mini!?!)) I II ill ill ill 11 ; 111 111 1 liil HI! ! 5 i f li! 11 II llli'tt (lit llij I II 1 1 ill I 1 ' I I I I i I li 11 Hiili! II i iiil! >i ! Iff It i Ml BT 4 1 Memoirs Carnegie Museum, Vol. v. Frontispiece. The Kaietevk Falls. Height 741 feet. MEMOIRS OF THE CAKNEGIE MUSEUM Vol. v. (Publications of the Carnegie Museum: Serial No. 67). W. J. HOLLAND, Editor. PITTSBURGH Published by the Autiioeity of the Board of Trustees of the CARNEGIE INSTITUTE JUNE, 1912. Press of the nev. era printing comp*nf Lancaster. Pa. °) \ OU-O^lr THE FRESHWATER FISHES OF BRITISH GUIANA, INCLUDING A STUDY OF THE ECOLOGICAL GROUPING OF SPECIES AND THE RELATION OF THE FAUNA OF THE PLATEAU TO THAT OF THE LOWLANDS. By C. H. EIGENMANN, Ph.D. Dean of the Graduate School of the Indiana State University, and Curator of Ichthyology in the Carnegie Museum. in " PREFATORY NOTE The following pages, constituting the Fifth Volume of the Memoirs of the Carnegie Museum, embody the results of the investigations made by Dr. C. H. Eigenmann upon the piscine fauna of the rivers of British Guiana. Having been granted temporary leave of absence from the Indiana State Uni- versity, where he is the Dean of the Graduate School, Dr. Eigenmann went to British Guiana in the fall of 1908, having the assurance of the financial support of the Carnegie Museum, both in meeting the expenses of his expedition and in pub- lishing the results thereof. The typical portions of the collections made by him are preserved in the Carnegie Museum, including all unica. A second set, includ- ing, so far as possible, cotypes of the new species, was reserved for the Indiana State University. It has been arranged to dispose of the remainder of the collec- tion by exchange with the leading museums of America and Europe. The Editor, in presenting this work to the scientific public, is animated by the hope that it may prove of value to the students of biology in all parts of the world, and tend to shed at least a little light upon some of the countless problems which confront the student of vital phenomena. W. J. Holland. Pittsburgh, Dec. 1, 1911. LETTER OF TRANSMITTAL Dr. W. J. Holland, Director Carnegie Museum, Pittsburgh, Pa. My dear sir: By accident you and I came to occupy contiguous seats in an excursion from Colorado Springs to Cripple Creek, after the close of the Denver meeting of the American Association for the Advancement of Science in 1901. I am transmitting one of the indirect results of that pleasant ride. Since that chance meeting we have frequently discussed South America, a continent in which a growing series of converging interests meet, the continent which served as the training ground of Spix, Martius, Castelnau, Miiller, Darwin, Bates, Wallace, Agassiz, and Hatcher, a land in which there is a great wealth of many sorts of animals and plants, and in which fresh-water fishes have evolved in the greatest profusion, a continent which will in the centuries to come furnish homes to count- less millions of men. I have been interested in South American fresh-water fishes for twenty-five years. A little over two years ago it became necessary in continuance of my work to visit British Guiana. The Trustees of Indiana University gave me leave of absence. I applied to various other institutions for assistance, but met everywhere with refusals. When I visited you in August, 1908, on my way to Guiana, you pledged the Carnegie Museum to help. Both the authorities of Indiana University and you have done everything possible to bring my report to a successful conclusion. It is a great satisfaction to me that the assistance furnished by the Carnegie Museum has brought it such rich returns in valuable specimens. Since my return from Guiana in December, 1908, all of the material collected has been examined, and the results of the examination together with an account of the Guiana Expedition are herewith transmitted for publication. This paper was prepared in the Zoolog- ical Laboratory of Indiana University and forms "Contribution No. 114" of that laboratory. Cordially yours, C. H. ElGENMANN. December 1, 1910. TABLE OF CONTENTS. Prefatory Note ............ v Letter of Transmittal vii Table of Contents ............ ix List of Figures in Text xi List of Plates xiii Errata and Corrigenda xix Memoir: The Freshwater Fishes of British Guiana. Introduction 1^4 Chapter I. British Guiana 5-14 Chapter II. Earlier Works on the Fishes of British Guiana . . . 15-29 Chapter III. General Account of the Expedition 30-59 Chapter IV. Geographical Distribution of the Freshwater Fishes of British Guiana ......... 60-86 Chapter V. Ecological Combinations of Species 87-93 Chapter VI. The Ichthyic Fauna of the Pbtaro River and of the Guiana Plateau 94-105 Chapter VII. General Considerations 106-115 Chapter VIII. Systematic Account of the Freshwater Fishes of British Guiana 116-529 Chapter IX. Bibliography 530-554 Index 555-578 LIST OF FIGURES IN TEXT. Page. Mouth of the Demerara River at Georgetown, British Guiana 30 A trench in Georgetown 31 View of the lowlands looking seaward from the railway near Mahaica 32 View of the drainage canal at Cane-Grove Corner 33 Photograph of Carnegiella strigala in aquarium 35 View on the right bank of the Demerara River 36 View of creek filled with brush-wood above Wismar at low tide. Indians are poisoning the creek 37 Fish-fence made of reeds and banana-leaves on creek above Wismar. Indian with bow and arrow ready to shoot larger fishes brought up by poison 38 View on bayou back of Christianburg. Indian fishing from his corial 39 Indian women pounding leaves in a hollow on the ground preparatory to using the pulp for poisoning a stream on Gluck Island 40 Small stream on Gluck Island dammed by Indian women before putting in the poison- ous leaf-pulp 41 Edge of bayou between right bank of Essequibo and outlying sand-bar at Rockstone . 42 Looking across the Rapids of the Potaro River, at Tumatumari. Papaya-trees in foreground 44 Albert, one of the Indian bearers, transporting goods at Kangaruma 46 View looking up the Potaro River in the early morning. Glimpse of the Guiana Plateau in the distance 47 Seining on a sand-bar below the Amatuk Cataract 48 View on the Potaro River looking up stream at the point where the first glimpse of the Kaieteur Falls is caught 49 Exposed left side of the bed of the Potaro River at Tukeit in the dry season 50 Looking up the Potaro Valley from the brink of the Kaieteur Falls 51 Looking down the Canyon of the Potaro River from the brink of the Kaieteur Falls . 52 View looking up the dry bed of Shrimp Creek. Figure of man in middle distance shows height of rocky steps 53 Diagrammatic representation of the contour of the Kaieteur Falls. (After Brown.) . . 56 Table showing distribution of the fishes taken by C. H. Eigenmann in the Potaro River above the Kaieteur Falls 101 xi Xll LIST OF FIGURES IN TEXT Bunocephalus amaurus Eigenmann. Type 127 Agmus lyriformis Eigenmann. Type 128 Teeth of Selenaspis herzbergi (Bloch) 141 Teeth of Selenaspis passany (Cuvier and Valenciennes) 142 Outlines of occipital processes of various individuals of Notarius grandicassis (Cuvier and Valenciennes) 143 Outlines of occipital processes of various individuals of Notarius par?nocassis (Cuvier and Valenciennes) 144 Outlines of occipital processes of three specimens of Notarius stricticassis (Cuvier and Valenciennes) 145 Outline of premaxillary band of teeth in Megalonema platycephahim Eigenmann 150 Outline of Pseudopimelodus villosus Eigenmann. Type 152 Chasmocranus longior Eigenmann. Type 161 Plecostomus hemiurus Eigenmann. Type 225 Pseudancistrus nigrescens Eigenmann. Type 234 Ancistrus lithurgicus Eigenmann. Type 241 Tetragonopterus argenteus Cuvier 318 Tetragonopterus chalceus Agassiz 319 Maxillary bone of Rhinosardinia serrata Eigenmann 445 LIST OF PLATES. Frontispiece: The Kaieteur Falls (Facing p. 1) Plate. I. Aspredo sicuephorus Cuvier and Valenciennes; Bunocephalus gronovii Bleeker. II. Bunocephalus chamaizelus Eigenmann; Bunocephalus amaurus Eigenmann. (Type.) III. Agmus lyriformis Eigenmann. (Type.) IV. Agmus scabriceps Eigenmann and Eigenmann. (Type.) V. Sciadeichthys flavescens (Cuvier and Valenciennes); Sciadeichthys proops (Cuvier and Valenciennes); Sciadeichthys parkeri (Traill). VI. Sciadeichthys proops (Cuvier and Valenciennes) ; Dorsal view (the so-called "monk"); Palatal view (the so-called "crucifix"); Lateral view. VII. Selenaspis herzbergi (Bloch); Selenaspis passany (Cuvier and Valenciennes). VIII. Notarius grandicassis (Cuvier and Valenciennes); Notarius parmocassis (Cuvier and Valenciennes); Notarius stricticassis (Cuvier and Valen- ciennes.) IX. Sciadeichthys emphysetus (Miiller and Troschel); Arius spixi (Agassiz); Hexanematichthys rugispinis (Cuvier and Valenciennes). X. Pseudopimelodus villosus Eigenmann. (Type); Megalonema platycephalum Eigenmann. (Type.) XL Pseudopimelodus albomarginatus Eigenmann. (Type); Brachijglanis melas Eigenmann. (Type.) XII. Brachyglanis phalacra Eigenmann. (Type); Microglaiiis poecilus Eigen- mann. (Type.) XIII. Brachijglanis phalacra Eigenmann. (Type) ; Leptoglanis essequibensis Eigen- mann. (Type.) XIV. Myoglanis potaroensis Eigenmann. (Type); Chasmocranus longior Eigen- mann. (Type.) XV. Chasmocranus brevior Eigenmann. (Type); Pimelodella megalops Eigen- mann. (Type.) XVI. Pimelodella macturki Eigenmann. (Type) ; Pimelodus heteropleurus Eigen- mann. (Type.) XVII. Leptodoras linnelli Eigenmann. (Type); Trachycorystes obscurus (Giinther). (Type); Br achy platy stoma vaillanti (Cuvier and Valenciennes); Doras granulosus Valenciennes. Xlll XIV LIST OF PLATES XVIII. Leptodoras linnelli Eigenmann. (Type); Hemidoras microstomus Eigen- mann. (Type.) XIX. Hemidoras leporhinus Eigenmann. (Type) ; Hemidoras notospilus Eigen- mann. (Type.) XX. Centromochlus aulopygius Kner; Pseudauchenipterus nodosus (Bloch); Tympanopleura piperala Eigenmann. (Type.) XXI. Auchenipterus demerarce Eigenmann. (Type) ; Ageneiosus guianensis Eigen- mann. (Type.) XXII. Ageneiosus marmoratus Eigenmann. (Type) ; Helogenes marmoratus Gunther. XXIII. Hemicetopsis macilentus Eigenmann. (Type) ; Hemicetopsis minutus Eigen- mann. (Type.) XXIV. Hoploslernum littorale (Hancock); Hoplosternum thoracatum (Cuvier and Valenciennes) ; Corydoras punctatus (Bloch.) XXV. Plecostomus hemiurus Eigenmann. (Type); Pseudancistrus nigrescens Eigenmann. (Type) ; Ancislrus lilhurgicus Eigenmann. (Type.) XXVI. Plecostomus rvatwata (Hancock) ; Lithogenes lillosus Eigenmann. (Type) ; Neoplecostomus granosus (Cuvier and Valenciennes). XXVII. Corymbophanes andersoni Eigenmann. (Type.) XXVIII. Hemiancistrus braueri Eigenmann. (Type) ; Psexidancistrus barbatus (Cu- vier and Valenciennes.) XXIX. Lithoxus lithoides Eigenmann, cf (Type) ; Lithoxus Uthoides Eigenmann, 9 and d\ Ventral surface. XXX. Loricariichthys microdon Eigenmann. (Type) ; Loricariichthys griseus Eigen- mann. (Type) ; Loricariichthys brunneus (Hancock) ; Loricariichthys platyurus (Muller and Troschel) ; Loricariichthys stewarti Eigenmann. (Type) ; Harltia platystoma (Gunther.) XXXI. Harttia platystoma (Gunther); Loricariichthys platyurus (Muller and Troschel) ; Loricariichthys brunneus (Hancock.) XXXII. Loricariichthys microdon Eigenmann. (Type) ; Loricariichthys griseus Eigen- mann. (Type) ; Farlowella hargreavesi Eigenmann. (Type.) XXXIII. Bivibranchia protractila Eigenmann. (Type.) XXXIV. Curimatus spilurus Gunther; Curimatus microcephalics Eigenmann and Eigenmann; Curimatus morowhannos Eigenmann. (Type); Curimatus issororoensis Eigenmann. (Type.) XXXV. Curimatus schomburgki Gunther ; Prochilodus maripicru Eigenmann. (Type) ; Tylobronchus maculosus Eigenmann. (Type) ; Chilodus punctatus Muller and Troschel. XXXVI. Parodon bifasciatus Eigenmann. (Type) ; He?niodus quadrimaculatus Pelle- grin; Hemiodus semitamiatus Kner; Nannostomus marginatum Eigenmann. LIST OF PLATES XV (Type); Nannostomus minimus Eigenmann. (Type); Nannostomus simplex Eigenmann. (Type.) XXXVII. Poecilobrycon harrisoni Eigenmann. (Type) ; Poecilobrycon Irifasciatus (Stein- dachner); Poecilobrycon erythrurus Eigenmann. (Type); Poecilobrycon ocellatus Eigenmann. (Type) ; Archicheir minutus Eigenmann. (Type) ; Characidium laterale Eigenmann. (Type and Cotype.) XXXVIII. Characidium vintoni Eigenmann. (Type); Characidium blennioides Eigen- mann. (Type) ; Characidium catena-turn Eigenmann. XXXIX. Characidium fascialum Reinhardt; Characidium pellucidum Eigenmann. (Type) ; Characidium pteroides Eigenmann. Type and Cotype. XL. Anostomus anostomus Linnaeus. XLI. Anostomus anostomus (Linnseus); Anostomus trimaculalus (Kner); Anosto- mus plicatus Eigenmann. (Type); Schizodontopsis laticeps Eigenmann. (Type.) XLII. Leporinus nigrotoeniatus (Schomburgk) ; Leporinus arcus Eigenmann. (Type.) XLIII. Leporinus allernus Eigenmann. (Type); Leporinus maculatus Miiller and Troschel; Leporinus granli Eigenmann. (Type); Leporinus friderici (Bloch.) XLIV. Poecilocharax bovallii Eigenmann. (Types c? and 9 ) ; Odontostilbe melandelus Eigenmann. (Type); Aphyocharax erythrurus Eigenmann. (Type); Aphyodite grammica Eigenmann. (Type.) XLV. Creagrulus melanzonus Eigenmann. (Type); Bryconamericus hyphesson Eigenmann. (Type); Pristella riddlei (Meek); Pristella aubynei Eigen- mann. (Type.) XLVI. Mamkhausia profunda Eigenmann. (Type) ; Fowlerina orbicularis (Cuvier and Valenciennes); Mamkhausia oligolepis (Giinther). XLVII. Ctenobrycon spilurus (Cuvier and Valenciennes); Moenkhausia grandi- squamis (Miiller and Troschel) ; Moenkhausia browni Eigenmann. (Type) ; Moenkhausia shideleri Eigenmann. (Type.) XLVIII. Hemigrammus unilineatus Gill; Hemigrammus erythrozonus Durbin. (Type) ; Hemigrammus rodwayi Durbin. (Type) ; Hemigrammus ocellifer (Steindachner) ; Hemigrammus orthus Durbin. (Type.) XLIX. Hemigrammus iota Durbin. (Type); Hyphessobrycon minimus Durbin. (Type); Hemigrammus cylindricus Durbin. (Type); Hyphessobrycon gracilis (Reinhardt); Hyphessobrycon minor Durbin. (Type); Hemi- grammus analis Durbin. (Type) ; Hyphessobrycon sticlus Durbin. (Type.) L. Hyphessobrycon rosaceus Durbin. (Type) ; Hyphessobrycon eos Durbin. (Type); Creatochancs affinis Giinther; Creatochanes melanurus (Bloch); Creatochanes caudomaculatus Giinther. XVI LIST OF PLATES LI. Astyanax guianensis Eigenmann. (Type); Astyanax essequibensis Eigen- mann. (Type) ; Astyanax mutator Eigenmann. (Type) ; Astyanax mucro- natus Eigenmann. (Type.) LII. Astyanax wappi (Cuvier and Valenciennes); Poecilurichthys polylepis (Giinther) ; Poecilurichthys abramoides Eigenmann. (Type) ; Poecilurich- thys potaroensis Eigenmann. (Type.) LIII. Deuterodon potaroensis Eigenmann. (Type) ; Deuterodon pinnatus Eigen- mann. (Type) ; Phenacogaster megalostictus Eigenmann. (Type) ; Phena- cogasler microstictus Eigenmann. (Type.) LIV. Holobrycon pesu (Muller and Troschel) ; Brycon falcatus Muller and Troschel ; Brycon siebenthalce Eigenmann. (Type.) LV. Chalcinus rotundatus (Schomburgk) ; Carnegiella slrigata (Giinther) ; Gastero- pelecus sternicla (Linnaeus). LVI. Serrasalmo rhombeus (Linnseus) ; Pygocentrus bilineatus Eigenmann. (Co- type) ; Caloprion mento (Cuvier) . LVII. Metynnis maculatus (Kner) ; Myloplus rubripinnis (Muller and Troschel) ; Myloplus asterias (Muller and Troschel), cf. LVIII. Myloplus rhomboidalis (Cuvier). LIX. Myleus pacu (Schomburgk) . LX. Exodon paradoxus Muller and Troschel; Roeboides thurni Eigenmann. (Type) ; Charax gibbosus (Linnseus) ; Asiphonichthys hemigramrnus Eigen- mann. (Type.) LXI. Acanthocharax microlepis Eigenmann. (Type); Heterocharax macrolepis Eigenmann. (Type) ; Acestrorhynchus falcatus (Bloch) ; Aceslrorhynchus nasutus Eigenmann. (Type.) LXII. Hoplias macrophthalmus (Pellegrin); Hoplias malabaricus (Bloch); Rhino- sardinia serrata Eigenmann. (Type) ; Stolephorus guianensis Eigenmann. (Type.) LXIII. Rivulus breviceps Eigenmann. (Type); Rivulus holmice Eigenmann, cf (Cotype), 9 (Type); Rivulus waimacui Eigenmann, d" (Cotype), 9 (Type) ; Rivulus stagnatus Eigenmann, c? (Cotype), 9 (Type). LXIV. Rivulus lanceolatus Eigenmann. (Type); Rivulus frenatus Eigenmann. (Type); Poecilia vivipara Bloch and Schneider, d\ 9 ; Acanthophacelus melanzonus Eigenmann, c? (Type), 9 (Cotype). LXV. Acanthophacelus reticulalus (Peters), d\ 9 ; Acanthophacelus bifurcus Eigenmann, d" (Cotype), 9 ; Tomeurus gracilis Eigenmann, d 1 (Type), 9 (Cotype) ; Nannacara anomala Regan. LXVI. Nannacara bimaculata Eigenmann. (Type); JFquidens potaroensis Eigen- mann. (Type); Geophagus surinamensis (Bloch). LXVII. Geophagus jurupari Heckel; Group of young. LIST OF PLATES XV11 LXVIII. Heterogramma ortmanni Eigenmann. (Type); Heterogramma steindachneri Regan; Crenicichla alta Eigenmann. (Type.) LXIX. Cichla ocellaris Bloch and Schneider; Young, half-grown. LXX. Apionichthys unicolor Gunther; Soleonasus finis Eigenmann. (Type.) LXXI. Map : Distribution of the Aspredinidse. LXXII. Map : Distribution of the Pimelodinse. LXXIII. Map : Distribution of the Pimelodinse continued. LXXIV. Map: Distribution of the Pimelodinse continued. LXXV. Map : Distribution of the Doradinse. LXXVI. Map: Distribution of the Doradinse continued. LXXVII. Map: Distribution of the Auchenipterinse. LXXVIII. Map: Distribution of the Callichthyidse. LXXIX. Map : Distribution of the Plecostominse. LXXX. Map: Distribution of the Plecostominse continued. LXXXI. Map : Distribution of the Plecostominse continued. LXXXII. Map: Distribution of the Loricariinse. LXXXIII. Map : Distribution of Bimbranchia and Curimatus. LXXXIV. Map : Distribution of the Hemiodinse and Prochilodinse. LXXXV. Map: Distribution of the Nannostomatinse. LXXXVI. Map: Distribution of Charicidium of the Nannostomatinse. LXXXVIL Map : Distribution of the Anostomatinse. LXXXVTII. Map : Distribution of Leporinus. LXXXIX. Map : Distribution of Mamkhausia. XC. Map : Distribution of the Tetragonopterinse. XCI. Map: Distribution of Hemigrammus: XCII. Map : Distribution of the Tetragonopterinse continued. XCIII. Map : Distribution of Crealochanes. XCIV. Map: Distribution of the Tetragonopterinse continued. XCV. Map: Distribution of the Tetragonopterinse continued. XCVI. Map: Distribution of various genera of Characidae. XCVII. Map : Distribution of the Characidse. XCVIII. Map: Distribution of Characidse continued. XCIX. Map : Distribution of the Erythrininse. C. Map: Distribution of the Cichlidae. CI. Map: Distribution of the Cichlidse continued. CII. Map: Distribution of the Cichlidse continued. CIII. Map: Distribution of the Cichlidse continued. MEMOIES 01' THE CAKNEGIE MUSEUM. VOL. V. NO. l. THE FRESHWATER FISHES OF BRITISH GUIANA, INCLUDING A STUDY OF THE ECOLOGICAL GROUPING OF SPECIES, AND THE RELATION OF THE FAUNA OF THE PLATEAU TO THAT OF THE LOWLANDS. By C. H. Eigenmann, Ph.D. INTRODUCTORY. I had two purposes in view in the trip to Guiana: first, to observe, photograph, and incidentally collect as many species as possible for my monograph of the characins; second, in connection with my general faunal study of the fishes of South America to determine, if possible, the relation of the fish-fauna of the Guiana plateau to that of the lowlands, more particularly the relationship existing between the faunas of the upper and the lower Potaro. The two reaches of the PotarO are separated from each other by the Kaieteur Fall, over which the water leaves the Guiana Plateau by a drop of seven hundred and forty-one feet. Although all other things were sacrificed to the two purposes mentioned, I cannot claim that I accomplished them to my entire satisfaction. The conditions were all so novel, the difficulties of travel so great, the heat so intense, the fauna so rich, the time and the money at my command so limited, that I now occasionally regret that at this or that point I did not use different means, or devote more time to the objects in view. But to offset this regret I have many solid satisfactions. If seeing and recording a lot of "specimens," which have been disintegrating for longer or shorter periods in alcohol, can be called acquaintance, I have been acquainted with South American fresh-water fishes for many years. In contrast 2 MEMOIRS OF THE CARNEGIE MUSEUM to such an acquaintance I recall standing one exciting morning on the brink of a small pool, which my Indians said contained fishes. It was not more than fifty feet across and was back-water left by the receding Essequibo. The Indians pounded poisonous hiari roots, tied them into bundles, and the boys then swam through the pool with them over their backs, and thus mixed in the poison. Soon one species, then another, and still others, which I had only known as mummies, were resurrected from the depths of that pool, and I danced about its margin with delight to see them in their vivid living colors and incidentally to embalm them in their turn for future reference. On another glorious working day, which lasted from 6 A.M. to 12:30 A.M. of the day following, I secured over seventy species of fresh-water fishes, more than forty of which were characins. Robert Schomburgk noted eighty-three species of fishes in his travels about Guiana, only thirty of which came from the region covered by this paper. I secured more than twice as many species in a few hours. Richard Schomburgk reported a total number of one hundred and forty-eight species, of which about eighty were recorded from the area covered by this report. In all, only one hundred and sixteen species have been hitherto recorded from the rivers of British Guiana, discharging into the Atlantic, while three hundred and sixty species are recorded in the present paper. Of these 1 myself collected all but twenty- seven. Of the twenty-seven species I have examined all but Rhamdia laukidi, Rhamdia arekaima, and Pimelodus altipinnis. Rhamdia laukidi and arekaima are probably forms of R. clarias. No one has seen them since Schomburgk's day, and he did not preserve specimens. Pimelodus altipinnis was described from a small specimen said to have come from Demerara. The species said to be P. altipinnis is abundant in the lower Amazon. With the exception of the few cases in which labels were lost in transit, all of the specimens here recorded are assigned to exact localities. Very few definite localities had been previously put on record. "Demerara," "Essequibo," "Rupununi," or "British Guiana," have been the usual designations. Nannostomus beckfordi Giinther and Heterogramma steindachneri Regan are among the few species the definite habitats of which were known. Science owes a debt to a number of persons in and out of Guiana, who very generously assisted me in various ways. The officers of the Quebec Steamship Co., Ltd., transported our collecting outfit and our collections free of all charge, and made generous reductions from the regular passenger tariff from New York to Georgetown and return. Messrs. Sprostons, Ltd., conveyed us and our baggage free to all points within the colony reached by their boats, and put EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 3 us under still greater obligations by instructing their various agents to help us in every possible way. The officers of the railway between Georgetown and New Amsterdam gave us similar help. Mr. Lyman Jones and Mr. St. Aubyne of the Lama Water Conservancy placed boats, a house, servants, and a number of men at our disposal, and the collections from Lama Stop-Off, Maduni Stop-Off, and Cane Grove Corner are altogether due to the disinterested help of these gentlemen. The most generous interest in the success of the expedition was displayed by Dr. Edwin Bovallius and Mr. George Linnell, who furnished me with boats and a crew of sixteen Indians, who conducted me from Kangaruma to the Chenepowo River at Holmia and to the Aruataima Cataract. Mr. William Grant, the Indian guide and interpreter on this trip, became an enthusiastic fisherman, and has sent me many new things since my return. 1 At Holmia the quarters of the Essequibo Exploration Company were placed at my disposal, and here I was able to recuperate from a fever, while my crew were gathering poison and fishes. The success in the upper Potaro is entirely due to the generous help of Messrs. Bovallius and Linnell. At Tumatumari Mr. Brummel, at Wismar Mr. J. D. Spence, and at Georgetown Messrs. J. B. Mitchell, James Rodway, C. W. Anderson, E. A. V. Abraham, and Professor J. G. Harrison aided me in various ways. Mr. B. S. Conrad of George- town, to whom I had a letter of introduction, greatly assisted the expedition with advice and guidance. He not only generously devoted much of his own valuable time to the interests of the expedition, but introduced mc to other gentlemen, who aided me in a variety of ways. I am indebted to Mr. E. S. Shideler, who acted as volunteer student assistant on this trip. He collected with me most of the time, and the collections from Malali, Bartica, the Botanic Garden, and the northwest are entirely due to his efforts. The summer of 1910 was devoted to examining the types of fishes from Guiana in European museums. Drs. G. A. Boulenger and C. T. Regan of the British Museum, Drs. Th. W. van Lidth de Jeude and C. M. L. Popta of the Leyden Museum, Director Dr. Max Weber of the Amsterdam Museum, Director Dr. A. Brauer and Dr. P. Pappenheim of the Royal Museum in Berlin, Intendant Hofrat Dr. Fr. Steindachner of the K. K. Naturhistorisches Hofmuseum in Vienna, and Dr. J. Pellegrin of the Musee d'Histoire Naturelle in Paris all cordially cooperated by placing the types in their keeping at my disposal. 1 Mr. Grant collected in the following places not reached by me: Nickaparoo (or Nickaparu) Creek, branch of the Ireng; Maripicru, a branch of the Ireng between Wontyke and Karakara, above the Karona falls; Chipoo, a tributary of the Ireng between Karakara and the Rupununi; Papan, near Eworora; Twoca Pan, between the Rupununi and Pununike; Rupununi, opposite Massara Landing; creek between Rapoo and the lower falls; Packeoo should possibly be "Pacu" Falls, in the Rupununi; Gatuck Creek, Potaro Highland; Yakeatonuk Fall, Potaro River. Some of these names are spelled in several ways on the different labels. 4 MEMOIRS OF THE CARNEGIE MUSEUM Most of the drawings illustrating the paper were prepared by Mr. W. S. Atkinson of Stanford University. The photo-drawings are for the most part the combined work of Professor W. Cogshall of Indiana University, Miss Maud Sie- benthal, and myself. The final arrangement of the figures on the plates was the work of the Editor, to whom I am indebted for much other kind help and crit cism. In the preparation of the account of the Ariince I had the assistance of Mr. Owen Frazee. Mr. Elmer Deem assisted with the Doradince, and Miss Lola Vance with the Pimelodince. The account of the Gymnotidce was prepared by Dr. Max Ellis. I have given no figures of the Gymnotidce, having reserved these for the monograph of the family by Dr. Ellis, which is ready to go to press. The Scicenidce were described by Mrs. Marian Durbin Ellis, and the accounts of Hemigrammus and Hyphessobrycon are extracted from the monograph of these genera by Mrs. Ellis, which is still in manuscript. Mrs. Rosa Smith Eigenmann has been of the greatest assistance in the prepara- tion of this paper. Deprived of the pleasures and responsibilities of more formal collaboration, she has critically read the manuscript and overseen the work of the artists. CHAPTER I. BRITISH GUIANA. In speaking of the country I cannot do better than to reproduce in part the excellent account from the pen of Wilgress Anderson, F.R.G.S., contained in a work upon the goldfields published a couple of years ago by Mr. J. B. Harrison. The pages of the original are indicated in brackets: 2 General Physical and Topographical Features of British Guiana. [p. 9.] "Situation and Extent. — The region called Guiana, or Guyana, stretches along the northern coast of South America from the mouth of the Orinoco River to that of the Amazon River, and inwards to Brazil. . . . " The only European possessions in South America are three in number, and are situated on the central portion of this territory, which is divided into the colo- nies of the British, Dutch, and French Guiana. " Of these colonies, the most westerly is that of British Guiana, which extends from the eastern limits of Venezuela, westward to Dutch Guiana, and north of Brazil to the coast on the Atlantic Ocean, its extreme limits touching the parallels of 0° 41' and 8° 33' 22" north latitude, and the meridians of 56° 20^' and 61° 23' 24.7" west longitude. " British Guiana has a seaboard of about 270 miles trending in a southeasterly direction, with a mean depth of about 500 miles, and is equal in extent to the com- bined size of England, Scotland and Wales, the area being about 90,000 square miles, most of wdrich is densly covered with exuberant primeval forest, but in some parts there are broad open flats and undulating grassy plains, or savannahs, and mountainous grass-clad country. "Physical Features. The Alluvial Belt. — The colony may be divided broadly into two low-lying belts near the coast and a hilly and mountainous hinterland which constitutes by far the largest area. "The coast-lands are flat and for the most part swampy, being slightly de- pressed below the level of ordinary spring-tides, so that sea-walls and other defences 2 The Geology of the Goldfields of British Guiana, by J. B. Harrison, 8vo, pp. i-ix, 1-320, 43 plates. London, Dulau & Co. 5 b MEMOIRS OF THE CAENEGIE MUSEUM have to be constructed to protect the settled parts of the coast-lands from being flooded by high tides. They form part of an alluvial belt which rises gradually from the sea-level and extends inland for a distance varying from ten to forty miles, and which is composed of variously colored clays with intermediate layers of sand and peat, the latter being locally known as pegass. "The margins of this formation along the sea and rivers are covered with a dense growth, consisting principally of mangrove (Rhizophora mangle) and of courida (Avicenna nitida), which form natural sea-defences, the former being found along the western and the latter along the eastern parts. Behind this growth are flat grassy savannahs [p. 10] interspersed with forest, consisting mostly of Aeta and Trooli palms (Mauritia fiexuosa and Manicaria sacdfera), whilst in some parts the land is covered with a dense jungle. It is on this belt that all the sugar estates and by far the greater part of the cultivated areas are situated. " The Sand and Clay Belt. — The alluvial belt is succeeded by a slightly elevated and undulating belt composed of sandy and clayey sedimentary soils, derived from the disintegration of the various country rocks in situ, and traversed in some places by sand-dunes, which rise from fifty to about one hundred and eighty feet above the sea-level. This second belt commences at the Waini River, in the north- western district, and gradually increases in width as it extends toward the eastern boundary of the Courantyne, in the vicinity of which it attains its greatest depth at about one hundred miles inland. Grass-covered downs occur on the banks of the Berbice and Courantyne Rivers, but the greater part of this tract consists of high forest, and along the river margins and in the valleys mora trees (Dimorph- andra mora) grow plentifully. "The Hinterland. — Beyond these belts, southward, the country rises between the river valleys, which are in many parts swampy, and as it approaches the sources of the larger rivers attains a height of about nine hundred feet above the sea-level at the source of the Takatu, the western boundary, and about four hundred feet above the sea at the source of the Courantyne, the eastern boundary. This more elevated portion occupies about eleven-twelfths of the area of the colony. It is diversified by numerous low hills and valleys, and contains three principal mountain ranges, several irregularly distributed smaller ranges, and in its southern and eastern parts many scattered isolated mountains, none of the last mentioned being more than fifteen hundred feet above sea-level. "The eastern portion is almost entirely forest-clad, yet the country on the western side of the colony, between the Rupununi and Ireng Rivers and extending southwards from the Pakaraima Mountains to the Kanuku Range consists of an EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 7 almost flat grass-clad plain or savannah, elevated about three hundred feet above sea-level, in which, in the vicinity of and bordering upon the many streams by which it is watered, are patches of woodlands. From the Kanuku Mountains southwards to about six miles from the source of the Takutu and from that river eastwards to a considerable distance beyond the Rupununi there is an extensive and undulating elevated savannah with similar patches of woods along the valleys of the many streams by which it is drained. Beyond this the extreme southern part of the colony is entire'y forest-clad. [p. 11.] "Rivers. — Of the numerous river-systems there are six principal ones, viz.: — (1) The Essequibo and its principal tributaries, the Mazaruni and Cuyuni; (2) The Courantyne and its tributary, the New River; (3) The Berbice and its tributary, the Canje; (4) The Waini and its tributary, the Barama; (5) The Demerara; (6) The Barima. "These, together with the following smaller ones, the Abary, Mahaicony, Mahaica, Boerasirie, Pomeroon, and Maruka Rivers, flow to the Atlantic Ocean. "In addition to the above mentioned rivers, there are the Takatu and its tributary, the Ireng, which meet together at 3° 34' north latitude and form the Brazilian boundary. The Takuta River flows thence to the Rio Negro, the waters of which join the Amazon. "The Essequibo, the largest river in the colony, rises in 0° 41' north latitude, about eight hundred and fifty feet above the sea-level and flows in a northerty direction for some six hundred miles. It is joined at Bartika, about forty miles from its mouth, by the Mazaruni River, a tributary, which is itself joined at Cartabo, five miles above Bartika, by another tributary called the Cuyuni River, all these combining to form an estuary with a width of about three miles below their junction, and which expands to a width of fourteen miles at the mouth, con- taining, as it approaches the sea, three islands, each of which is about twelve miles in length, and in addition many smaller ones. The river is navigable for large vessels as far as Bartika, and for small launches to the foot of the first rapids, eighteen miles above that point. Beyond this its course is broken by many rapids and cataracts, and about five miles above the junction of the Rupununi the Esse- quibo river is practically unnavigable on account of the many long series of cataracts and falls, which obstruct its course. [p. 12.] "The Rupununi River joins the Essequibo in 4° 2' 52" north latitude, 8 MEMOIRS OF THE CARNEGIE MUSEUM and gives access during the rainy seasons to the elevated grass-clad plains, or savannahs, on which at the present time a large number of cattle are being raised. During the height of the dry seasons the river becomes very shallow, its course being impeded by many sandbanks and rapids, and during this time it can only be ascended with great difficulty and much loss of time, or rather not at aU. Its many inlets and large lake-like ponds on both banks form a feature common also to the upper parts of the Berbice River. Its largest tributary is the Rewa, or Illiwa, which is itself joined by the Quitaro, and both flow through country covered with high forests." [p. 14.] "The Demerara River, although commercially the most important and best known of all the rivers in the colony, is, compared with some of those already described, a small one. As the greater depth at the bar admits of large vessels entering this river with more security and ease than is the case with any of the other rivers in the colony, Georgetown, the capital and principal port of the colony, hSs been established on its east bank at its mouth, which is there three-quarters of a mile wide, and furnishes a safe harbor for the many steamers and sailing vessels which frequent the port. The Demerara River takes its rise in the small mountain- range called the Maccari, which is really an offshoot of the great Pakaraima range. It has a generally northerly course, and flows between the Essequibo and the Berbice Rivers. The river is navigable for steamers for nearly eighty miles upwards from its mouth, and beyond this for launches for about twenty-four miles further up, as far as the Malali Rapids, where the influence of the tide ceases. Above Malali the river is again navigable for launches as far as Kanaimapoo, above which are the Kumaparu Rapids, where the Demerara River approaches nearest in its course to the Essequibo River. The first great cataract on this river is situated a short distance above Kamaparu, in latitude 5° 18' north, and is known as the Oruru- Malali, or Great Falls. Beyond Oruru-Malali the river is sluggish, and is again navigable for boats as far as the Cannister Cataracts, where it divides into two streams. Its forest-clad banks are flat as far up as the second or sand and clay belt, where the sand hills occur and form the first high land." "Mountain Ranges. — One of the most prominent features of the country is the great central mass of mostly flat-topped mountains, known as the Pakaraima group or chain, which occupies the most western [p. 15] portion of the colony, and stretches southward from the Cuyuni River to within thirty miles above the mouth of the Ireng River, and eastward to the Essequibo River right across the colony as far as the Courantyne River." "The bulk of these mountains forms a successive series of terraces and broad, EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 9 undulating plateaus, with bold, and in some cases, perpendicular sandstone escarp- ments, varying in height from about twelve hundred to over two thousand feet, and eventually forming a large undulating table-land at an average height of about thirty-five hundred feet above the sea-level. In many parts of the mountains and table-lands great and deep gorges have been eroded by the rivers and streams which traverse them. They attain their greatest height at Mount Roraima and Mount Kukenaam, both of which rise over eight thousand five hundred feet above the sea. The portion of this range which extends westward and down the southern bank of the Mazaruni River to the vicinity of the Teboco Cataracts retains the striking flat-topped features and is known as the Merume Mountains." "The elevated table-land (8635 feet) of Mount Roraima is about twelve square miles in area, and on it the boundaries of the colony with those of Venezuela and Brazil meet at a common point. This very remarkable mountain, together with Mount Kukenaam, is a part of one of the most extensive sandstone formations on the globe, and they both rise with perpendicular cliffs of sandstone two thousand feet in height above the base of the surrounding country." "The highest plateaus, such as Mount Roraima, are mostly bare, exposed expanses of rock, between the crevices of which grow many rare and curious orchids and other flowering plants, besides some low bushes and extremely stunted trees." General Geology. 3 [p. 19.1 " The Coast-Lands. — The alluvial deposits are of considerable, but unknown thickness. As they rest upon beds of pipe-clay or impure kaolin it is a matter of great difficulty to decide whether the borings for underground waters, which have from time to time been made in various parts of the coast-lands, have been wholly in the alluvium, or have, as they certainly have done in some cases, penetrated through these beds into the underlying residuary clays. In places, however, the alluvial deposits have been proved for depths of over two hundred feet, and it is possible that in many places their thickness far exceeds this. The cores of the borings show that the alluvial deposits consist of beds of more or less indurated marine muds and sands which have been laid down so as to form beds of clay, of mixed clay and very fine siliceous sand, locally known as 'caddy.' and of siliceous sands varying much in texture — some beds consisting of sand of ex- tremely fine texture, others of coarser grain, while others again approach in char- acter fine grits or gravels. In places some of the beds contain considerable quantities 3 The notes upon the Geology are derived from the same work hitherto quoted, and from the pen of Mr. J. B. Harrison, the Government Geologist of British Guiana. 10 MEMOIRS OF THE CARNEGIE MUSEUM of decomposing vegetable debris, and these, when drilled into during deep well- sinking operations, in some cases give off inflammable mixtures of gases containing marsh gas in considerable quantities. In places the effusion of the gases has been accompanied by that of small quantities of petroleum, a decomposition product of the organic matters. The geological age of these beds is uncertain, the lower parts may be of late Tertiary or Pleistocene age, while the parts now bordering the coast- line are undoubtedly recent. I am inclined to think that their age is, in part at least, similar to that of the Moruga sands of Trinidad. The sand-beds of these deposits are not unfrequently exposed in the cultivated parts of the coast-land, where they are known as sand-reefs. These form in places oval patches of land raised a few feet above the general level of the surrounding [p. 20] argillaceous soils, and in many others gives rise to long narrow ridges somewhat raised above the general level of the land which they traverse. Their mode of occurrence indicates that they are purely local modifications of the alluvial deposits, — sands separated from the mass of the marine silt by the action of local currents and of the waves,— and thus the sand-beds form more or less lenticular beds occupying, as a rule, no great area. They are, in my opinion, very distinct from the beds of sand which characterize various phases in areas where the land is either rising or falling to any extent in the vicinity of a shallow sea. " The general evidence indicates that British Guiana occupies one of the most stable areas of the earth's surface, — one which has been very slowly rising through long ages* — this slow movement having given rise to the low rapids which usually mark the termination of the tide-way in the rivers, and possibly which has so altered the contour of parts of the continent on which the colony is situated as to change the main lines of drainage, and thus to make the rivers relatively small streams traversing the deeper parts of the courses and valleys eroded by their predecessors in earlier periods. During a stage in this slow upheaval the low hills already mentioned as occurring in a few places in the alluvial coast-land were in turn rocks and small islands in the shallowing sea which then surrounded them, as now they are surrounded by an apparently unbounded expanse of forest or of marsh. "A remarkable feature in parts of the alluvial coast-land is the occurrence of extensive beds of a kind of peat. This is locally known as 'pegass,' and consists of the more or less altered remains of ferns, mosses, and sedges, and of other marsh- loving plants. It resembles in its general character the upper layers of the vegetable matter which are found in peat-bogs in temperate climates. As far as my observa- * The Italics are mine. C. H. Eigenrnann. . . EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 11 tions go it is never as compact as true peat. This is probably due to the deposits of it being seldom more than from two to four feet in thickness. "As pointed out by Sir Charles Lyell in his 'Principles of Geology,' a large portion of the sand and clays of the alluvial deposits has been brought by the cur- rents from the mouth of the Amazon River; the burden brought by the present rivers of the colony from the higher districts through which they flow having been, during recent periods, a very subordinate factor in the accumulation of this wide- spread formation; although perhaps in earlier times, before the land had risen to its present level, the river-borne silt may have contributed a large quota to the mass. "The Forest Lands and Residuary Deposits. — The alluvial strata extend to depths varying from five to, in places, as much as thirty-five miles from the coast- line and rest upon beds described by C. B. Brown as 'sand and clay deposits.' " [p. 21.] "These residual deposits cover and hide the true country over vast areas of the lower-lying parts of the colony, and form the characteristic sub-soils and soils of our forest regions. "The parts of the areas covered by these residuary deposits which abut upon the true alluvial beds are in many places traversed by long ranges of sand-dunes, giving rise to low hills, which, as in the case of the range traversed by the Demerara- Essequibo Railway, may attain a height of somewhat over two hundred feet. As a rule, their heights do not exceed one hundred or one hundred and twenty feet. The sand, of which the upper parts, at any rate, of these dunes (it is possible that in many cases they cover ridges of the residuary deposits) consist, is glistening, white, quartz sand, the grains of which are usually uniform in size over relatively large areas, the majority being well rounded, thus accentuating the wind-blown origin of the dunes. "C. B. Brown notices that the beds form a low escarpment at the southern limit of the fluvio-marine deposit, and that this has been taken for a ridge running parallel to the coast. I have not had opportunities of repeating this observation, but, accepting its accuracy, I consider that the ridge approximately marks the shore- line, which existed at the commencement of the deposit of the present fluvio-marine [p. 22] alluvium. Observations made in the forest regions since the time of Brown's geological reconnaissances of the interior of the colony have shown that the resid- uary deposits cover the country not alone on the plains of the lowlands, but along the great river-valleys and on the lower forest-covered parts of many of the hills and mountains. The heavy, at times torrential, tropical rains have carved out of the residuary coverings deep ravines and valleys; and the gravels, sands, and 12 MEMOIRS OF THE CARNEGIE MUSEUM silts derived from the eroded material have been laid down in the wide valleys along the courses of parts of the rivers as fluviatile loams, gravels, and sands. " The Sandstone and the Diabase Intrusions. — Large areas of the interior of the colony are occupied by a thick stratified formation of sandstone and conglomerate. Just as the basal igneous rocks are, so is this, pierced and traversed by dykes of diabase, hence the latter rock must be of later origin than all except the sedimentary coverings and the fluvio-marine deposits. The blue-grey rock varies much in depth of color and texture, and its varieties will be described in the chapters dealing with the petrography of the colony. "The diabase intrusions occur in belts, generally stretching across the colony in a north-westerly and south-easterly direction. The intrusions vary from narrow dykes, only exposed in the courses of the rivers during very dry seasons, some being not more than from two to three feet across, to low hills and to mountain ranges, some of which — for example, the Eagle mountains in the Potaro gold district — exceed in height two thousand feet. The tops and sides of the hills and mountains, except where they have suffered great denudation, are covered with ironstone gravel, while the lower parts of the district in which diabase forms the [p. 23] country are covered up with strata of laterite, frequently over one hundred feet in depth, and in places interspersed with nests of secondary quartz, or traversed by veins and stringers of quartz, or, less often, by lenticular layers of secondary quartz, closely resembling, when cut through by mining shafts, tunnels and trenches, — true quartz- reefs. The quartz rock in all these forms is not unfrequently auriferous, the metal being dispersed through it in a very irregular manner, especially in the large lenticu- lar layers, which in many parts are nearly, or even entirely, barren of gold, and in others are "bonanzas" carrying at rates from twenty to, in places, several hundred ounces of the precious metal to the ton of the rock. Unfortunately hitherto these bonanzas have proved few and far between; but there is no reason for assuming that they will not be found in many places in the enormous area of the laterite deposits which up to the present has not been prospected, as they have been in similar places at intervals in the past. Gold also occurs as paint gold, as gold dust, and as nuggets of very varying sizes in the laterite. "Of earlier age than the diabase is the sandstone and conglomerate series. It constitutes the greater portion of the Pacaraima mountains, and spreads west- wardly into Venezuela. A similar formation occurs in Brazil, and in all probability is part of the same massif as the Guiana one. Wherever it occurs it appears to be unfossiliferous, and hence we have no paleontological evidence with regard to the geological period at which it was deposited. Two conjectures have been made EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 13 as to this. C. B. Brown arrived at the conclusion on what appears to me to be somewhat defective evidence (its, in parts, reddish color, its unfossiliferous nature, and its being penetrated by masses, dykes, and sills of greenstone — diabase — as are sandstones of Triassic age in North America), that it is an equivalent of the New Red sandstone. In Venezuela its relationship to rocks of known age is said to be recognizable, and it is stated to be of Cretaceous age. A like conclusion that the northern parts of the format'on are of Cretaceous age has been arrived at in Brazil. If these views are correct, the later outbreaks of diabase, which are, directly or indirectly, the causation of many of the auriferous deposits of British Guiana, must be either of Cretaceous age, or belong to the Tertiary or to a later period. And as there is a very great resemblance in the magmatic character of the Guianan diabase, and of the lavas of the West Indian province, whose outbreaks are clearly of Tertiary and of present age, the assumption of the relatively recent age of the diabase is a plausible one. As will b? mentioned in a later chapter, the diabase shows no signs of the effects of the regional metamorphism which has materially affected many of the rocks underlying the sandstone formation. "The only evidence available in this colony with regard to the sandstone and the geological period of its formation is that wherever its base has been seen it occupies an analogous position to the Torridonian sandstones of the Scottish High- lands, to which the sandstone has a close resemblance in constitution. It lies invariably on the presumably Archean rocks of the colony; and its constituents, as far as I have been [p. 24] able to examine them, show no signs of having, even in part, been derived from later rocks. If it is of Cretaceous age it offers an interesting example of the recurrence of similar formations in widely divided geological ages, when the conditions affecting their formation and deposition are identical. Per- sonally I am not prepared from my own observations and studies to accept any statements of its geological age further than that shown by its relationship to the underlying gneiss, porphyries, felsite and schists derived from them. "The sandstone formation spreads eastwardly through the colony, crosses the Essequibo River in a low narrow belt at Comuti Mountain, gives rise to the Maccari Mountain in Demerara, and crossing the Berbice River near Marlissa Rapids, is seen forming a low mountain range at Itabru near that river. It passes into Dutch Guiana across the Courantyne River near its union with the Cabelebo River, and also in its higher reaches. The formation consists of beds of coarse conglomerate, red and white sandstones of very varying textures, and in places of strata of red shale. "High mountains occur in the sandstone formation, which consist of coarse 14 MEMOIRS OF THE CARNEGIE MUSEUM textured diabase or of rather fine-grained gabbro. This rock shows signs of meta- morphism, in places being granulitic in structure and in others being changed to a considerable extent, either by the development in it from augite of a dark-brown secondary biotite, or the pyroxene is altered from an almost colorless mineral to a brown-colored strongly dichroic one. "Mr. C. Wilgress Anderson, who in 1895 spent several months in traversing the sandstone district while nquiring into the alleged occurrence of beds of aurifer- ous conglomerate in it, and has since crossed it repeatedly during the Boundary Surveys, is of opinion that the diabase gabbro is of greater age than the sandstone, the latter formation in places resting on or abutting against it, and this view is upheld by its structure. The hills were probably small islands in the shallow seas in which the sandstone formation was laid clown. C. B. Brown, on page 14 of the 'Reports on the Geology of British Guiana,' mentions the occurrence of great layers of conglomerate in the neighborhood of 'greenstone,' and this is confirmatory of Mr. Anderson's view. The possible existence of 'greenstone' of two distinct geological ages and modes of occurrence does not seem to have struck Brown and Sawkins, but it offers an intelligible explanation of the facts recorded by them in their reports. These surveyors estimated the total thickness of the sandstone on the assumption that it is traversed by three layers of greenstone at about three thousand feet. As, however, it is probable that some at least of the latter diabase, as, for instance, that at Roraima, is in the form of laccolites, and during intrusion has elevated great tracts of the sandstone country, probably the formation has not the total thickness deducible from C. B. Brown's figures, and may at present not anywhere exceed in thickness that shown at Roraima — about two thousand feet. As a rule the sandstone lies nearly horizontally, dipping somewhat to the north, and few faults are seen in it, although in places near where diabase has intruded into it [p. 25] there are well-marked local disturbances in its dip. Many of the beds of sandstone of finer texture show well-marked current-bedding." CHAPTER II. EARLIER WORKS ON THE FISHES OF BRITISH GUIANA. The first notice of the fishes of the area covered by this report, which I have been able to discover, is Van der Lott's "KortBericht von den Congeraal, ofte Drilvisch," Verhandl. Holl. Maatsch., Harlem, 1762. Van der Lott was at the time a surgeon on the Essequibo, who attributed so many "medical properties" to the electric eel that he "acquired no increase in reputation therefrom in this colony," as Ban- croft says. In his "An Essay on the Natural History of Guiana," London, 1769, Edward Bancroft publishes a letter dated "Demerary, Aug. 15, 1766," in which he mentions or describes (p. 188) the "Lowlow," the "Bafroketa" (Arapaima?), (p. 189) "Pen" (Serrasalmo?) , (p. 190) "Saw-fish, Flounders, Brasilian Soles, Surinam Mackarel, Drummers, Old-wifes, Mullets, and a species of Anchovies." He devotes several pages, 190-202, to the "torporific eel," attributing the shock to electricity. This was several years before Hoist (1772) demonstrated that the peculiar power of the torpedo is due to electricity. Hillhouse in his "Indian Notices," a book which I have not seen, is said to enumerate " wenty-six species as peculiar to the coast, estuaries, and rivers of Guiana." Hancock in his "Notes on some species of Fishes and Reptiles from Demerara, presented to the Zoological Society by John Hancock" (Zoological Journal, Vol. IV, pp. 240-247, 1829; also I sis) describes: Doras costatus = Doras hancockii Cuv. & Val. Callichthys littoralis = Hoplosternum littorale (Hancock). Hypostomus watwata = Plecostomus watwata (Hancock). Hypostomus muUiradiatus = Pterygoplichthys vndtiradiatus (Hancock). Loricaria brunnea = Loricariichthys brunneus (Hancock). But the first real work in gathering fishes and knowledge concerning them began with the explorations of Robert Hermann Schomburgk, conducted during the years 1835-1839 under the auspices of the Royal Geographical Society and the British Government. The explorations were continued by Robert Hermann 15 16 MEMOIRS OF THE CARNEGIE MUSEUM Schomburgk and his brother, Richard Schomburgk, during the j^ears 1840-1844. Robert in the second expedition went as an explorer for the British Government. Richard Schomburgk was sent along by the king of Prussia at the instigation of Alexander Von Humboldt. Robert H. Schomburgk during his first expedition collected but few specimens, to which we may now refer. But he had many drawings made and furnished notes concerning the species drawn. The notes were edited, the drawings named, and the whole published by Jardine as "Fishes of British Guiana," Parts I and II, forming Volumes XXXIX and XL of "The Naturalists Library." Schomburgk not only explored the rivers flowing northward, but made a long tour across the head of the Orinoco, down the Rio Negro, and up the Rio Branco. Fifty-three of the eighty-three species noted in the volumes seem to have come from the basin of the Rio Negro, only thirty being definitely ascribed to streams flowing northward. I long ago expressed the opinion that some of the drawings are composites, and that in some cases a wrong combination of figure and description was made by the editor. 