.-XT' ^^^ MCZ Memoir V.14 No. 2 '-/ :'^ '-^> ""^^n] •^._-,.^,j^. -^> 4>s: t^llNkk. lilirai-jj of ibi Mtustum OF COMPARATIVE ZOOLOGY, AT HARVARD COLLEGE, CAMBRIDGE, MASS. JFountirli bu pribatc subscviptian, in 1861. Deposited by ALEX. AGASSIZ. Jmoirs of t^« Hustum of Comparalifaj ^oblogg AT HARVARD COLLEGE. Vol. XIV. No. 2. THE DISCOBOLI. CYCLOP TEBIB^, LIPAROPSID^, AND LIPARIDIDJE. By S. GARMAN. ""with thirteen plates. CAMBRIDGE, U.S.A.: Prtnteti for tije Jluscum. "'April, 1892. CONTENTS. Page INTRODUCTION 5 DiSTUIBUTION 7 History in General 9 History of Liparis 9 " LlPARIDID^ 11 " CtCLOPTERUS AND LUMPUS 12 " Cyclopterid.e .... 13 " LiPAROPSIDiE 14 " Discoboli 14 Relations 15 Literature 85 DISCOBOLI Gthr 17 Families and Genera 19 CYCLOPrERIDiE Bon 19 Genera and Species 20 Ctclopterus Linn 20 C. lumptis Linn 21 " Early Stages .... 31 Eumicrotremus spinosus Milll. ... 34 K orbis Gthr 36 Cyclopteroiues gen. n 37 C. gyrinops sp. n 37 LIPAROPSID^ fam. n 40 CYCLOPTERicnTiiYS Steiiul 40 C. ventricostis Pall 41 C. amissiis Vaill 42 LiPAEOPS gen. n 42 L. Stelleri Pall 42 Page LIPARIDIDtE Gill 43 Genera and Species 45 LiPARix^ Gthr 45 Liparis Art 45 L. Montagui Don 47 L. niucosus A\r. 52 i. calliodon Pall 54 L. lip>aris Linn 57 L. antarctica Put CI Zi. Agassizii Put 62 L. tunicatus Reinb 65 i. Steineni Fisch 66 L. pulchellus Ayr 67 L. pallidas Vaill 70 Careproctus Kro}- 71 C. microjnis Gthr 72 C. major Fabr 72 (J. gelatinosus Pall 76 C. Reinhardi Kroy 78 C. longifilis Gann 9 Amitrinje J. & G 80 Paraliparis Coll. . 80 P . rosaceus GiWo '.80 P. bathybius Coll 81 P. Uparinus Goode 82 P. membranacens Gthr. .... 83 P. finibriatiis Garm 9 Explanation of the Plates .... 93 THE DISCOBOLI. CYCLOPTERIDyE, LIPAROPSID^E, AND LIPARIDIDiE. Though there are several of the more rare Discoboli unrepresented in the Museum's collections, it possesses so many duplicates of certain species, in addition to rare or undescribed types of others, that, because of the uncertainty and confusion existing in the literature as to the validity of spe- cies and genera or their affinities, the facilities here provided for a study of the group are in themselves sufficient inducements for the undertaking. But, besides these, many of the drawings now published had been made long ago, by Mr. Eoetter, to illustrate a projected work for the institution to have been written by Prof. F. W. Putnam. At the instance of the Director of the Museum of Comparative Zoology, this paper has been pre- pared to utilize those drawings, and to make the work the more complete he kindly permits the use of a number of di'awings from his own work on the Young Stages of Osseous Fishes. Excepting the outline sketches, the remainder of the illustrations are by the pencil of Mr. J. H. Blake, from specimens herein described. The notes made by Professor Putnam for his intended work were published by him, under the title " Notes on Liparis and Cyclopterus," in the Proceedings of the American Association for the Advancement of Science, Vol. XXII., page 337, 1874. IN GENERAL. The species belonging to this group are provided with a disk, the trans- formed ventral fins surrounded by a marginal fold of the skin, below the anterior portion of the body, for the purpose of attaching themselves to rocks or other objects to prevent being dashed about by the waves or the currents. This adhesive apparatus is usually spoken of as the sucking 6 THE DISCOBOLI. or sticking disk. Its presence on species known to Cuvier induced him to designate them as Discoboles, or disk-bearers. One of the most familiar forms is that of the common Liunp-fish, Cyclopterus lumpus, also called Lumpsucker, Sea Owl, Paiclle, or the male Cock Paddle or Red Lump, and the female Hen Paddle or Blue Lump. In other languages we meet with such names as Lompe, Gras-mollet, See-hase, Bauchsauger, Snotdolf, Havpadde, Steenbider, Rognkjaekse, and Nepisa, applied to the same fish. In shape the Lump is rather short and clumsy, blunt, thick and high forward, slender in the caudal region, roughened all over with tubercular scales, as if for protection when at the mercy of the billows among the rocks, and, as if still more to suit the surroundings, it has flesh that might be described as gelatinous or mucilaginous, and a skeleton that is hardly more than cartilage. Lump-fishes reach a length of a couple of feet, and a weight of twenty pounds. In places they are numerous at particular seasons, but, though found to agree with the stomachs of some people, they are not objects of much consideration in the markets. Several species have been recorded. Others of the Discoboles are more elongate than the Lumps, more like slender tadpoles in shape, with a longer and stronger caudal region, and consequently more force in swimming. The common species of Liparis, L. liparis and L. Montagui, are good representatives of this section of the group. Vernacular names for these are Sucker, Sea Snail, Seeschnecke, and Ringbug. These have so many features in common with the Lumps that their kinship is not called in question. Closely allied in shape and gen- eral structure to the Liparis are a number of species that possess neither sucking disks nor ventral fins, yet the agreement in the other important details of structure, as exemplified by such items as the stomach, the ca?ca, the suborbitals, and the opercles, is so close that there is no doubt of the propriety of retaining them in the group. They are deep-sea species, and it is probable that, descending from the shoals to the ooze-covered floor of the ocean, in leaving the rocks and currents they left behind them the need of the disk, and lost it through disuse. However the loss may have been brought about, it is very evident that we have species without disks that must be included among the Discoboles. Though the Discoboli are not the only disk-bearers among the fishes, they are easily distinguished from others, Gobiidte and GobiesocidiB, which, through possession of an organ similar in location and utility, apparently THE DISCOBOLI. 