The Familial Phylogeny of the Tetraodontiformes (Acanthopterygii: Pisces) as Evidenced by Their Comparative Myology

RICHARD WINTERBOTTOM

m 4 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY • NUMBER 155 SERIAL PUBLICATIONS OF THE SMITHSONIAN INSTITUTION The emphasis upon publications as a means of diffusing knowledge was expressed by the first Secretary of the Smithsonian Institution. In his formal plan for the Insti- tution, Joseph Henry articulated a program that included the following statement: "It is proposed to publish a series of reports, giving an account of the new discoveries in science, and of the changes made from year to year in all branches of knowledge." This keynote of basic research has been adhered to over the years in the issuance of thousands of titles in serial publications under the Smithsonian imprint, com- mencing with Smithsonian Contributions to Knowledge in 1848 and continuing with the following active series: Smithsonian Annals of Flight Smithsonian Contributions to Anthropology Smithsonian Contributions to Astrophysics Smithsonian Contributions to Botany Smithsonian Contributions to the Earth Sciences Smithsonian Contributions to Paleobiology Smithsonian Contributions to Zoology Smithsonian Studies in History and Technology In these series, the Institution publishes original articles and monographs dealing with the research and collections of its several museums and offices and of professional colleagues at other institutions of learning. These papers report newly acquired facts, synoptic interpretations of data, or original theory in specialized fields. These pub- lications are distributed by mailing lists to libraries, laboratories, and other interested institutions and specialists throughout the world. Individual copies may be obtained from the Smithsonian Institution Press as long as stocks are available.

S. DILLON RIPLEY Secretary Smithsonian Institution SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY • NUMBER 155

The Familial Phylogeny of the Tetraodontiformes (Acanthopterygii: Pisces) as Evidenced by Their Comparative Myology

Richard Winterbottom

SMITHSONIAN INSTITUTION PRESS City of Washington 1974 ABSTRACT Winterbottom, Richard. The Familial Phylogeny of the Tetraodontiformes (Acanthopterygii: Pisces) as Evidenced by Their Comparative Myology. Smith- sonian Contributions to Zoology, number 155, 201 pages, 185 figures, 1974.—The myology of 46 species of the order Tetraodontiformes is described, the species representing the 11 families initially recognized for the purposes of this study. The data gathered from the descriptions has been interpreted phylogenetically, using only those characters of a shared specialized (synapomorph) nature. This analysis suggests that the Triacanthodidae and Triacanthidae are sister groups, and separated off first from the ancestral stock. The remaining families are divided into two lines. On the one hand, the Balistidae form the sister group of the Monacanthidae, and together form the sister group of the Aracanidae and Ostraciidae. In the other lineage, the Triodontidae separated off first, fol- lowed by the Molidae. The Diodontidae form the sister group of the Tetra- odontidae plus Canthigasteridae. In the final analysis, such fossils as are considered pertinent to the under- standing of tetraodontiform phylogeny are considered. The information pro- vided by these fossils, combined with the results of the myological interpretations, suggest the following systematic changes in the hierarchical classification of tetraodontiforms in order to bring it into line with the proposed phylogeny: (1) that two suborders be provisionally recognized, the first containing the Triacanthodidae and Triacanthidae, the second consisting of two superfamilies; (2) that a superfamily Balistoidea be recognized to contain the families Balistidae, fSpinacanthidae, and Ostraciidae; (3) that the families Monacanthidae and Aracanidae be treated as subfamilies of the Balistidae and Ostraciidae respec- tively, and that the fSpinacanthinae of Tyler (1968) be removed from the Triacanthodidae and placed as the sister group of the Balistidae, with equal rank (as fSpinacanthidae); (4) that a superfamily Tetraodontoidea be formed to contain the families tEoplectidae, Triodontidae, Tetraodontidae, Diodontidae, and Molidae, and, should -fZignoichthys prove to be further along the gymnodont line than fEoplectus, as hinted by Tyler (1972) and tentatively preferred here, the fZignoichthyidae; (5) that the family Canthigasteridae be reduced to sub- familial rank within the family Tetraodontidae. and that the fEoplectinae of Tyler (1972) be transferred from the Triacanthodidae to the superfamily Tetrao- dontoidea and raised to familial rank, in keeping with its phylogenetic position.

OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is recorded in the Institution's annual report, Smithsonian Year. SI PRFSS NUMBER 4823. SERIES COVER DESICN: The coral Montastrea cavernosa (Linnaeus) .

Library of Congress Cataloging in Publication Data Winterbottom, Richard, 1944— The familial phylogeny of the Tetraodontiformes. (Smithsonian contributions to zoology, no. 155) 1. Tetraodontiformes—Anatomy. 2. Phylogeny. 3. Muscles. 4. Fishes—Evolution. I. Title. II. Series: Smithsonian Institution. Smithsonian contributions to zoology, no. 155. QL1.S54 no. 155 [QL639] 591'.08s [597'.58] 73-2684

For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price $2.80 (paper cover) Contents

Page Introduction 1 Acknowledgments .. 3 Materials, Methods, and Terminology 4 Myological Descriptions of Representative Triacanthodids 5 Subfamily Triacanthodinae 6 Triacanthodes anomalus (Schlegel) 6 Tydemania navigatoris Weber 11 Macrorhamphosodes platycheilus Fowler 12 Halirriochirurgus centriscoides Alcock 12 Subfamily Hollardiinae 12 Parahollardia lineata (Longley) 12 Hollardia hollardi Poey 13 Summary of Triacanthodidae 13 Myological Descriptions of Representative Triacanthids 13 Trixiphichthys weberi (Chaudhuri) 13 Triacanthus biaculeatus (Bloch) 18 Tripodichthys oxycephalus (Bleeker) 18 Summary of Triacanthidae 18 Myological Descriptions of Representative Balistids 18 Balistes vetula Linnaeus 19 Balistapus undulatus (Mungo Park) 24 Melichthys niger (Bloch) . 24 Rhinecanthus aculeatus (Linnaeus) 25 Odonus niger (Riippell) 25 Summary of Balistidae 25 Myological Descriptions of Representative Monacanthids 25 Pervagor melanocephalus (Bleeker) 26 Stephanolepsis auratus (Castlenau) 31 Paramonacanthus barnardi Fraser-Brunner 31 Chaetoderma spinosissimus (Quoy and Gaimard) 31 Paraluteres prionurus (Bleeker) . 31 Cantherhines pardalis (Riippell) 32 Oxymonacanthus longirostris (Bioch and Schneider) 32 Pseudalutarius nasicornis (Schlegel) 33 Ahiterus monoceros (Linnaeus) 33 Aluterus heudeloti Hollard 34 Aluterus schoepfi (Walbaum) 34 Anacanthus barbatus Gray 34 Summary of Monacanthidae 35 Myological Descriptions of Representative Aracanids 35 Kentrocapros aculeatus (Houttuyn) 35 Capropygia unistriata Kaup 39

ui Page Myological Descriptions of Representative Ostraciids 40 Ostracion tuberculatus Linnaeus 40 Lactoria cornuta (Linnaeus) 44 Tetrosomus concatenatus (Bloch) 44 Acanthostracion quadricornis (Linnaeus) 45 Summary of Ostraciidae 45 Myological Description of the Representative Triodontid 45 Triodon macropterus Lesson 45 Myological Descriptions of Representative Tetraodontids 50 Lagocephalus lunaris (Bloch and Schneider) 50 Amblyrhynchotes honckenii (Bloch) 56 Colomesus psittacus (Bloch and Schneider) 56 Xenopterus naritus (Richardson) 57 Arothron immaculatus (Bloch) 57 Tetraodon schoutedeni Pellegrin 58 Summary of Tetraodontidae 58 Myological Descriptions of Representative Canthigasterids 58 Canthigaster margaritatus (Riippell) 58 Canthigaster valentini (Bleeker) 63 Myological Descriptions of Representative Diodontids 63 Diodon hystrix Linnaeus 63 Chilomycterus orbicularis (Bloch) 68 Chilomycterus schoepfi (Walbaum) 68 Summary of Diodontidae 68 Myological Descriptions of Representative Molids 69 Mola mola Linnaeus 69 Ranzania laevis (Pennant) 72 Phylogenetic Significance of Observed Myological Variation 73 Interpretation 74 Discussion 90 Alternative Interpretations 94 A New Classification of the Tetraodontiformes 97 Literature Cited 100 List of Abbreviations 104 Figures 1-185 106-201

IV The Familial Phylogeny of the Tetraodontiformes (Acanthopterygii: Pisces) as Evidenced by Their Comparative Myology

Richard Winterbottom

Introduction changes appear to be of a grosser and less subtle nature (and thus more obvious) than those This paper represents an attempt to elucidate the involving other facets of the functional systems. familial phylogeny of the order Tetraodontiformes, Since this paper represents one of the first using a comparative myological approach. Past attempts at detailed phylogenetic interpretation classifications, usually based on external characters, from a purely myological standpoint, it seems neces- osteology or both, have often been vague concern- sary to outline certain factors influencing myology. ing the precise interrelationships of the families, The evolutionary mechanism of muscular change although certain family groupings have been almost may be by the subdivision of new sections from universally accepted. Since the aim is to resolve the existing muscles, the reduction and ultimate loss familial phylogeny, the eleven families recognized of a muscle, or change in fiber direction, since by Tyler (1970a) are provisionally accepted rather these alterations have functional significances. than the seven families proposed by Greenwood, There are two extreme muscle types (pinnate and et al. (1966). This approach is used in order to parallel), with a complete gradation between them. facilitate recognition of evolutionary lines that In pinnate muscles, the fibers converge on a central may become obscured in a more compressed axis (usually a tendon), while they lie parallel to classification. one another in a parallel fibered muscle. The two For a group supposedly derived from a perciform types are apparently capable of the same amount or preperciform stock, the Tetraodontiformes (of of work (Alexander, 1968:19). Perhaps more impor- which there are some 320 Recent species) are very tant than work in an evolutionary context is (are) diverse in size, shape, habits, habitats, and diet. the advantage (s) conveyed by the different fiber The morphological changes associated with these arrangements. Parallel fibered muscles provide variations can be expected to be reflected in the maximum excursion, and can be arranged as thin anatomy, particularly in the myology. Myological sheets. Pinnation allows the localization of forces, the accommodation of the muscle in an otherwise Richard Winterbottom, J. L. B. Smith Institute of Ichthyol- ogy, Rhodes University, Grahamstown, Republic of South unsuitable place (e.g., a narrow bony canal), and Africa. the possibility of moving the tendon in two planes SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY by asymmetric contraction (Gans and Bock, 1965: of Tyler (1962a, 1962b, 1962c, 1963a, 1963b, 1968, 141-142). There is no evidence as to which of the 1970a, 1973) for information on this topic. two types (if either) is phylogenetically the older in Numerous papers contain sparse observations on vertebrates. Muscles affect a multidimensional the muscles of tetraodontiforms, with varying de- system, and therefore a change in any one muscle grees of accuracy. These include the works of is likely to produce changes in other muscles serving Cuvier (1817), E. Geoffroy Saint-Hilaire (1818); I. the same system. The synergistic interactions of Geoffroy Saint-Hilaire (1827), Meckel (1828), bones, nerves, blood vessels, and ligaments with Goodsir (1841), Cleland (1862), Owen (1866), each other and with the muscles are also subject Thilo (1879, 1896, 1899, 1914, 1920), S0rensen to this interpretion, of course, although not quite (1883, 1884), Wiedersheim (1887), Borcea (1906), so obviously. The apparent ease with which muscle Maurer (1912), Lubosch (1917, 1938), Takahasi fibers can migrate and rearrange themselves (as (1917), Soucte (1932a, 1932b, 1932c, 1935), Greg- evidenced, for example, by individuals with de- ory and Raven (1934), van Dobben (1935), Edge- formed limbs or jaws), coupled with the rigidity worth (1935), Nishi (1938), Raven (1939a, 1939b), necessary for the application of muscle force at Gabriel (1940), van Roon (1942), Kesteven (1943), the most suitable point of the lever system makes Monod (1950, 1959a, 1959b, 1960), Le Danois comparative myology a particularly useful and (1958), and Winterbottom (1970). rewarding field of study. Of the more detailed papers, Winther (1877), Although the data gathered here for phylogenetic Lubosch (1929), and van Roon and ter Pelkwijk interpretation can be utilized in a number of other (1940) described the jaw muscles, Dietz (1914) the ways (e.g., functional evolution, systematic defini- jaw and branchial muscles, Grenholm (1923) the tion), this lies beyond the scope of this study. The muscles of the fins, Rosen (1913), Parr (1927), and view that interpretations of phylogeny are only Winterbottom (1971a) aspects of the body muscu- meaningful if based on the presence of shared, lature, while Willem (1941, 1942, 1944, 1945, 1947a, specialized (synapomorphic) characters is finding 1947b, 1949) and Sarkar (1960) considered the increasing acceptance among biologists. This functional workings of the tetraodontiform head. process inevitably results in assumptions about the Le Danois (1955, 1959, 1961) has made extensive probable homology and character state (i.e., plesio- observations on the myology of the "Orbiculates" morph vs. apomorph) of the characters, and the (a group containing the gymnodonts and ostra- weighting of these characters is a necessary and cioids). Few of these papers contain any reasoned, integral part of the process. While the utilization of phylogenetic remarks. Winther (1877:462-463) ac- synapomorphic characters in the elucidation of cepted Cuvier's (1817) Sclerodermi (balistoids and phylogeny is unlikely to produce much disagree- ostracioids), and suggested that the Triacanthidae ment, the incorporation and reflection of that were specialized relatives. He felt, however, that phylogeny in a hierarchical classification is still the gymnodonts were most closely related to the avoided by certain biologists. I agree with Mayr "Discobali" (=Cyclopteridae), a conclusion which (1969:78) when he states, "The evolutionist believes has not, as far as I am aware, been supported since that a classification consistent with our reconstruc- then. Lubosch (1929) made certain general con- tion of phylogeny has a better chance of meeting clusions, but these suffer from the wide scope of these objectives, [i.e., maximum information con- the work in relation to the number of species tent with maximal ease of retrieval] than any other examined. The placement of the Acanthuridae method of classification." I have therefore attempt- between the Balistidae and Triacanthidae in the ed to reorganize the classification of the tetraodonti- text may indicate that Lubosch felt them to be forms to reflect the phylogenetic picture derived tetraodontiforms, but this is not explicitly stated. from the myology, using the basic methods set out Grenholm (1923:245), while stating that the tetrao- by Hennig (1966) and Brundin (1966, 1968). dontiforms were a well-defined group (and prob- Although muscles are described in terms of the ably closely related to the Acanthuridae), felt that bones they attach to (and hence perhaps an osteo- the separation of the order into scleroderms and logical summary should be included), the reader gymnodonts was difficult to maintain (". . . schwer is referred to the detailed osteological investigations afrecht zu erhalten"), although he gave no reasons NUMBER 155 for this opinion. Willem, in his (1942:19-21) paper, ACKNOWLEDGMENTS.—This paper represents the concluded that among the scleroderms, the acan- major portion of a doctoral dissertation submitted thurids were the most generalized, followed either to the School of Graduate Studies, Queen's Uni- by the triacanthids or the monacanthids, with the versity, Kingston, Ontario, Canada. balistids the most specialized. Later (1944), Willem I wish to express my deepest gratitude to Dr. placed the chaetodontid fishes in the tetraodonti- James C. Tyler of the Academy of Natural Sciences, forms (=his "Plectognathes"). Philadelphia (ANSP), whose stimulating interest and In her definition of the "Orbiculates"—the helpful advice has formed the mainspring of this ostracioid and gymnodont fishes—Le Danois study. He has freely made available his detailed (1955) regarded the following myological characters knowledge of the tetraodontiforms, and supplied as primitive, and used them to define the suborder: much unpublished osteological information as well musculature complex, subcutaneous and trapezoi- as most of the specimens dissected. Dr. Tyler has dal muscles present. As discussed above, it has generously commented on both the dissertation and recently been realized that primitive (symplesio- much of the manuscript, although he is not neces- morph) characters cannot be used to link groups sarily in agreement with the phylogenetic relation- phylogenetically. Additionally, complexity of mus- ships as presented here, and certainly disagrees with cles is a sign of specialization, not of antiquity; the classificatory approach taken by the author. subcutaneous muscles are found elsewhere (e.g., To Dr. S. H. Weitzman of the Smithsonian Insti- Antenariidae), and certainly could not exist where tution, Washington, D.C., I express my thanks for discreet, overlapping scales (surely a "primitive" his encouragement to complete the manuscript and condition for the teleosts) are present. "Trapezoi- to maintain it in its entirety, and for his comments dal" muscles present somewhat of an enigma. The on the text. anteroventral muscle (the "trapeze inferieur" of I thank my Ph.D. supervisor, Dr. J. Allen Keast, Le Danois, 1959) is the protractor pectoralis (de- for his guidance and advice throughout the study rived from the levator series of the branchial arches, period, and for making travel funds readily avail- and innervated by the vagus nerve), while her able to me. Various members of staff and graduate "trapeze superieur" is derived from the anterior students at Queen's University Biology Department trunk myomeres, and is innervated by an occipito- were always ready to discuss my problems with me spinal nerve. The protractor pectoralis occurs (particularly Drs. A. E. Downe and C. H. Hood, regularly throughout the Euteleostei (Division III) and Messrs. Amundrud, McLaughlin, and van der of Greenwood, et al. (1966, 1967) except in most, if Kloet), and their help is greatly appreciated. not all, members of the Protacanthopterygii (sensu In addition, dissection material and discussions Rosen and Patterson, 1969). The levator pectoralis are gratefully acknowledged from the following: has only been reported as a separate muscle from Drs. T. Abe of the University of Tokyo and H. certain families of Rosen and Patterson's (1969) Britski of the Universidade de Sao Paulo (for Series Percomorpha. Thus, the presence of two donating the specimens of the rare Triodon), J. G. "trapezoidal" muscles in the "Orbiculates" is more Field of the University of Cape Town, P. H. Green- in keeping with an alignment of these fishes with wood and G. Palmer of the British Museum (Nat- groups of at least the percomorph level of organi- ural History) (BM (NH)), D. E. McAllister and C. G. zation, rather than with the "malacopterygian" Gruchy of the National Museum of Canada (NMC), level (to which Le Danois felt they were related). J. E. Randall of the Bernice P. Bishop Museum, Parenthetically, it should be noted that the last M. M. Smith of the J. L. B. Smith Institute, J. H. two of these three myological characters are not Wallace of the Oceanographic Research Institute, found in ostracioids. The comparative reader will Durban, R. McN. Alexander, W. C. Freihofer and find certain discrepancies between Le Danois' H. Jessen. For the many hours they spent discussing myological descriptions and those given here. The phylogeny, the manuscript, and fishes with me, I phylogenetic conclusions reached by Tyler in his would like to thank Drs. G. J. Nelson and D. E. various publications on tetraodontiform osteology Rosen of the American Museum of Natural History. will be reviewed in the discussion section of this To my wife Irina I extend my gratitude for her paper. patience in dealing with my frequent unintelligible SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY soliloquys. dae, Monacanthidae, Aracanidae, Ostraciidae, Support during this study came from R. Samuel Triodontidae, Tetraodontidae, Canthigasteridae, McLaughlin Scholarship and Fellowships (Queen's Diodontidae, and Molidae. The following group University), a National Research Council Post- terms are used provisionally to contain more than graduate Scholarship, and a Jessup Fund Award one family: balistoid (Balistidae + Monacanthidae), from the Academy of Natural Sciences, Philadel- ostracioid (Aracanidae + Ostraciidae), tetraodon- phia, for the summer of 1969. The dissertation was toid (Tetraodontidae + Canthigasteridae), and rewritten in part for publication during the tenure triacanthoid (Triacanthodidae + Triacanthidae). of a Postdoctoral Research Fellowship at the The latter group may not be phylogenetically valid, Smithsonian Institution, Washington, D.C. and will be used for convenience only. The Cuvierian term "gymnodont" will be used initially to designate the Triodontidae + Molidae + Dio- Materials, Methods, and Terminology dontidae + tetraodontoids. The following is a list of the material examined, The nomenclature of the muscles used in this arranged alphabetically by family. The number in paper is taken from a descriptive synonymy of parentheses indicates the number of specimens teleostean muscles at present in preparation by the examined. Catalog abbreviations are clarified in author. In the case of the muscles to the rudimen- "Acknowledgments." tary balistoid pelvic fin, the nomenclature is based ARACANIDAE.—Capropygia unistriata Kaup: on the assumption that the fin ray elements rep- ANSP 109574, (1), 67 mm SL, Kingston, South resent the pelvic spines. In the descriptive section, Australia. Kentrocapros aculeatus (Houttuyn): the muscles have been grouped according to con- ANSP 117531, (2), 91 and 92 mm SL, no data, venient body regions. received from T. Abe. Osteological nomenclature is based on Tyler BALISTIDAE.—Balistapus undulatus ( M u n g o (1962a, 1970a) except for the hyoid arch, where Park): ANSP 108832, (1), 159 mm SL, Ressource Nelson (1969) has been followed. The suture lines Islands, Indian Ocean; ANSP 117598, (1), 83 mm shown in the figures are approximate, and are SL, Inhaca Island, Indian Ocean. Balistes vetula included for reference purposes only. For accurate Linnaeus: ANSP 105508, (1), 111 mm SL, Estado outlines of the bones, see Tyler's various publica- Sucre, Venezuela; ANSP 117535, (1), 137 mm SL, tions on tetraodontiform osteology. no data, aquarium specimen. Melichthys niger Specimens for the most part were dissected under (Bloch): ANSP 117515, (1), 153 mm SL, no data, a Wild M-5 stereoscopic research dissecting micro- aquarium specimen. Odonus niger (Riippell): scope. Rough pencil drawings were made on bristol ANSP 102779, (1), 257 mm SL, Amirante Islands, board with the aid of a Wild M-5 camera lucida Indian Ocean. Rhinecanthus aculeatus (Linnaeus): attachment. Detail was filled in freehand from the ANSP 117546, (1), 159 mm SL, Johnston Island, dissection, the drawing then being inked in. Bone Hawaii; ANSP 117544, (1), 175 mm SL, no data, and cartilage—stippled; muscles—lined with long received from J. H. Wallace. and short dashes; connective tissue—filled with CANTHIGASTERIDAE. — Canthigaster margaritatus wavy lines; tendons, aponeuroses and ligaments— (Ruppell): ANSP 117517, (1), 52 mm SL, no data, short dashes; coelom—crosshatched. Bones are only received from M. M. Smith. Canthigaster valentini labeled in the figures of species dissected in detail (Bleeker): ANSP 117534, (1), 45 mm SL, Seychelles for each family; the muscle abbreviations being Islands, Indian Ocean. underlined for clarity in these figures. Caudal fin DIODONTIDAE.—Chilomycterus orbicularis (Bloch): rays are numbered from the midline, those above ANSP 117510, (1), 80 mm SL, no data, received being prefixed by "D", those below by "V". Nu- from M. M. Smith. Chilomycterus schoepfi, (Wal- merical differences arising from the number of fin baum): ANSP 117518, (1), 138 mm SL, beach at rays in any fin are not considered. Kitty Hawk, North Carolina. Diodon hystrix Lin- The musculature of representatives of the follow- naeus: NMC 68-997, (1), 77 mm SL, Tres Maria, ing families, and in the following order, are Mexico; ANSP 117553, (1), 291 mm SL, Kaneohe described: Triacanthodidae, Triacanthidae, Balisti- Bay, Hawaii. NUMBER 155

MOLIDAE.—Mola mola Linnaeus: BM (NH) mm SL, no data, received from J. H. Wallace. 1938.6.23.38, (1), 287 mm SL, off Chousi, Japan; Ostracion tuberculatus Linnaeus: ANSP 117542, NMC 67.298, (1), 500 mm SL, off Newfoundland. (2), 102 and 157 mm SL, no data, received from Ranzania laevis (Pennant): ANSP 106723, (1), 93 }. H. Wallace. Tetrosomus concatenate (Bloch): mm SL, Hawaii. ANSP 117555, (2), 196 and 200 mm SL, no data, MONACANTHIDAE.—Aluterus heudelotii Hollard: received from M. M. Smith. ANSP 103224, (1), 110 mm SL, Guinea, West TETRAODONTIDAE. — Amblyrhynchotes honckenii Africa. A. monoceros (Osbeck): ANSP 117498, (1), (Bloch): ANSP 117547, (1), 106 mm SL, Cape Prov- 160 mm SL, R/V Oregon (Station 10523); ANSP ince, South Africa. Arothron immaculatus (Bloch): 117554, (2), 345 and 350 mm SL, no data, received ANSP 117528, (1), 68 mm SL, no data, received from J. Field. A. schoepfi (Walbaum): ANSP 101843, from M. M. Smith. Colomesus psittacus (Bloch and (1), 106 mm SL, Caribbean, off Columbia. Anacan- Schneider): ANSP 98887, (1), 117 mm SL, off thus barbatus Gray: ANSP 117516, (1), 166 mm SL, Surinam. Lagocephalus lunaris (Bloch and Schnei- Gulf of Thailand. Cantherhines pardalis (Riip- der): ANSP 113639, (3), 118, 124, and 137 mm SL, pell): ANSP 117529, (1), 61 mm SL, Inhaca Island, Malaysia. Tetraodon schoutedeni Pellegrin: ANSP Indian Ocean; ANSP 117530, (1), 126 mm SL, no 117501, (1), 62 mm SL, Congo River at Leopold- data, received from M. M. Smith. Chaetoderma ville. Xenopterus naritus (Richardson): ANSP spinosissimus (Quoy and Gaimard): ANSP 117514, 117495, (2), 102 and 120 mm SL, Bay of Bengal. (1), 42 mm SL, Gulf of Thailand. Oxymonacanthus TRIACANTHIDAE. — Triacanthus biaculeatus longirostris (Bloch and Schneider): ANSP 117496, (Bloch): ANSP 111536, (1), 111 mm SL, off Kerala (1), 72 mm SL, Zanzibar; ANSP 104791, (1), 72 State, India. Tripodichthys oxycephalus (Bleeker): mm SL, Amirante Islands, Indian Ocean. Paraluteres ANSP 89397, (1), 82 mm SL, Paknam, Siam. Trixi- prionurus (Bleeker): ANSP 106644, (1), 57 mm SL, phichthys weberi (Chaudhuri): ANSP 117492, (2), Farquhar Islands, Indian Ocean. Paramonacanthus 115 and 122 mm SL, Bay of Bengal. barnardi Fraser-Brunner: ANSP 117506, (1), 78 mm TRIACANTHODIDAE.—Halimoch irurgus cent riscoi- SL, Natal, South Africa. Pervagor melanocephalus des Alcock: ANSP 117494, (1), 114 mm SL, Bay (Bleeker): ANSP 117500, (2), 66 and 85 mm SL, of Bengal. Hollardia hollardi Poey: ANSP 117491, Taiwan; ANSP 117512, (1), 87 mm SL, Mocam- (2), 53 and 92 mm SL, off Honduras. Macrorham- bique Island, Indian Ocean. Pseudalutarius nasi- phosodes platycheilus Fowler: ANSP 117533, (2), 79 comis (Schlegel): ANSP 100830, (1), 45 mm SL, and 88 mm SL, Bay of Bengal. Parahollardia line- Bay of Bengal. Stephanolepis auratus (Castlenau): ata (Longley): ANSP 101031, (1), 68 mm SL, Gulf ANSP 117526, (1), 64 mm SL, Delagoa Bay, South of Mexico. Triacanthodes anomalus (Schlegel): Africa; ANSP 117521, (2), 69 and 82 mm SL, Natal, ANSP 101256, (1), 74 mm SL, Honshu, Japan; South Africa. ANSP 117493, (1), 87 mm SL, Honshu, Japan. OSTRACIIDAE.—Acanthostracion quadricornis (Lin- Tydemania navigatoris Weber: ANSP 117532, (1), naeus): ANSP 98848, (1), 126 mm SL, Atlantic 73 mm SL, Bay of Bengal. Ocean, off South America. Lactoria cornuta (Lin- TRIODONTIDAE. — Triodon macropterus Lesson: naeus): ANSP 117540, (1), 50 mm SL, Natal, South ANSP 103902, (3), 270, 350, and 350 mm SL, Africa; ANSP 117538, (1), 92 mm SL, no data, Ryukyu Islands, Pacific Ocean; ANSP 98916, (1), received from M. M. Smith; ANSP 117539, (1), 176 382 mm SL, Bonin Island, Pacific Ocean.

Myological Descriptions of Representative Triacanthodids

The general body outline of these fishes is rep- Caribbean (with one exception presently recorded resented in Figure 1. The sample size of the two only from HaAvaii), while the triacanthodins extend subfamilies is four out of nine genera for the broadly from the east coast of Africa to the western Triacanthodinae and two out of two genera for Pacific. Triacanthodids are usually benthic or bathy- the Hollardiinae. pelagic fishes, living well below 100 fathoms. The latter subfamily is found mainly in the They appear to feed predominantly on marine SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY invertebrates, with the exception of Macrorham- cavities are expanded (e.g., when feeding), since phosodes, in whose gut nothing but scales of other it is situated dorsal to the articulation of the fishes has yet been found. palatal arch with the skull. This medial movement The rounded caudal fin and rather deep body of locks the cranial condyle of the maxilla in position, these fishes implies that they are not fast swimmers, which in turn prevents the posterior movement of while the protractile jaws (of at least the more the premaxilla (since the articular process of the conservative genera) evidences that their prey is premaxilla jams against the premaxillary condyle active. The jaws, however, are not protrusile in of the maxilla). The mouth can thus be closed with the percoid manner, since the process here is com- the premaxilla in the protruded position (were pletely passive (i.e., no active part is played by the processes not present, the premaxillae would be the maxilla in protruding the premaxilla). They retracted during mouth closing—see Alexander, are apparently incapable of closing the mouth with 1967, for details). In the triacanthoids, medial the premaxillae still in the protruded position (the movement of the palatal process would put some advantages of such a maneuver are discussed by pressure on the maxilla, but since none of the Alexander, 1967). above processes are present, this would be insuffi- Figure 48 shows the overall structure of the jaws cient to lock them in position during closure of the of Triacanthodes anomalus and Triacanthus biacu- mouth. leatus (a triacanthid). The maxilla (Figures 4Sc,d) Of the two subfamilies, the Hollardiinae are is a fairly broad plate, the dorsal region curving apparently the most generalized (Tyler, 1968:31- inward toward the midline. The fact that it is a 33). However, Tyler (1968:30, fig. 6) also indicates broad bone and not rodlike indicates that it is in effect, that they, the triacanthids and the balis- incapable of rotating about its vertical axis, a toids, form the sister group of the triacanthodins. movement essential in the protrusion of the percoid jaw. The dorsomedial region of the maxilla has Subfamily TRIACANTHODINAE developed an external and internal flange. These two flanges together form a groove containing the Triacanthodes anomalus (Schlegel) lateral margin of the ascending process of the premaxilla. From the lateral face of the inner flange FIGURES 49-56 of the maxilla there arises a continuous sheet of tough ligamentous tissue (the maxillopremaxilla MUSCLES OF THE CHEEK ligament) which passes dorsomedially to attach to the ventromedial face of the ascending process of ADDUCTOR MANDIBULAE (Figure 49: A 1, A 2«, the premaxilla. Neither the posteroventral face of A 2P; see also Figure 68a of Hollardia).—The adduc- the maxilla, nor the anterior surface of the ethmo- tor mandibulae is well separated into two portions vomer region where the maxilla articulates show by the path of ramus mandibularis V. The postero- any evidence of the cartilagenous pad found there dorsal section (A 2a) originates from the antero- in percoids. This indicates that the maxilla can dorsal face of the hyomandibular and from a only move in one direction, since in most cases of sheetlike aponeurosis covering the dorsal surface joints with two degrees of freedom (a higher pair) of the section. Beneath the eye, A 1 begins to there is a cartilagenous meniscus or pad allowing develop, increasing in size as it passes anteriorly. this type of movement (Alexander, 1967). It seems, The fibers from A 1 grade into a tendon which then, that the maxilla cannot function as it does receives contributions from A 2P and A 3 ventrally. in percoids. From the anterodorsal region of the This composite tendon then divides, the dorsal maxilla the maxilloethmoid ligament passes poste- part passing anterodorsally to insert near the medial riorly to the ethmoid bone. Immediately beneath face of the cranial condyle of the maxilla, beneath its attachment to the maxilla is the anterior process its articulation with the palatine, while the ventral of the palatine, which rests in a concavity in the part inserts on the anteromedial face of the dentary maxilla (Figure 486). In perciform fishes with pro- in the Meckelian fossa. The lateral portion of the trusile jaws, the anterior process of the palatine ventral fibers contributing to the tendon mentioned moves medially when the buccal and opercular above forms A 23, which originates from the dorsal NUMBER 155 preopercle and quadrate and the anterior face of lateral and ventral surfaces of the anterohyal. The the hyomandibular. The ventromedial section, A 3, muscle attaches ventrally to the ventrohyals by a originates from the lateral faces of the quadrate, pair of ligaments passing anterolaterally from the syraplectic, and metapterygoid. anteromedial border of the fibers. These are similar, LEVATOR ARCUS PALATINI (Figure 49: L.A.P).— although somewhat weaker, than those figured for The muscle originates from the ventral face of the Hollardia (Figure 65). postorbital process of the sphenotic, and inserts on HYOHYOIDEI ABDUCTORES (Figure 49: H.AB.).— the dorsal and anterodorsal faces of the hyomandi- There are four sections of this muscle, one to each bular. It is a pyramidal muscle, whose apex lies of the ventral four branchiostegal rays. The section dorsally. to the first ray originates as a long tendon from the DILATATOR OPERCULI (Figure 49: D.O.).—Origin is posteroventral face of the ventrohyal, that to the from the ventrolateral sphenotic and pterotic and next ray from the base of the first ray. The two the dorsal tip of the hyomandibular. The fibers following sections arise from the base of the second pass ventrolaterally to insert on the tip of the dorsal ray. All insert on the medial faces of the rays distal process of the opercle. It is a conical muscle, with to their bases. the apex pointing ventrally. HYOHYOIDEI ADDUCTORES (Figure 49: H.AD.).— LEVATOR OPERCULI (Figure 49: L.O.).—The mus- The muscle passes between the dorsalmost four cle consists of two portions. The anterior part branchiostegal rays and continues above them as a originates from the ventrolateral face of the pterotic thin layer of fibers in the membrane lining the and inserts on the dorsomedial face of the posterior opercular cavity, fading out beneath the opercle flange of the opercle. The posterior section arises and subopercle. There is a fairly well-developed from the lateral pterotic and anteromedial supra- section along the posterior border of the opercle in cleithrum and passes ventrally to insert on the the tissues of the opercular valve. Fibers also con- posteromedial flange of the opercle just behind the tinue inward ventrally along the floor of the oper- insertion of the anterior part. cular cavity for a short distance. ADDUCTOR ARCUS PALATINI (Figure 49: A.A.P.).— STERNOHYOIDEUS (Figures 49, 50: STH.).—A well- This parallel fibered muscle originates from the developed muscle, consisting of three bundles of ventrolateral surface of the parasphenoid and the fibers separated by myocommata. The posterolat- lateral prootic. It inserts on the dorsomedial faces eral region is continuous with the anterolateral of the metapterygoid, mesopterygoid, and hyoman- fibers of the obliquus inferioris. Origin is from the dibular. anteroventral deithrum and insertion mainly on ADDUCTOR OPERCULI (Figure 50: AD.OP.).—Origin the posterior and lateral faces of the urohyal. The is from the ventral face of the pterotic, the fibers dorsomedial fibers tend to pass dorsally rather than inserting on the dorsomedial face of the opercle. anterodorsally. The more anterior of these fibers attach to the fascia separating the halves of the MUSCLES OF THE HYOID REGION sternohyoideus, while the more posterior ones are to a greater or lesser extent continuous with the INTERMANDIBULARIS (see Figure 65 of Hollardia, fibers of the sternobranchialis. win.).—This small muscle connects the halves of the STERNOBRANCHIALIS (Figure 50: STB.).—The lower jaw, attaching to the dentaries just below the muscle originates from the anteroventromedial face Meckelian fossae. of the cleithrum. It passes dorsally, the anterior fi- PROTRACTOR HYOIDEI (Figure 49: PR.HY.).—Origin bers intermingling with those of the sternohyoideus, is from the inner face of the dentary, both dorsal and grades into a broad, flat aponeurosis which and ventral to the intermandibularis. The fibers fuses with that from the other side. The two sheets pass posteriorly and split into left and right halves. then separate and divide into anterior and posterior Each section inserts on the lateral face of the an- components. The latter passes dorsally, lateral to terohyal as far back as the third branchiostegal ray. the rectus communis, to insert on the ventral faces HYOHYOIDEUS INFERIORIS (Figures 49, 50: HY. of the third hypobranchial and fourth cerato- IN.).—The muscle is relatively well developed, and branchial. From the posteromedial face of the passes from a median ventral raphe to the antero- anterior portion a tendon arises, fusing briefly 8 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY with its antimere before diverging to insert on the These four cylindrical muscles insert aponeuroti- ventral surface of the third hypobranchial. It cally on the posterodorsal faces of the four epi- passes medial to the rectus communis. The rest of branchials. They originate from the lateral prootic, the anterior sheet continues upward, lateral to the and a small part of the pterotic, increasing in size rectus communis. to insert on the ventral surfaces while attaining a more lateral origin as one pro- of the first three hypobranchials. ceeds posteriorly in the series. LEVATORES INTERNI II-III (Figure 52: L.INT.).— Both muscles originate from the prootic. Levator VENTRAL BRANCHIAL MUSCLES II inserts on the dorsomedial face of the infrapharyn- PHARYNGOCLAVICULARIS EXTERNUS (Figures 50, 51: gobranchial 2, while levator III (which originates PHC.E.).—Origin is from the anterodorsal face of the more dorsolaterally on the prootic) inserts on the medial region of the cleithrum, the fibers passing anterodorsolateral face of the infrapharyngobran- dorsomedially to insert on the anteroventral face chial 3. of ceratobranchial 5. OBLIQUUS DORSALIS HI (Figure 52: OBL.D.).—The PHARYNGOCLAVICULARIS INTERNUS (Figures 50, 51: muscle passes from the anterodorsal face of epi- PHC.L).—The fibers arise from the anteromedial branchial 2 to the posterodorsal face of infra- face of the cleithrum and pass forward to insert on pharyngobranchial S. the ventral border of ceratobranchial 5. The inser- OBLIQUUS POSTERIOR (Figure 52: OBL.P.).—The tion is medial to that of the pharyngoclavicularis muscle connects the anterodorsal face of cerato- externus. branchial 5 to the posterodorsal face of cerato- OBLIQUI VENTRALES I-III (Figure 51: OBL.V.).— branchial 4. These muscles pass between the ceratobranchials TRANSVERSI DORSALES II-III (Figures 50, 52: TR. and hypobranchials of the first three arches. The D.^.—The anterior muscle originates partly from the first is small, the second being somewhat better ventrolateral surface of the parasphenoid and part- developed. The third muscle attaches to the ventral ly from a raphe with its antimere behind this. It process of the third hypobranchial. The more inserts on the dorsal face of epibranchial 2. The posterior fibers of this section attach to the arch- posterior muscle arises from a raphe in the dorsal shaped ligament between the left and right hypo- midline and courses laterally to insert on the pos- branchial processes. terodorsal face of epibranchial 3. TRANSVERSI VENTRALES rv-v (Figure 51: TR.V.).— RETRACTOR DORSALIS (Figure 52: D.RETR.).—This The muscles pass transversely across the midline, is a well-developed muscle originating from the connecting the ventromcdial faces of the fourth and ventrolateral surfaces of the basioccipital and the fifth ceratobranchials, respectively. first vertebra. It passes anteroventrally to insert on RECTI VENTRALES I AND iv (Figure 51: RECT.V.).— the dorsal face of infrapharyngobranchial 3. The first of these muscles connects the anterolateral ADDUCTORS IV-V (Figures 50, 52: AD.).—Adduc- face of hypobranchial 1 to the posterolateral face tor IV is a small muscle connecting the medial of the dorsohyal. Rectus IV arises from the ventral faces of the fourth epi- and ceratobranchials. Ad- face of ceratobranchial 4 and passes anteromedially ductor V is also small, and passes from the antero- to attach to the posterior surface of the arch-shaped dorsal tip of ceratobranchial 5 to the distal tip of ligament between the third hypobranchials. epibranchial 4. It lies lateral to the obliquus RECTUS COMMUNIS (Figures 50, 51: R.COMM.).— posterior. Origin is from the posterodorsolateral face of the SPHINCTER OESOPHAGI (Figures 51, 52: s.o.).— urohyal. The fibers pass posteriorly, the muscle be- The fibers of the muscle surround the esophagus, coming aponeurotic about the level of the third attaching mainly to the posteromedial faces of the hypobranchial. It inserts on the ventrolateral face fifth ceratobranchials. In the ventral midline, it lies of ceratobranchial 5. dorsal to transversus ventralis V, and develops a raphe at its extreme anterior margin. It is continu- DORSAL BRANCHIAL MUSCLES ous across the dorsal midline, and is attached LEVATORES EXTERNI i-rv (Figure 50: L.EXT.).— anteriorly to epibranchial 4 by connective tissue. NUMBER 155