4 Nevertheless the two little volumes form a notable contribution to the knowledge of the fishes of Guiana. Schomburgk presented to the Jardin des Plantes and to the British Museum several specimens, which were made the types of new species. Whether they were collected during his first or second expedition I do not know. Schomburgk's first journey was described in his "Reisen in Guiana und am Orinoco wahrend der Jahre 1835-1839," published by O. A. Schomburgk, Leipzig, 1841. The journey was divided into several longer or shorter excursions. 1. He left Georgetown September 21, 1835, and ascended the Essequibo to the Cuyuni, up which he went a short distance, and then continued on up the Esse- quibo to the Rupununi and up this river to near its source. He then returned to the Essequibo and went up this stream to the William IVth Cataract, then down to the Siparuni, up which he went some distance, returning March 28, 1836, to Georgetown. 2. On September 2, 1836, he again left Georgetown and ascended the Couran- tyne to some large cataracts. 3. On November 25th he started up the Berbice River and went to the parallel 4 Schomburgk's difficulties may be appreciated from liis statements, Vol. I, 82-83: "But with the ex- ception of my Indian friends and Dr. Fleming's Philosophy of Natural History I had nothing to guide me in my researches" and "The first specimen of any fish, a drawing of which we did not previously possess, served generally to sketch its outward forms and general colors on the paper; and when we were fortunate enough to secure a second specimen those delicate hues were painted in, which are only visible immediately after the fish comes out of the water." EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 17 of the junction of the Rupununi and Essequibo and started over to the Essequibo and returned by way of the Berbice. 4. On September 12, 1837, he again ascended the Essequibo and Rupununi to the Rewa, up which he went and crossed over to the Cuyuwini, an upper tribu- tary of the Essequibo. He descended it to the Essequibo, which he ascended to its source. He then returned to the Rupununi about Lake Amucu, 6 whence he went to Fort St. Joaquin on the Takutu. He left this place on September 20, 1838, sailing up the Takutu to its tributary, the Mahu, swinging from here northwestward to Roraima, thence southwestward to the Parima, up this river, and then, following the mountains, to Esmeralda on the Orinoco, along the Cassiquiare to the Rio Negro, down this river to the Rio Branco, and up this stream back to Fort San Joaquin on April 22d. He returned thence to Georgetown on June 20, 1839. The species he figured and noted, together with the identification of those dealt with in the present work, are given in the following list. Those from the northern slope are marked with an asterisk; those which have been found on the northern slope since Schomburgk's day are marked f. 1. Acanthicus histrix Spix, p. 131, pi. 1. t 2. Loricaria cataphracta Linnaeus, p. 136. 3. Hypostoma plecostomus Valenciennes, p. 139. * 4. Hypostoma squalinum Schomburgk, p. 142, pi. 2. 5. Hypostoma punctatum Schomburgk, p. Vol. I. Rio Branco at Fort San Joaquin. Locality? = Loricaria cataphracta (Linnaeus). Rio Branco, 8 in. = Plecostomus plecostomus (Linnaeus). Rio Branco, Rio Negro, Essequibo. 13J4 in- = Plecostomus emarginatus (Cuv. & Val.). Rio Branco. 63^ in. 150, 145, figure. 6. Hypostoma barbatus Cuv. & Val., p. 147. Loc ? Q}i in. = Pseudancistrus barbatus (Cuv. & Val.). Pools, marshes, and creeks. 8" in. = Hoploslernum thoracatum (Cuv. & Val.). Curassarraka on the Rupununi. 12 in. = Doras costalus (Linnteus). Rio Negro. = Doras cataphractus (Linnaeus). Pasawiri. 7 in. =? Doras cataphractus (Linnaeus). * 7. Callichthys longifilis Cuv. & Val., p figure. * 8. Callichthys ? p. 152, figure, t 9. Doras costatus Lacepede, p. 155. flO. Doras cataphractus Linnaeus, p. 158. 11. Doras caslaneo-ventris Schomburgk, p. 161, pi. 3. *12. Doras brunnescens Schomburgk, p. 163. 13. Doras niger Valenciennes, p. 165. 14. Doras, p. 166. 6 In recent works this word is spelled Amacu. Upper Essequibo. 5 in. = ? Doras cataphractus (Linnaeus). Esmeralda. = Oxydoras niger (Val.). Orooporary on the Essequibo. 18 MEMOIRS OF THE CARNEGIE MUSEUM *15. Phractocephalus hemilopterus Agassiz, All rivers of Guiana. 4 ft. = Phractocephalas p. 169, figure. hemiliopterus (Bloch & Schneider). 16. Arius oncina Schomburgk, p. 173, pi. 4. Rio Padauiri. 10 in. = ? 17. Arius obesus Schomburgk, p. 174. Rio Branco. 8 in. = ? *18. Pimelodus (Bagrus) maculatus Lacepede, Most of the rivers. 12 in. = Pimelodus clarias p. 175, figure. (Bloch). *19. ? ? ? p. 176. Demerara and Essequibo. 18-20 in. *20. Pimelodus arekaima Schomburgk, p. 178, Upper Essequibo and Rio Branco. 2 ft. 3 in. = pi. 5. Rhamdia arekaima (Schomburgk). f21. Pimelodus insignis Schomburgk, p. 180, Rio Branco. 18 in. = Pimelodella cristata (Miiller pi. 6. & Troschel). 22. Pimelodus notatus Schomburgk, p. 181, Rio Branco. 3 feet. pi. 7. *23. Pimelodus pirinampu Spix, p. 183. Rivers of Guiana. 3 feet or more. = Piniram- pus pirinampu (Spix). 24. Platystoma tigrinum Valenciennes, p. 185, Most rivers of Guiana. = Pseudoplatystoma fas- pi. 8. datum (Linnaeus). 25. Platystoma planiceps Agassiz, p. 187. Rio Branco. 26. Platystoma vaillanti Valenciennes,' p. 188. Guiana. To 2 ft. 3 in. = Brachy platystoma vaillanti Valenciennes. 27. Hypophthalmus dawalla Schomburgk, p. Rivers of Guiana. 2J^ft. = Ageneiosusbrevifilis 191, pi. 9. (Cuv. & Val.). *28. Lau-Lau, p. 193, figure. Rivers of Guiana. 10-12 ft. 200 lbs. = Platy- stoma vaillanti Valenciennes. *29. Sudis gigas Cuvier, p. 198, pi. 17. Rupununi, Rio Branco, Rio Negro. 15 ft. 410 pounds = Arapaima gigas (Cuvier). *30. Osteoglossum aroivana Schomburgk, p. 205, Rupununi, Essequibo, Rio Branco, Rio Negro. pi. 10. = Osteogloss^l7n bicirrhosum Vandelli. 31. Chalceus rotundatus Schomburgk, p. 209, Padauiri. 6-7 in = Chalcinusrotundatus (Schom- figure. burgk). *32. Chalceus tceniatus Schomburgk, p. 210. Essequibo, Rio Negro, Rio Branco. 15-18 in. = ? 33. Chalceus labrosus Schomburgk, p. 212, pi. Rio Paduiri. 7 5J4 in. =? Brycon falcatus Miiller 13. & Troschel. *34. Chalceus nigrotceniatus Schomburgk, p. Loc. ? 14-16 in. = Leporinus nigrotceniatus 213, pi. 13. (Schomburgk). 35. Chalceus latus Schomburgk, p. 214. Padauiri, tributary of the Rio Negro. 4 in. 36. Chalceus fasciatus Schomburgk, p. 215. Padauiri. 18 in. *37. Chalceus macrolepidotus Cuvier, p. 216, pi. Essequibo. 15 in. = Chalceus macrolepidotus 14. Cuvier. 6 This species is given on the authority of Valenciennes. Its identity with the Lau-lau was not appre- ciated by Schomburgk. 7 Padauiri? EIGKNMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 19 f38. Anodus notatus Schomburgk, p. 218, pi. 15. f39. Serrasalmo piranha Spix, p. 221, pi. 16. *40. Serrasalmo stagnatilis Schomburgk, p. 222. ?41. Serrasalmo punctatus Schomburgk, p. 223, pi. 17. ?42. Serrasalmo ? p. 224. *43. Serrasalmo niger Schomburgk, p. 225, pi. 18. ?44. Salmo emarginatus Schomburgk, p. 231, pi. 19. 45. Salmo undulatus Schomburgk, p. 232, figure. 46. Serrasalmo scotopterus Schomburgk, p. 233. *47. Myletes pacu Schomburgk, p. 236 pis. 20 &21. 47a, b, c. Morocoto* cartabac, palometo, pp. 239, 240. *48. Tetragonopterus lotus Schomburgk, p. 241. 49. Tetragonopterus schomburgkii Jardine, p. 243, pi. 22. • *50. Xiphostoma ocellatum Schomburgk, p. 245, pi. 23. Rio Negro. = Anisitsia notatus (Schomburgk). Rio Branco. 10-11 in. = Pygocentrus piraya (Cuvier). Upper Essequibo, in pools. 8 in. = ? = Pygopristis denticulatus (Cuvier). All rivers of Guiana. 16 in. = Pygocentrus niger (Schomburgk). Loc. ? = Metynnis sp.? Padauiri. 6 in. = ? Rio Branco. Dikes of Essequibo. 24 in. = Myleus pacu (Schomburgk). All rivers of Guiana. = IMetynnis, sp.? Rio Negro. 2 ft. Hy- Essequibo, Rios Negro and Branco. drocynus cuvieri (Agassiz). 8 "The morocoto, cartabac, and a species of pacu which we found in the river Parama, and which differed only from the common pacu in its colour being black, constitute a group of fishes which resemble each other by structure, teeth, habits, and their being phytivorous. One of the most delicious among this division is the morocoto or osibu of the Warraus; it inhabits only the estuaries, and does not occur in fresh water; it would fall, therefore out of the limits of the present descriptions; but as it is so closely allied with the pacu, I shall mention at least its dimensions and general appearance. The teeth, which consist of fourteen in the upper jaw, and are placed in a double row in the fore part, are all distinctly molar or grinding teeth. It attains a length from about twenty-five to twenty-eight inches, and is twelve inches in depth. The gill covers consist of three strong bones, the dorsal fin of sixteen rays, the ventral of eight, and the anal of twenty-four, the caudal is compressed and thin; in every other respect it resembles the pacu, and is extremely fat and delicious. During the month of August, when they feed upon the fruit of the caramacata, a tree of large size and very hard wood, and the bark, leaves, and fruit of which is extremely bitter, their flesh has a bitter taste, but otherwise it is much sought after, and large numbers of it are occasionally brought from the mouth of the Orinoco to George- town. It forms the chief support of the Warrau Indians who inhabit the coast regions in the vicinity of the estuaries of the Orinoco and the months of the Guainia and Barima." "The palometo, which is about fourteen inches long and seven inches in depth, its body compressed and flat, with a thin sharp belly, is equally well flavoured as the morocoto and pacu, and frequents similar haunts as the morocoto." (Fishes of British Guiana, pp. 239-40.) I have not identified these species which are mentioned only by their Indian names. 20 MEMOIRS OF THE CARNEGIE MUSEUM *51. Hydrocyon microlepis Schomburgk, p. 247, Essequibo, Rios Negro and Branco. = Acestro- pl. 24. rhynchus microlepis (Schomburgk). *52. Hydrocyon armatus Schomburgk, p. 249, All rivers of Guiana. 10-12 lbs. = Acesirorhyn- pl. 25. chus falcirostris (Cuvier). 53. Schizodon fasciatus Spix, p. 252, pi. 26. Rio Branco. = Schizodon fasciatus Spix. *54. Erythrinus macrodon Schomburgk, non Berbice, beyond the cataract, Itabru, Cuyuni. Agassiz, p. 254, pi. 27. 4 ft. = Hoplias monophthalmus Pellegrin. *55. Huri, Cauhui, Tari-ira,* p. 256. In every river of Guiana. = Hoplias malabaricus (Bloch). *56. Prochilodus rubrotamiatus Schomburgk, p. Essequibo, Rios Branco and Negro. 2 lbs. - 258, pi. 28. Prochilodus rubrotceniatus Schomburgk. 57. Prochilodus binotatus Schomburgk, p. 260, Rio Branco. 13}^ in. pi. 29. 58. Prochilodus insignis Schomburgk, p. 261, Rio Branco. llj^ in. pi. 30. Vol. II. 59. Belone guianensis Schomburgk, p. 131, pi. Padauiri. 15 in. = Potamorhaphis guianensis 1 . (Schomburgk) . *60. Sciccna rubella Schomburgk, p. 133. Most rivers. 2 feet. = Plagioscion squamosis- simus Heckel. *61. Corvina grunnicns Schomburgk, p. 136, Essequibo. = Pachypops grunniens (Schom- pl. 2. burgk). *62. Cychla labrina Agassiz, p. 139, pi. 3. Upper and lower courses. 6-7 inches. = Cre- nicichla saxatilis (Linnagus). 63. Cychla fasciata Schomburgk, pi. 141, p. 4. Loc? = Crenicichla Johanna Heckel. 64. Cychla rutilans Schomburgk, p. 142, pi. 5. Rio Branco. = Crenichla lugubris Heckel. 65. Cychla fiavomaculata Schomburgk, p. Rio Negro and Padauiri. 2 ft. = Cichla ocellaris 145, pi. 6. (Bloch & Schn.). 66. Cychla nigro-macidata Schomburgk, p. 147, Rio Negro and Padauiri. 18 nches. = Cichla pi. 7. ocellaris (Bloch & Schn.). *67. Cychla argus Valenciennes, p. 149, pi. 8. Essequibo; Rio Branco; Rio Negro. = Cichla ocellaris (Bloch & Schn.). 68. Cychla trijasciatus Schomburgk, p. 151, Rio Negro; Padauiri. = Cichla ocellaris (Bloch & pi. 9. Schn.). 69. Cychla ? rubro-ocellatus Schomburgk, p. Rio Negro. = ? 153, pi. 10. 70. Centrarchus cychla Schomburgk, p. 157, Rio Negro. = ? pi. 11. 71. Centrarchus niger Schomburgk, p. 159, pi. Rio Negro. 6%. = ? 12. 72. Centrarchus notatus Schomburgk, p. 160, Loc? = Cichlasoma severum (Heckel). pi. 13. 9 This species is not listed under its scientific name. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 21 73. Centrarchus f vittatus Schomburgk, p. 161, Loc? = ? pi. 14. 74. Centrarchus ft rostralus Schomburgk, p. Rio Negro. 4.4 inches. = Acaropsis nassa 163, pi. 15. (Heckel). *75. Centrarchus f cyanopterus Schomburgk, p. Essequibo. 3 in. = ? Cichlasoma bimaculatum 165, pi. 16. (Linnaeus). 76. Pomotis f fasciatus Schomburgk, p. 169, Rios Padauiri and Negro. 8.5 in. pi. 17. *77. Pomotis f bono Schomburgk, p. 171, pi. 18. All rivers, and in pools and marshes. 6.5 in. = Mquidens tetramerus (Heckel). 78. Gymnotus eleclricus Linnaeus", p. 173, pi. Rio Negro. = Electrophorus electricus (Linnaeus). 18. 79. Gymnotus fasciatus Pallas, p. 174, pi. 19. Rio Branco. = Gymnotus carapo Linnaeus. 80. Trigon histrix f D'Orbigny, p. 180, pi. 20. River Roowa. 12 inches. = Potamotrygon hys- trix (Muller & Henle). 81. Trygon garrapa Schomburgk, p. 182, pi. Rio Branco. = Potamotrygon hystrix (Midler & 21. Henle). 82. Trygon strongylopterus Schomburgk, p. Rio Branco. = Paratrygon strongylopterus 183, pi. 22. (Schomburgk). 83. Elipesurus spinicauda Schomburgk, p. Rio Branco. 184, pi. 23. 84. Silurus parkeri Trail, p. 188, pi. 24. = Sciadeichthys parkeri (Trail). The brother, Richard Schomburgk, in the expeditions which have been men- tioned, ascended the Demerara River to near the "Great Fall," He also threaded the Pomeroon, Waini, and Barima Rivers, which discharge their waters west of the Essequibo. He ascended the Essequibo to' the Rupununi, following the latter to near its sources. From Pirara he ascended the Takutu to its source and de- scended it to Fort San Joaquin, whence he went to Roraima. His last collections, made on the trip to Roraima, were all lost. An account of his journeys was published in two volumes under the title "Reisen in British Guiana in den Jahren 1840-1844 im Auftrag Sr. Majestat des Konigs von Preussen." His collections were enumerated in a third volume. The new fishes were largely described in the "Horse Ichthyologicse," by Miiller & Troschel, who also prepared the account of the fishes in this third volume. He gives many notes on the habits of the fishes in the first two volumes. Mtiller & Troschel enumerate one hundred and forty-one species as represented in his collections. 10 Most of these I had an opportunity of examining in the Zoological Museum in 10 To these Schomburgk adds ten species known to him, but not recognized by Miiller and Troschel. Some of these, like Myletes pacu, as Schomburgk himself noted, were given by Miiller & Troschel under other names. 22 MEMOIRS OF THE CARNEGIE MUSEUM Berlin during July of 1910. On account of lack of time no effort was made to verify the marine or estuarine species. The list of species with the identification of those dealt with in the present volume is given in the following table. 1. Centropomus undecimalis Cuvier & Val- Coast. 1-2 ft. = Centropomus undecimalis enciennes, p. 620. (Bloch). 2. Serranus galeus Muller & Troschel, p. 621. Coast. 3. Pomotis calesbei Cuvier & Valenciennes, p. 621. 4. Otolithus toe-toe Cuvier & Valenciennes, Coast. 8-10 in. = Cynoscion acoupa (Lac6- p. 621. pede). 5. Otolithus leiarchus Cuvier & Valenciennes, Coast. 12-14 in. p. 621. 6. Ancylodon jaculidens Cuvier & Valen- Coast. 6-8 in. = Macrodon ancylodon (Bloch & ciennes, p. 621. Schneider). 7. Micropogon lineatus Cuvier & Valen- Coast. 1-2 ft. = Micropogon furnieri (Desrnar- ciennes, p. 621. est). 8. Micropogon trifilis Muller & Troschel, p. Coast. 16-18 in. = Pachypops furcrwus Lace- 622. pede. 9. Polycentrus schomburgkii Muller & Tros- Essequibo. = Polycentrus schomburgkii Eigen- chel, p. 622. mann. 10. Gerres rhombeus Cuvier & Valenciennes, p. Coast. 10-12 in. 622. 11. Acharnes speciosus Muller & Troschel, p. Mouth of Essequibo. 6-8 in. = Cichla ocellaris 622. Bloch & Schneider. 12. Chorinemus salie?is Cuvier & Valen- Coast. 2-6 ft. ciennes, p. 623. 13. Caranz carangus Cuvier & Valenciennes, Coast. 2-6 ft. p. 623. 14. Mugil liza Cuvier & Valenciennes, p. 623. Mouths of rivers. 18-20 in. 15. Mugil curema Cuvier & Valenciennes, p. Mouths of rivers. 16-18 in. 623. 16. Gobius bacalaus Cuvier & Valenciennes, p. Coast. 623. 17. Eleotris guavina Cuvier & Valenciennes, p. Mouths of rivers. 8-10 in. = Guavina guavina 623. (Cuvier & Valenciennes). 18. Batrachus surinamensis Bloch & Schnei- Coast. 8-10 in. der, p. 623. 19. Monochir maculipinnis Agassiz, p. 624. Coast. 20-26 in. 20. Acara margarita Heckel, p. 624. Amucu Swamps. 8-10 in. = Cichlasoma bi- maculaia (Linnaeus). 21. Acara nassa Heckel, p. 624. Lake Tapacuma. 4-6 in. = Acaropsis nassa Heckel. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 23 22. Acara tetramerus Heckel, p. 624. Lakes Tapacuma, Capoye, and Amucu. 4-6 in. = Mquidens tetramerus (Heckel). 23. Acara heckelii Miiller & Troschel, p. 624. Swamps and savannahs. 4-6 in. = Acarichthys heckelii (Miiller & Troschel). 24. Chcetobranchus flavescens Heckel, p. 625. Amucu. 6-8 in. = Cha>tobranchus flavescens Heckel. 25. Geophagus jurupari Heckel, p. 625. Amucu, swamps and savannahs. 8-10 in. = Geophagus jurupari Heckel. 26. Geophagus surinamensis Miiller & Tros- Lakes Tapacuma, Capoye, and Amucu. 4-6 in. chel, p. 625. = Geophagus surinamensis Miiller & Troschel. 27. Geophagus leucostictus Miiller & Troschel, Amucu. 4-6 in. = Geophagus jurupari Heckel. p. 625. 28. Geophagus pappaterra Heckel, p. 625. Amucu (Tributary of Branco). 4 in. = Geophagus jurupari Heckel. 29. Cichla ocellaris Bloch & Schneider, p. 625. All rivers. 2J^ ft. = Cichla ocellaris Bloch & Schneider. 30. Crenicichla saxatilis Heckel, p. 626. All rivers. = Crenicichla sazatilis Heckel. 31. Crenicichla vittata Heckel, p. 626. Essequibo, Lakes Tapacuma and Capoye. = Cren- icichla lugubris Heckel. 32. Crenicichla lugubris Heckel, p. 626. Essequibo and neighboring swamps. = Crenicichla lugubris Heckel. 33. Tylosurus guianensis Miiller & Troschel, Coast. 1-2 ft. = Tylosurus almeida Quoy & p. 626. Gaimard. 34. Bagrus mesops Cuvier & Valenciennes, p. Waini and Barima. 2 ft. = Selenaspis herz- 627. bergii (Bloch). 35. Bagrus proops Valenciennes, p. 627. Waini and Barima. 18-20 in. = Sciadeichthys proops (Cuvier & Valenciennes). 36. Bagrus passany Valenciennes, p. 627. = Selenaspis passany (Cuvier & Valenciennes). 37. Bagrus clarias Miiller & Troschel, p. 627. Waini & Barima. 12-14 in. = Pimelodus clarias (Bloch). 38. Bagrus coelestinus Miiller & Troschel, 627. Waini & Barima. = Selenaspis herzbergii (Bloch). 39. Bagrus emphysetus Miiller & Troschel, p. Waini and Barima. = Sciadeichthys emphysetus 627. (Miiller & Troschel). 40. Platystoma tigrinum Valenciennes, p. 627. Nearly all rivers. 2Yi ft. = Pseudoplaty stoma fasciatum (Linnseus). 41. Platystoma platyrhynchus Valenciennes, p. Rupununi. = Hemisorubim platyrhyncos Cuvier 628. & Valenciennes. 42. Galeichthys gronovii Valenciennes, p. 628. Waini and Barima. = Felichthys bagre (Linnseus) . 43. Pimelodus sebce Valenciennes, p. 628. All rivers. 8-10 in. = Rhamdia sebce (Cuvier & Valenciennes). 44. Pimelodus raninus Valenciennes, p. 628. All rivers. 6-8 in. = Pseudopimelodus villosus Eigenmann. 24 MEMOIRS OF THE CARNEGIE MUSEUM 45. Pimelodus cristatus Muller & Troschel, p. Takutu and Mahu. 16-18 in. = Piinodella cris- 628. lata (Muller & Troschel). 46. Pimelodus foina Muller & Troschel, p. 628. Takutu. 7-8 in. = Rhamdella foina (Muller & Troschel. 47. Pimelodus eques Miiller & Troschel, p. 628. All rivers. 16-18 in. = Goeldiella eques (Muller & Troschel). 48. Pimelodus stegelichii Muller & Troschel, p. Forest brooks. 10-12 in. = Rhamdia sebce Cu- 628. vier & Valenciennes). 49. Calophysus macropterus Muller & Tros- Essequibo. 12 in. = Callophysus macropterus chel, p. 629. (Lichtenstein). 50. Auchenipterus maculosus Valenciennes, p. Essequibo. 4-6 in. = Trachycorystes galealus 629. (Linnaeus). 51. Auchenipterus furcatus Valenciennes, p. Essequibo. 6-8 in. = Pseudauchenipterus nodo- 692. sms (Bloch). 52. Doras armatulus Valenciennes, p. 629. Rupununi and Awaricuru. 10 in. = Doras costa- tus (Linnseus). 53. Doras niger Valenciennes, p. 629. All rivers. 10-12 in. = Oxydoras niger (Valen- ciennes). 54. Doras carinatus Valenciennes, p. 629. Essequibo. 10-12 in. = Hemidoras carinatus (Linnaeus). 55. Doras maculatus Valenciennes, p. 629. Essequibo. 2 ft. = Doras grayiulosus (Valen- ciennes). 56. Callichthys coelatus Cuvier & Valenciennes, Trenches. 4-6 in. = Callichthys callichthys p. 630. (Linnaeus). 57. Callichthys exaratus Muller & Troschel, Trenches. 4-6 in. = Hoplosternum thoracatum p. 630. (Cuvier & Valenciennes). 58. Callichthys pictus Muller & Troschel, p. Trenches. 4-6 in. = Hoplosternum thoracatum 630. (Cuvier & Valenciennes). 59. Aspredo lozvis Valenciennes, p. 630. Waini. = Aspredo aspredo (Linnaeus). 60. Aspredo tibicen Temminck, p. 630. Coast. 10-12 in. = Aspredinichthys tibicen (Temminck). 61. Acanthicus hystrix Spix, p. 630. Takutu and Rio Branco. 3-3^ ft. = Acanthicus hystrix Spix. 62. Loricaria cataphracta Linnaeus, p. 631. Rupununi. 8-10 in. = Loricaria cataphracta Linnaeus. 63. Loricaria acuta Valenciennes, p. 631. Rupununi. 8-10 in. = Loricariichthys microdon Eigenmann. 64. Loricaria platyura Muller & Troschel, 631. Rupununi. 8 in. = Loricariichthys platyura (Muller & Troschel). 65. Hypostomus commersonii Valenciennes, p. Takutu. 6-8 in. = Plecostomus verres (Cuvier 631. & Valenciennes). 66. Hypostomus itacua Valenciennes, p. 631. Takutu. 3—4. = Hemiancistrus braueri Eigen- mann. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 25 67. Hypostomus temminckii Valenciennes, p. Takutu. 2-3 in. = Ancistrus temminckii Cuvier 631. & Valenciennes. 68. Hypostomus nudiceps Miiller & Troschel, Takutu. 2-3 in. = Xenocara gymnorhynchus p. 631. (Kner). 69. Anableps letrophthalmus Bloch, p. 632. Mouths of streams. 6-8 in. = Anableps ana- bleps (Linna?us). 70. Anableps microlepis Miiller & Troschel, Mouths of streams. 4-6 in. = Anableps micro- p. 632. lepis Miiller & Troschel. 71. Pcecilia vivipara Bloch & Schneider, p. 632. Georgetown canals and others. J-^ in. = Poecilia vivipara Bloch & Schneider. 72. Erythrinus unitamiaius Spix, p. 632. Swamps, streams, and brooks of the Canuku Mts. 8-10 in. = Hoplerythrinus unitceniatus (Spix). 73. Erythrinus salmis Agassiz, 632. Small forest streams and standing water. 10-14 in. = Hoplerythrinus unitceniatus (Spix). 74. Macrodon trahira Miiller, p. 632. Generally distributed in British Guiana. = Hop- lias macrophthalmus Pellegrin. 75. Macrodon brasiliensis Miiller, p. 633. Generally distributed, British Guiana. 12-14 in. = Hoplias malabaricus (Bloch). 76. Anodus alburnus Miiller & Troschel, p. Lake Amucu and swamps of the savannah. Up to 633. 10 in. = Curimatella alburna (Miiller & Tros- chel). 77. Anodus ciliatus Miiller & Troschel, p. 633. Lake Amucu. 6-8 in. Curimatus ciliatus (Miiller & Troschel). 78. Pacu nigricans Spix, p. 633. Generally distributed, British Guiana. 6-8 in. = fPygocentrus piraya (Cuvier). 79. Hemiodus unimaculatus Miiller & Troschel Essequibo. 6-8 in. = Anisitsia notata Schom- p. 633. burgk". 80. Piabuca argentina Cuvier, p. 633. Lake Amucu and swamps of the savannah. 3^4 in. = Piabucus dentatus Kcelreuter. 81. Chilodus punctatus Miiller & Troschel, 634. Still water of the savannah. 3-4 in. = Chilodus punctatus Miiller & Troschel. 82. Schizodon fasciatus Agassiz, p. 634. Upper Rupununi, Rio Branco, Takutu, and the swamps near these. = Chilodus punctatus Miiller & Troschel. 83. Leporinus fasciatus Miiller & Troschel, p. Pirara and nearby swamps. = Leporinus fasci- 634. atus Bloch. 84. Leporinus nigrotozniatus Miiller & Tros- Abundant in upper Pomeroon and its tributary chel, p. 634. forest brooks. 4-6 inches. = Leporinus nigro- taniatus (Schomburgk) . 85. Leporinus maculatus Miiller & Troschel, p. Abundant in Rupununi and Awaricuru. 6-8 in. 634. = Leporinus maculatus Miiller & Troschel. 86. Leporinus frederici Agassiz, p. 634. Pomeroon. 12-14 in. = Leporinus friderici Bloch. Bloch. 26 MEMOIRS OF THE CARNEGIE MUSEUM 87. Tetragonopterus argenteus Artedi, p. 634. Lake Amucu. 4-6 in. = Tetragonopterus argen- teus Artedi. 88. Tetragonopterus maculatus Miiller & Tros- Rupununi and Essequibo and the swamps and chel, p. 634. morasses near by. 3-4 in. = Pcecilurichthys bimaadatus (Linnaeus). 89. Tetragonopterus melanurus Miiller & Tros- Upper Rupununi. 4-6 in. = Creatochanes caud- chel, p. 635. omacidatus Giinther. 90. Tetragonopterus tmiiatus Jenyns, p. 635. Trenches and swamps along the coast. 1-2 in. = Moenkhausia oligolepis (Giinther). 91. Chalceus angulatus Spix, p. 635. Essequibo and Rupununi. 6-8 in. = Chalceus rotundatus (Schomburgk). 92. Brycon macrolepidotus Miiller & Troschel, Lower Essequibo and Mazaruni. 6-8 in. = p. 635. Chalceus macrolepidotus Cuvier. 93. Brycon falcatus Miiller & Troschel, p. 635. Generally distributed, British Guiana. 6-8 in. = Brycon falcatus Miiller & Troschel. 94. Brycon schomburgkii Miiller & Troschel, Plentiful in lower Essequibo. 6-7 in. = Brycon p. 635. falcatus Miiller & Troschel. 95. Brycon pesu Miiller & Troschel, p. 635. Lower Essequibo and Mazaruni. 4-6 in. = Holobrycon pesu (Miiller & Troschel). 96. Exodon paradoxus Troschel, p. 635. Plentiful in creeks of the upper Rupununi. 4-6 in. = Exodon paradoxus Troschel. 97. Epicyrtus gibbosus Miiller & Troschel, p. Plantation drains and lower Essequibo. 4-6 in. = Charax gibbosus (Linnseus). 98. Xiphoramphus falcatus Miiller & Troschel, Essequibo and Pomeroon. 6-8 in. = Acestro- p. 635. rhynchus falcatus (Bloch). 99. Xiphoramphus microlepis Miiller & Tros- Pomeroon, upper Essequibo, Rupununi, and chel, p. 636. Takutu. = Acestrorhynchus microlepis (Miiller & Troschel). 100. Hydrolycus scomberoides Miiller & Tros- Generally distributed. 2-3 ft. = Hydrolycus chel, p. 636. scomberoides (Cuvier). 101. Agoniates halecinus Miiller & Troschel, p. Cuyuni. 6 in. = Agoniates halecinus Miiller & 636. Troschel. 102. Xiphostoma cuvieri Spix, p. 636. Upper Essequibo, Rupununi, and Takutu. 2 ft. = Hydrocynus cuvieri (Agassiz). 103. Pygocentrus piraya Miiller & Troschel, p. Generally distributed in British Guiana. 10-12 in. 636. = Pygocentrus piraya (Cuvier). 104. Pygocentrus nigricans Miiller & Troschel, Generally distributed. 11-12 in. = Pygocentrus p. 636. piraya (Cuvier). 105. Pygocentrus niger Miiller & Troschel, p. Generally distributed, British Guiana. 16-20 in. 636. = Pygocentrus niger (Schomburgk). 106. Pygopristis denticulatus Miiller & Troschel, Essequibo, Rupununi, Takutu, and near-by p. 637. swamps. = Pygopristis dentiadatus (Cuvier). EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 27 107. Pygopristis fumarius Miiller & Troschel, p. 637. 108. Serrasalmo rhombeus Miiller & Troschel, p. 637. 109. Serrasal?7io aureus Spix, p. 637. 110. Catoprion mento Miiller & Troschel, p. 637. 111. Myletes rubripinnis Miiller & Troschel, p. 637. 112. Myletes schomburgkii Miiller & Troschel, p. 637. 113. Myletes hypsauchen Miiller & Troschel, p. 637. 114. Myletes latus Miiller & Troschel, p. 638. 115. Myletes asterias Miiller & Troschel, p. 638. 116. Myleus setiger Miiller & Troschel, p. 638. 117. Osteoglossum bicirrhosum Spix, p. 638. 118. Arapaima gigas Miiller, p. 638. 119. Megalops atlanticus Valenciennes, p. 639. 120. Elops saurus Linnaeus, p. 639. 121. Engraulis thrissoides Miiller & Troschel, p. 639. 122. Gymnothorax ocellatus Agassiz, p. 638. 123. Gymnotus eleclricus Linnaeus, p. 639. 124. Sternopygus virescens Miiller & Troschel, p. 639. 125. Sternopygus lineatus Miiller & Troschel, p. 640. 126. Rhamphichthys rostratus Miiller & Troschel, p. 640. Rupununi, Essequibo, and near-by swamps. 4-6 in. Pygopristis denticulatus (Cuvier). Rupununi, Takutu, and near-by swamps. 12-14 in. = Serrasalmo rhombeus (Linnaeus). Essequibo and Rupununi. 6-8 in. = ? Serrasal- mo gymnogenys Giinther. Lake Amucu. 3-4 in. = Catoprion mento (Cuvier). Lower Essequibo. 4-6 in. = Myloplus rubri- pinnis (M. & T.). Rupununi, Takutu, Zuruma, as well as swamps of the savannah. 4-6 in. = Myleus pacu (Schom- burgk) . Standing water, Tapacuma Lake. 3^ in. = Metynnis hypsauchen (Miiller & Troschel). Generally distributed in British Guiana. 10-12 in. = Myloplus rhomboidalis (Cuvier). Essequibo and Mazaruni near falls. 2 ft. = Myloplus asterias (Miiller & Troschel). Essequibo near cataracts and in swift water. 10- 12 in. = Myleus pacu (Schomburgk). Still water of Rupununi, Takutu, Rio Branco and near-by swamps of savannah, and in Lake Amu- cu. J/2 ft. = Osteoglossum bicirrhosum Van- delli. Generally distributed in streams of British Guiana. 8-10 ft. = Arapaima gigas Cuvier. Along the coast. 2 ft. = Tarpon atlanticus (Cuvier & Valenciennes). Along the coast. 12-14 in. Cuyuni. 4-7 in. = Stolephorus sjrinifer (Cuvier & Valenciennes). Plantation drains and standing water. 3-4 ft. = Lycodontis ocellatus (Agassiz). All fresh water of British Guiana. 7 ft. = Elec- trophorus electricus (Linnaeus). Lake Amucu and forest streams. 18-20 in. = Eigenmannia virescens (Valenciennes). Small forest brooks. 6-10 in. = Eigenmannia virescens (Valenciennes). Demerara River. 4-6 ft. = Rhamphichthys ros- tratus (Linnaeus). 28 MEMOIRS OF THE CARNEGIE MUSEUM 127. Sternarchus oxyrhynchus Miiller & Tros- chel, p. 640. 128. Synbranchus marmoratus Bloch, p. 640. 129. Chelichthys punctatus Miiller & Troschel, p. 641. 130. Chelichthys psittacus Miiller & Troschel, p. 641. 131. Chelichthiji asellus Miiller & Troschel, p. 641. 132. Syngnathus pelagicus Linnaeus, p. 641. 133. Carcharias (Prionodon) henlei Valencien- nes, p. 641. 134. Carcharias (Prionodon) oxyrhynchus Miil- ler & Henle, p. 642. 135. Sphyrna tudes Miiller & Henle, p. 642. 136. Pristis pectinatus Latham, p. 642. 137. Tamiura motoro Miiller & Henle, p. 642. 138. Trygon garapa Schomburgk, p. 642. 139. Trygon strongyloplerus Schomburgk, p. 642, 140 Hypostomus plecostomus Valenciennes, p. 643. 141. Hypostoynus squalitus Schomburgk, p. 642. 142. Phractocephalus hemiliopterus Agassiz, p. 643. 143. Pimelodus arekaima Schomburgk, p. 643. 144. Pimelodus insignis Jardine, p. 643. 145. Hypophthalmus dawalla Schomburgk, p. 643. 146. Myletes pacu Schomburgk, p. 644. 147. Prochilodus rubro-tceniatus Schomburgk, p. 644. 148. Trygon histrix Schomburgk, p. 644. Lower Essequibo. 16-18 in. = Sternarchorhyn- chus oxyrhynchus Miiller & Troschel. Drains of the plantations. 2-3 ft. = Synbran- chus marmoratus Bloch. Sand banks of the coast. 12-15 in. Mouths of the Waini and Barima and sand-banks of the coast. 4-6 in. = Colomesus psittacus (Bloch & Schneider). Fresh water, Barima. 4 in. = Colomesus psit- tacus (Bloch & Schneider). Salt water near the coast. 10-12 in. Found on whole coast, especially near the mouth of the Demerara. 4-6 ft. Coast, especially near the mouth of the Demerara. 6-8 ft. Whole coast, especially mouth of Demerara. 4-6 ft. Coast. 4-6 ft. Mouth of the Zuruma. 6-8 in. = Potamotrygon hystrix (Miiller & Troschel). Takutu and Rio Branco. 8-10 in. = Potamotry- gon hystrix (Miiller & Troschel). Rupununi, Takutu, and Rio Branco. 8-10 in = Paratrygon orbicularis Bloch & Schneider. Takutu and Rio Branco. 8 in. = Plecostomus. plecostomus (Linnaeus). Essequibo, Rio Branco, and Takutu. = Plecos- tomus emarginatus Cuvier & Valenciennes. Generally distributed in British Guiana. 2 ft. 11 in. to 4 ft. 2 in. = Phractocephalus hemiliop- terus Agassiz. All rivers of the savannah. 2-3 ft. = Rhamdia arekaima Schomburgk. Takutu and Rio Branco. = Pimelodella cristata (Miiller & Troschel). Rivers of British Guiana. 2-3 ft. = Ageneiosus brevifilis Cuvier & Valenciennes. = Myletes pacu Schomburgk. Essequibo and tributaries. 18-20 in. = Pro- chilodus rubrotomiatus Schomburgk. Rupununi, Takatu, and Rewa. = Potamotrygon hystrix (Miiller & Troschel). eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 29 The Lau-lau is mentioned without referring it to any other name. One other collection of note was made in British Guiana by a Mr. A Ehrhardt. He collected at Mocco Mocco and Arisaro, spending about two months on the Esse- quibo. Part of his collection is in the British Museum, part in the Berlin Museum. The portion in the British Museum was reported upon by Giinther in his " Catalogue of Fishes in the British Museum," and in a short paper " On New Species of Fishes from the Essequibo," published in the Annals & Magazine of Natural History, December, 1863. The new things described are: Acara punctulata = Crenicara punctulata (Giinther) ; Pimelodus holomelas = Rhamdia holomelas (Giinther) ; Auchenipterus obscurus = Trachycorystes obscurus (Giinther); Helogenes marmoratus Giinther; Crenuchus spilurus Giinther; Leporinus megalepis Giinther = Leporinus maculatus Mutter & Troschel, and Xiphoramphus ferox Giinther = Acestrorhynchus falcatus (Bloch). As I have already stated in the introduction, in none of these papers was a definite locality given. In recent years Mr. T. Sidney Hargreaves collected fresh-water fishes in British Guiana. They were deposited in the Georgetown Museum. Few have definite locality labels. Mr. Hargreaves published a series of articles in "The Argosy," a local paper, which were reprinted in book form. The booklet is only of local interest, and adds but little which is of a sufficiently specific nature to be quoted in a scientific work. CHAPTER III. GENERAL ACCOUNT OF THE EXPEDITION. Accompanied by Mr. S. E. Shideler, I sailed from New York on August 23, 1908, arriving in Georgetown on September 6th. From August to December is the long dry season in Guiana. In consequence the upper portions of the rivers are lowest in October and November, and the fishes are then concentrated in the channels of the streams. We had rain during the first week of our stay in Guiana, but later were only once interrupted by rain or high water. While on the Guiana plateau at Holmia a rain lasting a day and a night caused the river to rise many feet. ' mmmi Fig. 1. Mouth of the Demerara River at Georgetown, British Guiana. The two main objects of the expedition have been outlined in the Introduction. I desired to secure as many characins as possible and to compare the fauna of the plateau with that of the lowland. The former became an incident in the study of the latter question. To get an idea of the lowland fauna a series of collections. in fresh water was made at sea-level within tidal influence from Lama Stop-Off to Morrawhanna, Wismar, Malali, and Bartica. Above tidal influence collections were made at Rockstone, Crab Falls, Ivonawaruk, and Warraputa, in the Essequibo and along the Potaro from its mouth to the Kaieteur. The fauna of the plateau was studied in the Potaro from the Kaieteur to the Aruataima Cataract. 30 EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 31 Our equipment consisted of two barrels of alcohol, ten gallons of formalin, a Baird collecting-net, having a quarter of an inch mesh, fifteen feet long and five feet deep, a similar net forty-five feet long and seven feet deep, a net one hundred and fifty feet long with one-inch mesh in the wings and half-inch mesh in the other seventy-five feet, several gross of vials, several glass stoppered bottles, numerous empty one-pound coffee-tins, several five-gallon galvanized iron cans with screw- tops, one large galvanized iron can, buckets, etc. Kerosene tins were used in shipping specimens home. Immediately upon my arrival I rented a house of two rooms named "Charity," which stood between two others called "Faith" and "Hope." The preliminaries of attending to the customs, moving in, making several trips to the bank and the office of the Consul before my draft was accepted, making calls, and securing transportation, consumed Monday and Tuesday, September 7th and 8th. On Wednesday at 5 A.M. I visited the market, and from that time forward until Fig. 2. A trench in Georgetown. my return from the Kaieteur every calory of energy was consumed for the one object of making the trip successful. Georgetown is but a few inches above high tide; parts of it, in fact, are below tide. Many of the streets are provided with a wide central trench for drainage. Some of these are provided with tide-locks. One canal brings water from the Lamaha Conservancy, concerning which a little more later. The trenches are 32 MEMOIRS OF THE CARNEGIE MUSEUM planted with various native and exotic water-plants, one with Victoria regia, one with lotus, and so forth. The Lamaha trench is abundantly supplied with in- digenous water-plants. We began seining in the trenches of Georgetown as soon as our baggage was cleared and other preliminaries were attended to, and devoted four days to them. Toward the end of our stay in the country Mr. Shideler devoted himself again to the trenches, and particularly to those of the Botanic Garden, one of which he drained of its water, the results being given in a list published in a later chapter. We took altogether thirty-nine species (to which Mr. Ellis has added another) from the trenches, four of which we found nowhere else in Guiana: Plecostomus watwata, the sea-hassar, Ctenobrycon spilurus, Pcecilia vivipara, and Acanthophacelus melanzonus. The first is found from Georgetown along the coast to Para. The second has been taken only in Georgetown and Surinam. A close relative is exceedingly Fig. 3. View of the lowlands looking seaward from the railway near Mahaica. abundant in the Amazons. The third is widely distributed in the West Indies and South America, while the last has not been taken elsewhere. The water in most of the trenches in which we collected is fresh; that in a muddy pond at the lower end of the town may be at times contaminated with brackish water. Few of the species characteristic of brackish water were taken in the trenches. There were no EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 33 marine cat-fishes or Banjamans, both groups excessively abundant in the brackish water. The species most abundant were Pristella riddled, a highly colored little characin first reported from the Orinoco and also found at Wismar; Hemigrammus rodwayi, another little characin confined to Georgetown and the northwest coast; the widely distributed characins Pcecilurichthys bimaculatus and Charax gibbosus, the latter of which was most abundant of all; the eels, Sternopygus macrurus, Eigenmannia virescens, the poecilid " Millions " and the cichlid Cichlasoma bimaculatum. «i yank* fflfc. , w^^ SH ^I^Hi^^iiH ' Fig. 4. View of the drainage canal at Cane-Grove Corner. The fauna of the trenches as a whole is poor as compared with that of the interior. This character the trenches share with the streams of the northwestern territory near the coast, and the waters about Lama Stop-Off . Many of the species are so small that our nets having one-quarter inch mesh permitted them to get through without difficulty. This may in part account for the apparent restriction of some of the species in the different trenches. Of Pristella riddlei we took with our nets but eight specimens in the Georgetown trenches, while in the trench of the Botanic Garden, which was drained, we obtained two hundred and thirty-three. Having packed the specimens from Georgetown we left on the 15th for Lama 34 MEMOIRS OF THE CARNEGIE MUSEUM Stop-Off. It may be reached either by boat directly through the canal, or by rail and boat by way of Mahaica. Arrangements were made for us to go via Mahaica. The railroad runs through a flat country grazed by cattle, often standing up to their knees, or even deeper, in water. In places cocoanut-plantations abound and at frequent intervals canals run across the country to the ocean. Burnt clay is used in ballasting the railroads and in building the roads. Between the station at Mahaica and Cane-Grove Corner are a number of sugar-estates, all traversed by canals. 11 At Cane-Grove Corner the Lama Water Conservancy begins. It is a large tract of swampy savannah converted into a pseudo-lake, or water reservoir, by surrounding it with a ditch or canal and an outer embankment. At Lama Stop-Off and Maduni Stop-Off two streams, the Lama and the Maduni, tributaries of Mahaica Creek, formerly draining the savannah, are " stopped off " or dammed. 11 The following from Rodway gives a picture of a single plantation: " One of the principal estates situated on the east coast of Demerara is two hundred roods on facade by the full depth of two thousand two hundred and fifty roods, i. e., about half a mile wide by five and a half deep. In front is the seashore, to protect which mangrove and courida bushes are allowed to grow, inside of which a dam of earth is thrown up, the excavation alongside forming a drain for carrying off any salt water that may come over during high tides. At a short distance within the front dam comes the public road, which extends along the coast, and which is kept up at the expense of the estate owner, as far as it extends through his property. Beyond the road, which with its two canals at the sides forms a second dam, comes about a mile of grassy land which is used for pasturing cattle, horses, and mules belonging to the plantation. Then comes the railway, near which is the draining engine and kokers or sluices of the canefields, which commence immedi- ately behind this third defence. Beyond a mile or so of pale green sugar-cane come the plantation buildings, which consist of the sugar factory, manager's residence, house for the overseers, hospital, school-house, one or two shops, and the labourers' cottages, which last are very numerous. This group of buildings forms, to all intents and purposes, a self-contained village, the manager's house, standing in the midst of a fine garden, representing the mansion of the squire, while round liim live as many mechanics and labourers as are necessary to carry on the cultivation and factory. Beyond this village come interminable fields of canes as far as the cultivation extends, where a back dam protects it from floods. The plantation under review had some years ago nine hundred and thirty-five acres, or exactly half its area, planted with canes, and produced over fifteen hundred hogsheads of sugar annually. The remaining portion comprised three hundred and fifty-eight acres in pasture and bush, one hundred and sixty-two acres not then empoldered, one hundred and ninety-eight acres in dams, parapets, and trenches, one hundred and sixty-two acres of swamp in front, above low water mark, but outside the sea dam, and fifty-three acres covered by the buildings, garden, public road, railway, etc. From the number of acres in dams and trenches it may be seen how important this part of the economy of the plantation must be. On every hand is an earthen dam with corresponding canals, these latter being cut off from outside by flood-gates, so that no water from sea or swamp can penetrate, while the rainfall of the plan- tation itself is run off through the sluices at low water, or in very heavy weather, by means of the draining engine. These draining canals are connected with other trenches between every field, and these again with the ditches of each bed of canes. With such a perfect system of canals it has naturally followed that sugar canes are brought to the factory by water, and to complete the communication a middle dam and two canals are car- ried through the center of the plantation to the factory and thence up to the railway, or to the shipping-trench, where the droghers take the produce to the port by sea." EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 35 We embarked on the canal a Cane-Grove Corner and went about nine miles to Lama Stop-Off, where we made our headquarters, as the guests of Mr. St. Aubyne from the 15th to the 19th of September. Lama Stop-Off, according to a map at hand, is about twenty-two miles from Georgetown in a straight line. Lama Creek below the dam is influenced by the tide so that the top of the dam is but a few feet above Georgetown. Fishes are exceedingly abundant at Lama Stop-Off, although the number of species does not seem to be large. Undoubtedly several more species might have been secured if we could have used dynamite or poison. We arrived at nine in the evening and fishes could be heard jumping out of the water in all directions. We were awakened in the morning by the howling of monkeys. The entire force at the command of Mr. St. Aubyne was put to work to use every device known to him, or brought by us, to gather everything there was in the water. A naked negro baby caught the enthusiasm, walked into the edge of the water with a market basket, and made a dip, catching a specimen of Carnegiella! It Fig. 5. Photograph of Carnegiella strigata in aquarium. was the first time the fish had been caught since it was described by Gunther from his poor material without a known locality. Naturally I was very anxious to catch more. In spite of all our efforts we did not catch a second specimen at this place. We found it later at Maduni Stop-Off and in other places. Cichlids were very abundant here, and the delicious luckananee (Cichla) was more abundant here than we found it elsewhere. It gave me particular pleasure, in honor of Mr. St. Aubyne, to apply the name Pristella aubynei to the species from this place, of which we collected more specimens han of any other. We seined both in the canal and below the dams at Maduni Stop-Off and Lama Stop-Off. In all we secured forty-nine species, five of which were not taken elsewhere: Rhamdia holomelas, Ageneiosus brevifilis, Nannostomus simplex, Pristella aubynei, and Hyphessobrycon minimus. The first three are certainly found elsewhere 36 MEMOIRS OF THE CARNEGIE MUSEUM in the colony, and the fifth is a minute species imperfectly known from the few specimens secured. Although the Georgetown trenches are a direct continuation of the canal at Lama Stop-Off, we secured twenty-eight species at the latter point which we did Fig. 6. View on the right bank of the Demerara River. not get at Georgetown. Part of the twenty-eight have come from below the dam at Lama Stop-Off, but even this reach of water is but recently disconnected from the general system. From Lama Stop-Off we returned to Georgetown on the 19th of September, packed and forwarded the fishes collected to the United States, and prepared to go inland. Mr. Shideler left for Wismar on the 23d and I followed on the morning of the 24th. The steamer left Georgetown at 8 A.M. and reached Wismar at 4:30 P.M. The water is muddy until Berlin is approached and becomes blackish further up. Wismar is about sixty-five miles above Georgetown in a straight line. The Demerara is navigable for ocean-going sailing-vessels to this point, and is affected by the tide to the first cataract at Malali, about one hundred miles from Georgetown, in a direct line. The entire region from Georgetown to Wismar is flat, except for occasional sand-hills. Creeks enter the Demerara from both sides about Wismar. At Christianburg a creek has been dammed and a canal brings the water to the sawmill on the river. We collected in the Demerara river at Christian- burg and at Wismar, in the Christianburg canal, and in the creeks emptying into the EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 37 Demerara. It had become evident that during the heat of the day but few fishes could be taken in the shallow water of the main stream, so that from this time on the daytime was used to explore for likely places, which were then seined at night. Most of the mouths of the creeks about Wismar are provided with partial fences built of poles and palm or banana leaves. The center is ordinarily left open for the flow of the tide. A mat can be placed in the gap, which will prevent fishes from coming out of the creek. When the tide is high at night and fish have Fig. 7. View of creek, filled with brush- wood above Wismar at low tide. Indians are poisoning the creek. left the main stream and entered the creek the mat is put in place. In the morning when the tide is out the fishes trapped are either killed with a cutlass, or poisoned. Many of our specimens were obtained in this way. The creeks are so full of brush that all ordinary methods of fishing are out of the question. Trap-nets could probably be used effectively instead of the fence in the mouths of the creeks. Except in a very few favorable places the banks and shallows of the Demerara are so profusely overgrown with Caladium arborescms that the seine could not be used. I engaged fishermen to collect for me some distance below Wismar and had a creek poisoned at Kumaka, several miles above Wismar. From all of these places I secured ninety species, five of which were not taken elsewhere. The speci- mens not taken elsewhere are: Ageneiosus guianensis, Poecilobrycon harrisoni, 38 MEMOIRS OF THE CARNEGIE MUSEUM Acanthophacelus bifurcus, Archicheir minutus, and Steatogenys elegans. Of these A. minutus is known by but one specimen. The first three are new, and S. elegans is found in the Amazons. Mr. Shideler went up to the cataract at Malali but secured only twenty-three species, of which Myloplus asterias was the only one I Fig. 8. Fish-fence made of reeds and banana leaves on creek above Wismar. Indian with bow and arrow ready to shoot larger fishes brought up by poison. did not get elsewhere. A list of the fishes which do not extend so far inland and another list of those which do not extend so far down stream as Wismar are given later. On September 29th, at 5 :30 P.M., we took the train for Rockstone. Rockstone consists of an hotel, terminal station of the steamer connecting with the railroad, a small store, and a number of cottages for workmen. All goods for the gold-mines, and all rubber coming from the interior are transferred here to avoid the cataracts in the Essequibo below Rockstone. The Essequibo is here divided into two chan- nels by Gluck Island. At the time of our visit rocks were exposed at the stelling and a short distance below Rockstone. At the bend of the river above Rockstone there was an extensive sand-bar exposed. We collected at the Rockstone stelling, in the railroad-cut running into the river just below the town, in a small woodland creek on Gluck Island, and on the sand-bank at the bend above Rockstone. The EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 39 richest ground was found in the brook on Gluck Island and the slough between the sand-bank and the shore of the river. A haul made at night on the sand-bar was also very successful. A portion of a letter describing my first day at Rockstone follows : Fig. 9. View on bayou back of Christianburg. Indian fishing from his corial. September 30th, 1908. " Very early in the morning I engaged two Indians who were on a balata boat waiting for a crew to go up the Rupununi. They did not know English, neither did they know how to fish, and I got exasperated, till Shideler went in to show them. Hereafter I shall always dress to go into the water myself. We worked faithfully along the stelling and below, with the poorest success I ever had anywhere. We could see fishes galore, one especially (Chalceus ?nacrolepidotus) lustrous plum- beous, with the most gorgeous, maroon colored fins, flaunted its colors in my face, but it was im- possible to get at it. At ten we came to the hotel, I completely dead beat, for this was the place where we were told we should 'catch fishes.' "As we were waiting for breakfast a band of Indians came along, a man and about six girls and women. After parleying it turned out that they were going to poison a creek. We asked them to 40 MEMOIRS OF THE CARNEGIE MUSEUM wait at the station till we could join them. We rushed through breakfast, caught our two Indians, who were just ready to eat their breakfast, and made them take it along, and then went after the fishing Indians, who started as soon as they saw our boat come out of the creek. After skirting Gluck Island some time they stopped at a creek so small that I thought it could have no fishes. ..... HI'. MWpfeSH -/ - if '1 ' »■ \ ■ t 5 : #' f &%\9 1 . '** * V-:f ;:■■ f > \i^* '■Bg]l s ,4** " It" • \ - * . -"V 3 ■-'-■• 5 : . '■»*' H^ & *** S**-- ■j- - - | '^" — i ■P^- j" ft ; Fig. 10. Indian women pounding leaves in a hollow on the ground preparatory to using the pulp for poisoning a stream on Gluck Island. Two of the Indian women scraped a small depression into the ground, cut two sticks and used them as pestles and the depression as a mortar in which they pounded a basketful of leaves into a pulp. 12 They then built a fence across the creek with palm leaves, scraped the mud from their mortar into balls and squeezed them into the water some distance up the creek. The Indians and myself were soon knee-deep in water and mud, picking up the fishes which came to the surface. The little ones died in numbers on the banks, the bigger ones revived. I had a set-to with the Indian women because they did not want to sell me all the catch. We finally compromised, and I took all I wanted, giving them the larger ones. I supposed they wanted several dollars, but they asked only two shillings. I gave them three and we were both happy. After my transaction with the women was completed, the man gleefully held up a fine luckananee he had shot with an arrow. 