ov ates ma or ne tt Pa e a me es ae 5 Ans Minty Warr Lip Uy peopl s “ we oy ge n Y # 4 4 TT f : . ~ f. - > pas ” NMV$Y Le, Cf1UF AGUAS? F es — - F 7 es ho an Pe SOU bf thm J if THE PYGIDIIDAE, A FAMILY OF SOUTH AMERICAN CATFISHES IATA By Cari H. EKIGENMANN MBL/WHOI A A 0 0301 OO17bLLO 8 Reprinted from the Memorrs or THE Carnecige Museum, Vol. VII, No. 5, September, 1918 19is8 Sats rns a rN MEMOLRS CARNEGIE MUSEUM. Vou. VII. No. 5. THE PYGIDIIDH, A FAMILY OF SOUTH AMERICAN CATFISHES.! By Cari H. HIGENMANN. (PLates XXXVI-LVI.) INTRODUCTION. The Pygidiide are a family of South American ecatfishes distinguished exter- nally by the absence of an adipose fin and by the posterior position of the dorsal. Most of them are even more readily distinguished by the presence of spines or thorns on the opercle and interopercle, by twin barbels at the angle of the mouth, and by the absence of all mental barbels. Other characters of the catfishes may be present or absent, and by addition, subtraction, or modification of characters, various subfamilies have been formed. The genus Pariolius may be related to Nematogenys, but it is more likely to be related to Phreato- bius, Heptapterus, Myoglanis, Leptorhamdia (for Leptoglanis which is pre-oecupied), ete. The only speci- men recorded has been lost. I. PARIOLIUS* Cope. Pariolius, Corr, Proc. Acad. Nat. Sci. Phila., 1871, p. 289. Type.—Pariolius armillatus Cope. Similar to Pygidiwm; no nasal barbel; a single barbel at the angle of the mouth; two pairs of mental barbels; no armature on the opercles; gill-openings wide; teeth brush-like; origin of the dorsal behind that of the ventrals; anus under dorsal; anal short. Little can be said about the relationship of this genus until its skull and air-bladder are examined. It appears to be closely related to some members of the Pimelodinw. There are no specimens available for examination. It is known only from the type of the species, and that has been lost. Habitat.—Basin of Peruvian Amazons. Pariolius armillatus Corn, l. c. (Ambyiacu); Era@enmann & E1GENMANN, Proc. Cal. Acad. Sci. (2), H, 1889, p. 50; Occasional Papers Cal. Acad. Sci., I, 1890, p. 324; Proc. U. S. Nat. Mus., XIV, 1891, p. 36; EiGENMANN, Reports Princeton Univ. Exped. Patagonia, III, 1910, p. 398. “Head flat rounded, eyes small, superior, covered by the skin. Head 4.5 times in length to basis of caudal fin. Depth at D. I. one-half length to basis pectoral fin; width of head two-thirds the same dis- tance. Interorbital width 3.66 times in length of head. Maxillary and external mental barbels extending beyond basis of pectoral; inner mental barbel one-half the same. Radii D. 7; P. 8; V. 6; A. 11; caudal acuminate. Skin entirely smooth.”’ *T am not sure of the origin of this generic name. Is it from the proper name Parioli, or from mapawoniv = to trick, hence a trickster or simulator, or from rapa = with, and ados = speckled? EIGENMANN: THE PYGIDIIDA, A FAMILY OF SOUTH AMERICAN CATFISHES. 261 Nearest the Pygidiine are the Pareiodontine, which lack a nasal barbel. The teeth are very peculiar and in a single series (See Fig. 21) and the gill-membranes are attached. But one species is known. The Nematogenyinw, Pygidiine, and Pareiodontine are free-living, and have a terminal or subterminal mouth and pointed or incisor teeth. The Stegophiline, Vandelliine, and Tridentine differ widely from the members of the above-mentioned subfamilies in structure and habit. They are all small or minute; the mouth is inferior; the head flat below; the lower barbel at the angle of the mouth is minute. The jaws are weak, the teeth absent or slender. The gill- opening, in all but Acanthopoma, is greatly restricted, which, put in terms of habit, means that the mouth is suctorial. Some of them are parasites, or commensals. The Tridentine differ in having the anal fin much longer than the others. Nothing is known of their habits and they are so small (the largest known specimen is but 27 mm. long) that it is a wonder that any of them have arrived in the bottles of the naturalist. In the Vandelliine the teeth are reduced to a minimum, and the rami of the lower jaw do not meet in the middle. The differences between the genera are minute, but well marked. The habits of these fishes, as well as those of the next subfamily, are discussed below. The Stegophiline have a very large number of minute teeth in definite series in both jaws. The rami of the lower jaw meet in the middle. The genus Acantho- poma stands out in that its gill-membranes, while united, are free from the isthmus. The genus, Henonemus, is well marked by the small number of opercular spines. Ochmacanthus has numerous accessory caudal rays above and below, which make the tail look like that of a tadpole. The remaining genera, Stegophilus, Homo- dietus, and Pseudostegophilus, might well be united. They differ from each other largely in the position of the ventrals, the shape of the caudal, and in the number of accessory rays. Hapits. The habits, as well as the distribution, of various members of the Pygidiide have been derived from the general tendency of the catfishes to get under banks, under logs, out of the way, and out of sight. This general tendency has been modified into the specialized, insinuating habit of the Pygidiide, for which the opercular spines and the eel-like body are adaptations. On the plains of Bogota the Indians secured the largest specimens of Hremo- philus by thrusting their arms