7 have as much claim to the designation. The Discobole has a suborbital bone that reaches back across the cheek to the preopercle ; it has slender opercular bones and numerous pyloric cceca; and its disk, not always present, includes the transformed ventral rays. Neither Gobioid nor Gobi- esocoid has either the suborbital bone or similarly transformed ventral rays. The adhering organ of the Discobole is formed by modifying the ventral fins; that of tlie Goby is secured by the addition of a dermal fold, a sac, in front of the ventral rays and below the bases of the fins ; and that of the Gobiesocoid is obtained by the growth of a dermal disk between the ventrals and behind the coracoids. It is because of greater variety in their uses that the ventrals are more subject to differentiation than the other fins of these and other fishes. Elsewhere I have pointed out special modifications of the pelvis and ventrals of Selachians, notably on the Potamotrygons of South American rivers, or, more recently, in the union of the two fins to form a single one on Balistes vetula, and may in this place note the peculiar tube-like mittens for the reception of the ventrals of Chaunax pictus when the abdomen is distended. As is well known, on some fishes these fins have become foot-like ; on others, they are reduced to mere filaments ; rarely they are scale-like ; a few have but one of the pair, and others have, like Para- liparis, lost both. The food of the species with which w^e are at present more interested is in part that of most carnivorous fishes, — small fishes, Crustacea, mollusks, worms, and the like. The long intestine of the Lumps suggested that at some portion of the year they might be accustomed to more or less of a vegetable diet ; the contents of the stomachs confirm the idea, and show that vegetation forms a portion of their subsistence. DISTRIBUTION. All of the species are found in the colder waters. Their approaches to the shoals are made early in the spring, for the purpose of depositing their spawn, and they would seem to have returned to the deptlis again before the wintry temperature has been greatly changed by approaching summer. While it is pretty well established that the Lump occurs in the Mediterranean, it certainly is more at home northward from France, along the shores of England, Scotland, Iceland, and Greenland. On our own coasts it has been found as far to the south as New York. A second species, 8 THE DISCOBOLI. called the Spinous Lump, from the North Atlantic, descends to considerable depths. A third and a fourth species of the family Ci/clopteridic are de- scribed from the North Pacific. A couple of peculiar Lumps with a single dorsal fin, Liparopsidce, were discovered in the North Pacific, and recently Professor Vaillant has indicated a third from 'the Straits of Magellan. Of the genus Liparis, of the Liparididce, three species occur in the North Atlantic, four in the North Pacific, and three in the Antarctic regions, near the southern end of South America. Carcprodus, of the deep-sea forms in the family, has the disk much reduced in size ; three of its species are from the North Atlantic, and one from the North Pacific. Another of the genera from great depths, Paraliparis, of the same family, has entirely lost the disk, and with it all traces of the ventral fins have disappeared ; three of its species, as at pre.sent recognized, are Northern Atlantic, and one has been reported from the North Pacific. From the foreo-oino- it might be shown that the Discoboles are more numerous toward the Arctic regions, but the fact that the small amount of investigating thus far done at the far South has discovered so large a proportion of the whole number of species noted makes it appear very hazardous to assume that they are any less abundant in that direction. If, in connection with this, we take into account recent discoveries in deep-sea work, we can, to say the least, hardly avoid admitting a possible distribution of the Discoboles from one of the icy zones to the other through the frigid waters of great depths. The conditions far below the ocean surfoce in the torrid zone, as to temperature, pressure, and existence of animal food, not being essen- tially different from those obtaining in localities in which some of the forms are known to occur, there seems to be no apparent reason why we should not expect to meet with representatives of one or another of the genera, or with allied forms hitherto unknown, in the results of future explorations of the sea bottom in the equatorial regions. With an ascertained existence in the North and in the Soutli, and again in portions of the deep sea, antici- pation of a distribution of these fishes that shall be more or less extensive throughout the greater depths of the ocean from the Arctic to the An- tarctic does not appear at all unreasonable.* In the surface waters of the * In collections placed in my hands by Professor Agassiz for investigation since this work has been in press, I find sufficient grounds for announcing the confirmation of these suggestions, and the fact of the existence of a snbequatoVial distribution of the Discoboli. The material is of that recently gathered in the Pacific by the Steamer "Albatross" of the United States Fishery Commission. Among the fishes collected by this vessel there are species of both Careproctus and Paraliparis that THE DISCOBOLI. 