12 1 regret to say that I did not get the name of this poison. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 41 " On the way home we stopped on some rocks, dikes that run across the river, and secured crabs, and another Loricaria. " I ate dinner with a somewhat better feeling, but determined to use our big hundred-and-fifty- foot net at the sand-bank after dinner. The porters were all gone when we got to the station, but I was able to pick up a couple of negroes, and took one of our Indians. Mr. Kingsland, the agent at Wismar, went along. The crew played the most interesting tunes with their paddles. When- ever he felt like it, the leader, by a peculiar stroke with the paddle, would get all of them to hit the boat during a definite part of the stroke. It produced surprising results and varied the monotony of the long row. After we had gone what I thought twice the distance, we discovered that by staying on the wrong side of the river we had overshot our mark and had to go back. Soon our boat got stuck in the mud at the upper end of the bank, but finally landed on a place that must have been made for us. It was a shallow bay on the upper end of the sand bank, a hundred and fifty feet across with a sandy bottom. We stretched the big net across and hauled out at the head of the bay. Fig. 11. Small stream on Gluck Island dammed by Indian women before putting in the poisonous leaf-pulp. Fish flopped in every direction, dozens went over the net, one of them went at one of the men and made him jump. At the critical moment the enthusiasm got the better of even our Indian, and he ran ashore with the top of the net and let half the catch out. As it was we had two buckets full of specimens. I gave Mr. Kingsland a luckananee weighing seven pounds, and the crew had enough to make the haul historic for all time to come. We got home at 12:30 A.M." 13 13 1 recognized seventy-one species on the day following. The number was probably nearer ninety, for in the final examination I found I had taken sixty species in two or three hours out of the small brook on Gluck Island. 42 MEMOIRS OF THE CARNEGIE MUSEUM The first day of October until evening was devoted to sorting and preserving fishes. Prochilodus gave us much trouble. Full strength alcohol and formaline injected did not keep these specimens from beginning to decay. In the evening we seined on the rocks of the stelling and in the railroad-cut mentioned above. At the stelling we caught so many Hemidoras carinatus and allies with erected spines that it took us a long time to untangle them from the net. Each pectoral spine of these catfishes is provided with retrorse hooks, the spines are erected when the fish feels himself caught, and each spine must be individually disentangled from the net. On the second of October we went to the Rockstone sand-bar with our two Indians. We were soon joined by seven porters who came from Rockstone to get sand and who helped us pull the large net at the lower end of the sand-bar. The most important captures we made consisted of specimens of Geophagus carrying young in their mouths. The outer edge of the bar was almost barren, but yielded a few minute, translucent specimens of Characidium, which so closely resemble our sand- burrowing darters that they amply repaid for the water-hauls. But the greatest Fig. 12. Edge of bayou between right bank of Essequibo and outlying sand-bar at Rockstone. success was obtained in a bayou between the upper half of the bar and the land. Here we collected a large number of small fishes. The Indians took half a bushel. There is a great general similarity between the contents of the net here and one drawn at any similar locality in the Mississippi valley, although not a single species EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 43 or genus occurring in the Mississippi valley was found in the catches. Here we secured the only specimen, greatly mutilated, of the widely distributed Sy?nbran- chus, and the young of many species of large fishes. Before starting for the sand- bank I had an opportunity of securing a lau-lau, but in the hurry of getting off, and on account of a momentary fit of penuriousness, I took only the head. On October third I returned to Wismar to make purchases for our trip to Tumatumari, and incidentally arranged to have a creek "stopped." I returned to Rockstone in the evening and sent Mr. Shideler to bring over the catch the next day at 2 P.M. Among other things he brought the rare Rhamphichthys. It is a long- snouted, sword-shaped, gymnotid eel. On October 4th I watched some of the natives dynamite about the Rockstone stelling and packed most of our catch for shipment. In the work of enumeration, which has taken two full years since my return, I find that Rockstone, where our fishing began so discouragingly, is the richest in species of all the localities examined. This was no doubt due to the fact that con- ditions for collecting were favorable. The water was low and we fished exhaustively in a variety of places. No doubt many channel fishes living here we did not get. Pseudoplaty stoma, Phractocephalus, and others, should be found there. Altogether we got one hundred and thirty-three species, eighty-three of which were characins. Of the one hundred and thirty-three species fourteen were not taken elsewhere: 1. Xenacara gymnorhynchus. 2. Hemiodontichthys acipenserinus 3. Aphiocharax melanotics. 4. Aphyocharax erythrurus. 5. Mcenkhausia megalops. 6. Hemigrammus iota. 7. Hyphessobrycon rosaceus 8. Hyphessobrycon riddlei. 9. Hyphessobrycon gracilis. 10. Phenacocharax hemigrammus. 11. Acestrorhynchus nasutus. 12. Rivulus lanceolatus. 13. Rivulus frenatus. 14. Crenicara punctulata. Some of these are known to have a very wide distribution. Number 2 for instance, extends to Paraguay; Nos. 6 and 9 are found in the Amazon; No. 8 in the Orinoco. Others are known only from the specimen collected. A critical examination of this list of uniques thus shows that these lists mean nothing except what is on the surface. On Monday, the fifth of October, we started for Tumatumari, but on account of various accidents to the boats we had to tie to the shore over night. Being overtaken on the following day by another boat, we went on with it to Tumatumari. which we reached at 7:00 P.M. I called on Mr. Edward Bovallius, the representa- tive of the Essequibo Exploration Company, and on Mr. Brummel, one of the government officers, and began seining on the following morning. 44 MEMOIRS OF THE CARNEGIE MUSEUM Tumatumari owes its existence to a cataract in the Potaro River. The goods brought up by the launch to the lower landing are transported by cart to the upper landing, and this transport gives employment to nearly all the inhabitants. At the time of our visit the stream was confined to the northern channel, the southern channel being entirely dry. We made headquarters in the Sproston's rest-house, from which point we had a view over the cataract. We collected on sand-bars Fig. 13. Looking across the Rapids of the Potaro River at Tumatumari. Papaya-trees in foreground. above and below the cataract, in the cataract itself, and in a little stream emptying from the north just below the cataract. Our experience in fishing may again be quoted from letters sent home. "We found a brook and went up it. I enjoyed the water, it was nice and cool. The water we have to drink is in an iron tank out in the broiling sun and no ice is to be had nearer than forty- seven miles. We fished upstream until we came to a deep pool. The nondescript helping us did not know "how," and stepped on a spiny palm branch besides, so I got into the water to take his place, and told him to take a big club and beat the water to drive the fishes down. He did this slowly. Shideler and I then took up the lead line of the net, for the banks were so steep and full of snags there was no place to haul the net ashore. We were walking down stream with the bag of the net in the water to a place where we could land, when Shideler said, "I believe we have an electric eel, for I have had two slight shocks." eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 45 " I envied him, but too soon, for just then I got a good shock from ankle to knee and I jumped and yelled, not so much from pain as from the unexpectedness of the shock. " We found we had five eels in the net, the largest three feet long, and it took maneuvering to get them into the buckets. I received several more slight shocks before it was accomplished. " We rested in the afternoon, so as to be ready to go out at night. It is scarcely possible to catch anything in the river in daytime. We were simply looking about to see where we could haul at night. In crossing the river at night I think our boat struck half the sunken rocks, until I in- sisted that the water be baled out before we risked another shock. On the sand-bank across from the lower landing we caught two more electric eels in a net well filled with fishes. It was surprising how soon everything became quiet in the flopping netful of fishes with such customers in their midst. I opened one of the eels and found small fishes in its stomach. I put a twig through the gill of the largest eel, for we proposed to eat it; the other eel we wanted to take home alive to see some sport, and I placed it in a live-net. I had the eel in one hand, and in order to pick up the net put the lantern in the same hand, but as soon as the lantern touched the eel I got a shock through the handle. It was not a heavy shock, but I did not know how much heavier it might become, and so gave up that way of managing. " When I came to pick up the net containing the other eel, I got another slight shock, and con- cluded I needed help to carry them. We ate part of the largest; the electric organ was pasty and the rest was so full of bones that we did not succeed with it. 14 "Thursday morning Shideler went to fish above the falls preparatory to fishing at night. I took care of the fishes, but by night we were both so tired we postponed the fishing till Friday night, when we caught a seven pound luckananee, which, profiting by previous experience, we skinned and ate. On one haul our net was again full of Hemidoras. These became so tangled up in the net that it took an hour or more to get them out. One of them was new (Leptodoras linnelli)." On the 10th of October T went to Potaro Landing to make arrangements to ascend to the Kaieteur. Mr. Shideler went up, the next day to watch some poison- ing, but returned in the afternoon with Mr. Linnell, who had come down from Holmia with a crew of Indians. Mr. Linnell, the representative of the Essequibo Exploring Company at Holmia, and Mr. Bovallius, the representative at Tumatumari, after consultation, lent me the crew of Indians which brought down Mr. Linnell, under the Indian captain William Grant. They also placed their bateaux between Kangaruma and 14 The "numb fish" early excited the interest of naturalists and thus directed attention to the fish- fauna of the Guianas. The first notice of the fishes of the Essequibo was an account of the doings of the electric eel, and in the second paper Bancroft attributed the shocks delivered by the electric eel to electricity. Humboldt described how his assistants drove some horses into the water to exhaust the eels. It became generally accepted that this was the usual method of fishing, although it is doubtful whether this method was ever tried except on the occasion when Humboldt did it. A comprehensive account of the electric eel was published by Sachs as a result of a trip to Venezuela for the special purpose of studying it. " Aus den Llanos," Leipzig, 1899, and " Untersuchungen am Zitteraal," Leipzig, 1881. 46 MEMOIRS OF THE CARNEGIE MUSEUM Holmia at our disposal. Our success on the expedition to the upper Potaro was largely due to the cooperation of Messrs. Linnell and Bovallius. We started on a launch from Tumatumari on the 14th of October at nine o'clock in the morning. We were met at Potaro Landing by part of our crew, who carried a portion of our outfit about seven miles to Kangaruma. The crew returned to Potaro Landing the following day for the remainder of the goods. The latter part of the trip between Potaro Landing and Kangaruma was through the forest, the first two miles through a hot sandy road. I felt a slight fever on arriving Fig. 14. Albert, one of the Indian bearers, transporting goods at Kangaruma. at Kangaruma. The portage from Potaro Landing to Kangaruma, which can be made in two and a half to three hours, obviates the engagement of a series of cataracts between the two points. We left Kangaruma in the afternoon by a peculiar bateau. A tarpaulin covered the center of the boat where our goods were stored; in front sat three pairs of paddlers; behind several more paddlers. William was captain and steersman as well, until we picked up a corial at Amatuk and EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 47 "Trenchan" became steersman. Eddie, my cook, spent his time largely with the dog "Sunday." Part of our Indians were jolly, naked savages from near Holmia; the others were surly fellows dressed in shirt and trousers and had come from Brazil. After about two hours rowing we camped beside a creek, Erukin. I attempted Fig. 15. View looking up the Potaro River in the early morning. Glimpse of the Guiana Plateau in the distance. a little fishing at night, but with the very poorest success. On our return trip we were more successful at this point. On Friday, the 16th, we started at 6 A.M., and by eight were at Amatuk, where we remained till breakfast. We attempted to do some collecting on the sand- bank, but, as usual in the daytime, with very little success. At Amatuk the goods had to be carried to a boat above the twenty-five foot cataract. I had felt fine all the morning and enjoyed especially the lazily flopping Morphos crossing the river, and the flying fishes. The latter would dart up in front of us, cut through the water, leaving the breast or tail in the water and beating the water with their pectorals. One of the flying-fishes would cut through the water for forty feet or more and then leave the water entirely for five or ten feet. At the end of its flight it would fall sidewise into the water. At first it looked like a long, slender fish, but by watching near the end of the flight, when the momentum was gone, the disc-like shape could easily be made out. To make sure I asked the Indians 48 MEMOIRS OF THE CARNEGIE MUSEUM to point out the fish when we seined. William pointed out a long, slender Creato- chanes, but the naked hunter shook his head, and with thumb and forefinger made a circle. We caught none in the Potaro, where we saw them frequently, but got them in abundance in the lower Demerara. Gasteropelecus is more apt in its flight than Carnegiella. Whole schools will sometimes leave the water and shoot over the surface. After breakfast we rowed up through the gorge, which the Potaro has cut through the table-land beginning at Amatuk. The edges of the gorge have been . yte a . ""* irfflSaf*-^™ i\ -rotifer rfitft *" a* SHS^^BB :^.W>j V . ^ 1 bS 1 'Ai^>^ Hff ■( jk ^S3s ~^HI W^~ " "*" W^ ' ^ Fig. 16. Seining on a sand-bar below the Amatuk Cataract. carved in a variety of ways which give them the appearance of high mountains. The valley is quite broad, indicating the great age of 1 he gorge, which is a thousand feet deep. We took with us a corial from Amatuk, planning to send Mr. Shideler back in it from the Kaieteur. The Indians shot a baboon on the way up to Waratuk, the next portage. We camped rather early in the day above Waratuk and during the night I had a particularly nasty case of chills and fever. By midnight I could scarcely stand, but enjoyed the great variety of new noises from the water and forest. We started at 7:00 A.M. on the 17th and stopped at 8:00 to get our first glimpse of the pride of Guiana, the Kaieteur. The Kaieteur was hidden in mist EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 49 so early in the morning. We camped shortly afterwards at Tukeit, called "Tukui," or humming-bird, by the Indians, after the waterfall coming from the plateau opposite the camp. Our hunters killed four peccaries across the river, and young wild pork was a pleasant change from the canned meats. At Tukeit there is another cataract in the Potaro, and above it several more towards the foot of the fall. It is one of my regrets that time did not permit me to walk up to the Kaieteur. We collected in the Potaro at Tukeit and the following morning at eight started to ascend the plateau. The path leads back from the river for a time, crosses Shrimp Creek, and then ascends very steeply to the top of the plateau. Here it is comparatively level again and runs through the woods to the edge of the sav. nnah, or treeless tract, immediately about the fall. Fig. 17. View on the Potaro River looking up stream at the point where the first glimpse of the Kaieteur Falls is caught. After breakfast, at the edge of the savannah, Mr. Shideler and I, with William and another Indian, walked to the edge of the precipice and to the fall, while the rest of the Indians went on to pitch the camp and get the boats of the upper Potaro ready. The scenery about the fall is unique. Looking down the stream one sees 50 MEMOIRS OP THE CARNEGIE MUSEUM the U-shaped gorge cut out by the Potaro in the level plateau. The Potaro is visible from time to time as it crosses from one side of the valley to the other. The best view of the fall itself can be had by climbing down on a ledge of rock at the edge of the precipice. I not only climbed down, but, all excitement with the fever, Fig. 18. Exposed left side of the bed of the Potaro River at Tukeit in the dry season. the steep climb, and the superb view, set up my camera on the ledge and took numerous photographs. I confess to feeling distinctly dizzy when I placed my head under the focusing cloth, knowing that if something should happen I and the camera would land on the rocks a thousand or so feet below. Not that I could find a finer place to die, but I was reluctant to start to " kingdom come " on such a heavy down grade. After making about thirty exposures under varying conditions, we went to the camp in the bush some distance up the river. The fall is caused by an exces- sively hard conglomerate which overlies a softer sandstone. The savannah above the fall is in large part this naked conglomerate. In places bushes grow from cracks in the rocks. Many bunches of grass, or flowers grow from a little accumula- tion of soil that can be kicked from place to place along the surface of the rock. The afternoon of the 18th and part of the 19th were devoted to fishing and packing. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 51 Mr. Shideler started down for Georgetown at 11:00 A.M. of the 19th, with three of the Indians. He collected about Potaro Landing, in the Botanic Garden, at Bartica and near Morawhanna, Issorora, etc., of the northwest coast. The crew who remained with me was divided between two boats. One boat, the "balahoo," was dispatched at once with the goods, while a few of the Indians, with William and myself, stayed till the following morning. We started after the other boat on the 20th, at 6 A.M. It rained hard. All hands pulled with all their might on this, the home stretch. We reached the regular camping site at 3:00 P.M., but all the Indians were anxious to go to Holmia, which they said they could make in three hours. I was willing. By a supreme effort the three hours were cut down to two hours and twenty minutes, and at 5:20 P.M. we reached Holmia. Holmia is the trading camp of the Essequibo Exploration Company. The *^m •fjitd^A Fig. 19. Looking up the Potaro Valley from the brink of the Kaieteur Falls. company has a store and depot surrounded by a few Indian huts. It is situated on the Potaro at the entrance of the Chenapowu river. My crew of Indians went out at once to collect poison, the root of a plant called "hiari" (?Lonchocarpus) under the guidance of the local Indian, Jordan. The Indians of the surrounding regions brought me fishes and we ourselves poisoned a small creek just below the 52 MEMOIRS OF THE CARNEGIE MUSEUM houses. Unfortunately, it rained heavily, so that the Chenapowu and the Potaro rose many feet and made fishing in them not profitable for some time. We went up the Potaro a distance further and poisoned two creeks just below the Aruataima cataract. In the cataract itself we could do nothing on account of the high water. William later collected in the cataract and sent me two new genera and three new Fig. 20. Looking down the Canyon of the Potaro River from the brink of the Kaieteur Falls. species, from which it would seem that further collecting would prove profitable at this point. The character of the fauna of the plateau is discussed in detail in another place. It seemed that each creek we examined contained some one dominant form and a few stragglers. The dominant forms varied in different creeks. I started from Holmia on October 27th, fully intending to return with Mr. Linnell, but I found at Tumatumari that he had gone to England, and I did not return. That I could go no further on the Potaro, could not cross over to the Ireng, is a lingering regret. We reached Savannah Landing at 12:00 and walked over to the Kaieteur to take a few more photographs. We also poisoned another creek and collected in a swamp above the landing. We started down for Tukeit at 1 :30 on the 29th of October. In crossing Shrimp Creek, which seemed quite impossible as a fish habitat, I caught a Rivuhis with EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 53 my hand. I was surprised by the catch and more so by the sequel. The fish jumped out of my hand and stuck to a vertical rock by its tail and jumped from this to another point higher up on the same rock. It was lost on the way down to camp, so I sent William back the next day to poison the creek. He secured two species of Rivulus {waimacui and breviceps) and a Characidium (vintoni), all of them new. We collected about Tukeit in the afternoon, poisoning the creek just below Fig. 21. View looking up the dry bed of Shrimp Creek. Figure of man in middle distance shows the height of the rocky steps. Follow indices on cut. the landing with considerable success. The striking thing about the fauna at Tukeit is its evident contamination by the fauna of the plateau. Helogenes and Poecilocharax, fishes of the plateau, are found at Tukeit below the Kaieteur, but were not taken lower down. On the 30th we started down the river. While the Indians were transporting the goods at Waratuk I experimented a while with poison in a little side branch of the cataract. I was so successful that I proposed to make a more systematic attempt at Amatuk, where we stayed for the night. We poisoned above the cataract on an island, but without great success. Below the north branch of the cataract we tried once more on the morning of the 31st. Some of the water loses 54 MEMOIRS OF THE CARNEGIE MUSEUM itself here under rocks, and we placed the poison above the point where it runs into the rocks. The branch was entirely too large to enable us to kill the fishes, but the poison drove them out from under the rocks so dazed that we could pick them up with our dip-nets. In all we took forty- two species, of which five were not found elsewhere : 1. Brachyglanis frenata, 2. Brachyglanis phalacra, 3. Hemicetopsis minutus, 4. Pseud- ancistrus nigrescens, 5. Characidium laterale. All of these were new. Here I also secured a series of related little catfishes that were either confined to, or most abundant in, the cataracts of the lower Potaro. These are Brachyglanis frenata and phalacra, mentioned above; Myoglanis potaroensis and Chasmocranus longior and brevior. Other Pimelodinw, Doradinw, and Auchenipterinoe were ab- sent. Lithoxus, a little loricariid catfish, was very abundant. It is flat and clings to the rock, which is greatly resembles. They were especially abundant in rather deep pools and could only be seen when the poison brought them fluttering to the surface. Loricaria was not observed. Deuterodon pinnatus and potaroensis were also abundant in the cataract, and Po?cilurichthys abramoides and Mquidens potaroen- sis occurred in a shallow pool near the margin. Tetragonopterine characins were scarce, but one representative of the Ser- rasalminai was taken. Of the Characina? there were none. The peculiar Sternar- chorhynchus oxyrhynchus, Porotergus gymnotus, and Sternarchus leptorhynchxis were seen here for the first time. They were among the rocks and were driven out by the hiari. The following letter preserves the impressions of this region in their freshness : Kangaruma, October 30, 1908. " I think I have written everybody and answered all questions, which need answering, so I may continue my record. "I awoke before five, and as some of the Indians were stirring I began to sing "balahoo," re- peating the word to an old college tune. It has been the reveille and march for the crew ever since we first reached the Kaieteur. I learned several tricks yesterday in poisoning Waratuk that I proposed to put in execution today at Amatuk. We first tried poisoning a little branch above the fall at Amatuk, and got some things. Then we tried a more ambitious scheme of poisoning a big branch below the fall. I found that the poison will drive some fishes out before it kills. We had three men pound hiari and wash it into the branch of the northern part of the fall. William and another Indian stood a long distance below after the water had flowed among and under rocks. I at first stayed near a pool where they were poisoning. Soon Plecostominw began to come up. They were new to me and I dipped with enthusiasm until I fell in. This broke the ice for me, for I then waded from rock to rock, securing eighty-seven specimens of the new genus Lithoxus. William came with a dip-net full of fishes, among them long curved-snouted Gymnotidce I had not seen. "We poisoned and waded, gathering in all half a bucketful of small fishes, all valuable as speci- mens. It was rare sport, and I did not realize that it was ten o'clock and that I was played out eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 55 We had soup, rice, tea, and jam for breakfast, and started at 1 1 :30 on the home-stretch. Dead fishes were floating about the little bay from which we started, so the poison kills. "Most of the Indians had done nothing but swing in their hammocks all the morning, so they paddled with a swish and swing in great contrast to yesterday's dilly-dallying. At one P.M. we were at Erukin, a sandy, clear creek that I wanted poisoned. We had fished here with poor success at night on the way up. William thought it was too big and swift to poison, but I had them stretch the net across the mouth, sent the pounders up-stream, and William, some other Indians, and myself took up stations at intervals obliquely across the river. I stood in a patch of sunlight where every grain of sand could be seen at the bottom. I was in sleeveless undershirt, pants, and a pair of socks. My tennis-slippers had gone to pieces on the rocks at Amatuk. Soon fishes came down the stream in distress, and when the poison was exhausted, we found we had a number of novelties in our dip- nets and the fifteen-foot net at the mouth of the creek had caught most things as they came down. At 2:30 we were moving again, and when near 4:00 o'clock William cried out "Kangaruma," all paddles stopped for a moment and then dashed on, and we landed here at 4 P.M. 16 I read let- ters, not all of them thoroughly, till supper time. I must wrap and pack my fishes before we start on our two-and-a-half hours' walk to-morrow, for I don't want to lose the day's catch during the walk. Saturday morning. " For about a month now I have not slept out of hearing of the roar of the cataracts. Tumatu- mari, Kangaruma, Amatuk, Waratuk, Tukeit, Kaieteur, and Holmia are all on cataracts or falls. In fact they are all places where goods have to be transported on account of cataracts. I have so much stuff that it took more than one trip for the twelve carrying Indians that have come down with me. But then their food and personal effects add a little. My personal effects make about one load now. The trip has been a phenomenal one. I can't say that I could swear that I have every- thing, but every effort, seconded in each case by William and sometimes suggested by him, has been made. I ought to have poisoned the cataract above Holmia, but it rained a night and a day so the river rose five feet, and it could not have been done even' if I had at that time learned the trick." We left Kangaruma in the morning of November 1st for the walk through the woods, and arrived at Potaro Landing in time to catch the boat for Tumatumari. Here I parted with my crew, who returned to Holmia with Mr. Bovallius. I packed fishes and had the fever again. I broke it, and left on the 4th for Crab Falls. The Potaro Gorge is one of the remarkable features of the world. The river is lined with trees so tall they could only thrive in a region free from strong winds. The sides of the gorge are rugged, and the whole recalls the Rhine, or the Yosemite. We saw no one from the time of striking into the woods behind Potaro Landing till we reached Holmia. The entire stretch is utterly uninhabited and very few tourists pass this way to get the inspiring view of fall and gorge from its upper brink. The region between Savannah Landing and Holmia is level, but mountains 15 Mr. Macturk had brought a batch of letters from Georgetown and left them here. 56 MEMOIRS OF THE CARNEGIE MUSEUM can be seen in the distance, toward the south, when the view is not obstructed by trees. The fauna of the Essequibo becomes attenuated as we go up the Potaro. Its upper course below the fall undoubtedly contains intrusives from the plateau which decrease in number as one gets further and further from the Kaieteur. At Tumatumari I boarded the launch, taking with me a negro, Mr. Cum- mings, with a bateau. We landed at the head of an island a short distance below the mouth of the Potaro and just above Crab Falls. There was an Indian settle- ment here. We slung our hammocks under the shelter of one of their huts. My WIDTH 552ft 600 FT WIDTH 369 FT 1320 FT ^■822 FT Fig. 22. Diagrammatic representation of the contour of the Kaieteur Falls. (After Brown.) own hammock, that of Mr. Cummings, and that of an Indian woman radiated from the same center pole, at the base of which a monkey was tied, — a cosmopolitan quartet. Cummings and the Indians went out with the net at night to fish on the sandbanks. I remained in my hammock to recuperate from the fever. On the fifth I sent several of the Indians out to dig hiari roots while I fished about the rocks of Crab Falls. The Essequibo is very wide at this point, divided by an island, and falls over a dike running squarely across the river just after it has made a turn. On the sixth, Cummings, myself, and four Indians went with the bateau up the Essequibo to shoot pacu at the Warraputa Cataract. Above the mouth of the Potaro the Essequibo is broken by a large number of rocky islets, fragments of a dike crossing the river. Other dikes cross the river further up, the water rushing through the gaps. Through some of the gaps the Indians succeeded in paddling the bateau, through others they dragged the boat after being driven back several times by the current. A dike extends across the Essequibo near the mouth of the Konawaruk. Opposite the Konawaruk and below the dike there is a lagoon separated from the river by a sandy and partly wooded spit of land, but connected below with the eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 57 river. Near the head of the bayou and connected with it by a few inches of water there is a pool toward the river side. It is not more than fifty feet in diameter and perhaps six feet deep in its deepest part. Trees overhang it from the river side. The Indians pounded hiari roots into shreds. They were tied into bundles and two Indians boys swam through the pool with them to mix in the poison. I have described the effect in the introduction. First one species and then another came to the surface, and then they came indiscriminately. A stingaree came fluttering to the surface toward the last, while the little Corydoras punctatus with- stood the poison to the end. Catoprion mento, a Mylinid characin with a projecting chin, was particularly acceptable. The fish that created the greatest interest was Mcenkhausia dichrourus, of which I had seen dozens of specimens, hailing from dif- ferent localities all the way from Paraguay to Para, and all of them conveyed about as much idea of the appearance of the living fish as a dead and plucked Baltimore Oriole would give of the living bird. The base of the tail is bright canary-yellow, the lobes are crossed with jet-black bands and the tips are milk-white. I do not know how long we stayed here, not over two hours, in which we caught fifty-five different species of fishes, six of which were not secured elsewhere. The uniques were Bunocephalus amaurus, Ochmacanthus flabelliferus, Odontostilbe melanditus, Aphyodite grammica, Hyphessobrycon minor, Dormitator gymnocephalus, all of them new. Several other new or rare things were taken in only one other locality: Porotergus gymnotus, Catoprion mento, Hemiodus semitceniatus. We continued our row to the Warraputa Cataract. The river is divided here into several branches by wooded islands. The two older Indians started out to shoot pacu, but bagged nothing. The rest of the crew and myself set to work to poison a branch of the cataract, where we were again quite successful, securing a series of specimens that recalled the Amatuk cataract. We also got a series of the young of the pacu, the first which were obtained or recognized as such. The color changes in the young can be seen in plate LIX. After the poisoning Cum- mings went to the sand-bank to prepare supper, while I browsed about the cataract with the two young Indians. On my return at dusk I found that no provision had been made for my ham- mock. The Indians declined to go into the bush at night to secure palm- branches and I did not like to risk a drenching rain so soon after having had the fever. I insisted that the Indians either build me a shelter or take me back to their camp. Giving them an alternative was a mistake. They de- cided to take me back. It did not rain that night. It was all very well to come up through the gaps in the dikes in the day time. It was quite another matter 58 MEMOIRS OF THE CARNEGIE MUSEUM to safely guide the bateau down through them at night. We shot through at a tremendous rate and once the boat touched something. The boys shouted with glee, while I came to a sober realization that it would have been better to trust the sky than the rapids. But we got safely back to the huts early in the evening. On the seventh of November I collected about the rocks just above the fall. Here I succeeded again with hiari in getting fishes which could have been secured in no other way. At one point the bank is piled with huge blocks of stone. To dislodge the fishes from between them would have been impossible in any ordinary way. We pounded some hiari roots and washed them in the swift current that was flowing towards the rocks. At once some species came to the surface, straight up, without attempting to escape. Several species were dislodged including an electric eel. We repeatedly got it into our dip-net, and it as often got out again, without, however, making any coordinated movement to escape from the reach of the net. It proved too slippery, however, to hold in the net, and it got away. When the launch came by from Tumatumari we loaded my effects into it and I left the region of the Potaro and upper Essequibo. I landed at Rockstone in the afternoon and took the train for Wismar the next morning. After packing fishes all clay I went to Christianburg as the guest of Messrs. Spence and Brummel. On the ninth I rested at Christianburg. The tenth Mr. Brummel took me to Kumaka, where I made arrangements to have a creek poisoned, and on the morning of the eleventh I took the steamer for Georgetown. Visiting the markets, preserving and packing fishes, suffering a relapse, re- covering from the fever, and enjoying the hospitality of friends at Georgetown consumed the time until sailing for New York. Everything collected arrived at home safely. Method of Preserving. The method of preserving the fishes, not one of the twenty-five thousand of which was lost through decay in this tropical region, was as follows: In the Essequibo and the lowlands all fishes except large catfishes were dropped alive into a can of 25-35 per cent, alcohol in which they were killed. Minute fishes were placed in vials or in small bottles instead of in the tank. Shortly after reaching the camp or home all the fishes above two inches were injected with 95 per cent, alcohol, freed from the coagulated slime, and placed in a can with about 50 per cent, alcohol. The first alcohol, if less than 25 per cent., was thrown away; if more than 25 per cent, it was allowed to stand and was decanted to be used again. A day or two following the catch the specimens were transferred into 75 per cent, alcohol and the larger ones again injected. The following day they were roughly eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 59 sorted, wrapped in cheesecloth and dropped into 85-95 per cent, alcohol. As soon as sufficient material was at hand to fill a kerosene can or two, they were filled with fishes packed tight as sardines and the can soldered. A small hole was then punched and the can filled to the top with 95 per cent, alcohol, and again soldered. Two cans were usually packed in an original kerosene-can box and shipped at once to New York. Some of the cans were punctured and the alcohol leaked out on the way home. The fact that these also brought the fishes through in perfect shape showed that ordinarily greater precautions were taken than necessary. Because some one who knew nothing of the Indians told me they would steal the alcohol, I took only formalin beyond Tumatumari. I regretted it much, because the precaution was not at all necessary and it would have been better to have had all the scaled fishes preserved in the same manner. In wrapping the fishes for packing the usual precaution was taken. A locality label was added to each little bundle and the fishes were so wrapped that a layer of thin cheese-cloth was between every two. Care was always taken not to tie the bundles too tightly, and minute delicate things were placed in a small vial after being wrapped Smaller specimens after being wrapped were packed in empty one-pound coffee-tins to keep them from being crushed in the larger kerosene cans. Naked catfishes in general were preserved in formalin and transferred to alcohol at home. They do not have metallic tints which are dissolved by the for- malin, and retain their plumpness better than in alcohol. But the difficulty in opening the jaws and the danger of damaging the fins probably fully dis- counts for all the gain. Color notes upon the characins were preserved. It was not possible to do more at the rate I was collecting. The difficulty with Prochilodus was mentioned above. CHAPTER IV. THE GEOGRAPHICAL DISTRIBUTION OF THE FRESHWATER FISHES OF BRITISH GUIANA. Every effort was made to secure a complete series of the fishes of the area covered, and reasonable success may be claimed, the ichthyic fauna at large being well represented by the eighteen thousand three hundred odd specimens collected. Inasmuch as several localities were examined in each of the following districts, (1) the lowlands, (2) the middle Essequibo, (3) the lower Potaro, (4) the upper Potaro, we may also conclude that the faunas of these stretches of territory are well illustrated by the material collected. Furthermore, the work on such units of environment as the trench in the Botanic Garden, the woodland creek and sand- bank at Rockstone, the Konawaruk pool, and the Amatuk Cataract was entirely satisfactory. On the other hand, it may be doubted whether the lists of specific localities are in many other cases exhaustively complete. I have called attention to this matter in connection with the species taken at Rockstone and nowhere else. It was found that many of these are known to have a wide distribution and must occur in other localities in the area under consideration, although they were not taken elsewhere. Furthermore, it must be regretted that time did not permit me to make a more exhaustive study of the fauna in the Potaro between Amatuk and Tukeit and" above the Kaieteur. General Plans for Distributional Work. A question of greater interest to the next expedition than to the general account of the present trip is the length of time that should, and could to advantage, be given to any one place, and how far apart stations should be selected. To graphically demonstrate conclusions I have prepared two tables. - On p. 84 is a summary showing the total number of species taken at each place, and on p. 85 a summary of the number of species taken at but one locality. The latter shows how much would have been lost from the total number of species, had any particular locality been skipped. Only Kangaruma yielded nothing new, and fishing here was but a minor incident. There is very great inequality. The omis- sion of Konawaruk, where two hours were spent, would have entailed nearly as 60 EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 61 great a loss as the omission of Tumatumari, where a week was devoted to fishing. Amatuk also, where fishing was attempted twice in passing up and down the Potaro, yielded a large number of unique species. At Konawaruk we stopped for only about two hours on our way up the Essequibo. Our success here is chargeable altogether to the fact that everything was gathered that was contained in a small sand- and mud-lined pool. A week's stay would not have yielded more. No doubt the list for this locality would have been vastly increased if we had as thoroughly cleaned out one of the rapids, a sand- beach, a fallen tree-top, a woodland brook, a rock-lined pool, the channel, etc. As far as possible one of each of such different units should be exhaustively fished at each locality if the limits of the distribution of species is sought. Superficial fishing yields only the ordinary, the cream is obtained by stripping. The length of time that should be devoted to any one locality depends, therefore, on the variety of units to be found at one place. Rockstone shows a larger number of species in the preceding lists because a greater variety of units were examined. Further collecting at this place would have been profitable. Again, to determine the limits of the distribution of species, localities should be no more than twenty-five miles apart in a stream like the Essequibo or lower Potaro. The distance between stations should of course be much less where the stream is descending rapidly. It may be greater if the largest number of species in the shortest time is the object. Rockstone and Crab Falls are about forty-five miles apart in a straight line. Sixteen species taken at Crab Falls were not taken at Rockstone or below Rockstone. Their downward limit was therefore not determined. About twenty species taken at "Rockstone do not extend as far as Crab Falls, and their upward limit was therefore not determined. It is doubtful whether anything else of consequence was missed, because the stations were so far apart. Over twenty-five species taken at Rockstone were not seen at Crab Falls, but must occur there, for they were taken farther up the stream, a fact which again demonstrates the necessity of exhaustive work. The general recommendation for future work is, then, that representative localities be selected twenty-five to fifty miles apart,. and that collecting be done exhaustively in each sort of environment to be found at each locality. A week at the least in each locality should be the minimum time-allowance. On a large river like the Amazon a month or more would be required in a favorable season to secure a representative collection from any one point (In an unfavorable season it would not be worth while to attempt fishing.) Such collecting could advan- tageously be supplemented by exhaustive collecting in isolated favorable units. Here the time requirements would be much less. 62 memoirs of the carnegie museum The Ichthyological Position of British Guiana. In the Reports of the Princeton University Expedition to Patagonia, Vol. Ill, I discuss, among other things, the distribution of the fishes in the Americas south of the tropic of Cancer. Five distinct regions of unequal value are recognized: (1) Transition, (2) Mexican, (3) Brazilian, (4) Andean, and (5) Patagonian. In the Brazilian region the following "provinces" were enumerated: (1) Central America, (2) Pacific, (3) Magdalena, (4) Amazon, (5) Guiana, (6) Trinidad, (7) 16 San Francisco, (8) Coastal, (9) La Plata, the latter divided into the Paraguay and Parana-La Plata. Concerning the Guiana Province, including French, Dutch, British, and Vene- zuelan Guiana, the following is said (p. 319) : "The Guiana Province, including one of the oldest land-masses, is drained by the Cachipur, Oyapoc, Cayenne, Mana, Maroni, Surinam, Corentine, Essequibo (Mazaruni, Cuyuni, Rupununi and other tributaries of the Essequibo), Demerara, Berbice, by the Orinoco and its eastern and southern tributaries, the Caroni, Caura, Ventuari, and by the Rio Branco, and the northern tributaries of the Amazon east of the Branco. "The feature distinguishing this region is the known or reported connection between neighboring streams. The Cassiquiari connects the Orinoco with the Rio Negro. The Atabapo is said during the rainy season to be connected with the Guaina, and the Rupununi of the Essequibo basin with the Tacutu of the Rio Branco basin. It is said that the Essequibo is also connected with the Rio Trom- betas through the Apini and the Oyapoc, Cachipur, and Araguary with the tribu- taries of the Yari, which empties into the Amazon. "The lower course of the Essequibo of British Guiana at least is connected with the lower Orinoco by natural canals, so that these form part of the Orinoco- Amazon-La Plata meshwork and contain the same types. The Eastern Guianas (French) have a less varied fauna. "The fauna of this region is of the greatest importance to theoretical chorology, since this is one of the two old land-masses, and since it was by a continuation of this area that South America is supposed to have been connected with Africa. "Our knowledge of the fish fauna of this region is derived from Muller & Troschel's account of the fishes of British Guiana; Bleeker's Silure de Surinam; Vaillant's notes on the fishes of French Guiana, and his account of the Berbice and the general work of Cuvier & Valenciennes, Gunther, Eigenmann & Eigenmann, Regan and Pellegrin. 16 "East Brazilian Plateau," comprising 7 and S, was given a separate heading in the paper quoted. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 63 "Unfortunately little is known of the fauna of the streams of the table land of Guiana, the region where the remnant of the original fauna -may be expected to persist. Schomburgk's collections made in those streams were largely lost. He states that it was rich in species (over 30). "Andre ('A Naturalist in the Guianas,' p. 205) says: 'In fact the falls of Para [of the Caura] appear to constitute an effective barrier between distinct forms of river life', and that the fauna above the falls is different from that below. "The most promising field for scientific results, if not number of species, seems to me to be offered by the rivers of this region, which should be explored above and below falls that are impassable barriers for the ascent of fishes. "The rivers of this region, exclusive of the northern tributaries of the Amazon, concerning which not much is known, contain a total of about 298 recorded species. Of these about 60 per cent, are also found in the Amazon; as of these about 16 species are from the Rio Branco basin and not from the other streams and the Rio Branco belongs to the Amazon system, this number must be reduced by 16, which still leaves over 50 per cent, of the fauna identical with that of the Amazon." The following table is derived from the one succeeding the above quotation. It is condensed laterally by including the tributaries of the Branco and the rivers of French and Dutch Guiana in one column each, and is expanded vertically by add- ing the species discovered by me. It will especially show the relation of the fauna of the Essequibo to that of adjoining regions. Columns 1, 2, 5, 8, 9 and 10 offer an opportunity for a com- parison of the lowland faunas from west to east, i. e., of the Orinoco, Barima, lower Essequibo, Demerara-Mahaica, Surinam, and French Guiana. Columns 3, 4, 5, 6 and 7 offer an opportunity for a comparison of the fauna of the Lower Essequibo (5) from its mouth to the Warraputa Cataracts with that of the Upper Essequibo and its tributary the Rupununi (4) and with that of the lower Potaro (6) and upper Potaro (7), and also the Rio Branco (3) with its tributaries, the Ireng, et al. The details of the distribution of the British Guiana species are taken up later. Those species found also in the Amazon are preceded by an A. Those species preceded by an L are also found in the La Plata system. Those peculiar to the Guianas are marked with an *, and the genera peculiar to the region are marked with a double **. The estuarine species are preceded by an M. In the following table column 1 represents the Orinoco basin, 2 the streams emptying between the Orinoco and the Essequibo, 3 the Rio Branco basin, Takutu, Ireng et al., 4 the upper Essequibo with the Rupununi, 5 the Essequibo below Warraputa, 6 the lower Potaro, 7 the upper Potaro, 8 the Demerara, 9 Dutch Guiana, especially Surinam, and 10 French Guiana. 64 MEMOIRS OF THE CARNEGIE MUSEUM "to 03 PQ o o o a "C o 1 a o| o ja o.s m »- 2 a "5a a « o o a a u n o B 3 a 3 a 3 Qc 3 A 4 6 '3 a a 0Q o h a> a 0. P 7 08 08 eo a p 8 08 a 08 '3 O J3 o -*j 3 Q 9 C3 a 03 '3 O V a CJ 10 X X X X 2. A L Potamotrygon hystrix (Midler and Henle) X X X X X X X X X X 9. * Chamaigenes filamentosus (Cuv. and Valenciennes) X X X X X x X X 12 * Aspredo sicuephorus (Cuvier and Valenciennes). . X X X X X X 18. M Felichthys bagre (Linnaeus) X X X X X X X X 19. M Felichthys marinus (Mitchell) X 20 M Sciadeichthys flavescens (Cuv. and Valenciennes). X 22. M Sciadeichthys emphysetus (Miiller and Troschel). X X X X X 23. M Sciadeichthys proops (Cuvier and Valenciennes) . . 24. M Sciadeichthys luniscutis (Cuvier and Valenciennes) 25. M Sciadeichthys albicans (Cuvier and Valenciennes) X X X X X X X X X Y Y 28. M Selenaspis passany (Cuvier and Valenciennes) . . . V 29. M Notarius grandicassis (Cuvier and Valenciennes) 30. M Notarius parmocassis (Cuvier and Valenciennes) X X x X Y X 35. M Arius spizi Agassiz ... X X X Y 36. M Arius fissus (Cuvier and Valenciennes). . Y Y X x X X X Y X 44. A Pseudop'unelodus raninus (Cuv. and Valenciennes) X X X X X X X X X X X X X Y X EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 65 ,0 a m o CJ o a 'C o 1 03 a o-g SO §^= o-~ is 2 a '55 m o u a 03 ffl O s 3 a 3 a 3 a 3 4 6 "3 03 m CO H a & o 5 6 t- 03 o ft u o & o ►J 6 6 u o Ph u 'C 2 a 'Si a O o C 03 u m o £ 3 c s a 3 a 3 K 4 6 'B C 0J m 01 w is o ►J 5 6 -^> o Ph u OJ is o ►J 6 6 u 03 -^> o u a a a t> 7 C3 u 03 u a a) Q 8 oi a a '3 a ■a u 3 p 9 o! a 03 '3 a a u fi 10 273. * H emigrammus microplerus Meek. . . X 274. * Hyphessobrycon minor Durbin X V 275. * Hyphessobrycon rosaceus Durbin. . . 