to the bottom of holes in the banks of streams. At Honda I found one species buried in the sand in the bottom of the stream, It ° 262 MEMOIRS OF THE CARNEGIE MUSEUM. would dart from its hiding place as I raked my fingers through the sand, to dart into the sand again much like a lancelet or young lamprey, or to dodge under a rock. Mr. KE. B. Williamson wrote me that he noticed another species clinging to the vertical sides of a waterfall. It looked like a water-weed, but he found by watching closely that every little while the supposed weed would move up the wall a short distance, and by using his butterfly-net he secured specimens. It is this habit assisted by the opercular spines that accounts for the fact that the species of the genus Pygidium are found in every mountain-stream. The habit of insinuating themselves into crevices is undoubtedly also the starting point of the habit of resorting to the gill-cavities and probably other organs of larger fishes. There is a widely distributed belief among the Indians of the Amazon Valley, that fishes called ‘Candirti’”’ enter the urethree of bathers. Some travelers who have had this habit reported to them have simply dismissed the matter as absurd. Others have made attempts to identify the fish with results that have not always been fortunate. The native name, Candirt, is applied to some fishes (Cetopsis of the Cetopside) at least a foot long, and at least two inches thick, as well as to minute slender fishes, species of Vandellia, which might enter the urethra without violating the law that the greater cannot enter the less. The habit has been attributed to the large Cetopsis, to Pareiodon, more moderate in size and yet too large, and to some species of Vandellia and to Acanthopoma. It is, of course, possible that the young of the larger Candirtis have the urinophilous habit. It is also possible that the Indians consider the small Candirtis (species of Vandellia) as the young of the larger Candirtis, members of the genus Cetopsis, which according to the classification adopted, belong to a different family. The habit is also phys- ically possible for the species of Tridens, of Miuroglanis, of Paravandellia, Stego- philus, Branchioica, and for some of the minute species or young of Pygidiwm. However, these have not been indicated as being Candirtis. As far as I am able to find, the first notice of the peculiar habit is given by Spix (Selecta Genera et Species Piscium, 1829, p. vill), who says of Cetopsis: De alio pisce hominibus infesto nonnulla afferre debeo, quem Brasilienses CANpIRU, Hispani in provincia Maynas degentes CANpRO nuncupant. Singulari enim instinctu incitatur in ostia excretoria corporis humani intrandi, quae quum igitur in iis, qui in flumine lavant, attingit, summa, cum violentia irrepit, ibeque carnem morsu appetens, dolores, imo vite periculum affert, Urine odore hi pisciculi valde alliciuntur, quam ob causam accole intraturi flumen amazonum, cujus sinus hac peste abundant, preeputium ligula constringunt, et a mingendo abstinent. Pertinet hic piscis ad Cetopseos, quod depinximus, genus; at nescio, an descriptarum specierum (C. candiru et C. cecutiens) individua juniora, an tertiz cujusdam speciei minoris individua crudeli hoe instinctu a natura sint donata. EIGENMANN: THE PYGIDIIDM, A FAMILY OF SOUTH AMERICAN CATFISHES. 263 I am indebted to Professor Selatie E. Stout for the following translation: I should briefly mention another fish which is dangerous to man. The Brazilians call it Canptrt; the Spaniards in Maynas’ call it CaAnERo. It is impelled by a curious instinct to enter the excretory openings of the human body. Whenever it comes in contact with these openings of persons bathing in the stream, it violently forces its way in, and having entered, it causes constant pain, and even danger of life, by biting the flesh. These fishes are greatly attracted by the odor of urine. For this reason, those who dwell along the Amazon, when about to enter the stream, whose bays abound with this pest, tie a cord tightly around the prepuce and refrain from urinating. This fish belongs to Cetopsis, a genus which I have already described. But I do not know whether it is the younger individuals of the two species which I have described (C. candiri and C. cacutiens), or whether a third species of smaller fishes has been given this cruel instinct by nature. The habit here described by Spix in reality belongs to fishes of which he did not secure specimens. In 1808, Domingo Vandelli, professor of natural history at Lisbon, sent Lacé- péde three small fishes, which he placed with the Loricariidw. They were described by Valenciennes (Cuvier & Valenciennes, Hist. Nat. Poiss., XVIII, 1846, p. 386, pl. 547) as Vandéilia cirrhosa, and placed in their Esoces. Nothing was said of the habits, and even the habitat of the specimens was unknown. The identification of Vandellia with the urinophilous habit came later. Castelnau in his Animaux d’ Amérique du Sud, Poissons, 1855, says of his Trichomycterus pusillus = Pareiodon microps Kner: Cette espéce est, de la part des pécheurs de l’Araguay, l’objet d’un préjugé des plus singuliers, ils prétendent qu’il est trés dangereux d’uriner en riviére: car, disent ils, ce petit animal s’élance hors de l’eau et pénétre dans l’uréthre en remontant le long de la colonne liquide. As this species reaches a length of at least six inches and a corresponding thick- ness, Castelnau was probably mistaken in the species acting in this remarkable manner. It seems that Paul Marcoy (Voyage a travers Vv Amérique du Sud, Vol. II, p. 145-147) gives an account with a figure of a Candirti. I have not seen this book, but Liitken says: ‘‘ Etude de Candirti signeret med den Rejsendes Initialer, er en fuldstandig Umulighed, hvad den saa skal forstille’’(Vidensk. Meddel. Naturh. Foren. Kjobenhavn, 1891, p. 60). Lange ‘In the Amazon Jungle,” p. 214, says: In fact, throughout the Amazon this little worm-like creature, called the kandiroo, is so omnipresent that a bath-house of a particular construction is necessary. The kandiroo is usually 3 Probably Maind, an Igarapé tributary to the Amazon, near the Rio Negro; or a province of Peru with Moyobamba for its capital. 