9 tropical and subtropical seas, the Gobioids and Gobiesocoids are abundant, but there are apparently no representatives of the Discoboli. HISTORICAL. Knowledge of the Discoboli previous to the time of Gesner is so indefi- nite and uncertain that it can hardly be said to have existed. In tracing the origin of one of the generic names now in use, it becomes necessary to make references to earlier writers ; but the term had no application among the fishes properly included in the group before the appearance of Wil- lughby's work. To eliminate confusion as far as may be from the gen- eral history, that of each of the different subdivisions has been arranged separately under a distinct heading. Liparis. — In book thirty-second, chapter eleventh, " Animalium om- nium in mari viventium, centum septuaginta sex genera esse," of Pliny's Natural History, this name is given as that of one of the fishes in the list of marine animals. In Holland's translation the spelling is changed to Lyparis. There is no sufficient clue to the identity of the species. Salviani, 1554, Gesner, Artedi, and others, give references to Pliny for the name, without attempting identification. Belon, 1553, describes and figures one of the herrings (Clupeidte) as Liparis. Eondelet, 1554, be- stows the title on a very different fish, the figure of which bears some resemblance to the more slender genera of the Sparidse, such as Boops, Smaris, and Mtena. The only agreement between the forms to which these authors have applied the designation appears in an elongate shape, and especially in being very fat, from which the word "liparos" (XiTrapd?) derives its applicability. Neither has any other claim to present con- sideration than lies in the possession of the name indicative of fiitness or oiliness. Gesner, 1558, copies from both Belon and Rondelet ; Aldro- were secured southwest of Panama, in depths of more than 1,700 fathoms, and in temperatures of about 36° Fahrenheit. The species are described and figured in the forthcoming ichthyological report on these researches of the Commission. One species is named Careproclus longifilis, because of the slender prolongations of the upper rays of its pectoral fins. The type is black, much compressed, very slender, and in a total length of more than four inches and a quarter the head and body together measure only an inch, of which the head alone is more than half. The disk is very small, one sixth of an inch in length, little larger than the eye; the nostrils are tubular; and the caudal is acuminate. A second species is given the name Paraliparis Jimbriatus, on account of the prominent fringes formed by the lower parts of the pectoral fins, like those of P. membranaceus. Its type is black anteriorly, and be- hind the body it shades into light grayish. In total length it is more than four and three fourths inches, of which the head and body occupy an inch and five eighths. As in the first species, the nostrils are tubular, and the caudal is acuminate. The head is more depressed and broadened. 2 10 THE DISCOBOLI. vandi, 1613, also draws from them. Jonston, 1649, copies from Aldrovandi, makes a reference to Pliny, and, agi'eeing with Belon, fixes the name on a herring. Charleton, 1677, refers to Eondelet. Willughby, 1686, page 135, Plate H 6, Fig. 8, is indebted to the same author; he also, page 115, applies the name to a species of Gunnellus. In the Appendix is to be fomid the first use of the name in connection with a member of the group in which we are at present interested : " Liparis nostras, Dimehnensihis et Eboracen- sibiis, Tlie Sea Snail, i. e. Limax mariniis dictus, Piscis Rondeletiana Lipari forte idem," page 17, in all likelihood refers to one or both of the species now called Liparis Uparis and L. Montagid. The figure, H 6, Fig. 1, has no ventral disk, and otherwise is scarcely recognizable ; it has more like- ness to species of Lycodes than to any of the disk-bearers. The descrip- tion by Dr. Johnson does not satisfactorily fix the type ; most of what he gives relates to the common name, the disk, and the condition of the flesh. Ray, 1713, notes Liparis Rondeletii and L. nostras of Johnson. Artedi, 1738, was no doubt inspired by Willughby and Ray in establish- ing the genus Liparis with L. nostras as the type. He adds, with its synonymy, the Liparis of Rondelet as a second form, /3. That he used Liparis as a generic name is sufficiently evident from the introduction of the two forms, and the cai'e with which he separated his references, giving Johnson, Willugh. Append., p. 17, for the first, and Rondelet and Willugh., pp. 135, etc., for the second. Gronow, 1756, unable to reconcile Belon, Rondelet, and Artedi, renamed the genus Ci/clogaster, making L. nos- tras the type, and using the name given by Artedi as a synonym. Linne made no mention of any species of the Liparids until 1766, when he made Artedi's Liparis a species under Cyclopterus. Scopoli, 1777, adopted the genus from Artedi, and gave a pertinent diagnosis. Duhamel, 1777, Vol. II. p. 492, resuscitates the Liparis of Belon, correctly placing it with the sardines and herrings. La Cepede, 1800, formed two subgenera, Avithout Latin names, under Cyclopterus, the second of them to contain C. liparis and C. lineatus. Rafinesque, 18 15, provides the name Lipariits, without characterization, possibly intended for Liparis. Oken, 1816, placed Liparis as a subgenus under Cyclopterus. Cuvier, 1817, retained it in this rank. Fleming, 1822, makes it a genus, containing only species now included. Bonaparte, 1846, follows Fleming. Gunther, 1861, recognized the genus to be the type of a subfamily, Liparidina, under the family Discoboli. Gill, 1801, made a similar disposition, naming the subfamily Lipariuce, THE DISCOBOLI. 11 which he changed to Liparidmcc, 1864, and later to Liparididce, 1872, when, in his Gobiesocoidea, with the families Gobiesocidse and Cyclopte- rida3, he used it as that of a distinct family. In 1873, Gill rearranged the group, giving it the name Cyclopteroidea instead of Discoboli, and ranking it as a superfamily to contain the Cyclopteridoe, with one sub- family, the CyclopteriniB, and the LiparididfB with two, the Liparidinse and the Careproctina3 ; but in his latest publication on the subject, 1891, he suppresses his subfamily Careproctinse, and adds that of Jordan, 1882, the Amitrina3, but changes the name to Paraliparidina3. Giinther, 1887, places all the genera in a single family. Discoboli, without indicating other distinctions than those recognized as generic. Liparididoi. — The typical genus of this section is Liparis of Artedi, founded on the species afterward named Cyclopterus liparis by Linne. Canprodns was characterized by Krbyer, 1862, from a form named by him C. Reinhardi. It had previously been noted by Eeinhardt as Cyclopterus gelatinosus of Pallas, 1769, from which satisfactory distinctions have not yet been pointed out. Its