MUSCLES OF THE PECTORAL GIRDLE the hyomandibular fossa, the fibers passing ven- trally to insert on the anterodorsal face of the ABDUCTOR SUPERFICIALIS (Figure 49: ABD.S.).— cleithrum. The fibers originate from the posteroventral face LEVATOR PECTORALIS (Figure 50: TR.).—The of the cleithrum, and pass posterodorsally to insert muscle arises from the pterotic just behind the by a series of tendons on the dorsolateral bases of protractor pectoralis. The fibers fan out to insert the 12 principal fin rays. on the medial face of the supracleithrum. ABDUCTOR PROFUNDUS (Figures 49, 50, 53: ABD. P.).—Origin is from the posterior face of the cleithrum and the anterior face of the coracoid. MUSCLES OF THE PELVIC GIRDLE The fibers pass dorsally to insert tendinously on the lateral faces of the posteroventral flanges of the 12 ARRECTOR DORSALIS PELVICUS (Figures 50, 53: principal fin rays and the lateral half of the base A.D.P.).—Origin is from the anterolateral face of of the vestigial fin ray. the anterior part of the pelvis, and is bordered ARRECTOR VENTRALIS (Figure 49: ARR.V.).—The ventrally by the lateral flange of the bone. It inserts muscle originates just dorsal to the abductor super- on the anterolateral aspect of the pelvic spine, ficialis on the cleithrum. It courses posterodor- dorsal to the articulation with the pelvis. sally to insert tendinously on the base of the medial ARRECTOR VENTRALIS PELVICUS (Figure 53: half of the vestigial fin ray. A.V.P.).—The muscle arises from the anteroventral ADDUCTOR SUPERFICIALIS (Figure 53: ADD.S.).— surface of the anterior part of the pelvis. It inserts Origin is from the posterior face of the cleithrum on the anterolateral aspect of the pelvic spine, and the ventromedial scapula. The fibers insert ventral to the articulation with the pelvis. by a series of 12 tendons on the dorsomedial faces ADDUCTOR SUPERFICIALIS PELVICUS (Figures 50, 53: of the principal fin rays. The more medial and A.S.P.).—Originating on the posteromedial face of dorsal portion of the muscle serves the more ventral the anterior part of the pelvis, the fibers pass postero- fin rays, attaching at right angles to the longitu- laterally to insert on the dorsal surface of the dinal axes of the rays. The more ventral fibers attach dorsal flange of the pelvic spine. to the more dorsal rays with an increasingly parallel ADDUCTOR PROFUNDUS PELVICUS (Figure 53: orientation. The muscle has thus effectively "folded A.P.P.).—There are two slips of this muscle, which under" itself. originate from the dorsomedial face of the pelvis ADDUCTOR PROFUNDUS (Figure 53: ADD.P.).—Ori- posterior to the adductor superficialis pelvicus. gin is from the medial cleithrum and coracoid, the Each slip inserts on the dorsomedial surface of a muscle inserting tendinously on the ventromedial pelvic ray, the larger and more anterior section flanges of the bases of the 12 principal fin rays. inserting on the more anterior of the two rays. ARRECTOR DORSALIS (Figure 53: ARR.D.).—The muscle originates from the medial face of the scapu- MUSCLES OF THE DORSAL FIN la and cleithrum, and passes posterodorsally to insert on the medial half of the vestigial fin ray. INCLINATORES DORS ALES (Figures 49, 55a: INC.).— CORACORADIALIS (Figure 53: COR.R.).—A small The second to fourth dorsal spines and their muscle, arising from the dorsal face of the posterior pterygiophores receive a sheet of muscle fibers process of the coracoid, and inserting on the ventro- arising from the fascia overlying the epaxialis. The medial face of the last (vend-almost) radial. inclinator to the sixth spine and those to the soft ADDUCTOR RADIALIS (see Figure 62 of Parahol- dorsal fin rays are separate, inserting only on the lardia, ADD.R. ;.—The fibers arise from the ventro- bases of the spine or ray. The muscles to the rays medial faces of the middle two radials and insert are well developed, particularly anteriorly. on the bases of the ventralmost two fin rays. It ERECTORES DORS ALES (Figures 49, 54, 55a: courses at right angles to the coracoradialis, and lies EREC. )—All spines and fin rays receive an erector. lateral to the adductor profundus. The fibers arise from the lateral faces of the PROTRACTOR PECTORALIS (Figure 50: P.P.).—Ori- pterygiophores and associated neural spines, and gin is from the ventrolateral pterotic just behind insert on the anterolateral bases of the spines or 10 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY rays. The first erector is extremely powerful and is the dorsal face of the posterior tip of the pelvis. It bipinnate. The size of these muscles in both the consists of two bundles of fibers separated by a spiny and the soft dorsal fins decreases posteriorly. myocomma, just posterior to which is a small knot DEPRESSORES DORSALES (Figures 49, 54, 55a: of fibers passing dorsally to join the ventral surface DEPR.).—All spines and rays receive a depressor, of the obliquus inferioris. which originates from the lateral pterygiophores INFRACARINALIS POSTERIOR (Figure 56: INF.p.).— and neural spines in between the erectors. The Fibers arise from the posteroventral face of the last muscles are smaller than the erectors, and insert on anal pterygiophore and attach to the ventrolateral the posterolateral bases of the spines and rays. They tip of the haemal spine of caudal vertebra 10. The also decrease in size posteriorly in both fins. medial fibers of the muscle attach to the lateral tips of the haemal spines of the two intervening SUPRACARINALIS ANTERIOR (Figures 49, 54: s. vertebrae (8 and 9). ANT.).—Origin is from the lateral and dorsal faces of the supraoccipital, the fibers inserting on the anterodorsolateral face of the first pterygiophore of MUSCLES OF THE CAUDAL FIN the dorsal fin. INTERRADIALIS (Figure 56a,6: INT.).—The mus- SUPRACARINALIS POSTERIOR (Figure 56: S.POST.).— cle interconnects the caudal fin rays, with a section Fibers arise from the lateral surface of the last inserting on the ventrolateral face of each of the pterygiophore of the dorsal fin, and course poste- six dorsal rays and on the dorsolateral face of each riorly. The muscle attaches to the dorsolateral tips of the six ventral rays. The two strips of muscle of the neural spines of the caudal vertebrae, termi- thus formed mingle between the rays on either side nating on the neural spine of caudal vertebra 10. of the midline. HYPOCHORDAL LONCITUDINALIS (Figures 56a,6,: H.L.).—Origin is from the lateral face of hypurals MUSCLES OF THE ANAL FIN 1 and 2, the fibers coursing posterodorsally to in- sert on the lateral bases of the second to fourth INCLINATORES ANALES (Figure 556: INC.).—Each (D 2-4) dorsal fin rays. anal fin ray receives an inclinator on the lateral FLEXOR DORSALIS (Figure 566: F.D.).—The muscle face of the fin base, which originates in the fascia originates from the lateral surfaces of the neural overlying the hypaxialis. The anterior muscles are spines and dorsolateral surfaces of the centra of better developed than the posterior ones. There is the ninth to eleventh caudal vertebrae, and the an additional, residual inclinator muscle originat- dorsolateral face of the urostyle. The fibers pass ing in front of the first normal inclinator and posterodorsally to insert on the lateral bases of inserting on the ventrolateral face of the first anal rays D 2-5. fin pterygiophore. FLEXOR DORSALIS SUPERIOR (Figure 56: F.D.S.).— ERECTORES ANALES (Figure 556: EREC).—Each Origin is from the neural spines of the tenth and fin ray possesses an erector, originating from the eleventh caudal vertebrae, the fibers passing poste- lateral faces of the pterygiophores and haemal riorly to insert on the lateral base of D 6. spines and inserting on the anterodorsal base of the FLEXOR VENTRALIS (Figure 566: F.V.).—Fibers ray. arise from the posterolateral faces of the haemal DEPRESSORES ANALES (Figure 556: DEPR.).—These spines and ventrolateral surfaces of the centra of muscles originate just posterior to the erectors and the ninth to eleventh caudal vertebrae, and the insert on the posterolateral bases of the fin rays. ventrolateral face of the urostyle. Fibers pass pos- INFRACARINALIS ANTERIOR (Figures 49, 50, 53: teroventrally to insert on the bases of rays V 1-6. INF.A.).—Fibers arise from the posteroventral ex- FLEXOR VENTRALIS EXTERNUS (Figure 566: F.V.E.).— tremity of the cleithrum and fan out to insert on The muscle originates from the posterolateral the anterolateral face of the pelvis. faces of the haemal spines of caudal vertebrae 9-10, INFRACARINALIS MEDIUS (Figure 49: INF.M.).— and passes posteriorly to insert on the lateral bases The muscle originates from the anteroventral mar- of rays V 1-2. It lies directly beneath the hypaxial gin of the first anal pterygiophore and inserts on body musculature. NUMBER 155 11

FLEXOR VENTRALIS INFERIOR (Figure 56: F.V.I.).— The anteroventral portion of this section arises Origin is from the posterolateral tips of the haemal from the ventrolateral face of the cleithrum and the spines of caudal vertebrae 10-11. Insertion is on posterolateral pelvis. The anterior fibers are con- the base of ray V 6. tinuous with the sternohyoideus, most of the fibers TRANSVERSUS CAUDALIS (Figure 56c: TR.C).—This behind this attaching to the anteroventral margin fan-shaped muscle originates from the dorsolateral of the postcleithrum. The anterodorsal fibers arise face of hypural 5 and the posterolateral surface from the fascia over the midline and epipleurals of the urostyle. It courses posteroventrally beneath and course anteroventrally to the posterodorsal sur- the hypochordal longitudinalis to insert on the face of the postcleithrum. The fibers become in- dorsal base of ray V 1. creasingly horizontal posteriorly, and the section separates somewhat from the obliquus superioris before finally fusing with it in the peduncular LATERAL BODY MUSCLES region. The fibers from this part attach to the bases EPAXIALIS (Figures 49, 56a: EPAX.).—The epaxial of caudal fin rays V 1-6. The muscle is made up of musculature consists of nineteen myomeres, all ver- sixteen myomeres, the more anterior of which are tebrae except the urostyle thus being represented. parallel with the muscle fiber direction. There are two major directions of the myocom- mata, and a third direction is present. anterodor- Tydemania navigatoris Weber sally. The first two merge in the posterior region of the body, where they insert by a series of broad FIGURE 57 aponeuroses on the lateral bases of rays D 1-6. The Many of the differences in musculature between aponeuroses to rays D 1-3 pass medial to the this and the preceding species result from the hypochordal longitudinalis. Anteriorly, the mass of more elongate form of Tydemania. The sections of the muscle attaches to the posterior surface of the the adductor mandibulae are somewhat reduced in skull. dorsoventral extent. A 3 is absent, having either OBLIQUUS SUPERIORIS (Figures 49, 50, 56a: OBL. fused with A 23 or having degenerated. The origin s.).—This section forms the main mass of the hyp- of A 23 thus now includes the lateral faces of the axialis, and consists of eighteen myomeres. It is symplectic, mesopterygoid and metapterygoid. The poorly developed anteriorly, where it attaches to tendon of A 1 is better developed. The origin of the extreme posterodorsal tip of the cleithrum. the dilatator operculi is reduced, while that of the (The anterior section probably contains part of anterior part of the levator operculi is larger. Ow- the lateralis superficialis, which usually forms a ing to the more dorsal position of the opercle and thin sheet covering both epaxial and hypaxial the smaller size of the dilatator operculi, a hiatus musculature near the anterior midline. The fibers has appeared between that muscle and the anterior apparently always pass anteroposteriorly, with little portion of the levator operculi. dorsal or ventral angulation. Owing to the extreme The intermandibularis is poorly developed, con- difficulty of observing the nature of this subdivision sisting of only a few fibers, and the hyohyoideus in preserved (or even fresh) material, it will not be inferioris is not as large as in T. anomalus. A few considered further. Suffice it to say that a detailed, fibers of the sphincter esophagi overlie the retractor accurate description of the lateral body muscles in dorsalis. Origin of the adductor radialis is restricted general, and that of the lateralis superficialis in to the ventral face of the third radial, the fibers particular, would require serial sectioning of the inserting only on the ventralmost pectoral fin ray. material. Greene (1913), in his diagnosis of the The adductor profundus pelvicus, which inserts lateralis superficialis, relies on color and histologi- on the pelvic fin rays in T. anomalus, is absent in cal criteria.) The obliquus superioris broadens out all other members of the Triacanthodinae examin- posteriorly, forming the dorsal and posterior bor- ed. The erectores and depressores dorsales serve ders of the abdominal cavity. It contributes fibers only the first three dorsal spines. In this species to the compound section of the hypaxialis inserting and the remaining members of the subfamily, the on the lateral bases of caudal rays V 1-6. inclinatores dorsales are present only in the soft OBLIQUUS INFERIORIS (Figures 49, 56a: OBL.L).— dorsal fin. 12 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

Macrorhamphosodes platycheilus Fowler is as for M. platycheilus. A single tendon passes FICURES 58-59 anteriorly, inserting only on the posteromedial face of the dentary. The protractor hyoidei is similar The larger specimen had the mouth twisted some to that of M. platycheilus, but there is no connec- 55 degrees to the left, while that of the smaller tion to the dentary. Instead, the fibers fade out in individual was twisted to a lesser extent to the the connective tissue between the interopercles, right. No differences that could be correlated with about halfway up the length of the snout. The the side to which the mouth was twisted were comments made in connection with M. platycheilus found, except in the adductor mandibulae. Here for the remaining muscular systems also apply to A 1 appears to have completely replaced A 2a, and includes the anterior faces of the preopercle and H. centriscoides. hyomandibular in its sites of origin. A 3 is absent. In both examples a single tendon passes anteriorly Subfamily HOLLARDIINAE and splits into two smaller ones. One of these in- serts on the anteroventromedial face of the dentary, Parahollardia line at a (Long ley) while the dorsal division attaches to the postero- medial face of the maxilla. The tendon to the FIGURES 61-63 maxilla is divided into two sections on the side The muscles of the cheek are very similar in to which the mouth is twisted, but not on the disposition and proportion to those described for opposite side. T. anomalus. The hiatus between the dilatator Owing to the considerably modified snout, the operculi and anterior part of the levator operculi, origin of the protractor hyoidei has shifted mainly which was apparent in all the triacanthodins ex- to the ventral faces of the interopercles and the amined except T. anomalus, is present in the two thick connective tissue surrounding them. The species of Hollardiinae dissected. The ligaments origin extends a third of the way along the snout attaching the hyohyoideus inferioris to the ventro- before grading into a single medial tendon which hyal are somewhat better developed. The adductor courses anteriorly between the interopercles to at- radialis inserts on the three ventralmost pectoral tach broadly to the ventromedial faces of the den- rays. The levator pectoralis inserts more on the taries. The intermandibularis and hyohyoideus dorsomedial surface of the cleithrum than on the inferioris are as described for Tydemania. supracleithrum. Figure 59 gives a general view of the branchial The arrector dorsalis pelvicus is virtually con- arch muscles not originating on the arches them- tinuous with the infracarinalis anterior anteriorly, selves, and is included to illustrate topographical being separated only by a myocomma. This applies differences too slight to be discussed in the text. to the arrector ventralis pelvicus as well. Two The adductor radialis originates only from the sec- adductor profundus pelvicus muscles are present, ond radial, and inserts on the lowermost pectoral the section to the second ray passing ventrolaterally ray. Inclinatores dorsales to the spiny dorsal fin are as opposed to posteroventrolaterally. An adductor absent, and only the first two spines have erector superficial pelvicus originates from the dorsal pel- and depressor muscles. The interradialis is better vis between the two slips of the adductor profundus developed, and the flexor dorsalis inserts on caudal pelvicus, and inserts on the tip of the posterodorsal rays D 1-6. There is less continuity between the flange of the first pelvic ray. sternohyoideus and the anteroventral component of the obliquus inferioris, mainly because the The inclinatores dorsales do not form a continu- ventrolateral border of the cleithrum and the dorso- ous sheet anteriorly, and insert variously on the lateral surface of the pelvis are very close to one pterygiophores and spine bases, the precise relation- another. ships differing on the left and right sides of the fish. All the spines receive erector and depressor muscles. Halimochirurgus centriscoid.es Alcock The infracarinalis medius consists of four myo- FIGURE 60 meres separated by myocommata. The ventrolateral The origin of the adductor mandibulae sections surface of the first anal pterygiophore receives NUMBER 155 13 fibers from two residual inclinatores anales, which but not P. lineata) in the smaller (53 mm) fish, lie anterior to the first inclinator inserting on a and of three sections in the larger (92 mm) exam- fin ray. The body and caudal fin muscles are as ple. A single anterior residual inclinator analis is was described for T. anomalus. present. Other muscles are as for P. lineata, or, if not Hollardia hollardi Poey mentioned in connection with that species, as described for T. anomalus. FIGURES 64-65, 68a The powerful tendon of the adductor mandib- Summary of Triacanthodidae ulae which inserts in the Meckelian fossa develops a few muscle fibers on its dorsal surface. These pass Certain regions of the body are more prone to anterodorsally to insert along the medial face of variation than others. Among the triacanthodids the dentary dorsal to the fossa. The adductor ra- dissected, significant variations are found in the dialis consists of two muscle slips to the ventralmost adductor mandibulae, intermandibularis, protrac- two pectoral rays. tor hyoidei, levator and dilatator operculi, ad- The first pelvic fin ray receives an adductor ductor radialis, the muscles of the pelvic fin rays, profundus pelvicus muscle, all other fin ray and the erector, depressor, and inclinator muscles muculature being absent. of the spiny dorsal fin. The last three of these Five inclinatores dorsales are recognizable, most of muscles involve numerical differences, and seem to them overlapping from the base of the spine onto be following a reductive trend. Apart from those the surrounding pterygiophore. The last two spines outlined above, the triacanthodid muscle plan do not have erector and depressor muscles. appears to be rather stable, in spite of the modifi- The infracarinalis medius is made up of two cations of the two long-snouted genera (Macro- sections (as in the other triacanthodids examined, rhamphosodes and Halimochirurgus).

Myological Descriptions of Representative Triacanthids

The general body outline of this family is rep- ure 686 of Triacanthus biaculeatus: A 1, A 2, A resented in Figure 2. One species from three out of 3).—The sections of this muscle are not well sep- the four recognized genera was dissected. They are arated, particularly in the posterior region. A 1 fairly slender fishes, with a rather long snout re- arises partly from the thick ligamentous fascia gion and a maximum body depth between the bases covering the muscle beneath the eye, and partly of the dorsal and pelvic fins. The tapering caudal from the anterodorsal face of the hyomandibular peduncle ends in a lunate caudal fin. They are (a flange of which overlies the posterodorsal region evidently capible of some speed and, unlike other of the muscle). A small portion of this section tetraodontiforms (except possibly Triodon), use inserts on the dorsomedial face of the maxilla. The the tail as the primary unit of propulsion (see A 2 section originates from the dorsal and lateral Wickler, 1963). quadrate and preopercle, and the anteroventral Triacanthids are relatively shallow-water fishes, hyomandibular. It forms a broad sheetlike tendon their food apparently being composed of soft- and anteriorly, which receives a substantial contribution hard-bodied marine invertebrates. They are usually from A 1, and inserts in the Meckelian fossa in considered to share a common ancestor with the the dentary. A 3 is possibly the most distinct sec- Triacanthodidae, but to be somewhat more tion, and originates from the lateral quadrate and specialized. preopercle. It lies medial to the tendinous plate in the middle of the adductor mandibulae, and Trixiphichthys weberi (Chaudhuri) joins the tendon to the Mecklelian fossa anteriorly. FIGURES 66, 67. 69-75 LEVATOR ARCUS PALATINI (Figure 66: L.A.P.).— The muscle is fairly small, and conical in shape. MUSCLES OF THE CHEEK It originates from the ventral face of the postorbital ADDUCTOR MANDIBULAE (Figure 66; see also Fig- process of the sphenotic and inserts on the antero- 14 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY dorsal face of the hyomandibular, including the ventral raphe with its antimere. The anterior face flange overlying section A 1 of the adductor of each side gives rise toa short ligament which mandibulae. passes anteromedially to attach to the ventrohyal. DILATATOR OPERCULI (Figure 66: D.O.).—Origin HYOHYOIDEI ABDUCTORES (Figure 66: H.AB.).— is from the sphenotic-pterotic suture region, the mus- There are four sections of this muscle, which insert cle coursing ventrally to insert on the tip of the on the anteromedial faces of the ventral four dorsal process of the opercle. It is relatively well branchiostegal rays. The first muscle is well de- developed. veloped, and originates from the posterior face of LEVATOR OPERCULI (Figure 66: L.O.).—The muscle the ventral flange of the ventrohyal. The second consists of two sections. The anterior portion origi- muscle arises from the base of the first ray, the nates from the pterotic and inserts on the dorso- two dorsal muscles originating from the fascia medial flange of the opercle. The posterior part covering the bases of the first two rays. arises from the dorsomedial face of the supraclei- HYOHYOIDEI ADDUCTORES (Figure 66: H.AD.).—A thrum, and inserts just behind the anterior portion well-developed muscle, interconnecting the four on the dorsomedial opercle. dorsalmost branchiostegal rays. It continues dor- ADDUCTOR ARCUS PALATINI (Figures 66, 67: sally, but soon fades out in the connective tissue A.A.P.).—Origin is from the lateral surface of the beneath the opercle and subopercle. It also con- prefrontal and the ventral flange of the parasphe- tinues inwards on the floor of the opercular cavity noid. It inserts on the dorsomedial faces of the for a short distance. mesopterygoid, metapterygoid, and hyomandibular. STERNOHYOIDEUS (Figures 66, 69: STH.).—The The muscle is well developed, and extends anterior muscle originates from the anterolateral face of the to the orbit. Here the origin of the fibers becomes ventral region of the cleithrum. In the extreme relatively more dorsal than the insertion, so that anteroventral area there may be some fusion of the fibers pass anteroventrolaterally. fibers from the two sides of the body. The postero- ADDUCTOR OPERCULI (Figures 69, 71: AD.OP.).— ventrolateral fibers are continuous with the ob- Fibers originate from the ventral surface of the liquus inferioris. The muscle, which is made up of pterotic just behind the origin of the levatores three myomeres, inserts on the lateral, posterior, interni. The muscle courses ventrolaterally to in- and ventral surfaces of the urohyal. sert on the dorsomedial face of the opercle. STERNOBRANCHIALIS (Figures 69, 70: STB.).—Ori- gin is from the anteroventral cleithrum. The fibers pass dorsally, and form a broad, sheetlike aponeu- MUSCLES OF THE HYOID REGION rosis which fuses briefly with that of the opposite INTERMANDIBULARIS.—This is a small muscle con- side before diverging again. The sheet then divides necting the halves of the lower jaw. It attaches to into a large anterior section and a smaller posterior the dentary below the Meckelian fossa, just anterior part. The former inserts on the posterodorsal to the quadrate. urohyal and the three hypobranchials. The poste- PROTRACTOR HYOIDEI (Figure 66: PR.HY.).—Origin rior sections cross over to the ventral midline and is from the lower jaw both dorsal and ventral to the insert on the fourth ceratobranchial of the other intermandibularis, the fibers passing posteroven- side. The aponeurosis from one side passes through trally. It separates into two halves in front of the the middle of the aponeurosis of the other side in ventrohyal, and inserts on the lateral anterohyal. It the course of this crossing over. The anterior also overlies the ventral two branchiostegal rays, aponeurosis is thickest in the region of the junction some fibers arising from the lateral face of the between the first and second hypobranchials. third ray. HYOHYOIDEUS INFERIORIS (Figure 66: HY.IN.).— VENTRAL BRANCHIAL MUSCLES The muscle arises from the ventrolateral face of the PHARYNGOCLAVICULARIS EXTERNUS (Figure 69: anterohyal, lying for the most part beneath the PHC.E.).—Fibers arise from the anterodorsal face first two branchiostegal rays, but with a small sec- of the cleithrum and pass slightly posterodorsally tion overlying the base of the first ray. It lies medial to insert on the ventrolateral face of the cerato- to the protractor hyoidei, and arises from a mid- branchial 5. NUMBER 155 15

PHARYNGOCLAVICULARIS INTERNUS (Figures 69, 70: LEVATORES INTERNI II-III (Figure 71: L.INT.).— PHC.I.).—The muscle passes anteriorly from its ori- These muscles originate from a depression in the gin on the anteromedial face of the cleithrum. It posterodorsal surface of the prootic. Levator II, splits into a series of tendons (of which the ventral- which has a slightly more lateral origin, inserts on most is largest) which insert on the ventral surface the dorsomedial face of infrapharyngobranchial 2, of ceratobranchial 5. while levator III inserts on the dorsolateral face OBLIQUI VENTRALES I—m (Figures 69, 70: OBL. of infrapharyngobranchial 3. v.).—The oblique muscles connect the ventral sur- OBLIQUUS DORSALIS HI (Figure 71: OBL.D.).—The faces of the cerato- and hypobranchials of the first muscle passes from the dorsal surface of epibranch- three arches. Obliquus I is well developed. Part of ial 3 to the dorsomedial face of infrapharyngo- obliquus III inserts directly on the third hypo- branchial 3. branchial. The most ventral fibers contribute to a OBLIQUUS POSTERIOR (Figure 71: OBL.P.).—The ligament which passes anteriorly to join that from fibers connect the anteromedial face of cerato- the other side and attach to the posterodorsal face branchial 5 to the posterior face of epibranchial 4. of the urohyal. TRANSVERSI DORSALES II-III (Figures 69, 71: TR. TRANSVERSI VENTRALES IV-V (Figure 70: TR.V.).— D.).—The anterior of these two muscles originates These two muscles connect the anteromedial faces from the ventrolateral face of the parasphenoid, of the fourth and fifth pairs of ceratobranchials and courses laterally to insert on the dorsolateral across the ventral midline. face of epibranchial 2. A few of the more posterior RECTI VENTRALES I, ii, AND iv (Figures 69, 70: fibers may fuse with their antimeres. Transversus RECT.V.).—The first of these muscles connects the III arises from a middorsal raphe with its antimere, posterolateral face of the dorsohyal to the antero- and passes laterally to insert on the posterodorsal lateral face of hypobranchial 1, the second passing face of epibranchial 3 beneath obliquus dorsalis III. from the dorsolateral surface of the urohyal to the RETRACTOR DORSALIS (Figure 71: D.RETR.).—Ori- ventral region of ceratobranchial 2. Rectus IV gin is from the ventrolateral surface of the first arises from the anteroventromedial face of cerato- vertebra, although a few of the more posterior branchial 4 and passes anteromedially to attach to fibers may arise from the anterodorsal tip of the the arch-shaped ligament between the ventral proc- swim bladder. It passes anteroventrally to insert on esses of the third hypobranchials. There is a partial the posteromedial face of infrapharyngobranchial 3. crossing over of the more medial fibers in this ADDUCTORES IV-V.—The anterior muscle connects region. the dorsomedial face of ceratobranchial 4 to the RECTUS COMMUNIS (Figure 70: R.COMM.).—The ventrolateral face of epibranchial 4. Adductor V muscle originates from the posterodorsal urohyal. arises from the anterodorsal tip of ceratobranchial It passes posteriorly, and grades into a tendon at 5 and passes anterodorsally to the posterolateral the level of the third hypobranchials. It inserts on face of epibranchial 4, lateral to the obliquus the anterolateral surface of ceratobranchial 5, posterior. lying lateral to the sections of the pharyngoclavic- SPHINCTER OESOPHAGI (Figures 70, 71: s.o.).— ulares. Ventrally the fibers attach to the ventromedial faces of the fifth ceratobranchials, there being a midventral raphe. The fibers pass dorsally around DORSAL BRANCHIAL MUSCLES the esophagus, attached to the fifth ceratobranch- ials, the more dorsal fibers attaching to the fourth LEVATORES EXTERNI I-IV (Figure 69: L.EXT.).— epibranchials. A small section overlaps the retractor This series of four muscles originates from the dorsalis anterodorsally. lateral surface of the prootic. The muscles become larger, and have a more lateral origin as they pass MUSCLES OF THE PECTORAL REGION posteriorly in the series. They insert aponeuroti- cally on the posterodorsolateral faces of epibranch- ABDUCTOR SUPERFICIALIS (Figure 66: ABD.S.).— ials 1-4, the second levator having a small inner Fibers originate from the posterior rim of the connection to epibranchial 3. cleithrum and course posterodorsally to insert on 16 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY the dorsolateral bases of the 13 principal fin rays. ARRECTOR VENTRALIS PELVICUS.—The muscle is ABDUCTOR PROFUNDUS (Figures 66, 69: ABD.P.).— only partially separated from the infracarinalis The muscle originates beneath the abductor super- anterior, and originates from the anteroventral ficialis from the posteroventrolateral deithrum and face of the pelvis. Insertion is on the anteroventro- the anterolateral coracoid. It inserts on the postero- lateral face of the pelvic spine. lateral bases of both the principal and the vestigial ADDUCTOR SUPERFICIALIS PELVICUS (Figure 69: fin rays. A.S.P.).—The muscle originates from the postero- ARRECTOR VENTRALIS (Figure 66: ARR.V.).—The medial face of the cleithrum and passes postero- fibers arise from the posterior face of the cleithrum ventrally and somewhat laterally to insert on the just above those of the abductor superficialis, and posterodorsal face of the dorsal flange of the pelvic insert on the lateral base of the medial half of the spine. vestigial fin ray. ADDUCTOR SUPERFICIALIS (Figure 73: ADD.S.).— MUSCLES OF THE DORSAL FIN Origin is from the posteromedial face of the clei- INCLINATORES DORSALES (Figures 66, 74: INC.).— thrum. The fibers of the dorsomedial part of the There are three inclinators to the spiny dorsal fin, muscle serve the more ventral fin rays, those of the arising in the fascia over the epaxialis and insert- more superficial, ventral section inserting on the ing on the posterolateral bases of the first three more dorsal fin rays. Tendons insert on the dorso- pterygiophores. Each fin ray of the soft dorsal fin medial faces of the principal rays a little distal to receives an inclinator attaching to its lateral base, their bases. the fibers arising in the spaces between the lateral ADDUCTOR PROFUNDUS (Figure 73: ADD.P.).—The flanges of the pterygiophores. The inclinators to muscle originates from the medial faces of the the spiny dorsal fin are flat and fan-shaped, those coracoid and cleithrum ventral to the adductor to the soft dorsal fin are smaller and more conical. superficialis and inserts on the ventromedial bases ERECTORES DORSALES (Figures 72, 74: EREC).—The of the principal fin rays. first five spines of the dorsal fin have erector mus- ARRECTOR DORSALIS (Figure 73: ARR.D.).—Fibers cles, of which the powerful, bipinnate muscle to arise from the medial face of the cleithrum be- the first spine is very much the largest. The tween the two adductor sections. Insertion is on erectors of the spiny fin originate from the lateral the dorsomedial face of the base of the vestigial fin faces of the pterygiophores (including the exoccip- ray. ital in the case of the first muscle), while those CORACORADIALIS (Figure 73: COR.R.).—The muscle to the soft fin arise from the lateral faces of the originates from the fascia covering the adductor pterygiophores and the neural spines. profundus and from the posteromedial coracoid. DEPRESSORES DORSALES (Figures 66, 72, 74: The fibers insert on the posteromedial process of DEPR.).—Depressors serve the first five spines and all the fourth radial. the fin rays, inserting on the posterolateral bases ADDUCTOR RADIALIS.—This small muscle origi- of these elements. Origin is from the lateral faces nates from the medial faces of the middle two of the pterygiophores for the muscles serving the radials and passes posteriorly to insert on the bases spines, and also includes the lateral faces of the of the ventralmost two pectoral fin rays. neural spines for the sections to the finrays . PROTRACTOR PECTORALIS (Figure 69: P.P.).—Fibers SUPRACARINALIS ANTERIOR.—The muscle is short arise from the ventrolateral region of the pterotic and powerful. It originates from the posterior face and insert on the anterodorsal face of the cleithrum. of the supraoccipital and inserts on the anterior face of the first basal pterygiophore of the spiny MUSCLES OF THE PELVIC GIRDLE dorsal fin. SUPRACARINALIS POSTERIOR (Figure 756: S.POST.).— ARRECTOR DORSALIS PELVICUS (Figure 69: A.D.P.).— The muscle arises from the posterior face of the Origin is from the lateral surface of the anterior last dorsal fin pterygiophore and passes posteriorly, region of the pelvis, the fibers inserting on the attaching to the tips of the neural spines of the anterodorsolateral face of the pelvic spine. The seventh to tenth caudal vertebrae, ending on the muscle is bipinnate. tenth neural spine. NUMBER 155 17

MUSCLES OF THE ANAL FIN H.L.).—Origin is from the lateral surface of the fused lower hypurals (1 and 2). The fibers pass INCLINATORES ANALES.—Each anal fin ray posses- posterodorsally to insert on the lateral bases of rays ses an inclinator muscle, originating in the fascia D 3-4, lateral to the epaxialis. overlying the hypaxial body musculature and in- FLEXOR DORSALIS (Figure 756: F.D.).—The muscle serting on the lateral base of the ray. An anterior, arises from the lateral faces of the penultimate and residual inclinator muscle is present, inserting on ural centra and the anterolateral region of fused the ventrolateral base of the first anal fin pterygio- hypurals 3 + 4. Insertion is on the dorsolateral phore. bases of rays D 1-6. ERECTORES ANALES.—Each fin ray has an erector FLEXOR DORSALIS SUPERIOR (Figure 75: F.D.S.).— muscle, which originates from the lateral face (s) of The muscle originates from the lateral and dorsal the pterygiophore, haemal spine, or both, and faces of the neural spines and centra of the ninth inserts on the anterolateral base of the ray. to eleventh caudal vertebra. It inserts tendinously DEPRESSORES ANALES.—The depressor muscles on the anterolateral base of ray D 6. have the same sites of origin as the erectors, and FLEXOR VENTRALIS (Figure 75: F.V.).—Origin is insert on the posterolateral bases of all the finrays . from the posteroventral face of the flange of the INFRACARINALIS ANTERIOR (Figure 69: INF.A.).— ural centrum, the ventrolateral faces of the ural The muscle arises from the -ventromedial face of centrum, and the fused hypurals 1+2. The muscle the posterior coracoid and from the fascia attached inserts on the lateral bases of rays V 1-6. to the posteromedial face of the cleithrum. It FLEXOR VENTRALIS EXTERNUS (Figure 756: F.V. inserts on the anteroventral face of the pelvis. E.).—The fibers arise from the anteroventral sur- INFRACARINALIS MEDIUS (Figure 66: INF.M.).—The face of the eleventh caudal vertebra and insert on fibers connect the posterodorsal face of the pelvis the lateral bases of rays V 1-2. with the ventrolateral tip of the first anal pterygio- FLEXOR VENTRALIS INFERIOR (Figure 75: F.V.I.).— phore. The muscle originates from the lateral and ventral INFRACARINALIS POSTERIOR (Figure 756: INF.P.).— faces of the ninth to eleventh caudal vertebra, and The fibers arise from the posterior face of the last inserts on the anteroventral base of ray V 6. anal fin pterygiophore, and course posteriorly. The muscle ends on the anterior tip of the haemal spine LATERAL BODY MUSCLES of the tenth caudal vertebra, the fibers attaching to the haemal spine tips of the intervening vertebra EPAXIALIS (Figures 66, 75a: EPAX.).—Two main (7-9). directions of the myocommata are evident, with a third anterodorsally. In the anterior region, the MUSCLES OF THE CAUDAL FIN fibers attach to the posterior face of the skull. Posteriorly the muscle fibers become intermingled INTERRADIALIS (Figure 75a: INT.).—The muscle with aponeuroses, which gives rise to a tough is made up of slips of fibers arising from the ventral fascia covering the intrinsic caudal fin musculature. bases of the dorsal rays and the dorsal faces of the Beneath this fascia the fibers grade into aponeuroses ventral rays, the resulting muscle bundles inter- which insert on the lateral bases of caudal rays mingling in the midlateral line. All fin rays possess D 2-6, and lie, for the most part, medial to the a slip of this muscle. hypochordal longitudinalis. PROXIMALIS (Figure 756: PROX.).—The muscle OBLIQUUS SUPERIORIS (Figures 66, 69, 75a; OBL. originates from the lateral face of the centrum of s.).—The section is poorly developed anteriorly, but the tenth caudal vertebra. It passes posteriorly to forms the main mass of the hypaxial body muscu- insert on the lateral flange of the ural centrum. lature posteriorly. In the region of the caudal (The flange arises from the posterior region of peduncle, the fibers become intermingled with the centrum, and continues posteriorly onto the aponeuroses, which join those of the epaxialis to anterior part of the fused third and fourth hy- give rise to the fascia mentioned above. Beneath purals.) this, the fibers mingle with those of the obliquus HYPOCHORDAL LONGITUDINALIS (Figure 75a: inferioris and grade into a series of aponeuroses 18 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY inserting on the lateral bases of rays V 1-6. Antero- The spinalis is less well developed than in the ventrally, some fibers arise from the facia overlying preceding species. the dorsal surface of the swim bladder. In the extreme anterior region, the fibers attach to the Tripodichthys oxycephalus (Sleeker) posterodorsal tip of the cleithrum. OBLIQUUS INFERIORIS (Figure 66: OBL.I.).—Antero- FIGURE 77 ventrally this section originates from the posterior coracoid and posterior and lateral cleithrum, being The adductor arcus palatini is well developed, continuous anterolaterally with the sternohyoideus. being of quite substantial thickness beneath the Most of these fibers insert on the anteroventral face eye. The adductor operculi possesses a small ante- of the postcleithrum. Above this, fibers arise from rior section which inserts on the posterodorsomedial the midline at the level of the epipleurals and face of the hyomandibular. The hyohyoidei adduc- pass ventrally to attach to the posterodorsal face tores are as described for T. biaculeatus, and the of the postcleithrum. The fibers become more ventralmost muscle of the hyohyoidei abductores horizontally oriented in the posterior region, and is well developed. There is no fusion of the fibers begin to mingle with those of the obliquus supe- of the sternohyoideus with those of the sterno- rioris at about the level of the anal fin, giving rise branchialis. The aponeurotic sheet of the latter to the aponeuroses to the caudal fin rays. muscle is continuous, and the portion inserting on ceratobranchial 4 does not cross over with the SPINALIS (Figure 72: SPIN.).—This muscle origi- section from the other side. The adductor radialis nates from the posterior face of the strutlike first is very small and inserts only on the lowermost dorsal pterygiophore and passes posteriorly. It pectoral ray. soon grades into a long tendon which attaches to the dorsolateral tips of the neural spines of the There are four inclinatores dorsales, one to the second to fifth vertebrae. It is a bilaterally sym- dorsolateral region of each pterygiophore. The metrical muscle, and may represent the first lateral flange on the ural centrum is less well myomere. developed, so that the proximalis is scarcely visible superficially. The flexor dorsalis does not have a section to ray D 6. Triacanthus biaculeatus (Bloch)

FIGURES 686, 76 Summary of Triacanthidae

There are few differences in the musculature. Little variation is apparent in the single species Owing to the shorter snout of T. biaculeatus, the of each of the three genera examined. The inclina- overall shape of the adductor mandibulae is some- tores dorsales of the spiny dorsal fin do vary con- what different (compare Figures 66 and 76), al- siderably, however, but these muscles appear to though the subdivisions are the same. The hyohy- have no functional significance since they are very oidei adductors are better developed, some fibers small and insert on the bases of the dorsal pterygio- continuing dorsally behind the opercle to attach to phores. There is also some variation in the extent the supracleithrum. There are only two inclinator of development of the hyohyoidei adductores and dorsales to the spiny dorsal fin, which insert on the the adductor radialis, but the family is basically lateral bases of the first two basal pterygiophores. conservative.