264 MEMOIRS OF THE CARNEGIE MUSEUM. three to four inches long and one-sixteenth in thickness. It belongs to the lampreys, and its particular group is the Myxinos or slime-fish. Its body is coated with a peculiar mucus. It is dangerous to human beings, because when they are taking a bath in the river it will approach and with a swift, powerful movement penetrate one of the natural openings of the body, whence it can be removed only by a difficult and dangerous operation. A small but hard and pointed dorsal fin acts as a barb and prevents the fish from being drawn back. While I was in Remate de Males the local doctor was called upon to remove a kandiroo from the urethra of a man. The man subsequently died from the hemorrhage following the operation. The Candirti does not belong to the lampreys and its particular group is not Myxinos. The lampreys are not found in the Amazon Valley. Its dorsal fin is neither hard nor pointed, and hence cannot act as a barb to prevent the fish from being withdrawn. The retrorse spines on the interopercle and opercle are the obstacles which would prevent it from being withdrawn. The question naturally arises: Is Lange more trustworthy in his account of the habit than in his account of the structure and relationship of the Candirti? The only known specimen of Acanthopoma annectens, another Candirt, seems to have been collected by Gustav Wallis. In an article, “ Mittheilung von C. Miller iiber die Reise von Gustav Wallis” in Die Natur, Zeitung von P. Ule u. K. Miller, XIX, No. 23, p. 180, mention is made of the habits of presumably this species, though it may have been drawn from the general report given the traveler concern- ing the Candiri. Liitken quotes: In diesen noch so wenig bekannten Gewiissern, namentlich im Hualliga, beobachtete der Reisende (G. W.) einen Fisch, den ich der Aufmerksamkeit der Wissenschaft ganz besonders empfehlen will. Man nennt ihn dort den Candiru und fiirehtet ihn mit Recht ebensosehr fiir das Gebiet des Wassers, wie man fiir das des Landes die Moskitos und Ameisen fiirchtet. An sich selbst ist es nur ein kleines, kaum .75 Spannen langes Ding von welsartigem Korperbau, mit breitem, abgerundetem Kopfe, auf dem die beiden Augen ziemlich dicht neben einander liegen, wiihrend die beiden Brustflossen fliigelartig dicht unter ihm sich ausbreiten und der wtbrige Korpertheil keilformig zulauft. Den Riicken ziert eine dunklere Farbung mit undeutlich ver- laufenden Flecken, so dass das Geschépfehen an sich selbst kaum irgendwie durch eine hervor- ragende Eigenthiimlichkeit ausgezeichnet ist. Eine umso schrecklichere Plage ist es fiir den Badenden, eine Art Blutegel nimlich, der mit unglaublicher Schwimmfertigkeit jenem zu Leibe eeht, ihm tiberall schrépfkopfiihnliche Wunden beibringt und, wenn es ihm gelungen, sich dadureh an dem Kérper festzusetzen, in der Wunde ein Nadelbiindel ausspreizt, an dem er wie an Wider- haken sich derart festklammert, dass er nur durch eine schmerzhafte Operation aus dem Kérper entfernt werden kann. Diese Unart des Fisches ist umso grésser und gefihrlicher als er am lieb- sten die geheimsten Kérpertheile aussucht; man erzihlt sich Fille, die bei der Operation mit dem Tode endeten. Ich werde dafiir Sorge tragen, dass dieser seltsame Fisch, den ich in Spiritus vor mir habe, in die rechten wissenschaftlichen Hinde gelangt und seinen wissenschaftlichen Namen empfingt, den er noch nicht hat. EIGENMANN: THE PYGIDIIDM, A FAMILY OF SOUTH AMERICAN CATFISHES. 265 Boulenger (Proc. Zodl. Soc. London, 1897, p. 901) says of Vandellia cirrhosa, the urinophilous Candirt, par excellence: The “Candyrt,” as the fish is called, is much dreaded by the natives of the Jurua district, who, in order to protect themselves, rarely enter the river without covering their genitalia by means of a sheath formed of a small coconut-shell, with a minute perforation to let out urine, maintained in a sort of bag of palm-fibers suspended from a belt of the same material. The fish is attracted by the urine, and when once it has made its way into the urethra, cannot be pulled out again, owing to the spines which arm its opercles. The only means of preventing it from reaching the bladder, where it causes inflammation and ultimately death, is to instantly amputate the penis; and at Tres Unidos, Dr. Bach had actually examined a man and three boys with amputated penis as a result of this dreadful accident. Dr. Bach was therefore satisfied that the account given of this extraordinary habit of the ‘““Candyrti”’ is perfectly trustworthy. Mr. Boulenger further showed a photograph, taken by Dr. Bach, of two nude Indians wearing the protective purse. It is to be noted here that this evidence is only circumstantial. Dr. Bach did not himself operate or help to operate to remove the Candirti and a much simpler operation than amputation would be sufficient to remove it. The literature on the evil repute of members of the Pygidiide has been re- viewed by Pellegrin. In Bulletin Société Philomathique de Paris (10), I, 1909, pp. 101-104 [5-8 of the reprint], he says: Le Dr. C. Jobert qui accomplit au Brésil, en 1877, un voyage ot il rassembla des matériaux ichthyologiques considérables, a consacré 4 la question du Candiru un mémoire des plus docu- mentés, oti il n’admet pas sans réserve les déclarations du practicien américain cité par G. A. Boulenger. ‘Le Dr. Bach,’ écrit-il, ‘n’a pas vu le petit Poisson in situ; la chose est regrettable et, cette fois encore, nous ne sortons pas du cercle de la légende.’ Toutefois, le Dr. Jobert rapporte les dires d’un médecin trés estimé de Belem (Para), le Dr. Castro, qui lui affirma avoir extrait de luréthre d’une négresse un petit Candiru qui y avait pénétré pendant la miction, alors qu’elle se baignait en riviére. Mais ce qui fait le grand intérét de Varticle du Dr. Jobert, ce sont les renseignements qu'il a pu lui-méme recueillir sur place au Brésil au sujet des Candirus. Les Paraenses en distingueraient deux espéces, l'une petite, qui s’introduirait dans Vuréthre des baigneurs, l’autre de plus grande taille, ‘trop grande pour tenter ces mémes opérations, mais redoutable par les blessures qu’elle fait sur n’importe quelle partie du corps. On donne a cette derniére le nom de Candiru de Cavallo et les indigénes prétendent qu’elle attaque les chevaux pendant la baignade.’ Au sujet de celle-ci il rapporte en outre les faits suivants: ‘Un jour, 4 un mille environ en aval de Para, je voulus me baigner sans souci des Candirus qu’on m’assurait étre trés abondants en cet endroit. Je n’étais pas dans l’eau depuis cing minutes que je ressentis dans le région lombaire, au ventre, sur le cétés de la poitrine, comme de légers coups de griffes qui se succédaient rapidement. Voyant leau se teinter de rouge autour de mot, je me hatai de regagner le rivage et je constatai que, dans le région ou j’avais éprouvé la sensation 266 MEMOIRS OF THE CARNEGIE MUSEUM. de ces coups de griffe, le sang s’échappait de blessures en scarifications paralléles, qui eussent pu étre attribuées A un instrument, tant elles étaient réguliéres; elles constituaient des groupes de 5 A 6 lignes, longues d’un centimétre au plus et trés rapprochées; je n’ai pas cherché a apprécier a profondeur, mais ces blessures trés étroites saignaient abondamment.’ Les Poissons qui ont attaqué ainsi le Dr. Jobert appartiennent suivant moi, incontestablement au genre Vandellie, peut-étre méme A l’espéce Vandellia Wieneri. Si l’onse reporte a la descrip- tion donnée plus haut de la bouche et de l’appareil operculaire, on s’expliquera ainsi facilement le fonctionnement de ces divers organes; on comprendra aisément que la demi-couronne de dents en crochet placée en avant de la bouche, dents susceptibles d’un certain degré d’érection et au nombre de 5 46 principales produit ces scarijications paralléles, réguliére et en groupe de 5 a 6 lignes. Les épines interoperculaires du dessous de la téte, aussi un peu érectiles, peuvent également, dans une certaine mesure, déchirer les téguments, mais elles doivent sourtout servir 4 la fixation. Quant aux épines operculaires du dessus de la téte, elles me semblent plutot, étant donnée la direction de leur pointe, ‘destinées 4 faciliter la progression de |’animal et 4 empécher tout recul lors-qu’il s’engage dans un conduit étroit, par example entre les lamelles branchiales des Platy- stomes. Sans vouloir trancher la question de la pénétration des Vandellies dans l’uréthre, pour laquelle je ne puis apporter des documents nouveaux, il me parait tout au moins démontré en rapprochant les détails anatomiques que j’ai pu constater sur les Vandellia Wieneri, des observations faites sur lui-méme au Brésil par le Dr. Jobert, que les Candirus, véritables Poissons-sangsues, ne sont pas, ainsi que le pensait Giinther, de simples commensaux des grands Siluridés sur lesquels ils vivent habituellement; leurs dents et leurs épines operculaires et interoperculaires permettent. non seulement de se fixer sur les branchies de leur héte, mais aussi de faire des blessures amenant un écoulement de sang abondant qu’une disposition spéciale leur permet d’ingurgiter. Enfin a état libre, comme la constaté le Dr. Jobert, les Vandellies ne craignent pas de s’attaquer a Homme, dont elles percent les téguments, ce qu’elles font aussi certainement sur certains Mammiféres domestiques. I] y a lieu en terminant de noter que les dents volumineuses peu nombreuses, en forme de crochets acérés de la machoire supérieure, sont particuliéres au genre Vandellia, qu’elles sont absentes dans les genres voisins Stegophilus Reinhardt et Acanthopoma Liitken, ot elles sont remplacées par une bande de trés nombreuses petites dents acérées.* Les Vandellies représentent donc, chez les Siluridés, le dernier terme de la spécialisation en vue d’un parasitisme des plus caractérisés. That fishes found in the Amazon Valley and called Candirtis are a nuisance is certain. Whether the widely prevalent belief that the Candirti is tropic to urine, and consequently has a tendency to enter the urethra, or whether the Candirti’s tendency to burrow leads it accidentally to enter the urethra, are all matters that must for the present remain in debate. A very interesting subsidiary question is, whether, if Candirtis are tropic to urine they do not also enter the 4 While members of the Stegophilini have bands of minute teeth uniform in size in the upper Jaw there are frequently a few elongate, slender teeth in the middle of the upper jaw, which are similar and correspond to those of Vandellia. EIGENMANN: THE PYGIDIIDM, A FAMILY OF SOUTH AMERICAN CATFISHES. 