Q11U563CRLSSI C\^u>\tvcci^i^' SMITHSONIAN INSTITUTIONUNITED STATES NATIONAL MUSEUBulletin 150 REVISION OF THE FISHESOF THE FAMILY LIPARIDAE BYVICTOR BURKEOj the State College of WashingtonPullman UNITED STATESGOVERNMENT PRINTING OFFICEWASHINGTON : 1930 For sale by the Superintendent of Documents, Washington, D. Ci Price 45 cents Cop^ ADVERTISEMENTThe scientific publications of the National Museum include twoseries, known, respectively, as Proceedings and Bulletin.The Proceedings, begun in 1878, is intended primarily as a mediumfor the publication of original papers, based on the collections ofthe National Museum, that set forth newly acquired facts in biology,anthropology, and geology, with descriptions of new forms andrevisions of limited groups. Copies of each paper, in pamphletform, are distributed as published to libraries and scientific organi-zations and to specialists and others interested in the different sub-jects. The dates at w^hich these separate papers are published arerecorded in the table of contents of each of the volumes.The Bulletin, the first of which was issued in 1875, consists ofa series of separate publications comprising monographs of largezoological groups and other general systematic treatises, (occasion-ally in several volumes), faunal works, reports of expeditions, cata-logues of type specimens, special collections, and other material ofsimilar nature. The majority of the volumes are octavo in size,but a quarto size has been adopted in a few instances in which largeplates were regarded as indispensable. In the Bulletin series appearvolumes under the heading Contrihutions from the United StatesNational Herbarium, in octavo form, published by the NationalMuseum since 1902, which contain papers relating to the botanicalcollections of the Museum.The present work forms No. 150 of the Bulletin series.Alexander Wetmore,Assistant Secretary, Smithsonian Institution.Washington, D. C, April 19, 1930. Ill CONTENTS PageIntroduction 1Genera 1Species 2Descriptions 3Part 1. General 4Collecting 4Preservation 5Historical 5Definitions and modifications of structure 7Sexual dimorphism 7Body 7Head 8Eye 9Nostrils 9Pores 9Gill slit 10Teeth 10Pseudobranchiae 11Pyloric coeca 12Branchiostegals 12Prickles 12Dorsal fin 12Anal fin 18Caudal fin 18Pectoral fin 19Disk 20Vent 20Coloration 20Habits 21Distribution 22Genera 22Species 23Regions 24Center of dispersal 26Bathymetrical distribution 28Temperature 34Summary 34Explanation of table 38Part 2: Systematic 38Family Liparidae 38Family description 38Relationships of the genera 39Key to genera 42Genus Liparis 43V VI CONTENTSPart 2?Continued. PageModification of characters 43Body 43Nostrils 43Eye 44Pores 44Teeth 44Gill slit 45Dorsal fin 45Anal fin 47Caudal fin 47Pectoral fin 47Disk 48Vent 48Pseudobranchiae 48Pyloric coeca 49Prickles 49Coloration 50Habits 50Food 50Summary 50Distribution 51Limits of distribution 51Asiatic and American species 53Regions 54Distribution of closely related species 54Bathymetrical distribution 55Center of dispersal 55Key to species of Liparis 56atlanticus 59montagui 61rutteri 61curilensis 62caUyodon 62grebnitzkii 64mucosus 64micraspidophorus 67liparis 67cyclopus 69frenatus 70fucensis 71bristolense 72megacephalus 73tunicatus 73herschelinus 75agassizii 76dennyi 77gibbus 79cyclostigma 79ochotensis 80ingens 81rhodosoma 81steineni 84 CONTENTS VII Part 2?Continued.Genus Liparis ? ^^seantarctica ?^tanakae ?^owstoni ?'pulchellus ^^ontessellatus ?^major -- ^^Q1Genus Polypera ^'Key to species of Polypera 92greeni ^2beringianus ^^simushirae. 94Genus Careproctus 9?Modification of characters. - ^^Body -- 95Head -- 95 GiU slit.Teeth. 9797Dorsal fin 98Anal fin. 99Caudal fin 99Pectoral fin 99Disk.Vent. 100100Pseudobranchiae ^"^Pyloric coeca ^^^PricklesColoration.Habits. 1011011021 02Summary ^"^Distribution ^^^Asiatic and American species 1^^Regions ^"^Closely related species ^^'*Bathymetrical distribution 105Center of dispersal ^"^Key to species of Careproctus 105 ^^^^^^113 pycnosomacurilanusectenessimus 1mollis 114smensis ?segaliensis H^bowersianus ^^'^falklandica ?- H^pallidus 11^attenuatus H"homopterus H' VIII CONTENTSPart 2?Continued.Genus Careproctus?Continued. Pagerhodomelas 118ranula 118reinhardi 121micropus 121georgianus 122phasma 123spectrum 123melanurus 125furcellus 126cypselurus 127abbreviatus 128bathycoetus 129longifilis 130ovigerum 131opisthotremus 133longipinnis 134pellucidus 134acanthodes 135rastrinus 136trachysoma 137gilberti 138ostentum 140roseofuscus 141coUetti 142entomelas 145entargyreus 146Genus Temnocora 146Candida 146Genus Crystallichthys 147cyclospilus 147mirabilis 148Genus Crystallias 150matsushimae 150Genus Gyrinichthys 152minytremus 152Genus Elassodiscus 154tremebundus 154Genus Paraliparis 154Modification of characters 155Body.. 155Head 155Nostril 156Eye 156Pores 156Gill slit 156Teeth 156Dorsal fin 157 CONTENTS IXPart 2?Continued.Modification of characters?Continued. PageAnal fin. 157Caudal fin 157Pectoral fin... 157Vent 159Pseudobranchiae 159Pyloric coeca 159Prickles 159Coloration 159Habits 159Summary 160Distribution 160Regions 161Bathymetrical distribution 161Key to species of Paraliparis 162dactylosus 164angustifrons 166fimbriatus 167deani 168entochloris 169liparina 170ulochir 171membranaceous 172bathybius 172australis 172melanobranchus 173latifrons 174holomelas 175garmani 176cephalus 177mento 179atramentatus 180copei 181rosaceus 182grandiceps 183attenuatus 185Genus Rhinoliparis 185barbulifer 185attenuatus 187Genus Acantholiparis 188opercularis 188Genus Nectoliparis 189pelagicus 189Appendix 191Liparis osborni. _ 191Gymnolycodes edwardsi . 191Careproctus dubius 191burkei 192jordani 192 X CONTENTSAppendix?Continued. PageCareproctus sarasa 192punctulatus 192okadae 192mederi 192Paraliparis albescens 193caudatus 193antarcticus 193terrae-novae 193wildi 193Lipariscus nanus 194Bibliography and Literature 194Index .. 201LIST OF ILLUSTRATIONSFigure 1. Diagrams showing parts of fish 82. Liparis mucosus. Prickles from specimen from Pacific Grove,Calif 653. Liparis micraspidophorus. A typical tide pool species, withnotched dorsal and free caudalfin 664. Liparis micraspidopliorus. Prickles from type 665. Liparis micraspidophorus. Teetli from type 676. Liparis cyclopus. Teeth from a specimen from Puget Sound.. 707. Liparis frenatus. Teeth from type 718. Liparis fucensis. Teetli from type 719. Liparis bristolense. Prickles and teeth from cotype 7310. Liparis tunicatus. Teeth from specimen No. 34168, U.S.N.M.from Labrador 7411. Liparis ocliotensis. Prickles from type 8012. Liparis ingens. Teeth from type 8113. Liparis ingens. Prickles from type 8214. Liparis antarctica. Tooth from type 8515. Liparis tanakae. Teeth and prickles from type 8616. Liparis owstoni, showing tlie modified head, differing in thisrespect from the tide pool species 8817. Liparis owstoni. Prickle from type 8818. Liparis tessellatus. A species that in respect to the unnotclieddorsal, union of dorsal, anal, and caudal and in appearanceapproaclies the Careproctus type 9019. Liparis major. Teeth from several specimens 9020. Polypera greeni. Teeth from tj'pe 9321. Polypera beringianus. Type 9422. Polypera beringianus. Teeth from type 9423. Polypera simushirae. Teeth from type 9524. Careproctus pycnosoma. Teeth from type 10925. Careproctus curilanus. Teeth from type 10926. Careproctus ectenes. Teeth from type 11127. Careproctus ectenes. Side view of teeth from specinic n No.64043, U.S.N.M 11128. Careproctus simus. Teeth from type 11229. Careproctus mollis. A typical species of the genus in shapeand general appearance 11330. Careproctus mollis. Teeth from type 114 CONTENTS XIPageFigure 31. Careproctus sinensis. Teeth from specimen No. 53812,U.S.N.M 11432. Careproctus segaliensis. Teeth from type 11533. Careproctus bowersianus. Teeth from cotype 11634. Careproctus attenuatus. Type, showing greatly attenuatedbody, approaching the condition found in Rhinoliparis 11735. Careproctus attenuatus. Teeth from type 11736. Careproctus homopterus. Teeth from type 11837. Careproctus rhodomelas. Ventral view showing great reduc-tion of vetral disk 11938. Careproctus rhodomelas. Teeth from type 11939. Careproctus ranula. Teeth from specimen No. 9556, StanfordUniversity Zoological Museum, showing variation fromtype 1 2040. Careproctus ranula. Teeth from type 12041. Careproctus reinhardi. Teeth from specimen No. 28812U.S.N.M 12242. Careproctus phasma. Teeth from type 12343. Careproctus phatma. Teeth from specimen No. 53813,U.S.N.M 12344. Careproctus spectrum. Teeth from type 12445. Careproctus melanurus. Teeth from type 12546. Careproctus furcellus. Teeth from cotype 12647. Careproctus cypselurus, showing forked caudal 12748. Careproctus cypselurus. Teeth from specimen taken offKamchatka 12749. Careproctus abbreviatus. Teeth from type 12850. Careproctus bathycoetup. Teeth from type 12951. Careproctus longifilis. Teeth from type 13052. Careproctus ovigerum. Teeth from type 13253. Careproctus opisthotremus. Teeth from type 13354. Careproctus pellucidus. Teeth from type 13455. Careproctus acanthodes. Teeth from type 13556. Careproctus acanthodes. Prickles from type 13557. Careproctus rastrinus. Type, showing the great developmentof the head region characteristic of some species of thegenus 13658. Careproctus rastrinus. Teeth from type 13659. Careproctus rastrinus. Prickles from type 13760. Careproctus trachysoma. Teeth from cotype 13861. Careproctus gilberti. Teeth from type 13962. Careproctus gilberti. Teeth from cotype 13963. Careproctus ostentum. Teeth from type 14064. Careproctus ostentum. Prickles from specimen No. 3023,Stanford University Zoological Museum 14065. Careproctus roseofuscus. Teeth from type 14166. Careproctus coUetti. Teeth from type 14367. Careproctus coUetti. Teeth from a specimen from AlbatrossStation 5029 14468. Careproctus entomeles. Teeth and prickles from type 14569. Careproctus entargyreus. Prickle from type 14670. Temnocora Candida. Type, showing the notched dorsal andslitlike pupil 146 XII CONTENTS PageFigure 71. Temnocora Candida. Teeth from type 14772. Crystallichthys cyclospilus. Type, showing the pecuharpupil and blotched coloration 14873. Crystallichthys cyclospilus. Teeth from type 14974. Crystallichthys cyclospilus. Teeth from specimen 130 mm.in length, from Albatross Station 4779, Bering Strait 14975. Crystallichthys mirabilis. Teeth from type 15076. Crystallichthys mirabilis. Teeth from a young specimenfrom Albatross Station 4797, off Kamchatka 15077. Crystallias matsushimae. Introduced to show the barbelsdistinguishing the genus 15178. Crystallias matsushimae. Teeth from type 15279. Gyrinichthys minytremus. Teeth from type 15380. Elassodiscus tremebundus. Type 15481. Elassodiscus tremebundus. Teeth from cotype 15582. Paraliparis dactylosus. Teeth from type 16583. Paraliparis angustifrons. Teeth from type 16684. Paraliparis fimbriatus. Teeth from type 16785. Paraliparis deani. Teeth from type 16886. Paraliparis deani. Prickles from type 16987. Paraliparis entochloris. Teeth from type 16988. Paraliparis liparina. Teeth from specimen No. 46002,U.S.N.M 17089. Paraliparis ulochir. Teeth from specimen No. 48699,U.S.N.M 17190. Paraliparis melanobranchis. Teeth from type 17391. Paraliparis latifrons. Teeth from type 17492. Paraliparis holomelas. Teeth from type 17593. Paraliparis garmani. Teeth from type 17794. Paraliparis cephalus. Teeth from specimen No. 5785, StanfordUniversity Zoological Museum 17895. Paraliparis mento. Type, showing peculiar shape of head 17996. Paraliparis mento. Teeth from type 18097. Paraliparis atramentatus. A typical species of the genus.Type 18198. Paraliparis atramentatus. Teeth from type 18199. Paraliparis copei. Teethfrom specimen No. 46009, U.S.N.M. 182100. Paraliparis rosaceus. Teeth from type 183101. Paraliparis grandiceps. Teeth from type 184102. Rhinoliparis barbulifer. Teeth from specimen from Alba-tross Station 5043, off Hokkaido, Japan 186103. Rhinoliparis attenuatus. Teeth from type 187104. Acantholiparis opercularis. Type, showing opercular spinesfound in no other species of the family 188105. Acantholiparis opercularis. Teeth from cotype 189106. Nectoliparis pelagicus. Pectoral girdle with five brancheo-stegal rays 190107. Nectoliparis pelagicus. Type, showing location of vent andpeculiar pectoral _ 190108. Nectoliparis pelagicus. Teeth from type 191109. Paraliparis caudatus. Type. One of the shorter, deeperbodied species resembling species of Careproctus 193110. Lipariscus nanus. Type 194 REVISION OF THE FISHES OF THE FAMILY LIPARIDAE By Victor Burke0/ the State College of Washington, Pullman INTRODUCTIONThe following report consists of a biological and taxonomic study ofthe fishes belonging to the family Liparidae. The writer, as an assist-ant upon the United States Bureau of Fisheries steamer Albatrossduring the Japanese cruise of 1906, became interested in the manynew and peculiar species of Liparidae that were continually beingbrought up by the dredge. Later, when assisting in the preparationof reports on these fishes, it became evident that the published ac-counts of the Liparidae were very unsatisfactory and that with theacquisition of so many new species a thorough review of the familywould be most timely. It was seen also that the Liparidae, repre-sented as it is at all depths from the tide pools down to 2,000 fathoms,is a faverable group in which to study the modification of structure andcolor due to the enviornment of the deep sea. This led to the exten-sion of the subject matter of the report into two parts.Part 1 consists of (1) a general account of the modification ofstructure and color throughout the family; (2) the distribution, bothhorizontal and bathymetrical ; (3) the relation between the bathy-metrical distribution and the coloration; (4) summary.Part 2 consists mainly of a taxonomic review of the family. Thedescriptions of those species which have been adequately describedelsewhere have been reduced to a brief synopsis of important char-acters together with an account of the distribution and a discussionof relationships. Keys to the genera and species are included. Inaddition the taxonomic value and modification of the specific char-acters of the three genera, Liparis, Careproctus, and Paraliparis, aredescribed.Genera.?Thirteen genera are recognized in this report. A num-ber of the old genera have been reduced to synonomy. I have fol-lowed Garman and other writers in not recoginzing Neoliparis Stein-dachner. The notch in the dorsal fin, upon which this genus wasbased, is not of specific value in Liparis dennyi and other inter-1 2 BULLETIN 150, UNITED STATES NATIONAL MUSEUM mediate forms. Prognurus Jordan and Gilbert and BathyphasmaGilbert are reduced to the synonomy of Careprocius. Prognurus wasbased upon the forked caudal, but with the discovery of Careprociusfurcellus, a species with a slightly forked caudal, the gap betweenPrognurus and Careprocius is closed up. BatJiypTiasma was basedupon the character of the teeth, but these in no way differ from theteeth of several species of Careprocius. Amitra Goode and Hil-gendorjia Goode and Bean are referred to the synonomy of Para-liparis. Amitra is described as differing from Paraliparis solely in thepresence of pseudobranchiae. I have been unable to demonstratethe presence of pseudobranchiae in Amitra liparina and doubt theirpresence in any of the specialized species of the famUy. Hilgen-dorjia is supposed to differ from Paraliparis in the forward extensionof the dorsal ridge and possibly in the character of the pectorial fin.These characters do not appeal to the writer as being of genericvalue. Crystallichthys Jordan and Gilbert and Gyrinichihys Gilbertare provisionally retained as genera. CrysiallicMhys was describedas differing from Liparis chiefly in the single nostril. I have shownthat it agrees with Careprocius in the character of nostril and teeth.It is retained as a genus because of the peculiar slitlike pupil and thecoloration. Gyrinichthys differs from Careprocius solely in the re-duced gill slit and will have to be reduced to synonmy upon the dis-covery of a species of Careprocius with a gill slit but slightly smallerthan that of a number of known species.One new genus (Temnocora) is described in this report. Fivenew genera have recently been described and should be mentionedhere. Gilbert and Burke (1912) describe the three remarkablegenera, Elassodiscus, Acantholiparis, and Necioliparis. Burke (1911)describes the genus Polypera. Gilbert (1915) describes the genusLipariscus.Species.?In the present report 114 species (including the appendix)are recognized. Specimens of 90 of these species have been examinedby the writer. Many of the species are represented by but a single orfew individuals; 29 species of those examined are represented by asingle individual, 13 species by 2 specimens each, and 64 species by 5or less specimens each. This lack of material has caused to be leftundecided the fate of a number of doubtful species. Other writersmay disagree as to the importance of certain variations and of thevalidity of some of the species recognized here. Several doubtfulspecies have been recognized and so indicated. When a name is inprint it is best to recognize it until it can be reduced to synonomywith absolute certainty. With a series of types before one, thespecies which can not, or only doubtfully, be identified by the de-scriptions appear distinct. The more we know of the species and themore species we know the greater the difficulty of distinguishing REVISION OF THE FISH FAMILY LIPARIDAE 6them by keys and descriptions. No one familiar with these fishesshould expect to identify them easily without correctly labeled spec-imens at hand.The serial arrangement of the species differs from that presentedin any previous report. The large amount of material available hasmade possible a careful study of the taxonomic value of all specificcharacters. This has led to a clearer conception of the relationshipof the species. Many of the species, however, are based upon sucha combination of characters that it is difficult to trace out the differentlines of development. This accounts to a large extent for the differentviews presented by writers concerning the relationship of variousspecies.Descriptions.?The specific descriptions are written in a uniformmanner in order to facihtate comparisons.In the measurements the length is given in millimeters and refersto the distance between the tip of the snout and the tip of the caudalfin. The other measurements are given in tenths of the length ofthe body minus the caudal fin and of the length of the head. Thecomparative measurements of certain parts proved to be of little or novalue. Of these some have been included and others omitted. Ofthose omitted we refer to the length of the caudal fin, snout to dorsal,snout to vent, pectoral to dorsal, interorbital width and disk to vent.We have omitted some measurements in the species of one genus andincluded them in the species of other genera because they proved tobe of greater value. See diagram illustrating the method of takingmeasurements. The number given after "Locality" refers to Alba-tross stations unless otherwise stated.These fishes are somewhat difficult to describe on account of thetexture of the body. Quite a few mistakes have been made in identifi-cation and description. Many of the United States NationalMuseum specimens were found to be incorrectly labeled, and it wasnot uncommon to find more than one species, and sometimes twogenera, in one bottle. This collection included eight undescribedspecies and the very aberrant genus Nedoliparis. In the deep-seaforms the dorsal and anal fins have to be dissected, sometimes underwater, in order to make accurate counts. The teeth should bemounted on a slide and examined with the microscope. The micro-scope and a powerful hand lens are necessary when working withthese fishes. Some of the important characters are lost with themaceration of the skin.The distribution of the species as given is based solely upon thematerial examined unless otherwise stated. As practically all thematerial in America has been examined by the writer, records otherthan those given here should be viewed with suspicion as possibly theresult of misidentification. 4 BULLETIN 150, UNITED STATES NATIONAL MUSEUMThe synonomy given is based mainly upon the material examined.This is the only satisfactory method of dealing with these fishes.It is extremely difficult or impossible to decide upon the specimenslisted by others. The synonomy of the American and Asiatic speciesis easily traced out, but that of the European species is impossiblewithout an examination of the material in the museums of Europe.The material upon which this report is based consists of the collec-tions in the American Museum, the Museum of Comparative Zoology,the United States National Museum, United States Bureau of Fish-eries, and the Stanford University Zoological Museum. Altogetherabout 830 specimens have been examined, and these represent 90species. The collections well represent the family and form a satis-factory basis from which to build a monographic report.The drawings of the teeth were made by W. L. Atkinson, under agrant from Stanford University, from camera lucida drawings bythe writer.A bibliography ?onsisting of the most important papers relatingto these fishes is appended.I wish to acknowledge my indebtedness to Dr. David Starr Jordan,Dr. C. H. Gilbert, Dr. Bashford Dean, and Dr. Samuel Garman, forassistance and advice received during the preparation of this report.Dr. Peter Schmidt of the Imperial Museum of St. Petersburg, andShigeho Tanaka, of the Imperial University of Tokyo, very kindlysent material for examination Mr. Barton A. Bean, of the UnitedStates National Museum, gave the writer every assistance possible inthe examination of the collection in his charge.PART 1. GENERALThe Liparidae or sea snails comprise a group of softbodied degen-erate fishes related to the Cyclopteridae and Cottidae. They agreewith the Cyclopteridae and differ from the Cottidae in having theventral fin modified into a sucking disk. In some of the deep-seagenera the disk has become lost. The Liparidae and Cyclopteridaehad a common origin and closely resemble each other. The formercan be distinguished from the latter by the more slender, shortertrunk and the more elongate caudal region.These fishes are specialized for a bottom existence both in the shallowwaters and the deep sea. Only one species is known to belong to theplankton. This species, Nectoliparis pelagicus, is found in mid-ocean between 300 and 600 fathoms. The most familiar species arethose found in the tide pools, where they rest under or cling to therocks.Collecting.?The tide-pool species are sluggish and easily captured.The turning over of rocks along the shore frequently exposes specieswhich can readily be seized. Baling out tide pools or sprinkling in REVISION OF THE FISH FAMILY LIPARIDAE O chloride of lime brings these and other fishes into view, when they canbe picked up with a dip net or a strong pair of forceps.Preservation.?Special care should be taken in hardening thesefishes. If not properly treated they soon macerate and identificationthen becomes difficult. Many of the museum specimens examinedwere in a deplorable condition and practically valueless. The bestresults are obtained by hardening overnight in a 3 to 5 per centsolution of formalin, then washing off in water and running throughalcohol up to 75 per cent. The formalin should not be strong enoughto swell the tissues. If placed in strong alcohol, without first harden-ing, the specimens quickly shrivel and lose their normal appearance.If permanently preserved in formalin a more lifelilie appearance isretained. This is especially true of the deep-sea species, which soonlose their gelatinous appearance if preserved in alcohol.Historical}?The most comprehensive account of these fishes inrecent years is the work of Garman on the Discoboli, 1892. Thishas been followed by a review of the deep-sea species by Goodeand Bean in Oceanic Ichthyology, 1895. Jordan and Evermanngive a review of the North American species in their Fishes of Middleand North America, 1898. The few known Asiatic species are de-scribed by Peter Schmidt in Pisces Marium Orientalium, 1904.A large number of smaller papers contain notes and descriptionsdealing with these fishes The Europeans, lacking the wealth ofmaterial found in America, have contributed but few descriptionsand notes.Jordan and Evermann, 1898, say that there are "Genera 9; speciesabout 40." Since that time the number of genera has been increasedand the number of species more than doubled. In the present workthe writer recognizes 13 genera and 1 14 species. In 1906 the Albatrosscollected 4 new genera and 32 new species. With the discovery ofAcantholi'paris and Nedoliparis our knowledge of the amount ofmodification within the family has been greatly extended. It seemssafe to predict that with further exploration of the Antarctic and deep-sea regions the number of genera and species wUl be considerablyincreased.In order to facilitate a comparison of the species described by Gar-man in the Discoboli, 1892, and those recognized by the writer thefollowing table of species is given. It indicates the writer's conceptionof the species listed by Garman. It refers solely to the specimens inthe museum and not to the synonomy. It was a surprise to find thatGarman was able to produce such splendid work with such a smallamount of material. It seems advisable also to give a table of thespecies listed by Evermann and Goldsborough ui their Fishes of Alaska, 1 For a more complete historical account, see Gill, 1891, and Garman, 1892.91668?30 2 6 BULLETIN 150, UNITED STATES NATIONAL MUSEUM1907, as their report is full of errors and misleading as to distributionof the species.Discoboli: 2 LiparidaeLiparis montagui Liparis monlagui (European).Liparis atlanticus (American, pi. 7).Liparis mucosus Liparis mucosus (California, pi. 5, figs. 1-5).Liparis callyodon (Alaska and Siberia).Liparis callyodon fucensis (pi. 6, figs. 1-5).Liparis liparis Liparis liparis (European).Liparis atlanticus (American, pi. 7).Liparis antarctica Liparis antarctica (pi. 6, figs. 6-10).Liparis agassizii Liparis agassizii (Japanese, pis. 1-2-3) . Liparis gibbus (some of Bean's specimens).Liparis tunicaius Liparis tunicatus.Liparis steineni Liparis steineni.Liparis pulchellus.. Liparis pvlchellus (pi. 4, figs. 6-8).Careproctus major Liparis major.Liparis pallidus Careproctus pallidus.Careproctus micropus Careproctus micropus.Careproctus gelatinosus Careproctus, not recognized.Careproctus reinhardi Careproctus reinhardi (Arctic Ocean).Careproctus ranula (Halifax Harbor).Paraliparis rosaceus Paraliparis rosaceus.Paraliparis bathybuis Paraliparis bathybuis.Paraliparis liparinus Paraliparis liparina.Paraliparis membranaceus Paraliparis memhranaceus.Fishes of Alaska: *Neoliparis rutteri Ltpam r??eri (fig. 99).Neoliparis callyodon Liparis callyodon.Liparis mucosus.Liparis rutteri.Liparis beringianus.Liparis fucensis Liparis fucensis (Station 4302) . Liparis denmji (4208, 4220, 4270, 4271, 4272,4289, 4302, 4256, Hoods Canal).Liparis agassizii Liparis bristolense (Station 3247).Liparis gibbus (St. Paul Island, fig. 111).Liparis herschelinus - ^ Liparis herschtiinus (pi. 18).Liparis cyclostigma Liparis cyclostigma.Crystallichthys cyclospilus (pi. 19).Liparis pulchellus Liparis pulchellus.Crystallichthys mirabilis Crystallichthys mirabilis (pi. 20).Careproctus simus Careproctus simus.Careproctus colletti Careproctus colletti (Station 3338).Careproctus spectrum (Station 4295).Careproctus phasma Careproctus phasma.Careproctus spectrum Careproctus spectrum.Careproctus ostentum Careproctus ostenfum.Careproctus ectenes Careproctus ectencs.Prognurus cypselurus Careproctus cypselurus (pi . 20) . ' Qarman, Discoboli, 1892. 3 Evermann and Qoldsborough, Fishes of Alaska, 1907. REVISION OF THE FISH FAMILY LIPARIDAE /Gyrinichlhys minytremus Gyrinichthys minytremus.Paraliparis holomelas Careproctus gilberii (Stations 4292, 4293).Paraliparis deani (Stations 4194, 4203, 42514253, 4255, 4292, 4293).Nedoliparis pelagicus (Stations 4252, 42574258).Paraliparis cephalus Paraliparis cephalus.Paraliparis ulochir Paraliparis ulochir.Rhinoliparis barbulifer Rhinoliparis barb ulifer.DEFINITIONS AND MODIFICATIONS OF STRUCTUREThe structure of the Liparidae has been greatly modified as thespecies became adapted to the deep sea. The extent of these changesin structure throughout the family are briefly outlined in the succeed-ing pages. The definitions of the terms used and method of takingmeasurements are best illustrated by the following diagram of theexternal anatomy of a Liparid. For a discussion of the taxonomicimportance of the specific character see under each genus.Sexual dimorphism.?Few of the species are known to show sexualdifterences. In Liparis atlanticus and Liparis rutteri the males, andpossibly the females to a less degree, at the breeding season have theanterior dorsal fin elevated and the male of the former species has thebody covered with "thumb-tack" prickles. Of the other specieswith prickles some have them in both sexes but more commonly theyappear to be confined to the male. The various color phases shouldbe investigated in relation to sex.Body.?^The shape of the body varies widely. The species whichlive in the tide pools or very shallow water usually have a short,depressed, and comparatively firm body. The species of the greaterdepths typically are compressed, more elongate, and less firm. Itmay be said in a very general way that in the species of the veryshallow waters the body is stout and depressed and the vertebraeare stout and few in number; that in species from somewhat greaterdepths the body is moderately stout, less firm, deepened and com-pressed and the vertebrae weakened and increased in number; andthat in the greater depths the skeleton becomes greatly weakened,the flesh very soft and reduced, the body elongate and the verte-brae slender and in stUl greater numbers. The number of vertebraevaries from about 35 to 70. The genera, which in a general wayillustrate these phases of modification, are Liparis, Careproctus,and Paraliparis with Rhinoliparis.Careproctus best illustrates the production of pseudotissue whichenvelops the body and fins just beneath the skin. In these forms theskin is loosely attached to the body. In preserved specimens itfrequently becomes detached and appears as a transparent sackabout the body and bases of the fins. 8 BtJLLETIN 150, UNITED STATES NATIONAL MUSEUMHead.?The head is typically blunt though sometimes distinctlypointed as in Liparis rhodosoma. The snout is usually truncate,rarely projecting as in Acantholiparis opercularis and Allinectesectenes. The lower jaw is typically included and the mouth horizon-tal. In Paraliparis cephalus and Nectoliparis pelagicus the lower jawprojects and the mouth is oblique. The projecting shovel-shaped e:}^tS^- ' supra branchial pores rudimentary pores snout pores :-.' snout to disk notchlower lobe of pectoral disk disk to vent dorsal to oaudal Tent to anal '?/ caudal rskln-oovered base of caudalcenter or base margin or flapFigure 1.?Diagrams showing parts of fishsnout of Acantholiparis opercularis suggests a mud-bottom environ-ment. The species of Liparis typically have the width of the headgreater than the depth, the reverse being true of the majority of theremaining species, in which the head is compressed and the cheeksvertical. One species, Acantholiparis opercularis, has the operculararms projecting as spines from the sides of the head. The lower REVISION OF THE FISH FAMILY LIPARIDAE 9 surface of the snout is split in Crystallias matsusTiimae. Barbels arepresent on the snout in this species and also in Rhinoliparis.Eye.?The eye is distinctly larger in the deep-sea species. Theaverage number of times the eye is contained in the head in Liparisis 6.5, in Careproctus 4, and in Paraliparis 3.8.The color of the eye varies from black to silver. The silverycolor apparently is confined mainly to the tide-pool and moderatelydeep-water species of Liparis and Careproctus.The pupil is typically round or slightly oval. An exception isfound in Temnocora Candida in which the pupil is reduced to a slit.In Orystallichthys mirahilis and related species the pupil is a slit withthe upper margin notched. These characters are constant, as thepupil of fishes is only slightly, if at all, subject to muscular control.With preservation or .maceration the shape of the pupil may bemodified.A^ostrils.? The nostrils are double in Liparis and Polypera andsingle in all the other genera, the posterior nostril being lost in thelatter groups. The anterior nostril tube leads directly from theolfactory cavity to the surface and usually projects above. Theposterior tube extends backward beneath the skin and opens abovethe anterior part of the eye. In some forms it projects above thesurrounding surface and may have a fingerlike projection in front.The two tubes are sometimes nearly equal in length though one isentirely below the surface of the skin. In certain specialized speciesof Liparis, as in L. major and L. tessellatus, the posterior nostrilopening is reduced and thus approaches the condition found inVareproctus. In Liparis and the more generalized species of Careproc-tus the anterior nostril tube typically projects high above the surfaceof the head; but in the more specialized forms of Careproctus and inParaliparis the tube opens at the surface or projects slightly above.Pores.?The lateral line in this family is reduced to 1 or 2 openpores above the gill slit, 2 on the snout, a series of 6 extending alongthe lower margin of the snout and eye to above the posterior marginof the eye, and a series of 7 extending from the tip of the lower jawto the opercle. The pore formula may be designated as 2-6-7-2.In many species rudimentary pores are present on the sides andocciput. The rudimentary pores are present in many species ofLiparis and Careproctus. Their presence in many of the morespecialized members of the family has not been demonstrated.They may occur in a single series extending back from the supra-branchial pores or be somewhat scattered on the sides anteriorly andon the nape. Each consists of a small papilla, the apex of which maybe indented or split, but there is never a distinct opening as in theother pores. 10 BULLETIN 150, UNITED STATES NATIONAL MUSEUMThe taxonomic importance of the open pores has not beenthoroughly investigated. They offer a promising field for furtherresearch and may be found to be of value in working out the geneticrelationships of the genera. All of the species of Liparis and mostof the species of Careproctus have two suprabranchial pores; a fewof the species of Careproctus and all the species of Paraliparis seemto have but a single suprabranchial pore. All the genera derivedfrom Paraliparis have but one suprabranchial pore. In certainspecies of Careproctus there seems to be a variation of from 2 to 1suprabranchial pore. The pore formula of some species of Paraliparisis 2-5-6-1, as though the upper pore of the last three series had beenlost. In certain species of Liparis, as in L. rliodosoma, the lowersuprabranchial pore is reduced, but in some species of Careproctusit is the upper pore that has been reduced.The position of the pores varies considerably. The suprabranchialpores are in some species separated by a distance greater than theeye, and in others they are very close together. The upper poreon the snout in Liparis is typically some distance in front of thenostril; in Careproctus it is closer to the nostril; in Paraliparis itmay be between and in Nectoliparis above the nostrils. This appearsto be due to a shifting of both the pores and the nostrils. Theanterior snout pore lies either in front of the upper pore or in frontand lateral to it. The snout pores and the nostril form the apicesof variously shaped triangles. The anterior mandibular pores incertain species of Careproctus and Paraliparis have a commonopening.Gill slit.?In each of the three large genera, Liparis, Careproctus,and Paraliparis, we find the gill slit undergoing the same amountof variation from a position above the pectoral to one extendingdown in front of it. There is a difference, however, in the frequencewith which the gill slit extends down in front of the pectoral. InLiparis two-thirds of the species have the gill slit enlarged andextending in front of the pectoral. The reduced gill slit is restrictedmainly to the tide-pool species, the enlarged slit to the deeper-waterand more specialized forms of the genus. About one-third of thespecies of Careproctus have the gill slit extending in front of thepectoral, and these are found in both the most generalized andspecialized members of the genus. The gill slit is widest in themost specialized forms of this genus. An extreme reduction of thegill slit is found in GyriniclitJiys in which it is high above the pectoraland no larger than the diameter of a dissecting needle. In onlyone species, ISlectoliparis pelagicus, is the gill slit confined to thefront of the pectoral.Teeth.?The teeth exhibit a wide range of variation in shape, sizeof pulp cavity, and arrangement. The teeth may be either trilobed REVISION OF THE FISH FAMILY LIPARIDAE 11 or simple, and some species have both types. Species in 10 of thegenera have trilobed teeth. In three of the genera the teeth rangefrom trilobed to simple. In another genus, Liparis, (major), there isan approach to the same condition. The trilobed condition of theteeth is most typical of the shallow-water species. All the species ofLiparis have teeth of this type. The simple teeth are more typicalof the deep-water species. More than half the species of Careproctusand nearly all the species of Paraliparis have simple teeth. Theteeth may be either thick and blunt or lanceolate. The thick, bluntteeth may be either trilobed or simple. The heavy crushing teethof Careproctus rTiodomelas are formed by the three lobes becomingthickened and united. In Paraliparis atramentatus the heavy, bluntteeth are the result of an increase in the diameter of the simple teeth.In nearly all the species with simple teeth the teeth are round incross section. In the Paraliparis rosaceus group of species, however,the teeth are compressed and wedge shaped at the tip.The teeth are typically arranged in pavementlike bands. Thewidth of the bands and the number of oblique rows of teeth consti-tuting the bands vary. In some of the species of Careproctus andParaliparis the bands of teeth are very narrow and the number ofoblique rows of teeth is reduced. Looking directly into the mouthfrom in front the rows of teeth are seen to diverge posteriorly.These are the oblique rows of teeth referred to in the descriptions.If, however, we view the teeth from the opposite direction we findthem arranged in oblique rows which diverge anteriorly.The bands of teeth usually present an oblique cutting surface.This is due to the anterior teeth being smaller and extending onthe anterior surfaces of the mandibular and maxillary bones. Astriking exception to this condition is found in the Careproctuscolleiti group of species. Here the teeth are set on a more horizontalsurface, the anterior teeth are larger, and the inner teeth recurved.When the jaws are separated the cutting surface appears horizontaland the anterior teeth appear as large as the inner teeth. Whendissected out the anterior teeth are found to be somewhat smaller.The teeth are arranged in a single series in the Paraliparis rosacevsgroup of species and in Nectoliparis pelagicus.?In the Paraliparisrosaceus group of species the teeth are set close together and the tipsare wedge shaped and form a sharp cutting edge. In Nectoliparispelagicus the teeth are minute and difficult to examine, but theyappear to be conical and more widely spaced than in Paraliparisrosaceus.Pseudohrancliiae.?The pseudobranchiae consist of about five fil-aments and are known with certainty to exist only in the genusLiparis. Goode, (1880, p. 478), records their presence in Paraliparisliparina, but this has been questioned by later writers. 12 BULLETIN 150, UNITED STATES NATIONAL MUSEUMPyloric coeca.?Pyloric coeca are present in all the species exceptAcantholiparis opercularis (cf. C. ectenes). The number varies inLiparis from about 110 in L. antardica and L. liparis to 70 or more inL. agassizii and other species. In the genus Polypera the coeca arefine and thicldy matted about the pyloris and number about 300.In the more specialized genera the average number of coeca is less.This is readily seen by an examination of table on page 35. Whenthe number of coeca is not more than 10 they are arranged in asingle series which may not completely encircle the duodenum. InLiparis the coeca are on the right side of the visceral cavity and inthe more specialized genera on the left side. This change of positionis due to a shortening and shifting of the U-shaped stomach.BrancJiiostegals.?The Liparidae have previously been described ashaving six branchiostegals. This is true for all but Nedoliparisand Lipariscus which have five, the interior one being absent in thesegenera.PricMes.-?Scales are represented by two types of prickles termed "thumb-tack" and "cactuslike" prickles. The former have beenfound in three genera ? Liparis, Careprodus, and Paraliparis. The "thumb-tack" prickles consist of short spines with round, flat headswhich are imbedded in the skin. The spines are lost with the epider-mis but leave shallow pits in the dermis which indicate their presencein the specimen. These pits are easily seen on the thick-skinnedspecimens of Liparis but are obscure on the thin-skinned species ofCareprodus. The "cactuslike" prickles are known to be present infour species of Careprodus and probably will be found in othergenera. They consist of groups of spines arising close together inthe dermis and bear a close resemblance to some found in the Cyclop-teridae. (See Jordan and Starks, 1895, pi. 94.) The number of spinesor prickles in each group may vary from 4 or 5 to 10 or more.Dorsal Jin.?In a number of the species the anterior dorsal raysare set off from the remainder of the fin by a notch. This notch ispresent in species of Liparis, Careprodus, Polypera, and Temnocora.It may be hardly evident or may extend to the base of the fin. Insome species, such as Liparis dennyi, the dorsal notch is present insome specimens and absent from others.The dorsal typically is connected to the caudal. In certain tide-pool species the connection is short, not extending beyond the skin-covered base of the caudal. The other extreme is found in Liparispulchellus, tessellatus, and owstoni, in which the connection is fullythree-fourths the length of the caudal. In Liparis the connectionwith the cadual is usually abrupt and frequently notched. In thetypical deep-sea genera the connection between the dorsal and thecaudal is fairly constant, varying from 0.3 to 0.6 of the length of thelatter. EEVISION OF THE FISH FAMILY LIPARIDAE 13The number of dorsal rays varies from about 28 or 30 in somespecies of Liparis to 65 or more in species of Paraliparis. An exam-ination of the table on page 35 indicates that the number of dorsalrays is typically increased in the species inhabiting the deep water.The number of rays varies in Liparis from 28 to 48, in Careproctusfrom 39 to 61, and in Paraliparis from 48 to 66.The anterior dorsal rays are primitively connected with the finmembrane, and the tips project above as in the pectoral fin. Inmany of the deep-water species, especially in Careproctus, the anteriorrays are buried in tissue beneath the skm and free from it. Such raysare weak, and their tips frequently extend, undulating, backward tothe succeeding ray.All the dorsal rays are simple; that is, unbranched. A varyingnumber of the anterior rays are also unsegmented. In the speciesexamined the number of unsegmented rays varies from 6 to 18, andappears to be fairly constant for each species. The species of themore highly modified genera do not appear to have more unsegmentedrays than the species of Liparis. (See following chart.) Some ofthe anterior rays in some of the species, especially those of Liparis^in addition to being unsegmented, are undivided, the two shafts orlateral halves of the ordinary ray being united to form a single rod.The relation between the dorsal notch and the unsegmented andundivided rays is of some interest. The number of unsegmented raysusually, if not always, exceeds the number of rays in front of thenotch. In some species only one or two of the rays behind the notchare unsegmented, while in other species there are as many or more un-segmented rays behind as in front of the notch. The number ofundivided rays is always less than the number of unsegmented rays.It appears that the notch, in some species at least, marks the divisionbetween the undivided and divided rays, the rays in front of thenotch consisting of a single rod and therefore spinelike.In all the species in which the dorsal notch is present the rays infront of it appear to be unsegmented and undivided. We will nowdescribe the condition of the anterior rays in those species in whichthe notch is absent. As previously mentioned, the anterior rays arealways unsegmented. The question of their being undivided inthose species in which the notch is absent has been investigated inbut a few species. In Liparis tunicatus and tanaTcae the anteriordorsal rays are undivided and spinelike. This may represent thecondition of these rays in all the species of this genus. In Care-proctus and Paraliparis, more degenerate genera, some of the specieshave all the dorsal rays divided.The significance of the dorsal notch and the unsegmented andundivided rays may now be considered. Does the dorsal notchmark the division between the ancestral spinous and soft dorsal?If so what is the significance of the unsegmented rays behind the 14 BULLETIN 150, UNITED STATES NATIONAL MUSEUM notch? In those species in which the notch is absent and all therays are divided have the primitive spines, such as are present infront of the notch in Liparis, been lost or have they become modifiedto resemble the remaining unsegmented rays?We can best approach the solution of the questions involved by anexamination of the ancestral or related families. In the Cyclop-teridae the dorsal notch appears to separate the spines from thesoft rays. In some of the species the spinous dorsal is absent. Thenotch is always present when the spines are present. When thespines are absent it is very evident from the position of the firstdorsal ray and the number of rays that the spines have been lostand not transformed into rays. In such forms the anterior dorsalray bears the same relative position in regard to the first anal rayas the first ray of the soft dorsal bears to the first anal ray in thespecies in which the spinous dorsal is present. Also the numberof dorsal rays in those species in which the spinous dorsal is absentis the same as the number of rays in the soft dorsal of those speciesin which the spinous dorsal is present. It also bears the same pro-portion to the number of anal rays. The number of anal rays re-mains fairly constant for both types of species. The dorsal notchin these fishes disappears, not as the result of the lengthening orshortening of certain rays, but as the result of their disappearance.The number of spines is small, usually eight or less. In the Agonidaealso we find some species with the spinous dorsal absent. In theCottidae the spinous dorsal is always present. In Psychrolutes,however, it is weak and buried in a ridge of skin. The number ofspines varies from 6 to 18. A notch marks the change from thespinous to the soft dorsal. In some of the genera, as in Ohitonotus,Hemilepidotus , Blepsias, and Hemitripterus , a second or anteriornotch divides the spinous dorsal. In the Scorpaenidae the numberof dorsal spines varies from 8 to 16. In the Hexagrammidae thereare as many as 26 dorsal spines. In this family the notch is some-times absent, as in Pleurogrammiis . From the above review of the condition of the dorsal fin in theprincipal families of the Loricati we are forced to consider the pos-sibility that in the Liparidae all the unsegmented rays are ho-mologous to spines such as are found in the Scorpaenidae. At leastthe large number of unsegmented rays, 15 or more, is not fatal toany such assumption. We have seen that species of both the Cot-tidae and Scorpaenidae may have as many as 16 or 18 and the Hexa-grammidae as many as 26 spines. These families are less modifiedthan the Liparidae and we may assume that a large number of spinesis a primitive condition for the Loricati. It therefore seems prob-ably that in the ancestry of the Liparidae there have been as many REVISION OF THE FISH FAMILY LIPARIDAE 15 spines as there are unsegmented rays in the Hving species. If so wemay be deahng with the retention of a primitive condition of thedorsal fin and the unsegmented rays may be homologous to or repre-sent the spinous dorsal of the hypothetical ancestor. There are,however, a number of serious objections to considering the unseg-mented rays of the dorsal as primitive spines. These objectionsare as follows:1. If the unsegmented rays represent spines we have the unusualcondition of the notch dividing the spinous dorsal. In no othergroup in the Loricati do we find that, when only one notch is present,it divides the spinous dorsal. We have seen, however, that in someof the Cottidae two notches may be present and that the anteriorone divides the spinous dorsal. The loss of the second notch wouldbring about the condition we are discussing the possibility of in theLiparidae. It is possible that this is what has occurred in the latterfamUy. The presence of the notch in the middle of the unsegmentedrays, while casting some doubt upon the theory that these raysrepresent spines, is not fatal to it.2. In the Cyclopteridae, a family which is generally considered tobe very closely related to and but slightly differing from the Liparidae,the dorsal notch separates the spinous and soft dorsal. The numberof spines is small, eight or less, and about equal to the number ofundivided rays in the Liparidae. Considering the close relationshipsbetween the Cylopteridae and Liparidae (some ichthyologists recog-nize but one family) and the apparent stability of the dorsal notchin the Loricati (if present in other families it divides the spinous andsoft dorsal), we should expect to find the dorsal notch bearing thesame relation to the spines and soft rays in these two families. Theagreement in the number of spines in front of the notch in bothfamilies favors this view. If the spines in front of the notch in theLiparidae are homologous to the spinous dorsal of the Cyclopteridae,we can reasonably assume that the rays behind the notch in theLiparidae represent the soft dorsal of the Cyclopteridae. Thereforewe can not consider the unsegmented rays behind the notch as repre-senting ancestral spines, but should look upon them as having beenmodified independently in regard to segmentation.3. The most important evidence contradicting the assumptionthat all the unsegmented rays represent a primitive spinous conditionis obtained from a study of the anal fin. We find that the numberof unsegmented rays in the anal fin varies from 1 to 12 or 16. Noneof the families of the Loricati have more than three anal spines.These spines are absent from the Cottidae and Cyclopteridae. Wecertainly can not homologize all the unsegmented anal rays withspines, but are forced to explain the loss of segmentation of these rays 16 BULLETIN 150, UNITED STATES NATIONAL MUSEUM as a newly acquired condition. If we are forced to explain the lackof segmentation of the anterior anal rays in this manner, it would beillogical to explain the condition of the unsegmented dorsal rays inany other manner without the support of very strong evidence.Such evidence is lacking except as regards the rays in front of thedorsal notch.The dorsal rays in front of and possibly forming the notch appearto be homologous to spines. They differ from the unsegmentedrays behind the notch in being undivided and rodlike. In the Cyclop-teridae and Cottidae the notch separates the spines from the softrays. We should expect to find the same condition in the Liparidae.We may now discuss the fate of the undivided rays or spines inthose species in which the notch is absent. We have seen that inthe Cyclopteridae the absence of the dorsal notch indicates that thespines have been lost. The same is true of the Agonidae. In thesetwo families the notch disappeai-s only with loss of the spines. Butin the Liparidae the notch may disappear and the spines be retained.This becomes very evident when we examine species like Liparisdennyi, in which the notch is faintly indicated or absent. In thespecies of Liparis, in which the notch is absent, it appears that theanterior rays remain undivided and spinelike. The absence of thenotch in this genus does not indicate that the spines have been lost.The fate of the undivided rays in Careprodus and Paraliparisis not so easily solved. In these genera, with a few exceptions,the dorsal notch is absent and the anterior rays are divided.* Theabsence of the notch in these species, judging from what has occurredin Liparis, does not indicate that the spines have been lost. Thereare two other criteria that we may use in determining whether thespines have been lost or become divided. These are the position ofthe origin of the dorsal and the number of rays.The position of the origin of the dorsal in Careprodus and Para-liparis favors the view that the spines have not been lost. In thespecies of Cyclopteridae and Agonidae, in which the spinous dorsalis absent, the origin of the dorsal is proportionally farther back onthe body. If the spinous dorsal has been lost in Careprodus andParaliparis we should expect to find the origin of the dorsal fartherback than in the species of Liparis, in which we know the spinousdorsal is present. The base of the spinous dorsal in Liparis is shorterthan in the Cyclopteridae, but if this fin disappeared there would be * We assume here that in the ancestry of these genera a dorsal notch and spines were present. In two ofthe linown species of Careprocttis a dorsal notch appears to be faintly indicated. The condition of the raysin these species has not been examined. We are justified in assuming, I think, that in the ancestry of thesegenera a dorsal notch and spines were at one time present. REVISION OF THE FISH FAMILY LIPARIDAE 17 a perceptible change in the position of the origin of the dorsal. Theshape of the body is somewhat modified in Careprodus and Parali-paris, but this would hardly obliterate any change that might occurin the origin of the dorsal. The origin of the dorsal in Careprodusand Paraliparis is not perceptibly farther back than in Liparis.This indicates that in these genera the spinous dorsal has not beenlost.The proportional number of dorsal and anal rays also favors theview that the spines have not been lost in Careprodus and Para-liparis. In Liparis the number of dorsal rays (counting the spines)varies from four to nine more than the number of anal rays. Of thesethe anterior 4 to 7 or 8 are undivided or spinous. If these spineshave been lost in Careprodus and Paraliparis the number of dorsaland anal rays should be about equal. But we do not find this to betrue. In Careprodus the number of dorsal rays varies from 4 to 9more than the number of anal rays. In Paraliparis the dorsal hasfrom 3 to 10 more rays than the anal. The spinous dorsal appar-ently has not been lost in these genera. If the spinous dorsal hasnot been lost in these genera it possibly has not been lost in the othergenera in which the notch is absent and the anterior t&js divided.From these general considerations we are led to believe that in theLiparidae the dorsal notch bears the same relation to the spines andsoft rays as in the Cyclopteridae, that the rays in front of the notch, orthe undivided rays, are homologous to ancestral spines and that theunsegmented rays behind the notch and the unsegmented anal rayscan not be so homologized, but represent soft rays which have lostthe segmented condition. The evidence concerning the fate of thespines in those species in which the notch is absent and the anteriorrays divided seems to indicate that the spines have not been lost buthave become divided. Assuming that spines have been formed fromrays by the obliteration of the union of the two lateral shafts and theloss of segmentation we have, in these degenerated fishes, in thefailure of the two shafts of the anterior dorsal rays to coalesce, anexample of a retarded or reverted condition of a structure. 