Myological Descriptions of Representative Balistids

The general body outline of this family is rep- possess a terminal mouth armed with incisiform resented in Figure 3. Representative species from teeth (and two canines in Odonus). The snout five of the eleven genera recognized by Berry and region is long, with the eye high up on the head. Baldwin (1966) were dissected. Members of this The first dorsal fin consists of three spines, the family are very conservative in body form. They heavy, large first spine being locked in position by NUMBER 155 19

the small second spine. Pelvic fins are reduced to and originates from the lateral metapterygoid and rudiments at the end of a long, shaftlike pelvic mesopterygoid, being completely overlain by A bone (Tyler, 1962b). The soft dorsal and anal fins la. The two sections insert by a common tendon on are well developed, consisting of many rays. A the posteroventral and ventromedial faces of the slightly narrowed caudal peduncle ends in a double maxilla. A 2 is also subdivided into two parts. A emarginate or rounded caudal fin, although other 2a is superficial and originates from the latero- shapes do occur. Scales are large and heavy, and ventral ethmoid, the prefrontal, the infraorbital may possess longitudinal ridges (for the functional ligament and the lateral hyomandibular. It is significance of which, see Wahler and Wahler, separated from A 23 by the path of ramus mandib- 1964). Balistids occur mainly on and around coral ularis V, which emerges beneath the maxilla. The reefs, and are diurnal. A few oceanic forms exist, section largely overlaps the subdivisions of A 1, primarily in tropical waters. The food of the fam- and also covers the more dorsal portion of A 23. ily is comprised of both soft- and hard-bodied The latter section originates from the anterior hyo- marine invertebrates (often corals), although cer- mandibular, anterodorsal preopercle, and the tain species show a preference for plant material lateral metapterygoid. A 2v develops along the (see, for example, Melichthys, in Randall, 1967). ventrolateral extent of the preopercle. The sections Fast swimming is accomplished by synchronous of A 2 insert broadly on the posterodorsomedial side-to-side beating of the dorsal and anal fins, the face of the dentary, with some fibers of the first caudal fin being used for bursts of speed. Slow two sections inserting on the maxillomandibular swimming is achieved by undulations of the dorsal ligament. A deep section, arising from the lateral and anal fins. When stationary, the beat passes faces of the symplectic, quadrate, and metaptery- posteriorly in the dorsal fin and anteriorly in the goid appears to represent A 3. It fuses with the anal fin. This can be shown by a parallelogram of tendon of A 23 anteriorly. forces to help in counteracting the larger gravita- LEVATOR ARCUS PALATINI (Figure 78: L.A.P.).— tional forces acting on the head region. The muscle is fairly small, and originates from the Balistids are supposed to be derived from tri- ventral face of the orbital process of the sphenotic acanthid-like fishes, and to have "given rise" to the and the anterodorsal pterotic. The fibers pass an- Monacanthidae. Using a more modern approach, teroventrally to insert on the anterodorsal face of they are apomorph relative to the former, and rep- the hyomandibular. resent the plesiomorph sister group of the latter. DILATATOR OPERCULI (Figure 78: D.O.).—Nor- However, recent work (Tyler, 1970a:31) indicates mally in the shape of an inverted triangle, this that the ostracioids may have to be included some- muscle has inverted its shape anteriorly, the some- where in this lineage, probably sharing a common what tendinous apex arising from the orbital proc- ancestor with the balistoids. ess of the sphenotic. Some of the medial fibers are continuous with the posterolateral fibers of the levator arcus palatini. The more posterior fibers Batistes vetula Linnaeus arise from the ventrolateral face of the sphenotic. Insertion is along the dorsolateral face of the FIGURES 78-85 opercle. LEVATOR OPERCULI (Figure 78: L.O.).—Origin is MUSCLES OF THE CHEEK from the anteroventrolateral pterotic and the ante- rior supracleithrum, the fibers inserting on postero- ADDUCTOR MANDIBULAE (Figures 78, 79; and Fig- dorsolateral face of the opercle. ure 87a of Balistapus undulatus: A la A 13, ADDUCTOR ARCUS PALATINI (Figure 79: A.A.P.).— A 2a, A 23, A 2Y, A 3).—Section A 1 is separable The fibers arise from the anterolateral ethmoid and into two portions, one superficial (A la) and the the lateral parasphenoid in front of the orbit. The other deep (A 13). The former originates from more anterior fibers course in an increasingly an- the lateroventral faces of the ethmoid and pre- teroventrolateral direction. These fibers are antero- frontal, the infraorbital ligament and the antero- dorsally continuous with the anteroventral fibers of dorsal face of the hyomandibular. A 13 is small, the retractor arcus palatini. Insertion is on the 20 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY medial and dorsal faces of the mesopterygoid, lying the anteroventral region of the urohyal and metapterygoid, and the hyomandibular. pass dorsolaterally to insert along the ventral face RETRACTOR ARCUS PALATINI (Figures 78, 79: of the anterohyal. R.A.P.).—The muscle originates from the ventral HYOHYOIDEI ABDUCTORES (Figure 78: H.AB.).—The surface of the lateral flange of the ethmoid, the muscles serve the ventral five branchiostegal rays. lateral face of that bone, and the lateral para- The muscle to the first ray is much enlarged and sphenoid. It consists of two subsections. The more sheetlike. The fibers from the other four sections anterior, superficial fibers insert on the shaft of join that from the first ray and attach anteriorly the palatine and the ectopterygoid. From the dorsal to the ventromedial faces of the anterohyal, dorso- process of the ectopterygoid, a tendinous fascia hyal, and ventrohyal. passes back to below the prefrontal, separating the HYOHYOIDEI ADDUCTORES (Figure 78: H.AD.).— subsections completely posteriorly. The postero- The muscle is in the form of a continuous sheet ventral subsection is covered by a thick fascia, to covering the wall of the opercular cavity. Fibers which many of the fibers attach. It inserts on the arise from the fascia over the sternohyoideus and dorsal faces of the mesopterygoid and metaptery- sternobranchialis and pass dorsolaterally, fading out goid. The fibers mingle with those of the adductor in the connective tissue beneath the opercle and arcus palatini anteroventrally and those of the subopercle. The branchiostegal rays are buried in dorsal subsection anterodorsally. Both subsections the same connective tissue as the muscle fibers, but are served by the same branch of the ramus buccalis there is little direct attachment. that serves the adductor arcus palatini. STERNOHYOIDEUS (Figures 78, 80: STH.).—The ADDUCTOR OPERCULI (Figure 80: AD.OP.).—Origin posterolateral fibers are continuous with the antero- is from the anterolateral pterotic and lateral lateral fibers of the obliquus inferioris. The more prootic. The fibers pass anteroventrolaterally to medial fibers arise from the anteroventral face of insert on the dorsomedial face of the opercle. the cleithrum. A partially separated bundle of ADDUCTOR HYOMANDIBULAE (Figure 80: A.H.).— fibers originates from the anteroventral tip of the The muscle originates behind and above the ad- cleithrum and passes almost dorsally to the postero- ductor operculi from the anteroventrolateral faces ventral tip of the urohyal. The muscle inserts of the pterotic and prootic. It inserts on the dorso- broadly on the urohyal, with some attachment to medial face of the hyomandibular, immediately be- the posteroventromedial face of the ventrohyal. low the articulation with the skull. STERNOBRANCHIALIS (Figures 80, 81: STB.).—The muscle originates from the anteroventral face of the cleithrum, and passes vertically upward to MUSCLES OF THE HYOID REGION split into three sections. The anteriormost inserts INTERMANDIBULARIS.—Lying immediately behind on the posteroventral tip of the urohyal, the middle the symphysis of the halves of the lower jaw, this section inserting on the ventromedial face of muscle (which is wholly dorsal to the protractor hypobranchial 3. The third section joins its anti- hyoidei) connects the posterodorsomedial faces of mere ventral to the rectus communis. the left and right dentaries. PROTRACTOR HYOIDEI (Figure 78: PR.HY.).—There VENTRAL BRANCHIAL MUSCLES appear to be two sections of this muscle. A super- ficial division originates completely from the skin PHARYNGOCLAVICULARIS EXTERNUS (Figures 80, 81: and joins the deeper section before inserting on PHC.E.).—Originating from the cleithrum immed- the posteroventral margin of the lower jaw. The iately behind the sternobranchialis, the muscle larger inner portion originates almost entirely from courses posterodorsally to insert on the anteroven- the lateral faces of the branchiostegal rays and the tral face of ceratobranchial 5. anterohyal. There is some separation into three PHARYNGOCLAVICULARIS INTERNUS (Figures 80, 81: longitudinal bundles posteriorly. The more ventral PHC.I.).—The muscle originates from the antero- fibers meet their artimeres in the ventral midline. medial face of the cleithrum. It passes anteriorly HYOHYOIDEUS INFERIORIS (Figure 78: HY.IN. ) to insert on the ventromedial bases of cerato- The fibers arise from a midventral raphe under- branchials 4 and 5. NUMBER 155 21

OBLIQUI VENTRALES I-III (Figure 81: OBL.V.).— to the third arch is not well developed. A fifth These muscles connect the ventral faces of the levator is present, inserting on the dorsal tip of the ceratobranchials and hypobranchials of the first fifth ceratorbranchial, the more anterior fibers three arches. In the case of obliquus HI, most of fusing with the posterior region of levator IV. This the fibers from the ceratobranchial attach to the muscle appears to be divided into two sections by arch-shaped ligament between the ventral processes a myocomma, which seems to cast aspersions on a of the third hypobranchials, the section actually theory postulating its development solely from inserting on the ventral process of hypobranchial levator IV. 3 being fairly distinct. LEVATORES INTERNI II-III (Figure 82: L.INT.).— TRANSVERSI VENTRALES IV-V (Figure 81: TR.V.).— These muscles originate one on either side of the The anterior muscle connects the anteroventrome- facial foramen on the ventral surface of the dial faces of the fourth ceratobranchials across the prootic. The anterior muscle inserts on the dorso- ventral midline. There are two sections to the pos- medial face of infrapharyngobranchial 2, the poste- terior muscle. The anterior of these lies between the rior one inserting on infrapharyngobranchial 3. extreme anteroventromedial tips of the fifth cera- OBLIQUUS DORSALIS HI (Figure 82: OBL.D.).—There tobranchials and is small. There is a small hiatus are two sections of this muscle. The anterior one between it and the posterior section, which lies connects the anterodorsal face of epibranchial 3 to ventral to the sphincter oesophagi. the anterior and lateral faces of infrapharyngo- RECTI VENTRALES I, ii, iv (Figures 80, 81: RECT. branchial 3. The posterior part passes from the dor- v.).—The first rectus joins the anteroventrolateral sal face of epibranchial 3 to the dorsomedial face face of ceratobranchial 1 to the posterodorsal face of infrapharyngobranchial 3. of the dorsohyal. Some of the posteromedial fibers OBLIQUUS POSTERIOR (Figure 82: OBL.P.).—The are continuous with those of obliquus ventralis I. muscle connects the dorsomedial face of cerato- Rectus II arises from the dorsolateral face of the branchial 5 with the posterolateral face of epibran- urohyal and attaches to the posterolateral face of chial 4. ceratobranchial 2. Rectus IV is represented by two TRANSVERSI DORSALES II-III (Figure 82: TR.D.).— sections. The more ventral one arises from the ven- The anterior muscle originates from the ventro- tral face of ceratobranchial 4 and passes antero- lateral parasphenoid, with some posterior fibers medially to attach to the posterior face of the from a raphe in the dorsal midline. The fibers arch-shaped ligament between the third hypobran- insert on the dorsal face of epibranchial 2. The chial processes. The other section arises from the posterior muscle arises from the dorsal midline and dorsolateral face of ceratobranchial 4 and attaches inserts primarily on epibranchial 3, with some to the dorsomedial face of hypobranchial 3. This fibers to epibranchial 4. latter muscle is apparently a derivative of obliquus RETRACTOR DORSALIS (Figure 82: D.RETR.).—A ventralis III. well-developed muscle, arising from the lateral and RECTUS COMMUNIS (Figures 80, 81: R.COMM.).— ventral faces of the basioccipital and inserting on Origin is from the dorsolateral urohyal. The fibers the dorsomedial region of infrapharyngobranchial pass posteriorly, grading into an aponeurosis at 3. about the level of the third hypobranchial. The ADDUCTORES IV-V (Figure 82: AD.).—The muscle tendon lies medial to the pharyngoclavicularis of the fourth arch spans the medial side of the externus and lateral to the pharyngoclavicularis articulation between the epibranchial and cerato- internus, and inserts on the ventrolateral face of branchial. Adductor V connects the anterodorsal ceratobranchial 5. tip of ceratobranchial 5 and the posterodorsal face of epibranchial 4. SPHINCTER OESOPHAGI (Figures 81, 82: s.o.).—The DORSAL BRANCHIAL MUSCLES muscle encircles the esophagus, attaching mainly LEVATORES EXTERNI I-V (Figure 80: L.EXT.).—The to the posterior margins of the fifth ceratobran- first four leva tors originate from the ventral sur- chials, and lying above transversus ventralis V. face of the prootic shelf and pass ventrally to the Dorsally it attaches to epibranchial 4, with a few dorsolateral faces of epibranchials 1-4. The muscle fibers passing to infrapharyngobranchial 3. There 22 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY is a small bundle of fibers crossing the dorsal mid- MUSCLES OF THE PELVIC GIRDLE line above the retractor dorsalis. ARRECTOR VENTRALIS PELVICUS? (Figure 84: A.V.P.).—The muscle originates in the ventral groove MUSCLES OF THE PECTORAL REGION in the pelvis. Just before inserting, the fibers divide into left and right tendons. These continue pos- ABDUCTOR SUPERFICIALIS (Figure 78: ABD.S.).— teriorly, dorsal to the first and second encasing Origin is from the posterolateral face of the scales to attach to the ventrolateral faces of the cleithrum, the fibers passing posterodorsally to fin ray element. insert on the dorsal bases of the principal fin rays. ADDUCTOR SUPERFICIALIS PELVICUS? (Figure 84: ABDUCTOR PROFUNDUS (Figures 78, 80: ABD.P.).— A.S.P. ).—This bilateral muscle originates in the The fibers originate almost exclusively from the dorsal groove of the pelvis. Posteriorly, the tendons anterolateral face of the coracoid, a few fibers pass through a tunnel in the dorsal lobe of the coming from the posterolateral cleithrum. The pelvis and insert on either side of the dorsal half muscle inserts on the ventrolateral bases of the of the fin ray element. principal fin rays and on the lateral base of the lateral half of the vestigial ray. MUSCLES OF THE DORSAL FIN ARRECTOR VENTRALIS (Figures 78, 80: ARR.V.).— The muscle originates from the posterolateral face INCLINATORES DORS ALES (Figure 78: INC.).—There of the cleithrum and inserts on the lateral face of are no inclinators to the spiny dorsal fin. Each ray the medial half of the vestigial finray . in the soft dorsal fin receives an inclinator, which ADDUCTOR SUPERFICIALIS (Figure 83: ADD.S.).— inserts on the lateral base of the ray just above Origin is from the anteroventral face of the post- the insertion of the erector and depressor muscles. cleithrum and the posteromedial face of the dorsal The fibers originate in the fascia overlying the region of the cleithrum. The fibers insert on the epaxialis. dorsomedial faces of the principal rays distal to ERECTORES DORSALES (Figures 78, 80: EREC).— their bases, the more medial fibers serving the more The first erector is very powerful and passes from ventral rays. the dorsal faces of the frontal and supraoccipital ADDUCTOR PROFUNDUS (Figure 83: ADD.P.).—The to the anterolateral base of the first dorsal spine. muscle passes from the posterior cleithrum and The second erector arises from the posterior epiotic medial coracoid to insert on the ventromedial and, passing medial to the strut supporting the flanges of the bases of the principal fin rays. first spine, inserts on the lateral base of the second ARRECTOR DORSALIS (Figure 83: ARR.D.). — The spine. There is no erector to the third spine. The fibers originate from the medial cleithrum between erectors of the soft dorsal fin originate from the the two adductor sections, and insert on the medial lateral faces of the posterior halves of the ptery- base of the vestigial finray . giophores (just behind the lateral flanges) and the CORACORADIALIS (Figure 83: COR.R).—The mus- neural spines. The fibers reach the centra, and in- cle originates largely from the fascia overlying the sert on the anterolateral bases of the fin rays. medial surface of the adductor profundus. It at- DEPRESSORES DORSALES (Figures 78, 80: DEPR.).— taches to the medial process of the ventral radial. The depressors of the first two spines arise from PROTRACTOR PECTORALIS (Figure 80: P.P.).—The the epiotic. The first passes medial to the lateral fibers fan out from their origin on the ventral tip strut of the pterygiophore before turning dorsally of the pterotic and insert on the anterodorsal face to insert on the posterolateral base of the spine, of the cleithrum. The more anteromedial fibers while the second muscle passes lateral to the strut spread out to cover the posteromedial wall of the to insert on the anterior face of the elongate, branchial chamber. posteroventrally directed base of the second spine, LEVATOR PECTORALIS (Figure 80: TR.).—Origin is ventral to its articulation with the lateral strut. from the posteroventral tip of the pterotic, the The depressor of the third spine originates from muscle passing posteriorly to insert on the antero- the anteroventral face of the first pterygiophore dorsal tip of the cleithrum. of the soft dorsal fin. It passes anterodorsomedially, NUMBER 155 23

through a foramen in the pterygiophore, to insert infracarinalis anterior anterolaterally. on the base of the third spine. The depressors of INFRACARINALIS POSTERIOR (Figure 85: INF.P.).— the soft dorsal fin originate from the anterolateral Fibers arise from the posteroventrolateral tip of the faces of the pterygiophores and the neural spines, last anal pterygiophore and pass posteriorly, at- and insert on the posterolateral bases of the rays. taching to the tips of the haemal spines of the SUPRACARINALIS ANTERIOR (Figure 78: S.ANT.).— eighth and ninth vertebrae before ending on the The muscle connects the posterodorsolateral face midregion of the anterior face of the haemal spine of the frontal to the anterolateral edge of the first of the tenth caudal vertebra. basal pterygiophore of the spiny dorsal fin. SUPRACARINALIS MEDIUS (Figure 80: S.MED.).— MUSCLES OF THE CAUDAL FIN The muscle attaches to the posterolateral face of the second pterygiophore and fans out to insert INTERRADIAUS (Figure 85a: INT.).—The muscle broadly on the anterior face of the first pterygio- consists of a number of bundles of fibers passing phore of the soft dorsal fin. between the rays and meeting at the midlateral SUPRACARINALIS POSTERIOR (Figure 85; s. POST.).— line. Fibers usually pass over two or three rays Fibers arise from the posterodorsolateral face of before reattaching. the last pterygiophore of the soft dorsal fin and HYPOCHORDAL LONGITUDINALIS (Figure 85a: the tip of the neural spine of the eighth caudal H.L.).—Origin is from the lateral faces of hypural vertebra. The fibers attach to the tips of the ninth 2 and the urostyle, the fibers coursing posterodor- and tenth neural spines. sally to insert on the lateral bases of the fin rays D 2-4. FLEXOR DORSALIS (Figure 856: F.D.).—The muscle MUSCLES OF THE ANAL FIN originates from the dorsolateral faces of the tenth INCLINATORES ANALES (Figure 78: INC.).—Each caudal vertebral centrum and the urostyle. Fibers fin ray possesses an inclinator arising from the insert on the anterolateral bases of the rays D 1-6. fascia overlying the hypaxial body musculature FLEXOR DORSALIS SUPERIOR (Figure 85: F.D.S.).—A and inserting on the lateral base of the ray. There well-developed muscle, originating from the dorso- is a large residual inclinator, which inserts on the lateral faces of the neural spines and centra of the base of the first anal pterygiophore. The indinators eighth and ninth caudal vertebrae, and inserting on decrease in size posteriorly. the anterodorsolateral base of fin ray D 6. ERECTORES ANALES.—Each fin ray has an erector FLEXOR VENTRALIS (Figure 856: F.V.).—Origin is muscle, which originates from the lateral faces of from the ventrolateral face of the hypural plate, the pterygiophore and the haemal spines, and in- the ural, and the penultimate centrum. The fibers serts on the anterolateral base of the ray. insert on the anterolateral bases of rays V 1-6. DEPRESSORES ANALES.—The depressor muscles FLEXOR VENTRALIS INFERIOR (Figure 85: F.V.I.).— originate on the lateral faces of the pterygiophores A well-developed muscle, arising from the ventro- and haemal spines, and insert on the posterolateral lateral faces of the haemal spines and vertebral bases of each finray . centra of the eighth and ninth caudal vertebrae, INFRACARINALIS ANTERIOR (Figures 78, 80: INF. with a few fibers from the tenth haemal spine. The A.).—The muscle is well developed. It arises from the muscle inserts on the anteroventrolateral base of posteroventral region of the cleithrum and inserts ray V 6. broadly over the lateral face of the pelvic bone. TRANSVERSUS CAUDALIS (Figure 85&: TR.C).— Posteriorly, it passes into a lateral tunnel in the Fibers originate from the dorsolateral face of the bone, to whose walls fibers attach. hypural plate beneath the urostyle and pass postero- INFRACARINALIS MEDIUS (Figures 78, 80: INF.M.).— ventrally to insert on the anterodorsal base of ray The fibers arise from the posterolateral face of the V 1. pelvic bone and pass posterodorsally to join with the fibers of the obliquus inferioris before acquir- LATERAL BODY MUSCLES ing attachment to the first anal pterygiophore. The muscle overlaps the posterior fibers of the EPAXIALIS (Figures 78, 85a: EPAX.).—Anteriorly 24 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

the fibers arise from the supracleithrum, posttem- thus exposing more of A la in lateral view. Ramus poral, pterotic, epiotic, and exoccipital. The muscle mandibularis V passes through A 2 a dorsal to the is fairly uniform as it passes posteriorly, the myo- division between the two upper A 2 sections. A 2 Y commata becoming more horizontally oriented. is better developed, and is bipinnate. The fibers grade into a fascia which inserts on the There is a hiatus between the extreme posterior lateral bases of rays D 1-6. fibers of the protractor hyoidei and the fibers OBLIQUUS SUPERIORIS (Figures 78, 85a: OBL.S.).— anterior to them, and the hyohyoideus inferioris Anteriorly the fibers lie beneath the anteroventral is visible externally through the gap. downgrowth of the fibers of the epaxialis. The The medial section of the sternohyoideus has section does not reach the pectoral girdle anterior- separated completely from the main mass of the ly, and these fibers are also overlain by the muscle. There is, in addition, another section which anterodorsal region of the obliquus inferioris. They has separated off from the main mass. It passes become exposed posteriorly, where they grade into horizontally from the cleithrum (lateral to the a fascia with the obliquus inferioris to attach to sternobranchialis) to insert on the urohyal just the lateral bases of rays V 2-6. Dorsomedially, the lateral to the insertion of the vertical section. A fibers are attached to the centra of the vertebrae. myocomma is present across the middle of the OBLIQUUS INFERIORIS (Figure 78: OBL.I.).—The length of this horizontal section. Rectus ventralis I anteroventral section of the muscle arises from the has completely separated from obliquus ventralis I. posterior face of the coracoid, the lateral fibers The first levator externus is better developed being continuous with the sternohyoideus. Most of and the fifth less well developed than in B. vetuln. the fibers from this region attach to the anteroven- Transversus dorsalis II originates only from the tral face of the postcleithrum. Above the postclei- lateral parasphenoid, with no fibers from a mid- thrum, the fibers arise from the fascia overlying the dorsal raphe. The arrector ventralis pelvicus is as region of the epipleurals and course anteroventral- for B. vetula, except that the tendons join again ly to attach to the posterodorsal face of the post- before inserting on the fin ray element. The trans- cleithrum. The more posterior fibers become versus caudalis is not as well developed as in B. increasingly horizontal, and the section apparently vetula. fuses with the obliquus superioris to insert com- poundly on rays V 2-6. Melichthys niger (Bloch) SPINALIS (Figure 80: SPIN.).—Fibers arise from the posterodorsal face of the epiotic and pass FIGURE 88 posteriorly to attach to the tips of the neural spines A 23 is more exposed posterodorsally, and A 2Y of the second and third abdominal vertebrae. is more obviously bipinnate. The hyomandibular crest is somewhat larger, with the result that there CUTANEOUS MUSCLES is more of the levator arcus palatini visible lateral-

TRANSVERSUS CUTANEOUS (Figure 78: TR.CUT.).— iy. The fibers of this muscle arise from the fascia The cutaneous portion of the protractor hyoidei overlying the anteroventral portion of the obliquus is much better developed and overlies a greater inferioris and pass ventromedially to attach to the part of the ventral branchiostegal rays, while the ventrolateral face of the pelvis. The posterodorsal inner and more ventral sections are reduced. There fibers attach to the anteroventral tip of the post- is some development of the horizontal section of the cleithrum. The section probably arose from the sternohyoideus described in B. undulatus. The infracarinalis medius as postulated by Rosen (1913: posterodorsomedial fibers of the sternobranchialis 3), although he did not distinguish the infracari- do not meet their antimeres. The pharyngoclavicu- nalis medius from the obliquus inferioris. laris internus has partially divided into two por- tions. Levator externus V is less well developed Balistapus undulatus (Mungo Park) than in either of the two previous species, and transversus dorsalis II is as for B. vetula. The arrec- FIGURES 86, 87a tor ventralis pelvicus and transversus caudalis are A 2o is not as well developed anterodorsally, as for B. undulatus. NUMBER 155 25

Rhinecanthvs aculeatus (Linnaeus) of the sternohyoideus are present, and the pharyn- goclavicularis externus is poorly developed. FIGURE 89 Only the ventral fibers of levator externus V are The levator arcus palatini is poorly developed, developed, and these grade into an aponeurotic with only a few fibers to the anterodorsal face of sheet dorsally, which fuses to the outer face of the the hyomandibular, while the levator operculi is midregion of levator IV. Both levatores interni are well developed, extending inward beneath the di- poorly developed, and transversus dorsalis II is as for latator operculi to the ventral process of the B. vetula. The protractor pectoralis is well develop- pterotic. ed. The arrector ventralis pelvicus is as for R. The cutaneous portion of the protractor hyoidei aculeatus, the movable portion of the pelvis being is well developed, and there is a fifth section of much reduced. The inclinator analis of the first the hyohyoidei abductors. The sternohyoideus is anal fin ray appears to consist of two sections, and as was described for B. undulatus. the transversus caudalis is small. The fifth levator externus is well developed, and partially fused to the fourth levator anterodorsally. Transversus dorsalis II arises only from the para- Summary of Balistidae sphenoid. In the balistids, there is some variation in the The posterior region of the groove in the pelvis relative development of the sections of the adduc- which contains the arrector ventralis pelvicus has tor mandibulae, and minor differences in the size closed over ventrally to form a tube. The tendon of the levator arcus palatini. The development of remains single throughout its length, but divides the cutaneous portion of the protractor hyoidei immediately before its insertion. differs, as does the relative size of the sections of the hyohyoidei abductores. Two subdivisions of Odonus niger (Ruppell) the sternohyoideus are usually present (one verti- cal and the other horizontal). The extent to which FIGURE 90 levator externus V develops varies, and transversus The adductor mandibulae is somewhat reduced dorsalis II may or may not have fibers arising from in size, particularly sections A 2<* and A 2P. A 3 the midline behind its main origin on the para- is also reduced, but A la and A IP are better sphenoid. There is some variation in the posterior developed than in the other balistids examined. region of the arrector ventralis pelvicus, and the The cutaneous portion of the protractor hyoidei transversus caudalis is usually poorly developed. is well developed, with some fibers of the more While there is more variation in the Balisti- dorsal portions originating from the dorsal three dae than in either of the two preceding fam- branchiostegal rays. The sections of the hyohyoidei ilies, it is usually of a minor nature and relatively abductores from the second to fifth rays are much unimportant at the level of analysis to be used reduced. The horizontal and vertical subdivisions here.