267 urethree of aquatic mammals and of large fishes. Further study may demonstrate that some species of Candirtis have become parasitic in the bladders of large fishes and aquatic mammals. These are all questions that may legitimately be taken up by future expeditions. The first of the commensals or parasites of this family to be described is the Stegophilus insidiosus of Reinhardt. Reinhardt secured all of his specimens from the gills of the giant catfish of the Rio das Velhas, a tributary of the Rio San Fran- cisco. Haseman secured one specimen of this fish from the sandy island opposite Januaria, near the mouth of the Rio San Francisco. The fish therefore may and does live in the open as well as in the gill-cavities of larger fishes. The account of Stegophilus insidiosus Reinhardt, given by the author of the genus and species, which was published in 1858 (Cf. Naturhistorisk Forenings Videnskabelige Meddelelser, Copenhagen, 1858, reprint, pp. 1-19, Pl. IT) possesses great interest. Professor Reinhardt having been repeatedly informed that a large species of catfish, belonging to the genus Pseudolatystomus and known by the natives as Sorubim, protects its young by carrying them in its gills, determined, if possible, to verify the statement. An English translation of a portion of his narrative is here elven: It deeply interested me to ascertain with exactness the circumstances under which this pecu- liar method of protection takes place, and also to examine the young at the time when they make use of it. I therefore offered the fishermen in the vicinity of Lagoa Santa, where I was staying at the time, a good sum if they would bring me a Sorubim with some of its young in the gill-cavities. Finally on February 27, 1852, a fisherman brought me one, in the gills of which he said there should be a little “young one.’’ On examination I indeed found there a young fish, hardly an inch long, which was already dead, although the Sorubim still showed faint signs of life. The little fish looked so unlike the big one that I was astonished, and upon finding out that the old fish was a male I was strengthened in my doubt as to their relationship. When the same fisher- man two days later again brought a male Sorubim with a young one, which looked exactly like the first, but was about three times longer, it became clear to me that these two small fishes could in no wise be what it was claimed they were. On the other hand they recalled to me the picture I carried in my mind of a Trichomycterus which I had obtained one year previously from the Rio das Velhas under the name of Cambeja, or Bagre molle. I naturally concluded that the fisherman in order to get the reward offered, had brought me the young of this Cambeja and was passing them off as the young of the Sorubim. I complained to his face about this procedure, and, though I did not obtain any confession from him, I nevertheless had no doubt that I had been made the victim of a swindle. During the few weeks I still remained in Lagoa Santa before start- ing on my homeward journey to Europe, nothing happened to induce me to think otherwise. Upon my return home, as soon as I could get access to the literature, and could make a direct comparison between the supposed young of the Sorubim and the Cambeja, I at once saw that I had made a mistake in assuming that the former were the young of the latter. In short these 268 MEMOIRS OF THE CARNEGIE MUSEUM. so-called young of the Sorubim were the little fishes which I have had the honor of exhibiting to the Society. The whole matter became more involved and enigmatical to me, because it appeared that the fisherman, if he had been really guilty of an intended fraud, had for this purpose made use of a fish which was so rare that I had never found it, although I had collected great quantities of the various small fishes in the waters around Lagoa Santa; in fact a fish which I was forced to conclude to be as difficult to obtain as the real young of the Sorubim. In 1854, when I again visited Brazil, the solution of the riddle was one of my especial aims. Soon after I arrived at Lagoa Santa in the latter part of November I indeed reached the solution much more quickly than I had expected, and in the following manner: A person from the vicinity of Lagoa Santa, but not the same one, who almost three years before had brought me the first: Stegophilus, came to the village on a Sunday in the middle of December to attend mass according to the custom of the country. He brought with him on this occasion a Sorubim, which before he went to church he sold to a Frenchman who had a shop in the town. When mass was over he returned to get his pay, and watched the shopkeeper cut the fish into pieces. He remarked that when the fish had been pulled out of the water there had been five young in its mouth, of which two had remained inside. The shopkeeper looked and actually found the remaining ‘“‘ young,” and was kind enough, as he knew the matter would interest me, to immediately bring them to me and