel thread. One other time I thought I saw a long strand of tissue studded with minute lights, but I am not certain, and so far as identification by barbels is concerned, my dives were quite ineffective. This may indicate that barbels in general subserve a tactile rather than a luminescent function.” (Beebe, 1934.2, Appendix G, p. 312.) The old suggestion that barbels in this family sometimes serve as luminous lures is more plausible than ever after these Bathysphere observa- tions. In other cases, however, it is possible that luminescence of the organ is a purely secondary matter, or an accidental byproduct, its primary func- tion being sensory. We have also found that, at least in the genus Eustomias, barbels differ in both form and color in males and females of the same species (see p. 211). In regard to sexual differences in barbel color, Eustomias bigelowi had the bulb and bulblets bright yellow in the male and brilliant bluish-green in the female. Unfortunately, these fish were not taken alive, so that comparison of luminescence in the two sexes could not be made. As might be expected, the genera Pachystomias, Grammatostomias and Bathophilus, which have in common long, filamentous, bulbless barbels prac- tically lacking in pigment, save for a few rudimentary photophores, showed no color tints and gave out no luminescence. Genera with more elaborate barbels, on the other hand, usually had the bulbs and bulblets brightly colored, often combining two or more contrasting hues. The structures were always more or less translucent, and the colors consequently wonderfully clear. They ranged from gleaming white and silver ( Chirostomias ) through 84 Zoologica: New York Zoological Society [XXIV :6 pinks, yellows, blues, greens and lavenders: only bright red was missing from the spectrum. In some, such as Echiostoma, the daylight tints were rather delicate, but in Eustomias and Photonectes they were often blazingly vivid, even in dead specimens. In the two last genera, the only ones with non-filamentous barbels on which we have observations on a number of species, we find that many colors are found in the same genus, although there is practically no variation (except sexual, as noted) within species. One of the most interesting results of the dark-room studies is that barbel luminescence does not, in the two species observed, correspond to the color of the barbel in daylight. In Chirostomias the silver-white barbel bulb gave off a steady pink glow anteriorly and a white one posteriorly. In a male Eustomias bibulbosus the bulbs were bright pink, but gave a distinctly green light in three brilliant flashes. In Echiostoma, the barbel, though highly developed and colored in the young, was never observed to be lumin- ous; similarly, there was no glow from that of Photonectes margarita. In Bathysphaera intacta, seen only from the Bathysphere, the proximal bulb glowed rosy red, the distal blue. Postorbital Light Organ : The color of the postorbital in freshly caught specimens varies greatly, and includes all the colors of the spectrum as well as white and silver-white. The luminescence of these lights is known only for Echiostoma, the adults giving a rosy glow and a blue or white flash, while the young also gave a rosy glow (the anterior part of the organ being pink) but a distinctly green-white flash (the posterior portion being green). Silver or silver-white with opalescent reflections seems to be a generic char- acter in the postorbital of Eustomias. Bathophilus, on the other hand, varies within the genus. Other multi-specific genera have not yet been suffi- ciently observed to draw any conclusions. The postorbital is definitely under the control of the fish, can be rolled down out of sight, and made to glow steadily or emit sharp flashes of a different color, at least in Echiostoma. In the new Tactostoma Bolin, 1939, it apparently rotates forward, or forward and downward, instead of the usual downward. The following summary has been made of observations on this organ from the Bathysphere: “The cheek lights seemed under con- trol, and were seen occasionally to blink. Their color, whenever a definite tint could be assigned, was yellow or red. Every time they were rolled up into sight, these organs illumined the fish’s eye and most of its head. Why the creature is not momentarily blinded by the light is a question which has always puzzled me.” (Beebe, 1934.2, Appendix G, p. 313.) Serial Photophores: Throughout the superfamily Gymnophotodermi (Astronesthidae, Melanostomiatidae, Malacosteidae and Idiacanthidae) the color of the serial photophores in fresh specimens in all the observed genera save one is invariably purple or violet, sometimes verging on scarlet. The known exception is Bathophilus in which the organs in the two species observed were always golden yellow. This apparent color may have been due however to the inconspicuousness of the organs themselves in this genus and the relatively large size of the tinselly gilt reflectors. The latter char- acters are visible only in perfectly fresh specimens and consist apparently of small crescents or circles of specialized iridescent skin, always giving off bright golden reflections, which cap or completely surround the organs of the lateral series, the ventral or both. Their function is doubtless to magnify the light of the photophores, exactly as do tin reflectors placed behind the individual lights on a Christmas tree. Due to their evanescence, we have not been able to determine whether the crescentic caps above the organs in some species are the rule, or whether only the upper part of a complete gilt circle has survived capture. The luminescence of serial photophores has been observed in the dark- room only in Echiostoma, in which case it was rosy to scarlet. They glowed steadily for a time, but were apparently under the control of the fish. From 1939] Beebe & Crane: Family Melanostomiatidae 85 the Bathysphere the serial photophores appeared as follows : “The two rows of lateral serial organs were usually distinct, though not brilliant, and as far as I could tell glowed steadily. I cannot generalize on their tint, except that they often seemed faintly yellowish. It is interesting to note that on freshly caught dead specimens these organs are always clear violet or purple.” (Beebe, 1934.2, Appendix G, p. 312.) In fresh specimens of Malacosteus and Bathophilus brevis, the only stomiatoids in which regular serial organs are lacking, cei'tain of the num- erous tiny, non-serial photophores were distinctly larger, brighter, and more violet than their fellows, and in roughly linear formation in the region where serial organs are usually found. The differences were not, however, clear enough for counting of these larger organs. Non-serial Photophores: Their color, in fresh specimens, ranges from red or pink, through violet to bluish, differing in tint considerably from that of the clear violet photophores, and fading much more quickly than the serial organs. From the Bathysphere their importance was apparent: “One unexpected observation was the brightness of the tiny-non-serial organs scattered in large numbers over the heads and bodies of these fish. In newly caught specimens these are very inconspicuous in comparison with the much larger serial organs, and usually show no vestige of color. Yet a number of times in the Bathysphere I noted Melanostomiatids with these tiny pinpricks of light glowing with considerable brilliancy.” (Beebe, 1934.2, Appendix G, p. 312). In the laboratory dark-room the non-serial lights of Echiostoma glowed with a rosy light. Other Luminous Areas: The luminous tissue in which pectoral fins are often more or less imbedded is usually creamy white in fresh specimens. The highly specialized luminous line or loop on the side of Grammatostomias is bluish to bright metallic green-violet in fresh specimens. Yellow, purple, pink and green snout and jaw spots have been observed in freshly dead Melanostomias, Photonectes and Bathophilus. The whitish longitudinal body bands of Echiostoma gave off a bluish-white luminescence under ultra- violet light. Some genera, such as Echiostoma and Melanostomias, also certainly have the bases of the teeth luminous, the material being de- posited in the soft portion which permits bending of the fang. Also, luminous granules are frequently present on the dorsal and anal fins, being sometimes more easily seen in the adolescent than in the adult. We have never observed luminescence in this region, however, in the dark-room. Bathysphere observations show that a luminous mucous, of which we see few traces in captured specimens, may be a general family attribute. “Another point I cannot explain is how I could see outline after outline of the fish when they were in absolutely black water, while their lights had very little reflecting power. Perhaps there was a general coating of lumi- nous mucous, as trawled specimens frequently exude a whitish slime, or a loose epidermal membrane.” (Beebe, 1934.2, Appendix G, p. 313). The table on pp. 86-87 gives in summarized form all that we know at present of the color and luminescence of melanostomiatid light organs. In all cases, unless prefixed by the word “luminescence,” colors refer to those found on freshly dead, apparently unfaded, specimens. Detailed color descriptions will be found in the discussions of the individual species. 2. Body Form and Proportions: With the exception of the highly specialized Bathophilus brevis (depth in length ca. 2.5), all melanostomiatids are moderately to excessively elongate in form ; some species of Leptostomias, for example, are of anguilliform slenderness, the depth being contained in the length up to 17 times. Adult melanostomiatids are deepest at the shoulder except in the more specialized Photonectes, in which the fish is deepest toward mid-body, as in juvenile forms. The head is usually small, there being a distinct correlation between large heads and deep bodies; the extremes of head lengths range from about 2.5 (for B. brevis) to 11 Table II. Summary of melanostomiatid color notes taken on Bermuda Oceanographic Expeditions. 86 Zoologica: New York Zoological Society [XXIV :6 o- 3 0 CO O 0 3d a d is* . 0 3d ^ CO s ^ a.s --5 c3 a § 8? I S £ o CO ^ cu _i_j 02 CP 3 ^ 02 -O 3 o m m y ■ M r» if - .35 “-o ■§ m L W W ® D o > c a ‘m S a> §.t2 -) -a 3d ^ d . ^ *-g e= £ © ® d=d 3 CO pq >» © o o >» pQ *3 1 * 02 ^ •II 02 Oh a> S- CO bJD-*-3 a* £ ^ =3 3 m ^ . 3d ^ g III* C3 i-. r- (D CO © 03 .2 j© ©-^P a> o £ © pq g . ^ 3 © 3d CD ^ rt<3 fl o a _q 02 ■g a cO cO Si Oh^ ja2 co p3 3 C l.s • -a ® o m3 >y ~g t>> [©3 > d x> m s fe m -a Is 3 3 O PQ >» cO © Oh © - &> 03 > O © 03 33 0 03 § SP 3d 03 33 Oh 3 Oh Pq 03 .5 3d 3d •s . * * Cd o3 c- 03 d o 2 £ ®i2 bfi.-g 3d ..-a 0 .. a I OJ fl bD o febP-i g? £ o m Co s "§ . >H g 'b ,g-> J> - *M © m ^ a> th S >H O O S 3 §~ Oh e. Hi •g S' 1939] Beebe & Crane: Family Melanostomiatidae 87 p-t Q«.-^ SJ, b 0 w © ,£? 53 § 3 3 © cx.12 o jz £ if GO ^ r§ ^ . a © > 'g- 3 3 M o- §-g b£)_g ° £ 13 >» ■r* t_ CO -^5 a 's ° 3 •m 3 3 £3 O O F4 £ Ph 3.2 . 3 >> p. -Q PQ jsm bO g • ^ a* •s-e c3 © rO 3 ° Is .& a £3 2 3 S O ® ja M ®* •■S "» & -O . PQ 15 b£> , ^ ! PQ PQ o ’b ^ fl PQ ^ o 2 Xi 3 PQ K- CUU ^D>h -f-> ■§ © > pqpq *b q»> fe) J© 13 PQ bi) « i 3.2 Of ® °* >» •g o 3 fe 5 bfi c .S ?3 a o S.S g^ Ci 03 Q5 »3 03 ■g-s 3 a §o 63 63 63 63 63 Table II. Summary of melanostomiatid color notes taken on Bermuda Oceanographic Expeditions. Genus and Species Postorbital Barbel Serial Organs Other Luminous Areas Chirostomias pliopterus (including 9 in life) d\ White; silver rim. 9. Bulb white and silver; greenish-yellow in yg. Lumines- cence pink anteriorly, white pos- teriorly. Violet; gold frames. Green bronze iridescence on shoulder. Pachystomias atlanticus (in life) Bright red. Stem translucent white. Purple; gold caps. Antorbital yellow-green. Leptostomias bermudensis (No o' taken). 9 . Bulb base lilac; bulb yellow'. Lateral, purple; ventral, maroon. Both with gold caps. Non-serial organs purple; pale blue spots on abdomen. L. gladiator (No cf taken). ? Violet. Melanostomias spilorhynchus d". Pink to purple. 9. Bulb greenish-yellow; flanges and bulblets pink or purple. Violet; gold caps. Non-serial organs, violet; base of teeth, pale blue; snout lights pink. Eehiostoma tanneri (in life) Yg. Pink in front; green behind; rosy glow; green-white flash. cP and 9 . Pink in front, white behind; rosy glow; blue or white flash. Yg. Bulbs and filaments green and purple, rufous core. Adult. Washes of pink. No lumi- nescence in either sex. Yg. Violet. Adult. Scarlet. Luminescence of both, rosy to scarlet. 3 whitish, lateral bands. Non-serial organs, purple to scarlet. Fins luminous. Photonecles dinema Yg. Silver-white with lower front corner, purple. Bulb pinkish-purple at both ends, blue in middle, filaments pale yellow. Violet; gold caps. Snout light purple. P. leucospilus Yg. Blue or silver-white with lower front corner, yellow. Bulb violet to violet-blue; laven- der basally. Violet; gold frames. Non-serial organs violet. P. mirabilis Yg. Golden-yellow. Bulb silver; bulblets golden- yellow. Violet; gold frames. Snout light golden-yellow. P. parvimanus Yg. White. Bulb green, center peacock blue; appendage yellow-green. Violet; gold frames. P. bifilifa | 1. Bulb lavender near tip. ? P. margarita (including o' in life) 9. Rosy, o'. Yellow Yg. Purple or magenta. Yg. Bulb violet with pink spot posteriorly. 9 - Bulb purple. Violet to purple. Shoulder spots pale blue. Flagelloslomias loured o'. Silvery white. o'. Bulb greenish-yellow. Purple; gold caps. Pectoral tips greenish-yellow. Grammatostomias dentalus d\ Silvery. Stem translucent white. Purple; gold caps. Non-serial organs pink; lumi- nous hook bluish-white. G. flagellibarba o'. Bright yellow. Stem translucent white. Purple. Luminous loop blue-green; pec- toral, milk-white. Bathophilus brevis Yg. Pale green with silver rim. Stem translucent white. — Non-serial organs bluish to violet. B. aUipinnis o'. Anterior part, deep red. Stem translucent white. 7 B. longipinnis o'. Bright yellow. 9 . Pinkish-silver with silver rim. Stem translucent white. Golden. Non-serial organs pink. B. melallicus 9 and Yg. Pale greenish to lemon- yellow. 9. Stem translucent white with pink photophores near tip of barbel. Golden. Non-serial organs lavender. Maxillary patch purple. Eustomias bibulbosus (in life) o'. Silver. o'. 2nd bulb pink giving off green light. Purple; gold frames. E. dubius 9. Silver. 9. Peacock blue and turquoise. Purple. E. obscurus 9. Silver. 9. Bulb violet basally, green distally. ? E. bigelowi c?. Opalescent silver. 9. Bulbs bluish-green, cf . Bright yellow. Violet; gold frames. o'. Antorbital violet-blue. E. fissibarbis 9 . Greenish silver. ? Lavender; gold frames. E. siheseens 9 . Opalescent white. 9 Bulbs and branches yellow and light green. Violet; gold frames. E. sehmidti o’. Silver-white, golden rim. o'. Bulbs pink, branches ochre. Purple. Non-serial organs pale, bluish- violet. Zoologies New York Zoological Society [XXIV:6 I 1939] Beebe & Crane: Family Melanostomiatiiae 88 Zoologica: Nezv York Zoological Society [XXIV :6 ( Leptostomias ) in the length, the average being about 6 to 8. In Pachy- stomias the head is very broad as well as long. In all cases the eye is small, compared with those of fish in general, although as an organ it is well developed. The snout is noticeably short, excessively so in Melano- stomias and related genera. In Eustomias and Pareustomias, and to a much lesser degree in Flagellostomias, the snout is protractile; traces of this tendency are also present in Grammatostomias. The thrust forward is ac- complished by the dislocation of the upper jaw bones, which project the fleshy tip of the snout forward beyond the ethmoid region (see section on Osteology). The body proportions of larval melanostomiatids are not greatly dif- ferent from those of adults, if the finfolds and pendulous guts are disre- garded. However, the trunk is usually slimmer (when the depth is measured exclusive of the coelomic organs), and the head and snout always longer throughout the larval and post-larval stages, the jaw angle being under or well in front of the eye instead of behind it. The deepest part of the body is toward the middle of the length, instead of at the shoulder. With the material at hand, we have not been able to deduce any satis- factory numerical ranges for these proportions in the various larvae which would be of help in identification. It may be remarked, however, that in Eustomias1 the snout (and with it the head) is relatively much longer than in other genera; the snout itself is more than half the length of the head, and of a flattened, duck-bill-like shape, very much as in ldiacanthus. The elongate, forwardly rotated eye, which is confined to the larval stage of stomiatoids alone, is of especial interest. Apparently no larvae in other groups possess just this characteristic. The closest are certain young alepocephalids and myctophids, and in these forms the eyes are merely elliptical, and not at all turned toward the front. The vast majority of fishes pass through all phases of development with the round eye, often very large, which they have upon hatching. It seems likely that this tran- sient, elongate eye of the stomiatoids is a phylogenetic character, reminis- cent of ancestors which, like Ichthyococcus today — one of the least spe- cialized of existing stomiatoids — had semi-telescopic eyes. Further evolu- tion in that direction led to Argyropelecus. 3. Barbel: This highly specialized organ is the most variable in the entire family. It ranges from simple forms similar to that of the astrones- thids, with moderate stems and small, simple bulbs, to all extremes: to the attenuated, bulbless barbel of Grammatostomias flagellibarba, seven times as long as the fish; to the complex, tree-like organ characteristic of the eustomiad subgenus Dinematochirus ; and, in the opposite direction, to the degenerate barbels of adult Echiostoma and Photonectes, and the almost atrophied organ of Tactostoma, which consists only of a minute black stem with the slightest of distal swellings. Variability is great not only between genera, but, sometimes, between the species of one genus, as in Eustomias and Photonectes. In species hav- ing more or less complicated barbels, as in Flagellostomias boureei, Lepto- stomias gladiator and many forms of Eustomias, individual variation is the rule. Lack of adequate material to show specific ranges in this character has been another common cause for the erection of invalid species. Never- theless, it is often true that closely related but valid species differ from each other solely in the form of the barbel; as such, this organ is the most useful single taxonomic character in the entire family. In the genus Eustomias we have found a distinct sexual dimorphism in the shape of the barbel in most of the species which we have been able to study. In the subgenus N ominostomias the distal filaments tend to be branched in the females, but unbranched in the males. In the subgenus Dinematochirus the median posterior branch is short and tipped with a prominent bulb in the females, while in the males it is longer, with the 1939] Beebe & Crane: Family Melanostomiatidae 89 bulb small or absent; the main barbel bulb is smaller in the female than in the male. The probable functions of barbels have already been discussed under the heading “Color and Luminescence” (p. 83). Completely absent in the larva, a stump in the post-larva, and roughly formed in the adolescent, the barbel is often not fully formed until the completely adult stage is reached. This has been found to be especially true of the genera Eustomias, Echiostoma and Photonectes. Because it us- ually appears fully pigmented and apparently completely defined in transi- tional adolescence, a number of species in various genera (notably the same ones mentioned above) have been described on the basis of this character, whereas it merely represents a different stage in the development of a pre- viously known species. It is interesting that in Odonotostomias, Echiostoma and the majority, if not all, species of Photonectes, the barbel bulb is rela- tively larger in young specimens than in older; in Echiostoma, at least, the entire barbel is apparently actually shortened in the adult. In general, the relative length of the barbel increases with that of the fish until the proportion characteristic of the species is attained, sometime during transitional adolescence. In a few species, however, the barbel is longer relative to the length in the transitional adolescent than in the adult; that is, it grows faster than the rest of the body. 4. Light Organs: Antorbital : This small photophore, situated at the lower front corner of the eye, is of phylogenetic interest, since it is well developed and functional in all the lower stomiatoids, while in adult melanostomiatids it is almost or completely atrophied. In young melano- stomiatids, however, it appears with the other photophores in the post-larval stage, quickly develops at least one luminous center and sometimes two, and for a while may even be functional. In adolescence, however, it ceases to grow, and usually atrophies during transitional adolescence. In Pachy- stomias however, it is well developed and brightly colored even in adult fish. Also, a transitional adolescent Eustomias bigelowi, had an antorbital with a luminous center tinted violet-blue. Subdermal traces of the organ are sometimes found in fully adult melanostomiatids. Postorbital Organ: The postorbital, which is small or absent in gono- stomids, sternoptychids, stomiatids and Chauliodus, is well or highly devel- oped in male melanostomiatids and sometimes in females. Unlike the barbel, this organ does not vary greatly in structure throughout the family, although there are large differences in its size and color, and in most genera it is reduced or entirely atrophied in the female. In general form it is a gigantic photophore, apparently always under the control of the fish, which nevertheless can also be rotated downward out of sight, so that the luminous face is replaced by the pigmented inner surface of the organ. The histology of the postorbital has been described by Brauer (1908, p. 87). In genera having very large postorbitals, such as Echiostoma, Photonectes and Gram- matostomias, an overhanging “eye-brow” of skin protects the eye from receiving the full glare of the light; nevertheless, it is hard to understand how the fish can see when this organ is fully illumined. Sexual dimorphism in the form of reduced or absent postorbital organs in females has previously been noted in the case of the malacosteid Photo- stomias (Regan & Trewavas, 1930, p. 134), of the melanostomiatid Odon- tostomias (Norman, 1930, p. 309) and of Idiacanthus (Beebe, 1934.1). We have found that it is present in varying degrees in the majority of genera of Melanostomiatidae. We have not, however, been able to check it in every genus, nor in all the species, due both to a lack of material, since many species are known only by immature examples which cannot be sexed, and to the impossibility of our examining many type specimens which are deposited in European museums. However, the evidence so far obtained gives the following results: 90 Zoologica: Netv York Zoological Society [XXIV:6 Genus No. Species Examined Male Light Female Light Chirostomias 1 out of 1 Moderate Atrophied Trigonolampa 1 out of 1 Large ? Pachystomias 1 yg. out of 2 ? ? Thysanactis 1 out of 1 Moderate ? Leptostomias 3 out of ca. 9 Probably moderate Atrophied Odontostomias 1 out of 2 Moderate Atrophied Melanostomias 3 out of ca. 12 Large Large Echiostoma 1 out of 1 Large Large Photonectes 3 out of ca. 15 Large Large Tactostoma 1 (yg.) out of 1 ? ? Opostomias ? ? Flagellostomias 1 out of 1 Moderate Atrophied Grammatostomias 2 out of 2 Large Almost atrophied Bathophilus 4 out of 16 Moderate Small but functional Eustomias 7 out of <50 Moderate to large Small or atrophied Pareustomias ? ? We surmise that Melanostomias, Echiostoma, Photonectes and possibly the closely related Tactostoma, will prove to be the only genera in which dimorphism in this organ is absent. A preliminary survey of related families shows that similar postorbital sexual dimorphism is absent in at least two genera of Astronesthidae ( Astronesthes and Neonesthes), present or absent in Malacosteidae, and present in Idiacanthidae. The atrophying of the organ in the females is most interesting to trace through the developmental stages. The organ is always apparent, and even, possibly, functional in adolescence. In transitional adolescence, how- ever, it ceases to grow with the fish, the protecting Transparent skin be- comes gradually pigmented, and the organ remains relaxed, rolled down almost out of sight. Finally, in adults in which it is completely atrophied, such as Chirostomias, the covering skin is indistinguishable from the sur- roundng epidermis, and dissection shows no trace of the organ except a cavity above the maxillary. In the male, however, its growth like that of the serial organs is steady and fast. Colored (and presumably functional) postorbitals have been noted in early adolescence. Serial Photophores : These organs are well developed and are ranged in regular rows except in Bathophilus, Tactostoma and some species of Photo- nectes, in which genera they are small and sunken (atrophied in B. brevis ) ; notably in B. irregularis and P. margarita, those of the lateral series are placed high on the sides, out of alignment. Also, in Pachystomias, Pareus- tomias and Eustomias obscurus the photophores of the upper series are grouped, a similar condition being found in certain gonostomids, in Aris- tostomias and in Pareustomias. The serial photophores appear almost simultaneously, with the lateral series a little in advance of the ventral, and the last A-C photophores a little behind the rest of the series, especially in the development of luminous cen- ters. Unpigmented anlagen of the photophores are sometimes visible in advanced larvae. Pigmented frames appear in early post-larvae, with luminous centers following almost at once. Fully formed, violet photophores, presumably capable of giving off light, have been noted in advanced post- larvae. Non-serial Photophores : The tiny organs scattered over the head and 1939] Beebe & Crane: Family Melanostomiatidae 91 body of melanostomiatids vary greatly in number and prominence in the different genera. Sometimes they are located in definite patterns and some- times no particular arrangement is discernible. Usually there are about three principal sizes of non-serial lights; the smallest of these are possibly not true photophores at all, but rather pores exuding luminous mucous. The number of lights on a good-sized melanostomiatid, such as a 12-inch Echios- toma, may total several thousand. They are sometimes found on the posterior side of the barbel and on the pectoral fins. Non-serial organs do not usually appear until adolescence, and are fully developed only in late transitional adolescence. Other Luminous Areas'. In most melanostomiatids, luminous tissue is not confined to barbels and typical photophores. Chirostomias, Trigonolampa, Flagellostomias and Grammatostomias all have large or small amounts of luminous tissue on the pectoral fins. Many species, chiefly of the genera Melanostomias, Echiostoma, Photonectes, Grammatostomias and Batho- philus, have characteristic spots, patches or bands of luminous tissue on the head and body. These areas usually appear in adolescence, but sometimes, as in the body bands of Echiostoma, they are apparent only in adults. It is likely that the exudation of luminous mucous is a family character. 5. Teeth and Gill-Teeth: The dentition of the jaws is exceptionally well developed in the Melanostomiatidae. It and the Idiacanthidae are the only families in which depressible teeth are developed. Except in Chirosto- mias, Trigonolampa and Pachystomias, at least one or two, and sometimes all of the teeth in premaxillaries and mandibles are depressible, an attribute which is undoubtedly of great aid in swallowing large fish. Usually the general proportion and arrangement of fixed and depressible teeth serve as useful generic distinctions, the more specialized genera having the most depressible teeth; in Eustomias, however, a highly specialized genus, the species run the gamut from almost all fixed to all depressible; in this genus too, the number and size of the teeth are very variable. In Opostomias and Flagellostomias the largest fang in the mandible is not depressible and fits, respectively, into a hole in the premaxillary or into a groove in the same bone. In Echiostoma, Melanostomias and Photonectes the teeth are strongly barbed or bicuspid ; the same character is found to a lesser extent in Flagel- lostomias and Grammatostomias. The unossified central portions of the tooth bases, which enable them to bend, are often filled with luminous matter (see Beebe, 1934, p. 199, fig. 67). The jaw teeth are in single rows except in Echiostoma and Tactostoma; sometimes, however, especially in Chirostomias, the row is irregular, outer teeth alternating with inner ones; in genera with some teeth depressible, the fixed teeth are usually external to the depressible ones. Vomerine teeth are altogether absent only in Gramma- tostomias, Bathophilus, Tactostoma, and most species of Eustomias. Pala- tine and basibranchial teeth vary greatly in development throughout the family. Replacement teeth frequently develop before the loss of the tooth to be superseded. The gill-teeth have turned out unexpectedly to be a character of great importance. They are paired or in groups of threes or fours in about half the genera in the family, including both primitive and specialized forms; single in other genera and altogether absent in Bathophilus, Eustomias and, presumably, Pareustomias. Whether paired or unpaired, they are strongest, most numerous and present on the most arches in the most primitive genera (see p. 105 ff.). Although they are almost invariable in their number and position within the genus, they are not identical in any two genera; hence they form a valuable generic character. Although the teeth on the posterior arches are often small, are almost covered by skin in large specimens, it is easy to determine their presence even in uncleared specimens. Along with the palatines, basibranchials and pharyngobranchials, the gill-teeth, except where much reduced, must be of considerable importance in gripping prey. 92 Zoologica: Neiv York Zoological Society [XXIV :6 Text-figure 3. Chirostomias pliopterus. Anterior view, showing abundance of teeth on floor and roof of mouth in a primitive genus. Standard length 205 mm. 1939] Beebe & Crane: Family Melanostomiaticlae 93 Text-figure 4. Eustomias bigelowi. Anterior view, showing paucity of teeth on floor and roof of mouth in a highly specialized genus. Standard length 134 mm. 94 Zoologica: New York Zoological Society [XXIV :6 vo Text-figure 5. Leptostomms gladiator. Jaws, hyoid, and branchial arches of larva, standard length 23 mm. The general proportions of the cartilaginous elements and the character of the jaw- and gill-teeth are typical of melanostomiatid larvae. Abbre- viations as in Text-fig. 18. ) Text-figure 6. Gonostoma elongatum. Jaws and first branchial arch, standard length 90 mm. For comparison with Text-fig. 5. The temporary teeth of the larva present a problem of much interest. Doubtless they help the young fish in catching its minute planktonic food, but their resemblance to the permanent teeth of certain gonostomids is noteworthy (Text-figs. 5, 6). On the other hand, they have their counter- part in other larval fish, notably the enormous fangs of leptocephali. The toothless post-larval stage, and the early periods of adolescence, when the permanent teeth are too few and too weak to be of any practical value, are, perhaps, accompanied by fasting on the past of the shrinking, metamorphosing young fish. In any case, these stages are probably of short duration. The growth of the permanent teeth is slow; maxillary teeth in particular often increase in number until very late in transitional adolescence. Larval gill-rakers have apparently not been observed before in this family. They lag behind the larval teeth in development, appearing as mere stumps when the temporary teeth are already strong, and reaching their maximum development at the very end of the larval stage, when the teeth are already falling out, or even during the post-larval period, when the jaws are toothless. The rakers on the first arch usually are equal in number and position to the gill-teeth, found in the adult, each raker corresponding to a single tooth, pair or group, i. e., the rakers are never paired, as is so often the case with the gill-teeth of adults. Also, they can generally be found on all five arches, although in the form of low, spiny mounds on the last two. We have found well developed gill-rakers in the larvae of Eustomias and Bathophilus, genera in which they are entirely lacking in the adult. These larval gill-rakers at maximum development are long and strong, with spines at irregular intervals, exactly like the rakers found in some adult gonosto- mids, such as Photichthys and Gonostoma. They are doubtless of use in straining out microscopic organisms from the water. 96 Zoologica: Neiv York Zoological Society [XXIV :6 < Text-figure 7. Leptostomias gladiator. Tooth from first branchial arch of larva, standard length 23 mm. < < & * 0 6. Branchiostegal Rays : As with the gill-teeth, the branchiostegal rays are in general most numerous and strongest in the less specialized genera. They are present on hypo-, cerato- and epihyals although often reduced on the first. They are of some use as a generic character, but, since the last three or four arise close together, and, moreover, tend to split and laminate, it is difficult to count them accurately except in cleared and stained speci- mens. Visible development begins in the post-larval period. 7. Fins : In the Melanostomiatidae the entire function of swimming and balancing must be relegated to the caudal peduncle and vertical fins, with balancing assistance from the pelvics, since pectorals are always modified beyond their usual function: they may be small (their usual condition) or absent, as in some Photonectes and Eustomias; imbedded in luminous tissue, as in Chirostomias , Grammatostomias , Trigonolampa, Thysanactus and Fla- g ellostomias ; or one or more rays may be long as in the tactile fins of cer- tain bottom-dwelling fish, examples being Echiostoma and Bathophilus. The number of rays varies greatly often even in the same genus. The maximum is 47, in Bathophilus nigerrimus; however, 2 to 10 rays are most commonly found. Rudimentary, subdermal rays are often present in cleared specimens. The pelvic is far less variable; in all genera except Echiostoma and some Eustomias, which have 8 rays, and Bathophilus, which numbers up to 26, the pelvic is 7-rayed. The posterior rays are sometimes very long, and all seem to be fully webbed in well-preserved specimens. The dorsal and anal, being almost continuous with the caudal, form a powerful swimming organ ; sometimes, as in Chirostomias, Grammatostomias and some Photonectes, the rays are covered almost to their tips with the thick, black skin of the body. The caudal fin, like the peduncle, is very short, usually between one- eighteenth and one-twentieth the length of the fish. The lower lobe is always considerably longer than the upper. The dorsal finfold persists longer than the anal fold, remains of one or both being present throughout the post- larval stage. The genera which are known at the present time to have the largest larval finfolds are Flag ellostomias and Bathophilus. The typical larval pectoral fins are large, even when, as has been noted, the pectoral is much reduced or absent in the adult. When the fin is highly specialized, as in Grammatostomias, the full length of the ray or formation 1939] Beebe & Crane: Family Melanostomiatidae 97 of the luminous material is not completed until transitional adolescence; when the fin is normal, as in Melanostomias, it is fully formed in the post- larval stage. Pelvic anlagen are often discernible in advanced larvae, but the rays are not differentiated before the post-larval stage; the fin does not reach its full length until adolescence or later. The majority of the rays of both dorsal and anal are distinguishable in the larvae. In Flagel- lostomias and Eustomias, in which the anal originates conspicuously in front of the dorsal, the anterior rays are the last to form. The caudal, even in late larvae, leaves the typical larval heterocercal form behind, goes through a homocercal stage, and immediately afterwards passes into the final phase, in which the lower lobe is longer than the upper. The entire fin, though small in the larva, is relatively much longer in the post-larva and adolescent than in the adult. 8. Epidermal Grooves: More or less well developed depressions are usually present along the isthmus to receive the barbel, or at least the basal part of its stem, when it is laid back. In long-barbeled forms, such as Bathophilus, a median groove runs to the anus and continues along one side of the anal fin. Similar grooves behind the pectoral insertion are the rule when the rays support luminous material, as in Chirostomias and Gram- matostomias. 9. Osteology: In the accompanying diagrams of osteology, we do not include, except for Bathophilus which will serve as an example, any dorsal views of skulls, since they have been figured already for most genera by Regan & Trewavas (1930). The same is true of the vertebral column. We have verified their findings in so far as possible, although the boundaries of skull bones are difficult to determine in cleared and stained examples of this family, the majority of which are immature. Our only major differ- ence is that we have found the post-temporal to be present in a number of genera in addition to Chirostomias and Trigonolampa (see below). The following remarks and comparisons are necessarily derived only from cleared specimens of genera in the present collection, combined with the observations of Regan & Trewavas on apparently uncleared examples. The skeleton of melanostomiatids is moderately well developed, but with apparently little calcium phosphate deposit and a great deal of calcium carbonate, judging from its usually feeble reaction to calcium phosphate stain. In all general features the skeleton of Idiacanthus is typical also of that of the melanostomiatids, especially of the group including Melanosto- mias. We refer, therefore, to the detailed- description of Idiacanthus already published in this series (Beebe, 1934). Due to the unsatisfactory reaction of most of the fish to alizarin stain, we can say little about relative degrees of ossification, both between genera at different growth stages and within the species. It is obvious, nevertheless, that as usual in deep-sea fish the jaws are the only really strongly ossified parts of the body, the gill-arches usually come next, then the tip of the caudal peduncle, while the skull proper, the rest of the vertebral column and the supports of the vertical fins are ossified very late, and then usually weakly. No ossification is ever found before transitional adolescence. Head : The lack of parietals in most genera; the union of the frontals by suture; the mesethmoid usually with lateral expansions; the upward, median projection of the premaxillary in all except Leptostomias; the reduced, laminar mesopterygoid ; and the weak opercular apparatus are the principal characteristics of this family. The skull, especially in Melanosto- mias and Photonectes, is very short in comparison with the length of the jaws. The hyoid and gill-arches in some genera, especially Eustomias and Photonectes, are also very short. The single supramaxillary (not differen- tiated from the maxillary in Text-fig. 11, varies in its boundaries and degree of attachment to the maxillary; the relative lengths of premaxillary 98 Zoologica: New York Zoological Society [XXIV :6 Bathophilus metatticus. Skull. Upper, dorsal view; lower, lateral view. Standard length 105 mm. Boundaries of bones approximate. General facies typical of the Melanostomiatidae. art, articular; clt, cleithrum ; dn, dentary ; epiot, epiotic ; exoc, exoccipital ; fro, frontal ; hyom, hyomandibular ; iop, interopercle ; lat eth, lateral ethmoid: max, maxillary; meseth, mesethmoid ; metptg, metoptery- goid ; op, opercle ; pal, palatine; pas, parasphenoid ; pop, preopercle; premax, premaxillary; ptero, pterotic ; ptg, pterygoid ; qu, quadrate ; soc, supraoccipital ; sop, subopercle ; sphen, sphenotic ; supclt, supracleithrum ; supmax, supramaxillary. 1939] Beebe & Crane: Family M elanostomiatidae 99 and maxillary bordering on the gape also vary; the premaxillary is shortest in Chirostomias, Leptostomias and Photonectes margarita, longest in Eusto- mias. In the latter genus, the palatine and ectopterygoid are loosely attached by ligaments to the mesethmoid and quadrate respectively, but firmly fastened to the upper jaw; this arrangement permits the forward projec- tion of the jaw. There are 3 hypohyals. Pectoral Girdle : According to Regan & Trewavas (1930) post-temporals are present only in Chirostomias and Trigonolampa. We have found small ones, however, brightly stained, but not connected with the skull, in Flagel- lostomias, Leptostomias and Echiostoma. In addition vestigial slivers (some- times absent on the opposite side of the same specimen) are occasionally found in Photonectes. The supra-cleithrum is reduced in Melanostomias and Photonectes and entirely absent in Eustomias. The mesocoracoid is usually present, but sometimes has the upper arm reduced, and is entirely absent in Eustomias and Photonectes. Upper and lower coracoids usually well devel- oped, laminar. The actinosts tend to be reduced in number and of peculiar shapes and positions, corresponding to the various modifications of the pectorals. The rays themselves are laminate and strongly ossified basally in Grammatostomias , doubtless to support the weight of the luminous tissue. Vertebral Column : Regan & Trewavas have already described and figured the extraordinary modifications in the anterior part of the vertebral column in some of the melanostomiatids. Chirostomias and Trigonolampa alone are completely unmodified, with the first centrum fixed firmly to the skull. Most genera have the first one or two centra represented only by spinal nerves, there being a tough fibrous sheathing around the notochord. Leptostomias has similar modifications of seven vertebrae and Eustomias is the extreme with 9 or 10 specially adapted. The use of the modification in all, of course, is for the increase of the gape in grasping prey; the separa- tion of the post-temporals from the skull and their atrophy are closely con- nected with this adaptation. Posterior Part of Vertebral Column and Caudal Fin: The last 2 or 3 vertebrae before the urostyle are the first to be markedly modified, with prolonged neural and haemal arches reinforced with laminar expansions. There are 6 hypurals, 3 dorsal and 3 ventral to the median axis. Each gives rise to from 2 to 5 rays, the fifth and third hypurals being usually broadest, and supporting the most rays. This part of the vertebral column often becomes ossified before the rest. The sequence of raylets and rays, counting from the anterior dorsal raylet around to the corresponding ventral one, is as follows: 5 to 10 + 9 to 11 + 9 to 10 + 3 to 5 ( i . e., 18 to 21 true rays). 10. Coelomic Organs: The general plan of the body cavity of melano- stomiatids is identical with that found in female I diacanthus (Beebe, 1934.1, p. 218, fig. 75). The stomach varies from about 20% to 45% of the length of the fish, being shortest in Leptostomias, longest in Echiostoma. A straight intestine with two pyloric caeca is the rule, although in Chirostomias and Pachystomias the caeca are rudimentary, and in Opostomias, Flagellosto- mias, Thysanactis and Leptostomias an anterior pouch gives rise to a single caecum; in Odontostomias there is a second caecum in addition to the pouch structure. The great posterior extension of the gut, free of the body during the larval stage, is apparently confined to and characteristic of all of the Gym- nophotodermi. This phenomenon is probably a specialization connected with the lack of the yolk sac at this stage plus the absence of a stomach; there- fore the gut is large in diameter and projects behind, so as to give addi- tional absorptive surface. Sometimes it is more or less gathered in humps in the middle of its length, which accordingly presents still more digestive area. Subsequently, during the post-larval shrinking period, it is gradually absorbed. The stomach is certainly never large enough to be used until the 100 Zoologica: Neiv York Zoological Society [XXIV :6 Text-figure 9. Bathophilus metallicus. End of vertebral column in A, post-larva, standard length 25 mm.; B, adolescent, 30 mm.; and C, transitional adolescent, 96 mm. Principal characteristics typical of the Melanostomiatidae. 