276. * H i) phessobrycon riddlei Meek X X 277. A Hyphessobrycon gracilis Reinhardt. . 278. * Hyphessobrycon minimus Durbin. . . X 279. * Hyphessobrycon eos Durbin. . . X X 2S1. A Hyphessobrycon belotti (Steindaehner) 282. ** Dermatocheir catablepta Durbin . . . X X V 283. A Creatochanes affinis Glint her X X X X X X X X V 2S4. * Creatochanes melanurus (Bloch) 285. * Creatochanes cawlomaculatus Giinther. . . X X v 2S6. * Creagrutus melanzonus Eigenmann. . . X X X X X X X 287. * Bryconarnericus hyphcsson Eigenmann. . 288. * Astyanax guianensis Eigenmann X X 289. * Astyanax essequibensis Eigenmann. 290. * Astyanax mutator Eigenmann. . . 291. * Astyanax mucronatus Eigenmann . . 292. * Astyanax wappi (Cuvier and Valenciennes) . . ? X X 293. * Pwcilurichthys polylepis (Giinther). X X ? X 294. * Poecilurichthys abramoides Eigenmann . . X X X X X X X X V 296. A L Poccilurichthys abramis (Jenyns). . . 297. * Pcecilurichthys potaroensis Eigenmann. . . X 29S. * Clenobrycon spilurus (Cuvier and Valenciennes). 299. A Ctenobrycon hauxwellianus (Cope) X X X X X X x 301. * Deuterodon pinnaius Eigenmann. . . X X X X 302. * Phenacogister megatostictus Eigenmann. 304. A Holobrycon pesu (Miiller and Troschel) x X 305. * Brycon lonqiceps Steindaehner 306. * Brycon siebenthali Eigenmann. . . X X X 307. * Brycon jalcatus Miiller and Troschel.. . X X 308. A * Brycon lucidus (Kner).. . X 309. A Chalceus macrolepidotus Cuvier X X X X 310. A Fowlerina orbicularis (Cuvier and Valenciennes). 311. * Poptella longipinnis (Popta).. . . X X 312. A L Chalcinus angulatus Spix 313. A Chalcinus rolundatus (Schomburgk). X X X 314. A Chalcinus clonqatus Giinther X 315. A Carneqiella strigata (Giinther) . X X X X X X X 316. A Gasteropelecus sternicla Linnaeus. . . V X x 319. F L Serrasalmo marqinalus Valenciennes 320. A L Serrasalmo svilopleura Kner.. . X X X X X X X X X 9 X X X X 321. A * Serrasalmo rhombeus (Linnteus) X X X X 323. A * Pygocentrus niger (Schomburgk) X 1 X X X ? 327. * Pygocentrus bilineatus Eigenmann X 70 MEMOIRS OF THE CARNEGIE MUSEUM a '55 03 m o o _a "C O 1 00 a o-g ■so - — 2 a "53 m o u a 03 u pq o « 3 '8 a 3 a 4 6 '3 c CO m 11 H u V is o ►J 5 6 s o Ph u o ►J 6 6 u 03 o Ph 03 a a. t> 7 d u o3 u s a 8 03 c 2 '3 o o -+J 3 Q 9 03 c '3 o J3 - CJ u 10 X X X X X X 330. A Mylesinus schoniburqki Cuvier and Valenciennes . 331. A Acnodon oliqocunlhus (Miiller and Troschel) . X X 332. A Piaraclus macropomus (Cuvier) 333. A Metynnis hypsauchen (Miiller and Troschel).. . X X X X 334. A Metynnis tnaculat-us (Kner).. . X X V X x 336. A Myloplus schomburqki (Jardine) X X X X X 337. A Myloplus torquatus (liner) 338. * Myloplus kneri (Steindachner) . X 339. A Myloplus discoideus (Kner).. . X 340. * Myloplus rubripinnis (Miiller and Troschel). X X X X X X V 341. L * Myloplus asterias (Miiller and Troschel) X 342. A Myloplus rhomboidalis (Cuvier). . X V 343. A L Mylosoma aureum (Spix) X X X X X 346. A Hydrolicus scomberoides (Cuvier) 347. A Exodon paradoxus (Miiller and Troschel) . . X X X X X X 348. A Raboides affinis Giinther 349. * Rwboid.es thurni Eigenmann X X 350. R Raboides microlepis (Reinhardt) 351. A L Charax qibbosus (Gronow). . . X X X X X X X X 352. * Charax rupununi Eigenmann. . 353. * * Asiphonichthys hemigrammus Eigenmann. . X X X X 356. * * Acanthocharax microlepis Eigenmann. . X X X X X 359. A Acestrorhynchus falcotus (Bloch). . . X X X X X X X X X X X X X X X X X X X X X X X X V 360. A Acestrorhynchus microlepis (Schomhurgk). 361. A Acestrorhynchus falcirostris (Cuvier). . 362. * Acestrorhynchus nasutus Eigenmann 363. A Hydrocynnus ciwieri (Agassiz) 364. A L Hoplias malabaricus (Bloch) X X X X X X X X ? X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 366. A L Hoplcrythrinus unitwniatus (Spix). X X X X X V 367. A Erythrinus erythrinus (Bloch and Schneider) . X X X X X X V 369. A L Gymnotus carapo (Linnaeus) 370. A Sternopyqus macrurus (Bloch and Schneider) .... X X X X X X X X X X X 373. A Eiqenmannia line.atus (Miiller and Troschel) 19 . . X X X 374. A Steatoqenys eleqans (Steindachner). . . X X X X X X V X X V 379. A L Rhamphichthys marmoralus Castelnau 380. * Sternarchorhynchusoxyrhynchus (Mull. &Trosch.) 381. * * Stemarchogiton sachsi (Peters'! X X X V X X X X X 19 No. 373 = No. 272. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 71 a 'm 03 m o o o _a X o 1 03 a — 03 o-g •so O -J +^ *c 2 a 03 m O o a 03 S) O (3 3 a 3 a ■3 p. 3 « 4 6 c u £ 10 384. * Porotergus gymnotus Ellis X X X X X 385. * Sternarchus leptorhynchus Ellis. . . 386. A L Symbranchus marmoratus Bloch X X X v 3S7. M Tarpon atlanticus (Cuvier and Valenciennes) .... X 388. M * * Rhinosardinia serratn Eigenmann. . . V 389. M A L Ilisha flavipinnis (Valenciennes) .. . X X X 390. M * Odontognathus mucronatus Lac^pede 391. MA Pristigaster cayanus Cuvier 392. M A Stolephorus spinifer (Cuvier and Valenciennes). X 393. M A Stolephomts clupeoides (Swainson).. . . X 394. M * Stolephorus guianensis Eigenmann X X X X 395. M A Stolephorus surinamensis Bleeker X 396. MA Ptcrengraulis atherinoides (Linnaeus).. . 397. M A Lycengraulis grossidens (Cuvier) 39S. A Osteoglossutn bidrrhosum Agassiz. . . X X X X X X 399. A Arapaima gigas (Cuvier) V V 401. A Rivuhis urophthalmus Giinther. . . V 402. A Rivuhis micropus Steindachner V 403. A Rivuhis obscurus Garman X 404. * Rivuhis breviceps Eigenmann. . . X 405. * Rivuhis holmice Eigenmann X 406. * Rivuhis ivaimacui Eigenmann X 407. * Rivuhis stagnatus Eigenmann. . . X 408. * Rivuhis lanceolatus Eigenmann. X X 409. * Rivuhis frenatus Eigenmann X X X X X X X X X 411. A Anableps microlcpis M tiller and Troschel . . 412. A L Poecilia vivipera Bloch and Schneider X X X Y 413. A Acanthophacelus reticulatus (Peters) X 414. * Acanthophacelus melanzonus Eigenmann. . 415. * Acanthophacelus bifurcus Eigenmann. . . 416. * * Tomeurus gracilis Eigenmann X 417. M Tylosurus microps (Giinther). . . X X 418. M Tylosurus almeida Quov and Gaimard 419. A L Potamorrhaphis guianensis (Schomburgk) X X X X y X ? ? ? ? 421 . M Mugil brasiliensis Agassiz ? ? •> •> •? ? ? 9 X ? ? X X ? ? ? ? X ? 422. M Mugil cephalus Linnaeus . . ? 423. M Mugil incilis Hancock ? ? 425. M Cynoscions acoupa Lac^pede 20 , 21 Y 426. M Cynoscion leiarchus (Cuvier and Valenciennes).. . Y 427. M Cynoscion virescens (Cuvier and Valenciennes) 22 . . X X X 428. M Cynoscion steindachncri (Jordan). 429. M Cynoscion microlepidotus (Cuvier and Val.) 23 X X X 430. M Macrodon ancylodon (Bloch and Schneider) 2 - 1 . . X Y 431. M Sagenichthys ancylodon (Bloch and Schneider)... . Y 20 Several other species of Scia?nidae may enter rivers. 21 Lake Maracaibo. 22 Porto Alegro, Brazil. 83 Brazil. 24 Panama, Rio Grande do Sul, Montevideo, West Indies. 72 MEMOIRS OF THE CARNEGIE MUSEUM _d "53 C3 PQ o o o O 1 C3 a o-g ■so 6| 2 a '53 ca m o a 03 a o 3 3 3 a 3 a 3 « 4 6 & 3 o- a> en H u St is o ►J 5 6 03 o c^ u to is o >1 6 d u 03 o Ph u CJ p. a 7 u o3 u s Q 8 03 c 03 '3 o u -*^ 3 Q 9 03 c 03 '3 o a 0J u 10 432. M Bairdiella ronchus (Cuvier and Valenciennes) 25 . . X X X 433. M Nebris microps Cuvier and Valenciennes 26 X X X 434. A Plagioscion squamosissinius (Heckel) 27 . . . X X 435. A Plagioscion auratus Castelnau 28 . . 436. Plagioscion heterolepis (Bleeker) X X 437. A Plagioscion surinamensis (Bleeker) 29 . . . 438. M Stelliferus rastrifer Jordan. . . X X X 439. A Pachypops fwcrceus (Lacepede) X X X 440. M Pachypops tri fills (Mtiller and Troschel). 441. * Pachypops grunniens (Schomburgk). . X X 442. ALM Micropogon furnieri (Desmarest) 30 . . . X X X X X X X X X 443. M Lonchiurus lanceolata Bloch 444. M Centropomus undecimalis (Bloch). 445. M Centropomus ensiferus Poey .... 44fi. A C hwtobranchus flavescens Heckel. . X X X X X 447. A Acaropsis nassa (Heckel). . X y 448. * * Nannacara anomala Regan. . . X X 450. * /Equidens maronii (Steindaehner) X X X X 451. A /Equidens vittata (Heckel). X 452. * /Equidens polaroensis Eigenmann. . X X X X X 453. A ^Equidens telramerus (Heckel).. . X X X X X Y 454. * .Equidens geayi (Pellegrin). X V 456. A L Astronolus ocellatus (Agassiz). X 457. A L Mesonauta festivum Heckel X X X X X X X 458. A Cichlasoma bimaculatum (Linnaeus). X X V 460. A Cichlasoma scvcrum (Heckel). X X 461. A Cichlasoma psittacum (Heckel). X 462. A Acarichthys heckelii (Miiller and Troschel). X X X 463. A Biotodoma cupido (Heckel) . . . X X 465. A Geophagus surinamensis (Bloch).. . X X X X X X X X X X 466. A Geophagus jurupari Heckel. V 468. * Hetcrogramma ortmanni Eigenmann. X X X X X X X X X 469. * Hetcrogramma steindachneri Regan. X X X X 470. A Cichla ocellaris Bloch and Schneider. . . X X X X X 471. A Cichla temensis Humboldt. 472. A Crenicara punctulata (Gunther) X X ? X X X 473. * Batrachops punctulatus Regan. 474. A Batrachops semijasciatus Heckel. X 475. A Crenicichla saxaiilis (Linnaeus). X X X X 476. * Crenicichla alta Eigenmann . X X X X 477. A ? Crenicichla icallacei Regan. X X 4S0. A Crenicichla lentieulata Heckel X '' Maracaibo. 26 Panama. 27 Rios Crixas & Araguay. 28 Rio Ucayale. 20 Venezuela, Cauca. 30 Havana. 31 Guiana. EIGENMANN! THE FRESHWATER FISHES OF BRITISH GUIANA 73 a "ffi 03 pa o o O a '£ O 1 03 a u- 03 •so 61 Sffl 2 'S3 03 ffl O u a S o 2 3 3 3 C 3 a 4 6 I w w H a o ►J 5 X X X 6 u 03 o Ph u a 'f o 6 s 03 O P^ u a) P. ft t> 7 03 03 0) o 8 03 a '3 O A o ■♦a 3 Q 9 03 a 03 '3 O j= u a 03 u s 10 X X X X 482. A Crenicichla Johanna Heckel X X X 483. A Pkrophyllum scalare (Cuvier and Valenciennes). X X 485. * Polycentrus schomburgki Mtiller and Troschel . . . X X X X ? X X ? 488. M Guavina guavina (Cuvier and Valenciennes) .... ? X X X X X X 493. A Colemosus psittacus (Bloch and Schneider) X X X Totals 89 - nn 85 83 184 149 22 167 118 75 Total No. of species in Essequibo Basin. 266 Of the four hundred and ninety-three species enumerated in the preceding list all but twenty-three, which are recorded from the Rio Branco only, inhabit the northward-flowing streams from the Orinoco on the west to the Brazilian boundary on the east. Of these two hundred and forty, or over 50 per cent., are also found in the Amazon. In the Rio Branco, the only stream of the Guianas flowing southward which has been examined, twenty-three species have been taken, aside from those which it has in common with the upper Essequibo. The more readily to show the relation of the Rio Branco to the Rupununi and upper Essequibo, the fauna of the Essequibo is divided into (4) the fauna of the Essequibo above the Warraputa Cataract, (5) the fauna below the Warraputa, (6) the fauna of the lower Potaro, (7) the fauna of the upper Potaro. From the Essequibo basin two hundred and sixty-six species have been taken, and this number may serve as a measure of our information, or lack of information, about the inhabitants of the other streams. From the large Orinoco basin but eighty-nine species have been recorded. Distribution of the Species Obtained in the Area Examined. The maps, given in Plates LXXI-CIII inclusive, graphically show the points at which the commoner species characteristic of the fauna were taken. A refer- ence to these maps will, I trust, be useful to the student who is interested in the question of distribution. 74 MEMOIRS OF THE CARNEGIE MUSEUM In the following table the distribution of the species obtained is given. Species not obtained by me but reported by others have their approximate locality indicated by an asterisk. The number of specimens of each species obtained is given in the appropriate column under the several localities specified, thus indicating for each locality the relative abundance of the respective species found thereat, as well as the localities where each species was most numerous. For several unique localities the lists here given are repeated on subsequent pages. The columns are not quite of equal value. Under the head of Lama Stop-off are given all specimens obtained between Maduni Stop-off and Cane Grove Corner. Few came from the latter place, most of them from Lama Stop-off, which is between the other two. The specimens from the Georgetown trenches include those from the Botanic Garden. The specimens from the northwestern coast include all those taken in and near Morawhanna, i. e., Mora Passage, Aruka River, Koriabo Rubber Plantation, Issorora Plantation. The Kumaka column includes specimens from the Demerara River and its tributaries between Kumaka and Christianburg, a distance of several miles, and from some distance farther down the Demarara itself, at some mud-flats. The Rockstone column contains specimens from Rockstone itself, from the sand-bank just above Rockstone, and from a stream on Gluck Island across from Rockstone. The columns from Crab Falls, Tumatu- mari, Potaro Landing, Tukeit, Savannah Landing and Holmia contain similar groups of habitat. In all of these cases a radius of a mile would probably include all of them. I do not know how much Rupununi includes, but certainly Twoca Pan, and a creek opposite Massara Landing. All the specimens from this region were collected by Mr. Grant. Where a species was abundant in the Georgetown market the sign of infinity is substituted for the definite number of specimens preserved. The water of the lower Demerara (see Figure 1) and of the ocean for some distance out is muddy. This muddy water is inhabited by a peculiar fauna. It is rich in Aspredinince, Ariince, Mugil, Anableps, Scicenidce, several sharks, and a sting- ray. In so far as these have any relation to the strictly fresh-water fauna they are included in the present report. No sharp line has been or can be drawn. No species has been omitted which should be included, but some of the species included should probably be excluded. This is notably true of some of the Scicenidce, Centro- pi mi idee, and Mugilidce which are generally distributed in the West Indies. The Aspredinince, Ariince, and Anableps more clearly belong to the South American estuarial fauna and all properly belong here. Most of the fishes brought to the Georgetown market come from the estuaries or coast. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 75 •OOJKClB^OIN '3f33JO OOdjqJ 03 '( S IIM ^aoj -BJJ 8AOqB) TUOldUBJ^ ■BUIIB-JBnjy EtuqOJJ o. •StnpnB'i qEauBAttg 'l! 05 l n J, •^rutjjB^ •^ri)Bcny tn^rug ■T:tnnii;Stre}j SuipnE'j oje}oj ■uBinnjRiunx uuj eoomj, 'rannnclnji •ooajpEj H •BindxjjjB^ '^njBMuaojj ' S IIM 'F-'O ■pi ^on]Q 'auo^ssiooy ttDi^jBa '1F113IM 'SjnqmMjsuqo •^STJOQ }S9Mq}JOf^ i* 888 : : : : 88^^888* 888 saqonajx nAvo^Sjoag O ■janjoQ 3aojq gut!Q 'jgo-do^g ranp 03 I •8 S a s. i ^€0 _OD OS S. S. s OS oo e 3 e e £ S S c -i-i ■£ _ fti to >f gq co cq cq cq 3; . 55. oo .S -.2 3 e .?? <■„• :|4f s s : e g-3 e 2 oo g. So 33 E "ta E-B " 5| 8 & H C 3 § 3 3 S5. S =2 2 03 to -C ci feife; Anus Hexan Callop Megal e g CDs e §• :.s 1-2 a g'S 2 S 3 -° S C c X c d T3'« &3 CO S b S ; CO C3 s CO &5 CO c CO J- O 'i 33 cc s s • — --> CO c 25 CO ,3j Si =5 a. o — C3 , "o •< o S a •a 9 CB SI ft EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 77 78 MEMOIRS OF THE CARNEGIE MUSEUM a ■2 o N s < OOJEdB^Ot^ •3[83J3 oodiqQ CM ■ ■ • ^H ■ - • • -Bj£ 3A0qB) ruaiducjv \ • co.-i ■ • *— < ■ CI ; ; 6 03 -+J O Ph u i 3 1 n X . . CM r^:co-H;;;;;;:;: . ira •3[rnBJc y^ ■sp^Buiy : '. '. oo ; ; co : ;^k:co:;;:::'-<:: ■np(rua . . . -h • ■ ~H - - • ■ ••■ CO *o [njBMBUOJ£ : : ■* : : c < :cc : : :■- , i— i . . ... — 1 •• • ■— 1 CI ^J • CO ... . . j^i . . . . CM i— I CN • r-- • • CM • • ■ rt ■ BaijiBg : ; : id :::::::::::: :o : ............ .0* . M a u O a ■jppw CO <*H ^H Sjnqa^psuqQ puB jbuisi .^ 'Bjpjuinji c O • 00 CM • ■ H • . . ■ iO • • l^ -CI >o • '. CO ■ ■ ■ • ■ i-l ■ -ci • -CO ■03 ; ; GO 1 CD as 4^ CO OS o o '^SBOQ ^S8Mq}J0fv[ ■ • • CI 'O CO 00 • • ■ m co • • • i— i • i-i ■•* ■ ^H joqiBjj pire ■saqonajx aA\o)33.ioag _i . • . t^ • id ■ ■ • ■ •* • • •JGaJOQ 9A0JQ aaBQ 'jgo-do^g ranp •0 CO i-H ■ ■co ; Prochilodus rubrotwniatus Tylobronchus maculosus CI o a 3 a 3 EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 79 • o ■ <— ( ■ rH - ^ | rH co '■^& •00 CO o IN t~ • CT> to CO CM CO to "OS :- ^^ o : CM . CM 00 o CM CM 3 >o • 3 CM — 1 CD O CM CM • I> • CM .SJ ."5 CO •s £. S s -g a &~ co S © to s» r- -co ,c to & Oh CO 1 o e o CO s CO a, oo 3 c "oo CO | CO s ~= CO e .§ Q 6- "co CO s ."3 co «r- O .N to 00 Co .c; .00 CO Q o o .M cj &2 CO &. 53 CO O co CO ."3 S Q Cr e % "eS> *«. O CO oo s co c g oo © •ra 5 &~ & CO g O Co Cu ~5 CO Si .g ©o CO V. CO © oo CO CO g s = "3 S =0 ■ — o c d O CO i o .CJ e o M CO 1 CO CO 1 CO a O - 2 e CO s o o C oo = CO CO CO r~ CO CO o co- CS ~T Eh s. Co co ■Sr .53 CO = --. oo •ri «r^ CO ft b CO g e S3 t5 _ -f. ~? C3 r. CO .e CO C3 (3 CO r- C3 r~ C3 oo "to .53 CO rS* -~ o S e oo C e Q S •"2 CO ,e S3 e ^; ^c S 8 -- =0 .cj CO o C o V. e § "^ CO = 5 c" CO CO S b -~ — ; rt V, .CO cc e [S co 5 a .5 CO e •ra s oo CO a -t! ."o 1 b 1 &5 a - - o 3 80 MEMOIRS OF THE CARNEGIE MUSEUM C3 « a c N 6 < •oojredEjp!^ •3[3ajQ oodnjQ '( s lI B .i «noi -G}J 3AOqE) TUOldUBJ^ o CO - o ■- 03 -*J O u V tt a - GraiB^Biuy 1 05 Birajojj j m •SnipuBi qBnaBAGg oo tN CO 6 o Ph u o hi ?! 33 I n X 2 co CO •jfmujBjVi i> co •3[niBoiy | CD oo CO f~ CD CD' Ttl 00 ■np(nja | CN CN 'BcnoreSuGjj CN ^H 00 ■Snipirei oibjoj 1 iO CN CN CN ■trecarnBiimj, 1^ Oi co ^h C5 r^. CD rHQM CN •^ Oi co I> O ^H CO tM ooo CN OS 6 .o '3 a OJ 0Q CO H ubjboomx 'rannndnjj * oogjpEj , '-' "BjndBjjB^ i-H LO CN ■sjibj qBJ I-H CD r-i o i-H m cn «# CO (N : OS , T— 1 •]sj 3[onio 'aaojsjpoy cn 2 -3 CN co I— I 1— 1 CTiCN CN co co hqOOOh CN •BopiBg; ^H CN ^H CC f-i 1 •!I B I«IM • ■* O CO SjnquuijsuqQ pae jBins; ^ 'B!(Buin^j CN o CSOlN CM i-H -* ^ GO a 03 CU ~ -*> "3 4J /. 03 O o •}SB03 ^S3:Viq}JO>J CO CN 1—t joqjBH puB }33(JBJ^ n.MO}38j03Q •saqoaajx ua\o}33io3q (N CN o o CO J3UJ03 3AOJQ 3QB3 'jjo-do^g ranp ~ B W 'flo-do^g buigi CO CD (M C5 CO co co "- 1 E c i ft ft $~ 1 ;S 5 ; § ft- .J "■§ s c ft- i s 5 | It a e v ; s c 5. o a t> : v <: t, P - S 3 s C ft e 5 c ft, v. a S 1 § a c c r> a -c v. <4. -c I. ? o a e t- c a c ^ a -c v. -s - S i a • .ft. Z c & ft. C 1 5 e 5. a e IS s "; S ft, J. ) ? S s •T- ? c tr C ft v. { u « c c. c i a i s- ft. (5 1 c . p c o •I c .t 1 -^ ft. 1 ft, a 1 ft ft. a c: c v. ft. a - c 's c ft: -^ ft. ft •r- ■> V s c c &. K ~; "6 ■p. c c e a CO c 3 5 C3 •S -S 3 o EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 81 cr. T— CC w— CN CO Tf CO H LO CM CM • H co CM O C33 ~ *. l- CO Tf CO Tf CM CO CM * i-H co : o o CO * ^H r-( * r co CM - .-H ^H CO iH CD i— 1 1 ^H r+ T-< ^H ^H CM CO o CM - CM t-H l> I-H CM ^H O ^H CC co TjH CO ^h CM — ■' M CM o CM CO 00 —1 ^H U5 co ^i GC CM - -: • CD CM * CN i-H 1— ■ i-H :t ^H IO OS CM ^Ht^CM^i— lt~CM"-(Ci0^-llOrti-l CM t-1 CN M Tf CM CO co ; 1— 1 i-H # i-H ^H • T— l ^H (M - Tf CM O ^H i-H LO IC i—l CO - CN o Tf LO CM T-H ^-H oo CM C 1- i— f -t o I- o -t CN I- CM CM CM T-H T-H oo T-H CO CO CM CO o CM CM O i-H CM ft S c = ~— ; ft s e j s c s- ft 5 .S 1 •- ft 6 !C* e 1 _2 1 & ft ft 5 a £ "ft c G s "1 ft C £- S 1 -. • : c c ft I | 1 C z B C j 5 ! 5 e c ft § ^ a ft. z c ci "J V S z 1 5 c s •a. ? j ■ 5 ft X. •*• 5 O ! 1 1 1 E ft ft s .= * s - : 3 ■ — c J 1 c V -1 c ft P ; .0 S K - P & - '1 ft s 1 &■ '1 ■> s % » ; - ~ c a - - X IS c ft: ft = ft c - = | I ft = i c ; ft ~- c s •5 = ■a. 2 ft = ; I c i c 5 ,5 E c e = h c s "3 Cv Ok 1 E J g ?: . a — C c | J - - 1*1 1; s c i a c ft i ,1 c: ft c s 1 s i a 1 1 1 K C c c c ft ; : s ft s t - C c ft p c C - K . J ■s cs g c 1 1 Si. ? : E C >. ^. "oo 1 O -° , = 'S o » a. o So I o •3^ a* o » g Ph O . o .-, a §g i cd o3 n. n ^h <3 3 ^ . I? 3 O CT' — <" e 5 O o & o CJPH | S 3 O - C3 K^ ° 3 13 ?H 82 MEMOIRS OF THE CARNEGIE MUSEUM oojBdBspii^ •3J33JQ oodiqQ '( s II b jI viol -BJJ 3AOqB) IUOldlJBJ^ 03 is 3 o PL, o BaitBjBruy bihijoij SaipUB-J qBUUBABg •^la^nx •JJIUBJ'BAi ■^njBmy upiiua BinruBSuBjj SaipaBq ojb^oj •UBiun^ainx nBjBaoMj, 'tunundny • CO •CO OOa^DBJ -B}ndEjjBj\Y H •^tUBAlEUOJI s ii b j q^o x pi ^on[Q 'ano}B3[ooy ■s =1 ■s 3 W BDiiJua HB|BJ\[ SjnqtrepsuqQ • o ■ N ■ IC <-H »C •jsboq 'jsaMqjjoi^ ■ CO ■ Gi ■ C^ joqaBjj pus 13![JBJ\[ nMOjaSjoaQ ■ H ■ 3 S o s~ Si.g S 5 s 3-5 ; s s -J ft.g" ' Si-e ■ -> Oi C35 a;, cr, O =0 Co S 8 " c S © e c ^3 02 S ^ =•'2 •s s S "S lit : •§'£•§ £- C3i 00 =0 *^> e 5 co . 8 s » g •S ai s . 2 R. ^"S g s s W c^ C3 cu ci> S'S s. " e £ — o o S 5 ^ LV,a,o-jcc^o^t-sQ4 IS, Sa S 2 2 g co m oo co oo R. & m „^ ^ ^ ^j Cu ^ S .35ooo.g.» . so ■ eo • s • . CD co 2 IS .co ""^ ~ -?; Cj ^ 00 " « o* S cj b ~JS c the q o '3 a" 5 5jS5 o L^ co co ^2 C3 cj ^• S w ?> Si Si ^ EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 83 CM ' 00 lO oo CO co as ^^ Cl tH CM ^H CI 00 t^ oo i— I r-H ^H m rH CD C) ■* f-H - — ■-1 w 00 00 CO CD Ci 03 ■* C ^H CO CD i-H GO CJiCOliH CM CO CO oo io • co CO t^ CI CD j £ ■— 1 > e a g Q CJ O CJ oo >g 5 e g s = J CO c 1 5 C3 J. g c 5 ? 4 2Q 3 .£ .0! > - a s co K % £ 60 00 00 O £ o CC S ,c 00 .<; > s •J? s. 00 3 g CJ s. s- CO 3 = a ft. g g 3 5 oo p ' J CJ a ftn CO i V. k. c o 00 1 ■ s s s CO - - s U 60 60 g \ 00 00 a g .°! C - e s - : 1 s c -S e -^ e Cj s < *o3 e e s. - E 00 "J - 5 «9 R C c p 6C =: 00 «5 s £ S e S s O g "o: oc C oo O! Oo . V. e e o [ni3 •BtanjB3nBjj ■SaipaB'j ojB^oj ■UBain^Buinx cm CM • 6 '3 O" 01 00 03 H •nBj Boo.vix 'ranundn^j •ocrajpBj ■B^IldBJJB^ •^iubmbuoj]; •snBj: qBjQ -: ■JSI J[0niQ '3UOJS5[DOJJ ■* BOI^JBg SI a p ■JPFH ■3.mqueijsuq3 pnB 'jbiusj\y 'BjjBamjj; m in ea O O "JSBOQ ^S3Mq}JOJVJ CM •joq-reg puB }35[iBJ\[ UM0)33J090 CN 8 •satpuajx nM.0533j03Q i— 1 sdbo 'jgo-do}S ranp ~ V W 'flo-do'jg BlUB^ CO ft CD ,«D -c CO s o o CO "5 CO - c ■ — CM - rH rH ; -h ; i— t ^h ri -r — :- :cm I - * ; f-h \ i—i | ^H ; o ^h 1 rH :cm CO • Cvi rH • CO CO T-H •CM | i-H :cm ^h ICM '. CO IrH co : ■* ^H ; co ^ -h 1 rH ; cm ICO :>o Ol — CO -r CM ) rH : hh co -r T^ ^H '-' : co --I ;— - : co CO CM ^H | ^H ; oh .CM jrH t^ ^H ^H t— I O ^ CM ^ I'* 1 CO »— I Oi HHIOCO l— t " 1— 1 '00 :o oo cm ^H ^H ~ ^ r* T— 1 - 1 i - :t :o « CM rH 00 ^ CO rH CN .'r-l CM CN CM ^H CM ^ rH co T-H HHIO i-H m co CM --i -H CO CO CC 00 rH CO rH 1— 1 O rH 3 03 1C r-H hhoio CM >-l i-( CM ^ t-H IrHCM o ■* CO CD ~ : : : Ol rH CM CO C5 CM • • • CM rH CM ■* t^ CM Tt< CM "* . i-H rH CO : HHHH O t^ ■ HMrt • CM - - ; - ; CO rH a cj ►> CO Q — — a CQ & ■7. — o p. B a — a - B a '8 >■ Ph a — - S3 cj o 8J - a o & -r c a T3 p. - > 02 a CD r s o - — - < a o a — ■c o "3 — 03 ± T z - a _o CD - p. = o Q a 1 a .1 "It 02 r. C P. c 3 g s - 'C CD CD Z Ph a — o O a -r "o 02 a — c o -a o - CO H CO "3 -fH EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 85 •oojEdB5[o;f.*Helogenes marmoraius C PyGIDIIDjE. Cetops ince. 24.*Hemicetopsis maciientus A 25.*Hemicetopsis minutus B Pygidiince. 26.*Pygidium conradi B C 27.*Pygidium gracilior B 28.*Pygidium guianensis CallichthyidjE. 29. Callichthys callichthys (A B C?) 30. Corydoras punctatus A B LORICARIID/E. Plecostomince. 31.**Lithogenes villosus 32. Plecostomus hemiurus (A?) B C 33.**Corymbophanes andersoni 34. Hemiancistrus megacephalus B C 35.*Pseudancistrus nigrescens 36.*Lithoxus lithoides (A?) B C Loricariince. 37. Loricariichthys brumieus A 38. Loricariichthys platyurus A Characid^e. Curimatinm. 39. Curimatus spilurus A (B?) C 40. Curimatus microcephalus ABC Prochilodinm. 41. Prochilodus rubrotseniatus A Hemiodince. 42. Hemiodus quadrimaculatus ABC 43. Anisitsia notata A Chilodince. 44. Tylobronchus maculosus A B Nannostomatince. 45.*Characidium laterale (A?) B 46.*Characidium vintoni C 47. Characidium blennioides ABC 48.*Nannostomus minimus B 49. Pcecilobrycon erythrurus B 50. Pcecilobrycon ocellatus A Anostomirue. 51. Anostomus anostomus A B 52. Anostomus plicatus A B 53. Schizodon fasciatus A Leporinince. 54.*Leporinus arcus ABC 55. Leporinus nigrotsniatus A 56. Leporinus friderici A B 57. Leporinus maculatus ABC 58. Leporinus alternus A (B?) C 59. Leporinus fasciatus A Pyrrhulinince. 60. Pyrrhulina filamentosa (A? B C) Crenuchince. Ql**Poecilocharax bovallU C Iguanodectince. 62. Iguanodectes tenuis A Tetragonopterince. 63.*Tetragonopterus argenteus A 64. Tetragonopterus calceus B EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUTANA 97 §5* Mcenkhausia oligolcpis A B (C?) 66. Mcenkhausia grandisquamis A B 67. Mcenkhausia chrysargyrea A (j8*Mcenkhausia browni ABC 69. Mcenkhausia shideleri A 70. Mcenkhausia lepidurus A B 71. Mcenkhausia cotinho ABC 72. Moenkhausia colletti A B 73.*Hemigrammus erythrozonus B 74. Hemigrammus orthus (A B?) C 75. Hemigrammus cylindricus A 76. Hyphessobrycon eos B C 77.*Dermatocheir catablepta A 78. Creatochanes affinis ABC 79. Creatochanes caudomaeulatus A 80. Creagrutus melanzonus A 81.*Bryeonamericus hyphesson A 82. Pcecilurichthys polylepis A 83. Pcecilurichthys bimaculalus (A B C?) 84. Pcecilurichthys abramoides A B 85. Pcecilurichthys potaroensis B C 86. Astyanax mucronatus A (B?) C 87. Astyanax essequibensis A 88. Astyanax guianensis A 89.*Astyanax mutator 90. Deuterodon pinnatus (A?) B C 91.*Deuterodon potaroensis A B 92. Phenacogaster microstictus A 93. Phenacogaster megalostictus ABC Bryconince. 94. Holobrycon pesu A 95. Brycon falcatus (A B?) C Chalcininm. 96. Chalceus macrolepidotus A Brachycalcinince. 97. Fowlerina orbicularis A B Gasteropelicince. 98. Carnegiella strigata A 99. Gasteropelicus sternicla (A B C?) SerrasalmiruE. 100. Serrasalmo gymnogenys A 101. Serrasalmo rhombeus A Mylince. 102. Myloplus rubripinnis A 103. Myloplus rhomboidalis A B 104. Myleus pacu A Characince. 105. Charax gibbosus A 106. Cynopotamus essequibensis A 107. Acanthocharax microlepis A A ceslrorhynchince. 108. Acestrorhynchus falcatus ABC 109. Acestrorhynchus microlepis A B Erythrinince. 110. Hoplias macrophthalmus A (B?) C 111. Hoplias malabaricus ABC 112. Hoplcrythrinus unitceniatus (A B C?) 113. Erythrinus erythrinus A (B?) C GYMNOTIDiE. Electrophorince. 114. Electrophorus electricus A Gymnotince. 115. Gymnotus carapo (A B?) C Hijpopomince. 116. Sternopygus macrurus ABC 117. Hypopomus brcvirostris (A B C?) 118. Eigenmannia macrops A 119. Eigenmannia virescens A B 120. Rhamphichthys rostratus A 121. Gymnorhamphichthys hypostomus A. Stemarchince. 122. Sternarchorhynchus oxyrhynchus (A?) B 123. Porotergus gymnotus B 124. Sternarchus leptorhynchus B 125. Sternarchus albifrons A PCECILIID.E. 126.*Rivulus breviceps C 127.*Rivulus waimacui B C 128.*Rivulus holmiae BELONIDjE. 129. Potamorrhaphis guianensis A 98 MEMOIRS OF THE CARNEGIE MUSEUM Sclenid^e. 138. Heterogramma steindachneri A 130. Pachyurus schornburgki A 139. Cichla ocellaris A Cichlimi 140. Crenicichla alta ABC 131. Nannocara bimaculata B 14L Crenicichla wallacei A 132.*iEquidens geayi A B 142 - Crenicichla lugubris ABC 133. Mquidens potaroensis ABC 143 - Crenicichla Johanna A 134. Cichlasoma severum B Soleidjs. 135. Geophagus surinamensis A 144. Achiurus lineatus A 136. Biotodoma cupido A 145.**Seleonasus finis A 137. Heterogramma ortmanni (A?) B (C?) Of these, Poecilurichthys bimaculatus, Hypopomus brevirostris, and Callichthys callichthys have not been taken in the Potaro below the Kaieteur, but, inasmuch as they are found on the plateau and in the Essequibo, they probably also occur in the lower Potaro, and are so rated. The table on page 99 summarizes the different categories in the preceding list. In this table the first column represents the species found below the Kaieteur, and it will be seen that while there are one hundred and forty species found in this part of the river, there are but twenty-three above the Kaieteur. It is, however, not fair to compare the fauna of the plateau with the fauna of the entire lower Potaro, because the fishes of the Essequibo can easily ascend to Tumatumari. The cataract at Tumatumari is probably not an effective barrier to many fishes. There is a more difficult series of cataracts between Potaro Landing and Kangaruma, and another cataract at Amatuk. The cataract at Waratuk, like that of Tumatumari, may be left out of account. Hy comparing the fauna as determined in the three lower divisions of the Potaro it is seen that only seventy- six species (B and C) are found above Kangaruma. The rest, with very few excep- tions, are fishes of the lower and middle Essequibo. Evidently species that are not able to ascend the Potaro Landing cataracts need not be expected on the plateau. By further restricting the vision we find that only thirty-six species have so far been taken above the Amatuk Cataract. This is certainly not all of the species found in this area, but the number indicates that there is a distinct reduction in the number of species as compared with the Potaro at Tumatumari. Further investigations should be made of the fauna of the Potaro above the Amatuk cataract and above the Kaieteur. Eight species found above the Kaieteur and in the Essequibo, but not taken in C, should probably be added to the thirty-six actually captured. It would seem from these lists that the groups extending to the base of the Kaieteur, but not found above it, are the Curimatinse, Hemiodinse, Nannostomatinae, Leporininae, Bryconinse, Gasteropelicinse and Acestrorhynchinse. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 99 Aspredinidae Siluridae: Pimelodinae Doradinae Auchenipterinae . . Ageneiosinae Helogeneidae Pygidiidae: Cetopsinae Pygidiinae Callichthyidae Loricariidae: Plecostominae Loricariinae Characidae: CurimatinsB Prochilodina? Hemiodinae Chilodinae Nannostomatinae Anostomatime. . . Leporinae Pyrrhulininse . . . . Crenuchinae Iguanodectinae . . . Tetragonopterina? Brycoiiina? Chalcinina? Brachycalcininse . . Gasteropelicinae . . Serrasalminae . . . . Mylina? Characinae Acestrorhynchinae Erythrininae Gymnotidae: Electrophorinffi. . . Sternarchinae . . . . Hypopominae . . . . Gymnotinae Poeciliidae Belonidae SciaDnidae Cichlidae Achiridas Species Below the A B c Species Above the Kaieteur. Kaieteur. 1 1 1 13 10 5(6?) 2(3?) 1 o 4(5?) 1 2 1 1 1 1 1 2 1 1 2 2 1 1 2 1 1 1 4 (2?) 4 3 2 2 2 2 1(2?) 2 1 1 2 2 1 1 1 1 1 6 2(3?) 5 1 3 3 2 6 6 3 (or 4?) 3 1 (1?) 1 1 1 1 1 1 1 1 30 23(26?) 12(15?) 9(11?) 5 2 1(2?) 0(1?) 1 1 1 1 2 1(2?) (1?) (1?) 2 3 1 3 2 2 2 1 4 3(4?) 1(4?) 2(4?) 3 1 4 1(2?) 3 6 5(6?) 2(3?) 1(2?) 2 1 (1?) (1?) 1 1 2 1 2 1 1 1 1 1 13 10(11?) 7 3(4?) 3 2 2 140 87(102?) 58(71?) | 36(44?) 23 2. Whence did the fauna of the plateau come? A. Is it the nucleus of the original fauna of Guiana? Or B. Have the faunas on the plateau and the lowland developed from a common nucleus? Or C. Is it a relict of a more abundant modern fauna? Or D. Does it consist of recent immigrants? Or E. Is it a mixture? To assist in answering these questions I give the species of the plateau with their distribution in the Potaro and also their general distribution in South America. 100 MEMOIRS OF THE CARNEGIE MUSEUM s e3 a 05 a 5 H 3 3 -^ 03 B 3 3 03 "53 3 H si c ^3 a eS J= 03 a a oj > 03 03 | M si S 3 "5 3 1 3 h 'a, 03 % a O d u 03 a. 03 1 c 3 03 03 s 03 Hi 6 CO '3 3 03 a 03 03 1 X 13 19 1 41 15 3 8 X X 2. Helogenes marmoratus ? 14 15 7 in 1 1 X X 1 1 35 6. Corymbophanes andersoni. . . 7. Pyrrhulina filamentosa 365 95 40 2 7 63 7 9 27 50 220 41 104 3 41 14 x 9. Mcenkhausia oligolepis 8 1 47 1 3 1 X ? 2 6 69 38 1 6 4 5 8 45| 9 5 30 1 x 12. Poecilurichthys bimaculatus. . 233 X X V 121 4 2 14. Hoplias malabaricus 15. Hoplerythrinus uiiita3niatus. . 72 15 7 4 12 1 35 18 8 15 11 1 5 1 12 1 9 S 2 10 9 15 3 1 16 32 4 X X X X X X X X X X X X 16. Ervthrinus erythrinus 2 10 30 3 37 ... 34 12 4 181 1 1 2 5 V 18. Hypopomus brevirostris ?, 18a. Eigenmannia virescens 2 1 1 3 V 19. Rivulus holmise 18 33 1 9 2 16 35 1 4 1 3 46 2 17 1 25 27 5 X 9 2 4 2 8 The numbers in the respective squares show the number of specimens taken and thus indicate the relative abundance of the species at each place. An analysis of this list shows three groups of species. a. Very widely distributed species: 1, 4, 7, 9, 11, 12, 14, 15, 16, 17, 18, 21.- All of these are found in the neighboring river Ireng, of the Amazon basin. All of these are also found below the Kaieteur, but in different proportions. The per- centage of the total fauna of the plateau formed by this group is 56.60. b. Peculiar species of very widely distributed genera. (1) Confined to the plateau: 3, 13, 19 = 13.01 per cent. (2) Found also below the Kaieteur: 10, 20, 22 = 13.01 per cent. c. Peculiar genera and species. (1) Peculiar to the plateau: 5, 6 = 8.69 per cent. (2) Found both above and below the Kaieteur, but confined to the Potaro (?):2, 8 = 8.69 per cent. A. If this fauna is the nucleus of the original undifferentiated fauna of South America we are at once confronted by the fact that 56.60 per cent, of the species are identical with species now found everywhere, and 26.02 per cent, consist of species belonging to genera of the widest distribution in South America. The EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 101 1 5 id in j> Id ■ ■— p c E- 1 u p "1 +2 re o 5 „ E "5 8 5_ 3 c in ■ a- c Rhamdia auelen Helogjenes marmoratus __7J a l C(2)2 y^ auiancnse bd)3 CaflicMtya caflicfillwa Lithogenes viliosuS H ? ? a 4 C(0 5 CorymbophaneS andersom FyrrDulina plsmentoaa PoecWobrycon Dovallii Co)6 a 7 c{2)8 HH ManktauSi'a oligolepis Moenkhausia brown! a 9 b(i)10 Creatocnanes af finis Astyanax bimaculaTua /^styanax mutator Hoplias maiaDancus Hop1ery1nrinU3 unitaeniaujs Erythrmus eryfhnnus Gymnotus carapo Hypopomus brevirobtriS a n ? 9 EauidenS poferoensis HeTerogramma ortmanni Crtn'iocMa alfa ■i_^ | ? ^ a si ^4— 1 W)22 i Fig. 23. Table showing Distribution of the Fishes taken by C. H. Eigenmann in the Potaro River above the Kaieteur Falls. (No. 18a, mentioned in the table above, should have been added. It was overlooked in preparing this figure.) great improbability that so many species should have retained their identity since the early tertiary, or should have diverged only in minor specific characters, dis- poses of the theory that this fauna as a whole is the nucleus of the original South American fauna. The number of species peculiar to the plateau forms a larger percentage than a corresponding list for the entire Potaro below the Kaieteur, but not too large to be readily accounted for by the easy access to the Lower Potaro pos- sessed by all the species of the Essequibo. Such easy access tends to lower the per- centage of uniques. The percentage of peculiar species above the Potaro Landing is about the same as for the plateau, and in actual numbers almost equal to the entire fauna of the plateau. 102 MEMOIRS OP THE CARNEGIE MUSEUM B. The same facts also dispose of the notion that this fauna as a whole may have developed concomitantly with the lowland fauna from a common nucleus. If it had so developed, we could not expect so large a percentage to be identical. C. Is this poverty-stricken fauna the relict of a more abundant and divergent fauna, dating from a time when the plateau may have been easily accessible to the present diverse fauna of the lowland? If it is such a relict the fact implies: (a) That at some time in the past, when the plateau was more easily acces- sible than now, in fact formed part of the general level, it shared the fauna of the region with the general level. (6) That the fauna became isolated by the gradual elevation and formation of the plateau, and of the falls in its rivers. A simple isolation will not account for all the facts. To account for the poverty we must assume either that the isolation took place before the origin of the present variety in the lowland, or that conditions altered so that most of the great variety of the lowland became extinct. The first horn of this dilemma assumes that the fauna as a whole is a relict of the original fauna, which we have shown to be most improbable. I am not prepared to satisfactorily deal with the other horn of the dilemma. The elevation at the present time is not sufficient to exclude these various lowland types. I do not know of any evidence that other unfavorable conditions have obliterated the fauna of the plateau. D. Does the fauna consist of recent immigrants? If it consists of recent immigrants only, we must assume that the plateau became isolated before the origin of the fauna of the lowland, or again, that at some time it became elevated to such a height that practically the entire fauna was wiped out, and later restored by immigrants. Assuming that one or the other of these alternatives is a fact, we may discuss the possibility of the immigration of species. In this connection the word "migration" as applied to fishes needs defini- tion. It is applied to such journeys as that of the eel to the ocean, or that of the salmon toward the headwaters, for purposes of reproduction. Such trips, although very long, probably only incidentally influence the dispersal of fishes. On the other hand, some species are natural pioneers, found always in the rivulets of the headwaters, as far up as they can get. If the headwater is advanced a few feet, they advance with it. Their migration from one system to another is not miraculous. To find such fishes on opposite sides of a low water-shed presents no mystery. If by a sudden freshet two rivulets join for but a short time, these species are present eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 103 to take advantage of the temporary union, and frequently do take advantage of it. Such species usually have a very wide distribution. In other regions where falls as high as the Kaieteur are the only means of migration, they are effective barriers to the downward as well as upward migration of fishes. It would seem, however, from examining the distribution in the preceding list of species 2, 8, 9, 10, and 20, that in some way or other fishes succeed in getting down from the plateau. Some of these species, especially 2, 8, and 10, are abundant above the fall, while below it they were taken at Tukeit only, or but a few miles farther down the river. They are evidently fishes of the plateau which in some way have made their way down. In contrast to these we have the species numbered 1, 4, 7, 9, 11, 12, 14, 15, 16, 17, 18, 18a, and 21, abundant everywhere below, but also found on the plateau. These species probably came up from below. They certainly did not ascend the Kaieteur. The fact that a number of species got down may imply unknown means of ascent and descent near the Kaieteur. Whether any of them ascended from the side of the Amazon I cannot say. It is quite probable, as all are found in the Amazon. The method of ascent of vertical walls by members of the genus Rivulus, one of which is found on the plateau, is of great interest. One of these fishes taken in Shrimp Creek jumped against the vertical face of a huge rock and clung by the adhesion of its tail. From this point by another flop it made and clung to a point much higher up the face of the rock. I do not know how high this genus would be able to climb a vertical wall. It certainly would not be able to ascend the face of the Kaieteur, but there may be smaller rivulets up one of which it might have made its way. The fact that a very large part of the fauna consists of species of the widest distribution, or of local species of genera of the widest distribution, seems to make the recent settlement of the plateau by some of the species certain. E. Whether or not the fauna is a mixture, the genera and species of the remain- ing group "c," i. e., Lithogenes villosus, Corymbophanes andersoni, Helogenes marmo- ratus, and Poccilocharax bovallii should answer. Now Corymbophanes differs from Plecostomus in trifling characters only. The genus Plecostomus is one of the widely distributed genera of South America, being found on both slopes of the Andes and from Panama to Buenos Aires. Corymbo- phanes may therefore be a local modification of a comparatively recent immigrant to the plateau. Helogenes marmoratus, on the other hand, is found in the Essequibo basin 104 MEMOIRS OF THE CARNEGIE MUSEUM only. It is a very aberrant Nematognath, with no relatives in other parts of South America. It may be a remnant of the original forms. Pcecilocharax stands alone in the Characince, its only near relative being Crenuchus, found in the Essequibo and Amazon. It also may be a left-over. This leaves Lithogenes villosus the only genus of the Loricariidce with the armature reduced to a few prickles. Whether we regard this nakedness as primitive or as secondarily acquired it points to a long separation from the other Loricariidce, and Lithogenes may also be a left-over. The evidence seems, then, to favor the conclusion that the larger part of the fauna of the upper Potaro has been more or less recently acquired, and that a smaller part is composed of relicts of the original fauna of the Guiana Plateau. This conclusion is, however, but tentative. Nothing is known about the fauna or physical condition of the rest of the Guiana plateau except what follows under the next head. The only other collections made were those of Richard Schomburgk, and they were all lost. It is highly desirable, therefore, to collect about Roraima, in the streams flowing in different directions. Fishes of the Ireng River. The following list may serve as a cross-reference to the preceding list of the Upper Potaro. The names of species common to the Potaro and the Ireng are here italicized. The Maripicru is a branch of the Ireng between Wontyke and Karakara above the Karona Falls. Chipoo is between Karakara and the Rupununi and is probably below the Karona Falls. The location of Nickaparoo was not given me. The total list of species suggests that we may expect to find a number of additional species in the Upper Potaro. Maripicru. Chipoo. Nickaparoo. 1. Chasmocranus longior 3 2. Rhamdia quelen 3 8 3. Trachycorystes galeatus 4 4. Cdttichthys callichthys 1 1 .5. Hoplosternum thoracatum 1 6. Plecostomus hemiurus 2 7. Ancistrus temmincki 1 2 8. Ancistrus cirrhosus. 1 9. Loricariichthys steward 12 10. Prochilodus maripicru 3 1 1 . Parodon bifasciatus 1 12. Chilodua punctatus 1 13. Charicidium fasciatum 21 14. Pyrrhulina filamentosa 14 15. Leporinus granti 9 1 Near Holmia. 1 eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 105 16. Aphyocarax erythrurus 1 17. Mocnkhausia oligolepis 1 18. Creatochanes affinis 35 19. Poecilocharax bimaculatus 1 20. Hoplias macrophthalmus 2 21. Hoplias malabaricus 3 3 22. Hoplerythrinus unitmnialus 23. Erylhrinus erythrinus 16 24. Gymnotus carapo 1 32 25. Eigenmannia virescens 3 5 26. Hypopomus brerirostris 2 4 27. ^Equidens geayi 1 28. Mquidens potaroensis* ' 29. Cichlasoma bimaculatum 2 30. Heterogramma steindachneri 2 31. Crenicichla alia ^ 1 Near Holmia. CHAPTER VII. GENERAL CONSIDERATIONS. A number of interesting and important facts, bearing on general biological problems, are incidentally dealt with in the systematic portion of this paper. Inasmuch as it is not at all to be expected that anyone would read through the mass of details to discover them they are extracted and presented here. "Unnatural Natural History." The skeletons of a variety of the Ariince are prepared and sold as "crucifix fishes." The one which is most frequently prepared is Sciadeichthys proops (Plate VI). The dorsal surface of the skull and dorsal plate are pointed out as resembling a hooded monk with outstretched arms. The ventral surface resembles the cross. Fancy pictures the dorsal spine as a representation of a spear, while the otoliths, which rattle when the skull is shaken, are the dice with which the soldiers cast lots for the garment of our Lord. Skulls of the Aspredinince are put together to form a "crown of thorns." Jordan's Law. If we should find specimens of any group with such slight differences as there are between Crenicichla saxatilis and alta or Mquidens vittatus and potaroensis generally present at any locality we should probably consider them as variations without special significance. The differences must, however, be of distinct sig- nificance, if thej- are coordinated with the environment in such a way that one can predict what the form wiU be at any locality. In the present paper such forms have been given specific designation. Thus Mquidens vittatus of the lowlands is replaced by Mquidens potaroensis in the Potaro River, both above and below the Kaieteur. Similarly Crenicichla alta replaces Crenicichla saxatilis, the two forms overlapping at Rockstone. Similar replace- ments of lowland species by highland species are abundant. Rhamdia quelen of the upper Potaro replaces Rhamdia sebce of the lowland. Helogenes marmoratus replaces Hypophthahnus edentahis. Plecostomus hemiurus replaces Plecostomns 106 EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 107 watwata of the lower courses. Moenkhausia browni replaces Moenkhausia grandi- squamis, Crcatochanes affinis replaces Creatochanes melanurus, Chalcinus rotundatus replaces Chalcinus elongatus, Charax rupununi replaces Charax gibbosus, Pinielodella cristata and P. megalops replace P. macturki. In a number of other places the replacement is not so conspicuous. In several of the above mentioned cases the species overlap. In some of the cases there is quite evident genetic connection between the two mutually excluding forms. In other cases the genetic connection, while evident, is certain y remote, as in the case of Crenuchus and Pcecilocharax, for instance, and in still other cases the relation may be apparent rather than real, i. e., one or both forms may have immigrated to adapted localities, and not have become differentiated in their present habitat. In so far as two of the above pairs represent local adaptations, that is, in- so far as one or both members of the pair are autochthons, thus far do they give evi- dence in support of Jordan's Law. Where, however, we find such cases in support of the law, there are many others where the more closely related species occupy the same territory and are contrary to it. We may cite Phenacogaster megalostictus and rnicrostictus; Chasmocranus longior and brevior; Doras hancocki and cataphracta; Curi- matus microcephalics and spilurus; Characidium pellucidum, C. pteroides and C. cate- natum; Pozcilobrycon trifasciatus, P. erythrurus and P. ocellatus. It may be said in favor of Jordan's Law that in some of the cases in which we are certain that the forms are immediately related {Crenicichla and JEquidens) its postulates are undoubtedly sustained. Mutation. Two instances of apparent mutation may be mentioned. 1. Moenkhausia profunda differs from Fowlerina orbicularis apparently only in the generic characters. I say apparently, for it is to be borne in mind that in comparing "specimens" of fishes we are put on a level with the ornithologist who might be compelled to compare plucked specimens of the Orchard and Baltimore Orioles. It is doubtful whether he would draw correct conclusions as to their song and nesting-habits from such data. Similarly Moenkhausia profunda and Fowlerina orbicularis may differ from each other much more than the characters of speci- mens seem to indicate. Granting that the two species are as similar as they appear to be it is quite probable that the former is a recent mutant of the latter. Fowlerina is exceedingly abundant in the Essequibo. Moenkhausia profunda is known by but two specimens from the northwestern coast. (Plate XLVI, figs. 1 and 2.) 2. The second instance is complex and indicates simultaneous, orthogenic mutation in a number of not closely related species of the same family. 108 MEMOIRS OF THE CARNEGIE MUSEUM In different species of the Nannostomatince, Tetragonopterinas and the Chara- cince the pectoral has at times been arrested in its normal development. It has retained the embryonic form. Inasmuch as I had never seen a similar condition in other fishes the first specimen was described as a distinct genus. Mrs. Ellis at once found another. Mrs. Ellis then made a systematic search for similar mu- tants in the large collections made, not only by myself, but also among the Tetra- gonopterince collected by Mr. Haseman in various parts of Brazil. She reports : The specimens with an archaic pectoral were forty-four in number, belonging to eight different species of five different genera and three different subfamilies, as follows: Archicheir minutus Eigenmann, one specimen, 26 mm., Christianburg (Car- negie Museum Catalog of Fishes No. 1186). Dermatocheir catablepta Durbin, one specimen, 18 mm., Tumatumari, British Guiana (C. M. Cat. No. 1198). Hyphessobrycon parvella Mrs. Ellis, one specimen, 18 mm., Riberao, Azula Lagoa (C. M. Cat. No. 2930); one specimen, 12 mm., Rio Tiete (C. M. Cat. No. 2931), with archaic pectorals. Thirteen specimens, 13 to 29 mm. (No. 2932), Aqua Quente, Alagoinhas, Rio Catu, Queimadas, Rio Itapicuru, have normal pectorals. Hyphessobrycon lutkeni (Boulenger), four specimens, 12, 16, 16, and 19 mm. (without the caudal), Jacarehy (C. M. Cat. No. 2933), have archaic pectorals. Three specimens, 19, 19, and 20 mm. (without the caudal), Jacarehy, (No. 2937), have normal pectorals. The pectorals are also normal in thirteen hundred and seventy-four specimens, 23 to 65 mm., from Rio Grande do Sul, Paraguay basin, and the coastwise streams of southern Brazil. Hyphessobrycon bifasciatus Mrs. Ellis, twenty-seven specimens, 15-25 mm., Cacequy (C. M. Cat. No. 2935); and five specimens, 18 to 27 mm., Munez Freire (C. M. Cat. No. 2936), have archaic pectorals, while one hundred and eighty-four specimens, 25 and 29 to 47 mm., from Campos, Sao Joao da Barra, Xiririca, Porto Alegre, Morretes, Munez Freire, Lagoa Feia, and Cacequy have pectorals of the adult type. Hasemania maxillaris Mrs. Ellis, one specimen, 29 mm., from Porto Uniao, Rio Iguassu (C. M. Cat. No. 2937). Hasemania bilineatus Mrs. Ellis, four specimens, 14 to 16 mm., from Mogy das Cruzes, Rio Tiete (C. M. Cat. No. 2938) have archaic fins; five specimens, 20 to 41 mm. (2939), from Mogy das Cruzes, Alto da Serra, Sao Paulo, have normal pectorals. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 109 Charax gibbosus (Linmeus), four specimens, 29 to 33 mm., Botanic Garden, near Georgetown, British Guiana (C. M. Cat. No. 2940). Asiphonichthys hemigrammus Eigenmann, two specimens, 27 mm., Gluck Island (C. M. Cat. No. 2138, 1. U. Cat. No. 12044); one specimen, 25 mm. (without the caudal), Gluck Island (I. U. Cat. No. 12645), have archaic pectorals. One specimen, 33 mm. (with the caudal), Gluck Island (C. M. Cat. No. 2941), has a normal pectoral. In contrast to the species of Hyphessobrycon enumerated above we have one thousand nine hundred and ninety-nine specimens of Hyphessobrycon gracilis be- tween 19-30 mm., all with normal fins. Whether any of the forms here enumerated are permanent, or simply abnormal individuals, doomed to elimination, the interesting fact remains that a very unusual mutation is arising simultaneously in widely different members of the same family. This is of special interest in the case of the characins, for it seems quite certain that other characters have independently made their appearance in different subfamilies. A pair of conical teeth in the lower jaw belongs to this category. They are found in widely distinct subfamilies both in Africa and America. A triple series of teeth in the maxillary is another character that has been several times and inde- pendently derived from a double series, and the double series from a single series. The steps in this process have been preserved in part. The process in the latter cases is again orthogenic, but it is doubtful whether progress is made by mutation. Changes of Color and Shape with Age. Marked changes in color take place with growth in Cichla. Only melanism is taken into account. In the smaller specimens there is a series of three conspicuous spots on the sides (Plate LXIX, fig. 1). These become ocellated a little later, at the same time that vertical dark bands appear (fig. 2). With growth the two anterior spots disappear entirely, the last one being retained as a caudal ocellus (fig. 3). The sides are marked at this stage with three heavier, and a number of fainter, cross-bands. With full growth the bands become concentrated into ir- regular spots surrounded by lighter (fig. 4). Changes in the color of Crenicichla alta are sufficiently described in the diagnosis of the species. The changes in color in Myloplus rhomboidalis and Myleus pacu are indicated in Plates LVIII and LIX. In the figure of the adult M . pacu irregular black blotches, which are frequently present, are not indicated. In both these species there is a notable change in shape with age. The young are much slenderer, depth 2.6 in 110 MEMOIRS OF THE CARNEGIE MUSEUM the length in the young M. pacu, as compared with 1.6 in the adult; 2.4 in the length in the young Myloplus rhomboidalis as compared with 1.5 in the adult. Similar changes in shape take place in Serrasalmo rhombeus and probably in all of the other characins with an extremely deep body In this connection figures 2 and 3 of Plate LII should also be examined; they represent the change in shape in another type of characins. On the Utilization of Different Structures to Obtain the Same End. (Plates V-IX.) In a number of Siluridce, especially the marine Ariince, the nuchal region is heavily armored. Granting that this armor is an adaptation, it is of interest that it is formed of either the occipital process, the so-called dorsal plate, or combinations of the two in varying proportions. In Sciadeichthys flavescens (Plate V, fig. 1) the entire armature is furnished by the dorsal plate. In proops (Plate V, fig. 2) the occipital process takes a slight part, the major portion of the armature being formed by the dorsal p'ate. In parkeri (Plate V, fig. 3) the occipital process is still more prominent. In Sciadeichthys emphysetus (Plate IX, fig. 1) the dorsal plate and occipital process are of about equal value. In the two species of Selenaspis, S. herzbergii and S. passany (Plate VII) the occipital process is slightly larger than the dorsal plate. In Arius spixii and Hex- anematichthys rugispinis (Plate IX) the dorsal plate forms a negligible portion of the armature. It is still further reduced in the three species of Notarius (Plate VIII) in which the occipital process differs progressively from the linear form in N. stricticassis to the leaf-shaped in N. grandicassis. In the latter three species there is very great variation. The series outlined above is not a genetic series. It is simply a series showing the inversely proportioned coordinated development of the structures to accomplish the arming of the nuchal region. They illustrate the dictum of Weismann that "adaptations arise if they are at. all possible" and that the structures utilized in bringing out the adaptation are not material. Natural selection would be satisfied with the arming of the nuchal region regardless of the structure that did the arming. 