prominent features are curved simple teeth and a long tenuous caudal, which lose some importance when studied in such types as L. major. AcUnochir of Gill, 1864, was a genus based on one of the forms described by Fabricius as Cyclopterus liparis ; it is not sep- arated from Careproctus. Neoliparis of Steindachner, 1875, was intended to include Liparis mucosus as a subgenus. The characters assigned were a low membranous connection of dorsal and anal with caudal, and a sup- posedly distinct anterior portion of the dorsal, neither of which seems to furnish sufficient reason for separation from L. Montagui or L. liparis, on which species the dorsal rays, continuous under the skin, acquire a peculiar development at particular times. Paraliparis of Collett, 1878, was distin- guished by the absence of the ventrals ; its species are deep-sea forms. Amitra of Goode, 1881, was discarded by its author for Monomitra, 1884 ; its type, M. liparina, was identified by Giinther, 1887, as a Paraliparis. Gymnohj codes of Vaillant, 1888, was disestablished by that author, with the remark that its type, G. Edwardsii, was closely allied to L. micropus Gthr., if not the same. Enantioliparis of Vaillant, 1888, included an An- tarctic form, E. pallidus, which we may not separate from the species of Liparis. The genera of this section of the Discoboli, those here adopted, include Liparis, Careproctus, and Paraliparis. Liparis, however, is made to contain as subgenera Liparis proper and Careliparis. 12 THE DISCOBOLI. Ci/cloptenis and Lnmptis. — " De Orbe scutato," the third chapter in book fifteen of Rondelet's work, De Piscibus, 1554, may possibly have been intended for the Lump-fish. The figure is a caricature. Dorsal and anal fins are absent, and the ventrals are doubtfully indicated by faint lines on the thorax that with the aid of some imagination may represent a disk. The fish was a stout-bodied one, with four rows of tubercles on each flank ; these, with the statement of its occurrence on the coast of Holland, and the remote resemblance of the given vulgar names with those there used, increase the probability of identity. Gesner, 1556, under Schnottholf refers to Eondelet. His reference to Albertus Magnus is of very questionable application. Though he was not dealing with the speci- mens, his text leaves no doubt that he included the Lump with the Dio- dons and Tetrodons. This is more evident in the Historia Animalium, Liber IV., 1558, page 747. On page 745 Gesner copies Orbis of Ron- delet, but is not misled by it. In the same volume, Paraliporaena N and Oj page 1284, he gives figures of Lumpus anglonmi and Orbis gibbosus, which, with the text, fix their identity with the Lump. Both are copied by Aldrovandi, 1613. Schonevelde, 1624, under the name Le2ms marinis, gives a better idea of the fish and the uses of the disk. Philemon Hol- land, 1635, in his translation of Pliny's Natural History, for Orbis and Orihragoriscus of the original, has " The Lompe, Paddle, or Sea Owle, and the grunting Molebout." Jonston, 1649, copies from Gesner. A third figure, Leo marinis, Plate XL. Fig. 8, is given, which also bears resem- blance to one of the Discoboles, Cyclopterichthys ventricosus, but the lack of text leaves the matter imdecided. Merrett, 1667, makes refer- ence to the Lump. Willughby, 1686, copies Rondelet's Orbis from Gesner. With Lumpus anglorum of the latter, he has a better description than any previous one ; the ventral fins are said to coalesce in the disk, which is likened to that of the Gobies. In the Appendix to this work Tyson makes an attempt at the anatomy. Ray, 1713, treats of Lumpus anglorum, from Gesner, as also Dale, 1732, and others. Duhamel, 1782, adds very good figux-es under the same name. Cycloplerus, as applied to this fish, appears first in the earliest edition, 1735, of Linne's Systema. It was placed in all the editions previous to the tenth as one of the Branchiostegi, with Lophius, Ostracion, and Ba- listes. In the tenth, 1758, another species was added, C. nudus (not now included), and the genus was made the first in the extensive Group VIL, THE DISCOBOLI. 13 the Thoracici. In the twelfth edition, 1766, he added a third species, C. liparis, the Liparis of Artedi, and transferred the genus to the Am- phibia nautes, where it was retained by Scopoli, 1777. By Gmelin, 1789, it was replaced in the Branchiostegi, as enlarged by the addition of Mor- myrus, Tetrodon, Diodon, Syngnathus, Pegasus, and Centriscus. This author brought together in the genus five of the species still recognized as Discoboles : lumpus, spiiiosus, vciitricosii-s, gclatimsus, and liparis. The genus Cjdo-pterus was adopted from the first by Artedi, Grouow, Scopoli, Forster, Gouan, Bloch, and others. Klein, 1744, gave it the name Oncotion. La Cepede, 1800, divided it into two subgenera, calling them first and second ; the former contained C. lumpus, C. spinosus, C. minutus, and C. gelatinosus, with Gobiesocidae ; the latter, C. liparis and C. lineatus. Rafinesque, 1815, gives three names to what was known of the Cj'clojiteri by previous authors, Lumpus, Cyclopterus, and Liparius ; but his sub- divisions amount to nothing, being without characterization. Oken, 1816, makes three subgenera of the genus, or four, counting Cyclopterus as one; to do this he uses Liparis Art., Lepadogasier Gouan, 1770, and Gohie- sox La Cepede, 1800. Cuvier, 1817, removing the Gobiesoces, retains in the genus the subgenera Cyclopterus and Liparis; Bonaparte, 1831, has the same arrangement. Nilsson, 1832, retains the subgenera, but changes one of their names to Limiptis. Storer, 1839, uses Lumpus for the name of the genus, and Cyclopterus for the subgenus. Fleming, 1822, restored both Liparis and Cyclopterus to full generic rank, and retained in each the typical species only. Cydopteridce. — This family, as established by Bonaparte, 1831, contained the Gobiesoces with Cyclopterus and Liparis. It was an improvement on Cuvier's family, the Discoboles, 1817, because it no longer was burdened with Echeneis and Ophicephalus. Rafinesque's family, Plecopodia, 1815, was still more