1939] Beebe & Crane: Family Melanostomiatidae 101 D Text-figure 10. Melanostomias spilorkynchus. Diagrams showing relative lengths of stomach and intestine, and position of anal fin, in respect to standard length. A, larva, standard length 17 mm.; B, post-larva, 24 mm.; C, adolescent, 31 mm.; D, transi- tional adolescent, 35 mm.; E, adult, 222 mm. This series is typical of the Melanostomiatidae. adolescent stage. It does not reach its full pigmentation and length until sometime during transitional adolescence — often very late in the stage (as in Leytostomias and Flagellostomias ) and is one of the most useful means of determining immaturity. The gonads also develop very tardily, and we have not found it possible to distinguish sex earlier than transitional adolescence. When the gonads have developed to more than transparent ribbons of tissue, it is easy to determine the sex, since the eggs of the female are distinct from the very first, whereas testicular tissue is in contrast superficially homogeneous. However, in adults near breeding condition ripe testicles can easily be mis- 102 Zoologica : New York Zoological Society [XXIV :6 taken for unripe ovaries by workers not familiar with the group, since the testicles have a semi-granular appearance similar to that of partially devel- oped ovaries in certain other fishes. However, dissection and microscopical examination at once disclose their true nature. On the other hand, under low power, the ovaries of immature fish often resemble testicles; therefore a small piece of gonad should always be dissected and high power used, to avoid mistakes in this important subject. 11. Shrinking: Reduction in length during the post-larval, metamor- phosing stage appears to be relatively slight in the melanostomiatids, com- pared with that found in Chauliodus, in Stomias and in eels. We have found no evidence in the present material that more than 10 mm. in length is lost during this period. Whereas the reduction in length in eels takes place during early adolescence, in this family it occurs chiefly during the post-larval period. It may be stated here that good-sized melanostomiatids, around 300 mm. in length, shrink up to 20 mm. after preservation in 70% alcohol, and specimens of other lengths in proportion. There is a corresponding loss of depth, usually greater in proportion than the length; the head, eye and snout shrink little, however. Measurements given in the following pages, unless otherwise stated, are made from preserved specimens. E. Ecology. 1. Horizontal Distribution: Only 7 of the 16 genera of Melanosto- miatidae have been taken outside the Atlantic Ocean : Opostomias and Tactostoma, known only from Australia and the eastern Pacific, respectively; Leptostomias, from Hawaii and both North and South Atlantic; Pachysto- mias, from Australia and the North Atlantic; Photonectes, from Japan and the North Atlantic; Melanostomias, from the Indian Ocean and both North and South Atlantic; and Bathophilus, from the Indian Ocean and both North and South Atlantic. Odontostomias, Echiostoma, Flag ellostomias and Eustomias are known from the North and South Atlantic. The remaining genera — Trigonolampa, Chirostomias, Thysanactis, Grammatostomias and Pareustomias — have so far been taken only in the North Atlantic. Some genera will doubtless be found to have a wider distribution when intensive trawling is carried out in other oceans. Ten of the genera or 62V2% have been taken by the Bermuda Expedi- tions. These include every genus previously recorded from the western Atlantic except Thysanactis and Trigonolampa, the former being apparently a tropical form and the latter boreal. Of the remaining four genera, two ( Odontostomias and Pareustomias ) are known from the eastern Atlantic only; while Opostomias and Tactostoma, as remarked above, have been taken only in the Pacific. For observations on the distribution of species, see Regan & Trewavas, 1930, p. 34. The absence of Trigonolampa, Bathophilus pawneei and several species of Melanostomias in the Bermuda collection is added evidence to their suggestion that these forms are tropical and Antillean, rather than subtropical, in distribution. It seems worthwhile to reemphasize the fact that the 250 specimens, 32 species and 10 genera composing the present collection, and forming respec- tively more than a sixth, a third and five-eights of the known specimens, species and genera, were all obtained in what is scarcely more than a drop of water in the Atlantic Ocean: in an area 5 miles south of Nonsuch Island, Bermuda, 8 miles wide and 1 mile deep. 2. Vertical Distribution : As is the case with other families of deep-sea fish, the melanostomiatids around Bermuda seem to live at greater depths than elsewhere. Excluding a few colorless larvae, members of the family were not taken in the trawling nets above 300 fathoms (549 metres), and 1939] Beebe & Crane: Family Melanostomiatidae 103 most were taken far below this level, between 500 and 1,000 fathoms (914 to 1,829 metres, whereas other expeditions, notably the Dana, took a great number of specimens “with nets fishing at 200 metres or less below the surface” (Regan & Trewavas, 1930, p. 34). Thanks to our tests with the pressure gauge (Beebe, 1930, p. 244), we are convinced that the great majority of the Bermuda specimens were taken at the level trawled, and not when the net was on the way to the surface. Because of Bathysphere observations, however, it is also clear that these families are not absent from the upper layers here, but can merely avoid the net better, because of the fact that some light penetrates to these depths. From the Bathy- sphere, in its dives between the surface and 3,028 feet, members of the family were recognized 26 times, between 750 and 2,750 feet (125 and 458 fathoms or 228 and 835 metres), ranging in length from one inch to six feet (counting the six-foot Bathysphaera Intacta). Larvae and post-larvae were taken in the nets between the surface and 1,000 fathoms; it is likely that the few taken at the greater depths were among the minority caught on the way upward. 3. Abundance: About 1,450 specimens of Melanostomiatidae have been taken, including the 250 in the present collection. The total number is dis- tributed among 16 genera and, in the light of the synonymies proposed in the present paper, about 115 species. We are certain, however, that many of these, especially in the genera Leptostomias and Eustomias, will prove to be invalid, so that the total number of true species known at present comes actually to considerably under 100. In numbers of individuals, Bathophilus and Eustomias , with about 500 and 400 specimens, respectively, are the most abundant; Echiostoma, Melanostomias and Photonectes are each known from between 100 and 160 specimens, and Leptostomias from 49 ; less than 25 examples have been taken of every remaining genus. The best known species are Eustomias obscurus and Bathophilus metallicus, of which about 200 and 185 specimens have been taken, respectively. Melanostomiatids are among the rarest groups of deep-sea fishes taken off Bermuda. In contrast to the thousands of Cyclothone, myctophids and Sternoptyx taken, a total of only 250 melanostomiatids came up in the nets, belonging to 10 genera and 32 species. Of these, Melanostomias spilorhyn- chus, of which we have 51 specimens, was the most numerous, Photonectes dinema (26 specimens) next, Bathophilus metallicus (22 specimens) third, and Leptostomias gladiator (20 specimens) fourth. Of the remaining spe- cies, a dozen are represented only by single examples. In regard to number of species, as opposed to individuals, the Mela- nostomiatidae are surpassed in the collection only by the Myctophidae, of which about 57 species were taken, as opposed to the 32 species of the present family. It is interesting to note that in number of individuals it is the plankton eaters — Cyclothone, myctophids and sternoptychids — that are numerically far ahead of eaters of fish and shrimps, such as the large-toothed stomiatoids ( Stomias , melanostomiatids, Chauliodus, astronesthids, Idiacanthus) , the lyomerids, large-mouthed pediculates, Chiasmodon, etc., just as on land, large carnivores are surpassed in numbers by their herbivorous prey, such as rodents and ungulates. 4. Food and Enemies : The food of melanostomiatid larvae is, of course, confined to small organisms such as diatoms and copepods. Toothless, trans- forming, post-larvae and adolescents, however, probably do not eat at all for a short while; at least, we have found no food in their intestines. Transi- tional adolescence, however, is again a period of growth, the stomach is well developed, and the food represents on a small scale the food of adults, namely myctophids and other small fish, and good-sized shrimp. 104 Zoologica: New York Zoological Society [XXIV :6 Although more than 40 stomachs of transitional adolescent and adult melanostomiatids were examined, less than half contained any food at all, although finely digested matter was usually present in the intestine. The likelihood is that these strong fishes digest their food rapidly, even just before death. Myctophids were present in nine stomachs, unidentified small fish in three, Luciosudis in one, Cyclothone microdon in one, a shrimp in one, and ostracods in one. Without exception the fish were swallowed whole, head first, and measured one-half to five-sixths the length of the melanostomiatid. We have not found melanostomiatids in the stomach of any Bermuda fish. A specimen, the second known, of Trigonolamya was, however, found in the stomach of a swordfish (Parr, 1933, p. 178). 5. Activity : From the Bathysphere these fish appeared agile and eel-like, with rather slow twistings in progression. None of them seemed to be affected by the search-light. Usually only one of these fishes was seen at a time, but occasionally two or three appeared swimming together. When brought up alive they swam about and snapped with all the accuracy of balance and swiftness of surface fish. As with other living deep-sea fish, they would try to burrow downward, bumping their snouts against the bottom of the pan. Always they could be greatly revived by being placed in a pan of ice-cold salt water in the refrigeratoiv A young Pachystomias atlanticus, 37 mm. long, was the smallest member of the family, and, in fact, the smallest deep-sea fish, taken alive. F. Phylogeny. We agree with Parr (1927, p. 4) and Gregory & Conrad (1936, p. 23, fig. 2) that an astronesthid-like form or foi'ms were the ancestors of the Gymnophotodermi, and hence of the Melanostomiatidae. In common with the more primitive stomiatoids, the Astronesthidae have fixed teeth, unspecialized fins and vertebral column, and an adipose fin. Yet they have the barbels, well developed postorbital organs and naked black skin of the Gymnophotodermi. Parr pointed out the variability of the position of the dorsal fin in the Astronesthidae. The later work of Regan & Trewavas (1929) showed the diversity of other characters in the same family — char- acters which are found in similar diversity in the Melanostomiatidae, such as the form of teeth on the maxillary and gill-arches. All erect maxillary teeth are found in several astronesthid genera, exactly as in the melanostom- iatid Chirostomias, whereas all oblique teeth occur in others — as in the more specialized melanostomiatids. Similarly, both double and single gill- arch teeth occur in the family, just as in the Melanostomiatidae. Double gill-teeth are found as well far down the stomiatoid scale in Photichthys, whereas single, raker-like teeth and actual spiny rakers occur in such genera as Gonostoma, as well as in melanostomiatid larvae. In summary, the existing genera of Astronesthidae show all the elements needed by a hypothetical ancestor of the Melanostomiatidae — the various types of maxillary and gill-teeth, a variable dorsal fin trending backwards, and un- specialized finrays, combined with the already specialized naked black skin, barbel and well-developed postorbital organ. The remaining Gymnophoto- dermi— the Idiacanthidae and Malacosteidae — are doubtless off-shoots of the Melanostomiatidae. As a family, the Melanostomiatidae have specialized in slenderness, with increased numbers of vertebrae and elongation of the stomach, in elaborate and elongate barbels, in sexual dimorphism in the development of the postorbital cheek light, in specialized luminous or elongate pectoral fins, in the acquisition of depressible teeth, in the reduction and loss of parietals, in the reduction of the opercles and pectoral girdle, in modifica- Eustomias fissibarbis. Photonectes margarita. Eustomias fissibarbis. Text-figure 11. Jaws, hyoid and branchial arches, and pectoral girdles of typical Melan- ostomiatidae. Depressible jaw-teeth unshaded; divisions between palatine and pterygoid, and maxillary and supramaxillary not shown; fourth basibranchial, always unossified, unshaded. For the same illustrations, enlarged and labeled, see Text-figs. 13, 18, 23, 28, 33, 34, 45, 50, 55 and 64. 1939] Beebe & Crane: Family Melanostomiatidae 105 tions of the anterior part of the vertebral column, and in the shortening of the caudal peduncle and fin. Primitive, Specialized and Adventitious Characters : In determining the relationship of the genera to one another, the following characters may unquestionably be considered primitive, with their roots far back in the stomiatoid stock, since they are the rule among the more primitive of existing stomiatoids: Fixed, barbless teeth; an adipose fin; pectoral girdle well developed with a strong post-temporal and a related lack of modification in the anterior part of the vertebral column; pelvics near the middle of the body; parietal present; single or double gill-arch teeth strongly developed, present on all five, or at least four, arches, including hypobranchials and epibranchials as well as ceratobranchials. The following characters, on the other hand, prove to be of almost no use in determining relationships, since each of them varies greatly, not only in closely related genera, but even within the same genus: body proportions, notably depth and head length; (examples: Leptostomias , Bathophilus ) ; barbel length and form (example: Eustomias) ; pectoral de- velopment (example: Bathophilus, Eustomias ); presence and distribution of superficial luminous tissue (examples: Photonectes, Bathophilus) . Also, these variable characters crop out in specialized form in the most primitive genera, and vice versa. For example, Chirostomias, unquestionably the most primitive genus in the family in fundamental structural characters, is equipped with a highly complex barbel and pectoral fin. Similarly, Bathophilus and Tactostoma, two of the most specialized end-genera, have simple or degenerate barbels. Several distinct characters may be termed adventitious, since they occur sporadically throughout the family and, indeed, throughout the stomiatoids as a whole. Such are grouped serial photophores, which are common among the gonostomids and sternoptychids, and pi’esent in the astronesthid Heterophotus, in the malacosteid Aristostomias and in the melanostomiatids Pachystomias, Eustomias obscurus and Pareustomias. A more or less elongate anal fin, present in Flag ellostomias and Eustomias in the Melanostomiatidae and in other genera scattered through related fami- lies, is a similar character. An exceptionally high number of vertebrae is a third; among melanostomiatids Leptostomias, Tactostoma and Eus- tomias, although they have practically nothing else in common besides general family characters, all have many more vertebrae than the average of 50 to 60. The atrophy of the postorbital cheek light in females is a most puzzling character, since it does not, as far as is known, occur in lower stomiatoids, and yet is the rule in otherwise primitive melanostomiatids, while in females of the most specialized genera the organ is again functional and even (as in Melanostomias, Echiostoma and Photonectes) as large as in males. Generic Interrelationships : We support in general the groupings of genera suggested by Regan & Trewavas (1930). That is, we agree that Chirostomias and Trigonolampa are closely related and the most primitive known genera; that Leptostomias, Thysanactis, Flag ellostomias and Opos- tomias along w,ith Norman’s Odonotostomias (1930) form a natural group intermediate in degree of specialization; that Bathophilus is close to Grammatostomias; that Melanostomias, Echiostoma and Photonectes are closely related; and that Pachystomias and Eustomias are both aberrant. In addition, with our study of characters, such as gill-arch teeth and larval stages, other than those emphasized by these authors, we are able to give a tentative but plausible sketch of the relationships of the groups to each other. As we see it, Bathophilus could not possibly have come from a Chirostomias- like form except insofar as such a form was probably ancestral to the entire family, nor Eustomias from an Echiostoma- like fish — 106 Zoologica: New York Zoological Society [XXIV :6 Echiostoma itself being one of the most specialized genera, and not at all on the same line of development as that followed by Eustomias. Both of these suggestions have been made by Gregory & Conrad (1936, p. 26). Parr’s (1927) pioneering suggestions as to relationships within the group have of course been largely superseded by the osteological study of the large Dana collection. From the accompanying diagram (Text-fig. 12) it will be seen that we recognize a number of generic groups, with Tactostoma on the one hand and Bathophilus and Pareustomias on the other as the most specialized end forms in the Melanostomiatidae. The family shows a natural division into genera in which the gill-arch teeth are paired, and those in which they are single or absent. Another main division is into the primitive forms with all of the jaw teeth fixed, and the remaining genera, in which at least a few are depressible. We will now consider the various groups in detail. Chirostomias, Trigonolampa: These genera, both having paired gill- arch teeth, are, as has been said, the most primitive. Their basic, generalized characters are fixed teeth, well-developed parietals, numerous gill-teeth, unspecialized vertebral column, short stomach, strong post-temporal and pectoral arch and, in Chirostomias , an adipose fin. In contrast, the variable characters of barbels and pectoral fins are highly specialized. The post- orbital organ is completely atrophied in the female in Chirostomias ; its condition in female Trigonolampa is unknown. Pachystomias, with its small, fixed teeth and unspecialized fins, is placed next on the tree. As Parr, judging only from external appearances, suggested (1927), we think it is close to the line of malacosteid develop- ment, with its somewhat similar suborbital lights; exceptionally large, mas- sive head; strong, backwardly-extended jaws joined by only a thin mem- brane; its general shape and fin arrangement; and its single-gill-arch teeth. In the malacosteids the gill-teeth are single or absent, teeth in the jaws are fixed, the jaw membrane is entirely absent, and the grouped photo- phores of Aristostomias are similar in pattern to those of Pachystomias. The absence of parietals, post-temporals and pyloric caeca are other char- acters in common, and although the skulls are dissimilar, they present no significant differences to preclude relationship. The development of the postorbital light organ in females is unknown in Pachystomias. Oclontostomias, Thysanactis, Leptostomias, Opostomias, Flagellostomias : These genera form the somewhat heterogeneous base for the remaining melanostomiatids, just as the dissimilar Astronesthidae form a similar base for the more specialized of the Gymnophotodermi. The five genera have in common the primitive and semi-primitive characters of parietals, massive head and jaws, few depressible teeth, a pair of teeth on the vomer, many gill-teeth on four or five arches, moderately short stomachs, rudi- mentary caeca, and, with the exception of Thysanactis and Leptostomias, a vertebral column almost unmodified anteriorly, but usually with many vertebrae. We have found small but well ossified post-temporals, not con- nected with the skull, in both Flagellostomias and Leptostomias ; these bones were not discernible in the Dana specimens described by Regan & Trewavas (1930), probably because those specimens were not cleared and stained ; the bone may prove to be present in all five genera. The post- orbital organ is atrophied in the adult female at least in Odontostomias, Leptostomias and Flagellostomias. It is probable that the members of this group attain maturity at a greater length than others in the family. There is a natural division into two sub-groups, the first ( Odonoto - stomias, Thysanactis and Leptostomias ) having the gill-teeth paired or in threes and fours on all arches except the last, and the second ( Opostomias and Flagellostomias) with all the gill-teeth single. The ancestors of the first division probably gave rise to the Echiostoma-Melanostomias-Photo- nectes-Tactostoma group, while from Opostomias-Flagellostomias roots came 1939] Beebe & Crane: Family Melanostomiatidae 107 Text-figure 12. Suggested phylogeny of the Melanostomiatidae. 108 Zoological New York Zoological Society [XXIV:6 offshoots ending today in Grammatostomias and Bathophilus and in Eus- tomias and Pareustomias. Echiostoma, Melanostomias, Photonectes, Tactostoma3: These four genera have in common large but usually slender jaws with numerous teeth, almost all depressible, and with strongly barbed tips; paired or grouped teeth in reduced numbers on only two or three gill-arches; similar, very short skulls without parietals; post-temporals reduced or absent and the rest of the pectoral girdle weakened; anterior part of vertebral column slightly modified; a long stomach with two well-developed caeca; and pelvic inserted far behind the middle of the length. In Melanostomias , Echiostoma and Photonectes the cheek light is exceptionally and equally large in both males and females, apparently the only genera in the family of which this is true, and the larvae are very similar. (Neither of these characteristics is known as yet in Tactostoma, except that the postorbital is not large in the known specimens). Echiostoma and Tactostoma are the only genera in the family having the teeth multi-rowed, although the single rows of other genera are rarely perfectly regular). The teeth of adult Photonectes and of Tactostoma are small, but those of immature Photonectes bear a striking resemblance to those of adult Melanostomias (Text-fig. 9). In all except the latter genus the barbel tends to reduction, being better developed in the young than in the old, and vestigial in Tactostoma. Grammatostomias, Bathophilus , Eustomias, Pareustomias : The first two and the last two form closely related sub-groups; in fact, Pareustomias may prove to be a sub-genus of Eustomias. Grammatostomias is the only genus in which any gill-teeth at all are found, and these are few and single. Parietals and post-temporals are absent, teeth (except in some Eustomias) are mostly depressible, moderately large and sometimes with rudimentary barbs; stomach moderately elongate; two caeca; postorbital light organs usually large in male, always smaller, sometimes atrophied, in female. Bathophilus is highly specialized in development of fins and body depth, with vestigial serial organs. The semi-primitive genus Flag ellostomias has the beginning of the protractile snout which is carried to such high develop- ment in Eustomias. There is a questionable trace of the same character in Grammatostomias. All the genera have a tendency toward elaborate barbels, either through simple elongation or through the development of ornate branches and filaments. Comparison of Specialized End-Genera : Comparison of the two groups of offshoots from the five central genera is interesting. Group A (Echiosto- ma, and its allies) has kept paired gill-teeth, while Group B (ending in Bathophilus and Pareustomias) has single, Flag ellostomias- like gill-teeth or has lost them altogether. In both A and B the arch teeth, whether paired or unpaired, are reduced in number, being present at most on the first three ceratobranchials and first epibranchial. In Group A the jaw teeth become all depressible, more or less barbed and very numerous, the extreme being reached in the multi-rowed teeth of Tactostoma, and both the vomerine teeth* and an erect series on the maxillary are kept; in Group B a number of fixed teeth is always kept, the jaw teeth are never numerous, barbs are rudimentary or absent, and both vomerine teeth and erect maxillary teeth are lacking. In Group A the skull is reduced, although there is no jaw reduction, while in Group B the skull remains little shorter than the jaws, as in primitive genera. In both groups the pectoral girdle is reduced, and the anterior part of the vertebral column modified to allow a backward and upward movement of the head, increasing the gape. Sometimes there is a forward thrusting of the lower or upper jaw in capturing food. In Photo- nectes the curved lower jaw is dislocated and thrust forward, while in Eustomias it is the snout and upper jaw. The pelvics are inserted far back in Group A, while in Group B they have remained near the middle of the 3 Osteological and internal characteristics not yet known in Tactostoma. 4 Except in Tactostoma. 1939] Beebe & Crane: Family Melanostomiatidae 109 body. In Group A the barbel tends toward secondary reduction approaching atrophy, while in Group B are found the most elaborate and elongate barbels in the family. Both groups have elongate stomachs and two well-developed caeca. On the chart the Idiacanthidae are shown as a highly specialized offshoot of the main melanostomiatid stock. The skull and teeth are similar to those of Melanostomias, while the general form, lack of gill-arch teeth, shape and pigmentation of the larva, and unequal development of the postorbital organ in males and females show some resemblance to the Flag ellostomias-Eusto- mias axis. Conclusions : Most of the specializations of the Melanostomiatidae, both beyond those of the lower stomiatoids and within the family, are in the direction of increased efficiency in the capture, swallowing and digestion of large, living prey. To this end, the body becomes elongate and stream- lined with the vertical fins forming a single, powerful, swimming organ. The jaw teeth are enlarged, and efficiency is further increased through the development of depressibility and of barbs. Teeth on the vomer, palatines, basibranchials and gill-arches assume great functional importance. The gape is enlarged both by the flexibility of the jaw angle and by the modification of the anterior vertebrae and the related disconnection of the pectoral girdle from the skull, which enables the head and upper jaw to be swung backward and upward; through these devices the mouth can often be opened to an angle of fully 180 degrees. In addition, the upper jaw with its strong fangs can sometimes be thrust forward ; in other cases a similar movement can be made with the lower jaw through the swinging forward of the elements of the hyomandibular arcade. In the genera having these highly specialized modifications, all teeth except those in the jaws themselves are reduced in size and number, in direct ratio: the greater the depi'essibility of the teeth, the modification of the vertebral column, and the distensibility of snout or mandible, the fewer the teeth on the roof and floor of the mouth. The stomach becomes elongate for the reception of large, whole prey, which is invariably swallowed head first. Some specializations of luminous organs are probably also concerned in increased efficiency in the hunt for food, but the development of the post- orbital organ, at least, is unquestionably of sexual significance, while the same is sometimes true of the barbel. The following key to the genera of Melanostomiatidae attempts both to be of practical use and to indicate relationships. G. Synopsis of the Genera. A. All teeth in jaws firmly fixed. B. Teeth on first gill-arch in pairs; parietal present; post-tem- poral present. C. Adipose present Chirostomias (p. 111). CC. Adipose absent Trigonolampa. BB. Teeth on first gill-arch single, parietal absent; post-temporal absent Pachystomias (p. 117). AA. Some teeth in jaws depressible. D. Gill-teeth present at least on first 4 arches, and usually on all 5; lower jaw with only 1 or 2 depressible teeth; parietal present; post-temporal small or absent. E. Gill-teeth on first arch mostly in pairs. F. An isolated pectoral ray; gill-teeth present on 5th arch Thysanactis. FF. No isolated pectoral ray. 110 Zoologica: Neiv York Zoological Society [XXIV:6 G. Gill-teeth present on 5th arch; 32-35 O-V photo- phores Odontostomias. GG. Gill-teeth absent on 5th arch ; 39-47 O-V photo- phores Lcptostomias (p. 121). EE. Gill-teeth on first arch single; an isolated pectoral ray. H. Mandibular fangs perforating premaxillaries; dorsal and anal commencing at same vertical Opostomias. HH. Mandibular fangs not perforating premaxillaries; anal commencing well in front of dorsal Flagellostomias (p. 179). DD. Gill-teeth never present on more than first 3 arches; sometimes absent; lower jaw with more than 2 teeth depressible (or, if only 2, the gill-arches are toothless) ; parietal absent; post- temporal rudimentary or absent. I. Teeth on first gill-arch paired, or in groups of 3 or 4; vomerine teeth present; some erect teeth on maxillary; jaw teeth slightly or sharply bicuspid. J. Pectoral of 5 or 6 normal external rays; post-temporal sometimes present Melanostomias (p. 142). JJ. Pectoral of less than 5 rays. K. Pectoral of 4 external rays, the first isolated and produced; cleft of mouth straight; post-temporal present Echiostoma (p. 130). KK. Pectoral of 0 to 3 rays ; cleft of mouth more or less strongly curved upward. L. Jaw teeth in a single row; post-temporal sometimes present Photonectes (p. 154). LL. Jaw teeth in many rows or groups; post- temporal? Tactostoma. II. Gill-teeth all single or entirely absent; no erect maxillary teeth; vomerine teeth absent; post-temporal absent; jaw teeth slightly or not at all barbed. M. Gill-teeth present; a line or loop of luminous tissue on side Grammatostomias (p. 185). MM. Gill-teeth absent; no line or loop of luminous tissue on side. N. Dorsal and anal beginning at same vertical; upper jaw not protractile; several teeth on pala- tines; supracleithrum present Bathophilus (p. 196). NN. Anal originating well in advance of dorsal ; upper jaw protractile; palatines toothless; supraclei- thrum absent. 0. Premaxillary normal Eustomias (p. 210). 00. Premaxillary free of maxillary, curving up- ward above jaw line Pareustomias. 1939] Beebe & Crane: Family Melanostomiatidae 111 H. Report on the Collection of the Bermuda Oceanographic Expeditions, Including Revisions of Genera and Species. Genus Chirostomias Regan & Trewavas, 1930. (See also pp. 71, 73, 82-86, 90, 91, 96-99, 102, 104-106, 109). (Text-figs. 3, 11, 12, 13-16 incl.). General Discussion. Upon reexamination of the type specimen of Chirostomias lucidimanus Beebe, 1932, and comparison with immature specimens in the same Bermuda collection, we have decided to synonymize it with C. pliopterus Regan & Trewavas, 1930, the only other species which has been described. Our rea- sons for this step are as follows: I. The barbel differs from that of C. pliopterus only in a manner con- sistent with growth. The largest specimen described by Regan & Trewavas was 115 mm. long; C. lucidimanus measures 205 mm.; the next largest Bermuda specimen measures 118 mm. and is immature, with a barbel inter- mediate between that of lucidimanus and pliopterus; barbels of all small specimens (35 to 41 mm.) in the Bermuda collection agree excellently with typical pliopterus. The greater length of the stem in the largest specimen, and the greater number of anterior bulb filaments, are both perfectly normal growth differences; it is possible that the third, rather surprising difference — that of the smaller number of posterior bulb filaments — is due to their being literally rubbed gradually away, by contact of the barbel bulb with the isthmus, when the barbel is laid back in its groove. Minor details may also be sexual characters, but since we have only one male, an immature 118 mm. specimen, more material is needed in order to settle the question. 2. A recount of the dorsal fin rays in the type specimen of C. lucidi- manus gives 18 rays, as in pliopterus, instead of 16, as stated in the descrip- tion of C. lucidimanus. 3. Although the eye of C. lucidimanus is relatively smaller than in pliopterus, this characteristic again may logically be attributed to the difference in size; in our intermediate, 118 mm. specimen, the proportionate size of the eye is as in typical pliopterus. Distribution : Chirostomias pliopterus, the single known species, has been taken only in the Atlantic Ocean north of 20°, at depths between 55 and 700 fathoms. Known altogether from 13 specimens, including the present series. Generic Characters. (Since only one valid species is known in this genus, the following char- acters are also those of the unique species, C. pliopterus Regan & Trewavas). Color (from freshly caught immature male and two immature and 1 adult females, the adult being alive) : General color, velvety black with greenish bronze iridescence on shoulder; iris black; postorbital organ of male white, rimmed dorsally with silver; barbel bulb white and silver with pinkish luminescence anteriorly, white posteriorly (see below under “Barbel”) ; this organ is greenish-yellow and silver in young females, with no glow observed; serial photophores violet, the lights of the lateral series having gilt caps divided into five successive small sections and, below the lights, single, undivided smaller patches of gilt. Proportions: Moderately elongate melanostomiatids with adipose fin; depth in length 6 to 7.7 (13% to 16.7%); head in length 5 to 6 (16.7% to 20%) ; eye in head 4 to 5.8 (2.9% to 3.9% of length) ; snout to pelvic in length 1.7 to 1.9 (53% to 59%). 112 Zoological Neiv York Zoological Society [XXIV :6 I - yiGill AR.GMES 1- hypobranchiab 2- ce,ratobranchial6 3- epibranGhi.als 4- pharyngobmnchi ab Text-figure 13. Chirostomias pliopterus. Jaws, hyoid and branchial arches and pectoral girdle of transitional adolescent, standard length 118 mm. Explanation and abbrevia- tion as in Text-fig. 18. Barbel : Shorter than head, with stout black stem ending in a large, ovate swelling, which is black spotted with luminous tissue, and cleft dis- tally; a terminal, anterior series of simple and compound translucent projections, and a posterior, translucent, fringed flange. 1939] Beebe & Crane: Family Melanostomiatidae 113 Text-figure 14. Cldrostomias pliopterus. Supporting bones of pectoral fin. A1, A2, A3, A *, A5, actinosts; CL, cleithrum, L, lower coracoid; M, mesocoracoid; U, upper coracoid; fin-rays in solid black; short ray rudimentary, invisible externally. From a tran- sitional adolescent, standard length 118 mm. The barbel of the largest specimen known, a 205 mm. Bermuda female in breeding condition, differs slightly from those of previously known, younger examples, and hence a description is given in full below. This specimen was formerlv described as C. lucidimanus Beebe, 1932 (see under “Discussion,” p. 111). The bulb is blue-black, elongate and somewhat compressed. The ter- minal part is cleft, forming two large, tubular divisions, the anterior the broader, each tipped with a pair of sharp, tooth-like structures opening toward one another. From the front of the anterior division arises a tuft of 7 filaments from a single base, the upper ones being longest, longer than the bulb itself. Under the ultra-violet light, while the fish was still feebly alive, these filaments gave off a pinkish glow. At the tip of the anterior division, in the position of the club-shaped appendage found in smaller specimens, is a thick, beaded tubercle, while from each of the two extreme distal teeth-like structures arise two or three short filaments. The posterior surface of the bulb shows a number of isolated spots of luminous tissue which consolidate into a thick, luminous, white comb or flange, with only a few very short filaments at the tip in place of the bushy fringe found in smaller specimens ; the most distal of these is, however, as usual, a longer, beaded structure. The luminous tissue, which gave off a white glow in this area, dies out on the surface of the mid-bulb in an ever-thinning mass of scattered spots and dots. There is at least a single muscle at the tip of the bulb, which has the power of separating the terminal structures widely, the four tooth-like pro- tuberances showing up strongly through the translucent pink luminous tissue. 114 Zoologica: New York Zoological Society [XXIV :6 Chirostomias pliopterus. A, adult female, standard length 205 mm.; B, same, pectoral fin, showing luminous material on fourth and fifth rays. Light Organs : Postorbital moderately large in male, completely atro- phied in female. Serial photophores with the following counts: ventral series, I-P 9, P-V 26 to 28, V-A 18 to 20, of which 5 to 6 are above the anal, A-C 10 to 11; lateral series, O-V 23 to 24, V-A 18 to 20. Pectoral fin with luminous tissue. Teeth : Cleft of mouth straight; all teeth fixed; premaxillary and mandible with about 7 to 13 curved fangs each, set in 2 irregular rows; maxillary with 7 to 12 fangs and several small, oblique teeth at end; a pair of teeth on the vomer; a series of 7 to 9 on each palatine, extending onto the ectopterygoid. Usually 5 to 6 pairs of teeth on basibranchials. Teeth in pairs present on all gill-arches except fourth : on first and second hypo- branchials; on first, second, third and fifth ceratobranchials; and on first, second and third epibranchials ; 11 to 12 pairs present on first cerato- branchial. Branchiostegal Rays : 22. 1939] Beebe & Crane: Family Melanostomiatidae 115 Text-figure 16. Chirostomias pliopterus. A, end of barbel in adolescent, standard length 41 mm.; B, end of barbel in transitional adolescent male, 118 mm.; C, pectoral fin of same; D, end of barbel in adult female, 205 mm. Fins: Pectoral of 6 slender rays, the longest more than twice length of head, all branched distally, with one or more luminous swellings; in- serted far forward and very low, under opercle; pelvic 7, inserted slightly behind middle of length at about 29th myomere; dorsal 18 to 20; anal 22 to 26; dorsal and anal beginning at same vertical, but anal extending farther back; adipose present. Epidermal Grooves: There is a deep groove in the isthmus for the reception of the barbel, and one in the side for the pectoral fin. 116 Zoologica: New York Zoological Society [XXIV :6 Osteology : Parietals present; mesethmoid with lateral processes; post- temporal present; supracleithrum and cleithrum strong; upper coracoid rudimentary; lower coracoid large; upper arm of mesocoracoid rudimentary, lower arm large; actinosts 5; vertebrae about 54 (myomeres to end of anal about 56) ; anterior vertebrae unmodified, the first centrum articulating with skull. Coelomic Organs'. Stomach 26.5 of length of fish, not reaching pelvic origin; 2 pyloric caeca. Nearly ripe ovarian eggs, preserved in alcohol, measure .5 mm. in diameter. Sexual Dimorphism: Postorbital light organ well developed in male, atrophied in female. Size: The largest known specimen measures 205 mm. in length (225 mm. when freshly caught), and is a female near breeding condition; the next largest is a male 118 mm. long, which is immature; both were taken by the Bermuda Oceanographic Expeditions. Development: Larva and post-larva unknown. Adolescent with no traces of dorsal subdermal pigment blotches; barbel stem short; anterior filaments on barbel bulb few and short; filaments forming fringe on posterior comb of bulb more numerous than in adult; postorbital light organ of females gradually atrophying: In an adolescent measuring about 35 mm. the postorbital organ, although already covered with partly pigmented skin, shone through in the fresh specimen and was blue-white in color; in an older adolescent it was not visible externally, but a small organ, rolled downward, was found upon dissection; in the largest (adult) female, the organ is completely atrophied, leaving a gaping hole beneath the skin, well separated from the eye, surrounded only by muscle fibers. Sex cannot be determined by examination of the gonads of adolescents; we have stated that the specimens in question are females on the basis of their atrophying postorbitals. Viability: The large female lived for half an hour after reaching the laboratory. Chirostomias pliopterus Regan & Trewavas, 1930. (See also p. 111). Specimens Taken by the Bermuda Oceanographic Expeditions. 8 specimens; May to August, 1929 to 1931; 300 to 700 fathoms; from a cylinder of water 8 miles in diameter (5 to 13 miles south of Nonsuch Island, Bermuda), the center of which is at 32° 12' N. Lat., 64° 36' W. Long. ; standard lengths from 35 to 205 mm. Specimens Previously Recorded. 5 specimens; ca. 55 to 273 fathoms; eastern and western North Atlan- tic, between 20° and 44° N. Lat.; standard lengths from 33 to 115 mm. Description of Adult. With the characteristics of the genus. Development. Material: The Bermuda collection of Chirostomias pliopterus is di- vided as follows: 6 adolescents; 35 to 41 mm.; 300 to 700 fath. ; June, July; females. 1 transitional adolescent; 500 fath.; August; male. 1 adult; 500 fath.; August; female, near breeding condition. 1939] Beebe & Crane: Family Melanostomiatidae 117 All are typical representatives of their respective growth stages, (see pp. 000-000). The specific characters of the adolescents have already been given on page 000. Study Material. The following list gives the catalogue number, depth in fathoms, date, length and growth stage of each specimen of Chirostomias pliopterus taken by the Bermuda Oceanographic Expeditions. All were caught in the cyl- inder of water off the Bermuda coast described in Zoologica, Vol. XVI, No. 1, p. 5 and Vol. XX, No. 1 p. 1. No. 10,738; Net 194; 600 F.; June 20, 1929; 41 mm.; Adolescent. No. 10,956; Net 219; 700 F.; June 25, 1929; 35 mm.; Adolescent. No. 11,464; Net 297; 500 F.; July 13, 1929; 37 mm.; Adolescent. No. 11,752; Net 316; 600 F.; July 23, 1929; 40 mm.; Adolescent. No. 15,053; Net 587; 500 F.; May 17, 1930; 35 mm.; Adolescent. No. 21,259; Net 1071; 300 F.; July 10, 1931; 39 mm.; Adolescent. No. 22,029; Net 1143; 500 F.; Aug. 7, 1931; 118 mm.; Trans. Adolescent. No. 22,200; Net 1157; 500 F.; Aug. 10, 1931; 205 nun.; Adult. Synonymy and References. Chirostomias pliopterus : Regan & Trewavas, 1930, p. 54; pi. I, fig. 1; text-figs. 6B, 8B and 30. (5 specimens; 33 to 115 mm.; 150-1,000 m. wire; Atlantic between 20° and 40° N. Lat.). Beebe, 1933.1, p. 180. (Preliminary list of Bermuda specimens). Beebe, 1937, p. 199. (Preliminary list of Bermuda specimens). Chirostomias lucidimanus : Beebe, 1932.2, p. 52. (4 specimens from Bermuda included in the present account of C. pliopterus) . Genus Pachystomias Gunther, 1878. (See also pp. 70, 73, 82, 83, 86, 88, 90, 91, 99, 105, 106, 109). (Text-figs. 12, 17). General Discussion. Only 4 specimens belonging to this genus are known at the present time, including the single, 37 mm. example taken by the Bermuda Oceano- graphic Expeditions. Two species have been erected, P. microdon (Gunther, 1878), for the 215 mm. specimen taken by the Challenger off Australia, and P. atlanticus Regan & Trewavas, 1930, for the 165 mm. Dana specimen taken from the Caribbean Sea. The latter species is distinguished from P. microdon by the longer teeth, broader interorbital region and longer barbel. Differences in photophore and fin counts are very small, and would normally fall within the range of variation of one species. The third specimen, measuring 90 mm. and taken off Nova Scotia, is recorded by Roule & Angel, 1933 (p. 17), without comment, save that it is in poor condition; they refer it to P. microdon. The present young specimen taken off Bermuda differs from both the described species in a number of ways — the depth is less, head and eye both larger, snout longer, interorbital broader, basibranchial teeth fewer, grouping of serial photophores different, and barbel relatively much longer (1.3 times head instead of two-thirds of it). All except the last two characters are regular characteristics of immature melanostomiatids. Since 118 Zoologica: New York Zoological Society [XXIV :6 the grouping of the photophores is different even on the two sides of the present specimen, these distinctions cannot be called specifically im- portant. Finally, it is known that in some other genera the barbel grows relatively more rapidly than the standard length during adolescence, hence this peculiarity does not seem a basis for the establishment of a new species; also, it is likely that the delicate barbel is broken in the previously known specimens. It is probable that the Atlantic and Australian forms will prove to be conspecific when more material has been acquired. For the present, how- ever, we shall regard them as separate species, referring the Atlantic example recorded by Roule & Angel and our own specimen to P. atlanticus. Generic Characters. Color (from Murray’s observation on the living type specimen of P. microdon and notes by the present authors on a living immature P. atlan- ticus) : General color, velvety brownish-black; both antorbitals rose ( micro- don ) or greenish-yellow ( atlanticus ) ; postorbital red ( microdon ) ; serial photophores violet to red. Proportions (from the 2 largest known specimens) : Moderately elongate melanostomiatids with large, broad heads and large eyes; depth in length 5 (20%) ; head in length 4.5 to 4.7 (21% to 22.3%) ; eye in head 4; snout a little shorter than diameter of eye; snout to pelvic in length ca. 1.7 (58%); interorbital width 4 to 6 in head. Barbel : Simple, slender, tapering, apparently shorter than head in adult. Light Organs : A large mass of luminous tissue forming a cushion on each side of palate, and appearing externally as a small luminous patch anteriorly, in the usual position of an antorbital, and a much longer, spindle shaped organ immediately behind, falling directly beneath the eye. Regular postorbital organ small, well separated from eye. Serial photo- phores in groups with the following counts: ventral series, I-P 8 (1+2+- 2+3), the last group beginning between the diverging pairs of the pre- ceding, P-V 16 to 17 (in 2 or 3 groups) V-A 14 to 15 (9 to 10+5, the last 5 in a close-set series above or ending at the vent), A-C 9; lateral series O-V 17 to 18 (in 4 groups), the first 2 ascending obliquely from the isthmus, V-A 13 to 14 (in 2 or 3 groups) ending above vent. Teeth : Cleft of mouth slightly curved; jaws slender; premaxillary and mandible with rather small, slender, curved acute, unequal teeth, all fixed; maxillary with small oblique teeth; vomer toothless; 2 to 5 teeth on each palatine; well developed teeth on basibranchials ; single teeth present on gill-arches, arrangement not recorded. Text-figure 17. Pachystomias atlanticus. Transitional adolescent, standard length 37 mm. 1939] Beebe & Crane: Family Melanostomiatidae 119 Fins : Pectoral 5 to 6; pelvics 7, inserted a little behind middle of length, about equidistant from eye and base of caudal; dorsal 21 to 24; anal 25 to 27 ; dorsal and anal beginning at same vertical, but anal extend- ing farther back. Osteology : Parietals absent; mesethmoid with lateral processes; post- temporal absent. Coelomic Organs : Pyloric caeca represented by 2 pouches. Sexual Dimorphism: Unknown. Size and Development : Larva, post-larva and adolescent unknown; a specimen of 37 mm. is in transitional adolescence; the other known speci- mens measure 90, 165 and 215 mm., respectively. Pachysfomias atlantieus Regan & Trewavas, 1930. (See also p. 117). Specimen Taken by the Bermuda Oceanographic Expeditions. 