Myological Descriptions of Representative Monacanthids

The general body outline of representatives of normally two (the second being able to lock the this family is given in Figure 4. Out of the 22 first in position), vary in position from behind the genera recognized by Fraser-Bunner (1941a), species eye to the anterior snout region. The pelvis may representing 10 genera have been examined. Body be in evidence externally, as in balistids, or may form is basically conservative, and similar to balis- simply be a movable strut of bone completely tids, although there are numerous exceptions such covered by skin. Locomotion is mainly by undula- as the long snouted Oxymonacanthus and the tre- tions of the soft dorsal and anal fins, turning being mendously elongate Anacanthus. The scales are achieved by the pectoral fins and flexion of the spinose. The dorsal spines, of which there are body. The caudal fin does not seem to be used at 26 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY all in Pseudalutarius (Wickler, 1962), although it ventral fibers originate from the extreme ventral may be brought into play in escape movements. flange of the preopercle, but do not form a distinct The teeth are much the same as in the balistids, section as in the balistids. A 2a and A 23 insert by but the general body plan is less robust. a common tendon on the medial face of the den- The food of monacanthids is rather varied. Some tary. A 3 is completely separate from A 2 except for species are totally herbivorous, others omnivorous, the shared tendinous insertion on the dentary. It while yet others ingest primarily animal matter. originates from the lateral faces of the metaptery- They are usually found around coral reefs and on goid and quadrate. the sea-grass flats of tropical waters, although some LEVATOR ARCUS PALATINI.—The muscle is poorly species do occur in colder regions. developed, and no longer visible in superficial lateral view. It originates from the ventromedial face of the orbital process of the sphenotic and the Pervagor tnelanocephalus (Sleeker) anterolateral prootic. The fibers fan out to insert on the posterodorsal and posterolateral faces of the FIGURES 87b, 91-99 hyomandibular, the two regions being separated by MUSCLES OF THE CHEEK a thin ridge. DILATATOR OPERCULI (Figure 91: D.O.).—Origin ADDUCTOR MANDIBULAE (Figures 876, 91, 92; see is from the ventrolateral face of the orbital process also Figure 101 of Stephanolepis auratus: A la, of the sphenotic and the pterotic, the fibers insert- A IP, A 1Y, A 2a, A 23, A 3).—The A 1 section ing on the anterolateral face and dorsal rim of the of the adductor mandibulae has become consider- opercle. ably modified. A la and A la' arise from the lateral LEVATOR OPERCULI (Figure 91: L.O.).—The muscle face of the ethmoid (and the ventral surface of the originates from the anterodorsal supracleithrum lateral flange) as well as from the prefrontal and and anterolateral posttemporal and inserts on the the infraorbital ligament (between the prefrontal posterodorsal ridge of the opercle. and hyomandibular). A la' (Figure 91) grades into ADDUCTOR ARCUS PALATINI (Figure 92: A.A.P.).— an aponeurotic sheet which fades into the tissues of Fibers arise from the ventrolateral faces of the the upper lip. It is a fairly distinct subsection, vomer and parasphenoid and insert on the dorso- although there is some intermingling of fibers pos- medial hyomandibular and metapterygoid and teriorly. The main mass of A 1 is formed by A la, dorsal mesopterygoid. The anterior region of the which inserts on the posteroventrolateral face of pterygoids underlies the parasphenoid, with which the maxilla by a broad aponeurosis. Medial to this it is in contact dorsally. The muscle is not devel- are two other sections, arising from the infraorbital oped in the posterior half of the floor of the orbit. ligament. The dorsal of these, A h (see Figure RETRACTOR ARCUS PALATINI (Figures 91, 92: R.A. 101) soon grades into a flat aponeurosis which P.).—The muscle has completely separated from the attaches to the lateral process of the ethmoid sup- adductor arcus palatini, and the division of the porting the palatine, and the palatine itself. Its fibers into anterior and posterior sections as in function is obscure. Beneath this muscle is A IP' balistids is not nearly as striking (the ventral (Figure 101), a smallish, somewhat flattened muscle aponeurosis of the anterior portion being absent). which develops a tendon anteriorly joining that of A small anterolateral bundle arises from the lateral A la laterally and that of A 13 medially. A IP process of the ethmoid. Fibers from this region originates from the lateral faces of the mesoptery- insert mainly on the cartilagenous plug between goid and metapterygoid and inserts on the medial the ectopterygoid, maxilla, and vomer, the rest of face of the maxilla. A 2 consists of two subdivisions. the muscle inserting predominantly on the ecto- A 2a originates from the prefrontal, infraorbital pterygoid. Origin is entirely from the lateral and ligament, and the hyomandibular. A 23 is well ventral surfaces of the dorsal flange of the ethmoid. developed and is separated from A 2a by ramus ADDUCTOR OPERCULI (Figure 95: AD.OP. ).—This is mandibularis V. It arises from the anterior hyo- a well-developed muscle originating from the later- mandibular, the preopercle, symplectic, quadrate, al prootic and pterotic, and passing lateroventrally and posterior metapterygoid. Some of the antero- in front of the ventral process of the pterotic. It NUMBER 155 27 inserts on the dorsomedial margin of the opercle position in balistids, and is probably a homologous posterior to that bone's articulation with the structure. hyomandibular. STERNOBRANCHIALIS (Figures 94, 95: STB.).—Ori- gin is from the anteroventromedial cleithrum. The MUSCLES OF THE HYOID REGION fibers pass vertically upward, meeting briefly with their antimeres. The tendon divides into two sec- INTERMANDIBULARIS.—The muscle crosses trans- tions, the anterior of which inserts broadly on the versely between the halves of the lower jaw just posterodorsal urohyal while the posterior part in- behind the symphysis. It lies completely dorsal to serts on the ventral face of hypobranchial 3. the protractor hyoidei. RETRACTOR INTEROPERCULI (Figure 93: R.I.).— PROTRACTOR HYOIDEI (Figures 91, 92: PR.HY.).— Origin is from the medial face of the posteroventral The fibers arise from the posteroventrolateral face flange of the preopercle. The fibers converge to of the anterohyal. The dorsal fibers pass anteriorly insert on the anterior part of the interopercle- to the ventromedial face of the dentary. The more opercle ligament. Just in front of this, the inter- ventral fibers originate at an increasingly transverse opercle attaches to the posterohyal and the medial angle and meet their antimeres in the midventral preopercle. The muscle is probably a derivative of line. There is an obvious myocomma which passes the hyohyoidei adductores. anteriorly as it proceeds toward the midline in the dorsal region. VENTRAL BRANCHIAL MUSCLES HYOHYOIDEUS INFERIORIS (Figures 91, 95: HY. IN.).—Origin is from the entire ventral surface of PHARYNGOCLAVICULARIS EXTERNUS (Figures 94, 95: the anterohyal, the fibers passing ventrally to a PHC.E.).—The fibers arise from the anterior face of midventral raphe and attachment on the ventro- the cleithrum, and diverge as they pass dorsally. The lateral face of the urohyal. anterior section inserts on the extreme anterolateral HYOHYOIDEI ABDUCTORES (Figure 91: H.AB.).— tip of ceratobranchial 5, while the tendon of the Only the three ventral rays possess these muscles. posterior part joins the dorsal tendon of the cora- The more dorsal two are small, passing from the cobranchialis internus to insert on the anteroventral medial faces of the rays to the fascia on the medial face of ceratobranchial 5. face of the anterohyal. The ventralmost muscle is PHARYNGOCLAVICULARIS INTERNUS (Figures 94, 95: broad and flat, arising anteriorly in the region of PHC.I.).—The muscle originates from the antero- the urohyal and ventrohyal and inserting along the medial face of the dorsal cleithrum, and passes dorsomedial border of the first ray. anteriorly. It grades into a tendon which divides HYOHYOIDEI ADDUCTORES (Figures 91, 92: H.AD.).— into a dorsal section to ceratobranchial 5 (joining The fibers arise from the fascia overlying the dor- the posterior tendon of the pharyngoclavicularis sal surface of the sternohyoideus, and pass postero- externus) and a main anterior tendon which inserts dorsally, medial to the branchiostegal rays and the on the anteroventromedial faces of the fourth cera- preopercle, to fade out in the tissue beneath the tobranchials. opercle and subopercle. OBLIQUI VENTRALES I, HI (Figure 94: OBL.V.).— STERNOHYOIDEUS (Figures 91, 95: STH.).—The Both muscles are small, spanning the joints be- fibers arise mainly from the posterior myocomma tween the hypobranchials and ceratobranchials of separating the muscle from the obliquus inferioris, their respective arches. Obliquus II has apparently with only a few medial fibers from the anteroven- been modified into a section of rectus II. tral cleithrum. There are two myocommata, the TRANSVERSI VENTRALES III-V (Figure 94: TR.V.).— fibers inserting anterodorsally by a tendon to the The anteriormost of this series connects the ventral posteroventral face of the ventrohyal and antero- faces of the third ceratobranchials, and is single. ventrally on the posterolateral face of the urohyal. In the cases of transversi IV and V, certain changes Ventromedially, a small bundle of fibers arises from have occurred. Both these muscles consist of two the anteroventral tip of the cleithrum and inserts bundles of fibers, crossing each other in the midline on the posteroventral tip of the urohyal. It is very at some 30 degrees. The anterior muscles both con- similar to the vertical bundle found in the same nect the anteromedial faces of the fourth cerato- 28 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY branchials across the midline, the posterior muscles branchial 2 and the anterodorsal face of epibran- connecting the fifth ceratobranchials in a like man- chial 3, lying beneath transversus dorsalis II. The ner. In both pairs, the fibers passing from anterior smaller, posterior portion connects the third infra- left to posterior right lie ventral to those passing pharyngobranchial to epibranchial 3. in the opposite direction. OBLIQUUS POSTERIOR (Figure 96: OBL.P. ).—A RECTI VENTRALES I, ii, iv (Figures 94, 95: RECT. small muscle, which passes from the anterodorsal v.).—Rectus I connects the ventrolateral surface of tip of ceratobranchial 5 to the posterodorsal face basibranchial 1 (with some fibers to the postero- of epibranchial 4. dorsal face of the dorsohyal) to the anterolateral TRANSVERSI DORSALES II-III (Figure 96: TR.D.).— face of ceratobranchial 1. The main mass of rectus The anterior muscle originates from the lateral II lies between the anteroventral face of cerato- parasphenoid and passes out laterally to insert on branchial 2 and the dorsolateral face of the urohyal. the posterodorsal face of epibranchial 2. Trans- A very small lateral section attaches to the postero- versus III crosses the dorsal midline between the ventromedial face of ceratobranchial 1. The major dorsal faces of the third epibranchials, with a few section of rectus IV passes between the ventro- of the more anterior fibers arising from the para- medial face of hypobranchial 3 and the antero- sphenoid. lateral tip of ceratobranchial 4. The posteroventral RETRACTOR DORSALIS (Figure 96: D.RETR.).—Ori- section found in triacanthoids and balistids is gin is from the posterodorsolateral surface of the presumably represented by a slip of fibers diverg- basioccipital, the fibers coursing anteroventrally to ing from the posterior face of transversus ventralis insert on the dorsomedial face of the infrapharyngo- III, and attaching to the dorsal surface of the tendon branchial 3. of the rectus communis. ADDUCTORES IV-V (Figure 96: AD.).—Adductor IV RECTUS COMMUNIS (Figures 94, 95: R.COMM.).— passes across the angle between the medial faces The muscle originates from the posterodorsal uro- of the fourth epibranchial and ceratobranchial. hyal and grades into a tendon which inserts on the Adductor V connects the anterodorsal face of cera- anterolateral face of ceratobranchial 5. As described tobranchial 5 to the posterodorsal face of cerato- above, the tendon receives a small bundle of fibers branchial 4. from the posterior face of the transversus ventralis SPHINCTER OESOPHAGI (Figures 94, 95, 96: s.o.).—; III. The muscle encircles the esophagus, attaching to ceratobranchial 5, epibranchial 4, and infrapharyn- gobranchial 3. A fairly large section passes antero- DORSAL BRANCHIAL MUSCLES dorsal to the retractor dorsalis. LEVATORES EXTERNI i-rv (Figure 95: L.EXT. ).—The four muscles originate from the anteroventral face MUSCLES OF THE PECTORAL REGION of the prootic, the origins becoming more lateral posteriorly. Levator HI inserts on epibranchials 3 ABDUCTOR SUPERFICIALIS (Figure 91: ABD.S.).— and 4, the other three levators inserting on the Origin is from the ventrolateral and posteroventral epibranchials of their respective arches. The third cleithrum. The fibers insert on the anterolateral levator is partially fused to the fourth, and a small bases of each of the principal fin rays. aponeurosis connects the latter muscle to the dorsal ABDUCTOR PROFUNDUS (Figure 95: ABD.P.).—The tip of ceratobranchial 5. fibers arise from the anterolateral face of the cora- LEVATORES INTERNI II-III (Figure 96: L.INT.).— coid and insert on the bases of the principal fin Origin is from the prootic posterior to the hyoman- rays in the region of the posteroventral flange. dibular foramen. Levator II passes anteroventrally ARRECTOR VENTRALIS (Figure 91: ARR.V.).—The to insert on the medial face of infrapharyngobran- muscle originates from the lateral cleithrum be- chial 2, while levator III inserts on the dorsolateral hind the ala laminaris and above the abductor face of infrapharyngobranchial 3. superficialis. The fibers insert on the lateral base OBLIQUUS DORSALIS HI (Figure 96: OBL.D.).—There of the vestigial fin ray, of which only the medial are two sections, the anterior of which passes half remains. between the anterodorsal face of infrapharyngo- ADDUCTOR SUPERFICIALIS (Figure 97: ADD.S.).— NUMBER 155 29 Origin is from the dorsomedial face of the clei- lateral bases of the fin rays, decreasing in size thrum, the fibers inserting on the dorsomedial faces posteriorly. of the principal rays, a little distal to their bases. ERECTORES DORSALES (Figures 91, 92: EREC).—The The more dorsal fibers serve the more ventral rays, erector to the first spine is large, and originates and lie more medially. from the anterolateral face of the pterygiophore, ADDUCTOR PROFUNDUS (Figure 97: ADD.P. ).—The the frontal, and the supraoccipital. It inserts on fibers arise from the medial faces of the cleithrum the anterolateral base of the spine. The erector to and coracoid, and insert on the posteroventro- the reduced second spine arises from the frontal medial flanges of the principal fin rays. and inserts on the lateral surface of the main body ARRECTOR DORSALIS (Figure 97: ARR.D. ).—The of the spine. The erectors to the fin rays originate muscle originates from the medial face of the from the posterolateral halves of the pterygiophores cleithrum and passes posterodorsally to insert on and the anterolateral faces of the neural spines. the posteromedial flange of the vestigial finray . They insert on the anterolateral bases of the rays. CORACORADIALIS (Figure 97: COR.R.).—This small DEPRESSORES DORSALES (Figure 91: DEPR.).—The muscle arises from the anterior face of the medial first depressor passes from the frontal to the postero- flange of the coracoid. It passes laterally to insert lateral base of the spine, while that to the second on the medial face of the ventral (fourth) radial. spine arises from the frontal and lateral face of the PROTRACTOR PECTORALIS (Figure 95: P.P.).—The pterygiophore and inserts on the distal tip of the muscle connects the posteroventral tip of the ventrolateral process. The muscles to the fin rays pterotic to the anterodorsal face of the cleithrum. originate from the anterolateral halves of the LEVATOR PECTORALIS (Figure 95: TR.).—A some- pterygiophores and posterolateral faces of the what smaller muscle, arising from the ventral proc- neural spines and insert on the posterolateral bases ess of the pterotic and inserting on the extreme of the rays. anterodorsal tip of the cleithrum. SUPRACARINALIS MEDIUS (Figure 91: S.MED.).— The muscle passes from the posterior tip of the first basal pterygiophore to attach in a long, verti- MUSCLES OF THE PELVIC GIRDLE cal groove in the anterior face of the first ptery- ARRECTOR VENTRALIS PELVICUS (Figure 98: A.V. giophore of the soft dorsal fin. Some of the more P.)—The fibers arise from the midventral groove in medial fibers attach to the tips of the neural spines the pelvis, the structure becoming somewhat bi- of the second and third vertebrae. pinnate posteriorly. A tendon develops (and may SUPRACARINALIS POSTERIOR (Figure 99: S.POST.).— be split into left and right components), which The muscle connects the posterodorsal face of the inserts on the anteroventral tip of the cartilagenous last pterygiophore of the dorsal fin with the antero- plug. There did not appear to be a ventral fin ray dorsal tip of the last neural spine. The medial part element, but the specimens were not alizarin of the muscle attaches to the dorsolateral tips of stained. the neural spines of the antipenultimate and penul- ADDUCTOR SUPERFICIALIS PELVICUS (Figure 98: timate vertebrae. A.S.P. ).—The muscle consists of left and right halves, and originates in the trough on the postero- MUSCLES OF THE ANAL FIN dorsal face of the pelvis in front of the dorsal lobe. The tendons pass posteriorly through a bony canal INCLINATORES AN ALES (Figure 91: INC.).—The and insert on the fin ray element, the right tendon enlarged, residual first inclinator passes from the also attaching partly to the cartilagenous plug. fascia overlying the obliquus inferioris to attach to the anteroventral face of the first anal pterygio- phore. All the other inclinators, one of which in- MUSCLES OF THE DORSAL FIN serts on the lateral base of each ray, arise in the INCLINATORES DORSALES (Figure 91: INC.).—These fascia between the lateral pterygiophore processes muscles serve the soft dorsal fin only. They arise and the hypaxial body muscles. in the fascia between the lateral processes of the ERECTORES ANALES.—The erectors originate from distal ends of the pterygiophores and insert on the the posterior halves of the pterygiophores and in- 30 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY sert tendinously on the anterolateral base of each the former and the hypural plate. The muscle in- fin ray. serts on the bases of rays V 1-6. DEPRESSORES ANALES.—The fibers arise from the FLEXOR VENTRALIS INFERIOR (Figure 99: F.V.L).— anterolateral halves of the pterygiophores and in- The muscle originates from the ventrolateral faces sert on the posterolateral bases of the fin rays. of the centra of the third and fourth last vertebrae INFRACARINALIS ANTERIOR (Figures 91, 95: INF. and their haemal spines and inserts on the base A. ).—The muscle originates from the ventromedial of ray V 6. coracoid and posteroventral cleithrum. Insertion TRANSVERSUS CAUDALIS (Figure 996: TR.C).—Ori- is in the large ventrolateral groove in the pelvis, gin is from the dorsolateral face of the hypural the fibers almost reaching the posterior margin of plate. The fibers pass posteroventrally to insert on the bone. the dorsal base of ray V 1. INFRACARINALIS MEDIUS (Figure 91: INF.M.).—This is a thin, sheetlike muscle extending posteriorly LATERAL BODY MUSCLES from the lateral face of the pelvis. It joins the ventral fibers of the obliquus inferioris, the more EPAXIALIS (Figures 91, 95, 99a: EPAX.).—The dorsal fibers reaching the posteroventral tip of the fibers attach anteriorly to the epiotic, exoccipital, postcleithrum, while the posteroventral fibers at- and supraoccipital. Posteriorly, the angle of the tach to the ventrolateral face of the first anal myocommata increases to a sharp V, whose apex pterygiophore. lies posteriorly. Only the section closest to the mid- INFRACARINALIS POSTERIOR (Figure 99: INF.P.).— line contributes fibers to the insertion tendons Fibers arise from the posteroventral tip of the last serving rays D 1-5. The dorsal portion fades into anal pterygiophore. They attach to the tips of the a fascia covering the whole region, joining with a next two haemal spines before ending on the an- similar fascia from the obliquus inferioris. A small teroventral tip of the last haemal spine. slip of the ventral section attaches to epural 2. OBLIQUUS SUPERIORIS (Figures 91, 99a: OBL.S.).— The anterior region does not reach the skull, aris- MUSCLES OF THE CAUDAL FIN ing from the lateral surface of the vertebral column. The size of the muscle increases posteriorly, the INTERRADIALIS (Figure 99a: INT.).—The muscle is last myomere inserting on the bases of rays D 1 and made up of slips which pass from the lateral faces V 1-5, with a small ventral slip to the postero- of the rays toward the midline. They usually attach lateral face of the first hypural. to the bases of the rays lying nearer the midline, OBLIQUUS INFERIORIS (Figures 91,95, 99a: OBL.L).— but cross over each other in the case of the fibers The anteroventral portion of the muscle arises from from rays V 1 and D 1. the fascia overlying the lateral coracoid-cleithrum HYPOCHORDAL LONGITUDINALIS (Figure 99a: H.L.).— region, and is anteriorly continuous with the sterno- The fibers arise from the ventrolateral face of the hyoideus. Some fibers arise from the posterolateral hypural plate and divide into three tendons, which face of the coracoid. Most of the fibers from this insert on the ventrolateral bases of rays D 1-3. region attach to the anteroventral face of the post- FLEXOR DORSALIS (Figure 996: F.D.).—Origin is clei thrum. A number of myomeres arise from the from the dorsolateral faces of the 18th and 19th lateral midline, the fibers passing anteroventrally to centra, the neural spine of the former, the urostyle, insert aponeurotically on the posterodorsal face of and the epurals. The fibers divide into six tendons, the postcleithrum. Fiber direction gradually be- inserting on the bases of rays D 1-6. comes horizontal posteriorly, and the muscle grades FLEXOR DORSALIS SUPERIOR (Figure 99: F.D.S.). into a vague aponeurosis which joins that from the Origin is from the dorsolateral faces of the third dorsal portion of the epaxialis to cover the caudal and fourth last vertebrae and their neural spines, peduncular region. the fibers inserting on the base of ray D 6. SPINALIS.—The muscle originates from the poste- FLEXOR VENTRALIS (Figure 99b: F.V.).—The fibers rior face of the epiotic and the first dorsal ptery- arise from the ventrolateral faces of the antipenulti- giophore and inserts on the dorsolateral tips of the mate and penultimate centra, the haemal spine of neural spines of the second and third vertebrae. NUMBER 155 31

CUTANEOUS MUSCLES well developed, and visible anterior to A 23. The levator operculi originates more from the pterotic TRANSVERSUS CUTANEOUS (Figure 91: TR.CUT.).— and posttemporal than from the supracleithrum. This is a small muscle arising from the middle of Rectus ventralis II has developed extra fibers on the anteroventral face of the postcleithrum and its medial surface which join their antimeres in the passing downward to insert on the ventrolateral midline. The fibers are separated from the more face of the pelvis. anterior section by the anterior tendon of the ster- nobranchialis to hypobranchial 2 (see Figure 111 of Stephanolepis auratus (Castlenau) Aluterus for the extreme of this condition). The ventral portion of rectus ventralis IV is well de- FIGURES 100-101 veloped. In both the transversi ventrales, the rela- A 23 is slightly more bulky, and the levator tive positioning of the sections crossing over each arcus palatini is just visible in lateral view. The other is reversed; i.e., the anterior left section lies ventral portion of rectus ventralis IV is smaller, dorsal to the section arising from the anterior and the arch-shaped ligament between the ventral right. The muscle fibers of the arrector ventralis processes of the third hypobranchials is lost. The pelvicus and the adductor superficialis pelvicus are coracoradialis is less well developed, and the medial virtually absent, but fairly well-developed tendons projection for its insertion on the fourth radial is remain. The residual inclinator of the anal fin is not as prominent. The adductor superficialis pel- aponeurotic, with no muscle fibers visible. The vicus is single anteriorly but bilaterally symmetrical spinalis is poorly developed, and the transversus posteriorly. The depressor of the second spine is cutaneous does not reach the postcleithrum dor- larger, and is partially overlain by the epaxialis sally. which attaches to the lateral surface of the ptery- giophore between the two spines. The anterodorsal Paraluteres prionurus (Sleeker) portion of the obliquus inferioris above the post- cleithrum inserts musculously on that bone. The FIGURE 104 spinalis is small and inserts only on the neural spine of the second vertebra. A la is fairly small, originating primarily from the ethmoid with only a few fibers from the pre- frontal. A la' is well developed, while A h has Paramonacanthus barnardi Fraser-Brunner enlarged to take over much of the site of origin of A la. It now originates from the prefrontal and FIGURE 102 the entire anterodorsal half of the infraorbital The adductor mandibulae and levator arcus ligament. The insertion has also expanded to in- palatini are as for S. auratus, and the dilatator clude the dorsolateral face of the maxilla as well operculi is more parallel fibered. The retractor as the ethmoid and palatine. A 13 is larger and interoperculi is better developed. The adductor su- visible in superficial lateral view, while A 13' is perficialis pelvicus is single, as is the arrector ven- small. Another section, A 13", has separated from tralis pelvicus. The erector dorsalis to the first spine its dorsomedial face and inserts on the maxilla near is contained in a fairly deep depression and is par- A la. Both the A 13 sections mentioned originate tially overlain dorsally by a lateral flange of the from the anterior hyomandibular, A 13' also re- pterygiophore. The anterodorsal portion of the ceiving a few fibers from the infraorbital ligament. obliquus inferioris is as for S. auratus. A 2a is much reduced, but A 23 and A 3 are more normal. The adductor arcus palatini is tremendous- ly well developed, and inserts primarily on the en- Chaetoderma spinosissimus (Quoy and Gaimard) tire dorsomedial half of the metapterygoid. All the fibers of the hyohyoideus inferioris attach FIGURE 103 to the ventrolateral face of the urohyal, and the A la' is rather poorly developed, and A 1Y is dorsal fibers of the hyohyoidei adductores attach to visible laterally between A la and A IP'. A 3 is the dorsomedial face of the opercle. The ventro- 32 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY medial section of the sternohyoideus is poorly pelvis from the canal in which it lies. The adductor developed, and the insertion of the sternobranch- superficialis pelvicus is present as a tiny paired ialis is confined to the ventral faces of hypobran- muscle on the dorsal surface of the pelvis. It passes chials 2 and 3. through the canal in the dorsal lobe, and both ten- The small lateral section of rectus ventralis II dons insert on the posterior face of the cartilagenous attaches more to hypobranchial 1 than to cerato- plug. Some of the more ventral fibers of the infra- branchial 1, and the medial section is as for C. carinalis medius attach to the thickened dermis spinosissimus. The ventral portion of rectus ven- around the anus. tralis IV is absent, and transversus ventralis IV The anterodorsal portion of the obliquus infer- consists only of a single transverse fiber bundle. ioris is made up mainly of an aponeurotic sheet, the The arrector ventralis pelvicus and adductor super- muscle fibers being confined to a narrow band in ficialis pelvicus are absent, as they are in all the the middle of the sheet. remaining species except Cantherhines pardalis. The infracarinalis anterior is small, and the in- Oxymonacanthus longirostris fracarinalis posterior and supracarinalis posterior (Bloch and Schneider) do not attach to the tips of the haemal and neural FIGURE 106 spines respectively, apart from the spine on which they insert. The last three vertebrae possess well- A la is rather small, passing from the ventral developed lateral flanges, and the origins of the face of the maxilla to its origin at the anterior flexor dorsalis and flexor ventralis have shifted par- junction of the frontal and prefrontal. Above A la', tially onto these surfaces. The spinalis is well a third subdivision, A la", is recognizable. It origi- developed, and originates mainly from the ex- nates along the ventrolateral surface of the ethmoid occipital. ridge and passes into the connective tissue of the lips dorsal to A la'. A IP' has acquired some origin from the prefrontal, and partially overlies A lv Cantherhines pardalis (Riippell) ventrally. A 1Y itself is well developed with a broad insertion. A 2a is also well developed, its FIGURE 105 origin stretching from the prefrontal to the hyo- A la' is poorly developed, while A 2? has ex- mandibular and including the infraorbital liga- panded its origin dorsally to reach the prefrontal. ment. At about the middle of its origin is a gap A IP' inserts by a broad, aponeurotic sheet over in the fibers, through which passes a sensory branch the posterolateral face of the maxilla ventral to the of nerve V. There are no fibers of A 2P from the palatine. A IP is well developed, while A 1Y has ventrolateral flange of the preopercle (and this is a somewhat restricted insertion on the ethmoid true for all the remaining species). A 3 is poorly and palatine. A 3 is partially fused to A 2P antero- developed. The ridge of bone on the posterodorsal ventrally. There are no fibers of A 2P from the surface of the hyomandibular is well developed, ventrolateral flange of the preopercle. the fibers of the levator arcus palatini inserting All six branchiostegal rays may receive a slip on either side of it. The lateral wall of the fossa of the hyohyoidei adductores, but that to the dorsal- containing the origin of this muscle has broken most ray may be absent. The halves of the sterno- down, and the muscle is visible laterally between branchialis do not meet in the midline. The the frontal and sphenotic. A few fibers of the pharyngoclavicularis externus is bifid, rectus ven- levator operculi arise from the pterotic. The ad- tralis II is as for C. spinosissimus, and the ventral ductor arcus palatini is visible in lateral view, portion of rectus ventralis IV is well developed. and stretches to the rear of the orbit. The retractor The abductor profundus is considerably exposed arcus palatini is poorly developed, with no fibers to ventrally beneath the abductor superficialis. Both the palatine. the adductor profundus and the protractor pecto- There is no ventromedial section to the sterno- ralis are well developed. The arrector ventralis hyoideus, and the aponeuroses of the sternobran- pelvicus appears to be represented by a tendon chialis are more consolidated. The retractor which just reaches the posteroventral surface of the interoperculi is absent. NUMBER 155 S3

The pharyngoclavicularis externus is single, and dorsal part of the insertion extending into the inserts on ceratobranchial 5, while the pharyngo- tissues of the lips. The origin has expanded enor- clavicularis internus has been reduced to a strap- mously to include not only the anteroventral pre- like, parallel fibered muscle. Obliquus ventralis I is frontal, but the entire infraorbital ligament as well. well developed, being almost the same size as rectus A 13' is long and thin (Figure 109), while A 13 ventralis I. The medial portion of rectus ventralis and A 1Y are more like the generalized condition. II is as described for C. spinosissimus, and the A 2a (Figure 108) is much reduced dorsally, there ventral portion of rectus ventralis IV is absent (as being no fibers from the infraorbital ligament. A in P. prionurus). 3 is small. The adductor arcus palatini (Figure 110: The depressor dorsales of the first spine has A.A.P.) is divided into two sections by the junction broken through beneath the lateral strut of the between the metapterygoid and the parasphenoid. pterygiophore supporting the base of the spine, and The retractor arcus palatini is not subdivided at has extended its origin into the fossa originally all. occupied only by the erector dorsalis to the first Most of the posteroventral fibers of the protractor spine. Erector dorsalis II passes posteromedially to hyoidei arise from the fascia over the anteroventral its insertion, while depressor dorsalis II courses tip of the cleithrum. These fibers are almost at right posterolaterally to the base of the second spine. angles to the ones originating from the postero- Both penetrate the tunnel beneath the flange, but ventral face of the anterohyal. do not extend beyond its anterior margin. The Rectus ventralis II (Figure 111: RECT.V.) is well hypochordal longitudinalis inserts only on rays D 2 developed. There is a lateral section to hypobran- and 3, and the transversus caudalis is poorly de- chial 1, an anterior section to the dorsolateral uro- veloped. The transversus cutaneous fans out from hyal, and a transverse section across the midline. the posterodorsal face of the pelvis, and attaches to The anterior tendon of the sternobranchialis in- the coracoid anteriorly. serts on hypobranchial 2 between the two latter sections. The arrangements of the two bundles of Pseudalutarius nasicornis (Schlegel) fibers forming transversus ventralis IV are as for P. melanocephalus, but the relative positions of the FIGURE 107 bundles is reversed in transversus V. The pharyngo- The adductor mandibulae complex is very much clavicularis externus is bifid, with tendons to like that of O. longirostris, although the snout is ceratobranchials 4 and 5. not as elongated. A la' inserts more on the maxilla The anterior extension of the prootic shelf has than the lips, and A IP' completely overlies A 1Y. carried the origins of the levatores externi with it Where O. longirostris differs from P. melanoceph- (Figure 112: L.EXT.). There is some variation in the alus, the descriptions also apply to P. nasicornis, mode of insertion in levators II and IV. They may with the following exceptions. insert by a continuous aponeurosis on epibranchials There is a larger section of the levator operculi 3 and 4, or separately. If the latter is the case, originating from the pterotic, and the retractor the anterior tendon may send a small branch to interoperculi is present. Obliquus ventralis II is epibranchial 4. The aponeurosis of levator II sends apparently absent. a small section to epibranchial 3. The levatores The general positioning of the dorsal branchial interni retain their origin from the prootic beneath levators seems more primitive than in P. melano- the hyomandibular foramen. Some fibers of trans- cephalus, since all the origins are from the postero- versus dorsalis III overlap onto the dorsal face of lateral face of the prootic, and none from the poorly epibranchial 4. Obliquus dorsalis III is somewhat developed prootic shelf. The small size of the shelf modified. Fibers from infrapharyngobranchial 2 could, of course, be secondary. pass to three different points of attachment. The medial fibers pass posteriorly to the dorsal face Aluterus monoceros (Linnaeus) of infrapharyngobranchial 3. The more lateral fibers attach to epibranchial 3 (as in P. melano- FIGURES 108-113 cephalus), while some of the ventromedial fibers A la (Figure 108) is a single section, the more loop around epibranchial 2 and reattach to infra- 34 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY pharyngobranchial 2. of the latter section soon become aponeurotic, and The adductor superficial is well developed (Fig- insert on the ethmoid and parasphenoid posterior ure 113: ADD.S.). The supracarinalis medius is to the retractor arcus palatini. A 2a and A 23 are present as a pair of cylindrical muscles, rather than not separated posteriorly, but are so anteriorly, the more usual triangular muscle. The erectores where the ramus mandibularis V passes between and depressores dorsales do not reach the vertebral them. The orbital process of the sphenotic almost column ventrally. The infracarinalis medius forms reaches the hyomandibular. The levator operculi a broad tendon posterodorsally before grading into and levator arcus palatini are very small, but the the fascia beneath the postcleithrum. Both the dilatator operculi remains well developed. The ad- transversus caudalis and the spinalis are small, the ductor arcus palatini is confined to a region in latter muscle arising from the posterodorsal face of front of the orbit, but behind the retractor arcus the first vertebra. In view of the fact that A. monoc- palatini. The latter muscle is very well developed. eros exhibits certain characters not found in any Both the hyohyoidei abductores and the hyohy- of the other monacanthids examined (particularly oidei adductores are well developed, the latter mus- in the form of the adductor mandibulae complex), cle attaching to three of the four branchiostegal two additional species of the genus were dissected. rays. The hyohyoideus inferioris meets its antimere in the ventral midline, and is not attached to the urohyal. Aluterus heudeloti Hollard Rectus ventralis II is single, with the fibers pass- ing anteromedially to the midline, where they join The only difference noted was that there were to form two tendons to the urohyal. Obliquus ven- tough connective tissue strands from the postero- tralis III is absent, as is the ventral section of rectus dorsolateral face of A 2a to the prefrontal. IV. The dorsal ends of the gill arches lie beneath the Aluterus schoepfi (Walbaum) prootic shelf at the level of the prefrontal. Leva- tores externus IV inserts only on epibranchial 4. The findings were the same as for A. heudeloti. The levatores interni arise from the prootic just In the smaller specimen (57 mm SL), A 2a has before the start of the shelf at the rear of the orbit. relatively larger dorsal extent than in any other They course the length of the orbit before insert- Aluterus examined, the posterodorsal fibers almost ing on the infrapharyngobranchials. reaching the infraorbital ligament. The levator pectoralis is very well developed, originating from the elongated ventral process of the pterotic and sphenotic, the medial fibers aris- Anacanthus barbatus Gray ing from the posteromedial face of the hyomandib- ular. It inserts on the supracleithrum and FIGURE 114 anterodorsal cleithrum. The protractor pectoralis is Various modifications are apparent in the adduc- less well developed, but also more anterior in tor mandibulae complex. A la' originates from the position. lateral and medial ethmoid. It is partially covered There is a single dorsal spine, the depressor dor- by A la posteriorly, and has a long, thin aponeu- salis being larger than the erector dorsalis. The rosis which crosses the upper jaw to insert in the supracarinalis medius is musculous at its attach- anterior tissues of the barbel. A la arises from the ment sites, but aponeurotic in the middle two- ethmoid behind A la' and from the prefrontal. It thirds, where it attaches to the tips of the neural also develops a long aponeurosis, which fades into spines. The infracarinalis anterior has partially the tissues of the barbel at its posterior base. These fused in the midline. The residual inclinator of sections are only connected to the upper jaw by the anal fin is very well developed, grading into connective tissue. A 13 is fairly well developed, and the infracarinalis medius anteriorly and the obli- has its normal relationships. A IP' is absent, and quus inferioris dorsally. The interradialis is well A h is poorly developed, originating from the developed, but both the hypochordal longitudinalis prefrontal and the infraorbital ligament. The fibers and the transversus caudalis are absent. NUMBER 155 35

Summary of Monacanthidae ber of the sections of the branchiostegal ray musculature. The amount of separation and The monacanthids exhibit a wide range of var- consolidation of the ventromedial section of the iation and progressive specialization, which may sternohyoideus varies, as does the insertion of the or may not be obviously related to the body form. sternobranchialis. The retractor interoperculi is The adductor mandibulae is very variable in both sometimes absent. the number of subdivisions and in the relative de- The pharyngoclavicularis externus is often bifid, velopment of these sections. The basic plan is, and rectus II shows progressive specialization and however, fairly consistent, with two subdivisions subdivision. The ventral fiber bundle of rectus ven- of A la and A IP, and a single section of un- tralis IV may be absent, and the relative position- known function, A 1Y. Section A 2 usually has two ing and development of the sections of transversus subdivisions, separated by the path of ramus man- ventralis IV differs. dibularis V, the posterodorsal surface of A 2a some- The position of origin of the levatores externi times reaching the prefrontal. The levator arcus depends on the location of the prootic shelf, while palatini may be well developed or almost absent, that of the levatores interni seems independent of or it may be visible laterally owing to a breakdown this variable. The muscles serving the fin ray ele- in the lateral wall of the dilatator fossa. The ad- ments of the pelvis are variously developed or ab- ductor arcus palatini may be well or poorly devel- sent, the ray elements themselves being absent in oped, confined to the region in front of the orbit, certain genera. Muscles attaching to the dorsal stretch half way across it, or it may reach the rear spines vary in sites of origin and relative develop- of the orbit. The retractor arcus palatini is well ment. The anterodorsal portion of the obliquus separated from the above muscle, and may be well inferioris may be musculous or largely aponeurotic, or poorly developed. and the sites of origin of the spinalis vary. The protractor hyoidei may have a superficial At least some of the variation in monacanthids section arising from the fascia overlying the ventral is attributable to the large number of species and tips of the cleithra, and the hyohyoideus inferioris genera and the great variety of body form. They may arise partially, totally, or not at all from the would appear to have undergone "explosive" evo- ventrolateral face of the urohyal. There are minor lution, and form the apomorph sister group of the variations in the extent of development and num- balistids.

Myological Descriptions of Representative Aracanids

The general body form of these fishes is illus- The aracanids are usually considered, in effect, trated in Figure 5. In his review of the genera of as the plesiomorph sister group of the next family, ostracioid fishes, Fraser-Brunner (1935) recognized the Ostraciidae, and derived originally from a com- six genera in what was originally a subfamily, but mon ancestor with the balistids (although it has which he later raised to familial rank (1941b). also been suggested that they arose from a neotenic Representatives of two of these six genera have molid ancestor). been dissected for the present study. The fishes are encased in a bony cuirass of modified scales Kentrocapros aculeatus (Houttuyn) from the snout to the level of the posterior margin of the dorsal and anal fins. Neither this family, FIGURES 115, 116a, 117-124 nor the succeeding one (Ostraciidae) show any evi- dence of pelvic fins or girdle. They are presumably MUSCLES OF THE CHEEK slow swimmers, but virtually nothing is known about ADDUCTOR MANDIBULAE (Figures 115, 116, 117: their life habits. They appear to live in deepish A 13, A 2a, A 23, A 3).—A 2a is well devel- water (down to 100 fathoms). The gut contents oped, originating from the dorsolateral parasphe- of the specimens dissected for this study indicate noid, the interior face of the prootic shelf, and the a diet of small crustacea and worms. anterodorsal margin of the hyomandibular. Fibers 36 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY also arise from the tendinous sheath on the dorso- palatini. The fibers arise from the anterodorsal medial face of the muscle. The more medial fibers parasphenoid and pass anteroventrally to insert on in this region are not completely distinct from the the posterodorsal face of the metapterygoid. lateral fibers of A 13" (Figure 117) posteriorly. ADDUCTOR OPERCULI (Figures 119, 121: AD.OP. ).— A 23 arises from the preopercle and hyomandib- Fibers arise from the lateral face of the basioccipital ular, the fibers passing anterodorsally. They tend and anteroventral exoccipital, and insert on the to pass more medially than do those of A 2a, but dorsomedial face of the opercle by a long tendon. join to form a common tendon before inserting ADDUCTOR HYOMANDIBULAE (Figures 119, 121: on the inner face of the dentary. Ramus mandibu- A.H.).—The muscle lies just above the adductor laris V emerges from beneath A IP" and passes operculi, and passes more anteriorly to insert on lateral to the other sections of A IP. It appears the dorsomedial face of the hyomandibular. Its superficially from beneath the anterior fibers of A origin is poorly separated from that of the adduc- 23. The sections of A 1 are not well separated tor operculi, the muscle being derived from the either from each other or from the overlying A 2. more posterodorsal fibers. A 13" is the dorsalmost section and originates from the anterior face of the prootic shelf. Beneath it, A IP' (Figure 117) arises from the dorsolateral MUSCLES OF THE HYOID REGION metapterygoid and anterior hyomandibular. It is INTERMANDIBULARIS.—This is a small transverse poorly separated from A IP, which arises from the muscle spanning the symphysis between the den- anterolateral face of the hyomandibular. The lat- taries, and being closely associated with the antero- ter section partly overlies and intermingles with dorsal fibers of the protractor hyoidei. A 3. Anteriorly, these four sections come together PROTRACTOR HYOIDEI (Figures 115, 117, 118: to form a tendinous aponeurosis which divides into PR.HY.).—The muscle originates from the medial a strong dorsal component to the medial face of face of the dentary near the symphysis and inserts the maxilla and a weaker ventral one to the medial on the lateral faces of the anterohyal, posterohyal, face of the dentary. and the bases of the dorsal four branchiostegal rays. LEVATOR ARCUS PALATINI (Figures 115: L.A.P.).— The two halves remain separate for most of their This is a short, conical muscle originating from length, being joined by a midventral raphe just the ventral face of the orbital process of the sphe- below the jaws. A myocomma arises in the region notic and fanning out to insert on the anterolateral of the raphe, and passes posterodorsally. Some of face of the hyomandibular. the more dorsal fibers attach to the interopercle. DILATATOR OPERCULI (Figure 115: D.O.).—The HYOHYOIDEUS INFERIORIS (Figures 115, 118: fibers form a long, thin, bipinnate muscle which HY.IN.).—The muscle is rather small, and originates arises from the dilatator fossa (mainly in the sphe- from the posteroventral anterohyal and anteroven- notic) and inserts on the dorsal process of the tral posterohyal beneath the second branchiostegal opercle. ray. The fibers pass down to the midline, where LEVATOR OPERCULI (Figure 115: L.O.).—The mus- they meet their antimeres in a midventral raphe. cle originates from the ventrolateral posttemporal HYOHYOIDEI ABDUCTORES (Figures 115, 118: and inserts on the posterodorsal face of the opercle. H.AB.).—These muscles insert on the medial faces ADDUCTOR ARCUS PALATINI.—The muscle is short of the first four branchiostegal rays. The section to and thick. It originates from the ventrolateral face the first ray arises from the ventromedial face of of the parasphenoid and inserts mainly on the the ventrohyal, that to the second ray originates dorsomedial face of the metapterygoid. The an- from the base of the first ray, while the slips to terior fibers pass outward to their insertion at an the other rays arise from the base of the second ray increasingly anterolateral angle, and grade imper- and the surrounding anterohyal. ceptibly into the retractor arcus palatini. The mus- HYOHYOIDEI ADDUCTORES (Figures 115, 117, 118: cle lies wholly in front of the orbit. H.AD.).—The muscle lies mainly in the lateral tis- RETRACTOR ARCUS PALATINI (Figure 117: R.A.P.). sues of the opercular cavity. It attaches dorsally to The muscle is fairly well developed, and, as stated, the sphenotic and prootic shelf. Ventrally, fibers not completely separate from the adductor arcus attach to the fascia covering the sternohyoideus NUMBER 155 37 and the ventral surfaces of the branchial arches. ventral surfaces of the ceratobranchial-hypobran- The more dorsal branchiostegal rays are partially chial regions of the first and second arches respec- embedded in the fibers. A small slip of fibers arises tively. Rectus I attaches to the posterodorsal face from the anteroventral region of the cleithrum and of the dorsohyal, while rectus II passes antero- courses dorsally to join the main mass in the vicin- medially, fusing partly with its antimere and partly ity of the opercular valve. with the ligament from the ventral process of hypo- STERNOHYOIDEUS (Figures 115, 118, 119: STH.).— branchial 3. Rectus IV is well developed, and con- The muscle consists of two sections. The main nects the ventral face of ceratobranchial 4 to the lateral portion arises from the myocomma with arch-shaped ligament between the third hypobran- the obliquus inferioris lateral to the abductors of chials. Some fibers intermingle with those of the the pectoral fin. It inserts on the posteroventral face other side in the midline. of the ventrohyal. A small, medial section of the RECTUS COMMUNIS (Figures 119, 120: R.COMM.).— muscle originates from the anteroventral tip of the The muscle arises from the posterior face of the cleithrum and inserts on the lateral face of the urohyal and inserts on the ventrolateral face of small urohyal. It lies medial to the main body of ceratobranchial 5. the muscle. STERNOBRANCHIALIS (Figures 118, 119, 120: STB.).— DORSAL BRANCHIAL MUSCLES The fibers originate from the anteroventral region of the cleithrum and pass dorsally to sepa- LEVATORES EXTERNI I-IV (Figure 119: L.EXT.).— rate into two tendons. The anterior of these inserts The series is well developed. All the muscles orig- on hypobranchials 2 and 3 and ceratobranchial 3, inate from the ventral surface of the prootic shelf, while the posterior tendon inserts on ceratobran- and insert on the dorsal faces of the respective chials 3 and 4. epibranchials. Levators III and IV are completely fused dorsally, the bulk of the fibers affecting the tendon serving the fourth epibranchial. The tendon VENTRAL BRANCHIAL MUSCLES to epibranchial 3 arises from the anteromedial sur- PHARYNGOCLAVICULARIS EXTERNUS (Figures 119, face of this bundle. 120: PHC.E.).—This muscle is very well developed. LEVATORES INTERNI II, HI (Figures 115, 119, 121: It originates from the anterior and ventromedial L.INT. ).—These are both enormously developed, surfaces of the cleithrum, dorsolateral to the origin bipinnate muscles, which insert on infrapharyngo- of the medial section of the sternohyoideus. The branchials 2-4. Levator II originates from the fibers pass upward to insert on the ventral surface posterolateral prootic, ventral pterotic, lateral exoc- of ceratobranchial 5. cipital, and ventrolateral epiotic, and inserts on the PHARYNGOCLAVICULARIS INTERNUS (Figures 119, posterior face of infrapharyngobranchial 2. Levator 120: PHC.I.).—The fibers arise from the anterodor- III overlies the posteroventral surface of the an- somedial face of the cleithrum and insert on the terior muscle, and includes the pterotic, epiotic, ventral face of ceratobranchial 5. and exoccipital in its sites of origin. Anteriorly, the OBLIQUUS VENTRALIS HI (Figures 119, 120: OBL. muscle divides into two tendons, the anterior of v.).—The muscle is small, and connects the ventral which inserts on infrapharyngobranchial 3, the pos- face of ceratobranchial 3 to the tip of the ventral terior one attaching to the medial tip of epibran- process of hypobranchial 3. The process is embedded chial 4 and the tough connective tissue which in a flat ligament which passes anteriorly to attach appears to contain the cartilagenous remains of in- to the posterior face of the urohyal. frapharyngobranchial 4. The posterolateral origin TRANSVERSI VENTRALES IV-V (Figure 120: TR.V.).— of this muscle is visible in superficial lateral view. These two muscles connect the anteroventromedial OBLIQUUS DORSALIS HI (Figures 119, 121: OBL.D.).— faces of the ceratobranchials of the fourth and fifth The muscle connects the dorsal face of epibran- arches across the ventral midline. The posterior chial 3 to the dorsal face of infrapharyngobranchial muscle is well developed. 3, passing between the two tendons of levator in- RECTI VENTRALES I, ii, rv (Figures 119, 120: ternus III. RECT.V.).—The two anterior recti arise from the OBLIQUUS POSTERIOR (Figure 121: OBL.P.).—This 38 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY is a well-developed muscle, passing from the dorso- cleithrum. The muscle inserts on the anteromedial medial face of ceratobranchial 5 to the posterior faces of the principal rays distal to their bases. The face of epibranchial 4. more ventral fibers pass more dorsolaterally to serve TRANSVERSI DORSALES n, iv (Figure 121: TR.D.).— the more dorsal rays. Both muscles originate from middorsal raphes with ADDUCTOR PROFUNDUS (Figures 115, 118, 123: their antimeres. Transversus II passes out laterally ADD.P. ).—A well-developed muscle originating from behind infrapharyngobranchial 2 to insert on the the medial coracoid and posteromedial cleithrum. dorsal face of epibranchial 2. The posterior muscle It inserts on the posteroventromedial flanges of the inserts on the posterodorsal face of epibranchial 4. principal fin rays. RETRACTOR DORSALIS (Figures 119, 121: D.RETR.).— ARRECTOR DORSALIS (Figure 123: ARR.D. ).—The Origin is from the lateral face of the basioccipi- muscle originates from the medial face of the clei- tal, the muscle passing anteriorly beneath Baude- thrum, lying lateral to the adductor profundus and lot's ligament and the medial process of the medial to the adductor superficialis. It inserts on supracleithrum to insert in the connective tissue the ventromedial base of the vestigial finray . surrounding the cartilagenous remains of infra- CORACORADIALIS (Figure 123: COR.R.).—A small pharyngobranchial 4 near the medial tip of epi- muscle, consisting of only a few fibers from the branchial 4. posterodorsal face of the coracoid to the postero- ADDUCTOR IV.—This is a poorly developed mus- ventral face of the fourth radial. cle, crossing the angle between the medial faces of the fourth epibranchial and ceratobranchial. It is embedded in thick connective tissue. MUSCLES OF THE DORSAL FIN SPHINCTER OESOPHAGI (Figures 119, 120, 121: s.o.).—The muscle surrounds the esophagus, at- ERECTORES DORSALES (Figure 122: EREC).—These taching to the posterior faces of ceratobranchial 5 are greatly expanded muscles, whose origin covers and epibranchial 4. A pair of ligaments arise from the centra and neural spines of the fourth to the ventral surface and pass posterolaterally to twelfth vertebrae as well as the pterygiophores of attach to the medial faces of the cleithra. the dorsal fin. Each erector is partially overlain posterolaterally by the depressor muscle of the same fin ray. The muscles decrease in size posteriorly, MUSCLES OF THE PECTORAL REGION and insert tendinously on the anterolateral bases of the fin rays. ABDUCTOR SUPERFICIALIS (Figures 115, 118: ABD. s.).—Origin is from the lateral and posterior faces of DEPRESSORES DORSALES (Figure 122: DEPR.).— the cleithrum and the fascia overlying the abductor These muscles share their sites of origin with the profundus. The fibers insert on the anterolateral erectors, but are somewhat more laterally situated. bases of the principal fin rays. They insert tendinously on the posterolateral bases ABDUCTOR PROFUNDUS (Figures 115, 118, 119: of the fin rays, and decrease in size posteriorly. ABD.P.).—A well-developed muscle, arising mainly SUPRACARINALIS ANTERIOR (Figures 115, 122: from the broad, lateral surface of the coracoid. The S.ANT.).—Origin is from the posterodorsal face of tendons insert on the posteroventral bases of the the supraoccipital and the lateral face of its spine. principal fin rays. There is a section to the lateral The fibers pass posteriorly to insert on the tips of base of the vestigial ray that arises from the medial the neural spines of the third and fourth vertebrae surface of the muscle mass serving the dorsalmost and the anterior face of the "supraneural" (= basal ray. pterygiophore of the now absent spiny dorsal fin). ARRECTOR VENTRALIS (Figure 115: ARR.V.).—The SUPRACARINALIS POSTERIOR (Figures 122, 1246: muscle originates from the anterolateral face of the S.POST.).—The muscle arises from the posterodorsal cleithrum and inserts on the anterolateral margin face of the neural spine of the thirteenth vertebra of the medial half of the vestigial finray . and courses posteriorly, attaching to the neural ADDUCTOR SUPERFICIAL^ (Figure 123: ADD.S.).— spines of the succeeding vertebrae, to end on the Origin is from the posterodorsomedial face of the neural spine of the penultimate vertebra. NUMBER 155 39