64 A similar case is furnished by the genera Charax, Asiphonichthys, Acanthocha- rax and Heterocharax. In these a spinous armor for the lower gill-region, or breast, may be considered an adaptation. In the first two genera the clavicle furnishes the spine. Its lower edge rises blade-like and ends in a spine anteriorly and pos- 64 Dr. C. R. Eastman calls my attention to the fact that a similar armoring has developed in Coccosteus, Dinichthys, and other extinct fishes. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 111 teriorly. In the latter two genera the clavicle is free from spines, but the preopercle is provided with a spine at its lower angle. Granted that the presence of a spine was advantageous at this point, it is evident that natural selection might pre- serve two different lines of individuals which varied in this direction. Mutilation. Mutilated specimens which would not have survived, if there were the close individual selection and elimination which we have imagined, are not rare. A Potamotrygon evidently had a piece bitten out of the side by a " Perai." An Astyanax abramoides had a mutilated snout of long standing, although it was the largest specimen of the species obtained. A Loricaria brunnea had lost the posterior portion of the body and had re- generated a caudal fin. A Trachycorystes obscurus had suffered a similar injury and regenerated the caudal. Several specimens of Char ax gibbosus and Eigen- mannia virescens had suffered a variety of injuries. A detailed study of the mutila- tions in the gymnotid eels will be published in a monograph on the Gymnotidoe, which has been prepared by Dr. Max Ellis, and which will shortly appear in the Memoirs of the Carnegie Museum. The Characins. This family will be fully dealt with in another place, but a few of the new points brought out by the present expedition may be mentioned. More than half of the strictly fresh-water fauna of Guiana is composed of characins. In all there are one hundred and fifty-one species. In all localities, except Lama Stop-Off, the Georgetown trenches, Waratuk, the Packeoo Falls, and the Rupununi and Aruataima, they form more than half the number of species. It was found that some of the small species (Plate XXXIX, figs. 3-7) burrow on the sand-banks like some of the North American darters, and that some species actually fly for a short distance after scooting along the surface of the water for a much longer distance. The characins are thus found in the bottom below the river and in the air above it. It has long been known that different members are adapted to, and live in, all the regions between the two. The organs of flight, which have been the subject of study by one of my students, consist of the enlarged pectorals and the hypertrophy of the attached muscles. To furnish points of origin for the muscles the coracoids are enormously expanded and united below into a sternum. The entire anatomy of the anterior half of the fish has been modified to become adjusted to this peculiar structure. (Plate LV.) 112 MEMOIRS OF THE CARNEGIE MUSEUM I have elsewhere called attention to the fact that the characins parallel most other fishes. For this reason they were distributed by the earlier naturalists among various families of fishes. Pcecilocharax, a small species about the Kaieteur, looked so nearly like a Pceciliid that only careful scrutiny after my arrival at home enabled me to place it where it belongs. (Plate XLIV.) Not only do they parallel other species of fishes, but they parallel each other. Moenkhausia lepidurus "mimics " Creatochanes caudomaculatus, which frequently lives with it. The young of Ano- stomus anostomus (Plate XLI, fig. 1) so closely resemble the young of Leporinus arcus (Plate XLII, fig. 2) that a most skilled ichthyologist pronounced them the same when the snouts of the two specimens were covered. Some idea of the diversity in this family may be gathered by glancing at the plates from Numbers XXXIII to LXI. The most interesting discovery was probably Bivibranchia, a characin with a protractile upper jaw (Plate XXXIII) and of nearly equal interest was the dis- covery of the sexually dimorphic Pcecilocharax (Plate XLIV, figs. 1 and 2). Of more general interest was the finding of the young of the "Pacu" in the Wara- puta Cataract and elsewhere. The breeding-place of this important food-fish was not known, and it had been supposed that it bred on the overflowed land during the rainy season. Sexual Dimorphism. Color differences are as common between the sexes of fishes in Guiana as else- where, but it is not the intention to take up these. Red and yellow are frequently present. 65 Usually if yellow and red are present in a species the red replaces in the male the yellow of the female. The secondary sexual differences of the Pmciliidce are varied and great. They have often been described, but none of those hitherto described approaches the modification in the male of Tomeurus. (Plate LXV, figs. 7 and 8.) The anal fin in this species has been moved further forward and is more highly modified than in any other member of the family. In Rivulus (Plate LXIII) the female has an ocellus on the base of the upper caudal lobe. (See also Plates LXIV and LXV for other instances of sexual differences in the Poeciliidos.) Great secondary sexual characters are also found in the "Pacu," Myleus, and probably in all of the related species of Myloplus. The anal fin in the male is bilobed (Plate LIX, fig. 5), the middle rays longer than those either just in front or behind them. The anal of the female is falcate (Plate LIX, fig. 6). 65 In Hyphessobrycon and Hemigravimus the red of the caudal frequently encroaches on the sides of the body. The red markings are sometimes very abrupt, as in the caudal of Cichla ocellaris, and in the fins of Chalceus macrolepidotus. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 113 Th? difference between the male and female of Poecilocharax are indicated on Plate XLIV, figures 1 and 2. In some of the Loricariidie the male is more abundantly supplied with spines. In Pseudancistrus barbatus, for instance, the cheeks of the males are provided with long bristles. In Lithoxus lilhoides the pectoral spines are much longer in the male (Plate XXIX, fig. 4) than in the female (fig. 3) and more profusely provided with bristles. The peculiar barbels, much more profusely developed on the head of the male Ancistrus than on the female, have long been known (Plate XXV, fig. 3). New Genera and Species. The new genera described during the progress of the work are: 1. Chamaigenes. 15. Paecilobrycon. 2. Agmus. 16. Archicheir. 3. Megalonema. 17. Poecilocharax. 4. Microglanis. 18. Aphyodite. 5. Brachyglanis. 19. Dermatocheir. 6. Leptoglanis. 20. Camegiella. 7. Myoglanis. 21. Acanthocharax. 8. Chasmocranus. 22. Heterocharax. 9. Tympanopleura. 23. Gymnorhamphichthys. 10. Lithogenes. 24. Porotergus. 11. Corymbophanes. 25. Rhinosardinia. 12. Lithoxus. 26. Tomeurus. 13. Bivibranchia. 27.- Acarichthys. 14. Tijlobronchus. 28. Soleonasus. The new species of fresh-water fishes, based on the collections made, are: 1. Bunocephalus amour us. 14. Chasmocranus brevior. 2. Bunocephalus chamaizelus. 15. Pimelodella megalops. 3. Agmus lyriformis. 16. Pimelodella macturki. 4. Megalonema platycephalum. 17. Pimelodus heteropleurus. 5. Pseudopimelodus villosus. 18. Leptodoras linnelli. 6. Pseudopimelodus albomarginatus. 19. Hemidoras microstomia. 7. Microglanis poecilus. 20. Hemidoras micropaeus. 8. Brachyglanis frenata. 21. Hemidoras leporhinus. 9. Brachyglanis 7nelas. 22. Hemidoras notospilus. 10. Brachyglanis phalacra. 23. Auchenipterus demerara. 11. Leptoglanis essequibensis. 24. Auchenipterus brevior. 12. Myoglanis potaroensis. 25. Tympanopleura piperata. 13. Chasmocranus longior. 26. ^IgeneiosMS guianensis. 114 MEMOIRS OF THE CARNEGIE MUSEUM 27. Ageneiosus marmoratus. 28. Hemicetopsis macilentus. 29. Hemicetopsis minutus. 30. Pygidium guianense. 31. Pygidium conradi. 32. Pygidium gracilior. 33. Plecostomus hemiurus. 34. Lithogenes villosus. 35. Corymbophanes andersoni. 36. Hemiancistrus braueri. 37. Pseudancistrus nigrescens. 38. Lithoxus lithoides. 39. Loricariichthys microddn. 40. Loricariichthys griseus. 41. Loricariichthys stewarti. 42. Farlowella hargreavesi. 43. Bivibranchia protractila. 44. Curimatus morowhannce. 45. Curimatus issororoensis. 46. Prochilodus maripicru. 47. Tylobronchus maculosus. 48. Paradon bijasciatus. 49. Nannostomus marginatus. 50. Nannostomus minimus. 51. Nannostomus simplex. 52. Pacilobrycon harrisoni. 53. Pcecilobrycon erythrurus. 54. Pcecilobrycon ocellatus. 55. Archichcir minutus. 56. Characidium laterale. 57. Characidium vintoni. 58. Characidium blennioides. 59. Characidium pellucid um. 60. Characidium pteroides. 61. Characidium catenatum. 62. Anosfowus plicatus. 63. Schizodontopsis laticeps. 64. Leporinus arcus. 65. Leporinus granti. 66. Leporinus alternus. 67. Poecilocharax bovallii. 68. Aphyocharax melanotus. 69. Aphyocharax erythrurus. 70. Aphyodite grammica. 71. Mcenkhausia profunda. 72. Mamkhausia brown i. 73. Mcenkhausia shidleri. 74. Pri stella aubynei. 75. Hemigrammus erythrozonus. 76. Hemigrammus rodwayi. 77. Hemigrammus iota. 78. Hemigrammus orthus. 79. Hemigrammus cylindricus. 80. Hemigrammus analis. 81. Hyphessobrycon minor. 82. Hyphessobrycon rosaceus. 83. Hyphessobrycon minimus. 84. Hyphessobrycon eos. 85. Hyphessobrycon strictus. 86. Dermatochcir catablepta. 87. Crcagrutus melanzonus. 88. Bryconamericus hyphessus. 89. Astyanax guianensis. 90. Astyanax essequibensis. 91. Astyanax mutator. ■ 92. Asmal spines spike-like, separate from each other. Air-bladders minute, one on either side of the coalesced vertebra? and surrounded by a bony capsule, in contact with the skin at a long narrow slit in the temporal plate and at the last of a series of slits below the longer one. Mouth terminal; teeth villiform. Callichthyidae, VII. 120 MEMOIRS OF THE CARNEGIE MUSEUM aaa. Derm more or less completely covered with several series of bony plates; caudal vertebra? compressed, the neural and haemal spines expanded, forming a continuous ridge above and below. Air- bladder minute, one on either side of the coalesced vertebra 1 , and surrounded by a bony capsule, in contact with the skin at a notch in the posterior margin of the temporal plate at the beginning of the lateral line. Mouth inferior, the lower lip reverted, the lips disk-like; teeth (if present) in many series, a single series erect Loricariidae, VIII. Family II. ASPREDINID.E. The "Banjamans" (banjoman). = Aspredinoidei Bleeker, Nederl. Tijclschr. Dierk., I, 1863, 117. = Aspredinina Gunther, Catalogue, V, 1864, 266. = Aspredinidce Gill, Arrangement of the Families of Fishes, 1872, 19. = Bunocephalidw Eigenmann and Eigenmann, Am. Nat., 1888, 647; Occasional Papers Cal. Acad. Sci., I, 1890, 12. Characters given in the key to the families, p. 119. Key to the Genera of Aspredinid^e. a. Tail long, with a median dorsal fold; distance of vent from tip of snout less than half its distance from base of caudal. A. 50-60. (Aspredinina.) b. A series of barbels from the corner of the mouth along the breast and anterior part of abdomen; maxillary barbel with a smaller barbel at its base. c. Tip of snouth smooth, ethmoid without hooks; a single pair of mental barbels (rarely wanting). Chamaigenes. cc. Tip of snout (nasals and ethmoid) with four broad spines; a pair of mental and a pair of post- mental barbels Aspredinichthys. bb. Breast and head not margined with tentacles. Maxillary barbel adnate. d. Maxillary barbel with a smaller barblet near its base Aspredo. dd. Maxillary without a barblet at its base Platystacus. aa. Tail short, vent nearly equidistant from tip of snout and base of caudal. Anal with 11 rays or fewer. ( BunocephalincB.) e. Dorsal well-developed, of five rays, the last of which is adnate. Barbels six. /. Head depressed, its greatest depth about half its width; coracoid and coracoid processes not prominent; no prominent tentacles about head; humeral process extending past cora- coid process Bunocephalus. //. Head about as deep as wide; coracoid and coracoid processes prominent; head externally tuberculate; coracoid process extending past humeral process Agmus. Subfamily Aspredinina. Chamaigenes 3 gen. nov. Chamaigenes Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 380 (name only). » xatmiyevriSj earth-born. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 121 Type, Aspredo filamentosus Cuvier and Valenciennes. Distinguished by the marginal tentacles of the head and breast, and by the absence of prominent hooks on the ethmoid. Nasals each with a hook, which is partially concealed. A single pair of mental barbels. 3. Chamaigenes filamentosus (Cuvier and Valenciennes). Aspredo filamentosus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 437, pi. 450 (Cayenne). — Gunther, Catalogue, V, 1864, 270 (Demerara).— Eigenmann and Eigenmann, Proc. U. S. Nat. Mus., XIV, 1891, 26. Platystacus filamentosus Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), II, 1889, 50; Occasional Papers Cal. Acad. Sci., I, 1890, 24. Chamaigenes filamentosus Eigenmanx, Eepts. Princeton Univ. Exp. Patagonia, III, 1910, 380. Several specimens; largest about 215 mm. Georgetown. (C. M. Cat. No. 1553a-e; I. U. Cat. No. 11971.) Distance from tip of snout to dorsal plate about 4.5 in the length to base of caudal; depth of head about half its width, which is almost equal to its length to upper angle of gill-openings. A. 51-56; D. 5; last ray adnate for half its length, the first ray extremely elongate, considerably longer than its distance from the tip of the snout in the adult. Eye small, about one-fourth the interorbital; maxillary barbel reaching gill- opening or pectoral, scarcely adnate; patches of teeth of the two premaxillaries forming a single band; width of mouth about one-third the width of the head; a small barbel on the maxillary barbel opposite the corner of the mouth; a series of barbels directly back of it past the base of the pectoral; a pair of post-mental barbels, corresponding in position to the post-mentals in tibicen; no mental barbels. A round pectoral pore under the tip of the humeral process; pectoral spine with hooks on its inner margin, increasing in size toward the tip; those of the outer margin small, directed at right angles to the spine. Chocolate, without blotches; lower surface plain; a dark streak back from eye. Anterior part of anal hyaline; ventrals dusky, other fins blackish. Aspredinichthys Bleeker. Aspredinichthys Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 118 (tibicen). Type, Aspredo tibicen Temminck. Readily distinguished by the four nasal spines and the marginal tentacles of the head and breast. 122 MEMOIRS OF THE CARNEGIE MUSEUM 4. Aspredinichthys tibicen (Tcmminck). Aspredo tibicen (Temminck) Cuvier and Valenciennes Hist. Nat. Poiss., XV, 1840, 438 (Surinam). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 630 (coast of Guiana) .— Gunther, Catalogue, V, 1864, 270 (British Guiana). — Eigenmann and Eigenmann, Proc. U. S. Nat. Mus., XIV, 1891, 26. Platystacus tibicen Eigenmann and Eigenmann, Proc. Cal. Acad. ScL, (2), II, 1889, 50; Occasional Papers Cal. Acad. Sci., I, 1890, 24 (Caruca; RioMuria). Aspredinichthys tibicen Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 118. — Eigen- mann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 381. Large series of individuals up to a length of almost 215 mm. from the George- town market. Evidently the most abundant of the banjomans. (C. M. Cat. No. 1552a-o; I. U. Cat. No. 11970.) Distance from snout to dorsal plate nearly 5 in the length to the caudal; width of head about equal to its length to the upper angle of the gill-opening; depth of head about half its width; D. 5, the last ray adnate, the first much pro- longed, its length in the adult equal to its distance from the eye or longer, not much produced in the young. A. 51-58, the last ray adnate. Head pointed, width of mouth one-third width of head, snout produced be- yond the mouth for about one-third the width of the latter; eye rather large, almost half the width of the interorbital; maxillary barbels about reaching gill- openings, adnate for about one-fourth their length; maxillary barbel with a barbel opposite the corner of the mouth; a series of barbels behind it on the lower surface of the head and breast to about the base of the pectoral. A pair of mental barbels nearly equidistant from each other and from the mouth, not reaching the post-mental barbels, which are nearly twice as far apart; lower surface of head warty; a round pectoral pore below the tip of the humeral process; inner margin of pectorals with spines increasing toward the tip, those of the outer margin pointed toward its tip. Slaty, irregularly marked with squarish darker blotches on the back; ventrals and anterior part of anal hyaline, other fins slate or blue-black. I examined two specimens of this species in the Museum at Leiden from Surinam. Aspredo Bleeker. Aspredo (ex Linnseus, Mus. Adolphi Fred., 1754, 73) Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 117 {batrachiis) . Type, Aspredo batrachus Gronow = Aspredo aspredo (Linnseus). Distinguished by the absence of marginal tentacles on the head and breast and by having a basal barblet on the maxillary. Snout without hooks. Two species, distinguished thus: eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 123 Key to the Species of Aspredo. a. Head to tip of nuchal plate more than 4 in the length to base of caudal aspredo. na. Head to tip of nuchal plate 4 or less in the length to the caudal sicuephorus. 5. Aspredo aspredo (Linnaeus). Silurus aspredo Linn^us, Syst. Nat., ed. 10., 1758, 304; ed. 12, 1766, 502.— Bonna- terre, Tabl. Enc. Ichth., 178S, 150.— Lacepede, Hist. Nat. Poiss., V, 1803, 78. Platystacus aspredo Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), II, 1889, 50 (Para; Arary); Occasional Papers Cal. Acad. Sci., I, 1890, 23. Aspredo aspredo Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 380; Proc. U. S. Nat. Mus., XIV, 1891, 26. Platystacus Icevis Bloch, Ausl. Fische, VIII, 1794, 58. — Bloch and Schneider, Syst. Ichth., 1801, 373. Aspredo Icevis Muller and Troschel, in Schomburgk, Reisen, III, 1848, 630 (Waini). Aspredo batrachus (ex Linnseus, Mus. Adolphi Fred., 1754, 73) Gronow, Cat. Fish, ed. Gray, 1854, 137. — Bleeker, " Silures de Suriname," 1864, 93 (Surinam). — Gunther, Catalogue, V, 1864, 268, part (British Guiana; Cayenne; Surinam). No specimens of this species were secured. It has been recorded from Guiana by Muller and Troschel and by Gunther. Head greatly depressed, spatulate; interorbital width 3.5 in the distance from base of pectoral to tip of snout. Mouth broad, the snout projecting almost its entire length; each jaw with two patches of small teeth. Maxillary barbel reach- ing to base of pectoral, an accessory barbel in front; mental barbel placed near the lip, extending to the post-mental, which is equal to the interorbital. Coracoid processes slightly diverging backward, the length equal to the space between them. A minute pectoral pore at the extremity of the coracoid process. Humeral process a little longer than the coracoid process and overlapping it. Distance of dorsal fin from tip of snout 3.8 in the length; first dorsal ray rarely prolonged. Outermost caudal rays slightly produced. Pectoral spine as in A. tibicen. Dorsal surface uniform purplish brown, the ventral surface plain light, shading into light purple or pinkish; dorsal fin dusky, usually with a dusky median stripe on the interradial membrane; caudal dark, except the outermost rays; pec- torals smutty, ventrals usually white, the posterior half sometimes dusky; anal fin white anteriorly, becoming dusky backward; maxillary barbels more or less dusky. Greatest width before pectoral fins 5.66-6 in the length; A. 51-55. In the specimens recorded by Cuvier and Valenciennes as Icevis the head is 4 or less than 4 in the length to the tip of the caudal. 124 MEMOIRS OF THE CARNEGIE MUSEUM 6. Aspredo sicuephorus Cuvier and Valenciennes. (Plate I, fig. 1.) Aspredo sicuephorus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 439 (Mana). — Gunther, Catalogue, V, 1864, 269 (copied). — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 380. Platystacus sicuephorus Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), II, 1889, 50; Occasional Papers Cal. Acad. Sci., I, 1890, 24 (Curuca; Rio Muria). Aspredo Icevis Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 431 (Guiana). Aspredo batrachus Gunther, Catalogue, V, 1864, 269, part. Abundant in the brackish water about Georgetown. The largest of the Aspredinidse secured. Many specimens, the largest 340 mm. No young were secured. (C. M. Cat. No. 1551a-e; I. U. Cat. No. 11969.) Distance from snout to predorsal plate 4 to tip or to base of caudal; width of head equal to its length to upper angle of gill-opening, its depth little more than a fourth of its width; D. 5, the last ray adnate. A. 54-57, the last ray adnate. Eye minute; maxillary barbel adnate for nearly half its length, reaching the pectoral spine, a small barbel near its base reaching the mental barbel, the latter to the post-mental, the three in a nearly straight line, the mental barbel just behind the angle of the mouth; patches of teeth on the two premaxillaries separate in the middle; mouth broad, half the width of the head; the snout projecting beyond the mouth for half the width of the latter; first dorsal ray more or less prolonged; pectoral spine heavy, with nearly straight teeth on its inner margin, teeth on its outer margin smaller, pointed toward the tip of the spine; no axillary pore. Uniform light brown to slate above, ventral and anterior part of anal hyaline; anal and caudal dark. I have also examined the specimens in the Leiden and the British Museums. It is very probable that aspredo and sicuephorus are identical. Giinther's specimens "e" and three of " a-d" are sicuephorus, while one has the head more than 4 in the length and represents aspredo. Platystacus Bloch. Platystacus Bloch, Ausl. Fische, VIII, 1794, 52 (sp.). — Bleeker, Nederl. Tijdschr- Dierk., I, 1863, 118. Type, Platystacus cotylephorus Bloch. Sternal and abdominal region without tentacles; maxillary barbel simple, without tributary barbels at its base; snout without spines. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 125 7. Platystacus cotylephorus Bloch. Platystacus cotylephorus Bloch, Ausl. Fische, VIII, 1794, 54, pi. 372. — Bloch and Schneider, Syst. Ichth., 1801, 372. — Bleeker, " Siluresde Suriname," 1864, 95 (Surinam). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), II, 1889, 50 (Vigia; Para; Tajapuru); Occasional Papers Cal. Acad. Sci., I, 1890, 21 (Vigia; Para; Tajapuru). — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 300. Silurus cotylephorus Lacepede, Hist. Nat. Poiss., V, 1803, 78. Aspredo cotylephorus Gunther, Catalogue, V, 1864, 269 (Surinam). Silurus hexadayctlus Lacepede, Hist. Nat. Poiss., V, 1803, 82. Aspredo sex-cirrhis Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840,441. Aspredo spectrum Gronow, Cat. Fish, ed. Gray, 1854, 137. I did not observe this species myself, but several specimens were taken by Mr. Ellis. A. 53-57; greatest width, before pectoral spine, 5.5-6.5 in the length; length of head to gill-opening 9 in the length, with the caudal fin. Head greatly depressed, narrowed forward; interorbital width 3.5 in the dis- tance from base of pectoral to tip of snout; width of mouth equals the interorbital plus the orbits, the snout projecting half its length. Premaxillaries with two rhom- boidal patches of villiform teeth, lower jaw with wider patches of similar teeth. Maxillary barbels reaching to gill-opening, mental not to post-mental barbels, which are as long as or a little longer than the width of the mouth. Coracoid processes slightly diverging backward, the length of the processes 1.25 in the distance be- tween their bases. Humeral processes co-extensive with but not overlapping the coracoid processes. Pectoral pore below the tip of humeral process. Distance of dorsal fin from tip of snout 3.5-3.75 in the length; first dorsal ray scarcely prolonged. Outermost caudal rays prolonged. Four rows of papillae on the sides. Dark brown, blotched with lighter. Subfamily Bunocephalin^e. Bunocephalus Kner. Bunocephalus Kner, SB. Akad. Wiss. Wien, XVII, 1855, 95. — Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 118 (verrucosus). Type, Platystacus verrucosus Bloch. Tail short; dorsal well-developed; barbels 6; head depressed, without prom- inent knobs. Key to the Species of Bunocephalus. a. A. 6; distance from snout to dorsal 2.75-3 in the length with the eaudal ; width 3.5 in the length gronovii. 126 MEMOIRS OF THE CARNEGIE MUSEUM aa. A. 7; distance from snout to dorsal considerably more than 3 in the length with caudal; width more than 3 in the length; general color dark chocolate amaurus. aaa. A. 8 or 9, rarely 7; distance from snout to dorsal more than 3 in the length with the caudal; width less than 3 in the length; general appearance sand-like chamaizelus. 8. Bunocephalus gronovii Bleeker. (Plate I, figs. 2-2a.) Aspredo verrucosa Gronow, Cat. Fish, ed. Gray, 1854, 137 (not of Bloch). Bunocephalus verrucosus Kner., SB. Akad. Wiss. Wien., XVIII, 1855, 96 (Barra do Rio Negro). Bunocephalus gronovii Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 329 (based on Gronow, Mus. Ichth., II, 1756, 5, No. 153, pi. 5, fig.3). — Gunther, Cata- logue, V, 1864, 266 (Demerara). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), II, 1889, 48; Occasional Papers Cal. Acad. Sci., I, 1890, 17 — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 380. The following description by Gunther is based on a specimen (figs. 2-2a), 102 mm. long collected by Hancock. The distance between the end of the snout and the origin of the dorsal fin is one-third, or nearly one-third, of the total length (with the caudal fin) ; the length of the head to the gill-opening is rather more than one-seventh. Upper jaw but little prominent. The maxillary barbel extends to the third fifth of the pectoral; the dorsal is inserted nearer to the end of the snout than to the root of the caudal. Hind portion of the tail as high as broad. Brown ish, clouded and spotted with darker. Head to gill-opening 5 in the length; greatest width 3.5; distance of dorsal from tip of snout 2.25; caudal 4; D. 5; A. 6. 9. Bunocephalus amaurus sp. nov. (Plate I, fig. 2.) Bunocephalus amaurus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 380 (name only). Type, 69 mm. Konawaruk. (Carnegie Museum Catalog of Fishes No. 1555.) Cotype, 53 mm. Konawaruk. (I. U. Cat. No. 11973.) Head to gill-opening 5.66 in the length; greatest width 3.5; distance of dorsal from tip of snout 2.33 in the length; caudal nearly 5. D. 5; A. 7; depth of head 1.5 in distance from pectoral to tip of snout; length of coracoid process about equal to the distance between their bases; anterior margin of coracoid with an externally perceptible crest. Head verrucose but without knobs; nuchal crest prominent, a notch between it and the knobbed dorsal plate. Maxillary barbels reaching pectoral; pectorals not quite to ventrals; humeral EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 127 process reaching last fourth of the pectoral spine; half of the last dorsal ray adnate; caudal peduncle terete; everywhere tuberculate, nine conspicuous series of tubercles on the tail. Dark chocolate, a light saddle between the dorsal and nuchal crest; back be- hind the dorsal with lighter blotches; dorsal, caudal, and anal fins black, their margins white, first rays of dorsal and anal and outer rays of caudal banded; ventrals mottled, margined with white, darkest just below the light edge; pectorals black, last rays with light mottlings; barbels banded. 10. Bunocephalus chamaizelus sp. nov. (Plate II, fig. 1.) Bunocephalus chamaizelus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 380 (name only). Type, 36 mm. Erukin. (Carnegie Mu- seum Catalog of Fishes No. 1556.) Cotypes, five specimens, 24-44 mm. Erukin. (C. M. Cat. No. 1557a and 6; I. U. Cat. No. 11974.) Cotypes, seven specimens, 34-39 mm and b; I. U. Cat. No. 11975.) Cotypes, six specimens, 30-37 mm. Gluck Island and b; I. U. Cat. No. 11976.) Head to gill-opening 5 in the length; greatest width 3.16; distance of dorsal from snout 2.36 in the length; caudal 4.4-6; D. 5; A. 8 or 9, rarely 7; depth of head about 1.5 in distance of pectoral from tip of snout; length of coracoid process about one-half the distance between its bases; anterior margin of coracoid without an externally visible crest; caudal peduncle terete; everywhere tuberculate, the regular rows of the tail somewhat obscured by the intervening tubercles; tubercles on the head variously developed. A faint ridge from eye to occiput; occipital with a slight knob, nuchal plate with two, dorsal plate with one, the four in a straight line. Maxillary barbel reaching pectorals; pectorals to ventrals in the smaller speci- mens; humeral process reaching third fifth of pectoral spine. Lighter or darker sand-color; a dark streak from pectoral forward to mouth Fig. 24. Bunocephalus amaurus Eigen- mann. Type. C. M. Cat. No. 1555. Tumatumari. (C. M. Cat. No. 1558a (C. M. Cat. No. 1559a 128 MEMOIRS OF THE CARNEGIE MUSEUM or chin, sometimes expanded downward behind gill-opening; a more or less evident dark streak across the nuchal plate ; a conspicuous dark band across the back and dorsal fin, extending forward below to the axil; one or two more or less evident dark blotches on the back of the tail, the anterior over the anal; fins colored like the adjoining portions of the body. Agmus 4 gen. nov. Agmus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 379. Type, Agmus scabriceps Eigenmann and Eigenmann (Plate IV). This genus differs from Bunocephalus in the great development of tubercles about the head, which at the occiput is almost as high as it is wide; the depth of the head at this point equals the distance of the pectoral from the snout. This genus reaches the limit of divergence from Aspredo. 11. Agmus lyriformis sp. nov. (Plate III.) Agmus hjriformis Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 379 (name only). Type, 56 mm. Gluck Island. (Carnegie Museum Catalog of Fishes No. 1554.) Cotype, 49 mm. Rupununi. (I. U. Cat. No. 11972.) The only other species of the genus is A. scabriceps Eigenmann and Eigenmann from Jutahy, differing from lyriformis in its converging coracoid processes, mottled fins, and extremely tubercular head. Head to end of occipital process 2.66 in its length ; depth of head at occipital pro- cess about equal to its width at gill-open- ings. D. 5; A. 6. Nuchal plate with two high knobs; dorsal plate with two small knobs, one on either side in front of the dorsal; a trans- verse ridge in front of the nuchal crest, curved backward on the sides and ending in a small knob; a knob over each eye, con- tinued backward as a low crest with a small knob near its middle and a larger one in front of the large occipital knob, the two Fig. 25. Type. C. M Agmus hjriformis Eigenmann. Cat. No. 1554. 4 dynbs, a craggy place. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 129 orbital knobs connected by a low ridge, forming a lyre-shaped figure with the crests extending backward from the orbital knobs, which approach each other be- hind; low crests extending forward from the orbital knob and meeting on the snout, each with a small knob forward of the eye; humeral process extending to the middle of the pectoral spine, a small knob on the sides above and behind its tip; coracoid process extending to the last fifth of the pectoral spine, but slightly converging behind; a transverse ridge joining the bases of the processes, the cora- coid outlined by another ridge. Maxillary barbels extending slightly beyond base of pectoral; mental barbels not reaching post-mentals, post-mentals not equal to interorbital. Body and head everywhere covered with warts, those on the tail in longitudinal series, those along the lateral line longest. Dark chocolate, mottled, some of the warts light; first rays of dorsal, anal, ventrals, and pectoral banded, the rest of these fins black, margined with white. Outer rays of the rounded caudal banded, the rest of the fin black, except the extreme tips of the rays, which are white; barbels banded with black and white. Ventrals in the cotype mottled. Family III. SILURID^. "Skin-fishes." Characters as given in the key, page 119. Key to the Genera of Silurid^e. a. Gill-membranes free from the isthmus. Or the nares approximate in Arius. b. Nares approximate, the posterior with a valve (Ariince). c. Mental barbels two; maxillary barbel broad, band-like. Pectoral spine, and usually the dorsal spine, with long band-like filaments Felichthys. cc. Mental barbels four. Palatine teeth fixed. Head and occipital process covered with very thin skin, granular. '/. Posterior nostrils connected by a membrane Selenaspis. ■Id. Posterior nostrils not connected by a membrane. e. Palatine patches of teeth with a backward projecting angle on the inner margin. /. Dorsal plate enlarged, the occipital process correspondingly reduced. Sciadeichthys. //. Dorsal plate small, crescent-shaped. Occipital process linear or leaf-shaped. Notarius. ee. Palatine patches of teeth without a backward projecting angle. Eye above the level of the mouth; gill-rakers twenty-five or fewer. h. Teeth on palate granular; no teeth on vomer; gill-membranes united, a narrow membrane free Arius. hh. Teeth on palate villiform, vomer with or without teeth Hexanematichthys. 130 MEMOIRS OF THE CARNEGIE MUSEUM bb. Nares remote; barbels six; adipose fin well-developed. i. Teeth incisor-like, in two series in the upper, in a single series in the lower jaw. First dorsal and pectoral rays not spine-like; adipose long (Callophysina?) Callophysis. ii. Teeth villiform, in bands ( Pimelodinas.) j. Vomer without teeth or with teeth in minute patches. k. Snout broad and produced, spatulate. Adipose short. First ray of the dorsal and pectoral fins articulate, not pungent, prolonged in filaments; barbels flat, not fringed Megalonema. kk. Adipose fin very long, barbels flat, ribbon-like; dorsal and pectoral spines not pungent, the dorsal spine produced Pinirampus. kkk. Snout not produced, barbels not flat. I. Orbit without a free margin. m. Dorsal and pectoral spines strong, pungent; body short. n. Dorsal surface of skull covered with thin skin; head about as broad as long; origin of ventrals behind middle of body, except sometimes in the young; sometimes a small circular fontanel at base of occipital process; frontal fontanel not extending backward behind the eye. o. Premaxillary patch of teeth with a backward projecting angle. Pseudopimelodus. oo. Occipital crest very short, not nearly reaching dorsal; premaxillary teeth without a backward projecting angle on the sides. Species with a light band across the occiput. . . Microglanis. nn. Dorsal surface of skull covered with a thick layer of muscle; occipital crest short ; origin of ventrals in front of middle. . . Brachyglanis. mm. Pectoral spines strong but short; dorsal spines small or replaced by a soft ray; skull covered with a thick layer of muscle; anal 17-21, body long, slender; occipital crest short. 5 p. Caudal rounded; an occipital, no frontal fontanel; ventrals under posterior half of dorsal Leptoglanis. pp. Caudal forked; ventrals under posterior half of dorsal; a median ridge from the frontal fontanel to the occipital process; caudal without accessory rays, lower caudal lobe longer; adipose reaching caudal Myoglanis. 6 mmm. No dorsal or pectoral spines; upper surface of head covered with the skin only; fontanel a narrow slit continued to base of occipital process; ventrals entirely under the dorsal or partly in front of it. much nearer snout than to caudal, body long and slender. q. Caudal rounded Heptapterus.' 5 Skull covered with thin skin; A. 9. Imparfinis. 6 It is probable that Heptapterus colletti Steindachner "with the ventrals under the origin of the dorsal; upper caudal lobe longer; caudal with fulcra; adipose not reaching caudal" should be the type of another genus. 7 The genus Heptapterus Bleeker, so far as known with certainty, is confined to southeastern tropical America. It consists of the following species: Heptapterus mustelinus Valenciennes, from Rio Grande do Sul and the lower La Plata, and Heptapterus eigenmanni Steindachner, from Maldonado. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 131 qq. Caudal forked. r. Anal of eighteen or more rays Acentronichthys. 8 rr. Anal of about eleven rays. Premaxillary band of teeth with a backward projecting angle Chasmocranus. II. Orbit with a free margin. Head distinctly longer than broad, s. Occipital process not reaching the dorsal plate. t. Fontanel not continued behind the eye; a small pit or fontanel some- times at the base of the occipital process Rhamdia- tt. Fontanel continued to base of occipital process. A dorsal and a pectoral spine Rhamdella.' ss. Occipital process reaching the dorsal plate; adipose fin longer than high. u. Occipital process narrow; fontanel to base of occipital process; humeral process spine-like Pimelodella. uu. Occipital process tapering; fontanel not continued behind the eyes; humeral process broad, not spine-like. v. Caudal deeply forked Pimelodus. vv. Caudal obliquely rounded Gceldiella. jj. Teeth on vomer in moderate or large patches. w. Barbels not band-like. .r. Head as broad as long. D. 1,7. Upper surface of head with vermiculating ridges. Vomerine patch of teeth large, pentagonal, contiguous to the palatine patches. Occipital process large, semicircular, not meeting the reniform dorsal plate, upper half of the adipose fin usually rayed. Phractocephalus. xx. Head longer than broad; inner surface of gill-covers usually with one or more dermal pouches. D. 1,6-8. Caudal forked or rounded. y. Inner teeth of the upper jaw slender and freely movable, the band scarcely narrowed in the middle; teeth on vomer and palatines much smaller, (the palatine teeth smaller than those on the vomer), villiform. Adipose fin equal to or longer than the anal fin; caudal deeply forked, the lobes pointed Brachyplatystoma. yy. Teeth all similar in the upper jaw and on the vomer. 2. Head narrowed forward, its width at the mouth scarcely if any more than two-thirds its greatest width; adipose fin longer than the anal; lower jaw produced beyond upper Hemisorubim. 8 The genus Acentronichthys Eigenmann and Eigenmann contains with certainty but one species, the type, Acentronichthys leptos Eigenmann and Eigenmann. Heplaplerus sxirinamcnsis Bleeker may belong to this genus. It is certain that the Heptapterus collettii placed here by Eigenmann and Eigenmann is not an Acentronichthys, for it has a strong pectoral spine. An examination of the type of Heptapterus surinamensis in Leiden gives the following: Skull covered with thin skin, fontanel a narrow slit to base of occipital process; first dorsal ray just like the second, not spinous; last ventral ray and third dorsal ray about equidistant from the tip of the snout; A. 18; caudal with numerous accessory rays, especially on the ventral side; middle rays radiating, all broken, probably rounded; teeth on each pre-maxillary in a band one-half wider than deep. It is therefore a Heptapterus if the caudal was rounded or an Acentronichthys if it was forked. 9 No dorsal spine; first pectoral ray spinous at its base only; upper caudal lobe longer. Rhamdia- glanis of southeastern Brazil. 132 MEMOIRS OF THE CARNEGIE MUSEUM zz. Head depressed, its width at the mouth about equal to its greatest width. Adipose fin shorter than the anal; caudal forked, the rays much branched. Maxillary band of teeth much narrowed in the middle; teeth on the vomer separated on the median line, closely joined to the palatine patches, together forming, on either side, a figure somewhat like a comma. Top of head osseous; the occipital process produced, meeting or nearly meeting the dorsal plate; fontanel extending from the middle of the snout to behind the eyes, and continued to the occipital as a groove Pseudoplatystoma. aa. Gill-membranes united and joined to the isthmus; nares remote, without barbels. A. A series of bony plates along the sides. (Doradince.) B. Eye small, in anterior portion of head; snout depressed; teeth well-developed; humeral process much longer and stronger than the coracoid process; anterior nares near the upper lip Doras. BB. Eye in middle or behind the middle of the head. C. Barbels all simple, teeth none Oxydoras. CC. Maxillary and mental barbels fringed. D. Teeth none; a single large pectoral pore Leptodoras. DD. Teeth in one or both jaws; numerous pectoral pores Hemidoras. A A. No plates along the sides. E. Maxillary and mental barbels present. (Auchenipterinm.) F. Mental barbels in two pairs. Adipose fin shorter than the anal fin. G. Anal short, 7-11; mouth terminal; jaws equal; caudal forked; V. 6. . . Centromochlus. GG. Anal 19-41. //. Caudal obliquely truncate or slightly emarginate; outer margin of pectoral spine serrate; mouth terminal, jaws equal or the lower longer; no promi- nent bony orbit. V. 6-10, A. 19-40 Trachycorystes. HH. Caudal deeply forked; outer margin of pectoral spine smooth or granular. V. 8; A. 17-21 Pseudauchenipterus. FF. Mental barbels arranged in a series near the symphysis; adipose fin short; dorsal fin well-developed, 1,6. Teeth villiform. V. 12-15 Auchenipterus. EE. Maxillary barbels only. (Ageneiosince.) I. Air-bladder projecting into the abdominal cavity, naked laterally, the skin over it forming a large pseudo-tympanum; snout short, about equal to the eye. Tympanopleura. //. Air-bladder minute, concealed under peritoneum and largely covered with bone; no pseudo-tympanum. Snout much longer than eye Ageneiosus. Subfamily Ariinje. Felichthys Swainson. 10 10 A specimen of Felichthys 295 mm. long marked filamentosus is in the Leiden Museum, from Ind. Occid., probably collected by Schomburgk. Dorsal from snout 3 in the length; highest anal ray 6+ in the length, base of anal 7; pectoral spine 4 in the length, dorsal spine 4.S; anal lobe with numerous black dots; A. 19; distance of adipose from dorsal 2.6 in the length. This specimen would therefore be bahiensis. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 133 Gaff-topsail Catfishes. Breviceps Swainson, Class. Fishes, Amph., and Rept., I, 1838, 328 (bayre). Felichthys Swainson, Class. Fishes Amph., and Rept., II, 1839, 305, substitute for Breviceps Swainson, preoccupied (bagre). Ailurichthys Baird and Girard, Proc. Acad. Nat. Sci. Phila., 1854, 26 (marinus). Mlurichthys Gill, Proc. Acad. Nat. Sci. Phila., 1863, 172, emendation. Mystus Gronow, Cat. Fish, ed. Gray, 1854, 165 (carolinensis), name preoccupied. Pimelodus Bleeker, " Silures de Suriname," 1864, 65 (bagre), not of Lacepede, as restricted by Gill. Nostrils close together, separated by a valve; lower jaw with only two barbels, band-like; dorsal and pectoral spines prolonged into filaments. Key to the Species of Felichthys. a. Distance of dorsal fin from tip of snout 3.33 in the length; distance of adipose from the dorsal fin 2.6 times in the length, or longer; base of anal 4.66 in the length, or longer. Anterior lobe of anal with minute dots; highest anal ray less than half the length of the base of the anal; vomerine and palatine patches of teeth separate; anal rays 32-35 bagre. aa. Distance of dorsal from tip of snout 3 in the length; distance of adipose from dorsal fin 3 in the length; highest anal ray about as high as the base of the anal, which is 5.8-6.5 in the length; anal rays 20-24 marinus. 12. Felichthys bagre (Linnaeus). Silurus bagre Linnaeus, Syst. Nat., ed. 12, I, 1766, 505. — Gmelin, Syst. Nat., I, iii, 1788, 1360. Pimelodus bagre Lacepede, Hist. Nat. Poiss., V, 1803,93, 98 (Brazil). — Bleeker, " Silures de Suriname," 1864. 66 (Surinam)". Mlurichthys bagre Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Ailurichthys bagre Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 148 (Sao Matheos; Santos; Para; Curuca; Bahia; Pernambuco; Brit. Guiana); Occasional Papers Cal. Acad. Sci., I, 1890, 33. Felichthys bagre Gill, Proc. U. S. Nat. Mus., XIII, 1891, 354. — Jordan and Ever- mann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 116. Galeichthys gronovii Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 40 (Guiana; Maracaibo; Mana; Cayenne; Bahia). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 628 (Waini and Barima). — Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 392 (Cajutuba; Para). Mlurichthys gronovii Gunther, Catalogue, V, 1864, 178 (Demerara; West Indies). Galeichthys eidouxii Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 43 (Guayaquil). 134 MEMOIRS OF THE CARNEGIE MUSEUM Mlurichthys eydouxii Jordan, Proc. U. S. Nat. Mus., VII, 1884, 40 (note on type). Bagrus macronemus Ranzani, Nov. Com. Acad. Sci. Inst. Bonon., V, 1842, 334, pi. 28 (Brazil). Mystus carolinensis Gronow, Cat. Fish, ed. Gray, 1854, 156. Pimelodus longifilis " Mus. L. B." (fide Bleeker). Eight specimens, 194-354 mm. Georgetown market. (C. M. Cat. No. 1270, 1271; I. U. Cat. No. 11773.) With the characters given in the key. 13. Felichthys marinus (Mitchill). Silurus marinus Mitchill, Trans. Lit. and Philos. Soc. N. Y., I, 1814, 433. Galeichthys marinus DeKay, Nat. Hist. New York, Zoology, IV, 1842, 178, pi. 37, fig. 118 (New York). Ailurichthys marinus Baird and Girard, Proc. Acad. Nat. Sci. Phila., 1854, 26. — Girard, U. S. and Mex. Boundary Survey, II, 1859, 31, pi. 14 (Indianola, Tex.). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 148 (Rio de Janeiro; Para; Bay of Balaxy; Mobile Bay; Pernambuco; Victoria); Occasional Papers Cal. Acad. Sci., I, 1890, 36. jElurichthys marinus Gunther, Catalogue, V, 1864, 178 (North America). — Goode, Proc. U. S. Nat. Mus., II, 1879, 119 (St. Johns River, Florida) .— Stein- dachner, " Flussfische Siidamerika's," i, 1879, 10 (Orinoco near Ciudad Boli- var). — Jordan and Gilbert, Proc. U. S. Nat. Mus., V, 1882, 246 (abun- dant from Pensacola, Florida, to Galveston, Texas), 584 (Charleston, S. C). — Jordan, Proc. U. S. Nat. Mus., VI, 1883, 106 (Key West, Florida) .—Jordan and Gilbert, Synopsis Fishes N. Am., 1883, 111 (Cape Cod to Mexico). — Jordan, Cat. Fishes N. Am., 1885, 16 (name only); Proc. U. S. Nat. Mus., IX, 1886, 26 (Beaufort, N. C). Felichthys marinus Gill, Proc. U. S. Nat. Mus., XIII, 1891, 354. — Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 118. Galeichthys parrce Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 33 (New York; Charlestown; New Orleans; Rio Janeiro). — Castelnau, Anim. Am. Sud, Poiss., 1855, 37 (Bahia). — Hyrtl, Denkschr. Akad. Wiss. Wien, XVI, 1859, 17 (vertebra? 13 + 7 + 30). Several specimens, Georgetown market. (C. M. Cat. No. 1484.) With the distinguishing characters given in the key. Sciadeichthys Bleeker. Sciades Muller and Troschel, Horae Ichth., Ill, 1849, 8 (emphysetus et pictus). eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 135 Sdadeichthys Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 62, 66 {emphysetus), not Sciadeichthys Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 99. Type, Bagrus emphysetus Muller and Troschel. This genus is distinguished from other South American Silurids by the approxi- mate nares, the enlarged dorsal plate, the backward projecting angle of the palatine patches of teeth, and the absence of an internarial membrane. The species are indiscriminately called " gillbacker." Only four species were secured, but several others found at Cayenne may occur occasionally or seasonally in Guiana. Key to the Species of Sciadeichthys. a. Teeth on the palate villiform or bluntly conical. 6. Dorsal plate pointed in front, entering the notched occipital process. c. Eye 8 in interorbital; dorsal plate as wide as the cranium behind the eye; granulations not extending forward to eyes flavescens. cc. Eye 4 in interorbital; dorsal plate much narrower than the cranium behind the eye, its sur- face striate; maxillary barbel reaching past origin of ventrals emphysetus. 66. Dorsal plate notched in front, receiving the point of the occipital process. Eye nearer snout than to preopercle; jaws subequal; occipital process broader than long, mucronate at tip. Eye 1.25 in snout proops. aa. Teeth on the palate granular parkeri. 14. Sciadeichthys flavescens (Cuvier and Valenciennes). (Plate V, fig. 1.) Bagrus flavescens Cuvier and Valenciennes, Hist. Nat. Poiss., XIV, 1839, 462 (Cayenne).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 386 (locality ?). Arius flavescens Gunther, Catalogue, V, 1864, 151 (copied). Galeichthys flavescens Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Tachisurus flavescens Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 141 (name only); Occasional Papers Cal. Acad. Sci., I, 1890, 53. Sciadeichthys flavescens Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 123. The head of a specimen of this species was secured in the Georgetown market. It measures 305 mm. to the dorsal. (C. M. Cat. No. 1265.) 15. Sciadeichthys emphysetus (Muller and Troschel). (Plate IX, fig. 1.) Bagrus (Sciades) emphysetus Muller and Troschel, Horse Ichth., Ill, 1849, 8 (Surinam). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 627 (Waini and Barima). Arius emphysetus Gunther, Catalogue, V, 1864, 150 (copied). Sciades emphysetus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 91 (name only). 13G MEMOIRS OF THE CARNEGIE MUSEUM Galeichthys emphysetus Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559. Tachisurus emphysetus, Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 144; Occasional Papers Cal. Acad. Sci., I, 1891, 53. Arius physacanthus Vaillant, Bull. Mus., d'Hist. Nat., V, 1899, 155; Nouv. Arch. Mus. d'Hist. Nat., (4), II, 1900, 128, pi. 7, fig. 1-lc (French Guiana). I have examined the type of emphysetus in the Berlin Museum and five speci- mens, 250-500 mm., Georgetown market. (C. M. Cat. No. 1487, 1735, 2486; I. U. Cat. No. 12110.) I have not been able to find any specimens in the Berlin Museum from Guiana. The type from Surinam is undoubtedly identical with the specimens described below and with Vaillant's physacanthus. Head 3.8; depth 5.75; D. I, 7; A. 17 or 18; adipose fin 9.5 in the length; eye 3 in the snout, 4 in the interorbital, 10 in the head. Head depressed forward, the snout broad; width of the head at the rictus equal to the greatest depth; dorsal plate shield-shaped, its pointed anterior end fitting into the notched occipital process, its width 1.3 in the width of the skull behind the eyes; occipital process about half as wide as broad; surface of dorsal plate pitted; occipital process with rows of nodules, the median row being most prominent; skull striate to in front of eye; maxillary barbels reaching past base of ventrals; outer mental barbel to middle of pectoral. Teeth all small, those of the roof of the mouth in four contiguous patches, the palatine patches large, subcircular. Basal half or more of the dorsal spine swollen, its length equal to the width of the head at the eyes or the opercle; anterior half of the swollen part of the dorsal spine tuberculate, the part beyond with recurved notches both in front and behind; pectoral spine about equal to the dorsal spine, the outer margin tuberculate to near the tip, where it is notched, the posterior surface serrate; caudal deeply forked, the upper lobe longer, 4 in the length; anal emarginate. Bluish gray above, white below; tips of dorsal, pectorals, ventrals and anterior part of anal dusky. 16. Sciadeichthys proops (Cuvier and Valenciennes). (Plate V, fig. 2; Plate VI, figs. 1-3.) Bagrus proops Cuvier and Valenciennes, Hist. Nat. Poiss., XIV, 1839, 457 (Antilles; Guiana; Surinam; Porto Rico). — Muller and Troschel, in Schom- burgk, Reisen, III, 1848, 627 (Waini and Barima). — Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 386 (locality ?). Arius proops Gunther, Catalogue, V, 1864, 148 (copied). Netuma proops Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 67; " Silures de Suri- name," 1864, 62, pi. 7, pi. 12, fig. 2 (Surinam). EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 137 Galeichthys proops Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Tachisurus proops Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1, 1888, 141 (Pernambuco) ; Occasional Papers Cal. Acad. Sci., I, 1890, 57. Sciadeichthys proops Jordan and Evermann, Bull. U. 8. Nat. Mus. No. 47, I, 1896, 123. Several specimens, 245-680 mm. Georgetown market. (C. M. Cat. No. 1263, 1264, 1280; I. U. Cat. No. 11774.) I have also examined the specimens in the Leiden and Berlin Museums men- tioned in the synonymy. Head 4-4.33; depth 7; D. 1,7; A. 18. Eye 1.25-1.5 in snout, 5.5-8 in head, 1.75-2.66 in the interorbital. Slender and elongate, broader than deep. Head depressed, its width 1.33 in its length, its depth 2, width at mouth 2; anterior portion of the head flat above; top of the head, humeral process, and dorsal plate coarsely granular, the granules arranged in series along the fontanel. Occipital process mucronate, broader than long; dorsal plate large, butterfly-shaped. Opercle striate; fontanel 1.5 times as long as the eye, its center in front of the middle of the eye, continued as a shallow groove. Jaws subequal; teeth all villiform, the intermaxillary band very wide and shallow; teeth on the roof of the mouth in six contiguous patches. Gill-membranes meeting in an angle, forming a broad fold across the isthmus. Gill-rakers 5 + 10. Pectoral pore large; vertical series of pores present. Distance of dorsal spine from snout 2.8 in the length; the dorsal spine granular in front, striate on the sides, weakly serrate behind, its length 1.25-1.5 in the head. Space between dorsal and adipose fins 2.75-3 in the length, the adipose fin little shorter than the dorsal, the posterior margin free. Caudal deeply forked, its upper lobe longer, 4-4.5 in the length. Anal emarginate, as high as long, 2-2.33 in the head. Ventrals 2 in the head. Pectoral spine rough or granular in front, serrate behind, 1.2-1.33 in the head. Plumbeous above, with blue lustre, white below; maxillary barbels dark, the mental barbels white; fins all more or less dotted with brown. Abundant at Georgetown. The skull is prepared and sold in souvenir stores as the "Crucifix Fish." 17. Sciadeichthys parked (Traill). (Plate V, fig. 3.) Silurus parkeri Traill, Mem. Wern. Soc, VI, 1832, 377, pi. 6, fig. 1 (muddy water of rivers of Guiana). — Schomburgk, Fishes Brit. Guiana, I, 1843, 188 (Guiana). 