1 specimen (Cat. No. 17,769, Net 836) ; September 3, 1930; 500 fathoms; from a cylinder of water 8 miles in diameter (5 to 13 miles south of Nonsuch Island, Bermuda), the center of which is at 32° 12' N. Lat., 64° 36' W. Long.; standard length 37 mm. Specimens Previously Recorded. 2 specimens; between 0 and 2,200 fathoms; Caribbean Sea and off Nova Scotia; standard lengths 90 and 165 mm. Specific Characters. (From the description of the type specimen, 165 mm. in length). With the characters of the genus. This species differs from the 215 mm. type specimen of P. microdon Gunther from Australia in the following particulars: teeth in jaws longer; 2 teeth, not 5, on each palatine; 2 pairs, not 5, on first basibranchial; 4 pairs, not 2, on second; interorbital breadth contained 5, not 6 times in length of head; barbel longer, two-thirds, not one-fourth, length of head; pectoral 6, not 5; dorsal 21, not 24; anal 25, not 27. Serial photophores with following counts: ventral series, I-P 8 (l+2+2+3), P-V 16 (4+12), V-A 15 (10+5), A-C 9; lateral series, O-V 18 (2+4+4+8), V-A 13 (9+4). Development. The transitional adolescent from Bermuda will be described in detail. Color in Life: Skin velvety brownish-black; both antorbital light organs yellow-green; postorbital organ blazing dark red; serial photophores purple. Luminescence not observed in dark-room, although all organs appeared fully capable of functioning. Barbel colorless, translucent. Measurements and Proportions: Total length 41 mm.; standard length 37 mm.; depth 5.3 mm. (in length 7 or 14.2%) ; depth in front of vertical fins 3.5 mm. (in length 10.6 or 9.4%) ; head 11.7 mm. (in length 3.1 or 32%) ; eye 3.1 mm. (in head 3.8 or 8.4% of length) ; snout 3.1 mm. (in head 3.8 or 8.4% of length) ; interorbital 2.5 (in head 4.7 or 6.8% of length) ; 120 Zoologica: Neiv York Zoological Society [XXIV :6 interocular 4.8 mm. (in head 2.45 or 13% of length) ; snout to pelvic 20 mm. (in length 1.85 or 54%) ; anterior orbital light 1 mm.; median cheek light 2.5 mm.; posterior cheek light 1.1 mm.; barbel 15.5 mm. (1.3 times head, 42% of length). The outstanding characters of the fish are the very large, thick head, much thicker than the body, and the large, protuberant eyes, which project 1.2 mm. from the side of the head. The frontal crests divide the interorbital space into thirds. Barbel : The barbel, contained 2.4 times in the length, is slender, unpig- mented and tapers gradually to a thread-like tip. Light Organs : The first of the three orbitals is directly beneath the lower anterior part of the eye, and is a small oval ; the second is two and one- half times as long, and slightly broader, starting under the posterior corner of the first, and ending just before the vertical from the posterior border of the orbit; the third, corresponding to the usual postorbital of melanostomi- atids, is very slightly larger than the first and is obliquely set behind the end of the second. Serial photophores with the following counts and groupings: ventral series, both sides of fish, I-P 8 (1+2+2+3), P-V 17 (5+5+7), V-A 14 (9+5), A-C 9; lateral series, O-V right side 18 (1+5+4+8), left side 18 (1+5+5+7), V-A right side 14 (6+3+5), left side 14 (4+4+6). Numer- ous small organs are scattered over head and trunk. Teeth: The teeth, all fixed, are arranged as follows: 13 moderate-sized teeth in each premaxillary, the first, second, third and eighth somewhat enlarged, and the first, second and fifth external. There are about 20 small, oblique denticles on each maxillary; each mandibular ramus holds 16 or 17 teeth similar to those of the upper jaw, the first tooth a fang, the largest in the mouth, the third next largest and external ; there are one or two replacement teeth; 2 teeth on right palatine, 3 on left. One pair of teeth on first basibranchial, IV2 pairs on second. No teeth yet developed on gill- arches, but their anlagen appearing beneath the gill-arch skin. Fins : The pectorals are placed far forward on the body, beneath the posterior part of the maxillary, and are moderately short. The pelvics are well developed, originating behind the middle of the length and extending about two-thirds of the distance between their origin and that of the anal fin. The dorsal and anal have their proximal portions encased in a thick black membrane. Their rays are long, those in the posterior part of the fin reaching well beyond the caudal origin when laid back. Coelomic Organs: The stomach is fully pigmented, but short, 5.5 mm. long (15% of length of fish), and reaching only about three-fifths of distance between pectoral and pelvic insertions. There is no trace of the two rudimentary caeca described for adult fish of this genus by Regan & Trewavas (1930, p. 44). The gonads are slender, transparent threads. Viability: This specimen was alive when caught, but died a few minutes after its arrival in the laboratory, about one hour and a half after capture. This is the smallest deep-sea fish ever brought up alive by the Bermuda Expeditions. References and Synonymy. Pachystomias atlanticus : Regan & Trewavas, 1930, p. 70; pi. VI, fig. 1; (1 specimen; 3,500 m. wire; Caribbean Sea west of St. Lucia; standard length 165 mm.). Beebe, 1937, p. 199. (Preliminary listing of Bermuda specimen). Pachystomms microclon ( non Gunther) : Roule & Angel, 1933, p. 17. (1 specimen; 0 to 4,000 m. ; southeast of Nova Scotia; standard length 90 mm.). 1939] Beebe & Crane: Family Melanostomiatidae 121 Genus Leptostomias Gilbert, 1905. (See also pp. 70, 73, 75, 79, 80, 81, 85, 86, 88, 90, 97, 99, 101-103, 105,106,110). (Text-figs. 2, 5, 7, 11, 12, 18-22 incl.). General Discussion. Some authors have included the genus Leptostomias, with or without subgeneric standing, in Melanostomias. Regan and Trewavas, however, in their monograph on the family (1930) established the validity of the genus, and worked out its relationships. N ematostomias Zugmayer, 1911, is a synonym of Leptostomias. Eleven species of Melanostomiatidae are at present referable to Lep- tostomias, two, L. macropogon Norman and L. bermudensis Beebe, having been described since the publication of the above mentioned monograph. The delineation of all of the species is most unsatisfactory. Only two characters have been found to be specifically important, the number of P-V photophores and the form of the barbel; both are of dubious value, since the barbel has been found to be very variable in the only species, L. ramosus, of which more than four examples have hitherto been taken, and the number of photophores serves merely to separate groups of species; slight differences in finray counts and proportions cannot be regarded as useful until much more material is obtained. Finally, individuals of this genus do not mature until they attain a considerably greater length than others of the family, judging from the largest specimens we have been able to examine, namely, the holotypes of L. problematicus (Parr) and L. bermudensis Beebe, and the largest examples in the type series of L. longibarba Regan & Trewavas and L. ramosus Regan & Trewavas. These fish measure from 172 to 270 mm. in length, yet their gonads are so undeveloped as to be almost invisible in all but the 250 mm. specimen of L. longibarba, while the stomachs ai'e all short and distally unpigmented. Since only a single other large specimen has been recorded (the type of L. gladiator, 255 mm.), and the majority are much smaller, probably no fully adult Leptostomias has ever been taken. It is almost certain that a number of the species will prove to be synonymous, with the apparently specific differences being due instead to individual variation, age and, possibly, sex. For the convenience of future students, we include the following annotated list of species, although we can attempt no complete revision our- selves with available material, and feel justified at present in synonymizing only L. gladiator, L. problematicus and L. ramosus. 1. L. macronema Gilbert, 1905, p. 607. Known only from the type specimen, 74 mm. long, from Hawaii. We have examined it in the U. S. National Museum and, in addition to the pair of filaments at the base of the barbel figured by Gilbert and described by Parr (1927, p. 49), we have found a pair of simple, short filaments on the stem near the bulb, and a single, unpaired one just above the bulb. The bulb itself is somewhat damaged, but, as both Gilbert and Parr have observed, is apparently simple; the tip is narrowed, almost papilla-like. There is no question, however, in spite of the presence of these hitherto unnoticed filaments, that the species, which is the only one so far reported from the Pacific, is distinct from the known Atlantic forms. 2. L. gladiator (Zugmayer, 1911.1, p. 76). Hitherto known from 2 specimens from the eastern Atlantic in the Monaco collection, the 255 mm. type and a 70 mm. example recorded by Route & Angel (1933, p. 17). Re- ferred originally to the genus N ematostomias , erected for it; to Melanos- tomias by both Parr (subgenus Leptostomias ) (1927) and Roule & Angel; correctly placed in Leptostomias by Regan & Trewavas, 1930, p. 61. Not seen 122 Zoologica: New York Zoological Society [XXIV :6 by us, but there seems to be no reason for not considering L. problematicus and L. ramosus synonymous with it. The type is a male, judging from the postorbital organ “of moderate size,” described and figured in the type description. (See p. 127). 3. L. problematicus (Parr, 1927, p. 46). Known only from the type specimen, 172 mm. long, from the western Atlantic. We have examined it in the Peabody Museum and find that the barbel appears as figured, save for the following details: there are a number of short, unbranched filaments scattered along the stem, distal to the large basal filament described and figured in the type description. The tip of the barbel bulb is certainly broken off, so that at least one distal papilla is missing, and possibly a good-sized bit of the bulb itself. One of the distal papillae is not at the extreme end, as shown in the figure. Altogether, in spite of a slightly longer barbel, slimmer form, and 1 or 2 more dorsal and anal rays, it appears certain that this form and L. ramosus (species no. 9, below) are identical, and that both of them should be synonymized under L. gladiator (species no. 2, above). Distally the barbel stem has lost a considerable amount of skin, taking, per- haps, one or more filaments with it. The postorbital is almost completely atrophied externally, and though it is impossible to determine the sex from the condition of the internal organs, due to their immaturity, the specimen is doubtless a female, judging from the atrophy of the postorbital organ. There are 21, not 20, V-A photophores in both lateral and ventral series. There is a small pair of teeth in the vomer. Placed originally in Melanos- tomias, subgenus Leptostomias. 4. L. haplocaulus Regan & Trewavas, 1930, p. 59. Known from a single specimen from the North Atlantic, 100 mm. long. Not seen by us. 5. L. gracilis Regan & Trewavas, 1930, p. 59. Known from the 4 speci- mens in the type series from the North Atlantic, 70 to 75 mm. long. Not seen by us. 6. L. longibarba Regan & Trewavas, 1930, p. 60. Known from the 3 specimens in the type series from the North Atlantic, 53 to 250 mm. long. A 250 mm. immature female examined by us. 7. L. leptobolus Regan & Trewavas, 1930, p. 60. Known from the 2 specimens in the type series, 68 and 95 mm. long, from the Noi’th Atlantic. Not seen by us. 8. L. analis Regan & Trewavas, 1930, p. 61. Known from a single specimen, 168 mm. long, from the North Atlantic. Not seen by us. 9. L. ramosus Regan & Trewavas, 1930, p. 61. Known from 12 speci- mens in the type series from the North Atlantic, 56 to 180 mm. long. We have examined three, including the largest. All are immature, but females, judging from the atrophying postorbital organs. The variability of the barbel is adequate proof that this species and L. problematicus should be synonymized with L. gladiator. 10. L. macropogon Norman, 1930, p. 311. Known from a single speci- men from the South Atlantic, 165 mm. long. Norman suggests it may prove to be identical with L. gracilis. Not seen by us. 11. L. bermudensis Beebe, 1932, p. 59. Known from a single specimen from Bermuda, 270 mm. long, in the present collection. May prove to be synonymous with L. longibarba, L. gracilis or L. macropogon. (See p. 125). Distribution : Leptostomias is one of the seven genera in the family which have been recorded outside the Atlantic Ocean. L. macronema is known only from Hawaii, and L. macropogon only from the South Atlantic. All the rest have been taken in the North Atlantic. The depth range of the genus as known at present is undefined, between 0 and 1,500 fathoms. Includ- ing the present series, 49 specimens have been recorded, 35 of which are referred to Leptostomias ramosus. 1939] Beebe & Crane: Family Melanostomiatidae 123 Generic Characters. (The immaturity of all specimens from which these characters are de- rived must be kept in mind). Color (from fresh specimens of L. gladiator and L. bermudensis ) : General color dark brownish-black; barbel bulb yellow; serial photophores violet to maroon. Proportions : Very elongate melanostomiatids with short heads. Depth in length 10 to 17 (5.9% to 10%) ; head in length 7.5 to 11 (9.1% to 13.3%) ; eye in head 4.5 to 6; snout longer than eye; snout to pelvic in length ca. 1.5 to 1.6 (62.5% to 66.5%). Barbel : One-fourth to as long as fish; stem with or without filaments; bulb an elongate oval from which arise short filaments and papillae in vary- ing numbers, combinations and arrangements. Light Organs : Postorbital probably completely atrophied in adult female, moderate in male ; serial photophores with the following counts : ventral series, I-P 10 to 11, P-V 39 to 48, V-A 20 to 23, A-C 11 to 14; lateral series, O-V 39 to 48, V-A 20 to 23. Accessory light organs inconspicuous. Teeth : Cleft of mouth straight, jaws massive; premaxillaries and mandible with a few moderate fangs, each tapering to a point, mostly fixed ; first premaxillary tooth moderate, fixed; second long and depressible, fol- lowed by several smaller fixed outer teeth and 1 or 2 inner, depressible teeth; maxillary with a few small erect teeth followed by a series of about 20 to 30 oblique denticles. Mandible with a strong, fixed fang followed by a depres- sible tooth and several smaller fixed teeth. A pair of teeth on the vomer and 0 to 1 on each palatine; 1 to 2 pairs (sometimes rudimentary) on the basibranchials ; short, stout teeth present in pairs, three’s and four’s on first 4 gill-arches: on first and second hypohals, first 4 ceratohyals, and first 3 epihyals; about 9 groups present on first ceratobranchial, the last being elongate and single. Branchiostegal Rays : 17 to 19. Fins: Pectoral with 10 to 11 simple, moderately short rays; pelvics 7, placed well behind middle of length ; dorsal 16 to 21 ; anal 20 to 28, beginning below or very slightly in advance of dorsal and extending farther back. Epidermal Grooves : A pronounced groove in the isthmus for grasping the stem of the barbel. Osteology : Premaxillary without median process, but with well devel- oped lateral flanges instead; parietals present; mesethmoid with lateral processes; post-temporal present (at least in both Bermuda species), but weak, not attached to skull; supracleithrum and cleithrum strong; upper coracoid moderate, lower large; mesocoracoid with rudimentary upper arm but well developed lower arm; vertebrae about 75 to 83; first 7 vertebrae highly specialized, their centra being absent, and there being only 6 parapophyses ; first neural arch represented by a pair of small bones articulating with exoccipital; second neural arch much enlarged; third through eighth smaller, but larger than succeeding ones, and directed straight upward, instead of backward. Coelomic Organs : Stomach 19% to 23% of length, not nearly reaching pelvic fins ; distal portion lightly and incompletely pigmented in all specimens examined; gonads rudimentary. Intestine with an anterior pouch and a single caecum. Sexual Dimorphism : Postorbital light organ completely atrophied in females of 250 mm. and over; male postorbital of moderate size, judging from type description and illustration of L. gladiator (Zugmayer 1911 nl III, fig. 5). Size : The largest known specimen of Leptostomias is L. bermudensis, taken by the Bermuda Expeditions, and measuring 270 mm. long (285 mm. when fresh). The gonads are so immature as to be barely distinguish- 124 Zoologica: Neiv York Zoological Society [XXIV :6 hyphyi#2 brr pill arches' i:hypbr 2: cerbr 3:epibr h: pharbr Text-figure 18. Leptostomias bermudensis. Jaws, hyoid and branchial arches, and pectoral girdle of transitional adolescent, standard length 270 mm. Depressible jaw-teeth unshaded; divisions between palatine and pterygoid, and maxillary and supramaxillary not shown ; fourth basibranchial, always unossified, unshaded ; act, actinost ; basibrs, basibranchials ; brr, branchiostegal ray ; cerbr, ceratobranchial ; c.erhy, ceratohyal ; clt, cleithrum ; cor, coracoid ; dn, dentary ; epibr, epibranchial ; epihy, epihyal ; gloshy, glossohyal ; hypbr, hypobranchial ; hyphy, hypohyal ; inky, interhyal ; mo,x, maxillary ; pal-ptg, palato-pterygoid ; pharbr, pharyngobranchial ; premax, premaxillary; ptm, post-temporal; supclt, supracleithrum ; urohy, urohyal ; vo, vomer. Beebe & Crane: Family Melanostomiatidae 125 1939 Text-figure 19. Leptostomias bermudensis. Supporting bones of pectoral fin. From the tran- sitional adolescent holotype, standard length 270 mm. Abbreviations as in Text- fig. 14. able in this specimen, in the 180 mm. Dana example of L. ramosus, in the 170 mm. type of L. problematicus and in all smaller examples examined; these organs are only slightly more developed in a 250 mm. specimen of L. longibarba, although eggs are distinctly visible. Metamorphosis also occurs when the young are larger than usual (between 30 and 50 mm., instead of between 20 and 30). The indications are, therefore, that members of this genus attain relatively large size. Development : A series of growth stages of L. gladiator has been taken by the Bermuda Expeditions, which is probably typical of the genus. The characteristics in brief are, a large number of myomeres (ca. 75 to 78) between nape and end of anal (the count is probably greater in other species with higher photophore counts) ; pigment in a row of blotches immediately below the dorsal mid-line, and from 3 to 10 rows of compact pigment spots, the number of rows depending on the length of the fish; anal fin beginning slightly behind dorsal origin. Leptostomias bermudensis Beebe, 1932. (See also p. 122) . Type. (The unique specimen). Department of Tropical Research No. 20,826; Bermuda Oceanographic Expeditions of the New York Zoological Society; Net 1015;. June 15, 1931; 7% miles southeast of Nonsuch Island, Bermuda; 500 fathoms; standard length 285 mm. when fresh, 270 mm. after being pi'eserved in alcohol. Description. With the characteristics of the genus. Color (from fresh specimen) : General color dark, brownish-black; barbel stem proximally pigmented, distally lilac; barbel bulb bright, clear picric yellow, filaments white. Photophores of ventral series maroon with gilt caps, lateral series pale purple, also with gilt caps; small, non-serial photophores pale purple. 126 Zoologica: New York Zoological Society Posterior Aspect [XXIV :6 Leptostomias bermudensis. End of barbel in holotype, standard length 270 mm. Measurements and Proportions (at present): Total length 281 mm.; standard length 270 mm.; depth 17 (in length 15.8 or 6.2%) ; head 26 (in length 10.4 or 9.6%) ; eye 4.3 (in head 6, or 1.6% of length) ; snout 7.2 (in head 3.6 or 2.7% of length) ; upper jaw 17 (in head 1.5, in length 15.9 or 6.3%) ; snout to pelvic 176 (in length 1.53 or 65.4%). Barbel : Length 200 mm. (1.35 in length or 74%). Stem without fila- ments near base, but with at least three short ones, far separated from each other, in the distal three-fourths of stem. The latter is black for about 4/5 of the proximal portion, then this pales and grays, and changes into brilliant lilac with a dark core running through it. The bulb, in the fresh specimen, was abruptly bright, clear, picric yellow, and the filaments trans- lucent white with a scattering of black specks. The bulb arises abruptly from the stem, the lilac and the dark center ceasing at once. The bulb is slender, slightly curved, tapers gently from its center, and resembles in shape a diminutive cucumber. It narrows abruptly near the distal end, forming an elongate, rounded, terminal stem. There are three short, thin, median filaments given off, one from the back of the stem, and the other two from the proximal part of the bulb. Halfway down the bulb a pair of larger filaments arises, one from each side. Still farther a single one appears from the right side and, at the point of narrowing into the terminal stem, there arises a final pair of filaments, the longest of all, about 4 mm. in length. Light Organs : The postorbital is atrophied. Serial photophores: ven- tral series, I-P 10, P-V 48, V-A 21, A-C 12; O-V 47 to 48, V-A 22. Teeth: Premaxillary with 6 moderately large teeth, the second largest, the second and fourth depressible; maxillary with 5 to 6 tiny erect teeth followed by about 21 oblique denticles on the left side.; right side with addi- tional denticles discernible above the erect maxillary teeth; mandible with a moderately large fixed fang followed in turn by a slightly smaller depres- sible fang and 3 still smaller fixed teeth. Vomer with 1 pair of teeth; palatine toothless; 1 pair of rudimentary teeth on basibranchials. Branchiostegal Rays: 19. Fins: Pectoral rays 11 (not 12 as in type description), 12 mm. long; pelvic rays 7 ; 33 mm. long ; dorsal rays 20, anal rays 25, commencing oppo- site dorsal and continuing slightly behind it. Discussion. This species closely resembles L. mncropogon Norman and L. longibarba Regan & Trewavas, but it may be distinguished from both by the structure of the bulb of the barbel and by the presence of 48 P-V photophores. As has been said, however, it is extremely likely that, when additional material is secured, some of the described species will prove synonymous; the present species will perhaps be included among them. 1939] Beebe & Crane: Family Melanostomiatidae 127 References. Leptostomias bermudensis: Beebe, 1932.2, p. 59, fig. 10. (Type description, slightly amended and amplified in the present paper) . Beebe, 1937, p. 199. (Record of above specimen in preliminary Bermuda list) , Leptostomias gladiator (Zugmayer, 1911). (See also pp. 121-122). Specimens Taken by the Bermuda Oceanographic Expeditions. 20 specimens; May to September, 1929 to 1931; 50 to 1,000 fathoms; from a cylinder of water 8 miles in diameter (5 to 13 miles south of Nonsuch Island, Bermuda), the center of which is at 32° 12' N. Lat., 64° 36' W. Long.; standard lengths from 12 to 52 mm. Specimens Previously Recorded. 15 specimens; 0 to 2,680 fathoms; North Atlantic between 20° and 40°; standard lengths from 56 to 180 mm. Description. With the characteristics of the genus. Color (from fresh, immature, Bermuda specimens) : General color blackish-brown; barbel bulb creamy yellow; photophores violet. Proportions: Depth in length 10 to 13 (7.7% to 10%) ; head in length 7.5 to 8.5 (11.8% to 13.4%) ; eye in head 5 to 6 (16.7% to 20%). Barbel: About % length of fish; stem black with luminous patches; 1 or 2 pairs of filaments and usually an unpaired one at base; others scattered variably and irregularly along stem, including a pair slightly above bulb; bulb unpigmented, more than half length of head, slender, scarcely com- pressed, usually with narrowed, papilla-like tip and bearing 1 or 2 pairs of stout filaments near the base and 1 to 4 short, swollen filaments, often a pair near the tip and another a little behind them. Light Organs : Serial photophores with the following counts : ventral series, I-P 10 to 11, P-V 39 to 43, V-A 21 to 22, the last 4 being above the anal fin, A-C 11 to 12; lateral series O-V 39 to 42, V-A 20 to 22. Fins: Dorsal 19 to 22; anal 23 to 26. Development. Material: The Bermuda collection of Leptostomias gladiator is com- posed entirely of immature fish, a well-graduated series ranging from young larvae to transitional adolescents and measuring between 12 and 52 mm. in length. They are distributed as follows: 9 larvae; 12 to 30 mm.; 50 to 1,000 fath. ; May to Sept. 8 post-larvae; 38 to 45 mm.; 300 to 1,000 fath.; June to Aug. 1 adolescent; 45 mm.; 500 fath.; July. 2 transitional adolescents; 50, 52 mm.; 600, 1,000 fath.; Sept. All are typical representatives of their respective growth stages (see pp. 76-79). Their special characteristics are as follows: Myomere Counts: To end of anal 75 to 78; from nape to pelvic rudi- ment (when present) 42 to 44; from pelvic rudiment to anal origin 16 to 18. Pigment : Larval pigment is traceable subdermally even in early transi- tional adolescence; it reaches its most advanced condition not in the larva 128 Zoologica: New York Zoological Society [XXIV :6 A Text-figure 21. Leptostomias gladiator. A, larva, standard length 12 mm.; B, post-larva, 42 mm.; C, adolescent, 45 mm.; D, transitional adolescent, 52 mm. See also Text-fig. 2, A and B. but in the post-larva. In the larva there are four major longitudinal rows on each side: (1). There is a series of stellate blotches along each side of the dorsal profile from nape to immediately beyond the vertical from the end of the anal fin; each spot occupies an entire myomere and almost fuses with its neighbors, so that the general effect is of a dorsal band of pigment. (2). Well below this, just above the lateral line, are small, compact chromatophores in a sometimes single, sometimes double, line; they extend from the last gill-arch to the vertical from the end of the dorsal fin. (3 and 4). Between the lateral line and the intestine, which lies beneath the myomeral body, are two rows of chromatophores, similar, but longer, lying on the lines of myomeral demarcation so that each spot in the upper row is slightly ahead of the corresponding one below. Anteriorly the majority are dendritic, and there is a small amount of pigment between the rows throughout the series. Small chromatophores are sparsely scattered on the crown of the head, on the posterior portion of the ventral finfold, on the 1939] Beebe & Crane: Family Melanostomiatidae 129 dorsal surface of the end of the gut and on the anal fin; the isthmus is rather densely pigmented. In the post-larva the 3 lateral rows of subdermal pigment spots are increased to 8 or 9, while the dorsal profile series of spots is supplemented by an irregular line of much smaller dots between and above them. All of the lateral spots follow the outlines of the myomeres, as in the larvae, and the main profile series still contains one blotch to each myomere. There is, in addition, a light general coating of pigment. In subsequent stages these chromatophores gradually disappear, the dorsal series remaining longest. In our 50 and 52 mm. transitional adolescents the dorsal row is still visible externally through the skin, and, when the skin is removed, this row alone is sharply marked; the lateral rows are traceable only as dull blotches. Leptostomias gladiator. Typical barbel of late transitional adolescent (after Regan & Trewavas). Larval Teeth : Premaxillary with 7 pairs of teeth all directed straight forward; the maxillary holds 18 teeth, all erect; they are minute and close- set anteriorly, but increase in size posteriorly with progressively larger spaces between; in each half of the mandible are 12 teeth, in the front of the jaw only, placed close together; they are all directed straight outward and increase in size posteriorly. Larval Gill-rakers : Long, spiny rakers present on first 3 arches, and low, spiny mounds on last 2 ; 8 or 9 rakers on first ceratobranchial. Fins: Dorsal and anal rays not of complete number in larva, the anal beginning under the second third of the dorsal, instead of beneath its origin or slightly in advance, as in mature fish. Finfolds moderately deep. Study Material. The following list gives the catalogue number, depth, in fathoms, date, length and growth stage of each specimen of Leptostomias gladiator taken by the Bermuda Oceanographic Expeditions. All were caught in the cylinder of water off the Bermuda coast described in Zoologica, Vol. XVI, No. 1, p. 5 and Vol. XX, No. 1, p. 1. No. 10,197; Net 131; 800 F.; May 27, 1929; 13 mm.; Larva. No. 10,481; Net 171; 600 F.; June 6, 1929; 41 mm.; Post-larva. No. 10,992; Net 218; 600 F.; June 25, 1929; 45 mm.; Post-larva. No. 11,751; Net 316; 600 F.; July 23, 1929; 45 mm.; Post-larva. No. 11,942; Net 341; 700 F.; July 30, 1929; 39, 42 mm.; Post-larvae. No. 13,105a; Net 424; 600 F.; Sept. 5, 1929; 50 mm.; Trans. Adolescent. No. 13,396; Net 456; 1,000 F.; Sept. 10, 1929; 52 mm.; Trans. Adolescent. No. 13,888; Net 518; 1,000 F.; Sept. 28, 1929; 30 mm.; Larva. No. 16,174; Net 725; 500 F.; June 26, 1930; 18 mm.; Larva. No. 16,928; Net 776; 500 F.; July 5, 1930; 45 mm.; Adolescent. No. 16,865; Net 794; 800 F.; July 9, 1930; 13 mm.; Larva. No. 17,743; Net 834; 400 F.; Sept. 4, 1930; 14 mm.; Larva. No. 20,863; Net 1021; 600 F.; June 16, 1931; 42 mm.; Post-larva. No. 21,316; Net 1075; 50 F.; July 11, 1931; 23 mm.; Larva. No. 21,314; Net 1077; 300 F.; July 11, 1931; 38 mm.; Post-larva. No. 21,340; Net 1079; 50 F.; July 14, 1931; 12, 16 mm.; Larvae. No. 21,508; Net 1095; 600 F.; July 24, 1931; 20 mm.; Larva. No. 24,052; Net 1209; 1,000 F.; Aug. 20, 1931; 45 mm.; Post-larva. 130 Zoologica: Neiv York Zoological Society [XXIV :6 Synonomy and References. Nematostomias gladiator: Zugmayer, 1911.1, p. 76; pi. Ill, fig. 5. (1 specimen; 270 mm.; 4,900-0 m. ; eastern North Atlantic). Melanostomias problematicus : Parr, 1927, p. 46, figs. 26, 27, 28 A. (1 specimen; 172 mm.; 7,000 ft. wire; Bahamas). Examined by present authors. Melanostomias gladiator: Parr, 1927, p. 48, fig. 28 (Resume of type description of N. gladiator). Roule & Angel, 1933, p. 17. (1 specimen; 70 mm.; 0-250 m. ; Monaco Deep) . Fowler, 1936, p. 210. (Resume of type description of N. gladiator). Leptostomias problematicus : Regan & Trewavas, 1930, p. 61, fig. 41A. (Resume of type description of M. problematicus). Leptostomias gladiator : Regan & Trewavas, 1930, p. 61, fig. 41B. (Resume of type description of M. gladiator) . Leptostomias ramosus : Regan & Trewavas, 1930, p. 61, figs. 10B, 11C, 12B, 42. (12 specimens, 56 to 180 mm.; 150 to 1,000 m. wire; north Atlantic between 20° and 40° N. Lat.). Several specimens, including the largest, examined by present authors. Beebe, 1933.1, p. 180. (Preliminary list of Bermuda specimens). Beebe, 1937, p. 199. (Preliminary list of Bermuda specimens). Genus Echiostoma Lowe, 1843. (See also pp. 70, 73, 75, 79, 80, 82, 84-86, 88-91, 96, 99, 102, 105, 106, 108, 110). (Text-figs. 2, 11, 12, 23-27 inch). General Discussion. Five species of melanostomiatids properly referable to Echiostoma have been described, namely, E. barbatum Lowe, 1843; E. tanneri Gill, 1883; E. ctenobarba Parr, 1927; E. guentheri Regan & Trewavas, 1930; and E. callio- barba Parr, 1934. In addition, Parr described a subspecies, E. ctenobarba ramifera, in 1934. The type specimens of E. barbatum and E. guentheri differ from the others in having a single, unbranched, terminal appendage at the distal end of the barbel. From each other they differ chiefly in that E. guentheri has longer, more numerous and more extensive stem filaments and a more swollen bulb. Specimens referred to E. tanneri are distinguished by the presence of two well-developed bulbs and very long stem filaments. E. ctenobarba and E. calliobarba were erected for specimens each having only one terminal bulb, or none at all, and relatively short stem filaments. E. calliobarba and the subspecies of E. ctenobarba were differentiated on the basis of details of the extent and length of the lateral filaments, of the branching of the terminal filaments, and of the apparent presence or absence of a whitish body near the end of the barbel, and of the relative size of the postorbital light organ. As our series of barbels shows (Text-fig. 26), all of the barbel characters merge into one other, and may logically be laid to 1939] Beebe & Crane: Family Melanostomiatidae 131 individual variation, and to the different lengths at which individuals lose their juvenile characters. Variation, and, to a small extent, sexual dimorph- ism, is also found in the relative length of the postorbital organ. The Bermuda specimens divided themselves very readily into this ctenobarba-calliobarba form and into typical E. tanneri. When the internal organs were examined, however, it was found that every one of the 8 cteno- barba-calliobarba- like examples (measuring between 268 and 355 mm. in length) was an adult near breeding condition, while the 5 typical specimens of E. tanneri (between 60 and 195 mm. in length) were transitional adolescents with scarcely developed gonads and short, partially pigmented stomachs. Furthermore, our 268 mm. specimen, the smallest that could be called an adult, showed distinct remains of 2 barbel bulbs and stem filaments intermediate in length between those of the largest tanneri example and the other adults. Similarly, the type and other specimens of E. tanneri at the U. S. National Museum are all small specimens measuring under 200 mm., and the largest Dana example of the species is recorded as being 223 mm. long. In addition, Dr. Parr has permitted us to open his series of E. cteno- barba and E. calliobarba, measuring 275 to 297 mm., with the result that they prove to have adult characters. In view of the identity of proportions, counts, teeth and osteological characters (allowing only for differences, such as number of maxillary teeth, which are definite age characters), we think it unquestionable that these three species are identical, and hence synonymize them under E. tan- neri, the oldest name. An exactly similar case of a reduction of the barbel bulb in adults is found in Photonectes margarita, (see p. 177), and of the reduction of lateral filaments in Chirostomias pliopterus (p. 111). By analogy, it seems certain that E. guentheri Regan & Trewavas, 1930, is a young specimen of E. barbatum Lowe, 1843. E. microdon Gunther, 1878 is the genotype of Pachystomias; E. richardi Zugmayer, 1911, and E. margarita Goode & Bean, 1895, belong in the genus Photonectes (see p. 155). Distribution-. E. barbatum (including E. guentheri) is known only from 2 specimens, both taken at Madeira. E. tanneri, of which more than 100 specimens have been captured, occurs in both north and south Atlantic, between about 65 and 959 fathoms. Generic Characters. Color (summarized from observations on 6 living adult specimens of E. tanneri taken by the Bermuda Oceanographic Expeditions; postorbital light organ color of E. barbatum recorded by Lowe, 1843, rosy red) : Gen- eral color blackish-brown. Postorbital rosy red anteriorly, white posteriorly. End of barbel bulb pinkish. Serial and non-serial photophores violet to scar- let; longitudinal luminous bands, bluish-white. Proportions : Moderately elongate melanostomiatids; depth in length 5.7 to 8 (12.5% to 17.5%) ; head in length 6.2 to 7 (14.3% to 16%) ; eye in head 5 to 7 (2.1% to 2.6% of length) ; snout less than twice length of eye; snout to pelvic origin in length 1.7 to 1.8 (56% to 59%). Barbel : Shorter than head with a row of filamentous or papilliform proc- esses (sometimes almost atrophied) on each side of distal part of stem; two bulbs well developed in young, almost or completely atrophied in adults ; one or more moderately thick terminal filaments. Light Organs : Postorbital (measured as length of area of transparent skin) about 1% to 2 times diameter of eye in both sexes, contained 3 to 5 times in head. Serial photophores with the following counts : ventral series, I-P 8+2, P-V 25 to 27, V-A 16 to 18, of which only the last one is above the anal fin, A-C 11 to 13; lateral series, O-V 24 to 26, V-A 16 to 18. Tiny, non-serial photophores well developed. 132 Zoologica: New York Zoological Society [XXIV :6 — ■ 0Ct i. _2^^_ Ay basibrs _ cerhy urohy. hyphy is 2 brr. pill arches 1: hypbr 2: cerbr 3: epibr pharbr Text-figure 23. Echiostoma tanneri. Jaws, hyoid and branchial arches, and pectoral girdle of adult, standard length 325 mm. Explanation as in Text-fig. 18. Teeth : Cleft of mouth straight. Premaxillaries and mandible with close-set, curved, bicuspid fangs, all depressible except 1 to 3 in each jaw; mandibular teeth in 2 or 3 rows posteriorly; maxillary with 5 to 10 erect teeth and a long series (up to 50) of oblique denticles, 1939] Beebe & Crane: Family M elanostomiatidae 133 I L Text-figure 24. Echiostoma tanneri. Supporting bones of pectoral fin in adult, standard length 325 mm. Abbreviations as in Text-fig. 14. the anterior ones usually placed above the last few erect teeth; a pair of teeth on the vomer; a series of 3 to 12 teeth on each palatine. Usually 4 pairs on the basibranchials. Teeth, practically all in pairs, present on first and second gill-arches only : on first and second ceratobranchials and on first epibranchial ; about 10 pairs on first ceratobranchial. Branchiostegal Rays: 13 to 15. Fins: Pectoral with 4 rays apparent externally; 1 isolated and pro- duced, and 3 short rays; cleared and stained specimens, however, show that the isolated ray is composed of two rays united by a common sheath ; in addi- tion there is a rudimentary, subdermal ray in front of the isolated pair; the fin is inserted far forward; under the opercle. Pelvic 8, inserted well behind the middle of the length at about the 32nd vertebra. Dorsal 12 to 16; anal 15 to 19; dorsal and anal beginning at same vertical, but anal extending farther back. Epidermal Grooves: There is a pronounced groove in the isthmus for the reception of the barbel. Osteology: Mesethmoid without lateral processes; frontal ridges and pterotic canal-and-pore system strongly developed, with superficial patches of ossification around nostrils and behind pterotics; parietals absent; post- temporal rudimentary; supra-cleithrum and cleithrum moderately strong; all coracoid elements large except upper arm of mesocoracoid ; actinosts 3; vertebrae 57 to 59 ; first vertebra represented only by a fibrous ring, shorter than a centrum, enclosing the notochord and by a spinal nerve. Coelomic Organs: Stomach 45% of standard length, reaching well be- yond pelvic oi’igin; 2 pyloric caeca. Apparently ripe ovarian eggs, pre- served in alcohol, measure .72 mm. in diameter. Sexual Dimorphism: Postorbital light organ of female, although well developed and functional, slightly smaller than that of male. Size: The largest known specimen measures 355 mm. in length (375 mm. long, weight 12 oz., when freshly caught), and is a female in or close to breeding condition, taken by the Bermuda Oceanographic Expeditions. 134 Zoologica: New York Zoological Society [XXIV :6 Development : Larva and post-larva unknown. Adolescent of E. tanneri with subdermal series of pigment spots as follows: Each myomere has, typically, a dorsal, dendritic blotch immediately below the dorsal mid-line, and an obliquely vertical row of 3 to 5 dots, following the myomeral boundary, between the lateral mid-line and the series of photophores. Adolescent and transitional adolescents have two large barbel bulbs which almost or com- pletely disappear in the adult; some of the barbel stem filaments similarly often partially or completely atrophy. Synopsis of the Species. The following key may be adopted: A. Barbel with a single, simple, terminal appendage E. barbatum. AA. Barbel with one or more compound terminal appendages. B. Barbel with two distinct bulbs; stem filaments well-developed E. tanneri, young. BB. Barbel without bulb, or with a single, small one; stem filaments reduced in length, and often also in number . . . E. tanneri, adult. Echiostoma tanneri (Gill, 1883). (See also p. 130) . Specimens Taken by the Bermuda Oceanographic Expeditions. 