comprehensive than that of Cuvier, since it included the GobiidiB, with others not retained by the latter ; it was formed by uniting the Plecopodes of Dumeril, 1806, with the Plecopteres of the same author. Dumeril's Teleohranches Plecopteres contained Cyclopterus and Lepadogaster. Rafinesque's Plecopodia comprised the genera Gobius, Plecopodus (Gobioides LaC), Lepadogaster, Piescephalus, Cyclopterus, Lumpus, and Liparius. Bonaparte, 1837, in his Systema Vertebratorura, made the Cycloptetidw the first family of his order Cycloidei ; he also 14 THE DISCOBOLI. established the subfamily Ci/dojiterhn, an apparent equivalent of the family itself. Lcs Chondrosies Ptero-podes of Dumeril, 1856, were made up of Lepadogaster, Cyclogaster, Gobiesox, Cyclopterus, Chironeotes, Lopliius, and Malthea. Bleeker, 1859, adopted the order Ctjdopteri, and placed within it the families Cy dopier oidd, with the single genus Cyclopterus, and the Gobicsodoidei, with Gobiesox and its properly related allies, and also including Liparis. Gunther, 1861, made Liparis the typical genus of his subfamily Liparidina, which with the Cyclopterina constituted the family Discoboli. Gill, 1861, did not differ in regard to the contents of the family, but named it (Jydopteroidi^. His* subfamily names were Cyclop- terinfB and Liparinse ; of the latter he made Liparidinae in 1864, and LiparididiB in 1872, on raising it to family rank. In Gill's Gohiesocoidca, 1872, he included the Gobiesocidte, the Liparididae, and the Cyclopteridae. The Ci/dopteroidea of Gill, 1873, contained only the C^^clopteridce and the Liparididce, the Discoboli of Gunther. In the present revision the family Cyclopteridae is restricted to the genera Cyclopterus, Eumicrotremus, and Cyclopteroides, the rank of the second being somewhat doubtful. Cydoptcrus of Linne, 1735, was estab- lished on the common Lump of the North Atlantic. Eumicrotremus of Gill, 1864, was based on C. spinosus Fabr,, an Arctic and deep-sea form. C. orbis Gthr. belongs with this species; it is obtained in the North Pacific. Cijdopteroides is a new genus, characterized below, from the North Pacific. Liparopsidce. — This family is here arranged for a couple of genera, at present known from the North Pacific, Cyclopterichthys and Liparops. Cydopten'chthys was characterized by Steindachner, 1881, from what he took to be a new species, C. glahcr (not C. glaber from Steller, 1831), but which was identified by others with C ventricosiis of Pallas, 1769. Liparops is based on Cyclopterus Stelleri of Pallas, 1831, the C glaber of Steller's manuscript. Discoholt. — Cuvier, 1817, brought together Lepadogaster, Gobiesox, Cyclopterus, Liparis, Echeneis, and Ophicephalus in a group to which he gave the name Discoholes. At the hand of Latreille, 1825, the name was written Discobola, and at that of Griffith, 1834, it became Discoboli. The division Cyclopodi of J. MUller, 1843, contained the following : " die THE DISCOBOLI. 15 Gobioiden, die Discoboli (Cjclopterus, Liparis, Gobiesox, Sicyases, Coty- lis, Lepadogaster) und die Echeneiden." Muller's family GoUoidei, 1846, is the same as his Cyclopodi of 1843. Richardson's Discopodcs, 1860, contained the Discoboli and the Gobiesocidge. The Discoboli of Giinther, 1861, are Muller's Discoboli after the elimination of the Gobiesocida3. In 1880 an arrangement was published by Giinther, in which the Discoboli and Gobiidae were brought together to make up a division, to which was given the name Acanthopterygii Gobiiforraes. As specified above, the Gobiesocoidea of Gill, 1872, included all of GUnther's Discoboli, and also the Gobiesocidae, but his Cyclopieroidea, 1873, contained only the Discoboli as limited by Giinther, 1861. EELATIONS. The adoption of the Discoboli in the present essay, with the limita- tions assigned by Giinther, is not to be interpreted as an agreement with the expression of relationship he indicates by placing the group, a family in his system, together with the Gobiidce for a separate division. The affinities existing between these families do not seem to be of a character that warrants such an arrangement. More recent systems, in which the Gobies are placed farther away, and the Cottoids brought nearer, better express the natural order. It may be doubted whether Pallas had a clear idea of the relationships of the Cyclopteridse, but at any rate his comparisons were not far out of the way. He closes his description of the anatomy of Cyclopterus ventricosus, 1769, after the enumera- tion of three or four particulars, with the sentence, " ReliquJB fere ut in Cottis quibusdam." The hint contained in this led to nothing, for we find that Hermann, 1783, after quoting Pallas and discussing the matter in several places, retains the Discoboli in the Branchiostegi near Diodon and the Gobies, while he puts Cottus in the Thoracici with Echeneis. Eetz, 1800, in his edition of the Fauna Suecica, placed Cyclopterus between Gobius and Cottus, manifestly a better arrangement than that originally adopted by Linne. It was Pallas again, in 1831, in his descrip- tion of Cyclopterus callyodon, who first pointed out the clue by which the closer alliances have finally been traced. His statement, " A maxilla, superiore utrinque processus sub cute tenuis osseus, quasi zygoma, per discum operculorum," called attention to a feature previously unnoticed in these fishes, which, whether the disk is present or not, readily serves 16 THE DISCOBOLI. to connect them each with the other and with the members of the large group containing the Cottoids, Scorpaenoids, and allies, the cheek-boned fishes {Pareiostichtht/es), all of which possess a common character in the prolongation of one of the suborbitals backward toward the preoperculum. It was not, however, until Putnam, 1874, began the study of the group, that the full significance of this cheek bone was developed and the real affinities with the Cottoids determined. The publication of the outcome of his investigations, " Notes on Liparis and Cyclopterus," has brought about a more accurate understanding of the relations of the group as a whole, and of its position in