18 BULLETIN 150, UNITED STATES NATIONAL MUSEUMTable showing number of unsegmented dorsal and anal rays in specimens examined Name Liparis callyodon.DoDoDo...DoDoDo .Liparis tunicatusDo .Liparis herschelinusLiparis ailanticusDoDoLiparis rutteriLiparis mucosusLiparis dennyiPolypera greeniDoLiparis agassiziiDoDo...-Liparis majorCarepractus gilbertiDoCareproctus melanurus..Careproctus sinensisCareproctus ranulaCrystallichihys mirabilis.Paraliparis copeiDoDoParaliparis deaniDoParaliparis cephalusNectoliparis pelagicus Lengthin mm. 5340789542554860100?706595?826512523092125165105 Unseg-menteddorsalrays 15141414141512121110172627101511161697815151291514?18?139111018 Unseg-mentedanalraj's 323324?312? i2412321257&334121681044612 1 5 in first dorsal. ' 4 in first dorsal.Anal Jin.?The anal fin bears a close resemblance to the dorsal.The anterior rays appear to be spinous?that is, unsegmented. Thenumber of unsegmented rays is greater in the more specialized gen-era. The connection with the caudal is nearly always greater than thedorsal connection with the caudal. * The dorsal and anal are nevercontinuous, the caudal always being distinct.Caudal Jin.?The posterior margin of the caudal is typically formedof about 10 ra3^s, though there may be as many as 12 or 14. In addi-tion to these there are in most species a number of rudimentary raysat the base of the caudal which may increase the number to 20.The average number of caudal rays is slightly reduced in Care-proctus and further reduced in Paraliparis. The greatest reductionof the caudal is found in RMnoliparis attenuatus, where it is represented REVISION OF THE FISH FAMILY LIPARIDAE 19by a single elongate ray which is connected for about half its lengthto the dorsal and anal. It is thus seen that the change in numberof the caudal rays is opposite to that in the dorsal and anal.The caudal is the most powerful in the shallow-water forms. Therays are stout in these forms but become very slender and sometimeswavy at the tip in the deep-sea species. The fin is typically truncateor slightly rounded. It is never acuminate as has been described. InCareprodus cypselurus and related species it is forked.Pectoral fin.?The process of specialization in the pectoral, as in thecaudal, includes the reduction of the number of rays. The numberranges from 42 in Liparis to 21 in Careprodus and 14 in Paraliparis.In contrast to the trend of modification in the family the change inLiparis is to increase the number in the deep-water species. Inthis genus the pectoral typically has more rays than the anal, whilein the other genera it has less.The outer margin of the pectoral is typically notched. Thisnotch is usually shallow but may extend to the girdle from whichthe intermediate rays appear to have been lost. The notch is notpresent in certain species of Liparis, Careprodus, Acantholiparis, andParaliparis.In Liparis the rays are nearly equally spaced on the girdle, those atthe notch being little more widely spaced than those above or below.The separation of these rays reaches its greatest extent in Parali-paris. Here also we find a resemblance to the primitive condition,for P.fimhriatus has the rays evenly spaced on the girdle.The lower margin of the lower pectoral lobe in Liparis and Care-produs consists of a series of rays regularly graduated in length. Thenumber of rays in the lower lobe of the pectoral in Paraliparis isreduced and the lobe shaped as though the short anterior rays havebeen lost. The lower lobe also becomes elongate, the rays sometimesseparated to the base and coiled at the tip. The greatest elongationof the pectoral, however, is found in the Careprodus longifilis inwhich species it is fully half the length of the bodj^. In other species ofCareprodus, as in C. colletti, the lower lobe is longer than the head.In some species the length of the lower lobe increases with age.The upper edge of the pectoral in Cydogaster is on a level with theeye, and the symphj^sis is on the lower surface of the head behindthe eye. In some species of Paraliparis the pectoral is lowered and as-sumes a more horizontal position. The extreme is found in P. mento,in which the upper edge ef tho pectoral is below the angle of the mouthand the symphysis is far forward on the chin in front of the eye.The greatest reduction of the pectoral occurs in Nedoliparis, apelagic genus, in which the two lobes are separated and greatlyreduced. It probably has no function except that of balancing. Anextremely delicate type of fin not seen by the writer has been 20 BULLETIN 150, UNITED STATES NATIONAL MUSEUMdescribed for Paraliparis memhraneceous . The rays are very fine andthe margin of the fin fringed. In some of the fragile forms the skinconnects with the rays some distance from their base.Pislc.?The disk reaches its greatest development among the tide-pool species, where it is of most service to the individual. The averagefor the number of times the disk is contained in the head in Liparis is2.3 and in Careprodus 4.5. In Careprodus ostentwn the disk is ofminute size, yet perfect in structure. In Elassodiscus the disk isminute and imperfect, the rays having been lost. This represents theintermediate stage toward the loss of the disk.The disk is primitively flat. After it becomes functionless andreduced in size it becomes cupped in some forms as in Careprodusabbreviatus and the Careprodus colletti group. In some of thesecupped disks the margin is stiffened and the base set deeper into thebody.The shape of the disk is round or slightly oval and indented on thesides anteriorly. In a number of species, as in the Careprodusspedrum group and Careprodus rhodomelas, the disk is triangular.This is due to the posterior margin being folded over and the sides,anteriorly, being further indented or folded over. This conditionmay represent the intermediate stage toward the cupped conditionas the completion of the folding over of the margin woidd producethe cup.Vent.?One of the most striking modifications of a character in thisfamily is the change of position and direction of the vent. In somespecies of Liparis the vent is nearer the anal fin than the disk. In thisgenus and Careprodus the vent moves forward until it comes to besituated at the posterior margin of the disk. In Paraliparis the ventmoves into the place once occupied by the disk. In Nedoliparisit is on the throat in front of the symphysis of the pectoral. In allthe genera except Nedoliparis the vent opens downward. In thelatter genus the vent opens forward horizontally above a ridge on thethroat.Coloration.?The coloration in a general way is correlated with theenvironment. The prevailing ground color of the tide-pool speciesand their nearest relatives of the moderately deep water is brown,olive, and slate; of the species from the region of little or no lighttranslucent, pink and black. The deeper-water forms range fromtranslucent to pure black in the greater depths. Light-coloredspecies exist far into the zone of darkness and certain species arefound both in the regions of little light and complete darkness. Theblack species apparently are never found above 400 fathoms. Someof the tide-pool species, as Liparis rutteri, are quite dark, but neverblack. Light-colored species are taken from a black mud bottom,but how far above the bottom they exist is unknown. REVISION OF THE FISH FAMILY LIPARIDAE 21The shallow water forms are typically variegated, being markedwith stripes, bars, blotches, and mottliags. These markings vary agreat deal in the species, and their significance and importance areunknowb, as the variations have not been shown to be correlatedwith sex, season, or age. They can be used to a limited extent indifferentiating species.The deeper water forms, exclusive of Liparis, with but two excep-tions, are never variegated. The two exceptions are Crystallichthyscyclospilus and mirabilis, which have pink blotches on the head andbody. A large number of the species from the intermediate depthshave the posterior part of the body darker than the anterior part.This suggests that the black coloration appears first in the posteriorpart of the body and encroaches on the anterior part as the speciesbecomes wholly black. The pigment is not confined to the dermisbut frequently occurs in the flesh.Bars on the fins of species of Liparis are very common, but arenever found on the fins of the other genera. In some species thesebars extend onto the body, but the latter is never strongly barred.Longitudinal stripes are common among the species of Liparis butare unknown in the other genera. These stripes are white to grayand sometimes show traces of blue. They may have dark marginsand divide posteriorly. Blotches with definite margins are knownin but a few species of Liparis and Careproctus. Indefinite blotchesand markings are common in Liparis. The giant species of Liparisfrom Japan have a reddish lining to the dermis.The internal membranes vary from white to black. The peri-toneum is primitively white with scattered black dots. It is silveryin many of the intermediate forms and black in the majority of thedeep-sea species. The color of the stomach varies as much as thatof the peritoneum, but independently, and is not black in so manyspecies. The mouth and gill cavity also vary from white to black.(For a discussion of the relation between coloration and bathymet-rical distribution see Burke, 1911.)Habits.?The primitive species of Liparis doubtless inhabited shal-low waters, either resting upon the bottom or clinging to the rocks.Many of the present species have the same habits, but a few of themore specialized seem to have acquired free-swimming habits whichare of advantage where the food is scarce. Nectoliparis pelagicus isthe only known pelagic form. It is probable that some species ofParaliparis approach this condition.The eggs of the species are deposited and develop, as far as known,within the general habitat of the species. Liparis atlanticus has beendescribed as approaching the shore to deposit its eggs. Such a mi-gration must be very short. The eggs of the deep-sea species mustbe endemic, as the larvse are unknown from the plankton. Whether91668?30 3 22 BULLETIN 150, UNITED STATES NATIONAL MUSEUM or not these species approach the shallowest waters of their range todeposit their eggs is unknown. The great vertical range of some of thespecies may be accounted for in this way. Some of the deep-sea speci-mens have the body distended with eggs, but they do not furnishsufficient data to indicate that the regular breeding season has beenmodified. Careprodus ovigerum is known from a single specimen,male, and "contained in its mouth, when captured, a spherical massof eggs apparently of the same species. " The specimen was obtainedon the 3d of September and the eggs were, "well along in their de-velopment, the embryos distinctly visible through the very toughegg membranes." Putman (1873, p. 339) records that the spawn-ing season for Liparis is in March. Ehrenbaum (1905) states thatLiparis liparis spawns from November to February. Smith (1898)describes Liparis liparis as full of spawn in December and January.The very extensive dredging in the North Sea has demonstratedthat the young of Liparis montagui can be obtained in March, orpossibly earlier. These facts indicate that the tide-pool species of thefamily probably spawn during the winter months. Whether this istrue of the deep-sea species remains to be investigated.DISTRIBUTIONThe Liparidae is a boreal family. The large majority of the spe-cies are found in the North Pacific, North Atlantic, and the ArcticOcean. A few species have been taken in the great depths of thetropical Pacific and in the shallower cold waters of the Antarcticregions. It is highly probably that the family is represented in thecold depths of the tropical Atlantic and Indian Oceans.Of all the regions inhabited by these fishes the North Pacific isthe most favored. All of the 13 genera and about three-fourths ofthe species are recorded from this region. Only three of the generaand one-sixth of the species inhabit the North Atlantic. The samethree genera and six species are recorded from the Antarctic regions.Seven of the genera are monotypic and when better known their dis-tribution may be extended. Further exploring of the North Atlantic,Antarctic, and greater depths of the tropical regions will probablygreatly extend our knowledge of the distribution of these fishes.Genera.?The genus Liparis is represented in the cold shallow watersof the Northern and Southern Hemispheres. The species of thisgenus are the most numerous in the North Pacific. The genusis limited in its distribution toward the equator by the summerisotherm of about 60? F. The distribution of Careprodus, the nextlargest genus, differing from that of Liparis in being absent from thetide pools and, descending to greater depths, is continued into thetropical Pacific. The distribution of Paraliparis, the third largestgenus, resembles that of Careprodus. The remaining ten genera REVISION OF THE FISH FAMILY LIPARIDAE 23 are either monotypic or composed of a few species and confined tothe North Pacific. Nedoliparis, consisting of a single pelagic species,ranges widely over the North Pacific. Polypera consists of threeshallow-water species distributed along the shores of British Columbia,Aleutian, and Kurile Islands. Temnocora consists of a single speciesfrom Bering Sea. Crystallichthys is represented by two speciesfrom Bering Sea and Pacific Ocean off Kamchatka. Crystalliasis a monotypic Japanese genus. Gyrinichihys consists of a singleBering Sea species. Elassodiscus and AcantJioliparis are representedby a single species from southeast of Kamchatka. RMnoliparis con-sists of two species which are distributed from California throughBering Sea to Japan.Species.?It is generally said that deep-sea fishes are more cos-mopolitan in their distribution than shallow-water forms. It isassumed that the environment of the deep sea is quite uniform foreach depth and that the species will not be prevented from spread-ing widely over the bottom of the ocean Many of the specieshave a great vertical range which would enable them to migratemore readily. A great vertical range suggests that either the en-vironment changes vary gradually with depth or that the speciesare adapted to a varied environment.At the present time there appears to be a tendency to believethat deep-sea fishes are not so widely distributed as we have beenaccustomed to suppose. This view is substantiated by the deep-sea Liparids. These species appear to be little more cosmopolitanthan their relatives of the shallow water. In considering this state-ment it should be borne in mind that we know little about thesefishes. The deep-sea species are sluggish and, with but one or a fewexceptions, live upon or near to the bottom. The majority of thespecies are known from but a single locality or region. None arecommon to the Atlantic and Pacific Oceans or to the north and tropi-cal Pacific. The majority of the American and Japanese speciesare distinct. Only one species of RMnoliparis and two species ofCareproctus are found in both the American and Japanese faunas.The maximum range in latitude for the deep-sea and shallow-water species is about the same. A few species of Liparis, Care^proctus, Paraliparis, and RMnoliparis range through about 20? oflatitude. The majority of the species have a more restricted range.The distribution of the species of Liparis, Careproctus, and Parali-paris appears to be quite comparable. This similarity may disappearwhen we come to know the deep-sea species better.The giant species of the family are confined mainly to the north-west Pacific from Bering Sea through the Okhotsk Sea to northernJapan. This is especially true of the large species of Liparis, Care-proctus, and CrystallicMhys. The largest species of Paraliparis, a 24 BULLETIN 150, UNITED STATES NATIONAL MUSEUMmore typical deep-sea genus, are more widely distributed throughoutthe range of the genus. In addition to the largest species the largestspecimens of such species as extend from the Alaskan coast to theOkhotsk Sea are obtained in the latter region. The coldest tempera-tures for the Pacific Ocean were taken in the southern part of theOkhotsk Sea. A temperature of 30? to 31? F. is common in thisregion in depths of 100 fathoms. The Arctic current sweeps throughinto the Japan Sea, marine life is abundant, and with the extremelow temperature the Liparids find a favorable environment for anincrease in size. It will be interesting to note whether the largestspecies of other northern families are found in the Okhotsk Sea andneighboring regions.^Regions.?As the species are quite localized in their distributionwe can readily divide the area of the globe occupied by these fishesinto regions each of which is distinguished by a group of species notfound in the other regions. A few of the species are found in two orthree regions, but the majority are known from but a single region.We have found it convenient to designate seven regions, as enu-merated below. These regions are not of equal value, but are con-venient for illustrating the peculiarities of the distribution of thefamily.1. Asiatic region: This embraces the southern coast of Kamchatkaand the region southward to Japan. This region may be subdivided,for the Okhotsk Sea and the Japan Sea appear to have a fauna dis-tinct from that of the east coast of the Japanese Islands.2. Bering Sea region: This region includes Bering Sea and its shoresto the northward, not including the Aleutian Islands.3. Pacific American region: Aleutian Islands and southward toCalifornia.4. Atlantic American region: Coast of New England to Greenland.5. European region: Coasts of northern Europe.6. Tropical Pacific region: Depths of the Pacific Ocean between theTropic of Cancer and the Tropic of Capricorn.7. Antarctic region: Regions south of the Tropic of Capricorn,The following table designates the species found in each region: ' For a more detailed discussion see under Liparis and Careproctus. REVISION OF THE FISH FAMILY LIPARIDAE 25 26 BULLETIN 150, UNITED STATES NATIONAL MUSEUM Name NorthPacific pq NorthAtlantic aOiaou Nedoliparis pelagicusCareprodus gilberiiParaliparis dadylosusParaliparis ulochirParaliparis cephalusRhinoliparis attenuaiusLiparis rutteriLiparis mucosusLiparis grebnifzkiiLiparis micraspidophorus.Liparis pulchellusLiparis fucensisLiparis dennyiPolypera beringianusPolypers greeniCareprodus spedrumCareprodus melanurusCareprodus abbreviatusCareprodus ovigerumAcantholiparis opercularis.Paraliparis deaniParaliparis mentoParaliparis rosaceusLiparis atlanticusLiparis tunicatusCareprodus ranulaParaliparis liparinaParaliparis garmaniParaliparis copeiLiparis liparisCareprocius micropusLiparis majorLiparis moniaguiCareprodus reinhardiCareprodus longipinnisParaliparis membranaceus . Paraliparis bathybiusParaliparis grandicepsParaliparis longifilisParaliparis angushfrons- _Paraliparis fimbriatusParaliparis latifronsParaliparis attenuatusLiparis antardicaLiparis steineniCareprodus falklandicaCareprodus pallidusCareprodus georgianus XX?X?X?XX X? X? X? XXXXXXX?X?X 0-0-92-284- 300- 609102- 48229fr- 482406-1, 005284- 540576++++0-f-2121230+0+no8766251,588682-1, 97330- 274685984++365480540594+55- 976?8- 116?0+263- 658702400540- 6581,5881,8235111,7721,793902++8- 8215107 11?328-300-300- Center of dispersal.?The Liparidae has reached its greatest develop-ment in the North Pacific. Either the Okhotsk Sea or the BeringSea appears to be the center from which the present groups have REVISION OF THE FISH FAMILY LIPARIDAE 27 radiated. Some of the criteria ^ generally used in determining thecenter of dispersal of a group point to one of these regions and someto the other as the center of dispersal for the family. We shall notattempt to decide between these two regions, but shall simply citethe criteria which point to the northwest Pacific as the center fromwhich the existing groups have dispersed.1 . Most primitive species : This resolves itself into a discussion ofthe center of dispersal of the most generalized genus Liparis. Thecriteria utilized here, when applied to Liparis, indicate the NorthPacific as the place of origin of the genus. The primitive species ofLiparis are in the North Pacific.2. Greatest differentiation of type: Apparently all of the generaoriginated in the North Pacific. Only three of the 13 genera arefound in any other region. The most generalized species of thesethree genera exist in the North Pacific. It follows that the generaare most closely related in this region.3. Giant species: The largest species of all the genera exceptParaliparis are found in the North Pacific. The specific attain thegreatest size about northern Japan. Those species which extendfrom the American to the Asiatic coast attain the greatest size in thelatter region. The fact that the species attain the greatest size inAsiatic waters does not necessarily indicate that this region representsthe center of origin. It may simply mean that the conditions hereare the most favorable for growth.4. Dominance or abundance of species: More than half of thespecies exist in the Bering Sea and southward to Japan. Three-fourths of the species are recorded from the North Pacific.5. Geographical consideration: We may readily account for thepresent distribution of the family if we assume that the groupsdispersed from the North Pacific. The only barrier that we have tocontend with is the equatorial region. This effectively acts as abarrier to the shallow-water species. We have already attempted anexplanation of the presence of tide-pool species in the Antarcticregion. We have assumed that the species were able to cross theEquator during the glacial period. Whether or not this explanationis satisfactory we must account for the crossing of the equatorialregion if we should assign to any other region the center of dispersalfor the family.6. Lines of convergence: In tracing out the lines of convergencefor a family we are concerned mainly with the genera. If all thegenera have originated in one region we are justified in designatingthat region as the center of dispersal for the family. It is statedunder "Greatest differentiation of type" that aU the genera arerepresented and probably arose in the North Pacific. ? See C. 0. Adams, 1902, and A. Q. Ruthven, 1908. 28 BULLETIN 150, UNITED STATES NATIONAL MUSEUMBathymetrical distribution.?The Liparidae are found at all depthsfrom the tide pools down to 1,973 fathoms. The species are quitecommon in the tide pools and down to 800 fathoms. Only 15 of the114 species are recorded from below 800 fathoms, and 10 below 1,000fathoms. Perhaps the number of species taken below 1 ,000 fathomsbears the same relation to the number taken above this level as thenumber of dredge hauls below 1,000 fathoms bears to the numberabove. Doubtless the proportionate number of deep-sea forms willbe greatly increased.The three large genersi?Liparis , Careproctus, and Paraliparis?while overlapping in their vertical distribution, are confined mainlyto different depths. The genus Liparis is represented down to 250fathoms, but it is typically a tide-pool genus. Careproctus is typicallya genus of the dimly lighted regions. It is represented at all depthsfrom 30 or less to 1,823 fathoms, but half the species are found aboveand half below the 300-fathom level. For convenience we havedesignated this level as the center of population for the genus. Para-liparis is typically a deep-water genus. It is represented at alldepths from 30 to 1,793 fathoms, but the center of population is atabout 500 fathoms, or 200 fathoms below that for Careproctus.Polypera is closely related to Liparis and is confined to the same region.Temnocora, CrystallicTitJiys, Crystallias, and GyrinichtJiys are confinedto the same general region as Careproctus. Elassodiscus is confinedto greater depths. Acantholiparis and RJiinoliparis inhabit the samedepths as Paraliparis.The bathymetrical distribution of the species varies widely. Thetide-pool species have a vertical distribution of a few feet or fathoms,while the bathybial species range through several hundred fathoms.The maximum vertical distribution for species of several of thegenera is as follows : Rangefathoms DepthLiiparis dennyi 123 0- 123Liparis fucensis 213 0- 213Careproctus cypselur us 377 510- 887Careproctus mollis 347 135- 482Careproctus gilberti 380 102- 482Careproctus colletti 341 284- 625Careproctus melanurus 592 284- 876Acantholiparis opercularis 1, 291 682-1, 973Paraliparis ulochir 599 406-1, 005Paraliparis holomelas 1, 219 406-1, 625Rhinoliparis barbulifer 359 192- 551The vertical distribution of these species is of interest in connectionwith the region between 250 and 500 fathoms. Somewhere withinthis region is supposed to be the indefinite borderland between twodifferently colored faunas. The red light rays fail to penetratebelow 273 fathoms. Whether this level or some other is the most REVISION OF THE FISH FAMILY LIPARIDAE 29important in marking the region between two differently coloredfaunas is discussed under another heading. What we wish to notehere is that none of the species are found on both sides of the 250 to500 fathom region. Certain species extend from far above into thisregion, others appear to be confined to it, and still others extend fromit to the regions far below. The facts seem to indicate that thisregion marks a distinct change in the environment. It is a dimlylighted region and favorable to the development of the Liparidae.The region at from 20 to 30 fathoms separates the typical tide-poolspecies of Liparis from the shallow-water species of Careproctus andParaliparis. It is doubtful if any of the tide-pool species extend intothis region. A number of species of Liparis cross it, but none ofthem are known with certainty to inhabit the tide pools. Threespecies of Careproctus? C. falHandica, C. pallidus, and C. ranula?are reported from depths of less than 20 fathoms. The remainingspecies of the deeper-water genera extend from 29 fathoms downward.The following table indicates the bathymetrical distribution ofmost of the species of Liparidae. 30 BULLETIN 150, UNITED STATES NATIONAL MUSEUM -fa REVISION OF THE FISH FAMILY LIPARIDAE 31oeo00 00 00 00 OOU5OOtJ^O ooococo C<1(N +++OOO ??^S s s s o o e ?o. fc S. la COO '? _o NCO ci CO CO CO ^ o S9J0NNp>-l?oo>'0'? 22?-l>0 OC4COCOOO>H Boaoo 3CO oo?oOl00os^^?Dooo?ooot^co?o^o^^^^o^^c^_l_c^c^oJMOOCOOJ^Oi-HCO C^COt-l C^COrHtOMCOCOCOCOCO-HjIriOX"'^'^ ^ ?oooo CO t*?1^ M C< J;3 IBpnBo N 1 1 I lO ONCMOCOO 000000s[Bjojoej 00 CI 00 OS o 1?( c^ ?-* "^ I-* ?o h* 00 1^ o t- 1^ o> o d ?-< o c< ?-< o 1^ ?o >o t>- 1^ o f^^eoioeo^t>IN CO Ceo pj e^ CO CO CO CO CO ^ CO CO CO CO CO CO CO CO*** CO CO *'?' CO CO CO cocO'C cocococociweo IBuy . >0 TJH to ^^ CD r- ^ OS O ^H 00 C^l CD ^ tC -^ CS * M* CD t^ t^ -^ CO 10 iC OS t^ O 1-4 ^H ^ h- CO ?-? (^ t^ C^ COC^ C^ C^ CO CO CS CO CO CO CO CO CO CO CO CO CO CO CS CO CO CO CO CO ^ COCOCO C0^^^'^?0'*C^C^NC^C^C^C^C^C^COCOCSCOCOCOCOC C C C C C C C IBSJO a <*4 *-H 00 -^ (M 'sf CO i-H M< lO t^ lO ao CO c^ -^ c^ OS c^ o ?o 10 iT) 00 ''t* CO CO 00 r^ 00 000;^ cs OS r--. *-! t^ GO COCO CO C^ CO CO CO CO CO CO CO CO CO CO * 'J" * "S" CO * 'J" ?????>'"?' M ^ '?">>** * ^CO* ??H -^ * 10 't" "5 "5 q?d9a +2++0^00^ ooocsoc^es J I 00 ?-i I I Oco c^j L ososoe^ooscT^lOO I ' r CSC^?0 00 0?-HCO'-i'-i 000 cqocoYf^"-"s^?^ "-1 00 saennoadg _l_i^Ni-t^ l^rHC0C^C0?-tr*(Ni-f?0C0CDi-tu3iHWC0*-t?OOT^ COi-H -I rH M iCO"-l<-l(N g i|.. ss , V. , I I >> 1 I I I I I I I ' I I I I IiA -.M ', t > < < -M > ?S'^ ,^ I , , , , , , .^ . . ; . Ioaioa>ooooo"o o w oomii>ooooooo"oo?oonri^nFL, 1 ; ; ; jm ; \pn iqph ; < l J J jn ? Ji^ 0000"]M^ 3t3 ~ >? ' 00000^0 >>? J o >> ' 0000 g" ' o o ^ ^ J g'_o ^ tiC ^ bfl 1 '1 I 1 1 bo 1 3 be c8 I S ^ I I ? f 3 tj) ', ', bo bc bc 3 \ to;3p3 : : 1 i :3 ip3S ;q3 ! 1 : ;p3 1 i333Q ; 3 'J2 08, bJ) bO5 : :33 to ^ i i i i-H^ ooo^Sja o^ I 1 lE-i 1 ! I 1m&^noijoenuoolepriBO-osjoo: ^eo J^?o5^j:^e<5 0?o p-'j'M JC'*>o-o (N Srjrw "?? CD lO U5 ?0eXa t^ W31OC0CC CO^ ^STQ N t^t^eoc^^,^ CO '?** CO 00 01 10 OS ?* to 00 00 10 _i_ic CO CO 00 CO dM ^ CO (N 00 CO^^ C^CO^cd^iOCOCD^C^COTl5t--'^t-^t^a>05COCOCOC^'^CO coco o*i-H I IS9J0J j:rPPC^ ??, B0300 tOO--<'>*'5D C05DlepnBQ 000 I ICO icDOaO I 100 lOO 100000(0X000 IBJO}09daa OiOC^O'L0i000G0?00st^C00St~-i0OC000CDCSW000l05C^O0s0S00CDOi0iO<-HCO'^OCDu:iOOOOit^C^GOC^OJC3Cp?OCO*-t tliTt^^cOCDcDi-1 i I OOi-iOOCOQOcO'OOt-i-HCO'-HT+J-^j^eOi-HO-^ '^^j^ clcoi ^^-i- -L -jr-r-r_r J. A J. J, A A-L -1 12 A3COC^aio4"*COOc^ r- ,-? o o 1?? 00 -> J p ,q3 5iJ o o I I ! I lO I I ) Ho'do lO ^lo lo-^pwppppco "3 00 10CO lous^j;^ connwj I I I I I I I I I I I I I I I i I I I Ig^rtW-H^g-HrtpgW^g-Irt,-!^ 1 38 BULLETIN 150, UNITED STATES NATIONAL MUSEUMExplanation of table.?The list of specimens refers, in nearly allcases, to the specimens examined by the writer. When only a fewspecimens are recorded for a species they are usually included. Thedepth refers to the fathoms from which the species have been taken.The depth for the tide-pool species is given as + . The gill-slitcolumn refers to the number of pectoral rays in front of which thegill slit extends. The pores refer to the number of pores above thegill slit. The disk refers to the number of times the diameter of thedisk is contained in the head. The eye refers to the number of timesthe diameter of the eye is contained in the head. The dorso-caudalconnection refers to the proportional length of the connection betweenthe dorsal and caudal fins and the length of the caudal fin. Thevariegated coloration means that the body or fins are speckled, barred,or striped; the uniform coloration that the body and fins are notvariegated; the fight coloration that the body is whitish or pinkish;the light to dusky or black coloration means that the body is whitishor pinkish anteriorly and dusky or black toward the caudal.PART 2. SYSTEMATICFamily LIPARIDAEFamily description.?Body cavity short ; tail more or less elongate ; scales absent or represented by prickles; lateral line reduced to oneor two pores above the gill slit; third suborbital styHform, joined tothe preopercle as in the Cottoids and related families ; teeth tricuspidor simple, in bands or single series, absent from the vomer and pala-tines; premaxillaries sHghtly protractile; opercular bones unarmed,the opercular arms sometimes projecting as spines; interopercle ray-like, overlying the branchiostegals ; gill opening reduced, varyingmuch in size, never connected across the isthmus; gills 3J^, no slitbehind the last; pseudobranchiae rudimentary or absent; pyloriccoeca numerous to absent; branchiostegals 5 or 6; no air bladder;ventral fins present or absent, when present 1-5, united to form asucking disk; dorsal fin continuous, sometimes notched, the spinesflexible; anal similar to the soft dorsal; pectoral fin broad, in somespecies divided into an upper and lower lobe; vertebrae 35 to 70.Genera 13; species 114.The Cyclopteridae and Liparidae may be distinguished as follows:A '. Body cavity elongate, about as long as the caudal region; anal fin with lessthan 20 rays Cyclopteridae.A 2. Body cavity short, shorter than the caudal region; anal fin with more than20 rays Liparidae . REVISION OF THE FISH FAMILY LIPARIDAE 39RELATIONSHIPS OF THE GENERAUPABISThe species of the genus Liparis bear the closest resemblance tothe Cyciopteridae, and we may reasonably assume that this genus isthe most generaUzed one in the family and the one from which theother genera have directly or indirectly been derived. It is amongthe tide-pool species of the genus that we find the closest resemblanceto the Cyciopteridae. These species have the smallest number ofvertebrae and fin rays, the shortest connection between the dorsals,caudal, and anal fin, and the dorsal notched. L. callyodon is a goodexample of this type of species. The characteristics which this specieshas in common with some or all the other species of the genus andwhich point to this genus as being the most primitive are the smallnumber of vertebrae, the dorsal notch, the short connection betweenthe vertical fins, the trilobed teeth, the variegated coloration, and thetwo nostrils.It is, of course, possible that the most primitive species is bestrepresented by some of the deeper-water species such as L. megace-phalus. This species has a very wide gill slit. Prickles, presumablya primitive character, are most common among the deeper-waterspecies.It seems more likely that the genus originated in the tide pools andthat representatives descended to greater depths, retaining certainprimitive characters as the enlarged giU slit and prickles and becomingmodified in other ways, such as the increase in the number of vertebraeand the connection between the vertical fins. The species remainingor originating in the tide pools doubtless retained a larger number ofprimitive characteristics though becoming modified in some ways.They may have developed more compact, shorter bodies, larger disks,and smaller gill slits and thus became better fitted to withstand thebeating of the waves on our rugged coasts. It appears to us that thetide-pool species, such as L. callyodon and L. mwcosws, in their aggregateof characters most nearly resemble the ancestors of the group and theCyciopteridae. POLTPEKAThe genus Polypera differs from Liparis solely in the greatly in-creased number of pyloric coeca. All the species of Polypera have morethan 200 closely matted pyloric coeca and none of the species ofLiparis have 100. The pyloric coeca in Polypera are in a thick massand can be distinguished at a glance without counting from thelarger coeca of the species of Liparis. In the distribution, small gillslit, the notched dorsal, and the short connection between the verticalfins Polypera most closely resembles the tide-pool species of Liparis.In the increased number of fin rays Polypera bears a closer resemblanceto the deeper-water species. 40 BULLETIN 150, UNITED STATES NATIONAL MUSEUM CAREPR0CTTJ3Careproctus and all the remaining genera of the family differ fromLiparis and Polypera in the single nostril and the lack of a variegatedcoloration. It is not among the deep-water species of Liparis butrather among the shallow-water species with the notched dorsal thatwe shall look for the primitive Careproctus stem. In certain char-acters such as coloration?that is, black peritoneum or silvery abdo-men?and reduced posterior nostril Liparis major and Liparistessellatus approach the condition found in Careproctus. The formerspecies also has some simple teeth. These similarities, however,are only the result of parallelism or convergence. The primitiveCareproctus arose from an ancestor with a notched dorsal. Thedorsal notch is slightly evident in Careproctus pycnosoma and ectenesand possibly in Careproctus attenuatus. Also the genus Temnocora,agreeing with Careproctus in the single nostril and coloration andprobably derived from the same stem, has the dorsal notched. Care-proctus pycnosoma, in addition to the notched dorsal, appears to be themost primitive member of the genus in the compact body, smallnumber of fin rays, and the strongly developed anterior dorsal rays.In the character of the notched dorsal and the number of fin raysCareproctus pycnosoma resembles Liparis dennyi. We must, however,if we are to consider Temnocora as having been derived from the samestem as Careproctus, picture our hypothetical ancestor as having amuch deeper dorsal notch than Liparis dennyi.The species of the genus Careproctus vary widely in a number ofcharacters, and we may well inquire if the genus is not polyphyletic.We have seen that Liparis major and Liparis tessellatus approachCareproctus in the reduced posterior nostril and the coloration. Withthe closure of the posterior nostril and the loss of the variegatedcoloration these species would resemble typical species of Careproctus.There is hardly sufficient evidence to cause us to more than brieflyconsider the possibility of Careproctus being polyphyletic.TEMNOCORATemnocora difiers from Careproctus in the slitlike pupil and the moreprominent dorsal notch. It agrees with Careproctus in the singlenostril and the coloration and probably has been derived from thesame stem. Temnocora Candida has no near relatives among theknown species of Careproctus. In the shape of the body and thecoloration it resembles Careproctus pJiasma and related species. Inthe character of the pupil the species bears a resemblance to Crystal-licJitJiys and Crystallias. This resemblance may indicate geneticrelationships or only convergence. REVISION OF THE FISH FAMILY LIPARIDAE 41CEYSTALUCHTHY3Crystallichthys differs from Careprodus in the blotched colorationand the peculiar slitUke and notched pupil. It apparently has beenderived from the Careprodus stem, possibly after the loss of the dorsalnotch. CRYSTALUASCrystallias has apparently been derived from a Crystallichthysmirahilis-like ancestor. It resembles this species in coloration,shape of head and body and the folds on the snout. It differs fromC. mirahilis in the presence of barbels on the snout and the absenceof the notch in the pupU or the presence of a flap on the upper edgeof the iris. Kudiments of barbels are possibly present in Crystal-lichthys mirahilis in the folds of skin between the pores on the snout.GYRINICHTHY3The genus Gyrinichthys differs from Careproctus solely in thegreatly reduced gill slit. It apparently has been derived from thespecialized simple toothed species of Careprodus. With the discoveryof a species of Careproctus having the gill slit but little smaller thanin some of the known species the genus Gyrinichthys will have to bereduced to synonomy. ELASSODISCUS ^ Elassodiscus differs from Careprodus in the rudimentary conditionof the ventral disk, the rays being absent. The nearest approachto this condition is found in Careproctus ostentum, in which the diskis minute but perfect in structure. Elassodiscus, however, has beenderived from a different line of development than C ostentum forthe teeth are stronglj^ trilobed in the former and simple in the latterand related species. The species of Careproctus with trilobed teethall have the disk well developed. We must then conclude thatthere is a considerable gap between Elassodiscus and the generalizedspecies of Careprodus with trilobed teeth from which it must havedeveloped. PARALIPARISParaliparis apparently has been derived from the Elassodiscusstem. The only difference between the genera is the absence ofthe rudimentary disk in Paraliparis. The genus could not hav?arisen from an ostentum-like ancestor, for the most primitive species,P. dactylosus, has trilobed teeth. It is possible that the Careproctusostentum branch gave rise to species without disks and that Parali-paris is diphyletic. RHIN0UPARI3Rhinoliparis differs from Paraliparis solely in the presence ofbarbels on the snout. The presence of lobed teeth in Rhinoliparisharhulijer suggests that the genus arose from the primitive Parali-paris stem. 42 BULLETIN 150, UNITED STATES NATIONAL MUSEUMACANTH0UPAEI8AcanthoUparis differs widely from any known genus, and itsorigin is in doubt. It differs from Paraliparis in the absence ofpyloric coeca and the projection of the opercular arms from the sideof the head as spines. The only character which AcanthoUparisshares with Paraliparis and not with Careproctus is the loss of thedisk. It appears most likely that AcanthoUparis has been derivedfrom the Paraliparis stem. NECT0LIPARI3 8NectoUparis differs widely from all the other genera in havingonly five branchiostegals, the gill slit restricted to the front of thepectoral, and the vent situated on the throat and opening forward.^The absence of the disk suggests that the genus may have been de-rived from Paraliparis. The single series of teeth in N. pelagicusresembles the condition found in P. copei and related species.In this discussion it has been assumed that the trilobed teeth anddorsal notch wherever found are primitive and not secondarilyacquired and that the loss of the posterior nostril has taken placealong only one line of development and indicates genetic relationships.KEY TO GENERAA*. Nostril double; pseudobrauchiae present; pectoral typically with more raysthan anal.B'. Pyloric coeca less than 150, probably always less than 100 Liparis.B*. Pyloric coeca more than 150, probably always more than 200 Polypera.A*. Nostril single; pseudobranchiae apparently always absent; coloration notvariegated; pectoral typically with less rays than anal.C*. Dorsal notched.D >. Coeca present Temnocara.C Dorsal unnotched, or if notched the pupil round.E'. Disk present.F'. Disk perfect.G ^ Snout without barbels.H ?. Gill slit well developed.J'. Pupil round or but slightly oval; no color blotches on thebody Careproctus.J'. Pupil reduced to a horizontal slit; body with large round-ish or irregular blotches Crystallichthys.H*. Gill slit reduced to a pore Gyrinichthys.G*. Barbels present on the snout Crystallias.F^. Disk rudimentary, the rays absent Elassodiscus. ? In a recent article Johnsen (1919) questions the validity of this genus. We are unable to agree withhis opinion. ' A new genus, Lipariscus, has been described since the above was written. Its relationships withNectoUparis will be found discussed in the appendix on page 194. REVISION OF THE FISH FAMILY LIPARIDAE 43 ^?. Disk absent.Ki. Branchiostegals 6; vent vertical; gill slit at least partlyabove the pectoral fin.L^. Opercular arms not projecting as spines; coeca present.M^ Snout without barbels Paraliparis.M2. Snout with barbels Rhinoliparis.L2. Opercular arms projecting as spines; coeca absent.Acantholiparis.K2. Branchiostegals 5.O'. Vent forward on throat, opening forward;gill slit restricted to the front of the pectoralfin; pectoral lobes separate Nectoliparis.O*. Vent posterior in position, between pectorallobes; gill slit above pectoral; pectoral lobesconnected by widely speced rays-Lipariscus.Genus LIPARIS ArtediLiparis Artedi, 1738, Genera, 117. ? Scopoli, 1777, Introd. Nat. Hist., p. 453. ? And of various writers.Cyclogaster Gronow, 1763, Zooph. Girard, 1858, Pac. R, R. Rept., vol. 10,Fishes, p. 131; Int. Comm. zool. Nomen., Op. 20. ? Lonnberq, 1899, NotesSwed. Arct. Exp. ? Ehrenbatjm, 1902. ? Gilbert and Burke, 1912, a, 6.Burke, 1911, 1912, a, b.Actinochir Gill, 1864, Proc. Acad. Nat. Sci. Phila., vol. 16, p. 193 (major).Neoliparis Steindachner, 1875, Ich. Beitr., vol. Ill, p. 54 (mucosus).Careliparis Garman, 1892, Mem. Mus. Comp. Zool., vol. 14, No. 2.Lyoliparis Jordan and Evermann, 1896, Rept. U. S. Fish Comm., 1895, p. 451(pulchellus) . Disk present; nostrils 2; teeth trilobed; suprabranchial pores 2;anterior dorsal rays spinous; pyloric coeca present, less than 150;pseudobranchiae present; branchiostegals 6.MODIFICATION OF CHARACTERSBody.?The typical tide-pool species have a short, stout bodywhich is rather wide and depressed anteriorly. The shape of thebody in some of the deeper water species differs from that of thetide-pool species. In Liparis owstoni the body is greatly compressedand deepened, in Liparis rhodosoma more elongate and slender, andin Liparis megacephalus and major shortened, deepened, and com-pressed. In a few exceptional cases the shape of the body may help todistinguish the species.The body is typically firm in this genus. In some of the species,mainly the deeper-water forms, such as Liparis cyclopus, owstoni,and rJiodosoma, the skin and flesh are somewhat softer and approachthe condition found in Careproctus.Nostrils.?Both nostrils are always present in Liparis. In Liparismajor and tessellaius the posterior nostril is reduced in diameter and 44 BULLETIN 150, UNITED STATES NATIONAL MUSEUMforecasts the closing of the opaning. The posterior nostril opensabove the anterior part of the eye. The tube never projects as highabove the surface of the head as the anterior nostril tube. In a fewspecies, as in Liparis owstoni, the rim of the tube is distinctly raisedabove the surface of the head. In Liparis rutteri, callyodon, dennyi,and several other species the rim anteriorly is raised into a fingerlikeprojection. The anterior nostril tube always projects above thesurface of the head. The length of the tube varies somewhat, butit is doubtful if this variation can be utilized to any extent in separatingspecies. Some specimens of Liparis rutteri and major have a veryshort tube, there being a variation within the species. In Liparistessellatus the anterior tube projects hardly more than the posteriortube in some of the other species.Eye.?The proportionate size of the eye varies considerablythroughout the genus and is of value in distinguishing species. Asthe size of the eyes varies with age too much stress must not be placedupon a difference in the sizes of the eyes of two specimens unless theseare of the same length. The eye varies in the different species from4 to 10.5 in the length of the head. In some species, such as Liparismegacephalus and agassizii, the lower half of the eye is silvery. ThesUvery pigment frequently disappears in preserved specimens.Pores.?The pore formula for the genus is 2 on the snout, 6 in themaxillary series, 7 in the mandibular series, and 2 above the gill slit.These numbers are constant for all the specimens examined by thewriter. The position of the snout pores in relation to the anteriornostril, the maxillary pores in relation to the eye, and the supra-branchial pores in relation to the gill slit and to each other varies,but these variations appear to be of little practical value in separatingclosely related species.In many of the species rudimentary pores can be seen in the regionof the lateral line and scattered in the region of the gill slit and nape.These are always closed, the lips can sometimes be seen at the tipsof small papillae. The papillae are always white and frequentlysurrounded by a dark ring. When the skin is transparent and laxit is difficult to detect these pores. They have not been demon-strated in all the species. Their taxonomic importance has yet tobe investigated.Teeth.?All the species of Liparis have trilobed teeth. In Liparismajor and ingens the inner teeth are simple or weakly trilobed. Theteeth are arranged in oblique rows and form broad bands. Thereare usually more oblique rows in the upper jaw. The oblique rowsare usually distinct though in some of the giant species they becomeirregular and obscure. In some of the species the number of obliquerows appears to be the same in the young and adults and in otherspecies different. REVISION OF THE FISH FAMILY LIPARIDAE 45The taxonomic value of the number of teeth has yet to be thoroughlydemonstrated. The writer counted the rows of teeth in the majorityof the species but neglected to investigate the amount of variationwithin a species. It is possible that certain species may be separatedby the number of oblique rows of teeth.Gill slit.?The size and position of the gill slit are of great taxonomicvalue and are utilized for distinguishing the species in both the keysand descriptions. The gill slit, when small, is restricted to the sideof the head above the base of the pectoral fin. When large, the gillslit extends down in front of the base of the pectoral but is neverconfined to this region. The gUl slit varies in position from abovethe pectoral to down in front of 16 or more pectoral rays. The sizeof the gill slit varies little with age and within a species never variesmore than in front of four or five pectoral rays. As the gill slit isconstant in size and position for each species and has become modifiedin a heterogenetic manner in the different branches of the genus,it can be utilized to great advantage in distinguishing species.As we should expect, we find certain species, such as LiparisherscTielinus , in which some specimens will have the gill slit abovethe pectoral fin while others will have it extending down in front ofthe upper pectoral rays, the normal condition being either above thefin or extending down in front of two or three rays. In some speci-mens the gill slit on one side will be above the fin and on the otherside extending down on front of the upper ray. Such specimensare difficult to identify by this character.The gill slit is typically above the pectoral fin in the tide-poolspecies. The deeper water species typically have a larger gill slit.Dorsal Jin.'