13 specimens; May to September, 1929 to 1931; 500 to 900 fathoms; from a cylinder of water 8 miles in diameter (5 to 13 miles south of Nonsuch Island, Bermuda), the center of which is at 32° 12' N. Lat., 64° 36' W. Long. ; standard lengths from 61 to 355 mm. Specimens Previously Recorded. More than 100 specimens; between ca. 65 and 959 fathoms; north and south Atlantic; standard lengths from 20 to 297 mm. Description of Adult. With the characteristics of the genus. Barbel with one or more, com- pound, terminal filaments. Papillae or filaments on stem numbering from one to about a dozen pairs. Three longitudinal bands of luminous material running the entire length of the body from opercle to tail; one, the broadest, immediately below the dorsal mid-line; one, narrow, between the lateral 1939] Beebe & Crane: Family Melanostomiatidae 135 and ventral series of photophores; and one, also narrow, between the ventral series and the ventral mid-line. Color, Luminescence and Behavior : These three topics are so closely connected that it is advisable to treat them together. Five adults, including both males and females, and the largest transitional adolescent (192 mm. long) were captured alive and, with the aid of the refrigerator and, in two cases of adrenalin injections, lived between two and six hours after reaching the laboratory. In the dark-room an ultra-violet lamp was used to aid in the observation of luminous areas. Several fish were in especially good condition, twisting and snapping continually during the entire time they remained alive. Between periods of observation, the fish were placed in the refrigerator, which always revived them greatly. Excellent moving pictures were made. In the dark-room the fish gave a most wonderful display of lights, from which the following deductions were made. In general, direct correlation was found between the color of the various organs in daylight, and the luminescence given off by them. General Skin Color: In daylight, blackish-brown. A broad band of whitish tissue along each side of upper part of body, and two duller, narrow stripes on lower part of side (see below). Barbel: The barbels of adults showed very little color, only faint washes of pink through the unpigmented portions. On the other hand, the barbel bulbs of all of the transitional adolescents taken were emerald green with a buffy sheath, while the stem and terminal filaments were violet blue tipped with bright mallow purple. No luminescence was observed from this organ in young or adults, although the least touch, or even a slight stirring of the water near the barbel, would arouse the fish to the utmost, so that it thrashed about and snapped, striving to reach and bite the source of irri- tation. Again and again we proved the astonishing sensitiveness of this organ. Obviously the barbel is primarily an organ for detecting vibrations in the water. Postorbital Light Organ : In the adults, in daylight, the upper anterior portion of this light, and sometimes the entire anterior half, was invariably bright phlox pink and the remainder creamy white, the whole organ having a glistening, waxy look. In the dark room, however, both with and without the use of the ultra-violet lamp, this organ in two cases gleamed dully with a pinkish glow. In two other individuals it gave forth strong bluish or bluish-white flashes, sometimes at long intervals, sometimes almost between winks, both with and without stimulation. In the living transitional adolescent, on the other hand, while the anterior portion of the organ was pink, as in adults, the posterior was dis- tinctly silvery green, instead of creamy white. Whenever the fish twisted and turned and snapped, the cheek lights blazed out. Eight times this hap- pened and eight times there shot forth a strong, clear, greenish-white light which momentarily lighted up all our faces. Twice we saw a distinct rosy or deep pink light from the same organ. As two-thirds of the photophore is pink in color, there must be an extremely delicate and localized control of the area, and of the color of the light. The cheek light did not revolve down into its socket, although this move- ment is possible, in order to obliterate or interrupt a steady gleam, but, as in all the other body lights, the luminescence was thoroughly under the control of the fish. The overhanging, pigmented “eave” of the organ pro- tects the eye from the direct glare. When viewed from directly above, both cheek lights could sometimes be seen full on at once. In one case, the most powerful flashes occurred immediately before death. Serial and Non-serial Photophores: The organs in daylight were violet in the young with silver or gold caps, and scarlet in the adults. The luminescence in all, however, was rosy to deep scarlet, deepest in the adult. 136 Zoologica: New York Zoological Society [XXIV :6 Text-figure 26. Echiostoma tanneri. Lateral views of barbels. A, transitional adolescent, standard length 39 mm. ; B, same, 61 mm.; C, same, 116 mm. (female) ; D, same, 192 mm. (male) ; E, F, G, H, adult males, 268 mm., 273 mm., 285 mm. and 302 mm., respectively ; I, J, K, adult females, 315 mm., 333 mm. and 375 mm., respectively. A, after Regan & Trewavas ; others from specimens in present collection. A and B, X 9.6 ; C-K, inch, X 3-9. 1939] Beebe & Crane: Family M elanostomiatidae 137 " ■ — A Echiostoma tanneri. Diagrams of barbels, anterior views. A, adolescent and early transitional adolescent; B, late transitional adolescent; C, adult. In the adult the barbel becomes literally reduced in size and is actually smaller than that of moderate-sized transitional adolescents (see Text-fig. 26, D and F). The small cephalic and trunk photophores of all were rosy red to deep purple. Every light was directed downward. See p. 138 for a description of the distribution of these small organs. Luminous Tissue: The body light seemed to be of two kinds, illumina- tion of the photophores, which we could see with a hand lens placed close to the surface of the water, and a general illumination of the whole skin by a greenish-white glow, although in the young the actual back was always dark, without a trace of the dorsal luminous bands observed in adults. When the adults were placed under the fluorescent screen of the ultra- violet lamp, the luminescent areas were very distinct. Down the middle of the back two widely separated lines of bluish-white extended from nape to tail. These were broad and very irregular, narrowing and expanding and sending out short lines at irregular intervals. Below these were two addi- tional pairs of the same type of line, but much narrower and more linear; one pair extended between the lateral and ventral series of photophores, and the second pair close together, on each side of the ventral profile. Over the cheek were scattered small, irregular blotches formed of grayish-white. All of the fins were unquestionably luminous, and all the rays were per- meated with scarlet blood vessels, the corpuscles moving regularly. Even the short webs of the pectoral were luminous, while the long, isolated ray was shining bluish-white, except for an external, broad band of skin which was, in daylight, brown with a series of a dozen or more photophores on its proximal portions; the second or third of these photophores was at least twice as large as the others. The membranes connecting the three short rays contained one or two layers of white, oval, luminous granules ; down each ray were scattered about a dozen small glands, probably giving out mucous. Touching or moving the pectorals drew no apparent response from the fish, although a touch on the side of the body, especially near the tail, usually aroused an immediate reaction. The bases of the teeth appeared pale blue in daylight, and were faintly luminous in the dark-room. 138 Zoologica: New York Zoological Society [XXIV :6 One specimen gave out a brilliant flash, at least a third as bright as that from the postorbital organ, from an undetermined spot near the begin- ning of the anal fin. Special Activity Observations'. All of the living specimens were vigor- ous, swimming strongly and snapping until a short while before death. In the most lively, the breathing was at a rate of 2 to 3 respirations a second; when feeble, only half as fast. The mouth was not seen to close, and the gill openings were never quite shut, the gills protruding beyond the opercle. Viewed from above, when the fish was swimming normally upright, the gill covers were seen to open widely with all the gills showing expanded and deep red, being much more visible than is the case with any ordinary, adult fish. When adrenalin was injected, the point of puncture became luminous at once, and little by little the illumination — -yellowish in this single case — spread along the side. The small organs even along the dorsal profile became distinctly luminous and the hyoid line especially so, while the round granules along the short pectoral rays and all of the pelvic rays shone out clearly. (See Harvey, 1931, p. 67 ff.). In the light of the above experiences with this species, Gill’s only specific description of the fish, although not very scientific, was exceedingly vivid and quite characteristic: “a black fish with formidable teeth, which was so lively when brought to the surface that it twisted itself around in its attempt to bite the commander of the vessel, Captain Tanner.” Arrangement of Small, Non-serial Photophores: An adult male, in or near breeding condition, and measuring 325 mm. in length, was depig- mentized in potassium hydroxide solution for about 36 hours, so that even the smallest organs were clearly visible in the now pallid skin. Since the exact arrangement and distribution of these lights has never been described, the following detailed account is given : Head: In addition to the minute, apparently atrophied, suborbital and the large postorbital photophores, the light organs of the head are of four types : 1. On each side there is an irregular, broken line of about 14 small, dark organs, each about a third the size of a nostril; the line extends, just above the level of the eye, from the nostrils to the upper base of the opercular flap. The separate organs of the two sides correspond to one another in position. 2. The second type, forming the majority of the cephalic photophores, consists of organs also dark, but much smaller, the largest being not more than a fourth as large as the preceding. They are scattered without discern- ible pattern over the sides of the snout, along the margins of both upper and lower jaws, on the inside of the maxillary, on the roof of the mouth at least as far back as the palatines, around the eye in a close-massed ring (except for a short section of the anterior upper margin), on the cheeks and on the opercles. The organs are densest along the snout and mandible, scarcest in a small area on the posterior portion of the cheek, and completely absent on the top of the snout and head between the two series of larger organs (type 1, above). There are altogether about 300 of these small organs on each side of the head, not counting those on the inside of the mouth. They vary slightly in size and are not found in exactly the same relative position on the two sides. 3. The third variety is infinitesimal in size and only visible to the naked eye as vertical, pale golden streaks foi’med by the collection of innum- erable organs into irregular lines of varying length, the individuals in each line being set very close together. These are found all over the head, jaws and opercles. 4. Finally, there are numbers of small, whitish luminous patches scat- 1939] Beebe & Crane: Family Melanostomiatidae 139 tered over the whole head (with the exception of the cheeks), jaws and opercles. These are thickest and most conspicuous on the snout, between the eyes and around the mandible. Trunk Organs: (Above the regular, lateral series of photophores) : The organs of the trunk may be divided into the same groups as those of the head, except that there are none corresponding in size to those of the first group described above. Those corresponding to the second group (small, dark, and visible to the naked eye) are arranged in about sixty vertical series which extend from the opercles to the caudal, and from about a fifth of the distance from the dorsal mid-line to immediately above the lateral series of photophores — one series to each myomere. Each series consists of one or two dorsal photo- phores separated by a considerable gap from the 10 to 20 organs below them. The latter are sometimes arranged in a single, unevenly spaced line, but more often in an irregularly double row. The number in each series decreases toward the tail, and it is there that single rows are dominant. The smallest, light-colored organs, as on the head, are innumerable, and in general are confined to outlining the myomeres, although there are a few odd ones scattered in the interspaces, and along the back. The luminous patches seem confined to the broad, longitudinal band described earlier in this account (p. 137). Ventral Organs: The same groups are found as on the trunk. The small, dark organs are closely massed in a roughly double row down the midline of the isthmus. Between the lateral and ventral rows of serial photophores they are extremely numerous and quite irregular, the only semblance of arrangement being in the semi-circular arch of six or seven organs above each photophore of the ventral series. Each of the serial photophores is surrounded by a clear space. The smallest, light-colored organs are comparatively few, and are grouped into short, broken lines similar to those on the head and without apparent pattern, except that a line of them runs between each two photo- phores. Two bands of the small, whitish luminous patches present below lateral and ventral series, respectively, as described on p. 000. Fins: A single row of small, dark organs, the second or third at least twice as large as the others, extends out along the isolated ray of the pectoral fin for a distance at least equal to the length of the head. The webbings of all the fins contain one or more layers of whitish, luminous granules. It is probable that the smallest organs described above on both head and body are not strictly photophores, but glands giving off luminous mucous. Development. The Bermuda collection contains no very young specimens, the youngest being five transitional adolescents measuring between 61 and 192 mm., and differing from adults in the usual details typical of transitional adolescence (see p. 000). Their own particular characteristics are the pronounced de- velopment of 2 barbel bulbs and the great number and length of the stem fila- ments, which pass through a stage when they are literally longer than in the adult. Also, there is an actual shrinking in length of the whole barbel after the specimen becomes adult. The remaining specimens in the collection consist of 5 males between 268 and 325 mm. in length, and 3 females between 315 and 355 mm. All may be counted as adults, since the gonads are very well developed, those of the 2 largest males and 2 largest females being apparently in full breeding condition. The smallest counted as an adult (268 mm.) shows unmistakable connections in the form of the barbel with the 140 Zoologica: New York Zoological Society [XXIV :6 double-bulbed transitional adolescents (Text-fig. 26E). We were able to de- termine sex in all specimens of 115 mm. and over. The Dana collection fortunately contains specimens as small as 20 mm. in length, and Regan & Trewavas (1930, p. 117) give the following details of the growth of barbels and maxillary teeth in adolescents and transitional adolescents: postocular small (from about 1/20 of head, or half diameter of eye) ; no filaments above barbel bulb in specimens under 90 mm., or, between 75 and 90 mm., the distal pair may be represented by buds; 2 or 3 pairs of short filaments in specimens of 95 and 96 mm.; at 133 mm. there are 3 or 4 long filaments on each side and 4 to 6 buds; in larger specimens, 142 to 153 mm., there are 12 or 13 pairs of filaments, and in the 223 mm. specimen 15 to 18, several being branched. Thus the maximum barbel devel- opment occurs during late transitional adolescence. It will be noted that in adults it is the more distal stem filaments which persist; often vestiges of the more proximal pairs can be detected subdermally. Maxillary with a few oblique teeth, 6 to 9 in examples of 35 mm.; in specimens of 75 to 95 mm. there are 3 or 4 fangs and 15 oblique teeth; in specimens of 130 to 223 mm., 6 or 7 fangs and 18 to 25 oblique teeth. As has been said, in our largest female (355 mm.) there are more than 50 oblique maxillary teeth. Thanks to the kindness of Dr. Norman, we have been able to examine two of the smallest of the Dana series from the British Museum, and to determine the presence of subdermal larval pigment spots. Each myomere is marked by one large stellate blotch immediately below the dorsal mid-line, and an oblique row of 3 to 5 small dots, following the myomeral boundary, between the lateral mid-line and the lateral series of photophores. (Text-fig. 2H). Ecology. Vertical and Season Distribution : The 13 specimens were taken singly from May to September between 500 and 900 fathoms. Specimens apparently in full breeding condition were taken in July and August. Food : Only two of the stomachs or intestines showed any food. The first, a specimen of 115 mm., contained one fish eye and a piece of crustacean cuticulum; the second, 154 mm. long, held a 67 mm. Lampanyctus polyphotis Beebe, described in 1932 and previously known from the Bermuda type alone, a specimen only 40 mm. in length. Since six of the specimens were alive and active after trawling, it is reasonable to suppose that at least some of the others may have lived for a time in the net; this would have allowed time for digestion of food before death, which seems a better explanation for the lack of food than the alterna- tive one that these active, well-armed fishes feed only at long intervals. This is especially likely since the abdominal wall is very thick and not as greatly distensible as in some forms, such as Chiasmodon, which can swallow such enormous fish that the food supply thus obtained is presumably adequate for long periods. Enemies : A specimen of Echiostoma has been taken from the stomach of a swordfish ( Xiphias gladius ) (Kingsley, 1922, Science, N. S., Vol. LVI, pp. 225-226). Parasites : Several small round worms, probably nematodes, were usual- ly present in the intestine. Study Material. The following list gives the catalogue number, depth in fathoms, date, length and growth stage of each specimen of Echiostoma tanneri taken by the Bermuda Oceanographic Expeditions. All were caught in the cylinder of water off the Bermuda coast described in Zoologica, Vol. XVI, No. 1, p. 5 and Vol. XX, No. 1, p. 1. 1939] Beebe & Crane: Family Melanostomiatidae 141 No. 10,125; Net 116; 900 F.; May 18, 1929; 315 mm.; Adult Female. No. 10,882; Net 212; 600 F.; June 24, 1929; 302 mm.; Adult Male. No. 11,183; Net 245; 800 F.; July 1, 1929; 333 mm.; Adult Female. No. 12,976; Net 412; 800 F.; Sept. 3, 1929; 154 mm.; Trans. Adolescent Male. No. 15,054; Net 587; 500 F.; May 17, 1930; 115 mm.; Trans. Adolescent Female. No. 15,651; Net 657; 700 F.; June 2, 1930; 192 mm.; Trans. Adolescent Male. No. 17,792; Net 815; 900 F.; Aug. 28, 1930; 285 mm.; Adult Male. No. 20,141; Net 839; 700 F.; Sept. 3, 1930; 273 mm.; Adult Male. No. 21,603; Net 1102; 500 F.; July 25, 1931; 325 mm.; Adult Male. No. 22,528; Net 1194; 700 F.; Aug. 18, 1931; 82 mm.; Trans. Adolescent. No. 22,559; Net 1194; 700 F.; Aug. 18, 1931; 355 mm.; Adult Female. No. 22,798; Net 1228; 500 F.; Aug. 27, 1931; 61 mm.; Trans. Adolescent. No. 22,974; Net 1243; 700 F.; Aug. 31, 1931; 268 mm.; Adult Male. Synonymy and References. Hyper choristus tanneri : Gill, 1883, p. 256. (1 specimen; 959 fathoms; 660 miles northwest of Bermuda) . Examined by present authors. Ecliiostoma barbatum : Goode & Bean, 1895, p. 109; pi. XXXV, fig. 130. (“Numerous speci- mens;” east of New Jersey and Old Bahama Channel). Several examples examined by present authors. Parr, 1927, p. 53, fig. 31. (3 specimens, 28 to 48 mm.; 800, 8,000 ft. wire; south and southeast of Nassau, Bahamas). Examined by present authors. Borodin, 1931, p. 65 {part.) (1 female, 255 mm.; 1,500 m. ; off Ber- muda). Examined by present authors. Echiostoma tanneri : Regan & Trewavas, 1930, p. 117; fig. 113. (92 specimens; 20 to 223 mm.; 50 to 2,000 m. wire; Gulf of Mexico, Caribbean and North Atlantic). 2 small examples examined by present authors. Norman, 1930, p. 314. (2 specimens; 170, 200 mm.; 850 to 950 m. wire; off Cape Town, South Africa). Beebe, 1937, p. 199. (Preliminary note on specimens treated in the present paper). Echiostoma ctenobarba : Parr, 1927, p. 55, figs. 32 and 33. (1 male; 285 mm.; 4,000 to 7,000 ft. wire; Bahamas, southeast of Nassau). Examined by present authors. Regan & Trewavas, 1930, p. 117, fig. 112c. (Resume of type descrip- tion). Harvey, 1931, p. 67. (Results of stimulation by adrenalin of 2 speci- mens in the present collection). Parr, 1934, p. 16, fig. 4a. (Supplementary description of the type speci- men) . Echiostoma ctenobarba ramifera : Parr, 1934, p. 17, fig. 4b. (1 female; 297 mm.; 1,050 to 1,100 m. ; off Bahamas). Examined by present authors. Echiostoma calliobara : Parr, 1934, p. 15, fig. 4c. (1 female; 290 mm.; 610 to 930 m. ; off Azores) . The two small specimens (70 to 75 mm.) described by Borodin (1931, p. 65) under the heading Echiostoma barbatum prove, upon examination by us, to be Photonectes margarita (see p. 000). 142 Zoologica: New York Zoological Society [XXIV :6 Genus Melanostomias Brauer, 1902. (See also pp. 70, 72, 73, 75, 79, 81, 82, 85, 86, 88, 90, 91, 97, 102, 103, 105, 106, 108-110). Text-figs. 2, 10, 11, 12, 28-32 incl.). General Discussion. Nineteen species of Melanostomiatidae have been referred in original descriptions to the genus Melanostomias. In addition, the two species for which Regan & Trewavas erected the genus Haplostomias (1930, p. 109) should also be referred to Melanostomias, as will be shown below. Hence a total of 21 species of Melanostomias have been described. In addition to our study of the specimens taken by the Bermuda Oceanographic Expeditions, all of which prove to be M. spilorhynchus and M. biseriatus, we have examined all of the specimens of Melanostomias de- posited in American museums, as well as two on loan from the British Museum. Difficulties in the delineation of species are greater than usual in this genus, because only 2 characters have been found to be specifically sig- nificant, namely, the number of P-V photophores, which usually serves only to separate groups of species, and the form of the barbel. The latter is one of the last organs to achieve adult shape, since it often continues development far into transitional adolescence; also, it is frequently variable. We are convinced that many of the so-called specific differences between barbels in Melanostomias will prove to be due to growth stage characteristics and to individual variation ; it is also likely that sex may be a controlling factor in barbel form, as in the genus Eustomias. Unfortunately, however, an ade- quate revision of the genus is impossible until additional material has been acquired, particularly fully adult examples, and specimens in European museums examined. The following annotated list of species described up to the present may be of help to future investigators. 1. M. melanops Brauer, 1902, p. 284. Six specimens have hitherto been referred to this species: the 183 mm. type from the Indian Ocean, 4 small specimens from the Bahamas (Parr, 1927, p. 42), and 1 specimen, 242 mm. long from the Caribbean (Regan & Trewavas, 1930, p. 114). We have examined Parr’s material and decided that his nos. 2,066 and 2,067 are rightly referred to M. melanops, while the other two are M. melanopogon (species no. 15 below). In addition, M. albibarba (species no. 13), of which we have examined one of the type series, is in all probability the transi- tional adolescent phase of M. melanops. 2. M. valdiviae Brauer, 1902, p. 285. Three specimens have hitherto been referred to this species: 2 from the Indian Ocean (55 and 165 mm.), and one taken by the Dana in the Caribbean (24 mm.). Not seen by us. Probably M. melanocaulus, M. heteropogon, M. stewarti and M. vierecki (species nos. 10, 11, 18 and 20 below) will prove to be synonymous with M. valdiviae. If they are thus synonymized, a hitherto unrecorded female, 101 mm. long, in the U. S. National Museum taken by the Albatross off Brazil (U. S. N. M. No. 2,761) also belongs to this species. 3. M. braueri Zugmayer, 1913, p. 3. Already rightly referred to the genus Photonectes by Regan & Trewavas, 1930 p. 121. 4. M. niger Gilchrist & von Bonde, 1924, p. 6. Known only from the type specimen, 220 mm. long, from South Africa. Not seen by us. 5. M. bartonbeani Parr, 1927, p. 45. Described from a specimen in the U. S. National Museum in which the specifically important barbel is broken off above the bulb. We have examined it, and find that it is impossible to tell whether it is conspecific with M. spilorhynchus, or with M. valdiviae, or is actually a different species. 1939] Beebe & Crane: Family Melanostomiatidae 143 6. M. problematicus Parr, 1927, p. 46. Already rightly referred to the genus Leptostomias by Regan & Trewavas, 1930, p. 61. 7. M. tentaculatus (Regan & Trewavas, 1930, p. 109). It was for this and the following species that the genus Haplostomias was erected. Seven specimens have been referred to this species, 6 measuring between 20 and 100 mm. in the type series, from the North Atlantic and Caribbean, and one, 204 mm. long, from the South Atlantic (Norman, 1930, p. 314). We have examined the latter specimen. By definition, the newer genus differs from Melanostomias chiefly in having the fangs “simple, or with a rudimentary cusp;” the proportions, photophores and finray counts all fall within the limits set by typical species of Melanostomias; also, the barbel is of the same general type — mid-way, in fact, between the simple, terminal bulb of M. niger and the more complicated form, with luminous bodies before and behind the terminal axis, found in other species, since in the present species luminous bodies occur only behind the terminal axis. Our examination of the 204 mm. specimen shows that a number of the fangs have cusps considerably more pronounced than shown by Regan & Trewavas, p. 110, fig. 105 a, although smaller than in typical Melanostomias; also, the tips of a number of the fangs are obviously broken, as is often the case in bicuspid-fanged genera, so that it seems altogether probable that in adult specimens small cusps are present on all the teeth. Also, the number and arrangement of the teeth in both jaws and gill-arches are typical of those occurring in true Melanostomias. Finally, the proposed genus Haplostomias is so much closer to Melanostomias than to any other genus, and the dif- ferences so much slighter than those between any other two genera in this family, that the advisability of uniting them seems unquestionable. Hence, we propose to place Haplostomias in synonymy with Melanostomias so that its two species will become Melanostomias tentaculatus and M. bituberatus, respectively. 8. M. bituberatus (Regan & Trewavas, 1930, p. 110). See preceding species (no. 7). Known from a single specimen, 20 mm. long, taken in the tropical North Atlantic. Not seen by us. 9. M. spilorhynchus Regan & Trewavas, 1930, p. 112. All except 4 of the specimens in the Bermuda collection are referred without question to this species, the best known in the genus; it appeal’s to be the species typical of the subtropical, as opposed to the tropical Atlantic. A full discussion begins on p. 148. M. bulbosus Beebe, 1933, is a synonym of M. spilorhynchus (see species no. 17, below). 10. M. melanocaulus Regan & Trewavas, 1930, p. 113. Known only from the type specimen, 55 mm. long, from the Caribbean Sea. Not seen by us. It is likely that this will prove to be conspecific with M. valdiviae. 11. M. heteropogon Regan & Trewavas, 1930, p. 113. Known only from the 2 or 3 specimens in the type series, measuring up to 62 mm. long, from the tropical and subtropical west Atlantic, including a station near Bermuda. Not seen by us, but we think it most likely that this species will also prove to be a synonym of M. valdiviae. 12. M. biseriatus Regan & Trewavas, 1930, p. 113. Known only from 4 specimens, 20 to 25 mm. long, from a single station east of Bermuda, and from 4 Bermuda post-larvae and adolescents in the present collection (see p. 152). M. margaritifer (species no. 14, below) or an allied species will probably prove to be a more advanced stage of M. biseriatus. 13. M. albibarba Regan & Trewavas, 1930, p. 114. Known from 11 specimens, 20 to 60 mm. long, in the type series, chiefly from the tropical, rarely the subtropical, Atlantic. Regan & Trewavas also refer to this species two of the series identified by Parr (1927, p. 42, nos. 2064 and 2065) to M. melanops. We have examined the latter specimens, as well as one of the series described by Regan & Trewavas, and are fairly certain that M. albibarba represents merely the young of M. melanops. Our basis for this 144 Zoologica: New York Zoological Society [XXIV :6 conclusion is the fact that in small examples of M. spilorhynchus the rounded luminous bodies of the barbel end are much larger and more distinct than in adult specimens; the same is true of barbel bulbs in related genera (e.g. Echiostoma and Photonectes) . 14. M. margaritifer Regan & Trewavas, 1930, p. 115. Known only from the 2 specimens, 52 and 80 mm. long, in the type series, from the north Atlantic. Not seen by us. May prove to be a more advanced stage of M. biseriatus. 15. M. melanopogon Regan & Trewavas, 1930, p. 115. Known from 3 or 4 specimens, 27 to 153 mm. long, in the type series from the North Atlantic, and by 2 specimens, 66 and 95 mm. long, from the Bahamas. The latter examples were recorded by Parr (1927, p. 42, nos. 2066 and 2067), who referred them to M. melanops. We have examined no. 2067 (66 mm. long), and find that without question it should be referred to M. melanopo- gon, as suggested by Regan & Trewavas. 16. M. macrophotus Regan & Trewavas, 1930, p. 115. Known from the 9 specimens, in the type series, 20 to 62 mm. long, from the north Atlantic. Not seen by us. 17. M. bulbosus Beebe, 1933.2, p. 166. A reexamination of the type, the unique specimen, shows that it is in reality an example of M. spilor- hynchus in which the barbel has been broken between the distal end of the pigmented swelling and the luminous, terminal expansion. The lateral series of V-A photophores actually numbers about 13, as in typical spilor- hynchus, instead of 11, as stated in the type description, the first 2 or 3 in the series being rudimentary in this specimen on the left side of the fish, which is in all other respects normal. 18. M. stewarti Fowler, 1934, p. 262. Known from a single specimen, 215 mm. long, from the Philippine Islands. After examining this form in the U. S. National Museum, we conclude that M. stewarti should be synony- mized with M. valdiviae. Fowler differentiates it from M. valdiviae chiefly because of the presence of apparently only one pair of teeth, not 2, on the basibranchials. We find, however, that there is a small second pair under- neath the skin. The barbel agrees with the figure given by Regan & Tre- wavas (1930, p. 112, fig. 108 A), except that the posterior luminous bodies are relatively larger with respect to the anterior one, and there is a minute, second, anterior body immediately in front of the tip of the axis. Finally, there are 5, not 4, pectoral rays. The specimen is somewhat shriveled, the measurements not entirely agreeing with those given in the description. 19. M. globulifer Fowler, 1934, p. 263. Known from a single specimen, 180 mm. long, from the Philippine Islands. Examined by us in the U. S. National Museum, and found to be valid, as far as present knowledge goes. Except for the tiny ovoid white bodies along the stem the barbel is, how- ever, exceedingly close to those of M. valdiviae and M. melanocaulus (see Regan & Trewavas, 1930, p. 112, fig. 108 A and C) : at the distal end of the barbel there are 2 luminous bodies behind the axis, the proximal slightly the longer, and 1 in front, short but very slender. Contrary to the statement in the type description, there is a small terminal filament, arising as usual from the posterior translucent portion. Another correction is that when the V-A series is counted, in the customary fashion, to include the 3 to 5 organs which continue the row above the anal fin, there are 13, not 11 organs in the ventral V-A series, and 10, not 13, in the A-C series; similarly, there are 12, not 10, in the lateral V-A series. We count 15, not 14, dorsal rays. 20. M. vierecki Fowler, 1934, p. 265. Known from a single specimen, 118 mm. long, from the Philippine Islands. After examining this form in the U. S. National Museum, we see no reason why it should not be referred to M. valdiviae. The barbel is exactly as figured by Regan & Trewavas (1930, p. 112, fig. 108 A), except that the posterior bulbs are relatively larger, as in M. stewarti, the distal being considerably larger than the proximal; 1939] Beebe & Crane: Family Melanostomiatidae 145 there is a short, posterior filament in the usual position. There is no trace of a tiny, anterior, terminal body, such as is found in M. stewarti. There are 5, not 4, pectoral rays. Counting the photophores in the customary man- ner gives the following results, perfectly concordant with the counts for M. valdiviae : ventral series, I-P 8+2 or 3, P-V 27 or 26, (depending upon in which series a subpectoral organ is counted), V-A 12, A-C 10; lateral series as given in the type description (O-V 25, V-A 12). The specimen is some- what shriveled, and the present measurements do not entirely agree with those given in the description. 21. M. pauciradius Matusubara, 1938, p. 39. Known from a single specimen, 98 mm. long, taken off Japan. Not seen by us. Apparently valid. Distribution : Melanostomias is one of the 7 genera in the family which have been recorded outside the Atlantic Ocean. Two of the species, M. valdiviae and M. melanops, occur in both the Atlantic and Indian Oceans. Another, M. pauciradius, has been taken only off Japan. The depth range as known at present is between approximately 22 and 1,100 fathoms. Includ- ing the present collection, 160 specimens of Melanostomias have been taken, of which 99 are referred to M. spilorhynchus. Generic Characters. Color (summarized from observations on about 15 transitional adoles- cent and 2 adult male specimens of freshly caught Melanostomias spilorhyn- chus) : General color brownish-black; antorbital organ yellow; postorbital and snout lights pink to purple; end of barbel with greenish-yellow core, and pink or purple flanges and luminous bodies; serial photophores violet with gold caps; bases of teeth pale blue. Proportions 5: Elongate melanostomiatids with very short, rounded snout; depth in length 8.5 to 11 (9.1% to 11.8%) ; head in length 5 to 8 (12.5 % to 20%) ; eye in head 4 to 6; snout no longer than eye, sometimes shorter; snout to pelvic in length ca. 1.55 to 1.7 (59% to 67%). Barbel: Five-sixths to 3 times length of head, with a simple stem end- ing in an ovate bulb or a flattened, terminal expansion enclosing one or more luminous bodies; terminal filament present or absent. Light Organs : Postorbital (measured as length of area of transparent skin) ca. .75 to 1.5 times diameter of eye in both sexes. Serial photophores with the following counts: ventral series, I-P 8+1 to 3, P-V 23 to 29, V-A 12 to 15, of which 3 or 4 are above the anal fin, A-C 9 to 11 ; lateral series, O-V 23 to 28, V-A 12 to 15. Tiny, non-serial photophores well developed. Luminous matter in bases of teeth and on fins usually conspicuous. Teeth: Cleft of mouth straight, or slightly curved at symphysis; pre- maxillaries and mandible with depressible, curved, fangs arranged in a single row in 2 or 3 series, the teeth of each series progressing in size posteriorly; barbs usually strongly developed, sometimes weak; 1 or 2 pairs of small, fixed teeth in anterior part of each jaw; maxillary with about 3 to 9 erect teeth and a long series (up to about 45) of oblique denticles; a pair of teeth on the vomer; a series of 3 to 6 teeth on each palatine. Typi- cally 2 pairs of teeth on the basibranchials. Teeth, in pairs, with a few individual teeth single, and in threes and fours, present on first 3 gill-arches only: on first, second and third ceratobranchials and on first epibranchial; 9 or 10 groups on first ceratobranchial. Branchiostegal Rays: ca. 12 to 13. Fins: Pectoral with 3 to 6 rays, normal, short. Pelvic 7, much longer than pectoral, inserted far behind middle of length. Dorsal 13 (10?-M. niger) to 16; anal 16 to 20; dorsal and anal beginning at the same vertical, but anal extending farther back. 5 The proportions given include those of some obviously immature specimens upon which species have been founded. 146 Zoologica: New York Zoological Society [XXIV :6 VO" " pal-ptq pre m ax _fr — ^ \r supclt max - — . _clt hyphy i s2 u brr_ archest /’ i: hypbr 2: cerbr 3: epibr :pharbr Text-figure 28. Melanostoviias spilorhynchus. Jaws, hyoid and branchial arches, and pectoral girdle of adult, standard length 222 mm. Explanation as in Text-fig. 18. Epidermal Grooves : There is a shallow groove in the isthmus for the reception of the barbel stem. Osteology. Mesethmoid with lateral processes; parietals absent; post- 1939] Beebe & Crane: Family Melanostomiatidae 147 Text-figure 29. Melanostomias spilorhynchus. Supporting bones of pectoral fin in adult, standard length 222 mm. Abbreviations as in Text-fig. 14. temporal absent; supracleithrum and cleithrum moderately strong; all cora- coid elements large; actinosts 2; vertebrae about 50 to 55; first vertebra represented only by a fibrous ring, shorter than a centrum, enclosing the notochord, and by a spinal nerve. Coelomic Organs : Stomach ca. 40% of standard length, reaching almost to the pelvic origin; 2 pyloric caeca. Sexual Dimorphism : Apparently none, but should be watched for in structure of barbel, and for slight differences in relative size of postorbital. Size: The largest known specimen is an M. melanops 242 mm. long, taken by the Dana in the Caribbean. Examples of 4 or 5 other species, which measure more than 200 mm. in length, have been taken. Sex can be determined in transitional adolescents measuring 100 mm. or more. Male Bermuda specimens of M. spilorhynchus of 222 and 240 mm. appear to be adult, although not in full breeding condition. The same is true of the 204 mm. M. tentaculatus , also a male, in the British Museum. Development : A series of growth stages of M. spilorhynchus, from late larva to adult, has been taken by the Bermuda Expeditions; the pigment spots of the early stages are identical with those in post-larval and adoles- cent M. biseriatus and with those remaining subdermally in a borrowed Dana specimen of M. albibarba (probably young M. melanops — see p. 143) in the adolescent stage, and of those in young Echiostoma: each myomere has, typically, a dorsal spot immediately below the dorsal mid-line, and an obliquely vertical row of three dots, along or near the myomeral boundary, between the lateral mid-line and the series of photophores. The length of the barbel apparently increases with the length of the fish. 148 Zoological New York Zoological Society [XXIV :6 Melanostomias spilorhynchus Regan & Trewavas, 1930. (See also p. 143) . Specimens Taken by the Bermuda Oceanographic Expeditions. 51 specimens; May to September, 1929 to 1931; 400 to 1,000 fathoms; from a cylinder of water 8 miles in diameter (5 to 13 miles south of Nonsuch Island, Bermuda), the center of which is at 32° 12' N. Lat., 64° 36' W. Long. ; standard lengths from 17 to 240 mm. Specimens Previously Recorded. 48 specimens; ca. 18 to 275 fathoms; North Atlantic; Bermuda and eastward to Azores, Madeira and Canaries; standard lengths from 23 to 206 mm. Description of Adult. With the characteristics of the genus. This species is distinct among the members of the genus in having a characteristic barbel (see below), only 23 to 25 P-V photophores, and 3 luminous spots on the snout. Color (from observations on about 15 transitional adolescent and 2 adult males, all freshly caught) : General color brownish-black. Barbel stem and proximal part of swollen end black; usually 1 or 2 white spots at base of stem; barbel bulb with greenish-yellow core and flanges and luminous bodies phlox pink or purple. Antorbital pale yellow; postorbital and snout lights bright pink to true purple. Serial photophores bluish-violet or violet, with gilt caps. Luminous bases of teeth pale blue. Proportions : Depth in length 8.5 to 11 (9.1% to 11.8%); head in length 6 to 8 (12.5% to 16.7%); eye in head 5 to 6 (2.3% to 2.7% of length) ; snout as long as or slightly shorter than eye. Barbel: 1.2 to 1.5 times length of head with the flattened, terminal expansion enclosing the straight, central axis and two strips of loose, luminous tissue, one in front and one behind the axis; a small ovoid body at proximal end of anterior strip, a second at distal end of posterior strip, at the base of a tapering terminal appendage which usually ends in a fila- ment; a few speckles of pigment at distal end of axis. Stem black, with a row of photophores down posterior side; distal part of stem swollen im- mediately above junction with the terminal expansion. Light Organs : Antorbital tiny but apparently functional in transitional adolescent, atrophied in adult; postorbital 1.2 to 1.5 times diameter of eye. Serial photophores with the following counts: ventral series, I-P 8+2 or 3, P-V 22 to 25, V-A 13 to 14, the last 2 or 3 being above the anal fin, A-C 9 to 10; lateral series, O-V 22 to 24, V-A 13 to 14. The non-serial organs of both head and trunk are of two sizes, the first considerably smaller than the serial photophores but clearly visible to the naked eye, the second microscopic. A third variety of luminous organ is the luminous matter. All three kinds are as found in Echiostoma (see p. 138). Head: The photophores of the larger type (Group A) are scattered without special arrangement all over the head, except on the crown and interorbital region. They are most dense on the lower jaw. The second type (Group B) are everywhere, always arranged in short, irregular lines. On top of the head they run transversely; on the cheeks and opercles they con- tinue transversely (vertically), but less regularly and more obliquely. They are least conspicuous on the lower jaw. There is an unpaired luminous patch in the middle of the snout and a pair of similar ones, almost as large, one in front of each eye, in addition to about six pairs of smaller, but exactly 1939] Beebe & Crane: Family Melanostomiatidae 149 similar spots, in front of and below the eye and close to the front teeth; usually a conspicuous unpaired luminous patch in center of crown of head, behind the level of the eyes. Trunk Series (above the regular lateral photophore series) : In the middle of each myomere is a group of A-type organs in a roughly irregular double line. They commence on a level with the opercle and extend to the serial photophores. Between 20 and 30 lights are usually found in each myomeral group. The dorsal part of the fish lacks them entirely. The B-type group of microscopic organs outline in solid, single rows the lines of demarcation between myomeres. Short rows of them similar to those found on the head are scattered through the myomeres themselves. These organs continue without interruption across the dorsal profile. Ventral Series: A-type organs are massed without special arrangement on the isthmus. Midway between the lateral and ventral series runs a fairly continuous line, extending from opercles to the anal fin, dying out at this point. Above and below this line many other organs of the same type are distributed without apparent arrangement except that scallops are formed below each large, lateral light and above each large ventral one, a short tongue of the small lights thrusting up between each two serial organs. In the ventral midline this same scalloped formation is found but shows no special arrangement except that the lines of segmentation between the myomeres are devoid of organs. B-type organs are here distributed as on the trunk, between myomeres and in short strings. In general it may be noted that the concentration of organs from snout to caudal is ventral, and that the organs are sparser toward the caudal fin. On each side of the fish there are about 50 to 70 organs in the areas of deepest concentration. Fins: Pectoral 5 to 6; dorsal 14 to 15; anal 18 to 20. Text-figure 30. Melanostomias spilorhynchus. A, post-larva, standard length 24 mm.; B, adolescent, 25 mm.; C, adult, 222 mm. See also Text-fig. 2 D. 150 Zoologica: Neiv York Zoological Society [XXIV :6 Text-figure 31. Melanostomias spilorhynchus. Barbels, lateral views. A, post-larva, standard length 24 mm.; B, adolescent, 25 mm.; C, ti’ansitional adolescent, 36 mm.; D, adult, 222 mm. (end of barbel only). Development. Material : All stages are represented in the Bermuda collection, and are distributed as follows: 1 larva; 17 mm.; 500 fath. ; July. (In poor condition; identification certain from myomere counts and pigment, but impossible to take body measurements) . 12 post-larvae; 21 to 32 mm.; 400 to 1,000 fath.; July to Sept. 8 adolescents; 22 to 31 mm.; 500 to 900 fath. July to Sept. 27 transitional adolescents; 24 to 36 mm.; 400 to 1,000 fath.; Aug., Sept. 3 adults ; 222 to 240 mm. ; 500 to 900 mm. ; May to Sept. ; males. All are typical representatives of their respective growth stages (see pp. 77). The special characteristics of their young stages are as follows: Myomere Counts : To end of anal 49 to 52 ; from nape to pelvic rudiment (when present) 29 or 30; from pelvic rudiment to anal origin 10 to 12. 