138 MEMOIRS OF THE CARNEGIE MUSEUM Arius parkeri Gunther, Catalogue, V, 1864, 153 (copied). Galeichthys parkeri Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Tachisurus parkeri Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1, 1888, 141 (name only); Occasional Papers Cal. Acad. Sci., I, 1890, 65. Selenaspis parkeri Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 123. Arius guadriscutis Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 111 (Cayenne; Mana).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 389 (Para). Netuma quadriscutis Bleeker, "Silures de Suriname," 1864, 59, pi. 8, pi. 13, fig. 2 (Surinam). Several specimens, 238-408 mm. Georgetown market. (C. M. Cat. No. 1272-1279; I. U. Cat. No. 11775.) I have also examined the specimen of quadriscutis mentioned by Bleeker. Head 3.6; depth 5.33-5.5; D. 1,7; A. 15-18. Body comparatively stout, the greatest width equaling the greatest depth. Head large, flatfish above; profile descending; width of head 1.17 in its length; width at the mouth 2-2.4, its depth at the base of the occipital process scarcely less than its greatest width. Top of head coarsely granular in young, the granules becoming finer and more regularly arranged in the adult; opercles smooth; humeral process with radiating lines of granules. Dorsal plate large, emarginate in front, receiving the pointed occipital process. Middle of the fontanel above the posterior margin of the eye. No skinny flap connecting the posterior nostrils. Maxillary barbels extending little beyond the base of the pectoral or shorter; mental barbels short. Teeth in jaws coarse, conical, those on the palate and vomer finely granular, the patches separate in the young, but united and covering most of roof of mouth in the adult. Gill-membranes forming a broad marginal flap across the isthmus. Gill-rakers 3^1 + 7-9. Distance of dorsal from tip of snout 2.33 in the length; the spine 1.33-1.6 in the head, granular in front, serrate behind. Distance of adipose fin from the dorsal 4-4.5 in the length, the adipose fin twice as long as high, adnate, as long as the dorsal fin. Caudal forked, the upper lobe longer, 4.33-4.5 in the length. Anal fin about as long as high, 2.15-2.25 in the head. Ventrals 1.75-2 in the head. Pectoral spine stout, 1.25-1.16 in the head, granular in front (serrate in young), striate on sides, serrate along inner margin. Yellow in life. The most abundant catfish of the Georgetown market. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 139 Selenaspis Bleeker. Selenaspis Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 62 (herzbergii) . Leptarius Gill, Proc. Acad. Nat. Sci. Phila., 1863, 170 (dowii). Type, Silurus herzbergii Bloch. This genus is distinguished from Sciadeiehthys by the internarial membrane. The character is scarcely of generic importance, especially since in the young of S. proops a slit or incipient membrane is often present between the posterior nares. Two species are found on the Atlantic coast of America, and a third one on the Pacific coast of Panama. Key to the Species of Selenaspis. a. Upper jaw distinctly longer than lower; snout 3.5 in the head; palatine teeth forming a U-shaped figure herzbergii. aa. Upper jaw distinctly shorter than the lower; snout about 7 in the head; palatine teeth forming a transverse patch without backward projecting angles passany. 18. Selenaspis herzbergi (Bloch). (Plate VII, fig. 1.) Silurus herzbergii Bloch, Ausl. Fische, VIII, 1794, 33, pi. 367 (Surinam). — Bloch and Schneider, Syst. Ichth., 1801, 383 (Surinam). Bagrus herzbergii Cuvier and Valenciennes, Hist. Nat. Poiss., XIV, 1839, 453 (Mana; Cayenne).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 386 (Para). Selenaspis herzbergii Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 63. — Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, 1, 1896, 125. — Eigenmann and Bean, Proc. IT. S. Nat, Mus., XXXI, 1907, 659 (Amazon). Arius herzbergii Gunther, Catalogue, V, 1864, T44 (British Guiana; Demerara).— Vaillant, Nouv. Arch. Mus. d'Hist. Nat,, (4), II, 1900, 124 (Mahury, French Guiana). Netuma herzbergii Bleeker, " Silures de Suriname", 1864, 61, pi. 9, pi. 13, fig. 3, (Surinam). Galeichthys herzbergii Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Tachisurus herzbergii Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 141 (Para; Curuca; Bahia); Occasional Papers Cal. Acad. Sci., I, 1890, 59. Pimelodus argenteus Lacepede, Hist. Nat. Poiss., V, 1801, 94, 102. Bagrus pemecus Cuvier and Valenciennes, Hist. Nat. Poiss., XIV, 1839, 456 (Cayenne) . Bagrus mesops Cuvier and Valenciennes, Hist. Nat, Poiss., XIV, 1839, 456.— Muller and Troschel, in Schomburgk, Reisen, III, 1848, 627 (Waini and 140 MEMOIRS OF THE CARNEGIE MUSEUM Barima).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 384, pi. 1, fig. 2 (Para). Arius mesops Gunther, Catalogue, V, 1864, 145 (copied). Galeichthys mesops Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Tachisurus mesops Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 141 (name only); Occasional Papers Cal. Acad. Sci., I, 1890, 57. Sciadeichthys mesops Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 123. Bagrus coelestinus Muller and Troschel, Horse Ichth., Ill, 1849, 7 (Guiana).— Muller and Troschel, in Schomburgk, Reisen, III, 1848, 627 (Waini and Barima) . Hexanematichthys hymenorhinus Bleeker, Versl. en Med. Akad. Wet. Amsterdam, XIV, 1862,377 (Surinam); " Silures de Suriname," 1864, 57, pi. 11, fig. 2, pi. 13, fig. 4 (Surinam). Netuma dubia Bleeker, Versl. en Med. Akad. Wet. Amsterdam, XIV, 1862, 382 (Surinam); ''Silures de Suriname," 1864, 63, pi. 15, fig. 2, pi. 13, fig. 5 (Surinam). Arius dubius Gunther, Catalogue, V, 1864, 144 (copied). Galeichthys dubius Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Tachisurus dubius Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 141 (name only). Many specimens, 230-335 mm. Georgetown market. (C. M. Cat. No. 1266- 1269; I. U. Cat. No. 11776.) Several specimens, to 197 mm. Mahaica. (C. M. Cat. No. 1733). Head 3.66-3.75; depth 5-6; D. 1,7; A. 18. Eye 1.75-2.5 in snout, 2.33-4 in interorbital, 5.5-8 in head. Elongate, the width equal to or greater than the depth. Width of the head 1.25-1.33 in its length, at the angle of the mouth about 2; its depth 1.66-1.75 in its length. Humeral process, dorsal plate, and top of head to between the eyes, granular. Occipital process wider than long, scarcely keeled. Fontanel not continued behind the eyes and without backward projecting groove. Posterior nares connected by a membrane. Barbels flattish, those of the maxillary reaching to near the ventrals (to middle of pectorals in older individuals) ; post-mental to or beyond base of pectoral, mental to gill-opening. Teeth villiform, vomerine and palatine patches of about equal size and shape in the young, a separate patch behind the palatines being developed sooner or later. The teeth of the palate differ more than in any other species. Gill-membranes meeting in an angle, forming a fold across the isthmus. Gill- rakers 6 + 10. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 141 Distance of dorsal spine from snout 2.5-2.75 in the length; dorsal and pectoral spines subterete, the outer margin roughened, the sides striate; the dorsal spine slightly serrate behind, a little shorter than the pectoral spine, 1.4-1.6 in the head; pectoral spine strongly serrate behind. Space between the dorsal and adipose fins 3.6-4 in the length. Adipose fin as long as the dorsal. Upper caudal lobe the longer, about 4 in the length. Anal as high as long, 2 in head. Ventral 1.6-2 in head. Pectoral pore minute; sides with vertical series of pores. Plumbeous above, silvery on sides, fins dusky. I have examined the types of coelestinus, hymenorhinus, dubius, and mesops. They are all identical with herzbergi. In the types of Bagrus coelestinus Miiller and Troschel in the Berlin Museum, 215-275 mm. long, from Guiana, there is a distinct but narrow streak between the posterior nostrils. The teeth are just as in a specimen of herzbergi 270 mm., col- lected by me. The mesops of Miiller and Troschel, about 320 mm., is like coelestinus, but the posterior patches of teeth are not so highly developed as in the latter. Fig. 26. Teeth of Selenaspis herzbergi (Bloch). Total length of specimens the teeth of which are fig- ured in order from left to right: 140 mm.; 240 mm.; 300 mm. , Three specimens of herzbergi, hymenorhinus, and dubius, 255, 207, and 170 mm. long, in the Leiden Museum, have the dorsal aspect of the head alike, and the maxil- lary barbels extend respectively not quite to the end of the pectoral, to the ventral, and a little beyond the origin of the ventral. The internasal membrane is most devel- oped in the largest and least in the smallest specimen. In the second in size the posterior palatine patches are not developed, corresponding with a specimen 210 mm. long collected by me and now in the Berlin Museum. In the smallest the two palatine patches are close together and the posterior patches much larger in proportion. 19. Selenaspis passany (Cuvier and Valenciennes). (Plate VII, fig. 2.) Bagrus passany Cuvier and Valenciennes, Hist. Nat. Poiss., XIV, 1839, 458 (Cayenne). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 627 (Waini and Barima). 142 MEMOIRS OF THE CARNEGIE MUSEUM Arius passany Gunther, Catalogue, V, 1864, 149 (copied). Galeichthys passany Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Tachisurus passany Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 141; Occasional Papers Cal. Acad. Sci., I, 1890, 58. I have examined the specimen about 543 mm. long collected by R. Schomburgk and now in the Berlin Museum, No. 2968. It is very similar to herzbergii but can easily be distinguished by its short snout or rather abbreviated upper jaw and the peculiar dentation of the palate. Fig. 27. Teeth of Selenaspis -passany (Cuvier and Valenciennes). Not arius Gill. Arius Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 91 (grandicassis) , not Arius Bleeker, 1858. Notarius Gill, Proc. Acad. Nat. Sci. Phila., 1863, 171 (grandicassis). Type, Arius grandicassis Cuvier and Valenciennes. Distinguished by its peculiar and variable occipital crest. Dorsal plate small ; occipital crest not wider, usually much narrower at its base than towards its middle or tip. But three species of this genus are known. They range along the coast of South America from near Bahia to Guiana. All three are abundant in the George- town market. Key to the Species of Notarius. a. Occipital process narrowly and deeply constricted at base, its margins regularly and strongly convex. grandicassis. aa. Occipital process broadly and deeply constricted, its margins describing a double curve; convex toward tip, concave toward base parmocassis. aaa. Occipital process narrowed, its margins scarcely convex stricticassis. 20. Notarius grandicassis (Cuvier and Valenciennes). (Plate VIII, fig. 1.) Arius grandicassis Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 54, pi. 427 (Guiana?).— Gunther, Catalogue, V, 1864, 153 (copied). Galeichthys grandicassis Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only) . eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 143 Tachisurus grandicassis Eigenmann and Eigenmann, Proc. Cal. Acad. ScL, (2), I, 1888, 141 (Maranhao; Bahia) ; Occasional Papers Cal. Acad. Sci., I, 1890, 65. Netuma grandicassis Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 126. Several specimens, 335-490 mm. Georgetown market. (C. M. Cat. No. 1254-1259; I. U. Cat. No. 11777.) Fig. 28. Outlines of occipital processes of various individuals of Notarius grandicassis (Cuvier and Valenciennes). Total length of specimens the processes of which are figured in order from left to right: 335 mm.; 435 mm.; 400 mm.; 490 mm. Head 3.4-3.75; depth 5.66-6; D. 1,7; A. 18. Eye 3-3.5 in snout, 4-4.5 in interorbital, 8.5-10 in head. Body cylindrical in front, tapering to a slender caudal penduncle. Head greatly depressed, profile almost straight, descending, the width of the head 1.33-1.4 in its length, its depth 1.8-2 in its length. Occipital process with a deep constriction where it joins the occiput, shaped like a clover leaflet, much longer than broad, sometimes keeled. Center of the fontanel over the middle of eye, the fontanel not continued backward as a groove. Occipital process, top of head, and humeral process granular; interorbital region with four ridges, the inner ones bounding the fontanel, the outer ones running obliquely backward from near the posterior nares. Maxillary barbels reaching to the base of the pectorals, mentals to gill-opening, post-mentals a little longer. Upper jaw projecting a diameter of the eye or more, the lip very wide, espe- cially in front, making the nose pointed; teeth of both jaws rather large, those on the palate somewhat smaller; the depth of the premaxillary band 7-9 in its width; the mandibulary band very shallow; vomerine teeth present in adult ; pala- tine patches triangular, produced backward; four patches in adult, the two additional ones placed just back of the two triangular ones in front. Gill-membranes meeting in an angle, forming a fold across the isthmus. Gill- rakers 6 + 10. Distance of dorsal spine from snout 2.33-2.5 in the length. Distance of adipose fin from the dorsal 3.6-4 in the length; adipose fin at least as long as the 144 MEMOIRS OF THE CARNEGIE MUSEUM dorsal fin, adnate. Caudal fin forked, the upper lobe longer, about 5 in the length. Anal fin longer than high. Ventrals small. Pectoral pore large. Light brown above, somewhat smutty below. 21. Notarius parmocassis (Cuvier and Valenciennes). (Plate VIII, fig. 2.) Arius parmocassis Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 57 (Bahia). — Gunther, Catalogue, V, 1864, 154 (copied). Tachisurus grandicassis parmocassis Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 141 (Bahia; San Matheos; Maranhao); Occasional Papers Cal. Acad. Sci., I, 1890, 68. Netuma stricticassis Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 126, part. Fig. 29. Outlines of occipital processes of various individuals of Notarius parmocassis (Cuvier and Valenciennes). Total length of specimens the processes of which are figured in order from left to right: 376 mm.; 418 mm.; 480 mm.; 435 mm.; 375 mm.; 232 mm. Several specimens, 232-480 mm. Georgetown market. (C. M. Cat. No. 1250-1253; I. U. Cat. No. 11779.) This species differs from grandicassis in having the occipital process separated from the occiput by a broader constriction and the process more elongate-ovate. Vomerine teeth are present in the adult. No patches of teeth are found back of the two triangular palatine patches. 22. Notarius stricticassis (Cuvier and Valenciennes). (Plate VIII, fig. 3.) Arius stricticassis Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 58 (Cayenne). — Gunther, Catalogue, V, 1864, 154 (copied). — Bleeker, " Silures de Suriname," 1864, 55, pi. 5, pi. 12, fig. 4 (Surinam). Tachisurus grandicassis stricticassis Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 141 (Bahia; Maranhao); Occasional Papers Cal. Acad. Sci., I, 1890, 68. Netuma stricticassis Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 126. Several specimens, 280-405 mm. Georgetown market. (C. M. Cat. No. 1260-1262; I. U. Cat. No. 11778.) EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 14.") I have also examined the specimens in the Leiden Museum. This species differs from Notarius parmocassis in having the occipital process still narrower, the margins being little convex. No teeth on vomer; no patches behind the two triangular palatine patches. Fig. 30. Outlines of occipital processes of three specimens of Notarius stricticassis (Cuvier and Valenciennes). Total length of specimens the processes of which are figured in order from left to right: 280 mm.; 400 mm.; 405 mm. Arius Cuvier and Valenciennes. Arius Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 53. — Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 62, 67 (arius). Type, Pimelodus arius Buchanan. Palatine patches of teeth without a backward projecting angle; teeth of the palate granular, none on the vomer; gill-membranes united to the isthmus, with a very narrow free margin across the latter. 23. Arius spixi (Agassiz). 11 (Plate IX, fig. 2.) Pimelodus abidus Spix, Selecta Gen. et Spec. Pise. Bras., 1829, pi. 7, fig. 1 (Equa- torial Brazil). Pimelodus spixii Agassiz, Selecta Gen. et Spec. Pise. Bras., 1829, 19. Arius spixii Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 76 (copied). Tachisurus spixii Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1, 1888, 146 (Maranhao; Bahia; Rio Janeiro; Para; Santos in Sao Paulo; Abrolhos); Occasional Papers Cal. Acad. Sci., I, 1890, 88. — Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 131. Arius arenatus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 106 (Cay- enne). — Gunther, Catalogue, V, 1864, 172 (copied). — Bleeker, " Silures de Suriname," 1864, 53, pi. 4, fig. 2 (Surinam). Galeichthys arenatus Jordan, Proc. U. S. Nat. Mus., IX, 1886, 558 (name only). Arius nuchalis Gunther, Catalogue, V, 1864, 171 (British Guiana). Galeichthys nuchalis Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). 11 Two specimens of fissus from Cayenne (from the Paris Museum) and one from Surinam (all in the museum at Leiden) have head 3.4 and 3.6 in the length; otherwise they are very similar to the specimens of arenatus ( = spixi) with head 3.9 in the length. The palatine patches of teeth are also smaller. 146 MEMOIRS OF THE CARNEGIE MUSEUM Tachisurus nuchalis Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 145 (name only); Occasional Papers Cal. Acad. Sci., I, 1890, 86. — Jordan and Evermann, Bull. U. S. Nat. Mus. No. 47, I, 1896, 131. There is no doubt that the specimens recorded below are identical with the nuchalis of Giinther, the types of which I have examined. Whether or not they are identical with spixi is not quite so certain. Several specimens, 128-235 mm. Georgetown market. (C. M. Cat. No. 1287a-c; I. U. Cat. No. 11781.) Several specimens, 140 mm. Mahaica. (C. M. Cat. No. 1732.) Head 3.6-4; depth 5-5.5; D. 1,7; A. 21. Eye 1.5-2 in the snout, 5-6.5 in the head, 2.75-3 in the interocular. Body compressed, especially towards the caudal fin, the depth greater than the width. Head narrowed forward, its greatest width 1 .33 in its length, its greatest depth 1.5; width at the mouth 2.2 in the length of the head. Top of the head granular in the young, the granules becoming more or less united in the adult, forming fine reticulating ridges, especially on the occipital process; occipital process longer than broad, with a blunt median ridge, the margins concave. Fontanel narrowing without interruptions, continued as a deep tapering groove to near the base of the occipital process; interorbital area with four ridges; opercles and hu- meral process rough, covered with skin; sides of the head and snout with reticulating mucous canals. Maxillary barbels varying in extent, from about the middle of the pectoral to the base of the ventrals; post-mental barbels extending to base of pectoral or to near its tip; mentals to edge of gill-membrane or to beyond base of pectoral. Upper jaw projecting; lips more or less papillose; teeth on the premaxillary and the outer ones of the mandible villiform; the inner series of the mandible and the palate granular; the palatine patches of teeth small, subovate, sometimes contiguous in front. Gill-membranes united, joined to the isthmus, a very narrow free margin across it. Gill-rakers 6 + 11-13. Pectoral pore moderate. Distance of dorsal spine from snout 2.33-2.64 in the length; the spines 1.2-1.5 in the head, serrated on its inner margin, granular or almost smooth on its outer margin. Distance of adipose from the dorsal fin 3.2-3.6 in the length, the adipose fin shorter than the dorsal fin, free posteriorly. Caudal forked, the upper lobe slightly the longer, 4-5 in the length. Anal fin scarcely longer than high, its highest ray about 2 in the head. Ventral fin 1.66-2 in the head. Pectoral spine strong, about as long as the dorsal spine, serrate on its inner margin, granular or scarcely roughened on the outer margin. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 147 Color brownish above; sides and ventral surface silvery, sometimes with brown dots. Hexanematichthys Bleeker. Hexanematichthys Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 62, 65. Type, Bagrus sondaicus Cuvier and Valenciennes. Similar to Arius, but the palatine teeth villiform. The species are found largely on the Pacific Coast. But one species has been taken in British Guiana. 24. Hexanematichthys rugispinis (Cuvier and Valenciennes). (Plate IX, fig. 3.) Arius rugispinis Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 77 (Cayenne).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 388 (Para).— Gun- ther, Catalogue, V, 1864, 156 (copied). — Vaillant, Nouv. Arch. Mus. d'Hist. Nat., (4), II, 1900, 124 (Carsevenne, French Guiana). Galeichthys rugispinis Jordan, Proc. U. S. Nat. Mus., IX, 1886, 559 (name only). Tachisurus rugispinis Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 145 (Para); Occasional Papers Cal. Acad. Sci., I, 1890, 83. Several specimens, 235-420 mm. Georgetown market. (C. M. Cat. No. 1281-1286; I. U. Cat. No. 11780.) Head 3.5-4; depth 5.5-6; D. I, 7; A. 19-21. Eye 3 in snout; 3.5 in inter- orbital, 10 in head. Slender, tail compressed. Head broad and depressed, tapering forward; width of the head 1.5-1.4 in its length, at the angle of the mouth 2.4-2.5; depth of head 1.6-2; profile rather steep. Top of head, humeral process, front and sides of the spines, and dorsal plate granular, the granulation not extending forward to above middle of cheeks. Occipital process triangular, about as long as broad, the median ridge not very prominent. Middle of the fontanel behind the eye, the posterior portion separated by a bridge, not continued backward as a groove; interorbital region with four ridges. Barbels villiform; maxillary barbel reaching to or beyond base of pectoral, post-mental to gill-opening, mental barbels much shorter. Mouth inferior, lower jaw included, lips thick; teeth villiform, the anterior ones in the jaws longer, depth of the premaxillary band 4 in its width; palatine patches of teeth one diameter of the eye apart, the width of the patches less than one diameter of the eye. Gill-membranes meeting in an angle, forming a fold across the isthmus. Gill- rakers 6 + 11. Pectoral pore none; vertical series of pores present. Distance of dorsal spine from the snout 2.33-2.5 in the length. Space between dorsal and adipose fins 4-4.66 in the length. Adipose fin adnate, as long as the anal fin. Ventrals 2.5 in the head; pectoral spine serrate behind. 148 MEMOIRS OF THE CARNEGIE MUSEUM Subfamily Callophysin^e. Callophysus Miiller and Troschel. Callophysus Miiller and Troschel, Horse Ichth., Ill, 1849, 1 (sp.). — Bleekeh, Nederl. Tijdschr. Dierk,, I, 1863, 101 (macropterus). Pimelotropis Gill, Proc. Acad. Nat. Sci. Phila., 1859, 196 (lateralis = macropterus) . Pseudocallophysus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 102 (ctenodus = macropterus) . Type, Pimelodus macropterus Lichtenstein. One series of teeth in the lower jaw, two series in the upper; dorsal and pectoral, without spines; adipose long. A single species. 25. Callophysus macropterus (Lichtenstein). Pimelodus macropterus Lichtenstein, Wiedem. Zool. Mag., I, iii, 1819, 59 (Brazil). Callophysus macropterus Muller and Troschel, in Schomburgk, Reisen, III, 1848, 629 (Essequibo); Horse Ichth., Ill, 1849, 1 (Brazil; Guiana). — Gunther, Cata- logue, V, 1864, 137 (copied).— Peters, MB. Akad. Wiss. Berlin, 1877, 470 (Apure). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 120 (Obidos; Lake Jose Assu; Cameta; Rio Negro; Santarem; Tonantins; Mana- capuru); Occasional Papers Cal. Acad. Sci., I, 1890,95. — Eigenmann and Bean, Proc. IT. S. Nat. Mus., XXXI, 1907, 659 (Amazon).— Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 382. Pimelodus ctenodus Agassiz, Gen. et Spec. Pise. Bras., 1829, 21, plate 8a (Equa- torial Brazil). — Cuvter and Valenciennes, Hist. Nat. Poiss., XV, 1840, 186 (copied). — Castelnau, Anim. Am. Sud, Poiss., 1855, 35 (Amazon). Callophysus ctenodus Muller and Troschel, Horse Ichth., Ill, 1849, 2 (Brazil). — ? Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 422 (locality?).— Gunther, Cata- logue, V, 1864, 137 (copied). Pimelodus instants Schomburgk, Fishes Brit. Guiana, I, 1841, pi. 6 (not descr.). Pimelotropis lateralis Gill, Proc. Acad. Nat. Sci. Phila., 1859, 196 (Amazon). Callophysus lateralis Gunther, Catalogue, V, 1864, 136 (copied). — Steindachner, "Ichthyologische Beitriige," v, 1876, 105 (Santarem; Tabatinga; Montalegre; Obidos; Rio Negro; Tonantins; Lake Manacapuru; Jose Assu). — Cope, Proc. Am. Philos. Soc, XVII, 1878, 676 (Peruvian Amazon). — Vaillant, Bull. Soc. Philom., (7), IV, 1880— (Calderon). No specimens were secured. I have examined the specimen collected by Schomburgk and now in Berlin. EIGKNMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 149 Head 4-4.66; depth 5-6; Br. 8-9; D. 7; A. 12; eye 3-4 in the snout, 2-3 in the interorbital, 7-8 in the head. Barbels flat, those of the maxillary reaching to the end of the adipose or beyond. Color uniform light brown, or with spots on the sides and adipose. Subfamily Pimelodin^e. Pinirampus Bleeker. Pinirampus Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 198 (pirinampu). Pirinampus Gunther, Catalogue, V, 1864, 135. Type, Pimelodus pirinampu Spix. Barbels broad, flat with a membranous border behind; first dorsal and pec- toral rays articulate, not pungent; adipose fin much longer than the anal fin; no teeth on the vomer. A single species. 26. Pinirampus pirinampu (Spix). "Mairipak." Pimelodus pirinampu Spix, Gen. et Spec. Pise. Bras., 1829, 20, pi. 8 (Brazil). — Schomburgk, Fishes Brit. Guiana, I, 1841, 183 (Guiana) .— Castelnau, Anim. Am. Sud, Poiss., 1855, 35 (Amazon). Pinirampus pirinampu Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 121 (Cameta); Occasional Papers Cal. Acad. Sci., I, 1890, 104.— Eigen- mann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 383. Pimelodus pirinampus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 196 (Brazil).— ?Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 416 (locality?). ? Pimelodus barbancho Humboldt, Rec. Obs. Zool., II, 1833, 172 (Venezuela). Pinirampus typus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 100 (name only). Pirinampus typus Gunther, Catalogue, V, 1864, 135 (copied). A single specimen of this species 770 mm. long from the falls of the Mazaruni was secured for me by Mr. Fowler, Commissioner of Lands and Mines. (C. M. Cat. No. 2490.) Head 4.66, depth about 6; D. 1,6; A. 11; eye 5.5 in the snout (3.5 in young), 10.75 in the head (7.5 in young), 5 in the interorbital (2.5 in young); adipose 2.2 in the length. Head depressed, snout parabolic, body sub terete, slightly compressed above; caudal peduncle subterete. First dorsal ray prolonged, longer than the head. Pectoral spine equal to the head in length, not prolonged; adipose dorsal beginning near middle of last dorsal ray; maxillary barbel extending to end of ventrals. Steel blue above, white below. 150 MEMOIRS OF THE CARNEGIE MUSEUM Megalonema gen. nov. Megalonema Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 383 (name only). Type, Megalonema platycephalum sp. nov. Pimelodines without teeth on vomer, the pectoral and dorsal spines prolonged into filaments, articulate, not pungent; barbels flattened, not fringed; premaxillary band of teeth without backward projecting angles; occipital process narrow, not reaching dorsal plate; caudal deeply forked; eye in middle of head. 27. Megalonema platycephalum sp. nov. (Plate X, fig. 2.) Megalonema platycephalum Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 383 (name only). Type, 173 mm. Tumatumari. (Carnegie Museum Catalog of Fishes No. 1684a.) Cotypes, four specimens, 37-65 mm. Tumatumari. (C. M. Cat. No. 1685a; I. U. Cat. No. 12060.) Head 3.66; depth 5.5, D. 1,6; A. 11; eye 2 in snout, 5 in head, equal to interorbital. Triangular in section at the dorsal, becoming oval at the caudal peduncle; profile nearly straight descending to the snout, ventral profile straight. Fig. 31. Outline of premaxillary band of teeth in Megalonema platycephalum Eigenmann. Head rounded at the occiput, flat between the eyes; occipital process spine- like, not quite reaching the dorsal plate; a pair of ridges from in front of the max- illary barbels converging to the base of the occipital process; fontanel not continued to the posterior margin of the eye; upper jaw projecting the width of the premax- illary band of teeth; snout broad, depressed, width of mouth equals half the length of the head; width of premaxillary band of teeth 7 in the length of its outer margin, its outer ends rounded. Maxillary barbel reaching tip of anal; outer mental barbel reaching tip of inner pectoral ray; the mental barbels some distance in advance of the outer or post- mentals, reaching base of pectorals; gill-membranes broadly overlapping, separate to below the angle of the mouth. Dorsal spine slender, articulate above, as long as the head, the rays decreasing EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 151 in height, the second extending past the tip of the last; space between the dorsals 6.5 in the length, base of adipose 4.25; caudal lobes slender, longer than head, the middle rays about one-third the length of the lobes; anal emarginate, the anterior rays extending past the tip of the last; ventrals not reaching anal; pectoral to ventrals. Straw-color in alcohol ; a pair of hidden spots at the base of the caudal, lower lobe dusky. In the specimen 65 mm. long the first pectoral ray is enormously prolonged, reaching past origin of the anal; first dorsal ray to the adipose; in the younger specimens they are shorter. In the young there is a faint lateral streak, an area along the base of the anal and along the middle of the back, four streaks on the snout, one from mouth up to base of maxillary barbel and then back to eye, and a pair along the middle of the snout, diverging in front and behind. Pseudopimelodus Bleeker. Pseudopimelodus Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 196, 207 (sp.); Nederl. Tijdschr. Dierk., I, 1863, 101 (raninus). Batrachoglanis Gill, Ann. Lye. Nat. Hist. N. Y., VI., 1858, 389. Type, Pimelodus raninus Cuvier and Valenciennes. Distinguished from Microglanis by the backward projecting angles in the premaxillary patches of teeth. In 1890 I identified Pimelodus bufonius with the Pimelodus zungaro of Hum- boldt. Humboldt's species is known only from his figure and description, which are said to have been taken from a fish three feet four inches long and reported to attain seven feet. It is very doubtful whether the zungaro of Eigenmann and Eigenmann from Goyaz is identical with Humboldt's species. It is also very doubtful whether Pimelodus mangurus Valenciennes from the La Plata is syn- onymous with zungaro. This elimination of synonyms leaves the genus Pseudo- pimelodus, as here understood, consisting of raninus Cuvier and Valenciennes (the type), bufonius Cuvier and Valenciennes, acanthocheira Eigenmann and Eigenmann, cottoides Boulenger, and the two species here described. Key to the Species of Pseudopimelodus. a. Occipital crest not meeting the dorsal plate; dorsal, caudal and ventrals in adult spotted, the latter two not margined with white; dorsal with a median or submedian light bar on the last rays, its margin sometimes white; no humeral spine, pectoral teeth not as wide as the spine .. villosus. aa. Occipital crest meeting the dorsal plate; a humeral spine. b. Teeth on anterior margin of the pectoral spine as wide as the spine or wider. 152 MEMOIRS OF THE CARNEGIE MUSEUM c. Nape usually without a distinct light band; dorsal, caudal, anal and ventrals edged with white. d. Maxillary barbels not quite reaching gill-opening; dorsal spine rough on both edges. albomarginatus. dd. Maxillary barbels reaching a little beyond tip of humeral process raninus. (cc. Nape with a zigzag cross-band; anal margined with white acanthochira. 66. Teeth on anterior margin of the pectoral less than half the width of the spine cottoides and bufonius.) 28. Pseudopimelodus villosus sp. nov. (Plate X, fig. 1.) ? Pimelodus raninus (not of Cuvier and Valenciennes) Muller and Troschel, in Schomburgk, Reisen, III, 1848, 628 (all rivers). Pimelodus (Pseudopimelodus) raninus Peters, MB. Akad. Wiss. Berlin, 1877, 470 (Apure). Pseudopimelodus villosus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, 148 mm. Potaro Landing. (Carnegie Museum Catalog of Fishes No. 1677.) Cotypes, two specimens, 148 and 142 mm. Potaro Landing. (I. U. Cat. No. 12056.) Fig. 32. Outline of Pseudopimelodus villosus Eigenmann. Type. C. M. Cat. No. 1677. Cotype, one specimen, 38 mm. Kumaka, Demerara. (C. M. Cat. No. 1680.) Cotype, one specimen, 52 mm. Wismar. (C. M. Cat. No. 1679.) Head 3-3.2; depth 5-5.3; D. 1,6; A. 11; eye 3 in snout, 10 in head, 5 in distance between the eyes. Everywhere covered with fine hair-like filaments, these especially abundant above the pectoral. Width of head scarcely less than its length; depth of head a little more than half its length; head flat above; anterior nostrils with very short tubes; jaws equal; width of mouth equaling length of head without snout; depth of premaxillary bands of teeth 6 in their width; a sharp backward prolongation at the outer angle; maxillary barbel reaching middle of pectoral, outer mental barbel eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 153 to the base of the last ray; inner mental barbels considerably in advance of the outer. Distance of dorsal spine from snout 2.4 in the length, the spine about half the length of the head, roughened with antrorse notches in front towards the tip, smooth behind; dorsal rays nearly all of the same height, equal to head less snout; caudal notched, the upper lobe longer, nearly equal to the length of the head; and rounded, reaching caudal; ventrals to anal; pectorals not to ventrals; pectoral spine strong, about equal to head without snout, with retrorse teeth along its entire posterior margin, its anterior margin with antrorse teeth which, in the adult, are nearly as long as the spine is wide. Marbled, dark and light brown, without distinct markings on the body, the area between the dorsals and below the posterior part of the adipose lighter in the younger specimens. Margin and a variable median band on the posterior rays of the dorsal light; caudal, anal, ventrals and pectorals profusely spotted, the margins hyaline. In the specimens mentioned by Miiller and Troschel and by Peters the premaxil- lary band of teeth have a backward projecting angle. Both are in such a bad state of preservation that I am not sure whether they are to be referred to this species or not. 29. Pseudopimelodus albomarginatus sp. nov. (Plate XI, fig. 1.) Pseudopimelodus albomarginatus Eigenmann, Repts. Princeton Univ. Exp. Pata- gonia, III, 1910, 384 (name only). Type, 98 mm. Tukeit. (Carnegie Museum Catalog of Fishes No. 1680.) Cotypes, twenty-one specimens, 33-108 mm. Tukeit. (C. M. Cat. No. 1681 a-d; I. U. Cat. No. 12057.) Cotype, one specimen, 90 mm. Waratuk. (C. M. Cat. No. 1682.) Head 2.66-2.9; depth 4.66; D. 1,6; A. 10; eye 4 in snout; 12 in head, 5 in space between the eyes. Snout much depressed, head flat above; width of head but very little less than its length, its height equal to half its width; anterior nostrils with a tube projecting slightly beyond the upper lip; jaws equal, width of mouth equals length of head without the snout; width of premaxillary bands of teeth about 6 in their anterior margin, a long pointed backward projecting angle; maxillary barbel not quite reaching gill-opening; outer mental barbel scarcely shorter. Distance between snout and dorsal spine 2.25-2.3 in the length, the spine slightly rough on the upper half of its margins; dorsal rays of nearly equal height, 154 MEMOIRS OF THE CARNEGIE MUSEUM 2 in the head; caudal notched, the upper lobe much longer, nearly or quite equal to the length of the head; anal reaching lobe of caudal, ventrals to anal; pectorals not to ventrals; pectoral spine less than half the length of the head, its inner margin with strong recurved hooks, its outer margin with still larger teeth, antrorse near the tip, retrorse near the base. Black, with markings of brown, the lighter color forming a blotch on the sides, below the space between the dorsal and a few smaller irregular ones in front of it; adipose and upper surface of caudal peduncle and a spot on or near the lower surface of the peduncle also light brown. Dorsal black, its margin and a wedge entering the posterior rays hyaline; caudal with a median spotted area near its base, the margin white, the rest of the fin black; anal, ventrals and pectorals each with a light, clouded area at the base, and a broad black band and white margin, which is very narrow on the pectoral. 30. Pseudopimelodus raninus (Cuvier and Valenciennes). Pirnelodus raninus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 157 (Mana; Rio Janeiro).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 421 (Barra do Rio Negro; Guapore; Mattogrosso). Pseudopimelodus raninus Steindachner, " Flussfische Siidamerika's," iv, 1882, 4 (Rio Huallaga). — Eigenmann and Eigenmann, Occasional Papers Cal. Acad. Sci., I, 1890, 111. Batrachoglanis raninus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 383 (name only). No specimens of this species were secured by me. The specimens mentioned by Giinther as from the Essequibo do not belong to this species. I examined the types. Three specimens, 91-110 mm. " De la Mana, Leschenault," in the Jardin des Plantes. Dorsal plate long, nearly touching the occipital process. Dorsal spine smooth; pectoral spine with strong teeth in front and behind, as wide as the spine. Maxillary barbel reaching a little bej^ond the tip of the humeral spine; a band be- tween the gill-openings, a submedian band on the last dorsal rays, tips of rays hyaline, anal similar, base and subterminal band of caudal dark. Two of the specimens are nearly uniform dark brown, one has a very distinct band between the gill-openings, and all have the entering wedge of light on the dorsal and anal and have the vertical fins margined with white. Villose as in villosus. All have backward projecting angles to the premaxillary bands of teeth. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 155 Microglanis gen. nov. 11 Type, Microglanis pcecilus sp. nov. Small Pimelodines, reaching a maximum length of 110 mm., with the head as wide as long, the skull covered by skin only; the occipital crest small; frontal fontanehnot extending much if any behind the eye, sometimes a minute occipital fontanel. Eye without a free orbital margin; dorsal and pectoral spines well- developed; premaxillary patches of teeth without backward projecting angles. The species of this genus parahybce and pulcher are variegated and marked with three more or less well-defined cross-bands, one over the nape, one behind the dorsal, and one across the caudal peduncle. 31. Microglanis pcecilus sp. nov. 12 (Plate XII, fig. 2.) Eight specimens, 22-37 mm. Below Packeoo Falls. (C. M. Cat. No. 1676 a and b; I. U. Cat. No. 12055.) The larger of the two specimens in the Carnegie Museum (a) may be considered the type. Head as wide as long, 3.4 in the length; depth 6; D. 1,6; A. 9; eye 2.5 in snout, 8 in head, 3.5 in distance between the eyes. Head depressed, snout rounded; barbels banded, the maxillary barbel reaching tip of humeral process, post-mental to base of last pectoral ray; mental barbel considerably in advance of the post-mentals; mouth wide, but its angle in front of the eye; lower jaw slightly projecting; anterior nares not tubular. Distance of dorsal from snout 2.4 in the length, the spine about twice as long as the eye; pectoral spine strong, with large teeth along the entire length of both margins, half of those of the outer margin retrorse, the other half antrorse; ventrals inserted behind the vertical from the last dorsal ray, just reaching the rounded anal; base of dorsal, adipose, and anal about equal in length; upper caudal lobe longer than head. A light, wavy band across the nape from the pectoral to pectoral; a light spot at base of dorsal spine; a light band downward and backward from behind the dorsal, joining another light band which extends upward but not to the dorsal spine; an oval light area in front of adipose; a light band across caudal peduncle; some light vermiculations about the snout; the light areas sometimes bordered by white, the extent of light and dark brown varying greatly, sometimes one and sometimes the other predominating. Pectoral light, with an oblique dark band; dorsal black, 11 filxpos, small, 7Xai'is J cat-fish. 12 The duplicates of this new species were distributed as Batrachoglanis raninus. 156 MEMOIRS OF THE CARNEGIE MUSEUM its margin and a wedge near the middle of the posterior rays hyaline; caudal spotted, sometimes a narrow subterminal black band; anal and ventrals spotted; lower surface profusely but faintly spotted. Brachyglanis 13 gen. nov. Breviglanis Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, Brachyglanis frenata sp. nov. Dorsal and pectoral spines well-developed, pungent; skull covered with a thick layer of muscle; ventrals under posterior half of dorsal; caudal forked, the lobes short and about equal; adipose fin not joined to the caudal; anal short; occipital process very short, the skull with a median ridge to near the eye, the fontanel short; eyes small, not strictly superior, without a free orbital notch; premaxillary patch of teeth without a backward projecting angle. Key to the Species of Brachyglanis. a. Maxillary barbel in the adult not reaching gill-openings when laid back, to the pectoral in the young; a dark streak from anterior nares through eyes to gill-openings; a light streak in front of dorsal. frenata. aa. Maxillary barbel extending to below the dorsal. b. Color uniform melas. bb. A broad quadrate light spot on the back in front of the dorsal phalacra. 32. Brachyglanis frenata sp. nov. Breviglanis frenata Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, 49 mm. Amatuk. (Carnegie Museum Catalog of Fishes No. 1670.) Cotypes, two specimens, 41 and 68 mm. Amatuk. (I. U. Cat. No. 12052.) Head 3.5; depth about 5; D. 1,6; A. 7 or 8; eye 2 in the snout, 7.5 in the head, 1.5 in the distance between the ej-es. Tail compressed, depth nearly uniform from caudal peduncle to occiput; head tapering to the flat snout; head slightly rounded above, its width equal to its length exclusive of the part in front of the posterior nares, its depth equal to the postorbital part of the head; angle of mouth below front margin of eye; the tubular anterior nares projecting beyond the upper lip; maxillary barbel in the smallest extending to the pectoral spine, not to gill-opening in the largest; inner mental barbels but slightly in advance of the outer. Dorsal spine equal to snout and eye or a little less, the dorsal rays of nearly 13 /Spax^s, short, y\dvis, cat-fish. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 157 equal height, 2 in head; distance from snout to dorsal 2.7 in the length; adipose fin beginning at the tip of the dorsal, its margin rounded, free posteriorly; caudal short and very broad; anal short, rounded, its origin near the vertical from the middle of the adipose; pectoral fin reaching ventral, the spine short, pointed, with recurved hooks on both margins. Light brown, spotted and marbled with darker; a clavate light area in front of the dorsal; a dark streak from the anterior nares to the upper angle of the gill- opening; bases of dorsal and caudal opaque, then abruptly hyaline, the margin of the opaque area of the caudal blackish. 33. Brachyglanis melas sp. nov. (Plate XI, fig. 2.) Breviglanis melas Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, 60 mm. Crab Falls. (Carnegie Museum Catalog of Fishes No. 1672.) Cotypes, nine specimens, 36-53 mm. Crab Falls. (C. M. Cat. No. 1673a-6; I. U. Cat. No. 12053.) Cotype, one specimen, 44 mm. Georgetown market. (C. M. Cat. No. 1674a.) Cotypes, two specimens, 32-48 mm. Warraputa. (C. M. Cat. No. 1675a; I. U. Cat. No. 12054.) Head 3.8; depth 5.5-8; D. 1,6-9; A. 8, rarely 7. Maxillary barbel extending to below the dorsal; outer mental barbel to base of last pectoral ray or to gill- opening; origin of adipose fin some distance behind the tip of the dorsal, 5-5.5 in the length. Light brown to blue-black; no markings; fins as in frenat a. 34. Brachyglanis phalacra sp. nov. (Plate XII, fig. 1, and Plate XIII, fig. 1.) Breviglanis phalacra Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, 81 mm. Amatuk. (Carnegie Museum Catalog of Fishes No. 1671.) Very similar to frenata. D. 1,8; eye 3 in snout, 11 in head, 2 in distance between the eyes; maxillary barbel reaching to below the middle of the dorsal, outer mental barbel to the gill- opening. Light brown, a dark margin to the opaque portion of the caudal; a large quad- rate yellow spot just behind the head, connected with a light bar extending down over the opercle. Otherwise as in frenata. 158 MEMOIRS OF THE CARNEGIE MUSEUM Leptoglanis 14 gen. nov. Leptoglanis Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, Leptoglanis essequibensis sp. nov. First dorsal ray soft or a minute spine; pectoral spine short but strong; origin of ventrals under posterior half of dorsal, far in advance of the middle; adipose fin long and low, continuous with caudal; caudal rounded; anal long; top of head covered with a thick layer of muscle ; an oval fontanel at the base of the occipital process; premaxillary patch of teeth subrhomboidal, with the outer posterior angle prolonged backward. 35. Leptoglanis essequibensis sp. nov. (Plate XIII, fig. 2.) Leptoglanis essequibensis Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, 156 mm. Crab Falls. (Carnegie Museum Catalog of Fishes No. 1652.) Cotypes, four specimens, 90-170 mm. Crab Falls. (I. U. Cat. No. 12041.) Cotypes, five specimens, 50-170 mm. Warraputa. (C. M. Cat. No. 1653a; I. U. Cat. No. 12042.) Cotype, one specimen, 66 mm. Konawaruk. (C. M. Cat. No. 1654a.) Head 4-4.9; depth 9.5; D. 7 or 1,6; A. 15-17; eye 2.5-3 in snout, 8-11 in head. Elongate, depth of caudal peduncle equals depth at eyes; depth behind dorsal equals depth of head, which is equal to half its width; occipital crest very short, not nearly reaching dorsal; eye superior; anterior nostrils tubular, extending be- yond upper lip; posterior nostrils nearer the eyes than to the anterior; head, and especially the snout, depressed; mouth terminal, the lips thin, plicate; premaxillary patch of teeth subrhomboidal, the outer angle much prolonged, its depth at the middle about 1.5 in its outer edge; an oval fontanel at the base of the occipital process; no frontal fontanel. Angle of the mouth considerably in advance of the eye; maxillary barbel reach- ing to middle of dorsal in the adult, farther in the young; outer mental barbels on a line with the angles of the mouth, reaching the opercle; inner mental barbels con- siderably in advance of the outer, their distance from the edge of the lower lip equal to the distance between the outer barbels; four large pores on the lower lip. Pectoral spine strong, its length equal to half the length of the head, with teeth along the basal parts of both edges, much stronger in the young; pectorals 14 XorrAs, slender, yXdvis, cat-fish. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 159 not reaching ventrals; distance from snout to dorsal 3.3-3.4 in the length, the spine evident in the young, less than twice the length of the eye, becoming soft with age; adipose dorsal long and low, connected with a forward extension of the caudal, 3.5 in the length; origin of ventrals near vertical from middle of dorsal or farther back. Ashy to steel-blue above; fins hyaline. Myoglanis 15 gen. nov. Myoglanis Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384. Type, Myoglanis potaroensis sp. nov. First ray of dorsal soft; pectoral spine very strong; skull entirely covered above by a thick layer of muscle; ventrals under posterior half of dorsal; caudal forked, the lower lobe longest ; adipose long and low, continued to the caudal, there being no caudal fulcra; a frontal fontanel, scarcely extending beyond the minute eyes; skull behind the fontanel narrow, with a median ridge; occipital crest narrow and short; eyes very smaU, directed slightly sidewise, mostly upward, no free orbital margin; the two premaxillary patches of teeth forming a crescent; anal long. It is probable that the Acentronichthys collettii of Steindachner belongs to this genus. 36. Myoglanis potaroensis sp. nov. (Plate XIV, fig. 1.) Myoglanis potaroensis Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 381 (name only). Type, about 58 mm. Creek at Tukeit. (Carnegie Museum Catalog of Fishes No. 1664.) Cotypes, sixteen specimens, the largest 60 mm. Tukeit. (C. M. Cat. No. 1665a-d; I. U. Cat, No. 12048.) Cotypes, three specimens, 99-113 mm. Potaro Landing. (C. M. Cat. No. 1666a; I. U. Cat, No. 12049.) Cotypes, three specimens, 35-40 mm. Waratuk. (C. M. Cat, No. 1667a; I. U. Cat. No. 12050.) Cotype, one specimen, 33 mm. Erukin. (C. M. Cat. No. 1668a.) Cotypes, eight specimens, 32-80 mm. Amatuk. (C. M. Cat. No. 1669a-6; I. U. Cat. No. 12051.) Head 4-4.3; depth 5.3-8; D. 6 to 8; A. 16-21; width of head equals its length without snout; depth of head 2 in its length; eye 14 in head, 3.5 in snout, 5 in space between the eyes. i6 m5s, muscle, TXdm, cat-fish. 160 MEMOIRS OF THE CARNEGIE MUSEUM Compressed behind, depth from occiput to middle of anal nearly equal; head not greatly depressed, rounded above; occipital process short, its tip a little nearer to the dorsal than to the line between the posterior margins of the eye: anterior nostril very close to the lip; lips very thin, smooth; angle of snout below the eye; maxillary barbel reaching tip of pectoral in the adult, somewhat farther in the young; inner mental barbel in advance of the line joining the outer. Distance from dorsal to snout 2.6 in the length; tip of dorsal about reaching adipose, which is about 3 in the length; caudal frequently divided to the base, the lower lobe the longer; pectoral spine strong, not quite half the length of the head, with recurved spines on both margins; anal long, the tips of some of the rays reaching caudal, its base 3.5-4 in the length. Purplish brown, lighter below, the rayed fins lighter. No distinct markings. Chasmocranus 16 gen. nov. Chasmocephalus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only, preoccupied). Type, Chasmocranus longior sp. nov. First rays of pectoral and dorsal soft ; origin of ventrals under origin of dorsal or but slightly behind it, far in advance of the middle; caudal forked; top of skull covered with skin only; fontanel narrow, extending to the base of the occipital interrupted over the eyes and again some distance behind the eyes; occipital crest short and narrow; eyes superior, without a free orbital margin; adipose fin low, not connected with the caudal; premaxillary patch of teeth subrhomboidal, its outer posterior angle extended backward; anal short. Closely allied to Heptapterus and Acentronichihys. Key to the Species op Chasmocranus. a. Adipose fin 3.5-4.5 in the length longior. aa. Adipose fin about 6 in the length brevior. 37. Chasmocranus longior sp. nov. (Plate XIV, fig. 2.) Chasmocephalus longior Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, 110 mm. Amatuk. (Carnegie Museum Catalog of Fishes No. 1655.) Cotypes, twelve specimens, 40-93 mm. Amatuk. (C. M. Cat. No. 1656a-c; I. U. Cat. No. 12043.) Cotypes, three specimens, 95-129 mm. Maripicru ? (C. M. Cat. No. 1657a; I. U. Cat. No. 12044.) 16 x«m«, a gaping, xpimr, head. EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 161 Cotype, one specimen, 89 mm. Waratuk. (C. M. Cat. No. 1658a.) Cotypes, two specimens, 67 and 68 mm. Konawaruk. (C. M. Cat. No. 1659a; I. U. Cat. No. 12046.) Cotypes, twelve specimens, 38-83 mm. Warraputa. (C. M. Cat. No. 1660 a-c; I. U. Cat. No. 12046.) One specimen, 60 mm. Warraputa. (C. M. Cat. No. 1661a.) This may prove to be distinct. Very similar to Leptoglanis. Fig. 33. Chasmocranus longior Eigenmann. Type. C. M. Cat. No. 1655. Head 4.5-5; depth 9-10; D. 7; A. usually 11 or 12, rarely 10; width of head less than its length, its depth 2 or 3 in its length; eye 6-7 in head. 2-2.66 in snout, 1-1.33 in distance between the eyes. Head depressed, tail compressed ; tip of occipital crest about equidistant from snout and dorsal; anterior nostril nearer the snout than to posterior nostril; lips thin, plicate; mouth terminal, its angle about midway between anterior and pos- terior nostrils; depth of premaxillary patch of teeth about 1.5 in the length of its outer margin; maxillary barbel extending to or a little beyond origin of pectoral; outer and inner mental barbels in a straight line, their distance from the edge of the lower lip equal to the distance between the inner barbels. Distance of dorsal from tip of snout 2.5-2.66 in the length, the rays from the second to the sixth of about equal height; ventrals and dorsal equidistant from tip of snout; pectorals rounded, equal to head less snout in length; origin of anal under origin of adipose, its base long, last but fourth ray highest or of equal height with those immediately in front of it; caudal slightly forked, the lobes rounded, of nearly equal length, considerably less than length of head. Dark brown to black; sometimes a light streak across back from gill-opening to gill-opening; a light spot at base of first dorsal ray; fins light brown or blackish with white margins. 