?The dorsal fin presents three important specific char-acters, the number of rays, presence or absence of the notch, andthe extent of the connection with the caudal fin. The number ofrays varies from 28 to 48. The number never varies more thanfour or five within a species. The dorsal notch is more characteris-tic of the shallow water than of the deep water species. The dorsalmay be free from the caudal or connected with three-fourths thelength of the latter.The dorsal notch has formerly been used to separate the genusNeoliparis from Liparis. The notch, however, is not of specificvalue in Liparis dennyi and hristolense and other species. The dorsalnotch is present in some specimens and absent from others. Theanterior dorsal rays in the specimen from which the notch is absentdo not increase regularly in length but are about equal and form ahorizontal outline.The dorsal notch apparently bears no relation to the unsegmentedrays. The number of unsegmented rays also appears to be inde-pendent of age. The number of rays in front of the notch is seldom 46 BULLETIN 150, UNITED STATES NATIONAL MUSEUMmore than six. The following table presents the facts from whichthe above conclusions have been drawn. Name EEVISION OF THE FISH FAMILY LIPARIDAE 47 water species Liparis owstoni and tessellatus have the dorsal connectedto fully three-fourths of the caudal. In some of the species the lastdorsal ray is shortened, this forming a notch between dorsal andcaudal. This notch appears to be of little taxonomic value becauseof great variabiUty. It may be constant for some species but thewriter never found it so. When the notch is present or when thedorsal connection to the caudal is short the outline of the posteriorpart of the fin descends abruptly to the caudal. In Liparis pul-chellus and tessellatus, in which the dorsal is broadly connected tothe caudal, the outline of the dorsal slopes gradually and indistinctlymerges into the caudal outline.Anal Jin.?The important characteristics of the anal fin are thenumber of rays and the extent of the connection with the caudal fin.The number of rays varies from about 24 to 40. The specific varia-tion is seldom more than four?that is, two on each side of theaverage number. The connection with the caudal is, in practicallyall of the species, shghtly greater than the dorsal connection with thecaudal. The anterior anal rays, in some species if not in all, resemblethe anterior dorsal rays in being unsegmented but differ in beingsplit to the tips.Caudal Jin.?The caudal fin is alwaj^s strong and broad. It varieslittle among the various species, either in shape or number of rays,and does not present characters which can be frequently utilized indistinguishing species. The average number of rays forming themain body of the fin is 10 or 12. There appears to be little variationwithin a species. There are in many species short rays on each sideof the base which appear only when the fin is dissected. Theseshort rays appear to be more numerous in such species as Lipariscallyodon, in which the caudal is practically free from the dorsal andanal. There may be as many as 8 or 10 rudimentary rays, whichbring the total number of caudal rays up to 18 or 20. The discrep-ancies between the descriptions of the caudal of a species are probablydue to the short rays which may or may not have escaped observation.In the specific descriptions in this work the number of caudal raysgiven refer to the full-length rays.Pectoral Jin.?The pectoral fin presents important characters.The number of rays is of the most importance. The absence of thepectoral notch separates a few of the species from all the rest. Thelength of the lower pectoral lobe varies and can be utilized to aslight extent in separating species.The number of pectoral rays ranges from about 28 to 41. Thespecific variation is about four. In nearly all of the species thenumber of pectoral rays is less than the number of dorsal rays andgreater than the number of anal rays. In one species, LiparisJucensis, the pectoral has more rays than the dorsal. The pectoral of 48 BULLETIN 150, UNITED STATES NATIONAL MUSEUMLiparis pulchellus, tessellatus, and steineni has an equal or smallernumber of rays than the anal. Liparis major has a smaller number ofpectoral than anal rays and in this character, as well as the simpleteeth, reduced posterior nostril, and black peritoneum, approachesthe condition found in Careproctus.The majority of the species have the pectoral notched. This iscaused by the rays near the middle of the fin being shorter than someof those below. In Liparis owstoni, tanakae, and possibly aniarcticaand steineni these rays are not shorter than those below and the mar-gin of the fin is not notched. Whether or not this is true of thepectoral in the young is unknown. There is some evidence that theyoung of Liparis tanakae have a Ughtly notched pectoral.In Liparis the lower pectoral lobe is never elongate. The lengthvaries among the species and with age. In some species it does notextend beyond the disk and in others reaches nearly to the anal. Thevariation in length of either the upper or lower lobe between relatedspecies is so slight as to be of little taxonomic value. In all thespecies the rays of the lower lobe are somewhat thickened, partly free,and graduated in length.The upper edge of the pectoral fin is at about the same level on thebody in all the species.Dislc.?The size of the disk is one of the important specific char-acters. The diameter of the disk is contained from 1 .5 to 3.2 times inthe head. The proportional size of the disk varies with age, beinglarger in the young individuals.The distance from the tip of the lower jaw to the disk varies with thelength of the head and the size and position of the disk. It may begreater or less than the diameter of the disk. The number of times itis contained in the head varies with the species and, in a few cases,serves to distinguish them.The width of the flap varies slightly, but not enough to be of specificvalue. It may be greater or less than the diameter of the center of thedisk. The disk is round or shghtly oval, but never triangular as insome species of Careproctus. The long diameter of the disk is usuallyparallel with the body. The margin of the disk is indented at thesides anteriorly.Vent.?The position of the vent varies but little among the differ-ent species. It is close to the disk, but usually about midway betweendisk and the origin of the anal fin. The distance between disk andvent varies from 1 .6 to 4.8 in the head. The amount of variation with-in a species may be due to age.Pseudohrancliiae.?The writer investigated the pseudobranchiaesolely for the purpose of demonstrating their presence in the species.It was noted however that there vv^ere usually about five filaments.Pseudobranchiae were seen on the following species : gihhus, cyclopus,agossizii, fucensis, tunicatus, lierschelinus, tanakae, and pulchellus. REVISION OF THE FISH FAMILY LIPARIDAE 49Pyloric coeca.?The number of pyloric coeca varies from about 10to 90. There is considerable variation in the number of coeca in someof the species. The variation in L. callyodon is from 42 to 64 and inL. agassizii from about 60 to 90. The evidence at hand indicates thatwhen the average number of coeca is 40 or 50 the amount of variationwill be 10 or more and possibly as high as 30. In spite of this varia-tion, the number of coeca constitutes an important specific character.When more thoroughly investigated it ^vill doubtless be found thatthe average number of coeca can be used to considerable advantagein designatmg species.The length of the coeca varies considerably, but this character hasnot been sufficiently studied by the writer to warrant any conclusionsas to its value.In this genus the coeca are always on the right side and more easilyexamined through an incision on this side.PricMes.? In 1904 Jordan and Synder proposed a new genus,Trismegistus, based on the presence of "thumb-tack" prickles. Thisis good evidence that these structures had escaped serious considera-tion previous to this time. They had been noted and figured, howeverby Liitken (1886) and Garman (1892). Jordan and Evermann,1898 (p. 2107) describe the color of L. atlanticus as, "with smallscattered light or bluish dots over the body." These dots are thebases of the prickles which they failed to notice. The writer hasdescribed their presence in a number of the genera. The species ofLiparis upon which the writer has noticed prickles are L. atlanticus,mucosus, micraspidopliorus , hristolense, ochotensis, ingens, megacepTialus,rhodosoma, tandkae, and owstoni. Liitken records them in L. major.The significance of these prickles is unknown. In some species, asL. atlanticus, the pricldes appear to be found only on the male andduring the breeding season. They are not solely a male character,however, for the writer has examined them from the bodies of femalespecimens of L. megacepTialus, hristolense, and ocTiotensis. Schmidt,1904 (p. 191), records what he considers to be a specimen of L. owstoni(probably tanalcae) as a female and "covered with thumb-tacklikeplates which make it rough." All the 12 specimens of L. hristolensehave thumb-tack pricldes. Some species, such as L. callyodon,never have prickles. It is possible that in some species prickles arealways present on both sexes, in others occasionally present on bothsexes, m others occasionally, seasonally, or always present on themale. Other possibilities suggest themselves but the above-mentionedcover the field of probability.The prickles are usually or always absent from the lower surfaces ofthe body. They may be confined to a limited area, such as the nape,or distributed over the upper and latral surfaces of the head and body 50 BULLETIN 150, UNITED STATES NATIONAL MUSEUMand onto the fins. It is not known that in any species the pricklesare localized or restricted to a definite region of the body.Coloration.?The species of Liparis are all dull colored. Thecoloration consists of bars, lines, blotches, and mottlings of brown,olive, slate, and white. In the majority of the species the fins arespeckled or faintly barred with brown. The coloration of some of thedeeper-water forms has been slightly modified. The varigated color-ation is retained but in addition, as in L. rhodosoma and L. tessellatus,the body becomes somewhat pinkish and translucent. In L. tessel-latus the peritoneum becomes silvery and in L. major black. InL. owstoni and L. tanalcae the lining of the dermis is reddish.The coloration is not always uniform for a species. Some of thespecies exhibit a wide range of color variation. Certain specimens ofX. pulchellus, dennyi, agassizii, and other species have longitudinal,wavy, or straight lines on the head and body. One of the specimensof L. rhodosoma and several of L. dennyi have irregular oblong pink orwhitish blotches on the head and body. These may represent modi-fied stripes. These peculiar types of coloration have not beenassociated with sex or age. Many of the species exhibit distinctcolor markings by which they can be easily identified. In some of thespecies certain specimens can be identified by their coloration whilethe remaining specimens differently colored have to be identified byother means.The membranes of the mouth, gill cavity, and peritoneum areusually a dull white. The peritoneum is frequently sparsely dottedwith black or brown. In L. dennyi the peritoneum is undotted.In L. tessellatus it is silvery. In L. major it is black or heavily pig-mented with black.Habits.?The tide-pool species remain stationary on the bottom orcling to the rocks and seaweed by means of the disk. When attachedto seaweed they may be carried some distance from the shore. Someof the deeper-water forms, judging from the shape and texture of thebody may move about more than the tide-pool species. It is unlikelythat any of the species swim continuously and at a considerable dis-tance from the bottom.Food.?The food of these species consists mainly of Isopods andCopepods. Small fishes are sometime taken.Summary.?^In order to present a guide to the study of the specificcharacters of these fishes the following outlines are given. Thefirst list suggests the characters which the writer has found to be themost valuable in detecting and separating species. The second listinvolves those characters which are of less value or have not beenadequately studied. KEVISION OF THE FISH FAMILY LIPARIDAE 51List 1Body Shape.Depth.Length.Texture.Head Length.Depth.Width.Nostril Flap.Eye Size.Gill slit In relation to pectoralDorsal Number of rays.Notch.Elevated raj^s.Connection to caudal. Anal Number of rays.Connection to caudal.Pectoral Number of rays.Notch.Length of lower lobe.Disk Size.Snout to disk.Coeca Number.Color Body.Peritoneum.Distribution. List 2Head MaxUlary.Nostril Length of tube.Eye Color.Pores Position of.Rudimentary.Teeth Number of rows.Dorsal Origin of.Segmentation of rays. Caudal Number of rays.Rudimentary rays.Prickles.Color Stripes.Pyloric coeca... Length. DISTRIBUTIONThe genus Liparis is confined to the shallow cold water ot theNorthern and Southern Hemispheres. The majority of the knownspecies are from northern regions. Up to the present time only twospecies are known from Antarctic regions. Many of the species arefound in the tide pools. At least 21 of the 30 species have been takenin less than 10 fathoms. Only 5 species have been taken from depthsbelow 100 fathoms and 3 from below 200 fathoms. The greatestdepth at which a specimen has been taken is 250 fathoms. We mayprovisionally place the lower margin of the bathymetric distributionof the genus at about this level.Limits of distribution.?The tide-pool species extend farthest southon the Pacific coast of America. Species of Liparis mucosus areinfrequently taken in the deep tide pools at Pacific Grove, Calif., orjust north of parallel 36? N, On the Asiatic coast specimens ofLiparis agassizii have been taken at Mizako, north of parallel 39? N.^?The species is not a typical tide-pool species, and the specimensrecorded may have all been taken with seines below the tide pools.The writer did considerable collecting in Hakodate Harbor butobtained no specimens of Liparis from the tide pools. The Japanese 'D Smith and Pope (1906) record specimens of Liparis agassizii obtained from the museum at the FisheryExperiment Station, Shiogama, Matsushima Bay, Japan. 52 BULLETIN 150, UNITED STATES NATIONAL MUSEUM were observed to catch specimens of Liparis agassizii with hook andline in shallow water. The southernmost locality in Asiatic waterfrom which typical tide-pool species of Liparis have been obtained isSimushir Island, at about latitude 45? N. It is very likely thattide-pool species exist on the northern coast of Hokkaido. Thetide-pool species apparently extend 3 or more degrees of latitudefarther south on the American than on the Asiatic side of the Pacific.The deeper-water species have been recorded farthest south inAsiatic waters. On the American coast none of the deeper-waterspecies are recorded as far south as the tide-pools species. Thereverse is true of the Asiatic species. The Albatross obtained specimenof Liparis tessellatus at Station 4867, Japan Sea, 36? 31' N., depth150 fathoms, temperature 33.4? F. This is at about the same latitudeas Pacific Grove and represents the southernmost distribution of thegenus in the Northern Hemisphere. ^^We may account foi the difference in distribution between theAmerican and Asiatic species as due to the difference in temperature.This in turn is governed by the warm Japan current from the southand the cold Arctic current from Bering Sea. The warm currentflowing northward meets the cold current in the latitude of Hako-date, 42? N., and is deflected to the eastward. It is cooled by thetime it reaches the American coast and flows southward at a lowertemperature. The surface temperature of the Japanese waters, dueto the Japan current, is higher than the temperature of the surface ofthe California coast. This may account for the difference in distri-bution of the tide-pool species.The cold Arctic current from Bering Sea cools the northern coastof Hokkaido and makes veritable ice chests of the bottom of theOkhotsk and Japan Seas. The deeper waters of Japan are coolerthan the waters of the same depth and latitude on the Americancoast. The Japanese specimens so far collected are from colderwaters than the American specimens. The Asiatic species inhabitthe same depths as the American species but on account of the differ-ence in temperature are enabled to find a favorable environmentfarther south.The Okhotsk and Japan Seas offer a favorable environment for thedevelopment of Liparids and when further explored will doubtless befound to contain many more species. A temperature record of 29.7?F., the lowest for the Pacific Ocean, was taken in the southern part ofthe Okhotsk Sea at a depth of 64 fathoms. In the Japan Sea the tem-perature at a depth of 150 fathoms is less than 34? F. and at 200 fath-oms less than 33? F. Species of Liparis can descend to this depth and "Schmidt (1904) records a large specimen ol" Liparis owstoni" from the Nagasaki fish market. Wherethis specimen was collected is unknown. REVISION OF THE FISH FAMILY LIPARIDAE 53find a very favorable environment. In fact a temperature of 40? F. isfavorable to most of the species.The favorable environment offered by the cold waters of this regionmay account for the gigantic size attained by some of the species. Thefour largest species of the genus, ranging in length from 430 to 485 mm.are found in Japanese waters. One of these comes from a temperatureof 32.7? F., another from 33.9? F. The other two specimens lackrecords. The largest species from the American coast is L. greeni andreaches a length of 244 mm.The tide-pool species of the Atlantic coast of America extend as farsouth as the coast of Massachusetts or possibly of Connecticut. Thisis about the same latitude as the north of Hokkaido, their southern-most distribution on the Asiatic coast. In both regions a northward-moving warm current raises the temperature of the surface waters andproduces an unfavorable environment for these Arctic fishes. TheGulf Stream is deflected to the northeast off Nova Scotia and warmsthe shores of Europe to such an extent that these fishes are restrictedstill further to the northward. Only 5 species are known from theNorth Atlantic, in contrast to the 14 American species and 10 Asiaticspecies in the North Pacific.The tide-pool species are, in a general way, limited in the southerndistribution by the summer isotherm of 60? F. The deeper-waterforms are not, or but rarely, found in a bottom temperature of 50? F.and are usually found in less than 45? F. This temperature, with thefactors governing the vertical distribution, marks the southern limitfor the distribution of the deep-water species of the genus.Two species of Liparis are known from the Southern Hemisphere.Both are from the cold waters of South America. Liparis antarctia isfrom Eden Harbor and Liparis steineni from South Georgia. Theyinhabit an environment similar to that of their northern relatives.The distribution of the genus in the Northern and Southern Hemi-spheres bears the same relation to the marine regions of Gill (1875).In the north the genus is distributed from near the middle of thePararctalia northward and in the south from near the middle of theNotalia southward.Asiatic and American species.?The Asiatic and American speciesare distinct. It can not be said with certainty that any species isfound on both sides of the deep Kamchatka Channel. Three speciesare doubtfully recorded from both sides of this channel. Two imma-ture and poorly preserved specimens from Plover Bay, Siberia, arelisted as L. callyodon. Two immature specimens from Petropavlovskare recorded as L. cyclopus. L. gihhus is recorded from both shores ofBering Sea. Of these species L. callyodon is distinctly a tide-pool spe-cies. The other two descend to deeper water. It appears that there91668?30 5 54 BULLETIN 150, UNITED STATES NATIONAL MUSEUMhas been little or no migration across the Kamchatka Channel inrecent times. The deep water acts as a barrier to the adults, and theeggs are probably endemic or fastened to the seaweed or rocks.It is interesting to note that the majority of the American speciesinhabit the tide pools and that a large majority of the Asiatic speciescome from greater depths. This may be explained as only apparentbecause of the lack of a thorough exploration of the Siberian coast, or,that failing, to the longer stretch of favorable coast line reaching fromcentral California to the Aleutian Islands and northern Alaska.Regions.?We may conveniently divide the species of this genus intofour groups which inhabit different geographical regions. Theseregions may be termed Asiatic, Pacific American, North Atlantic, andSouth American. The species of each region, with a few possibleexceptions, are distinct. The Asiatic region contains 10, the PacificAmerican 14, the North Atlantic 5, and the South American 2 species.The Pacific American region can be divided into two sub regions ? (1) northern California to the Aleutian Islands, (2) coast of Alaskanorth of the peninsula to the Arctic Ocean. The first of these regionsconforms with the first three regions given by Evermann and Golds-borough (1907-222), and the second in a general way with the fourthand fifth. Certain species extend nearly throughout the first subregionor practically 20? of latitude. L. mucosus inhabits the shore line fromPacific Grove to Kodiak Island; L. pulchellus from San Francisco Bayto Bristol Bay; L. callyodon, P. heringianus, and P. greeni from thenorthern coast of Washington to Bering Island. None of the Asiaticor North Atlantic soecies are known to range through so many degreesof latitude.The Alaska Peninsula apparently serves as a barrier to the deeper-water species. None of these are known from both sides of thisbarrier.The deeper-water species are not known to have a greater rangethan the tide-pool species. Liparis fucensis and L. dennyi have thegreatest range and extent from the coast of Washington to the Gulfof Alaska.Distribution of closely related species.?Our knowledge of the rela-tionships of the species of Liparis is too incomplete to draw any finalconclusions concerning the distribution of closely related species, butwhat evidence there is is of interest to anyone testing the law statedby Jordan as, "given any species in any region, the nearest relatedspecies is not likely to be found in the same nor in a remote region,but in a neighboring district separated from the first by a barrier ofsome sort." The genus Liparis can be divided into groups of closelyrelated species and species which are isolated in their relationships.The groups of related species and their distribution are as follows : EEVISION OF THE FISH FAMILY LIPARIDAE 55 callyodon?Washington to Bering Island.curilensis?Simushir Island, Kurile Group.mucosus?Pacific Grove to Kodiak Island.micraspidophorus?Agattu and Bering Island.greeni?Washington to Bering Island. Not in tide pools.beringianus?Washington to Bering Island. Tide pools.simushirae?Simushir Island. Tide pools.dennyi?Washington to Gulf of Alaska.gihha?Bering Sea.cyclostigma?Bering Sea.ochotensis?Okhotsk Sea.rhodosoma?Okhotsk Sea.ingens?Japan Sea.brisiolense?Bristol Bay.cephalus?Bristol Bay.tunicatus?Labrador to Greenland.herschelinus?Arctic coast of Alaska.agassizii?northern Japan.Certain of these species may be combined or divided when betterknown. The distribution of some will be extended and of othersrestricted. When this is done it will doubtless be seen that theprmciple stated by Jordan applies to the distribution of the species ofLijparis.Bathymetrical distribution.?The species of Liparis, as has been said,are most common in the tide pools and do not descend below 250fathoms. The greatest vertical distribution for any one species is175 or possibly 212 fathoms for Liparis fucensis. Liparis dennyiis known to range through about 100 fathoms. The majority of thespecies range through less than 50 fathoms. It may be found that,on account of the greater difference in temperature betw^een thesurface and bottom waters on the Japan coast, the species do nothave as great vertical distribution as on the American coast.Center of dispersal.?When we attempt to apply the criteria usedby Adams (1902) and Ruthven (1908) in finding the center of dis-tribution of a group we find that, of those applicable, certain onespoint to northern Japan and the remainder to Bering Sea. Atpresent we shall not attempt to distinguish between these two regionsbut shall designate Bering Sea as the center of dispersal for the genus.At the time when there was a land connection between Asia andAmerica (see J. P. Smith 1907) and the Arctic Current was preventedfrom entering Bering Sea the waters of the North Pacific were tem-pered by the warm Japan Current more than at present. Thespecies of this genus must have retreated to Bering Sea and may havebeen exterminated there. When the land bridge subsided and thecold Arctic Current again chilled the North Pacific the species ofLiparis remaining in Bering Sea or others from the Arctic Oceanmigrated southward along the Asiatic and American shore lines. Thesubsidence of the land bridge across Bering Strait was at a sufficiently 56 BULLETIN 150, UNITED STATES NATIONAL MUSEUM remote period to allow specific differentiation to take place. Therehas been no recent intermigration between Asia and America, for thespecies of the two regions are distinct.The distribution of the few North Atlantic species can readily beaccounted for by assuming that they migrated from Bering Seaalong the Arctic coast of America to the North Atlantic. Certainspecies crossed over to northern Europe and others migrated downthe northeast coast of America to Connecticut. The species of thetwo shores are different. Liparis liparis, a European species, hasbeen recorded from the New England coast, but the writer has beenunable to verify this record. Liparis atlanticus extends from Con-necticut to Labrador, Liparis tunicata from Labrador to Greenland,Liparis major from Greenland to the Kara Sea. The latter speciesmay be circumpolar. Liparis liparis and Liparis montagui areEuropean species.Assuming that the genus originated in the North Pacific the mostreasonable way by which we can account for the presence of species atthe southern end of South America is by migration along the Americancoast line at some remote period when the tropical waters were suf-ficiently cooled to offer a favorable environment. At the presenttime a depth of 150 to 200 fathoms along the west coast of tropicalAmerica affords a sufficiently cool temperature for these fishes. Butthe genus is not represented there and is seems unlikely that the mi-gration to the Antarctic has taken place under present conditions.During the Ice Age, conditions were more favorable than now andit was probably during this period that species of Liparis made theirway across the Equator. KEY TO SPECIES OF LIPARIS "A'. Peritoneum pale or silvery, usually with sparsely scattered brown or blackdots; pectoral in most of the species with more rays than the anal.B^ Caudal free from the dorsal or connected for not more than one-fifth itslength; pectoral rays more numerous than the anal rays.Ci. Dorsal notched; gill opening either above the pectoral or extending|i.' I down in front of not more than 6 rays.D'. Anal less than 30; dorsal not more than 35; coeca less than 150.E'. Gill slit above the pectoral, sometimes appearing to extend downin front of the upper ray.F'. Disk 2 or less in the head.G'. Pyloric coeca less than 50 (15-37); anterior dorsal rays some-times elevated. American.HI. Pectoral 26-29; no white bar at the base of the caudal.Atlantic montagui, atlanticus.H2. Pectoral 30-33; a white bar across the base of the caudal.Pacific rutteri.G". Pyloric coeca more than 50 (65-70) ; anterior dorsal rays neverelevated. Asiatic curilensis.F^. Disk more than 2 in the head callyodon.grebnitzkii. la See supplementary key for North Atlantic species. REVISION OF THE FISH FAMILY LIPARIDAE 57 E^. Gill slit extending down in front of 3 to 5 pectoral rays.Ji. Disk usually more than 2 in the head in the adult; greatestdepth of the body usually behind the origin of the seconddorsal; color brownish or striped mucosus.J2. Disk 2 or less in the head; color slaty or olive gray; bodydeepest at the origin of the dorsal micraspidophorus.C*. Dorsal unnotched, of if notched the gill slit extends down in front ofmore than 6 pectoral rays.M'. Pectoral with not more than 35 rays, less than thenumber of dorsal rays; gill slit in front of 10 or lesspectoral rays.N^ Pectoral 29-32; gill slit extending down in frontof 3-5 pectoral rays cyclopus.N2, Pectoral 34; gill slit in front of 9 pectoral raysfrenatus.M2. Pectoral 3S-43, with more rays than the dorsal;gill slit in front of 13-16 rays fucensis.B*. Caudel connected to the dorsal for more than one-fifth and less than three-fourths its length, or if connected for three-fourths its length thepectoral is unnotched; the number of pectoral raj^s greater than thenumber of anal rays.O^ Pectoral notched in adult and young.P^ Gill slit extending down in front of not morethan 10 pectoral rays.Q*. Gill slit extending down in front of 6 orless pectoral rays; pyloric coeca less than55. American.R'. Dorsal 38-40; anal 32-35; prickles ap-parently always present- _bristolense.R2. Dorsal 41-44; prickles never present.S'. Pyloric coeca 36-48; gill opening infront of 3-6 pectroal rays_tunicatus.S^. Pyloric coeca 18-20 (?); gill openingin front of 0-3J^ pectoral raysherschelinus.Q2. Gill slit extending down in front of_6-9pectoral rays; coeca more^ than r 55;Asiatic agassizii.P^ Gill slit extending down in front of morethan 10 pectoral rays; coeca less than 55.T^. Head blunt; width of head usuallygreater than the depth ; head andbody resembling the tide-poolforms; snout to disk usually lessthan 7.5 in the|^length.^ Fleshnot pinkish.U^ Dorsal 37-40; no prickles; dorsalfrequently notched.'*; Coast ofWashington and southeastAlaska dennyi.U^. Dorsal 40-44; no prickles. Ber-ing Sea cyclostigma, gibbus. 58 BULLETIN 150, UNITED STATES NATIONAL MUSEUM U*. Dorsal 45; prickles frequentlypresent. Asiatic.\K Inner teeth stout, stronglytrilobed; jaws nearly equal;body stout. Saghalin, Ok-hotsh Sea ochotensis.V^ Inner teeth slender, weaklytrilobed; upper jaw project-ing and exposing nearly all ofthe teeth in front. Southernpart of the Japan Sea_ingens.T2. Head pointed; width of head aboutequal to the depth; head andbody not resembling the tide-pool forms, having more theappearance of deep-water forms;snout to disk not over 7 in thelength. Flesh pinkish. OkhotshSea rhodosoma.T*. Head exceptionally heavy and blunt;snout not projecting, broadlyrounded; width of head aboutequal to the depth; head andbody not resembling the tide-pool forms; snout to disk 7.8-8.4in the length of the body withoutthe caudal. Flesh not pinkish.Bering Sea megacephalus.O*. Pectoral unnotched in the adult and possiblyin the young.W. Dorsal 28 antarctica.W^. Dorsal 40 or more.X'. Pectoral 32 steineni.X2. Pectoral about 40.Y'. Head flattened in thenasal region, snoutlow and projectingtanakae.Y^. Head convex in thenasal region, not de-pressed; snout deepand short, not pro-jecting owstoni.B*. Caudal connected to the dorsal for three-fourths or more of its length;pectoral notched; the number of pectoral rays equal to or less than thenumber of anal rays. Z^ Gill opening abovethe pectoral or infront of 1 to 4rays pulchellus.Z^. Gill opening in frontof about 16 pectoralrays tessellatus. REVISION OF THE FISH FAMILY LIPARIDAE 59 A^. Peritoneum black or densely pigmented; pectoral with less rays than theanal major.SUPPLEMENTAET KET TO SPEOES OF LIPAEI3 FBOM THE NORTH ATLANTIC(Ai) Peritoneum silvery or white, with few or no black dots.(B') Anal less than 30; dorsal usually notched.(C) Pectoral less than 30.(D^) Anterior dorsal rays sometimes elongate; prickles sometimes present;dorsal more than 30. Coeca usuallj' more than 20. Americanatlanticus.(D2) Anterior dorsal rays never elongate; prickles never present; dorsal30 or less. European montagui.(C^) Pectoral more than 30; coeca less than 20 liparis.(B^) Anal more than 30 (34-37); dorsal never notched; dorsal 41-44; anal34-37 tunicatus.(A^) Peritoneum usually black, sometimes heavily pigmented with black. D. 48;A. 40; P. 34. _._ major.LIPARIS ATLANTICUS Jordan and ErermannLiparis montagui Garman, 1892, p. 47 (part, pi. 7, figs. 6-20, not of Donovan)?Liparis liparis Garman, 1892, p. 59 (part, pi. 7, figs. 1-5, 21-22).Neoliparis atlanticus Jordan and Evermann, 1898, No. 47, p. 2107.Type.?Male, No. 47215, U.S.N.M. Godboiit, Quebec, 1885.Length 109 mm.Distribution.?Specimens examined, 36 in number, are from thecoast of Connecticut northward to Quebec. The records of thespecimens examined are incomplete and do not indicate the verticalrange of the species. It has been recorded from the tide pools downto 50 fathoms.Relationships.?L. atlanticus has frequently been confused withmontagui Donovan of Europe. Jordan and Evermann consider thatthe former species differ from the latter in the smaller head and themore distinct and sometimes elevated anterior dorsal fin. Boulenger(Jordan and Evermann, 1898. p. 2107) says that the anterior dorsalfin of L. montagui is always distinct but never elevated. I havebeen unable to make a satisfactory comparison of these two speciesbut I am inclined to believe that they will be found to be distinctmorphologically as well as geographically. In comparing L. atlan-ticus with the Pacific species we find that in most respects it bearsa close resemblance to L. rutteri. The resemblance is more strikingwhen we consider that these are the only two species in which theanterior dorsal fin is sometimes elevated. The only differences oftaxonomic value that we find between L. atlanticus and L. rutteriare in the coloration and the number of rays in the pectoral fin. Thedisk may be found to be slightly larger in the latter species. L.rutteri apparently never has thumb-tack prickles. These are not 60 BULLETIN 150, UNITED STATES NATIONAL MUSEUM always present in L. atlanticus and so fail to serve as a constantmeans by which the species can be distinguished.Description of the type.?D. 34; A. 27; P. 28; pyloric coeca 30.Depth in length, 4.4; head, 4.5. Eye in head, 5.6; disk, 1.8.Body heavy anteriorly, moderately elongate. Head short, rathersmall; cheeks swollen; profile slightly depressed over the eyes. Mouthsmall, terminal; maxillary scarcely reaching beneath the front of thtpupil. Teeth in broad bands, strongly trilobed; eight or nine obliquerows in the half of each jaw. Snout projecting. Anterior nostril ina prominent tube; posterior nostril not prominent, with a fingerlikeprojection in front. Eye small. Gill sUt, small, either above thepectoral or extending down in front of the upper ray. Pores on thehead and body normal; suprabranchial pores separated by the diam-eter of the eye. "Thumb-tack" prickles scattered over the head andbody.Dorsal with the five anterior spines elongate; second spine longest,a little longer than the head; tips of the elongated spines free; seconddorsal rays increasing rapidly in length. Caudal truncate, con-nected to the anal for less than one-fifth its length; last dorsal andanal rays shortened, forming a notch. Pectoral notched; the lowerlobe of six thickened, partly free rays, reaching midway betweendisk and vent. Disk large, less than 2 in the head. Vent separatedfrom the disk, by 0.6 the diameter of the disk.Color.?Body reddish brown; the bases of the "thumb-tack"prickles appear as scattered pale dots; peritoneum pale, faintlydotted with brown.Synopsis.?Dorsal, 32-34; anal, 25-27; pectoral, 26-28; pyloriccoeca, 19-37, typically more than 20. Eye 5.6-6.5 in the head;disk 1.8. Gill slit above the pectoral or extending down in frontof one ray. Spinous dorsal usually distinct, sometimes hardlyevident; the spines sometimes elevated. Caudal typically con-nected to the dorsal for less then one-fifth its length. Pricklespresent or absent. Color reddish brown to light olive brown; thecaudal faintly barred; in some specimens bars extend from thedorsal and anal onto the body. A small species reachiug a lengthof slightly more than 100 mm.Remarlcs.?One of the Salem specimens has "thumb-tack" pricklesscattered over the body as in the type. Five of the specimensexamined have the spinous dorsal elevated. In our specimens theprickles and elevated spines are associated only with the males.Both of these characters are said to be typical of the male duringthe breeding season. The dorsal notch sometimes hardly evident.As some specimens of L. liparis have a shallow dorsal notch we cannot depend on this character to separate the two species. REVISION OF THE FISH FAMILY LIPARIDAE 61L. atlaniicus may be the only described species of Liparis on ourAtlantic coast. (See L. tunicatus.) Putman and later Garman (forother records see Kendall 1908 and Tracy 1909), record L. liparisfrom this region. Specimen No. 12958; Garman 1892 (pi. 7, figs. 1 , 2,3, 4), appears to be distinct from L. liparis. The number of coeca isgreater and the disk is larger, both agreeing with L. atlaniicus. Thespecimen is in such condition that the other characters can not beascertained though the anal appears to be as figured, and the colora-tion striped. Someimmature specimens from Block Island and Noankfound in pectens seem to differ from L. atlaniicus in the increasedconnection between the dorsal and the caudal. They probably, withNo. 12958, represent some other species, possibly an undescribed one.These specimens have probably been separated from L. atlaniicusand placed with L. liparis because of the absent or obscure dorsalnotch and the extent of the connection between dorsal and caudal fin.Specimens of L. atlaniicus one-half inch in length have been takenon the 31st of July. Putman (1873, 339), records that the spawningseason is in March. UPARIS MONTAGUI (Donovan)Cyclopierus montagui Donovan, 1805, pi. 68.Liparis montagui Gunther, 1861, p. 161.Distribution.?Coasts of northern Europe, recorded from the northof France northward. No specimens examined.Relationship.?A discussion of the possible differences between L.montagui and L. atlaniicus are given in the description of the latter.UPAKIS RUTTERI (Gilbert and Snyder)Neoliparis ruiteri Gilbert and Snyder, 1898, in Jordon and Evermann, 1898, p.2108. ? Evermann and Goldsborough, 1907, p. 331, fig. 99.Types 3, No. 5701, S. U. Z. M., Uyak Bay, Kodiak Island, Alaska;Rutter, 1903. Length 65 mm.Distribution.?Known from the southeast coast of Alaska and theAleutian Islands; tide pools to 16 fathoms. Twenty-four specimensexamined.Relationsliip . ?L. ruiteri closely resembles L. atlaniicus. The onlydifferences between the two species are found in the coloration and thenumber of pectoral rays. L. ruiteri appears to have a slightly largerdisk than L. atlaniicus and never has prickles. L. ruiteri can bedistinguished from all the Pacific species by the large disk combinedwith the coloration.Synopsis.?Dorsal 31-32; anal 23-26; pectoral 30-33; pyloric coeca23-31 . Disk 1 .4-1 .7 in head. Gill slit above the pectoral. Spinousdorsal usually distinct but sometimes hardly evident; the spines some-times elevated. Caudal typically connected for less than one-fifthits length to the dorsal. Disk large, less than 2 in head. Prickles 62 BULLETIN 150, UNITED STATES NATIONAL MUSEUM apparently never present. Color variable, usually distinct in thepresence of a white bar across the base of the caudal and the posteriormargin of the dorsal and anal fins. A small species, the largest speci-men being 63 mm. in length.Remarks.?The elevated dorsal is not always present in the male,for one of our specimens, a male from Agattu, has a very low dorsal.None of the female specimens exammed have an elevated dorsal.Four of the specimens exhibit the striped color pattern; of these 1 isfrom Karluk, 2 from Loring, and 1 from Station 4205, AdmiraltyInlet. Apparently no morphological differences can be correlatedwith the striped color pattern. The specimen from Station 4205 isthe only one taken in the dredge. It differs from the others in thepaler coloration. In addition to the stripes?the color of this speci-men is as follows: Light ashy brown, darker posteriorly; margin ofdorsal and anal dusky; caudal finely crossbarred. The white bar atthe base of the caudal apparently is distinctive of the species thoughabsent in one specimen and faint in others.LIPARIS CURILENSIS (Gilbert and Burke)Cyclogaster curilensis Gilbert and Burke, 19125, p. 353.r?/2?6.? Female, No. 73326, U.S.N.M. Simushir Island, Japan.Albatross, 1906. Length 110 mm.Distribution.?Known only from the type locality. Thirty-twospecimens examined.RelationsM'p.?In general appearance and in coloration this speciesresembles L. callyodon. It can be distinguished from the latter bythe much larger ventral disk. In the size of the disk L. curilensisagrees with L. rutteri. It is readily distinguished from the latter bythe coloration and number of pj'loric coeca. The coloration doesnot differ materially from that of L. callyodon though usually the crossbars are much more pronounced. None of our specimens exhibitedthe spotted coloration sometimes shown by L. callyodon.Synopsis.?Dorsal 34-35; anal 26-27; pectoral 29-30; pyloric co-eca in one specimen 65. Disk large, 1.5-2 in head. Gill slit abovethe pectoral or extending down in front of the upper pectoral ray.Dorsal notched. Dorsal and anal barely or not at all connected withthe caudal. Coloration as in L. callyodon.LIPARIS CALLYODON (Pallas)Cyclopterus callyodon Pallas, 1811, p. 75.Liparis mucosus Garman, 1892, p. 52 (part, not of Avers).Nsoliparis callyodon Jordan and Evermann, 1898, p. 2110, fig. 760. ? Evermanmand GoLDSBOROUGH, 1907, p. 332. fig. 100.Distribution.?L. callyodon is common in the tide pools along thesoutheastern coast of Alaska and the Aleutian Islands. The collec- REVISION OF THE FISH FAMILY LIPARIDAE 63tions examined contained specimens from points as far south as PortTownsend, Wash., and as far north as St. Paul Island, Bering Sea.The author doubtfully records three small and poorly preservedspecimens form Petropavlosk and Plover Bay, Siberia. This speciesis not found in tide pools in its southern range.Relationsliijp . ?L. callyodon appears to be most closely related toL. curilensisoi the Kurile Islands. It appears to differ from the latterspecies mainly in the smaller size of the ventral disk. In other re-spects the two species appear similar.Synopsis.?Dorsal 33-35; anal 26-27; pectoral 29-31; pyloric coeca42-66. Disk moderate, 2.2-3 in the head. Gill sht either above thepectoral or extending down in front of the upper pectoral ray. Dorsalfin distinctly notched. Dorsal and anal connection with the caudalnot extending beyond the skin covered base of the latter. Pricldesnever present. Coloration variable; for a description of the severaltypes see remarks. A small sized species reaching a length of about100 mm.Remarks.?The gill slit in this species varies somewhat. It mayappear to be above the pectoral or extending down in front of theupper ray. The following types of coloration are exhibited by speci-mens in the United States National Museum.(1) Purphsh above, varying to olive-brown on the sides, palerbelow; fins finely crossbarred or not.(2) Light olive to an olive brown; fins finely barred.(3) Tawny to olive brown, paler below, punctulate with dark dots;fins faintly barred.(4) Olive brown mottled -with ash; the sides sometimes withround dark spots about the size of the pupil.In addition to the variation in the coloration there is considerablevariation in shape and depth of body and head in specimens fromdifferent localities. The variation in the number of fin rays and sizeof disk falls within narrow limits and the author is unable to satis-factorily designate local races. With the acquisition of more datait may be possible to divide the species into local races based on colorand form. Other tide-pool species of Bering Sea, such as Pallasinacarhata, exhibit local peculiarities. The islands of Bering Sea offera splendid opportunity to study the effect of deep channels as bar-riers to the ready dispersal of the tide-pool species and if they actas barriers, the eft'ect of isolation in the formation of species. ;jSome of our specimens from Bering Island, collected by Greb-nitzki, agree fairl}' well with the description of L. grelnitzkii Schmidt.L. grebnitzkii appears to be a color form of L. callyodon. 64 BULLETIN 150, TJNITED STATES NATIONAL MUSEUMLIPARTS GREBNITZKII (Schmidt)Neoliparis grebnitzkii Schmidt, 1904a, p. 165, pi. 6, fig. 4.Type.?Female, No. 8855, museum of St. Petersburg, Bering Island.Grebnitzki. Length 83 mm.Distribution.?Known only from Bering Island. No specimensexamined.Relationsliip.?Schmidt (1904a) considers this species to be mostclosely related to L. rutteri. I can not agree with him and believe L.grebnitzkii to be most closely related to L. caUyodon and possiblyidentical with it. I am unable to detect any important differencesbetween the two species.Synopsis.?Taken from the original description. Dorsal 32; anal27; pectoral 29; caudal 10; pyloric coeca about 30; eye 6.6; snout3.1; disk 2.2. Gill slit above the pectoral. Color tawny; head andback brown; fins grayish; pectoral yellowish, without spots.LIPARIS MUCOSUS (Ayers)Liparis mucosus Ayres, 1855, vol. 1, p. 24; 1873, vol. 1, p. 22. ? Garman, 1892,p. 52, pi. 5, figs. 1-5 (part). ? Jordan and Starks, 1895, ser. 2, vol. 5, p. 832,pi. 95. ? Jordan and Evermann, 1898, p. 2111, fig. 761.Neoliparis florae Jordan and Starks, 1895, ser. 2, vol. 5, p. 830, pi. 96. ? Jordanand Evermann, 1898, p. 2111, fig. 762.Type.?Destroyed. San Francisco.Distribution.?Northwest coast of America from Pacific Grove,Calif., to Kodiak Island, Alaska. A tide-pool and shallow-waterspecies. Twelve specimens examined.Relationships.'?L. mucosus is most closely related to a twin species, L.micrasjndopliorus, from the Aleutian Islands. It can be distinguishedfrom the latter species by the coloration and the shape of the head andbody. In addition to the above differences the disk is slightly smallerin L. mucosus. The coloration of the body and peritoneum, the shapeof the head and body, and the size of the gill slit and disk readilydistinguish this species from all other species with an equal number offin rays inhabiting the same regions.Synopsis.?Dorsal 31-33; anal 25-27; pectoral 30-33 ; pyloric coeca48?-70. Disk 2.1-2.4, in the head. Gill slit extending down in frontof from 1 to 6 pectoral rays, usually in front of more than 3 rays.Dorsal notch distinct, the anterior dorsal fin sharply pointed. Dorsaland anal connections with the caudal very short, little, or not at all,extending beyond the skin-covered base of the latter. Prickles presentor absent. Head and anterior part of body appearing depressed;greatest depth of body usually at the end of the first third of the seconddorsal fin. Color dark brown to light yellowish brown, paler below,sometimes with longitudinal wavy stripes; fins colored as the body orfaintly barred; first dorsal sometimes a dark blotch. A small-sized species, reachmg a length of 125 mm. REVISION OF THE FISH FAMILY LIPARIDAE 65RemarJcs.?In this species the head is usually thrown upward,giving the body the appearance of being depressed in front of the firstdorsal fin. In a normal specimen the depth increases gradually tothe end of the first third of second dorsal ; sometimes the depth is asgreat at the origin of the first dorsal, but it never exceeds that underthe second dorsal; dorsal outline of body convex under middle of sec-ond dorsal. The gill slit sometimes extends down in front of onlyone pectoral ray and then resembles the gill slit in some specimens ofL. collyodon.The coloration varies from a drab brown to a light yellowishbrown; it lacks the slaty cast of L. callyodon and in this respect morenearly resembles L. rutteri. The fins may be colored as the body orfaintly barred; the tip of the first dorsal sometimes with a darkblotch. A small speciftien 55 mm. in length has light wavy lines on Figure 2.?Liparis mucosus. Prickles from specimen from Pacific Grove, Calif.the body and head as in L. pulcJiellus. The disk of this specimen islarge and is contained 1.7 in the head. One of the Pacific Grovespecimens has the body and the top of the head and cheeks coveredwith "thumbtack" prickles; these extend onto the fins. Thesepricldes are not a sex character, unless seasonal, for one of the PacificGrove specimens, a male, does not possess them. The pricldes inthis species are club shaped?that is, heavier near the tip than the base.The anterior nostril tube is sometimes very short and the rim dark.The specimens from Karluk and Neah Bay are very similar to thetype of '' Neoliparis florae" and easily recognized as belonging to thesame species. The Pacific Grove specimens are larger and presentslight peculiarities which are not of specific value.Jordon and Starks (1895, p. 832) describe what they consider to bea specimen of " Neoliparis mucosus" as follows: Head 4; depth 4.3.Dorsal 6-26; anal 26; pectoral 29; caudal 12. Eye 7; disk 1.5.Body holding its depth well past the middle. Nostrils not endingin tubes. Gill slit not extending below the upper edge of the pec- 66 BULLETIN 150, UNITED STATES NATIONAL MUSEUMtoral. Color olive-brown, light below; indistinctly mottled; dorsaland anal darker at the margin; pectoral uniform dark brown; cau-dal light, with indistinct cross lines; lips dark.This specimen has a very large disk and is here doubtfully referredto L. mucosus. The gill slit is described as being above the pectoraland the nostril tube absent. In some of our specimens of L. mucosus .y'^ "^^ .?=^ /i~Figure 3.?Liparis micraspidophorus. Tvpe. A typical tide pool species, with notched dorsalAND FREE CAUDAL FINthe gill sht extends down in front of only one pectoral ray and thenostrU tube is very short. It seems likely that the Jordan andStarks specimen belongs with this species.The following note is from the origmal description by Ayers : Head4 in the total length; eye nearly 6 in the head. Dorsal and anal ter-minate at the base of the caudal. Dorsal notched. Gill opening asin L. pulcJiellus. (The gill slit in L. pulcTiellus is described by Ayres Figure 4.?Liparis micraspidophorus. Prickles from type as being above the pectoral, but it extends down in front of two orthree rays.) Anterior nostril tubular. Disk as in L. pulchellus.(The disk in L. pulchellus is contained more than two times in thehead.) Color plain greenish olive, lighter below.There is little in this description to indicate that the species de-scribed by Ayres as Liparis mucosus is different from the speciesdescribed by Jordan and Starks as " Neoliparis florae." I believe REVISION OF THE FISH FAMILY LIPARIDAE 67the species are identical and here reduce florae to the synonomy ofmucosas. LIPARIS MICRASPIDOPHORUS (Gilbert and Burke)Cyclogaster micraspidophorus Gilbert and Burke, 1912a, p. 71.Type?Male, No. 74379, U.S.N.M., Nikolski, Bering Island. Al-hatross, 1906. Length 73 mm.Distribution.?Known only from Agattu and Bering Islands. Atide-pool species. Four specimens examined.RelationsJiips.? L. micraspidopliorus appears to be most closely re-lated to L. mucosus. It differs from the latter species in the shape ofthe body, greater depth, more swollen cheeks, slightly larger disk, inthe shape of the prickles, and in the coloration. In other respectsthe two species appear to be similar.Synopsis.?Dorsal 31-32; anal 25-27; pectoral 30-32; pyloric coeca55-65; disk 1.7-2 in the head. Gill slit extending down in front of4-5 pectoral rays. Dorsal notch deep. The connection between thedorsal and the caudal equal to the skin-covered base of the latter.Prickles present or absent. Body slate coloredabove, paler below; vertical fins indistinctlyspeclded and barred; upper half of pectoral <^speclded. In life uniform brownish red. A [^small sized species, none of our specimens reach-ing a length of 100 mm. ^ figure 5.-liparis mc-RemarJcs.?The swollen occiput is not so no- rashdophorus. teethticeable in the cotype, and the color varies toan olive gray somewhat resembling the typical coloration of L.callyodon. The lower lobe of the pectoral appears shorter than inL. mucosus. LIPABIS LIPARIS (Linnaeua)Cyclopterus liparis Linnaeus, 1776, p. 414.Distrihution.?Coast of northern Europe. Recorded from thenortheastern coast of America and Japan. The possibility of thisspecies being on the American coast is discussed in the account ofL. atlanticus. The Japanese record, (Franz, 1910), is very doubtfuland will doubtless be found to be that of another species. (Seeremarks.)RelationsMp . ?The relationship of L. liparis with the other knownspecies is obscure. In the small number of pyloric coeca it agreeswith L. antarctica but the two species differ widely in other charac-ters. L. liparis is readily distinguished from the other species of theNorth Atlantic by the number of fm rays, absence of a distinct dor-sal notch, the distinct connection between the dorsal and anal andthe caudal fin, and the small number of pyloric coeca.Description.? Description of a number of specimens from theCheshire coast of Ens-jand. 68 BULLETIN 150, UNITED STATES NATIONAL MUSEUMDorsal 34-35; anal 29; pectoral 32-33; disk a little more than 2 inthe head.Body heavy, short. Head heavy, blunt, depth equal to or greaterthan the width. Cheeks nearly vertical; occiput high. Mouthbroad, with little lateral cleft. Teeth strongly tri-lobed, arranged inabout 12 oblique rows in the half of each jaw. Snout bluntlyrounded; lower jaw included. Eye small, silvery. Gill slit small,extending down in front of one to three pectoral rays. Anteriornostril in a short tube. Coeca 10-13, in a single row about the gut.Dorsal usually with a shallow notch setting off the first five rays.Dissection of the fin will show that in most of the specimens thefifth and sixth rays are slightly shortened. Sometimes they appearto be of the same length as the preceding rays. In none of our speci-mens do the rays increase regularly in length. Caudal truncate, of12 rays, connected for a little more than, or about one-fifth its lengthto the dorsal. Pectoral notched; the lower lobe reaching nearly to thevent. Disk more than 2 in head. Vent separated from the disk bythe diameter of the latter.Color.?Slaty to brownish; the fins speclded to mottled or barredwith brown. Peritoneum silvery, with few or no dots. A few of thespecimens have pale stripes along the top and sides of the head.Synopsis.?Dorsal 34-35; anal 29; pectoral 32-33; pyloric coeca10-13, typically less than 20. Disk more than 2 in the head. Gillslit extending down in front of from one to three pectoral rays. Dorsalfin appearing to be unnotched, but upon dissection the fifth and sixthrays will usually be found to be slightly shortened. Caudal connectedfor a little more than, or about, one-fifth its length to the dorsal fin.Prickles absent. Color variable; many color varieties described;our specimens slaty to brownish; the fins speckled, mottled, or barredwith brown; some of the specimens with pale stripes on head andbody.Remarks.?Many color varieties of this species have been described.This species and in fact all the species of Europe need a thoroughrevision based upon a large series of specimens. It is likely thatother species have been confused with L. liparis.Ehrenbaum (1905) records the spawning season from November toFebruary, Smith (1897) records this species as common in winterand full of spawn in December and January. (In the discussion ofL. atlanticus I have intimated that the specimens from the NewEngland coast recorded as L. liparis will doubtless be found to bethat of another species.)Franz (1910) records two specimens of Liparis liparis from Japan.Franz has the following to say about these specimens : Dorsal 32; anal 26. Nach vergleichung meiner Exemplare mit solchen vonHelzoland kann ich keine spezifischen Untersheide den letzteren gegeniiber REVISION OF THE FISH FAMILY LIPARIDAE 69finder. Flossen strahlenzahlen wie die des vorliegenden Exemplares komeenauch bei europaischen, wenn auch als unteres Extrem, vor. Die Formen desFlossen, die Dimensiones der Korpers stimmen genau mit denen der hiesigenTiere iiberein, und auch die Farbung fallt innerhalb des sehr grossen variations-bereishes der europaischen Form. Die Zeichnung des einen Exemplares ist indes Figur angegeben. Grundton purpurrot.Ein Exemplar von Fubuura, 7.5 cm. lang, coll. Haberer; ein Zweiter vonMisake, 8.6 cm. lang, coll. Doblein.The figure (pi. 9, fig. 76) presents the following: Gill slit not distinctbut appearing to be above the pectoral fin. The connection betweenthe dorsal and caudal fins short, shorter than in L. agassizii. Bodywith longitudinal light stripes.The combination of fin rays, unnotched dorsal, short connectionbetween dorsal and caudal, and the coloration serves to distingidshthese specimens from all the other Pacific species. The number offin rays as given by Franz is less than typical for L. liparis; and this,with the distribution, leads me to believe that Franz has erred inreferring his specimens to L. liparis.LIPARIS CYCLOPUS GuntherLiparis cyclopus Gunther, 1861, p. 163. ? Jordan and Starks, 1895, p. 834,pi. 97. ? Gilbert, 1896, p. 446. ? Evermann and Goldsborough, 1907, p.332, pi. 18.Type.?Esquimault Harbor, Vancouver Island. Length 45 lines.Distribution.?The specimens examined are from Petropavlovsk (seeremarks), Bering Island; Puget Sound; Port Angeles, Wash.; PointMoller, Alaska: Albatross Station 3230, Bering Sea. If the Petro-pavlovsk specimens prove to be identical with the American species^L. cyclopus will be one of the few shallow-water species existing onboth sides of the deep Kamchatkan Channel. L. cyclopus is notcommon in the shallow tide pools but prefers slightly deeper water.Relationships.?L. cyclopus does not appear to be closely related toany known species. It, with L. frenatus and L. fucensis, presentstransition characters between the tide pool and more highly modifiedspecies of the genus. In these three species the dorsal fin is unnotched,or the notch but faintly indicated, the number of fin rays is inter-mediate, the gill slit extends down in front of the pectoral fin, and theconnection between the dorsal and caudal is very short. L. cyclopusis distinguished from L. frenatus and L.jucensis by the smaller numberof pectoral rays and the narrow gill slit. The shape of the body inL. cyclopus resembles that of L. mucosus, a species extending fartherto the south. The two species differ in the presence or absence of thedorsal notch and the number of fin rays, and do not appear to beclosely related.Description.?The following description is of a specimen fromPoint Moller, Alaska, No. 24007, U.S.N.M.91668?30 6 70 BULLETIN 150, UNITED STATES NATIONAL MUSEUMDorsal 35; anal 29; pectoral 29; pyloric coeca 39. Depth 5 inlength without caudal; head 5.25. Eye 6.2 in head; disk 2.5.Body shaped somewhat as in L. mucosus; the depth retained tobeneath the second dorsal. Head low, wider than deep; occiputnot swollen; profile rising gradually, depressed over the eyes. Mouthwide, terminal; maxillary reaching vertical in front of eye. Snoutlow, jaws equal. Teeth slender, recurved; the lateral lobes small.Anterior nostril in a prominent tube; posterior nostril with a lowrim. Gill slit extending down in front of 5 pectoral rays; 3.8 in head.Dorsal unnotched, originbehind middle of pectoral ; caudal of 10 rays.Dorsal and anal connection to the caudal very short, extending little, if atall, beyond the skin-covered base of the latter. Pectoral notched, thelower lobe of six rays, reaching a little more than halfway between diskand vent. Disk moderate. Vent about midway between disk and anal.Olivaceous above, paler below; body and fins speckled with olivebrown. Peritoneum pale, with sparsely scattered black dots.Synopsis.?Dorsal 35 ; anal 29 ; pectoral 29 ; py-loric coeca 39. Disk 2.5in the head. Gill slitextending down in frontof five pectoral rays.Dorsal fin unnotched,the connection with thecaudal not extendingFigure 6.?Lifaris cyclopus. Teeth from a specimen from bevond the skin-COVeredPuGET Sound , i- ^ i i ^ ^ -ntbase 01 the latter. JNoprickles. Body somewhat depressed anteriorly, the greatest depthretained to beneath the second dorsal. Color olivaceous above,paler below; body and fins speckled with light olive brown. Asmall-sized species.RemarJcs.?L. cyclopus is a flabby appearing species with a broad,low head. It is easily recognized at sight in the adult stage. Thespecimens from Petropavlovsk are immature and may represent anundescribed species. The specimens from Kamchatka may have asmaller gill slit and the snout more evenly rounded than the Americanspecimens of the species.The following details are from the larger Kamchatka specimen.Head 4 in length without caudal; depth 4.7. Dorsal 37; anal 30;pectoral 32. Eye 5 in head; snout 3; disk 2; gill slit extending downin front of 3 pectoral rays.LIPARIS FRENATUS (Gilber. and Burke)Cyclogastcr frenatus Gilbert and Burke, 19126, p. 356.Type.?Female, No. 73329, U.S.N.M. Albatross Station 4809,Japan Sea; depth 207 fathoms. REVISION OF THE FISH FAMILY LIPARIDAE 71Distribution.?Japan Sea, only the type known, apparently a deep-water species.Relationship.?L. frenatus is not closely related to any knownspecies. It can be distinguished from L. cyclopus, which it resemblesin many characters, by the increased number of fin rays, the largegill slit, and the narrow head.Synopsis.?Dorsal 37; anal 31; pecto-ral 34; pyloric coeca 21; disk 2.5 in thehead. Gill slit extending down in frontof nine pectoral raj^s. Dorsal fin un-notched, connected to the caudal finfor less than one-fifth of the length ofthe latter.RemarJcs.?L. frenatus has an unusu- -n ., ,?' . . FlQUEE 7.?LiPARIS FRENATUS. TeETHally narrow head and in this respect from typeresembles the tj'pical forms of Care-proctus. The reddish color also suggests an approach to the conditionfound in Careprodus. UPARIS FUCENSIS GilbertLiparis fucensis Gilbert, 1896, p. 447. ? Jordan and Evermann, 1898, p. 2119.Liparis callyodon Garman, 1892, p. 54, pi. 6, figs. 1-5.NeoUparis fissuratus Starks, 1896, p. 560. ? Jordan and Evermann, 1898,p. 2113. Figure 8. ? Liparis fucensis. Teeth from typeType.?Femsile, No. 48600, U.S.N.M., Straits of Juan de Fuca,Albatross Station 3451; depth 106 fathoms. Length 110 mm.Distribution.?Coast of Washington and southeastern Alaska,Albatross Stations 2865, 3451, 3452, 3461, 4302, and from Port Ludlow,Wash.; depth, shallow water dovm to 212 fathoms, at present notknown to inhabit the tide pools. Thirteen specimens examined.Relationship.?L. fucensis apparently is not closely related to anyknown species. In the character of gill opening and increasednumber of pectoral rays it represents a transition stage between 72 BULLETIN 150, UNITED STATES NATIONAL MUSEUMthe lower and higher members of the genus. It still retains inreduced form the primitive notch in the dorsal fin. The short con-nection between the dorsal and caudal fins also is primitive. Itdiffers from all other species in the family in having more rays in thepectoral than in the dorsal fin. The teeth also appear to have dis-tinctive characteristics, rather short and stout; the lateral lobesnot in the same plane as the central lobe; the lobes flattened andthin near the margins.Synopsis.?Dorsal 33-35; anal 27-29; pectoral 38-43; pyloriccoeca 26?