the system. In the present study various comparisons have been made for the pur- pose of determining the degrees of relationship existing between the several families of the Discoboles, and also between them and allied groups. There is nothing in the results obtained that declares with any emphasis against the conclusion that the closest affinities are with the Cottoids. Within the group it is found that the nearest approach is made by the Cyclopteridae, and that the Liparididse have diverged more from the progenitors common to these disk-bearers and the Cottidoe. Excepting, in cases, the shape and the disk, the agreement in general features with the latter is rather close. This is shown by comparison in the fins and their attachments, particularly in the connection of carpals and pubes with the claviculars, in the elements of the skulls (Plates IX. and X.), especially in the suborbital stay and the opercles, in the branchihyals, in the teeth, in the gills, in the stomach, in the caeca, and in the urogenital apparatus. The general correspondence is very noticeable in regard to the glossohyal and basibranchial elements, which are either suppressed in part or greatly reduced in size. The apparent difference in shape and the softness of the bones notwithstanding, it is Cyclopterus that is to be placed nearest to Cottus. Without going far into details, reasons for this will be found by comparing the opercular bones and the expansion and connection of the suborbital process (Plate X. Figs. 28, 30, 32, 33, 19"). In the Liparididce the other opercular bones are more reduced, while the preoperculum is more expanded in the upper limb, and the suborbital process is more slender and crosses the preoperculum to its hinder edge. Better agree- ment also exists between the brain of Cyclopterus and that of Cottus; in both the brain is shortened, and cerebellum and hemispheres are small as compared with the optic lobes. Liparis has a more elongate brain, and THE DISCOBOLI. 17 the optic lobes are comparatively much smaller. A number of particulars may be mentioned in which Cyclopterus differs from Cottus, as shape, disk, obsolescent pleurapophyses, toothless vomer, cephalic spines, smooth gill rakers, multiplicity of cajca, elongation of intestine, simple bladder, etc., but their character is not such as weighs greatly against those tending toward closer relationship. The branching bladder and the unmodified ventrals of Cottus lose much of their apparent importance as distinguish- ing characters on closer investigation. General form, firmness of skeleton, toothed gill rakers, smaller number of caeca, shorter intestine, well devel- oped pleurapophyses, and other features in which Liparis and Cottus agree better, can hardly be considered of sufficient importance to counterbalance those tending to place Liparis farther away than Cyclopterus. The adhesive apparatus on the ventral surface of certain Gobiidte and on the Gobiesocidas formerly caused these forms to be located with the Lumps and their allies. The disk of the Gobies differs greatly from that of the Gobiesoces, and in either case the organ is structurally very different from that of the Discoboles, while there is little in the balance of the anatomy to suggest close affinities with the latter. In fact, modern ich- thyologists, recognizing the lack of anything in skeletal or other features on which to predicate an approach to near relationship, have generally agreed in placing all of the mentioned types at a considerable distance from one another. DESCRIPTIONS. In the majority of Discoboli the shape resembles that of the larvae of toads and frogs. That portion of the body including the visceral cavity is broad and massive, while that behind the chamber is narrowed and slender, as in tadpoles. In one section of the group, the anterior portion is the longer ; in another, it is the shorter. Though the application of the name was caused by the disk, the latter is not always present ; some species have lost it so completely as to be without evidence of former possession of either disk or ventrals. When present, the adhesive apparatus is formed of the ventral fins, the six rays in each being transformed into ribs for the central portion, which is surrounded by a dermal fold as a margin. The skin varies in the different genera from tough and covered with tubercles to very tender and slim}'. Ordinarily the flesh is described as gelatinous, and the attachments of the skin are not of the most firm, and in some cases are of 3 18 THE DISCOBOLI. the slightest. Certain species possess two dorsal fins ; on others the two are more or less perfectly confluent, and on some the spinous dorsal would seem to have been lost. The head is short and broad, the snout blunt and rounded, and the mouth, small to moderate, is anterior. The teeth are small, either subconical or tricuspid, and the number of series in function varies from a couple to half a dozen or more forming a pavement. In some cases the eyes are very small, in others they are of moderate size ; their direction is lateral or obliquely upward. Quite generally there are six branchiostegal rays ; so far as we know there are no exceptions to this number. The gill openings are rather narrow to very narrow ; sometimes they extend down- ward upon the bases of the pectorals, but often they are situated entirely above the fins. The gill membranes are united with the isthmus and the shoulders. Usually there are pseudobranchisB in addition to the three gills and a half; they are said to be absent in a particular species. None of the Discoboles possess air bladders. The stomach is siphonal, and bears a cluster of caeca around the side of the pylorus. The skeletal structure is almost cartilaginous in Cyclopterus, the osseous matter being small in quantity and arranged in thin plates, forming chambers as in Lophius ; from this it varies to more solid and firm in species of Liparis. Throughout the genera the operculum, the suboperculuin, and the interoperculum are reduced in size ; most so perhaps in some Liparididae, where these bones are slender and spine-like. On Cyclopteridte they are thin, but broader and more like blades. A common feature of the Discoboli is the prolongation of the third suborbital bone to the preoperculum across the cheek. The Lumps have this suborbital rather broad, thin, flexible, and slightly drooping backward ; on certain Liparids it is slender, styliform, nearly straight, and extends to the hinder edge of the preopercle. The following synopsis includes the known members of the group. THE DISCOBOLI. 