162 MEMOIRS OF THE CARNEGIE MUSEUM 38. Chasmocranus brevior sp. nov. (Plate XV, fig. 1.) Chasmocephalus brevior Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 384 (name only). Type, 56 mm. Waratuk. (Carnegie Museum Catalog of Fishes No. 1662a.) Cotypes, two specimens, 69 mm. Waratuk. (I. U. Cat. No. 12047.) ? One specimen, 23 mm. Amatuk. (C. M. Cat. No. 1663a.) This species differs from longior in having more pointed caudal lobes, and especially in the length of the adipose fin, which is contained six times in the length. Rhamdia Bleeker. Pimelodus Lacepede, Hist. Nat. Poiss., V, 1803, — (sp.). — Cuvier, Regne Animal, II, 1817, 203 (sp.).— Gunther, Catalogue, V, 1864, 114 (sp.). Pteronotus Swainson, Class. Fishes, Amph., and Rept., II, 1839, 309 (5-tentaculatus) , preoccupied in mollusks. Rhamdia Bleeker, Icth. Arch. Ind. Prodr., I, 1858, 197, 207 (sp.); Nederl. Tijdschr. Dierk., I, 1863, 101 (queleni). Pimelonotus Gill, Ann. Lye. Nat. Hist. N. Y., VI, 1858, 391 (yilsoni). Notoglanis Gunther, Catalogue, V, 1864, 136 {multiradiatus) . Type, Pimelodus quelen Quoy and Gaimard. Eye with a free orbital margin. Teeth on vomer none or minute; occipital process short, not reaching the dorsal plate; no parietal fontanel; head covered with skin, not granular; caudal forked; barbels terete. Key to the Guiana Species of Rhamdia. a. D. 1,6 or 7. b. Premaxillary band of teeth rounded laterally. c. Maxillary barbel not or very rarely extending to middle of adipose. Space between the eyes 2.2-2.6 in the head, both caudal lobes rounded, not quite as long as head, the sixth ray of the upper lobe from the median cleft longest; eye 3 in the snout, 6.5-7 in the head, 3 in the space between the eyes; adipose dorsal 2.5 in the length quelen. cc. Maxillary barbel extending past middle of adipose; upper caudal lobe pointed; the fifth ray of the upper lobe from the median cleft longest: space between the eyes about 2.33 in the head sebae. bb. Premaxillary band of teeth with a backward projecting angle laterally. Maxillary barbel not extending to middle of adipose; space between the eyes 3 in the head; caudal cleft, both lobes rounded, the third or fourth ray from the cleft in the upper lobe longest; adipose dorsal 2-2.4 in the length holomelas. aa. D. 1,10; A. 6; maxillary barbel to tip of ventrals arekaima. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 163 39. Rhamdia quelen (Quoy and Gaimard). Pimelodus quelen Quoy and Gaimard, Voy. Uranie, Zool., 1824, pi. 49, figs. 3-4. Rhamdia quelen Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 126 (Santa Clara; Rio Mucuri; Juiz de Fora; Campos; Rio Jequitinhonha; Mendez; Rio de Janeiro; Macacos; Sao Matheos; Rio Parahyba; Canna- vierias; Rio Grande do Sul); Occasional Papers Cal. Acad. Sci., I, 1890, 127; Proc. U. S. Nat. Mus., XIV, 1891, 28.— Berg, An. Mus. Nac. Buenos Aires, IV, 1895, 133.— Eigenmann and Norris, Rev. Mus. Paulista, IV, 1900, 350 (Taubate). — Eigenmann and Kennedy, Proc. Acad. Nat. Sci. Phila., 1903, 499 (Estancia la Armonia; Asuncion; Campo Grande). — Eigenmann and Bean, Proc. U. S. Nat. Mus., XXXI, 1907, 660 (Amazon).— Eigenmann, Ann. Car- negie Mus., IV, 1907, 113 (Corumba); Repts. Princeton Univ. Exp. Pata- gonia, III, 1910, 386. Rhamdia queleni Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 101 (name only). Pimelodus queleni Gunther, Catalogue, V, 1864, 123 (Brazil). — Boulenger, Boll. Mus. Zool. Anat. Comp. Torino, XV, 1900, — (Carandasinho). Pimelodus {Rhamdia) queleni Steindachner, " Siisswasserfische Siidostlichen Brasilien," iii, 1876, 64 (Rio Parahyba, near Juiz de Fora; Campos; Rio Doce; Porto Alegre; Cannavierias; Amazon, near Para; Bahia). Heterobranchus sextentaculatus Agassiz, Selecta Gen. et Spec. Pise. Bras., 1829, 28, pi. 11 (12 in. long; locality?). Pimelodus sellonis Muller and Troschel, Horse Ichth., Ill, 1849, 2 (Brazil). ? Pimelodus bahianus Castelnau, Anim. Am. Sud, Poiss., 1855, 35, pi. 16, fig. 2 (Bahia). Pimelodus sebce Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 417, fig. 19 (Maar- bitanos) . Silurus sapipoca (ex Natterer, MS.) Kner, SB. Akad. Wiss. Wien., XXVI, 1857, 418. Pimelodus wuchereri Gunther, Catalogue, V, 1864, 123 (Bahia). Pimelodus (Rhamdia) queleni cuprea Steindachner, "Siisswasserfische Siidostlichen Brasilien," iii, 1876, 65 (Juiz de Fora). Pimelodus (Rhamdia) cuyabce Steindachner, " Siisswasserfische Siidostlichen Brasilien," iii, 1876, 75, footnote (Cuyaba). Six specimens, 112-169 mm. Gatuck Creek, Potaro Highlands. (C. M. Cat. No. 1594a-b; I. U. Cat. No. 11999.) Forty-one specimens, 71-230 mm. Aruataima. (C. M. Cat. No. 1595a-i; I. U. Cat. No. 12000.) 164 MEMOIRS OF THE CARNEGIE MUSEUM Six specimens, 107-184 mm. Yackeatonuk Fall, Potaro River. (C. M. Cat. No. 1596a-&; I. U. Cat. No. 12001.) Eight specimens, 86-168 mm. Nickaparu Creek. (C. M. Cat. No. 1604a-c; I. U. Cat, No. 12002.) Fifteen specimens, 80-219 mm. Guiana, label not legible, collected by Indians in the upper Potaro or upper Rupununi. (C. M. Cat. No. 1605a-e; I. U. Cat. No. 12005.) One specimen, 215 mm. Potaro Landing. (C. M. Cat, No. 1597a.) Thirteen specimens, 78-164 mm. Holmia, creeks. (C. M. Cat. No. 1598a-e; I. U. Cat, No. 12003.) Three specimens, 205-242 mm. Chipoo Creek, branch of the Ireng. (C. M. Cat, No. 1599a; I. U. Cat, No. 12004.) This species inhabits the plateaus of Guiana. It has not been recorded from Guiana before. It is very closely allied to sebce, from which it differs principally in its longer maxillary barbel. Head 4.16-4.5; depth 5-4.3; D. 1,6; A. 11. Head about an orbital diameter longer than broad; eye in middle of head, 6.5 in head in adult, 2.5 in interorbital ; maxillary barbel reaching middle of adipose in the young, not to middle in the old; outer or post-mental barbel not reaching beyond pectoral; dorsal spine equaling snout in adult; pectoral spine a little longer; spines shorter in the young; adipose fin half of the length without the head; caudal cleft to its base, the upper lobe narrower, both lobes rounded, equal to the head. Ashy above, blotched with darker; base of dorsal opaque at base, then hyaline, darker above; other fins uniform. 40. Rhamdia sebae (Cuvier and Valenciennes). " Rhamdia ou bagre de rio " Marcgrave, Hist, Rer. Nat. Bras., IV, 1648, 149. - Seba, Locupl. Rer. Nat, Thes. Ace. Descr., Ill, 1748, pi. 29, fig. 5. " Mystus " No. 83, Gronow, Mus. Ichth., 1, 1754, 34; No. 384, Zoophyl., 1763, 125. Pimelodus sebce Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 169 (Surinam; Cayenne; Rio Janeiro; Buenos Aires). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 628 (all Guiana rivers). — Hyrtl, Denkschr. Akad. Wiss. Wien, XVI, 1859, 16 (vertebrae 11 + 2 + 26).— Quoy and Gaimard, Voy. Uranie et Physicienne, Zool., 1824, 228, pi. 49, figs. 3 and 4.— Gunther, Catalogue, V, 1864, 119 (Demerara; British Guiana; Brazil). Pimelodus [Rhamdia) sebm Steindachner, " Siisswasserfische Siidostlichen Brasilien," iii, 1876, 68 (Demerara; Essequibo; St. Martha, mouth of eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 165 Magdalena) ; " Fisch-Fauna Magdalenen-Stromes," 1878, 17 (Magdalena) ; " Fisch-Fauna des Cauca," etc., 1880, 7 (Cauca). Rhamdia sebce Eigenmann and Eigenmann, Proc. Cal. Acad. Sci. (2), I, 1888, 126 (Tonantins; Gurupa; Rio Janeiro; Bahia; Xingu; Santa Cruz; Cudajas; Sao Matheos; Rio Doce; Serpa; Tabatinga; Goyaz; Para; Teffe; Surinam; Villa Bella); Occasional Papers Cal. Acad. Sci., I, 1890, 123; Proc. U. S. Nat. Mus., XIV, 1891, 28.— Eigenmann, Repts. Princeton Univ. Exp. Pata- gonia, III, 1910, 385. Pimelodus stegelichii Mulle and Troschel, in Schomburgk, Reisen, III, 1848, 628 (forest brooks); Horae Ichth., Ill, 1849, 3 (Surinam). — Gunther, Catalogue, V, 1864, 121 (Demerara; Surinam). Pimelodus musculus Muller and Troschel, Horse Ichth., Ill, 1849, 4 (America). Pimelodus mulleri Gunther, Catalogue, V, 1864, 119 (River Capin; Para; Surinam). Rhamdia queleni Bleeker (not of Quoy and Gaimard), " Silures de Suriname," 1864, 75 (Surinam). Twenty-nine specimens, 96-246 mm. Mud creek in Aruka River. (C. M. Cat. No. 1588a-e; I. U. Cat. No. 11194.) One specimen, 119 mm. Below Packeoo Falls, Essequibo. (C. M. Cat. No. 1589a.) Seven specimens, 118-226 mm. Creek in Barima River. (C. M. Cat. No. 1590a-6; I. U. Cat. No. 11995.) Five specimens, 150-280 mm. Creek in Mora Passage. (C. M. Cat. No. 1591o-6; I. U. Cat. No. 11996.) Fourteen specimens, 129-228 mm. Gluck Island, Rockstone. (C. M. Cat. No. 1592a-e; I. U. Cat. No. 11997.) One specimen, 153 mm. Kumaka. (C. M. Cat. No. 1593.) One specimen, 294 mm. Lama Stop-Off . (I. U. Cat, No. 11998.) One specimen, 200 mm. Georgetown, mud-flats. (C. M. Cat. No. 1601.) One specimen, 220 mm. Botanic Gardens. (C. M. Cat. No. 1602.) One specimen, 244 mm. Wismar. (C. M. Cat. No. 1603.) I have also examined the types of stegelichii in Berlin. Head 4-4.25; depth 3.3-5.16; D. 1,6; A. 10-12; eye in the middle or slightly in advance of the middle of the head, 6-6.5 in the head in the adult, 2-2.5 in the interorbital; maxillary barbel reaching beyond middle of adipose, to tip of caudal in some young; post-mental barbels to tip of pectoral or beyond origin of ventral. Dorsal spine somewhat longer than snout in adult (shorter in young) ; pectoral spine somewhat longer than snout and eye in the adult; adipose fin nearly half the 166 MEMOIRS OF THE CARNEGIE MUSEUM length from eye to caudal. Upper caudal lobe more pointed than in quelen, equal to the length of the head. Ashy to black, with darker mottlings, or plain, other- wise as in quelen. One of the types, 262 mm. long, in the Jardin des Plantes has the maxillary barbel extending to the second third of the adipose only (the specimen mentioned by Cuvier and Valenciennes from Guayaquil is another species), upper caudal lobes rounded, the fourth ray longest. 41. Rhamdia holomelas (Gunther). Pimelodus holomelas Gunther, Ann. and Mag. Nat. Hist., (3), XXII, 1863, 442 (Essequibo); Catalogue, V, 1864, 120 (Essequibo). I have examined the types in the British Museum and seven specimens, 275- 345 mm. Lama Stop-Off . (C. M. Cat. No. 1600« and 2227a-c; I. U. Cat. No. 12413.) Head 4-4.2; depth 5.5-5.66; D. 6; A. 11-13; adipose dorsal 2-2.4 in the length. Head about an orbital diameter longer than broad, its depth half of its length; interorbital flat, the snout distinctly depressed, the upper jaw longer; premax- illary band of teeth five and a half times as long as deep, a distinct backward pro- jecting angle laterally; maxillary barbel extending beyond the base of the pectoral but not to its tip, post-mental a little beyond the base of the pectoral. Caudal cleft to the base, one short ray on either side of the cleft; the second ray of the upper lobe nearly as long as the third and fourth ; both lobes rounded ; pectoral spine equal to the postorbital portion of the head or a little longer, with retrorse hooks in front, slightly rough near the middle behind. Black, with obscure marblings. 42. Rhamdia arekaima (Schomburgk) . Pimelodus arekaima Schomburgk, Fishes Brit. Guiana, I, 1841, 178 (not pi. 5), (Upper Essequibo). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 643 (all rivers of the savannah). Pimelodus multiradiatus Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 414 (Borba on the Rio Madeira; Forte do Rio Branco on the Rio Takutu). Notoglanis multiradiatus Gunther, Catalogue, V, 1864, 136 (copied). Rhamdia multiradiatus Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 126 (name) ; Occasional Papers Cal. Acad. Sci., 1, 1890, 130. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 386, part. Habitat, Amazon and its tributaries, and northward. This species is placed here on the authority of Schomburgk. The plate of Schomburgk 's arekaima is apparently Pimelodus clarias and the description does not fit it. The specimen from which the drawing was made was said to be two feet three inches long, of fine flavor, and called "Tiger fish." EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 167 Rhamdella Eigenmann and Eigenmann. Rhamdella Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 129 {eriarcha) . Type, Rhamdella eriarcha Eigenmann and Eigenmann. This genus differs from Rhamdia in the presence of a long slit-like parietal fontanel. Most of the species are found in southeastern Brazil. 43. Rhamdella foina (Midler and Troschel). Pimelodus foina Muller and Troschel, in Schomburgk, Reisen, III, 1848, 628 (Takutu); Horse Ichth., Ill, 1849, 5 (Guiana). — Gunther, Catalogue, V, 1864, 130 (copied). Rhamdia foina Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 126 (name); Occasional Papers Cal. Acad. Sci., I, 1890, 126; Proc. U. S. Nat. Mus., XIV, 1891, 28. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 386. This species has been taken but once before; I have examined the type in Berlin and five specimens, 141-201 mm. Warraputa Cataract. (C. M. Cat. No. 1587a-b; I. U. Cat. No. 12007.) Head 4; depth 5-6.5; D. I, 6; A. 10. Head much depressed, its depth equal to eye and postorbital part of head, its width equal to the length behind the maxillary barbel; head evenly arched from its lower margin to its lower margin, the eyes directed upward and outward, 4 in the head, two-thirds of the interorbital; occipital fontanel long and narrow ; occip- ital process very short, extending about one-fourth of the distance to the dorsal spine. Premaxillary teeth fine, in a band of about equal width throughout. Maxillary barbel extending slightly beyond tip of pectoral; post-mental a little beyond their base. Dorsal spine two-fifths the length of the head; base of dorsal a little less than its distance from the adipose, which is 3.5 in the length. Caudal forked, not cleft to its base, the lower lobe rounded, the upper lobe pointed, .4 longer than the lower, 3.5 in the length; depth of caudal peduncle about half its length. Uniform blue-black to mottled ashy; dorsal uniform or its base darker. Pimelodella Eigenmann and Eigenmann. Pseudorhamdia Steindachner, " Siisswasserfische SiidostlichenBrasilien," iii, 1876, 46 (lateristriga) , not Pseudorhamdia Bleeker. Pimelodella Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 131 (cristatus) . 168 MEMOIRS OF THE CARNEGIE MUSEUM Type, Pimelodus cristatus Miiller and Troschel. Similar to Pimelodus, but with a parietal fontanel persistent throughout life; the occipital process not tapering, but of nearly uniform width, in contact with the dorsal plate behind; no vomerine teeth. Species usually slender and of small size. Key to the Guiana Species of Pimelodella. a. Adipose fin 2.33-2.5 in the length. A narrow lateral band, conspicuous in young, becoming obscure with age. Maxillary barbels to below middle of adipose or beyond tip of caudal cristata. aa. Adipose fin 3 or more in the length. 6. Eye 2.5 in the head, longer than snout; interorbital 2 in the eye; maxillary barbel reaching end of adipose or caudal; caudal lobes long, slender, the lower 3 in the length; pectoral spine little shorter than head megalops. 66. Eye 3.5 in head, shorter than snout; tip of dorsal dusky; maxillary barbel reaching middle of adipose; pectoral spine about equal to snout and eye macturki. 44. Pimelodella cristata (Miiller and Troschel). ? Pimelodus insignis Schomburgk, Fishes Brit. Guiana, 1, 1841, 180 (not plate). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 643 (Takutu and Rio Branco). Pimelodus cristatus Muller and Troschel, in Schomburgk, Reisen, III, 1848, 628 (Takutu and Mahu Rivers); Horae Ichth., Ill, 1849, 4 (Guiana, in Esse- quibo). — Gunther, Catalogue, V, 1864, 117 (Guiana; Essequibo; River Capin, Para).— Vaillant, Bull. Soc. Philom., (7), IV, 1880, 152 (Calderon) .— Stein- dachner, " Flussfische Siidamerika's," iv, 1882, 4 (Rio Huallaga). — ? Perugia, Ann. Mus. Genova, (2), X, 1891, 631 (Tucuman). Pimelodella cristata Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1, 1888, 132 (San Goncallo; Avary; Villa Bella; Jutahy; Tapajos; Rio Mucuri; Tabatinga; Hyavary; Coary); Occasional Papers Cal. Acad. Sci., I, 1890, 150.— Eigenmann and Bean, Proc. U. S. Nat. Mus., XXXI, 1907, 660 (Amazon).— Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 388. Pimelodus agassizii Steindachner, " Icthyologische Beitrage," v, 1876, 99 (Peruvian Amazon; Hyavary). Pimelodus (Pseudorhamdia) wesselii Steindachner, " Stisswasserfische Siidostlichen Brasilien," iii, 1876, 56, footnote (Essequibo). ? Pimelodella wesselii Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 132 (Cudajas; Para; Marajo; Rio Madeira; Rio Puty; Santarem); Occasional Papers Cal. Acad. Sci., I, 1890, 152. Pimelodus ophthalmicus Cope, Proc. Am. Philos. Soc, XVII, 1878, 675 (Upper Ama- zon). eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 169 Without opening the question whether the specimens recorded by Eigenmann and Eigenmann as cristatus and wesselii belong to distinct species, I am inclined to consider the wesselii of Steindachner as a synonym of the cristatus of Miiller and Troschel. I have examined the type of cristatus and: Twelve specimens, 81-137 mm. Tumatumari. (C. M. Cat. No. 1686a-d; I. U. Cat, No. 12062.) One specimen, 145 mm. Below Packeoo Falls. (C. M. Cat, No. 1687.) One specimen, 105 mm. Rockstone. (C. M. Cat. No. 1688.) Two specimens, 115-150 mm. Konawaruk. (C. M. Cat. No. 1689; I. U. Cat, No. 12063.) Twelve specimens, 160-205 mm. Creek below Potaro Landing. (C. M. Cat. No. 1690a- f /; I. U. Cat. No. 12064.) Two specimens, 133-158 mm. Twoca Pan. (C. M. Cat, No. 1691a; I. U. Cat. No. 12065.) Essequibo and creeks of the interior, most abundant in the lower Potaro River. Head 4.5; depth 4.5-5.75; D. 1,6; A. 13-15; adipose fin 2.25-2.4; eye 3.5-4 in the head, interorbital 4.25-5; half the eye or all of it in the posterior half of the head. Occipital crest more than three times as long as wide, reaching the dorsal plate; fontanels of about equal width, the posterior becoming very narrow behind; maxillary barbels extending to the base of the caudal or slightly beyond its tip; outer mental barbels reaching ventrals or in the largest specimens a little shorter. Dorsal rounded, the spine slender, about equal to the fourth ray in height; anterior margin of the dorsal spine with recurved notches on its distal half or less, posterior margin with small recurved hooks; interspace between the dorsals about equal to the eye; caudal deeply forked, frequently split to the base, the lower lobe much the broader; ventrals not reaching anal; pectoral not to ventral; pectoral spine with small hooks along the posterior margin, its anterior margin rough, with recurved notches on distal half. A narrow blue-black stripe from below origin of dorsal, disappearing near caudal, and becoming obscure with age; dorsal with the usual hyaline band. 45. Pimelodella megalops sp. nov. (Plate XV, fig. 2.) Pimelodella megalops Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 389. Type, 100 mm. Tumatumari. (Carnegie Museum Catalog of Fishes No. 1692.) Cotypes, fifty-six specimens, 67-96 mm. Tumatumari. (C. M. Cat. No. 1693a-e; I. U. Cat. No. 12066.) 170 MEMOIRS OF THE CARNEGIE MUSEUM Three specimens, 56-90 mm. Crab Falls. (C. M. Cat. No. 1694; I. U. Cat. No. 12067.) Allied to macturki, but with a distinctly larger eye and a longer and more slender lower caudal lobe. Head 4.75-5; depth 6.5-7; D. 1,6; A. 11 to 13; adipose fin 3.33 in the length; eye 2.5 in the head, its center a little behind the middle of the head; interorbital 5.5 in the head. Width of occipital crest 2.5-3 in its length, reaching dorsal plate; posterior fontanel considerably wider than the anterior for a short distance, narrowed back- ward; maxillary barbel reaching to near end of adipose or base of caudal; outer mental barbels reaching the ventrals. Dorsal spine slender, equal to the third or fourth ray in height ; a few scarcely evident recurved notches near the tip in front, and fine recurved teeth for nearly its entire length behind; space between the dorsals considerably greater than the large eye; caudal very widely forked and sometimes split to the base, the lower lobe longer than the upper by nearly the length of the eye, 2.75-3 in the length; pectoral spine but little shorter than head, with minute straight teeth along its anterior margin to the tip, where they are replaced by recurved hooks, posterior margin with about eighteen recurved spines near the tip, the spines largest near the middle. A dark median band; dorsal hyaline, a narrow streak of chromatophores along the front of each ray. 46. Pimelodella macturki sp. nov. (Plate XVI, fig. 1.) Pimelodella macturkii Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 389. Type, 69 mm. Creek in Mora Passage. (Carnegie Museum Catalog of Fishes No. 1695.) Cotypes, fourteen specimens, 46-76 mm. Creek in Mora Passage. (C. M. Cat. No. 1696a-e; I. U. Cat, No. 12068.) Cotypes, two specimens, 66 and 65 mm. Georgetown trenches. (C. M. Cat. No. 1697; I. U. Cat, No. 12069.) Cotypes, fifteen specimens, 49-71 mm. Choca trenches at Morowhanna. (C. M. Cat. No. 1698a-rf; I. U. Cat. No. 12070.) This species takes the place of cristata along the coast from Georgetown to Morowhanna. It differs notably in the length of the adipose and barbels, the color of the dorsal, and the serration on the pectoral spine/ EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 171 Head 4-4.5; depth 5.25; dorsal 1,6; A. 10-13; adipose fin 3.5-4 in the length; eye 3.33-4; interorbital 4.5; eye about in middle of the head. Occipital crest more than three times as long as wide, reaching the dorsal plate; fontanels of about equal width, the posterior becoming narrower backward; maxillary barbel reaching origin or end of base of anal; outer mental barbel to near tip of pectoral or a little shorter. Dorsal low, rounded, the spine equal to the third or fourth ray, rough near the tip in front and on the distal half of the posterior margin; space between the dorsals much greater than the eye; caudal deeply forked, the lower lobe broader, a little longer than the head, 3.5 in the length; ventrals not reaching anal ; pectorals not to ventrals; pectoral spine equal to snout and eye in length or a little longer, with minute teeth along its anterior margin to near the tip, with twelve long re- curved spines along the middle of its posterior margin, the larger nearest the tip. An obscure lateral stripe; tip of caudal dusky. Pimelodus Lacepede. Pimelodus Lacepede, Hist. Nat. Poiss., V, 1803, — (species of several genera). — ■ Cuvier, Regne Animal, II, 1817, 203 (species having a single band of teeth in the upper jaw). — Swainson, Class. Fishes, Amph., and Rept., II, 1839, 305 (quadrimaculatus) . — Lutken, Dan. Vidensk.-Selsk., Skr., (5), XII, 1875, 163 (macidatus = clarias). Pseudariodes Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 99 (clarias). Pseudorhamdia Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 101 (macidatus = clarias) . Pseudorhamdia Lutken, Dan. Vidensk.-Selsk. Skr., (5), XII, 1875, 49, 169 (fur). Type, Pimelodus macidatus Lacepede. Eye with a free orbital margin; teeth on vomer none or minute; occipital process reaching the dorsal plate; no parietal fontanel; head covered with thin skin, granular; caudal forked; humeral process broad; barbels terete. Key to the Guiana Species of Pimelodus. a. Adipose dorsal about 4 or more than 4 in the length. b. Occipital process broad at base, tapering to the dorsal plate. c. Pectoral spine smooth or nearly smooth in front. Dorsal without a conspicuous spot; sides spotted, striped or plain; snout narrowed, its width less than half the length of the head; posterior nares large; upper jaw but little projecting; dorsal spine reaching considerably beyond tip of last ray clarias. cc. Pectoral spine with antrorse teeth in front and retrorse teeth behind. Dorsal with a con- spicuous black blotch near the middle of the front half of the fin; a light streak from 172 MEMOIRS OF THE CARNEGIE MUSEUM dorsal spine to above ventral and then back to the middle caudal rays, another light streak above it; snout broad, depressed, the width of the mouth about half the length of the head; posterior nares small; upper jaw considerably projecting; dorsal spine not nearly reaching tip of last ray ornatus. bb. Occipital process narrow at the base, scarcely tapering to the tip; pectoral spine with retrorse teeth on the basal half of the posterior margin; without conspicuous markings, the chroma- tophores in an area between the anal and dorsal aggregated along the septa. .heteropleurus. aa. Adipose dorsal very long, 2.4-2.6 in the length; caudal lobes produced, twice as long as the head, maxillary barbels extending much beyond the tips of the caudal altipinnis. 47. Pimelodus clarias (Bloch). Silurus clarias Bloch, Ausl. Fische, 1795, pi. 35, figs. 1-2 = Silurus clarias Linnaeus in part; not Silurus clarias Hasselquist, which is Synodontis clarias from the Nile. Pimelodus clarias Lacepede, Hist. Nat. Poiss., V, 1803, 93 (? 8 dorsal rays); Cas- telnau, Anim. Am. Sud, Poiss., 1855, 34 (Crixas; Araguay; Ucayale; Ama- zon). — Steindachner, "Fisch-Fauna Magdalenen-Stromes," 1878, 15 (Mag- dalena River); "Flussfische Siidamerika's," xiv, 1882, 4 (Rio Huallaga).— Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 134.— Eigenmann, Ann. N. Y. Acad. Sci., VII, 1894, 633 (Rio Grande do Sul) .- Eigenmann and Norris, Rev. Mus. Paulista, IV, 1900, 353 (Iguape). — Eigen- mann and Kennedy, Proc. Acad. Nat. Sci. Phila., 1903, 499 (Asuncion; Arroyo Trementina). — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 388. Bagrus (Ariodes) clarias Muller and Troschel, in Schomburgk, Reisen, III, 1848, 627 (Waini and Barima). Ariodes clarias Muller and Troschel, Horse Ichth., Ill, 1849, 10 (British Guiana).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 413. Pseudariodes clarias Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 99 (name only). — ■ Lutken, Vid. Med. Naturhist. For. Kjobenhavn, 1874, 194, 199 (Guiana). Silurus callarias Bloch and Schneider, Syst. Ichth., 1801, 379, part. Pimelodus maculatus Lacepede, Hist. Nat. Poiss., V, 1801, 94, 107 (Rio Plata). — Valenciennes, in d'Orbigny, Voy. Am. Mer., V, ii, 1847, pi. 1, figs. 1-3 (La Plata to Mexico, Lake Maracaibo). — Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 192 (Cayenne; Maracaibo). — Schomburgk, Fishes Brit. Guiana, 1, 1841, 175 (Rivers of Guiana generally; Rio Negro; Amazon). — Stein- dachner, " Icthyologische Notizen," vi, 1867, 32 (La Plata) ; ix, 1869, 6 (Monte- video). — Hensel, Archiv fur Naturg., I, 1870, 69 (Jacuhy). — Lutken, Dan. Vidensk.-Selsk. Skr., (5), XII, 1875, 163, fig. (Rio das Velhas). — Steindachner, "Siisswasserfische Siidostlichen Brasilien," iii, 1876, 40 (La Plata ; Rio San Fran- eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 173 cisco; Rio das Velhas; Amazon between Para and Santarem; Rio Grande do Sul; Porto Alegre).— Peters, MB. Akad. Wiss. Berlin, 1877, 470 (Calabozo) — Boulenger, Proc. Zool. Soc. London, 1891, 233 (Rio Grande do Sul). — Peru- gia, Ann. Mus. Genova, (2J, X, 1891, 630 (Rio Durazno; Rio de La Plata; Paraguay; Parana; Montevideo; Buenos Aires). — von Ihering, Slisswasser- fische Rio Grande do Sul, 1893, 17.— Lahille, Rev. Mus. la Plata, VI, 1895, 271 (very abundant every where in La Plata) . — Boulenger, Boll. Mus. Zool. ed Anat. Comp. Torino, XII, 1897 (Mission San Francisco). — Perugia, Ann. Mus. Genova, (2), XVIII, 1897, 149,— (Puerto 14 de Mayo).— Pellegrin, Bull. Mus. d'Hist. Nat., V, 1899, 157 (Apure).— Vaillant, Nouv. Arch. Mus. d'Hist. Nat. (4), 1900, 124 (Carsevenne, French Guiana). — Steindachner and von Bayern, Denkschr. Akad. Wiss. Wien, LXXII, 1902, 135. — Eigenmann and Bean, Proc. U. S. Nat. Mus., XXXI, 1907, 668 (Amazon) ; Eigenmann, Ann. Carnegie Mus., IV, 1907, 115 (Porto Murtinho; Bahia Negra; Corumba). Pimelodus rigidus Agassiz, Selecta Gen. et Spec. Bras., 1829, 19, pi. 7, fig. 2. Pimelodus blochii Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 188 (Cayenne; (Surinam). Paramutana blochii Gunther, Catalogue, V, 1864, 111 (copied). — Vaillant, Bull. Soc. Philom., (7), IV, 1880, 152 (Calderon). ? Pimelodus arekaima Schomburgk, Fishes Brit. Guiana, I, 1841, 178, plate, not descr. (Essequibo). Mystus ascita Gronow, Cat. Fish, ed. Gray, 1854, 156 (based on Mus. Ichth., I, 1754, 35; Zoophyl, I, 1763, 125, No. 385). - Pimelodus schomburgkii Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 208 (for B. maculatus Schomburgk). Pseudorhamdia ascita Bleeker, Versl. en Med. Acad. Wet. Amsterdam, XIV, 1862, 384 (Surinam). Pimelodus macronema Bleeker, " Silures de Suriname," 1864, 79, pi. 14 (Surinam). Pseiidorhamdia piscatrix Cope, Proc. Am. Philos. Soc, XI, 1870, 569 (Pebas); Proc. Acad. Nat. Sci. Phila., 1872, 262 (Ambyiacu River) ; Proc. Am. Philos. Soc, XVII, 1878, 674 (Peruvian Amazon). Pseudariodes pantherinus Lutken, Vid. Med. Naturhist. For. Kjobenhavn, 1874, 192, 199 (Venezuela). Pseudariodes albicans Lutken, Vid. Med. Naturhist. For. Kjobenhavn, 1874, 194, 198, not Aims albicans Valenciennes (La Plata and its tributaries). — Stein- dachner, " Fisch-Fauna Magdalenen-Stromes," 1878, 61, note. Piramutana macrospila Gunther, Ann. and Mag. Nat. Hist. (5), VI, 1880, 10, pi. 2 (Rio Plata). 174 MEMOIRS OF THE CARNEGIE MUSEUM The following is a general description, applicable to all the varieties. Head 3.75-3.85; depth 4-4.2; D. 1,6; A. 12 or 13. Top of head, occipital process, dorsal plate and humeral process granular; profile steep, rising with more or less of an angle from the base of the occipital process. Dorsal spine striate on the sides, with weak retrorse teeth near its tip on the anterior margin; posterior margin with a few similar hooks near the tip; spine about equal to the head in length, the last ray two-fifths the length of the spine. Base of dorsal slightly greater than its distance from the adipose, which is slightly longer than the dorsal; its height 2.5 in its length. Upper caudal lobe two-fifths longer than the lower; anal emarginate; pectoral spine striate above and below, its outer margin nearly smooth in the adult, retrorse spines along its inner margin. Humeral process broad, obliquely truncate behind, not spine-like. The following varieties were collected. One specimen, 98 mm. (maxillary barbel 115). Koriabo Rubber Plantation. (C. M. Cat. No. 1617.) One specimen, about 127 mm. (maxillary barbel 150). Issorora Rubber Planta- tion. (I. U. Cat, No. 12009.) Ten specimens, 53-68 mm. Creek in Mora Passage. (C. M. Cat. No. 1683 a-d; I. U. Cat. No. 12059.) One specimen, 238 mm. (maxillary barbel 270). Lama Stop-Off. (C. M. Cat. No. 1608.) Here belongs also the macronema of Bleeker. Maxillary barbel extending beyond tip of caudal; vomerine teeth present; upper caudal rays 2.5 in the length; eye 1-1.4 in the interorbital. Ashy, lateral line light; no spots. B. Four specimens, 115-253 mm. (maxillary barbel 133-238). Wismar. (C. M. Cat. No. 1609; I. U. Cat. No. 12010.) Three specimens, about 67-108 mm. (maxillary barbel 138). Locality ? (C. M. Cat. No. 1610; I. U. Cat. No. 12011.) 114 specimens, 50-173 mm. Bartica Sand-bank. (C. M. Cat. No. 1611a-i; I. U. Cat. No. 12012.) Two specimens, 92 (maxillary barbel 100) to about 96 mm. (maxillary barbel 100). Rockstone. (C. M. Cat. No. 1612; I. U. Cat. No. 12013.) eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 175 Two specimens, 89-93 mm. (maxillary barbel 110-115). Crab Falls. (C. M. Cat. No. 1613; I. U. Cat. No. 12014.) Two specimens, 111 mm. to base of lower caudal lobe. Rupununi. (C. M. Cat. No. 1614; I. U. Cat. No. 10215.) Vomerine teeth present; upper caudal lobe 2.66 in the length; eye greater than interorbital to 1.5 in interorbital; color in young ashy to below the lateral line, lateral line and a narrower line half-way between it and the back light, the dark streak below the lateral line breaking into spots at times, the light line on the upper part of the sides becoming irregular; a dark predorsal spot; a darker streak along humeral spine to above the ventral; dorsals spotted; color especially bright in Nos. 1614 C. M. Cat. and 12015 I. U. Cat.; color of adult uniform ashy, without spots. A single small specimen, 50 mm. long, from Tumatumari, probably belongs here. Two specimens, 275 (maxillary barbel 171, not reaching to caudal)-330 mm. (maxillary barbel 235, reaching to caudal). Botanic Gardens. (C. M. Cat. No. 1615; I. U. Cat. No. 12016.) Five specimens, 190-172 mm. Lama Stop-Off. (C. M. Cat. No. 1616; I. U. Cat. No. 12017.) Steel-blue, with light reticulations breaking the blue into small spots; lateral line white. Maxillary barbel usually not reaching the caudal. Eye 1.5 in interorbital. 48. Pimelodus ornatus Kner. Pimelodus ornatus Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 411, fig. 18 (Surinam; Rio Negro; Cujaba). — Gunther, Catalogue, V, 1864, 116 (River Capin, Para).— Peters, MB. Akad. Wiss. Berlin, 1877, 470 (Calabozo) — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1, 1888, 134 (Goyaz). — Eigenmann and Kennedy, Proc. Acad. Nat. Sci. Phila., 1903, 499 (Asuncion; Arroyo Trementina). — Eigenmann, Ann. Carnegie Mus., IV, 1907, 115 (Corumba); Repts. Princeton Univ. Exp. Patagonia, III, 1910, 388. Pseudorhamdia ornata Bleeker, " Silures de Suriname," 1864, 77 (Surinam). Silurus megacephalus (ex Natterer, MS.) Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 413. Five specimens, 140-283 mm. Creek below Potaro Landing. (C. M. Cat. No. 1618a and 6; I. U. Cat. No. 12018.) Twenty-six specimens, 45-167 mm. Tumatumari. (C. M. Cat. No. 1619a- e; I. U. Cat. No. 12019.) 176 MEMOIRS OF THE CARNEGIE MUSEUM Three specimens, 114-144 mm. Rockstone. (C. M. Cat, No. 1620; I. U. Cat. No. 12020.) One specimen, 55 mm. Crab Falls. (C. M. Cat. No. 1621.) Head 3.3-3.5; depth 4.4-5.5; D. 1,6; A. 12-13. Profile nearly straight, interorbital slightly depressed; surfaces of head covered, scarcely granular in the adult, Dorsal spine slender, smooth or slightly rough in front, not quite equal to snout and eye; last dorsal ray about half the length of the highest, the spine not reaching beyond the middle of the last ray when depressed, usually shorter; base of dorsal a little greater than its distance from the adipose, which is longer than the dorsal; caudal lobes slender, the upper 3.5-4 in the length; posterior margin of anal subtruncate; pectoral spine about equal to the snout, with antrorse teeth in front and retrorse teeth behind. Maxillary barbel reaching to base or to tip of caudal, post-mental to middle of pectoral; no teeth on the vomer; humeral process spine-like. Dorsal with a conspicuous black blotch near the middle of the front half of the fin; a light streak from dorsal spine to above ventral and then back to the middle caudal rays, and above this again another light streak. 49. Pimelodus heteropleurus sp. nov. (Plate XVI, fig. 2.) Type unique, 46 mm. Rupununi Pan. (Carnegie Museum Catalog of Fishes No. 1734.) Head 4; depth 5.3; D. 1,6; A. 11; adipose 4 in the length; eye 3.3 in the length, but little shorter than the snout, one-fourth longer than width of interorbital. Form of Pimelodus clarias, the occipital process narrow at its base, scarcely tapering to its tip, which just reaches the dorsal plate; fontanel not quite reaching to the posterior margin of the eye, a bridge over the middle of the eye, the part back of it narrower, not continued as a groove; jaws equal, the teeth in bands of equal width in the two jaws; maxillary barbels reaching a little beyond middle of adipose, outer mental barbel just beyond base of pectoral. Dorsal spine slender, equal to head without snout in length, about equal to the distance between the dorsals; anal rounded; ventrals not reaching anal; pec- toral spine a little longer than the dorsal spine, rough in front, with retrorse teeth behind which increase in size to the middle of the spine, where they cease abruptly; humeral process spine-like. Sides punctate, the chromatophores in the median area between the dorsal and anal gathered along the septa; a median series of chromatophores along the sides; back along base of dorsal and between the dorsals dark. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 177 50. Pimelodus altipinnis Steindachner. Pimelodus altipinnis Steindachner, " Icthyologische Notizen," i, 1864, 14, pi. 2, figs. 3 and 4 (Demerara) ; " Icthyologische Beitrage," iv, 1876, 55, pi. 11 (Para; Santarem; Cameta). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1, 1888, 135 (Para); Occasional Papers Cal. Acad. Sci., I, 1890, 180 (Para). — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 388 — ? Perugia, Ann. Mus. Genova, (2), XVIII, 1897, 18 (Rio Beni).— Eigen- mann and Bean, Proc. U. S. Nat. Mus., XXXI, 1907, 660 (Amazon). The type of this species came from Demerara. It is 86 mm. long. No speci- mens were secured by me. The species has the general characters of a Pimelodella, and it would not be surprising if the small type in the Vienna Museum should prove to be a Pimelodella and distinct from the specimens subsequently referred to the same name. Gceldiella Eigenmann and Norris. Goeldiella Eigenmann and Norris, Rev. Mus. Paulista, IV, 1900, 353 (eques). Type, Pimelodus eques Miiller and Troschel. This genus differs from the related genera Pimelodus and Pimelodella in the character of the caudal fin, which is unequally lobed. 51. Goeldiella eques (Miiller and Troschel). Pimelodus eques Muller and Troschel, in Schomburgk, Reisen, III, 1848, 628 (all rivers of Guiana); Horse Ichth., Ill, 1849, 5 (Guiana). — Gunther, Catalogue, V, 1864, 116 (copied). — Steindachner, "Icthyologische Beitrage," v, 1876, 99 (Amazon near Fonteboa; Teffe; Obidos; Villa Bella; Jose Fernandez; XingujTonantins; Hyutahy; Lake Hyanuary ) . — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1888, 134 (localities given by Steindachner); Occa- sional Papers Cal. Acad. Sci., I, 1890, 166; Proc. U. S. Nat. Mus., XIV, 1891, 29.— Pellegrin, Bull. Mus. d'Hist, Nat., V, 1899, 405 (Manaos). Goeldiella eqties Eigenmann and Norris, Rev. Mus. Paulista, IV, 1900, 353. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 388. Eleven specimens, 91-148 mm. Rupununi. (C. M. Cat. No. 1606o-c; I. U. Cat. No. 12008.) One specimen, 205 mm. Rockstone. (C. M. Cat. No. 1607.) Head 3.5-3.75; depth 4.75-5.25; D. 1,6; A. 10 or 11. Profile in front of dorsal nearly straight, lower outline nearly straight from caudal to below eye; interorbital flat; bones of the head from the nares to the dorsal plate and the dorsal plate with reticulated ridges; occipital fontanel a narrow slit in specimens up to a length of at least 150 mm. 178 MEMOIRS OF THE CARNEGIE MUSEUM Eye 4.5-5.5, 1.3-1.5 in interorbital, somewhat behind the middle of the head; premaxillary teeth in a narrow band of the same width throughout. Maxillary barbel extending to the middle of the caudal or farther, post-mental not quite to tip of pectoral. Dorsal spine with retrorse notches in front, roughened behind, about equal to snout and eye in length; pectoral spine strong, striate above and below, with antrorse teeth in front and retrorse teeth behind; adipose fin nearly or quite half as long as the body without the head; caudal peduncle half as high as long. Caudal slightly notched, the upper lobe shorter than the lower, projecting for about one-fourth of its length beyond the fork. A dark saddle downward and forward from front of dorsal, covering nearly the entire opercle; an irregular mottled dark band along the lateral line and just below it; base of caudal dark, the fins irregularly mottled. Lower surface white. Phractocephalus Agassiz. Phractocephalus Agassiz, Selecta Gen. et Spec. Pise. Bras., 1829, 22 (bicolor = hemilopterus) . Large. Vomer with a large pentagonal patch of teeth in contact with the palatine patches. Occipital process large, semicircular, not meeting the reniform dorsal plate. Upper half of adipose usually rayed. Barbels subterete. Caudal forked; head as broad as long. Type, Silurus hemiliopterus Bloch and Schneider. 52. Phractocephalus hemiliopterus (Bloch and Schneider). Silurus hemiliopterus Bloch and Schneider, Syst. Ichth., 1801, 385. Phractocephalus hemiliopterus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 3, pi. 421. — Schomburgk, Fishes Brit. Guiana, I, 1841, 169 (Guiana, everywhere). — Muller and Troschel, in Schomburgk, Reisen, III, 1849, 643 (all rivers of Guiana). — Castelnau, Anim. Am. Sud, Poiss., 1855, 47, pi. 15, fig. 4 (Rio Crixas; Araguay; Amazon). — Gunther, Catalogue, V, 1864, 110 (River Cupai).— Vaillant, Bull. Soc. Philom., (7), IV, 1880, 152 (Calderon).- CoPE,Proc. Am. Philos. Soc, XVII, 1878, 674 (Peruvian Amazons). — Eigen- mann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 135 (Xingu; Coary; Teffe; Manacapuru; Obidos; Lake Hyanuary); Occasional Papers Cal. Acad. Sci., I, 1890, 188. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 390. Phractocephalus bicolor Agassiz, Selecta Gen. et Spec. Pise. Bras., 1829, 23 (Amazon). eigenmann: the FRESHWATER PISHES OF BRITISH GUIANA 179 Pirarara bicolor Spix, Gen. Spec. Pise. Bras., 1829, 23 (Amazon). No specimens of this species were preserved. Parts of a head were seen in the Georgetown market. Schomburgk states that it reaches a length of four feet and is "common to all rivers." The description is takan from Eigenmann and Eigenmann. Head 3.33; depth 4.75; Br. 9; D. 1,7; A. 9; eye 9 in head, 3 in snout, 5 in inter- orbital, 2 diameters behind the rictus. Body rapidly tapering towards caudal; head heavy, broad, flattened between the eyes; greatest width of the head equals its length; width at the angle of the mouth 2 in its length. Maxillary barbel on edge of lip, opposite anterior nostril, reaching beyond the tips of the pectoral fin; mental barbels two-thirds to one-half as long as post- mental barbels, which reach the pectoral. Lower jaw included; teeth all alike, those on the intermaxillaries in a broad band of equal depth throughout; vomerine teeth in a much broader patch; palatine teeth in narrower, wedge-shaped patches contiguous to the vomerine patch. Gill-membranes separate to below anterior margin of the eye. Gill-rakers short and fleshy, 4 + 15. Bones behind eye variously grooved and granulated; the occipital process broadly rounded behind, not meeting the reniform dorsal plate. Dorsal spine midway between snout and tip of adipose dorsal, and between bases of pectoral and ventral fins; last dorsal ray over base of ventral, the length of the spine 2.25 in head. Adipose dorsal short, high, its upper portion generally transformed into true rays; its base longer than that of the anal. Caudal broad, slightly emarginate. The rays of the dorsal and caudal thick and terete, once or twice branched. Ventrals extending for half their length beyond the vent. Pectoral spine two or three times as thick as the dorsal spine, broad lamellae in front and sharp recurved teeth behind; its length 2 in the head. Between pectorals and anal dark brown; lower part of head, a narrow band above pectorals and along sides, spreading over the lower three-fourths of the tail, white (yellow in life), region above this brownish; a round white spot on each side of dorsal spine; orbit bordered with white above, head and ante-dorsal region with darker spots or vermiculations; sometimes all of the lower parts are white. Brachyplatystoma Bleeker. Platystoma Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, -- (sp.). Brachyplatystoma Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 97 (vaManti). 180 MEMOIRS OF THE CARNEGIE MUSEUM Piratinga Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 100 (reticulata). Malacobagrus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 100 (filamentosus). Type, Platystoma vaillanti Cuvier and Valenciennes. Vomer and palate with villiform teeth; inner teeth of the jaws slender and freely movable; caudal forked; head covered with skin; occipital process short, concealed, not reaching the dorsal in the adult. 53. Brachyplatystoma vaillanti (Cuvier and Valenciennes). (Plate XVII, fig. 3.) "Laulau." Platystoma vaillanti Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 21, pi. 423 (Cayenne and Surinam) .— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 397 (Para). — Gunther, Catalogue. V, 1864, 108 (Demerara). — Peters, MB. Akad. Wiss. Berlin, 1877, 469 (Calabozo). Brachyplatystoma vaillanti Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 97 (name only) ; ' ' Silures de Suriname, ' ' 1 864, 70 (Surinam) . — Eigenm ann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1, 1888, 136 (Tabatinga; Para; Porto do Moz; Arary; Rio Puty; Juiz de Fora); Occasional Papers Cal. Acad. Sci., I, 1890, 196. — Eigenmann and Bean, Proc. U. S. Nat. Mus., XXXI, 1907, 662 (Amazon).— Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 390. Platystoma affine Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 24 (Brazil). — Castelnau, Anim. Am. Sud, Poiss., 1855, 40 (Araguay).— Gunther, Catalogue, V, 1864, 109 (copied). I have examined the specimen mentioned by Bleeker and: One specimen, head only. Rockstone. (C. M. Cat. No. 1288.) Sixteen specimens, 178-290 mm. to tip of middle caudal rays. Georgetown market. (C. M. Cat. No. 170a-d; I. U. Cat, No. 12083.) Head 3.4-3.5; depth 5.5; D. 1,6; A. 13; eye 4 in snout; 10 in head, 2.5 in interorbital in a specimen 280 mm. long to tip of middle caudal rays, 4 in snout, 9 in head, and 2.5 in interorbital in a specimen 178 mm. long to tip of middle caudal rays; adipose fin 4.4-4.5 in the length. Subtriangular in section at the dorsal fin, the snout much depressed, the tail moderately compressed; occipital crest narrow, not much tapering, its width near the middle one-fourth its length in the Georgetown specimens, reaching but not join- ing the dorsal plate. An elongate depression along middle of the head ; fontanel nar- row, extending to the posterior margin of the eye; mouth lunate from below, the upper jaw scarcely projecting ; maxillary barbel extending beyond tip of middle caudal rays in all the Georgetown specimens; mental barbels opposite angle of EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 181 mouth, the post-mentals reaching beyond the ventrals. Dorsal and anal emar- ginate, the first ray in each reaching beyond the tip of the last; dorsal and pectoral spines slender, with recurved hooks on the posterior margin, most marked toward the tip; pectoral and ventral falcate; caudal deeply forked, the outer rays normally about 4 times as long as the middle ones, but prolonged in the young to equal the length from snout to caudal. Pale below, darker above. The young of this species can be readily distinguished by its long barbels and prolonged caudal tips. Hemisorubim Bleeker. Platystoma Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 27 (sp.), pre- occupied in Diptera. Hemisorubim Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 97 (platyrhynchos). Type, Platystoma platyrkynchos Cuvier and Valenciennes. Upper jaw thin and truncate, shorter than the lower jaw; snout narrowed; the width at the angle of the mouth 1.5 times in the greatest width of the head; mental barbels approximate and near the edge of the lip; premaxillary band of teeth much shallower at the middle than at the ends; the palatine patches large. Occipital process shorter than the dorsal plate and meeting it; postorbital portion of the head striate and granulate; skin on the sides of the head and snout reticulate. 54. Hemisorubim platyrhynchos Cuvier and Valenciennes. Platystoma platyrhynchos Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 27 (no locality). — MtJLLERand Troschel, in Schomburgk, Reisen, III, 1848, 628 (Rupununi). — Castelnau, Anim. Am. Sud, Poiss., 1865, 40 (Amazon).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 398 (Barra do Rio Negro). Hemisorubim platyrhynchos Bleeker, Nederl. Tijdschr. Dierk., I, 1863,97 (name only). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 138 (Manacapuru; Rio Negro; Montalegre; Rio Puty; Lago Alexo; Obidos; Coary; Tabatinga; Hyavary; Tonantins; Sao Paolo) ; Occasional Papers Cal. Acad. Sci., I, 1890, 206.— Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 390. Hemisorubim platyrhynchus Gunther, Catalogue, V, 1864, 109 (copied). — Peters, MB. Akad. Wiss. Berlin, 1877,470 (Calabozo).— Cope, Proc. Am. Philos. Soc, XVII, 1878, 674 (Peruvian Amazon).— Vaillant, Bull. Soc. Philom., (7), IV, 1880, 152 (Calderon). No specimens of this species were obtained. 182 MEMOIRS OF THE CARNEGIE MUSEUM I have been able to examine a specimen collected by Schomburgk and now in the Berlin Museum. Head 3; depth 6-7; D. 1,6; A. 10; Br. 11; eye 7 in head, 3 in snout, 1.75 in interorbital. Head depressed, rounded on occiput; eye directed largely upward. Lower jaw projecting and entering profile. Maxillary barbels extending past dorsal fin; mental barbels approximate, extending to below eyes; post-mentals to pectorals. Teeth in the lower jaw in a narrow band. Vomerine and palatine teeth close behind the premaxillary teeth, the vomerine teeth in a single patch separate from the palatine patches. Dorsal spine weak and slender, 2.5 in head, its distance from the snout greater than its distance from the posterior margin of the adipose fin, much nearer base of ventrals than base of pectorals; with retrorse teeth behind. Adipose fin longer than anal. Lower caudal lobe wider and longer than upper, rounded in adult, pointed in young. Ventrals extending two-fifths their length beyond the vent. Pectoral spine 1.6 in head, usually with equally strong teeth in front and behind. White below, olivaceous above, with a few jet-black spots scattered on sides, and usually a similar spot at base of upper caudal lobe; fins plain. Pseudoplatystoma Agassiz. Platystoma Agassiz, Selecta Gen. et Spec. Pise. Bras., 1829, 23 (sp.). Sorubim Spix, Selecta Gen. et Spec. Pise. Bras., 1829, pis. 12-15 (sp.). Pseudoplatystoma Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 97 (fasciatum). Hemi-platystoma Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 97 (tigrinum). Type, Silurus fasciatus Linnseus (in part). Upper jaw little longer than lower; teeth of the jaws alike; vomerine and palatine patches of teeth more or less united, the two forming a comma-shaped patch on each side of the palate; gill-rakers short, spine-like, not overlapping; caudal deeply lobed, the lobes rounded (except in very young), the rays very much branched, giving the fin a leathery texture. Branchiostegals 14 or 15; barbels short. 55. Pseudoplatystoma fasciatum (Linnseus). Tiger-fish. Silurus fasciatus Linnaeus, Syst, Nat., ed. 12. I, 1766, 505.— Gmelin, Syst. Nat., I, iii, 1788, 1359.— Bonnaterre, Tabl. Enc. Meth., Ichth., 1788, 154,252.- Bloch, Ausl. Fische, VIII, 1794, 30, pi. 366.— Bloch and Schneider, Syst. Ichth., 1801,382. EIGENMANM : THE FRESHWATER FISHES OF BRITISH GUIANA 183 Pimelodus fasciatum Lacepede, Hist. Nat. Poiss., V, 1803, 94, 99, 100. Platystoma fasciatum Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 14 (copied).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 401 (Surinam).- Gunther, Catalogue, V, 1864, 107 (Essequibo; Surinam; River Capin, Para).— Peters, MB. Akad. Wiss. Berlin, 1877, 469 (Calabozo). — Steindachner, "Fisch-Fauna Magdalenen-Stromes," 1878, 15 (Magdalena River). — Cope, Proc. Am. Philos. Soc., XVII, 1878, 674 (Peruvian Amazon). — Stein- dachner, ''Fisch-Fauna des Cauca," etc., 1880, 5 (Cauca); "Icthyologische Beitrage," viii, 54 (Surinam); "Flussfische Stidamerika's," iv, 1882, 4 (Rio Amazonas ; Iquitos) . Pseudoplaty stoma fasciatimi Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 97 (name only); "Silures de Suriname," 1S64, 72 (Surinam). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 138 (Obidos; Coary; Hyavary); Occasional Papers Cal. Acad. Sci., I, 1890, 208.— Steindachner and von Bayern, Denkschr. Akad. Wiss. Wien, LXXII, 1902, 136 (Rio Lebrija, Colombia). — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 39. Platystoma truncation Agassiz, Selecta Gen. et Spec. Pise. Bras., 1829, 27, pi. 13a (Rio Japura and Solimoens). — Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 20 (Brazil).— Hyrtl, Denkschr. Akad. Wiss. Wien, XVI, 1859, 17 (vertebra? 13 + 33).— Gunther, Catalogue, V, 1864, 108 (copied). Platystoma tigrinum (not of Cuvier and Valenciennes) Schomburgk, Fishes Brit. Guiana, 1, 1841, 185 (most of the rivers). — Muller and Troschet., in Schom- burgk, Reisen, III, 1848, 627 (nearly all rivers). Platystoma punctifer Castelnau, Anim. Am. Sud, Poiss., 1855, 40, pi. 19, fig. 2 (Amazon). Four specimens, 245-295 mm. Rupununi Pan. (C. M. Cat. No. 1699a; I. U. Cat. No. 12082.) One specimen, 455 mm. Wismar. (C. M. Cat. No. 1295.) Head 2.66-2.75; depth 6.5-7.25; D. 1,6; A.ll-14; eye 6 in snout, 12 in head, 2.5 in interorbital. Snout much depressed; body slender, subterete at the dorsal; groove of the fontanel not continued to the occipital process; a deep groove across the head at base of the occipital process; maxillary barbel reaching to near tip of dorsal. Dorsal spine nearer tip of adipose than to tip of snout. Dark brown above, abruptly white on the lower part of the sides; about ten black cross-bands, margined with white in front; some dark spots along the line of 184 MEMOIRS OP THE CARNEGIE MUSEUM the lower ends of the bars; vertical fins, and sometimes tips of ventrals, spotted. The specimen in the Leiden museum mentioned by Bleeker agrees with the above. Subfamily Doradin^e. Doras Lacepede. 