-55. Disk 2.5-2.8 in the head. Gill slit extending downin front of 13-16 pectoral rays. Dorsal fin unnotched or with a shal-low notch which becomes evident upon dissection. More raysin the pectoral than in the dorsal fin. Dorsal connection wdth thecaudal very short, little if any extending beyond the skin-coveredbase of the latter. No prickles. Color olive brown to brownish.A small-sized species reaching a length of 110 mm.Remar-ks.?The dorsal fin in this species is usually if not alwaysnotched. In all the specimens examined, with the exception of amutilated individual, the notch in the dorsal fin could be detected bydissection. In the mutilated indivdual the anterior dorsal raj^s werebroken and some of them absent, so that the presence or absence of anotch could not be ascertained. In the type or Neoliparis JissuratusStarks, the first seven dorsal rays are unsegmented. ,N . fissuratus , ashas been pointed out by Gilbert and Thompson, 1905, is identicalwith L. Jucensis.A notch is usually present between the caudal and anal fins. Thelower lobe of the pectoral in some specimens reaches halfway betweendisk and vent, while in others it reaches to the vent. The colorationvaries from that described for the tj^pe to a plain olive brown dustedwith dark dots. LIPARIS BMSTOLENSE (Burke)Cyclogaster brisiolense Burke, 1912a, p. 568.Tijpe.?Female, No. 53790, U.S.N.M.; vicinity of Bristol Bay?Bering Sea; Albatross Station 3514.Distribution.?Southeastern Bering Sea, Albatross Stations 3247^3301, 3514, 3518; depth 17 to 36 fathoms. Twelve specimensexamined.RelationsMp.?The distribution, shape of body and the presence ofprickles suggest a close relationship between Liparis bristolense andLiparis megacepJialus . The latter species appears to be distinct in.having a larger number of dorsal rays, a larger gill slit, and thedorsal notch absent. Liparis bristolense differs from Liparis tunicatus-and Liparis Jierschelinus in the smaller number of dorsal rays and thecoloration. It differs from Liparis agassizii in the smaller number ofdorsal rays and pyloric coeca, the coloration, and the smaller gill slit. REVISION OF THE FISH FAMILY LIPARIDAE 73Synopsis.?Dorsal 38-40; anal 30-35; pectoral 33-37; pyloric coeca16-22. Disk 1.7-2.2 in head, typically less than 2. Dorsal fin usuallywith a shallow notch. Connection between dorsal and caudal usuallytwo-fifths the length of the caudal. Color variable.LIPARIS MEGACEPHALUS (Burke)Cyclogaster megacephalus Burke, 1912a, p. 569.Type.?Female, No. 53791, U.S.N.M.; Southeastern Bering Sea,Albatross Station 3519; depth 37 fathoms. Length 145 mm.Distribution.?SoutheasteTB. Bering Sea, Albatross Stations 3518,3519, 3520; depth 36 to 38 fathoms. Three specimens examined.Relationship.?Liparis megacephalus resembles Liparis major in theheavy head and body but is otherwise distinct. For comparison withLiparis bristolense see description of latter species. Figure 9.?Liparis bristolense. Prickles and teeth from cotypeSynopsis.?Dorsal 43-44; anal 36; pectoral 36-38; pyloric coeca "29-31. Disk 2 in head. Gill sht extending down in front of 12 pectoralrays. Dorsal unnotched. Dorsal connected to the caudal for two-fifths caudal. Prickles present.LIPARIS TUNICATUS ReinhardtCyclopterus liparis minor Fabricitjs, 1780, p. 135.Liparis tunicata Reinhardt, 1836, p. cxi.Liparis arctica Gill, 1864, p. 191.Liparis tunicatus Garman, 1892, p. 65.Distribution.?Apparently common in Arctic waters about Green-land and Labrador. One of our specimens, which is doubtfully placedwith this species, is from Nauset Beacon, Mass. A tide-pool andshallow-water species. Thirty-one specimens examined. 74 BULLETIN 150, UNITED STATES NATIONAL MUSEUMRelationship.?L. tunicatus is closely related to L. Tierschelinus. Thetypes of L. Tierschelinus are small and pressed out of shape, so that anaccurate description is impossible. The author has not comparedspecimens of the same size of the two species, but it appears that inL. tunicatus the gill slit extends farther down in front of the pectoralfin and the number of pyloric coeca is greater than in L. herschelinus . Description.?Description of several specimens from Labrador,collected by L. M. Turner.Dorsal 41-44; anal 35-37; pectoral 35-37; pyloric coeca 38-48.Depth 4-4.7, in length without caudal; head 3.6-3.8. Eye 5.3-6.3 inthe head; disk 2.2-2.4.Body robust, deepest at front of first dorsal. Head heavy andbroad; occiput slightly swollen; cheeks swollen; profile depressed overeyes. Mouth broad, nearly terminal; maxillary reaching vertical frompupil. Teeth short, stout, arranged in about 11 oblique rows in thehalf of each jaw; the lateral lobes appearing nearly as prominent as thecentral lobe. Eye small. Gill slit ex-tending down in front of three to sixpectoral rays in adults, in young speci-mens apparently not extending as fardown in front of the pectoral fin.Pricldes absent.Dorsal fin unnotched; the originover the end of the first third ofpectoral. Caudal slightly rounded, of10 rays, connected for about onep. , ,n T,?.?,= .??. rr?^^? quarter its length to the dorsal; dorsalFigure 10.?Lipaeis tunicatus. Teeth ^ ? _ _ 'FROM SPECIMEN No. 341G8, u.s.N.M. and anal connection with the caudal,FROM LABRADOR gradual, the last rays not shortened andforming a notch. Pectoral fin notched; the lower lobe of seven rays,reaching halfway between disk and vent or beyond. Disk ratherlarge, with a broad flap. Distance from disk to vent, 2.5 to 3.4 in head.Color: The specimens examined exhibit two types of coloration:(1) Represented by No. 34163, upper parts of head and body brown-ish; a pair of pale stripes extending from tip of snout through nostrUsand along top of head, uniting at the origin of the dorsal, then deflecteddownward along the sides of the body to the base of the caudal; asecond extending backward from the snout below the eye and acrossthe gill slit, hence it is deflected downward on the sides to near thebase of the anal and then backward to the base of the caudal; peri-toneum silvery, black dotted; outer half of vertical fin darker; (2)body gvsiy, slaty brown; skin finely dusted with brown dots, palerbelow; dorsal and anal with the base pale and the outer half darkbrown; caudal uniform brown or speckled and crossbarred withbrown; the base unusually pale; pectoral pale or speckled with brown. REVISION OF THE FISH FAMILY LIPARIDAE 75Synopsis.?Dorsal 41-43; anal 34-35; pectoral 35-38; pyloriccoeca 36-48. Disk 2.2-2.4. Gill slit extending down in front of fromthree to five pectoral rays. Dorsal fin uimotched. Dorsal fin con-nected to the basal fourth of the caudal fin. No prickles. A mediumsized species, our largest specimen 121 mm. in length.Remar-ks.?Specimen No. 31533, U.S.N.M., from Nauset Beacon,Mass., is doubtfully placed with the species. The gill slit extendsdown in front of the pectoral fin as in L. tunicatus, but the fin rayscan not be counted. It is possible that this specimen belongs to thesame species as the specimen doubtfully recorded as L. atlanticus.It is to be regretted that these specimens are in such condition that wecan not ascertain their relationships.LIPARIS HERSCHELINUS ScofieldLiparis herschelinus Scofield, 1899, p. 504, pi. 7. ? Evermann and Golds-borough, 1907, p. 333, pi. 18.Type.?^o. 5601, S. U. Z. M.; Herschel Island, Arctic Ocean. Thetype bottle contains 17 small specimens; length 37-64 mm.Distribution.?Ejiown onlj" from the tide pools of Herschel Island,Seventeen specimens examined.Relationships.?L. TierscJielinus is very closely related to L. tunicatus.The former species appears to have a smaller number of pyloric coecaand a smaller gill slit than the latter. These differences may not befound to hold, as the type specimens of L. herschelinus are small and inpoor condition and do not permit of satisfactory comparison to bomade \vT.th L. tunicatus. It is possible that the two species areidentical.Description of the types.?Dorsal 44; anal 35; pectoral 37; p34oriccoeca 18. Depth 4.4 in length without caudal; head 3.5. Eye 4.3 inhead; disk 2.2.Head and bod}^ distorted and not satisfactory for a description.Teeth stout, in about 9 oblique rows in the half of each jaw. Snoutslightly projecting. Anterior nostril in a short tube; posterior nostrilwithout a tube. Eye small. Gill slit in some specimens appearing tobe above the pectoral fin and in others extending down in front of 1 to2 or 3 pectoral rays; the distorted condition of the bodies probablyaccounts for some of this variation. Pyloric coeca in one specimen 18,Dorsal fin unnotched; about 10 rays unsegmented. Caudal slightlyrounded, connected for about one-fourth its length to the dorsal; nonotch between dorsal and caudal or anal and caudal; anal connectionto the caudal a little greater than dorsal connection. Pectoral fin witha shallow notch; the lower lobe short, reaching a little past disk.Disk moderate. Vent nearer anal fin than disk, separated from diskby diameter of disk. 76 BULLETIN 150, UNITED STATES NATIONAL MUSEUMSkin transparent, dusted above and on sides with dark punctula-tions, pale or straw colored beneath; anal crossed by faint blotches;outer half of the dorsal fin darkened posteriorly; caudal with a darkbar near the base and a faint narrow bar near the tip of the fin. Peri-toneum pale, with sparsely scattered black dots.Synopsis.?Dorsal 44; anal 35; pectoral 37; pyloric coeca 18. Disk2.2 in head. Gill slit above the pectoral or extending down in front of1 to 2 or 3 pectoral rays. Dorsal fin unnotched. Dorsal fin connectedto the basal fourth of the caudal. No prickles. A small-sized species,our specimens reaching a length of 64 mm.LIPARIS AGASSlZn (Putnam)Ldparis agassizii Putnam, 1874, p. 339. ? Garman, 1892, p. 62, pi. i-iii, (part).Type.?Sagahlin, Channel of Tartary. Pierce and Smith. Length250 mm.Distribution.?L. agassizii is the common species of northern Japan.The limits of its distribution to the northward are unknown but itcertainly is not found in Bering Sea as has been recorded. It extendsdown on the east coast of Hondo as far south as Same and Miyako.Smith and Pope (1906), record specimens obtained at the Shiogamafisheries station, Matsushima Bay; these specimens may have beencollected elsewhere and may also belong to some other species, pos-sibly L. tanalcae. The specimens examined were obtained at Otaru,Aomori, Hakodate, Tomakomai, Same, Miyako, and Albatross Station4808, Tsugaru Straits. A shallow-water species but apparently nottypical of the tide-pool fauna, at least in its southern ranges extendingdown to a depth of 47 fathoms. L. agassizii is a common market fishin northern Japan, being taken by the fishermen with hook and lineand with the seine. Thirty-two specimens examined.RelationsMp.?In many respects L. agassizii resembles L. tunicatusand L. herschelinus. Besides being isolated geographically it can bedistinguished from these two species by the larger gill opening, theincreased number of pyloric coeca, and the coloration.Description of No. 49827, U.8.N.M.?Dorsal 43; anal 34; pectoral36; pyloric coeca 67. Depth 4.8 in length without caudal; head 3.5.Eye 7 in head; disk 2.4.Body short and heavy anteriorly, tapering gradually to the caudalfin; flesh flabby. Head broad and heavy; occiput slightly swollen;width of head greater than depth of head; profile depressed over theeyes. Mouth broad; maxillary reaching vertical from posterior halfof eye. Teeth rather small and slender, arranged in about 20 obliquerows in the half of each jaw, the lateral lobes well developed. Snoutbroadly rounded; upper jaw slightly projecting. Anterior nostril in ashort tube; posterior nostril without tube. Eye small, the lower half REVISION OF THE FISH FAMILY LIPARIDAE 77 silvery. Gill slit extending down in front of nine pectoral rays. Noprickles.Origin of dorsal fin over gill flap; the rays increasing gradually inlength. Caudal rounded; connected for nearly one-half its lengthto the dorsal fin, connected a little more than one-half its length to theanal fin; a shallow notch between the dorsal and the caudal. Pectoralnotched; lower lobe of eight rays, reaching to nearly halfway betweendisk and vent.Several types of coloration are exhibited by the specimens at hand : (1) Skin translucent, with a purplish or brownish cast; fins dusky; eyeblack. (2) Upper parts brownish; a series of white spots along thesides, these bordered with dark brown in front and behind; finsspeckled and barred with brown; tips of rays whitish; eyes silvery.(3) Body brownish, blotched with crossbars on body and fins; shortstripes radiating out from eye. (4) Brownish with purplish stripesextending from snout to base of caudal; eye silvery ; peritoneum silverywith scattered black dots.Synopsis.?Dorsal 42-43; anal 33-35; pectoral 34-39; plyoric coeca64-90. Disk 2.1-2.7 in the head. Gill slit extending down in front offive to nine pectoral rays. Dorsal fin unnotched. Dorsal fin con-nected to the basal half of the caudal. No pricldes. Color variable.A medium to large sized species, our largest specimen 161 mm. inlength.Remarks.?L. agassizii exhibits considerable variation in color andgeneral appearance, but there is little difl&culty in separating thisspecies from all others from the same region. The gill slit varies con-siderably, extending down in front of from five to nine pectoral rays,the normal condition being in front of seven or eight rays. A shallownotch is usually present between the dorsal and the caudal fin.LIPARIS DENNYI Jordan and StarkaLiparis dennyi Jordan and Staeks, 1895, p. 835, pi. 98.^ ? Jordan ana Evermann,1898, p. 2124, fig. 766.Type.?Mole, No. 3703, S. U. Z. M. ; Admiralty Inlet, Puget Sound;Young Naturalists Society of Seattle.Distribution.?Coasts of Washington and northward to the Gulf ofAlaska; shallow waters down to 123 fathoms, rarely if ever found inthe tide pools. One hundred specimens examined.Relationships.?L. dennyi closely resembles L. gihhus of BeringSea. It can be distinguished from the latter species by the smallernumber of dorsal rays and sometimes by the presence of the dorsalnotch. The adults of L. dennyi appear to have a less projectingsnout and a deeper head than in L. gihhus. At present the regionsinhabited by the two species appear to join but not overlap, theAlaska Peninsula serving as a barrier between the two. 78 BULLETIN 150, UNITED STATES NATIONAL MUSEUMSynopsis.?Dorsal 37-40; anal 31-34; pectoral 37-39; pyloriccoeca 19-31. Posterior nostril apparently always with a fingerlikeprojection in front. Disk 2.2-2.4 in head. Gill slit extending downin front of from 10 to 15 pectoral rays. Dorsal fin notched or theanterior rays nearly equal in length. Dorsal fin connected to thebasal two-fifths of the caudal. No prickles. Coloration variable.A medium-sized species, reaching a length of 198 mm.Remarks.?The posterior nostril in this species apparently alwayshas a fingerlike projection in front. The jaws are equal or nearly so;usually a few of the upper teeth are exposed when the jaws are closed.In the young the maxUlary extends to beneath the middle of the pupUor beyond. Sometimes rudimentary pores, appearing as dark spotswith white centers, extend backward from the suprabranchial pores.Numbers of specimens from East Sound differ from the typicalspecimen of this species in the length of the lower pectoral lobe andthe coloration. These specimens may represent an undescribedspecies though we can not at present decide this point. For thepurposes of comparing these specimens with the typical specimensNo. 4009, S. U. Z. M., we present the following:In the East Sound specimens the lower lobe of the pectoral reachesthe vent or beyond; in nearly all the specimens, No. 4009, S. U. Z. M.,it does not reach the vent. These two groups of specimens differ alsoin the coloration. Of the specimens, No. 4009, one resembles thetype in coloration; the remainder are plain olive brown or withnarrow longitudinal white lines on head and body; the peritoneumsilvery, undotted. The East Sound specimens present "Four differenttypes of coloration: (1) Nearly plain dark bro\vn with obscure duskymottlings, the pectoral fin crossbarred with black and white; chinspeckled, margin of median fins nearly black. (2) The entire upperparts, including the dorsal fin, marked ^^ith paralled wavy lightstreaks with darker margins, the intervals between the streaks duskyolive, pectoral more obscurely barred. (3) Entire upper parts darkolive, thickly covered with small white dots less than the diameter ofthe pupU. (4) Lighter olive, marked with few whitish or silverywhite narrow streaks Qr bars, which are black margined. One seriesof these cross the dorsal fin, another the anal, both continued a variabledistance on the body, sometimes meeting. Other streaks and spotsoccupying the top and sides of the head and may be symmetricallydisposed on the two sides of the same individual, although not agree-ing in different specimens." Caudal faintly or broadly barred, peri-toneum silvery, dotted.In this species the dorsal fin is notched in fully half the specimens.In specimens No. 60200, U.S.N.M., consisting of young individuals20 to 28 mm. in length, the dorsal notch is some times evident. Someof these specimens, in which the notch was not evident, were dissected REVISION OF THE FISH FAMILY LIPARIDAE 79and the sixth and seventh rays found to be shorter than the precedingrays. Two of these specimens showed pale stripes extending back-ward from the snout above and below the eye to the caudal. Theremainder of these specimens were speckled with brown and the caudalbroadly barred. Fully half the specimens No. 21314 S.U.Z.M.,have the dorsal notch evident without dissection.LIPARIS GIBBUS BeanLiparis gibbus Bean, 18816, p. 148.Liparis agassizii, Jordan and Evermann, 1898, p. 2721, fig. 765 (part) .^Ever-MANN and GoLDSBOROUGH, 1907, p. 333, fig. 101, (part).Type.?Msxle, No. 24047, U.S.N.M.; St. Paul Island, Bermg Sea.Length 68 mm.Distribution.?St. Paul Island, Bering Sea, and coast of Siberia.Four specimens examined.Relationship.?In the region extending from Washington north-ward through Bering Sea and southward through the Okhotsk andJapan Seas is a series of six closely related species of Liparis. Ofthese L. dennyi is found from the coast of Washington to the Gulf ofAlaska, L. gihhus and L. cyclostigma in Bering Sea, L. ochotensis andL. rJiodosoma in the Okhotsk Sea, and L. ingens in the Japan Sea.The two Bering Sea species are not readily separated. I haverecognized the two species as distinct. However, the slight differencesbetween the types may be due to age. The three Japanese speciesappear to have a larger number of dorsal and anal rays and prickles.Synopsis.?Dorsal 42-44; anal 35-36; pectoral 37-38; pyloric coeca45. Posterior nostril apparently without a fingerlike projection infront. Upper jaw projecting. Eye moderate, 4 to 5.5 in head.Disk 2-2.3 in head. GiU slit extending down in front of 14 pectoralrays. Dorsal unnotched. Dorsal connected to caudal for nearlyone-half the length of the latter. No prickles.UPARIS CYCLOSTIGMA GilbertLiparis cyclostigma Gilbert, 1896, p. 446, (part).Cyclogaster cyclostigma Gilbert and Burke, 1912a, p. 73, fig. 18.Type.?Male, No. 48621, U.S.N.M., vicinity of Bristol Bay,Bering Sea, Albatross Station 3252; depth 295 fathoms.Distribution.?Bering Sea, Albatross Stations, 3252, 4777, 4779,4789, 4795, 4796 ; depth 43 to 295 fathoms. Six specimens examined.Relationships.?L. cyclostigma closely resembles L. gibbus. (Seedescription of latter species.)Synopsis.?Dorsal 40-44; anal 33-35; pectoral 39-43. Posteriornostril without projection. Disk 2.3 in head. Gill slit extendingdown in front of 12 to 14 pectoral rays. Dorsal connected to caudalfor nearly one-half the length of the latter. No prickles. A largespecies reaching a length of 357 mm. 80 BULLETIN 150, UNITED STATES NATIONAL MUSEUMLIPARIS OCHOTENSIS SchmidtLiparis ochotensis Schmidt, 1904o, p. 163.Cydogaster ochotensis Gilbert and Burke, 19126, p. 359.T?/^es.?Male, No. 12963 and 12964, Museum of the ImperialAcademy of Sciences, St. Petersburg. Saghalin Island.Distribution.?Shallow waters about Saghalin Island, OkhotskSea, our specimens from Cape Terpenis and Anuva Bay, Saghalin;depth 21-43 fathoms. Six specimens examined.Relationships.?L. ochotensis appears to be most closely related toL. ingens. It differs from the latter species in the characters of theteeth, equal jaws, and stouter body. It is readily distinguished fromL. agassizii by the larger gill slit and smaller number of pyloric coeca.See descriptions of L. gibhus, and L. rhodosoma for comparisons withthese species.Description of type.?No. 12963, male, length 195 mm. Dorsal45, anal 36; pectoral 39; pyloric coeca 25. Depth 3.5 in lengthwithout caudal; head 3.3. Disk 2.6 in head.Body rather deep, dorsal outline sloping gradually. Head broad,depth greater than mdth;profile gradual, slightlydepressed over the eyes;occiput not swollen.Mouth broad; maxillaryreaching pupil. Teethtrilobed, recurved; thelateral lobes not so prom-inent on the larger teeth;in broad bands, arrangedin about 17 oblique rows in the half of each jaw. Snout short, de-pressed; jaws nearly equal. Anterior nostril in a tube; posteriornostril with a prominent rim. Eye small, 6.1 in head. Gill slit ex-tending down in front of 18 pectoral rays, equal to snout. A fewminute prickles on head, these consisting of a small round base anda short thick conical spine.Dorsal fm high, the rays increasing in length to middle of fin orbeyond, at this point 1.5 times the length of the snout. Anal resem-bling dorsal, the rays not quite so long. Caudal slightly rounded,connected for nearly one-half its length to the dorsal. Pectoralbroad; the lower lobe of six rays, reacliing three-fourths the distancefrom disk to vent. Disk large, with a broad flap. Vent separatedfrom disk by nearly diameter of disk.Color grayish; narrow, paired, dusky stripes on head, body, andbase of dorsal; these sometimes united anteriorly to form singlestripes; margin of dorsal and anal mottled and dusky; caudal mottledand barred with dusky marldngs; margin of pectoral dusky; peri-toneum pale, with scattered dots. Figure 11. ? Liparis ochotensis. Prickles from type KEVISION OF THE FISH FAMILY LIPARIDAE 81Synopsis.?Dorsal 45; anal 36-38; pectoral 39-42; pyloric coeca23-38; posterior nostril without a projection in front. Jaws aboutequal. Teeth not so slender or recurved as in L. ingens, arrangedin 14-25 oblique rows in the half of each jaw, the number in upper andlower jaws about equal. Disk 2.5-2.6 in head. Gill slit extendingdown in front of from 12 to 18 pectoral rays. Dorsal fin unnotched.Dorsal connected to near the middle of the caudal. Prickles some-times present. A large-sized species, reaching a length of 465 mm.LIPARIS INGENS (Gilbert and Burke)Cyclogaster ingens Gilbert and Burke 19126, p. 360.Type.?Msde, No. 73330, U.S.N.M.; Albatross Station 4863,southern part of the Japan Sea; depth 250 fathoms.Relationships.?L. ingens appears to be most closely related to L.ochotensis. It can be distinguished by the teeth, the projecting upper Figure 12.?Liparis ingens. Teeth from typejaw, and the slenderer body. The coloration of the type is distinctbut this character is a variable one among the species of Liparis.Synopsis.?Dorsal 45; anal 37; pectoral 42; pyloric cocea 37.Posterior nostril without projection in front. Upper jaw projecting,exposing nearly all the upper band of teeth; about 20 oblique rows inthe half of the lower and 30 in the half of the upper jaw; the innerteeth slender and strongly recurved. Gill slit extending down infront of 16 pectoral rays. Dorsal fin unnotched. Prickles present,A large-sized species, reaching a length of 485 mm.LIPARIS RHODOSOMA, new speciesCyclogaster ochotensis Gilbert and Burke, 19126, p. 359 (part).Type.?Male, No. 22271, S.U.Z.M. Okhotsk Sea, off Saghalin,Albatross Station No. 5023; depth 75 fathoms. Length 200 mm.Distribution.?Southwestern part of the Okhotsk Sea off the coastof Saghalin, Albatross Stations 5016, 5017, 5020, 5021, and 5023;depth 64 to 75 fathoms. Five specimens examined. 82 BULLETIN 150, UNITED STATES NATIONAL MUSEUMRelationships.?L. rhodosoma resembles L. ochotensis in most re-spects, differing in the shape of the head and the coloration. Thesetwo species inhabit the same geographical region, but up to thepresent time all the specimens listed as L. rhodosoma are from greaterdepths than any of the specimens of L. ochotensis. It seems unlikelythat L. rhodosoma represents a color phase of L. ochotensis.Description oj type.?Dorsal 45; anal 37; pectoral 41; pyloric coeca37; depth of body 3.7; head 3.5. Eye 7 in head; disk 2.7. Figure 13.?Lipaeis ingens. Prickles from typeBody more elongate and tapering more gradually than is typical ofthe tide-pool species. Head distinctly pointed; profile rising grad-ually from snout to occiput, a slight depression over the eyes; sidesof head sloping outward slightly; the cheeks little swollen. Mouthmoderate; maxillary reaching vertical from middle of eye.* Teethwith well-developed lateral lobes; the smaller teeth blunt but the|obes indicated; the larger teeth rather slender and recurved, withateral lobes well developed; arranged in 17 oblique rows in the half REVISION OF THE FISH FAMILY LIPARIDAE 83^ of the lower jaw, 20 in the half of the upper jaw. Snout overlapping^the mouth for one-half the eye, not so pronounced in the youngindividual; upper teeth exposed when the jaws are closed. Eye small.Gill slit extending down in front of 15 pectoral rays. "Thumb-tackprickles" on top of head and body, extending back slightly beyondorigin of dorsal; the spine slightly conical. Pyloric coeca 37.Dorsal normal; the origin just back of the base of the pectoral; 12or more rays unsegmented. Two anal rays unsegmented. Caudalslightly rounded; the dorsal and anal connection abrupt; a slightnotch between the anal and the caudal, none between the dorsal andthe caudal; anal connected to the caudal for more than one-half thelength of the caudal; the dorsal connection less. Pectoral notched,the lower lobe reaching nearly to vent. Disk large, with a broad flap.Color pinkish, slightly gelatinous, with a few faint roundish or ir-regular brown spots over body and fins; pectoral with faint bars;caudal with 2 bars near the base, 3 farther out, and 1 near the tip.Peritoneum white, undotted.Synopsis.?Dorsal 45-46; anal 37; pectoral 40-42; pjdoric coeca33-44. Posterior nostril without projection in front. Upper jawdistinctly overlapping the lower jaw, upper teeth exposed when thejaws are closed. Disk 2.1-2.7 in head. Gill slit extending down infront of 14 to 15 pectoral rays. Dorsal fin unnotchfed. Dorsal con-nected to the caudal for about one-half of the length of the latter,Pricldes sometimes present. A medium to large species, reaching alength of 200 mm.Remarks.?In the young individual the snout does not project so far,thus bearing a closer resemblance to the typical Liparis. Thesmallest individual has 13 oblique rows of teeth in the half of eachjaw. Color variable; the larger individuals more pinkish and gela-tinous than the smaller ones. An individual 105 mm. long has thehead and body covered with double dark-brown stripes which divergeposteriorly; dorsal and anal variegated with dusky markings andfaint bars; caudal with two dark bars and several blotches. Theadult specimens exhibit two types of coloration? (1) that describedfor the type, sometimes the markings more pronounced; (2) palepinkish blotches with dark margins on top of head and sides of body,as follows: 4 on the dorsal and anal, extending on the sides of thebody and nearly uniting at the median line; 1 extending back fromgill opening, those on the head obscure; 1 extending from the occi-put down onto the sides of the head; 1 running back from the lipthrough nostril and uniting across the top of the head with its mate.Prickles are absent in all but the type. 84 BULLETIN 150, UNITED STATES NATIONAL MUSEUMTable of measurements of Liparis rhodosomaLocality?stations 5016 5021Length 144 191Depth of body 4.3 4.0Snout to disk 7.0 6.6Disk to anal 5.9 6.2Depth, head 4.7 4.5Width, head 4.7 5.0Head 3.6 3.6Eye in head 6.0 6.5Snout 2.3 2.6Internostril 3. 6 4. 2Gill slit 2.5 2.3Disk 2.1 2.5Disk to vent 2.3 2.6Pyloric coeca 44 39Dorsal 45 45Anal 37 37Pectoral 7-35 8-34Pectoral rays behind gill slit 14 15Sex Female. Female. 5023 REVISION OF THE FISH FAMILY LIPARIDAE 85UPARIS ANTARCTICA PutnamLiparis antardica Putnam, 1873, p. 339.?Garman, 1892, p. 61, pi. 6, figs. 6-10.Enantioliparis antardica Gill, 1891, p. 365.Type.?No. 12972, M. C. Z. Eden Harbor. Hassler Expedition,1872. Length 32 mm.Distribution.?Known only from the type locality, Eden Harbor,South America.Relationships.?L. antardica is widely divergent from any of thenorthern species. The distinctive characters are the small numberof fin rays and coeca and the unnotched pectoral fin. L. antardicaand L. liparis have the least pyloric coeca of any of the species ofLiparis. In the character of the unnotched pectoral fin L. antardicaagrees with L. tanalcae and L. owstoni of Japan.Description oj type.?Dorsal 28; pyloric coeca 10; depth 3.6 inlength without caudal; head 3.2. Eye 4 in the head; disk 1.9. Gar-man (1892) records the anal as with 24 and the pectoral with 30 rays.Body short and heavy as in L. callyodon. Head deep ; depth about equal to the width; occiput swollen as in theyoung of L. dennyi; cheeks slightly swollen. Mouthterminal with little lateral cleft; maxillary reaching be-neath the front of the pupil. Teeth stout, stronglytrilobed, in broad bands, arranged in oblique rows.Snout abrupt, not projecting, lower jaw included. An-terior nostril in a short tube; posterior nostril withouttube. Eye black. Gill slit above the pectoral. Nopricldes. Pyloric coeca in a single row, about 10. figure h-Dorsal apparently without a notch. Caudal broad, arctic a'.connected for nearly one-half its length to the anal. toothfromTYPEPectoral somewhat mutilated, apparently unnotched,possiblj" as figured. Disk large, round; snout to disk 8 in the lengthwithout caudal. Vent close to the disk. Color pale, brown to slate;peritoneum pale.Synopsis.?Dorsal 28; anal 24?; pectoral 30?; pyloric coeca 30.Disk 1.9 in head. Gill slit above the pectoral fin. Dorsal fin un-notched. Dorsal connected to the caudal for nearly two-fifths of thelength of the latter. Pectoral unnotched. A small-sized species.LIPARIS TANAKAE (Gilbert and Burke)Liparis owstoni Schmidt, 1904a, p. 189 (part, not of Jordan and Snj'der). ? Tanaka,1908, p. 45, pi. 3, (part, not of Jordan and Snyder).Cydogaster tanakae Gilbert and Burke, 19126, p. 357.Type.?No. 21417, S. U. Z. M. Vries Island, Sagami Sea, Japan.Length 368 mm.Distribution.?L. tanakae apparently exists in the Japan Sea and offthe northeast coast of Hondo. I have examined specimens of the91668?30 7 86 BULLETIN 150, UNITED STATES NATIONAL MUSEUM =.^^.'''?Ji:'.'^st^.'r',..'i's?'z,^. FIGURK 15.-LIPARIS TANAKAE. Teeth AND PBICKLES PKOM TTPEof Noto, and another from the TnWn fioi. i ^ . REVISION OF THE FISH FAMILY LIPARIDAE 87 recorded as Liparis owstoni) in the Nagaski fish market. Nine speci-mens examined.Relationships.?L. tanakae agrees with L. owstoni in the broad con-nection between the dorsal and the caudal and in the unnotchedpectoral. It probably represents an intermediate stage between L.owstoni and the typical Liparis. It is distinguished from L. owstoni bythe broad, low, projecting snout, the broad head, and the coloration.Synopsis.?Dorsal 41-45; anal 34-36; pectoral 40-42; pyloric coeca55-62. Disk 2.5-2.7 in head. Gill slit extending down in front of6 to 11 pectoral rays. Snout broad, low, projecting. Pectoral finunnotched in the adult, possibly slightly notched in the young.Dorsal fin unnotched. Dorsal connected to more than the basaltwo-thirds of the caudal. A gigantic species, reaching a length of430 mm.Remarhs.?Two specimens, collected by H. M. Smith at Shiogama,Matsushima Bay, present certain peculiarities. They are not fullygrown and may not represent adult conditions but appear to belongto this species.The smaller specimen (113 mm.) agrees with the type of L. tanalcaein the shape of the body. The pectoral is slightly notched, when thefin is folded the seventh, eighth, and ninth rays slightly exceed thetenth, eleventh, and twelfth. Teeth arranged in about 15 rows in thehalf of each jaw. There is no white line on the caudal as in the type.The top of head and side of body with pale gray stripes separated bynarrow brown stripes. Maxillary reaching middle of eye. Pricklesabsent, no pits in the skin.The larger specimen (218 mm.) differs from the type of L. tanaJcaein a shorter snout and deeper nasal region. The body is shaped morelike L. owstoni. The maxillary reaches the middle of the eye. Ventseparated from disk by diameter of disk. The tenth, eleventh, twelfththirteenth, and fourteenth pectoral rays ending at same level. Stripeson the body obscure. This specimen shows more resemblance toL. owstoni than does the smaller specimen but the wide head andprojecting snout readily distinguish it from L. owstoni.LIPARIS OWSTONI (Jordan and Snyder)Trismegistus owstoni Jordan and Synder, 1904, p. 238, pi. 58.Liparis owstoni Schmidt, 19046, p. 189, figs. 1-2, (part, confused with L. tanakae).?Tanaka, 1908, (part, confused with L. tanakae).Cyclogaster owstoni Gilbert and Burke 19126, p. 358.Type.?Male, No. 8385, S. U. Z. M., Enoshima, Sagami Bay,Japan. Length 429 mm.Disirihution.?Sagami Bay, from deep water. One specimen, thetype examined.RelationsJiips . ?L. owstoni dift"ers from all the remaining species ofLiparis in the peculiar shape of the head. It agrees with L. tanalcae 88 BULLETIN 150, UNITED STATES NATIONAL MUSEUMin the broad connection between dorsal and caudal and the unnotchedpectoral.Synopsis.?Dorsal 43; anal 35; pectoral 40; pyloric coeca 50. Disk2.3 in head. Snout very short, broad, blunt, in profile rounded, not I'v *.'(-" -'A ^?u^j^r:-: = w-Jici.- "^Figure 16.?Liparis owstoni, showing the modified head, differing in this respect from theTIDE POOL speciesprojecting. Gill slit extending down in front of 10 pectoral rays.Pectoral fin unnotched. Dorsal fin unnotched. Dorsal connectedto more than the basal half of the caudal. A gigantic species, reachinga length of 429 mm.Remarks.?The shape of the headand the anterior position of thedisk agree with some species of Care-produs. The color of the dermis,being reddish violet, resembles thatof L. tanakae, L. ochotensis, andL. ingens. Tanaka (1908) andSchmidt (1904) have confused thisspecies with L. tanakae.LIPARIS PULCHELLUS AyresLiparis pulchellus Ayres, 1855, vol. 1, p.23.?Garman, 1892, p. 67, pi. 4, figs. 6-8.Cyclogaster 'pulchellus, Girard, 1858 p. 132.Type.?Lost. San Francisco fishmarket.Distribution.?A shaUow-water species extending from San Fran-cisco Bay to Bristol Bay, Alaska. Only one of our specimens is fromBristol Bay, and this record may be questioned. The species cer-tainly is not common in Bering Sea.Relationship.?L. pulchellus can be distinguished from all otherAmerican species by the extremely long connection between thedorsal and the caudal. It appears to be allied to L. tessellatus ofJapan. It can be distinguished from the latter species by the muchsmaller gill slit. Fifty-nine specimens examined. Figure 17.?Liparis owstoni.from type Prickle REVISION OF THE FISH FAMILY LIPARIDAE 89Description.?^A specimen from Point Reyes, Calif. Dorsal 48;anal 39; pectoral 37; pyloric coeca 32. Depth 4.6; head 5.6.Body soft, thick, and elongate. Head broad, flat, and blunt;occiput not swollen. Mouth broad; maxillary reaching verticalfrom front of pupil. Teeth short, strong, slightly recurved, stronglytrilobed, widely spaced, about 12 rows in the half of each jaw, theteeth in front appearing little smaller than those behind. Snoutbroad, bluntly rounded; jaws equal or the upper slightly projecting.Gill slit either above the pectoral or extending down in front of one tofour rays, more distinctly down in front of the fin in adults. Pricklesabsent.Dorsal normal, the rays increasing gradually in length. Caudalslender, connected to dorsal and anal for nearly its whole length, theconnection gradual, not abrupt as in nearly all the other species ofLiparis. Pectoral notched; the lower lobe reaching halfway betweendisk and vent. Disk large, 2.4 in head. Vent about midway betweendisk and anal, separated from disk by more than diameter of disk.Color light brown. Other specimens have the upper surfaces ofhead and body covered with narrow, wavj^ lines; lower parts white;in others dark spots and vermiculations replace the lines; verticalfins dusky; dorsal and anal with longitudinal stripes and mottlings;pectoral with crossbars; peritoneum silvery with dark dots.Synopsis.?Dorsal 48; anal 39; pectoral 37; pyloric coeca 32; disk2.4 in head. Gill slit either above the fin or extending down in frontof one to four pectoral rays. Dorsal fin unnotched; dorsal connectedto near the tip of the caudal, the connection gradual and obscure.No prickles. A small to medium sized species, reaching a length of123 mm. or more.UPARIS TESSELLATUS (Gilbert and Burke)Cydogaster tessellatus Gilbert and Burke, 19126, p. 355.Type.?Male^ No. 73328, U.S.N.M.; south coast of Hokkaido,Albatross Station 5042 ; depth 61 fathoms.Distribution.?South coast of Hokkaido to coast of Korea, Seaof Japan, Albatross Stations 4867, 5041, and 5042; depth 61-150fathoms. Ten specimens examined.Relationships.?L. tessellatus resembles L. pulchellus in the extremeconnection between the dorsal, anal, and caudal fins, and in thenumber of fin rays. It can be distinguished from L. pulchellus bythe much wider gill slit and the more pointed and compressed head.Synopsis.?Dorsal 46-48; anal 37-38; pectoral 35-38; pyloriccoeca 27-32. Disk 3-3.2 in head. Gill slit extending down in frontof 1 6 to 1 7 pectoral rays. Dorsal fin unnotched. Dorsal connectedto near the tip of the caudal, the two fins appearing almost continuous 90 BULLETIN 150, UNITED STATES NATIONAL MUSEUM as in L. pulchellus. A small to medium sized species, reaching alength of 187 mm. -jatjU'^ iiO^ ' if^lKH ?' Figure 18.?Liparis tessellatus. Ttpe. A spehes that in respect to the unnotched dorsal,UNION OF DORSAL, ANAL, AND CAUDAL AND IN APPEARANCE APPROACHES THE CaREPROCTUS TYPERemarks.?In specimens less than 135 mm. in length the head isnot so pointed and the snout does not project so for beyond the lowerjaw. LIPARIS MAJOR (Gill)Cyclopterus liparis major Fabricius, 1780, p. 136.Actinochir major Gill, 1864, p. 193; 1873, p. 193.Liparis fabriciiLvTKF.N, 1886. ? Gunther, 1887, p. 66. ? Gill, 1891, p. 376, pi. 29.Careprodus major Garman, 1892, p. 72Cyclogaster fabricil LoNNBTURG, 1899, (part).Distrihution.?Greenland and the Arctic Ocean to the eastward,probably circumpolar. Recorded by Liitken to range in depth from46 to 106 fathoms, by Lonnberg from 14 to 140 meters. Ten speci-mens examined. Figure 19.?Liparis major. Teeth from several specimensRelationships.?L. major is not closely related to any other knownspecies. In the appearance of the heavy head and body this speciessomewhat resembles L. megacephalus though the head is muchnarrower. L. major differs from all the other species of Liparis in REVISION OF THE FISH FAMILY LIPARIDAE 91the dark peritoneum and the dentition, the teeth being simpleor weakly trilobed.Description.?Body deep at union with head, short, taperingrapidly to caudal, much compressed. Head very heavy, flat betweenthe eyes, profile sloping rapidly to snout, cheeks nearly vertical.Mouth wide, the lateral cleft extending to vertical from middle ofeye; maxillary reaching beyond orbit. Teeth arranged in about 13oblique, widely spaced rows in the half of each jaw; inner teeth muchlarger, slender, recurved, simple, or with the lateral lobes but faintlyindicated; smaller teeth trilobed. Snout short, deep, rising abruptly;jaws equal. Anterior nostril in a short tube; posterior nostril reducedto a small pore. Eye large and prominent. Gill slit large, extendingdown in front of 10 pectoral rays, the flap broad. No prickles onany of the specimens examined. (See Liitken, (1886), for presenceof prickles in this species.) Pyloric coeca reduced in number, about20. Pores 2-6-7-2; rudimentary pores on the sides.Dorsal rather high, the origin far forward, in front of the base ofthe pectoral. Anal connected to more than the basal half of thecaudal; the dorsal connected to slightly less than the basal half of thecaudal; the connection between the fins abrupt, sometimes a notchpresent. Pectoral deeply notched, the upper edge on a level with thelower margin of the eye; the lower lobe of six rays, reaching vent.Disk large, with a narrow flap. Vent midway between disk and anal,separated from disk by two-thirds diameter of disk.Color a uniform dull slaty brown or the skin translucent and dustedwith fine brown dots, these forming dusky streaks along base ofdorsal; muscles puntulate; peritoneum pure black or heavily pig-mented, differing in this respect from all other species of Liparis;stomach and ovary dusky, gill lining speckled.Synopsis.?Dorsal 48; anal 40; pectoral 35; pyloric coeca 20.Disk 2.7 in head. Gill sHt extending down in front of nine pectoralrays. Dorsal fin unnotched. Dorsal connected to slightly less thanthe basal half of the caudal. Prickles present or absent. Perito-neum black or heavily pigmented. Reaching a length of 144 mm. ormore.Remarks.?This species resembles members of the genus Care-proctus in the large head, deep body, simple teeth, large eye, the colorof the peritoneum, the reduced number of pyloric coeca, the compara-tive number of pectoral rays, and the reduced posterior nostril.Genus POLYPERA BurkeNeoUparis Jordan and Starks 1895, (greeni).Polypera Burke, 1912a, (greeni).Disk large; dorsal fin notched; nostrils 2; teeth weakly trilobedand simple; pyloric coeca numerous, more than 200; branchioste-gals, 6. 92 BULLETIN 150, TJ^TITED STATES NATIONAL MUSEUMThe genus Polypera differs from Liparis solely in the number ofpyloric coeca. None of the 30 species of the latter genus has asmany as 100 pyloric coeca. The pyloric coeca in Polypera arematted closely together and can be distinguished without countingfrom those of any of the species of Liparis. The dentition in Poly-pera is distinct from that found in any of the species of Liparis inwhich the dorsal fin is notched. The larger teeth in Polypera greeniare simple or have the lateral lobes but faintly indicated. Thenearest approach to this type of dentition found amoung the speciesof Liparis is in Liparis major (Gill). In the latter species some ofthe larger teeth are simple, but in other characters the species is sowidely divergent from Polypera greeni that we are led to behevethat the similarity in dentition is due to parallel development.KEY TO SPECIES OF POLYPERAA*. Dorsal 40 or less; anal 30 or less; pectoral 37.B'. Color pale gray; gill opening above the pectoral or in front of the upper ray.P. beringianus.B^. Color light brown with the epidermis removed; gill slit either above thepectoral or in front of one to four rays P- greeni.A*. Dorsal 44; anal 34; pectoral 40 P- simushirae.POLYPERA GREENI (Jordan and Starks)Neoliparis greeni Jordan and Starks, 1895, p. 829, pi. 96. ? Jordan and Ever-MANN, 1898, p. 2112, fig. 763.lAparis tunicatus Bean and Bean, 1896a, p. 243.Liparis callyodon Bean and Bean, 1896a, p. 243 (part, No. 47561).Liparis hercshelinus Jordan and Gilbert, 1899, p. 476 (part.).TV^e.?Female, No. 3019, S. U. Z. M. Victoria Harbor, BritishColumbia. Length 244 mm.Distribution.?Apparently extending northward from British Co-lumbia to the Aleutian Islands. The specimen from Bering Islandappears to belong to this species. Depth unknown; the type wasdredged in Victoria Harbor, and the records of the Bering Islandspecimens are incomplete. Four specimens examined.Relationships.?The three species of Polypero recognized in thiswork appear to be very closely allied and I am not certain that allthree species should be recognized. More material must be obtainedbefore we can demonstrate the validity of all three species. P.simushirae is based upon a single specimen. The horizontal dis-tribution of P. greeni and P. beringianus is practically identical;it is possible that the vertical distribution of the two species may differ.P. greeni appears to come from deeper water than either P. beringia-nus or P. simushirae. As P. beringianus is known only from smallspecimens it is possible that it may represent the young of eitherof the other species. We will indicate here the apparent differencesbetween the three species. P. beringianus has a paler coloration and REVISION OF THE FISH FAMILY LIPARIDAE 93perhaps a smaller gill slit than P. greeni. P. simushirae appears todiffer from P. greeni solely in the larger number of fin rays. P.heringinaus differs from P. simushirae in the Ughter coloration,smaller gill slit and the smaller number of fin rays. Minor differ-ences in the proportional measurements appear to exist.Synopsis.?Dorsal 37-40; anal 31-32; pectoral 33-37; pyloriccoeca more than 200. Disk 2.3-2.8 in the head. Gill slit typicallyextending down infront of 1 to 4 pec-toral rays, some-times appearing tobe above the fin.Teeth simple andtrUobed, elongate,recurved. Dorsalfin deeply notched, . 1 , 1 Figure 20.?Polypera greeni. Teeth from typeconnected to lessthan the basal fifth of the caudal fin. Color a light brown. A largespecies, reaching a length of 244 mm.RemarJcs.?The Bering Island specimens differ in certain respectsfrom the type as wUl be seen by the following notes on these specimens.Head wider than deep; cheeks swollen. Maxillary reaching posteriormargin of eye. Spout somewhat depressed, the upper jaw slightlyoverlapping. Greatest depth of body at front of first dorsal. ThegUl slit is either above the pectoral or extending down in front of oneor two rays. Vent separated from the disk by less than the diameterof the disk. Origin of dorsal above middle of pectoral and midwaybetween the disk and vent. A shallow notch between the dorsal andcaudal fins is present in the large specimens as figured in the type,absent in the smaller specimen. The two large specimens colored asoft brown as the type. The smallest specimen faded out to an ashybrown. Peritoneum pale with scattered dots,POLYPERA BERINGIANUS (Gilbert and Bnrke)Liparis callyodon Starks, 1911, p. 196.Cyclogaster beringianus Gilbert and Btjrke, 1912a, p. 72.Tt/PC?Female, No. 74380 U.S.N.M.; Nikolski, Bering Island,Albatross, 1906; tide pools. Length 65 mm.Distribution.?'Aleutian Islands and southward to Port Townsend,Wash. The specimens examined are from Bering Island; MedniIsland; Agattu Island; Unalaska; Kodiak Island; and Port Townsend,Wash. A tide-pool species. Thirty-four specimens examined.Relationships.?P. beringianus closely resembles P. greeni and P.simushirae. A discussion of the relationships of these three species isgiven under P. greeni. 94 BULLETIN 150, UNITED STATES NATIONAL MUSEUMSynopsis.?Dorsal 38-39; anal 31-32; pectoral 36-37; pyloriccoeca more than 200. Disk 2.4-2.9 in the head. Gill slit typically-above the pectoral fin, sometimes appearing to extend down in front ofthe upper ray. Teeth as in P. greeni or the lateral lobes appearingsomewhat stronger. Dorsal fin notched, barely connected to thecaudal. Color pale gray. A small sized species, reaching a length of126 mm.Remarks.?