1939] Beebe & Crane: Family Melanostomiatidae 151 Pigment : The characteristic larval pigment spots of the genus (p. 147) are visible externally or subdermally even well into transitional adolescence. Larval Teeth: In each premaxillary of the single larva are 7 pairs of larval teeth, all minute, and all directed straight outwards; the maxillary holds 15 teeth, increasing slightly and progressively in size toward the pos- terior part of the jaw; in each half of the mandible are 7 teeth, correspond- ing to those of the premaxillary, directed outward and set in the anterior part of the jaw only. Larval Gill-rakers : Long, spiny rakers present on first 3 arches, vestigial or absent on fourth and fifth, continuing throughout post-larval stage; on the first ceratobranchial they number 11 or 12. Fins : Dorsal and anal rays of full number in larva, but in this stage and early post-larvae the anal commences slightly behind the dorsal, under about the fourth dorsal ray. The barbel of adolescents and transitional adolescents differs from that of the adults noticeably in the relatively large size and conspicuousness of the 2 ovoid, luminous bodies. The stem is only lightly pigmented, and the entire barbel length relatively less than in fully grown fish. The postorbital organ grows slowly, being smaller than the eye well into transitional adolescence. Ecology. Seasonal Distribution : The majority of specimens, practically all of which are young, having been taken in August and September, a summer breeding period may be indicated. Abundance : The 51 specimens in the collection indicate that this species is by far the most common melanostomiatid off Bermuda. About 1 in 5 of the melanostomiatids was an M. spilorhynchus, the rest being distributed among more than 30 species. Food : Of 12 stomachs examined, only 1, that of a transitional adolescent 34 mm. long, contained food, a Myctophum laternatum measuring more than half the length, and almost twice the thickness, of the Melanostomias. Study Material. The following list gives the catalogue number, depth in fathoms, date, length and growth stage of each specimen of Melanostomias spilorhynchus taken by the Bermuda Oceanographic Expeditions. All were caught in the cylinder of water off the Bermuda coast described in Zoologica, Vol. XVI, No. 1, p. 5 and Vol. XX, No. 1, p. 1. No. 10,235; Net 136; 700 F.; May 30, 1929; 222 mm.; Adult Male. No. 11,418; Net 277; 1,000 F.; July 9, 1929; 222 mm.; Adult Male. No. 11,867; Net 311; 600 F.; July 22, 1929; 25 mm.; Post-larva. No. 11,877; Net 315; 500 F.; July 23, 1929; 26 mm.; Post-larva. No. 11,919; Net 330; 900 F.; July 27, 1929; 25 mm.; Adolescent. No. 11,920; Net 334; 500 F.; July 29, 1929; 23 mm.; Post-larva. No. 11,936; Net 339; 500 F.; July 30, 1929; 25 mm.; Adolescent. No. 12,110; Net 355; 600 F.; Aug. 8, 1929; 31 mm.; Adolescent. No. 12,339; Net 373; 500 F.; Aug. 15, 1929; 30 mm.; Trans. Adolescent. No. 12,346; Net 374; 600 F. ; Aug. 15, 1929; 26 mm.; Trans. Adolescent. No. 12,403; Net 378; 500 F.; Aug. 16, 1929; 30 mm.; Trans. Adolescent. No. 12,464; Net 384; 500 F.; Aug. 17, 1929; 28, 28 mm.; Trans. Adolescents. No. 12,475; Net 385; 600 F.; Aug. 17, 1929; 35 mm.; Trans. Adolescent. No. 12,559; Net 391; 600 F.; Aug. 19, 1929; 25 mm.; Adolescent. No. 12,815; Net 397; 700 F.; Aug. 31, 1929; 36 mm.; Trans. Adolescent. No. 12,953; Net 410; 600 F. ; Sept. 3, 1929; 25, 26 mm.; Post-larva, Adolescent. No. 12,969; Net 412; 800 F.; Sept. 3, 1929; 26, 32 mm.; Trans. Adolescents. No. 13,049; Net 417; 600 F.; Sept. 4, 1929; 24, 32 mm.; Post-larva, Trans. Adolescent. No. 13,098; Net 423; 500 F.; Sept. 5, 1929; 27, 27 mm.; Trans. Adolescents. 152 Zoologica: New York Zoological Society [XXIV :6 No. 13,105; Net 424; 600 F.; Sept. 5, 1929; 28 mm.; Trans. Adolescent. No. 13,158; Net 430; 500 F.; Sept. 6, 1929; 28, 36 mm.; Trans. Adolescents. No. 13,211; Net 437; 500 F.; Sept. 7, 1929; 26, 31 mm.; Trans. Adolescents. No. 13,368; Net 451; 500 F.; Sept. 10, 1929; 22 mm.; Post-larva. No. 13,376; Net 452; 500 F.; Sept. 10, 1929; 23, 24 mm.; Post-larvae. No. 13,589; Net 481; 800 F.; Sept. 20, 1929; 26 mm.; Trans. Adolescent. No. 13,766; Net 500; 900 F.; Sept. 24, 1929; 24 mm.; Trans. Adolescent. No. 14,969; Net 574; 500 F.; May 14, 1930; 240 mm.; Adult Male. No. 17,189; Net 804; 500 F.; July 16, 1930; 17 mm.; Larva. No. 17,874; Net 842; 600 F.; Sept. 4, 1930; 34, 36 mm.; Trans. Adolescents. No. 20,140; Net 868; 900 F.; Sept. 10, 1930; 22 mm.; Adolescent. No. 18,349; Net 869; 1,000 F.; Sept. 10, 1930; 21 mm.; Post-larva. No. 18,494; Net 882; 700 F.; Sept. 12, 1930; 28 mm.; Trans. Adolescent. No. 21,676; Net 1113; 400 F.; July 29, 1931; 28 mm.; Post-larva. No. 22,307; Net 1170; 800 F.; Aug. 12, 1931; 26 mm.; Adolescent. No. 22,752; Net 1181; 600 F.; Aug. 15, 1931; 24 mm.; Adolescent. No. 22,696; Net 1187; 400 F.; Aug. 17, 1931; 27 mm.; Trans. Adolescent. No. 22,796; Net 1227; 400 F.; Aug. 27, 1931; 24, 24 mm.; Post-larvae. No. 24,053; Net 1228; 500 F.; Aug. 27, 1931; 29 mm.; Trans. Adolescent. No. 23,290; Net 1287; 1,000 F.; Sept. 10, 1931; 28 mm.; Trans. Adolescent. No. 23,307; Net 1291; 600 F.; Sept. 12, 1931; 29 mm.; Trans. Adolescent. No. 23,461; Net 1312; 400 F.; Sept. 16, 1931; 26 mm.; Trans. Adolescent. Synonymy and References. Melanostomias spilorhynchus : Regan & Trewavas, 1930, p. 112, fig. 107 ; pi. X, fig. 1. (48 specimens, the type series, 23 to 206 mm.; 65 to 1,000 m. wire; North Atlantic, Bermuda and eastward to Azores, Madeira and Canaries). Beebe, 1937, p. 199. (Preliminary list of Bermuda specimens). Melanostomias bulbosus : Beebe, 1933.2, p. 166. (Description of a single Bermuda specimen, synonymized with M. spilorhynchus in the present paper). Beebe, 1937, p. 199. (Record of the above specimen in a preliminary list of Bermuda specimens). Melanostomias biseriatus Regan & Trewavas, 1930. (See also p. 143). Specimens Taken by the Bermuda Oceanographic Expeditions. 4 specimens; September only, 1929; 500 to 900 fathoms; from a cylinder of water 8 miles in diameter (5 to 13 miles south of Nonsuch Island, Bermuda), the center of which is at 32° 12' N. Lat., 64° 36' W. Long.; standard lengths from 21 to 25 mm. Specimens Previously Recorded. 4 specimens; ca. 22 to 42 fathoms; north Atlantic, east of Bermuda; standard lengths from 20 to 25 mm. Description. (From the description of the type series, none of which can be more advanced than early transitional adolescence; serial photophore counts sup- plemented by those of our younger examples). Proportions : Depth in length about 10 (10%); head in length 5 to 6 (16.7% to 20%) ; eye in head 5. Barbel: About twice as long as head; proximal 2/5 pigmented; distal 1939] Beebe & Crane: Family Melanostomiatidae 153 part a white axis, with a narrow, translucent band in front and behind, each including a series of white luminous bodies; the posterior series starting proximally with a rather large white body, and ending with a larger one that curves the axis; beyond this white bulb and the end of the axis is an expansion without a filament. Light Organs: Postorbital longer that diameter of eye. Serial photo- phores with the following counts : ventral series, I-P 8+2 to 3, P-V 26 to 28, V-A 13 to 14, A-C 9 to 11: lateral series, O-V 26 to 27; V-A 12 to 13. Fins: Pectoral 5; dorsal 13 to 16; anal 16 to 17. Development. The 4 Bermuda specimens of M. biseriatus, though comparable in size to those of the type series, are obviously even younger, 2 (23, 25 mm.) being in the post-larval stage and 2 (21, 24 mm.) in the adolescent. They are typical of their stages in every way, (see p. 77), and their pigmentation is typical of the genus (see p. 147). They differ from corresponding stages of M. spilorhynchus in the following details: P-V and O-V photophores number 26 to 28, not 22 to 25 ; myomeres to pelvic origin number 33 to 34, not 29 to 30 ; myomeres to end of anal number about 55, instead of 49 to 52 ; the barbel of the adolescents shows clearly the specific characters of M. biseriatus distally (it differs in lacking stem pigment and in having the proximal luminous bodies very rudimentary and scarcely distinguishable, although the entire distal half of the stem is noticeably expanded; also, the barbel in the most advanced is only a little longer than the head, instead of twice the head length). The post-larvae, too, have the single, oblong, terminal luminous body, anteriorly directed end of barbel core, and non-filamented distal tissue. In this species the barbel seems to develop specific characters earlier and to grow faster than in M. spilorhynchus. The present post-larvae could be referred to some closely allied species with the same photophore counts, but they appear to form a continuous series with the adolescents in every particular. It is, of course, probable that M. biseriatus will prove to be the young of some other species, also described by Regan & Trewavas, such as M. margaritifer, in which case M. biseriatus will take precedence. Melanostomias biseriatus. Barbel, lateral view, of adolescent, standard length 24 mm. Ecology. It is interesting to note that all 4 specimens of M. biseriatus were taken in the single month of September, 1929. Study Material. The following list gives the catalogue number, depth in fathoms, date, length and growth stage of each specimen of Melanostomias biseriatus taken 154 Zoologica: New York Zoological Society [XXIV :6 by the Bermuda Oceanographic Expeditions. All were caught in the cylinder of water off the Bermuda coast described in Zoologica, Vol. XVI, No. 1, p. 5 and Vol. XX, No. 1, p. 1. No. 24,172; Net 444; 500 F.; Sept. 9, 1929; 23 mm.; Post-larva. No. 13,376a; Net 452; 500 F.; Sept. 10, 1929; 25 mm.; Post-larva. No. 13,758; Net 499; 800 F.; Sept. 24, 1929; 24 mm.; Adolescent. No. 13,772; Net 502; 900 F.; Sept. 24, 1929; 21 mm.; Adolescent. Reference. Melanostomias biseriatus : Regan & Trewavas, 1930, p. 113, fig. 109A. (4 specimens, the type series, 20 to 25 mm.; 80 and 150 m. wire; north Atlantic east of Bermuda). Genus Photonectes Gunther, 1887. (See also pp. 70, 73, 75, 79, 81-91, 96, 97, 99, 102, 103, 105, 106, 108, 110). (Text-figs. 2, 11, 12, 33-44 inch). General Discussion. The first species referable to this genus was P. albipinnis, described by Doderlein in 1882 (p. 26) for a specimen taken off Japan, for which he erected the genus Lucifer. Since the latter name was preoccupied, Gunther in the Challenger report substituted the name Photonectes. Afterwards, several species were originally referred wrongly to other genera, Photonectes braueri to Melanostomias, and P. richardi and P. margarita both to Echios- tomu. Parr (1927) and Regan & Trewavas (1930) in their monographs correctly relegated all of these to Photonectes. The genus as now understood includes some species, usually recognized as members of different subgenera, which are so diverse that we considered for a time separating Photonectes into at least two distinct genera. Study of the family as a whole, however, shows that all of the species of Photonectes are so much more closely related to each other than they are to even their closest generic relatives (e.g. Echiostoma, Melanostomias and Tactostoma ) that we agree with the above mentioned authors that only a single genus should be recognized, although the maintenance of two sub- genera is convenient. Nineteen species of Photonectes have been recorded, including the two new species taken by the Bermuda Expeditions, which have been described (Beebe, 1933.2) since the publication of the Regan & Trewavas monograph. As in the case of a number of other genera, we are convinced that some of these species represent merely growth stages or variations of others. We have been able to prove this in several species. The following annotated list, in chronological order, gives the standing of the species as far as known. We are as usual hampered by the immaturity of the majority of recorded specimens. 1. P. albipinnis (Doderlein, 1882, p. 26). Known only from the type specimen, 240 mm. long, from Japan. Not seen by us. 2. P. margarita (Goode & Bean, 1895, p. 109). Only the type speci- men, 320 mm. long, from the Gulf of Mexico and one, 340 mm. long, recorded by Borodin (1931, p. 66), from the western Atlantic, have been previously referred to this species. We have examined both, as well as the unique specimen of P. flagellatus (species no. 6, below), and several specimens of P. intermedius (species no. 7). All of these we have compared with our own representative series of specimens referable to these species, and with the descriptions of P. richardi (species no. 5) and P. monodactylus (species no. 17), with the result that we are certain that the differences between P. margarita, P. flagellatus, P. richardi and P. monodactylus are due solely to 1939] Beebe & Crane: Family Melanostomiatidae 155 size, individual variation, and damaged barbels, while P. intermedius repre- sents only the early transitional adolescent stage of the same species (see Text-fig. 44 for complete barbel series, and p. 175 for a description of the species and the details in regard to our examination of the specimens in other museums). P. margarita was referred originally to Echiostoma. 3. P. gracilis Goode & Bean, 1895, p. 112. Known only from the type, 167 mm. long, in the U. S. National Museum, and one, 170 mm. long, in the Peabody Museum, recorded by Parr (1927, p. 113). In both the barbel is broken off short. The type is temporarily not available, but we have ex- amined the second specimen, and find that only the extreme tip of the barbel is missing. In addition to the characteristic, conspicuous, horizontal band of metallic blue, there are also definite traces of similar markings in exactly the regions where they occur in P. achirus, P. caerulescens and P. mirabilis. P. gracilis is definitely set off from related species, however, by the forward position of the pelvic fins. In P. gracilis, the only available large example of the genus which does not belong to the subgenus Trachinostomias, the premaxillary and mandibular teeth are relatively small and even, exactly as in large P. margarita (subgenus Trachinostomias) , as opposed to the very uneven, relatively large, Melanostomias- like form of the same teeth in young transitional adolescents throughout the genus, (cf. Text-figs. 32 and 33). 4. P. braueri (Zugmayer, 1913, p. 3). Only the type, 115 mm. long, from the eastern Atlantic, has been previously referred to this species, which was originally placed in the genus Melanostomias. We have a smaller speci- men which, by the intermediate form of its barbel, teeth, etc., shows that the 11 specimens, 22 to 31 mm. long, of P. ovibarba (species no. 11) should also be referred to P. braueri. (For description, see p. 165). 5. P. richardi (Zugmayer, 1913, p. 4). Only the type, 170 mm. long, from the eastern Atlantic, has been referred to this species, which was originally placed in Echiostoma. Not seen by us, but from the descriptions we consider it identical with P. margarita (see species no. 2, above and p. 175). 6. P. flagellatus Parr, 1927, p. 107. Only the type, 280 mm. long, from the Bahamas has been referred to this species. We have examined it and consider it identical with P. margarita (see species no. 2, and p. 175). 7. P. intermedius Pari', 1927, p. 109. Previously known from the two specimens in the type series and 22 recorded by Regan & Trewavas (1930, p. 125), all measuring between 20 and 86 mm. in length from the north Atlantic. Unquestionably these represent the early transitional adolescent stage of P. margarita (see species no. 2 and p. 175). The type specimens, as well as one from the Dana collection and two from the Atlantis collection have been examined by us. 8. P. mirabilis Parr, 1927, p. 111. Known only from five transitional adolescents up to 60 mm. in length, including three in the present Bermuda collection. Type specimen examined by us. 9. P. dinema Regan & Trewavas, 1930, p. 120. Known from 36 speci- mens, in the type series and the present Bermuda collection, all young transi- tional adolescents measuring between 23 and 51 mm. in length, from the North Atlantic. (See p. 162). 10. P. leucospilus Regan & Trewavas, 1930, p. 121. Known from 14 specimens, in the type series and the present Bermuda collection, all young transitional adolescents measuring between 25 and 50 mm. in length; from the North Atlantic. (See p. 164). 11. P. ovibarba Regan & Trewavas, 1930, p. 121. Known only from the type series, 11 specimens, 22 to 31 mm. long, from the North Atlantic. We have examined one of the series and find that these represent unquestion- ably the young transitional adolescent stage of P. braueri (see species no. 4 and p. 166). 156 Zoologica: New York Zoological Society [XXIV :6 12. P. achirus Regan & Trewavas, 1930, p. 122. Known only from the type series, four specimens, 20 to 62 mm. long, from the western North Atlantic. We have examined the largest specimen and are sure that this series represents the young transitional adolescent stage of P. caerulescens (species no. 13), the barbel and barbel bulb being larger in these more juvenile forms, as in P. margarita and P. braueri (species nos. 2 and 4). 13. P. caerulescens Regan & Trewavas, 1930, p. 122. Known from a single specimen, 116 mm. long, from the Caribbean Sea. Not seen by us, but from the figure and description it seems certain that this specimen is a more advanced stage of P. achirus, above. 14. P. phyllopogon Regan & Trewavas, 1930, p. 122. Known only from the type specimen, 21 mm. long, from the Caribbean Sea. Not seen by us. 15. P. parvimanus Regan & Trewavas, 1930, p. 124. Known from 17 specimens in the type series and the present Bermuda collection, all young specimens from larvae to transitional adolescents, measuring between 14 and 55 mm. in length; from the North Atlantic. It is possible P. fimbria (species no. 16) and P. bifilifer (species no. 18) will prove to be synonymous with P. parvimanus. (See p. 170). 16. P. fimbria Regan & Trewavas, 1930, p. 125. Known only from the type specimen, 55 mm. long, from the North Atlantic. Not seen by us. From the description and figure it differs from P. parvimanus only in details of the barbel bulb and in having 30 to 34, not 34 to 38 P-V and O-V photo- phores. Our series of P. parvimanus shows specimens with barbels inter- mediate in form, therefore the photophore distinction alone remains. 17. P. monodactylus Regan & Trewavas, 1930, p. 127. Known only from 5 specimens in the type series, 180 to 255 mm. long, from the North Atlantic. Not seen by us, but from the description and figure we consider this species identical with P. margarita (see species no. 2, and p. 175). 18. P. bifilifer Beebe, 1933.2, p. 167. Known only from the type speci- men, 245 mm. long, from Bermuda, in the present collection. Allowing for the difference in size and growth stage, this adult upon reexamination differs from P. parvimanus (species no. 15, above) only in the lack of a crest on the barbel bulb and in the great length of the 2 pectoral rays. Both may prove also to be growth characters, or the barbel crest may have been torn away. Until intermediate stages are secured, however, it seems best to keep the two forms separate (see p. 173). 19. P. cornutus Beebe, 1933, p. 169. Known only from the type speci- men, 19 mm. long, from Bermuda, in the present collection. In poor condi- tion, and exceedingly close to P. mirabilis, but apparently distinct. (See p. 169). It is possible that sexual differences will be found to have significance in the proper delineation of species. A synopsis of the species, as now understood, will be found on p. 161. Distribution : Photonectes is one of the 7 genera in the family which have been recorded outside the Atlantic Ocean, P. albipinnis having been taken off Japan. Otherwise it is known only from the North Atlantic. The depth range indicated at present is between approximately 25 and 1,400 fathoms. A total of 146 specimens of Photonectes are now known, of which almost half (71) were taken by the Bermuda Expeditions. Generic Characters. Color (summarized from observations on about 20 freshly caught, including one living, transitional adolescent and adult specimens of 6 species) : General color blue-black, the skin being very fragile in adult speci- mens of the subgenus Trachinostomias ; barbel bulbs blue, pink, purple, yel- low, green or silvery; postorbital organ white, yellow, rose or violet; serial organs violet to purple with gold caps or frames; non-serial organs violet; 1939] Beebe & Crane: Family Melanostomiatidae 157 luminous patches on snout and jaws, when present, yellow or purple; luminous shoulder patches, when present, pale blue. Proportions 6: Melanostomiatids of moderate to excessive slenderness, with very short snouts and strongly curved jaws, the mandibles projecting in front of snout. Depth in length 6.5 to 15.5 (6.4% to 15.4%) ; head in length 6.5 to 9 (11.1% to 15.4%) ; eye in head 4 to 7.7; snout about as long as, or a little longer than, diameter of eye; snout to pelvic in length 1.5 to 1.9 (52% to 67%). Barbel : Usually shorter than head, rarely as long, or slightly longer, almost always unbranched above the bulb, but usually with a distal ap- pendage, varying in shape and degree of branching with the species; the bulb itself may be almost or completely atrophied in the adult. Light Organs : Postorbital (measured as length of area of transparent skin) .67 to 1.8 times diameter of eye in both sexes, probably always larger than eye in fully adult specimens. Serial photophores with the following counts: ventral series, I-V 30 to 48, V-A 11 to 18, A-C 9 to 15; lateral series, O-V 17 to 36; V-A 11 to 17. Non-serial photophores usually not conspicuous. Luminous tissue usually present in one or more of the following areas: in spots on snout, jaws, or shoulder; in bands or stripes on body. Teeth : Cleft of mouth strongly curved upward anteriorly; jaws remark- ably slender; the lower projecting far in front of upper premaxillaries and mandible with numerous close-set curved, slightly barbed teeth, all depres- sible, not large in adults; they are arranged in a single row in 3 to 6 series in each jaw, the teeth of each series increasing in size posteriorly; up to about 25 erect teeth on the maxillary followed by up to 50 very fine, oblique denticles; 1 or 2 pairs of teeth on the vomer; a series of 1 to 7 teeth on each palatine. Six to 10 pairs of teeth on the basibranchials. Teeth, in pairs, with a few individual teeth singly or in threes, present only on first 2 or 3 gill- arches, and only on the ceratobranchials. About 5 to 8 pairs on first ceratobranchial. Branchiostegal Rays: ca. 14 to 15. Fins : Pectoral with 0 to 3 rays, either short or long; pelvic 7, well developed, long; inserted far behind middle of length; dorsal 15 to 22; anal 15 to 24; membranes of dorsal and anal in some species very thick and black with only the tips of the rays visible in the adult; dorsal and anal com- mencing opposite each other, but anal extending farther back. Epidermal Grooves: There is a well defined groove in the isthmus for the reception of the barbel stem. Osteology: Mesethmoid with lateral processes; parietals absent; skull exceedingly short, less than half the length of the upper jaw; mandible with an elongate posterior process projecting behind the end of the maxillary; hyoid and gill-arches short; pectoral girdle much reduced, the postemporal usually absent, rarely vestigial, the supracleithrum and cleithrum weak, sometimes separated from each other, and the mesocoracoid absent; upper and lower coracoids well developed, laminar; actinosts 1 to 4; vertebrae 49 to 64; first vertebra represented only by a fibrous ring, enclosing the notochord, and by a spinal nerve. Coelomic Organs: Stomach ca. 33% of standard length, not reaching pelvic origin; 2 pyloric caeca. Sexual Dimorphism: Apparently none. Size: The largest known specimen is a specimen of P. margarita about 340 mm. in length taken by the Atlantis (Borodin, 1931, p. 66) ; the measure- ment is ours, 100 mm. more than given in the record which is probably a mis- print; the example is a female apparently in full breeding condition. Other specimens of the same species of both sexes, examined may be counted as 6 The proportions given include those of some obviously immature examples upon which species have been founded. 158 Zoologica: New York Zoological Society [XXIV :6 vo pal-ptq_ V 'v premax supclt max__ clt basibrs cerhy epihy inhy r.hypbr 2: cerbr 3: epibr : pKarbr Text-figure 33. Photonectes dinema. Jaws, hyoid and branchial arches, and pectoral girdle of transitional adolescent, standard length 38 mm. Explanation as in Text-fig. 18. adult from about 250 mm. on. A 170 mm. specimen of P. gracilis is a slightly immature female; P. bifilifer, 245 mm. long, is an adult male, not ready for breeding. Small specimens belonging to about 8 species, measuring less than 1939] Beebe & Crane: Family Melanostomiatidae 159 supclt jz tyT ■5 00 00 oo “ +3 co co co •- C I I I ,-Q a) t>- 00 OS 'dcococo o cS a> .2 00 co c3 CL CO v 00 _ 00 3 00 ' co G co ScoO w I *.-1 I g I CO CH c3 o 03 1 1 5-i 1 CD 1 linpT^soj ;inpv 10 0) > ◄ J?- Ofc°W c oo as*? ph<; 4) kps cS £ 00 £ cc i3 1 C5 t— KPS goo-P S'? £ §<“?w KPS'- 2 oo £op 2 p ->* cr • KPS ccS •4-5 S 00 S °o KPS T* as cS oo £ i rv\ 3 , 2 W £ S'? p M a a p'n ' O’ » E CO Ph N o LO LO ^inpy •uiuij o o H > H «■§ » § a; £ ui ^ ccS $ > ^3 rft CD £ > -I k M T3 US M £ 3 0) 5h C^ » G O' o G o be O G S’O il 3 ft a +-> C G £ j o, s 6 0) p Mortality Statistics, 1938. Deaths by Orders and ape groups, with causes. (For identification of postmortem numbers, see list preceding table). Class Reptilia1 Aves Mammalia Total Deaths2 550 | 43.2% 301 | 15.4% 128 | 25.3% International List Number Selected Causes3 H | Adult Orders and Postmortem Reference Number Imm. Adult Orders and Postmortem Reference Number Pre-adult *5 T3 c Post-adult Orders and Postmortem Reference Number 1 Infectious and parasitic diseases 5 15 4 23 4 15 Total: 66 1 Typhoid 1 Squamata R-26-38 2 Diseases due to Salmonella sp. 1 Squamata R-27-38 1 Pelecaniformes A-17-38 13a Amebic dysentery 1 Squamata R-52-38 (Helminthiasis) 23 Tuberculosis of respiratory system 1 Squamata R-44-38 32 Disseminated tuberculosis 1 Squamata R-64-38 1 1 1 1 Galliformes A-69-38 Gruiformes A-7-38 A-168-38 Psittaciformes A-66-38 1 1 1 Rodentia M-34-38 M-37-38 M-38-38 (Scabies) 36 Septicemia — purulent infection 1 1 Squamata R-19-38 (Trauma) R-22-38 (Drowning) 1 1 1 1 1 1 1 Pelecaniformes A-3-38 Falconiformes A-196-38 Galliformes A-79-38 Piciformes A-99-38 A-103-38 A-157-38 1 1 1 1 l Marsupialia M-5-38 M-130-38 Primates M-41-38 (Brain abcess) M-80-38 ll i I R-29-38 | I (Acariasis) 1 A-5-38 l M-52-38 i R-60-38 1 A-6-38 1 M-233-38 (Acariasis) 1 A-63-38 1 R-63-38 1 A-64-38 1 R-67-38 1 A-65-38 1 M-7-38 1 R-70-38 1 A-89-38 (Acariasis) R-88-38 1 1 A-124-38 A-126-38 pneumonia) 1 (Stomatitis) 1 M-74-38 (Enteritis) 38 Malaria 1 Squamata R-66-38 39 Sarcosporidia 1 Artiodactyla M-169-38 41 42 Hydatid cysts 1 Rodentia M-201-38 Diseases caused by helminths 1 1 1 Squamata R-35-38 R-39-38 R-74-38 1 Primates M-88-38 (Microfilaria) Carnivora 1 1 M-112-38 M-116-38 Rodentia — 1 M-127-38 Mycoses 1 Struthioniformes A-73-3S 1 Marsupialia M-ll-38 1 Gruiformes A-68-38 1 (Septicemia) M-22-38 — (Tracheal ob- struction) (Trauma) Other infectious and parasitic diseases 1 Squamata R-28-38 1 Anseriformes A-194-38 1 R-37-38 (Ixodiasis) (Botulism) Galliformes 1 R-95-38 1 A-199-38 (Acariasis) (Pox) Columbiformes 1 A-127-38 (Pigeon pox) i Including Amphibia. not i r^rponrtetf 'hi °thk 6teb I e" ep»r™ t^P ^ '-C8 ' «d including many specimens not sen 3 Based o nin terna t fo no 1 ju t r h “ , m«.Total., Deatk8 CO,Umn r5fer to ]osst* baaed on total num international list for human deaths and joint causes, with modifications. the Laboratory, of specimens in or not autopsied, and consequently each Department during the year. 272 Zoologica: New York Zoological Society [XXIV:10 I 1939] Schroeder: Mortality Statistics, 1938 273 Glass || Reptilia II Aves II Mammalia 274 Zoologica: Neiv York Zoological Society [XXIV :10 ■o S ® C ^ o ^ & c s m o P ^ ^ c bi S 02 P <*> P T14->'H I3 £ w P 00 o co .& »o P*p ?H *"“< PJ< o§ P<*> -d °? 000 4-> l> !h "O O 00-2 > CO g 'S rH £ 50 45 3,LH U§- ^inpB-^soj mpv ^inpB-aaj -O S 0) S«o^ * O Jj'S ® £ S 5 S w 02 ^ o£K £ O P 00 w 1 — 1 GO •’-? OO P CO ^ d w >L CO O « .2 Q) I I •’-< LQ *d SSSiS»2| • f-l *- CO 02 . m w P (Lo ocS- £00 3 s? 3 co 43 ■ o HnPV •luatj "© 2 OJ P 0) o ^ P ^ C 0) 2 2 2 Is i CD P *h g|5 o So P « P h bx) 5 Jh < O 5 "3 CD 5(5 2'3 5 rl a P P | P CO S o §l I O : -O 05 * O 3 O O aaquxnjvi puot^uja^ui 1939] Schroeder : Mortality Statistics, 1938 275 | Edentata M-146-38 Total: 4 Artiodactyla M-189-38 (Nutritional) M-265-38 (Nutritional) Total: 2 Primates M-239-38 Total: 15 Artiodactyla M-196-38 Marsupialia M-203-38 ( Salmonella enteriditis) Artiodactyla M-202-38 ‘ Primates M-150-38 Pinnipedia M-3-38 Carnivora M-56-38 - rH CM rH rH - - - CO rH j rH rH - tH - rH Ciconiiformes A-80-38 (Intramuscular hemorrhage) Anseriformes A-37-38 Galliformes A-170-38 (Nutritional) Anseriformes A-33-38 Ciconiiformes A-50-38 Anseriformes A-26-38 (Aneurism) Psittaciformes A-8-38 Passeriformes A-131-38 Ciconiiformes A-140-38 Galliformes A-167-38 rH fM rH rH - rH rH rH ^ rH rH rH Squamata R-86-38 (Parasitic) Squamata R-50-38 R-56-38 CO rH rH rH I Scurvy Diseases of blood and blood- making organs Anemias Others Diseases of the nervous system Encephalitis Diseases of the circulatory system Pericarditis Acute endocar- ditis Diseases of myocardium 95 Other diseases of heart | 97 | Arteriosclerosis (Atheromatosis) 09 > 1 71 71b V! 00 > 06 91 93 Class Reptilia Aves Mammalia International List Number Selected Causes Imm. Adult Orders and Postmortem Reference Number Imm. Adult Orders and Postmortem Reference Number Pre-adult Adult Post-adult Orders and Postmortem Reference Number 44 Other infectious and parasitic diseases ( continued ) 1 1 1 Passeriformes A-13-38 (Tropical fowl mite) A-46-38 (Pediculosis) A-185-38 (Teniasis) II Cancers and other tumors 4 3 Total: 7 46 46f Cancer and other malignant tumors of digestive tract Pancreas 1 Carnivora M-155-38 48 Cancer of the uterus 1 Artiodactyla M-78-38 49 49a Cancer of genital organs Ovary 1 1 Galliformes A-ll-38 A-154-38 53 53e Cancer of un- specified organs 1 Galliformes A-16-38 (Melanoma) 1 Carnivora M-61-38 (Thyroid) 54 54e Nonmalignant tumors 1 Charadriiformes A-30-38 III Rheumatic, nu- tritional, endo- crines, general 2 1 Total: 3 58 Gout 1 Anseriformes A-34-38 (Visceral ) 60 IV Scurvy l Ciconiiformes A-80-38 ( Intramuscular hemorrhage) 1 Edentata M-146-38 Diseases of blood and blood- making organs 2 1 1 Total: 4 71 71b Anemias Others 1 1 Anseriformes A-37-38 Galliformes A-170-38 (Nutritional) 1 1 Artiodactyla M-189-38 (Nutritional) M-265-38 (Nutritional) VI Diseases of the nervous system 1 1 Total: 2 78 Encephalitis 1 Anseriformes A-33-38 1 Primates M-239-38 VII Diseases of the circulatory system 3 6 1 3 2 Total: 15 90 Pericarditis 1 Artiodactyla M-196-38 91 Acute endocar- ditis 1 1 1 Ciconiiformes A-50-38 1 1 Marsupialia M-203-38 ( Salmonella enteriditis) Artiodactyla M-202-38 93 Diseases of myocardium — 1 Squamata R-86-38 (Parasitic) Anseriformes A-26-38 (Aneurism) Psittaciformes A-8-38 95 Other diseases of heai't 1 1 Squamata R-50-38 R-56-38 1 Passeriformes A-131-38 1 1 Pi’imates M-150-38 Pinnipedia M-3-38 97 Arteriosclerosis (Atheromatosis) I 1 1 Ciconiiformes A-140-38 Galliformes A-167-38 1 Cai’nivora M-56-38 274 Zoologica: New York Zoological Society [XXIV :10 I 1939] Schroeder: Mortality Statistics, 1938 275 276 Zoologica: New York Zoological Society [XXIV : 10 Mammalia Orders and Postmortem Reference Number Total: 60 Primates M-55-38 (Septicemia) M-114-38 Carnivora M-113-38 Rodentia M-4-38 Primates M-27-38 Marsupialia M-44-38 M-87-38 Carnivora M-144-38 Marsupialia M-97-38 (Trichomo- niasis) Carnivora M-117-38 Rodentia M-118-38 M-190-38 M-195-38 Hyracoidea M-138-38 ^npu-^soj linpv 00 rH rH rH rH rH rH rH ^npn-aaj r- rH rH rH t— 1 rH rH rH Aves Orders and Postmortem Reference Number Psittaciformes A-219-38 Ciconiiformes A-12-38 Passeriformes A-70-38 A-172-38 A-181-38 Pelecanifoi-mes A-40-38 Ciconiiformes A-176-38 Anseriformes A-55-38 A-203-38 (Proventriculitis) F alconiformes A-9-38 Galliformes A-28-38 A-91-38 tlnPV 00 fM rH rH rH rH rH rH rH rH rH rH rH rH •umij - Reptilia Orders and Postmortem Reference N umber Squamata R-2-38 R-13-38 R-34-38 R-45-38 R-46-38 R-47-38 R-76-38 R-90-38 Squamata R-6-38 R-9-38 R-23-38 (Septicemia) R-25-38 ( Helminthiasis) R-51-38 R-84-38 R-87-38 (Pericarditis) R-100-38 4InPV rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH *UIUIJ Class Selected Causes Diseases of respiratory system Bronchopneu- monia Lobar pneumonia Pneumonia (All unspecified) Congestion, edema, embolism, hemorrhagic infarct, throm- bosis joqum]q ^sirj puoi^uuja^ui VIII o o rH 00 O rH O rH rH rH rH 1939] Schroeder : Mortality Statistics, 1938 277 Artiodactyla M-16-38 (Aspiration pn.) Total: 72 Hyracoidea M-197-38 M-208-38 I r- rH rH N A-179-38 (Laryngo- tracheitis) A-180-38 Gruiformes A-159-38 (Compound fracture) Psittaciformes A-43-38 A-53-38 A-144-38 (Trauma) Passeriformes A-19-38 A-22-38 A-27-38 A-29-38 A-36-38 A-51-38 A-58-38 A-62-38 A-75-38 A-87-38 A-115-38 Strigiformes A-24-38 (Pox) Falconiformes A-171-38 Passeriformes A-211-38 tH r~t rH rH i-H i-H HHHHH 22 rH 1 tH rH N fH Squamata R-40-38 R-42-38 R-43-38 R-55-38 R-59-38 R-62-38 R-102-38 34 rH rH rH rH rH rH rH Other respiratory diseases Diseases of the digestive system Diseases of buccal cavity and annexa Ulcer of stomach and duodenum Other diseases of stomach Class Reptilia Aves Mam malia — u o ^ Orders and Orders and 3 a Orders and z g S' Postmortem S 3 Postmortem TJ 3 Id Postmortem Causes fc Reference M Reference i 4-> Reference Number Number u &, £ Number VIII Diseases of 28 respiratory system 16 i 7 8 Total: 60 107 Bronchopneu- Psittaciformes Primates monia 1 A-219-38 1 M-55-38 (Septicemia) 1 M-114-38 Carnivora 1 M-113-38 Rodentia 1 M-4-38 108 Lobar pneumonia 1 Primates M-27-38 109 Pneumonia Squamata Ciconiiformes Marsupialia (All unspecified) 1 R-2-38 1 A-12-38 1 M-44-38 1 R-13-38 Passeriformes 1 M-87-38 1 R-34-38 1 A-70-38 Carnivora 1 R-45-38 1 A-172-38 1 M-144-38 1 R-46-38 1 A-181-38 1 R-47-38 1 R-76-38 1 R-90-38 111 Congestion, Squamata Pelecaniformes edema, embolism, 1 R-6-38 1 A-40-38 1 M-97-38 hemorrhagic 1 R-9-38 Ciconiiformes (Trichomo- infarct, throm- 1 R-23-38 1 A-176-38 bosis (Septicemia) Anseriformes 1 R-25-38 1 A-55-38 1 M-117-38 ( Helminthiasis) 1 A-203-38 1 R-51-38 (Proventriculitis) 1 M-118-38 1 R-84-38 Falconiformes 1 M-190-38 1 R-87-38 1 A-9-38 1 M-195-38 (Pericarditis) Galliformes 1 R-100-38 1 A-28-38 1 M-138-38 1 A-91-38 114 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 A-179-38 (Laryngo- tracheitis) A-180-38 Gruiformes A-159-38 (Compound fracture) Psittaciformes A-43-38 A-53-38 A-144-38 (Trauma) Passeriformes A-19-38 A-22-38 A-27-38 A-29-38 A-36-38 A-51-38 A-58-38 A-62-38 A-75-38 A-87-38 A-115-38 1939] Schroeder: Mortali\ Other respiratory diseases 1 Artiodactyla ^ M-16-38 Co (Aspiration pn.) o IX Diseases of the digestive system 4 34 2 22 3 7 1 Total: 72 S’ 115 Diseases of buccal cavity and annexa 1 1 1 1 1 1 1 Squamata R-40-38 R-42-38 R-43-38 R-55-38 R-59-38 R-62-38 R-102-38 1 Strigiformes A-24-38 (Pox) , 1938 117 Ulcer of stomach and duodenum 1 1 Hyracoidea M-197-38 M-208-38 118 Other diseases of stomach 1 1 Falconiformes A-171-38 Passeriformes A-211-38 | -a -3 276 Zoologica: New York Zoological Society [XXIV:10 Class | Reptilia i| Aves || Mammalia 278 Zoologica: Neiv York Zoological Society [XXIV :10 £ cu 5 ® o ^ ts V d 0> « s 3 S S § 5 £*Z a ■£oc 3°? m o £ 05 Cu l C$ CS £ £ 00 M 4J O 2 O 00 O CO^OOCS c3 00 £ o C/2 I cj 1 CD C/2 ^ i P •*--< oo .^h -«-• t— c yr; 2 lo a» ^ ^ Oi fH W ^-| g eg P-H +* *4J (M C tlsblai:g=8i;si c3 _ Jh 00 00 aj O CO CO .i—i 00 00 £ 3 | :»mv o ' c S 8 - “ <01 ^ "poo £ co — P O I o< O O vi oo ^5 •geo — c.._ £ 00 S^tDtOlO-^ ■+■? 00 MHHCO^O ."rH T— I C/2 ^ T— H 1— I T— I T— I C/3 I 03 I I • I I CL,<2fa C o O ^ ®s t, cS m o ?* 'S _ 5 in ^oi>-ooooc50^aia5asaiOj> Mfafafafafafafafafafafafafafafafafafafa PifafaMfa 2oOOOOOOOOOOO-£m £cocoeocoeoco‘£ i i • I I » • I g <£> c$.£ CD •— < w a/ p£ 1939J Schroeder: Mortality Statistics, 1938 279 Total: 5 Marsupialia M-247-38 Carnivora M-102-38 Total: 2 Artiodactyla M-85-38 Total: 4 Carnivora M-29-38 Edentata M-6-38 Totel: 5 Primates M-234-38 M-264-38 Artiodactyla M-221-38 Total: 1 i Total: 9 ; Primates M-21-38 M-167-38 Carnivora 1 M-l-38 II 1 CM II rH rH rH CM II tH tH II tH o* H H rH Struthioniformes A-14-38 Cuculiformes A-214-38 Strigiformes A-20-38 Passeriformes A-215-38 Anseriformes A-56-38 Falconiformes A-189-38 Gruiformes A-190-38 CO rH tH rH - rH 1 CM rH H rH | Squamata R-78-38 R-103-38 II CM — i Diseases of the j genito-urinary system Nephritis Disease of pregnancy (Obstetrical diseases) Diseases of skin and cellular tissue Miscellaneous Diseases of bones and organs of locomotion Osteomyelitis and other diseases of bone Congenital malformations Congenital malformations Diseases of new-born Premature birth, injury at birth, atelectasis © H (M CO ^ CO CO CO CO HHHH X CO t- CO ^ ^ rH rH rH rH X 151 152 153 XIII rf LOO UOlO kO rH rH rH XIV 157 AX 159 160 161a 1 Class leptilia Aves Man imalia 5 J3 Orders and Orders and 3 3 Orders and Selected £ 'g Postmortem £ 'p Postmortem 3 Is Postmortem c £ Causes £ Reference £ Reference u -*-> Reference 5 w Number Number £ Number ►5 J 122 Diseases of Squamata Ciconiiformes Marsupialia 123 the intestine 1 R-l-38 l A-121-38 1 M-90-38 1 R-12-38 Falconiformes (Trichomo- 1 R-17-38 l A-88-38 niasis) 1 R-20-38 Galliformes 1 M-98-38 1 R-21-38 i A-98-38 (Trichomo- 1 R-30-38 l A-138-38 niasis) 1 R-31-38 i A-210-38 i M-171-38 1 R-33-38 Gruiformes Primates 1 R-38-38 l A-67-38 i M-26-38 1 R-41-38 Trogoniformes ( Ulcerative 1 R-53-38 1 A-81-38 colitis) 1 R-54-38 Coraciiformes Artiodactyla 1 R-61-38 i A-142-38 1 M-25-38 1 R-69-38 l A-143-38 (Pneumonia) 1 R-71-38 l A-222-38 1 R-85-38 Piciformes 1 R-89-38 l A-10-38 1 R-91-38 l A-78-38 1 R-92-38 Passei-iformes 1 R-93-38 l A-21-38 1 1 R-94-38 R-96-38 l A-178-38 1 R-97-38 1 R-98-38 124 125 Diseases of the liver 1 Squamata R-14-38 l Columbiformes A-l-38 1 Carnivora M-181-38 1 R-16-38 Psittaciformes 1 M-188-38 1 R-32-38 l A-188-38 1 R-48-38 Passeriformes 1 M-2-38 1 R-68-38 i A-41-38 1 R-99-38 l A-116-38 Amphibia i A-136-38 1 Am-1-38 i A-145-38 1 1 A-174-38 1 X Diseases of the genito-urinary system 1 3 2 Total: 5 130 131 132 133 Nephritis 1 1 1 Struthioniformes A-14-38 Cuculiformes A-214-38 Strigiformes A-20-38 1 1 Marsupialia M-247-38 Carnivora M-102-38 XI Disease of pregnancy (Obstetrical diseases) 1 Total: 2 143 145 147 149 1 Passeriformes A-215-38 1 I Artiodactyla M-85-38 XII Diseases of skin and cellular tissue 2 2 Total: 4 151 152 153 Miscellaneous 1 1 Squamata R-78-38 R-103-38 1 1 Carnivora M-29-38 Edentata M-6-38 XIII Diseases of bones and organs of locomotion 2 , 2 Total: 5 154 155 156 Osteomyelitis and other diseases of bone 1 1 Anseriformes A-56-38 Falconiformes A-189-38 1 1 1 Primates M-234-38 M-264-38 Artiodactyla M-221-38 XIV Congenital malformations 1 Total: 1 157 Congenital malformations 1 Gruiformes A-190-38 XV Diseases of new-born 9 Total: 9 159 160 161a Premature birth, injury at birth, atelectasis 1 1 1 1 Primates M-21-38 M-167-38 Carnivora M-l-38 278 Zoologica: New York Zoological Society [XXIV :10 I 1939] Schroeder: Mortality Statistics, 1938 280 Zoologica: New York Zoological Society [XXIV:10 a 15 £ Orders and Postmortem Reference Number Artiodactyla M-13-38 M-28-38 M-42-38 M-43-38 M-46-38 M-184-38 Total: 57 Carnivora M-17-38 M-18-38 M-20-38 M-67-38 Primates M-14-38 M-79-38 Rodentia M-143-38 Artiodactyla M-8-38 M-81-38 M-83-38 s cS s linpu-^soj CM llnPV CM CM H H rH rH rH rH qinpu-ajj r— 1 rH rH H H H m rH rH rH rH CO a> > Orders and Postmortem Reference Number l Psittaciformes A-169-38 Pelecaniformes A-149-38 Ciconiiformes A-175-38 Galliformes A-77-38 A-101-38 Psittaciformes A-153-38 Passeriformes A-54-38 A-82-38 A-90-38 A-147-38 linpy in rH rH rH rH rH rH rH rH rH ‘UIUIJ 1 rH teptilia Orders and Postmortem Reference Number Squamata R-ll-38 R-57-38 Squamata R-15-38 (Parasitism) Squamata R-24-S8 • HnPV ro rH rH rH •UIUIJ rH Class Selected Causes Premature birth, injury at birth, atelectasis (continued) Violent and accidental deaths Homicide Attack by venom- ous animal Other poisons Accidental traumatism jaquinM lsiq 1 03 o LO CD 1!^ {^UOI^BUja^UJ j rH rH XVII 175 176 179x 981 1939] Schroeder: Mortality Statistics, 1938 281 .2 13 *0.00 S'co in zo es c7 a u oooooo cS o oo oo OOOOOCOCOCO-^OO crtCOCO >COCO'*'-*^COrrt'' • 1-1 I I HOJO g I X ^ CO>I>OHC <1> LO .2 00 o Im CJ 00 co O 00 03 00 CO CO C 50 .2 50 LO 50 ^w^cohh £i> c t-H O) OhS — a oo OS CO .©§ H-> rH aJ ^ C o oo a> . > CO nd Ch ‘-H 03 ^ a> T3 r2 !>> 0000 as 00 % tj >)00 oo CO CO .-H co P as $5 - .sSS^o^aou^S co im "rt ooi®n^ mo .2 co -m dol DlOolNN-BlM Jh > -u CO *-< 3s^11gSSStfS&]L3s~ ■§5 CO G o o m h-< • y—i hh rH t-H t-H t-H t-H t-H 00 CO i CO t> t-H < Psittaciformes A-57-38 Passeriformes A-201-38 Struthioniformes A-206-38 Gruiformes A-25-38 Passeriformes A-15-38 t-H t-H t-H t-H rH t-H * Injuries by animals Hunger and thirst Excessive cold Violent deaths, nature unknown Destroyed 188 189 190 195 198 arthritis) Class Eteptilia Aves Mar nmalia •a S g ^ Orders and Orders and 3 3 Orders and £ Postmortem £ 3 Postmortem 3 Postmortem g Z Causes £ Reference £ Reference i J. Reference s « Number Number £ (2 Number £ J 159 Premature birth, Artiodactyla 160 injury at birth, 1 M-13-38 161a atelectasis 1 M-28-38 ( continued ) 1 M-42-38 1 M-43-38 1 M-46-38 1 M-184-38 XVII Violent and accidental deaths 1 3 1 15 13 22 2 Total: 57 175 Homicide Squamata Carnivora 1 R-ll-38 1 M-17-38 1 R-57-38 1 1 1 M-18-38 M-20-38 M-67-38 176 Attack by venom- Squamata ous animal 1 R-15-38 (Parasitism) 179x Other poisons Squamata Psittaciformes 1 R-24-88 1 A-169-38 186 Accidental Pelecaniformes traumatism 1 A-149-38 1 M-14-38 Ciconiiformes 1 M-79-38 1 A-175-38 Rodentia Galliformes 1 M-143-38 1 A-77-38 Artiodactyla A-101-38 1 M-8-38 Psittaciformes 1 M-81-38 1 A-153-38 Passeriformes 1 M-83-38 1 A-54-38 1 A-82-38 1 A-90-38 i A-147-38 1 1 l A-173-38 188 Injuries by animals 1 1 Psittaciformes A-57-38 Passeriformes A-201-38 1 | 1 1 1 1 1 1 1 1 — Marsupialia M-76-38 Carnivora M-49-38 M-77-38 M-101-38 M-219-38 M-266-38 Rodentia M-95-38 Artiodactyla M-180-38 M-204-38 189 Hunger and thirst 1 1 1 1 1 Carnivora M-66-38 Artiodactyla M-36-38 M-151-38 M-168-38 190 Excessive cold 1 Struthionifonnes A-206-38 1 Pi’imates M-170-38 (Teniasis) 195 Violent deaths, nature unknown 1 1 Gruiformes A-25-38 Passeriformes A-15-38 Artiodactyla M-120-38 198 Destroyed 1 1 1 1 1 1 1 Carnivora M-40-38 ( Purulent cellulitis and osteomyelitis) M-59-38 M-96-38 M-220-38 M-249-38 Rodentia M-250-38 (Fractured pelvis) Artiodactyla M-33-38 ( Arteriosclerosis and chronic arthritis) Zoologica: New York Zoological Society [XXIV :10 I 1939] Schroeder: Mortality Statistics, 1938 281 Class |j Reptilia || Aves || Mammalia 282 Zoologica: Neiv York Zoological Society [XXIV :10 ■oSqj C o cj cUf c 0) co o 9>-$ a; P ^ S t'-p'h 3 -L c/3 GO 00 00 00 CO CO O 00 »^'Sc?c?c?c?Aco >°? ®WOCO P CO £ O 05 ^ 05 CO GO c ^ JjTtHH.SH.p;HtD0OClHH RnPV qnpmaaj C aj o 133 -P c ^5^ O rv ^ p a 5 w £3 .2 a> 00^ o W £ 00-* S'? jS'ScO'o io o 10 "5 m ° -I pg 2 CO * ~ 05 -fi o 00 00 00 tH CO CO 00 ■£ rJ* I — < O 05 S»HO CO 2 rH D1 o«? <<-<•< up £3 GO 00 '£ 00 •hX 5h 00 CO CO rS 00 CO geo Oco^^ o CO Xi; CD Cvl -pf ^T—ILO r^C5^t>HH ^ CD t— 1 CO I C3 I 1 I _C I I f* <; o c a> o ir* £ a> cfl o ^ qj £ a> p ^ £ <4H ^ 2 cg a> ^ o,£«* T [l'lP'^ c3 -*-> a3 £ 00 00 00 00 ^ CO CO CO CO 3 CO Tf lO [-• pi pi pi pi 0000000000000000 OOCOCOCOCOCOCOCOCO C?OCD05 00l0(MC0i0 OOHCO^lOCDt-C^I> pi p^ pi pi pi pi pi*pi pi P tfl +3 3 S1 C6 3h . 33 33 fig 03 O 33 +* o +>r& ^ r* . o § 8 ’5 1939] 00 Q 00 00 00 00 00 CO 00 c3 oo CO CO CO COCOCO [> O r-H liOS-P^HWN Schroecler : Mortality Statistics, 1938 283 in o> o> 00 K) rH£ CM .Th ' w - i-H ?h CM ^ CM (J H H ^HcOCOtJ^^IOCDI^I>HHhHH 00 CO I CO CM 01 OH(MCOOO tOOOCOOOOOHH c3 -4— 1 cC £ 00 ^ CO -2o> C/1 PhC^PhPhPhP^PhHPh List Number II 1 1 R-77-38 1 A-2 18-38 1 l I 1 1 R-80-38 Psittaciformes l 1 1 R-81-38 1 A-109-38 1 1 R-82-38 Strigiformes 1 1 R-83-38 1 A-72-38 1 1 R-101-38 1 A-134-38 1 R-104-38 Micropodiformes 1 Testudinata 1 A-44-38 1 1 R-79-38 Coraciiformes 1 1 1 A-2-38 1 1 A-133-38 1 A-221-38 Piciformes 1 A-128-38 1 A-137-38 | Passeriformes 1 A-18-38 1 A-31-38 1 A-32-38 1 A-42-38 1 A-48-38 1 A-49-38 1 A-59-38 1 A-60-38 1 A-71-38 1 A-74-38 1 A-108-38 1 A-l 10-38 1 A-lll-38 1 A-l 18-38 1 A-156-38 1 A-223-38 i M-137-38 Edentata M-92-38 Artiodactyla M-47-38 M-125-38 M-267-38 M-268-38 Zoologica: New York Zoological Society [XXIV:10 I 1939] Schroeder: Mortality Statistics, 1938 Kresky & Barnett: Carcinoma of the Pancreas 285 11. Carcinoma of the Pancreas of Acinar Origin in a Bear.1 Philip J. Kresky, M. D. & Roy N. Barnett, M. D. (Plates I & II). Carcinoma of the Pancreas. Malignant tumors of the pancreas in animals are of sufficient rarity to warrant the report of an individual case. A review of the literature reveals only eleven well-authenticated cases of carcinoma of the pancreas in mam- mals; of these, five were in domestic animals, one in a horse (1), three in dogs (2), (3), and one in a cat (4). The six cases in captive wild animals included two in mice, (3), two in monkeys, one in a jackal and one in an Indian civet (5). Balozet & Chainet (1) reporting an adenomacarcinoma of the pancreas in a horse, state that malignant tumors of the pancreas are extremely rare in all the mammalian orders with the exception of carnivora and primates. This statement is confirmed by Slye, Holmes & Wells (6), who performed autopsies on 125,000 mice dying of natural causes. They found 20,000 tumors, of which only two were carcinomas of the pancreas. Ratcliffe (5), in reviewing autopsies on 3,400 mammals and 6,898 birds from the Philadelphia Zoo, found 96 tumors in mammals, 65 of which were malignant and 28 benign. Six were pancreatic tumors, two benign adenomas and four carcinomas. Of 82 tumors found in 6,898 birds, only one was a pan- creatic tumor. This was an adenocarcinoma of the pancreas in a fantail grackle. The only malignant pancreatic tumor in the reptilian class (4) was reported by Ratcliffe (7) in a Say’s pine snake; on histologic examination this proved to be an adenocarcinoma. Only two of the tumors mentioned (2), (5), had organ metastases. One of these was the case reported by Bru, in a dog with carcinoma of the pancreas, composed of epithelioid cords of large clear cells. He believed that the tumor had its origin in the Islands of Langerhans. Metastatic foci were found in the lung, heart and liver. The other case was that of a jackal discussed by Ratcliffe. This was a medul- lary carcinoma with liver metastases. With the exception of Bru’s case and the case of Nocard referred to by Kitt (3), which lacked a histologic descrip- tion, all of the malignant tumors were thought to have their derivation from either the ductal or acinar portions of the pancreas. Five were adeno- carcinomas (5), (6), two were scirrhous carcinomas (3), (4), one was a medullary carcinoma (5) and one was an adenocarcinoma with areas con- taining primitive cells of an embryonal type (1). In no instance was the diagnosis made before death. No clinical syn- drome, comparable to that seen in humans, can be constructed from the 1 From the New York Zoological Park and the Laboratories of The Mount Sinai Hospital. 