17 Doras Lacepede, Hist. Nat. Poiss., V, 1803, 116 (carinatus and costatus) — Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 267 (costatus). Centrochir Agassiz, Selecta Gen. et Spec. Pise. Bras., 1829, 14 (crocodili). Lithodoras Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 84 (lithog aster) . Pterodoras Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 84 (granulosus). Platydoras Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 86 (costatus) = Doras Cuvier and Valenciennes. Acanthodoras Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 86 (cataphr actus) . Astrodoras Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 86 (asterifrons) . Amblydoras Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 86 (affinis). Zathorax Cope, Proc. Acad. Nat. Sci. Phila., 1872, 271 (monitor) = Astrodoras Bleeker. Agamyxis Cope, Proc. Amer. Philos. Soc, XVII, 1878, 101 (pectinifrons) . Type, Silurus costatus Linnseus. Doradines with anterior nares on or near the upper lip ; the snout short, the eye in anterior half of the head; teeth well-developed; adipose fin usually with a well- defined anterior margin. Key to the Species of Doras. a. Caudal forked; dorsal spine serrate in front and behind, the serrae of the anterior margin antrorse. Sides of dorsal spine with striations (Doras). 6. Plates of the sides increasing in height to the caudal peduncle, the highest about .4 the height of the peduncle; caudal peduncle naked above and below, spotted, without lateral bands. granulosus. 66. Plates of sides decreasing in height regularly from the dorsal to the caudal, not meeting along a median line above or below; caudal peduncle with a series of plates above and below. Slate-colored above, with a yellow band from the fontanel along the sides to the tip of the middle caudal rays; ventral surface without ossifications costatus. aa. Caudal very slightly emarginate or rounded; dorsal spine without serrations either in front or behind, the sides of the spine with smooth striations; plates of the sides decreasing in height from the dorsal to the caudal, not meeting above and below; caudal peduncle largely naked above and below, the caudal fulcra continued about half-way to adipose and anal fins; chocolate brown with an irregular dark band below the median hooks of the lateral plates, and other darker blotches; nuchal region roof -shaped. (Amblydoras) hancocki. 17 Doras maculatus Valenciennes, Miiller and Troschel, in Schomburgk, Reisen, III, 1S48, 629 (Esse- quibo). EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 185 aaa. Caudal rounded; dorsal spine with a series of straight teeth along its anterior margin, its sides with a regular or several irregular (in old), series of straight teeth, its posterior margin without teeth; plates of sides of about the same height from dorsal to anal, those of opposite sides meeting along the median line on upper and lower surfaces of the caudal peduncle of the adult; a narrow light line from the upper margin of the eye to the caudal; a light line from dorsal spine forward to the fontanel; nuchal region broadly rounded. (Acanthodoras) cataphractus. 56. Doras granulosus Valenciennes. (Plate XVII, fig. 4.) Doras granulosus Valenciennes, in Humboldt, Rec. Obs. Zool., II, 1811, 184.— Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 392. Pterodoras granulosus Bleeker, Nedeii. Tijdschr. Dierk., I, 1863, 15 (name only) " Silures de Suriname," 1864, 36 (Surinam). Doras maculatus Valenciennes, in d'Orbigny, Voy. Am. Mer., V, ii, 1847, 7, pi. 5, fig. 3. — Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 281 (Buenos Ayres). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 629.— Steindachner, Denkschr. Akad. Wiss. Wien, XLI, 1879, 47 (Rio de la Plata). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sei., (2), I, 1888, 150 (Arary?); Occasional Papers Cal. Acad. Sci., I, 1890, 229 (Arary ?; Uruguay; Buenos Aires, Serfa). Doras murica (ex Natterer, MS.) Kner, SB. Akad. Wiss. Wien, XVII, 1885, 129 (Cujaba). Doras muricus Gunther, Catalogue, V, 1864, 202 (Demerara ?) . No specimens were secured. I have examined a specimen about 395 mm. long collected by Schomburgk in Guiana. Head 4, depth about 4; D. 1,6; A. 11; plates 35; eye 3.5 in the snout, 10 in the head, 3.5 in the interorbital. Lateral plates beginning on the vertical from just in front of the dorsal, the twelfth about equal to the eye in height, the highest on the caudal peduncle .4 the height of the peduncle, with a median but no lateral spines. Dorsal spine nearly as long as the head, the spines of the posterior margin much longer than those of the anterior; pectoral spine longer than the head, its posterior spines about twice as long as its anterior; caudal forked, the lower lobe longer. Humeral spine very slender, reaching to below the dorsal spine. Maxillary barbel reaching to the tip of the pectoral ; free margin of the nasal plate but little pectinate. Chocolate, marbled; the fins spotted. 57. Doras costatus (Linnaeus). Silurus costatus Linnaeus, Syst. Nat., ed. 12, I, 1766, 506. Cataphractus costatus Bloch, Ausl. Fische, VIII, 1794, 82, pi. 376. 186 MEMOIRS OF THE CARNEGIE MUSEUM Doras costatus Lacepede, Hist. Nat. Poiss., V, 1803, 116, part (South Amer- ica). — Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 268 (Guiana). — Castelnau, Anim. Am. Sud, Poiss., 1855, 48 (Amazon). — Gunther, Cata- logue, V, 1864, 201 (British Guiana; River Cupai). — Eigenmann and Eigen- mann, Proc. Cal. Acad. Sci., (2), I, 1888, 161 (Rio Preto; Rio Puty; San Gon- callo ; Xingu Cascade ; Obidos ; Gurupa ; Teffe) ; Occasional Papers Cal. Acad. Sci. , I, 1890, 231.— Perugia, Ann. Mus. Genova, (2), 1891, 34 (Villa Maria, Para- guay). — Kindle, Ann. N. Y. Acad. Sci., VIII, 1895, 251 (Trocera on the Tocan- tins). — Eigenmann and Kennedy, Proc. Acad. Nat. Sci. Phila., 1903,500 (Para- guay). — Eigenmann, Ann. Carnegie Mus., XXXI, 1907, 1 16 (Corumba; Laguna Ipacarai); Repts. Princeton Univ. Exp. Patagonia, III, 1910, 393. Platydoras costatus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 16 (name only); " Silures de Suriname," 1864, 38 (Surinam). Doras cataphractus (not of Linnaeus) Schomburgk, Fishes Brit. Guiana, I, 1841, 158 (Rio Negro). Doras armatidus Muller and Troschel, in Schomburgk, Reisen, III, 1848, 629 (Rupununi; Awaricura). Two specimens, 105 and 108 mm. Twoca Pan. (C. M. Cat. No. 1638; I. U. Cat. No. 12030.) Four specimens, 75-90 mm. Gluck Island. (C. M. Cat. No. 1639; I. IT. Cat. No. 12031.) One specimen in the Berlin Museum from Guiana and two from Calabozo. Head 3.75; depth 4.25; D. 1,6; A. 11; lateral plates 2 + 30 or 31. Eye 1.66-2 in snout, 4-5 in head; 1.5-2 in interorbital. Ventral surfaces of coracoid not exposed. Width at tip of humeral processes greater than the depth; nuchal region bluntly roof-shaped. Head granular to in front of posterior nares; nasal bone with blunt serration, scarcely raised; orbit with granular margin; maxillary barbel reaching to tip of humeral process, outer mental to base of pectoral. 18 Upper jaw longer; width of mouth less than half the distance between the gill-openings. Pectoral spine striate above and below, strongly serrate in front and behind, reach- ing beyond base of ventrals; dorsal spine shorter than pectoral spine, its sides striate, serrate on its anterior and posterior margin, the serrse on the posterior margin much smaller; teeth on front margin of dorsal spine antrorse; largest near the tip; humeral process reaching to fourth fifth of the pectoral spine; caudal forked; height of highest lateral plate about half the length of the head, each plate with numerous small backward-directed spines, in several vertical rows in front, 18 To its middle in the larger specimen at Berlin. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 187 in a marginal row behind; median spines strong, recurved; caudal peduncle entirely covered with a median series of plates above and half-way to the anal below. Slaty blue, a light band from the fontanel along the middle to the tip of the caudal; lower surface and margin of head to near the eye light. Barbels dark, dark streak back from the maxillary barbel; dorsal with a dark band across its upper part; center and margins of caudal light, a pair of submarginal dark bands; anal with a dark streak; ventrals hyaline; pectoral dusky. 58. Doras hancocki ( luvier and Valenciennes. " Saurauwari," "Yarauira." Doras costata (not of Linnseus) Hancock, Zool. Journ., IV, 1828, 242 (Demerara).— Schomburgk, Fishes Brit. Guiana, I, 1841, 156. Doras hancockii Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 279.— Gunther, Catalogue, V, 1864, 202 (Demerara; Rio Cupai). — Eigenmann and Eigenmann, Occasional Papers Cal. Acad. Sci., I, 1890, 234. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 393. One hundred and ninety-five specimens, 9-120 mm. Lama Stop-Off. (C. M. Cat. No. 1443a-t; I. U. Cat. No. 12033.) Ten specimens, 65-123 mm. Maduni Creek. (C. M. Cat. No. 1444a-6; I. U. Cat, No. 12034.) One specimen, 28 mm. Wismar. (C. M. Cat, No. 1445a). Nine specimens, 26-41 mm. Rockstone. (C. M. Cat. No. 1446a-6; I. U. Cat, No. 12035.) Twelve specimens, 28-36 mm. Gluck Island. (C. M. Cat, No. 1649a-6; I. U. Cat. No. 12038.) Four specimens, 68-98 mm. Tumatumari. (C. M. Cat. No. 1447a; I. U. Cat, No. 12036.) Eight specimens, 30-45 mm. Rupununi Pan. (C. M. Cat. No. 1448a-6; I. IT. Cat, No. 12037.) Head 4-4.3; depth 4.25; width at tip of humeral process 3.66; D. 1,6; A. 11- 13; lateral plates 1 + 25; eye 1.5 in snout, 4.5 in head. 1.5 in interorbital. Coracoid and its process exposed below, striate; nuchal area with a blunt median keel, roof-shaped; head granular to the anterior nares; movable nasal bone large, forming part of the orbit ; maxillary barbel reaching tip of humeral process, outer mental barbel to its middle; jaws equal, mouth terminal, its width less than half the space between the gill-openings. Pectoral spine strongly serrate on its anterior and posterior margins, its upper 188 MEMOIRS OF THE CARNEGIE MUSEUM surface with three sharp ridges, its lower striate, reaching past base of ventrals; dorsal spine shorter than pectoral spine, with four sharp ridges on each side and a median one in front, the latter sometimes broken up into teeth near the base of the spine; caudal truncate, very slightly emarginate or rounded; humeral process with lines of granules, which become stronger along the lower margin, and recurved spines towards its tip. Lateral plates regularly decreasing in height backward, each with a strong recurved hook, and ridges above and below which end in spines; caudal fulcra extending forward above and below on the caudal peduncle, but the latter largely naked. Ash-colored; head and body variously marked with black; an irregular black band along sides below the median hooks. A black spot at base of caudal; three more or less distinct black blotches on the back; one down and behind dorsal, one about adipose and one across caudal peduncle; dorsal and caudal, and to less extent the remaining fins, marked with black; lower surface variously marked. 59. Doras cataphractus (Linnaeus). Silurus cataphractus Linnaeus, Syst. Nat., ed. 10, 1, 1758, 307; ed. 12, 1, 1766, 506. Doras cataphractus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 276 (?).— Kner, SB. Akad. Wiss. Wien, XVII, 1855, 126 (Rio Guapore; Barra do Rio Negro). — Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 54. — Gunther, Catalogue, V, 1864, 204 (?). — Eigenmann and Eigenmann, Occasional Papers Cal. Acad. Sci., I, 1890, 234. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 393. Acanthodoras cataphractus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 17; " Silures de Suriname," 1864, 40 (Surinam). Cataphractus americanus Bloch and Schneider, Syst. Ichth., 1801, 107, pi. 28.— Lacepede, Hist, Nat. Poiss., V, 1803, 124, 127 (Carolina?). Doras blochii Cuvier and Valenciennes, Hist. Nat, Poiss., XV, 1840, 277 (copied). ? Doras castaneo-ventris Schomburgk, Fishes Brit, Guiana, I, 1841, 161, pi. 3 (Rio Pasawiri) . ? Doras brunnescens Schomburgk, Fishes Brit. Guiana, I, 1841, 163 (Upper Essequibo) . Doras polyramma et polygramma (ex Heckel, MS.) Kner, SB. Akad. Wiss. Wien, XVII, 1855, 126, 127. Callichthys asper Gronow, Cat. Fish., ed. Gray, 1854, 157. One specimen, 170 mm. Kangaruma. (C. M. Cat. No. 1640.) eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 189 One specimen, 84 mm. Georgetown. (C. M. Cat. No. 1641.) Twenty specimens, 53-112 mm. Gluck Island. (C. M. Cat. No. 1642a-c; I. U. Cat. No. 12032.) Head 3.75-4; depth equals head; width 3.16; D. 1,5; A. 10 or 11; lateral plates 2 or 3 + 22 to 27; eye 2.5 in snout, 9 in head, 3.5-4 in interorbital. Ventral surfaces of coracoid not exposed; head depressed, flat between the eyes, becoming slightly arched in the nuchal region, not roof-shaped, granulated to between the anterior nares; posterior nares just in front of the supraorbital ridge, scarcely protected by the low nasal bone; orbital margin, preopercle and opercle granular in the adult; maxillary barbel reaching to second third of humeral process in the adult, beyond its tip in the young; outer mental barbels scarcely shorter than the maxillary. Jaws equal or the upper but slightly longer; mouth strictly terminal, its width less than half the distance between the gill-openings. Pectoral spine strongly serrate on anterior and posterior margins; upper sur- face in young with a median series of straight teeth, in the adult with numerous irregularly placed teeth; lower surface rough in young, striate in adult; dorsal spine about two-thirds the length of the pectoral spine, its posterior surface smooth, its anterior surface with a series of straight spines, its sides with one or more series of similar spines; caudal rounded, humeral process reaching third fourth or fourth fifth of pectoral spine, with a series of large recurved spines in young, and smaller spines above them, the difference between large and small spines becoming less with age; lateral plates highest above origin of anal, those of opposite sides meeting along the median line from near the tip of the dorsal to the caudal and along the ventral line of the caudal peduncle; median spine of plates nearly straight, each plate from about the seventh with a series of teeth increasing in size toward the upper and lower margins; anterior plates with more numerous smaller spines. Black; a light median line along back of head and sometimes body; a light line from posterior nares following margin of skull and then along median line of spines. Caudal black, with hyaline bars; other fins black with hyaline blotches or bands; posterior margin of adipose light; lower surface and sides of head marbled. Oxydoras Kner. Oxydoras Kner, SB. Akad. Wiss. Wien, XVII, 1855, 115 (sp.); Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 14 (niger). Pseudodoras Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 53 (niger). Rhinodoras Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 14 (Orbignyi). Type, Doras niger Valenciennes. Doradines with the barbels simple; no teeth; eye in posterior half of head, snout long and pointed. 190 MEMOIRS OF THE CARNEGIE MUSEUM 60. Oxydoras niger (Valenciennes). Doras niger Valenciennes, in Humboldt, Ree. Obs. ZooL, II, 1811, 184. — Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 291 (?). — Schomburgk, Fishes Brit. Guiana, I, 1841, 165. — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 629 (Rivers of Guiana).— Bleeker, Nederl. Tijdschr. Dierk.. I, 1863, 14 (name only). Rhinodoras niger Gunther, Catalogue, V, 1864, 209 (Amazons). — Cope, Proc. Am. Philos. Soc, XVII, 1878, 678 (Nauta) .— Vaillant, Bull. Soc. Philom., (7), IV, 1880, 14 (Calderon). Oxydoras niger Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 159 (Teffe; Gurupa; Manacapuru; Coary; Obidos); Occasional Papers Cal. Acad. Sci., I, 1890, 247.— Kindle, Ann. N. Y. Acad. Sci., VIII, 1894, 251 (Para). — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 393. Doras humboldt i 'Agassiz, Selecta Gen. et Spec. Pise. Bras., 1829, 129, pi. 5 (Rio San Francisco, Brazil). — Agassiz, a Journey in Brazil, 1868, — . Corydoras edentatus Spix, Selecta Gen. et Spec. Pise. Bras., 1829, pi. 5. Rhinodoras prionomus Cope, Proc. Acad. Nat. Sci. Phila., 1874, 134 (Nauta); Proc. Am. Philos. Soc, XVII, 1878, 678 (Nauta). Rhinodoras teffeanus Steindachner, SB. Akad. Wiss. Wien, LXXI, 1875, 145, pi. 3 (Teffe). One specimen, 178 mm. Rupununi. (C. M. Cat, No. 1622.) Head 3.3; depth 4.5; D. 1,6; A. 11; lateral plates 23; eye 3 in interorbital, 4 in snout, 3 in preorbital part of head. Deepest in front of dorsal spine, the depth equal to the width; caudal peduncle depressed, its depth half the distance from anal to lower caudal rays. Head pointed, its width 1.4 in its length, almost equal to its depth. Opercle, top of head, and dorsal plate tubercular striate. Interorbital flat; fontanel narrow, in a wide groove continued beyond the fontanel, which does not extend beyond the middle of the eye. Lips and barbels thickly papillose, the barbels simple, free from the lips and not united; maxillary barbel extending a little beyond eye; distance between nares less than their distance from tip of snout or eye; distance between gill-openings equals width of mouth. Highest lateral scute less than one-fourth the length of the head, each scute with a very strong median hook, up to eight spines below the median hook and up to ten above it. Dorsal spine striate, with a few straight spines on its posterior surface and much more numerous and stronger antrorse hooks on its anterior margin. Pectoral EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 191 spines with hooks of equal size in front and behind, the anterior antrorse, the posterior retrorse. Adipose fin low, gradually merging into the back in front. Caudal short and broad, its length equal to snout and eye. Slaty black, with lighter markings below; fins black. Leptodoras Boulenger. Leptodoras Boulenger, Ann. and Mag. Nat. Hist., (7), II, 1898, 477. Type, Oxydoras acipenserinus Giinther. No teeth; maxillary barbel fringed; anal 15-18; eye large. Boulenger creates the genus Leptodoras on account of the longer body and the longer anal fin, 15-17 rays. In another place (Trans. Zool. Soc, XIV, 1898,423) he rejects the genus Hemidoras as indistinguishable from Oxydoras. As the type of Hemidoras has thirteen anal rays and Hemidoras brevis has thirteen or fourteen it might be doubted whether Leptodoras can be separated from Hemidoras on the score of anal rays. Since, however, Leptodoras has no teeth it maybe kept distinct. The extralimital species of this genus are acipenserinus, juruensis and probably stubelii. I have been able to examine the types of the two former in the British Museum. Key to the Species of Leptodoras. a. Dorsal spine short, pungent, serrate in front. 6. Eye 3.5 in the head; pectoral spine longer than snout and eye; reaching near middle of ventrals. linnelli. (66. Eye 5 in the head; pectoral spine longer, equal to snout and eye; reaching scarcely beyond origin of ventrals ". acipenserinus. aa. Dorsal spine prolonged, not pungent, roughened near its base; eye 6.5 in the head; pectoral spine reaching ventrals, head depressed juruensis.) 61. Leptodoras linnelli sp. nov. (Plate XVII, fig. 1; Plate XVIII, fig. 1.) Leptodoras linnelli Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 395 (name only). Type, 190 mm. Tumatumari. (Carnegie Museum Catalog of Fishes No. 1626.) Cotypes, twenty-seven specimens, 64-209 mm. Tumatumari. (C. M. Cat. No. 1627a-e; I. U. Cat. No. 12022.) Cotypes, two specimens, 64 and 182 mm. Rockstone. (C. M. Cat. No. 1628; I. U. Cat. No. 12024.) Cotypes, six specimens, 75-103 mm. Crab Falls. (C. M. Cat. No. 1629a; I. U. Cat. No. 12023.) 192 MEMOIRS OF THE CARNEGIE MUSEUM Cotype, one specimen, 174 mm. Georgetown market ? (C. M. Cat. No. 1630.) Head 3.5; depth 6; D. 1,6; A. 12-14; lateral plates 38-39; eye 3.5-4 in head, 1.75-1.8 in snout; interorbital 2 in eye. Slender; depth equal to width; caudal peduncle broad, depressed, its depth equal to half its width, less than the distance of the anal from the caudal ; fontanel continued as a groove to the tip of the occipital process. A small foramen on either side of the occipital process; no teeth; mouth nearly equal to the distance between the gill-openings. Maxillary barbel reaching gill-opening in the young, shorter in the adult. Lateral plates well-developed along the entire sides, each with a central hook and a series of marginal spines, those above the median hook close-set; median hook strongest on the plates, just behind the vertical from the anal; humeral process short, its upper margin greatly arched. Dorsal and pectoral spines of about equal length, little shorter than the head, the pectoral spine reaching to the middle of the ventrals or farther in the adult, not much beyond the base in the young; pectoral spine serrate on both edges, the teeth much stronger on the inner surface; dorsal spine with much feebler serrations; adipose free behind, its base nearly or quite equal to the base of the dorsal exclusive of the spine. A large pectoral pore. White beneath, uniform dark above; a pair of parallel bands on the caudal, the middle rays light. Hemidoras Bleeker. Doras Lacepede, Hist. Nat. Poiss., V, 1803, 116 (carinatus and costatus).— Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 13 (carinatus), not Doras Cuvier and Valenciennes. Oxydoras Kner, SB. Akad. Wiss. Wien, XXVII, 1855, 142 (sp.). Hemidoras Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 53 (stenopeltis) . Hassar Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), 1, 1888, 158 (orestes). Type, Doras stenopeltis Kner. Barbels fringed; eye large. One or both jaws with teeth; numerous pectoral pores. Key to the Guiana Species of Hemidoras. a. No foramen on either side of juncture between occipital process and dorsal plate. (Hemidoras.) b. Lateral plates of the anterior part of the body well-developed; teeth in the lower jaw small, in two patches. c. Groove of the fontanel not extending backward; maxillary barbel not reaching gill-opening. Eye 2.75 in head (1 in young); teeth very feeble, mouth half as wide as distance between gill-openings microstomus. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 193 cc. Maxillary barbel extending to base of pectoral ; eye 4.5 in the head in the adult (3 in the young), 2.75 in the snout (1.5 in young); teeth well-developed in both jaws; snout pointed; mouth 1-1.5 in distance between gill-openings carinatus. bb. Lateral plates of anterior part of body rudimentary or very small, increasing in size backward. d. Teeth of the lower jaw minute, in two separate patches; differing from carinatus largely in the rudimentary anterior plates micropoeus. dd. Teeth of the lower jaw large, brown-tipped, in a single median patch; maxillary barbel ex- tending beyond origin of pectorals; snout bluntly decurved; pectorals extending to ventrals; base of dorsal dark; a light band along the lateral hooks, bordered above and below by darker, which is most intense on base of caudal leporhinus. ao. A large foramen on either side of the juncture between dorsal plate and occipital process; dorsal with a conspicuous black spot. (Hassar.) c. Maxillary barbel extending to base of pectoral; orbit oval; snout long, conical; pectoral spine reach- ing ventrals; dorsal spine with hooks on the basal half of the anterior margin; humeral process elongate, its upper margin not greatly arched notospilus. (ee. Maxillary barbel to below posterior part of the eye; orbit ovate with a narrowing forward exten- sion; snout slender; origin of orbit behind the middle of the head; pectoral spine not reaching ventrals; dorsal spine with hooks along the basal two-thirds or three-fourths of its anterior margin; humeral process spatulate, its dorsal margin much curved. /. Scapula covered with skin; dorsal spot not extending to tip of the membranes orestes. //. Scapula granular; black spot extending to upper margin of dorsal membranes wilderi.) 62. Hemidoras microstomus sp. nov. (Plate XVIII, fig. 2.) Hemidoras microstomus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 394 (name only). Type, 52 mm. Rockstone. (Carnegie Museum Catalogue of Fishes No. 1650.) Cotypes, twelve specimens, 35-51 mm. Crab Falls. (C. M. Cat. No. 1651 a-b; I. U. Cat. No. 12040.) Cotype, one specimen, 52 mm. Rockstone. (I. U. Cat. No. 12039.) Head 3.75-4.2; depth 4.33-5.5; D. 1,6; A. 10-12; lateral plates 26-32; eye 2.5-2.66 in the head, 1-1.33 in the snout; interorbital 1-1.66 in eye. Profile from dorsal to above eye straight, then descending rapidly; the snout very blunt, rounded; lower jaw much shorter, included; fontanel elliptical, little longer than eye, not continued as a groove; top of head convex. Width of mouth equal to half the distance between gill-openings. Teeth very feeble, if present. Maxillary barbels reaching gill-openings, or a little shorter; distance of dorsal spine from tip of snout 2.75-3 in the length. Dorsal spine as long as the head or a little longer, with serrse on the basal half of the anterior margin and entire length of posterior margin. Pectoral spine a little longer or shorter than the dorsal spine, reaching the ventrals; base of adipose 194 MEMOIRS OF THE CARNEGIE MUSEUM less than diameter of eye. Lateral shields small, of uniform height, height 2.5-3.3 in eye, hooks highest above tip of anal fin. Humeral process four times as long as broad, spine-like. Sides and back covered with minute black dots, thickest at base of dorsal and caudal fins, white underneath. 63. Hemidoras carinatus (Linnseus). Silurus carinatus Linnaeus, Syst. Nat., ed. 12,1,1766,504. — Bloch and Schnei- der, Syst. Ichth., 1801, 108. Doras carinatus Lacepede, Hist. Nat. Poiss., V, 1803, 116 (Surinam). — Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 288, pi. 442 (Cayenne). — M tiller and Troschel, in Schomburgk, Reisen, III, 1848, 629 (Essequibo). — Bleeker, Ichth. Arch. Ind. Prodr., I, 1858, 54; Nederl. Tijdschr. Dierk., I, 1863, 13 (name only); " Silures de Suriname," 1864, 31 (Surinam). Doras (Oxydoras) carinatus Kner, SB. Akad. Wiss. Wien, XVII, 1855, 144 (Surinam). Oxydoras carinatus Gunther, Catalogue, V, 1864, 206 (Surinam ; Essequibo River) . — Vaillant, Bull. Soc. Philom. (7), IV, 1880, 154 (Calderon). Hemidoras carinatus Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 158; Occasional Papers Cal. Acad. Sci., 1, 1890, 258. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 394. Doras oxyrhynchus Valenciennes, in Humboldt, Rec. Obs. Zool., II, 1833, 184. One hundred and thirty-three specimens, 44-320 mm. Tumatumari. (C. M. Cat. No. 1631«-2; I. U. Cat. No. 12025.) One specimen, 95 mm. Bartica. (C. M. Cat. No. 1632.) Ninety-two specimens, 48-267 mm. Rockstone. (C. M. Cat. No. 1633a-t; I. U. Cat, No. 12026.) Three specimens, 45, 113, and 247 mm. Georgetown market. (C. M. Cat. No. 1634a; I. U. Cat. No. 12027.) One hundred and sixteen specimens, 39-140 mm. Crab Falls. (C. M. Cat. No. 1635a-i; I. U. Cat. No. 12028.) Head 3.16-3.33; depth 4.25-4.5; D. 1,6; A. 11-14; lateral plates 33; eye 3- 4.5 in the head, 1.5-2.75 in the snout, the snout proportionately longer in the old individuals; interorbital 1.3-1.5 in the eye. Heavy below the dorsal, the width 1.25 in the depth; caudal peduncle as wide as deep, its depth less than distance of anal from the lower caudal rays; occipital area roof-shaped, with a median groove which is not continued to the eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 195 fontanel; profile abruptly descending in front of the eye, the snout sharp; a pair of patches of small teeth in each jaw; those of the upper jaw sometimes wanting; mouth narrow, 1-1.5 in the distance between the gill-openings. Maxillary barbel reaching to the gill-openings. Lateral plates nearly equally well-developed along the entire length; humeral process truncate, its upper and lower margins nearly parallel, its depth one-third of its length. Distance between snout and dorsal 2.5 in the length, dorsal spine as long as snout and eye or shorter; pectoral spine longer than dorsal spine, reaching the ventrals; adipose fin low, its base equal to the base of the dorsal without the spine. Axillary pores numerous. White below, ashy above, no definite markings. In the specimen from Bartica the dorsal spine is margined with dark. 64. Hemidoras micropceus sp. nov. Hemidoras micropceus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 394 (name only). Type, 365 mm. Wismar. (Carnegie Museum Catalog of Fishes No. 1636.) Cotype, 217 mm. Wismar. Cotypes, two specimens, 250-270 mm. Lama Stop-Off. (C. M. Cat. No. 1637; I. U. Cat. No. 12029.) These specimens differ from carinatus in having the plates from above the ventrals forward rudimentary. 65. Hemidoras leporhinus sp. nov. (Plate XIX, fig. 1.) Hemidoras leporhinus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 394 (name only). Type, 56 mm. Tumatumari. (Carnegie Museum Catalog of Fishes No. 1624.) Cotypes, three specimens, 53-59 mm. Crab Falls. (C. M. Cat. No. 1625; I. U. Cat. No. 12021.) Readily distinguished by its peculiar leporine snout, and by its mouth, lower teeth, and coloration. Head nearly 4; depth 5; D. 6; A. 12; lateral plates 33 to 35; eye 2.4 in the head, equal to the snout; interorbital 1.6 in the eye. Depth equal to the width in front of the pectoral; caudal peduncle slender, its depth equal to the distance of the anal from the caudal fulcra; profile curved, 196 MEMOIRS OF THE CARNEGIE MUSEUM sharply so in front of the eye; snout pointed; fontanel continued as a groove to the tip of the dorsal plate; a small foramen on either side of the juncture between the dorsal plate and occipital process; mouth 2 in the distance between the gill-open- ings; five to seven teeth in each ramus of the lower jaw, grouped in front so as to form one continuous patch; maxillary barbel reaching at least to base of pectoral; humeral process more than three times as long as wide, rounded behind. Lateral plates deepest above end of anal, their depth four-tenths of the length of the eye. Distance of origin of dorsal from tip of snout 2.7 in the length. Dorsal and pectoral spines about equal in length and equal to the length of the head. Belly white; sides of head and body thickly dotted, darkest on dorsal surface of caudal peduncle; a light stripe along the lateral plates and middle caudal rays; bases of the caudal fulcra and a band above and below the central light band black; tip of upper lobe dusky; a dark area on base of dorsal, highest on the spine. 66. Hemidoras notospilus sp. nov. (Plate XIX, fig. 2.) Hemidoras notospilus Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 394 (name only). Type unique, 70 mm. Crab Falls. (Carnegie Museum Catalog of Fishes No. 1623.) Head 3.33; depth 5; D. 1,6; A. 12; lateral plates 33; eye 3 in head, 1.5 in snout; interorbital a little more than half the eye. Profile from dorsal to above front of eye nearly straight, descending rapidly in front of eye; snout pointed, lower jaw much the shorter; fontanel continued as a groove to the occipital, not to the tip of the process or the dorsal plate; dorsal plate bat-shaped, a large foramen on either side of it in front; interorbital and snout narrow; depth of head greater than its width, 1.5 in its length. Eye oval, 1.75 times as long as high. Maxillary barbel reaching pectoral, having numerous small barblets; mental barbels papillose. Distance of dorsal spine from tip of snout 2.5 in the length. Dorsal spine a little longer than snout and half the eye. Pectoral spine a little longer than the dorsal spine, reaching the ventrals. Base of adipose less than the length of the eye. Lateral shields highest above end of anal, their height 1.33 of the length of the eye. Humeral process more than three times as long as broad. Belly white, sides with an increasing amount of pigment toward the back; eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 197 base of dorsal and sides of the spine dotted, rest of fin hyaline, except a conspicuous black spot near the tip of the first three dorsal rays. Subfamily Auchenipterin^e. Centromochlus Kner. Centromochlus Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 430 (megalops = heckelii) . Glanidium (ex Reinhardt, MS.) Lutken, Dan. Vidensk.-Selsk. Skr., XII, 1874, 31 (albescens). Type, Centromochlus megalops Kner. Mental barbels in two pairs; adipose fin shorter than the anal; anal short, 7-11; mouth terminal; jaws equal; caudal forked; V. 6. 67. Centromochlus aulopygius Kner. (Plate XX, fig. 1.) Centromochlus aulopygius Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 432, pi. 8, fig. 25 (Rio Guapore). — Gunther, Catalogue, V, 1864, 198 (Essequibo). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 157 (Cudajas); Occasional Papers Cal. Acad. Sci., I, 1890, 270. — Pellegrin, Bull. Mus. d'Hist. Nat., V, 1899, 158 (Apure) —Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 395. One specimen, 65 mm. Creek below Potaro Landing. (C. M. Cat. No. 1727.) One specimen, 65 mm. Wismar. (I. U. Cat. No. 12643.) Head 3.75-4.25; depth 4-5; D. 1,4 or 1,5; A. 9 or 10; eye 5 in snout, 3 in head, 2 in interorbital. Head blunt, bullet-shaped; tail compressed; top of head finely granular; an ovate fontanel, its posterior margin over the last third of the eye; jaws subequal; teeth conical, fixed, in narrow bands; lower margin of gill-opening on a level with the upper margin of the pectoral spine; maxillary barbel reaching slightly beyond the middle of the pectoral spine; mental barbels reaching beyond base of post- mentals, the post-mentals not to base of pectoral. Dorsal spine a little shorter than the pectoral spine, as long as the head; anterior margin of dorsal with large antrorse teeth, the posterior margin smooth; pectoral spine serrate on both margins, the teeth of the posterior margin larger. Ventrals not reaching to anal; anal rays crowded, especially in the Wismar speci- men; adipose fin shorter than the dorsal. Chocolate brown above, shading to white on the belly; sides of body and caudal with numerous horizontally-oval light spots; lower fins light. 198 MEMOIRS OF THE CARNEGIE MUSEUM Trachycorystes Bleeker. Trachycorystes Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 88 (typus = trachy- corystes) . Parauchenipterus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 88 (galeatus). Type, Trachycorystes typus Bleeker. Mental barbels in two pairs; adipose fin shorter than the anal fin; anal 19-41; caudal obliquely truncate or slightly emarginate; outer margin of pectoral spine serrate. Key to the Species of Trachycorystes. a. Caudal oblique. 6. Dorsal spine longer than the pectoral spine, which is 1.5 in the head; head covered with smooth skin glaber. 66. Pectoral spine longer than dorsal spine, about as long as head; head granular, or covered with thin skin in the young; A. 22-28 galeatus. aa. Caudal emarginate; A. 19 or 20; dorsal spine serrate on its anterior edge; pectoral spine as long as the head obscurus. 68. Trachycorystes glaber (Steindachner). Auchenipterus glaber Steindachner, " Siisswasserfische Siidostlichen Brasilien," iii, 1876, 97, footnote (Demerara). Trachycorystes glaber Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 154 (name only); Occasional Papers Cal. Acad. Sci., I, 1890, 272, 275 — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 395. No specimens of this species were obtained. It is known from the types in the Vienna Museum. D. 1,5; A. 23; dorsal spine smooth on outer, weakly serrate on inner margin, longer than pectoral spine, 1.25 in the head; pectoral spine 1.5 in the head; caudal obliquely rounded. 69. Trachycorystes galeatus (Linnseus). Silurus galeatus Linnaeus, Syst. Nat., ed. 12, I, 1766, 503 (based on Seba, Locupl. Rer. Nat. Thes. Ace. Descr., Ill, 1748, pi. 29, fig. 7).— Gmelin, Syst. Nat., I, iii, 1788, 1357.— Bloch, Ausl. Fische, VIII, 1794, 39, pi. 369, fig. 1- Bloch and Schneider, Syst. Ichth., 1801, 384. Pimelodus galeatus Lacepede, Hist. Nat. Poiss., V, 1803, 97, 114 (South America). Auchenipterus galeatus Gunther, Catalogue, V, 1864, 196 (Guiana). — Peters, MB. Akad. Wiss. Berlin, 1877, 470 (Calabozo). Parauchenipterus galeatus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 88 (name only); " Silures de Suriname," 1864, 45 (Surinam). EIGENMANN: THE FRESHWATER FISHES OF BRITISH GUIANA 199 Trachycorystes galeatus Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 155 (Pernambuco; San Goncallo; Rio San Francisco, below the falls; Tabatinga; Teffe; Rio Puty); Occasional Papers Cal. Acad. Sci., I, 1890, 279. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 396. Auchenipterus maculosus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 216 (Cayenne). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 639 (Essequibo).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 425 (Mara- bitanos). — Gunther, Catalogue, V, 1864, 196 (Surinam; Essequibo). — Vail- lant, Bull. Soc. Philom., (7), IV, 1880, 154 (Calderon). Auchenipterus immaculahts Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 218 (Cayenne). Auchenipterus punctatus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 219 (Brazil?). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 629 (Essequibo).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 425 (Rio Branco). Auchenipterus lacustris Lutken, Dan. Vidensk.-Selsk. Skr., XII, 1875, 148, with fig. (Rio das Velhas). Auchenipterus robustus Gunther, Catalogue, V, 1864, 197 (Demerara). Trachycorystes robustus Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 156 (name only). One specimen, 110 mm. Mud-flats, Demerara River, below Wismar. (C. M. Cat. No. 1717a.) One specimen, 132 mm. Barima River. (,C. M. Cat. No. 1718a.) Five specimens, 116-160 mm. Aruka River. (C. M. Cat. No. 1719a; I. U. Cat. No. 12096a-6.) Eleven specimens, 99-195 mm. Lama Stop-Off. (C. M. Cat, No. 1720a-6; I. U. Cat. No. 12093.) Four specimens, 130-228 mm. Georgetown trenches. (C. M. Cat. No. 1721a; I. U. Cat. No. 12094.) One specimen, 167 mm. Maduni Creek. (C. M. Cat. No. 2229.) Four specimens, 91-140 mm. Chipoo Creek. (C. M. Cat. No. 1722a-6; I. U. Cat. No. 12095.) Head 4-4.5; depth 3.5-4; D. 1,6; A. 22-27, most frequently 24; V. 6; P. 1,6. Eye 1.5-2 in snout, 5.5-6 in head, 3.5-4 in interorbital. Head heavy, its width equal to its length, depressed, profile but little concave at the nape, rounded forward, the lower jaw entering the profile; upper surface of head finely granular in the old, covered with thin smooth skin in the young; 200 MEMOIRS OF THE CARNEGIE MUSEUM fontanel surrounded by bone; maxillary barbel reaching to or beyond tip of the humeral process, mental barbel 2.5-3 times as long as the eye. Dorsal spine equal to head without the snout or a little longer but not equal to the length of the head, rough in front, with recurved hooks for over half its length from the tip on the posterior margin; pectoral spine as long as the head, with antrorse teeth in front and larger retrorse teeth behind; caudal obliquely rounded. Color variable, much lighter in those inhabiting muddy water, with a light band across the head behind the eye, the dorsal plate also light, sides strigate; much darker to black in those inhabiting clear waters, the sides strigate. The specimens from Chipoo Creek, a tributary of the Ireng, differ from those near the coast in having the maxillary barbels a little shorter, reaching to or not quite to the tip of the humeral process, the mental barbels 1.5-1.75 times as long as the eye. 70. Trachycorystes obscurus (Giinther). (Plate XVII, fig. 2.) Auchenipterus obscurus Gunther, Ann. and Mag. Nat. Hist., (3), XII, 1863, 442 (Essequibo); Catalogue, V, 1864, 195. Trachycorystes obscurus Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 154; Occasional Papers Cal. Acad. Sci., I, 1890, 275. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 396. I have examined the types in the British Museum and two specimens in the Berlin Museum, all collected by Ehrhardt. No specimens of this species were secured by me. The species is readily distinguished by its emarginate caudal. D. 1,5; A. 19-20; maxillary and post-mental barbel extending to, or somewhat beyond, the tip of the humeral process; mental barbel 2 or 3 times as long as the eye; dorsal serrate along its anterior edge; pectoral serrate along both edges; dorsal spine shorter than the pectoral spine, which equals the head in length; caudal emarginate, the upper lobe scarcely longer than the lower. Uniform brownish black. Pseudauchenipterus Bleeker. Felichthys Swainson, Class. Fishes, Amph., and Rept., II, 1839, 305 (sp). — Swain, Proc. Acad. Nat. Sci. Phila., 1882, 281 (nodosus). Pseudauchenipterus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 88 (nodosus). Type, Silurus nodosus Bloch. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 201 71. Pseudauchenipterus nodosus (Bloch). (Plate XX, fig. 2.) Silurus nodosus Bloch, Ausl. Fische, VIII, 1794, 35, pi. 368, fig. 1 (Tranque- bar?). — Bloch and Schneider, Syst. Ichth., 1801, 383. Arius nodosus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 70 (copied). Auchenipterus nodosus MtJLLERand Troschel, Horse Ichth., Ill, 1849, 11. — Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 424 (Surinam).— Gunther, Catalogue, V, 1864, 194 (British, Dutch, and French Guiana). Pseudauchenipterus nodosus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 88 (name only); " Silures de Suriname," 1864, 43, pi. 11, fig. 1, pi. 13, fig. 6 (Surinam). —Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 396. Felichthys nodosus Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 154 (Para; Bahia); Occasional Papers Cal. Acad. Sci., I, 1890, 291. Auchenipterus furcatus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 211 (Guiana). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 629 (Essequibo) . Pseudauchenipterus guppyi Regan, Proc. Zool. Soc. London, 1906, 387 (Trinidad). Par auchenipterus paseadw Regan, Proc. Zool. Soc. London, 1906, pi. 23. Twenty specimens, 210-290 mm. Georgetown market. (C. M. Cat. No. 1723a-e; I. U. Cat. No. 12097.) Three specimens, 137-145 mm. Mahaica. (C. M. Cat. No. 1724a; I. U. Cat, No. 12098.) Two specimens, 212 and 227 mm. Mud-flats of Demerara River, near Wismar. (C. M. Cat. No. 1725a; I. IT. Cat. No. 12099.) Head 4.2; depth 45-4.75, D. 1,6; A. 21 or 22; V. 8; P. 1,7. Eye about 1 in snout, 4.25 in head, 2.5 in interorbital. Subtriangular in front, the head blunt, the tail compressed; maxillary barbel reaching posterior third of pectoral spine, mental barbel beyond base of pectoral, post-mental to middle of pectoral spine; head nearly as broad as long; frontal bones pitted, not much swollen; lower jaw included. Dorsal spine 3-3.6 in the length, feebly serrate behind; pectoral spine 3.5-3.75 in the length, its outer margin striate, its inner finely serrate. Caudal forked; anal slightly emarginate. Back and sides dark blue; an undulating white streak follows the lateral line, dividing with it on the caudal; area above anal lighter than the rest of the sides; upper and lower margin of caudal light, the hinder margin black; tips of remaining fins light. 202 MEMOIRS OF THE CARNEGIE MUSEUM Auchenipterus Cuvier and Valenciennes. Auchenipterus Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 207 {nuchalis) . Evanemus Muller and Troschel, Horse Ichth., Ill, 1849, 11 (colymbetes) . Type, Hypophthalmus nuchalis Spix. Mental barbels arranged in a line near the symphysis; a short adipose fin; anal long; ventrals with 12 to 15 rays; pectorals 1,11. Key to the Species of Auchenipterus. a. Mandibular barbels extending to the tip of the pectoral demerarae. aa. Mandibular barbels extending to the lower angle of the gill-opening brevior. 72. Auchenipterus demerarae sp. nov. (Plate XXI, fig. 1.) Type, 115 mm. Wismar. (Carnegie Museum Catalog of Fishes No. 1714.) Cotype, 115 mm. Malali. (C. M. Cat. No. 1713.) Cotypes, two specimens, 101 and 118 mm. Wismar and mud-flats just below Wismar. (I. U. Cat. No. 12091.) Similar to Auchenipterus nuchalis (Spix), but with a lateral band. Head 5.25; depth 4.75; D. 1,6; A. 41 to 44; V. 12; P. 1,11; eye 3 in head, 1.5 in interorbital, 1 in snout; depth of caudal peduncle 2 in the head. Head short, blunt, depressed; body much compressed; occipital process as long as wide, fontanel as long as the pupil, its anterior edge but slightly in advance of the posterior margin of the eye; lower jaw slightly included, the teeth in very narrow bands ; gill-membranes free to behind the lower margin of the pupil ; maxil- lary barbel extending to the tip of the pectoral, mandibular barbels to the middle of the pectorals; distance from tip of snout to dorsal 4.33 in the length, the spine equal to snout and eye, with a few recurved notches on the posterior margin near the tip; pectoral spine but little shorter than the head, smooth in front, with recurved teeth along its entire posterior margin; pectorals not reaching ventrals, ventrals to about the sixth anal ray ; highest anal ray equals snout and eye; middle caudal rays about half as long as the outer. Sides everywhere peppered; a dark median band from the gill-opening to the caudal; dorsal and caudal dusky, the tip of the upper caudal lobe sometimes quite dark; lower fins hyaline. 73. Auchenipterus brevior sp. nov. Type, 71 mm. Tumatumari. (Carnegie Museum Catalogue of Fishes No. 1715a.) eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 203 Cotypes, twenty specimens, 50-73 mm. Tumatumari. (C. M. Cat. No. 1716 a-d; I. U. Cat. No. 12108.) These specimens differ from demerarce notably in one observed particular : the mandibular barbels extend to about the lower angle of the gill-opening, not to the pectoral. The maxillary barbels are as long as in demerarce; depth 5.33. Subfamily Ageneiosin^e. Tympanopleura 19 gen. nov. Type, Tympanopleura piper ata sp. nov. Maxillary barbels only, short. Air-bladder projecting into the abdominal cavity, naked laterally, the skin over it forming a large pseudo-tympanum; snout short, about equal to the eye; profile very concave; first dorsal and first pectoral rays pungent; origin of anal equidistant from rictus and middle caudal rays. 74. Tympanopleura piperata sp. nov. (Plate XX, fig. 3.) Type, a male, 64 mm. Crab Falls. (Carnegie Museum Catalog of Fishes No. 1708.) Cotypes, two males and five females, 57 -61 mm. Crab Falls. (C. M. Cat. No. 1709a; I. U. Cat. No. 12090.) This is evidently a young fish. To what extent the large, protruding air- bladder and the large pseudo-tympanum are characters of immaturity I am unable to say. The short snout very probably is due to the age of the specimen. Head 4.5; depth 4.75; D. 1,5 or 6; A. 31; V. 7; P. 1,9; Br. 7; eye 1 in snout, 2.8 in head, 1 in space between the eyes below. Profile much concave; fontanel open in front, continued as a groove to above the posterior margin of the pupil; snout rounded, the gape short; maxillary barbel in the male with an osseous base extending to below the anterior margin of the eye, the fleshy tip a little farther; in the female, minute, fleshy, reaching the rictus; gill-opening extending to below the posterior margin of the eye; dorsal spine with minute teeth on its front margin, its posterior margin smooth in the male, with recurved teeth along its entire length in the female; pectoral spine smooth in front, with recurved teeth along its entire margin behind; a large pectoral pore. Distance of dorsal from snout 3.66 in the length; origin of anal equidistant from rictus and caudal; ventrals reaching past origin of anal; pectorals not to ventrals. 19 Tvniravov, tympanum, Tr\evpi, side. 204 MEMOIRS OF THE CARNEGIE MUSEUM Sides everywhere lightly peppered with ehromatophores ; an hour-glass-shaped dark bar across the base of the caudal. Ageneiosus Lacepede. Ageneiosus Lacepede, Hist. Nat. Poiss., V, 1805, 132 (armatus); IX, 167. Pseudageneiosus Bleeker, Nederl. Tijdschr. Dierk., I, 1863, 108 (davalla). Ageniosus Gunther, Catalogue, V, 1864, 191 (sp.). Type, Ageneiosus armatus Lacepede. Maxillary barbels only; air-bladder minute, concealed under the peritoneum and largely covered with bone; no pseudo-tympanum. Snout much longer than the eye. Key to the Guiana Species of Ageneiosus. a. Caudal deeply forked. 6. A. 45; back with obscure marblings; maxillary band of teeth comma-shaped, the teeth large; dorsal and pectoral spines slender, pungent; origin of anal a little nearer base of middle caudal rays than the rictus; pectorals not reaching ventrals. Head 4: P. 1,14; Br. 9. guianensis. aa. Caudal emarginate; dorsal and pectoral rays smooth on posterior margin, the pectoral ray not pungent. c. Dorsal spine strong, sinuate, rough or spinous in front; profile steep, concave, origin of anal equidistant from base of middle caudal rays and anterior margin of eye; pectoral falcate, the first ray reaching beyond origin of ventrals; head 3.66; P. 1,14; Br. 9; A. 34, dark above, lighter below brevifilis. cc. First dorsal ray not spinous; profile nearly straight and horizontal; origin of anal a little nearer origin of pectoral than to base of middle caudal rays; pectoral rounded, not reaching ven- trals; head 3.33; P. 16; Br. 10; A. 29. Conspicuously marked with light and dark. marmoratus. 75. Ageneiosus guianensis sp. nov. (Plate XXI, fig. 2.) Type unique, a female, 175 mm. Wismar. (Carnegie Museum Catalog of Fishes No. 1712a.) Head 4; depth 5.5; D. 1,6; A. 45; V. 8; P. 1,14; Br. 9; eye 3 in snout, 6 in head, 3.5 in space between the eyes below. Profile concave, not very steep, snout much depressed; width of head 1.4 in its length; length of snout less than interocular width by about an orbital diameter; fontanel reaching above middle of eyes, the groove considerably farther; cleft of mouth 3 in the length of the head; half the maxillary barbel osseous, its tip reaching rictus; premaxillary band of teeth comma-shaped, widest in front, its width half the length of the eye, the teeth comparatively large. Gill-openings extending to below the posterior margin of the eye, the space between them about equal to the diameter of the eye. eigenmann: the FRESHWATER FISHES OF BRITISH GUIANA 205 Distance from snout to dorsal 3.6 in the length; the spine slender, with minute wide-set teeth on the anterior margin, more numerous minute teeth on the posterior margin; pectoral spine slender, pungent, slightly rough in front, with small curved teeth along its entire inner margin; caudal deeply forked, the middle rays not half as long as the outer rays, which are 4.5 in the length; origin of the anal a little nearer to the base of the middle caudal rays than to the rictus; ventrals reaching beyond origin of anal; pectorals four-fifths the distance to the ventrals. Dark chocolate above, shading to the anal; ventral surface light; back ob- scurely marbled; dorsal, base and tips of caudal, upper surface of pectoral and ventral, except their last rays, dark. This species is near dentatus Kner and ucayalensis Castelnau, but differs in the width of the head, the length of the snout, the mouth, color, etc. 76. Ageneiosus brevifilis Cuvier and Valenciennes. Ageneiosus inermis (not Silurus inermis Linnaeus or Bloch) Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 240, pi. 440 (Surinam).— Castelnau, Anim. Am. Sud, Poiss., 1855, 48 (Amazon). Ageneiosus brevifilis Cuvier and Valenciennes, Hist. Nat. Poiss., XV, 1840, 242 (Cayenne).— Kner, SB. Akad. Wiss. Wien, XXVI, 1857, 438 (Rio Cujaba).- Gunther, Catalogue, V, 1864, 192 (River Capin, Para); Proc. Zool. Soc. 'London, 1868, 229 (Xeberos).— Cope, Proc. Am. Philos. Soc, XVII, 1878, 676 (Peruvian Amazon). — Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 150 (name only); .Occasional Papers Cal. Acad. Sci., I, 1890, 309. — Eigenmann, Repts. Princeton Univ. Exp. Patagonia, III, 1910, 397. Pseudageneiosus brevifilis Bleeker, " Silures de Suriname," 1864, 83, pi. 16, fig. 1 (Surinam) . Hypophthalmus dawalla Schomburgk, Fishes Brit. Guiana, I, 1841, 191, pi. 9 (Guiana). — Muller and Troschel, in Schomburgk, Reisen, III, 1848, 643. Ageneiosus dawalla Eigenmann and Eigenmann, Proc. Cal. Acad. Sci., (2), I, 1888, 150 (name only); Occasional Papers Cal. Acad. Sci., I, 1890, 309. Ageniosus sebce Gunther, Catalogue, V, 1864, 192 (copied). One specimen, a male, 445 mm. Lama Stop-Off. (C. M. Cat. No. 1711a.) Head 3.66; depth 5; D. 1,6; A. 34; V. 8; P. 1,14; Br. 9; eye 3.5 in snout, 6.5 in head, 5.5 in space between the eyes below. Profile steep, concave; head depressed, broad; fontanel reaching eye, the groove beyond the posterior margin of the eye; maxillary barbel osseous to near its 206 MEMOIRS OF THE CARNEGIE MUSEUM tip, tuberculate above, reaching the eye; premaxillary band of teeth .66 of the length of the eye, the teeth minute; gill-openings reaching to below the eye, the space between them equal to the eye. Distance from snout to dorsal 3.66 in the length; the spine an orbital diameter shorter than the head, with short sharp teeth along its concave anterior margin, the posterior margin smooth; caudal emarginate; origin of anal equidistant from base of middle caudal rays and the anterior margin of the eye, its first rays thick^ ened, osseous; ventrals emarginate, reaching to the third anal ray; pectorals sub- falcate, the first ray not pungent, smooth on both margins, reaching the first or fifth ventral ray. Steel-blue above, lighter below; dorsal spotted; upper surface of pectoral and anterior ventral rays steel-blue, the membranes white with dark spots; caudal margined with light ; a submarginal dark band shading into the color of the caudal peduncle. Bleeker's figure was based on his smaller specimen, and the markings are still visible on the fins. The barbel is however too long, not reaching the eye. 77. Ageneiosus marmoratus sp. nov. (Plate XXII, fig. 1.) Type unique, a female, 175 mm. Creek below Potaro Landing. (Carnegie Museum Catalog of Fishes No. 1710.) Readily distinguished by its conspicuous markings and short anal. Head 3.33; depth 5; D. 1,6; A. 29; V. 8; P. 16; Br. 10; eye 3.5 in snout, 6.3 in the head, 4 in the interocular space ventrad. Head wedge-shaped, much depressed, the profile nearly straight and but little ascending; snout parabolic, the gape of the mouth very long, about three times as long as the eye; fontanel extending to the anterior margin of the eye, its groove to the posterior margin; premaxillary band of teeth half as wide as the eye, of nearly uniform width to near the posterior end of the band; maxillary barbel not reaching the rictus; gill-membranes free to below middle of the eye, the distance between the clefts not equal to the eye; distance from snout to dorsal spine 3 in the length; caudal emarginate; origin of anal but very little nearer base of pectoral spine than to base of middle caudal rays; first dorsal and pectoral rays n