^In this species the depth of the body decreases verygradually, or remains the same from the first dorsal to well under the ? ^Figure 21.?Polypera beringianus. Type Figure 22.?Polypera beringianus. Teeth from typesecond dorsal. The gill slit is either above the pectoral or the upperpectoral ray points at a slight angle downward directly at the loweredge of the gill slit. In some specimens there are shallow notches be-tween the caudal and the dorsal and anal.POLYPERA SIMUSHIRAE (GUbert and Burke)Cyclogasier simushirae Gilbert and Burke, 19126, p. 354.Type.?Male, No. 73327, U.S.N.M.; Milne Bay, Simushir Island,Japan, Albatross, 1906; tide pools. Length 138 mm.Distribution.?Known only from the type specimen,species. Type examined. A tide pool REVISION OF THE FISH FAMILY LIPARIDAE 95Relationships.?P. simushirae is closely allied to P. greeni and P.heringianus. An account of the differences between these threespecies is given in the discussion of P. greeni.Synopsis.? Dorsal 44; anal 34; pectoral 40; pyloric coeca about 300.Disk 2.7 in the head. Dentition resembling that in P. greeni and P. Figure 23.?Polypera simushirae. Teeth from typeheringianus. Dorsal fin notched, connected slightly with the caudal.Color olive brown. Reaching a length of 138 mm.Genus CAREPROCTUS (Kroyer)Careproctus Kroyer, 1862, p. 252, (reinhardi) . EnantioUparis Vaillant, 18886, p. 22, (pallidus).Caremitra Jordan and Evermann, 1896, p. 452, (simus).Bathyphasma Gilbert, 1896, p. 442, (ovigerum).Allochir Jordan and Evermann, 1896, p. 452, (melanurus) . Allurus Jordan and Evermann, 1896, p. 452 (ectenes); 1898, p. 2136, (edenes).Allinedis Jordan and Evermann, 1898, p. 2866, {ectenes).Prognurus Jordan and Gilbert, 1899, Fur Seal Rept., pt. 3, p. 478 icypselurus) . Disk present, perfect, nostril single; teeth trUobed to simple;suprabranchial pores usually 2; dorsal notch present or absent;pyloric coeca present or absent; pseudobranchiae absent; branchios-tegals 6. MODIFICATION OF CHARACTERSBody.?The body varies from firm, short, stout, and depressed, asin C. pycnosoma, to the gelatinous, elongate, and compressed speciestypical of the genus. Many of the species have the body compressedand deepened in sharp contrast to the condition typical of the speciesof Liparis.Head.?The head of the species of this genus is typically compressed,the depth being greater than the width. We should note that the 96 BULLETIN 150, UNITED STATES NATIONAL MUSEUM reverse is true of the majority of the species of Liparis. An exceptionto the typical compressed head is found in Careprodus curilanus inwhich species the width of the head is greater than the depth, thecheeks slope outward, and the snout is depressed and ends in a shovel-like projection. The cheeks are typically vertical and the snoutshort. The profile of the snout may rise gradually or abruptly fromthe upper lip. The mouth is always horizontal. The interorbitalregion may be quite narrow, as in C. segaliensis, or broad and flat,as in the spectrum group of species. Of the above characters thedepth of the head and the shape of the snout can be the most readilyutilized in separating species.Nostril.?The posterior nostril is never present in this genus. Theanterior nostril typically projects above the surface of the head. Insome of the species, as in C. hathycoetus , the nostril tube hardly projectsand is less prominent than the posterior nostril in some species ofLiparis. The posterior margin of the tube is sometimes raised intoa projecting flap, as in Careprodus pycnosoma. The position of thenostril opening varies in its relation to the eye, depending to someextent upon the shape of the head. It may be above the front ofthe eye or directly in front of the pupil. The writer failed to makeany practical use of the character presented by the nostril. Thiswas partially due to the fact that in these fishes the nasal region isfrequently collapsed in preserved specimens.Eye.?The eye in the species of Careprodus is typically larger andmore prominent than in the species of Liparis. It is usually containedfrom three to five times in the head. In the giant species, such asG. colletti, the eye does not appear especiall}'- prominent. The pupilis round or slightly oval. The pupil varies from being reducedalmost to a point, as in C. entargyreus, to enlarged and comprisingthe major part of the eye. All of this variation can not be due tocontraction and expansion of the iris. The color of the eye rangesfrom black to silvery. In many species, as in C. spedrum, the upperhalf of the eye is black and the lower half silvery. The silverypigment frequently disappears in preserved specimens. The size ofthe eye and pupil and the coloration of the eye can be utilized indistinguishing certain of the species.Pores.?The pores are frequently difficult to study on account ofthe condition of the sldn. In some of the species the pore formulacould not be made out. The pore formula 2-6-7-2 appears to beconstant for the majority of the species of the genus. In C. sinensisand C. segaliensis, however, the pore formula is modified to 2-5-6-1.This may be true of other species in which the pores could not bestudied. The position of the pores varies among the species. Thesuprabranchial pores may be closed together or widely separated.The posterior pore in C. trachysoma is rudimentary. The upper REVISION OF THE FISH FAMILY LIPARIDAE 97pore on the snout may be high up and nearly between the nostrils,as in C. pellucidus and C. tracJiysoma, or far forward on the snoutnear the anterior pores, as in C. hathycoetus and C. roseqfuscus. Insome of the species the maxillary and mandibular pores are sunkenin pits. In C. sinensis and a few other species the anterior mandibularpores have a common opening. In C. hathycoetus, a black species, thepores appear as white dots. A more extensive study of the pores isdesirable as they will probably be found to be of considerabletaxonomic importance.Many of the species of Careproctus as of Liparis have rudimentarypores on the sides of the body and on the nape. These pores couldnot be detected on many of the species with a thin, lax skin and areprobably absent. They are the most readily seen on the generalizedspecies C. pycnosoma, C. curilanus, and C. howersianus.Gill slit.?In the majority of the species of Careproctus the gill slitis confined to the region above the base of the pectoral. For thisreason the gill slit is not of such taxonomic importance as in Liparis.In the latter genus the gill slit, in the majority of the species, extendsdown in front of the pectoral and varies greatly in size. The variationin the size of the gill slit, when confuied to the region above the baseof the pectoral fin, is so slight as not to be of taxonomic value.The extreme variation in the gill slit is almost as great as in Liparis.In C. ostentum the gill slit extends down in front of 14 pectoral rays.The gill membranes are frequently torn so that it is difficult to decidewhether the slit is confined to the region above or extends down infront of the fin. The variation within some species is such that thegill slit in some specimens is above the pectoral fin and in otherspecimens extends down in front of the upper one or two rays.Teeth.?The teeth may be either trilobed or simple. In some ofthe species both kinds of teeth are present. The teeth do not presentany characters of generic value. We can rouglily divide the genusinto three groups of species based on the character of the teeth butthese groups grade insensibly into each other. These three groupsmay be defined thus: (1) Teeth strongly trilobed; (2) teeth weaklytrilobed to simple; (3) teeth simple, lanceolated, recurved, the anteriorteeth appearing as large as the inner teeth.The arrangement of the teeth on the jaw varies among the differentspecies. They are always arranged in oblique rows. The inwardlydiverging rows are, in some of the species with simple teeth, veryoblique and difficult to count. In such species the band of teeth isusually narrow. A good example of this type of dentition is foundin C. gilberti. The cutting surfaces presented by the bands of teethare typically oblique. This is because the anterior teeth are smallerand placed lower on the outer surfaces of the jaw bones. A strikingexception to this type of dentition is found in C. colletti and related 98 BULLETIN 150, UNITED STATES NATIONAL MUSEUM species. In these species the cutting surfaces of the bands of teethappear horizontal when the jaws are opened. This is because theanterior teeth are not so much smaller than the inner teeth and notplaced so low on the jaw bones as among other species. Also theinner teeth are recurved which helps to flatten the cutting surfacesof the bands of teeth.The characters of the teeth are of little or no practical use in separat-ing closely related species but are of considerable value in indicatingthe general relationships of the species.Dorsal jin.?The rays of the dorsal fin are difficult to count onaccount of the delicate nature of these fishes. The rays appear asdelicate threads in the fin tissue. The writer found the most satis-factory method for counting the fin rays to be to place the specimenon a piece of glass or the dissecting microscope, slit the fin membraneat the margin of the fin and fold it back on the body, and, with apair of hooked needles, separate the rays as they are counted. It issometimes an advantage with very delicate specimens to place themin a glass vessel, cover with alcohol, and examine under the simplemicroscope, the light being properly controlled from below.The number of dorsal rays ranges from about 40 to 60. On accountof the difficulty of making accurate counts of the number of dorsal oranal rays it has not been considered advisable to depend upon thenumber of these rays to separate species. With the study of morematerial (many of the species are represented by a single specimen),it will be possible to determine the number and amount of variationof the dorsal and anal fin rays. Then it will probably be found thatthe number of rays will serve as an index to many of the species.The origin of the dorsal varies slightly but can not be used to advan-tage in distinguishing species at the present time. The length of thefirst dorsal ray, as compared with the diameter of the eye, serves toseparate some of the species. This is a character which deservesfurther investigation. In C. yycnosoma the anterior dorsal rays arerather stiff and project distinctly. This condition is characteristicof the species of Liparis. In many of the species the anterior dorsalrays are buried in pseudotissue and do not come in contact with thefin membrane. They can be seen only by dissecting the fin. Insuch fins the anterior rays extend undulatingly backward.The dorsal notch is absent or but faintly indicated in this genus.Two species, C. attenuaius and C. pycnosoma, appear to have the dorsalnotch persisting. In the specimens known the fourth, fifth, andsixth rays appear slightly shorter than the preceding one.The anterior rays are unsegmented but appear to be divided longi-tudinally and thus differ from the condition of the anterior rays in thespecies of Liparis. The following table indicates the species in which EEVISION OF THE FISH FAMILY LIPARIDAE 99the segmentation of the anterior dorsal and anal fin rays was investi-gated and the number of unsegmented rays in each fin. Dorsal AnalSpecies fin flngilberti 15 7Do 12 6melanurus 9 3sinensis 15 3ranula 14 4mirabilis (?) 12The connection between the dorsal and caudal is nearly constantfor the genus and is typically less then one-half the length of the caudal.AnalJin.?The number of anal rays varies from 32 to 57. What hasbeen said concerning the taxonomic value of the number of dorsalrays applies equally well to the number of anal rays. It is doubtfulif the number of dorsal or anal rays for the species varies more thansix or seven. The number of anal rays in a species is always less thanthe number of dorsal rays and, with a few exceptions, in which thenumbers are equal, greater than the number of pectoral rays. InLiparis the pectoral typically has a greater number of rays than theanal. A few of the anterior anal rays are unsegmented. The con-nection between the anal and caudal varies from three to six tenths ofthe length of the latter but is of little taxonomic value.Caudal Jin.?The number of caudal rays varies from 6 to 12 but istypically 8 or 10. There are few or no rudimentary rays at the baseof the caudal. All the rays typically enter into the body of the fin.The variation of the number of caudal rays within a species has notbeen studied, and until this is done we can not utilize the number ofrays in separating species.The caudal is typically truncate or slightly rounded. An exceptionis found in C. Jurcelliis, in which the caudal is slightly forked, and inC. cypselurus, in which it is deeply forked.Pectoral fin.?The pectoral presents a number of important char-acters. The number of rays, the extent of the pectoral notch, andthe length of the pectoral fin, especially of the lower lobe, can be uti-lized to a considerable extent in distinguishing the species.The number of pectoral rays varies from 21 to 37. The numberof rays for any one species apparently does not vary more than fouror five. The pectoral shows the greatest reduction in the more special-ized members of the genus. In Liparis, as we have already noted, thenumber of pectoral rays, typically exceeds the number of anal rays.In Careprodus the change has been in the opposite direction and thepectoral typically has a smaller number of rays than the anal fin.The pectoral notch may be absent or extend nearly to the base ofthe fin. In C. Jurcellus the notch is absent or hardly discernible andin C. cypselurus very shallow. It is typically distinct, however. 100 BULLETIN 150, UNITED STATES NATIONAL MUSEUMIn only one species, C. longifilis, are the middle pectoral rays rudimen-tary. In this species the notch extends nearly to the base of the fin.The length of the lower pectoral lobe varies considerably amongthe different species but is quite constant for any one species and isof considerable taxonomic importance. It apparently varies butslightly with age. In C. abhreviatus the lower pectoral lobe is con-tained two or more times in the head, and in C. rhodomelas and C.colletti it exceeds the length of the head, and in C. longifilis is greatlyelongate, extending past the middle of the body. The rays of thelower lobe are graduated in length and partly free. In C. collettithe rays are free nearly to the base. In none of the species of thegenus, however, do they become as free as in some species of Parali-paris. The tips of^the rays are sometimes coiled or wavy.The upper edge of the pectoral fin remains at about the same levelon the side of the body throughout the genus.Dislc.?The ventral disk becomes greatly modified both in sizeand shape in this genus. Its position also changes and the measure-ment "snout to disk" is of considerable importance. In the moregeneralized species the disk is large and oval or round and in everyrespect resembles the disk in Liparis. As we examine some of themore specialized species we find the disk becoming smaller. InC. gilherti it is contained seven to nine times in the head, and in C.ostentum it has become minute but remains perfect in structure.With the reduction in size the disk becomes triangular and cupped.The triangular shape is caused by the margin posteriorly and on thesides anteriorly folding over the center. When the whole margin isfolded over the center the disk is cupped. Whether or not the tri-angular shape always precedes the cupped condition is uncertain.The width of the margin is greatly reduced in some species. Examplesof a triangular disk can be seen on C. rhodomelas and C. longipinnis andof a cupped disk on C. abhreviatus, C. colletti, C. hatliycoetus , and others.In C. rhodomelas the margin of the disk is very narrow.The above enumerated characters of the disk are of considerable tax-onomic importance. The size and shape of the disk vary somewhatwithin a species but in general they aid in distinguishing many of thespecies.Vent.?The vent is typically close behind the disk and the distancebetween the two varies little in the majority of the species. There arethree species in which the vent is some distance from the disk. In C.attenuatus and C. pycnosoma the vent is about midway between thedisk and margin of the anal fin and in C. opisthotremus it is nearer theanal fin than the disk. ^PseudohrancMae.?Pseudohrsinchisie are absent from many if notfrom all the species of the genus. The following species were found tolack pseudobranchiae : gilherti, melanurus, cypselurus, howersianus, col- REVISION OF THE FISH FAMILY LIPARIDAE 101 letti, and spectrum. It is possible that pseudobranchiae may be pres-ent in the more generahzed species of the genus.Pyloric coeca.?The number of pyloric coeca varies from to 45.The majority of the species have less than 20. This is quite a reduc-tion from the number found in Liparis in which genus the numbervaries from 10 to more than 90 and only 2 or 3 of the species have lessthan 20. On account of the scarcity of material the taxonomic valueof the number of pyloric coeca has not been investigated. The coecain this genus became shifted over the left side of the abdominal cavitydue to a shortening of the alimentary tract. They are more readilyreached through an incision on the left side of the body.PricMes.?Two distinct types of pricldes, the "thumb-tack" andcactuslike, are present on some of the species of Careproctus. Onlytwo species ? C. entomelas and C. entargyreus?are known to have *' thumb-tack" prickles. It is very likely that other species, possiblyknown ones, will be found to have this type of prickle. Four species ? C. ostentum, C. rastrinus, C. trachysoma, and C. acanthodes?are knownto possess cactuslike prickles. The significance of these prickles isunknown. Whether or not they are of specific value remains to beinvestigated. They are present on only one of the specimens of C.ostentum but are present on all the specimens of the other three species.They are easily rubbed ofl; with the epidermis.Coloration.?The species of Careproctus are never variegated, mot-tled, barred, or striped as is typical of the species of Liparis. The spe-cies of Careproctus are uniformly colored whitish, pinkish, dusky, orblack, or a combination of these colors, but never in such a manner asto be variegated. White and pink predominate in the shallower waterforms and black in the greater depths. C. pycnosoma, one of the mostgeneralized species of the genus, has a uniform dusky gray color like afew species of Liparis. The peritoneum varies from silvery to dottedand pure black. The stomach varies from pale to black. It variesindependently of the peritoneum and may be black when the perito-neum is pale or pale when the peritoneum is black. The peritoneumis white or black, but apparently is never white or silvery when theepidermis is black. The color of the mouth and gill cavity variesfrom pale to black. When the skin is transparent the flesh is usuallystippled with black. The black pigment apparently encroaches firstupon the caudal region, peritoneum, gill cavity, and mouth. Theseregions are frequently black when the remainder of the body is lightcolored. All intermediate stages from white to pure black are foundin the genus. Rarely, if, ever is there more pigment in the dermisanteriorly than posteriorly. The genus represents the transition fromthe hemibathybial to the bathybial type of coloration; Liparis rep-resents the littoral type,91668?30 8 102 BULLETIN 150, UNITED STATES NATIONAL MUSEUMThe specific coloration appears to be quite constant and of consider-able taxonomic importance. We have hardly enough material todraw any definite conclusions concerning the color variation within thespecies. A few examples will suffice to indicate both our lack of knowl-edge concerning the variability of coloration in these fishes and thepossible importance of the color in distinguishing species. C. spectrumand C. gilherti resemble each other closely and formerly have been con-fused. Their coloration differs, however, for in C. spectrum the stom-ach is always white and in C. gilherti blacldsh. The difference in thecolor of the stomach is supplemented by anatomical differences whichhave formerly been overlooked. C. spectrum and C. melanurus areclosely related species. The former difTers from the latter in havingless pigment in the mouth, gill cavity, peritoneum, and the posteriorpart of the body. The two species differ also in distribution; theyinhabit neighboring regions. In some of the specimens of C. melanu-rus the peritoneum has little more pigment than is typical of thespecies C. spectrum. C. cypselurus varies somewhat in coloration.The posterior part of the body is always black, but the anterior partvaries from wliitish to purplish black. C. entomelas and C. entargyreuswere thought at first to differ solely in the amount of pigment in theperitoneum, but further study brought to light other differences.These few examples indicate that whenever we find two specimensdiffering in coloration it is possible that they represent differentspecies and that other differences should be searched for.Habits.?Our knowledge concerning the habits of these fishesmust necessarily be limited. The dredging records indicate in mostcases, nothing but the depth at which the species exist and the kindof bottom they inhabit. The structure of the body, from analogy,indicates in some cases what the habits may be. The species takenin intermediate hauls are of course free-swimming forms. Themajority of the species of the genus have the body compressed anddeepened and probably seldom rest upon the bottom. The rays ofthe lower pectoral lobe are elongate in a number of species andpossibly are used as feelers. C. pycnosoma and the few other specieswith depressed bodies probably come to rest upon the bottom as dothe species of Liparis. C. rJiodomelas, judging from the contents ofthe stomach and the character of the teeth, feeds mainly uponOphiurians and must remain in close proximity to the bottom. Thegenus is represented on all kinds of ooze and rock bottom.Summary.?In order to present a guide to the study of the specificcharacters of these fishes the following outlines are given. The firstlist suggests the characters which have been found to be the mostuseful in distinguishing the species. The second list involves thosecharacters which were studied but not used to any extent. With EEVISION OF THE FISH FAMILY LIPARIDAE 103further study some of these characters will doubtless be found to be ofconsiderable value in designating species. For modification of char-acters see table on page 35. List 1Body Shape.Depth.Length.Texture.Head Length.Depth.Width.Snout DepthProjecting.Eye Size.Pupil.Color.Gill slit Size.Teeth Arrangement.Trilobed.Size. Dorsal Notched.Buried.Caudal Forked.Pectoral Number of rays.Notch.Length lower lobe.Disk Size.Shape.Position.Vent Position.Coloration Body.Gill cavity.Mouth.Peritoneum.Stomach.Distribution.List 2Nostril.Pores. - Dorsal. Length tube.Position.Formula.Position.Rudimentary.Number of rays.Segmentation of rays.Origin.Connection to caudal. Anal Number of rays.Connection to caudal.Caudal Number of rays.Pectoral Position.Prickles.Pyloric coeca.. NumberColor.Length.DISTRIBUTIONThe genus Careprodus is widely distributed, as is usual with deep-water genera. It is represented in the moderately cold waters of thenorth and south temperate regions and at greater depths inHhetropical Pacific. It has not been recorded from the tropical Atlanticor the Indian Ocean but may reasonably be expected to be discoveredin these regions.The giant species of the genus, as is true of Liparis, are mostcommon in the North Pacific. The giant species and their distribu-tion are as follows: C. ovigerum, 318 mm.; off British Columbia,depth 1,588 fathoms; C. longipinnis , 270 mm., Arctic Ocean, depth702 fathoms, C. trachysoma, 263 mm., Japan Sea, depth 318 fathoms;C. cypselurus, 260 mm., coast of Washington to Okhotsk Sea, depth510-887 fathoms; C. coUetti, 248 mm.. Gulf of Alaska to OkhotskSea, depth 284-629 fathoms; C. rastrinus, 280 mm., Okhotsk Sea,depth 73-119 fathoms. Two species, C. cypselurus and C. colletti,range from the American to the Asiatic coast, but the largest specimens 104 BULLETIN 150, UNITED STATES NATIONAL MUSEUM of both species have been taken from the western end of their areasof distribution. Bering Sea, the Okhotsk Sea, and the northern partof the Japan Sea appear to be more favorable to the production oflarge species and their existence nearer the surface of the ocean.The highest temperature in which a species of Careproctus has beentaken is a doubtful one of 59? F. for C. ranvla. The average for thisspecies is 42.2? F. None of the other species have been taken intemperatures above 45? F. The majority of the species are found ina temperature of 40? F. The regions inhabited by these fishes arelittle or not at all affected by the change of the seasons. The recordsthat we have indicate that few, if any of the species, range throughmore than 10? of temperature.Asiatic and American species.?The Japanese and Americanspecies are mainly distinct. At present only two species, C. cypsel-urus and C. colletti, are known to be common to both the Americanand Japanese faunas.Regions.?Our records concerning the distribution of the speciesof this genus are too incomplete to offer a basis for generalizing.Many of the species are known from but a single locality. The datathat we have indicates that the North Atlantic species are differentfrom the North Pacific species. Only two species, C. cypselurusand C. colletti, are recorded from both sides of the Aleutian Islands;C. giTberti is a possible third species. None of the species are commonto the two sides of the Japanese Archipelago. The Gulf of Alaska,Bering Sea, Okhotsk Sea, and the east and west coasts of Japanappear to have distinctive species. None of the species of the tropicalPacific have been taken in the North Pacific. We should expect tofind that the Arctic Current has distributed the Bering Sea speciessouthward to Japan. Our records do not indicate that this hashappened to any great extent.Careproctus, as well as Liparis, has but a meager representationin the North Atlantic. Only four species are recorded from thisregion and adjacent portions of the Arctic Ocean. Other species willdoubtless be discovered during further explorations but apparentlythe genus is poorly represented in this region.When we compare the range of specific distribution of the speciesof Careproctus with those of Liparis we find a similarity. We haveseen that some tide-pool species of Liparis range through practically20? of latitude. None of the species of Careproctus are known torange through more than 20? of latitude. The records that we havedo not indicate that these deep-water species have a wider range ofdistribution than the shallow-water species.Closely related species.?The close relatives of the majority of thespecies have yet to be discovered. In no case can the most closelyrelated species be said to inhabit the same environment. The REVISION OF THE FISH FAMILY LIPARIDAE 105 nearest approach to this condition is that of 0. cypselurus, Bering Sea,depth 510 to 887 fathoms, and C. furcellus, among the AleutianIslands, depth 480 fathoms. A splendid example of closely relatedspecies inhabiting neighboring regions is that of C. phasma, BeringSea, depth 46 to 59 fathoms; C. spectrum, Gulf of Alaska, depth 92 to110 fathoms; and C. melanurus, British Columbia to California, depth284 to 876 fathoms.Bathymetrical distribution.?As has previously been said, speciesof Careprodus have been taken at depths between 30 and 1,823fathoms. The species are found nearer the surface in the northernregions. Two-thirds of the species are found between 100 and 500fathoms. Only four are found below 500 fathoms. The center ofpopulation for the genus is at about 300 fathoms. The red lightrays fail to penetrate below this level.Some of the species are known to have considerable vertical dis-tribution. C. cypselurus has a vertical distribution of 377 fathoms,C. reinhdrdi 395 fathoms, C. ranula 354 fathoms, C. mollis 347 fathoms,C. colletti 341 fathoms, and C. melanurus 592 fathoms.Center oj dispersal.?The genus Careprodus apparently has hadthe same place of origin and dispersal as Liparis. The regions ofdistribution of the two genera are quite similar except that Care-produs typically inhabits a lower level than Liparis. The NorthPacific appears to be the region from which representatives of thesetwo genera migrated to the North Atlantic and southward in thePacific. The most primitive species, the greatest number of species,the greatest amount of structural variation, and largest species arefound in this region. KEY TO SPEQES OF CAREPR0CTU3 ??A'. Dorsal with a shallow notch; teeth trilobed.B^. Body stout; pectoral notch shallow; dorsal 42; anal 37; pectoral 37, peri-toneum pale pycnosoma.B^. Body slender; pectoral notch deep; dorsal 48; anal 40; pectoral 34; peri-toneum black attenuatus.A'. Dorsal unnotched.CK Teeth short, stout, (cf. ectenes), strongly trilobed, (cf. mollis, simus, andsinensis, melanurus, and opisthotremus) ; peritoneum, except in attenu-atus, pale.D'. Anterior dorsal rays with their tips projecting above the fin membrane.E^ Depth less than 6; head less than 4; snout not projecting; gill slitabove the pectoral; dorsal 42; anal 37 pycnosoma.E^. Depth more than 6; head more than 4; snout distinctly projecting;dorsal 49; anal 43;F'. Disk less than 3 in head curilanus.F^. Disk more than 3 in head ectenes. " Species listed in the appendix not included. 106 BULLETIN 150, UNITED STATES NATIONAL MUSEUM D^. Anterior dorsal rays weak, their tips not projecting above the finmembrane, usually buried in pseudotissue.G^ Lower lobe of the pectoral shorter than the head; color pale ordusky.H^ Gill slit in front of 3 or 4 pectoral rays; depth less than 6.I^. Pectoral 33; snout projecting; eye 3 in the head simus.P. Pectoral 35; snout not projecting; eye 4.8 in thehead mollis,H^. Gill slit above the pectoral or in front of the upper ray.J'. Occiput high; depth 5 or less.K'. Snout projecting; skin on the head extremely lax; thepores on the chin and snout in pits; body muchcompressed.L^. Pectoral 30; disk 3.5 sinensis^U. Pectoral 24; disk 4.2 segaliensis.K*. Snout not projecting; skin not lax; pores on head notin pits; body stout, little compressed.M^ North Pacific: anal 46; pectoral 37; disk5 bowersianus?M^. South Atlantic; anal?; pectoral?N^. Pectoral notched; rays 31; disk 2.2_falklandica,N^. Pectoral unnotched? pallidas,J*. Occiput not high; depth distinctly more than 5.O'. Disk 3.7; lower lobe of the pectoral 1.5, firstdorsal ray elongate, equal to the orbit,,longer than the 3 or 4 succeeding rays;stomach pale; peritoneum black.attenuatus,O^. Disk 2.7; lower lobe of the pectoral 2.4; firstdorsal ray about equal to one-half the eye,shorter than the second ray; stomach black;peritoneum pale, dotted homopterus.G^. Lower lobe of the pectoral longer than the head ; teeth very stout,the lateral lobes nearly equal to the central lobe; color black.rhodomelas.C*, Teeth more elongate, slender (cf. mollis, melanurus, sinensis), simple,sometimes arrow shaped, the lateral lobes hardly evident, (cf. opistho-tremus, gilberii), frequently recurved; peritoneum pale or black.P'. Teeth rather short and stout, (see ovigerum) , the cutting surfaces oblique; peritoneumpale to black.Q'. Gill slit above the pectoral.R'. Pectoral more than 25.S*. Disk as large or larger than the eye,not much cupped.T*. Body slender; depth more than 5.Atlantic.U*. Eye less than 4 in the head;pectoral notch hardly evident.ranula.U*. Eye more than 4 in the head,v. Pectoral notch shallow; dor-sal 47; anal 41..micropus. REVISION OF THE FISH FAMILY LIPARIDAE 107 V2. Pectoral notch deep; dorsal55; anal 45; pectoral 33.reinhardi.T2. Body deep; depth less than 5.W. Atlantic.X^. North Atlantic; lowerlobe of the pectorallonger than the head;disk 4.6__longipinnis.X2. South Atlantic; lowerlobe of the pectoralshorter than the head;disk 2 georgianus.W2. North Pacific.Lower lobe of the pectoralshorter than the head;disk 3.3 phasma.S^. Disk smaller than the eye, rarelyequal to it; usually distinctlycupped or triangular.Y'. Caudal not forked;pectoral distinctlynotched.Z'. Peritoneum silvery,undotted; mouthand gill-cavitypale. . Lower lobe of the pectoral nearly equal to the head or longer than the head; noprickles.b'. Pacific, south of Alaska Peninsula spectrum.b^. Atlantic longipinnis.a*. Lower lobe of the pectoral about 2 in the head; Gulf of Tartary; cactuslikeprickles present acanthodes.T?. Peritoneum black,rarely silverj' anddotted; mouthand gill cavitydusky to black;no prickles; Brit-ish Columbia toCalifornia.melanurus.Y2. Caudal forked; thepectoral notchhardly evident orabsent.c'. Depth 4.5; pectoral slightly notched; pyloric coeca 29 cypselurus.c^. Depth 4; pectoral unnotched; coeca 45 furcellus.R2. Pectoral with less than 25 rays; diskdeeply cupped,d *. Upper pectoral ray not elongate; middle pectoral rays not rudimen-tary.e'. Color pale; tail short abbreviatus.e*. Color black; tail attenuate bathycoetus.d^. Upper pectoral ray greatly elongate; middle pectoral rays rudi- 108 BULLETIN 150, UNITED STATES NATIONAL MUSEUM mentary longifilis.Q.2 Gill slit extending down in front of thepectoral fin.f ^ Disk well developed, not over 4 in the head.gi. Body deep; depth not over 4 in the length; vent near thedisk ovigerum.g2. Body depressed, elongate; depth about 6; vent far back fromthe disk opisthotremus.f 2, Disk small, 6 or more in the head.h'. Gill slit in front of not more than 6 pectoral rays.i^ Peritoneum silvery, undotted.ji. Eye 3.4; dorsal, 55; anal, 49; pectoral, 33-35; no prickles,pellucidus.j2. Eye 4.6; dorsal, 59; anal, 52; pectoral, 34-37; cactuslikeprickles present rastrinus.12. Peritoneum silvery, dotted; dorsal 58; anal 53; pectoral31-32; cactuslike prickles present trachysoma.h^. Gill slit in front of 10 or more rays.k^ Disk small, 7 to 9 in the head gilberti.k*. Disk minute, more than 9 in the head ostentum.P2. Teeth slender, lanceolate, recurved; thecutting surface appearing horizontalwhen the jaws are opened, outer teethappearing little shorter than the innerteeth; peritoneum black except inentargyreus.V. Body deep; depth less than 5 in the length; lipspale; lower lobe of the pectoral about 2 in thehead, not reaching beyond the vent.roseofuscus.P. Body not so deep; depth more than 5 in specimensless than 120 mm. in length; lower lobe of thepectoral distinctly less than 2 in the head,reaching past the vent.m*. Prickles absent; lips dark or black; lower lobeof the pectoral about equal to the head,usually reaching midway between vent andanal fin coUetti,m2. Thumb-tack prickles present; lower lobe of thepectoral about 1.5 in the head, reachinglittle past the vent,n*. Peritoneum black; pupil moderate; depth144-428 fathoms entomelas.n^. Peritoneum pale, dotted; pupil minute; depth35-66 fathoms entargyreus.CAREPROCTUS PYCNOSOMA Gilbert and BurkeCareproctus pycnosoma Gilbert and Burke 19126, p. 372.Type.?No. 73340, U.S.N.M.; Albatross Station 4803, off CapeRollin, Simushir Island, Japan; depth 229 fathoms. Length 46 mm.Distribution.?Off Simushir Island, Japan, Albatross Station4803; depth 229 fathoms. Type specimen examined.Relationslhips . ?In the character of the anterior dorsal rays andthe firm body C. pycnosoma appears to be the least modified of any REVISION OF THE FISH FAMILY LIPARIDAE 109 of the species of the genus. The short, firm body, the number offin rays, the dentition, the truncate snout, and the coloration of theperitoneum serve to distinguish C. pycnosoma from the other specieshaving a notched dorsal. Figure 24. ? Careproctus pycnosoma. Teeth from typeSynopsis.?Dorsal, 42; anal, 37; pectoral, 37; pyloric coeca, 14.Depth of body 5.5 in length without caudal. Disk well formed,flat, 3 in the head. Gill slit above the pectoral fin. Teeth shortand stout, strongly trilobed. Snout not projecting. Dorsal finwith a shallow notch. Pectoral with a very shallow notch, the Figure 25.?Careproctus cueilanus. Teeth from typenotch hardly evident; the lower lobe 2.5 in the head. Color duskygray; peritoneum pale. A small species, reaching a length of 46 mm.CAREPROCTUS CURILANUS Gilbert and BurkeCareprochis curilanus Gilbert and Btjrke 19126, p. 373.Type.?Male, No. 73341, U.S.N.M.; Albatross Station 4803, offSimushir Island, KurU Group, Japan; depth 229 fathoms. Length72 mm. 110 BULLETIN 150, UNITED STATES NATIONAL MUSEUMDistribution.?Off Simushir Island, Japan; Albatross Station 4803;depth 229 fathoms. Two specimens examined.Relationships.? C. curilanus closely resembles C. edenes in thedepressed head, projecting snout, and distinct type of dentition.C. curilanus appears to differ from C edenes in the shorter body,larger disk and larger gill slit. In addition, C edenes appears to bedistinct in having the dorsal notch more or less developed, the pyloriccoeca absent, a smaller number of pectoral rays, and the vent fartherfrom the disk. The two species appears to be separated geographi-cally, but there is no evident barrier between. The differences wehave enumerated may fall within the range of individual variationand the two species be found to be identical though at present theevidence we have indicates that they are distinct.Synopsis.?Dorsal 49; anal 43; pectoral 33-34; pyloric coeca 6.Depth of body 6.7-6.8 in length without caudal. Disk 2.7-2.8 inhead. Teeth strongly trilobed. Snout depressed and distinctlyprojecting. Gill slit extending down in front of from one to threepectoral rays. Dorsal fin unnotched; the anterior rays projecting,about equal in length. Vent close to the disk. Color grayish.Reaching a length of 74 mm.CAREPROCTUS ECTENES GilbertCareprodus edenes Gilbert, 1896 p. 442.Careprodus (Allurus) edenes Jordan and Evermann, 1898, p. 2136.Type.?^o. 48618, U.S.N.M.; Bering Sea, Albatross Station 3331;depth 350 fathoms. Length 70 mm.Distribution.?Bering Sea, Albatross Stations 3331 and 3785;depth 270 to 350 fathoms. Five specimens examined.Relationships.? C. edenes closely resembles C. curilanus. Adiscussion of the similarities and differences between these two speciesis given in the discription of C curilanus. In the presence of thedorsal notch and the trilobed dentition C. ectenes resembles the mostprimitive members of the genus. In the reduction of the pyloriccoeca, the distinct dentition, the depressed head, and projectingsnout the species has diverged widely from the primitive type. Ifthe coeca were always absent the species would be sufficiently distinctto place in a separate genus.Description oj type.?Body slender, elongate, firm, depressed;depth 7.7 in length; width through base of pectoral greater thandepth. Head slender, depressed, 4.6 in the length, wider than deep;occiput not swollen; profile low, gradual. Mouth small, with littlelateral cleft under the projecting snout; maxillary nearly reachingvertical from middle of pupil. Teeth as in C. curilanus, stout, inbroad bands, distinct in that the margin of the lobes are flangelike,arranged in oblique rows; the inward diverging rows prominent, eas- REVISION OF THE FISH FAMILY LIPARIDAE 111ily counted. Snout depressed, bluntly rounded, overlapping mouthfor half the pupil, resembling the snout of G. simus. Nostrils single,far apart, with a raised rim. Eye large 3.2 in head. Gill slit small,5.2 in head, above the pectoral. Pores normal, not enlarged, two abovegill slit, the posterior one minute. No prickles. Pyloric coeca absent.Dorsal notched; the first ray elongate, about 3.2 in head; the tipprojecting up above the skin, the two or three succeeding rays short- FlGURE 26.?CAREPR0CTU3 ECTENES. TeETH FROM TYPEened, the following rays increasing gradually in length. Anterioranal rays with free tips. Caudal slender, truncate, of about eightrays; dorsal and anal connections to the coudal apparently not greaterthan basal third of caudal. Pectoral notched; the lower lobe of sevenrays graduated in length, reaching vent or beyond; upper edge ofpectoral on a level with lower part of the eye. Disk small, equal tothe eye or smaller, posterior edge under base of upper pectoral ray;the flap broad posteriorly. Ventdistant from disk by three-fourths diameter of disk, behinda vertical from the gill slit.Color nearly a uniform duskybrown, lighter on snout andhelly; mouth, gill cavity, andstomach white, peritoneumwhite, dotted.Synopsis.?Dorsal 48; anal 44 pectoral 30-32; pyloric coeca 0-6.Depth of body 7.1 to 8 in length without caudal. Disk 3.2-4.6in the head. Teeth strongly trilobed. Snout depressed and distinctlyprojecting. Gill slit above the pectoral. Dorsal fin typically notched,sometimes the anterior rays equal in length. Vent distant fromdisk by three-fourths diameter of disk. Color dusky brown to gray-ish. Reaching a length of 87 mm.Bemarks.?The presence of pyloric coeca in aU the specimens of thisspecies is somewhat doubtful. In the specimens examined the Figure 27.?Careproctus ectenes. Side viewof teeth from specimen no. 64043, u.s.n.m. 112 BULLETIN 150, UNITED STATES NATIONAL MUSEUMpresence of coeca could be detected in only one specimen, in whichsix very short and small coeca were found adhering closely to thepylorus. In another specimen some filaments may have representedcoeca.In this species the dorsal fin is more or less distinctly notched and theanterior rays are not buried but project distinctly. The anterior dor-sal ray is typically longer than the two or three succeeding rays. Inone or two specimens the anterior ray is about equal to the succeedingrays and thus resembles the condition found in C. curilanus. In No.64043, U.S.N.M., larger specimen, the first dorsal ray is contained3.4 in the head, the third ray 5 ; in the smaller specimen the first dorsalray is contained 2.6 in the head, the second ray 3.3, and the fourth ray5.5, the following rays increasing in length.The specimens No. 64043 are lighter than the type, being grayishand resembling the coloration of C. curilanus.CAREPROCTUS SIMUS GilbertCareproclus simus Gilbert, 1896, p. 444.Type.?'No. 51688, U.S.N.M.; Albatross Station 3331, Bermg Sea;north of Unalaska; depth 350 fathoms. Length 71 + mm. (Caudalabsent.)Distribution.?Bering Sea, north of Unalaska, Albatross Station3331; depth 350 fathoms.Relationships.?In most respects C. simus agrees with C. mollis^differing in the projecting snout and larger eye. In general appear-ance, especially in the broadhead, C. simus resembles C.spectrum and related species.It resembles C. curilanus inthe projecting snout, possiblyin the condition of the ante-rior dorsal rays and in othercharacters. The dentitionof these two species is quiteFigure 28.?Careproctus simus. Teeth from distinct. C simUS is knOWn ^"^^^ from a single unsatisfactoryspecimen and we can not be certain of all the characteristics ofthe species.Description.?Dorsal 47+ ; anal 41 + ; pectoral 33; pyloric coeca20. Depth of body 1.3 in head; eye 3; disk 3.8; snout 3.1; inter-nostril 2.9.Body rather deep at union with head, tapering rapidly for a shortdistance, then more gradually until the tail becomes attenuate. Headheavy, broad, flattened on top, resembling C. spectrum; depth of headgreater than width of head; profile descending gradually to tip ofsnout and then retracting to the mouth. Mouth broad, the angle REVISION OF THE FISH FAMILY LIPARIDAE 113 reaching past vertical from front of eye. Teeth short, strong, trilobed,the lateral lobes weakly developed, forming a shoulder; teeth are ineight oblique rows in the half of the lower jaw, nine in the upper jaw;the rows widely spaced; outer teeth smaller. Snout heavy, broadlyrounded, projecting beyond upper lip for diameter of pupil. Nostrilin a low tube, barely raised above the skin, the posterior margin raisedinto a rounded flap. Eye large, prominent, black, pupil oval, hori-zontal. Gill slit small, 4 in the head, extending down in front of threepectoral rays. Two suprabranchial pores; other pores normal. Noprickles. Pyloric coeca 20, about equal to the eye.Anterior dorsal rays apparently stiff, not filamentous and undulatingbeneath the skin; first ray elongate, as long as the second ray; theanterior rays little shortened; not buried in pseudotissue. Caudalabsent. Pectoral notched; the lower lobe of six partly free, thickened,and exserted rays, reaching a little past vent, 2.2 in head. Disk welldeveloped; distance from tip of lower jaw to disk 3 in head. Ventclose to disk, in front of gill opening. Figure 29.?Carepeoctus mollis. Type. A typical species of the genus in shape and generalAPPEARANCEColoration: Discolored, apparently dusky with the mouth and gillcavity pale; the peritoneum discolored, possibly pale.Synopsis.?Dorsal 47+; anal 41 + ; pectoral 33; pyloric coeca 20;disk not cupped, 3.8 in the head. Gill slit extending down in front ofthree pectoral rays. Teeth stout, the lateral lobes weakly developed,forming a shoulder. Snout distinctly projecting. Anterior dorsal rays*rather stiff, little shortened.CAREPROCTUS MOLLIS Gilbert and BurkeCareprodus mollis Gilbert and Burke 1912a, p. 77.Type.?'No. 74383, U.S.N.M. ; Albatross Station 4784, south of AttuIsland, Bering Sea; depth 135 fathoms. Length 85 mm.Distribution.?Albatross Stations 4781 and 4784, Bering Sea; depth135-482 fathoms. Five specimens examined.Relationship.? C. mollis appears to be related to C. bowersianus. Itdiffers from the latter species in the smaller eye, humped body, andlarger disk. For a comparison with C. simus see description of thelatter species. 114 BULLETIN 150, UNITED STATES NATIONAL MUSEUMSynopsis.?Dorsal 51; anal 47; pectoral 35; pyloric coeca 8. Disk:flat, 3.5 in the head. Gill slit extending down in front of threepectoral rays. Teeth stout, distinctly trilobed. Anterior mandibularpores united. Snout not pro-jecting. Eye 4.8 in the head.Occiput swollen and the bodyhumped at the nape. Colorpale. Our specimens reaching aFigure 30?Careproctus molus. Teeth tipFROM TYPE length 01 85 mm.CAREPROCTUS SINENSIS Gilbert and BurkeCareproctus sinensis Gilbert and Burke 1912&, p. 371.Type.?'No. 73339, U.S.N.M.; Albatross Station 4813, off SadoIsland, Japan Sea; depth 200 fathoms. Length 72 mm.Distribution.?Japan Sea, off Sado Island, Albatross Station 4813;.depth 200 fathoms. See remarks. Three specimens examined.Relationships.? C. sinensis resembles C. segaliensis in many respects.It can be distinguished from the latter species by the larger number ofpectoral rays, larger disk, and minor differences.Synopsis.?Dorsal 47?; anal 47; pectoral 33; pyloric coeca 21.Depth of body 4 in length without caudal; depth of head 4.3 ; width ofhead 7; head and bodygreatly compressed, resem-bling CrystallicJithys mira-bilis. Snout projecting asin the latter species. Gillslit above the pectoral fin.Teeth stout, weakly tri-lobed, the lateral lobesforming a shoulder One figure si.?careproctus sinensis, teeth from speci- , , . , * MEN No. 53812, U.S.N.M.suprabranchial pore; man-dibular pores enlarged, the anterior pair united. Anterior dorsalrays shortened, increasing rapidly in length. Disk normal, flat, 3.5in head. Lower lobe of pectoral short, 2.3 in the head, reachingsUghtly past vent. Vent close to disk. Color pale. Our largest spe-cimen 72 mm. in length.RemarTcs.?Specimens No. 53811 and No. 53812, from AlbatrossStations 2847 and 2851 , south of the Alaska Peninsula, differ from thetype in several respects. These specimens are in poor condition andprobably represent a new species. Body outline not abruptly descend-ing above pectoral, the dorsal outline gradual. Occiput not swollen.Eye small, the lower half silvery. Pyloric coeca 25. Anterior dorsalrays increasing gradually in length. Lower lobe of pectoral not reach-ing vent. Vent back of disk by diameter of disk. Two suprabranchialpores present, small. /] REVISION OF THE FISH FAMILY LIPARIDAE 115CAREPROCTUS SEGALIENSIS Gilbert and BurkeCareprodus segaliensis Gilbert and Burke, 19126, p. 367.Type.?1^0. 73336, U.S.N.M.; Albatross Station 5026, off CapePatience, Saghalin, Okhotsk Sea; depth 119 fathoms. Length 75 mm.Distribution.?Okhotsk Sea, off Saghahn, Albatross Station 5026;depth 119 fathoms.Relationships.? C. segaliensis bears a close resemblance to Crystal-licthys mirdbilis in the compressed head and projecting snout; in otherrespects it is quite distinct. See Careprodus sinensis for comparisonswith the latter species. Figure 32.?Careproctus segauensis. Teeth from typeSynopsis.?Dorsal 58; anal 52; pectoral 24; pyloric coeca 17.Depth 4.5 in length without caudal; depth of head 5.6; width of head8; head and body greatly compressed, gelatinous. Snout projectingas in Crystallichthys mirabilis. Gill slit above the pectoral fin. Teethstout, distinctly trilobed. One suprabranchial pore; pores on head inpits; anterior mandibular pores united. Anterior dorsal rays buriedin pseudotissue. Lower lobe of pectoral 2 in head. Disk small,cupped, 4.2 in head. Vent close to disk. Color pale, transparent.The type specimen the only one know^n, 75 mm. in length.CAREPROCTUS BOWERSIANUS Gilbert and BurkeCareprodus bowersianus Gilbert and Burke, 1912o, p. 76.Type.?^o. 74382, U.S.N.M. ; Albatross Station 4772, Bowers Bank,Bering Sea; depth 344 fathoms. Length 100 mm.Distribution.?Bowers Bank, Bering Sea. Albatross Stations 4771and 4772; depth 344 to 426 fathoms. Two specimens examined.Relationships.? C. bowersianus does not closely resemble any knownspecies. It is distinguished by a large, prominent eye and deep, stoutbody.Synopsis.?Dorsal 52-53; anal 46-48; pectoral 36-37; depth 4.7-5in length without caudal; depth of head 4.5-5; width of head 5.5-6.3;eye 3.7-4 in head. Snout not projecting. Gill slit above the pectoralfin or extending down in front of the upper ray. Teeth stout, strongly 116 BULLETIN 150, UNITED STATES NATIONAL MUSEUMtrilobed. Two suprabranchial pores; rudimentary pores on the sidesof the body. Anterior dorsal rays buried in tissue. Disk normal, 5 inOq Figure 33.?Cakeproctus bowersianus. Teeth from cotypethe head; disk to vent 5.4-7 in the head. Color pale; stomach blackish.Only two specimens known, the largest 113 mm. in length.CAREPROCTUS FALKLANDICA (Lonnberg)Liparis antarctica falklandica Lonnberg, 1905, p. 17.Careproctus falklandica Burke, 19126.Distribution.?Antarctic Ocean, Berkeley Sound, Falldand Islandsand Burchwood Bank; depth 16 to 150 meters. No specimensexamined.Relationships.?The single nostrU and the coloration indicate thatthe species belongs with Careproctus.Synopsis.?Pectoral 30 or 31. Depth of body about 3.7 in totallength. Disk about 2.2 in head. Gill slit probably above pectoral.Teeth strongly trilobed. Dorsal unnotched. Color pale.CAREPROCTUS PALLIDUS (Vaillant)Enantioliparis pallidus Vaillant, 1888, p. 22, pi. 4, fig. 3.Liparis pallidus Garman, 1892, p. 70.Careproctus pallidus Burke, 19126.Distribution.?Orange Bay, Tierra del Fuego; depth 28 meters.No specimen examined.KelationsJiips.?I have formerly (19126) given reasons for placingthis species with Careproctus.CAREPROCTUS ATTENUATES GUbert and BurkeCareproctus attenuatus Gilbert and Burke. 1912a, p. 79.Type.?No. 74386, U.S.N.M.; Bermg Sea, Albatross Station 4781;depth 482 fathoms.Distribution.?Bering Sea; Albatross Station 4781; depth 482 fath-oms. One specimen.RelationsTiips.?In many respects C. attenuatus resembles C.Tiomopterus. A discussion of the differences between the two speciesis given under C. Tiomopterus. The dorsal fin in C. attenuatus appearsto be slightly notched and in this respect resembles the dorsal fin inC. ectenes. REVISION OF THE FISH FAMILY LIPARIDAE 117Synopsis.?Dorsal 48; anal 40; pectoral 34; pyloric coeca about10. Eye 3.1 in the head; disk 3.7, not cupped. Gill slit apparently ;?i?L Figure 34.?Careproctus attenuatus. Type, showing greatly attenuated body, ap-proaching THE CONDITION FOUND IN RHINOLIPARISabove the pectoral fin. Teeth stout,strongly trilobed. Snout deep, notprojecting. Head depressed, nearlyas vdde as deep. Body slender,depth 5.9 in length without caudal.Dorsal fin apparently notched as inC. edenes; anterior dorsal ray elon-gate, longer than the two or threesucceeding rays. Lower lobe of thepectoral 1.5 in the head. Vent mid-way between disk and anal. Colorpale; peritoneum black. Only typespecimen known. Figure 35.?Careproctus attenuatus.Teeth from typeCAREPROCTUS KOMOPTERUS Gilbert and BurkeCareproctus homopterus Gilbert and Burke, 19126, p. 374.Type.?No. 73342, U.S.N.M.; Albatross Station 5029, southernpart of the Okhotsk Sea; depth 440 fathoms. Length 49 mm.Distribution.?Southern part of the Okhotsk Sea, Albatross Station5029; depth 440 fathoms. One specimen.RelationsJiips.? Careproctus homopterus resembles C. attenuatus inmany respects. It is distinguished from the latter species by theshape of the snout, the smaller eye, larger disk, the shorter first dorsalraj^, and in other characters.Synopsis.?Dorsal 55; anal 49; pectoral 32. Eye 3.7 in the head;disk 2.7, with ''a broad margin. Gill slit extending up from the baseof the upper pectoral ray. Teeth stout, strongly trilobed. Snout notso deep as in C. attenuatus, rising gradually from the mouth. Bodymoderately deep and compressed, depth 6 in length. Dorsal finunnotched; the anterior ray short, one-half the eye. Lower lobe ofthe pectoral 2.4 in the head. Vent separated from disk by less than91668?30 9 118 BULLETIN 150, UNITED STATES NATIONAL MUSEUMdiameter of disk. Color paie to dusky; dusky toward the caudal,peritoneum pale; stomach appearing black through the abdominalwall. Only the type specimen known.CAREPROCTUS RHODOMELAS Gilbert and BurkeCareprodus rhodomelas Gilbert and Burke, 19126, p. 365.Type.?M&le, No. 73334, U.S.N.M.; Albatross Station 4958, offthe Bungo Channel, Japan; depth 405 fathoms. Length 123 mm.Distribution.?Southeast coast of Japan; Albatross Stations 4958and 4980; depth 405-507 fathoms. Two specimens examined.Relationsliips.? C. rhodomelas is not closely related to any knownspecies. The heavy crushing teeth and the long lower pectoral lobedistinguish this species from all others with which it ma}^ be confused. FiGUEE 36.?Caeepeoctus homopteel's. Teeth feom ttpeSynopsis.?Dorsal 56; anal 48; pectoral 31; caudal 9; pyloriccoeca 12. Body slender; depth 5.5 in length without caudal; head5.4. Teeth coarsely trilobed, the lateral lobes nearly as large as thecentral lobe. Gill slit extending down in front of three pectoral rays.Anterior dorsal rays increasing gradually in length. Lower lobe ofthe pectoral elongate, reaching past the origin of the anal fin, four-thirds times the head. Disk small, triangular; 8.3 in the head.Color blackish; peritoneum and stomach black. Only two speci-mens known. CAREPROCTUS RANULA (Goode and Bean)Liparis ranula Goode and Bean, 1879, p. 46.Careprodus ramda Goode and Bean, 1895, p. 275, fig. 251. ? Jordan andEVERMANN, 1898, p. 2134.Careprodus reinhardi Garman, 1892, p. 78, (part, not of Kroyer). ? Jordan andEvERMANN, 1898, p. 2133, (part, after Garman).r7/2?e.?Female, No. 22310, U.S.N.M.; Speedwell Station 117, offHalifax Harbor; depth 52 fathoms. Length 56 mm. REVISION OF THE FISH FAMILY LIPARIDAE 119Distribution.?New England coast to Newfoundland; depth possibly1 1-365 fathoms. Fifteen specimens examined.Relationships.?'C. ranula belongs with the spectrum group of speciesin which the head is blunt and heavy, the interorbital region broad,the peritoneum silvery and the teeth rather strong but simple or thelobes faintly indicated.The three species of the North Atlantic, C. ranula (Goode andBean), C. reinhardi (Kroyer), and C. micropus (Giinther) closelyresemble each other. The validity ofall of these species has been questioned.By a number of writers, C. reinhardihas been regarded as identical withC. gelatinosus (Pallas). Garman(1892) refers C. ranula to the synon-omy of C. reinhardi. Lutken (1898)indicates his uncertainty concerningthe individuality of C. reinhardi and 6.micropus. I have examined but one of these species, C. ranula, andcan add but little toward the solution of the puzzle. I believe, how-ever, that the three species should be recognized until we can eliminatethem with certainity. From an examination of the descriptions andfigures C. ranula appears to differ from C. reinhardi in the larger eyeand shallower pectoral notch. C. micropus appears to differ fromC. reinhardi in the shallower pectoral notch and possibly in the smallernumber of fin rays. C. ranula and C. micropus agree in having a Figure 37.?Careproctus rhodomelas.Ventral view showing great re-duction OF VENTRAL DISK Figure 38.?Careproctus rhodomelas. Teeth from type shallow pectoral notch but appear to differ in the size of the eye andthe number of fin rays.Description.-?-The type is in such a poor state of preservation thatit will not admit of description. The following note is extractedfrom the original description.D.48; A. about 48; P. 15 + 12 (13). Length without caudal 52 mm.Depth 0.25; head 0.25; orbit 0.7; disk 10. Teeth villiform. Gill slitextends upon the upper part of the root of the pectoral. Pectoralwith 15 long rays and 12 or 13 short ones. Color uniform white. 120 BULLETIN 150, UNITED STATES NATIONAL MUSEUMDescription of No. 28812, U.S.N.M.; dorsal 53; anal 49; pectoral 28;pyloric coeca about 9. Head about 5.3 in length without caudal.Eye 3.4 in head; disk 3.2. Figure 39.?Careproctus ranula. Teeth from specimen No. 9556, Stanford UniversityZoological Museum, showing variation from typeBody distended with eggs, rather elongate and slender. Head low,broad, nearly as wide as deep; interorbital flat; occiput low; sides ofhead nearly vertical. Mouth broad; maxillary reaching posterior Figure 40.?Careproctus ranula. Teeth from typemargin of pupU. Teeth simple or the lateral lobes wealdy developed,rather slender and elongate, slightly recurved; the oblique rowsdifficult to count; about 10 or 12 in the half of each jaw; inner teeth KEVISION OF THE FISH FAMILY LIPARIDAE 121larger but not especially prominent. Snout low, broad. Nostril ina tube in front of eye. Eye moderate; pupU large, round. Gill slitabove the pectoral. No pricldes. Pyloric coeca 9 or more, 2.7 inthe head.Anterior dorsal rays free beneath fin membrane, increasing graduallyin length. Caudal of 8 or 10 slender rays, connected with its basalhalf to the anal. Pectoral with a shallow notch; the lower lobehardly differentiated, of six rays, reaching slightly past the disk, 1.5 inthe head; upper edge of pectoral on a level between eye and angle ofmouth. Disk moderate, flat. Vent close to the disk.Color pale; coeca, stomach, and peritoneum pale.Synopsis.? Dorsal 53; anal 49; pectoral 28; pyloric coeca about 9.Depth of body less than length of head. Eye 3.4 in head; disk 3.2.Teeth simple, rather slender and elongate, sometimes a light shouldernear the tip or the lateral lobes distinct. Gill slit above the pectoral.Pectoral with a shallow notch, the lower lobe hardly differentiated,reaching slightly past the disk, 1.5 in the head. Color pale, includingperitoneum and stomach.CAREPROCTUS REINHARDI (Kroyer)Liparis gelatinosus Reinhardt, 1842, p. 82, pi. 10.Liparis reinhardi Kroyer, 1862, p. 252. ? Collett, 1905, pi. 2, fig. 8. (part,confused with C. longipinnis) . Careproctus reinhardi Kroyer, 1862, p. 257. ? Collett, ISSO, p. 57, pi. 2, figs.15-16. ? Garman, 1892, p. 78, (part, confused with C. ranula).?Jordanand EvERMANN, 1898, p. 2133, (part, after Garman).Liparis gelatinosus GtJNTHER, 1887, p. 67, (part, not of Pallas).Distribution.?The types are from Greenland, Arctic Ocean;recorded from about Jan Mayen and Bear Island and off Arendaland in the Kara Sea; depth 263-658 fathoms.Relationships.? C. reinhardi appears to be closely related to C.ranula and C. micropus. For a comparison of these species seedescription of C. ranula.Synopsis.?Dorsal 54-55; anal 45-46; pectoral 32-33; depth ofbody less than length of head. Eye 5-6 in the head. Disk httlelarger than the eye. Teeth simple. Pectoral notched ; the lower lobeshorter than the head. Color reddish or whitish.CAREPROCTUS MICROPUS (Gunther)Liparis micropus Gxjnther, 1887, p. 66, pi. 12, fig. B. ? Lutken, 1898, p. 14.Careproctus micropus Garman, 1892, p. 72. Goode and Bean, 1895, p. 277.Distribution.?Giinther's specimens come from the Faroe Channel,depth 540 to 608 fathoms. Lutken records specimens from west ofGreenland and Iceland, north of Iceland, and south of the FaroeIslands. No specimens examined. 