286 Zoologica: New York Zoological Society [XXIV :11 available data. The only case in which jaundice was seen during life, was reported by Scott & Moore (4). This was a cat with a scirrhous carcinoma involving the head and body of the pancreas and constricting the common bile duct. Our case is the first in which a carcinoma of the pancreas was found in a bear. Ratcliff e’s (6) series included sixty bears on which autopsies were performed. He found four instances of carcinoma; these included a medul- lary carcinoma of the mammary gland, a basal cell carcinoma of the tongue, an adenocarcinoma of the adrenals, and hypernephroma of the kidney. The only other reported instance of a malignant tumor in a bear was reported by Perry (8). This was a 34-year-old animal with an adenocarcinoma of the kidney. Case Report. The large Kadiak bear, whose case is reported here, was captured as a cub in 1907 on Kadiak Island, Alaska, and presented to the New York Zoological Park. He grew and developed normally, reaching a weight of approximately 450 kilograms. He was well until February, 1938, when loss of appetite and weight were noticed. His course was steadily downhill until in August, 1938, he was too weak to stand. At no time was icterus or any other localizing sign or symptom noted. His weight had dropped to approxi- mately 160 kilograms when he was destroyed on August 4, 1938, at the esti- mated age of 31 years. Post Mortem Examination. The Mount Sinai Hospital Autopsy No. Z-300; New York Zoological Park No. m-155-38. Gross : The specimen was an emaciated, adult, male Kadiak bear, Ursus middendorffi Merriam, about eight feet total length and weighing approxi- mately 160 kilograms. His fur was lustreless, sparse, brittle and pulled out easily. Over the upper inner aspect of each buttock was a small shallow decubitus ulcer. A firm pedunculated tumor, about the size of a plum, hung from the outer aspect of the upper one-third of his right thigh. This had a narrow stalk and was covered by skin. The superficial fat was almost lacking. The heart and lungs were not unusual, except for the pale, gelatinous appearance of the epicardial fat. There was no ascites. In the head of the pancreas, behind the descending portion of the duodenum, was a large, elliptical, stony hard tumor, measur- ing 8 by 6 by 3 cm. (PI. I, Fig. 1). The common bile duct passed over the anterior aspect of the tumor, but was neither narrowed nor infiltrated. On cut section, the tumor was salmon-colored and the surface was traversed by interlacing bands of glistening white fibrous tissue. Many soft cystic areas, containing semi-liquified tissue, were found in the tumor. The body of the pancreas appeared normally lobulated. The pancreatic duct could not be found. The retroperitoneal lymph nodes, posterior to the tumor, were as large as hazel nuts, firm, and on section contained gray-white areas. The liver was huge, weighing 5,560 grams (PI. I, Fig. 2). The surface was studded with large round pink-white metastatic nodules, some umbilicated and ranging from one to eleven cm. in diameter. On section, much of the liver parenchyma was replaced by pink-white circumscribed round foci similar to those seen on the surface. Many of these metastatic nodules con- tained areas of necrosis and hemorrhage. No other tumor metastases were found. An incidental finding was the presence of numerous, large, facetted stones in the gall bladder, whose wall was somewhat thickened. The large biliary radicles were all patent. No unusual changes were noted in the remaining viscera. 1939] Kresky & Barnett: Carcinoma of the Pancreas 287 Microscopic : The tumor in the pancreas, liver and lymph nodes had essentially the same histologic appearance. There were two types of cell patterns. In one the cells were arranged in definite pseudo-acinar group- ings (PI. II, Figs. 3A and 4A) or in thick, tortuous cords having papillary projections. The cells had large vesicular hyperchromatic nuclei with promi- nent deep-staining nucleoli. Mitotic figures were infrequent. The cytoplasm was abundant and contained innumerable eosinophilic zymogen granules (PI. II, Fig. 3B). The zymogen granules in the pseudo-acinar structures were situated on the side of the cells next to the lumen, and pushed the nuclei to the base of the cell. The other predominating cell pattern consisted of epithelioid cords of small cells, many with dark pyknotic nuclei. (PI. II, Fig. 4B). The cytoplasm was scant and filled with vacuoles which took the Sudan III stain for neutral fat. Some of the cells in these cords contained zymogen granules. The stroma of the tumor consisted of a loose, delicate, highly vascular connective tissue framework. No portion of the carcinoma resembled the Islands of Langerhans. Sections of the pancreas immediately adjacent to the primary tumor showed atrophy of the parenchyma and dense interstitial fibrosis. Histologic study of the rest of the pancreas revealed slight lipomatosis. Comment. Ordinarily pancreatic tumors arise from either the ducts or the Islands of Langerhans. Histologic examination proved our tumor to be a paren- chymal cell carcinoma, having its origin in the acinar tissue of the pancreas. The most unusual feature was the presence of zymogen granules in the cells. This is a definite indication that active secretion and elaboration of pan- creatic enzymes was occurring, and on this basis the origin of the tumor in pancreatic acinar cells can be postulated. In some of the other glandular organs, carcinomas producing secretion are well-known. Hepatomas of the liver produce bile and have the same metabolic functions as normal liver cells. Symptoms of Grave’s disease may be associated with adenocarcinomas of the thyroid gland. However, in only one pancreatic carcinoma in the literature was acinar origin proved by the presence of secretion. This was the case of Sugiura, Pack & Stewart (9) who proved the presence of active enzymes in a human pancreatic adenocarcinoma. Their tumor was considered histologically characteristic of an acinar adenocarcinoma, but no mention of the presence of zymogen granules was made. They found that the tumor con- tained as much protease and lipase and more amylase than normal human pancreas. Summary. A case of primary parenchymal cell carcinoma of the pancreas with extensive liver metastases is reported. The significance of zymogen granules pointing to acinar cell origin is discussed. Bibliography. 1. Balozet, L. & Chainet, L.: Primary Epithelioma of the Pancreas in a Horse. Bull. Acad. Vet. 10:301, 1937. 2. Bru, P.: Cancer Langerhansien Generalise chez un Chien. Rev. Medico- Chirurgicale des Maladies du Foie, du Pancreas, et de la Rate. 2:40, 1927. 3. Kitt, T.: Lehrb. d. path. Anat. d. Haustiere, 1921, 600. 4. Scott, E. & Moore, R. A.: A case of Pancreatic Carcinoma in a Cat. J. of Cancer Research. 11:152, 1927. 5. Ratcliffe, H.: Incidence and Nature of Tumors in Captive Wild Mammals and Birds. A. J. Cancer. 17:116, 1933. 288 Zoologica: New York Zoological Society 6. Slye, M., Holmes, H. F. & Wells, H. G.: Comparative Pathology of Carcinoma of the Pancreas. A. J. Cancer. 23:81, 1935. 7. Ratcliffe, H.: Carcinoma of the Pancreas in Say’s Pine Snake. A. J. Cancer. 24:78, 1935. 8. Perry, C. B. : Carcinoma of the Kidney in a Bear. J. Comp. Path, and Therap. 42:133, 1929. 9. Sugiura, Pack & Stewart: A Study of the Enzyme Content of a Paren- chymatous Adenocai’cinoma of the Pancreas and a Comparison with the Normal Human Pancreas. A. J. Cancer. 26:351, 1936. EXPLANATION OF THE PLATES. Plate I. Fig. 1. Primary carcinoma of pancreas. D — Duodenum. S — Stomach. T — Tumor. Fig. 2. Liver with metastases. Plate II. Fig. 3A. Pseudo-acinar structure. Fig. 3B. Higher magnification, showing zymogen granules. Fig. 4A. More anaplastic area. Fig. 4B. Small-cell area, same magnification as 4A. KRESKY ft BARNETT. PLATE I. FIG. 2. CARCINOMA OF THE PANCREAS OF ACINAR ORIGIN IN A BEAR. KRESKY & BARNETT. PLATE II. FIG. 3A. FIG. 3B. FIG. 4A. FIG. 4B. CARCINOMA OF THE PANCREAS OF ACINAR ORIGIN IN A BEAR. Finkelstein : Goiter in a Dromedary 289 12. Congenital Macrofollicular Cystic Colloid Goiter in a Dromedary. Leonard E. Finkelstein, M.D.1 (Plate I). Congenital colloid goiter, especially with cystic change, is a rarity not only among the lower animals but in the human species as well (4,12). More infrequently is it a cause of intrauterine fetal death and of dystocia, as occurred in the following case at the Zoological Park. Case. An adult female dromedary, multiparous although without any record of previous pregnancies at the Zoological Park, went into labor following an uneventful full-term period of gestation. Parturition failed to proceed normally. A vaginal examination by Dr. Schroeder, Veterinarian of the New York Zoological Park, revealed that the fetus was in a dorso-sacral position with the hind legs retracted, so that the tarsi presented. In addition, early maceration was evident, indicating that the fetus was dead. A very large cystic mass was palpated on the neck of the fetus. This, together with the head and forelegs, was too large to pass through the birth canal. Embry- otomy was performed. Pathological Examination. The head of the fetus was shortened in the antero-posterior diameter, being a so-called “bull-dog head.” Except for this and the tumor of the neck there were no other congenital malformations. The tumor consists of two equal and similar masses which occupy the normal positions of the thyroid lobes, and is, in fact, a markedly enlarged thyroid gland. Each lobe measures 25 centimeters in length, 15 centimeters in width, and 8 centimeters in thickness. (The length of a normal thyroid in an adult dromedary is 3-4 inches (7-10 centimeters) according to Leese (5)). The enlargement appears diffuse and uniform. A dense gray fibrous capsule to which is adherent several strands of muscle surrounds each lobe. The gland is soft in consistency and feels cystic on palpation. The cut surface of the gland is moist, glistening, pale brownish in color, and gelatinous in appearance. It possesses a coarsely honeycombed struc- ture, being composed of numerous bulging, round and angulated, very thin walled cysts. (PI. I). These vary in size from about 0.5 to 2 centimeters in diameter, most of them being about 1 centimeter. The cystic spaces are all filled with translucent, pale brown colloid. In some areas extending in from the capsule there are broad gray fibrous septa. No normal thyroid structure is grossly observed. i Theodore Escherich Fellow in Pathology. From the Laboratories of The Mount Sinai Hospital, New York City, and the New York Zoological Park. 290 Zoological Neiv York Zoological Society [XXIV :12 On microscopic examination moderate autolysis is noted. The capsule is seen to be composed of a broad band of dense connective tissue on the external surface of which several striated muscle fibers are intimately attached. The glandular follicles are enormously dilated and cystic, being distended by eosinophilic homogeneous colloid. Some follicles beneath the capsule appear compressed and distorted. The lining epithelium is every- where compressed, the cells being very much flattened, so that it appears as a thin line between distended follicles. There ai-e no papillomatous spurs, and no distinct evidence of parenchymal proliferation. The cell nuclei are small and very basophilic. Except for occasional thick connective tissue septa, no interfollicular substance is noted. A few small compressed blood vessels are present in the capsule and some of the septa. The anatomic diagnosis of the specimen is congenital colloid goiter with macrofollicular cystic degeneration. The intrauterine death of the fetus was probably due to compression of the carotid arteries by the huge thyroid lobes (12). Autopsy on the Dam. About 16 months after this pregnancy the dam became blind and weak and was destroyed. At necropsy extensive acute and chronic infection of the pulmonary and intestinal tracts and of the meninges and uterus was found. The thyroid gland was moderately enlarged, the right lobe measuring 15X 7X2.5 centimeters, and the left lobe 11X1.5X2.5 centimeters. Upon examin- ation the gland is uniformly brownish-red in color, and on section it appears fleshy. There are no adenomatous nodules. Histologically striking hyper- plasia of the parenchymal element is observed. The epithelial cells are large, being high cuboidal in shape. The nuclei, too, are large and vesicular. The colloid content is considerably less than normal. Discussion. Simple, non-toxic goiter is found in a variety of animals, not only in adults, but in the newborn as well (2, 3, 4, 6, 7, 13). In fact, according to Marine (8), goiter may occur in any land or fresh water animal. It is said that wild animals never develop goiter even in regions where endemic goiter is prevalent (14). However, Fox (3) has noted the affliction in captive wild animals at the Philadelphia Zoological Garden. In acquired goiter, human as well as animal, a definite goiter cycle has been established wherein the hyperactive phase is associated with glandular hyperplasia. Colloid goiter represents a resting state in the cycle, a physio- logic return to normal which is expressed anatomically by an accumulation of colloid (7, 8). Joest (4) cites a case described by Johne, of a colloid struma in an adult dromedary. In congenital goiter, which per se is not infrequent, by far the usual pathologic picture is one of hyperactivity — parenchymal hyperplasia with very little colloid (2, 4, 7, 10, 12, 13, 14). Marine (8) explains this as a physiologic reaction in the fetal gland compensating for the increased de- mands on the maternal thyroid that often obtain during pregnancy. This causal relationship is well demonstrated in the experimental production of congenital goiter by almost completely extirpating the maternal thyroid (2, 9). Abbott & Ball (1) in a study of 100 fetal and newborn thyroid glands state that it is reasonable to assume that the pathologic changes in the fetal thyroid are induced by the same type of stimulus that exists in the adult. If the etiology of hyperplastic parenchymatous struma in the new- born is reasonably clear, the pathogenesis of a congenital colloid goiter is equally unclear. The latter is uncommon, the cystic form being extremely rare (4, 10, 12). In most cases of the colloid type there is a coexisting 1939] Finkelstein: Goiter in a Dromedary 291 hyperplasia. This, however, may be obscured by the colloid, so that one may be unable to affirm or deny its presence (11, 12). Since a paucity of colloid is a striking feature of the normal thyroid in the newborn, the asso- ciation of hyperplasia with colloid is significant. It lends color to the con- cept that congenital colloid goiter originates in a hyperplastic thyroid rather than that it develops de novo. In the case presented here, the enormous cystic dilatation of the fol- licles would readily obfuscate histologic evidence of preceding hyperplasia. It appears likely that the maternal thyroid was already in the hyperactive phase during the period of pregnancy, and initiated the fetal changes which went on to regression. One cannot say from this case alone whether a dysontogenetic hypersecretion of colloid ensued or whether some other fac- tors came into play to produce the final picture. Congenital goiter in the one-humped camel is not unknown. Leese (5) in a treatise on this species writes that the thyroid glands may be enormous- ly enlarged at birth, and that death from suffocation may occur shortly after delivery. At times the goiter may persist without change throughout life, or may even become larger, reaching the size of a man’s head and in- terfere with grazing from the ground. While Leese does not describe the anatomic type of goiter, such large masses would seem to be similar to the cystic colloid struma of this fetus. Leese emphasizes the occurence of congenital goiter among camels out of dams confined to zoos or to ships during long sea-voyages, and suggests that insufficient exercise of the pregnant dam may be a predisposing factor. However, we now know that goiter is due to a relative or absolute lack of iodine which leads to a work hypertrophy of the thyroid (Marine (8) ). The mediate factors involved include various dietary faults and the increased metabolic requirements in pregnancy and febrile, toxic states, any or all of which may be present in captive animals. Summary. 1. A case is reported of a congenital macrofollicular cystic colloid goiter occurring in a full-term, still-born dromedary. The goiter resulted in dystocia. 2. Such a struma is rare not only in the lower animals but in humans as well. 3. An autopsy performed on the dam 16 months after delivery dis- closed marked parenchymal hyperplasia of the thyroid together with severe acute and chronic infection of the pulmonary and intestinal tracts, meninges and uterus. 4. It is likely that the maternal thyroid was already hyperplastic dur- ing pregnancy, and initiated the fetal pathology. The cystic changes in the fetal gland would readily obscure evidence of parenchymatous hyperplasia— the usual picture in congenital goiter. The author is indebted to Dr. Charles R. Schroeder, Dr. Paul Klemperer and Dr. Sadao Otani for their aid in the preparation of this report. Bibliography. 1. Abbott, A. C., & Ball, R. P. 1931. Pathology of the thyroid gland of the human fetus and the newborn infant. Canad. Med. Assoc. Jour., 24, 347. 2. Carlson, A. J. 1914. On the cause of congenital goitre (thyroid hyperplasia) in dogs and cats. Am. Jour. Physiol., 33, 143. 292 Zoologica: New York Zoological Society 3. Fox, Herbert 1923. Disease in captive wild mammals and birds. J. B. Lippincott Co., Phila. 4. Joest, Ernest 1924. Handbuch der speziellen pathologischen Anatomie der Haustiere. Vol. 3, p. 29. Richard Schoetz, Berlin. 5. Leese, A. S. 1927. A treatise on the one-humped camel in health and disease. Haynes & Son, Stamford, Lincolnshire. 6. Loken, A. 1913. Congenital goiter of the dog. Am. Vet. Review, 43, 91. 7. Marine, David 1907. On the occurrence and physiological nature of glandular hyper- plasia of the thyroid (dog and sheep), together with remarks on important clinical (human) problems. Johns Hopkins Hosp. Bull., 18, 359. 8. Marine, David 1924. Etiology and pi’evention of simple goiter. Medicine, 3, 453. 9. Marine, D., & Lenhardt, C. H. 1909. Effects of the administration or the withholding of iodine con- taining compounds in normal, colloid or actively hyperplastic (par- emchymatous) thyroids of dogs. Some experiments on (congenital) prenatal thyroid hyperplasia in dogs: remarks on the clinical man- ifestations associated with marked thyroid hyperplasia. Arch. Int. Med., 20, 131. 10. Mitchell, D. S., & Struthers, R. R. 1933. Congenital hypertrophy of the thyroid gland with report of a case. Canad. Med. Assoc. Jour., 28, 27. 11. Redaelli, P. 1934. Struma congenita colloide, struma congenita telengettasica ed al- terazioni ipofisarie. Arch. Ital. Anat. e Istol. Patol., 5, 1050. 12. Wegelin, C. 1926. Schildriise. In Handbuch der speziellen pathologischen Anatomie und Histologie, Henke, F., and Lubarsch, O., Vol. 8, p. 150, 316. Julius Springer, Berlin. 13. Wegelin, C. 1928. Malignant disease of the thyroid and its relation to goiter in man and animals. The Cancer Review, 3, 297. 14. Wegelin, C. 1929. Endemic goiter, especially in Switzerland. Jour. State Med., 37, 480. EXPLANATION OF THE PLATE. Plate I. Cut surface of the fetal thyroid gland in a case of congenital macrofollicular cystic colloid goiter. Note the uniform enlargement, the glistening, translucent appear- ance, and the coarsely honeycombed structure. No normal thyroid tissue present. Each lobe measures 25 X 15 X 8 centimeters. (Photographed under water). FINKELSTEIN. PLATE I. CONGENITAL MACROFOLLICULAR CYSTIC COLLOID GOITER IN A DROMEDARY. Herman: Pentatrichomonas macropi from Kangaroos 293 13. Pentatrichomonas macropi Tanabe from Kangaroos. Carlton M. Herman, Sc.D. Hospital and Laboratory, New York Zoological Park (Text-figure 1). A Woodward’s wallaroo, Macropus robustus woodwardi, in the collection of the New York Zoological Park, died on July 5, 1938. It was autopsied at the Hospital and Laboratory of the Zoological Park within an hour after death. A microscopical examination of material obtained from the enlarged and distended caecum revealed a flagellated protozoan organism of the Tricho- monas type. Subsequently this parasite was obtained from the caecal con- tents or feces from a total of five species of kangaroos living at the New York Zoological Park. The original material collected from the caecum of the wallaroo was taken to the laboratory of the Mt. Sinai Hospital by Dr. L. Finkelstein and successfully grown on blood agar slants. It was maintained further, by trans- fers on this media, at the Laboratory of the Zoological Park. Observations were made from living specimens- — -both fresh and cultured material — and from fixed preparations. The fixatives employed in this study were Schaudinn’s fluid (plus 5% glacial acetic acid), osmic acid vapor and methyl alcohol. Organisms killed with either of the first two fixatives were stained with Heidenhain’s iron-alum haematoxylin. Organisms fixed in methyl alcohol were stained with Giemsa’s stain. The Giemsa method was found to be best for diagnostic purposes. The stain was used in the same concentration and for the same length of time as employed for blood smears. The morphology of this parasite agrees with the description of Penta- trichomonas macropi Tanabe (1926). Tanabe’s description of this parasite from a kangaroo (genus and species not given) was made entirely from cul- tured material. In the present study there seemed to be a much greater diversity in size in the fresh material and a greater variation of shapes in the cultured forms. The length of specimens from fresh material varied between 4.5 p. and 15 g while in the cultured forms the size tended toward an average between 7 and 10 g as reported by Tanabe. A diagrammatic sketch of Penta- trichomonas macropi made from observations on both fresh and cultured forms is included in this paper as a text-figure. The following hosts in the collection of the New York Zoological Park have been found to be infected with this parasite : 3 rock kangaroos, Macropus brunii. 2 Woodward’s wallaroos, Macropus robustus woodwardi. 3 black-faced kangaroos, Macropus melanops. 3 great gray kangaroos, Macropus g. giganticus. 3 black tree kangaroos, Dendrolagus ursinus. 294 Zoologica: New York Zoological Society [XXIV :13 Pentatrichomonas macropi Tanabe. From kangaroos (diagrammatic sketch). Pentatrichomonas macropi grows quite readily on a variety of culture media at 37° C. The greatest abundance of organisms can be found in orig- inal cultures between 48 and 72 hours after inoculation with fecal material. Subcultures were successfully made through seven transfers. The parasites tend to die out after 72 hours (becoming overwhelmed with bacteria) and even in subcultures the organisms never achieved the great numbers seen in the first tubes inoculated. A number of culture media were tested as to their ability to support a growth of Pentatrichomonas macropi. Fair growth was obtained with the following: undiluted blood serum, diluted blood serum (1 part serum plus 1 part dist. aqua), Loeffler’s serum-saline (0.5% Loeffler’s plus 0.75% serum), and Hogue’s ovo-mucoid (100 cc. physiological saline plus white of one egg). Good results were obtained with blood agar slants. The water of condensation did not prove sufficient to support a growth of the trichomonads but good results were obtained when the various liquid media listed above were added in small quantities. Distilled water or saline on blood agar slants also sup- ported a good growth. Charcoal agar with these various fluids did not sup- port as abundant a growth as the blood agar media. Horse blood was used throughout in the preparation of the culture tubes. The organisms seem to 1939] Herman: Pentatrichomonas macropi from Kangaroos 295 require a liquid medium but prefer a solid base such as a blood agar slant. The best medium obtained in these experiments was a fecal infusion on blood agar slants. Both rat feces and kangaroo feces were tried with equal success and were used for routine diagnosis of Trichomonas in most pf the kangaroos studied antemortem. Of five of the kangaroos that came to autopsy during this study, three had ulcers in the digestive tract. A careful examination of microscopic sec- tions of these ulcers in each case did not reveal any trichomonads associated with the necrosis. Whether or not there is any correlation with the presence of Pentatrichomonas macropi and the occurrence of ulcers in kangaroos it would be impossible to state from our present knowledge. Only one of these animals exhibited a diarrhea. Samples of intestinal contents from various areas seem to indicate that Pentatrichomonas macropi is primarily a parasite of the caecum. However, positive evidence of their presence in other portions of the large intestine and the posterior region of the small intestine was obtained both by direct smear and by the culture method. Summary. 1. A trichomonad parasite, Pentatrichomonas macropi Tanabe, is re- ported from five species of kangaroos in the collection of the New York Zoo- logical Park. 2. The parasites can be grown readily in vitro. The best culture media found was fecal infusion on a blood agar slant at 37° C. 3. Although three of the kangaroos infected with trichomonads were found to have intestinal ulcers, no parasites could be found in the tissues and it is doubtful if there is any correlative significance to this finding. Bibliography. Tanabe, M. 1926. Morphological studies on Trichomonas. Jour. Parasit. 12, 120-130. Kazimir off : Dental Pathology, 1938 297 14. A Report on the Dental Pathology Found in Animals that Died in the New York Zoological Park in 1938. Theodore Kazimiroff (Plates I-IX). Introduction. It is very seldom that one finds a description of oral pathology in autopsy reports of animals in captivity. This is due to several factors; primarily inability to recognize or interpret lesions either in the living state or on the autopsy table, and secondly, the masking of these hard tissue lesions by apparently normal-appearing soft tissues, which effec- tively cover and hide any lesions present. It is not a widely known fact that the majority of oral lesions invariably affect the hard structures around the mouth — the teeth, which may present lesions of the enamel, dentine or cementum, and the surrounding bone, which indelibly bears the lesions of various oral diseases. It is apparent that the study of oral lesions in the morbid state would materially augment autopsy reports. In addition this valuable information that would otherwise go unknown and undetected, may help solve some of the mysterious conditions often encountered. As far as preparation of the material is concerned, the best means is maceration, preferably at body temperature, although room temperature, while slower, is suitable. If rapidity of preparation is desired, the speci- mens may be cooked down in soap solutions. Neither way will affect the hard tissues, or produce unrecognizable artifacts. This paper covers the range of pathology found in the specimens col- lected at the New York Zoological Park Hospital during 1938. This in- cludes dental caries, injuries of the teeth, mal-positioning of the teeth, dento-alveolar abscess, periodontal pathology, impacted teeth and other conditions. The photography is the work of E. R. Osterndorff. The radiographs were taken by the author. 1. Dental Caries. Dental caries occurs in both wild and captive animals. Sir Frank Colyer has shown that the disease is more prevalent in captive animals than in wild animals. The term caries denotes a pathologic condition which results in the destruction of either enamel, dentine or cementum through the agency of specific, pathogenic micro-organisms. This condition is to be differentiated from loss of tooth structure by attrition or trauma. Although the cause of dental caries is still disputed, several important factors are to be considered: a, the action of micro-organisms; b, structural defects of 298 Zoologica: New York Zoological Society [XXIV :14 the teeth; c, traumatic fractures of teeth; d, mal-positioning of the teeth with resultant food impaction areas; e, a faulty diet, usually found to be rich in carbohydrate intake, and many other factors disputed or accepted by authorities. The following specimens illustrate some of these conditions. Primates. a. Macacus rhesus (immature). This specimen presents a condition of incipient caries of the enamel on the distal surface of the maxillary right central incisor. The caries occurs at the point of contact between this tooth and the lateral incisor. The mesio-incisal corner of the lateral had been chipped, creating a wedge- shaped food impaction area which undoubtedly contributed to the carious process. b. Hussar Monkey, Erythrocebus patas, M-234-38.1 This specimen shows caries that had occurred on the site of an old fracture of the maxillary left canine tooth. The carious process has hol- lowed out the crown of the canine. (Plate I, Fig. 1). The exposed pulp present may have been due either to the original fracture or to caries following the fracture. Alveolo-dental abscess of long standing is present, with fistulation through the buccal plate of bone in the apical region. Rodentia. a. Woodchuck, Marmota monax, M-37-38. Plate I, Fig. 2b. This specimen exhibits caries of the premolar and of the first and second molars. The disto-buccal cusp of the mandibular left premolar is completely destroyed, the carious process having extended below the cemento-enamel junction onto the disto-buccal root which has become ex- posed through alveoloclasia. The mandibular left first molar presents mesio- occlusal and disto-occlusal caries of the crown, and buccal caries at the cemento-enamel stagnation area. The mandibular left second molar like- wise presents caries at the cemento-enamel junction. The carious process starts at the mesial surface and ends on the mesio-buccal apect. In addi- tion there is deep and extensive pocket formation with much evidence of suppuration. (Plate I, Fig. 2). b. Woodchuck, Marmota monax, M-38-38. This specimen presents occlusal caries of the mesio-lingual cusp of the mandibular left first molar. The carious process is of the typical pit caries type, wherein the caries has penetrated and undermined the mesio-lingual cusp for a distance of two-thirds the length of the crown. The caries has not broken through the mesial plate of enamel, although a discoloration is seen through the enamel. Carnivora. In the wild state the Carnivora are relatively free from caries. How- ever the Ursidae and Procyonidae show an increased caries susceptibility in captivity. Raccoon, Procyon lotor, M-59-38. This specimen shows interproximal caries involving the distal surface of the mandibular left first molar, resulting from a food impaction area. The second molar is tilted at an angle of approximately 30° with the horizontal plane of occlusion, and is situated on the ascending anterior curve of the coronoid process. This created a wide, wedge-shaped, food impaction area between the distal of the first molar and the mesial of the second molar. Food, wedged into this interproximal space, and acted upon by the lactic acid-producing organisms, undoubtedly caused the caries. 1 This and subsequent numbers refers to the case history in the records of the Hospital and Laboratory, New York Zoological Park. 1939] Kazimiroff : Dental Pathology, 1938 299 (Plate I, Fig. 2a). In addition bone resorbtion and pocket formation resulted from this food impaction area. Hyracoidea. Hyrax, Procavia capensis, M-138-38. Plate II, Fig. 3. Heretofore, the Hyracoidea have been regarded as a caries-resistant type. Sir Frank Colyer records 300 specimens examined with none showing caries. Hence I believe that this specimen of hyrax presenting caries is the first to be described in the literature. This animal exhibits a very extensive and destructive type of caries, simulating the condition known as “rampant caries” in the human. All the maxillary premolars and molars are involved, producing a con- tinuous trough-like carious gutter running from the first premolar to the last molar. (Plate II, Fig. 3). The carious process has hollowed out the crowns, removing most of the coronal dentine, but has left the undermined buccal plate of enamel intact on each tooth. The lingual plate of enamel is missing from some of the teeth, but this appears to be due to the under- mined, weakened, enamel plates having chipped off. The carious process has attacked the interproximal areas and the transverse occlusal fissures as well. The mandibular teeth show seven of the twelve posterior teeth involved. The mandibular incisor teeth show incipient caries in the form of interproximal discoloration and etching of the enamel below the contact points. The two central incisors show caries of the enamel at the contact point. 2. Injuries of the Teeth. Plate II, Fig. 4. Carnivora. Coyote, Canis latrans, M-18-38. The left maxillary fourth premolar of this specimen was injured in some manner during the formative stage, producing several interesting sequellae. The injury affected the mesial portion of the tooth, resulting in a haphazard calcification of dentine and enamel. The crown appears to be completely denuded of enamel in some parts, and composed entirely of enamel in other parts. The dentine and enamel calcification had become indiscriminately intermingled. It appears that the power of growth of both ameloblasts and ondoblasts is not affected by traumatic injury as in this instance. The root portion was also affected: a gnarled, shortened, misshapen root resulted, with several “rootlets” having been created, particularly one rootlike spur on the buccal aspect which is 1 mm. wide and 2 mm. long. There was a definite ankylosis of root to alveolar bone. The apex of the mesial root had fused to the alveolar bone, and the mesio-buccal margin had fused to the buccal plate of bone. The X-ray indicates that a pulp was present in the distal portion of the tooth, both root and coronal pulp being present. The effect on the maxillary and mandibular dental arch is interesting. The side opposite that of the injury shows the more severe mal-positioning of the teeth. The mandibular right canine protrudes horizontally at right angles to the long axis of the other mandibular teeth. The left mandibular canine tooth is likewise misplaced but not as severely as the right canine. The maxillary right canine has changed its axial inclination; the crown has been pushed anteriorly, and the root displaced posteriorly. In spite of all this, the animal showed very little, if any, evidence of suppuration, and only slight alveoloclasia. 300 [XXIV :14 Zoologica: New York Zoological Society 3. Dento-alveolar Abscess. Dento-alveolar abscess is frequently found in wild animals kept in captivity. The usual causes are exposure of the pulp by means either of fracture of a tooth or severe attrition. Not all such cases result in abscess formation, however. Another cause of abscess formation, though relatively infrequent, is the result of infection of the peri-dental membrane progressing so far as to produce an abscess. Carnivora. Gray Wolf, Canis nubilus, M-61-38. Plate III, Fig. 5. This specimen shows abscess formation following a fracture of the upper right central incisor, exposing the pulp chamber. The resultant pulp infection was followed by apical involvement indicated by fistulation and rarefaction of bone. The extreme of alveoloclasia is shown in this case with complete destruction of all alveolar bone, as is seen in the illustration. The pocket formation encircles the root completely. An unusual situation re- sulted from this condition, hypercementosis and root resorbtion occurring simultaneously. Marsupialia. Kangaroo, Macropus robustus. This specimen exhibits abscess formation as a result of a fracture of the upper left second incisor. Fistulation took place with the pus pointing into the left nasal aperture. The fistulation resulted in extensive bone de- struction with much evidence of suppuration. Primates. Hussar Monkey, Erythrocebus patas, M-234-38. This specimen presents in addition to abscess formation a very unusual condition which will be fully described in the section dealing with perio- dontal pathology. The abscess formation followed as a result of exposure of the pulp of the maxillary left canine tooth. An old fracture of the crown was followed by caries. Whether the pulp exposure resulted from the fracture or from the caries is hard to say. However the pulp infection was followed by apical involvement and abscess formation. The resultant fistulous opening is in the bone of the maxilla at the apex of the canine. (Plate III, Fig. 7). Drainage was evidently by means of the soft tissues between the periosteum and the skin, into the oral cavity. Artiodactyla. Axis deer, Axis axis, M-221-38. This specimen presents one of the rare cases of abscess formation fol- lowing periodontal membrane infection. The entire condition will be dis- cussed under the section dealing with periodontal pathology. A severe food impaction area resulted in deep pocket formation be- tween the distal root of the left mandibular second molar and the mesial root of the third molar. Even after maceration, fodder fragments are present interproximally and in the pockets. Suppuration hollowed out the lingual bone and left a definite lateral sinus. (Plate IV, Fig. 9). Fistula- tion is on the lingual surface of the mandible and is accompanied by a peculiar, raised osteoporosis surrounding the fistulous opening. Drainage was probably through the soft tissues of the floor of the mouth. 4. Periodontal Pathology and the Resultant Bone Lesions. The majority of the specimens skeletonized exhibit evidences of perio- dontal disturbances. 1939] Kazimir off : Dental Pathology, 1938 301 Carnivora. Gray Wolf, Canis nuhilus, M-61-38. Plate III, Figs. 5, 6. This specimen shows a fairly well developed periodontal disturbance. Alveolar bone destruction is marked throughout, slightly more severe in the upper jaw. The bone shows the characteristic signs of suppuration, rarefying osteitis pronounced throughout. Abscess and pocket formation around the upper right central incisor have been described in a preceding section. There is a deep food pocket present between the distal of 'the maxillary right fourth premolar and the first molar. The maxillary fourth premolar has a large deposit of salivary calculus (visible in the illustration as the projecting white mass under the zygomer, posterior to the canine. Plate III, Fig. 5). There is also a severe food impaction area between the left mandibular first and second molars. The talonid or distal cusp of the mandibular left first molar has been broken off. The space created acted as a food impac- tion area, and the wedging of the food caused deep interproximal and lateral pocket formation. (Plate III, Fig. 6). There is much evidence of marginal suppuration in the form of a rarefying osteitis. Paradoxure, Paradoxurus jerdoni, M-77-38. Sir Frank Colyer reports a case of periodontal disease in Paradoxurus larvatus (masked paradoxure) which is similar to the condition found in M-77-38. The alveolar bone destruction around the maxillary teeth is highly advanced, more than one-half of the root surface being exposed. There is much evidence of a rarefying osteitis. The mandibular teeth, however, show a heavy marginal proliferation of bone instead of an alveoloclasia. This appears to be a slow response to a condition of long standing. (Plate V, Fig. 10). Artiodactyla. Axis Deer, Axis axis, M-221-38. This specimen presents the last stages of periodontal disease, the stage preceding loss of teeth. There is a complete destruction of almost all alveolar bone, lateral as well as interproximal bone. The only means of retention of the teeth is by extreme hypercementosis. This is a response to the destruction of the alveolar bone and is retentive in function. The hypercementosis is so severe that the distal roots of the maxillary first premolars have fused with the mesial roots of the second premolars. (Plate V, Fig. lib). The incisor teeth likewise exhibit this extreme deposition of cementum. The apical enlargements are thus two or three times the dimensions of the crowns of the teeth. (Plate V, Fig. 11a). The alveoli show signs of extensive suppuration, a complete rarefying osteitis being present. Fistulation into the maxillary sinus has taken place. (Plate VI, Fig. 12). An unusual condition has resulted from the drainage into the left antrum. The best description seems to be a “suppurative blow- out” of the posterior wall of the left antrum. This is well seen in the in- ferior view of the skull, marked by the arrows. (Plate IV, Fig. 8). The discharge of pus was into the inferior orbital and infra-temporal region. The pus had distended the posterior wall of the sinus, resulting in a paper- thin, globular, bony swelling that ultimately “blew out.” The degree of attrition and the amount of alveoloclasia is well depicted in the accompany- ing illustrations. (Plate VI, Figs. 12, 13). Primates. Hussar Monkey, Erythrocebus patas, M-234-38. This specimen presents a very unusual condition: a severe unilateral alveoloclasia. For some reason this animal developed a unilateral mastica- tion, the left side being used almost exclusively for chewing. As a conse- 302 Zoologica: New York Zoological Society [XXIV :14 quence, the degree of attrition is greater on the left side than on the right side. The left was in a fairly normal state, very slight alveolar bone destruction present. (Plate VII, Fig. 14). However, the right or atrophic side shows the extreme effect of alveoloclasia. As the illustrations show, there is complete destruction of all alveolar bone. (Plate VII, Fig. 15). The entire right side, maxillary and mandibular, exhibits a severe, progressive osteitis. The comparison of the photographs of the right and left sides presents the case graphically. The premolars and molars of the right side have their root apices pro- truding into the maxillary sinus. See illustration — the arrows point to the sinus openings. (Plate VII, Fig. 15). Plate I, Fig. 2, shows, beside the raccoon mandible with a carious molar, two examples of periodontal disease in the rodents. Rodentia. Woodchuck, Marmota monax, M-37-38. In addition to caries this specimen exhibits a lateral pocket or sup- purative sinus extending from the mesial of the left mandibular premolar to the distal of the second molar. There is a space of 3 mm. between the root surfaces and the buccal plate of bone. (Plate I, Fig. 2b). African Ground Squirrel, Geosciurus capensis. This specimen exhibits the end stages of periodontal disease. Only two of the teeth remain, the position of the lost teeth being indicated by the presence of an edentulous trough. (Plate I, Fig. 2c). 