MUSCLES OF THE ANAL FIN TRANSVERSUS CAUDALIS (Figure 1246: TR.C).—The ERECTORES AN ALES (Figure 122: EREC).—The first muscle is well developed, and originates from the three muscles arise from the anterior face of the entire dorsolateral surface of the hypural plate, laterally displaced pterygiophores, the other mus- reaching as far anteriorly as the ural centrum. It cles originating from the posterior and lateral faces inserts by three distinct tendons (the ventral one of the pterygiophores. Insertion is tendinous on of which serves two rays) on the dorsolateral bases the anterolateral bases of the finrays , the muscles of rays V 2-5. decreasing in size posteriorly. DEPRESSORES ANALES (Figure 122: DEPR.).—The LATERAL BODY MUSCLES muscles lie immediately behind the erectors of the EPAXIALIS (Figures 115, 122, 124a: EPAX.).—The respective finrays , and have the same origins. They two portions of the muscle are well separated an- insert on the posterolateral bases of the rays, de- teriorly, but grade into each other posteriorly. The creasing in size caudally. dorsal moiety attaches anteriorly to the posterior INFRACARINALIS POSTERIOR (Figures 122, 124&: faces of the supraoccipital and exoccipital. The INF.P. ).—The muscle arises from the posteroventral ventral section is larger, and arises anteriorly from tip of the last anal pterygiophore. It passes poster- the lateral faces of the anterior vertebrae. It in- iorly to attach to the anteroventral margin of the creases in size posteriorly, where it overlaps both haemal spine of the penultimate vertebra. centra and parts of the haemal arches. The muscle inserts on the lateral bases of the dorsal caudal MUSCLES OF THE CAUDAL FIN fin rays, and is poorly separated from the flexor INTERRADIALIS (Figure 124a: INT.).—The muscle dorsalis medially. courses dorsoventrally across the finrays , a slip of OBLIQUUS SUPERIORIS (Figures 122, 124a: OBL.S.).— the more posterior fibers attaching to the ventro- The muscle is poorly developed, particularly an- lateral (rays D 1—5) or dorsolateral (rays V 1-6) teriorly, where it arises from the ventrolateral face faces of the rays. of the epaxialis. Posteriorly, the lateral septum HYPOCHORDAL LONGITUDINALIS (Figure 124a: fades out, and it is not clear how much the obliquus H.L.).—The muscle originates from the lateral face superioris contributes to the musculature inserting of the ventral part of the hypural plate. It inserts on the ventral caudal fin rays. beneath the interradialis on the ventrolateral bases OBLIQUUS INFERIORIS (Figures 115, 118, 122: of rays D 3—5. OBL.L).—The anteroventral section of the muscle FLEXOR DORSALIS (Figure 1246: F.D.).—The fibers passes from the posterolateral face of the coracoid arise from the dorsolateral faces of the centra and to the anteroventral face of the postcleithrum. It neural spines of the last three vertebrae, and insert is partially overlain by the sternohyoideus antero- on the bases of rays D 1-4. It is poorly separated laterally. The anterodorsal section is much reduced. from the epaxialis laterally and the flexor dorsalis It originates from the ventrolateral ridge of the superior dorsally. epiotic (where it shows faint traces of segmenta- FLEXOR DORSALIS SUPERIOR (Figure 124: F.D.S.).— tion) and inserts on the dorsal face of the post- Origin is from the dorsolateral surface of the pe- cleithra. Posteriorly it grades into an aponeurotic nultimate vertebra, the muscle inserting on the fascia overlying the greatly expanded lobes of the anterolateral base of ray D 5. liver, which cover the obliquus superioris and FLEXOR VENTRALIS (Figure 1246: F.V.).—The mus- epaxialis in this region. Posteroventrally, the mus- cle originates from the centrum and haemal spine cle is small, and soon grades into an aponeurotic of the penultimate vertebra and the anteroventral sheet which overlies and intermingles with the border of the hypural plate. It inserts on the an- obliquus superioris. terolateral bases of rays V 1-5. FLEXOR VENTRALIS INFERIOR (Figure 124: F.V.I.).— Capropygia unistriata Kaup The fibers arise from the lateral tip of the haemal spine of the penultimate vertebra and insert on the FIGURE 125 anterolateral base of ray V 6. A 2 is poorly subdivided, with no clear-cut sep- 40 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY aration of parts. It is, however, bipinnate, and A single, small inclinator dorsalis passes to the could easily have arisen or given rise to the condi- ventrolateral face of the first pterygiophore. tion in K. aculeatus. There is no intermingling of A small inclinator analis arises from the fascia fibers with the underlying A 13. A 13" is better over the obliquus inferioris behind the postero- separated from the more ventral portions of A 1, ventral surface of the postcleithrum and attaches and originates exclusively from the prootic shelf. to the ventrolateral face of the first anal pterygio- The ventral fibers do not intermingle with those phore. Behind this, a thin tendon attaches to the of A 3, although they do share the archlike tendon anteroventral base of the first fin ray, and passes between the maxilla and dentary. The levator and posterodorsally. The fibers sweep dorsally in a large dilatator operculi are smaller. The retractor arcus semicircle to fade into the fascia beneath the pos- palatini is continuous with the adductor arcus terior half of the dorsal fin, overlying the epaxialis. palatini anteriorly and ventrally, and merely con- The muscle appears to represent an extreme modi- sists of a more anterior-posterior orientation of fication of the inclinator of the first anal fin ray. these muscle fibers. There is no adductor hyoman- The bulk of the body musculature consists of the dibulae. more ventral moiety of the epaxialis. This is well The sternobranchialis has completely separated separated from the more dorsal portion, and orig- into two sections, and its origin just overlaps the ventrolateral face of the cleithrum. Levator ex- inates anteriorly from the vertebral column and ternus III is not fused to levator IV, but is closely the fascia over the erectores and depressores dor- applied to it. Adductor IV is well developed. The sales. Posteriorly it forms a tendinous sheath at- adductor superficialis has some fibers originating taching to the bases of all the caudal fin rays from the ventral face of the medial process of the (although part of the fibers in the ventral region cleithrum. The adductor profundus is better de- may be derived from the obliquus superioris. At- veloped, and is partly visible posteroventrally be- tachment to the two outermost fin rays is mainly hind the coracoid. The supracarinales anterior are derived from the dorsal and ventral portions of fused in the midline, although a septum is present. the body musculature, and is tendinous.

Myological Descriptions of Representative Ostraciids

The normal body outline of members of this Ostracion tuberculatus Linnaeus family is illustrated in Figure 6. Fraser-Brunner (1935, 1941b) recognized seven genera of ostraciids, FICURES 1166, 126-134 which he divided (1941b) among two subfamilies. MUSCLES OF THE CHEEK Representatives of four of the seven genera were dissected. These fishes are rather similar to the ADDUCTOR MANDIBULAE (Figures 1166, 126, 127: aracanids, except that the carapace extends behind A 13, A 2a, A 23, A 3).—The section A 2 is the base of the anal fin, and, in all but one species, well developed, and superficial in position. It con- the dorsal fin as well. They are slow swimmers, liv- sists of a smallish dorsal section, A 2 a, which ing on or near coral reefs and sea-grass meadows. originates mainly from the region where the eth- Their food is comprised mainly of soft-bodied ma- moid, parasphenoid, prefrontal, and prootic meet rine invertebrates, although crabs and other harder each other. The main mass is formed by A 2P, bodied organisms are also eaten. The dorsal and which arises from the anterior margin of the anal fins are used for locomotion, being supple- prootic shelf, the hyomandibular, and the pre- mented by the caudal fin in emergencies. Steering opercle. The two subdivisions share a tendinous in- is achieved by the pectoral and caudal fins. sertion in the Meckelian fossa on the medial face As stated above, the ostraciids are considered of the dentary. A 1 consists of three poorly sepa- the more specialized (apomorph) sister group of rated subdivisions or "heads." A 13" (Figure the aracanids. 127) is best developed and originates from the NUMBER 155 41 anterior and ventral faces of the prootic shelf and that region. It passes anterolaterally to insert on the lateral parasphenoid. A IP' (Figure 127) the anterodorsomedial face of the hyomandibular. arises primarily from the metapterygoid, with a few fibers from the mesopterygoid. A 13 originates MUSCLES OF THE HYOID REGION from the hyomandibular, ventrolateral metaptery- goid, and symplectic. These subdivisions join to INTERMANDIBULARIS.—This is a small, short mus- insert on the medial face of the maxilla. The ten- cle which spans the midline just behind the sym- don curves ventrally in the form of an arch to join physis between the dentaries. It is closely associated with that of A 3. A 3 itself is fairly well developed, with the anterodorsal fibers of the protractor hy- originating from the quadrate and symplectic and oidei. inserting on the medial face of the dentary (ven- PROTRACTOR HYOIDEI (Figure 126: PR.HY.).—The tral to the insertion of A 2). muscle arises from the posteroventral corner of the LEVATOR ARCUS PALATINI (Figure 126: L.A.P.).— anterohyal and the bases of the dorsalmost four The muscle originates from the ventral face of branchiostegal rays. The more ventral fibers join the sphenotic and inserts on the posterodorsal face their antimeres in the midline, while the more of the hyomandibular. dorsal fibers sweep upward to insert on the pos- DILATATOR OPERCULI (Figure 126: D.O.).—Origin terior margin of the lower jaw. A myocomma is is from the dilatator fossa behind the orbit. The present in the anterior region. fossa is bounded variously by the sphenotic, ptero- HYOHYOIDEUS INFERIORIS (Figure 126: HY.IN.).— tic, prootic, and frontal. Insertion is along the This is a small muscle passing across the ventral dorsal rim of the opercle behind its articulation midline between the posteroventral margins of the with the hyomandibular. anterohyals. It lies lateral to the hyohyoidei abduc- LEVATOR OPERCULI (Figure 126: L.O.).—The tores, but medial to the branchiostegal rays. fibers arise from the anteroventral posttemporal HYOHYOIDEI ABDUCTORES (Figure 126: H.AB.).— and insert on the posterodorsal rim of the opercle. This is a series of four muscles inserting on the ADDUCTOR ARCUS PALATINI (Figure 128: A.A.P.).— medial faces of the ventral four branchiostegal rays. The muscle covers a fairly small area in front of The ventralmost originates tendinously from the the orbit, where it is well developed. Origin is ventrohyal, the others arising from the base of the from the ventrolateral parasphenoid, stretching second ray and the surrounding anterohyal. They from its suture with the vomer to the flange be- lie medial to the hyohyoideus inferioris. neath the prootic shelf. Insertion is on the dorso- HYOHYOIDEI ADDUCTORES (Figure 126: H.AD.).— medial faces of the mesopterygoid, metapterygoid, The muscle consists of a well-developed sheet of and anterior hyomandibular. fibers in the lateral wall of the opercular cavity, RETRACTOR ARCUS PALATINI (Figure 128: R.A.P.).— stretching from the fascia overlying the sternohy- Origin is from the lateral parasphenoid above oideus to the medial faces of the preopercle, sub- the adductor arcus palatini. Anteriorly, a tendon opercle, opercle, and on up to the prootic shelf. attaches to the dorsomedial face of the mesoptery- The more dorsal branchiostegal rays are attached goid. The dorsomedial face of the tendon is joined to the sheet by connective tissue. There is a small by a small bundle of fibers originating from the bundle of fibers—arising from the anteroventral lateral face of the vomer. cleithrum—which passes dorsally to attach to the ADDUCTOR OPERCULI (Figure 129: AD.OP.).—The main mass of the muscle in the region of the muscle has a fairly extensive horizontal origin from opercular flap. the lateral faces of the parasphenoid and basioc- STERNOHYOIDEUS (Figures 126, 127: STH.).—The cipital. The fibers converge to form a long, thin muscle is subdivided into two distinct sections. The tendon which inserts on the dorsomedial face of larger, lateral portion arises from the posterolateral the opercle. face of the coracoid and the posterior continuity ADDUCTOR HYOMANDIBULAE (Figure 129: A.H.).— with the obliquus inferioris (it has effectively in- This is a broad, short muscle originating from the corporated the anterior myomere of the latter mus- anterodorsomedial face of the cleithrum, as well cle). It inserts on the posteroventral face of the as Baudelot's ligament and the posttemporal in ventrohyal. The medial section originates from the 42 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY anteroventromedial face of the cleithrum and in- about the level of the third hypobranchial and serts on the ventromedial face of the ventrohyal. insert on the ventrolateral base of ceratobranchial Some of the more medial fibers attach to the small 5. urohyal. STERNOBRANCHIALIS (Figures 129, 130: STB.).— DORSAL BRANCHIAL MUSCLES There are two well-separated sections of this mus- cle. The larger and more ventral of these originates LEVATORES EXTERNI I-IV (Figure 129: L.EXT.).— from the fascia overlying the pectoral fin abductors The four muscles arise from the ventrolateral face and the anteroventrolateral face of the cleithrum. of the anterior part of the prootic shelf. Levator Insertion is on the ventromedial face of hypobran- II is overlain anterolaterally by levator I, and by chial 2. The dorsomedial section arises almost levators III and IV posterolaterally. The muscles exclusively from the anteroventral face of the clei- insert on the dorsolateral faces of epibranchials thrum and inserts on the ventral faces of hypobran- 1-4. Levator II is the smallest, and is poorly sep- chial 3 and ceratobranchial 4. arated from levator IV dorsally. LEVATORES INTERNI II, HI (Figures 129, 131: L. INT.).—Levator II is well developed and originates VENTRAL BRANCHIAL MUSCLES from the ventral prootic and pterotic and the lateral PHARYNGOCLAVICULARIS EXTERNUS (Figures 129, parasphenoid. It inserts on the posterodorsal sur- 130: PHC.E.).—The muscle originates from the ven- face of infrapharyngobranchial 2. Levator III is tromedial face of the cleithrum and passes dorsally smaller and more flattened, and arises just behind to insert on the ventral part of ceratobranchial 5. levator II. It includes the exoccipital in its sites PHARYNGOCLAVICULARIS INTERNUS (Figures 129, of origin, and inserts on the posterodorsal face of 130: PHC.L).—The fibers arise from the medial infrapharyngobranchial 3. region of the cleithrum and insert on the ventro- OBLIQUUS POSTERIOR (Figure 131: OBL.P.).—A medial face of ceratobranchial 5. well-developed muscle, which passes from the dor- OBLIQUUS VENTRALIS HI (Figures 129, 130: OBL. somedial face of ceratobranchial 5 to the posterior v.).—The muscle connects the anteroventral region margin of epibranchial 4. of ceratobranchial 3 to the ventral process of hypo- TRANSVERSI DORS ALES I-IV (Figures 129, 131: TR. branchial 3 and the arch-shaped ligament behind D.).—These four muscles connect the dorsal surfaces it. Posteriorly, the fibers mingle somewhat with of the left and right epibranchials of arches 1—4 those of rectus IV. across the dorsal midline. Transversus I passes be- TRANSVERSI VENTRALES rv, v (Figure 130: TR. tween the levatores interni, and also receives a few v.).—The anterior muscle passes across the midline fibers from infrapharyngobranchial 3. The third between the anteromedial faces of the fourth cera- and fourth muscles are not well separated. tobranchials. Transversus V is better developed, and RETRACTOR DORSALIS (Figures 129, 131, 133: connects the fifth ceratobranchials across the ven- D.RETR. ).—This is a well-developed muscle orig- tral midline. inating from the lateral faces of the first few RECTI VENTRALES, I, ii, rv (Figures 129, 130: vertebrae and inserting on the medial face of infra- RECT.VJ.—The first muscle connects the dorsohyal pharyngobranchial 3. to ceratobranchial 1. Rectus II passes from the SPHINCTER OESOPHAGI (Figures 129, 130, 131: anteroventral face of ceratobranchial 2 to a mid- s.o.).—The muscle encircles the esophagus attached ventral raphe, the anterior portion of which is to the posterior margin of ceratobranchial 5. It lies firmly attached to basibranchial 1. Rectus IV is a dorsal to transversus ventralis V and ventral to small muscle arising from the anteroventral region the retractor dorsalis. of ceratobranchial 4 and attaching to the posterior face of the arch-shaped ligament between the ven- tral processes of the third hypobranchials. MUSCLES OF THE PECTORAL REGION RECTUS COMMUNIS (Figures 129, 130: R.COMM.). ABDUCTOR SUPERFICIALIS (Figure 126: ABD.S.).— The muscle originates from the posterior face The mass of the muscle originates from the pos- of the urohyal. The fibers grade into a tendon at terolateral face of the anterolateral flange of the NUMBER 155 43 cleithrum. It inserts on the anterolateral bases of tween the bifid neural spines of the intervening the principal fin rays. vertebrae, and is partially attached to their tips. ABDUCTOR PROFUNDUS (Figures 126, 129: ABD.P.).— Origin is from the ventrolateral regions of the coracoid and cleithrum. The muscle inserts on the MUSCLES OF THE ANAL FIN ventrolateral bases of the principal fin rays, with a small tendon to the ventrolateral base of the ERECTORES ANALES (Figure 133: EREC.).—The vestigial fin ray. first four muscles arise from the anterior faces of the laterally displaced pterygiophores, the rest of ARRECTOR VENTRALIS (Figures 126: ARR.V).—This is a well-developed muscle, originating from the the muscles originating from the lateral faces of posterior face of the anterolateral flange of the the succeeding peterygiophores and the ventrolat- cleithrum and inserting on the anterior base of the eral vertebral column. They insert on the antero- vestigial fin ray (of which only the medial half is lateral bases of all the fin rays. present, the lateral half having apparently fused DEPRESSORES DORSALES (Figure 133: DEPR.).— with it). These muscles originate just behind the erector of each ray, lying somewhat more medially. They ADDUCTOR SUPERFICIALIS (Figure 132: ADD.S.).— insert on the posterolateral bases of the rays. The fibers arise from the posterodorsomedial face of the cleithrum and insert on the anteromedial INFRACARINALIS POSTERIOR (Figure 133: INF.P.)— faces of the principal rays distal to their bases. The The fibers arise from the posterior and lateral more ventral fibers serve the more dorsal rays, faces of the haemal spine of the fourteenth verte- and lie more laterally. bra and pass posteriorly to attach to the antero- ventral surface of the haemal spine of the eighteenth ADDUCTOR PROFUNDUS (Figure 132: ADD.P.).—A vertebra. well-developed muscle, originating over most of the posteromedial faces of the cleithrum and coracoid. It inserts on the ventromedial bases of the principal MUSCLES OF THE CAUDAL FIN fin rays. ARRECTOR DORSALIS (Figure 132: ARR.D.).—Ori- INTERRADIALIS (Figure 134a: INT.).—The muscle gin is from the posteromedial cleithrum, the mus- spans the caudal fin rays dorsoventrally, a short cle inserting on the medial base of the vestigial fin distance from their bases. Each ray gives rise to a ray. small bundle of fibers joining the main mass pos- teriorly, and the more medial rays receive a small fiber bundle anterolaterally. MUSCLES OF THE DORSAL FIN HYPOCHORDAL LONGITUDINALIS (Figure 134: ERECTORES DORSALES (Figure 133: EREC).—Each H.L.).—The muscle originates on the ventrolateral fin ray has an erector muscle inserting on its surface of the hypural plate, and inserts tendinous- anterolateral base. The fibers originate from the ly on the ventrolateral bases of rays D 3-5. lateral faces of the neural spines and centra of the FLEXOR DORSALIS SUPERIOR (Figure 134: F.D.S.).— seventh to thirteenth vertebrae and the lateral This is a fairly small muscle, originating from the pterygiophores. They tend to lie somewhat lateral dorsolateral region of the ural centrum and fused to the depressores dorsales, and decrease in size hypural-epural region and inserting on the antero- posteriorly. dorsal face of ray D 5. DEPRESSORES DORSALES (Figure 133: DEPR.).—The FLEXOR VENTRALIS INFERIOR (Figure 134: F.V.I.).— origin of these muscles is similar to that of the Origin is from the ventrolateral region of the ural erectors, but somewhat more medially. They insert centrum and fused hypural-parhypural region, the on the posterolateral bases of the fin rays. muscle inserting on the ventrolateral base of ray SUPRACARINALIS POSTERIOR (Figures 133, 134a: V5. S.POST.).—The muscle arises from the posterior and TRANSVERSUS CAUDALIS (Figure 1346: TR.C).—The lateral faces of the neural spines of the thirteenth fibers originate from the dorsolateral surface of the and fourteenth vertebrae and ends on the antero- hypural plate and insert tendinously on the dorso- dorsal face of the eighteenth vertebra. It passes be- lateral faces of rays V 2-4. 44 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

LATERAL BODY MUSCLES vertebra, and the muscle as a whole is considerably smaller than swim bladder "a." EPAXIALIS (Figures 126, 133, 134a: EPAX.J.—The The wall of the swim bladder beneath these muscle arises anteriorly from the posterodorsal re- muscles is very thin, and the bladder is firmly at- gion of the skull. It passes posteriorly, and is only tached to the ventral processes of the first six verte- attached to the vertebral column behind the dorsal brae (as described by Tyler, 1963b: 159). fin. Here the dorsolateral fibers grade into a ten- don that joins a similar tendon from the obliquus inferioris. The cruciform tendon thus formed in- Lactoria cornuta (Linnaeus) serts on the anterolateral bases of the uppermost FICURE 135 and lowermost rays of the caudal fin. The re- maining fibers of the muscle more or less fuse with Certain proportional differences are evident. The the obliquus superioris medially, and insert by a shorter snout and more down-tilted mouth are re- continuous tendinous sheath on the anterolateral flected in the shape of A 2 and the dilatator bases of the dorsal fin rays (D 1-5). operculi. The sections of A 2 are poorly separated, OBLIQUUS SUPERIORIS (Figures 133, 134a: OBL.S).— as are those of A 10. This is even more apparent The section is much reduced. Anteriorly it arises in the 50 mm specimen, where both A 1 and A 2 from the lateral faces of the centra above the swim appear as single, undivided muscles. The levator bladder. Posteriorly, in the urostyle region, the arcus palatini is well developed, and its fibers are fibers appear to fuse dorsally with the epaxialis. It not entirely separate from those of the dilatator inserts tendinously on the anterolateral bases of operculi posteriorly. the fin rays. In front of the anal fin the muscle There is a slip of the hyohyoidei abductores to is completely separate from the epaxialis dorsally the fifth branchiostegal ray. The interradialis is and the obliquus inferioris ventrally. very well developed, and the hypochordalis longi- OBLIQUUS INFERIORIS (Figures 126, 133, 134a: tudinalis inserts only on the base of D 5. The OBL.1.).—The muscle consists of a ventral portion transversus caudalis has a more restricted origin only. Anteriorly, it is laterally continuous with from the hypural plate dorsally, but a more ex- the sternohyoideus, the deeper fibers attaching to tensive one anteroventrally. Both the flexor dor- the posterior face of the coracoid. Posteriorly, it salis superior and the flexor ventralis inferior are grades into a tendon that forms the ventral com- better developed, and have a more extensive origin, ponent of the cruciform tendon. Most of the fibers which includes the neural and haemal spines of from this tendon pass dorsally to the base of the the penultimate vertebrae respectively. uppermost fin ray. There is little fusion with the Both the muscles to the swim bladder are better obliquus superioris, and the more ventrolateral developed, swim bladder "a" being more firmly at- fibers have some connection to the carapace. tached to the lateral surface of the vertebral col- umn medially.

SWIM BLADDER MUSCLES Tetrosomus concatenatus (Bloch) SWIM BLADDER "a" (Figure 133: s.BL.a.).—The main mass of the muscle passes across the dorsal FIGURE 136 surface of the anterior half of the swim bladder. The subdivision of A 2 is not at all clear, and The fibers originate from the lateral faces of the A IP' is hardly distinguishable from A 13". The centra and overlying fascia of the middle ab- levator arcus palatini is fairly small, and there is dominal vertebrae and pass posterolaterally to in- no dorsal bundle of fibers to the dorsomedial face sert on the dorsolateral face of the swim bladder. of the tendon of the retrator arcus palatini. The SWIM BLADDER "b".—The muscle lies beneath subdivisions of the sternobranchialis are not as swim bladder "a," near the midline. The fibers clear as in either of the previous species. Levator lie in a longitudinal plane, inclining slightly to- internus II is extremely well developed, with a ward the midline anteriorly. They lie very close broad insertion on infrapharyngobranchial 2. The to the ventral surface of the seventh abdominal obliquus posterior is somewhat reduced. The hy- NUMBER 155 45 pochordal longitudinalis inserts by a broad, un- on the middle eight caudal fin rays (D 1-4 and divided tendon on the bases of rays D 3—5, while V 1—4), there being no tendons to D 5 or V 5. the transversus caudalis inserts on rays V 3—5. The swim bladder musculature differs from O. tuberculatus. The lateral and more posterior fi- bers pass dorsally and medially from their insertion Acanthostracion quadricornis (Linnaeus) on the anterodorsal surface of the swim bladder to their origin on the dorsolateral face of the vertebral FIGURE 137 column. Anteriorly, the fibers tend to curve ven- The sections of A 2 are well separated, and there trally and then posteriorly from their origin on are two fiber directions in A 2a. A 13" is the vertebrae, and lie medial to the outer, more well separated from the more ventral fibers, but posterior fibers. This leaves a gap between the the other two portions are scarcely distinct. The ventromedial surfaces of the transverse posterior anterior bundle of fibers of the retractor arcus fibers and the dorsolateral surface of the anterior palatini is well developed. The dilatator operculi and more longitudinal fibers, giving rise to the has an entirely tendinous insertion on the tip of "hollow ball of muscles" (Tyler, 1963b: 159). the dorsal process of the opercle, and the levator arcus palatini is well developed. Summary of Ostraciidae The dorsolateral half of the protractor hyoidei passes lateral to the interopercle to insert on the The main variation in this family is found in dentary. the jaw and caudal fin musculatures. In the former, The interradialis partly overlies the tendons of the differences involve mainly the degree of sub- the hypochordal longitudinalis, and the transversus division of the adductor mandibulae, while in the caudalis inserts on rays V 3-5. The flexor dorsalis latter most variation is seen in the sites and number superior and the flexor ventralis inferior are better of rays involved in the insertion of the hypochordal developed than in O. tuberculatus. longitudinalis and transversus caudalis. Apart The epaxialis and the obliquus superioris insert from this, this family is conservative in its myology.

Myological Description of the Representative Triodontid

The outline of Triodon, the monotypic represent- Triodon is usually considered to be the pleiso- ative of this family, is given in Figure 7. The morph member of the gymnodonts (e.g., Tyler, species inhabits the deeper waters (150-300 meters) 1962c), and to form a morphological intermediate of the continental shelves of the Indo-Pacific. Al- between them and the triacanthoids + balistoids most nothing is known of the life history and + ostracioids. habits of this rare fish. Of the four specimens dissected for this study, only one had anything in Triodon macropterus Lesson its gut. This specimen (270 mm SL) contained the remains of a number of echinoids (PClypeasteroi- FIGURES 138, 139a, 140-145 dea), with occasional carapaces of shrimplike crus- taceans. The lunate caudal fin, slender peduncle, MUSCLES OF THE CHEEK and poorly developed dorsal and anal fins suggest ADDUCTOR MANDIBULAE (Figures 138, 139, 140: that Triodon is a fairly fast swimmer, but there is A 1, A 2a, A 23, A 3).—A 1 is a single, undivided no direct evidence. The pelvic girdle is repre- muscle which originates beneath the orbit as a flat, sented by a long, shaftlike bone, very similar to that aponeurotic sheath attached primarily to the antero- of the balistoids. The jaws are massive and beak- dorsal face of the hyomandibular. It inserts on like, and form a powerful crushing apparatus, as the posteromedial face of the ventral half of the they do in all the remaining families except the maxilla. A 2 is separated into two parts by the Canthigasteridae (although even here the basic passage of ramus mandibularis V. A 2a originates plan remains). exclusively from the anterodorsolateral face of the 46 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY hyomandibular—a few fibers may attach to the hyomandibular fossa ventral to, but dorsally con- ventral face of the aponeurosis of A 1 posteriorly. tinuous with, the origin of the adductor hyoman- The extreme posterior fibers are covered laterally dibulae. The fibers pass out laterally to insert on by the dilatator operculi. A 2P, which lies beneath the dorsomedial face of the opercle, just posterior A 2a, originates almost entirely from the dorsal to its articulation with the hyomandibular. and lateral faces of the preopercle. The postero- ADDUCTOR HYOMANDIBULAE (Figure 142: A.H. ).— dorsal fibers overlie the posteroventral region of The muscle originates from the ventrolateral A 2a. The two subdivisions insert together on the prootic just above the adductor operculi, and in- medial face of the dentary above the Meckelian serts as a more or less flattened sheet on the pos- fossa. A 3 is present at a thin sheet of fibers orig- terodorsomedial face of the hyomandibular. inating from the lateral metapterygoid and sym- plectic. It inserts medial to A 2 on the dentary by a somewhat sheetlike aponeurosis. It is partially MUSCLES OF THE HYOID REGION confluent with A 2P posteriorly. A a> is well de- VALVULUS.—The muscle is present as a thin, flat- veloped, and stretches along the medial surface of tened sheet of fibers, arising from the dorsolateral the quadrate and anterior metapterygoid. It in- face of the protractor hyoidei at about the middle serts by a flat aponeurosis on the dentary, lying of its length and grading into the tissues of the medial to the insertion of A 3. buccal valve anterodorsally. LEVATOR ARCUS PALATINI (Figures 138, 140: PROTRACTOR HYOIDEI (Figure 138: PR.HY.).—The L.A.P.)—From a straplike origin on the ventral fibers arise from the posteroventromedial dentary sphenotic at the rear of the orbit, the muscle and insert on the posteroventral anterohyal and passes downward to insert on the posterodorsolat- the adjoining posterohyal. It completely covers eral face of the hyomandibular. This point lies the first two branchiostegal rays, and is a long, medial to the dorsolateral flange of that bone, cylindrical muscle. which has developed in association with the origin HYOHYOIDEUS INFERIORIS.—The fibers of this mus- of A 2a. The fibers of the muscle fan out before cle are partially fused with those of the protractor their insertion, this being particularly marked hyoidei. It arises from the ventral face of the anteriorly. anterohyal, sharing, to some degree, aponeuroses DILATATOR OPERCULI (Figure 138: D.O.).—Origin with the medial region of the protractor hyoidei. is from the ventrolateral sphenotic and adjoining The fibers pass anteromedially and fuse with their pterotic. The muscle narrows to a tendon which antimeres. The muscle lies lateral to the origin inserts on the enlarged anterodorsal process of the of the infracarinalis anterior, and there is some opercle. intermingling of fibers with those of the protractor LEVATOR OPERCULI (Figure 138: L.O.).—This is a hyoidei. broad, well-developed muscle, originating in the HYOHYOIDEI ABDUCTOREs (Figure 138: H.AB.).—The groove between the pterotic and the large supra- muscle is made up of three slips. That to the first cleithrum. It inserts broadly on the dorsomedial ray originates from the medial face of the ven- face of the opercle. A small superficial aponeuro- trohyal and inserts broadly over the ventromedial sis is present on the midlateral face of the muscle. face of the ray. The slip to the second ray arises ADDUCTOR ARCUS PALATINI (Figure 140: A.A.P.| from the expanded dorsolateral surface of the first The muscle is confined to the floor and rear of the ray, while that to the third ray originates from the orbit. Origin is from the lateral surface of the bases of the first and second rays. parasphenoid. Anteriorly it inserts on the dorso- HYOHYOIDEI ADDUCTORES (Figure 138: H.AD.).— medial face of the mesopterygoid. There is a gap This is a well-developed sheet of muscle fibers. It where the fibers end in the tough connective tissue is firmly attached to the dorsomedial face of opercle lining the buccal cavity. Posteriorly, the fibers in- dorsally, and fades out in the wall of the opercular sert on the posterodorsal faces of the metaptery- chamber in the region of the dorsohyal antero- goid and hyomandibular. ventrally. ADDUCTOR OPERCULI (Figure 142: AD.OP.). The STERNOHYOIDEUS.—The muscle is small, its origin muscle originates from the prootic beneath the being confined to the anterodorsal tip of the ventral NUMBER 155 47 part of the cleithrum, including a small fossa in insert on the dorsal faces of epibranchials 1—4. that bone. Anteriorly, the fibers sweep up to insert The first two muscles are not very well developed, on the ventral face of the ventrohyal. the fourth levator being the largest. LEVATORES INTERNI II, HI (Figure 142: L.INT.).— The muscles originate from the prootic beneath VENTRAL BRANCHIAL MUSCLES the hyomandibular fossa, levator II having a more medial origin than levator HI. The former inserts PHARVNCOCLAVICULARIS EXTERNUS (Figure 141: partially on the medial face of infrapharyngo- PHC.E.).—This is a straplike muscle originating branchial 2 but mainly on the anteromedial face from the lateral face of the cleithrum and inserting of infrapharyngobranchial 3. Levator III inserts on the anteroventral face of ceratobranchial 5. The on the dorsal face of infrapharyngobranchial 3. fibers pass posterodorsolaterally from their origin. OBLIQUUS DORSALIS HI (Figure 142: OBL.D.).—Two PHARYNGOCLAVICULARIS INTERNUS (Figures 141, heads of this muscle arise from the anterolateral 142: PHC.I.).—Orgin is from the anterodorsomedial face of epibranchial 3. The fibers coalesce and face of the cleithrum, the fibers passing anteriorly pass medially to attach to the posterodorsal face of to insert on the posterior face of ceratobranchial 5. infrapharyngobranchial 3. It is a flattened muscle. OBLIQUUS POSTERIOR (Figure 142: OBL.P.).—Fibers OBLIQUI VENTRALES II, HI (Figure 141: OBL.V.).— pass from the dorsomedial face of ceratobranchial The anterior muscle connects the ventral faces of 5 to the posterior face of epibranchial 4. the ceratobranchial and hypobranchial of the sec- ond arch. Obliquus III joins ceratobranchial 3 to TRANSVERSI DORSALES II, HI (Figure 142: TR.D.).— The anterior muscle is well developed and partially the ventral process of hypobranchial 3 and the subdivided. The dorsal layer arises from infra- arch-shaped ligament. pharyngobranchial 2 and the surrounding connec- TRANSVERSI VENTRALES rv, v (Figure 141: TR.V.).— tive tissue. The fibers pass posteriorly and then The anterior muscle connects the anteromedial curve toward the midline just behind the basal faces of the fourth ceratobranchials across the ven- process of the parasphenoid, and join their anti- tral midline. Transversus V is well developed, and meres. Beneath this, the ventral layer passes out- joins the anteromedial faces of the fifth cerato- ward from the midline to attach to the anterolateral branchials. face of epibranchial 4, the two layers separating RECTI VENTRALES I, iv (Figure 141: RECT.V.).— completely. Transversus III arises from the dorsal Rectus I is a fairly long, straplike muscle which midline and passes laterally to insert on the dorsal connects the dorsal face of the dorsohyal and the face of the medial process of epibranchial 3. ventrolateral face of the first ceratobranchial and RETRACTOR DORSALIS (Figure 142: D.RETR.).—The hypobranchial. Rectus IV arises from the antero- muscle originates from the lateral face of the first ventral face of ceratobranchial 4 and passes antero- vertebra and passes anteriorly to insert on the medially to attach to the posterior surface of the ventromedial faces of infrapharyngobranchials 3 arch-shaped ligament between the ventral processes and 4. of the third hypobranchials. ADDUCTORES rv, v (Figure 142: AD.).—Adductor RECTUS COMMUNIS (Figure 141: R.COMM.).—The IV is well developed, and connects the medial faces muscle originates anteriorly from the posterior face of epibranchial 4 and ceratobranchial 4. Adductoi of the urohyal. The fibers pass posteriorly, grading V is less well developed, and passes from the an- into an aponeurosis at about the level of the third terodorsal tip of ceratobranchial 5 to the postero- basibranchial which inserts on the ventrolateral lateral face of epibranchial 4. face of ceratobranchial 5. SPHINCTER OESOPHAGI (Figures 141, 142: s.o.).— The anteroventral fibers of this muscle attach to DORSAL BRANCHIAL MUSCLES the ventromedial face of ceratobranchial 5. There are two fairly distinct layers dorsally. The ventral LEVATORES EXTERNI I—IV (Figure 142: L.EXT.).— one arises from infrapharyngobranchial 2, the fi- The four muscles originate from the prootic just bers passing posteriorly and then medially before anteroventral to the hyomandibular fossa. They fading out in the connective tissue of the esopha- 48 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY gus. This is covered posterodorsally by the main to insert on the extreme posterodorsal tip of the mass of the muscle, which does not extend anteri- cleithrum. The muscle is almost entirely covered by orly beyond the posteromedial face of epibranchial the supracleithrum, and is closely associated with 4. the fibers of the epaxialis dorsally.