122 BULLETIN 150, UNITED STATES NATIONAL MUSEUMRelationshi'ps . ? C. micropus appears to be closely related to C.ranula and C. reinhardi. (See description of C. ranula.)Synopsis.?Dorsal 47; anal 41. Depth of body less than lengthof head. Eye 5 in head; disk 4. Teeth simple. Gill slit apparentlyabove base of pectoral. Lower lobe of pectoral shorter than head. Figure 41.?Careproctus reinhardi. Teeth from specimen No. 28812,U.S.N.M.CAREPROCTUS GEORGIANUS LonnbergCareproctus georgianus Lonnberg, 1895, p. 41, pi. 3. ? Burke, 19126,Distribution.?Antarctic Ocean, South Georgias, depth 195 m.No specimens examined.Relationships.-?In many respects C. georgianus resembles C.pJiasma and related species of the North Pacific.Synopsis.?Dorsal 45-52; anal 42-46. Pectoral about 30. Deptha little more than 4 in total length. Disk 1.9-2.1 in head. Gillslit above base of pectoral. Teeth simple. KEVISION OF THE FISH FAMILY LIPARIDAE 123CAREPROCTUS PHASMA GilbertCareprodus phasma Gilbert, 1896, p. 443. ? Jordan and Evermann, 1898,p. 2132.Type.?No. 48604, U.S.N.M.; Bering Sea, Albatross Station 3254;depth 46 fathoms. Length 89 mm.Distribution.?Southeastern Bering Sea, Albatross Stations 3254,3256, 3530; depth 46 to 59 fathoms; temperature 34.9? to 36.2? F.Three specimens examined.Relationships.? C. pTiasmaclosely resembles C. spectrumin the broad head and in gen-eral appearance. It can bedistinguished from the latterspecies by the larger disk ; theskin appears to be thicker andmore opaque. C. phasma maybe confined to shallower waterthan C. spectrum.Synopsis.?Dorsal 53; anal 45; pectoral 34; caudal 8; pyloriccoeca 21. Eye 4 in head. Body heavy, deep, debth 3.7 in lengthwithout caudal. Head heavy, 3.7 in length; interorbital broad andflat; cheeks vertical. Teeth simple, slender. Snout blunt. Gil-slit above pectoral fin. Anterior dorsal rays buried in pseudoltissue. Lower lobe of pectoral 1.5 in head. Disk well developed,larger than eye, 3.3 in the head. Color pale or white; peritoneumsilvery. This species is represented by three specimens, all of whichare less than 100 mm. in length. Figure 42. ? Careproctus phasma.TYPE Teeth from I\ a Figure 43.?Careproctus phasma. Teeth from specimen No. 53813, U.S.N.M.CAREPROCTUS SPECTRUM BeanCareproctus spectrum Bean, 1890, p. 40 (part). ? Jordan and Evermann, 1898,p. 2133 (part). ? Evermann and Goldsborough, 1907, p. 333 (part).T?/2;e.?No. 45363, U.S.N.M.; Albatross Station 2848, north of theShumagin Islands, Alaska; depth 110 fathoms; temperature 41? F.Length 97 mm.Distribution.?Gulf of Alaska, Albatross Stations 2848 and 4295;depth 92-110 fathoms. Four specimens examined.Relationships.? C. spectrum is closely related to C. melanurus. Itdiffers from the latter species in having less pigment in the mouth, 124 BULLETIN 150, UNITED STATES NATIONAL MUSEUM gill cavity, peritoneum, and the posterior part of the body; the dis-tribution is more northern, and the depth inhabited possibly a littlegreater.Description of type.?Dorsal 52; anal 47; pectoral 32; caudal 10;pyloric coeca 21, Depth 4.3 in length without caudal; head 4. Eye3.3 in head; disk 4.Body heavy anteriorly, deep, short, tapering rapidly to the caudal.Head heavy; interorbital region broad and flat, cheeks nearly ver-tical, mouth broad; maxillary reaching vertical from middle of pupil.Teeth simple or one or both lateral lobes faintly developed, moder-ately stout to slender, hardly recurved ; inwardly diverging rows veryoblique and difficult to count, 10 or 12 in the half of each jaw; out-wardly diverging rows more distinct, close set. Snout short, about4 in head, broad, not projecting bej^ond mouth. Nostril in a shorttube. Eye large, prominent; pupil oval, horizontal. Gill slit smallabout 3.3 in head, extending up from the base of the upper pectoral Figure 44.?Careproctus spectrum. Teeth from type ray. Two suprabranchial pores present. Prickles absent. Pyloriccoeca 21, 2.4 in head, on the left side.Dorsal rays increasing gradually in length, the anterior rays ratherstiff, their connection Avith the skin uncertain. Caudal of 10 slenderrays, mutilated, connected for half its length to the anal. Pectoralnotched; the upper lobe of 25 rays, reaching anal; lower lobe ofeight slender, graduated rays, coiled at the tips, free nearlj^ to the base,reaching anal, nearly as long as the head. Disk moderate, the mar-gin rather thick, folded over posteriorly, the disk somewhat trian-gular, a little broader than long; distance from tip of lower jaw todisk 9.9 in length without caudal; front of disk just behind verticalfrom pupil. Vent next to disk.Coloration: Skin transparent, lax; muscles flesh colored; mouth,gill cavity, and internal organs pale; peritoneum silvery, withoutdots; abdomen silvery.Synopsis.? Dorsal 52; anal 47; pectoral 32; pyloric coeca 21.Depth 4.3 in length without caudal; head 4. E3'e 3.3 in head; disk REVISION OF THE FISH FAMILY LIPARIDAE 1254; smaller than eye. Body heavy, deep, short. Head heavy; in-terorbital region broad and flat. Teeth simple or with a slightshoulder. Snout blunt, not projecting. Gill slit extending up frombase of the upper pectoral ray. Anterior dorsal rays increasinggradually in length. Lower lobe of the pectoral reaching the originof the anal fin, about equal to the head. Disk somewhat triangular.Vent close to the disk. Color pale, little or no dark pigment any-where. Four specimens known, reaching a length of about 100 mm.CAREPROCTUS MELANURUS GHbertCareproctus mclanurus Gilbert, 1890, p. 56. ? Jordan and Evermann, 1898,p. 2135.Type? Male, No. 44285, U.S.N.M.; Albatross Station 2925, offsouthern California; depth 339 fathoms. Length 179+ mm.(Caudal absent.) aFigure 45.?Caheproctus melanurus. Teeth from typeDistribution.?British Columbia to southern California, AlbatrossStations 2860, 2892, 2925, 3112, and 3186; depth 284 to 876 fathoms;temperature 36.5? to 44.1? F. Eleven specimens examined.Relationships.?See description of C. spectrum.Synopsis.?Dorsal 54; anal 47; pectoral 31; pyloric coeca 20-27.Depth 4.2-4.5 in length without caudal; head 4.3-4.5. 'Eye 3.5-4 inhead; disk 6.4-6.9. Body rather deep and compressed. Headheavy; interorbital broad and flat. Teeth rather stout, simple orwith the lateral lobes faintly developed. Snout blunt, not projecting.Gill slit sbove the pectoral fin. Disk small, less than the eye, cupped.Lower lobe of the pectoral 1.3 in the head, reaching more than halfway to the orgin of the anal fin. Color pinkish or whitish ; pertioneummouth, gill cavity, caudal, and neighboring portions of dorsal and analand the posterior surface of the pectoral black or dusky; posteriorpart of the body more or less dusky; peritoneum rarely whitish withscattered black dots. Eleven specimens, reaching a length of about200 mm. 126 BULLETIN 150, UNITED STATES NATIONAL MUSEUMRemarks.?Note on specimen No. 604, U. S. B. F. Head not soflat as in the type; occiput rather deep. Caudal connected for fullyhalf its length to the dorsal. Pectoral low, the upper edge a httleabove the angle of the mouth; the lower lobe of graduated, elongaterays, reaching more than half way to anal, 1.3 in the head; the threeanterior rays free to the base, the remainder half free. Gill slit abovethe pectoral. Disk less than eye. Color whitish; posterior part ofbody, caudal, and neighboring parts of dorsal and anal dusky; mouth,gill cavity and peritoneum dusky; stomach white; inner surface ofpectoral dusk3^ In some specimens the peritoneum is light anddotted. CAREPROCTUS FURCELLUS Gilbert and BurkeCareprodus furcellus Gilbert and Burke, 1912a, p. 80.Type.?'No. 74387, U.S.N.M.; Albatross Station 4781, Bering Sea;depth 482 fathoms. Length 129 mm. Figure 46.?Careproctus furcellus. Teeth from cotypeDistribution.?Bering Sea, Albatross Station 4781; depth 482 fath-oms. Two specimens examined.Relationships.?See description of C. cypselurus. C.furcellus closelyresembled C. melanurus, differing in the larger disk, the shorter,lower pectoral lobe, and in other characters.Synopsis.?Dorsal 62; anal 57; pectoral 36; pyloric coeca 46.Depth 4 in length without caudal; head 3.7. Eye 3.7 in head; disk4. Body shorter than in C. cypselurus, the posterior part not so at-tenuate. Head broad, interorbital wide, flat. Teeth rather elongateand slender, the lateral lobes forming a shoulder or absent. Gill slitabove the pectoral fin. Caudal slightly forked. Pectoral fin un-notched. Disk weH developed, smaller than the eye. Color reddishor pale, posteriorly the body and fins black; gill cavity, abdomen andthe pectoral dusky; peritoneum black. Only two specimens known,reaching a length of 128 mm. or more. REVISION OF THE FISH FAMILY LIPARIDAE 127CAREPROCTUS CYPSELURUS (Jordan and Gilbert)Prognurus cypselurus Jordan and Gilbert, 1898, in Jordan and Evermann,1898, p. 2866; 1899, p. 478, pi. 77. ? Evermann and Goldsborough, 1907,p. 333, pi. 20.Careprodus cyptehirus Gilbert and Burke 19125, p. 362.Type.?Male, No. 48232, U.S.N.M.; Albatross Station 3634, northof Unalaska, Bering Sea; depth 664 fathoms. Length 211 mm. ''S-' Figure 47.?Careproctus cypselurus, showing forked caudalDistrihution.?North Pacific off the coast of Washington, BeringSea, and the Okhotsk Sea; Albatross Stations 3074, 3634, 4797, 5015;depth 510-887 fathoms. Four specimens examined.BelationsMps.? C. cypselurus bears a close resemblance to C. mela-nurus. The forked caudal of C. cypselurus serves to distinguish thetwo species. C. cypselurus bears a closer resemblance to C. furcellus.It can be distinguished from the latter species by the more elongatebody; the eye and disk appear to be smaller, the gill slit larger, thenumber of coeca less, and the pectoral slightly notched. Figure 48.?Careproctus cypselurus. Teeth from a specimen-taken OFF KamchatkaSynopsis.?Dorsal 58; anal 54; pectoral 33; pyloric coeca 29-32.Depth 4.4-5.2 in length without caudal; head 4.4-4.7. Eye 4-4.7 inhead; disk smaller, 4.8-6.1. Body moderately heavy anteriorly, at-tenuate posteriorly. Head broad and heavy. Teeth rather shortand stout, simple or a few with the lateral lobes faintly indicated.Gill slit above pectoral. Caudal deeply forked. Pectoral fin with avery shallow notch; the lower lobe 1.8 in the head. Disk triangular,the posterior margin folded over. Anterior part of body and headpale to dusky; posterior part of body black; mouth and gill cavitydusky; peritoneum blackish. Four specimens known, reaching alength of 260 mm. 128 BULLETIN 150, UNITED STATES NATIONAL MUSEUMRemarks.?Note from a specimen from Albatross Station 5015,Okhotsk Sea. Depth 4.4; head 4.7. Disk 6.1 in head. Teeth slen-der, a few with weak cups. Pyloric coeca 32, 2.3 in the head. Disktriangular; the flap curled over the center; tip of mandible to disk 3in head. Color purplish black, darker posteriorly; margin of pectoralblack; mouth dusky. In life, greater part of head, body, and finsdeep purplish indigo, the anterior part of the sides almost clear red,with little blue; head again becoming deeper blue, but with more redthan the posterior part of the trunk; basal portion of upper pectoralrays also with more reddish.The specimen from Station 4797 has the anterior part of the bod}'"pale and the posterior part black. Ninth and tenth pectoral raysslightly shortened.CAREPROCTUS ABBREVIATUS, new speciesType.?Female, No. 3082, Stanford University Zoological Museum;south of the Alaska Peninsula, Albatross Station 3338; depth 625fathoms. Length 53 mm. Figure 49.?C.vreproctus abbeeviatus. Teeth from typeDistribution.?Known only from type locality.RelationsMps.?Careprodus abbreviatus is readily distinguishedfrom all the species with which it might be confused by the smallnumber of fin rays. The type specimen was labeled 0. colletti andresembles this species in the shape of the disk and general appearance.Description of type.?Dorsal 39; anal 32; pectoral 21. Depth 4.3in length without caudal; head 4.3 Eye 3.4 in head, disk 3.8.Body short, moderately deep and compressed, tapering to thecaudal fin, not attenuate. Head deep, much compressed; occiputhigh; profile rather steep; cheeks vertical; mouth terminal; maxillaryreaching beneath the pupil. Teeth simple, conical, short, in a narrowoblique band. Snout truncate, deep. Nostril in a short tube directlyin front of the eye. Eye moderate, black; pupil round. Gill slitsmall, 3.8 in head, above the pectoral. No prickles.Dorsal normal; the anterior rays not buried in thick tissue, notprojecting. Caudal mutilated. Pectoral rays widely spaced; theupper lobe of 15 rays, not reaching the anal fin; the notch shallow; thelower lobe short reaching the vent, 2 in the head, the rays about half REVISION OF THE FISH FAMILY LIPARIDAE 129free. Disk very deeply cupped; the central portion hidden in a pitformed by the thick, stiff flap; tip of mandible to disk 7.9 in the lengthwithout caudal. Vent a short distance behind the disk.Coloration: Body pale; skin thin and lax; flesh becoming brownishtoward the caudal; mouth and gill cavity pale; peritoneum andstomach nearly black.Synopsis.?Dorsal 39; anal 32; pectoral 21. Disk deeply cupped,3.8 in head. Gill slit above the base of the pectoral. Teeth simple,conical, short. Body pale, becoming brownish toward the caudal;peritoneum nearly black. One specimen, 53 mm. in length.CAREPROCTUS BATHYCOETUS Gilbert and BurkeCareprodus baihycoetus Gilbert and Burke, 19126, p. 368.Type.?Fem&le, No. 73337, U.S.N.M.; Albatross Station 5030,southern part of the Okhotsk Sea; depth 1,800 fathoms. Length 181mm. Figure 50.?Careproctus bathycoetos. Teeth from typeDistribution.?Southern portion of the Oldiotsk Sea; AlbatrossStation 5030; depth 1,800 fathoms. One specimen known.Relationships.? C. hathycoetus is readily distinguished by the re-duced number of pectoral rays, the coloration, and the attenuatebody.Synopsis.?Dorsal 61; anal 55; pectoral 21; caudal 6; pyloric coeca9. Depth 5.7 in length without caudal; head 5. Eye 4.8 in head;disk 6.9. Body very attenuate posteriorly; the posterior fourth of thebody not as deep as the diameter of the eye. Teeth simple, ratherstout, recurved. Gill slit above the base of the pectoral fin. Anteriormandibular pores separated by the diameter of the pupil. Lowerlobe of the pectoral 1.5 in the head. Disk small, deeply cupped.Color black; lips of the pores and the ventral surface of the disk black;peritoneum black; stomach pale. 130 BULLETIN 150, UNITED STATES NATIONAL MUSEUMRemarks.?In the reduction of the pyloric coeca and pectoral raysand the slender body this species most closely resembles the typicalParaliparis. It could not, however, have given rise to the lattergroup of species because the primitive Paraliparis had trilobed teeth.CAREPROCTUS LONGIFILIS GarmanCareprodus longifilis Garman, 1892, p. 9; 1899, p. 115, pis. 27-29.Type.?No. 28703, M. C. Z.; Albatross Station 3374, Pacific Oceanoff Panama; depth 1,823 fathoms.Distribution.?Pacific Ocean, off Panama, Albatross Station 3374;depth 1,823 fathoms. One specimen known.Relationships.? C. longifilis is not closely allied to any knownspecies. It is readily distinguished from the other species of the genusby the extremely elongate upper and lower pectoral lobes and therudimentary condition of the middle pectoral rays. C. longifilisparallels certain species of Paraliparis in the reduction in length of themiddle pectoral rays. Figure 51.?Careproctus longitius. Teeth from typeDescription of type.?The type is mutilated beyond description.The following notes were taken from what remains.Body tapering rapidly to the caudal. Head broad and heavy as inC. melanurus; interorbital flattened. Mouth broad; maxillaryreaching vertical from pupil. Teeth rather stout and blunt, slightlyrecurved, in narrow bands; outer teeth smaller; the lateral lobesfaintly indicated. Snout broadly rounded; upper jaw slightly longerthan the lower. Nostril in a short tube. Eye small, black, about4.7 in head. Gill slit apparently above the pectoral fin. Poresapparently normal; the upper pores on the snout nearly on a line withthe nostrils.Caudal well developed, of 8 or 10 slender rays. Pectoral findivided to the base, thus differing from all other species of thegenus; the two lobes connected by four widely spaced rudimentaryrays; the upper pectoral ray extremely elongate, half the body lengthwithout caudal; the lower lobe of four elongate free rays, considerablylonger than the head, reaching past the front of the anal and beyond REVISION OF THE FISH FAMILY LIPARIDAE 131the middle of the upper pectoral ray. Disk small, cupped; aboutequal to the eye, 4.5 in the head; the flap narrow, suspended as inC. colletti; anterior edge of the disk under the posterior margin ofthe eye. Vent distant a little less than the diameter of the disk.Color black; peritoneum black; stomach absent.The following note is extracted from the orignal description.Dorsal 54; A. 49; P. 13+ 4+ 4; C. 9. Head less than one-fifth the totallength. Gill opening narrow, as wide as the eye, above the pectoral fin. Dorsaland anal overlapping the caudal for more than half its length. Longest pectoralray at the upper edge of the fin, equal to four-tenths the total length of the speci-men; the upper lobe of 13 rays, graduated in length, connected with the lowerlobe by four shorter, widely separated rays; the lower lobe contains four separatedra5's, of which the third and fourth, provided with long filaments, are nearlythree-tenths of the total length; the rays are more rigid than those of the speciesof Paraliparis and the ends are slender and flexible. Intense black, uniformover head, body, and fins. Total length 3M inches.Synopsis.?Dorsal 54; anal 49; pectoral 21; caudal 9. Headheavy and broad, less than one-fifth the total length. Eye about4.7 in the head; disk 4.5, cupped. Gill slit above the base of thepectoral fin. Pectoral fin divided to the base; the middle raysrudimentary; the upper lobe equal to half the length of the bodywithout the caudal; the lower lobe about one-third the body length.Color black; peritoneum black. One specimen, length 3^ inches.CAREPROCTU3 OVIGERUM (Gilbert)Bathyphasma ovigerum Gilbert, 1896, p. 448. ? Jordan and Evermann, 1898, p.2128, fig. 767.Type.?Male, No. 48622, U.S.N.M.; Albatross Station 3342, offBritish Columbia; depth 1,588 fathoms; temperature 35.3? F.Length 318 mm.Distribution.?Off British Columbia, Albatross Station 3342; depth1,588 fathoms. One specimen known.Relationships.? C. ovigerum does not appear to be closely allied toany Itnown species. It is distinguished by the simple, slender,lanceolate teeth, the heavy body, the gill slit extending down infront of the pectoral fin, and the large disk. In the width of thehead, the size of the eye, the large disk, and the shape of the bodythis species bears a close resemblance to the typical species ofLiparis. The simple, slender, teeth, the single nostril, the anteriorposition of the disk, and the coloration indicate a close kinship withCareproctus with which we place the species. Gilbert, (1896, p. 448),proposed the genus Bathyphasma to include this species and basedhis genius upon the simple, lanceolate teeth which he believed tohave been derived along a different line of development from thatwhich gave rise to the simple teeth of Careproctus. At present weare not justified in recognizing the genus Bathyphasma as the denti- 132 BULLETIN 150, UNITED STATES NATIONAL MUSEUMtion of ovigerum grades imperceptably into that of a number ofspecies of Careprodus. The genus Careprodus, as at present recog-nized, is probably a polyphyletic one but with our present knowl-edge we certainly are not justified in dividing it.Description of type.?Dorsal 43; anal 34; pectoral 34; caudal 10;pyloric coeca 19. Depth 3.7 in length without caudal; head 3.5.Eye 6 in head; disk 2.9.Body rather deep at the nape; moderately elongate and com-pressed, tapering gradually to the caudal. Head wide, heavy;width nearly equal to the depth; cheeks swollen; profile concave overeye, rising rapidly at the occiput; occiput slightly swollen; inter-orbital wide; internostril 3.7 in the head. Mouth broad; lateralcleft considerable; angle of mouth reaching vertical from front of Figure 52.?Careproctus ovigerum. Teeth from type eye; maxillary reaching vertical from posterior margin of eye. Teethin broad bands, simple, slender; oblique rows widely spaced, about15 in the half of each jaw; inner teeth lanceolate, distinctly recurved;outer teeth much smaller. Snout short, comparatively deep, abrupt,not projecting beyond mouth; upper jaw slightly overlapping lower.Nostril single, apparently in a short tube. Eye moderate, notprominent; pupil large, round. Gill slit wide, 2.4 in the head, extend-ing down in front of eight pectoral ra3^s. The number of pores onthe head uncertain as the skin is in poor condition; one supra-branchial pore detected. No prickles. Pyloric coeca 19, on theright side.Origin of dorsal above the base of the pectoral fin ; some of the anter-ior rays unsegmented. Caudal broad, connected for two-fifths itslength to the anal; the connections between the three vertical finsabrupt, not notched. Pectoral notched; the lower lobe reaching KEVISION OF THE FISH FAMILY LIPAKIDAE 133 nearly to the vent, 1.8 in head. Disk large; the flap broad but notequal to the center of the disk; disk well forward under the head;distance from tip of lower jaw to disk 7.2 in length without caudal.Vent separated from the disk by two-thirds diameter of disk, twice asfar from the anal fin.Color pale, dusted with light brown; abdomen dusky; peritoneumdusky, possibly silvery in life.Synopsis.?Dorsal 43; anal 34; pectoral 34; caudal 10; pyloriccoeca 19; body and head heavy, somewhat resembling that of Liparisagassizii; depth 3.7 in length without caudal; head 3.5. Eye 6in head; disk 2.9. Teeth simple, lanceolate. Gill sht extendingdown in front of eight pectoral rays. Dorsal fin unnotched. Pec-toral fin notched; the lower lobe reaching nearly to the vent, 1.8in the head. Disk large, well forward under the head. Color pale, Figure 53.?Careproctus opisthotkemus. Teeth from typedusted Avith light brown; peritoneum dusky, possibly discolored.One specimen, 318 mm. in length.CAREPROCTUS OPISTHOTREMUS Gilbert and BurkeCareproctus opisthotremus Gilbert and Burke, 1912a, p. 78.Type.?^o. 74385, U.S.N.M.; Alhatross Station 4780, among theAleutian Islands, Bering Sea; depth 1,046 fathoms. Length 50 mm.Distribution.?Among the Aleutian Islands, Alhatross Station 4780;depth 1,046 fathoms. One specimen known.Relationships.? C. opistliotremus does aot appear to be closelyallied to any known species. It is distinguished by the slender,depressed body, the position of the vent, and the gill slit extendingdown in front of the pectoral fin.Synopsis.?Dorsal 46?; anal 36?; pectoral 32; pyloric coeca 12.Body slender, elongate, head depressed; depth of body 6 in lengthwithout caudal; head 3.5. Eye 4.7 in head; disk 3.4, flat. Teethsimple, or few of the larger teeth with distinct lateral lobes. Gill slitextending down in front of five pectoral rays. Pectoral fin notched;91668?3C 10 134 BULLETIN 150, UNITED STATES NATIONAL MUSEUMthe lower lobe 2 in head, reaching midway between disk and vent.Vent nearer the anal fin than the disk; disk to vent 2.4 in the head.Color probably pale. One specimen, 50 mm. in length.CAREPROCTUS LONGIPINNIS BurkeLiparis reinhardi Lutken, 1898, p. 14, pi. 3, fig. 3.3a, (part, not of Kroyer). ? CoLLETT, 1905, p. 97, pi. 2, fig. 7, (part, not of Kroyer).Careproclus longipinnis Burke, 19126, p. 509.Distribution.?Arctic Ocean, Faroe Islands to Beeren Island.Type from "Ingolf Station 139, north of the Faroe Islands; depth702 danish fathoms." Figured by Liitkin as L. reinhardi in "DanishIngolf Expedition, vol. 2, pt. 1, The Ichthyological Results, pi. 3,figs. 3-3a." Figure 54.?Cabepboctus pellucidus. Teeth fbom typeRelationships.? C. longipinnis differs from all the other species ofthe genus from the same region in the deeper body, the longer lowerlobe of the pectoral, and the more deeply cupped disk.Synopsis.?Dorsal 53; anal 44; pectoral 30-31; caudal 10. Bodygelatinous, deep; depth of body greater than length of head. Head4.9 in total length. Eye 5 in head. Disk 4.6. Gill slit above thebase of the pectoral fin. Pectoral fin unnotched; lower lobe elongate,exceeding the length if the head by a little more than the diameter ofthe eye. Disk small, cupped and triangular; posterior margin foldedover. Vent close to disk.CAREPROCTUS PELLUCIDUS Gilbert and BurkeCareproclus pellucidus Gilbert and Burke, 19126, p. 366.Ty^pe.?Female, No. 73335, U.S.N.M.; Albatross Station 5048,off Kinka San, Japan; depth 129 fathoms. Length 128 mm.Distribution.?Northeast coast of Hondo, Japan, Albatross Stations5048 and 5049; depth 129-182 fathoms. Seven specimens examined,reaching a length of 128 mm. REVISION OF THE FISH FAMILY LIPARIDAE 135 Figure 55.?Careproctus acanthodes. Teeth from type Relationships.-? C. pellucidus bears a close resemblance to C. trachy-soma. It has a larger eye, the gill slit extending more distinctly-down in front of the pectoral fin, the fin rays different in number, theperitoneum undotted, and no prickles. It seems unlikely that thisspecies represents spec-imens of C. trachysomain which the pricldesare absent.Synopsis.?Dorsal \ \ jy I f\54-55; anal 48-49; pec- \ \ /V [ Gtoral 33-35; caudal I i // 11 /7-8; pyloric coeca 18-22. Depth 3.5 in lengthwithout caudal; head3.9. Eye 3.7 in the head; disk 7; lower lobe of the pectoral fin 1.2-1.4.Body short, deep and compressed. Teeth simple, and slender. Gillslit extending down in front of five or six pectoral rays. No prickles.Disk small, cupped, somewhat triangular. Color pale; peritoneumundotted. CAREPROCTUS ACANTHODESGilbert and BurkeCareproctus acanthodes Gil-bert AND Burke, 19126,p. 363.Type.?Ferasile, No.73332, V.S.l^.M.; Alba-tross Station 4997, Gulfof Tartary; depth 318fathoms; temperature32.8? F. Length 89 mm.Distribution.?Gulf ofTartary, Albatross Sta-tion 4997; depth 318fathoms. Five speci-mens examined, reach-ing a length of 89 mm.Relationships .? C.acanthodes closely resem-bles!C.^ras^rwws and C. spectrum; differing from C. rastrinus in theshape of the head, larger eye, smaller gill slit?, the silvery coloring ofthe eye?, and the smaller number of fin rays; differing from C. spectrumin the presence of cactuslike prickles and the much shorter, lower pec-toral lobe.Syiiopsis.?Dorsal 52-53; anal 45-46; pectoral 33-34; caudal 8;pyloric coeca 19. Depth 3.6-3.7 in length without caudal; head 3.5. Figure 56.?Careproctus acanthodes. Prickles from type 136 BULLETIN 150, UNITED STATES NATIONAL MUSEUMEye 3.5-4 in head; disk 4.5-4.8; lower lobe of the pectoral fin 1.8-2.Body rather deep and compressed. Teeth simple. Eye moderate,the lower half silvery. Gill slit either above the pectoral fin orextending down in front of the upper ray. Cactuslike pricldespresent. Lower pectoral lobe short, about 2 in the head, reaching Figure 57.?Careproctus rastrinus. Type. Showing the great development of the headREGION characteristic OF SOME SPECIES OF THE GENUS little past the vent. Disk small, cupped, somewhat triangular.Body translucent, dotted with brown; peritoneum silvery, undotted.CAREPROCTUS RASTRINUS Gilbert and BurkeCareproctus raslrinus Gilbert and Burke, 19126, p 362.Type.?Female , No. 73331, U.S.N.M.; Albatross Station 5026,southern part of Okhotsk Sea; depth 119 fathoms. Length 280 mm.Distribution.?Southern part of the Ok- ^.hotsk Sea, off the coast of Sakhalin. Alba-tross Stations 5021 and 5026; depth 73 and119 fathoms. Three specimens examined.RelationsTiips.? C. rastrinus agrees with C.acantJiodes and C, trachysoma in the presenceof cactuslike prickles and in most other re-spects. These three species inhabit neigh-boring or overlapping regions. (See descrip-tions of G. acanthodes and C. trachysoma).Description of type.?Eye moderate, not especially prominent,black; pupil small, round. Gill slit 3.4 in the head, extending downin front of four pectoral rays. Suprabranchial pores 2, close together;rudimentary pores on the top of the head and along the median lineof the sides; upper snout pores nearly between the nostrils; otherpores normal. Cactuslike prickles on the body and fins, morescattered toward the caudal, none on the chin and lips; these prickles Figure 58.?Careproctus ras-trinus. Teeth from type REVISION OF THE FISH FAMILY LIPAPJDAE 137may be in groups of as many as 12 or more. Pyloric coeca 1.7 inthe head, on the left side.Origin of the dorsal over the gill slit; the anterior third of the raysburied beneath the fin membranes; the rays increasing rapidly inlength. Anal similar to the dorsal. Caudal slightly rounded, com-posed of stout rays, connected for three-fifths of its length to the anal;dorsal and anal connections with the caudal abrupt. Pectoraldeeply notched; the lower lobe of thickened, partly free, exerted rays,reaching the anal and extending back nearly as far as the uper lobe,nearly equal to the head; in the small specimen shorter, 1.6 in thehead. Disk small, hidden in a depression between the lower pectorallobes, slightly cupped, triangular, as wdde as long; the flap thickenedand suspended, posteriorly folded overthe center; tip of mandible to disk 10.2in the length Avithout caudal. In thesmall specimen the disk is more prom-inent and slightly triangular. Ventnext to the disk.Color: Opaque white; pores on thesides yellowish; stomach and part ofthe intestines dotted; peritoneum sil-very; coeca dusky at the base.Synopsis.?Dorsal 58-59; anal 52;pectoral 34-37; caudal 8; pyloric coeca21-34. Depth 2.8-3.3 in length with-out caudal; head 3.5. Eye 4.6-4.8 inthe head; disk 3.7 in the young to 7.8in the adults ; lower lobe of the pectoral1.1 in the adults and 1.6 in the j^oung.Body deep, shorter than in C. acantJiodesand C.tracJiysoma. Teeth simple, mod-erately stout. Eye moderate, black inall our specimens. Gill slit extending down in front of four or fivepectoral rays. Cactuslike pricldes present. Lower lobe of the pec-toral fin reaching to the origin of the anal fin, nearly equal to the headin the adults, 1.6 in the head in the young. Disk small, slightly cup-ped and triangular; tip of mandible to disk 9.6 to 10.8 in the lengthwithout the caudal. Body opaque, white, pinkish in life, peritoneumsilvery, undotted; stomach pale, dotted or not. Tliree specimens,the largest 280 mm., the smallest 84 mm. in length.CAREPROCTUS TRACHYSOMA Gilbert and BurkeCareproctus trachysoma Gilbert and Burke, 19126, p. .364.Type.?Male, No. 73333, U.S.N.M.; Albatross Station 4982, JapanSea; depth 390 fathoms. Length 263 mm. Figure 59.?Carepkoctus rastrinus.Prickles from type 138 BULLETIN 150, UNITED STATES NATIONAL MUSEUMDistribution.?Japan Sea, and Gulf of Tartary, Albatross Stations4814, 4981, 4982, 4983, 4992, and 4997; depth 318 to 429 fathoms.Thirteen specimens examined, reaching a length of 263 mm.RelationsTiips.? C. trachysoma appears to be closely allied to C.acantJiodes and C. rastrinus; distinguished from C. acanthodes by thedotted peritoneum, smxaller disk, narrower head, more fin rays, andthe longer lower pectoral lobe; from C. rastrinus by the larger eye,the dotted peritoneum, the depth and shape of the body, and thesmaller number of pectoral rays; from C. melanurus in the largergill slit, coloration, and the presence of prickles. (See description ofC. pellucidus.)Synopsis.?Dorsal 58-60; anal 53-55; pectoral 31-32; caudal 7-8;pyloric coeca 20. Depth of body 3.5-4.1 in length without caudal;head 3.7-4.6; tip of mandible to disk 8.5-11. Eye 3.8-4.2 in head;disk 7.5-8.3 in large specimens, 5.8 in young; lower lobe of the pectoral Figure 60.?Caeeproctus trachysoma. Teeth from cotypefin 0.9 in adults, 1.3 in young. Body moderately deep, more elongatethan in C. rastrinus. Teeth simple. Eye moderate, the lower halfsilvery or black. Gill slit normally extending down in front of fromone to five pectoral rays. Cactuslike prickles present. Lowerpectoral lobe elongate, reaching nearly to the anal fin, about equal tothe head. Disk small, slightly cupped ; the margin in front and behindfolded over. Body dusky gray, bluish black posteriori; lips dusky;peritoneum silvery, dotted; stomach blacldsh.CAREPROCTUS GILBERTI BurkeCareprodus spectrum Bean, 1890, p. 40 (part, confused with type of C. spec-trum.?EvERMANN and GoLDSBOROUGH, 1907, p. 333 (part, not of Gilbert.)?Gilbert and Burke. 1912a, p. 76.Careprodus gilberti Burke, 1912a, p. 568.Type.?'No. 64110, U.S.N.M.; north of Kodiak Island, Alaska,Albatross Station 4292; depth 102 fathoms. REV ISION OF THE FISH FAMILY LIPARIDAE 139Distribution.?^Off British Columbia and southeastern Alaska,Albatross Stations 2848, 2862, 3480, 4292, 4293, and 4781; depth 102to 482 fathoms. Thirty specimens examined.Relationships.? Careproctus gilberti closely resembles Careproctusostentum. Specimens of Careproctus gilberti were found in the type Figure 61.?Careproctus gilberti. Teeth from typebottle of Careproctus spectrum indicating that these species closelyresemble each other. Careproctus gilberti can be distinguished fromC. ostentum by the distinctly larger disk and from C. spectrum by thesmaller disk, wider gill slit and the darkly colored stomach. Figure 62.?Careproctus gilberti. Teeth from cotypeSynopsis.?Dorsal 55; anal 46-48; pectoral 31; pyloric coeca10-12. Disk 7.3-9 in head, small and deeply cupped. Gill slitextending down in row of 14 pectoral rays. Teeth simple or a fewwith lateral lobes. Color pale or pinkish; peritoneum silvery;stomach blackish. 140 BULLETIN 150, UNITED STATES NATIONAL MUSEUMCAREPROCTUS OSTENTUM GilbertCareproctus ostentum Gilbert, 1896, p. 444.Type.?Fem&\e, No. 48619, U.S.N.M.; Albatross Station 3324,north of Unalaska, Bering Sea; depth 109 fathoms. Figure 63.?Careproctus ostentum. Teeth from type ^Distribution.?Bering Sea; Albatross Stations 3331 and 3324;depth 109-350 fathoms. Three specimens examined, all of small size.Relationships.? C. ostentum isdistinguished by the minute disk.It closely resembles C. gilberti buthas advanced a step farther to-ward the loss of the disk.Description of type.?Dorsal 54;anal 47; pectoral 32; caudal S.Depth 4.5 in length withoutcaudal; head 3.8. Eye 3.2 in head;disk minute, more than 9 in thehead.Body rather heavy at the nape,short, slender posteriorly. Headheavy; interorbital broad, flat;occiput slightly swollen; cheeksvertical. Mouth broad, lateralcleft reaching vertical from pupil : maxillary reaching vertical fromposterior margin of eye. Teethsimple, numerous, inner teethelongate, slender, recurved; theoblique rows difficult to count,about 10 in the half of the lower jaw. Snout low, profile retreatinggradually from the mouth; upper teeth partly exposed when thejaws are closed. Nostril in a very short tube. Eye large, promi- FiGURE 64.?Careproctus ostentum.from specimen No. 3023, StanfordsiTY Zoological Museum pricklesUniver- EEVISION OF THE FISH FAMILY LIPAKIDAE 141 nent, black, the lower half probably silvery in life; pupil large, round.Gill slit in front of 12 pectoral rays. Pores 2-6-7-2. Pyloric coecaabout 10. No prickles on type; cactuslike pricldes on specimen No.3023, these may be isolated or in groups of 10 or 12 or more.Anterior dorsal rays short, buried in tissue beneath the sldn.Pectoral fins very oblique, the symphysis under the front of the eye,the notch broad, bridged by 10 shortened, \videly spaced rays; 17T&js above the notch; the lower lobe of 5 delicate rays, reachingmidway to anal fin. Disk absent from the type, in specimen No. 695minute, perfect in structure, hidden between the pectoral lobes.Caudal slender, of eight rays. Figure 65.?Carlproctus roseofuscus. Teeth from typeColor pale; peritoneum silvery with black dots; mouth and gillcavity pale; stomach black.Syno2)sis.?Dorsal 54; anal 47; pectoral 32; caudal 8; pyloriccoeca 10. Depth 4.5 in length without caudal; head 3.8; lower lobeof the pectoral fin 1.8. Body rather heavy anteriorly, short. Teethsimple, slender, elongate and recurved. Gill slit extending down infront of 12 pectoral rays. Cactuslike prickle present or absent.Disk minute, apparent!}^ perfect in shape and structure, more than 9in the head. Body pale, peritoneum silvery, dotted; stomach black.CAREPROCTUS ROSEOFUSCUS Gilbert and BurkeCareproclus roseofuscus Gilbert and Btjrke, 19126, p. 389.Type.?No. 73338, U.S.N.M.; Albatross Station 5026, southernpart of the Okhotsk Sea; depth 119 fathoms; length 91 mm. 142 BULLETIN 150, UNITED STATES NATIONAL MUSEUMDistribution.?Okhotsk Sea, off Saghalin, Albatross Stations 5018and 5026; depth 100 and 119 fathoms. Five specimens examined,reaching a length of 114 mm.RelationsMps.-?Careprodus roseojuscus, having s similar type ofdentition, is allied to the C. coUetti group of species. It is distinguishedby the paler coloration, the greater depth of head and body and theshort lower pectoral lobe.Synopsis.?Dorsal 57-58; anal 48-49; pectoral 28-30; caudal 8pyloric coeca 21-22; depth 4 in length without caudal; head 4-4.2tip of mandible to disk 9.5-10. Eye 4.2-4.5 in head; disk 3.3-3.5lower lobe of pectoral 2.1. Body comparatively deep and compressed . Teeth simple, lanceolate, considerably recurved; the anterior teethappearing little smaller than the inner teeth; the cutting surface ofthe bands of teeth appearing horizontal. Gill slit above the base ofthe pectoral fin. No prickles. Lower lobe ofHhe pectoral fin short,reaching nearly to the vent, 2.1 in the head. Disk oval, cupped.Body pale or pinkish; caudal and neighboring portions of dorsal andanal dusky; peritoneum black. Five specimens, reaching a length of114 mm. CAREPROCTUS COLLETTI GilbertCareproctus colletti Gilbert, 1896, p. 442. ? Jordan and Evermann, 1898,p. 2131. ? Gilbert and Burke, 19126, p. 370.Type.?No. 48698, U.S.N.M.; Albatross Station 3338, south of theAlaska Peninsula; depth 625 fathoms; length 84 mm.Distribution.?IsioTth Pacific, Bering Sea, Okhotsk Sea and theJapan Sea; Albatross Stations 3325, 3338, 4982, and 5029; depth 284to 625 fathoms; temperature 35? to 38? F. Seventeen specimensexamined.Relationsliips . ? C. colletti bears a close resemblance to C. entomelas.It differs from the latter species in the absence of prickles, the darkercoloration, the longer lower pectoral lobe, and the shorter snout.Description oj type.?Dorsal 52; anal 50; pectoral 28; caudal 6;pyloric coeca 8. Depth 5.6 in length without caudal; head 4.5.E3^e3.5inhead; disk 3.2.Body elongate, deepest at union with head; tapering gradually.Head slender, compressed; occiput high; profile forming a straightline from tip of snout to occiput; interorbital flat. Mouth moderate,lateral cleft reaching past front of eye; maxillary reaching posteriormargin of pupil. Teeth lanceolate, recurved, the cutting surface ofthe bands horizontal; the rows very oblique, about 11 in the half ofeach jaw; outer teeth smaller but -with dissection, appearing almostas large as the inner teeth. Snout short, deep; jaws about equal, thelower slightly projecting. Nostril in front of the eye, the tube, veryshort, simply a raised rim. Eye large, black; pupil round. Gill slitsmall, 5.5 in the head, above the base of the pectoral fin. Supra- REVISION OF THE FISH FAMILY LIPARIDAE 143branchial pores 2. No prickles. Pyloric coeca 8, nearly equal to theeye, on the left side of the body cavity.Dorsal rays slender; the anterior rays not buried in pseudotissue,not projecting above the fin membrane. Caudal slender, truncate,connected for one-sixth its length to the anal, the connection gradual,without a notch. Upper lobe of pectoral reaching anal; the raysslender at the tips, partly free and coiled; the lower lobe of five raysgraduated in length, free nearly to the base, coiled at the tips, reachingthe anal, 1.1 in the head. In a small specimen the lower lobe does notreach beyond the vent. Disk oval, cupped, the margin stiff andsuspended; tip of mandible to disk 9 in length without caudal.Vent distant from disk by two-thirds diameter of disk.Coloration dusky; snout, lips, abdomen, gill cavity, and tonguedusky; caudal and dorsal and the anal black posteriorly; peritoneumblack; coeca pale.Synopsis.?Dorsal 54-57 ; anal 49-o2 ; pectoral 27-29; caudal 8; pyloric coeca8-18. Depth 3.6-6 in the length with-out the caudal, depending upon thesize of the specimen; head 4.2-4.8.Eye 3.4-5 in the head; disk 3.2-5.3;lower lobe of pectoral 0.8-1.2. Bodyelongate. Teeth simple, elongate, dis-tinctly recurved ; the bands horizontal.Gill slit above the base of the pectoralfin. Lower lobe of pectoral elongate,about equal to the head, reaching back-ward to the anal fin in adults, in some ^"^"^^ 'te"eth?""e """"'of the small specimens not extendingbeyond the vent. Disk small, oval, deeply cupped. No pricldes.Color dusky or gray, darker posteriorly; lips, mouth, and gill cavitydusky or black; caudal, margin of dorsal and anal, and the peritoneumblack. Seventeen specimens, attaining a length of 248 mm.Remarks.?In 1906 the Albatross collected a number of specimensof C. coUetti in the Okhotsk and Japan Seas. These specimens differsomewhat from the types as evidenced by the following notes.Note on small specimens less than 100 mm. in length from theOkhotsk Sea. Body elongate; depth more than 5 in the length.Teeth lanceolate, strongly recurved, the cutting surface of the bandshorizontal; about 15 oblique rows in the half of each jaw; teeth infront appearing little or not at all shorter than the inner teeth. Nos-tril in a short tube. Snout short, bluntly rounded, slightly project-ing; jaws equal. Pores 2-6-7-2; the 2 suprabranchial pores closetogether, the tubes of equal length; the upper pore on the snout at 144 BULLETIN 150, UNITED STATES NATIONAL MUSEUMthe tip of the snout, distant from the nostril by the diameter of thepupil.Caudal slender, truncate, connected for one third its length todorsal and anal; last dorsal and anal rays not shortened. Thelower lobe of the pectoral of elongate rays, these free nearly to thebase; the lobe variable in length, reaching from a short distance pastthe vent to the anal fin, usually reaching midway between vent andanal fin.Color dusky, darker toward the caudal; the skin transparent, lax,dusted with brown dots; lips black; flesh dotted; lower lobe ofpectoral fin darker than the upper lobe; mouth and gill cavity dusky;peritoneum black; stomach white.Note on several large specmiens over 200 mm. in length. Bodydeeper; depth 3.6 to 4.1 in length without caudal. Lower lobe ofpectoral fin longer than head, extending beyond the upper lobe and Figure 67.?Careproctus colletti. Teeth from a specimen from Albatross Station 5029past the origin of anal fin. Body gelatinous; anterior dorsal rayssurrounded by tissue; skin lax, less dusky than in the small specimens.Color pinkish; lips black; margin of dorsal and anal and the caudalblack; gill cavity dusky.The large specmiens from the Japan and Okhotsk Seas appear tobelong to the same species as the specmiens from Bering Sea and theGulf of Alaska. There are no large specimens from the latter regionswith which to make comparison. The large specimens differ consider-ably among themselves, especially in the size of the disk. The smallspecimens taken with the large specimens differ slightly, or not at all,from specimens of equal size taken off the coast of Alaska. It appearsfrom the evidence we have that the species ranges from the Gulf ofAlaska to the Japan Sea and that the specimens attain the greatestsize in the latter and neighboring portions of its range.The teeth of specimen No. 53033, U.S.N.M., appear to differ fromthe teeth examined in other specunens. The teeth of this specimenappear to be stouter and with the lateral lobes slightly developed. REVISION OF THE FISH FAMILY LIPARIDAE 145The specimen is immature and may not belong with this species. Theteeth may, however, be of this type in the very young.Color in life gray, everywhere very finely punctulate with black;head, body, and fins lightly tinged with red, most pronounced onfront of head and on pectoral fins; lower pectoral rays faintly barred;vertical fins anteriorly with a narrow black margin, which graduallybroadens posteriorly until it involves the entire height of the fins;lips, mouth, and gill cavity dusky; the peritoneum black.CAEEPROCTUS ENTOMELAS Gilbert and BurkeCareproctus entomelas Gilbert and Burke, 19126, p. 374.r^/pe.?No 73343, U.S.N.M.; Albatross Station 4983, coast ofHokkaido, Sea of Japan; depth 428 fathoms. Length 77 mm.Distribution.?J&-pa,n Sea; Albatross Stations 4838 and 4983;depth 144-428 fathoms. Two specimens examined, reaching a lengthof 77 mm. Figure 68.?Careproctus entomelas. Teeth and prickle from typeRelationsTiips.? C. entomelas is allied to C. colletti and C. entargyreus.The dentition of these three species is similar; they agree also in mostother respects. C. entomelas is distinguished from C. colletti bythe lighter coloration, the deeper body?, the presence of prickles,the more projecting snout, and the shorter lower pectoral lobe; fromG. entargyreus by the darker coloration, the larger pupil, and thedistribution.Synopsis.?Dorsal 53; anal 46; pectoral 28; pyloric coeca 19.Depth 5.5; head 4. Eye 3.5; disk 3.3; lower lobe of pectoral fin 1.4.Body elongate. Teeth lanceolate, recurved; the bands horizontal.Snout blunt, distinctly projecting. Eye moderate, pupil large,round. Gill slit above the pectoral fin or extending down in front ofthe base of the upper ray. Thumb-tack prickles present. Lowerlobe of the pectoral fin reaching little past the vent. Disk small,cupped. Body pale, slightly dotted, more pigment posteriorly;lips pale or slightly dusky; mouth and gill cavity slightly dusky;peritoneum black. 146 BULLETIN 150, UNITED STATES NATIONAL MUSEUMCAREPROCTUS ENTARGYREUS Gilbert and BurkeCareproctus entargtjreus Gilbert and Burke, 1912b, p. 375.Type?'No. 73344, U.S.N.M.; Gulf of Tartary, ^Z6a^ross Station4998; depth 66 fathoms. Length 73 mm.Distribution.?Guif of Tartary, Japan Sea; Albatross Stations 4998and 5003 ; depth 35 to 66 fathoms. Two specimensexamined, reaching a length of 77 mm.Relationships.? C. entargyreus appears to be closelyrelated to C. entomelas. It differs from the latterspecies in the smaller amount of pigment in theperitoneum, the smaller pupil and the verticaldistribution.Synopsis.?Dorsal 57; anal 49; pectoral 31; caudal8; pyloric coeca 13. Depth 5.7 in length withoutcaudal. Disk 2.8 in head. Gill slit above pectoral.Teeth as in C. coUetti, simple, elongate, slender, strongly recurred.Pupil reduced to a minute pore. Color pale, mouth and gill cavitypale; peritoneum pale with black dots.TEMNOCORA, new genusDisk, present; nostril single; teeth trilobed; suprabranchial pores2; pyloric coeca present; pupil slitHke; dorsal fin notched; branchi-ostegal rays 6.Genotype.? Temnocora Candida. Figure 69.?Care-proctus ENTARGYR-ous. Prickle fromTYPE /s. Figure 70.?Temnocora Candida. Type, showing the notched dorsal and slitlike pupilTEMNOCORA CANDIDA (Gilbert and Burke)Careproctus candidus Gilbert and Burke, 1912a, p. 77.Type.?No. 74384, U.S.N.M.; Albatross Station 4784, off AttuIsland, Bering Sea; depth 135 fathoms. Length 72.5 mm.Distribution.?Bering Sea, Albatross Station 4784; depth 135fathoms. Four specimens examined, reaching a length of 76 mm.Synopsis.?Dorsal 45-48; anal 39; pectoral 33-37; pyloric coeca20. Depth 3.7 in length without caudal; head 3.5-3.8. Eye 2.8 in Figure 71.?Temnocora Candida. Teeth from type REVISION OF THE FISH FAMILY LIPARIDAE 147head; disk 2.9-3.4; lower lobe of pectoral fin 1.5. Body short, deep.Teeth strongly trilobed. Eye large, prominent; pupil reduced to ahorizontal slit. Gill slit above the pectoral fin. Dorsal fin distinctlynotched. Disk well developed, flat. Color white or pinkish; peri-toneum pale.Genus CRYSTALLICHTHYS Jordan and GilbertCrystallichthys Jordan and Gilbert, 189S, Bull. U. S. Nat. Mus., No. 47, p. 2864.Disk present; nostril single; teeth trilobed; suprabranchial poresone or two; pyloric coeca present; pseudobranchiae absent; bran-chiostegal rays 6; pupil reduced to a slit; body with roundish orirregular blotches.Genotype.? Crystal-lichthys mirabilis.CRYSTALUCHTHYS CYCLO-SPILUS Gilbert and BurkeLiparis cyclostigma Gilbert,1896, p. 446 (part, con-fused with Liparia cyclos-tigma.?Jordan andEVERMANN, 1898, p. 21-25. ? Jordan and Gil-bert, 1899, (part, after Gilbert). ? Evermann and Goldsborough, 1907,p. 333, pi. 19 (part).Crystallichthys mirabilis Jordan and Gilbert, 1899, p. 476, pi. 76 (part).Crystallichthys cyclospilus Gilbert and Burke, 1912a, p. 74.Type.?'No. 74381, U.S.N.M.; Albatross Station 4779, Bering Sea;depth 54 fathoms. Length 200 mm.Distribution.?Bering Sea, Albatross Stations 3252, 3439, 3638,4777, and 4779; depth 29-52 fathoms. Twelve specimens examined,reaching a length of 200 mm.Relationships.? C. cyclospilus closely resembles C. mirabilis.These two species agree in most respects, particularly in having apeculiar pupil and blotched coloration, C. cyclostigma can bedistinguished from C. mirabilis by the shorter rounded snout, thebroader head, the more rounded pinkish blotches, and the snout notdivided on the lower surface.Synopsis.?DotsbI 48-50; anal 42-43; pectoral 33-35; caudal 10;pyloric coeca 36. Depth 3.1-3.8 in length without caudal; head3.6-3.8. Eye 6-7 in the head; disk 2.2-2.4. Body deep and muchcompressed, moderately elongate. Teeth stout, blunt, distinctlytrilobed. Snout short, deep, broadly rounded, not projecting. Eyesmall; pupil elliptical, partially divided dorsally by a projection fromthe iris. Gill slit above the pectoral fin. Disk normal, large. Bodywhitish in alcohol, pinkish in life, large roundish, pinkish blotches onhead, body, and fins; peritoneum pale. 148 BULLETIN 150, UNITED STATES NATIONAL MUSEUMCRYSTALUCHTHYS MIRABILIS Jordan and GilbertCrysialUchthys mirabilis Jordan and Gilbert, 1898, p. 2864 (part).?1899,p. 476, pi. 75 (part). ? Evermann and Goldsborough, 1907, p. 333, pi. 20. ? Gilbert and Burke, 1912o. p. 75.Type.?Female, No. 51466, U.S.N.M.; Albatross Station 3643, off.the southeast coast of Kamchatka; depth 100 fathoms. Length330 mm.Distribution.?Southeast coast of Kamchatka, Albatross Stations3643 and 4794; depth 58 to 100 fathoms. Two specimens examined,the largest 330 mm.Relationsliips.? C. mirahilis is closely related to C. cyclospilus. Itcan be distinguished from the latter species by the more compressedhead, the long conical divided snout, and the coloration.Description of type.?Body gelatinous, deep and compressed;depth 3.7 in length without caudal; body tapering very slowly, FiGURK 72.?Crystallichthys cyclospilus. Type, showing pecuuar pupil and blotchedCOLORATIONholding the depth past the middle then descending rapidly to thecaudal. Head much compressed, the sides nearly vertical; profile inan even curve; interorbital narrow; occiput not sw^ollen. Mouthbeneath the snout, small; maxillary reaching vertical from posteriormargin of pupil. Teeth stout, blunt, distinctly trilobed, in broadbands, close set; outer teeth smaller. Snout divided longitudinallyon the low^er surface, conical, a somewhat rigid fold within the cleft,tapering sharply and projecting beyond mouth. Nostril single, in aprominent tube. Eye moderate, 5 in the head; pupil elliptical,partially divided as in C. cyclospilus; eye black. Gill slit small, 4.6in head, above the pectoral fin. Pores on the snout and chin be-tween projecting folds of skin; suprabranchial pores 2. Pyloriccoeca absent from the type, 40 in a small specimen. No pricldes.Anterior dorsal rays buried in pseudotissue. Anal fin similarto the dorsal. Caudal truncate, connected for half its length to the REVISION OF THE FISH FAMILY LIPARIDAE 149 anal; the dorsal connection less, notched. Pectoral notched; thelower lobe reaching halfway from disk to vent. Disk large, 2.6 inthe head; the posterior margin under the gill slit; the flap thickand broad, equal to the center of the disk. Vent close behind disk.Color gra3''ish purple, probably translucent and pinkish in life,and with the dorsal region marked with large round reddish spotsas described; the type is discolored and no spots are visible.Synopsis.?Dorsal 53; anal 44; pectoral 30-33; caudal 12; pyloriccoeca 40. Depth 3.7 in length without caudal; head 4.7? (snout Figure 73.?Cbystallichthys cyclospi-Lus. Teeth from type Figure 74.?Crystaluchthys cyclospilus. Teethfrom specimen 130 mm. in length, from alba-TROSS Station 4779, Bering Straitinjured). Eye 5-5.5; disk 2.6-3. Body deep, greatly compressed.Head much compressed. Snout conical, piojecting, the lower sur-face cleft longitudinally. Teeth blunt, distinctly trilobed. Gill slitabove the pectoral fin. Pupil as in C. cyclospilus. Disk normal, large.Suprabranchial pores 1 or 2. Color graj^ish or whitish, translucentpinkish in life, pinkish blotches on head, body, and fins, these not sorounded as in C. cyclospilus, some irregular, other barlike.Remarks.?Note on a small specimen from Albatross Station4794. Head 3.5; depth 3.7. Dorsal 53; anal 44; pectoral 30; pyloric91668?30 11 150 BULLETIN 150, UNITED STATES NATIONAL MUSEUM coeca 40. Eye 5.5; snout 2.7; gill opening 7; disk 3. Snout clefton the lower surface as in the type, projecting beyond the upper lipfor a distance nearly equal to eye. Teeth trUobed, arranged in sevenoblique rows in the half of each jaw. The pupil resembles that inC. cyclospilus, elliptical, horizontal, partially divided from the dorsal FiGUBE 75.?Crystaluchthys mtrabius. Teeth from type side. Distance from tip of lower jaw to disk 3.3 in the head. Pores2-5-6-1. Body translucent; body, head, and fins with short, narrow,pale bars and small roundish spots; a bar extending forward from eyeand one above eye; the bars and spots, probably, pinkish in life,shaped somewhat as in Crystallias matsushimae. No notch betweend?rsal and caudal; caudal slightly rounded. /I A A Figure 76.?Crystaluchthys mirabilis. Teeth from a young speci-men FROM Albatross Station 4797, off KamchatkaGenus CRYSTALLIAS Jordan and SnyderCrystallias Jordan and Snyder, 1902, Proc. U. S. Nat. Mus., vol. 24, p. 349.Disk present; nostril single; teeth ti'ilobed; suprabranchial poresingle; pyloric coeca present; pseudobranchiae absent; branchiostegalrays 6; snout with barbels.Genotype . ? Crystallias matsushimae . CRYSTALUAS MATSUSHIMAE Jordan and SnyderCrystallias matsushimae Jordan and Snyder, 1902, p. 350, figure 2. ? Schmidt,1904a, p. 167, pi. 6.?Jordan, 19055, p. 218, figure 158.?Gilbert andBurke, 19126, p. 376. REVISION OF THE FISH FAMILY LIPARIDAE 151Type.?lSio. 49802, U.S.