relate the circumstances. At the very first glance at the so-called ““young” I saw to my surprise that again Stegophili had been brought me as the young of the Sorubim. That deception should again be at the bottom of the matter appeared in the highest degree improbable. It could hardly be thought of, except upon the assumption that the person who had sold the last Sorubim was in collusion with the fisherman who during my previous stay, three years before, had brought me the first two Steg- ophili. How could it be explained that both had conceived the idea of passing off the very same fish as the young of the Sorubim, and that a fish, which has no particular resemblance to the latter? But, even if there had been collusion, would it not have been more likely that the first party con- cerned would have come directly to me with his “Sorubim young,” instead of leaving it to be more or less of a chance whether or not they should fall into my hands? Even if a trick, prearranged to allay a possible suspicion, were thinkable, nevertheless it was hard to believe that under the existing conditions the parties involved would have taken the time and the trouble to deceive me, unless they had expected to reap advantage from their effort. If a trick had been planned in the present case it was entirely aimless, as no pay was either asked, or given, for these last “young Sorubim”’; and neither the last person, nor any one else, came at a later date to offer me “Sorubim young.”” There was therefore left for me no other alternative than to conclude that I had been unjust in my suspicion in the case of the fisherman who on the occasion of my previous stay had brought me the first Stegophili. In other words, this little fish in reality passes into and abides in the gill-cavities of the Sorubim. Its presence there has through an easily explained mis- interpretation on the part of the common people given rise in Minas to the story about the Sorubim’s care for its young. The second species, Branchioica bertonii, known to inhabit the gill-cavities of larger fishes, is recorded in the present volume. It really belongs to the Vandelliine. One specimen was sent me several years ago by Sn. A. de W. Bertoni from Puerto EIGENMANN: THE PYGIDIID®, A FAMILY OF SOUTH AMERICAN CATFISHES. 269 Bertoni, Paraguay. Later he sent me two more specimens, all three having been taken from the gills of a large characin, Piaractus brachypomus. Ribeiro, of the National Museum of Rio de Janeiro, caught another very similar member of this subfamily, Paravandellia, among the water-weeds of the stream near San Luis de Caceres, in the Upper Paraguay basin. With fishes as rare as these and as small as these, the question arises whether two species are really different, or whether the described differences are due to the fact that one worker uses a hand lens, and the other a binocular dissecting micro- scope with an are spot-light. The results of the two instruments are comparable to the effects produced by an old-fashioned cannon and a modern forty-two centi- meter howitzer. Branchioica and Paravandellia may prove to be synonymous. DISTRIBUTION (Plates XXX VI-XX XIX.) In considering the distribution of the fresh-water fishes of South America I found, among other things (Proc. U. S. Nat. Mus., XIV, 1891, p. 18) ‘“‘that genera of many species usually have a wide distribution, and conversely, genera of wide d distribution usually have many species.’ With one exception the number of species of any genus of the Pygidiida varies directly with the greatness of the area over which it is distributed. Some genera consist of but one species, and that restricted to but one, or a few neighboring localities. As far as known, Hremophilus is all but confined to the plateau of Bogots, Scleronema to the center of the Uru- guay basin, Acanthopoma to a part of the Huallaga basin, Stegophilus to the Upper San Francisco basin, Paravandellia to the Upper Paraguay basin, Branchioica to the Lower Paraguay basin. The genera with more than one species invariably have a wider distribution. Homodiatus, with two species, is limited to the lower and central La Plata basin, Henonemus, with four species, to the Amazon basin, Hatch- eria, With six species, to the Andes of central and southern Argentina and Chile, and Pygidium, with sixty-three species, is found in all the mountain streams from the Tuyra in southern Panama to central Chile and central Argentina, in the moun- tain streams from Rio Grande do Sul to the Rio Sao Francisco, and sparingly in the lowlands of Guiana and Brazil. The only exception to the general rule is Och- macanthus, with three species, ranging from Guiana to Paraguay. The Pygidiine are mountain forms, and while they are found in lowlands near the mountains, we find the optimum in the plains of Bogoté and in Lake Titicaca. They are sometimes the last species to succumb in the struggle with adverse con- ditions found in high altitudes, and they range further south (to latitude 47° 30’ ), than any other tropical American fishes. 