19 FAMILIES AND GENERA OF DISCOBOLI. Body cavitj' elongate ; caudal region short ; disk present ; teeth simple ; dorsals two ; skin tubercular, barbels none ; tubercles in rows, tubercles not in rows, barbels present, dorsal one, skin smooth ; dorsal short, skin tubercular ; dorsal elongate, Bod.v cavitj' short ; caudal region elongate, disk present, teeth tricuspid, simple in older Careprocti / caudal more or less distinct, vertebriE less than forty, vertebriB more than forty-five, caudal indistinct, tenuous ; teeth simple, at least in older stages, disk absent, no ventrals, caudal indistinct, tenuous, CYCLOPTERID^, Cyclopterds, eomicrotrem us , Ctclopteroides, LIPAROPSID^, CrCLOPTERICUTHYS, 40 LiPAROPs, 42 LIPARIDID^, 43 ZiIPASIN^, 45 PAGE 19 20 34 37 40 LiPARIS, 45 Jjiparis, 47 Uareliparis, 62 Careproctus, 71 {mitring, 80 Paraliparis, 80 CYCLOPTERIDffi. Anteriorly the form of the Lump-fishes is stout, thick, and deep ; behind the body cavity, which occupies the greater portion of the length, it rather abruptly becomes weak and slender. The head is short and broad, sub- quadrangular in transsection ; the snout is short and blunt ; the mouth is of moderate width, anterior, and opens slightly upward ; the teeth are small, subconical, and arranged in a band or card ; the eyes are of medium size and have a lateral outlook. All of the members of the family have pseudo- branchiae, three and a half gills, six branchiostegal rays, small gill openings, numerous pyloric casca, and an elongate intestine. In their lower portions the broad rounded pectorals extend forward under the throat, along the sides of the disk. The vertical fins are not of large extent ; the caudal and the two dorsals are quite separate. The disk is comparatively large. Early in life the skin is tender and naked ; later it grows tough and is covered with roughened or spine-bearing osseous tubercles. Semi-cartilaginous de- scribes the skeleton with tolerable accuracy ; the small amount of bony 20 THE DISCOBOLI. matter lies in thin plates, often forming cells and chambers similar to those to be noticed in the bones of Lophius. The thu-d suborbital is thin and broadens as it extends back to the preopercle. The interopercle is more blade-like than in the Liparididte. On one genus the pores around the mouth are tubular, and form barbels. When Bonaparte, 1831, first used the family name Cijchptendce, he in- cluded Liparis with the Lumps. In the present writing the family is made up of Giinther's subfamily of 1861, Gijchptenna, with the addition of a genus of recent discovery, first characterized in these pages. GENERA AND SPECIES OE CYCLOPTERID^. PAGE Barbels none ; disk anterior, sub-cephalic, Cyclopterds, 20 larger tubercles in rows, separated, C. lumpus, 21 larger tubercles not in rows, crowded, Eumicrotremus, 34 staped more lilie the Lump, E. orbis, 36 shaped more like a Diodon, E. spinosus, 34 Barbels present ; disk central, subabdominal, Ctclopteroides, 37 lateral tubercles minute, C. gyrinops, 37 CYCLOPTERUS. Body more or less compressed toward the back, somewhat triangular in a transverse section at the first dorsal, covered with conical, rough, bony tubercles ; head short, thick, subquadrangular in a cross section ; snout blunt, rounded ; moitth anterior, opening slightly upward ; teeth simple, small, arranged in a band ; eye moderate, lateral ; dorsals two ; caudal dis- tinct; disk moderately large, anterior. On the species properly belonging to Ci/clopteriis, the first dorsal fin with age becomes completely hidden by the skin, and the larger tubercles of the flanks, though in regular series, have a scattered appearance. On the larger specimens of the species belonging to Eumicrotremus the large tubercles are very irregularly disposed and close together. The history of Cyclopterus as a genus begins with Linne. Flem- ing, 1822, says of it: '-'This genus has been subdivided into (1) Cyclopterus, having the body furnished with ridges of tubercles, as C. lumpus, and (2) Leparis, having the body smooth, as L. vulgaris and Moiitagui." This subdivision left the genus as recognized at present. THE DISCOBOLI. 21 Cyclopterus lumpus, Plate VIII. Figs. 1-3, 15-17 ; Plate IX. Fig. 3 ; Plate X. Pig. B; Plate XI. Fig. 10 ; Plates XII., XIII. Cyclopterus Iwnpus Linne, 1735, Systema Natuvse, ed. 1, — 1738, Art. Syn., p. 87, Gen., p. 62, — 1740, Systema, ed. 2, .52, — 1740, ed. 3, 55,-1744, ed. 4, 81,-1747, ed. 5, 60, — 1748, ed. 6, 41, — 1748, ed. 7, 41,-1756, ed. 9, 43,-1758, ed. 10, I. 260,-1760, ed. 11, I. 260, — 1766, ed. 12, I. 414,-1767, ed. 13, I. 414 ; Poiitop., 1763, Atlas, I. 648; Beckmaim, 1772, Linn. Syst. Nat., I. 102 ; MUll., 1774, Linn. Natursyst., III. 330; Mull., 1776, Zool. Dan. Prodr., 39; Asoan., 1777, 1806, Icon., IV. 4, pi. xxxiv.; Fabricius, 1780, Fauna Groenl., 131 ; Bloch, 1780, Skr. Ges. Nat. Fr. Berl., I. 247,— 1784, Oecon. Natuvg. Fische Deutschl., III. 103, pi. 90; Hoslin, 1781, Leiii-Buch ub. Nat. Syst., I. 577; Mohr, 1786, Forsog, 61; Bonnat, 1788, Ichth., 26, pi. 20, fig. 63; Ginel., 1789, Linn. Syst., I. 1473; Walb., 1792, Art. Gen. Pise, 484,-1793, Klein. Ichtli. Enod., Index, 32; Donnd., 1798, Zool. Beitr., in. 812; Cuv., 1798, Tabl. Element, 326,-1817, R. An., II. 226,-1829, ed. 2, II. 346,-1836, ed. 3, I. 573; AVUh., 1799, Unterh. Naturg., IX. 125, pi. xvi. fig. 48; Retz, 1800, Fauna Suec. Linn., 327; Heppe, 1800, Abb. u. Beschr. Fische, 256, pi. 60, fig. 119; Georgi, 1801, Geog.