^M.; Albatross Station 3773, MatsushimaeBay, Japan; depth 78 fathoms. Length 113+ mm.Distribution.?Okhotsk Sea, Japan Sea, and Matsushimae Bay,Japan, depth 70-200 fathoms. Eight specimens examined, thelargest 221 mm.RelatiojisMps.? C. matsushimae resembles CrystallicMhys mirabilisin the cleft snout, shape of head and body, and the coloration;differing mainly in the presence of barbels and the character of thepupil.Description oj ^?/2?e.? Dorsal 56; anal 53; caudal 10; pectoral 31;pyloric coeca 71 (not from type). Depth 4.7 in length withoutcaudal; head 4.2. Eye 5 in head; disk 3.Body as in G. mirabilis, gelatinous, deep and much compressed;the depth about twice the width; cheeks nearly vertical; profilenearly straight from snout to occiput. Mouth narrow, Teethshort, stout, blunt, strongly trilobed; oblique rows widely spaced, Figure 77.?Ceystallias matsushimae. Introduced to show the barbels distinguishikgTHE genus. Typeabout 18 in the half of each jaw. Snout distorted; jaws equal. Thefollowing description of the snout and barbels is taken from twospecimens from Albatross Station 4855, Japan Sea. Snout projectingbeyond the upper lip for the diameter of the orbit. Barbels on eachside of the snout as follows : One in front of the upper snout pore andat the tip of the snout; sometunes a second lateral to and separatedfrom the former barbel by the anterior pore on the snout; three on thelower anterior surface of the snout between the maxillary pores,separated from those on the other side of the snout by the stiffvertical fold terminating and dividing the lower surface of snout.The upper lip with three and sometimes five short barbels. Lowerlip with a median and two lateral barbels on each side, separated bythe mandibular pores; sometimes a second series of two behind thefirst series. The longest barbel on the snout equals the eye; thebarbels on the chin equal to half the eye. The upper lip and lowersurface of snout divided, a stiff vertical fold of skin projecting fromthe base of the cleft, as in CrystallicMhys mirabilis. Nostril in a 152 BULLETIN 150, UNITED STATES NATIONAL MUSEUMprominent tube. Eye moderate; the lower half .silvery; pupilelliptical, horizontal, apparently not partially divided as in Crys-tallichtJiys. In some specimens the eye is black and the pupil dis-tended and round. Gill slit small, above the pectoral fin. Pores2-5-6-1 on a specimen from Albatross Station 4855; rudimentarypores on the sides of the body as in species of Liparis. Pyloriccoeca 71, in one specimen on the right side of the body cavity,1.8 in the head.Anterior dorsal rays buried in pseudotissue beneath the skin. Analsimilar to the dorsal. Caudal truncate, connected for a little morethan half its length to the anal fin. Pectoral fin notched; the lowerlobe more distinct than figured, reaching the vent, 2.4 in the head,composed of thickened rays as in Liparis. Disk large ; tip of mandibleto disk 11.2 in the length of the body without caudal. Vent closeto the disk.Color translucent, pinkish in life; narrow dark bars and spots onthe head and body; bars beginning near the median line on the sidesand extending onto the dorsal and anal fins,seven above and five below; one bar extendingbackward from the lip through the eye; twoacross the top of the head ; other spots and barson head and body smaller; peritoneum andstomach pale.Synopsis.?Dorsal 56; anal 53; pectoral 31;Figure 78.-crystalli as caudal 10; pyloric coeca 71. Depth 3.2-4.7;MATsusHiMAE. T E E T H j^ead 3 .9-4.2 . Eyc 5. 5 lu hcad ; dlsk 3 . Bodydeep and much compressed, gelatinous. Teethblunt, strongly trilobed. Snout distinctly projecting, cleft. Barbelspresent on chin and snout. Eye moderate; pupil oval. Gill slit abovethe pectoral fin. Disk large, flat. Body translucent, pinkish; narrowbars and spots on head and body as in CrystallichtJiys mirahilis;peritoneum pale. Genus GYRINICHTHYS GilbertGyrimci.thys Gilbert, 1895, Rept. U. S. Fish. Comm., 1893, p. 444.Disk present; nostril single; teeth simple; suprabranchial poresingle; pyloric coeca present; gill slit reduced to a minute pore abovethe pectoral fin; branchiostegal rsijs 6.Genotype.-?GyrinicJitJiys minytremus.GYRINICHTHYS MINYTREMUS GilbertGyrinichthis minytremus Gilbert, 1896, p. 444. ? Jordan and Evermann,1898, p. 2137.Type.?Fem&le, No. 48617, U.S.N.M.; Albatross Station 3331,north of Unalaska, Bering Sea; depth 350 fathoms. Length 72 mm.Distribution . ?Bering Sea, Albatross Station 333 1 , depth 350 fathoms. KEVISION OF THE FISH FAMILY LIPARIDAE 153Relationships.?G. minytremus differs from certain species of Care-produs solely in the reduced gill slit. See discussion of the relation-ships of the genera.Description of type.?Head 5.5 in length without caudal. Eye 3.3in head; disk 2.2 Pectoral 25; caudal 14.Body rather slender and elongate as in Careproctus edenes; abdomendistorted with large eggs. Head small, depressed, a little deeperthan wide; depth 7, width 8 in length without caudal; cheeks nearlyvertical; occiput not swollen; interiorbital flat, internostril 3.2 in thehead. Mouth small, terminal, with little lateral cleft; maxillaryreaching a vertical from slightly behind the front of the eye. Teethsimple, slender, concial, inner teeth prominent, much larger thanouter teeth ; about five widely spaced oblique rows in the half of eachjaw. Snout short, 3 in the head, rising abruptly from the mouth.Nostril single, the tube with a raised rim in front. Eye large, black;pupil round. Gill slit a minute pore that will not admit a largedissecting needle, high up above the base of pectoral fin. Tip of T Figure 79.?Qyeinichthys MiNVTREMrg. Teeth fkom type opercular arm buried beneath skin. One suprabranchial pore present,above and behind the gill slit; upper pores on the snout high; nearlybetween the nostrils. Pyloric coeca unknown. No prickles.Origin of dorsal fin back of the middle of the pectoral; the anteriorrays not distinctly seen but not buried in thick tissue. Caudaltruncate, of very slender rays, connected for half its length to theanal, the dorsal and anal connections with the caudal gradual.Pectoral fin not notched, with the upper edge on a level with the lowermargin of the orbit, of about 25 rays which increase gradually inlength. Disk well formed, oval ; snout to disk 8.5 in the length withoutcaudal. Vent close to the disk; tip of mandible to vent 4.3 in lengthwithout caudal.Color light brown; peritoneum white, dotted with browTi.Synopsis.?Pectoral 25; caudal 14. Body rather slender and elon-gate; head 5.5 in length without caudal. Eye 3.3 in head; disk 2.2.Teeth simple, slender and conical. Eye large, black; pupil round.Gill slit leduced to a minute pore high up above the base of the pecto-ral fin. One suprabranchial pore evident. Pectoral fin apparently 154 BULLETIN 150. UNITED STATES NATIONAL MUSEUM unnotched; the rays increasing gradually in length. Body slightlybrownish; peritoneum pale.Genus ELASSODISCUS Gilbert and BurkeElassodiscus Gilbert and Burke, 1912a, p. 81.Disk not perfect, reduced to a rudiment, the rays absent; nostrilsingle; teeth trilobed; suprabranchial pore single; pyloric coecapresent; pseudobranchiae absent; branchiostegal rays 6.Genotype.?Elassodiscus treinehundus.ELASSODISCUS TKEMEBUNDUS Gilbert and BurkeElassodiscus tremebundns Gilbert and Burke, 19120, p. 81.Tijpe.?No. 74388, U.S.N.M.; Albatross Station 4797, off Starits-chkof Island, Kamchatka; depth 682 fathoms. Length 214 mm.Distrihution.?Southeast coast of Kamchatka; Albatross Station4797; depth 682 fathoms. Numerous specimens examined. Figure 80.?Elassodiscus tremebundus. TypeSynopsis.?Dorsal 65; anal 60; pectoral 32; caudal 8; pyloriccoeca 16. Depth 4.4 in length without caudal; head 3.6. Eye 5.5in head. Body moderately elongate, deep and much compressed.Teeth stout, trilobed. Gill slit above the pectoral fin. One supra-branchial pore present. Disk rudimentary. Body pale; lips andgill cavity dusky; caudal and the marginal half of the dorsal andanal posteriorly black; peritoneum and stomach black. In lifetranslucent, reddish. Specimens numerous, reaching a length of214 mm. or more. Genus PARALIPARIS CollettParoKpam CoLLETT, 1878, Chra. Vid. Selsk: Forh., 1878, No. 14, p. 32, (bathybii).Amitra Goode, 1880, Proc. U. S. Nat. Mus., vol. 3, (liparina).Monomitra Goode, 1883, Proc. U. S. Nat. Mus., {Uparina}?.Hilgendorfia Goode and Bean, 1895, Special Bull. U. S. Nat. Mus., No. 2.,{membranaceus) . Amitrichthys Jordan and Evermann, 1896, Rep. U. S. Fish Comm., 1895, p. 453,(cephalus).Hilgendorfia Jordan and Evermann, 1896, Rep. U. S. Fish Comm., 1895, p. 453;1898, Bull. U. S. Nat. Mus., No., 47, (ulochir). REVISION OF THE FISH FAMILY LIPARIDAE 155Disk absent; nostril single; teeth trilobed to simple; suprabrancliialpores apparently single in all the species; pyloric coeca present, i^small numbers; pseudobranchiae absent; ^* branchiostegals 6.MODIFICATION OF CHARACTERSBody.?In the species of Paraliparis the body is never depressedas in the species of Liparis. It is compressed but never attains thedepth of some of the species of Careproctus, The shape of the bodyvaries among the different species. In P. copei it is moderately stoutand elongate, in P. liolomelas very heavy anteriorly and attenuateposteriorly and in P. mento very slender. Many of the species havethe posterior part of the body attenuate, the vertebrae being verysmall and evident through the skin.The bodies of the species of Paraliparis are typically frail and softbut lack the excessive amount of pseudotissue found in many species Figure Si.?Elassodiscus tremebundus. Teeth from coTrPE of Careproctus. Paraliparis is more typically a deep-sea genus thanCareproctus . Pseudotissue apparently is more common to the shal-lower water species of the two genera.The shape of the body and the proportional measurements of headand body have been used in but a general way in distinguishing thespecies.Head.?The head furnishes a number of useful characters. It istypically compressed. The width of the head is seldom if ever equalto the depth. The snout is usually deep and abrupt. It distinguishesa few of the species by projecting slightly. The profile of the headis typically rather low and rises gradually to the nape. P. cepJialusis distinguished by the occiput being greatly swollen. The mouth isusually horizontal, but P. cepTialus and a few other species are dis-tinguished by the mouth being at an angle. When the mouth is onan angle the lower jaw may be heavy and project as in P. mento.w See discussion of this character. 156 BULLETIN 150, UNITED STATES NATIONAL MUSEUMNostril.?The posterior nostril is absent in this genus. Theanterior nostrU opens close in front of the eye. There is never aprominent tube. In a number of species the tube appears to opento the surface by a pore. So far as known the nostril does notpresent any characters of specific value. The skin in many of thespecies was in such a condition that the nostril could not be ade-quately studied.Eye.?The characters presented by the eye have been utilizedbut to a slight extent. This is due to the laclc of good material.The eyes of many of the specimens examined were in poor conditionand would not admit of any importance being placed upon thedifferences they presented. The size of the eye among the differentspecies varies from 2.8 to 5.4 in the length of the head. The pupilis always round or slightly oval and usually large. The color ofthe eye varies from black to silvery. The silvery pigment is likelyto disappear in preserved specimens.Pores.?The pore formula for many of the species could not bemade out on account of the condition of the skin. The formulaappears to vary from 2-6-7-1 to 2-5-6-1. In none of the speciescould more than one suprabranchial pore be detected. The upperpores on the snout are close together and practically between thenostrils. Rudimentary pores were not observed in an}^ of the species.The pores are usually small but may be enlarged as in P. holomelas.In a few species the anterior mandibular pores are united, that is,have a common opening.Gill slit.? The size of the gill slit in a number of the species isunknown because of the torn condition of the skin. The gill slithas about the same amount of modification as in Careprodus. Inthe majority of the species it is confined to the region above the baseof the pectoral fin. In but a few of the species does the slit extenddo\\Ti in front of the pectoral. The largest gill slit is found in P.Tiolomelas in which species it extends down in front of 13 pectoralrays.Teeth.?The teeth in Paraliparis show a greater range of modifi-cation than in any other genus of the famLl3^ In P. dadylosus theteeth are trUobed, In the majority of the species they are simple,conical, and in narrow or broad bands; the oblique rows few or many,obscure or prominent; the inner teeth enlarged or not. In P. atra-mentatus the teeth are very stout and apparently suited for crushinghard objects. In P. rosaceus they are stout, wedge-shaped, arrangedin a single series, and their tips form a sharp cutting edge. Theteeth of the species of this genus deserve special stud}'" for they areof considerably taxonomic importance. The different types ofdentition described appear to grade into each other and can not be REVISIOX OF THE FISH FAMILY LIPARIDAE 157 of generic value. They are, however, of great value in distinguishingspecies or groups of species.I have not attempted to utilize to any great extent the modificationsin shape, size of pulp cavity, and numbers in the teeth of these species.When the time comes for a more intensive study of the fishes, theteeth will probably be found to offer a profitable line of research.The data that we have at present is not satisfactory for the reasonsthat we do not know the amount of variation in the teeth of theyoung and adults and between the two jaws and different sectionsof each jaw.Dorsal fin.?The dorsal fin in this genus has not been studiedsufficiently to warrant the use of the number of rays in distinguishingspecies. The number of dorsal rays varies from 48 to 66 or more.We have made no attempt to study the amount of specific variation.The rays are extremely delicate and difficult to count accurately.For these reasons the writer has made no attempt to use the numberof dorsal and anal rays in separating species. Dorsal fin is unnotched.The segmentation of the anterior dorsal rays was investigatedin a number of species. The anterior rays are divided but unseg-mented. The number of unsegmented rays varies among thedifi'ercnt species. The species examined, with the number ofunsegmented dorsal and anal rays, are listed below: Dorsal Analfin finco-pei 18 16Do (?) 8Do 13 10deani 9 4Do 11 4cephalus 10 6Anal fin.-? Wl\a.t has been said concerning the number of dorsalrays applies equalh' well to the number of anal rays. The numberof anal rays varies from 42 to 60.Caudalfin.?In studying the caudal fin attention should be dii'ectedto the number of rays and the connection with the dorsal and analfins. The number of rays varies from three or four to eight. Appar-ently there is little variation in the number for each species. Thisis a question which needs further study. Apparently there ai'e norudimentary rays at the base of the caudal as in Liyaris.The caudal fin is always distinct. The dorsal and anal are nevercontinuous as frequently described. The connection between theanal and the caudal varies from 0.3 to 0.6 of the length of the latter.The connection with the dorsal and anal is usually very gradual,though somewhat abrupt in a few species.Pectoral fin.?The pectoral fin presents some of the most importantspecific characters. These have been used freely in the key to the 158 BULLETIN 150, UNITED STATES NATIONAL MUSEUM species. The number of rays, the extent of the pectoral notch anddevelopment of the middle pectoral rays, the length of the lowerpectoral lobe, and the level at which the upper edge of the pectoralis attached to the body have been found to be of the most importance.The pectoral notch is distinct in all the species except P. fimbriatuSyin which it is hardly evident. P. fimbriatus is distinguished also bythe pectoral rays being evenly spaced on the girdle. The middlepectoral rays are, in all the other deep-sea species of the family, morewidely spaced than the rays above or below. In species like P.Tiolomelas and P. hathyhius in which the pectoral notch extends to ornearly to the girdle the middle rays are reduced to short filamentsand may not project into the fin membrane. In none of the speciesexamined by the writer, with the possible exception of P. rosaceus,in which one ray appears to have been torn away, is the middle of thegirdle bare of rays for an unusual distance. The distance betweenthe rudimentary rays may progressively increase toward the lowerlobe. The space between the lowest rudimentary ray and the upperray of the lower lobe was never found to greatly exceed the spacebetween the two lower rudimentary rays. Giinther describes P.hathyhius and Gilchrist P. australis as having the girdle opposite thenotch free of rays.The length of the lower pectoral lobe varies considerably. InP. melanohranchus the lower lobe is contained twice in the head andin P. mento it is longer than the head. The variation of the length ofthe lobe within the species can not be studied because of the lack ofmaterial.The lower pectoral lobe never consists of graduated rays as inCareproctus and Liparis. In some of the species the lower ray is notmore than one-half the length of the next ray but there is never aseries of rays beginning with a short one and increasing regularly inlength.The rays of the lower pectoral lobe are usually half free. In somespecies, as in P. liolomelas , they are free practically to the base. Theyare frequently coiled or wavy at the tips.The level on the side of the body at which the upper part of thepectoral fin is fastened is considerably lower in some of the speciesthan in others. In the more generalized species the upper edge of thepectoral is on a level with the pupil. In P. cepTialus it is below thelevel of the orbit and in P. mento below the angle of the mouth. Inthese species the pectoral girdle becomes more oblique and thesymphyses of the pectoral is carried forward onto the throat until, asin P. mento, it is in front of a vertical from the anterior border of theeye. Associated with the lowering of the pectoral is the projectionand enlargement of the lower jaw and the changing of the angle of themouth. EEVISION OF THE FISH FAMILY LIPARIDAE 159Giinther describes P. membranaceous as having a very delicatefringed pectoral. This species has not been examined by the writer.Vent:?The position of the vent varies from beneath the gill slit tobeneath the eye. The distance between the tip of the lower jaw andthe vent varies with the position of the vent and has been used tosome extent in separating species.PseudohrancTiiae.?The writer has been aunble to demonstrate thepresence of pseudobranchiae in any of the specimens examined.These specimens were in such condition that it can not be said thatpsuedobranchiae were not present. Goode has described pseudo-branchiae in P. liparina but I have been unable to confirm hisdescription.Pyloric coeca.?The number of pyloric coeca has been greatlyreduced in this genus and varies from 5 to 18. As in Careproctus thecoeca are on the left side. At present we can not estimate theirtaxonomic importance.PricHes.?Thumb-tack prickles have been discovered on only onespecies, P. deani. None of the species are known to have cactuslikeprickles.Coloration.?The range of coloration exhibited by the species ofParaliparis is similar to that of the species of Careproctus. The colorof the species varies from white or pinkish to black. Paraliparis indistribution, structure, and coloration is more of a deep-sea genus thanCareproctus. More of the species, fully 50 per cent, are black and theperitoneum is dusky or black in nearly all the species. (See deaniand entochloris.) The stomach, gill cavity, and mouth are frequentlydusky or black.As in Careproctus the pigment is more common to the deeper waterforms. None of the black species have been taken at a depth of lessthan 405 fathoms. P. rosaceus, a light colored species, is found at adepth of 984 fathoms but the majority of such species are restrictedto shallower water. When the body is partly black the pigment isusually confined to the posterior region. An exception is found inP. rosaceus, in which the pigment is most pronounced on the snout.The variations in the coloration of the skin, flesh, peritoneum,stomach, gill cavity, and mouth all deserve careful consideration forthey frequently indicate specific differentiation.Habits.?The habits of the species of Paraliparis are probablysimilar to those of the species of Careproctus. Species have been takenfrom all kinds of mud and rock bottom.Little is known concerning the food habits of these fishes. Thelack of material prevented an exa^mination of the contents of thestomachs. The teeth are greatly modified in some of the species andprobably indicate special food habits. 160 BULLETIN 150, UNITED STATES NATIONAL MUSEUMSummary.?In order to present a guide to the study of the specificcharacters of these fishes the following outlines are given. The firstlist suggests the characters which the writer has found to be the mostvaluable in separating and detecting species. The second list involvesthose characters which could not be adequately studied but whichgive promise of being of taxonomic importance and warrant furtherstudy. For modification of characters in the genus see table onpage 36. List 1Body Depth.Length.Head Depth, snout.Depth, occiput.Angle of mouth.Eye Size.Gill slit in relation to pectoral.Teeth Trilobed.Shape.Size.Bands.Single series. Pectoral. Coloration. Number of ra3's.Level of upper edge.Rudimentary rays.Depth of notch.Length of lower lobe.Spacing of rays.Separation of rays oflower lobe.Body.Peritoneum.Stomach.Gill cavity.Mouth.Distribution.List 2AnalCaudaL Coeca__Prickles. Number of rays.Number of rays.Shape.Connection with dor-sal and anal.Number. Head Nostril.Eye Pupil.Color.Pores Formula.Position.Size.United on chin.Dorsal Number of rays.Buried in tissue.Spinelike rays.DLSTRIBUTIONThe genus Paraliparis is represented throughout the regions occu-pied by Careprodus, that is, the cold northern and southern regionsof the Atlantic and Pacific and the depths of the tropical Pacific.The limits of the bathymetrical distribution of the two genera arepractically identical. Paraliparis has been taken in depths rangingfrom 30 to 1793 fathoms. The distribution of the two genera differin two particulars. A smaller percentage of the species of Paraliparishave been taken in the North Pacific and a larger percentage in thetropical Pacific. Also the genus typically inhabits greater depthsthan does Careprodus. About half of the species are found aboveand half below the 500 fathom level. The 300-fathom level holdsthe same relation to the distribution of the species of Careprodus. REVISION OF THE FISH FAMILY LIPARIDAE 161Paraliparis, in distribution as well as in structure, is more typicallya deep-sea genus than Careprodus.The giant species of the genus are not confined to the North Pacific.These species are: P grandiceps, 256 mm., Gulf of California, 1,588fathoms; P. hathyhius, 208 mm., Arctic Ocean, 640-658 fathoms;P copei, 178 mm., off New England, 300-594 fathoms; P latifrons,147 mm., off Panama, 1,793 fathoms; P rosaceous, 135+ mm., offCalifornia, 984 fathoms.The maximum temperature record for any of the species is that of52.8? F. for P. ceplialus. The minimum temperature is that of 28? F.for P. tathyhius. The majority of the species are taken in tempera-tures ranging from 35? to 45? F. and live in an average temperature of40? F. The greatest range of temperature recorded for any species is15? F. for P. ceplialus.Regions.?We know so little about the species of Paraliparis thatwe can only state the facts that we have without drawing any con-clusions as to the regions or the center of dispersal. The majority ofthe species are known from but a single dredge haul. None of thespecies are common to both the Atlantic and Pacific or to the Japa-nese and American coasts. The Aleutian Islands do not act as abarrier to these fish.We have yet to discover the extent of the range of these deep-seafish. We have practically nothing to indicate whether or not theyare cosmopolitan in their distribution. P. ulocTiir is recorded fromthe Gulf of California and Bering Sea. None of the other species arerecorded from such widely separated localities.Bathymetrical distribution.?Species of Paraliparis have been takenin depths ranging from 30 fathoms off British Columbia to 1,793fathoms off Panama. About half of the species have been takenabove and half below the 500-fathom level. The vertical distributionof some of the species appears to be excessive though we should becareful in accepting the records as accurate when the dredge is hauledup open. P. holomelas appears to have a vertical range of 1,219fathoms and P. ulocJiir of 599 fathoms. These may be extended orrestricted when the species beco*me better Icnown. While some of thespecies have considerable vertical range none of them extend fromthe well lighted regions down below the penetration of sunlight.The species may extend from the well lighted to the dimly lightedregions or from the dimly lighted regions to the dark regions. Noneof the species of the family are known to exist on both sides of theregion between 250 and 500 fathoms which may be considered as thedimly lighted regions of the oceanic depths. 162 BULLETIN 150, UNITED STATES NATIONAL MUSEUMKEY TO SPECIES OF PARAUPARIS "A'. Teeth trilobed; gill slit in front of 4 pectoral rays; coeca 13-18; pectoralrays 30 or more dactylosus.A^. Teeth simple.B'. Pectoral with more than 30 rays; pectoral notch very shallow .angustifrons.B^. Pectoral with less than 30 rays.C. Teeth in bands.D*. Teeth conical.E'. Mouth horizontal.F'. Middle pectoral rays not rudimentary, distinctly forming part ofthe fin outline; pores on the he^d not enlarged.G'. Pectoral normal.H'. Pectoral rays evenly spaced on the girdle; little shorterthan lower rays fimbriatus.H2. Middle pectoral rays widel}' spaced.J'. Gill slit extending down in front of the pectoral?K'. Pacific; dorsal 56; anal 46; gill slit in front of 10pectoral rays deani.K2. Pacific; dorsal 63; anal 56; gill slit? entochloris.K^. Atlantic; dorsal 59; anal 51; pectoral 20; gill slit prob-ably above the pectoral bathybius^*.J 2. Gill slit above the pectoral fin.L'. Atlantic.M'. Color pale liparina.M2. Color black bathybius.JJ. Pacific.N'. Pectoral 25; the upper edge on a level with theupper margin of the eye; color black, includ-ing mouth and gill cavity ulochir.N^. Pectoral 20; the upper edge on a level with thepupil; color pale, including mouth and gillcavity; gill slit unknown entochloris.N'. Pectoral 17; the upper edge on a level withthe pupil; color pale; mouth and gill cavitvblack melanobranchus.'*G^. Pectoral delicate, fringed membranaceous.'"F^. Middle pectoral rays rudimentary, short or absent.O'. Color black; rudimentary median pectoralraj's present; rays 23.P'. Stomach pale latifrons.P2. Stomach black holomelas.O^. Color black; pectoral rays 15-19?; no rudi-mentary rays? bathybius.'"0^. Color pale; pectoral 17; no rudimentary raysbetween the two lobes australis.E^. Mouth at an angle.Q'. Atlantic; depth 4.6; snout to vent 6; pec-toral 21 ; the upper edge of the fin above theangle of the mouth; the middle rays rudi-mentary; the symphysis behind the front ofthe eye; stomach pale; caudal S.garmani. " See appendix for additional species. '? In key in more than one place. REVISION OF THE FISH FAMILY LIPARIDAE 163Q'. Pacific; snout to vent 5.5; pectoral 14; theupper edge of the fin above the angle of themouth; the middle rays not rudimentary;the symphysis behind the front of the eye.cephalus.Q^. Pacific; depth 6; snout to vent 9.8; pectoral16; the upper edge bdlow the angle of themouth; the middle rays not rudimentary;the symphysis in front of the eye__inento.D^. Teeth blunt, not conical atramentatus.O. Teeth in a single series. R'. Atlantic; snout to vent 8.5; head 6.5copei.RK Pacific.S^ Snout to vent 6; head 5; color rose red;fins and snout largely black; dorsal58; anal 53; pectoral 20 rosaceus.S2. Color blackish; dorsal ca 56; anal ca 42;pectoral 24; gill slit unknown,grandiceps.S^. Color blackish; dorsal 66; anal 57; pec-toral 24; gill slit unknown,attenuatus.SUPPLEMENTARY KEY TO SPECIES OF PARALIPARISA'. Atlantic species.B'. Teeth in bands.C Color pale.D'. Mouth horizontal.E'. Pectoral normal.F'. Caudal 9; D. 48; P. 17; no rays bet^veen the two pectorallobes australis.'*F^. Caudal 4; D. 58; P. 22; 3 rudimentary rays between the twolobes of the pectoral Uparina.E2. Pectoral dehcate, fringed membranaceous.'*D2. Mouth at an angle garmani.CK Color blackish bathybius."*B2. Teeth in a single series copei.A^. Pacific species.G'. Pectoral 30 or more.Hi. Teeth trilobed; mouth horizontal dactylosus.H^. Teeth simple; mouth oblique; pectoral notch veryshallow angustifrons.G^. Pectoral 25 or less.J'. Teeth in bands.K'. Mouth horizontal.L'. Middle pectoral rays long; pores not enlarged.M'. Middle pectoral rays evenly spaced on the girdle;little pectoral notch fimbriatus.M^. Middle pectoral rays more widely spaced than tho&eabove and below. '8 In key in more than one place. 164 BULLETIN 150, UNITED STATES NATIONAL MUSEUMNi. Color pale; skin transparent.0^. Mouth and gill cavity pale; peritoneum dotted.Pi. D. 56; A. 46; gill slit in front of 10 pectoralrays; southeast coast of Alaska deani.P2. D. 63; A 56; giU slit?; Okhotsk Sea..entochloris.O^. Mouth and gill cavity black; peritoneum black.melanobranchus.N2. Color black.Q'. Teeth conical, sharp ulochir.Q-. Teeth stout, blunt atramentetus.L". Middle pectoral rays rudimentary; pores on the headenlarged.R^ Stomach pale; gill slit? latifrons.R*. Stomach black ; gill slit extending down infront of 13 pectoral rays_-.holomelas.K2. Mouth at an angle.S'. Symphysis of pectorals behind front ofeye cephalus.S-. Symphysis of pectoral in front of eye.mento.J^. Teeth in a single row, stout.T'. Color rose red, fins and snout largelyblack rosaceus.T2. Color blackish.U'. Gulf of California grandiceps.U-. Off Panama attenuatus.PARALIPAUIS DACTYLOSUS GilbertParaliparis dactylosus Gilbert, 1896, p. 469, pi. 34. ? Gilbert and Burke,1912a, p. 82.Ty^e.?^o. 48616, U.S.N.M.; Alhatross Station 3112, off Cali-fornia; depth 29G fathoms. Length about 90 mm.Distribution.?Coast of Cahfornia and among the Aleutian Islands,Albatross Stations 3112 and 4781; depth 296-482 fathoms. Thespecimen from Station 4781, Bering Sea, is doubtfully placed withthis species. Four specimens examined.RelaiionsMps.?P. dactylosus differs from all the other species of thegenus in the trilobed teeth. In the character of the teeth, number ofpectoral rays, and pyloric coeca, this species is the most generalizedof the genus.Description oj type (mutilated).?Body deepest at the nape; thedorsal outline descending very rapidly for a short distance, thengradually to the attenuate tail. Head moderate; cheeks nearlyvertical; profile gradual from occiput to snout, then vertical to thelips. Mouth narrow, the angle reaching under front of pupil. Teethshort, stout, blunt, weakly trilobed, in moderate bands, arranged ineight oblique rows. Snout deep, abrupt. No nostril tube evident. E3"elarge, 3.3 in head; black; pupil large, oval. Gill slit described asextending down in front of three or four pectoral rays. Pores?Pyloric coeca 13, equal to the eye. REVISION OF THE FISH FAMILY LIPAMDAE 165Dorsal of very slender rays. Caudal absent. Pectoral deeplynotched; the two lobes connected by short, delicate rays; the upperedge of the pectoral on a level with the lower margin of the eye; thelower lobe mutilated, of about six delicate, elongate rays. Vent farforward; tip of mandible to vent 1.5 in the head.Color of flesh pale, black dotted ; mouth, gill cavity and peritoneumblack; stomach pale.Synopsis.?Dorsal 64; anal 59; pectoral 33; pyloric coeca 13-18.Depth 5.7 in the length without the caudal; head 4.8. Eye 3.7 in thehead; gill slit 4; lower lobe of the pectoral fin 1.5. Body elongate,attenuate posteriorly. Teeth stout, weakly trilobed. Gill slitextending down in front of four pectoral rays. Pectoral fin deeplynotched, not divided to the base. Body pale, probably pinkish in FiGURE 82. ? Pakaltparis dactylosus. Teeth from type life; mouth, gill cavity, peritoneum, and the vertical fins posteriorlyblack. Four specimens, reaching the length of 93 mm. or more.Remarks.?The specimen from Bering Sea, the characteristics ofwhich are given below, apparently differs in no essential respect fromthe type, though a close comparison between the two specimens hasnot been made.Note on a specimen from Station 4781, Bering Sea. Dorsal 64;anal 59; caudal 6; pectoral 34; pyloric coeca 18, one-half the eye.Depth 5.7 in length without caudal; head 4.8. Eye 3.7 in head;lower lobe of pectoral 1.5; caudal 2.7; gill sht 4; tip of mandible tovent 2.7.Teeth weakh' trilobed, in narrow bands; eight oblique rows in thehalf of each jaw. Gill slit extending down in front of four pectoral rays.Upper edge of pectoral on a level with lower margin of orbit; the91668?30 12 166 BULLETIN 150, UNITED STATES NATIONAL MUSEUMlower lobe of six exserted rays, 1.5 in the head; seven widely spaced,well-developed rays connecting the two lobes. Caudal connectedfor nearlj^ one-third its length to the anal fin; the dorsal and analconnections with the caudal more or less abrupt.In life red, the vertical fins posteriorly black.PARAUPARIS ANGUSTIFRONS GarmanParalipaHs angustifrons Garman, 1899, p. 119, pi. 27, fig. 4; pi. 28, fig. 4; pi. 29,fig. 3.Types.?Two; No. 28699, M. C. Z.; Albatross Station 3394, offPanama, Pacific Ocean; depth 511 fathoms. Length 108 mm.Distribution.?Known only from the type locality. Two specimensexamined.Relationships.?Differing from P. dactylosus in the simple teethand oblique mouth and from all the other species in the large numberof pectoral rays.Description oj the types.?Body as in P. mento; deepest at the nape,sloping down for a short distance and then retaining an even depth <7Figure 83.?Paraliparis angustifrons. Teeth from ttpeto near the middle of the body, much compressed. Head muchcompressed; occiput slightlj^ swollen; profile convex; cheeks vertical.Mouth not greatly enlarged, pointing obliquely upward, the anglereaching beneath the eye; the maxillary reaching beneath the posteriormargin of the eye; teeth simple but with shoulder present, short,rather slender, conical, slightly recurved, in narrow bands arrangedin oblique rows. Snout low, not projecting; lower jaw heavy; point-ing obliquely upward; jaws equal or the symphysis of the lowerjaw projecting. Nostril without a tube. Eye moderate, 4 in thehead, black; pupil round. Gill slit uncertain. Pores not enlarged;upper pore on the snout nearly on a level with the nostrils. Noprickles. Pyloric coeca 6.Upper edge of the pectoral on a level with the angle of the mouth;the symphisis under the front of the eye; pectoral not divided to thebase, apparently with a very shallow notch as described; the lowerrays elongate. Caudal connected for at least one-half its length tothe anal.Color dusky, darker along the base of the fins; mouth and peri-toneum black; stomach pale. REVISION OF THE FISH FAMILY LIPARIDAE 167Synopsis.?From the original description: D. 57; A. 53; P. 37.Depth 5.5 in the total length; head 6.5. Eye 3.3 in the head. Pec-toral broad, without a gap at the bases of the rays, but with a notchin the lower half caused by shorter rays ; ends of the rays prolonged asfilaments. Head and abdomen black; fins blacldsh; body over themuscular portions somewhat grayish. Total length 4.5 inches.Two specimens. PARALIPARIS FIMBRIATUS GarmanParaliparis fimbriaius Garman, 1892, Mem. Mus. Comp. Zool., vol. 14, No. 2,p. 9.?1899, Mem. Mus. Comp. Zool., vol. 24, p. 116, pi. 29, fig. 1; pi. D.fig. 3.Type.?'No. 28702, M. C. Z.; Albatross Station 3381, off Panama,Pacific Ocean; depth 1,772 fathoms. Length 4 inches.Distribution.?Known only from the type locality.RelationsMps . ?P. jimbriatus does not appear to be closely relatedto any known species. It appears to be distinct from all the species Figure 84.?Paraliparis fimbriatus. Teeth from type with the exception of P. angustifrons, in having the middle pectoralrays no more widely spaced than those above and below. It isdistinguished from P. angustifrons by the smaller number of pectoralrays, the angle of the mouth and in other characters.Description of type.?The type is so badly mutilated as to bevalueless; what remains appears to agree with the original descrip-tion and the figure. The following notes were taken.Body slender ; the dorsal outline not descending rapidly. Head low,broad, the width equal to the depth; profile low, convex; occiputnot swollen. Mouth wide; maxillary reaching beneath the posteriormargin of the orbit. Snout low, retreating, slightly projecting;the lower jaw included. Teeth slender, conical, simple, not recurved,in very narrow bands, apparently in a few oblique rows; the pulpcavity small. Condition of the nostril, gill slit, pores, and pyloriccoeca uncertain. No prickles.Pectoral slightly notched; the lower lobe hardly distinguishable;all of the rays evenly spaced on the girdle, elongate, filamentous at 168 BULLETIN 150, UNITED STATES NATIONAL MUSEUMthe tip. Head and body anteriorly blackish, paler posterioriy;peritoneum black.Original description abbreviated; D. 49; A. 45; P. 24. Depth7 in the total length; head 5.5. Eye 4 in the head. Head flattenedon the crown; nape not high. Snout short, nearly as long as the eye.Pectoral broad, notched somewhat by the shorter fifth and sixth ray,.each ray prolonged into a soft filament. Length 4 inches.PARAUPARIS DEANI BurkeParaliparis holomelas Evermann and Goldsborotjgh, 1907, p. 334, (part, notof Gilbert).Paraliparis deani Burke, 1912o, p. 571. ? Gilbert, 1915, p. 355.Ty2)e.?No. 60570, U.S.N.M.; Stephens Passage, southeast coastof Alaska, Albatross Station 4253 ; depth 188 fathoms. Length 68 mm. Figure 85.?Pakalipakis deani. Teeth from typeDistribution.?Southeast coast of Alaska; Albatross Stations 4194,4203, 4251, 4253, 4255, 4292, and 4293; depth 30-274 fathoms.Twenty-two specimens examined.Relationships.?P. deani is one of the few species of the genus inwhich the gill slit extends down in front of the pectoral fin. It ispossibly allied to P. entocMoris of the Okhotsk Sea, apparently differ-ing in the smaller number of dorsal and anal rays and possibly in thestronger dentition and the presence of prickles.Synopsis.?Dorsal 56-57 ; anal 44-48; pectoral 18-21 ; pyloric coeca9. Depth of body 4.5-6 in length without caudal. Lower lobe ofpectoral 1.2 in head. Gill slit extending down in front of 10 to 13pectoral rays. Teeth simple but with a well developed shoulder nearthe tip, somewhat arrow-shaped at the tip, short, slender, slightlyrecurved. Thumb-tack prickles present. Body pale; peritoneum REVISION OF THE FISH FAMILY LIPARIDAE 169 silvery; stomach black. A small sized species, reaching a length ofless than 100 mm.Remarks.?The remaining specimens examined vary somewhatfrom the type. The gill slit extends down in front of from 10 to 13pectoral rays. The eye varies from silvery to black. The absenceof silvery pigment in the eye may be due to the preserving fluid.In the two cotypes, as in the type, the lower lobe of the pectoral fin Figure 86.?Paraliparis deani. Prickles from type retains the epidermis in which are found thumb-tack prickles. Noneof the specimens have prickles scattered over the body. As the epi-dermis is absent from the bodies of all the specimens, prickles may havebeen present on the sides of the body in life. The caudal fin is eithertruncate of slightly concave. Specimens No. 60671 differ from theother specimens and may represent another species. They are largerwith shorter, deeper bodies and a paler coloration. Figure 87. ? Paraliparis entochloris. Teeth from typePARALIPARIS ENTOCHLORIS Gilbert and BurkeParaliparis entochloris Gilbert and Burke, 19126, p. 378.Type.~No. 73347, U.S.N.M.; Albatross Station 5018, OkhotskSea, off the coast of Saghalin; depth 100 fathoms. Length 98 mm. 170 BULLETIN 150, UNITED STATES NATIONAL MUSEUMDistribution.?KJiown only from the type locality. Two speci-mens examined.Relationships.?-For comparisons with P. deani see description oflatter species. PARAUPARIS LIPARINA (Goode)Aniiira liparina Goode, 1880, Proc. U. S. Nat. Mus., vol. 3, p. 487. ? Goode andBean, 1895, Special Bull. U. S. Nat. Mus. No. 2, p. 278, fig. 252, (part).?Jordan and Evermann, 1898, Bull. U. S. Nat. Mus., No. 47, p. 2138.Monometra liparina Goode, Proc. U. S. Nat. Mus., 1883, p. 109.Paraliparis liparina Gunther, 1887, Challenger Rept., vol. 22, p. 68. ? Garman*1892 Mem. Mus. Comp. Zool., vol. 14, No. 2, p. 82.Type.?No. 26184, U.S.N.M.; Fish Hawk Station 891, southeastof Long Island, New York; depth 480 fathoms. Type lost.Distribution.?Atlantic Ocean, soiitheastof New York. Fish HawkStation 891 and Albatross Station 2586; depth 328-480 fathoms. Figure 88.?Paraliparis uparina. Teeth from specimen No. 46002, U.S.N.M.Relationship.?The author has been unable to demonstrate thepresence of pseudobranchiae in the species and places it with Parali-paris; a pale colored species with the gill slit above the pectoral.Description of No. 46002.?Dorsal 58; anal 55; pectoral 22; caudal4; pyloric coeca 4. Depth 5.5; head 5.3. Eye 3.5 in the head.Body slender, not attenuate, tapering gradually to the caudal.Head rather slender; occiput slightly swollen; cheeks vertical; pro-file concave over the eye. Mouth moderate; lower jaw included.Teeth simple but with a shoulder sometimes indicated, short andstout, very blunt, tuberclelike, not recurved, in moderate bands.Snout deep, abrupt, slightly projecting. Nostril without a tube.Eye large, black; pupil round. Gill slit imcertain. One supra-branchial pore present. No prickles. Pyloric coeca 6, on the leftside.Anterior 12 dorsal rays unsegmented. Caudal slender, truncate,connected for two-fifths of its length to the anal. Pectoral deeplydivided; the upper lobe of 15 rays; the lower lobe of four elongatedrays; the space between the two lobes divided by three rudimentaryrays. REVISION OF THE FISH FAMILY LIPARIDAE 171Color pale, pigmented with brown dots; these gathered about themouth, along the back and toward the caudal and causing these regionsto become duskj^; peritoneum black; stomach pale.From the .type description: Dorsal 67; anal 54; pectoral 23;caudal 6. Depth 5.5, head 6.7. Orbit 5; snout 3.7.PARALIPARIS ULOCHIR GilbertParaliparis ulochir Gilbert, 1896, p. 441; 1915, p. 354.Type.?Lost, U. S. National Museum Collection. Albatross Sta-tion 3010, Gulf of California; depth 1,005 fathoms.Distribution.?Gulf of California and Bering Sea; Albatross Stations3010 and 3332; depth 406 to 1,005 fathoms.Relationships.?The distinguishing characters of the species arethe liigh pectoral, large number of pectoral rays, the gill slit abovethe pectoral, and the black coloration.Description oj No. 48699 U. S. N. M.?A poorly preserved speci-men from Station 3332, Bering Sea. Figure 89.?Paraliparis ulochir. Teeth from specimen No. 48699, U.S.N.M.Body tapering rapidly into an attenuate tail. Head short; inter-orbital flat; occiput slightly swollen, prominent. Teeth simple,short, conical, with a broad base, strongly recurved, sharply pointed,in narrow bands, arranged in oblique rows. Snout abrupt, deep.Nostril without a tube. Eye medium. Gill sUt above the pectoral.One suprabranchial pore present. No prickles. Pyloric coeca onthe left side, about 8. The upper edge of the pectoral on a level withthe upper margin of the eye; the fin apparently not divided to thebase; the lower lobe of short rays, bound in membrane nearly tothe tip; the two lobes connected by four more widely spaced rays.Skin absent; thebody pigmented with brown dots; mouth, gill cavity,peritoneum and the stomach black.From the original description: Dorsal about 65; anal about 60.Head 5 in the length. Eye 3 to dji in the head. Upper lobe of thepectoral extending beyond the anal origin; four or five somewhatwidely spaced, well-developed rays connecting with the lower lobe; 172 BULLETIN 150, UNITED STATES NATIONAL MUSEUMlower lobe of nine rays; none of the rays are free. Uniformly black,including mouth and bronchial cavity.PARALIPARIS MEMBRANACEOUS GuntherParaliparis membranaceous Gunther, 1887, p. 69, pi. 12, fig. D. ? Garman, 1892,p. 83.Hilgendorfia membranaceous Goode and Bean, 1895, p. 280.Ty^pe.?British museum, Challenger Station 310, off Cape SaintVincent; depth 400 fathoms. Length 60 mm.Distribution.?Known onlj'- from type locality. Not examined.Synopsis.?Dorsal about 70; anal about 70; caudal 2 or 3. Gillslit above base of pectoral fin. Teeth simple. A broad median dorsalfold arises from the top of the snout and is continued backward; ante-rior dorsal rays buried in this fold; pectoral broad, delicate, fringed.Body pale; peritoneum black.PARALIPARIS BATHYBIUS (Collett)Liparis bathybii Collett, 1878, p. 32; 1880, p. 52, pi. 2, fig. 14.?Lutken, 1898,p. 17.?Collett, 1905, pi. 2, fig. 9.Paraliparis bathybius Gunther, 1887, p. 68, pi. 12, fig. C. ? Garman, 1892, p. 81.Type.?Female; Arctic Ocean, west of Bear Island, NorwegianNorth Atlantic Expedition, 1878, Station 312; depth 658 fathoms.Length 208 mm.Distribution.?Arctic Ocean, region of the Faroe Island. Nospecimens examined.Relationships.?P. iathyhius agrees with P. holomelas in the reduc-tion of the middle pectoral rays and the coloration.Remarlcs.?A number of specimens have been recorded and de-scribed as belonging to the species. The descriptions of these latterspecimens differ greatly from the original description as regards thepectoral fin.*^Synopsis.?Dorsal 59; anal 51; pectoral 19; caudal 8. Teethsimple. Gill slit probably above base of pectoral. Pectoral findeeply notched; the presence of rudimentary rays connecting thetwo lobes uncertain. Color black.PARALIPARIS AUSTRALIS GilchristParaliparis australis Gilchrist, 1904, p. 107, pi. 7.Type.?Two; off Cape of Good Hope; depth 300 fathoms. Length50 mm.Distribution.?Known only from type locality. No specimensexamined.Relationship.?P. australis resembles P. garmani and P. bathybiusof the North Atlantic in the loss or great reduction of the rays con-necting the two lobes of the pectoral fin. It can readily be distin- '' See Gunther, Lutkin, and Collett. REVISION OF THE FISH FAMILY LIPARIDAE 173guished from P. garmani by the horizontal mouth and from P.hathyhius by the lighter coloration.Synopsis.?Dorsal 48; anal 43; pectoral 14-3; caucial 9. Gill slitprobably above base of pectoral. The two lobes of the pectoralseparated by a space free of rays. Color pale; visceral mass black.Only the types known.PARALIPARIS MELANOBRANCHUS Gilbert and BurkeParaliparis melanobranchus Gilbert and Burke, 19126, p. 378.Type.?^o. 73346, U.S.N.M.; Albatross Station 5029, OkhotskSea; depth 440 fathoms. Length 84 mm.Distribution.?Known only from the type locality.RelationsMps . ?The distinguishing characters of this species are thesmall number of pectoral rays, the gill slit above the pectoral, and thepale coloration.Description of type.?Dorsal 60; anal 53; pectoral 17; caudal 4;pyloric coeca 7. Depth 5.1; head 4.7. Eye 3.7; gill slit 6.1. Figure 90.?Paraliparis melanobkanchus. Teeth from typeBody deepest at the nape, tapering into an attenuate tail as in P.cepfialus. Head comparatively heavy; occiput swollen, prominent;profile straight from occiput to snout. Alouth horizontal; angle ofmouth under pupil. Teeth simple, short, stout, blunt, little recurved,in broad bands, arranged in about 12 oblique rows in the half of eachjaw; inner teeth not prominent, little larger than anterior teeth.Snout deep, abrupt, not projecting; jaws equal. Nostril with araised rim. Eye large, pupil round. Gill slit small, above thepectoral. One suprabranchial pore present ; pores on the head normal.No prickles. Pyloric coeca 7, on the left side.Anterior dorsal rays short, free from the membrane. Increasinggradually in length, eight or more unsegmented. About three analrays unsegmented. Caudal truncate, connected for one-third itslength to the anal. Pectoral notched deeply but not to the base; thetwo lobes connected by three widely spaced rays; lower lobe of fournearly equal rays, reaching midway between vent and anal, two in thehead. Vent in front of gill slit. 174 BULLETIN 150, UNITED STATES NATIONAL MUSEUMSkin transparent, sparsely dotted; flesh dusted with brown dots;chin dusky; gill cavity, mouth, peritoneum, and stomach black. ? PARALIPARIS LATIFRONS GarmanParaliparis laiifrons Garman, 1899, Mem. Mus. Comp. Zool., vol. 24, p. 118, pis.27-28.Types.?Seven; No. 28698, M. C. Z.; Albatross Station 3382,Pacific Ocean, off Panama; depth 1,793 fathoms. Length 145 mm.or less.Distribution.?-Known only from the type locality. Types exam-ined.Relationship.?-In the shape of the head and body and the enlargedpores P. latifrons resembles P. holomelas; for differences see descrip-tion of the latter species.Description of one of the types.?-Dorsal 53 ; anal 47 ; pectoral about23; caudal 6; pyloric coeca 7. Depth 5.1; head 4.3. Ej^e 4.3 inthe head. Figure 91.?Parauparis latifrons. Teeth from typeBody heavy and deep at the nape, tapering rapidly into the slen-der compressed tail. Head heavy and deep; depth 5.1; width 6.5;cheeks vertical; occiput high. Mouth broad; the lateral cleft ex-tending beneath the middle of the eye; maxillary extending beneaththe posterior margin of the orbit. Teeth simple, slender, notrecurved, arranged in narrow bands, the inward diverging rows veryoblique; inner teeth longer and comparative}}'" blunt. Snout deep,abrupt, not projecting. Nostril uncertain. Eye large, black; pupillarge, round. Gill slit vmcertain, the membrane macerated. Poreson the head enlarged as in P. holomelas; suprabranchial pore uncer-tain. No prickles. Pyloric coeca short, on the left side.Origin of dorsal above the base of the pectoral. Caudal slender,of four or six rays. Pectoral divided nearly to the base; the upperedge below the level of the orbit; the two lobes connected by widelyspaced rudmientary rays; the lower lobe of five elongate rays; thefirst ray shorter. In two specimens the eggs are 4 mm. in diameter. REVISION OF THE FISH FAMILY LIPARIDAE 175Color: Remnant of skin black, apparently uniformly black; fleshpigmented; mouth, gill cavity, and the peritoneum black; stomachpale.From the original description: D. 55; A. 47-49 P. 18-1-1-4 or 5.Head nearly 6 in the total length. Eye 3.3 in the head. Bonesof the head thin and fragile. Upper portion of the pectoral withabout 18 rays, separated from the lower lobe by two short rays; thelongest rays nearly as long as the head; the majority of the rays havefilamentous tips. Black on the side, abdomen, fins, and the lowersurface of the head; remainder blackish to clouded brownish. Length5 inches or more. PARALIPAEIS HOLOMELAS GilbertParaliparis holomelas Gilbert, 1895, Rept. U. S. Fish Comm., 1893, p. 441. ? Jordan and Evermann, 1898, Bull. U. S. Nat. Mus., No. 47, p. 2140.?EvERMANN and GoLDSBOROUGH, 1907, Bull. U. S. Fish Comm., vol. 26.p. 334, (part). FiGURK 92.?PARAUPARIS HOLOMELAS. TEETH FROM TYPEType.?Female, No. 48637, U.S.N.M.; Albatross Station 3332,north of Unalaska, Bering Sea; depth 406 fathoms.Distribution.?Bering Sea; Albatross Stations 3332 and 3308; depth406 to 1,625 fathoms. Type examined.Relationships.?Apparently closely related to P. latifrons; possiblydistinguished by the black stomach and the larger number of fin rays.Description of type.?Mutilated. Body deepest at the nape, slop-ing rapidly to the elongate slender tail; the trunk short, not equalto the head. Head heavy, deep, occiput high, swollen; profde notconcave over the eyes; interorbital high; cheeks nearly vertical;mouth broad; the lateral cleft extends nearly to the posterior marginof the eye, maxillary reaching a vertical from behind the posterior mar-gin of the orbit. Teeth as in P. latifrons, simple, not recurved, few,moderately elongate, bluntly rounded at the tip, in narrow bands; theoblique rows obscure; inner teeth prominent. Snout abrupt. Nostrilclose in front of the eye ; the character of the tube uncertain. Eye large,black, 4 in the head. Gill slit large, 2.4 in the head, extending downin froHt of 13 pectoral rays. Pores on the head enlarged, oval; one 176 BULLETIN 150, UNITED STATES NATIONAL MUSEUM suprabranchial pore; snout pores close together; the upper one nearlybetween the nostrils; the anterior pore reduced in size. No prickles.Pyloric coeca 8, on the left side, short, about equal to the pupil.Dorsal rays very slender, the anterior rays increasing gradually inlength, apparently not buried in tissue, the membrane absent. Cau-dal absent. Pectoral 23, deeply divided; the two lobes connectedby two or three rudimentary, widel}'- spaced rays; the upper lobelong, reaching past the front of the anal; the lower lobe of five fila-mentous rays free to their base, the longest reaching the anal, 1.2 inthe head, the shortest not half the longest; the rays not regularlygraduated in length as in Xipam. Snout to vent 4 in the head. Th&type is a female with eggs. Color black; peritoneum and stomachblack; pyloric coeca white; mouth and gill cavity dusky to black.From the original description: Head about 5; depth about 6..Dorsal 58 to 61; anal 54. Eye 3.7 in the head. Very large mucousslits on head, five forming a series from tip of snout below eye andacross cheeks, six along mandible and preopercle. No pseudobran-chiae. Pectoral inserted low, the upper edge below the level of the eye ,?:the two lobes distinct, the interspace without free membraneous mar-gin, the skin of the abdomen directly continuous at this point withthat of shoulder girdle; beneath the integument the interspace betweenthe lobes is provided with two or three short widely spaced rays, as in.all other species examined by us.PARAUPARIS GARMANI BurkeAmiira liparina Goode and Bean. 1895, p. 278 (p.art, confused with P. liparina)^Paraliparis garmani Burke, 1912o, p. 572.Type.?No. 64129, U.S.N.M.; Albatross Station 2586, off NewEngland; depth 328 fathoms. Length 141 mm.Distribution.?Atlantic Ocean off New England. Albatross Sta-tions 2212, 2586, 2676; Fish Hawk Stations 898, 937, 952, 994, 997,1093; depth 300 to 542 fathoms. Fifteen specimens examined.Relationships.?P. garmani does not closely resemble any knownspecies of the North Atlantic Ocean. It has been confused with P.liparina by various writers but can readily be distinguished from the-latter species by the oblique mouth and the more elongate, slender,,and recurved teeth. In these two characters P. garmani resemblesP. cephalus of the Pacific Ocean. It differs from the latter species,however, in the larger number of pectoral rays and the more rr.di-mentary condition of the middle pectoral rays.Synopsis.?Dorsal 54; anal 49 ; pectoral 21 ; pyloric coeca 6. Depthof body 4.6 in length without caudal. Mouth oblique, symphysis oflower jaw projecting. Teeth simple, comparatively stout, sharplypointed, recurved, in broad bands. Gill slit above the pectoral.Middle pectoral rays rudimentary, hidden beneath the skin; lowerlobe of the pectoral 2.9 in the head. No prickles. Color pale, dusky EEVISION OF THE FISH FAMILY LIPAEIDAE 177posteriorly; mouth and stomach pale; peritoneum black. Reachinga length of 141 mm.Remarlcs.?The head of the young resembles that of P. cephalus,the occiput being greatly swollen, not so pronounced in the adult.The maxOlary reaches from beneath the middle of the eye to theposterior margin. Tail very slender and attenuate, appearing short,slender, and attenuate in the young when contrasted with the deephead and trunk. Pores normal for the genus. Pseudobranchiaeabsent. About 10 of the dorsal rays unsegmented. Caudal com-paratively broad, truncate, connected for half its length to the anal,the dorsal and anal connections nearly equal. Pectoral divided into Figure 93.?Parauparis garsiani. Teeth from typetwo distinct lobes ; the upper edge on a level with the lower margin ofthe eye or below; the upper lobe of about 14 rays; the lower lobe ofthree or four elongate rays, not graduated in length as in Liparis,equal to twice the eye, the tips sometimes coiled; the space betweenthe two lobes bridged by four rudimentary rays.PARALIPARIS CEPHALUS GilbertParnUparis cephalus Gilbert, 1891, Proc. U. S. Nat. Mus., vol. 14, p. 561. ? JoRDON and Evermann, 1898, Bull. U. S. Nat. Mus., No. 47, p. 2141.?Gilbert, 1915, p. 354.Type.?Apparently lost, U.S.N.M.; Albatross Station 2892, offsouthern California; depth 284 fathoms. The type not seen by thewriter. Four cotypes are in the Stanford Museum, No. 21. 