5. Impactions and Mal-positioning of Teeth. Occasionally teeth are mal-occluded, mal-positioned or impacted. The cause of these conditions is still obscure, and many factors may be in- volved. The specimen included in this section is one of unusual interest. A very brief case history and autopsy report may be of value. Primates. Woolly Monkey, Lagothrix humboldtii, M-216-38. Young adult, male. Partial right facial hemiplegia, cheek and lip principally — blindness — bilateral exopthalmia— weakness — unsteady gait — tarsi and carpi flexed — physically unable to extend limbs — oedema of extremities — eyes had much retro-bulbar fat surrounding optic nerve. The description of the specimen is augmented by photographs and radiographs. The alveolar process and ridge is about three times that of normal, presenting a swollen, hypertrophic, maxillary alveolar process. The molars were the only teeth in functional occlusion. There are several visible im- pactions: two maxillary right premolar impactions, and one left maxillary premolar impaction. The maxillary deciduous canines were retained, while the permanent canines remained impacted. (See X-rays, Plate IX, Fig. 18). The maxillary third molars are seemingly horizontally impacted. The mandible presents a somewhat similar picture. The permanent canines are impacted, but there are no retained deciduous canines. (Plate IX, Fig. 18a). The left mandibular second premolar is likewise impacted. The third molars are impacted horizontally, under the ascending portion of the coronoid process. (Plate IX, Fig. 18 b, c). Unfortunately, the cal- varium and some of the teeth were lost in the preparation of the specimen. However, an examination of the long bones ruled out what appeared at first to be a case of Paget’s disease. It would appear that this condition resulted from some complex glandular disturbance. 1939] Kazimir off : Dental Pathology, 1938 303 Summary. This paper is a report on the oral pathology observed in specimens received from the New York Zoological Park hospital for the year 1938. The specimens were skeletonized, and all observations were made from lesions as seen in the morbid state. This means of studying dental pathology should be invaluable in augmenting the autopsy reports of zoological col- lections. The pathology found includes dental caries, injuries of . i -1 * The number in parenthesis is the titer in thousands. Heterologous titers may be calculated from the per cent, of relationship. 316 Zoologica: Neiv York Zoological Society [XXIV :16 gether in Table 5. Serological tests have shown that the bloods of these two animals are very closely related and usually indistinguishable by the pre- cipitin method. Wolfe (1933) emphasized that when the bloods of two or more animals are very closely related their heterologous reactions should be similar. The data further verify this principle. The antisera of the two different species gave reactions that indicated the very close affinity of the tahr to them and also showed that the sable antelope, though not as closely related to the goat and sheep as is the tahr, has a closer affinity to them than do the other Bovidae or the Cervidae. The anti-sheep sera brought out this fact more clearly than did the anti-goat sera; this may be due to the fact that the former antisera were produced by the injection of a smaller quantity of antigen, which often results in more specific antisera. The anti-sheep sera also gave definite Bovidae-group reactions but the less specific anti-goat sera did not. Two anti-eland sera (55 and 64) of different specificity were produced by injections of minute quantities of antigen. The relationship values of these two antisera are recorded in Table 6 and they indicate that the ox and buffalo are more closely related to the eland than are some other members of the Bovidae. Serum 55 was the more specific antiserum and its reactions at the intervals of time at which the titers were read were greater with the ox and buffalo antigens than they were with other Bovidae or with the Cervidae. On the other hand, serum 64 showed a closer kinship of the ox and buffalo to the eland only at the 5, 10 and 20 minute titers while the 30 and 60 minute readings were not significantly different from those of other Bovidae or the Cervidae. Serum 55 gave weak reactions but of similar magnitude with the three species of Cervidae and the Bovidae except the ox and buffalo. On the other Table 7. Relationship values (in per cent.) of anti-black buck antelope antisera*. Antigen No. 45 No. 43 No. 72 No. 73 Diluted 1 :0.5 Undiluted Undiluted Undiluted Black Buck Antelope 100.(516)f 100.(256) 100.(256) 100.(256) Sable Antelope 6.25 0 12.5 6.25 Sheep 6.25 0 12.5 3.12 Goat 3.12 0 12.5 1.56 Tahr 3.12 0 6.25 1.56 Ox 3.12 0 3.12 3.12 Buffalo .78 0 3.12 1.56 Eland 1.56 0 3.12 .78 Elk .78 0 .78 .78 Axis Deer .39 0 1.56 .78 Virginia Deer .78 0 .39 .39 t The number in parenthesis is the titer in thousands. Heterologous titers may be calculated from the per cent, of relationship. * All are 60 minute values. 1939] Wolfe: Serologic Relationships 317 hand, with serum 64 the Cervidae reactions were very definite at 10 minutes, but the sheep, goat, tahr and black buck antelope reactions were not noticed until the 20 minute period. This difference in reaction was evident when one volume of the antiserum was diluted with three volumes of buffered saline but not if it were used undiluted or diluted with one volume of saline. These results may indicate that the bloods of the Cervidae used are more closely related to the eland than certain members of the family to which the eland belongs according to its present classification. Since only one of 2 antisera gave these results it is, of course, necessary to secure additional data. Table 8. Relationship values (in per cent.) of 3 species of Cervidae. Antiserum Antiserum No. Antiserum dilution Time of reading (min.) A xis Deer Elk Virginia Deer Axis Deer A12 1:1 5 100. (16) 100. 0 10 100. (64) 75. 50. 20 100. (192) 100. 75. 30 100. (256) 100. 75. 60 100. (512) 75, 75. 1:4 20 100. (8) 50. 50. 30 100. (128) 100. 3.12 60 100. (128) 100. 100. 35 1:2 5 100. (2) 25. 0 10 100. (128) 50. 1.56 20 100. (384) 75. 37.5 30 100. (1024) 25. 12.5 60 100. (1024) 25. 12.5 61 1:3 5 100. (64) 100. 50. 60 100. (512) 100. 25. Elk 46 1:3 5 100. 100. (256) 25. 60 50. 100. (1024) 25. 48 1:3 5 50. 100. (128) 25. 60 100. 100. (1024) 25. 1:5 5 (1) (2) (0) 60 50. 100. (128) 3.12 57 1:0.5 5 37.5 100. (128) 3.12 10 50. 100. (256) 25. 60 100. 100. (256) 75. 1:2 60 50. 100. (128) 12.5 Virginia Deer 62 1:1 5 50. 50. 100. (128) 60 75. 75. 100. (512) 1:3 10 0 0 100. (64) 20 50. 50. 100. (128) 60 50. 50. 100. (256) 1:4 10 0 0 100. (16) 20 0 0 100. (64) 60 50. 50. 100. (256) * The number in parenthesis is the titer in thousands. 318 Zoologica: New York Zoological Societg | XXIV : 16 Four antisera were produced by injections of minimum quantities of black buck antelope serum ; these were all very specific. The cross reactions with Bovidae and Cervidae were much weaker than the homologous tests. Antiserum 49, though showing a high homologous titer, failed to give reac- tions with any of the heterologous antigens. Further results indicate that the sable antelope, sheep, goat, and tahr were more closely related to the black buck antelope than were the other Bovidae or Cervidae. Beddard (1920) states, “It is exceedingly difficult to separate antelopes from the sheep, oxen and goats. Their inclusion along with these creatures in one family, Bovidae, shows that no differences of an important character exist. ... It is perhaps with the goats that the antelopes have their nearest affinities.” The correla- tion of present serological findings with the statements of Beddard based on morphological data is highly significant. The reactions of black buck antelope with the ox, buffalo and eland, though slightly greater, were not significantly higher than those with the Cervidae. It is possible that the less specific antisera might have demon- strated the relationships better than the antisera that were used. The high degree of specificity of the reactions has somewhat obscured the kinship involved and attempts should be made to further verify the present findings by the use of slightly less specific antisera. The Cervidae antisera reactions furnished the interesting results shown in Table 8. The only reactions listed are the homologous ones and the cross reactions with the bloods of the other two deer, although tests were also made with eight species of Bovidae. The three axis deer and three elk antisera yielded reactions that usually failed to distinguish these two species. The greater number of relationship values for the cross tests with the elk or axis deer were 50 to 100 per cent., only a few being 25 and 37.5 per cent., which is just outside the limits of error of the tests. On the other hand, the titers of certain of the antisera of these two species with the Virginia deer serum were often so much lower than the homologous titers that the latter could easily be distinguished from either the axis or elk. This fact is especially well shown with the anti-axis deer serum A12 diluted 1 :4, anti-axis deer serum 35 diluted 1:2, anti-elk serum 57 diluted 1:5 and anti-elk serum 48 diluted 1:2. Only one anti-Virginia deer serum was produced. This serum when diluted with 3 volumes of saline could differentiate the Virginia deer serum from that of either the elk or axis deer at the 10 minute reading. A 1:4 dilu- tion of the antiserum retarded the positive reaction with the two heterologous bloods to the 60 minute reading, while with the homologous antigen the reactions were positive though of low titer at 10 and 20 minutes. These results, which show an exceedingly close similarity of the blood of elk to that of the axis deer and a distinction of both of these from the Virginia deer, are very significant, for they can be correlated with the origin of these animals. According to Scott (1937, p. 322), the American elk (also called wapiti) and the Virginia deer are North American forms that have had a different origin. Scott states that, “. . . North American deer form two strongly contrasted groups, the northern and southern. In the northern group the deer are like those of the Old World — these include the Wapiti.” The southern group, which includes the Virginia deer, seems to have had a different ancestry and, according to Scott, probably originated from a “long line of American ancestry.” The per cent, values of the anti-deer reaction with the Bovidae were usually much lower than that with the homologous blood. In a small number of tests the titers were high enough so that a distinction was not possible, but most often the percentage of reaction was 12.5 per cent, or lower. The indications were that the ox and buffalo were more closely related to the deer group than were the other Bovidae. 1939] 319 Wolfe: Serologic Relationships Discussion. The precipitin test, a serological method, was used by the author to determine the relationships among eight Bovidae and three Cervidae. The value of the serological method in taxonomic studies is controversial. Zucker- man & Sudermann (1935) believe “the serum precipitin test is of limited value in tracing phylogenetic relationship.” Boyden, on the other hand, in his publications states that he considers the method an important and reli- able tool for phylogenetic studies. The data presented here offered evidence in support of the idea that the precipitin test can, at times, be of use in the corroboration of morphological facts. Secondly, it seems that it may be a better method for determination of the degree of interrelationships among animals in the smaller taxonomic group, such as an order or family. Especially is this true when certain fac- tors, such as protein concentration of the antigen solutions, are determined and the specificity of the antisera is controlled by injection methods and in vitro factors. Of the eight species of Bovidae whose bloods were tested it was shown that the sheep, goat and tahr were very closely related. Their classification into the subfamily Caprinae would be justified by the serological reactions. The bloods of three antelopes, the eland, black buck and sable (each of which is placed in a separate subfamily) did not give similar reactions with sheep, goat or tahr antisera, the sable antelope reactions being definitely higher, indicating its closer kinship to the sheep, goat and tahr. The degree of this relationship was less than that of the sheep, goat and tahr to each other. The reactions of the anti-eland sera made possible a distinction of the eland blood from that of all other Bovidae and of the Cervidae. Its relation- ship values were by far the highest with the ox and buffalo bloods, indicating a closer affinity of the eland to these two animals than to the other animals studied. Beddard states (1920, p. 308), “Such an antelope, however, as the eland, is very ox-like in habit.” This similarity is, of course, based on a very superficial characteristic, and the addition of the similar serologic qualities of the animals is important. There was some indication that the eland may be more closely related serologically to the three Cervidae than to the Bovidae, with the exception of the ox and buffalo. This would be contrary to their present taxonomic position. Several antisera were produced against ox and buffalo blood sera. The results were conflicting, however, and it was believed advisable to post- pone a report of the data until further research could ascertain the cause of the discrepancies. The ox and buffalo antisera reactions always showed the very close affinity of the ox to the buffalo, and the value of the cross reaction was almost always 50 to 100 per cent. On the other hand, the titers with the eland, other Bovidae, and the Cervidae were dissimilar and inconsistent. In most instances the eland and Cervidae showed a closer reiacionsnip to the ox and butialo than did the other Bovidae; less often the opposite results were secured. The antisera produced against the black buck antelope resulted in reac- tions that easily distinguished it from the other bloods tested. The degree of the heterologous tests suggests that this animal was sufficiently different from the other Bovidae tested to warrant its being classified as a definitive group within the family Bovidae. The results of the reactions of the anti-Cervidae sera were elaborated upon in the presentation of the data. It was conclusively shown that the American elk and axis deer, an old world deer, were very closely related to each other. The degree of this relationship was of the same order that was found among sheep, goat and tahr, or between ox and buffalo. In conclusion a brief serologic classification, based upon the values of the Bovidae and Cervidae antisera reactions, is presented. The homologous 320 Zoologica: New York Zoological Society [XXIV :16 serum is considered to belong to Group I and it may be subdivided into sub- groups A and B. This group includes, in addition to the homologous blood, those heterologous bloods that consistently gave the highest cross values. The bloods which were usually indistinguishable with the less specific antisera would belong in subgroup A (refer to cross reactions of serum 41 with sheep, goat and tahr) . Group II would contain the species whose reactions are consistently lower than those of Group I, and Group III would include the lowest titered reactions. Table 9 presents the provisional serologic classification of the eight Bovidae and three Cervidae. Table 9. A provisional serologic classification of eight Bovidae and three Cer- vidae. Classification based on antisera against sheep, goat and tahr bloods : Group I Subgroup A. Sheep, goat, tahr. Subgroup B. Sable antelope. Group II — Other Bovidae tested. Group III — Cervidae. Classification based on anti-eland sera : Group I Subgroup A — Eland. Subgroup B- — Ox and buffalo. Group II — Other Bovidae tested and Cervidae. Classification based on anti-black buck antelope sera : Group I — Black buck antelope. Group II — Sheep, goat, tahr, sable antelope. Group III — Other Bovidae tested and Cervidae. Classification based on anti-ox and anti-buffalo sera: Group I — Ox and buffalo. Others doubtful. Classification based on anti-deer sera : Group I Subgroup A — Axis deer and elk. Subgroup B — Virginia deer. Group II — Ox and buffalo. Group III — Other Bovidae tested. Summary. 1. Nineteen antisera against five species of Bovidae and three species of Cervidae were produced in rabbits. 2. The antigens of fifteen species of Artiodactyla and Perissodactyla were used as test antigens. 3. Many of the antisera which were produced by injections of minute quantities of antigen resulted in very specific antisera. 4. The advantage of the serologic method is that a quantitative rela- tionship could be determined for closely related bloods. 5. The serologic relationships usually agreed with the accepted mor- phological classification. 6. There were indications that the ox and buffalo were more closely related to the Cervidae than were the other Bovidae tested. 7. The tahr and sable antelope gave higher relationship values with the goat and sheep antisera than did the other Bovidae and Cervidae. 1939] Wolfe: Serologic Relationships 321 8. The ox and buffalo showed the closest affinity to the eland. 9. The axis deer and elk sera were found to be indistinguishable from each other but distinguishable from the Virginia deer. Bibliography. Beddard, Frank Evers. 1920. Mammalia. Cambridge Natural History. Macmillan & Co., Ltd., London. Boyden, Alan A. 1926. The precipitin reaction in the study of animal relationships. Biol. Bull., Vol. 50, No. 2, p. 73. Ehrhardt, Albert. 1929. Die Verwandtschaftbestimmungen mittels der Immunitatsreaktionen in der Zoologie und ihr Wert fur phylogenetische Untersuchungen. Ergebn. u. Fortschr. der Zool., Vol. 7, p. 279. Myers, W. 1900. On immunity against proteids. Lancet, Vol. 2, p. 98. Nuttall, G. H. F. 1904. Blood immunity and blood relationships. Cambridge Univ. Press, Cam- bridge, England. Nuttall, G. H. F., & Dinkelspiel, E. M. 1901. Experiments upon the new specific test for blood. Preliminary Note. Brit. Med. Jour., Vol. 1, p. 1141. Scott, Wm. B. 1937. A history of land mammals in the western hemisphere. Rev. ed. Mac- millan Co., New York. Uhlenhuth, P. 1901. Eine Methode zur Unterscheidung der verschiedenen Blutarten, in besondern zum differential-diagnostischen Nachweise des Menschen- blutes. Deutsche Med. W ochenschr., Vol. 27, p. 82. Wolfe, Harold R. 1933. Factors which may modify precipitin tests in their application to zoology and medicine. Physiol. Zool. Vol. 6, p. 55. 1935. The effect of injection methods on species specificity of serum pre- cipitins. Jour, of Immun., Vol. 29, p. 1. 1936. The specificity of precipitins for serum. Ibid, Vol. 31, p. 103. 1939. Standardization of the precipitin technique and its application to studies of relationships in mammals, birds, and reptiles. Biol. Bull., Vol. 76, No. 1, p. 108. ZUCKERMAN, S., & SUDERMANN, ANN E. 1935. Serum relationships within the family Cercopithecidae. Jour. Exp. Biol., Vol. 12, p. 222. Elek & Finkelstein: Multiceps serialis in a Baboon 323 17. Multiceps serialis Infestation in a Baboon1. Report of a Case Exhibiting Multiple Connective Tissue Cystic Masses. Stephen R. Elek, M.D. & Leonard E. Finkelstein, M.D.2 Mount Sinai Hospital, New York City (Plates I & II). Introduction. Sporadic reports have appeared in the literature concerning the occur- rence of the coenurus, Multiceps serialis, in human and sub-human primates. Infestation with related members of the family Taeniidae, the genera Taenia and Echinococcus, is fairly well known. Accordingly, a pathological analysis of a case in a monkey is of interest not only because of its “considerable potential significance in human medicine” (10), but also because of a dearth of detailed post-mortem study of the disease. This paper is a report of a coenurus, which produced a large intra- abdominal mass together with numerous connective tissue and intermuscular cystic tumor masses in a baboon, Theropithecus gelada. Case Report. The animal was a young unexhibited baboon about 4 years of age. It was born in captivity in nanover, Germany, ana was Drought to this country about two to three months before death. For a few months the animal ap- peared to be ailing, lost weight, was noticeably weak, and had developed several small ulcerations on the dorsum of the right foot. A large mass was palpable in the abdomen. A short time before death spastic paralysis of the right lower extremity was noticed, the leg being kept flexed against the body. The animal was destroyed by its owner and the necropsy per- formed by us on April 27, 1939, about twenty-four hours after death. Post-mortem Examination. The body is that of a well developed, cachectic baboon ( Theropithecus gelada ) about four years of age. Moderate post-mortem rigidity is present. In each pectoral region there is a subcutaneous cystic mass, which is ap- proximately the size and shape of a hen’s egg. The one on the left is slightly larger. The overlying skin is freely movable. Similar, but smaller cystic masses are palpable in the right mastoid region, in the left axilla and over- 1 From the Laboratories of The Mount Sinai Hospital and the New York Zoological Park. 2 Theodore Escherich Fellow in Pathology. Mount Sinai Hospital. 324 Zoologica: New York Zoological Society [XXIV :17 lying the upper border of the left scapula. An incision over the mass in the left pectoral region reveals a well encapsulated, soft, multiloculated cystic structure lying within the subcutaneous tissues. The capsule is fairly thin and gray. Upon section, numerous small round clear cysts (bladder worms) are extruded together with some clear colorless fluid. These bladder worms measure from 0.5 to 1 centimeter in diameter. The other above-mentioned cystic masses are morphologically identical. The lower portion of the abdomen, especially on the right side, bulges dis- tinctly and on palpation a large mass can be felt. The right lower extremity is partially flexed at the hip and completely flexed at the knee. On the dorsum of the right foot, overlying the metatarso-phalangeal joints and also over the tarsal bones, there are several shallow ulcers with even, regular edges. The majority of these are approximately 1 centimeter in diameter and about 0.4 centimeter in depth. A similar but larger lesion is present over the medial aspect of the hallux, measuring about 4 by 1 centimeter in diameter. The bases of the ulcers are grayish and have a dry slough. Thorax. Neither free fluid nor adhesions are present in either pleural cavity. In the left paravertebral region, just above the dome of the dia- phragm, and extending over the lower four ribs, there is a firmly adherent multilocular cystic mass which measures approximately 6 by 5 centimeters in diameter. This rests upon the costovertebral junctions and portions of the adjacent ribs. At its infero-lateral margin it lies anterior and internal to another cystic mass, which is situated within the lowermost intercostal space, bulging externally, and which measures about 5 by 6 centimeters. This tumor is 2 centimeters in thickness and is located within the intercostal muscle at this site The intra-pleural cystic mass penetrates the pleura to communicate with this intermuscular cystic mass. At no place do the cysts invade the ribs, vertebral column or spinal cord. The heart, mediastinum and diaphragm are grossly uninvolved. Abdomen. Upon opening the peritoneum a large cystic mass presents, occupying the entire right lower quadrant, its upper pole being at the liver edge and its lower pole being at the brim of the true pelvis. It is covered anteriorly by the posterior parietal peritoneum, and is entirely encompassed by a well defined translucent capsule. Posteriorly it is firmly adherent to the muscle and connective tissue overlying the right wing of the ilium and the lower portion of the spine. The peritoneal surfaces, elsewhere, are smooth and glistening and there is no free fluid in the peritoneal cavity. The viscera have their usual position and relationships except that the right kidney and ureter and the ascending colon are displaced to the left by the cystic mass. The mesenteric nodes are similar to the pulmonary nodes de- scribed below. Lungs. The lungs weigh 190 grams, have a smooth and glistening pleura, and retain their shape fairly well upon the table. The upper lobes are light pink in color, while the lower lobes are dark purple. Crepitancy is normal throughout. The entire surface is studded by discrete, pin-head sized, pale gray, slightly elevated nodules, which on section extend into the depth of the lung for a distance of about 1 millimeter. The largest of these occurs in the right middle lobe, measures 5 millimeters in diameter and feels firm and calcified; on section, it has a caseous core. The cut surface of the lung is grossly negative, except for the presence of small gray pin- head sized nodules which are distinctly visible throughout. The hilar and tracheo-bronehial nodes are slightly enlarged, firm and adherent to each other. On section, each presents a peripheral grayish-white thick zone, which in some is calcified, and a central cheesy-like yellowish mass. Liver. The liver surface contains a few pin-point to pin-head sized flat grayish nodules identical with those seen in the lung. A small firm nodule, 2 millimeters in size, is likewise present. It is translucent and contains within 1939] Elelt & Finkelstein: Multiceps serialis in a Baboon 325 it two pin-point opaque yellow dots. On cut surface the hepatic architecture is regular and distinct. The remaining abdominal viscera and the genitalia are grossly not re- markable. Abdominal Tumor. This large oblong cystic mass measures 21 centi- meters in length, 12 centimeters in width, 7 centimeters in depth and weighs 815 grams (PI. I, Fig 1). The anterior surface is smooth and presents several well demarcated lobulations, which contain clear colorless fluid and bladder worms (larval tapeworms). On opening into the mass, it is found to be multiloculated and honeycombed due to the presence of num- erous closed and intercommunicating chambers, which also contain many small and large bladder worms. The cystic mass infiltrates the underlying muscle, but does not invade the bone. Bladder Worms. The bladder worms vary in shape, most of them being elliptical, while others are sperical or entirely irregular and branching. Many of them occur in grape-like clusters; some are pedunculated. These bladder worms possess a pale, milky-white, transparent, delicate wall, and contain clear, colorless fluid together with multiple heads, both invaginated and evaginated (PI. II, Fig. 4A). A fully detailed description of this coenurus has been given by Schwartz (10). Extremities: Upper. Both upper extremities show similar changes. On the medial and volar aspect of each there is a large multiloculated, well en- capsulated, cystic mass, measuring approximately 12 centimeters in length and 5 centimeters in width, which extends along the entire length of the forearm, its upper pole being about 3 centimeters distal to the olecranon. It is adherent to the muscles and tendons of the forearm, and compresses and flattens them. Numerous bladder worms are visible in the loculations. Lower: Left. Between the larger muscle groups of the posterior aspect of the thigh there is an oval, translucent, well encapsulated cystic structure, approximately 8 by 4 by 3 centimeters, which compresses, flattens, and is bound to the muscles of the adjacent region. A cyst, 3 centimeters in diameter, is present in the connective tissue of the popliteal space. A similar mass is present in the lower leg and is adherent to the muscles and tendons immediately posterior to the tibia. It displaces the gastrocnemius posteri- orly, which has become a thin compressed band of muscle (PI. I, Fig. 2, and PI. II, Fig 3). Right. Cystic masses, similar in all respects to those de- scribed in the lower extremity, diffusely infiltrate the subcutaneous connec- tive tissue and the inter-muscular fascia, producing pressure atrophy of the muscles. The tail, on section, contains many very small cysts within the con- nective tissue. The brain and spinal cord are grossly uninvolved. Microscopic Examination. Typical fresh miliary tubercles are found in the liver, lung and spleen. A fibro-calcific area containing lymphocytes and giant cells is present in the liver just beneath the capsule. Smaller collections of large mononuclear cells are scattered in the sinusoids ; a moderate number of lymphocytes and plasma cells are found in the portal fields. In the spleen there were many plasma cells in the pulp and moderate erythro- and sidero-phagocytosis. The tracheo-bronchial and mesenteric lymph nodes show evidence of old and re- cent tuberculosis with calcification. The left axillary node shows caseation. Ziehl-Neelsen stain failed to reveal any acid-fast organisms. The connec- tive tissue adjacent to the right wing of the ilium and located just beneath the lower pole of the abdominal mass shows acute and chronic non-specific inflammation with focal necrosis. 826 Zoologica: New York Zoological Society [XXIV: 17 The muscle immediately adjacent to the cystic mass in the extremities reveals marked degeneration and atrophy, some edema, mild replacement fibrosis and focal collections of lymphocytes together with evidences of re- pair as indicated by the presence of many muscle buds. Section of the bone ( right wing of the ilium) is entirely normal. The contiguous muscle shows the changes described above. The bases of the ulcers of the right foot are formed by non-specific granulation tissue. Comment. It is to be noted that M. serialis infestation evokes no characteristic or specific inflammatory reaction in its sites of localization, and, indeed, very little inflammatory reaction altogether. The changes in the muscles result- ing from the presence of the interfascicular cystic masses are those due to pressure. In passing, it may be stated that calcification of tuberculous foci, as found in our case, is uncommon in monkeys. Laboratory Data. The fluid within the cysts gave a strongly positive (four plus) comple- ment fixation reaction for Echinococcus. X-ray of the extremities showed cystic masses, with streaks of calcification, not invading the bone. X-ray of the abdominal mass revealed only a multicystic tumor without calcifica- tion. Post-mortem roentgenogram of the lung showed numerous calcified para-tracheal nodes. Guinea pig innoculation for tubercle bacilli was un- satisfactory. A fresh preparation of a coenurus revealed the presence of many hooklets (PI. II, Fig. 4B). Pathologic Diagnosis : Widespread infestation by a coenurus (M. serialis) resulting in multiple connective tissue and inter-muscular cystic masses located in all extremities, retroperitoneally, intra-pleurally, in the anterior and posterior chest wall, right mastoid region, and the tail. Gen- eralized miliary tuberculosis of the lung, liver, spleen and old and recent tuberculosis of the axillary, hilar, tracheo-bronchial and mesenteric lymph nodes are present, with calcification of the latter two. Degeneration, atrophy and fibrosis of the involved muscles of the extremities. Trophic ulcers of the right foot. Discussion. The larval cestode described above was sent to the Bureau of Animal Industry3, U. S. Department of Agriculture, where it was identified by Mr. Allen McIntosh as Multiceps serialis theropitheci (Schwartz, 1927). M. serialis infestation in primates is rare. The literature on this sub- ject has been recently reviewed by Sandground (9), who listed five cases in primates and two instances in humans. The usual intermediate host is the rabbit. References to other intermediate hosts can be found in Meggitt (6) and Hall (5). The common definitive host is the dog. Schwartz (10) de- scribed a case of M. serialis infestation occurring in a baboon ( T . ohscurus) with a large cystic subcutaneous tumor in the right thoracic region similar to that found in our case, and gave a careful morphologic description of the larval and adult tapeworm. Schwartz adopted the name M. serialis var. theropitheci because, although the morphology of the cestode and of the tapeworm reared from a dog was similar to, if not identical with M. serialis, nevertheless there was a biological difference in that he was unable to obtain 8 We are greatly indebted to Mr. Allen McIntosh for kindly identifying the bladder worm for us and to Dr. E. W. Price, Acting Chief, Zoological Division, Bureau of Animal Industry, U. S. Department of Agriculture, Washington, D. C., for his cooperation. 1939] Elek & Finkelstein: Multiceps serialis in a Baboon 327 this cestode after feeding the adult worm to rabbits. In Scott’s (11) two cases, both occurring in Theropithecus gelada, one animal had cystic masses in the right upper arm, submental region, pericardium, mediastinum and right perirenal tissue, while in the other, a mass in the left pleural cavity had invaded and compressed the lower dorsal spinal cord. Sandground’s (9) case was atypical in location, the cyst being in the brain of a monkey ( Cercopithecus nic titans) .4 Railliet & Marullaz (8) reported cystic masses in the perineum of a monkey ( Macacus sinicus ) due to M. ramosus. Sand- ground (9), however, believes this is really M. serialis. In the two reported human cases (1, 2), one showed multiple subcutaneous tumors and the other a solitary cyst in the right buttock. In these cases, the coenurus was iden- tified by morphological study. According to Deve and the Registry of the Royal Australasian College of Surgeons (both cited by Godfrey (4), the incidence in humans of Echino- coccus infestation involving muscle and fascia is about 5%. Generally the diagnosis of Echinococcus disease is made by the mere finding of cysts without a detailed morphologic study of them and by various laboratory tests, including the complement fixation, intradermal and precipitin tests. These tests, however, are now recognized to be specific for related groups rather than for individual species (3, 7). Since in our case of M. serialis we obtained a positive complement fixation for Echinococcus, we are led to consider the possibility that infestation by M. serialis in man may be more frequent than commonly noted. This view gains credence since as Schwartz (10) has stated: “The question of the specific identity of the coenurus stage of the tapeworm genus Multiceps involves primarily the number, size and shape of the hooks and incidentally the other head structures, notably the suckers and rostellum.” Such an analysis is not a routine procedure in most hospital laboratories. Summary. 1. A case is reported of infestation in a baboon ( Theropithecus gelada ) by a cestode identified as Multiceps serialis. 2. There were numerous cystic masses in the subcutaneous and inter- muscular connective tissues and a large intra-abdominal, retro-peritoneal cystic tumor,, 3. This type of cestode infestation is rare in primates. So far as could be determined, this is the sixth such case to be reported in a primate. Two instances have been described in humans. 4. M. serialis infestation produced no specific or characteristic inflam- matory response in our case. 5. An incidental finding was a fibro-caseo-calcific tuberculosis with miliary dissemination. 6. Fluid obtained from the coenurus gave a strongly positive comple- ment fixation test for Echinococcus. It is suggested, therefoi'e, that some cases thought to be “Echinococcus" infestation in humans, with cystic masses in connective tissue and in muscle, may actually be due to M. serialis. The authors are indebted to Dr. Charles R. Schroeder for his kind cooperation; and Drs. Paul Klemperer and Sadao Otani for their valuable criticism and helpful aid in the preparation of this report. Bibliography. 1. Bonnal, G., Joyeaux, C., & Bosch, P. 1933. Un cas de cenurose humaine du a Multiceps serialis (Gervais). Bull. Soc. Exot., Paris, 26, 1061. 4 It may be mentioned here that some authors, notably Southwell (10), believe that no sharp distinction can be made morphologically between M. serialis and M. multiceps. B28 Zoologica: Neiv York Zoological Society 2. Brumpt, E., Duvoir, M. E., & Sainton, J. 1934. Un cas de cenurose humaine du au Coenurus serialis. Ann. de Parasit., 12, 371. 3. Culbertson, J. T. 1938. Recent contributions to the Immunology of Helminthic Infections. Arch, of Path., 25, 256. 4. Godfrey, M. F. 1937. Hydatid Disease. Arch. Int. Med., 60, 782. 5. Hall, M. C. 1919. The Adult Taenoid Cestodes of Dogs and Cats, and Related Carnivores in North America. U. S. Natl. Mus. Proc., 55, 1-94. 6. Meggitt, F. J. 1924. The Cestodes of Mammals. 282 pp. London. 7. Outeirino, J. 1935. Des recherches sur la pretendue specificite des reactions de “Ghedini- Weinberg” et de “Casoni” dans le diagnostic de l’echinococcose hu- maine. Ann. de Med., 38, 493. 8. Railliet, A., & Marullaz, M. 1919. Sur un Cenure nouveau du Bonnet chinois ( Macacus sinicus). Bull. Soc. Path. Exot., 12, 223. 9. Sandground, J. H. 1937. On a Coenurus from the Brain of a Monkey. J. Parasit., 23, 482. 10. Schwartz, B. 1927. A Subcutaneous Tumor in a Primate Caused by Tapeworm Larvae Experimentally Reared to Maturity in Dogs. J. Agric. Research, 35, No. 5, 471. 11. Scott, H. H. 1926. Report on Deaths Occurring in the Society’s Garden During the year 1925. Proc. Zool. Soc. London, 1, 240. 12. Southwell, T. 1930. Fauna of British India. Cestoda, Vol. II. Taylor and Francis, Fleet- street, London. EXPLANATION OF THE PLATES. Plate I. Fig. 1. Large intra-abdominal cystic mass, which measures 21 by 12 by 7 centi- meters. Weight 815 grams. Anterior surface covered by peritoneum. Fig. 2. Left lower limb showing cystic masses displacing muscle in thigh and leg. Plate II. Fig. 3. Left leg, higher power view, showing encapsulated cystic mass with muscular atrophy. Arrow points to muscle. Fig. 4A. Right forearm. Cystic masses displacing muscle and showing extruded bladder worms. Fig. 4B. Photomicrograph of bladder warm showing hooklets (fresh preparation). ELEK ft FINKELSTEIN. PLATE I. FIG. 1. FIG. 2. MULTICEPS SERIALIS INFESTATION IN A BABOON. ELEK a FINKELSTEIN. PLATE II. FIG. 3. FIG. 4A. FIG. 4B. MULTICEPS SERIALIS INFESTATION IN A BABOON. Knepp: Testes of Certain Mammals 329 18. A Quantitative Study of the Testes of Certain Mammals. Thomas H. Knepp. (Text-figures 1 & 2). A quantitative histological study of the testes of six mammals, four domestic and two wild, was undertaken in 1935, and the research problem initiated at that time has been continued through the cooperation of the Laboratory and Hospital of the New York Zoological Park in supplying me with testes of mammals which died in the Zoological Park. The testes were fixed in 10% formalin, except those of the gray squir- rel and cottontail rabbit which were fixed in Bouin’s picro-formoi. Each testis was weighed after fixation and hardening. Material was sectioned in paraffin, then stained with Heidenhain’s iron-hematoxylin. Microscopic measurements were made with an ocular micrometer. In finding the percentage of interstitial tissue and seminiferous tubules the paper method employed by Bascom (1925) was used. This method does not take into account the thickness of the tunica albuginea and mediastium in calculating the amount of seminiferous tubules and interstitial tissue. The average of 20 measurements was used in computing the diameter of the seminiferous tubules and the thickness of the tunica albuginea. In finding the average size of the sperms 25 measurements were taken. Testes from the polar bear, Punjab wild sheep, mouflon and wallaroo were from the Zoological Park; testes from the Virginia deer, gray squirrel and cottontail rabbit were from the fauna of Pennsylvania. In the table below, the date the specimen was secured is given. It remains to be proved whether the season of the year has anything to do with spermatogenesis and quantitative measurements. It is evident that the testicular weight bears no ratio to the gross weight of the animal. Tubular diameter is not as great in the carnivore (polar bear) as in the six herbivores; likewise the cross-section area of the tubules is smaller in the carnivore. The polar bear and gray squirrel are at the extremes in thickness of tunica albuginea and connective tissue. The smallest testis in weight has the thinnest tunica albuginea and con- nective tissue, but the greatest thickness is not found in the heaviest testis. The percentage of interstitial tissue is greatest in the polar bear, but relatively constant in the herbivores ; the inverse is true of the seminiferous tubules. The length of the seminiferous tubules is greater in the largest testes, the weight of the testis being due to the increased length of the tubules. In meters of tubules per gram of testicular weight, and in meters of tubules per gram of tubules, the carnivore is the extreme. Sperm sizes vary from the smallest in the Virginia deer to the largest in the wallaroo, the intermediate ones being fairly constant. The heads of the sperms of all the animals except the wallaroo are somewhat oval; those of the wallaroo appear 330 Zoologica: Neiv York Zoological Society [XXIV :18 to be pointed and thin, with the end of the head somewhat curved. In the wallaroo no loose sperms are seen in the lumen of the tubules ; the tails are in large masses, making the individual study of a sperm difficult. This may be due to the fact that the animal was immature. Polar Bear Punjab Wild Sheep Mouflon Virginia Deer Wallaroo Gray Squirrel Cottontail Rabbit Date secured 11/2/37 1/6/38 1/6/38 12/4/37 2/17/38 11/25/33 5/22/35 Weight of the testis with- out epididymus, grams 31.7 70.2 41.05 36.2 5.8 1.97 9.6 Mean tubule diameter, microns 105.8 176.4 149.5 139.4 141.1 169.68 173.88 Mean cross-section area of tubules, square mil- limeters 0.008 0.024 0.017 0.014 0.015 0.022 0.023 Mean thickness of tunica albuginea and connec- tive tissue, microns 708.3 555.2 473.7 168.0 164.6 99.9 374.64 Interstitial tissue, per- cent. 33.5 15.86 15.45 17.94 21.96 18.44 18.69 Seminiferous tubules, percent. 66.5 84.14 84.55 83.06 78.04 81.55 81.30 Length of seminiferous tubules, meters 2635 2461 2041 2121 301 73 339 Meters of tubules per gram of testicular weight 83.1 31.5 49.7 58.5 52.0 37.1 35.3 Meters of tubules per gram of tubules 125.5 41.7 58.8 71.4 67.1 45.6 43.5 Length of average sperm, microns 36.3 24.3 26.6 18.9 49.4 27.47 29.57 Text-figures 1 & 2 (gray squirrel and cottontail rabbit) illustrate the structural make-up of the respective testes. Definite pyramidal arrange- ment is evident in the rabbit testis, the base of the pyramid resting on the tunica albuginea, the apex pointing toward the center. (This arrangement is also evident in the common dog). From the Zoological Park’s Laboratory and Hospital there have been sent to me testes from the collared peccary, mouflon, white-tailed para- doxure, Hussar monkey, woolly monkey, tahr, pigmy hippopotamus, Kadiak bear, axis deer, black buck antelope and North and South African ostrich. Since this group includes carnivores, herbivores and omnivores, the data when prepared may show some interesting facts. Bibliography. Bascom, Kellogg F.: Quantitative studies of the testis. Anat. Rec., Vol. 30, No. 3. June. 1925. Bascom, Kellogg F., & Osterud, Hjalmar L.: Quantitative studies of the testes. Anat. Rec., Vol. 31, pp. 159-169. 1925. 1939 J Knepp : Testes of Certain Mammals 331 Text-figure 1. Section of Gray Squirrel testis. Tubules not in any definite arrangement. Curtis, G. M.: The morphology of the mammalian seminiferous tubules. Amer. Jour, of Anat., Vol. 24. 1918. Huber, G. G., & Curtis, G. M.: The morphology of the seminiferous tubules of the mammalia. Anat. Rec., Vol. 7. 1913. Knepp, Thomas H.: Comparative study of the testes of the dog, cat, sheep, bull, cottontail rabbit and grey squirrel. Pro. Pa. Acacl. of Sci,, Vol. X. 193fi. Knepp, Thomas H.: A quantitative study of the testes of five mammals. Pro. Pa. Acad, of Sci., Vol XII. 1938. Moore, Carl R. : The biology of the mammalian testis and scrotum. Quart. Rev. of Bio., Vol. 1. 1926. 