-Phys. Nat. Besclir., VII. 1967; Castel, 1801, Hist. Poiss. Bloch, VIII. 121 ; Schn., 1801, Bloch. Syst.Ichth., 197; Donov.,1802, Brit. Fish., pi. 10; Sonn., 1803, Hist. Poiss., V. 247; Shaw, 1804, Gen. Zool., V. 388, pis. 166, 167 ; Turt., 1806, Linn. Syst. Nat, I. 904 ; Quensel, 1806-24, Svensk Zool., 49, pi. 52 ; Oken, 1816, Lehrb., III. 135; Rathke, 1822, Meek. Arch. Physiol., VII. 493; Rosenth., 1822, Ichthyot, 16, pi. 19; Hofm., 1823, Tidsskr. Naturw., II. 373 ; Anslijn, 1828, Syst. Beschr., IV. 66, pi. 62 ; Faber, 1828, Tidsskr. Naturw., 251,-1829, Naturg. Fische Islands, 49 ; Flem., 1822, Phil. Zool., 387,— 1828, Brit An., 190; Fee, 1830, Linn. Syst. Nat., ed. prim. reed, curante Fee, 70; Pall., 1831, Zoogr., III. 72; Nilss., 1832, Prodr. Ichth. Skand., 61, — 18.5.3, Skand. Faun. Fisk., 232; Griff., 1834, An. King. Cuv., X. 501; Ekstr., 1834, Fisk. i Morko Skargard, — 183.5, Die Fische in den Scheeren von Jlorkd, 108; Kielsen, 1835, Icon. Pise., pi. xviii. ; Ross, 1835, Nat. Hist. Second Voy., p. xlvi ; Jen., 1835, Brit. Vert., 471 ; Schinz, 1836, Naturg. u. Abb. Fische, 258, pi. 86, fig. 4; Kaup, 1836, Thierreich, III. 67; Rich., 1836, Fauna Bor. Amer., III. 260; Yarr.. 1836, Hist Brit Fishes, H. 270, —ed. 2, II. 36.5, — ed. 3, II. 343; Templeton, 1837, Charlesw. Mag. Nat. Hist. (2), I. 412; Wilson, 1838, Ichth. in Encycl. Brit. ; Reinh., 1838, Ichth. Bidr. Groenl. Fauna, 108, 116 ; Swains., 1838, Nat. Hist. Fish, II. 339 ; Johnst., 1838, Berw Nat. Club, 1. 174; Fries, 1838, Vet Ak. Handl., pi. iv. fig. 1 ; Parn., 1838, Fish Firth of Forth, 380, Mem. AVern. Soc, VII.; Gmel., 1839, Gemein. Syst. Nat., Fische, 72, pi. 12, fig.21; Schill., 1839, Naturg. Fische, 90, pi. 14, fig. 3; Val., 1840, R. An., ed. ill., 310; Selys-Longch., 1842, Faune Beige, 240 ; Dub. & Kor., 1814, Vet. Ak. Handl., 116, pi. 3, fig. 8 ; Guerin-Menev., 1844, Icon., 35, pi. 62, fig. 2 ; Krby., 1845, Danm. Fiske, II. 489 ; Gaim., 1851, Voy. Isl. & Groenl., Poiss., pi. 8 ; Gron., 1854, Cat., ed. Gray, 38 ; Ham., 1854, Brit. Fishes, IL 250, 404, pi. 13, fig. 2 ; Thomp., 1856, Nat. Hist. Irel., IV. 215; Gill, 1861, Cat. Fish E. C. N. Amer., 47, — 1864, Pr. Phil. Ac, 190, — 1865, Can. Nat., Aug., —1873, Cat. Fish E. C. N. Amer., 21, — 1891, Pr. U. S. Mus., XIII. 368, pi. xxviii. fig 1 (fr. Goode), fig. 3 (fr. Borckert), pi. xxx. fig. 1 ; Gthr., 1861. Cat, HI. 155, — 1880, Introd. Study of F'ishes, 484; Schleg., 1862, Dier. van Nederl., Vi.s.sch., .58, pi. 6, fig. 1; Couch, 1863, Hist. Brit. Fishes, 183, pi. 105 ; Mcintosh, 1862-66, Pr. R Soc. Edinb., V. 614, — 1875, Fish St. Andrews, 174, — 1886, Ann. Nat Hist (5), XVIII. 81; Smith, 1868, Pr. R. Soc. Edinb., 1865-66, 355,-1888, P. R. Phys. Soc. Edinb., 143, pi. 7, fig. 8; Coll., 1875, Norges Fiske, 63 ; Hilgend., 1878, S. B. Nat. Fr., 156; Goode, 1879, Bull. Essex Inst , XI. extr. p. 11,-1884, Fishery Industr. U. S., 253; Bean, 1879, Bull. 15 U. S. Mus., 11.5,— 1880, Pr. U. S. Mus., III. 83; Winth., 1878, Nat. Tidsskr., 3 R, XI. 290, pi. 1, figs. 1-13,-1879, Prodr. Ichth. Dan., 19; Buckl., 1881, Brit. Fish., 125; Moreau, 1881, Poiss. Fr., III. 349; A. Agassiz, 1882, Pr. Am. Ac. Arts & Sci., XVII. 28G, with 2 pi. of young stages; Jord. & Gil., 1882, Bull. 16 U. S. Mus., 747 ; Dresel, 1884, Pr. U. S. Mus., VII. 2.50 ; Heusen, 1884, Viert. Ber. Comm. AViss. Unters. deutschen Meere, 301; Mob. & Hein., 1884, Viert. Ber. Deutsch. Meere, 204, 226, fig. ; Day, 1884, Fish. Grt. Br. & Ire., I. 179, pi. 55; Hans., 1885, Zool. Dan. Fiske, 39, pi. 6, figs. 11, 12 ; Jor., 1887, Rep. U. S. F. Com., 1885, 904; Borck., 1889, Anat.-phys. Unters. Haftscheibe, p. 7, pi. 1. Lumpus anglorum Gesner., 1.558, Hist. Anim., lA''. 1281; Aldrov., 1613, De Pise, et Cetis, 479, — 1623, De Pise, 180; Jonst., 1649, De Pise, et Cot., 40, pi. xiii. figs. 1,2; Sibbald, 1684, Scotia illustr., 22 THE DISCOBOLI. Part 2, II. 24; Willugh., 1686, Hist. Pise, 208, pi. N. 11, — i. gibbosus, N. 10, fig. 2 ; Ray, 1713, Syn. Pise, 77; Duham., 1782, Traite Gen., IV. 308, pi. 24; De Kay, 1842, Zool. N. Y., IV. 305; Stor., 1846, Synops., 481, — 1867, Hist. Fish. Mass., 208, pi. 32, fig. 2 (both pub. in Mem. Am. Acad.); Lumpus vulgaris McMurt., 1831, Cuv. An. King. ; Stor., 1839, Rep. Fishes Mass., 151; Knight, 1866, Fishes Nova Scotia, 16. Cydopterus minulus Pall., 1769, Spic. Zool., VII. 12, pi. 3, figs. 7-9 ; Fabr., 1780, Fauna Groenl., 135; Bonn., 1788, Ichth.,27; Gmel., 1789, Systema,I. 1475; Walb., 1792, Art. Gen. Pise, 485; Bound., 1798, Beitr., III. 815; Schn., 1801, Bl. Syst. Ichth , pp. xxiii, 3; Turt., 1806, Syst. Nat., I. 905; Cuv., 1817, R. An., II. 227, — 1829, R. An., II. 346, — 1836, R. An., I. 573; Griff., 1831, An. King , X. 502; Ross, 1835, App. to Sec. Voy., p. xlvi ; Rich., 1836, F. B. Amer., III. 262 ; Fries, 1838, K. Vet. Ak. Handl., 226; Val., 1840, R. An., ed. ill , Poiss., 311. Orhe gihbosus Gesn., 1558, Hist. An., IV. 747, 1285. Piscis gibbosus Aldr., 1613, De Pise, 480, — 1623, De Pise, 180. Lepus marinis Schonev., 1624, Ichth., 41. Orbis oceani Worm, 1655, Mus., 269 ; Cijchpterus sp. Art., 1738, Gen. Pise, 62, — Linn., 1746, Fauna Suec, 103, — Gron., 1754, Mus., I. 56, — Liun., 1759, Anim. Specierum, 91, — Gouan, 1770, Hist. Poiss., 223. Oncoliun Klein, 1744, Miss., IV. 49, — Ascan., 1777, Icon., pi. 34. Oncolion gibbosus Klein, 1744, Miss., IV. 50, and Oncotion (.3), pi. xi v. fig. 5, — Walb., 1792, Art. Gen. Pise, 583,-1793, Kleinii Ichth. Enod., 67. Cydopterus heptagunus Linn., 1754, Mus. Ad. Frid., L 57. Cydopterus Houtt, 1764, Nat. Hist, VII. 248, pi. 60, fig. 5. Cydopterus paooninus Shaw, 1797, Nat. Misc., IX. pi. 310,-1804, Gen. Zool., V. pi. 167, fig. 1. Cydopterus pyramidatus Shaw, 1804, Gen. Zool., V. 390, PI. 167, fig. 2. Cydopterus co!ruteus Mitch., 1815, Trans. Lit. & Phil. Soc. N. Y., I. 480. Le Cydoptere lompe LaC, 1800, Hist. Poiss., H. 52, pi. 3, fig. 1. Gobius minutus Mull., 1808, Zool. Dan , IV. 38, pi. 154 B, figs. 1-3, — Risso, 1810, Ichth. Nice, 159. Cydopterus coronatus Couch, 1823, Nat. Hist. Cornw., 47, — Ann'. Nat. Hist., II. 382, — Thomps., 1840, Ann. Mag., V. 216. Lump-sucker Pennant, 1776, Brit. Zool., III. 133, pi. 21, figs. 1, 2,— 1812, Brit. Zool., III. 176, pi. 24. Lump-Jlsh Low, 1813, Nat. Hist. Orkn., 177. B. 6 ; D. 6-8 / 11 ; A. 9-10 ; V. 6 ; P. 20 ; C. 12-14 ; Vert. 29 (11 + 18). Body m