178 BULLETIN 150, UNITED STATES NATIONAL MUSEUMDistribution.?Coast of California and northward to Bering Sea;Albatross Stations 2892, 3112, 3126, 3330, 3348, and from Karluk;depth 284-455 fathoms. Sixteen specimens examined.RelationsMps.?Not closely related to any known species. Forcomparision with P. mento see description of the latter species.Description of cotypes.?No. 21, S. U. Z. M. Body deep at theunion with the head; the dorsal outline descending rapidly to theattenuate tail; distance from tip of opercle to posterior end of bodycavity equal to the snout and orbit. Head 4.5 in the body length,deep, compressed; cheeks vertical; occiput greatly swollen. Mouthat an angle, pointing obliquely upward; maxillary reaching beneaththe posterior margin of the orbit. Teeth simple, slender, sharplypointed, recurved, in narrow bands; the inwardly diverging rows veryoblique and difficult to count. Snout short, rising abruptly to thenasal region; jaws nearly equal; the symphysis of the lower jaw sharp, Figure 94. ? Paraupakis cephalus. Teeth from specimen No. 5785, StanfordUniversity Zoological Museumprojecting. Nostril without a distinct tube. Eye moderate, 4 orless in the head. Gill slit described as being above the base of thepectoral, in all our specimens the membrane is torn. Pores appearingenlarged when the sldn is absent; only one suprabranchial pore pres-ent. No prickles. Pyloric coeca 9, on the left side.Origin of the dorsal over the end of the first third of the pectoralfin. Caudal very slender, elongate, of three rays, equal to the snoutand orbit, connected for one-third its length to the anal. Pectorallow, the upper edge below the orbit and above the angle of the mouth,deeply notched but not to the base; the lower lobe consists of threeexserted half free rays; the upper lobe consists of eight rays whichform the body of the lobe and three shortened, not rudimentary,widely spaced rays which span the notch; these three rays are equalto one-half the length of the longest ray in the upper lobe; symphysisof the pectoral under the posterior border of the orbit. Vent just in REVISION OF THE FISH FAMILY LIPARIDAE . 179front of the gill slit; distance from tip of lower jaw to vent 5.5 in thelength of the body.Flesh pinkish, dotted; mouth and gill cavity pale, dotted; skinthin, transparent; stomach and peritoneum black.Remarks.?Mouth and gUl cavity sometimes blackish ; peritoneumsometimes only heavUy pigmented, possibly black in hfe. TheKarluk specimens appear to have the mouth more horizontal andmay belong to a different species.PARALIPARIS MENTO GilbertParaliparis menlo Gilbert, 1891, Proc. U. S. Nat. Mus., vol. 14, p. 562. ? Jordan and Gilbert, 1898, Bull. U. S. Nat. Mus., No. 47, p. 2142.?Gil-bert, 1915, p. 354.Type.?^o. 44298, U.S.N.M.; Albatross Station 3071, coast ofWashington; depth 685 fathoms. Length 87 mm.Distribution.?Off Washington and California. Type examined. 'A ir Figure 95.?Paraliparis mento. Ttpe, showing pecuuar shape of headRelationships.?Not closely related to any known species; resem-bling P. cepTialus; distinguished from the latter species by the moreelongate body, the extremely low pectoral, and in minor characters.Description of type.?Dorsol 57; anal 52; pectoral 16; caudal 4;pyloric coeca 8. Depth 6; head 5.7 Eye 3; gill slit 2.5.Body slender, compressed, deepest at the union with the head,dorsal outline descending rapidly for a short distance, attenuateposteriorly; visceral cavity short; distance from tip of gUl flap tothe posterior end of the visceral cavity equal to the snout and one-half the eye. Head deep, compressed; occiput swollen; cheeksvertical, profile from upper lip to occiput almost straight. Mouthoblique, large; the angle beneath the e5^e; maxillary reaching beneaththe posterior margin of the pupU, a little more than one-half the head.Teeth simple, rather slender, with a broad base, lanceolate, recurved,in a narrow band in the lower jaw, arranged in two or three veryoblique rows; inner teeth larger and recurved; teeth in the upper jawin a moderate band, the oblique rows close set. Snout retreating fromthe upper lip; lower jaw heavy, projecting upward; the symphysisvery prominent. Nostril apparently without a tube projecting above 180 BULLETIN 150, UNITED STATES NATIONAL MUSEUMthe skin. Eye moderate, black; pupil large, round. Gill slit large,above the pectoral, extending obliquely forward and downward; theopercular arm supporting the gUl flap directed downward. Poresuncertain. No pricldes. Pyloric coeca 8, on the left side.Origin of dorsal above the gill slit; the anterior rays not buried intissue. Caudal very slender, connected for one-half its length to theanal. Pectoral very low, the upper edge below the angle of themouth; the base of the pectoral almost horizontal, the middle underthe pupil; the pectoral symphysis under the chin; the lower lobe ofelongate, partly free rays, coUed at the tips; the longest ray 0.9 inthe head; the anterior ray one-half the longest; the rays not regularly Figure 96.?Parai.ipakis MtiNio. Teeth from typegraduated in length as in Liparis; the upper lobe very slender,reaching the anal, the shortened rays widely spaced, distinct. Snoutto vent 9.8 in the length; vent below the pupil . Color: Skin absent; flesh pinkish; peritoneum black; color prob-ably pale or pinkish in life.PARALIPARIS ATRAMENTATUS Gilbert and BurkeParaliparis atramentatus Gilbert and Burke, 19126, p. 377.Type.?Mede, No. 73345, U.S.N.M.; Albatross Station 4971, eastcoast of Hondo, Japan; depth 649 fathoms. Length 76 mm.Distribution.?Known only from the type locality.RelationsJdp.?Apparently with the branch ending in the P. rosaceusgroup in which the teeth are stout and in a single row; dift'ering fromall the species with the teeth in bands by the extremely heavy teeth. REVISION OF THE FISH FAMILY LIPARIDAE 181PARALIPARIS COPEI Goods and BeanParaliparis copei Goode and Bean, 1896, Special Bull. U. S. Nat. Mus., No. 2,p. 279, fig. 253. ? Jordan and Evermann, 1898, Bull. U. S. Nat. Mus., No. 47,p. 2143.Type.?'No. 35637, U.S.N.M.; Albatross Station 2237, southeast ofLong Island; Depth 520 fathoms. Length 178 mm.Distribution.?Atlantic Ocean off the New England coast; FishHawk Station 898; Albatross Stations 2186, 2237, 2586, 2722; depth300-594 fathoms. Ten specimens examined. tf-c^^^^gi; 'A,.FiGVRE 97.?Paraliparis atramentatus. A typical species of the genus. Type Figure Paraliparis atramentatus. Teeth from typeRelationships.?P. copei closely resembles P. rosaceus of the Pacific.The two species appear to differ in the length of the head, size of eye,and the distance from snout to vent. The material is not sufficient formaking satisfactory comparisons.Description of type.?Dorsal 60; anal 55; pectoral 20; pyloric coeca6. Depth 6.5; head 6.5. Eye 3.6; gill slit 6.8.Body comparatively heavy, compressed, elongate^but not attenuate.Head wide, blunt; cheeks nearly vertical; occiput slightly swollen;interorbital comparatively wide; 2.9 in the head. Mouth with little91668?30 13 182 BULLETIN 150, UNITED STATES NATIONAL MUSEUMlateral cleft; maxillary reaching beneath pupil. Teeth simple, stout,wedge-shaped, in a single series in both jaws, forming a sharp cuttingedge, more widely spaced and weaker at the symphysis; apparentlyfewer teeth in the upper jaw; pulp cavity small. Snout deep, abrupt,little if at all projecting. Nostril with a raised rim. Eye large, black,pupil round. Gill slit above the base of the pectoral. One suprabran-chial pore present; upper pores on the snout nearly between thenostrils. No prickles. Pyloric coeca 3.2 in the head, on the left side.Origin of dorsal slightly behind the middle of the pectoral; theanterior rays buried in tissue, depressed. Caudal of 6 rays, connectedfor one-half its length to the anal: the dorsal and anal connectionsgradual. Pectoral notched ; the two lobes connected by rudimentarywidely spaced rays ; the lower lobe of three or four rays ; the lower rayone-half the second, elongate; the lower lobe reaching past the vent,1 .6 in the head. Vent under the base of the pectoral. Figure 99.?Parauparis copei. Teeth from specimen No. 46009, U.S.NiM.Body milky white anteriorly, light brown posteriorly; snout andchin blackish; gill cavity, a strip back of the vent, and the peritoneumblack; stomach pale. PARAUPARIS ROSACEUS GilbertParaliparis rosaceus Gilbert, 1890, Proc. U. S. Nat. Mus., vol. 13, p. 93. ? Garman1892, Mem. Mus. Comp. Zool., vol. 14, No. 2, p. 80. ? Jordan and Ever-MANN, 1898, Bull. U. S. Nat. Mus., No. 47, p. 2142.Type.?'No. 48918, U.S.N.M.; Albatross Station 2919, off southernCalifornia; depth 984 fathoms. Length 135 + mm.Distribution.?Known only from the type locality. Type examined.Relationships.?See description of P. copei; differing from P.attenuatus and P. grandiceps in the lighter coloration.Description of type.?Pectoral 20 ; caudal 4 ; pyloric coeca 5. Depth5.7; heads. Eye 4.8; gill slit 6.5.Body comparatively deep, compressed, moderately elongate, notattenuate, tapering gradually. Head not heavy or deep ; occiput notswollen; cheeks vertical. Mouth horizontal, small, with little lateralcleft; maxillary reaching beneath the pupil. Teeth short, stout,appearing more bluntly rounded than in P. copei, those in the lowerjaw in one row, forming a single cutting edge, teeth in upper jaw nearly REVISION OF THE FISH FAMILY LIPARIDAE 183 all absent, probably similar to the teeth in the lower jaw. Snoutdeep, rising steeply, not projecting. The nostril tube apparently notprojecting above the surface of the head. Eye small, black; pupilround. Gill slit above the pectoral. One suprabranchial pore.Pyloric coeca about 5.Origin of the dorsal over the middle of the pectoral; the anteriorrays buried in tissue beneath the fin membrane. Caudal multilated,apparently connected for one-half its length to the anal. The upperedge of the pectoral on a level with the lower margin of the eye; thetwo lobes connected by rudimentary rays, a short space of the girdlebare as though one ray had been torn away; the lower lobe of fourslender rays, 1.8 in the head. Snout to vent 5.9 in the length of thebody. Figure 100.?Paraliparis rosaceus. Teeth from typeColor pinkish; the head and fins black; mouth, gill cavity, and theperitoneum black ; stomach pale.PARAUPARIS GRANDICEPS GarmanParaliparis grandiceps Garman, 1899, Mem. Mus. Comp. Zool., vol. 24, p. 117,pi. 29, figures 4-4c,Type.?Female, No. 28701, M. C. Z.; Albatross Station 3434,Gulf of California; depth 1,588 fathoms. Length 10 inches.Distribution.?Known only from the type locally. Type examined.RelationsMps . ?P. grandiceps is closely related to P. attenuatus.See description of the latter species.Description.?Note on the type which is in a mutilated condition.Teeth resembling those in P. copei and P. rosaceus, on the lower jawin a single series, crowded together at the symphysis into two ormore rows, progressively larger at the sides ; the teeth in the upperjaw not seen.Pectoral divided nearly to the base; the space between the twolobes bridged by four widely spaced rudimentary rays; outline of finas figured for P. bathybius in Plate 12 of the Challenger Report. 184 BULLETIN 150, UNITED STATES NATIONAL MUSEUMPeritoneum black; stomach pale.From the original description: D. about 56; A. about 42; P. 15-1-1-1-1-5. Depth 5 in the total length; head about 6. Orbit lessthan three in the head. Evidently the form and proportion as in P.hathyhius. Body compressed; caudal section thick, deep anteriorlyand for some distance back of the body, then decreasing in depthrapidly and becoming thin and slender. Head as broad and deep aslong, subquadrangular in transsection across the orbit; interorbitalflattened. Cheeks swollen. Snout blunt, slighty overhanging the FiGUEE 101.?Parauparis grandiceps. Teeth trom typemouth. Mouth wide. Maxillary extending below the entire orbit.Teeth small, short, robust, blunt, subcorneal, with broad bases,firmly attached, in one series on each jaw except near tip of mandiblewhere a second row may be present, larger toward the angle of themouth.Upper lobe of the pectoral of 15 rays, connected with the lowerlobe by four short rays; lower lobe of five rays, nearly as long as theupper lobe; the four short rays occupying a space equal to the orbit,connected by membrane. Total length 10 inches. Color dark brownor black. BEVISION OF THE FISH FAMILY LIPARIDAE 185PARALIPARIS ATTENUATUS GarmanParaliparis atiemiatus Garman, 1S99, Mem. Mus. Comp. Zool., vol. 24. p. 118.Type.?'No 28700, M. C. Z.; Albatross Station 3364, off Panama,Pacific Ocean; depth 902 fathoms. Length 3 inches.Distribution.?Kno's\Ti only from the type locaUty. Type examined.Relationships.?P. attenautus resembles P. rosaceus in the charac-ters of the head, teeth, and body. It differs from the latter speciesin the darker coloration and from P. grandiceps apparently in the lar-ger number of dorsal and anal rays.Description.?Note on the type which is in a mutilated condition.Body slender, resembling P. rosaceus. Head as in P. copei, the profileevenly rounded; the occiput not swollen; depth of head greated thanwidth. Mouth small, horizontal; maxillary reaching below thepupil. Teeth apparently in single rows as in P. copei and relatedspecies. Snout not projecting. Eye moderate, about 4 in the head,black. The character of the gill slit and the pores can not be ascer-tained. No prickles. Pyloric coeca, few, short. Pectoral notchednearly to the base, the two lobes connected by three rudimentaryrays; the fin consisting of about 24 rays of which 4 are in the lowerlobe. Caudal narrow, acuminate.Skin absent, probably dusky or black; mouth, gill cavity, and theperitoneum black; stomach pale.Note from the original description: D. 66; A. 57; P. 17-1-1-1-4.Depth 6 in the total length. Head less than 7. Head one-fourthdeeper than wide, high at the nape. Mouth medium, horizontal.Teeth small, acuminate, in a single series. Color black, or lighteningto brown on the posterior half. Total length 3 inches or more.Genus RHINOLIPARIS GilbertRhinoliparis Gilbert, 1895, Rept. U. S. Fish Comm., 1893, p. 445.Disk absent; nostril single; teeth simple or trilobed; one supra-branchial pore; pyloric coeca present; barbels on the snout;pseudobranchiae absent; branchiostega rays 6.Genotype.?Rhinoliparis harhulifer.RHINOUPARIS BARBULIFER GilbertRhinoliparis harhulifer Gilbert, 1895, Rept. U. S. Fish Comm., 1893, p. 445. ? Gilbert and Burke, 19126, p. 379.T^/pe.?Female, No. 48576, U.S.N.M.; Albatross Station 3325,Bering Sea north of Unalaska; depth 284 fathoms. Length 110 mm.Distribution.?North Pacific, off California, east coast of Hondo,Okhotsk Sea and Bering Sea; Albatross Stations 3325, 3327, 3329,3331, 4540, 5019, 5028, 5039, 5043, 5044, 5045, 5050; depth 192-551fathoms. Forty specimens examined. 186 BULLETIN 150, UNITED STATES NATIONAL MUSEUMRelationships.?R. harhulifer is not closely related to any knownspecies.Description of type.?Pectoral 20 ; caudal 3 ; pyloric coeca 7. Depth6.6; head 5.3. Eye 3; gill slit 3.6.Body slender, becoming very attenuate posteriorly. Head com-paratively heavy; occiput slightly swollen; cheeks vertical; profilerising gradually from the snout. Mouth narrow; the lateral cleftextending to below the middle of the pupil; the maxillary reachingbeneath the posterior margin of the eye. Larger teeth more or lessstrongly trilobed, not close set, in broad bands; the inner teeth FiGTTBE 102.- -Rhinoliparis barbuluee. Teeth from specimen from Albatross Station5043, OFF Hokkaido, Japanlarger; nine oblique rows in the half of the lower jaw, distinct; theoblique rows in upper jaw indistinct. Snout low, projecting beyondthe upper lip for the diameter of the pupil; two barbels on the tipof the snout, separated by three-fifths the length of the barbels;the length of the barbels about one-half the eye. Nostril tube notprojecting above the skin. Eye large, black, far forward, the anteriormargin above the tip of the lower jaw; pupil large, round. Gillslit extending down in front of five pectoral rays. Pores not enlarged;one suprabranchial pore present; the upper snout pore above thebarbel ; the anterior snout pore low ; below and slightly to one side of REVISION OF THE FISH FAMILY LIPARIDAE 187the barbel. No prickles. Pyloric coeca short, three-fourths of theeye, on the left side.Origin of dorsal above gill slit; tips of anterior rays not connectedwith skin. Caudal of three rays, slender, truncate, 2 in the head,connected for one-fourth its length to the anal. The upper edgeof the pectoral on a level with the pupil; the fin deeply notched;the space between the two lobes with two very widely spaced rudi-mentary rays hidden beneath the skin; the upper lobe not quitereaching the anal; the lower lobe reaching halfway to anal, 2 in thehead, of three elongate, nearly equal rays, free nearly to the base,one short rudimentary ray in front, suggesting that the anteriorrays have been lost; two-fifths of the length of the pectoral appearsto be beneath the lax skin.Color: Skin transparent; flesh dotted, dusky on the nape and alongthe base of the dorsal and anal; peritoneum, mouth, and gill cavityblack; stomach dusky. Figure 103.?Rhinouparis attent;atus. Teeth from typeRHINOUPARIS ATTENUATUS BurkeRhinoliparis attenuatus Burke, 1912a, p. 573. ? Gilbert, 1915, p. 357.T?/pe.?No. 28377, M. C. Z. Bering Sea, Albatross Station 3326;depth 576 fathoms.Relationships.?RJiinoliparis attenuatus constitutes the secondspecies in the genus. It does not appear to be closely related toRhinoliparis harhulifer and can zeadily be distinguished from thelatter by the larger number of barbels on the snout and the morehighly modified type of dentition.Description oj type.?Body as in R. harhulifer, low, extremelyattenuate. Head 5.7 in the length of the body without the caudal,depressed; width of head greater than depth of head; profile low,nearly straight from snout to occiput; interorbital flattened. Mouthbroad; maxillary reaching vertical from posterior margin of eye. 188 BULLETIN 150, UNITED STATES NATIONAL MUSEUMTeeth in narrow bands, elongate, slender, slightly recurved, simple orthe lateral lobes faintly indicated on some of the teeth. Snoutdepressed, broad, projecting as in R. harhulijer, extending beyond theupper lip for three-fourths the diameter of the eye. The snout issomewhat mutilated but at least three barbels are present on eachside; one of these is situated near the tip of the snout and two near thelateral margin of the lower surface. Nostril with a raised rim. Eyelarge, 4 in the head. Gill slit apparently above the base of the pec-toral fin. Pyloric coeca about 12. One suprabranchial pore.Caudal fin mutilated but apparently reduced to a single elongateray, connected for a short distance with the dorsal and anal fins.Pectoral fin with a shallow notch; the middle rays well developed,widely spaced; the lower lobe of six elongate mainly free rays; theupper lobe of 16 rays.The dermis absent, probably dusky or blackish; flesh pinkish;mouth dusky; peritoneum black; stomach pale or slightly dusky.Genus ACANTHOLIPARIS Gilbert and BurkeAcantholiparis Gilbert and Burke, 1912a, p. S3.Disk absent; nostril single; teeth simple; one suprabranchial pore;pyloric coeac absent; pseudobranchiae absent; opercular arms pro- Figure 104.?Acantholiparis opercularis. Type, showing opercular spines found in noother specils of the familyjecting as strong spines from sides of head; gill flap supported by theposterior arm of the suboperculum; branchiostegal rays 6.Genotype.?Acantholiparis opercularis.ACANTHOLIPARIS OPERCULARIS Gilbert and BurkeAcantholiparis Gilbert and Burke, 1912a, p. 83.Type.~No. 74390, U.S.N.M.; Albatross Station 4797, off Starit-schkof Island, Kamchatka; depth 682 fathoms. Length 76 mm.Distribution.?North Pacific, Albatross Stations 4761 and 4797;depth 682 to 1,973 fathoms.Description of the type.?Dorsal 45; anal 39; pectoral 26; caudal 10;pyloric coeca absent. Depth 7; head 4.6. Eye 4.8; snout 3.Head and body as in C. ectenes; body slender, depressed. Headbroad, depressed, flat. Snout low, wide, projecting over the broadmouth. Mouth broad; maxillary reaching below the posterior mar-gin of the pupil. Teeth fine, conical, in moderate bands, the tip REVISION OF THE FISH FAMILY LIPARIDAE 189 slightly recurved. Nostril in a slender tube. Eye moderate; pupiloval. Gill slit above the pectoral. One suprabranchial pore; uppersnout pore far forward toward the tip of the snout. No prickles.Pyloric coeca absent in cotype, apparently absent in the species.Origin of dorsal a short dis-tance behind the base of thepectoral; caudal slender, strong,1.2 in the head; connected forone-fourth its length to the anal.Pectoral broad, reaching theanal, unnotched, anterior raysprogressively shortened as in thelower lobe of the fin in Liparis;anterior rays free nearly to thebase. Vent midway betweenthe tip of the lower jaw and the \anal fin.Color dusky to grayish, palerposteriorlv ; lower surface of the ^"^^"^^ ioo.-acaxthouparis opercularis. teeth . 1 ? 1 FROM COTYPEhead, snout,mouth, interor bital,pectorals, and the abdomen dusky; peritoneum dusky to black;stomach dusky to pale. In life dusky reddish throughout.Genus NECTOLIPARIS Gilbert and BurkeNcctoUparis Gilbert and Burke, 1912a, p. 82.Disk absent; nostril single; teeth simple; gill slit restricted to thefront of the pectoral; one suprabranchial pore; pyloric coeca present;vent horizontal, on the throat, opening forward; branchiostegal rays5; one of the anterior rays lost; pseudobranchiae absent.Genotype . ?Nedoliparis pelagicus . NECTOUPARIS PELAGICUS Gilbert and BurkeNedoliparis pelagicus Gilbert and Burke, 1912a, p. 82; 19126, p. 380.Type.?'No. 74389, U.S.N.M.; Albatross Station 4785, oft" AttnIsland, Bering Sea, depth 400 fathoms. Length 30 mm.Distribution.?North Pacific, southern California to Hokkaido,Japan; Albatross Stations 4252, 4257, 4258, 4333, 4539, 4541, 4765,4767, 4781, 4785, 4800, 5032, and 5039; depth intermediate, 300 to609 fathoms. Fifteen specimens examined.RelationsJiips . ?Resembling Paraliparis in the reduction of thepectoral rays, and coeca and the character of the teeth. Widelydivergent in the loss of one of the branchiostegals.Description of the type.?Dorsal 53; anal 48; pectoral 19; caudal 6;pyloric coeca 8; depth 4.8; head 4.2. Eye 2.7 (from a cotype). 190 BULLETIN 150, UNITED STATES NATIONAL MUSEUMBody comparitively robust, dorsal outline descending very grad-ually to the caudal. Head compressed; cheeks vertical; interorbitalflat; occiput deep; profile rounded. Mouth terminal, pointingobliquely upward; maxUlary reaching below the pupil. Teeth ex-tremely small, mere conical tubercles, difficult to see with a hand lens,in a single series, not close set; possibly absent at the symphysis; Figure 106.?Nectoliparis pelagicus. Pectoral girdle with five bkancheostegal rays not seen in upper jaw. Snout short, abrupt, 3 in the head; jawsnearly equal, the lower jaw appearing larger when the mouth is open.Nostril with a raised rim. Eye large, prominent; pupU round. ThegUl slit is confined to the front of the upper pectoral lobe and extendsin front of 14 rays. One suprabranchial pore; upper snout pore high, Figure 107.?Nectoliparis pelagicus. Type. Showing location of vent and pecularpectoralhigher than the nostril; nostril nearly between the two snout pores;six pores on the maxillary and about the eye, six or more in a series onthe mandible and cheeks. No prickles. Pyloric coeca 8, one-halfthe eye, on the left side.Dorsal origin far back, over the tip of the pectoral. Caudaltruncate, of four or six rays, connected for nearly one-half the length ^ A. JL REVISION OF THE FISH FAMILY LIPARIDAE 191to the anal. Pectoral divided to the base; the lobes far apart, sepa-rated by nearly the diameter of the eye; the upper lobe reachingnearly to the anal, of 13 or 14 well-developed rays and two or threerudimentary, widely spaced rays, these not bridging the gap betweenthe two lobes ; lower lobe of four rays, short, (variable in the cotypes, ofthree or four rays, sometimes elongate, as long as the upper lobe ; therays sometimes half free,not graduated in length),the anterior ray shorter.Vent tar lorward on tne figure los.?nectoliparis pelagicus. teeth from typechin; snout to vent 2 inthe head; vent in front of the base of the lower pectoral lobes, above ahorizontal forward projecting fold with the anal papUla at the tip,this fold filled by the rectum and the oviduct. In one specimen 18rays in the dorsal and 12 in the anal are unsegmented.Color pale, heavily pigmented with black dots; skin transparent,dotted on the head and the posterior part of the body; flesh moreheavily pigmented, sometimes very dark; abdomen and the cheekssometimes silvery; stomach and coeca pigmented, mouth, gill cavity,and the peritoneum black. Reaching a length of 64 mm.APPENDIXIncluded in the appendix is one species which is doubtfully placedwith the Liparidae and a number of species described since the presentreport was completed. The writer deems it advisable not to attemptto include these species in the body of the text because in some casesthe descriptions are incomplete and specimens are not available forexamination. Attempting to determine the relationships of theLiparids with incomplete descriptions and without specimens is ahopeless task. LIPARIS OSBORNI Townsend and NicholslAparis osborni Townsend and Nichols, 1925, p. 14, fig. 4.Described from California.GYMNOUCODES EDWARDSI VaillantGymnolicodes edwardsi Vaillant, 1888, pp. 313 and 387, pi. 26, fig. 3.?BotrL-ENGER, Zool. Record, 1888, Pisces, p. 18. ? Goode and Bean, 1895, p. 281-fig. 254.The relationships of this species are unknown. I do not believethat it belongs with the Discoboli. It differs from any known deep-sea Liparid in having the vent far back near the anal fin. The pectoralfin also differs widely from that of any known Liparid.CAREPROCTUS DUBIUS ZugmayerCareproctus dubius Zugmater, 1911, p. 9.Described from Spitzbergen. 192 BULLETIN 150, UNITED STATES NATIONAL MUSEUMCAREPROCTUS BURKEI Jordan and ThompsonCareprodus burkei Jordan and Thompson, 1914, p. 281, pi. 34, fig. 2 and 2a.Type.?No. 6457 ; Carnegie Museum, a female 92 mm. in total length,and a female cotype, 83 mm. long, both from Yokohama market,doubtless outside the heads of Awa and Bashu.CAREPROCTUS JORDANI, new nameCareprodus gilberti Jordan and Thompson, 1914, p. 282, pi. 34, fig. 1 and la.This species was described as Careprodus gilberti by Jordan andThompson in 1914. Unfortunately the present writer (1912a) hadused the same specific name for another species of Careprodus. Wetherefore submit the name jordani for this species and repeat Jordanand Thompson's original description:Description from the tj'pe. No. 6456, a specimen much shrunken by strongalcohol, from Misaki, Sagami Bay, Japan, 16 cm. in total length. The measure-ments are given in hundredths of body length to correspond with the compre-hensive paper recently published by Gilbert and Burke on the Japanese Cyclo-gasteridae.Head .25; width of same .17; centerocular width .10; width at angle of mouth .12; length of snout .09; eye .06; maxillary .12; width of gill slit .06; depth ofbody .19; distance from snout to disk .19; from snout to anus .34; snout toanal fin .42; snout to dorsal insertion .28; transverse diameter of disk .08;distance disk to anus 7; longest ray in upper pectoral lobe 17, in lower lobe 12;D. 45; A. 38; P. 32.Body elongate, compressed; head deeper than wide; dorsal profile moststrongly curved on snout; jaws subequal; maxillary extending to below anteriorborder of pupil; eyes high, reaching dorsal profile as in Careprodus curilanus;nostril at level of upper edge of pupil; teeth coarse, in bands in both jaws, simple,not triolbed, depressible. Gill opening extending to second pectoral ray. Pectoralnot deeply notched, rays of lower lobe slightly produced as filaments. Anteriordorsal rays half length of posterior rays; latter .10 of body length; former withtips slightly produced; first anal ray .07, last .10; dorsal and anal adnate tocaudal by one-fourth its length; last ray in each slightly shorter than precedingray. Disk beginning under, or slightly behind, posterior margin of eye and reach-ing to below point of opercle, separated by its own transverse diameter from anus,which resembles that of Careprodus pycnosoma in position.Color faded or lacking. Peritoneum jet black.CAREPROCTUS SARASA TanakaCareprodus sarasa Tanaka, 1916, p. 173.CAREPROCTUS PUNCTULATUS TanakaCareprodus pundulatus Tanaka, 1916, p. 173.CAREPROCTUS OKADAE TanakaCareprodus okadae Tanaka, 1916, p. 173.CAREPROCTUS MEDERI SchmidtCareprodus mederi Schmidt, 1915, p. 628, figs. 7-8.Okhotsh Sea. REVISION OF THE FISH FAMILY LIPARIDAE 193PARALIPARIS ALBESCENS GilbertParaliparis albescens Gilbert, 1915, p. 355.Type.?Cat. No. 75816, U.S.N.M.; a female with well-developedova, 63 mm. long, from Station 4515, Monterey Bay. Only the typeknown. PARALIPARIS CAUDATUS GilbertParaliparis caudatus Gilbert, 1915, p. 355.Tyjpe.?Csit. No. 75815, U.S.N.M.; 86 mm. in total length, (77 mm.to base of caudal), from Station 4527, Monterey Bay. Only the typeknown. Figure 109.?Paeauparis caudatus. Type. One of the shorter, deeper bodied speciesRESEMBLING SPECIES OF CaREPROCTUSResembling P. ulochir in the restricted gill slit and the wide pec-toral, but differing among other characters in the shorter, deeperbody, and the wider, many-rayed caudal fin. In P. ulochir, the trunktapers to an extremely slender tail, the fin composed of but four rays.In P. caudatus, the base of the caudal has an easily appreciable widthand the fin contains 10 or 11 rays.PARAUPARIS ANTARCTICUS ReganParaliparia antarciicus Regan, 1914, p. 11.PARAUPARIS TERRAE-NGVAE ReganParaliparis terrae-novae Regan, 1916, p. 129, pi. 1, fig. 6.The original description is as follows:A young fish, 35 mm. long, was taken in McMurdo Sound on January 16, 1912,at Station 332, 77? 15' S., 166? 0' E., 0-550 meters. It is very similar in mostrespects to Paraliparis antaclicus Regan, but differs notably in the fewer fin rays,(dorsal 55; anal 43) and in the form of the pectoral fin, which has no elongatelower rays. This is the second Antarctic species of the genus.PARALIPARIS WILDI WaiteParaliparis mldi Waite, 1916, pi. 4, fig. 1.The following is from the original description : Locality?The only specimen was taken at Station X., lat. 65? 6' S., long. 96?13' E.; off Shackleton Iceshelf, in 325 fathoms, the temperature being 1.65 C,and the bottom ooze. 194 BULLETIN 150, UNITED STATES NATIONAL MUSEUMRemarks.?This species differs from P. antadicus Regan, the only other speciesdescribed from Antarctica, in the following characters: The eye is slightly smallerand the snout longer; the pores on the head are more numerous, occurring on thesnout, none being so figured in P. antarcticus. The shape of the vertical finsis difi'erent, while the caudal is partly enclosed by the dorsal and anal membranes;the vent occupies a forward position, behind the lower portion of the pectoral, asis usual in these forms, while Regan's statement that these lower portions reachthe vent would indicate that the latter is placed quite near the origin of the analfin, and therefore, in a position normal to the majority of fishes. P. wildi appearsto be more nearly allied to the northern P. bathyhii CoUett, the type of the genus.Addendum.?P. terrae-novae Regan differs more from P. antarctica than doesP. wildi, possessing fewer fin rays, and having no elongate lower rays in the pec-toral. Three species are now known from the Antarctic.Type in the South Australian museum.Genus LIPARISCUS GilbertLipariscus Gilbert, 1915, p. 358.Allied to Paraliparis and Nedoliparis, agreeing in habit with theformer, and with the latter in having but five branchiostegal rays,while all other genera in this family have six. FiGUEE 110.?Lipariscus nanus. TypeNo trace of a ventral disk. Vent posterior in position, lying in thearea between the lower pectoral lobes. Pectoral fin greatly reducedbut the two lobes connected, not separate and distinct as in Nedol-iparis. Teeth simple, in narrow bands. Branchiostegals 5. GUIslit narrow, confined to the suprapectoral region.Genotype.?Lipariscus nanus Gilbert.LIPARISCUS NANUS GilbertLipariscus nanus Gilbert, 1915, p. 358.Type.?C&t. No. 75817, U.S.N.M.; 47 mm. long, from Station4461, Monterey Bay 285 to 357 fathoms.One paratype from the type locality, and three others from Station4468, Monterey Bay, 32 to 309 fashoms.BIBLIOGRAPHY AND LITERATUREThis bibhography is designed to include the most important worksrelating to the Liparidae. It is not the result of an attempt toinclude all the lists and discussions of the species described. It isreasonably complete concerning recent descriptions and notes. Fora more complete list of older publications see Garman, 1892. REVISION OF THE FISH FAMILY LIPARIDAE 195Adams, C. C.1902?Southeastern United States as a Center of Geographical Distributionof Flora and Fauna. Biol. Bull., vol. 3, No. 3, p. 115.Allis.1910?Cranial Anatomy of Mailed-Cheeked Fishes. Zoologica. Heft 57.Ayres, W. O.1855?Description of L. pulchellus and L. mucosus. Proc. Cal. Acad. Sci.,vol. 1, p. 24. Second edition, 1873, p. 22.Bean, T. H.1881a?Partial Bibliography of the Fishes of the Pacific Coast of the UnitedStates and of Alaska, for the Year 1880. Proc. U. S. Nat. Mus.,vol. 4, p. 312.1881&?Descriptions of New Fishes from Alaska and Siberia. Proc. U. S.Nat. Mus., vol. 4, p. 144.1881c?A Prehminary Catalogue of the Fishes of Alaska and AdjacentWaters. Proc. U. S. Nat. Mus., vol. 4, p. 241.1884?Notes on a Collection of Fishes made in 1882 and 1883 by Capt.Henry E. Nichols, U. S. Navy, in Alaska and British Columbia,with a Description of a New Genus and Species, Prionistius macellus.Proc. U. S. Nat. Mus., vol. 6, p. 353.1890?New Fishes Collected oflf the Coast of Alaska and the Adjacent RegionSouthward. Proc. U. S. Nat. Mus., vol. 13, p. 37.Bean, T. H., and Bean, B. A.1896a?Fishes Collected at Bering and Copper Islands by Nikolai A. Greb-nitski and Leonhard Stejneger. Proc. U. S. Nat. Mus., vol. 19,p. 237.18966?Notes on Fishes Collected in Kamchatka and Japan by LeonhardStejneger, with a Description of a New Blenny. Proc. U. S.Nat. Mus., vol. 19, p. 381.Burke, C. V.1911?The Relation between the Coloration and the Bathymetrical Dis-tribution of the Cyclogasteridae. Science, new ser., vol. 34, No. 875,October 6, 1911.1912a?A New Genus and Six New Species of Fishes of the Family Cyclo-gasteridae. Proc. U. S. Nat. Mus., vol. 43, p. 567.19126?Note on the Cyclogasteridae. Ann. Mag. Nat. Hist., 1912, ser. 8,vol. 9, p. 509.COLLETT, R.1878?Fiske fra Nordhavs-Expeditionens sidste Togt, Sommeren 1878.Chra. Vid. Selsk. Forh., 1878, No. 14.1879? Meddelelser om Norges Fiske 1 Aarene 1875-1878. Chra. Vid.Selsk. Forh., 1879, No. 1.1880?The Norwegian North Atlantic Expedition, 1876-1878. Fishes.1902? Meddelelser om Norges Fiske, Aarene 1884-1901, Chra. Vid. Selsk.Forh., 1902, No. 1.1905?"Fiske Under, Michael Sars," Togter i Nordhavet. 1900-1902.Rept. Norweg. Fish. Mar. Invest., vol. 2, No. 3.Dana, C.1853?On an Isothermal Oceanic Chart, Illustrating the GeographicalDistribution of Marine Animals. Amer. Journ. Sci. and Arts,ser. 2, vol. 16, p. 153, 314.DOLLO, L.1904?Resultats du Voyage du S. Y. Belgica en 1897-1899. Poissons.Anvers, 1904. 196 BULLETIN 150, UNITED STATES NATIONAL MUSEUMElGENMANN, C. H.1905?Divergence and Convergence in Fishes. Biol. Bull., vol. 8, p. 59,Contbr. Zool. Lab. Indiana Univ., No. 64.1909?Cave Vertebrates of America. Carnegie Inst, of Wash., Pub. Xo. 104.EvERMANN, B. W., and Goldsborough, E. L.1907?Fishes of Alaska. Bull. U. S. Bureau Fish., vol. 26, p. 219.Fabricius.1780?Fauna Groenlandica.Fischer, J. G.1S85?tjber Fische von Siid-Georgien. Jahrb. Hamburg, wissensch.Anst., vol. 2.Flint, J. M.1905?A Contribution to the Oceanography of the Pacific. Bull. U. S.Nat. Mus., No. 55.Franz, V.1910?Die japanischen Knochenfische der Sammlungen Haberer undDoflein. Verlag der K. B. Akademie der Wissenschaften. Mun-chen, 1910.Garman, S.1892?The Discoboli. Mem. Mus. Comp. Zool., vol. 14, No. 2.1899?The Fishes. Mem. Mus. Comp. Zool., Vol. 24.Gilbert, C. H.1890?A Preliminary Report on the Fishes Collected by the Steamer Al-batross on the Pacific Coast of North America during the year 1889.Proc. U. S. Nat. Mus., vol. 13, p. 42.1891?Descriptions of Thirty-Four New Species of Fishes Collected in 1888and 1889, Principally Among the Santa Barbara Islands and theGulf of California. Proc. U. S. Nat. Mus., vol. 14, p. 539.1896?Ichthyological Collections of the U. S. Fish Commission SteamerAlbatross during the years 1890 and 1891. Rept. U. S. Fish. Comm.1893, p. 393.1915?Fishes Collected by the United States Fisheries Steamer Albatrossin Southern California in 1904. Proc. U. S. Nat. Mus., vol. 42,p. 305.Gilbert, C. H., and Burke, C. V.1912a?Fishes from Bering Sea and Kamchatka. BuU. U, S. Bur. Fish.,vol. 30, 1910, No. 754.19126?New Cyclogasterid Fishes from Japan. Proc. U. S. Nat. Mus.,vol. 42, p. 351.Gilbert, C. H., and Thompson, J. C.1905?Notes on the Fishes of Puget Sound. Proc. U. S. Nat. Mus., vol. 28,p. 973.Gilchrist, J. D. F.1904?South African Fishes. Mar. Invest. South Africa, vol. 2.Gill, T.1864?Synopsis of the Cyclopteroids of Eastern North America. Proc. Phila.Acad. Sci., vol. 16, p. 189.1872?Arrangement of the Families of Fishes. Smith. Misc. Coll., No. 247.1873?Catalogue of the Fishes of the East Coast of North America. Smith.Misc. CoU., No. 283.1875?On the Geographical Distribution of Fishes. Ann. and Mag. Nat.Hist., ser. 4, vol. 15, p. 251. The Nation, vol. 24, pp. 27 and 42.1882?Bibliography of the Fishes of the Pacific Coast of the United Statesto the End of the Year 1879. BuU. U. S. Nat. Mus., No. 2. REVISION OF THE FISH FAMILY LIPARIDAE 197Gill, T.?Continued.1889?On the Classification of tlie Mailed-Cheeked Fishes. Proc. U. S.Nat. Mus., vol. 2, p. 567.1891?On the Relations of the Cyclopteroidae. Proc. U. S. Nat. Mus.,vol. 13, p. 361.1893?Families and Subfamilies of Fishes. Mem. Nat. Acad. Sci., vol. 6, p.125.1907?The Lumpsucker; Its Relationships and Habits. Smith. Misc. Coll.,vol. 50, pt. 2, p. 175.GiRARD, C.1858?Pac. R. R. Rept., vol. 10, pt. 4, Fishes.GooDE, G. B.1880?Fishes from the Deep Water on the South Coast of New EnglandObtained by the United States Fish Commission in the Summer of1880. Proc. U. S. Nat. Mus., vol. 3, p. 467.GooDE, G. B., and Bean, T. H.1879?Description of a New Species of Liparis (L. ranula) obtained by theUnited States Fish Commission off Halifax, Nova Scotia. Proc. U.S. Nat. Mus., vol. 2, p. 46.1895?Oceanic Ichthyology. Special Bull., U. S. Nat. Mus., No. 2.Gregory, W. K.1907?The Orders of Teleostomous Fishes. Ann. N. Y. Acad. Sci., vol. 17,pt. 2, No. 3, p. 437.GtJNTHER, A.1861?Catalogue of Fishes, vol. 3.1887?Deep-Sea Fishes. Challenger Report, vol. 22.Herdman, W. a.1896?Oceanography, Bionomics and Aquiculture. Smith. Rept., 1895, p.433.Holway.Cold Water Belt along the West Coast of the United States. Univ. Cal. Pub.Geol., vol. 4, p. 263.Jensen, A. S.1904?Fishes of East Greenland. Med. om Gronland, vol. 29.Johnsen, S.1919?Ichthyologiske Notiser 1. Bergens Museums Aarbok 1918-1919Naturvidenskabelig Raekke Nr. 6.Johnson, R.1907?The IndividuaUty and Variation of the Pyloric Coeca of the Cen-trachidae. Trans. Wis. Acad. Sci., vol. 15, pt. 2, p. 713.Jordan, D. S.1893?Temperature and Vertebrae. Wilder Quarter-Centuary Book;1901?The Fish Fauna of Japan, with Observations on the GeographicalDistribution of Fishes. Science, new ser., vol. 14, No. 354, p. 545.1905a?The Origin of Species Through Isolation. Science, new ser., vol. 22,p. 545.19056?Guide to the Study of Fishes.1908?Geminate Species. Amer. Nat., vol. 42, No. 494.Jordan, D. S., and Evermann, B. W.1898?Fishes of North and Middle America. Bull. U. S. Nat. Mus., No. 4791668?30 14 198 BULLETIN 150. UNITED STATES NATIONAL MUSEUMJordan, D. S., and Gilbert, C. H.1882?Synopsis of the Fishes of North America.1899?The Fishes of Bering Sea. Fur Seal Rept., pt. 3, p. 433.Jordan, D. S., and Snyder, J. O.1900?A List of Fishes Collected m Japan by K. Otaki, and by the UnitedStates Steamer Alhaiross, with Descriptions of Fourteen NewSpecies. Proc. U. S. Nat. Mus., vol. 23, p. 335.1901?A Preliminary Check List of the Fishes of Japan. Ann. Zoo!. Japan.,vol. 3, pts. 2 and 3.1902?A Review of the Discobolous Fishes of Japan. Proc. U. S. Nat. Mus.,vol. 24, 343.1904?On a Collection of Fishes made by Mr. Alan Owston in the DeepWaters of Japan. Smith. Misc. Coll., vol. 46, p. 230.Jordan, D. S., and Starks, E. C.1904?List of Fishes Dredged by the Steamer Albatross off the Coast ofJapan in the Summer of 1900. Bull. U. S. Fish Comm., vol. 22, p.577.Jordan, D. S., and Thompson, W. F.1914?Record of Fishes obtained in Japan in 1911. Mem. Carnegie Mus.,vol. 6. No. 4.Kendall, W. C.1909?Fishes of Labrador. Proc. Port. Soc. Nat. Hist., vol. 2, p. 207.1910?Report on the Fishes Collected by the Mr. Owen Bryant on a Trip toLabrador in the Summer of 1908. Proc. U. S. Nat. Mus., vol. 38,p. 503.Kroyer, H.1862?Nogle Bidrag til Nordisk Ichthyologi. De Gronlandske Liparis-Arten.Nar. Tidssar., (3), 1 B.LONNBERG, E.1899?Notes on the Fishes of the Swedish Arctic Expedition to Spitzbergenand King Charles Land in 1898. Bihang till K. Svenska Vet-Akad.Handlengor, vol. 24, Afd. IV, No. 9.1905?Fishes of the Swedish south polar expedition. Wissench. Ergeb.Schwed. Sud. polar Exped. 1901-1903. Vol. 5, Lieferung 6.1907?Fische. Hamburger Magalhaensesche Sammelreise.LtJTKEN, C. F.1862?Anledning of Hr. Professor H. Kroyers Kritik of mine Bemaerkningerom Liparis lineatus. Nat. Foren. Vid. Medd., 1861, pi. 7.1886?Et Bidrag til Kundskab om Kara-Havets Fiske. Dijmphna TogtetsZoologisk Udbytte.1898?The Ichthyological Results. Danish Ingolf Expedition, vol. 2 pt. 1.Merriam, C. H.1894?Laws of Temperature Control. Nat. Geogr. Mag., vol. 6.Murray, J.1898?The General Conditions of Existence and Distribution of MarineOrganisms. Smith. Rept., 1896, p. 397.Ortmann, a. E.1904?Origin of the Deep Sea Fauna. Internat. Geogr. Cong., vol. 8, p. 618.Pallas, P.1811?Zoogr. Rosso-Asiat., vol. 3.Putman, F. W.1874?Notes on Liparis and Cyclopterus. Proc. Amer. Assoc. Adv. Sci.,vol. 22, p. 337. REVISION OF THE FISH FAMILY LIPARIDAE 199Regan, C. T.1914?Diagnosis of New Marine Fishes Collected by?the British Antarctic("Terra nova") Expedition. Ann. Mag. Nat. Hist., ser. 8,vol. 13.1916?Larval and Post larval Fishes; British Antarctic ("Terra nova")Expedition, 1910. Zool., vol. 1, No. 4.Richardson, J.1836?Fauna Boreali Americana., pt. 3., The Fish. London.RUTHVEN, A. G.1908? Variations and Genetic Relationships of the Garter Snakes. Bull.U. S. Nat. Mus., No. 61.RUTTER, C.1899?Notes on a Collection of Tide Pool Fishes from Kadiak Island, inAlaska. Bull. U. S. Fish Comm., vol. 18, p. 189.Ryder, J. A.1886?On the Value of Fin Rays and their Characteristics of Development inthe Classification of the Fishes, Together with Remarks on theTheory of Degeneration. Proc. U. S. Nat. Mus., vol. 9, p. 71.Schmidt, P.1904a?Pisces Marium Orientalium.19046?On the Liparis (Trismegistus) owstoni Jordan and Snyder. Proc.U. S. Nat. Mus., vol. 28, p. 189.1915?Ichthyological notes, IL On a New Cj^clogasterid Fish with aRudimentary Disk. Ann. Mus. Zool. Russ., vol. 20.SCOFIELD, N. B.1906?List of Fishes Obtained in the Waters of Arctic Alaska. Fur SealRept., vol. 3, p. 493.Smith, H, M., and Pope, T. E. B.3 906?List of Fishes Collected in Japan in 1903, with Descriptions of NewGenera and Species. Proc. U. S. Nat. Mus., vol. 31, p. 459.Smitt, F. a.1893?A Historj' of Scandinavian Fishes, pt. 1. Stockholm.Starks, E. C.1896?List of Fishes Collected at Port Ludlow, Washington. Proc. Cal.Acad. Sci., ser. 2, vol. 6.Tanaka, S.1908a?Notes on Some Japanese Fishes, with Descriptions of Fourteen NewSpecies. Journ. Col. Sci., Imp. Univ. Tokyo, vol. 23, art. 7.19086?Notes on a Collection of Fishes made by Professor Ijima in theSouthern Parts of Saghalin. Ann. Zool. Japan., vol. 6, pt. 4.1916?Dobuts, Z. Tokyo, vol. 28, 1916, pp. 173-174.TowNSEND, C. H. and Nichols, L. T.1925?Deep-Sea Fishes of the Albatross Lower California Expedition. Bull.Amer. Mus. Nat. Hist., vol. 52.Vaillant, L.1888a?Miss. Sci. Cape Horn, 1882-83. Vol. 6, Zoologie-Poissons.18886?Expeditions Scientifiques du Travailleur et du Talisman, 1880-1883.Poissons.Vernon, H. M.1898?Relations between Marine Animal and Vegetable Life. Proc. Roy.Soc, vol. 63, p. 155.Waite, E. R.1916?Fishes, Australasian-Antarctic Expedition, 1911-1914. ScientificReports, ser. C, vol. 3, pt. 1. INDEX [The black-face numbers indicate specific headings]PageAciiDtholiparis - 2, 188Distribution 23Pectoral fin 19Relationships 42Acantholiparis opercularis 188Pyloric coeca - 12Amitra 2Anal fin {see Genera).-- ISAppendix -- 191Bathymetrical distribution 28Careproctus 105Liparis - 55Paraliparis 161Bathyphasma 2Bibliography -._ 194Body {see Genera) - 7Branchiostegals 12Careproctus - 2,7.9,95Anal fin 99Body 95Caudal fin - 99Coloration. 101Disk - 20,100Distribution 22,23,103Bathymetrical 28,105Closely related species 105Regions -.. 104Dorsal fin 13,98Dorsal notch 98Eye 96Giant species 103Gill slit 10,97Habits -.- 102Head 95Ivcy to species 105Modification of characters 95Nostril 9,96Pectoral fin 19,99Pores 9,96Prickles 12,101Pseudobranehiac 100Pyloric coeca 101Relationships 46Specific characters 103Suprabrachial pores _.. 10Teeth 11,97Temperature 34Vent 100Ciixeproctus abbreviatus 128Pectoral lobe 100Disk.. 100Ciireproctus aeanthodes 135Prickles 101Careproctus attenuatus 116Dorsal notch 98Vent 100 PageCareproctus bathycootus 139Disk- 100Nostril 96Pores 97Careproctus bowersianus 115Pores ..- -- - 97Pseudobranehiac - --- 100Careproctus burkei 192Careproctus colletti 148Disk 100Distribution 104, 105Eye 96Pectoral fin 19,100Pseudobranehiac 100Teeth - -- 11,97Careproctus curilanus --- 109Head -- 96Pores - -- 97Careproctus cypselurus 127Caudal fin 19,99Coloration - - 102Distribution 104,105Pectoral notch 99Pseudobranchiae 100Careproctus dubius 191Careproctus ectenes 119Careproctus entargyreus- -.- 146Coloration 102Prickles 101Pupil - --- 96Careproctus entomelas 145Coloration 102Prickles --- - 101Careproctus falklandica 116Careproctus furcellus.- _ 126Caudal fin - 99Pectoral notch 99Careproctus georgianus - 122Careproctus gilberti 138Coloration -.. 102Disk 100Distribution -. 104, 105Pseudobranchiae 100Teeth 97Careproctus homopterus 117Careproctus jordani 193Careproctus longifilis. 130Pectoral fin 19, 100Careproctus longipinnis 134Disk 100Careproctus mederi... 192Careproctus melanurus 125Distribution 105Pseudobranchiae ..., 100Careproctus micropis.. 121201 202 INDEXPageCareproctus mollis.. 113Distribution 105Careproctus okadae__ 193C areproctus opisthotremus 133Careproctus ostentum 140Disk 100Gill slit 97Prickles 101Careproctus ovigerum 131Habits 22Careproctus pallidus 116Careproctus pellucidus 134Pores 97Careproctus phasma 123Distribution _ 105Careproctus punctulatus 192Careproctus pycnosoma 108Body 95Coloration 101Dorsal notch 8,98Dorsal rays 98Nostril 96Pores 97Vent 100Careproctus ranula 118Distribution 105Careproctus rastrinus 136Prickles 101Careproctus reinhardi. 121Distribution 105Careproctus rhodomelas 118Disk 100Food 102Pectoral lobe 100Teeth 11Careproctus roseofuscus , 141Careproctus sarasa 192Careproctus segaliensis 115Interorbital 66Pores 96Careproctus simus 112Careproctus sinensis 114Pores 96Careproctus spectrum.. 123Coloration 102Distribution. 105Eye 96Pseudobranchiae 101Careproctus trachysoma 137Caudal fin (see Genera) 18Pores 101Prickles 96,97Center of dispersal ._ 26Liparis 56Collecting... 4Coloration (see Genera) 20,34Cottidae-. 14, 15Crystallias 150Relationships 41Crystalhas matsushimae 150Crystallichthys 2, 147Distribution 23Relationships. 41Crystallichthys cyclospilus 147Coloration 21Crystallichthys mirabilis 148Coloration 21 PageCyclopteridae 15,38Descriptions 3Diagram of characters 8Discoboli.. 5-6Disk (see Genera) _. 20Distribution 22Bathymetrical 28Center of dispersal 26Regions 24Species (see Genera) _ 23Dorsal fin 12Dorsal notch (see Genera) 13Elassodiscus 2, 154Distribution.. 23Relationships 41Elassodiscus tremebundus 154Eye 9Genera.. 1,5Key 42Relationships 39Gill sht -- 10Gymnolicodes edwardsi 191Gyrinichthys... 152Gill sht 10Relationships 41Gyrinichthys minytremus.. 153Habits.. 21Hexagrammidae 14Hilgendorfia 2Keys:Genera (see Genera) 42Lateral line 9Liparidae 4, 14,38Liparis 2, 43Anal fin. 47Body 43Caudal fin 47Coloration 50Disk-. 20,48Distribution 22, 51, 104Bathymetrical 54Closely related species 54Regions.. 54Dorsal fin 12,45Dorsal rays 13Eye 44Food 50Gill slit 10Habits -. 50Key to species 56Modification of characters .. _ . 43Nostrils 43Pectoral fin 19,47Pores 9,44Prickles 12,49Pseudobranchiae U, 48Pyloric coeca 12, 49Relationships 39Specific characters 51Suprabranchial pores 10Teeth 11,44Temperature 53Vent 20,48Liparis agassizii. ''6Coloration 50Distribution 51Eye .- ii INDEX 203PageLiparis agassizii:Pseudobranchiae 48Pyloric coeca 12,49Liparis antarctica. 85Pectoral fin 48Prickles.. 49Pyloric coeca.. 12Liparis atlanticus 59Distribution 56Dorsal rays 46Habits... 21Prickles 49Liparis bristolense 72Dorsal notch 45Prickles 49Liparis callyodon 62Distribution 53,54Nostril 44Prickles 49Pyloric coeca 49Liparis curilensis. 62Liparis cyclopus 69Body 43Distribution 53Pseudobranchiae 48Liparis cyclostigma 79Liparis dennyi.. 77Coloration 50Distribution 54,55Dorsal notch 46Nostril 44Liparis frenatus 70Liparis fucensis 71Distribution 54,55Pectoral fin 47Liparis gibbus 79Distribution 53Pseudobranchiae 48Liparis grebnitzkii. 04Liparis hersehehnus 75Pseudobranchiae !... 48Liparis ingeus 81Prickles 49Teeth 44Liparis liparis 67Distribution 56Habits 22Pyloric coeca.. 12Liparis major 9,90Body 43Coloration. 50Distribution 56Nostril 9,43,44Pectoral fin 48Posterior nostril -. 9Teeth 44Liparis megaeephalus 73Body 43Prickles. 49Lipai'is micraspidophorus 67Prickles 49Liparis montagui 61Habits 22Liparis mucosus 64Distribution 51, 54Prickles 49 PageLiparis ochotensis 80Prickles 49Liparis osborni.. 191Liparis owstoni ?7Body... 43Coloration 50Dorsal fin 47Nostril 43Pectoral fin 48Prickles 49Liparis pulchdUus 88Coloration 50Distribution 54Dorsal fin 46,47Pectoral fin. 43Pseudobranchiae 48I/iparis rhodosoma 10,81Body 43Coloration 50Prickles 49Liparis rutteri 61Dorsal rays 46Nostril 44Liparis steineni 84Pectoral fin 43Liparis tanakae 86Coloration 50Dorsal rays 13,46Pectoral fin... 43Prickles 49Pseudobranchiae 43Liparis tessellatus 9,89Coloration 50Distribution 52Dorsal fin 47Nostril 9,43,44Liparis tuaicatus 73Distribution 56Dorsal rays 13,46Pseudobranchiae... 48Lipariseus 2, 194Branchiostegals 12Lipariseus nanus 194Loricati 14,15Modifications in the deep sea 35Modification of characters 7Liparidae 7Anal fin 13Body 7Branchiostegals 12Caudal fin 18Coloration 20,34Disk 20Dorsal fin 12Eye 9Gill slit 10Head 8Pectoral fin 18Pseudobranchiae 11Pyloric coeca 12Teeth 10Vent 20Careproctus. 95Liparis 43ParaUparis 155 204 INDEXPageNectoliparis 2,189Branchiostegals.. 12Distribution... 22Pectoral fin 19Relationships... 42Suprabranchial pores.. 10Vent 20Nectoliparis pelagicus 189Gill slit . 10Habits 21Teeth... 11Neoliparis 1,45Number of genera and species 5Paraliparis. 7,9, 154Anal fin 157Body 155Caudal fin.... lo7Coloration. 159Distribution 22,23,160Bathymetrical 28-161Regions 161Dorsal fin 13Eye 156Giant species 161Gill slit 10, 156Habits 159Head 155Key to species... 162Modification of characters 155Nostril 9-155Pectoral fin 19-157Pores 10-156Pricliles 12-159Pseudobranchiae 11-159Pyloric coeca 159Relationsliips 41Specific characters 160Teeth 11-156Temperature 34-161Vent 20-159Paraliparis albescens 193Paraliparis angustifrons 166Paraliparis antarcticus 193Paraliparis atramentatus .-. 180Teeth 156Paraliparis attenuatus 185Paraliparis australis - 172Pectoral fin 158Paraliparis bathybius 172Pectoral fin 158Paraliparis caudatus 193Paraliparis cephalus 177Dorsal fin... 157Pectoral fin 158Paraliparis copei-- - 181Dorsal fin 157Paraliparis dactylosus 164Teeth 156Paraliparis deani 168Dorsal fin 157Prickles.. 159Paraliparis entochloris 169Paraliparis fimbriatus 167Pectoral fin 158Paraliparis garmani 176 PageParaliparis grandiceps 18SParaliparis holomelas 17SBody 155Gill slit 168Pectoral fin ijgPores 156Paraliparis latifrons 174Paraliparis liparina 170Pseudobranchiae 157Paraliparis melanobranchus 17$Pectoral fin. 158Paraliparis membranaceous 178Pectoral fin 19, 159Paraliparis mento ntBody 155Pectoral fin 168Paraliparis rosaceus 188Coloration 159Pectoral fin 158Teeth II, 156Paraliparis terrae-novae 19$Paraliparis ulochir 171Distribution 161Paraliparis wildi 19SPectoral (see Genera) J9Polypera 91Key to species 92Pyloric coeca 12Relationships 39Polypera beringianus 93Polypera greeni 98Polypera simushirao. 91Pores (see Genera) 9, 10Preservation 5Prickles (see Genera) 12Prognurus iPseudobranchiae (?ee Genera) 11Pseudotissue 7Pyloric coeca {see Genera) 12Regions 24Careproctus 104Liparis 64Paraliparis 161Rhinoliparis.. 7, 185Distribution 23Relationships 41Rhinoliparis attenuatus 187Rhinoliparis barbulifer 185Scorpaenidae 14Sexual demorphism. 7Species 2-5Distribution (see Genera).. 25Bathymetrical 30Giant forms 23Specific characters 8Careproctus 103Liparis --- 51Paraliparis.- 160Teeth (see Genera) 10Temnocora. - 2, 14SRelationships 40Temnocora Candida - 9, W6Temperature (see Genera). - 34Trismegistus 49Vent (see Genera) 20o SMITHSONIAN INSTITUTION LIBRARIES 3 9088 01421 2294