270 MEMOIRS OF THE CARNEGIE MUSEUM. The Stegophiline, Vandelliine and Tridentine are essentially lowland forms, although some species reach considerable elevations. CHRONOLOGY. The first species of the Pygidiide discovered was taken by Humboldt at Bogota, and described in 1805 (Recueil d’Observations de Zodlogie, ete., pp. 17-19, pl. VI) as Hremophilus mutisti. The habit of one of the species was next described by Spix in 1829, but attri- buted to a member of another family. See page 262. The most prominent genus was first described by Meyen (Reise, I, p. 475, Wiegm. Arch. Naturg., 1835, II, p. 269) as Pygidium. I have at diverse times defended the name, Pygidium, as against the name Trichomycterus and its varia- tions. The various generic names and their present equivalents are given in the fol- lowing table: Name Proposed. Proposed in Present Equivalent. Bremophilus, Humboldte asc. .: 2 = ete USD 3 6 oo< Eremophilus Humboldt. Thrichomycterus Cuvier & Valenciennes........- 1805 eee Eremophilus Humboldt. Trichomycterus Valenciennes...............--- 1833......Pygidiwm Meyen. Vandellia Cuvier & Valenciennes..............1846...... Vandellia Cuvier & Valenciennes. Thrychomycterus Cuvier & Valenciennes. .......1846...... Pygidium Meyen. Dhrichomyctenus Gibardsee nace se cies sels es coos ok Pygidium Meyen. IPanetodonSMenrsvcpn acta ashlar: 2 ae ee aes IC Waes ace Pareiodon Kner. Gentronhorus het eer heen reer er eee S55 e eet Pareiodon Kner. StegophiltussRembardtige erect es teint S58 meer Stegophilus Reinhardt. Astemomycterus Guichenot.................... S60 eee Pareiodon Kner. Parcodonn Gi here tras see yt) ys ster cine S647 -Ly: Pareiodon Kner. TrachypomaiGiebele anes ae sais eee ee eee Olle eine Eremophilus Humboldt. Tridens Higenmann & Eigenmann..........-.- 138 ORs Tridens Higenmann & Eigenmann. Pseudostegophilus Kigenmann & Higenmann ....1889......Pseudostegophilus Kigenmann & Kigenmann Miuroglanis Eigenmann & Higenmann.........1889...... Miuroglanis Kigenmann & Eigenmann. Alcanthoponcamlt Ke Oivartntitertsten cis err hiare USOT ye Acanthopoma Liitken. Homodietus Higenmann & Ward...............1907...... Homodietus Kigenmann & Ward. Henonemus Eigenmann & Ward..............- 190: Snes Henonemus Kigenmann & Ward. Hatcherta Bieenm anne este cei eh IQ ee eade Hatcheria Eigenmann. Ochmacanthus Higenmann................-.-.-- 1 eas ose Ochmacanthus Eigenmann. Gyrinurus Ribelrol seeeescr ee tee ee (OU 2 ees: Ochmacanthus Higenmann. Poravandellia Ribeiro. -ee ee ee ee eee TOU Pele ass Paravandellia Ribeiro. Cobitoglanis Howlers .- eo cee ee ae NOMA weer Henonemus EKigenmann. Urinophilus Higenmann..................-.-- Ig 5 6 oe Urinophilus Higenmann. Branchioica Bigenmanner. ss see eee eee a 1917......Branchioica Higenmann. EIGENMANN: THE PYGIDIIDH, A FAMILY OF SOUTH AMERICAN CATFISHES. 201 LOCATION OF THE TYPES AND SPECIMENS IN THE MUSEUMS OF THE WORLD. The species are, for the most part, but little known. Over forty of the ninety- five recorded species are known only from the types, which are widely scattered. Ten or twelve of the types are in Vienna, two are in Berlin, eleven or twelve in Paris, eleven in London, one in Torino, Italy, two possibly in Munich, one in the University of Leipzig, two in Copenhagen, three presumably in Santiago, Chile, three in Buenos Aires, five in Rio de Janeiro, two in Cordoba, Argentina, one in the Field Museum, two in the Philadelphia Academy of Sciences, eight in the Museum of Comparative Zodlogy at Harvard, eight in Indiana University, twenty-four in the Carnegie Museum, one in Princeton University. The Carnegie Museum pos- sesses forty-six species, Indiana University is next in line with thirty-two species, and the Museum of Comparative Zodlogy comes third with twenty species. The distribution of the known specimens in the various museums of the world is given in the following table: | l l Sie we | | . |8| & = | | BS 5| & : | q | | s | 35 S = Jj d S| Sti @ eI een] ene | EE Pill opi aese ode Sie meeacet | es a neces 3 I 3 o = ai {. S a = > a a o 5 a ) D — Ss mn 5 8 oe S 1 RD oD = 4 ~ ay oD S i>) a | é Z > 5 4 qa =| A 43 As} ro) 5 a) a S =I = Py < Ss da 3 3 oe = By Ss a S Pit beta 2 sao iS ast Zi ts a cS acs a EI = a | & g & . zi | cao as | by FI zB 2 eae ears || > = ZS |o|l 2 = = ieee al Pesci ese wll Gk WW \iesee |) mee “en (Pe coo Uh al Cera > Q ey 2) a 2) er 5 fe eq |) | ost & 5 1. Nematogenys inermis. type a | YI | | 2. Scleronema opercula- | | | WI co sade aelo Ono type 3. Hatcheria patagonien-| | SUE ley Reet | | | | heal ap || ak 4. Hatcheria maculata. .. type, ? |} + 5. ss titeombi ... == type 6. es areolatum. . type| + ar Sis ie 5 burmeistert.| = type, | 8. macret....| | | | type + “7° | | | 9. Pygidium marmor- | CHUM isp codoece cr type?) | 10. Pygidium pallewm ...| itype?| hen | 11. = tigrinum ..' 'type?| | 12. xf tenue®.....| | ibs eS corduvensis® aia 14. af Spegazzinir . type 1B}, i borellit™ | type ae | 16. - eichornia-_ | RUT Ur ga Ae oR eae | be | typel | ae 5 Type in Princeton University. 6 Types in Cérdoba, Argentina? 7 Types in Mus. Uniy. Torino, Italy? 272 MEMOIRS OF THE CARNEGIE MUSEUM. ial mole lees |e EI a | s| 8 & ; Bis sel cealltege Peel bulma yeaa ieee = LB 2 3 Bet a s 2 § i) So Ss 7 SRE A eerece le ee eee less | 2 SCM eae eee cs em S| ce eee eee se hs BO | eae Se aT Oe eed ee erate eect lta ee cs alla Salsa eee lee a) coos ee eel I Se SP eMeesh hE tS 17. Pygidium riojanum . type 18. “y heterodon- Wie oeobin arco mes type} - 19. Pygidium fuscum. . type 20. oe eigenmanni Ff type 21. vittatum. type 22. 2 dispar .... type + Doe punctula- tum . type f + + 24. P: ygidium taczanowskit type| + 25. rivulatum . type; + ; + ap |) SF 26. oe poeyanum . type Wife TR se, San type "28. 