332 Zoolog ica: New York Zoological Society Text-figure 2. Section of Cottontail Rabbit testis. Tubules in pyramidal arrangement, with bases of pyramids resting on tunica albuginea. Tiemeier: Os Opticus of the Bird’s Eye 333 19. A Preliminary Report on the Os Opticus of the Bird’s Eye. Otto W. Tiemeier University of Kansas Museum of Birds and Mammals, Lawrence, Kansas. (Text-figures 1-24). In the eyes of many birds there is a small U- or horseshoe-shaped bone, located in the rear portion of the cartilaginous sclerotic coat of the eye. The open part of the U is toward the upper part of the eye, and through this the optic nerve passes. In some instances the bone may be more or less circular in form and thus completely surround the optic nerve. Upon the suggestion of Mr. C. D. Bunker of the University of Kansas Museum of Birds and Mammals, an investigation of this bone was made. The purpose of the investigation was to consider its possible taxonomic value, the forms in which it was present, the extent of variation and its probable function. An examination of the literature revealed that the bone was first described by Gemminger (1853). He noticed that it was present in the eyes of woodpeckers. In his list of 20 forms of European birds he describes and illustrates it in a number that were not woodpeckers. Leydig (1855) adds 13 forms to this list. Later investigators have added a few more forms. Concerning the function of this element, Gemminger, whose work was mostly with the woodpeckers, writes that this bone is for protection of the optic nerve at its entrance into the eyeball. I am sure that the shock which the eyeball receives as these birds hammer holes in trees must indeed be considerable, but it appears to me that this does not completely answer its purpose. Gemminger failed to find the bone present in the nocturnal and diurnal birds of prey, in the gallinaceous birds, and what he calls the swamp and swimming birds. The fine collection of skeletons in the University of Kansas Museum of Birds and Mammals furnished most of the material. I have examined 6,500 skeletons in the above collection. Other material was obtained through the courtesy of Dr. C. R. Schroeder of the New York Zoological Park and from Mr. C. C. Sperry and Mr. Ralph H. Imler of the Food Habits Research Laboratory at Denver. The method used to secure this very small bone was the dfermestid beetle process. Specimens of birds were skinned and drawn, thoroughly dried and then left in the “bug room” to become infested. Within several months the skeletons were entirely cleaned by the beetles and the bones could be picked out of the debris. If, in particular forms, I was certain that the bone was present but I could not find it in the collection of skele- tons, the eyes of fresh specimens coming into the museum were saved and dried, and then placed in small glass containers with a number of small 334 Zoologica: New York Zoological Society [XXIV: 19 beetle larvae. A week later the inconspicuous bone could be separated from the debris of cast-off beetle remains, if it had ever been present in the sclera of the eye. Former investigators have described this bone under the misleading name of “the rear sclerotic ring,” which would seem to indicate an associa- tion with or a relation to the sclerotic ring. In reality there is very little relation between the two, except that they are both located in the sclerotic coat of the eye. I wish to propose the name “os opticus” which, I believe, will help to clarify and differentiate the two. Microscopic sections of the os opticus show that it can be differentiated from the sclerotic ring in that it has a marrow cavity which contains fat and marrow cells and blood vessels. In the sections that I have made of the bone I have always found the cells and vessels. Franz (1934) did not find them in Motacilla. The os opticus does not consist of plates like that of the sclerotic ring, although in some instances it has been described as con- sisting of two or three separate bony elements. Leydig writes that the origin of the two bony elements of the eye are different. The sclerotic ring, he states, is formed by calcification of the con- nective tissue, while that of the os opticus is formed by calcification of the hyaline cartilage of the sclera. I have not been able to verify this statement. There is a great variation in the size, shape and development of the os opticus. In general it is U-shaped or horseshoe-shaped with variations as to the development of the two heels. One heel may be long and well developed and the other one short. There is often very little symmetry and by a comparison of the bone in the right and left eye, Text-figs. 8 and 9, it would often be difficult to conceive of them as coming from the same bird, except for the ground pattern of the U that is easily identifiable. In others they may be exact mirror images and may be definitely designated as left and right. By securing the eyes of fifty game cocks and using the dermestid method on them, I have been able to obtain a very interesting series of bones from the eyes of this species, although Gemminger and Leydig stated that it was not present in the domestic fowl. In Text-figs. 1 to 6 I have drawn a series of these bones. Text-figs. 1 and 2 show the basal plate which is apparently the first portion to develop. As the series progresses the heels of the horse- shoe develop until finally they are in their mature forms. The bone was apparently not present in all of the eyes, although the birds were mature enough to have been entered in cock fights. This seems to indicate that the development of the os opticus is an age character, and may be formed after the birds are mature. Gemminger and Leydig report that they have found the bone in nestling woodpeckers that they examined. Possibly in forms like the woodpeckers the bone develops early in the life of the bird, and in those forms in which it is not so well developed it appears later in life. By comparing the different families of birds it appears that the os opticus is best developed in the Picidae, although it is also prominent in the Corvidae and the Fringillidae. In the birds which I have examined, the bone is largest in the Pileated Woodpeckers where the measurement is 1 cm. in the longitudinal axis and .73 cm. in the vertical axis. The measurement of the bone in the fully mature Brown Creeper is .16 cm. by .15 cm. and is the smallest of any of the species represented in my list. In many instances the bone forms a complete ring around the optic foramen. The optic foramen is a slit whose vertical distance is about twice that of the horizontal distance. This accounts for the U-shape of the os opticus. In those species in which the bone is well developed it is curved to correspond to the curvature of the eyeball. In woodpeckers, especially, there is quite frequently an enlarged develop- ment of bone at the upper portion of the heel. Text-figs. 7 and 9. I have not Fig. 1 Fig. 13 Fig. 21 Fig. 2 Fig. 6 Fig. 14 Fig. 18 Fig. 22 Fig. 15 Fig. 23 Text-figures 1-24. Fig. 12 Fig. 16 Fig. 24 I to 6. Gallus sp.( X 5.5; 7. Ceophloeus pileatus pileatus, X 3; 8. Corvas brachy- rhynchos brachyrhynchos, X 5; 9. Dryobates villosus villosus, X 4; 10. Buteo borealis calurus, X 4; 11. Colaptes auratus luteus, X 5; 12 & 13. Melanerpes erythrocephalus, X 5; 14. Falco sparverius sparverius, X 4; 15. Sturnella magna magna, X 5; 16. Hedymeles melanocephalus papago, X 6; 17. Tyrannus tyrannus, X 5; 18. Butorides virescens virescens, X 8; 19. Phasianus colchicus torquatus, X 6; 20. Piranga ludoviciana, X 6; 21. Toxostoma rufum, X 5; 22. Dumatella carolinensis, X 6; 23. Stumus vulgaris , X 5; 24. Dendroica fusca, X 7. 336 Zoologica: New York Zoological Society [XXIV :19 been able to determine whether this development is separate from that of the bone proper or whether it is only a lobular development of the bone itself. I am rather inclined to think that it is the latter because it is not uniform in position. Sometimes the lobe may be attached to the basal portion of one of the heels or in any position between these two extremes. In any event it does not seem to be a constant development. Gemminger reports the presence of this element in 20 forms of Euro- pean birds. Leydig later adds 13 forms to the list. In my material I find it to be present in 152 different forms of North American birds. I feel cer- tain that it is also present in a large number of species belonging to orders represented in my list but which are absent because of lack of material. I have included a list of the'forms in which Gemminger and Leydig have reported the bone, and a list of forms Reported by Gemminger Dry copus martinus Gecinus viridis Gecinus canus Picus minor Picus medius Picus major Apternus tridactylus Corvus corax Corvus cornix Corvus corone Corvus frugilegus Corvus monedula Pica caudata Garrulus glandaris Silla europaea Certhia familiaris Tichdroma muraria Parus ater Pyrrhula rubicilla Following is a list of forms in os opticus. Butorides virescens virescens Accipiter velox velox Buteo borealis borealis Buteo borealis calurus Falco sparverius sparverius Falco sparverius phalaena Gallics sp. Phasianus colchicus torquatus Syrmaticus reevesi Zenaidura macroura carolinensis Columbigallina passerina pallescens Megaceryle alcyon alcyon Colaptes auratus luteus Colaptes cafer collaris Colaptes chrysoides mearnsi Ceophloeus pileatus pileatus Centurus carolinus Centurus aurifrons Melanerpes erythrocephalus Sphyrapicus varius nuchalis Sphyrapicus thyroideus nataliae Dryobates villosus villosus Dryobates villosus monticola Dryobates pubescens medianus Dryobates pubescens pubescens Dryobates scalaris symplectus that I wish to report. Reported by Leydig Falco tinnunculus Muscipeta satelles Motacilla alba Turdus merula Sylvia phoenicurus Troglodytes gigas Passer domesticus Fringilla carduelis Fringilla caelebs Sturnus vulgaris Cassicus phoeniceus Trochilus Hirundo urbica which I have secured specimens of the Tyrannies vociferans Muscivora forficata Myiarchus crinitus boreus Sayomis phoebe Sayornis nigricans nigricans Empidonax minimus Empidonax difficilis difficilis Myiochanes virens Nuttallornis mesoleucus Otocoris alpestris leucolaema Otocoris alpestris praticola Hirundo erythrogaster Progne subis subis Perisoreus canadensis capitalis Cyanocitta cristata cristata Cyanocitta stelleri diademata Pica pica hudsonia Corvus corax sinuatus Corvus cryptoleucus Corvus brachyrhynchos brachyrhynchos Cyanocephalus cyanocephalus Penthestes atricapillus atricapillus Penthestes atricapillus septentrionalis Penthestes gambeli gambeli Baeolophus bicolor Sitta carolinensis carolinensis 1939] Tiemeier: Os Opticus of the Bird's Eye 337 Dryobates arizonae arizonae Tyrannus tyrannus Tyrannus verticalis Salpinctes obsoletus obsoletus Mimus polyglottos polyglottos Mimus polyglottos leucopterus Dumatella carolinensis Toxostoma rufum Toxostoma curvirostre curvirostre Oreoscoptes montanus Turdus migratorius migratorius Turdus migratorius propinquus Hylocichla mustelina Hylocichla guttata faxoni Hylocichla ustulata swainsoni Hylocichla minima aliciae Sialia sialis sialis Polioptila caei~ulea caerulea Regulus satrapa satrapa Corthylio calendula calendula Anthus spinoletta nibescens Bombycilla cedrorum Phainopepla nitens lepida Lanius ludovicianus migrans Lanius ludovicianus excubitorides Sturnus vulgaris vulgaris Vireo belli belli Vireo gilvus gilvus Vireo gilvus swainsoni Mniotilta varia V ermivora peregrina V ermivora celata celata Compsothlypis americana pusilla Dendroica aestiva aestiva Dendroica magnolia Dendroica coronata Dendroica auduboni auduboni Dendroica cerulea Dendroica fusca Dendroica striata Seiurus aurocapillus Oporornis formosus Geothlypis trichas trichas Icteria virens virens Passer domesticus domesticus Dolichonyx oryzivorus Sturnella magna magna Sturnella neglecta Agelaius phoeniceus phoeniceus Agelaius phoeniceus fortis Icterus spurius Certhia familiaris americana Thryothorus ludovicianus ludovicianus Heleodytes brunneicapillus couesi Icterus galbula Icterus bullocki Euphagus carolinus Euphagus cyanocephalus Quiscalus quiscula aeneus Molothrus ater ater Molothrus ater artemisiae Molothrus ater obscu'rus Piranga ludoviciana Piranga erythromelas Piranga rubra rubra Richmondena cardinalis cardinalis Richmondena cardinalis canicauda Pyrrhuloxia sinuata texana Hedymeles ludovicianus Hedymeles melanocephalus papago Guiraca caerulea caerulea Guiraca caerulea interfusa Passerina cyanea Spiza americana Carpodacus mexicanus frontalis Spinus pinus pinus Loxia sp. Loxia curvirostra pusilla Pipilo erythropthalmus erythropthalmus Pipilo maculatus arcticus Pipilo maculatus montanus Calamospiza melanocorys Ammodramus savannarum australis Passerherbulus caudacutus Pooecetes gramineus confinis Chondestes grammacus strigatus Aimophila cassini Junco hyemalis hyemalis Junco caniceps Spizella arborea arborea Spizella arborea ochracea Spizella pusilla pusilla Zonotrichia querula Zonotrichia leucophrys leucophrys Melospiza lincolni lincolni Melospiza georgiana Calcarius lapponicus alascensis Calcarius pictus Taeniogypio castanotis Serinus sp. To date I have not found the bone in a number of families of North American birds, namely: Gaviidae, Pelecanidae, Anatidae, Cathartidae, Tetraonidae, Perdicidae, Meleagrididae, Gruidae, Rallidae, Charadriidae, Scolopacidae, Laridae, Cuculidae, Tytonidae, Strigidae, Caprimulgidae, Micropodidae and Trochilidae. In some of these families the os opticus has been reported but because of the lack of material I am not including them. The most striking gaps in the list are the ones among the water-inhabiting birds and the owls. I have, however, found the bone in the Eastern Green Heron (Butorides virescens virescens ) and I believe that when I can secure more specimens some of these gaps will be filled. Possibly the development of the bone is a phylogenetic character in the development of the birds and therefore absent in the more primitive forms. Examination of the eyes of an ostrich and an emu revealed that the os 338 Zoologica: Neiv York Zoological Society opticus was not present in those particular specimens of the two forms. I am convinced that in the Order Passeriformes every family is represented by forms in which the bone appears. It is quite certain that the os opticus can not be used as a diagnostic character in the separation of species because there is too much variation within the same species. There may, indeed, be considerable variance in the two eyes of the same bird. Upon the assumption that the function of the os opticus is protection of the optic nerve at its entrance into the eyeball, it would be difficult to explain why it is so well developed in the flycatchers, swallows, and other birds whose methods of securing food are quite different from that of the wood- peckers. It would also be difficult to explain why in one order of birds, namely the Falconiformes, it seems to be universally present in the members of the Family Falconidae and only very seldom present in the Accipitriidae. It appears to be the concensus of opinion of former workers that one of the chief purposes of the pecten is for the nourishment of the vitreous and retina of the eye. I feel that there is a very definite relationship between the pecten of the eye and the os opticus other than that of spatial relation- ship. In many of the bones, that are exceptionally well developed, thei'e is a small opening on one side and at the base of the bone for the passage of the different blood vessels of the pecten. It is my intention to continue my investigation of the os opticus, par- ticularly the condition in the nestling, and if possible to ascertain specific changes due to age. I am greatly indebted to Mr. C. D. Bunker of the University of Kansas Museum who suggested this investigation and who placed at my disposal the extensive collection of bird skeletons for examination. I am also indebted to Dr. C. R. Schroeder of the New York Zoological Park, and to Mr. C. C. Sperry and Mr. Ralph H. Imler of the Food Habits Research Laboratory at Denver, for valuable study material. I wish to thank Dr. E. H. Taylor of the University of Kansas, for helpful suggestions and criticisms. References. Gemminger, M. 1853. Ueber eine Knochenplatte im hinteren Sklerotikalsegment des Auges einiger Vogel. Zeitschrift fur Wissenschaftliche Zoologie. Leipzig, 1853, pp. 215-220. Leydig, F. 1855. Der hintere Sclerotikalring im Auge der Vogel. Muller's Archives fur Anatomie. 1855, pp. 40-46. Franz, Victor 1934. Vergleichende Anatomie des Wirbeltierauges. In Handbuch der vergleichenden Anatomie der Wirbeltiere. Bei-lin, 1934, pp. 1133-1134. Herman & Schroeder: Dysentery in an Orang 339 20. Treatment of Amoebic Dysentery in an Orang-utan. Carlton M. Herman, Sc.D., & Charles R. Schroeder, D.V.M. Hospital and Laboratory . New York Zoological Park. On July 2, 1938, a male orang-utan ( Pongo pygmaeus) was purchased for the collection of the New York Zoological Park. The animal was a young male estimated about three years of age and weighing about 21 lbs. It had been captured in North Borneo, brought to Baltimore, Md., and finally bought by a New York dealer who sold it to the Park. During the first few weeks that this orang was in the Hospital and Laboratory for routine quarantine prior to being placed on exhibition, it had an intermittent diarrhea. Examination of a fecal sample on July 18, 1938, showed the presence of a few amoebic trophozoites with the morphology of Endamoeha histolytica. In a watery stool passed on July 20 both tropho- zoites and cysts were very numerous. On July 20 treatment was begun. Carbarsone (0.05 gms.) was admin- istered three times daily either mixed with milk or in a slice of banana. This treatment was continued for seven days. Smears of fecal material made during treatment seemed to indicate a marked decrease in the number of trophozoites present. By the end of the week of treatment the orang was passing well-formed stools in which no amoebae could be demonstrated by the direct smear method. Eleven days following the completion of the treatment the animal again became diarrhetic. A few trophozoites were demonstrable in a fecal smear. The Carbarsone treatment was repeated three times daily for three days, when the diarrhea had been corrected and no parasites could be seen in a fecal smear. Since this last treatment about a year ago fecal smears have been examined more or less periodically, particularly whenever a loose stool was passed, but no trophozoites or cysts of Endamoeha his- tolytica could be demonstrated. Since the animal arrived in the Park it has been kept under the best sanitary conditions possible with an exhibit primate. Its cage is cleaned at least daily and absorbent, dried, sugarcane stalks are used for bedding so that the risk of reinfection has been at a minimum. The authors believe this case represents complete elimination of Endamoeha histolytica from the orang-utan by treatment with Carbarsone. Olsen: Neiv Species of Cestode 341 21. Deltokeras multilobatus, a New Species of Cestode (Parauteri- ninae : Dilepiididae) from the Twelve-wired Bird of Paradise ( Seleucides m. melanoleucus (Daudin). : Passeriformes) .1,_ O. Wilford Olsen Minnesota Agricultural Experiment Station, St. Paul, Minnesota. (Plate I). A collection of cestodes obtained from a twelve-wired bird of paradise that had died at the New York Zoological Park was submitted to the author for study through the courtesy of Dr. Carlton M. Herman. The vial contained a large number of specimens, totaling about one hundred in all, approximately half of which are Hymenolepis brevicirrosa Fuhrmann, 1912, and the re- mainder an undescribed species of the genus Deltokeras Meggitt, 1927. The name Deltokeras multilobatus n. sp. is proposed, being suggested by the many- lobed condition of the ovary. Deltokeras multilobatus n. sp. Description : Strobilae up to 45 mm. long, proglottides nearly as thick as long, width of mature segments 470-525 p, length 292-324 p, thickness 243 p in posterior part; ripe segments 616-696 p wide by 393-486 p long. Scolex rounded, diameter 243-324 p, length 243-324 p; rostellum short, its length, 64-99 p, about equal to its width, 72-91 p, dome shaped with truncated part directed cephalad; because of their caducous nature only two rostellar hooks remained, length 17-19 p, ventral root short and with very broad base, its distal extremity provided with a large knob that is bisected, dorsal root with or without terminal knob. Suckers approximately circular, 87-102 p wide by 80-91 p long, located laterally near equator of scolex, unarmed. Neck dis- tinct, length 508-1060 p, width at base of scolex 146-190 p. Dorsal excretory canals small, about 4 p in diameter, ventral canals large, about 27 y in diam- eter, located immediately mesad from nerve cords, transverse canal in caudal part of segment. Genital pores unilateral, very small, 12 p in diameter, lo- cated near middle of lateral margin of proglottid. Genital ducts pass between excretory canals and dorsal to nerve cords. Cirrus pouches long and slender, extending anterio-mesad to a point beyond excretory canals, length 121-133 p, width 19-23 p; cirrus very slender, diameter 3 p, aspinose; cirrus and that portion of vas deferens within pouch looped; internal and external seminal vesicles absent; vas deferens forms a number of loops in anterior portion of proglottid before passing caudad over dorsal surface of ovary; testes pos- terior and lateral to ovary, 12-16 in number, diameter about 34 p, both testes and ovary persist until uterus is well filled with immature ova. Vaginal 1 Paper No. 1732 Scientific Journal Series Minnesota Agricultural Experiment Station, St. Paul. 2 In cooperation with the Minnesota Conservation Department, Division of Game and Fish. 342 Zoologica: New York Zoological Society [XXIV :21 opening caudad from male pore, vagina extends as a thin tube parallel with and on caudal side of cirrus pouch to beyond excretory canals where it bends meso-caudad, increasing greatly in diameter to form an elongated seminal receptacle that lies dorsal to ovary. Ovary located centrally in proglottid, multilobated, there being 8-9 pedunculated lobes filling greater portion of space between longitudinal excretory canals. Vitelline gland at caudal margin of ovary, oval, 61-76 g long by 38 g wide. Uterus with lobes, the interior being divided into compartments by extensions of the wall, fills entire intercanalular space ; parauterine organ very poorly developed, being represented by a very meager amount of tissue spread evenly over entire surface of uterine wall. Ova few, with three membranes, 53-57 g long by 30 g wide (sectioned speci- mens), embryoes 30-31 g long by 23-27 g wide, hooks of embryo 13 g long. Ova not encapsulated in uterine cavity. Host: Seleucides melanoleucus melanoleucus (Daudin). Habitat: Intestine. Locality: New York Zoological Park. Cotypes: U. S. Nat. Mus. Helm. Coll. No. 9291, others in Univ. Minn. Helm. Coll., New York Zool. Soc. Coll, and of author. Discussion. Deltokeras multilobatus n. sp. may be recognized by the characteristic lobation of the ovary, the lobes being on long pedunculated stalks. The ovary of D. ornitheios Meggitt, 1927, is sac-like and at most only slightly lobed; D. delachauxi Hsu, 1935, is strongly bilobed. In the case of D. campylometra Joyeux, Gendre & Baer, 1928, the description of the female glands is dismissed with the statement that they “ne presentent pas de parucularites,” which is taken to indicate that they are similar to those of D. ornitheios. The size and shape of the rostellar hooks serve to further differentiate the four species. In D. ornitheios, they are the largest, being 27-31 g long, and most similar in shape to D. multilobatus, while in both D. campylometra and D. delachauxi they are smaller than in D. multilobatus, being 10-15 g and 14-15 g long, respectively, as compared to 17-19 g long as in the case of the latter, as well as being very different in shape (cf. Figs. 2, 4, 6 and 7 ) . Meggitt (1927) stated that while the species of Biuterina have a parau- terine organ D. ornitheios does not, nor does the uterine wall show any of the characteristic fibrinous structure of that organ. Joyeux et al (1928) for D. campylometra and later Hsu (1935) for D. delachauxi reported a definite thickening of tissue surrounding the uterus and pointed out that it is a form of the parauterine organ. Hsu believed the presence of the parauterine organ a characteristic of the genus and emended the generic concept accordingly. These authors are of the opinion that the specimens of D. ornitheios w'ere too young to show the parauterine organ although Meggitt stated that while the oldest proglottides were not gravid “the most fully developed segment showed a lobed sac filled with eggs, occupying the former position of the ovary and extending to the anterior margin of the proglottis.” In considering the opin- ion of these authors in their belief that the specimens of D. ornitheios were too immature to show the parauterine organ, it is interesting to note that Joyeux et al found their specimens “ne sont pas assez mur pour nos permettre d’observer l’organe completement developpe;” even so it is figured as being very conspicuous and Hsii figured a sexually mature proglottid, not gravid, in which the organ is sufficiently well developed to be quite as obvious as the ovary and equal to it in size. This leads to the opinion that even though Meg- gitt’s specimens were not fully gravid they were undoubtedly sufficiently developed to show at least some indications of the presence of a parauterine organ, if it were to develop at all. The case of D. multilobatus appears to be intermediate between D. ornitheios, as described by Meggitt, on the one hand, and D. campylometra and D. delachauxi on the other. Here the parauterine 1939] Olsen: New Species of Cestode 343 tissue is present in fully gravid segments but very sparingly as shown in sec- tions, and furthermore, it is evenly dispersed over the entire uterus. This condition appears to be analogous to that which Joyeux et al noted for species having a parauterine organ which has not reached its full develop- ment. In view of the above discussion on the parauterine organ in Deltokeras, it is suggested that possibly here is a group of cestodes that represents a transitional stage between those genera having no parauterine organ and Biuterina and related genera having a well developed and specialized one. The species may be differentiated by means of the following key. Key to the species of Deltokeras. 1. Genital pores irregularly alternate; rostellar hooks 14-15 p long; 15-17 testes; ovary bilobed. Deltokeras delacliauxi Hsu, 1935 Genital pores unilateral 2 2. Ovary with 8-9 long pedunculated lobes; hooks 17-19 p long; 12-16 testes. Deltokeras multilobatus n. sp. Ovary not with long pedunculated lobes but sac-shaped 3 3. Hooks 27-31 p long, 80 in number, dorsal root longer than ventral root or blade, knob on dorsal or ventral roots relatively small in comparison to size of hook. Deltokeras ornitheios Meggitt, 1927 Hooks 10-15 p long, 46 in number, dorsal and ventral roots of equal length and with knobs which are extremely large in comparison to size of hook. Deltokeras campylometra Joyeux, Gendre and Baer. 1928 References. Hsu, H. F. 1935. Contributions a l’etude des cestodes de chine. 477-570. Rev. suisse Zool. 42 : Joyeux, Ch., Gendre, C., & Baer, J. G. 1928. Note sur les helminthes d l’Afrique occidentale frangaise. Coll. Soc. path. exot. Monogr. 2, 120 pp. Meggitt, F. J. 1927. On cestodes collected in Burma. Parasitology 19: 141-153. 344 Zoologica: New York Zoological Society EXPLANATION OF THE PLATE. Plate I. All drawings made with the aid of a camera lucida. Scale of enlargement 0.1 mm. except in Figs. 2 and 5, where it is 0.01 mm. or as otherwise noted. Fig. 1. Scolex of Deltokeras multilobatus. Fig. 2. Rostellar hooks of D. multilobatus. Fig. 3. Frontal section of gravid proglottid showing uterus with parauterine tissue. Partly reconstructed. Fig. 4. Hooks of D. ornitheios. After Meggitt, 1927. Fig. 5. Hook from intrauterine embryo of D. multilobatus. Fig. 6. Hook of D. campylometra. After Joyeux, Gendre & Baer, 1928. Fig. 7. Hooks of D. delachauxi. After Hsu, 1935. Fig. 8. Cross-section through anterior portion of mature segment. Fig. 9. Mature segment showing reproductive glands. Fig. 10. Cross-section taken near middle of same segment as in Fig. 8. Key to abbreviations. C cirrus CP cirrus pouch DE dorsal excretory canal EL external longitudinal muscle IL internal longitudinal muscles N nerve 0 ovary OD oviduct T testes V vitelline gland VD vas deferens VE ventral excretory canal OLSEN. PLATE I. DELTOKERAS MULTILOBATUS, A NEW SPECIES OF CESTODE FROM THE TWELVE-WIRED BIRD OF PARADISE, SELEUCI DES M. MELANOLEUCUS (dAUDIn). Jackson, Dring & Schroeder: Urinary Nitrogen of Carnivora 345 22. The Urinary Nitrogen Distribution of Representative Members of the Carnivora. Richard W. Jackson, Thomas J. Dring & Charles R. Schroeder.* The Department of Biochemistry, Cornell University Medical College, New York, and the Neiv York Zoological Park. (Text-figures 1-5). One of the many interesting chapters in our knowledge of comparative biochemistry is that dealing with the form in which nitrogen is excreted from the animal body. An excellent perspective of a considerable number of the quantitative investigations in this field is provided in a table com- piled by Needham (1931, pp. 1139-41). With few exceptions among both the invertebrates and vertebrates, ammonia, urea and uric acid have been shown to account for the bulk of the nitrogen eliminated, and according to Needham (1931, p. 1132), these three compounds appear to be the only substances “which are available in the animal kingdom for carrying away the nitrogenous waste resulting from protein breakdown.’’ There are to be found in the urine, of course, quite a number of other nitrogenous con- stituents, but their combined nitrogen content is usually only a small frac- tion of the total nitrogen. Baldwin has summarized the salient facts con- cerning the excretory products of protein and purine metabolism in verte- brates as shown in Table I. The fascinating evolutionary aspects of the biochemical differences briefly surveyed above, though outside the province of this paper, have been dealt with in several contributions (cf. Needham (1929; 1931, p. 1132), Smith (1932; 1935), Baldwin (1937) and Florkin (1935)). The mammalian class exhibits a monotonous regularity in the employ- ment of urea as the chief end-product of nitrogen metabolism. This holds even for the egg-laying mammal, Echidna aculeata (Neumeister (1898); Robertson (1923); Mitchell (1931)). One valuation of note in the urinary nitrogen partition pertains to the excretion of hippuric acid. Though the synthesis of this substance is by no means limited to the herbivorous ani- mals, the latter, after ingesting large amounts of hay and other benzoic acid-yielding feed, may excrete a considerable fraction of the total nitrogen as hippuric acid. Again, the diet of the herbivora is often predominantly base-forming and this leads apparently to a diminished production and elimination of nitrogen in the form of ammonium salts. A variation of more fundamental significance, inasmuch as it is related to specific metabolic processes rather than to dietary habits, is that evinced by various mam- malian species in the extent of oxidation of the purine bodies before elimina- * The authors wish to acknowledge the generous cooperation of the late Dr. Charles V. Noback in the early part of this investigation. 346 Zgologica: New York Zoological Society [XXIV :22 Table I. After Baldwin (1937, p. 61).t End-product of Protein metabolism Purine metabolism Mammalia Urea Allantoinf Aves Uric acid Uric acid Reptilia : Snakes, lizards Turtles Uric acid Urea Uric acid Allantoin ? Amphibia Urea Urea Pisces Elasmobranchii Teleostei Urea Ammonia Urea Urea t From “An Introduction to Comparative Biochemistry,” by Ernest Baldwin, by permission of Cambridge University Press, London. t Uric acid in man, higher apes and Dalmatian dog. tion. Man, the anthropoid apes, and to a degree the Dalmatian dog, excrete uric acid as an end-product of purine metabolism whereas the remaining mammalian forms, as far as data are available to show, carry the oxidation a stage further to allantoin. With regard to the excretion of creatinine and creatine, Hunter (1928, pp. 104-5) states that creatinine is never absent from the urine of mammals and further that “it is probable . . . that every adult mammal, when placed under standard conditions upon a creatine-free diet, excretes only creatinine.” The nitrogen distribution in the urine of the order Carnivora, as represented by the domestic cat and dog, conforms to what has been said above concerning the general character of mammalian urine. Urea is the dominant nitrogen-containing component, and the main product of purine metabolism is allantoin. Studies of the nitrogenous substances in the urine of some of the wild members of the Carnivora date back to the early part of the nineteenth century (see Milne Edwards (1862)) but these studies naturally were limited in scope and accuracy. Investigations dealing with non-domesticated Carnivora and involving the use of modern methods of urine analysis, insofar as revealed by our search of the literature, are as follows: of the coyote by Swain (1905) and by Hunter & Givens (1910-11) ; of the fox and coyote by Hawk (1910-11) ; of the weasel, raccoon dog ( Nyctereutes viverrinus), tiger, leopard and hyena by Fuse (1925); and of the seal by Smith (1936) \ Also, Hunter and associates (1914 and 1920) showed that the uricolytic index2 of the raccoon, black bear, badger, coyote and dingo, as of the domestic cat and dog, is high, the excretion of allantoin being in large excess over that of uric acid. Without going further into a detailed analysis of these contributions, it may be stated that they all point to a similar pattern in the urinary nitrogen distribution of the Carnivora. In contrast to this picture of uniformity was the finding of S. R. Benedict (1916) that the pure strain Dalmatian coach dog excretes an unusual amount of its purine end-product as uric acid. The uricolytic index is in 1 Swain & Rakes traw (1923) reported the presence of uric acid in the urine of the sea Hon. 2 This term is defined on a subsequent page. 1939] Jackson, Dring & Schroeder: Urinary Nitrogen of Carnivora 347 Text-figure 1. Exterior view of metabolism cage. the neighborhood of 30-40, intermediate between the much higher values found for other Carnivora and the very low figures in the case of man and anthropoid apes. This rather strange aberration in an otherwise fairly uniform series indicates that it is at least possible that other similar in- stances of significant variation in urinary nitrogen partition may await detection — even among the different families and species of one order as, for example, the Carnivora. We were therefore interested, when the opportunity arose in connection with experiments initiated for the purpose of studying the kynurenic acid excretion by representative species of the Carnivora, in determining as well the nitrogen distribution of the urines of these animals. In our work, analyses were made for total, urea, ammonia, creatinine and creatine nitro- gen, and as far as facilities would permit, for allantoin and uric acid nitrogen. The volume of urine voided in a given period, generally 24 or 48 hours, and the specific gravity of each sample, have a’so been recorded. As will be seen, some of the species included in our study have been investigated before. However not all of the previous analyses were made on samples collected over a definite period, and in some cases the analyses were not sufficiently extensive to account for the most of the urinary nitrogen. 348 Zoologica: New York Zoological Society [XXIV :22 Text-figure 2. Interior view of metabolism cage. Experimental. For the collection of urine, the genets and the skunk were confined in a small metabolism cage of the type ordinarily employed with rabbits. For the remainder of the animal subjects, excepting the bear, a special cage was constructed by modifying one of the cages available at the Zoological Park Hospital. The original cage is about 4 feet along each edge and is con- structed of heavy wire mesh mounted on angle iron and is bolted to an elevated concrete floor. The alteration of this cage for our purpose was accomplished by building and inserting a snugly fitting unit consisting of a urine collecting box with deep side walls to insure against loss of urine, and with a heavy false bottom to protect the lighter solid zinc metal bottom below and to hold back fecal material. The false bottom in position rests on lugs but may be raised free of these and, if desired, removed from the cage, to facilitate cleaning. All seams of the urine collecting box are soldered so that except for the door it is water tight. The door, like the sidewalls, is constructed of galvanized iron. It is inserted vertically through channels in the adjacent side walls and extends to a point below the door sill and also below the false bottom. This door is contiguous with the door of the cage proper. The complete metabolism cage is illustrated in Text-figs. 1, 2 and 3. It has served excellently for animals ranging in size from the raccoon and fox to the cheetah and hyena. In order to collect urine from the Tibetan bear, it was necessary to prepare still another cage (see Text-figs. 4 and 5). This was effected by placing a close-fitting, shallow, flat, galvanized iron pan on the bottom of a heavily constructed crate ordinarily used for shipping bears. This pan extends from the rear of the crate about three-quarters of the way forward 1939] Jackson, Dying & Schroeder: Urinary Nitrogen of Carnivora 349 Text-figure 3. Schematic drawing of metabolism cage. 350 Zoologica: New York Zoological Society [XXIV :22 Text-figure 4. Photograph of cage with pan, used in metabolism experiments on bear. so that with a male animal no urine is lost. In use, the cage is placed on blocks with the front end slightly elevated thus to promote the flow of the urine to the rear of the pan through the tubulature to the collecting bottle. By having the urine collecting pan of such length as not to extend forward beneath the head of the bear, it is possible to supply water periodically without danger of its being spilled into the urine. The predilection of our bear subject for using his front paws to bend up the front end of the urine-collecting pan was thwarted by inserting a heavy iron bar laterally through the cage and just over the front end of the pan. All of the animals studied by us were ingesting diets composed entirely or largely of meat or of fish. Water was supplied ad libitum unless the individual exhibited a persistent tendency to dislocate the water container with danger of diluting the urine. In this event, drinking water was given at intervals during the day. The sea lion was generally permitted to secure his supply of water from his generous daily quota of fresh (frozen) fish. However, in order to guard against physical discomfort of the sea lion while under experiment, the animal was occasionally moistened with a fine mist of water from an insect gun, or wet sheets were hung about the cage. All specimens of urine, with the exception of one which was secured from the bladder at autopsy, were collected from the metabolism cages di- rectly into bottles containing toluene. The samples were filtered at room temperature to remove any hair or occasional slight contamination of fecal material, and then immediately stored under toluene at a temperature of 5° C. preliminary to analysis. The analytical procedures employed were as follows: total nitrogen by the Kjeldahl method; urea plus ammonia nitrogen by the urease-aeration-titration procedure of Van Slvke & Cullen (1914 and 1916) ; ammonia by the method of Folin & Bell (1917) ; creatinine by Folin’s colorimetric method (see Hawk & Bergeim (1937)); creatine according to Benedict (1914); allantoin by Larson’s method (1932); and uric acid by the indirect precipitation method of Benedict & Hitchcock 1939] Jackson, Dring & Schroeder: Urinary Nitrogen of Carnivora 351 (1915). With the advice of Dr. Benedict, we incorporated two modifications of the procedure for determination of uric acid. The arsenophospho- tungstate reagent of Benedict (1922) was substituted for that of Folin & Denis, and 15% of urea was added to the sodium carbonate solution to prevent turbidity (cf. Folin (1930) and Christman & Ravwitch (1932)). Acidity of the urines was tested with litmus paper, and the specific gravity was determined by means of a urinometer. 352 Zoologica: New York Zoological Society [XXIV :22 Discussion. The results of our experiments on fifteen different species representing eight different families of the Carnivora are presented in Tables II and III. It is to be emphasized that the information given relative to the diets is, as stated, a rough estimate. The food was ordinarily not weighed and the amount of food intake was therefore subject to fluctuation. The diet com- positions show nevertheless that all the animals were on a comparable basis in the ingestion of a high-protein diet. In connection with a study of kynurenic acid excretion to be reported elsewhere, tryptophane was admin- istered in some of the experiments. Inasmuch as this procedure did not appear to alter the distribution of nitrogen, these experiments have been included in the tables. Again, it should be pointed out that the figures for urine volume and total nitrogen are for the stated collection period. In the case of the genet, for example, the values (Table II) are for two animals over a period of 45 hours. Furthermore, the urine volumes are those col- lected from the cage and cannot be taken as necessarily representing the exact amounts of urine produced during the experimental periods. The urines with few exceptions were acid to litmus. The exceptions were doubt- less the result of some conversion of urea to ammonia despite the precau- tions taken to prevent bacterial action. The distribution of nitrogen is much the same for all species studied by us, and is generally similar to that reported in previous experiments on members of the Carnivora. The majority of the values for the different nitrogenous constituents, expressed as nitrogen in per cent, of total nitro- gen, fall within the following ranges: urea, 80 to 86, ammonia, 2 to 5, urea plus ammonia, 83 to 89, creatinine, 1 to 2, creatine, 1.5 to 3, allantoin, 2 to 4, uric acid, 0.05 to 0.20. The extent of deviation from these ranges may be seen by inspection of the tables. The few values for ammonia nitrogen which are over 7% of the total nitrogen are very likely the result of some bacterial conversion of urea, inasmuch as either the urine was actually alkaline (Exp. 37, Table II) or there had been opportunity for soiling of the cage during the immediately preceding collection periods (Exps. 33a, 29a, 35c, and 32a, Table III). A part of our program was carried out during warm weather which, of course, would be especially conducive to ammonia production. However, in no instance was there evidence of any extensive decomposition of the urine specimen. The relatively low percentages for urea and the corresponding elevated values for some of the other constituents in the bear experiments (Table III) may owe their explanation to a relatively lower nitrogen intake or to the storage of nitrogen. The uricolytic index, that is, the per cent, allantoin nitrogen of total allantoin and uric acid nitrogen excreted, was determined by Hunter and his co-workers on animals which were either fasting or ingesting a diet low in purines. The purpose, of course, was to confine the criterion to the endogenous purine metabolism and thereby to eliminate the variable and disturbing influence of the exogenous metabolism of purines in the diet. Nevertheless, Hunter & Givens (1910-11) in their early studies on the coyote found that when the animal was ingesting a meat diet supplying relatively considerable quantities of purine material, the allantoin nitrogen constituted more than 95% of the total allantoin and purine (including uric acid) nitrogen excreted. The authors state, “Whether, therefore, endogenous or exogenous purines be concerned, it is evident that among the end-products of their metabolism allantoin plays an enormously preponderating part.” What Hunter & Givens report in regard to the allantoin-uric acid relation- ship for the coyote ingesting a meat diet, we have found essentially to be true also for the fox, dingo, cheetah, serval, civet, badger, Tibetan and grizzly bears, raccoon and sea lion, all likewise ingesting a diet wholly or mainly of flesh. In the case of the grizzly bear, the allantoin nitrogen constituted 84% of the total allantoin and uric acid nitrogen excreted, in the other cases, the values were 92% or higher. Fuse found a similar pre- anidae) . Family, tion of N in per cent of total N common name , species inine Creatine N Allantoin N I Uric acid N 1 Undeter- mined N Felidae Cheetah ( Acinonyx jubatus) 7 1.3 2.3 0.06 9.7 Cheetah (same animal) 3 2.3 2.2 0.07 10.9 Serval ( Felis serval) 7 1.4 2.0 0.08 5.0 Serval ( Felis serval) 2.0 0.04 Vivenridae Civet ( Civettictis civetta) 0.6 2.4 0.04 5.4 Civet (same animal) 1.7 2.8 trace 3.1 Civet ( Civettictis civetta) 2.2 Genet 2 animals ( Genetta India) - total = 2.5 Hyaenidae Hyena (Hyaena hyaena) 1.9 0.056 Hyena (same animal) 1.8 0.054 Cauidae Coyote (Canis latrans) ) 1.0 1.4 0.13 7.2 Coyote (same animal) S 1.4 1.6 0.12 7.1 Red fox (Vulpes fulva) 2.1 2.2 0.16 6.5 Red fox (same animal) 2.2 2.0 0.11 10.7 Dingo (Canis dingo) 1.4 3.2 0.12 4.5 Wolf (Canis nubilus) 2.5 Wolf (same animal) t Specimens not fully * The symbol T is for gm. given during the entire period whether 24 or 48 hours, etc. Table II. Distribution of Nitrogen in the urine of the Carnivora (Felidae, Viverridae, Hyaenidae and Canidae). Family, common name, species Weigh t and sex Exp. No. and date Daily diet * ( rough estimate) Period of collection Specific gravity at 15.6° Distribution of N in per cent of total N volume and reaction N Urea N Ammonia N Creatinine N Creatine N Allantoin N Uric acid N Undeter- mined N Kg. lb. hr. cc. gm. Felidae Cheetah (Acinonyx jubatus ) 40.5