MUSCLES OF THE PECTORAL REGION MUSCLES OF THE DORSAL FIN

ABDUCTOR SUPERFICIALIS (Figure 1S8: ABD.S.).— INCLINATORES DORSALES (Figure 144: INC.).—Each The muscle arises from the posteroventral surface fin ray possesses an inclinator which arises in the of the elongated cleithrum and inserts on the fascia between the skin and the epaxialis and in- anterolateral bases of the principal fin rays. serts on the lateral base of the fin ray a little dorsal ABDUCTOR PROFUNDUS (Figure 138: ABD.P.).— to the insertions of the erector and depressor mus- Origin is from the lateral face of the cleithrum. cles. The inclinators decrease in size posteriorly. The fibers insert on the posterolateral bases of the ERECTORES DORSALES (Figure 144: EREC).—There principal rays and on that of the lateral half of are two small, anterior erectors, which originate the vestigial finray . from the lateral faces of the first two pterygiophores ARRECTOR VENTRAUS (Figure 138: ARR.V.J. —The and insert on the bases of the spines of the spiny muscle lies above the abductor superficialis and dorsal fin. Each fin ray has an erector inserting on originates from the lateral face of the cleithrum its anterolateral base. These muscles arise from the beneath the flange. It inserts on the anterior face lateral faces of the pterygiophores and the neural of the medial half of the vestigial fin ray. There spines and centra of the eighth to fourteenth verte- may be a few fibers attaching to the lateral half of brae. the ray posteriorly. DEPRESSORES DORSALES (Figure 144: DEPR.).—The ADDUCTOR SUPERFICIALIS (Figure 143: ADD.S.).— first dorsal spine only has a depressor muscle, The fibers originate from the anteromedial face of which originates from the lateral base of the first the cleithrum and insert on the dorsomedial faces pterygiophore and inserts on the posteroventral of the principal rays distal to their bases. The base of the spine. Each ray in the soft dorsal finha s more ventrolateral fibers serve the more dorsal rays. a depressor inserting on its posterolateral base. Ori- ADDUCTOR PROFUNDUS (Figure 143: ADD.P.).—The gin is as for the erectors, the depressor to each ray muscle is well developed, and originates from the lying just behind the erector. medial face of the cleithrum and coracoid. It in- SUPRACARINAUS ANTERIOR.—This is a well- serts on the ventromedial bases of the principal fin developed muscle arising from the posterior and rays, the tendons to the ventral two rays being lateral faces of the supraoccipital. The fibers pass covered medially by the main mass of the muscle. posteriorly to attach to the anterior face of the first ARRECTOR DORSALIS (Figure 143: ARR.D.).—This is pterygiophore. There is some intermingling of a fairly small muscle, originating from the medial fibers from the left and right sides in the midline. face of the cleithrum and inserting on the base of SUPRACARINALIS POSTERIOR (Figures 144, 145: s. the medial half of the vestigial fin ray. It is partial- POST.).—The muscle stretches from the posterior ly covered medially by the adductor profundus. face of the last pterygiophore of the dorsal fin and CORACORADIALIS (Figure 143: COR.R.)._A some- the neural spine of the fourteenth vertebra to the what flattened muscle, arising from the dorsal face anterodorsal regions of the neural spines of the of the posterior process of the coracoid and insert- third and fourth last vertebrae. ing on the ventromedial face of the fourth radial. PROTRACTOR PECTORALIS (Figures 138, 142: P.P.). MUSCLES OF THE ANAL FIN The muscle originates from the ventrolateral tip of the pterotic and fans out posteroventrally to attach INCLINATORES ANALES (Figure 144: INC.).—An in- to the dorsal region of the cleithrum. clinator arises from the fascia over the obliquus LEVATOR PECTORALIS (Figure 138: TR.).—Origin is inferioris and inserts on the anterolateral base of from the posterior face of the pterotic. The fibers each fin ray. The medial fibers of the more anterior pass mainly posteriorly, but also somewhat laterally, inclinators attach to the myocommata of the NUMBER 155 49 obliquus superioris, making separation of the mus- FLEXOR DORSALIS (Figure 145c: F.D.).—Origin is cle difficult. In addition, there is a large, residual restricted to the lateral faces of the epural, the first inclinator inserting on the anteroventrolateral base uroneural and the dorsal crest of the ural centrum of the first pterygiophore. Its ventromedial region The tendons insert on the bases of rays D 2-6. is more or less continuous with the ventral fibers of FLEXOR DORSALIS SUPERIOR (Figures 1456/r: F.D. the obliquus inferioris. s.).—The muscle originates from the lateral faces ERECTORES ANALES (Figure 144: EREC).—These of the epural and the neural spines of the second muscles, which insert on the anterolateral bases of and third last vertebrae. Insertion is on the lateral the anal fin rays, originate almost entirely from base of ray D 7. the lateral faces of the pterygiophores, with only a FLEXOR VENTRALIS (Figure 145c: F.V.).—The mus- few fibers from the tips of the haemal spines of cle is fairly well developed, and lies beneath the the associated vertebrae. flexor ventralis externus. Origin is from the lateral DEPRESSOR re ANALES (Figure 144: DEPR.).—Origin faces of the centra and haemal spines of the last is as for the erectors, the depressor of each fin ray three vertebrae (including the ural centrum), the inserting on the posterolateral base of the ray. parahypural, and the hypurapophysis. The muscle INFRACARINALIS ANTERIOR (Figure 138: INF.A.).— inserts on the bases of rays V 1-7. This is a well-developed muscle which originates FLEXOR VENTRALIS EXTERNUS (Figure 145c: F.V. from the anteroventral tip of the cleithrum and E.).—The fibers arise from the lateral faces of the inserts in the anteroventral groove in the pelvis. centra and haemal spines of the third and fourth Evidence of its originally paired nature is visible last vertebrae and pass posteriorly to insert on the externally, but the deeper fibers of either side have bases of rays V 1-2. fused in the midline. FLEXOR VENTRALIS INFERIOR (Figure 1456^: F.V. INFRACARINALIS MEDIUS (Figure 138: INF.M.).—A L).—Origin is primarily from the lateral face of the fairly well-developed section, which attaches to the haemal spine of the penultimate vertebra, the ventrolateral face of the pelvis just anterior to the muscle inserting on the lateral base of V 7. middle of its length. It grades into the obliquus TRANSVERSUS CAUDALIS (Figure 145c: TR.C).—The inferioris at about the level of the postcleithrum, muscle is poorly developed. It originates from the the more posterolateral fibers intermingling some- lateral face of the fourth hypural and inserts on what with those of the residual inclinator analis. the dorsal base of ray V 1. INFRACARINALIS POSTERIOR (Figures 144, 145: INF. P.).—The muscle arises from the posterolateral LATERAL BODY MUSCLES faces of the last anal pterygiophore and the haemal spine of the fifteenth vertebra and attaches to the EPAXIALIS (Figures 138, 145a,b: EPAX.). — The anteroventral face of the haemal spines of the third subdivision of this muscle into dorsal and ventral and fourth last vertebrae. moieties is incomplete, particularly medially. Anteriorly, the dorsal section attaches to the pos- terodorsal faces of the frontal, epiotic, pterotic, MUSCLES OF THE CAUDAL FIN supraoccipital, and exoccipital. The myocommata INTERRADIALIS (Figure 145a: INT.).—The muscle become increasingly oblique posteriorly until they is developed between the middle twelve caudal fin are almost horizontal (behind the dorsal fin). An- rays. The fibers span a few rays between their sites teromedially, some of the fibers arise from the of attachment. Those of the dorsal rays pass antero- neural spines and centra of the vertebrae and the ventrally, while those of the ventral rays pass pterygiophores. Posteriorly, the fibers grade into a anterodorsally. flattened tendon which joins a similar tendon from HYPOCHORDAL LONGITUDINALIS (Figure 1456: the obliquus inferioris. Together, these two tendons H.L.).—The muscle is well developed, and originates form an aponeurotic sheath inserting on the bases from the dorsolateral face of the ural centrum, and of all the procurrent and principal caudal fin rays. the lateral faces of the hypurals 1 and 2, parhy- The ventral section of the epaxialis is virtually con- pural, and hypurapophysis. The muscle inserts on tinuous with the obliquus superioris. The fibers the bases of rays D 3-5. arise anteriorly from the posterior faces of the 50 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY pterotic and exoccipital. In the posterior region the thrum, and attaches to the anteroventral face of muscle inserts on the lateral bases of rays D 1-6, the postcleithra. Two vague myocommata are pres- the more ventral tendons passing medial to the ent, and occasional fibers are to be found in the hypochordal longitudinalis. connective tissue between the ventral border of OBLIQUUS SUPERIORIS (Figures 138, 145a,b: OBL. this section and the infracarinalis medius below s.).—This section arises anteriorly as a small slip it. A thin sheet of fibers passes obliquely upward of fibers from the posterodorsal tip of the clei- from the posterodorsal faces of the postcleithra thrum. It fans out posteroventrally, attaching to and fades into the fascia overlying the obliquus and between the pleural ribs, with the ventral superioris. Posteroventrally, the fibers become more fibers intermingling with the obliquus inferioris. horizontally oriented, and intermingle with those Behind the anal fin, these two sections are rela- of the obliquus superioris, infracarinalis medius, tively well separated. It inserts tendinously on the and residual inclinator analis. Myocommata, which bases of rays V 1-6. A large bundle, which appears are continuous with those of the obliquus superi- to be mostly continuous with the obliquus superi- oris, are present in the fibers behind the postclei- oris, but which also receives fibers from the epaxi- thra. From the ventromedial surface of the alis, passes somewhat dorsally in the peduncular posterior myocommata of the eleventh myomere, region to insert on the lateral crest of the ural and closely associated with the medial faces of centrum. It is slightly overlain posteriorly by the hypochordal longitudinalis. It is somewhat reminis- the infracarinalis medius and the rest of the ob- cent of a proximalis muscle, but is not separated liquus inferioris, there arises a short bundle of from the body muscles. fibers which passes posterodorsomedially to insert OBLIQUUS INFERIORIS (Figures 138, 144, 145a: tendinously on the anteroventral face of the haemal OBL.I.).—The muscle has a number of subsections. spine of the fourteenth vertebra. The fibers pass The anteroventral section arises from the posterior medial to the obliquus superioris but lateral to the and posteroventral faces of the coracoid and clei- erectores and depressores anales.

Myological Descriptions of Representative Tetraodontids

The members of this family (see Figure 8 for sea-grass meadows, sandy bottoms, and the adjoin- general outline) as recognized here contain ten ing coral reefs. genera (representatives of six genera were dissected The family is considered to be closely related to for this study), which were divided into four fam- the Canthigasteridae (which latter has often been ilies by Fraser-Brunner (1943). Although his fam- included in the Tetraodontidae), and also close ilies may well represent distinct phyletic lines, the to the Diodontidae. status accorded them by Fraser-Brunner has not been followed by other workers (e.g., Tyler, 1970a). Lagocephalus lunaris (Bloch and Schneider) The skin is smooth or covered with small prick- les, and the fish are capable of considerable FIGURES 1396, 146-155 inflation when threatened. Some forms have an almost lunate caudal fin (e.g., Lagocephalus), but MUSCLES OF THE CHEEK a rounded caudal is more common. Dorsal and anal fins are usually small, and normally beat ADDUCTOR MANDIBULAE (Figures 1396, 146, 147, synchronously from side to side. The pectoral fins 148, 149: A la, A IP, A 2a, A 2P, A 3)._A 1 con- are large, and often used in locomotion. The pow- sists of two well-developed sections. A IP has a erful beaklike jaws are well adapted for crushing complicated fiber structure. Fibers originating from the mollusks and brachyurans which form a major the lateral faces of the palatine, vomer, and eth- part of the puffer-fish diet. They are generally moid pass ventrolaterally. They form a bipinnate tropical to subtropical in distribution (with a few structure with the fibers arising from the anterior fluviatile forms), and are commonly found over face of the prefrontal (which pass ventromedially). NUMBER 155 51

Fibers from the ventrolateral tip of the prefrontal are almost parallel, but the muscle becomes bipin- course anteroventrally, the fibers from the ventral nate posteriorly. Origin is from the lateral faces face of that bone at first pass ventrally, and then of the frontal and sphenotic, and the dorsolateral (as they become more posterior in origin) antero- hyomandibular about its articulation with the ventrally. The ventral surface is largely aponeu- skull. Insertion is on the dorsolateral opercle, rotic. Beneath this, another bundle of fibers is mainly around the strong, dorsoventrally oriented present, arising from the lateral parasphenoid. It arista, but also on the anterolateral surface of the is well separated from the above-mentioned aponeu- bone. It is overlain dorsally by an aponeurotic rosis posteriorly, but grades into it anteriorly. The sheath from the epaxialis. Anterolaterally, a strap muscle inserts on the dorsomedial faces of the pre- of muscle fibers attaches to the posteroventral face maxilla and maxilla. A 1 a is smaller. It originates of the preopercle. from an aponeurosis on the floor of the orbit (at- LEVATOR OPERCULI (Figure 146: L.O.).—This is taching posteriorly to the hyomandibular medial a flattened muscle, which arises from the antero- to A 2 a) and from a small head on the postero- ventrolateral pterotic and the ventrolateral face of lateral preopercle. Insertion is on the posteromedial the posterior process of the frontal. It inserts on face of the maxilla. A 2a is very well developed the posterodorsal face of the opercle and its crista. and is expanded ventrally to originate from two Some of the ventral fibers are closely associated heads. The anteroventral one arises from the quad- with the dorsal fibers of the protractor pectoralis. rate, symplectic, lateral preopercle, and the antero- ADDUCTOR ARCUS PALATINI (Figures 147, 148: ventral tip of the hyomandibular. The other head A.A.P. ).—The muscle is extensively developed, arises from the dorsolatera"l preopercle and postero- particularly anteriorly. It originates from the an- lateral hyomandibular. There is a considerable teroventral prootic, lateral parasphenoid, and pos- variety of fiber direction. Anteriorly, a tough teroventrolateral vomer. The fibers beneath the aponeurosis joins A 23. The section lies lateral vomer stretch across the midline. Insertion is on the to ramus mandibularis V apart from a small fiber medial face of the palatine, and the dorsomedial bundle from the dorsolateral quadrate. A 23, faces of the ectopterygoid, metapterygoid, and hyo- which lies medial to the above nerve, has shifted mandibular. dorsally in origin, and lies above A 3. It originates ADDUCTOR OPERCULI (Figures 147, 150: AD.OP.).— from the lateral parasphenoid and prootic. The The muscle originates from the anteroventral face fibers pass laterally and anteroventrally to grade of the pterotic and passes outward as a thin sheet into a tendinous aponeurosis which joins that of to insert on the dorsomedial face of the opercle A 2 a. The two sections insert on the medial face just medial to the insertion of the levator operculi. of the dentary. A 3 is a distinct muscle arising from the lateral faces of the metapterygoid, quad- rate, and ectopterygoid and inserting by a flat, MUSCLES OF THE HYOID REGION tendinous aponeurosis on the dorsal surface of VALVULUS (Figure 149: VA.).—The muscle, Meckel's cartilage. A co is a small muscle (Figures which is not fully separated from the protractor 147, 149). It arises from the anteromedial edge of hyoidei posteriorly, arises from the anteroventral the ectopterygoid and inserts on the ventral rim surface of the interopercle. It passes anteriorly, of the Meckelian fossa. It lies medial to the tendon and then curves upward to face out in the tissues of A 3. of the mandibular buccal valve. LEVATOR ARCUS PALATINI (Figures 146, 147: PROTRACTOR HYOIDEI (Figure 146: PR.HY.).—Ori- L.A.P.).—The fibers of this muscle are continuous gin is from the ventral margins of the dentary, dorsolaterally with the dilatator operculi. Origin angular, anterior preopercle, and interopercle. The is from the ventrolateral tip of the frontal, and fibers insert on the lateral faces of branchiostegal the fibers fan out to insert in the fossa on the rays 2-4, and the anterohyal between these rays. posterodorsal face of the hyomandibular, just an- The more posterior fibers tend to course more terior to the articulation with the skull. toward the midline from their sites of insertion. DILATATOR OPERCULI (Figure 146: D.O.).—This HYOHYOIDEI ABDUCTORES (Figures 146, 150: H. muscle is very well developed. The anterior fibers AB.).—A bundle of fibers passes posteroventrally 52 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY from the base of the second ray to the medial faces STERNOHYOIDEUS (Figures 146, 150: STH.).—The of the third and fourth rays. The section serving muscle is somewhat, conical, and is poorly devel- the second ray arises from the posterior face of oped. It originates from the anteroventral tip of the lateral process of the enlarged first ray. The the cleithrum and inserts on the dorsomedial face muscle to the first ray is massively developed. The of the anterior region of the anterohyal. ray forms a flat, horizontal plate extending pos- teriorly from the anterohyal, which lies lateral to VENTRAL BRANCHIAL MUSCLES it. The muscle has a large number of fiber direc- tions. The ventral surface of the ray gives rise to PHARYNGOCLAVICULARIS EXTERNUS (Figures 150, a large bundle of fibers that pass anteriorly and 151: PHC.E.).—Origin is from the ventromedial face a little dorsally to originate from the ventral and of the cleithrum, the fibers passing dorsally to in- lateral anterohyal. The more lateral fibers, includ- sert on the anteroventral face of ceratobranchial 5. ing those from the lateral process of the ray, tend PHARYNGOCLAVICULARIS INTERN US (Figures 150, to sweep anterodorsally, and then medially to the 151: PHC.L).—The muscle arises from the medial midline, where they join their antimeres. A large face of the cleithrum, and consists of two sections. volume of fibers inserts on the dorsomedial face of The anterior section passes anterodorsally, medial the ray. They originate from the fascia overlying to the pharyngoclavicularis externus, to insert on the sternohyoideus, the lateral face of the first the posteroventral face of basibranchial S. The basibranchial and the medial anterohyal. Most of more medial fibers of this section join their anti- these fibers, however, arise from the midline, where meres in a midventral raphe, while some of the they join with their antimeres. Although undoubt- more dorsal ones arise from ceratobranchial 4 and edly functioning as two muscles, there is no physi- pass to basibranchial 3. The more posterior section cal hiatus, and the fibers from the two bundles arises from the anteroventromedial cleithrum and intermingle to a considerable extent. The postero- inserts on the ventral face of ceratobranchial 5. dorsomedial edge of the ray gives rise to a small OBLIQUUS VENTRAUS HI (Figure 151: OBL.V.).— bridge of fibers across the midline, which lies be- The muscle originates from the anteroventral re- neath the ventral branchial musculature at about gion of ceratobranchial 3 and the posteroventral the level of the third basibranchial. tip of hypobranchial 3. The fibers pass antero- HYOHYOIDEI ADDUCTORES (Figures 146, 147, 150: medially to join with the fibers of rectus IV in an H.AD.).—The muscle is somewhat subdivided. The insertion on the anteroventral process of hypobran- main mass passes dorsoventrally in the lateral wall chial 3. of the opercular cavity, with a dorsal bundle at- TRANSVERSI VENTRALES IV, v (Figure 151: TR.V.).— taching to the dorsomedial face of the hyomandib- These muscles connect the anteromedial faces of ular. Ventrally, the fibers (to which the medial the ceratobranchials of the fourth and fifth arches faces of branchiostegal rays 2-6 attach) pass an- across the ventral midline. Both muscles are small. teromedially to meet another sheet of fibers from RECTI VENTRALES i-rv (Figures 150, 151: RECT. the medial floor of the opercular cavity, the latter v.).—Rectus I is well developed, passing from the passing anterolaterally. Posterolaterally, a bundle dorsolateral face of the anterohyal to the antero- arises from the posteroventral face of the supraclei- lateral surface of hypobranchial 1. The lateral sur- thrum and a myocomma with the anterior part of face of the muscle gives rise to a bundle of fibers the obliquus inferioris. These fibers pass antero- passing posterodorsally, which grade into a tendon ventrally and then posteroventrally in a semicircle attaching to the ventromedial face of the hyo- to attach ventrally to the dorsolateral tip of the mandibular just lateral to its articulation with the coracoid. From the anteroventral margin of this skull. Recti II and III connect the hypobranchials semicircle, a bundle passes anteriorly to attach of the first and second and second and third arches along the ventrolateral surface of the cleithrum. respectively. Rectus IV passes from the anteroven- Anterodorsally, fibers from this arch are continuous tral face of ceratobranchial 4 to the posterior face with the rest of the muscle. Posterodorsomedially, of the ventral process of hypobranchial 3. The a bundle passes upward to grade into the posterior anterodorsolateral fibers are continuous with the fibers of the levator externus IV. anteromedial fibers of obliquus ventralis III. NUMBER 155 53

RECTUS COMMUNIS (Figure 151: RXOMM.).—The fibers almost intermingle with the anterior fibers muscle arises from the posterior face of the ventro- of the lateral bundle of the sphincter oesophagi. hyal. It passes posteriorly as a tendon, and develops The main mass of the fibers passes laterally and fibers at the level of hypobranchial 1, where it splits then curves anteriorly. Dorsolaterally, a strong into left and right halves. The fibers then pass pos- aponeurosis attaches to the medial face of levator teriorly, and grade into tendon at the level of internus II. The ventrolateral fibers attach to the ceratobranchial 4. The tendon inserts on the ven- dorsal faces of epibranchials 1 and 2 and infra- trolateral face of ceratobranchial 5. pharyngobranchial 1. The more anterior fibers pass medially to a middorsal raphe, and are not at- tached to the skull. Posteriorly, the muscle gives DORSAL BRANCHIAL MUSCLES rise to a bundle that passes laterally into the pos- LEVATORES EXTERN I I-IV (Figure 150: L.EXT.).— terodorsal wall of the opercular cavity. Transversus These muscles originate from the ventral faces of III is a well-developed muscle, which passes across the prootic and pterotic beneath the hyomandib- the dorsal midline between the third and fourth ular fossa. Levator I inserts on the dorsolateral epibranchials of either side. It lies ventral to trans- face of epibranchial 1, with a few posterior fibers versus II, which covers it both anteriorly and pos- passing to the anteroventral face of epibranchial teriorly. 2. Levator II inserts on the dorsolateral faces of RETRACTOR DORSALIS (Figure 152: D.RETR.).— epibranchials 2 and 3, and apparently represents Origin is from the lateral and ventrolateral faces fusion with the third levator. Levator IV is well of vertebrae 1 and 2. The fibers course antero- developed. It inserts along the dorsal length of ventrally, fuse briefly with their antimeres in the epibranchial 4, and continues in the wall of the midline, and insert on the posteromedial faces of opercular cavity behind the bone, reaching the infrapharyngobranchials 2 and 3. It is overlain dorsomedial face of the cleithrum. dorsally by a small bundle of the sphincter oesoph- LEVATORES INTERNI II, HI (Figure 152: L.INT.).— agi. These muscles have altered somewhat from the ADDUCTORES I-V (Figures 150, 151, 152: AD.).— usual condition. Levator II originates from the Adductors I—III all connect the anteromedial faces posteroventromedial face of the pterotic. It passes of ceratobranchials 1-3 to the anterolateral faces anteroventrally to insert partly on the dorsomedial of epibranchials 1-3. Adductor IV connects the face of infrapharyngobranchial 3, and partly on medial faces of ceratobranchial 4 and epibranchial the dorsolateral face of infrapharyngobranchial 2. 4. Adductor V passes from the lateral region of The medial face of the muscle shares a strong ceratobranchial 5 to the posterior face of epibran- aponeurosis with the dorsolateral fibers of trans- chial 4. versus dorsalis HI. Behind this, the muscle passes SPHINCTER OESOPHAGI (Figures 150, 151, 152: medial to levator III to its origin. Levator III s.o.).—The muscle has a very complicated form. originates from the ventrolateral sphenotic beneath Ventrally, two longitudinal bundles arise from the the hyomandibular fossa. It passes ventrally to posterior face of basibranchial 3. They fuse at the insert on the anterodorsolateral face of infrapha- level of ceratobranchial 5, and form a longitudinal ryngobranchials 2 and 3. sheath on the floor of the esophagus. Beneath this, OBLIQUUS DORSALIS HI (Figure 152: OBL.D.).—A a thick, almost transverse bundle connects the fairly well-developed muscle, which passes antero- medial faces of the fifth ceratobranchials. Ventral medially from the dorsomedial faces of epibran- to the latter bundle is a thin sheet of fibers, also chials 3 and 4 to the posterodorsal face of between the fifth ceratobranchials posteriorly, but infrapharyngobranchial 2. becoming longitudinally oriented and fading out OBLIQUUS POSTERIOR (Figure 150: OBL.P.).—The above transversus ventralis V. Dorsally, three muscle appears to be present, but is not clearly bundles pass anteriorly beneath the infrapharyngo- separated from the sphincter oesophagi. branchials, attaching to the ventromedial faces of TRANSVERSI DORSALES II, HI (Figure 152: TR.D.).— infrapharyngobranchials 1-3. They fade out poste- In transversus II, the fibers arise from the ventro- riorly in the tissue of the esophagus. The muscle lateral face of the parasphenoid. The posterior passes across the midline below the retractor dor- 54 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY salis posteriorly, except for a small bundle of fibers pass from the posteroventral face of the fibers lying above and in front of the retractor. pterotic to the extreme dorsolateral tip of the clei- thrum medial to the supraclei thrum.

MUSCLES OF THE PECTORAL REGION MUSCLES OF THE DORSAL FIN ABDUCTOR SUPERFICIALIS (Figures 146, 150: ABD. s.). — The muscle originates from the posterior, INCLINATORES DORS ALES (Figure 154: INC.).—The medial, and lateral faces of the cleithrum. It passes muscles are represented by a flat sheet of transverse posteriorly to insert on the lateral faces of the fibers. They have lost their association with the anterodorsal processes of the bases of the principal fin rays, and now form the dorsal part of the fin rays. transverse cutaneous system. ABDUCTOR PROFUNDUS (Figure 150: ABD.P.).—The ERECTORES DORSALES (Figure 154: EREC).—The muscle originates from the posteroventral clei- erectors originate from the lateral faces of the thrum and lateral faces of the coracoid and acti- pterygiophores and neural spines, and the dorso- nosts. It inserts on the ventrolateral bases of the lateral faces of the sixth to thirteenth centra. They principal rays and the lateral base of the lateral insert on the anterolateral bases of all the fin rays half of the vestigial ray. (including the small first ray). ARRECTOR VENTRALIS (Figure 146: ARR.V.).—Ori- DEPRESSORES DORSALES (Figure 154: DEPR. ).— gin is from the ventrolateral face of the dorsal These muscles arise in the same area as the erectors, region of the cleithrum. The muscle inserts on the and insert on the posterolateral bases of all the anterolateral base of the medial half of the vesti- fin rays. gial fin ray. SUPRACARINALIS ANTERIOR (Figures 146, 154: s. ADDUCTOR SUPERFICIALIS (Figure 153: ADD.S.).— ANT.).—The muscle is well developed, and arises The muscle originates from the dorsomedial clei- from the posterior faces of the supraoccipital and thrum. The fibers to the dorsalmost rays pass poste- epiotic. It passes posteriorly to attach to the antero- riorly, the slip to each succeeding ray passing lateral face of the first dorsal pterygiophore. increasingly ventromedially. Thus, the slip to the SUPRACARINALIS POSTERIOR (Figures 154, 155: s. ventralmost ray passes ventrally, and lies medial to POST.).—The fibers arise from the posterodorsal all the others. The ventral six rays have dorsal face of the last dorsal fin pterygiophore, and attach processes, to which the tendons attach, the more to the dorsolateral faces of the neural spines of the dorsal tendons inserting on the dorsomedial faces next three vertebrae to end on the neural spine of of the rays distal to their bases. the seventeenth vertebra. ADDUCTOR PROFUNDUS (Figure 153: ADD.P.).—The muscle originates from the medial faces of the coracoid and the radials, and inserts on the postero- MUSCLES OF THE ANAL FIN medial flanges of all but the six ventral rays, where insertion is on the ventromedial shaft of the ray. INCLINATORES ANALES (Figures 146, 154: INC.).— ARRECTOR DORSALIS (Figure 153: ARR.D.).—The The residual inclinator arises from the postero- muscle is small, arising from the medial face of ventral tip of the postcleithrum and attaches to the scapula, and inserting on the medial face of the lateroventral face of the first pterygiophore of the medial half of the vestigial fin ray. the anal fin. The first inclinator of the fin is well CORACORADIALIS (Figure 153: COR.R.).—The mus- developed, but all the others are poorly so, and do cle is poorly developed. It passes from the dorso- not insert on the fin rays, but on the cartilagenous medial face of the posterior coracoid process to the distal pterygiophores. These muscles arise from the posteroventral face of the fourth radial. fascia over the obliquus inferioris. PROTRACTOR PECTORALIS (Figure 146: P.P.).—From ERECTORES ANALES (Figure 154: EREC).—The its origin on the ventrolateral pterotic, the muscle muscles originate from the lateral faces of the passes ventrally to the anteromedial face of the pterygiophores and the haemal arches and spines supracleithrum and the posterodorsal cleithrum. of the ninth to thirteenth vertebrae, inserting on LEVATOR PECTORALIS (Figure 146: TR.).—The the anterolateral bases of all the fin rays. NUMBER 155 55

DEPRESSORES ANALES (Figure 154: DEPR.).—The trum and haemal spine of the penultimate vertebra. depressors arise just behind the erectors to any It inserts on the bases of rays V 3 and 4. particular ray, but insert on the posterolateral FLEXOR VENTRALIS EXTERNUS (Figure 155b: F.V. bases of the rays. E.).—The fibers arise from the ventrolateral faces INFRACARINALIS ANTERIOR.—The muscle is ap- of the centra and haemal spines of the second and parently represented by fibers arising from the third last vertebrae and insert on the bases of rays ventral tip of the cleithrum and passing posteriorly V 1 and 2. in the subdermal tissue to intermingle with those FLEXOR VENTRALIS INFERIOR (Figure 155: F.V.L).— of the infracarinalis medius, helping to form the Origin is from the ventrolateral faces of the centra longitudinal cutaneous system. of the thirteenth to fifteenth vertebrae. The fibers INFRACARINALIS MEDIUS (Figures 146, 154: INF. give rise to a tendon which meets its antimere M.).—It seems likely, as discussed above, that this before inserting on the base of ray V 5. muscle forms part of the longitudinal cutaneous TRANSVERSUS CAUDALIS (Figure 155fc: TR.C).—The system. It arises from the first pterygiophore of the muscle arises from the ventral face of the epural anal fin and passes anteriorly into the subdermal and the lateral surface of the upper part of the tissue. hypural plate. It passes posteroventrally as a thin INFRACARINALIS POSTERIOR (Figures 154, 155: INF. tendon to insert on the dorsolateral face of ray P.).—The muscle arises from the posteroventral VS. face of the last anal pterygiophore and its posterior process. Fibers attach to the ventrolateral faces of LATERAL BODY MUSCLES the haemal spines of the fourteenth to sixteenth vertebrae, ending on the spine of the seventeenth EPAXIALIS (Figures 146, 154, 155a: EPAX.).— vertebra. Anteriorly, the muscle is attached to the posterior faces of the supraoccipital, exoccipital, pterotic, epiotic, frontal, and supracleithrum. Fibers also MUSCLES OF THE CAUDAL FIN arise from the lateral faces of the neural spines INTERRADIALIS (Figure 155a: INT.).—The muscle and centra of the vertebrae. The dorsal moiety is passes toward the midline from the lateral faces well separated from the main mass at about the of the caudal rays, a little distal to their bases. The level of the dorsal fin. It passes posteriorly, becom- fibers are superficially continuous. ing increasingly aponeurotic, and eventually forms HYPOCHORDAL LONGITUDINALIS (Figure 155: a flat, sheetlike tendon which joins a similar ten- H.L.).—This muscle originates from the dorsal sur- don from the obliquus inferioris to insert on the face of the ridged parhypural and the lateral faces lateral bases of all the caudal fin rays. The anterior of the ventral fused hypurals, as well as from the region covers the anterior part of the obliquus ural centrum. It inserts on the ventrolateral faces superioris. Fiber direction superficially tends to of rays D 3 and 4. be from anteroventral to posterodorsal. A few of FLEXOR DORSALIS (Figure 155&: F.D.).—Origin is the ventromedial fibers attach to the peritoneum from the dorsolateral faces of the third and fourth overlying the swim bladder near the vertebral cen- last vertebral centra and the anterolateral face tra. Posteriorly, the fibers grade into a flat aponeu- of the epural. The tendons insert on the antero- rosis which inserts on the lateral bases of rays D lateral bases of rays D 1-4. 1-4. FLEXOR DORSALIS SUPERIOR (Figure 155: F.D.S.).— OBLIQUUS SUPERIORIS (Figures 146, 154, 155a: A well-developed muscle, originating from the dor- OBL.S.).—The muscle arises from the posterodorsal solateral faces of the thirteenth to fifteenth verte- tip of the cleithrum. It broadens out rapidly, bral centra. The fibers give rise to a tendon which becoming closely associated with the epaxialis dor- joins its antimere before inserting on the antero- sally and the vertebral centra and haemal spines dorsal base of D 5. medially. Posteriorly, it grades into an aponeurosis FLEXOR VENTRALIS (Figure 1556: F.V.). — The which inserts on the lateral bases of rays V 1-4. muscle originates from the ventrolateral face of OBLIQUUS INFERIORIS (Figures 146, 154, 155a; the hypural plate, the ural centrum, and the cen- OBL.1.).—The anteroventral section consists of two 56 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY well-separated parts. The dorsal one connects the the dorsal region of the prefrontal to the lips, where anteroventral face of the postcleithrum to the it is joined by another bundle arising from the posterior and ventral faces of the coracoid. Be- anteroventral corner of the eye. neath this, the ventral part passes back from the The longitudinal system is best developed ven- posteroventral cleithrum and coracoid to attach to trally, between the anus and the lower jaw. This the medial face of the postcleithurm via the third part of the system presumably arose from the fibers myocomma, the more posterior fibers being contin- of the infracarinales anterior and medius after the uous with the anterodorsal part. The anterodorsal loss of the pelvis, while the transverse system may portion arises from the fascia overlying the epaxialis, have been initially derived from the inclinators of being partly overlain itself by the superficialis the dorsal and anal fins. There is an enigmatic lateralis. The fibers attach to the posterodorsal bundle (present in this species only of those exam- face of the postcleithrum, the more medial ones ined) which arises from the peritoneum overlying being continuous with those of the anteroventral the swim bladder and passes anterolaterally be- section. Posteriorly, the fibers become increasingly tween the posterodorsal tip of the cleithrum and horizontally oriented, most of them attaching to the the anterior margin of the obliquus inferioris. It ventrolateral face of the first anal pterygiophore. grades into the connective tissue overlying the Some of the posterior fibers join to a myocomma of dilatator operculi. the deeper section of the obliquus inferioris. This deeper part is continuous anteriorly with the fibers Amblyrhynchotes honckenii (Bloch) of the infracarinalis medius, and now forms the FIGURE 156 longitudinal cutaneous system. A small bundle of A lo completely covers A 2o posterodorsolateral- fibers passes posterodorsomedially from this part of ly. The ventral bundle of A 13 is well developed, and the muscle. It soon grades into a tendon which reaches as far as the rear of the orbit posteriorly. courses lateral to the erector-depressor complex of The separation of the two bundles of A 2a is the anal fin to attach to the anteroventral tip of more distinct, and A 23 is not as well developed. the haemal spine of the twelfth centrum. The anterolateral bundle of the dilatator operculi inserts on the anterodorsal face of the opercle and CUTANEOUS MUSCULATURE subopercle. The pharyngoclavicularis externus consists of two This system, which lies immediately below the parts, the anterior inserting on basibranchial 3 skin, is very complicated and difficult to follow. (where it is partly confluent with the pharyngo- These difficulties arise largely because of the clavicularis internus), the posterior part inserting numerous elastic connective tissue fibers which in- on the anteroventral face of ceratobranchial 5. The termingle with the muscle fibers. Owing to the obliquus posterior is well developed, and distinct difficulty in following the system, only cursory from the sphincter oesophagi. The inclinator analis attention will be paid to it. The fibers pass in two to the first ray now inserts on the cartilagenous main directions, transverse and longitudinal, of distal pterygiophore and not on the base of the which the transverse is the more superficial. The first ray. The infracarinalis medius is somewhat fibers are concentrated in the regions where the distinct. It attaches anteriorly to the posteroventral spines or prickles are best developed—mainly along tip of the cleithrum, and becomes continuous with the stomach and back. The central bases of the the longitudinal cutaneous in the midregion. It spines are closely associated with the muscle separates from the cutaneous again posteriorly to fibers, but histological sections are necessary to attach to the first anal pterygiophore. The hypo- determine whether the fibers actually insert on chordal longitudinalis inserts on rays D 2-4, while these bases. Transverse fibers are also present be- the transversus caudalis inserts on rays V 2-3. hind the dorsal and anal fins, across the midline of the skull, and beneath the eye. The ring of fibers encircling the mouth in the tissues of the Colomesus psittacus (Bloch and Schneider) lips may also be derived from this system. A well- FICUIUE 157 developed bundle of fibers passes anteriorly from A la and A 13 are as for A. honckenii. A 2a, NUMBER 155 57

which is not particularly well separated into two face of the pterotic. A 2P is as for Colomesus, ex- heads, has a high dorsal origin from the ventral cept that the frontal secondarily roofs the dilatator surface of the lateral process of the hyomandibular, fossa, and some of the fibers arise from it. The and has increased in bulk. A 2P has expanded levator arcus palatini inserts on the hyomandibular across what was the dilatator fossa, and extends its by a broad, flat aponeurosis. origin posterodorsally onto the sphenotic, the The hyohyoidei adductores is poorly developed posterior tip of the frontal, the supraoccipital, in the outer wall of the opercular cavity, but with epiotic, and pterotic. The dilatator operculi is well the posterodorsal and posteroventral bundles still developed, and inserts on the anterodorsal face of well developed near their origins. The halves of the subopercle, posterodorsal interopercle, and the the sternohyoideus fuse in the anterior midline. A opercle. Origin has greatly expanded to include well-developed and distinct intermandibularis is the dorsal faces of the prefrontal, frontal, supra- present. It attaches in the fossa of the dentary below occipital, epiotic, and pterotic. The anterior fibers the insertion of A 2, lying anterodorsal to the cover the posterodorsomedial fibers of A 2P. The angular. hyomandibular is not included in the sites of ori- Levatores externi II and III are fused, and leva- gin. The levator arcus palatini is very small. tor internus II is not associated with transversus There is no section of the hyohyoidei abductores dorsalis II medially. Adductor I is small. The to the fourth ray. The posteroventral bundle of erectores dorsales and anales lie in a somewhat the semicircular portion of the hyohyoidei adduc- more lateral position than do the depressores dor- tores attaches to the ventrolateral face of the fourth sales and anales. Inclinatores anales are attached to radial. Two slips of fibers pass inward to the the bases of the first and the twenty-first (twentieth medial face of the posterodorsal bundle. One passes in the 124 mm specimen) to twenty-eighth fin rays. dorsomedially to attach to the posteromedial face The obliquus inferioris does not separate from of epibranchial 3, the other passes posteroventrally the obliquus superioris in the region of the anal to fade out lateral to the pharyngoclavicularis in- fin, and there is thus no anteroventral section ternus. Both lie lateral to the posterior fibers of joining the fibers arising from the posteroventral levator externus IV. tip of the postcleithrum. The posterodorsomedial There are no fibers from levator externus I to bundle of fibers attaching to the haemal spine of epibranchial 2, and only a few from levator II to the twelfth vertebra is absent. The extreme antero- the fascia between epibranchials 2 and 3. No other ventral fibers of the obliquus inferioris attach to fibers attach to epibranchial 3. Levator externus the lateral face of the first branchiostegal ray. A IV is well developed. No lateral bundle of trans- small aponeurosis from the obliquus superioris versus dorsalis II was found, but it may be rep- attaches to the base of ray D 1. The hypochordal resented by a dorsomedial bundle which passes out longitudinalis and flexor ventralis externus are as laterally to join the posterodorsal part of the for Colomesus, and the transversus caudalis inserts hyohyoidei adductores. on the bases of rays V 1-2. The hypochordal longitudinalis inserts on rays D 2-4, while the transversus caudalis inserts on rays Arothron immaculatus (Bloch) V 2-3. The flexor ventralis externus inserts on the

bases of rays V 1-3, and its origin is expanded FIGURE 159 distally onto the haemal spines. The posteroventral part of A IP is expanded along the floor of the orbit, and A la covers A Xenopterus naritus (Richardson) 2a laterally. A 2P is well developed, and A 3 is much reduced (particularly dorsally). The anterior FIGURE 158 part of the dilatator operculi inserts on the sub- The posteroventral portion of A IP is expanded opercle and opercle. dorsally and originates from the roof of the orbit, The posterodorsal bundle of the hyohyoidei ad- while the dorsal head of A 2a extends posteriorly ductores attaches to the anteroventral face of the beyond the hyomandibular to the anterolateral supracleithrum. 58 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY The fibers of levator externus I insert only on The flexor ventralis externus inserts only on V 1, epibranchial 1, and adductores II and V are absent. while the flexor ventralis attaches to V 2-5. The The residual inclinator analis has fused with the transversus caudalis inserts on V 2. obliquus inferioris, and the first and last anal fin rays have inclinators attaching to their bases. The hypochordal longitudinalis is as for A. Summary of Tetraodontidae honckenii. The single tendon of the transversus A number of regions of the body in this family caudalis attaches to the base of ray V 2. The flexor show myological variation. The subdivisions of the ventralis attaches to the bases of rays V 1-5. adductor mandibulae, while constant in presence The posterodorsomedial bundle of the obliquus and position, show a considerable variation in rela- inferioris fades into an aponeurotic sheet at the tive development. This is particularly true of the level of the anal fin, the sheet attaching to the hae- A 1 sections, and of A 20. The size and sites of mal arches of the thirteenth and fourteenth verte- origin of the dilatator operculi differ markedly, as brae. does the insertion of the anterolateral bundle of fibers of this muscle. The lateral fibers of the hyohyoidei adductores vary in size and position. Tetraodon schoutedeni Pellegrin The insertion of the levatores externi shows con- siderable variation, and there are other minor dif- FIGURE 160 ferences in the dorsal branchial musculature. The A 13 is well developed, and similar to that of number and form of the inclinatores anales differ, Xenopterus, the posteroventral portion having ex- and there is some variation in the relative develop- panded posterodorsally along the roof of the orbit. ment of the posterodorsolateral bundle of fibers of A 2a is reduced dorsally, and there is only a thin the obliquus inferioris. The number of insertion slip of muscle to the posterodorsolateral face of the tendons (and their relative positions) of the hypo- hyomandibular beneath the dorsal process of the chordal longitudinalis, flexor ventralis, flexor ven- opercle. The adductor operculi is very well develop- tralis externus, and transversus caudalis varies ed. A lateral bundle of the sphincter oesophagi at- widely. The remaining muscles are relatively con- taches to the anteroventral face of epibranchial 4. servative in the members of the family examined.