4 barbourt . . type| + 29. oroye..... +| type] + 30. quechuorum) type Gille sf laticeps®. . . |type? ar 32. uf stellatum . . + 33. oe chapmani . + 34. st tenium®. . . \type? ae 30: ue caliense... a 36. e latidens . .. type 37. f UNAMID 0 a0 6 type 38. i straminewm + 39. o unicolor. . . type 40. es [HOGG 350.0 type ote 41. 4 merid@ ... type aP 42. se bogotense. . ao + 43. Pygidium nigroma- culatum . ee type ain = 44. Pygidium banneaui.. ar 45. es sptlosoma . type 46. ‘i dorsostri- VOT [Oi CRE RIG ho Ba +P 47. Pygidium venulosum. | type 48. ne latistri- GLU eee eee eee 49. Py ygudium striatum type AF a 50. regani.... type 51. ee retropinne : type 52. guianense . Des sy conradi.. . alg 54. 3 gracilior . . 55. y amazoni- cum. . .| type 56. P ygidium hasemani. . ar 57. Nigricans . . type 58. fe theringi.. . type 59. se zonatum . . _ 60. fe proops.... type S618 a paolence . . 62. as reinhardtt . 63. ef davist..... ® Types possibly in Munich. EIGENMANN: THE PYGIDIIDA, A FAMILY OF SOUTH AMERICAN CATFISHES. 273 = |) q A 8 o = a ta oe 5 2 g eI aA 5 a S & e eS a les a ea Se eee etealines 2 lewceun ly oct Smee gee ete |i getl he cee call = aati) = = 5 a = a EI me | < 5 ail d= 2 8 - | isa) Ay i) n ies} ion] P cy Sechelt aia . 291 RePDIGRENUM ee Nee Was hoon yooh. oye a aes oot cee wie oa tees a 291 1 FEO en RE ee Ore ee (Text figure 7) 292 Na DECOTLOUIDENS Cas << «2s bak SSA Re HE Sa ee Bee (Text figure 8) 293 GUE ASDCQULZUUUL 5 ccs nah» Helge a tne eet os Be oe Deena atin 294 POM PING O TCH cob cca 6d oe ui SG IIE Fy ee a ne gee 22 CICNOTTAATIUNM <«..¢ «one sdi eee) ete es (Plate XLIV, fig. 3) 295 QR PR. riojanium Bere. +... ip ene «, valor ano ciegipans 4 aawiete ai: 295 396 MEMOIRS OF THE CARNEGIE MUSEUM. PAGE 10) Ps hetenodontumpasr eee oe ee eae Ae ne (Plate XLIV, fig. 4) 296 Nils Pe PUSCUIT oak aro SR te eerie ot oa eee 298 1D EPR ANI (2 (LTH Oe ae Bees SRP NRT Ate co och ls, God, 3s Re, Sees oe 298 UB Pe ULAR yee RS a el ae yee a ae ee ea ee 299 VAS PIS OTs Spin caver <1 nee ete iy ge: Meme Dee (Plate XLV, fig. 5) 299 15a de mu RCllOtIns o.com rt ea 7 oe ea (Plate XLV, fig. 4) 300 LG PM ACZO NO WSU prteee eee Peta ea (Plate XLVI, figs. 5-8) 300 Ue feked PANTONE Ahaha at osc oes oboe oid See eee ate (Plate XLV, figs. 2 and 3) 301 HS ee OC TOIL TING. tre Re aes ec: Ceres oe gee ee Seer Rae are 302 LQ Rye AU ORO Avtar ee ROR oR nous Aap Pe a (Text figure 9) 303 DAI i age (CESS oh ie ped Seg eon gat 1A Sr ER EERE ROE ie Od nt, EYES 303 PA) Sl EERO) ROI Re At ik A Pall aed ON eat GIR ower OR (Text figure 10) 304 DPMS CMRI ITAA UTI ROR Omer tee tk tok ait. Oaty MRE TEs - (Text figure 11) 305 PAB MNN CAA ICIS) AP as Bice ORE, oe a Pe PMN iso? PE (Text figure 12) 307 AREF ASTELLCLUN AS ce Eee he ere (Plate XLVII, fig. 1) 308 25.3. ChAPMANU «02... (Plate XLVI, figs. 2 and 3; Text figure 13) 309 DG DET AU TH eek ass he NG SSL) OU ee (Text figure 14) 310 DAR 1 ULSTER PO as Chee Oca oot Meet ot tc aS SE (Text figure 15) 311 LOM RICLUGONS A fine che tae NRO ea a eg (Plate XLVII, fig. 4) 312 DOO WEU Re opted ops React id cee ne AME aS os (Plate XLVII, fig. 5) 312 SMES SURMIMUNELLO oe. o Nie aN alae ee (Plate XLIX, fig. 1) 313 She dee: AL TU COU Olen a sae ect Sst La 9 ta oy ee eg ee inet tae 314 OEIC a LPG ae eA EE inten SEDAN Ae (Plate XLVI, figs. 1 and 2) 314 DS eoeg TENT O UE Orca ctl 5, Se eer ah a ee (Plate XLIX, fig. 2) 315 BL Wl MOOQOLCNMSE 7-3 ieke seer de nee aoe (Plate XLIX, figs. 3 and 4) 315 Sa eTLLOMONUACU LAT cies ee ey nae eee ara (Plate XLIX, fig. 5) 317 rime UO AILILCOIUT «woke pata eaN os cate Reem aren (Plate XLVIII, fig. 1) 318 Seek SDULOSONUE St siac oe ne ne Aan ine (Plate XLVIII, fig. 2) 319 SON ele LOTSOSUUALUTIUD =. cmci os ws Litera etepnde cae (Plate XLVIII, fig. 3) 320 39 oP DEN WIOSUNE: saree oe ee ceed maga eat, eae LL ee eee et 320 AQ UP a latistriGiliny aoe Gok: aaa eee ee (Plate XLVIII, fig. 4) 321 Ae IP SEAL 3s. OF A alkha 2s ee ee ee ee ee 321 ADP WEGUTU. so nee og oie Rs es ae (Plate XLVIII, fig. 5) 323 AB. Pr relrOpinne ee he Re ah ae os ce i SE eC Pa ee 324 AA. P > Guranense: sores nas ee ook ee a oe lait alist) ao2o ADS. IPS COTTOMUS sade oe ee ee are (Plate L, fig. 2) 325 AG. P= QTOCHMIOR 220 eas por ea a eee (Plate L, fig, 3) 326 EIGENMANN: THE PYGIDIID4, A FAMILY OF SOUTH AMERICAN CATFISHES. 397 PAGE Ale [POUT CULO ee (Plate XLVI, figs. 3 and 4) 326 ASS 12. QU SCTULTIG SS, Beta. 3 te ea (Plate L, fig. 4) 326 A OME PRUGTUGOITS ES Are aie ee fic alc Soghye ls S21 et OL Sats ee 329 SO, 15 Vag ian pe Be ees oO ee eee ee a a ee (Plate L, fig. 5) 330 BLS IPs AGRICUL O Speen conan Bie Ona ae Ea ee (Plate LI, fig. 1) 330 LSPS I PPS OVO Oise OA eee (Plate LI, fig. 2; Text figure 16) 331 DORM aD OLCTICOMIE A er, a a eee Ro oes tte te ue (Plate LI, fig. 3) 332 CONE OL IOS Ahk ts ue Gen tee San. eS ee (Plate LI, fig. 4) 333 De ee CLOLUIES TRO Ra eee oe) Oo eh neg to tr Sula 8 Fee SAREE (Plate LI, fig. 5) 334 HOSE UNTINACUHLADUM wate eee rat ease ee (Plate hos) 3o4 BY SE, WO UCUN THis oa hee Siceee he bab ood obos Dad (Plate LII, fig. 2) 335 EN Seam ee CLULCTIVCLEALTIVG AGAR tec ghee Ai che Pe ee eye a (Plate LI, fig. 3) 336 Dy OEI SG OGLE Ie eas Soci MR ane Seed SA ot ES anger La TM ee nets pana ee 337 CORP SbnOStensSCe cs ak oe oe ees (Text figures 17 and 18) 337 Ol laa a er ein ae See ae ee eae (Text figure 19) 339 ODE UG ULC UI Sch tc eee eek ee ee Ee (Plate LIT, fig. 4) 339 63: —P. punciatissimum...............+----..... (Plate XLV, fig. 1) 340 GARE TELL a sis oe Aik entre dae oo ele Be (Text figure 20) 340 65. P. sante-rite. Noe ue ..(Plate LI, fig. 5) 341 Wael CMOphiisits ae. 6212s ese (Plate XXXVI. Plate