Myological Descriptions of Representative Canthigasterids

The typical body outline of this monogeneric Canthigaster margaritatus (Riippell) family is illustrated in Figure 9. Fraser-Brunner (1943) recognizes a single genus with 15 species in FIGURES 161-167 this family. Representatives of two of the species were dissected for this study. MUSCLES OF THE CHEEK The head region is higher, narrower, and of less robust construction than in the tetraodontids. ADDUCTOR MANDIBULAE (Figures 161, 162: A 1 a, Other bodily proportions are similar, and some in- A IP, A 2a, A 2P, A 3).—A la inserts on the flation is possible. These fishes can also raise the posteroventral face of the maxilla. The ventral fibers arise from the posterodorsolateral preopercle. skin on the back into a ridge by contraction of the The dorsal region overlies A 2a (with some fibers underlying cutaneous musculature. They are arising from the fascia between them), and then usually brightly colored coral reef forms, feeding passes medial to that muscle to originate from the mainly on sea grasses and soft-bodied marine dorsal metapterygoid. A IP originates from the invertebrates. ethmoid and the ventral surface of the prefrontal. As outlined above, they are generally considered A ventral extension of this muscle arises from the to be closely allied to the Tetraodontidae. lateral face of the parasphenoid anterior to the NUMBER 155 59 orbit. The fibers insert on the posterodorsal face PROTRACTOR HYOIDEI (Figure 161: PR.HY.).—The of the maxilla, medial to A la. A 2a originates from muscle is well developed, connecting the anterior the lateral preopercle, metapterygoid, quadrate, and elements of the hyoid arch to the posteroventral hyomandibular. The muscle becomes tendinous an- face of the dentary near the symphysis. The mus- teriorly, and joins the lateral face of the tendon cle is single anteriorly, and diverges into two halves of A 23. A 2P arises from the anterior faces of before attaching to the lateral faces of the antero- the hyomandibular and prootic and the lateral hyal and ventral branchiostegal rays. face of the posterior region of the parasphenoid. HYOHYOIDEI ABDUCTORES (Figure 161: H.AB.).— The joint tendon with A 2a inserts on the medial The form of the muscle is very similar to that face of the dentary. A 3 originates from the lateral described for Lagocephalus, with the following faces of the quadrate, ectopterygoid, and meta- differences. The large section to the first ray is pterygoid, and inserts tendinously in the Meckelian faintly subdivided into ventrolateral and dorso- fossa. medial bundles. A small bundle attaches to the LEVATOR ARCUS PALATINI (Figures 161, 162: posteromedial face of basibranchial 1, just in front L.A.P. ).—A minute muscle, originating from the of the section which fuses with its antimere in the ventral face of the postorbital process of the frontal midline beneath hypobranchial 3. and inserting on the posterodorsolateral ridge of HYOHYOIDEI ADDUCTORES (Figure 161: H.AD.).— the hyomandibular. The form of this muscle is also very similar to DILATATOR OPERCULI (Figure 161: D.O.).—Origin that described for Lagocephalus, but differs in the is from the posterolateral frontal (primarily in following respects. The posterodorsal bundle the gutter formed by the vertical lateral ridge of attaches to the anteroventral face of the supraclei- this bone behind the orbit) and the anterior epi- thrum. The posterior wall of the branchial (oper- otic. The fibers insert on the posterodorsal face cular) chamber is formed by fibers from the of the interopercle, the anterior opercle, and the posterolateral face of epibranchial 3, which pass out dorsal process of the latter bone. laterally in front of epibranchial 4, and a ventral LEVATOR OPERCULI (Figure 161: L.O.).—Origin bundle from the ventromedial face of the clei- is from the anterolateral face of the lateral process thrum. of the pterotic, the fibers inserting on the pos- STERNOHYOIDEUS (Figures 161, 163: STH.).—The terodorsal face of the opercle. muscle is cone-shaped, with a fairly wide origin ADDUCTOR ARCUS PALATINI (Figure 162: A.A.P.).— on the anteroventral cleithrum and an aponeurotic The muscle originates from the lateral face of insertion on the ventrohyal. The halves fuse in the parasphenoid and passes laterally to insert on the anterior midline. the dorsomedial faces of the hyomandibular, meta- pterygoid, and ectopterygoid. Anteriorly, the fibers VENTRAL BRANCHIAL MUSCLES cross the midline beneath the anterior part of the parasphenoid and the vomer, attaching to the PHARYNGOCLAVICULARIS EXTERNUS (Figures 163, medial faces of the palatines. 164: PHC.E.).—Origin is from the ventromedial ADDUCTOR OPERCULI (Figure 163: AD.OP.).—The face of the cleithrum, the muscle dividing into muscle arises from the posteroventrolateral face three bundles. The anterior one inserts on the of the pterotic and inserts on the posterodorso- junction between basibranchial 2 and hypobran- medial face of the opercle. chial 2, the middle part inserts on basibranchial 3 and the posterior section attaches to the ventro- medial face of ceratobranchial 5. MUSCLES OF THE HYOID REGION PHARYNGOCLAVICULARIS INTERNUS (Figures 163, 164: PHC.L).—The muscle arises from the antero- VALVULUS.—The muscle originates from the an- ventromedial cleithrum. It passes anterodorsally teroventromedial face of the preopercle, and sweeps to insert on basibranchial 3 and ceratobranchial forward in a dorsoventral curve to grade into the 5. In between the two points of insertion, the tissues of the mandibular buccal valve. It lies fibers sweep across the midline to join their anti- medial to the interopercle. meres. 60 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

OBLIQUUS VENTRALIS HI (Figures 163, 164: OBL. from the dorsolateral prootic and inserts on the v.).—The fibers pass between the anteroventral anterolateral face of infrapharyngobranchial 3 and face of ceratobranchial 3 and the posteroventro- the dorsolateral edge of infrapharyngobranchial 2. lateral tip of the ventral process of hypobranchial 3. Levator II passes medial to levator III, and forms TRANSVERSUS VENTRALIS v (Figure 164: TR.V.).— an aponeurotic connection with transversus dorsalis This is a small muscle which crosses the ventral II. midline between the anteroventromedial faces of OBLIQUUS DORSALIS IV (Figure 165: OBL.D.).—Ori- the fifth ceratobranchials. There is no transversus gin is from the posterior and dorsal faces of epibran- IV, but it may have fused with the fibers of the chial 4, the fibers passing anteromedially to insert pharyngoclavicularis internus to form the trans- on the posterodorsal face of infrapharyngobran- verse part. chial 2. RECTI VENTRALES I-IV (Figures 163, 164: RECT. OBLIQUUS POSTERIOR.—The fibers of this muscle v.).—The first muscle consists of two parts. The are not well separated from the sphincter oesoph- more medial section links the dorsomedial face agi, but pass from the dorsal tip of ceratobran- of the anterohyal with the anteroventral face of chial 5 to the posteromedial face of epibranchial 4. hypobranchial 1. The lateral portion arises from TRANSVERSI DORSALES II, iv (Figure 165: TR.D.).— the dorsomedial surface of the anterohyal and The anterior muscle arises from the midline and sweeps dorsally to attach to the anteroventrolateral attaches to the connective tissue covering infra- flange of ceratobranchial 1. Recti II and III inter- pharyngobranchial 2 (there is no direct connec- connect the hypobranchials of the first and second, tion) and to the medial face of levator internus and second and third arches, respectively. Rectus II. Transversus IV crosses the dorsal midline di- IV passes from the anteroventral face of cerato- rectly beneath the posterior portion of transversus branchial 4 to the posteroventral tip of the ventral II, connecting the medial faces of the fourth epi- process of hypobranchial 3. branchials. RECTUS COMMUNIS (Figures 163, 164: R.COMM.).— RETRACTOR DORSALIS (Figure 165: D.RETR.).—The The muscle arises tendinously from the posterior muscle originates from the lateral face of the face of the ventrohyal. It passes ventral to basi- exoccipital and the ventrolateral face of the branchial 1 (lying medial to the basibranchial- basioccipital. It inserts on the posterodorsal face hypobranchial ligament), where is develops a belly of infrapharyngobranchial 2. of muscle fibers, before grading into a tendon ADDUCTORES I-V (Figures 163, 164, 165: AD.).— which attaches to the anteroventrolateral face of The muscles pass between the anteromedial faces ceratobranchial 5. of the ceratobranchials and epibranchials of the first three arches, the medial faces of these bones of the fourth arch, and from the dorsal tip of cer- DORSAL BRANCHIAL MUSCLES atobranchial 5 to the posterolateral face of cerato- LEVATORES EXTERNI i-rv (Figure 163: L.EXT.).— branchial 4. The muscles originate from the prootic and ptero- SPHINCTER OESOPHAGI (Figures 164, 165: s.o.). tic beneath the hyomandibular fossa. The first Dorsally, the muscle consists of a bundle of fibers muscle inserts on the dorsolateral face of epibran- across the midline, joined by a bundle from epi- chial 1, while the second inserts on the dorsolateral branchial 4. In addition, the muscle sends out a faces of epibranchials 2 and 3. There is no distinct bundle which attaches to the anterior face of cer- third levator. Levator IV is well developed, and atobranchial 4. Ventrally, there is a bundle across inserts along the entire dorsal surface of epibran- the midline behind transversus IV. Immediately chial 4. It continues posteriorly in the wall of above this, the fibers are transverse posteriorly, but the opercular cavity, the fibers attaching to the become increasingly longitudinally oriented ante- dorsomedial face of the clei thrum. riorly. This gives rise to a pair of fiber bundles LEVATORES INTERNI II, HI (Figure 165: L.INT.).— passing forward beneath the basibranchials to at- Levator II originates from the posteroventrolateral tach to the posteroventral face of basibranchial 1. face of the pterotic, and inserts on the anterodorsal The anterior section of the pharyngoclavicularis face of infrapharyngobranchial 2. Levator III arises internus inserts between the halves. NUMBER 155 61

MUSCLES OF THE PECTORAL REGION fin rays, arising in the subdermal fascia and insert- ing on the posterolateral bases of the rays. ABDUCTOR SUPERFICIALIS (Figure 161: ABD.S.).— ERECTORES DORSALES (Figures 161, 166: EREC. ).— Origin is from the posterior face of the lateral The muscles originate from the lateral faces of flange of the cleithrum, the muscle inserting on the pterygiophores and the neural spines of the the anterolateral bases of the principal fin rays. seventh to tenth vertebrae. They insert on the ABDUCTOR PROFUNDUS.—The fibers arise from the anterolateral bases of the rays, partially overlying lateral face of the coracoid and the ventrolateral the depressors anteriorly, and completely covering cleithrum. Insertion is on the posterolateral bases them posteriorly. of the principal fin rays and on the posterolateral DEPRESSORES DORSALES (Figures 161, 166: DEPR.).— face of the lateral half of the vestigial fin ray. These muscles originate behind and medial to the ARRECTOR VENTRALIS (Figure 161: ARR.V.).—This respective erectors, and insert on the posterolateral is a small muscle, originating from the postero- bases of the fin rays. lateral face of the cleithrum beneath the lateral SUPRACARINALIS ANTERIOR (Figures 161, 166: s. flange. It inserts on the lateral extension of the ANT.).—Origin is from the supraoccipital and its base of the medial half of the vestigial fin ray. spine and the exoccipital. The lateral fibers pass ADDUCTOR SUPERFICIAL^ (Figure 153 for Lago- posteroventrally to attach to the dorsomedial faces cephalus, ADD.S.).—Origin is from the dorsomedial of the bifid neural spines of the first four vertebrae. face of the cleithrum, the muscle inserting on the The dorsomedial fibers pass posteriorly to the an- dorsomedial bases of the ventral six rays, and the terolateral face of the first pterygiophore of the dorsomedial faces of the next nine rays some dis- dorsal fin. tance distal to their bases. SUPRACARINALIS POSTERIOR (Figures 166, 167: s. ADDUCTOR PROFUNDUS (as above, but ADD.P.).— POST.).—The fibers arise from the last pterygio- The fibers arise from the medial faces of the clei- phore of the dorsal fin and end on the anterodorsal thrum, coracoid, and scapula. Insertion is on the face of the neural spine of the penultimate verte- ventromedial bases of the rays. There is no aponeu- bra. They attach to the tips of the neural spines of rosis anteriorly on the body of the muscle as there the intervening vertebrae. is in Lagocephalus. ARRECTOR DORSALIS.—The muscle originates from the cleithrum beneath the adductor profundus, MUSCLES OF THE ANAL FIN and inserts on the base of the medial half of the INCLINATORES ANALES (Figures 161, 166: INC.).— vestigial fin ray. There are four of these muscles. The first residual CORACORADIALIS.—The muscle is better devel- one arises in the fascia overlying the obliquus in- oped than in the tetraodontids examined. It passes ferioris and attaches to the anteroventral tip of from the posterior process of the coracoid to the the first anal pterygiophore. The first and last two distal ventral tip of the fourth radial. fin rays receive inclinators inserting on their lateral PROTRACTOR PECTORALIS (Figures 161, 163: bases, the fibers arising in the fascia over the ven- p.p.).—Origin is from the posterolateral face of tral body muscles. the pterotic. The fibers fan out to insert broadly ERECTORES ANALES (Figure 166: EREC).—The on the dorsal face of the cleithrum and anteroven- muscles originate from the anterolateral faces of tral supracleithrum. the anal pterygiophores, anterolateral to the origins LEVATOR PECTORALIS.—The muscle arises from of the depressors, and insert on the anterolateral the pterotic behind the protractor pectoralis and bases of the fin rays. passes posteroventrally to insert on the dorsomedial DEPRESSORES ANALES (Figures 161, 166: DEPR.).— face of the cleithrum beneath the supracleithrum. The fibers arise posteromedial to those of the erec- tors, and insert on the posterolateral bases of the rays. MUSCLES OF THE DORSAL FIN INFRACARINALIS ANTERIOR.—The muscle is ap- INCLINATORES DORSALES (Figure 166: INC.).— parently represented by a bundle of fibers arising There are two inclinators to the two posteriormost from the ventral tip of the cleithrum, and passing 62 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY into the subdermal tissues as part of the longi- FLEXOR VENTRALIS EXTERNUS (Figure 1676: tudinal cutaneous. F.V.E.).—The fibers arise from the lateral face of INFRACARINALIS MEDIUS (Figures 161, 166: INF. the haemal spine of the penultimate vertebra, and M.).—The muscle arises from the anteroventral insert on the lateral bases of rays V 1-2. face of the first anal pterygiophore and grades pos- FLEXOR VENTRALIS INFERIOR (Figure 167: F.V.I. ).— teriorly into the longitudinal cutaneous together Origin is from the lateral faces of the haemal spines with the anteroventral bundle of the obliquus in- of the second to fourth last vertebrae, and the cen- ferioris. Its posterodorsal surface gives rise to a tra of the third and fourth last vertebrae. The mus- posterodorsomedial bundle of fibers attaching to cle inserts on the anterolateral base of ray V 6. the anterior face of the haemal spine of the twelfth TRANSVERSUS CAUDALIS (Figure 1676: TR.C).— vertebra. This bundle is probably derived from The fibers arise from the dorsolateral face of the the obliquus inferioris, which (in tetraodontids hypural plate, and insert on the anterodorsal faces and triodontids) gives rise to a similar bundle. of rays V 1 and 2. INFRACARINALIS POSTERIOR (Figures 166, 167: INF. P.).—The fibers arise from the ventrolateral face LATERAL BODY MUSCLES of the last anal pterygiophore, and insert on the anteroventral face of the haemal spine of the pe- EPAXIALIS (Figures 161, 167a: EPAX.J.—The nultimate vertebra. They contact the tips of the fibers of this muscle attach variously to the rear haemal spines of the intervening vertebrae. of the skull, the centra and neural spines, and the dorsal caudal fin rays. In the posterior region, the dorsal moiety becomes inclined at an increas- MUSCLES OF THE CAUDAL FIN ingly oblique angle to the main section, and gives INTERRADIALIS (Figure 167a: INT.).—The muscle rise to an aponeurotic sheet which, together with stretches between the fin rays, each ray receiving a a similar sheet from the obliquus inferioris, forms more or less distinct slip. a superficial aponeurosis to all the caudal fin rays. HYPOCHORDAL LONGITUDINALIS (Figure 167a: The main medial portion inserts posteriorly on H.L.).—The muscle originates from the lateral face the bases of rays D 2-5. of the fused lower hypurals and inserts on the OBLIQUUS SUPERIORIS (Figures 161, 167a: OBL.S.).— bases of rays D 1—4. The muscle is small anteriorly, where it arises FLEXOR DORSALIS (Figure 1676: F.D.).—The mus- from the posterodorsal tip of the cleithrum. It cle is not well developed. It originates from the broadens out posteriorly behind the abdominal lateral faces of the neural spines of the second cavity, attaching to the centra and haemal spines and third last vertebrae, the epural, and the antero- of the caudal vertebrae. Posteriorly, it inserts on dorsal tip of the urostyle, and inserts on the the bases of rays V 1-4. anterolateral bases of rays D 1—3. OBLIQUUS INFERIORIS (Figures 161, 166, 167a: FLEXOR DORSALIS SUPERIOR (Figure 167: F.D.S.).— OBL.L).—As in tetraodontids, the muscle has become Origin is from the lateral faces of the neural spines extensively subdivided. The main mass arises and centra of the third to fifth last vertebrae. The anteriorly from the coracoid and ventrolateral fibers sweep posterodorsally, grading into a tendon cleithrum. It passes posteriorly, the more dorsal inserting on the dorsolateral base of ray D 5. fibers attaching to the anteroventral faces of the FLEXOR VENTRALIS (Figure 1676: F.V.).—The postcleithra, and then continuing posterodorsally fibers arise from the lateral faces of the second to fade aponeurotically into the fascia overlying and third last vertebrae and the ural centrum. the midlateral septum. The anterodorsal fibers of The muscle inserts on the bases of rays V 1-5. The this section sweep up to attach to the posterolateral section to ray V 5 is fairly well separated from base of the first pterygiophore of the dorsal fin. the other part, the two bundles being subequal The more ventral fibers pass beneath the tip of in size. The fibers to V 1, and some of those to the ventral postcleithrum, joining the fibers from V 2 are partially separated from the main mass the posteroventral face of that bone to attach to of the muscle, lying on the medial side and pass- the anteroventral tip of the first anal pterygiophore. ing posterodorsally instead of posteroventrally. The posterior section, which is somewhat separated NUMBER 155 63 from the obliquus superioris, grades into an dontids, the anterior longitudinal bundle of the aponeurotic sheet joining that of the dorsal moiety snout region overlying A IP and attaching to the of the epaxialis as described above. Anteroventrally, dorsal face of the palatine. this section passes beneath the posterior region of the main anterior mass, and fades out in the subdermal tissue covering the abdominal cavity, Canthigaster valentini (Bleeker) where it contributes to the formation of the longitudinal cutaneous. A small bundle passes FIGURE 168 posteroventrally to join the dorsal surface of the The ventral portion of A 23 is expanded ven- infracarinalis medius. trally to almost completely cover A 3 laterally. The obliquus posterior is well developed, with some CUTANEOUS MUSCLES slight intermingling of fibers with adductor V at its site of attachment to ceratobranchial 5. The The cutaneous muscular system is as for tetrao- hypochordal longitudinalis inserts on rays D 2-4.

Myological Descriptions of Representative Diodontids

In his review, Fraser-Brunner (1943) recognized and premaxilla. The anterolateral face of the mus- three genera and about 18 species of diodontids cle gives rise to a short tendon which passes into (see Figure 10 for body outline). Representatives the tissues of the lips at the corner of the mouth. of two of these genera have been dissected. They A 13 (Figure 170) is much reduced. It originates are much like the tetraodontids, but covered with from the posterodorsal palatine and the ethmoid- spines. The head is broad and flat, and their in- vomer, and inserts by a sheetlike aponeurosis on flation capabilities often impressive. The jaws are the posteromedial face of the maxilla (Winterbot- even better developed than in the two preceding tom, 1971a). The remaining sections are greatly families, and the diet is primarily composed of modified, primarily through the enlargement of gastropods, hermit crabs, and echinoids. They oc- A 2. The fibers of A 23 originate from the ventral cur mainly in coral reef and sandy bay environ- faces of the frontal and prefrontal, and the lateral ments, and swim almost exclusively by undulations faces of the palatine, mesopterygoid, and parasphe- of the well-developed pectoral fins. noid (the latter three sites of origin being mainly Diodontids are usually considered to be closely concerned with the posteroventral extension of the related to tetraodontids, although they have oc- muscle). In addition, the fibers meet their anti- casionally been allied with the Molidae. meres in the midline across the anteroventral re- gion of the orbit. The posteroventral fibers are continuous with the main mass of A 2 (which Diodon hystrix Linnaeus probably represents A 2a), in which a great variety FIGURES 169-175 of fiber direction is found. Fibers originate from the lateral or anterior faces of the preopercle, hyo- mandibular, quadrate, mesopterygoid, pterotic, MUSCLES OF THE CHEEK sphenotic, frontal, pterosphenoid, prootic, and ADDUCTOR MANDIBULAE (Figures 169, 170: A la, parasphenoid. The ventromedial section is partially A IP, A 2a, A 23).—A 1 originates from the entire subdivided anteriorly, and the medial portion may lateral margin of the preopercle, the fascia over represent A 3. Insertion is on the medial face of the opercle-hyomandibular articulation, and the the dentary. A

A tendon inserts on the anterolateral base of each There is a single inclinator arising from the fascia of the principal fin rays. The ventral nine rays overlying the dorsal extension of the obliquus in- possess anterodorsal flanges which overlap the base ferioris. It inserts on the lateral face of the first of the preceding ray, and the tendons attach to dorsal fin pterygiophore. these flanges. ERECTORES DORSALES (Figure 169: EREC. ).—Origin ABDUCTOR PROFUNDUS (Figure 169: ABD.P.).—The is from the dorsal faces of the ventrolateral flanges muscle originates from the ventrolateral cleithrum, and centra of the tenth to fifteenth vertebrae, and, the lateral coracoid, and the ventrolateral faces posteriorly, from the ventrolateral faces of the of the ventral radials. Insertion is on the ven- pterygiophores. The muscles insert on the antero- trolateral bases of the principal rays, there being lateral bases of the fin rays, and are partly overlain no section to the vestigial ray. by the depressors anteriorly, being completely cov- ARRECTOR VENTRALIS (Figure 169: ARR.V.).—The ered by them posteriorly. muscle is fairly well developed, and originates from DEPRESSORES DORSALES (Figure 169: DEPR.).—The the posterolateral face of the cleithrum, beneath sites of origin are much the same as for the erec- the ala laminaris externus. It inserts on the antero- tors, but somewhat more lateral. The muscles in- lateral base of the medial half of the vestigial fin sert on the posterolateral bases of the fin rays. ray. SUPRACARINALIS ANTERIOR.—The muscle has ADDUCTOR SUPERFICIALIS (Figure 174: ADD.S.).— partially fused with the epaxialis. It is contained The muscle has the same form as that described in the trough formed by the bifid neural spines for tetraodontids. The ventral nine tendons insert of the anterior vertebrae, and passes from the on the dorsomedial flanges of the respective rays, supraoccipital to the first pterygiophore of the dor- the other slips inserting on the dorsomedial faces sal fin. of the rays distal to their bases. SUPRACARINALIS POSTERIOR (Figures 169, 1756: s. ADDUCTOR PROFUNDUS (Figure 174: ADD.P.).—The POST.).—The muscle connects the last pterygio- fibers arise from the medial faces of the coracoid phore of the dorsal fin to the anterodorsal face and cleithrum and insert on the ventromedial bases of the neural spine of the penultimate vertebra. of the principal rays. It lies between the tips of the neural spines of the ARRECTOR DORSALIS (Figure 174: ARR.D.).—The intervening vertebrae. muscle is poorly developed, arising from the medial face of the cleithrum (lateral to the adductor profundus), and inserts on the base of the medial MUSCLES OF THE ANAL FIN half of the vestigial finray . INCLINATORES ANALES (Figure 169: INC.).—There CORACORADIALIS (Figure 174: COR.R).—The mus- are two of these muscles, both arising in the fascia cle connects the enlarged medial flange of the over the obliquus inferioris and inserting on the fourth radial to the posteroventral flange of the ventrolateral face of the first anal pterygiophore. coracoid. It is joined by a bundle of the obliquus ERECTORES ANALES (Figure 169: EREC).—Origin inferioris which arises from the anteroventral face is from the ventrolateral faces of the ventrolateral of the ventral postcleithrum. processes of the tenth to fifteenth vertebrae. The PROTRACTOR PECTORIALIS (Figures 169, 170, 171: muscles insert on the anterolateral bases of the P.P.).—Origin is from the ventrolateral surface of fin rays. the winglike process of the pterotic. The fibers DEPRESSORES ANALES (Figure 169: DEPR.).—Origin pass posteroventrally to insert on the anterolateral is the same as for the erectors, the depressor of face of the dorsal region of the cleithrum. each ray arising behind the respective erector. The LEVATOR PECTORALIS.—The muscle arises from muscles insert on the posterolateral bases of the the posterior face of the lateral process of the ptero- rays, and arise lateral to the erectors. tic and inserts on the anterodorsal tip of the INFRACARINALIS MEDIUS (Figure 169: INF.M.). cleithrum. The fibers arise from the anteroventral face of the first anal pterygiophore, and join with the ventro- MUSCLES OF THE DORSAL FIN lateral section of the obliquus inferioris before INCLINATORES DORSALES (Figure 169: INC.).— grading into the subdermal tissue as part of the NUMBER 155 67 longitudinal cutaneous system. and pterotic, and from the lateral faces of the bifid INFRACARINALIS POSTERIOR (Figure 1756: INF.P.).— neural spines of the anterior vertebrae (where the This muscle connects the posterior face of the last fibers are poorly separated from those of the supra- anal pterygiophore and the supporting haemal carinalis anterior). Posteriorly, the ventral moiety spine to the anterior margin of the haemal spine inserts on the lateral bases of caudal fin rays D of the penultimate centrum. 1-5. The dorsal moiety is restricted to a section arising from the dorsomedial faces of the bifid neural spines of the fifteenth and sixteenth verte- MUSCLES OF THE CAUDAL FIN brae. It joins the aponeurosis from the posterior INTERRADIALIS (Figure 175a: INT.).—The muscle section of the obliquus inferioris to form an interconnects the caudal fin rays just distal to their aponeurotic sheet to all the caudal fin rays. bases. OBLIQUUS SUPERIORIS (Figures 169, 175a: OBL.S.).— HYPOCHORDAL LONGITUDINALIS (Figure 1756: The fibers arise from the ventrolateral region of H.L. ).—Origin is from the posteroventrolateral hy- the seventh vertebra, and increase in size pos- pural plate, the dorsal face of the flange of the teriorly. The muscle inserts on the lateral bases ural centrum and the centrum itself, and the neural of the ventral caudal fin rays. and haemal spines and centrum of the penultimate OBLIQUUS INFERIORIS (Figures 169, 174, 175a: vertebra. The muscle inserts on the bases of rays OBL.1.).—The muscle is considerably modified. The D 3-5. anterodorsal section arises from the fascia over FLEXOR DORSALIS (Figure 1756: F.D.).—Origin the epaxialis and from the dorsolateral tips of the is from the lateral face of the neural spine of the first and last dorsal fin pterygiophores behind this. penultimate vertebra, the muscle inserting on the Ventrally, a similar attachment occurs on the last lateral bases of rays D 3-4. anal fin pterygiophore. The more dorsal of these FLEXOR DORSALIS SUPERIOR (Figure 175: F.D.S.).— fibers pass to the posterodorsal face of the postclei- The fibers originate from the lateral faces of the thrum, the middle ones attach to the posteroven- neural spines of the second to fourth last verte- tral face of the coracoid, while the ventral fibers brae, and the dorsolateral faces of the epural and pass into the subdermal tissue, contributing to the the hypural plate. The muscle inserts on the dorso- formation of the longitudinal cutaneous. The an- lateral base of ray D 5. teroventral face of the postcleithrum gives rise to FLEXOR VENTRALIS (Figure 1756: F.V.).—The a short section which forms an aponeurosis with fibers arise from the ventral face of the flangeo f the coracoradialis (Figure 174). A ventral section the ural centrum, and the dorsolateral and ventro- passes posterodorsally, lateral to the erector-depres- medial faces of the ventrolateral flanges of the sec- sor complex of the anal fin, to attach to the ond to fourth last vertebrae. Insertion is on the anteroventral surface of the ventrolateral process lateral bases of rays V 2-4. of the sixteenth vertebra. This section fuses with FLEXOR VENTRALIS EXTERNUS (Figure 1756: the infracarinalis medius anteriorly, and fades into F.V.E.).—The muscle arises primarily from the the subdermal tissue. The posteriormost section fascia on the lateral surface of the flexor ventralis, arises from the ventrolateral process of the six- with some attachment to the lateral flange of the teenth vertebra, and soon grades into an aponeu- ural centrum. It inserts on the lateral bases of rays rosis which joins that of the dorsal moiety of the V 1 and 2. epaxialis before inserting on the lateral bases of FLEXOR VENTRALIS INFERIOR (Figure 175: F.V.L).— the caudal fin rays. Origin is from the ventrolateral faces of the third and fourth last vertebrae, the muscle inserting on SWIM BLADDER MUSCLES the ventrolateral base of ray V 4. In addition to the ventral fibers of the retrac- tor dorsalis to the anterodorsal surface of the swim LATERAL BODY MUSCLES bladder, the posterior border of the bladder is EPAXIALIS (Figures 169, 170, 173, 175a: EPAX.).— covered by a muscle which passes dorsoventrally Anteriorly, the fibers arise mainly from the epiotic across its surface (Figure 173: S.M.). 68 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

CUTANEOUS MUSCLES and the fibers are inseparable from those of A 2a. Insertion of the additional fibers is primarily on The cutaneous muscles are much like those of the ventrolateral face of the maxilla. Anteroven- tetraodontids, although the transverse system is trally, some of the fibers arise from the dentary not as well developed (particularly dorsally and and articular. A IP is also better developed, as posteriorly). The longitudinal system is better de- is the A 2 complex. veloped than in the former group. The posterodorsal section of the hyohyoidei ad- ductores has a tendon passing posteriorly over the Chilomycterus orbicularis (Bloch) cleithrum to attach to the postcleithrum. The pharyngoclavicularis internus is smaller, while the FIGURE 176 externus division is larger than in D. hystrix. What The adductor mandibulae is not as well devel- is apparently transversus ventralis IV is modified oped as in D. hystrix. There is no tendon from to form an aponeurotic sheet to basibranchial 2, the anterolateral face of A la to the lips. A 2a i.e., it is obliquely oriented, and the fibers do does not extend beyond the border of the orbit, not meet their antimeres. and A 2P does not meet its antimere across the The levatores externi are as for C. orbicularis midline at the front of the orbit. Since the pos- and transversus dorsalis II is very well developed terior section of A 2 does not extend behind the posteriorly, with a large section attaching to the orbit, the fibers of the levator arcus palatini no dorsomedial cleithrum. Adductor III is small, and longer pass through the muscle. the retractor dorsalis originates from the third The protractor hyoidei arises from branchioste- to seventh vertebrae. No inclinatores dorsales or gals 2 and 3, and possesses a distinct longitudinal anales were separable. The anterodorsal origin of myocomma about the middle of its width. The the flexor dorsalis superior is more extensive. The origin of the levatores externi is less from the pos- flexor dorsalis is also larger, with a thin aponeu- terodorsomedial face of the hyomandibular. Trans- rosis to ray D 5. The flexor ventralis is not sub- versus dorsalis II is better developed, particularly divided, and originates mainly from the medial posterolaterally. The retractor dorsalis arises from faces of the ventrolateral flanges of the second the lateral faces of the second to seventh vertebrae. and third last vertebrae. The ventralmost eight pectoral rays only possess The anteroventral section of the obliquus in- anterodorsal processes, with the concomitant change ferioris from the postcleithrum attaches to the in the insertion of the abductor superficialis. The coracoid rather than joining the coracoradialis. The flexor dorsalis inserts on the bases of rays D 3-4, section over the abdominal cavity is reduced, al- and the flexor ventralis is subdivided near its though not to the extent seen in C. orbicularis. The origin. The obliquus inferioris is much reduced, posterodorsomedial section is bifid, attaching to being composed of only a few bundles of fibers. the lateral flanges of the thirteenth and fourteenth In addition to the swim-bladder muscle men- vertebrae. The posteriormost section and the dorsal tioned in D. hystrix, there is a well-developed mus- moiety of the epaxialis are much reduced. cle from the anterodorsal surface of the bladder which passes forward and ventrally around the anterior end of the bladder as far as its antero- Summary of Diodontidae ventral tip. The muscle is largely contained in the The relative development of the A 1 and A 2 groove separating the halves of the bladder. sections of the adductor mandibulae varies fairly considerably, although the basic plan of the sub- Chilomycterus schoepfi (Walbaum) divisions remains constant. The sites of origin of the levatores externi and the retractor dorsalis FIGURE 177 vary, as does the relative development of the trans- A la is better developed dorsally than in D. versus dorsalis II. Other areas of the musculature hystrix, and has acquired attachment to the head showing significant variation are the flexors of of the hyomandibular and the anterolateral ptero- the caudal fin and the development of the obliquus tic. This region of origin is somewhat aponeurotic, inferioris. NUMBER 155 69

Myological Descriptions of Representative Molids

The body outline of members of this family are ventral and ventral prootic, and the dorsomedial illustrated in Figure 11. Two out of the three faces of the hyomandibular and metapterygoid. It genera (possibly all monotypic) were dissected for lies medial to ramus mandibularis V, and inserts this study, no specimens of Masturus being avail- together with A 2a and A 3 in the Meckelian fossa. able. The mouth is small, the dorsal and anal fins A 3 arises from the lateral quadrate, symplectic, very high, and rather stiff. The caudal peduncle mesopterygoid, and hyomandibular and passes for- is aborted, so that the rays of the caudal fin lie ward to join the insertion tendon of the A 2 com- in an almost vertical line behind the dorsal and plex. anal fins, and appear to grade into them. The LEVATOR ARCUS PALATINI (Figure 178: L.A.P.).— pectoral fins are short. In swimming, the dorsal This is a small muscle, originating from the ven- and anal fins flap synchronously from side to side, tral face of the sphenotic at the rear of the orbit steering apparently being accomplished by using and passing down to insert on the dorsal face of the caudal fin as a rudder. The body is high and the hyomandibular medial to the posterior exten- thin, the shape being ellipsoid or circular. Molids sion of A 2a. possess the powerful jaws typical of gymnodonts, DILATATOR OPERCULI (Figure 178: D.O.).—Origin although their food is reported to consist of coelen- is from the ventral face of the lateral process of terates and pelagic tunicates (Mola) or seaweeds the sphenotic. The fibers insert aponeurotically (Ranzania). They have a worldwide distribution on the dorsal margin of the opercle. (Mola itself occurring in both tropical and cold- LEVATOR OPERCULI (Figure 178: L.O.).—This is temperate waters). Molids are thought to be pelagic, a well-developed muscle, originating from the ven- although the possibility that they are bathypelagic trolateral face of the pterotic and inserting on the cannot be ruled out. posterodorsal face of the opercle. Phylogenetically, they are usually considered to ADDUCTOR ARCUS PALATINI (Figure 179: A.A.P.).— be derived from the gymnodont line before separa- The muscle is well developed, and passes ventro- tion of the tetraodontoid and diodontid stocks. laterally from its origin on the lateral face of the parasphenoid to insert on the dorsomedial faces of the palatine, mesopterygoid, metapterygoid, and Mola mola Linnaeus hyomandibular. Posteriorly, the origin is aponeu- FIGURES 178-184 rotic. Anteriorly, a very small bundle of fibers passes anterolaterally from the parasphenoid to the palatine. It possesses a ventral aponeurosis, and R. MUSCLES OF THE CHEEK maxillaris V passes through it. It bears a striking ADDUCTOR MANDIBULAE (Figures 178, 179: A la resemblance to the retractor arcus palatini of the A IP, A 2a, A 23, A 3).—A la is fairly small, balistoids and ostracioids. arising from the fascia over A 2 beneath the orbit, ADDUCTOR OPERCULI (Figure 180: AD.OP.).—The and inserting on the posteroventral face of the fibers arise from the ventral face of the pterotic maxilla. A IP originates from the posterolateral and insert on the posterodorsomedial face of the face of the palatine and the lateral parasphenoid. opercle. It passes anteroventrally to insert on the postero- dorsal face of the maxilla. A 2a originates from the dorsolateral preopercle, the anterior and dorsal MUSCLES OF THE HYOID REGION hyomandibular and by a small slip from the ven- VALVULUS.—Origin is from the ventromedial tromedial face of the pterotic. It passes anteriorly, face of the angular, the fibers sweeping antero- lateral to ramus mandibularis V, to insert with dorsomedially to fade into the tissues of the man- A 23 and A 3 in the Meckelian fossa. A 2P orig- dibular buccal valve. inates from the lateral parasphenoid (some fibers PROTRACTOR HYOIDEI (Figure 178: PR.HY.).—The arising as far back as the posterolateral margin fibers arise from the posteroventral face of the of the bone beneath the basioccipital), the antero- dentary and attach to the lateral faces of the an- 70 SMITHSONIAN CONTRIBUTIONS TO ZOOLOCY terohyal and posterohyal. The more posteroven- the cleithrum, but fades into an aponeurotic sheet tral fibers pass medially to a midventral raphe. on the inner wall of the opercular cavity, just lat- HYOHYOIDEI ABDUCTORES (Figure 178: H.AB.).— eral to the heart. A small ventral bundle continues There are three of these muscles. The dorsalmost anterodorsally to insert on the anteroventral face passes from the anterodorsomedial face of ray 2 of ceratobranchial 5. to the posteroventromedial face of ray 3, the sec- PHARYNGOCLAVICULARIS INTERNUS (Figures 180, ond has the same sites of attachment, but involving 181: PHC.I.).—A large bundle of the muscle in- rays 1 and 2, while the third muscle connects the serts on the posteroventral margin of ceratobran- ventromedial face of ray 1 with the ventromedial chial 5. Most of the fibers grade into the fascia regions of the anterohyal and ventrohyal. The formed by the pharyngoclavicularis externus and muscles are well developed. the hyohyoidei adductores. A small dorsal bundle HYOHYOIDEI ADDUCTORES (Figures 178, 180: continues posteriorly to attach to the medial face H.AD.).—The muscle is well developed in the lateral of the cleithrum. wall of the opercular cavity, where it attaches to OBLIQUI VENTRALES I-III (Figure 181: OBL.V.).— the medial faces of branchiostegal rays 2-6. A small The muscles span the articulations between the anterior slip passes ventrally to attach to the dor- ventral faces of the ceratobranchials and the respec- sal face of ray 1. Some of the dorsal fibers arise tive hypobranchials of the first three arches. from the medial face of the preopercle and the TRANSVERSI VENTRALES IV, v (Figure 181: TR.V.).— anterodorsal face of the cleithrum. Ventrally, the The anterior muscle crosses the midline between fibers attach to the anteromedial face of the clei- the anteroventromedial faces of the fourth cerato- thrum and the aponeurotic fascia covering the branchials. Transversus V spans the midline be- pharyngoclaviculares. Posterodorsally, fibers of the tween the anteroventromedial faces of the fifth muscle arise from the medial face of the hyoman- ceratobranchials, and is somewhat larger. dibular, ventral pterotic, and the fascia covering RECTUS VENTRALIS i (Figures 180, 181: RECT.V.).— the wall of the opercular cavity, which, in molids, The muscle connects the anterolateral faces of the extends dorsally to the region of the sixth to sev- first ceratobranchial and hypobranchial to the pos- enth vertebrae (medial to the posterior part of terodorsal face of the dorsohyal. the pterotic and the supracleithrum). These fibers RECTUS COMMUNIS (Figures 180, 181: R.COMM.).— form the dorsal and medial walls of the cavity be- The fibers arise aponeurotically from the postero- tween the pterotic and the fifth vertebra, ulti- ventromedial face of the ventrohyal. They grade mately fading out in the aponeurotic sheet which into a thin, aponeurotic sheet which inserts on attaches to the dorsomedial margin of the clei- the anteroventral face of ceratobranchial 5. thrum. HYOHYOIDEUS INFERIORIS (Figure 178: HY.IN.).— This muscle is well developed, passing dorsolat- DORSAL BRANCHIAL MUSCLES erally from the midline to the lateral face of the LEVATORES EXTERNI I, HI, iv (Figure 180: L.EXT.).— anterohyal and the base of the first branchiostegal Origin is from the dorsomedial face of the hyo- ray. mandibular, and the muscles are small. Levator I STERNOHYOIDEUS (Figures 178, 180: STH.).—Ori- inserts on the dorsolateral face of epibranchial 1 gin is from the anteroventral face of the cleithrum. and the ligament between the first and second The ventral fibers pass medial to the rectus com- ceratobranchials. Levator III inserts on the dorso- munis to the posterior face of the ventrohyal, lateral faces of epibranchials 2 and 3, with the while the dorsolateral fibers lie lateral to this mus- bulk of the fibers serving the latter bone. Levator cle, also attaching to the ventrohyal. IV is best developed, and inserts along the dorso- medial face of epibranchial 4. VENTRAL BRANCHIAL MUSCLES LEVATORES INTERNI II, HI (Figure 180: L.INT.). Both muscles originate from the posteroventral PHARYNGOCLAVICULARIS EXTERNUS (Figures 178, face of the prootic and pass anteroventrally. Leva- 180, 181: PHC.E.).—The muscle is well developed tor II inserts on the lateral face of infrapharyngo- near its origin from the ventromedial face of branchial 2, while levator III inserts on the NUMBER 155 71 dorsolateral faces of infrapharyngobranchials 2 and base of the vestigial ray. 3. ARRECTOR VENTRALIS (Figure 178: ARR.V.).—The OBLIQUUS DORSALIS HI (Figure 182: OBL.D.).— muscle arises from the lateral cleithrum dorsal to The muscle connects the posterodorsal face of the abductor superficialis, and inserts on the an- epibranchial 3 with the posterodorsal faces of terodorsal face of the medial half of the vestigial infrapharyngobranchials 2 and 3. fin ray. TRANSVERSI DORSALES II, iv (Figures 180, 182: ADDUCTOR SUPERFICIALIS (Figure 183: ADD.S.).— TR.D.).—Transversus II arises from the midline The fibers arise from the anterodorsomedial face beneath and including the ventrolateral surface of the cleithrum, and insert on the anteromedial of the parasphenoid. The more anterior fibers have faces of the principal rays some distance distal to migrated forward along the parasphenoid, so that their bases. this part of the muscle passes posterolaterally from ADDUCTOR PROFUNDUS (Figure 183: ADD.P.).— its origin. Insertion is on the dorsomedial face of Origin is from the medial faces of the cleithrum infrapharyngobranchial 2. Transversus IV passes and coracoid and the posterolateral face of the across the midline between the medial faces of latter bone. The muscle inserts on the postero- the fourth epibranchials. medial faces of the principal rays a little distal RETRACTOR DORSALIS (Figure 182: D.RETR.).— to their bases. Origin is from the posterodorsolateral faces of the ARRECTOR DORSALIS (Figure 183: ARR.D. ).—This exoccipital and basioccipital, and the neural spine is a small muscle, originating from the medial of the first vertebra. The muscle inserts on the face of the cleithrum and inserting on the medial posterodorsomedial faces of infrapharyngobran- face of the medial half of the vestigial fin ray. chials 2-4. PROTRACTOR PECTORALIS (Figures 178, 180: P.P.).— ADDUCTORES IV, v (Figures 180, 182: AD.).—Ad- The muscle originates from the posterodorsal tip ductor IV is well developed and connects the of the hyomandibular and the pterotic behind it. medial and posteromedial faces of epibranchial It attaches to the anterodorsal face of the clei- 4 and ceratobranchial 4. Adductor V is small, pass- thrum. The medial fibers intermingle somewhat ing from the anterodorsal tip of ceratobranchial with the lateral fibers of the levator pectoralis. 5 to the posterolateral face of epibranchial 4. The muscle lies lateral to the adductor operculi SPHINCTER OESOPHAGI (Figures 180, 181, 182: but medial to the levator operculi. s.o.).—The muscle surrounds the esophagus, at- LEVATOR PECTORALIS (Figure 180: TR.).—The taching primarily to the posterior faces of epi- fibers of this muscle pass medial to the adductor branchial 4 and ceratobranchial 5. Dorsally, some operculi, and arise from the fascia of the opercular fibers become longitudinally oriented, lying both cavity beneath the posterior extension of the hyo- lateral and medial to the retractor dorsalis. They mandibular. The fibers insert on the anterodorso- attach to the posteromedial faces of epibranchial medial face of the cleithrum. 4 and infrapharyngobranchial 3. Ventrally, some fibers also take on a longitudinal orientation, lying dorsal to transversus ventralis V and attaching to MUSCLES OF THE DORSAL FIN the anteromedial face of ceratobranchial 5. ERECTORES AND DEPRESSORES DORSALES (Figures 178, 184: EREC. + DEPR.).—Only the erectors and depressors of the first two fin rays have separate MUSCLES OF THE PECTORAL REGION tendons. In the next fourteen rays, the erectors ABDUCTOR SUPERFICIALIS (Figure 178: ABD.S.).— and depressors have fused, and insert by a single The fibers arise from the anterolateral face of tendon along the lateral face of each ray a little the cleithrum and insert on the dorsolateral faces distal to the base. Origin is from the lateral faces of the principal fin rays. of the vertebral centra and neural spines, the pos- ABDUCTOR PROFUNDUS (Figure 178: ABD.P.).— terodorsal surface of the skull, and the septum in Origin is from the lateral faces of the cleithrum the midline dividing the dorsal and anal finmus - and coracoid, the muscle inserting on the postero- culature. The ventral fibers are continuous with lateral faces of the principal rays and the lateral the surrounding fibers, particularly anteriorly. The 72 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY muscles to the more anterior rays tend to lie nearer The mass of this musculature arises from the the midline, those to middle rays being best de- lateral faces of the ventral caudal fin supporting veloped. The lateral fibers of the more anterior elements and the centra and haemal spines of the muscles also take origin from the medial face of fifteenth and sixteenth vertebrae. Insertion is the thick, subdermal fascia (represented as cut sur- tendinous on the anterolateral bases of the eight face in Figure 178, but not in Figure 184). ventral caudal fin rays. SUPRACARINALIS ANTERIOR.—This is a small mus- OBLIQUUS INFERIORIS (Figure 178: OBL.I.).—The cle which arises from the posterodorsal face of the muscle is reduced to two small fiber bundles. The supraoccipital. The fibers soon grade into a long, anteroventral one arises from the posteromedial thin tendon which attaches to the anterodorsal face of the coracoid, and sodn fades out in the face of the first dorsal pterygiophore. fascia overlying the abdominal cavity. The other section arises from the posteroventral tip of the postcleithrum and passes posteroventrally into the MUSCLES OF THE ANAL FIN subdermal tissue, fading out just above the first pterygiophore of the anal fin. It is just possible ERECTORES AND DEPRESSORES AN ALES. (Figure 184: that this section represents the residual first in EREC. + DEPR.).—The first two fin rays have sep- clinator analis, but it seems more likely that it arate erector and depressor tendons, the tendons is part of the obliquus inferioris. to the other fourteen fin rays being compound, owing to the fusion of the erector and depressor muscles. Insertion of these tendons is on the lateral Ranzania laevis (Pennant) faces of the rays a little distal to their bases. The fibers originate from the lateral septum, the centra FIGURE 185 and haemal spines of the seventh to sixteenth verte- A la is smaller, while A 13 originates predom- brae, and the basal pterygiophores of the anal fin. inantly from the ventrolateral prefrontal and The fibers do not encroach on the abdominal frontal. A 2a is smaller, without a posterior sec- cavity, and intermingle in the dorsal region. tion along the hyomandibular to the pterotic. A 2P does not reach the rear of the orbit, nor does LATERAL BODY MUSCLES it extend far in front of it. A 3 appears to be absent. The levator arcus palatini is better de- EPAXIALIS (Figures 178, 184: EPAX.).—The muscle veloped. The levator operculi originates more from is much reduced. Anteriorly, it arises from the pos- the supracleithrum than from the pterotic. There terodorsal tip of the epiotic. It passes posteriorly is no small, anterior bundle of the adductor arcus as a very thin sheet, just below the skin, and palatini. gradually becomes broader. It is separated from The valvulus arises from the anterodorsal face the erector-depressor complex by a tough fascia, of the interopercle. Only a small bundle of the from the medial surface of which the lateral fibers hyohyoidei adductores is developed in the dorso- of this latter complex arise. Vague myomeres are lateral wall of the opercular cavity above the sixth recognizable, there being some fifteen of them be- branchiostegal ray. Rectus ventralis II joins cerato- fore the fibers become continuous with the section branchial 2 to hypobranchial 1 and basibranchial serving the dorsal eight caudal fin rays. This main I. Both sections of the pharyngoclaviculares are posterior section arises from the lateral faces of more diffuse. the caudal fin supporting elements and the centra There are four levatores externi. Transversus and neural spines of the fourteenth to sixteenth dorsalis II is more complex, since some of the vertebrae. Insertion is tendinous on the antero- posterolateral fibers attach to the anteromedial lateral bases of the eight dorsal caudal fin rays. face of epibranchial 3. Obliquus dorsalis II passes OBLIQUUS SUPERIORIS (Figure 184: OBL.S.).—Pos- from the dorsal face of epibranchial 2 to the dorso- teriorly, vague fibers in the subdermal tissue are lateral region of infrapharyngobranchial 2. This continuous with the anterodorsolateral surface of may represent that portion of transversus dorsalis the muscle mass serving the ventral caudal fin rays. II which usually arises from the epibranchial, and NUMBER 155 78 which has in this case become separated by the cleithrum. The fibers virtually cover the abdom- dorsal expansion of infrapharyngobranchial 2. The inal cavity ventrally. The residual inclinator analis protractor pectoralis grades into a diffuse aponeu- arises from the anterolateral base of the first anal rosis attaching to the posterodorsal face of the pterygiophore, but fades out in the subdermal tis- preopercle, and passing lateral to the levator oper- sue before reaching the posteroventral tip of the culi. postcleithrum. There are no fibers from the post- The first dorsal fin ray only has both an erec- cleithrum. tor and a depressor. The broad tendons of the The epaxialis is restricted to a well-defined mus- rest of the compound muscles overlap onto the cle serving the dorsal ten caudal fin rays, with no base of the preceding ray. The last three tendons fibers from the subdermal tissue. The obliquus pass anterior to the bases of the last three rays, superioris is likewise well separated from the and meet their antimeres in the midline just above erector-depressor complex of the anal fin, and the platelike, cartilagenous pterygiophores. No inserts on the bases of the ten ventral caudal fin fibers arise from the subdermal fascia. rays. The obliquus inferioris is thus confined to The muscles to the anal fin rays are as described a small bundle of fibers from the posteroventral above for the dorsal fin. Their origin is more exten- face of the coracoid which fades out in the fascia sive, and includes the posteromedial face of the over the abdominal cavity.

Phylogenetic Significance of Observed Myological Variation

Each muscle is analyzed on the assumption that In the following section, the letters "R.T." indi- nothing is known of the phylogeny of the con- cate that there is some reason to believe that the tained units. This approach has certain drawbacks. phylogeny indicated by that muscle may be the It precludes, for example, the possibility of any result of a reductive trend. The term "R.T. pars" explanation based on parallel or "convergent" means that only some of the phyletic lines exhibit evolution, although the former phenomenon is a reductive trend in the character under consid- often traceable through the presence of reductive eration. These abbreviations are used whether the trends. Evidence from muscles which suggest phy- indicated phylogeny agrees or disagrees with that logenies other than the one postulated in Figure postulated in Figure 44. The Hennigian nomen- 44 (which is derived from the evidence of 64 per- clature to denote character state is used, which cent of the muscles for which phylogenetic interpre- roughly corresponds to more widely used terms as tation proved possible) is reinterpreted immediately follows: apomorph (specialized, derived); autapo- after the initial analysis of the muscle concerned, morph (specialized in one lineage only—denoted and I have attempted to provide plausible expla- by "A" in Figure 44); plesiomorph (primitive, nations for the discrepancies. In this connection, it generalized); symplesiomorph (shared primitive) is of interest to note that the phytogenies initially and synapomorph (shared specialized). The indicated by these muscles seldom agreed with each nuances of the Hennigian nomenclature differ other, which is a further justification for their slightly from the alternative terms given above— reinterpretation. plesiomorph meaning "primitive" for the group It should be stressed that in this interpretive under consideration only, and always defined rela- section, all statements concerning phylogeny and tive to apomorph. character state are based on a consideration of the Symbols used in the phylogenetic tree accom- muscle being reviewed at that time only, and are panying each muscle (Figures 12—43) are in the not intended as generalized or particularized state- "List of Abbreviations." In interpreting the phylo- ments of fact outside that context. It should also genetic trees accompanying the muscle analyses, be noted that one of the major difficulties in decid- the following conventions are used: oblong "boxes" ing character state in fish muscles is due to the indicate a common ancestor for all the families lack of reliable, published information on the whose lines arise from the "box" itself; lines which majority of fish groups. branch are to be interpreted as is usual for phylo- 74 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY genetic trees, i.e., they indicate the sequence of (which appears to be homologous with the balis- branching in relative time; "boxes" with one end toid A la) is very well developed in all families open connect up with the similar "box" with the except the molids. It reaches the height of its de- opposite end open, and indicate a common an- velopment in the diodontids. A IP (which does cestor for the families arising from both parts not seem to be homologous with that of balistoids (they only appear when the analysis of the muscle + ostracioids) is a dorsal subdivision of A 1, and, indicates a sequence of branching differing from in most cases, acquires attachment to the pre- the one postulated in Figure 44). frontal region of the skull. It is well developed in molids, and enormously enlarged in the tetrao- dontoids, where it extends posteriorly along the Interpretation parasphenoid beneath the orbit. It is small in 1. ADDUCTOR MANDIBULAE, SECTION A 1 (Figure diodontids, where most of the area of its origin in 12).—It is probable that the ancestral tetraodonti- the other families has been 'taken over' by A 23. form condition of this section was such as is found The small size is probably associated with the in most perciforms with protrusile jaws—i.e., the lateral migration of the oblique eye muscles (see section is not subdivided, and inserts high on the Winterbottom, 1971a). In these gymnodonts, the posterodorsal face of the maxilla (the anteroven- feeding emphasis is on grinding and crushing— tral region being ligamentously continuous with hence the massive jaws and well-developed muscu- A 2). As such, it is found in all the Triacanthodidae lature. examined except the two long-snouted genera Phylogenetically, the form of A 1 in the triacan- (Macrorhamphosodes and Halimochirurgus), where thodids is assumed to be plesiomorph. Triacan- the protrusibility of the jaw has been lost with the thids are autapomorphly specialized, while all the elongation of the snout. other families are linked by the broad insertion In triacanthids, the section is more intimately of A 1 g on the maxilla. Triodontids possess a associated with A 2, the tendon to the maxilla condition similar to what one would expect lying in an almost horizontal plane, and being in the common ancestor of these families. The much reduced. This is presumably associated with balistoids and ostracioids are linked by the a lesser degree of jaw protrusion and the more development of the deep A IP subdivision (see incisiform teeth, perhaps indicating that biting analysis above), and it appears that the balistoids and nibbling are important in feeding. The only form the plesiomorph sister group of the ostracioids other family in which A 1 is single is the triodon- (where A la is lost). The molids appear to form tids, where the muscle inserts broadly on the inter- the plesiomorph sister group of the diodontids and digitated elements of the upper jaw. tetraodontoids, in which latter A la is well devel- There is a fairly clear dichotomy between the oped and attaches to the preopercle laterally. Di- other families. In balistoids and ostracioids, a deep odontids do not have the well-developed A IP of subdivision to A 1 develops (A IP), lying, for the tetraodontids and canthigasterids, but this is the most part, medial to A 2. In monacanthids, probably an autapomorph condition. However, considerable further subdivision of both sections the enlargement of this subdivision (relative to has occurred, but not in balistids. Ostracioids have the condition in both diodontids and molids) in lost A la, and A 13 develops three heads of origin. the tetraodontoids does seem to be synapomorph, It seems likely that the common balistoid-ostracioid and thus valid as phylogenetic evidence. ancestor possessed a condition similar to that found 2. ADDUCTOR MANDIBULAE, SECTION A 2 (Figure in the extant Balistidae. The subdivisions pre- 13).—A 2 consists of two subdivisions in all the sumably allow more controlled nibbling move- families except the triacanthids and the two long- ments of the upper jaw. It is therefore not snouted genera of the triacanthodids (Macrorham- surprising to find their greatest development in phosodes and Halimochirurgus). In the more the more herbivorously inclined monacanthids than generalized tetraodontiforms, A 2a lies dorsola- in the other more omnivorous groups. terally and A 23 lies ventromedially. The sub- In the gymnodonts (exclusive of triodontids), divisions are separated by the? path of ramus there are again two subdivisions of A 1. A lo mandibularis V, which passes medial to A 2a and NUMBER 155 75 lateral to A 2P (except in the ostracioid families from the overlying sections of A 2. In the ostraci- where it lies medial to both sections, and may oids, however, the dorsal fibers are continuous indicate the sister group status of the two families). with the ventral region of A IP indicating that It is by no means certain whether the ancestral the Aracanidae and Ostraciidae are sister groups. condition of this muscle in the pretetraodontiforms 4. ADDUCTOR MANDIBULAE, SECTION A to (Figure was divided or undivided, and the usefulness of 14fr,c).—The section is possibly present in the it as a phylogenetic indicator among the more triacanthodid Hollardia (the fibers do not, how- basal groups is thus negligible. The generalized ever, reach the quadrate), and is fully formed in condition of section A 2 as described above is triodontids, diodontids, and tetraodontids. In the found in the triacanthodids, triodontids, balistoids, three latter families, the muscle originates from and ostracioids. The condition in triacanthids the palatal arch (anterior face of quadrate and appears to be autapomorph. Balistoids are linked ectopterygoid in the diodontids and tetraodontids, by the development of a ventral section, A 2o, al- and the medial face of the quadrate and metaptery- though this may be absent in certain specialized goid in triodontids). Since it seems likely that the monacanthid genera. The synapomorphic condi- triodontid condition is autapomorph, we can say tion of A 2 unites the gymnodonts (except the nothing further concerning the phylogeny of these triodontids). Here A 23 has expanded dorsally three families. and medially, above A 3, to include the para- The simplest inference is that the section devel- sphenoid and prootic in its sites of origin. The oped soon after the gymnodont line diverged, and molids show a lesser (and thus plesiomorph) was correlated with the powerful, crushing "beak" development of the section anteriorly than is of these fishes. Its absence in canthigasterids and found in the remaining families. Out of these molids ties in very nicely with their diet (canthi- last three families, only the diodontids show an gasterids are mainly herbivorous, or take soft- enormous, anterodorsal development of the sec- bodied marine invertebrates—molids apparently tion to the ventral roof of the orbit, but this is feed largely on medusoid coelenterates and salps), an autapomorph condition. where powerful jaws are not necessary. It is thus In functional terms, it seems that the much feasible that the section has been lost separately broader anterior portion of the diodontid cranium in the two lineages (Figure 14c). is an adaptation to the inequality in the jaw 5. LEVATOR ARCUS PALATINI (Figure 15a).— musculature. The muscles to the upper jaw are Little can be deduced about the detailed familial much weaker, and those to the lower jaw much interrelationships from this muscle. It is relatively more powerful than in the tetraodontoids. Thus, well developed in triacanthoids, triodontids, and there will be greater force tending to push the molids. It is rather poorly developed in all the upper jaw dorsally on the contraction of the jaw other families (except a few ostraciids), but in musculature, this force being in some measure bal- somewhat different ways. In balistoids and ostra- anced in tetraodontoids by the almost equal force cioids, as a result of the ventral extension of the applied to the upper jaw. Hence, a stronger brac- sphenotic region immediately behind the orbit, ing of the elements of the upper jaw against the the area containing the muscle is reduced. This skull is required in diodontids, and is reflected in may be correlated with the presence of the retrac- the broader anterior region. The same argument tor arcus palatini (which appears to perform a may apply in reverse to the canthigasterids, whose similar function anteriorly—see below). In diodon- jaw musculature is relatively weaker than that of tids and tetraodontoids, the hyomandibular ex- the tetraodontids and diodontids. Here the snout tends posterodorsally far behind the orbit. Here region is relatively more elongate than in the other again the muscle is small, but this time its reduc- two families. tion is probably associated with the decreased im- 3. ADDUCTOR MANDIBULAE, SECTION A 3 (Figure portance of the palatal arch in respiration, and Ha).—The section is of little use in phylogenetic the concomitant increase in the first branchiostegal interpretations. It is almost uniformly present in ray pumping mechanism. all the families (absent in three genera of triacan- 6. DILATATOR OPERCULI (Figure 156).—In its thodids), but is not always completely separated plesiomorph condition (as found in most perci- 76 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY forms), this muscle arises from the dorsolateral represents the pars posterior of the muscle and surfaces of the sphenotic and pterotic, and inserts the gymnodont condition the pars anterior, or tendinously on the tip of the dorsal process of else (and more likely) the muscle was reduced the opercle. As such, it is present in triacanthoids to a single entity in a postulated common ances- and triodontids. In molids, it inserts on the antero- tor, this then developing in the directions outlined dorsal rim of the much-reduced opercle. It is very above. The available evidence does not seem to well developed in all the other families (it is warrant a firm conclusion in this matter. The worth adding here that these are the groups in gymnodont condition is presumably associated with which the levator operculi is small), but there the increased use of the opercle and subopercle in again appear to be two lines. In balistoids and respiration. Both conditions are synapomorph rela- ostracioids, the origin is low on the skull, and tive to the triacanthoid state, and indicate a com- contained in the dilatator fossa (synapomorph), mon ancestor for each of the two groups which is which is particularly well covered laterally by the not shared by the triacanthoid fishes. sphenotic in ostracioids (also synapomorph). The 8. ADDUCTOR ARCUS PALATINI (Figure 166).— insertion has become specialized in balistoids, and The plesiomorph condition, in which the muscle the fibers attach to the dorsolateral surface of the occupies the floor of the orbit, is found in most opercle as well as the dorsal rim (the latter site perciforms, triacanthodids, and triodontids. Tri- being of less importance in monacanthids). In acanthids are somewhat specialized in that the ostracioids, the muscle attaches to the remains of muscle has expanded anterodorsally in front of the dorsal process of the opercle (the more so in the orbit. In balistoids, the muscle is largely in the aracanids), this apparently being the plesio- front of the orbit, in ostracioids it is completely morph condition. The muscle becomes very well so. In both, the anterodorsal fibers are to some developed (particularly in the dorsal region of degree continuous with those of the retractor arcus origin) in diodontids and tetraodontoids. The palatini. The muscle is also plesiomorph in molids. fiontals are included among the sites of origin, Mola (but not Ranzania) develops a small, longi- and insertion is all around the dorsal process of tudinal, anterior bundle which inserts on the pala- the opercle. Tetraodontids and canthigasterids are tine and mesopterygoid. The homology of this linked synapomorphly by the presence of an an- bundle is not clear (see below). In diodontids, the terolateral strap of muscle fibers to the interopercle muscle is small and confined to the orbit, while (canthigasterids), preopercle (Lagocephalus), or the fibers continue anteriorly in tetraodontoids to subopercle (most other tetraodontid genera). The connect the medial faces of the palatines across large development of the muscle in the latter three the midline beneath the vomer. Phylogenetic in- families is again presumably associated with terpretation is difficult. The triacanthid condition changes in respiratory mechanism, and would al- appears to foreshadow the development of the re- low for more forceful operation of the opercular tractor arcus palatini. It is thus possible that (1) suction pump. the anterior bundle of Mola is a "de novo" devel- 7. LEVATOR OPERCULI (Figure 16a).—Many opment, (2) molids were derived from the ances- perciforms and the triacanthoids exhibit an ap- tral stock before the separation of the triodontids, parently plesiomorph condition of this muscle. the molid sister group (i.e., all other gymnodonts) It consists of anterior and posterior parts, the losing the anterior bundle (in which case we can- former primarily arising from the pterotic, the not explain the absence of the bundle in Ranzania), latter from the posttemporal. In all other tetra- or (3) the incipient retractor arcus palatini has odontiforms, the muscle is single. In the balistoids been lost in three separate lineages—triodontids, and ostracioids, it is small, and arises from the Ranzania, and the diodontids plus tetraodontoids. secondary portion of the pectoral girdle (post- It seems that the acceptance of the "de novo" temporal and, in balistoids, the supracleithrum). development of the bundle in Mola is the most In the gymnodonts, however, the muscle is well parsimonious interpretation. However, the anterior developed, and originates from the pterotic and bundle can be of little functional significance some of the bones dorsal to it. There are two pos- at this size, and its morphological similarity to sible interpretations. Either the former condition the retractor arcus palatini of balistoids and ostra- NUMBER 155 77 cioids adds an element of incredulity to such an triacanthids (although not so pronounced). In acceptance. It seems better to leave this part of acanthurids (including the Zanclinae) a well- the interpretation open to question (although developed retractor arcus palatini is present. The atavism may be involved). Tetraodontoids exhibit homologies of this with the balistoid-ostracioid a synapomorph condition in the anterior exten- muscle are uncertain. In balistids, the posterodorsal sion of the muscle across the dorsal midline, while force of the muscle causes the arch to rotate out- there is a partial (balistoids) or complete (ostra- ward about its articulation with the palatine. In cioids) migration of the muscle in front of the orbit. monacanthids, the fibers are more vertical, and Functional interpretations pose as much of a rotation could occur about the cartilagenous ball problem as those of a phylogenetic nature. Pre- between the palatine, vomer, and ectopterygoid. sumably the triacanthid condition is a response In ostracioids, the movement seems to take place to the increased length of the snout, which is con- between the palatine and vomer. tinued further in the balistoids and ostracioids. 10. ADDUCTOR OPERCULI (Figure 176).—The However, it is not so easy to postulate a reason muscle is present in all the families examined. It for the difference between the tetraodontoid and originates from the pterotic, prootic, or both in diodontid conditions, and none will be attempted. all except the ostracioids, where the fibers arise 9. RETRACTOR ARCUS PALATINI (Figure 17a).— from the occipital region. This condition is apo- The muscle has only become distinct in the balis- morph, and links the two families. toids and ostracioids, although its development is 11. ADDUCTOR HYOMANDIBULAE (Figure 18a,fc).— apparently foreshadowed by the dorsal extension The muscle has a somewhat peculiar distribution. of the adductor arcus palatini in triacanthids (see It is present in balistids (but not monacanthids), above). The muscle is well separated from the in the aracanid Kentrocapros (but not Capropygia), adductor arcus palatini and inserts tendinously in ostraciids, and in triodontids. In perciforms on the dorsomedial face of the palatal arch in (?thus plesiomorph), origin is from the prootic in ostracioids. The muscle is less well separated with front of the adductor operculi. It is not certain a more vertical fiber direction, and inserts on the whether the tetraodontifonn condition is homol- dorsolateral face of the palatal arch in the balis- ogous with this, or, indeed, whether it is homol- toids. The balistoid condition is not considered ogous within the tetraodontiforms themselves. In to be wholly plesiomorph, since the muscle is very the ostracioids it originates from the exoccipital well developed. In this respect, the ostracioids are (Kentrocapros) or from the cleithral region (os- more plesiomorph; however, here the orientation traciids), and this posterior origin is apparently of the fibers, the tendinous insertion, and the com- apomorph. Its absence in gymnodonts (other than plete separation of the muscle from the adductor triodontids) is not suprising, since the hyomandib- arcus palatini seem to be synapomorph. The con- ular becomes increasingly interdigitated with the dition in Mola has been discussed above, but the chondrocranium, and thus less mobile. Its absence absence of a well-developed muscle in gymnodonts in triacanthoids may be due either to the fact is scarcely suprising in view of the interdigitation that they have independently lost the muscle, or, of the palatine with the ethmoid-vomer region of more likely, because the muscle only developed the skull. after they had split off from the other groups (in The development of this muscle is presumably which case the muscle is not homologous with its in response to the elongation of the ethmoid region. perciform counterpart). Its absence in monacan- Usually, the palatal arch is moved laterally by the thids could be due to subsequent loss. contraction of the levator arcus palatini. However, In reexamining the situation, I assume that the the elongation of the ethmoid region decreases the most generalized condition of the adductor arcus mechanical advantage of this muscle, since it is palatini is when it stretches across the posterior situated at the rear of the orbit. The development part of the floor of the orbit, and inserts on the of the retractor arcus palatini may alleviate this palatal arch, the hyomandibular, and the opercle deficiency. The anterior fibers of the adductor arcus (forming the adductor operculi of more advanced palatini in certain chaetodontins and pomacanthins fishes). These points of insertion are close together, show a somewhat similar orientation to those o£ and it is not unreasonable to assume a "polyphy- 78 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY letic" (i.e., independent) occurrence of the adduc- present as a "de novo" development in the gym- tor hyomandibulae, which occupies the area nodonts only. It may arise from the protractor between the adductor arcus palatini and adductor hyoidei as a diffuse sheet (triodontids), or from operculi. It seems probable that the triacanthoid the interopercle (tetraodontids and the molid Ran- condition, where the muscle is absent, is plesio- zania), the angular (diodontids and the molid morph for the tetraodontiforms. The common Mola), or the preopercle (canthigasterids). The balistoid + ostracioid-gymnodont ancestor devel- triodontid condition would seem to be plesiomorph. oped a section (most likely from the adductor 14. PROTRACTOR HYOIDEI.—Such variation as operculi, with which the present muscle is in con- occurs in this muscle cannot, at present, be used tact) which attached to the hyomandibular. In in phylogenetic interpretation, either because it the balistoid-ostracioid line, the insertion became is too sporadic, or because the character state can- tendinous (ostracioids), while the muscle was lost not be satisfactorily determined. in the monacanthids (the reason why this should 15. HYOHYOIDEUS INFERIORIS (Figure 20a,b).— be so is enigmatic). Among the gymnodonts, tri- In the apparently plesiomorph condition, the mus- odontids retain what is now the plesiomorph con- cle passes from the anterohyal beneath the first dition, while all the other families have lost the few branchiostegal rays to a midventral raphe. It muscle—probably because of the increasing inter- is connected to the ventrohyals by tendons arising digitation of the hyomandibular to the chon- from its anterior surface. As such, it is found in drocranium. Reinterpreted in this way, the triacanthoids. In all other families, the tendons phylogenetic tree changes to that shown in Figure are lost. The muscle is well developed in balis- 18b. toids (in most monacanthid genera it is partially 12. INTERMANDIBULARIS (Figure 19a).—The pres- attached to the urohyal) and poorly so (synapo- ence of this muscle is plesiomorph in the Teleostei. morph) in the ostracioids. It is partially fused to It is absent in the gymnodonts dissected, and this the protractor hyoidei in triodontids, but absent is presumably associated with the firm suture or in all other gymnodonts except molids, where it fusion of the halves of the lower jaw, precluding is well developed. the normal lateral movement of these elements. In diodontids and tetraodontids, the loss (or There is one exception to this general absence—it fusion) of the muscle is presumably associated with appears as a healthy, well-developed muscle in the the branchiostegal ray pumping mechanism. This tetraodontid Xenopterus. This is inexplicable in mechanism is not found in triodontids, but here normal evolutionary terms—it seems fatuous to the pectoral and pelvic girdles closely underlie assume (1) that Xenopterus represents a lone, evo- the hyoid arch, inhibiting the amount of lateral lutionary line which split off from the gymnodont movement of the arch, and thus decreasing the stock before any of the other derivatives or (2) usefulness of the muscle; that is, it does not neces- that the muscle was present in the tetraodontid sarily mean that triodontids share an ancestor with line up until Xenopterus split off, being subse- diodontids and tetraodontoids, in which groups quently lost in the other species. The situation is the reduction has been completed. Figure 206 perhaps analogous to that found in the pleural gives the resulting phylogeny. ribs of tetraodontiforms—they are found only in 16. HYOHYOIDEI ABDUCTORES (Figure 21a). Triodon and the monacanthid Pseudaluteres Many generalized perciforms possess these muscles (Tyler, pers. comm.). The single possible explana- to the ventral six branchiostegal rays. Their num- tion is that an unblocking of the genetic code ber is usually four in tetraodontiforms (five in governing the development of the muscle has oc- some balistids, three to six in monacanthids). Balis- curred, possibly as a result of selection pressures toids are apomorph in that the muscle to the acting on one of the other facets of that gene com- first ray is enlarged, and often passes more dorsally plex's pleiotropism (i.e., atavism, sensu Rensch, than the other muscles. Triodontids and molids 1959). Although this is not a particularly con- have three muscles, with the first broadly devel- vincing argument, it is certainly easier to accept oped. The first branchiostegal ray of diodontids than either of the above alternatives. and tetraodontoids is enormously enlarged, and 13. VALVULUS (Figure 196).—The muscle is this is accompanied by an equally enormous, apo- NUMBER 155 79 morph development of the muscle serving the ray. It is possible (but difficult to demonstrate) that Tetraodontoids have developed a medial fiber bun- the medial section of the sternohyoideus in balis- dle which crosses the midline beneath the bran- toids and ostracioids is involved in more complex chial arches above the ventral tip of the cleithrum. respiratory movements. The small size of the mus- This bundle is not as well developed or separated cle in gymnodonts is again presumably associated in the diodontids. with a different breathing mechanism. 17. HYOHYOIDEI ADDUCTORES (Figure 21&).—In 19. STERNOBRANCHIALIS (Figure 22b,c).—The most perciforms, the muscle is usually somewhat plesiomorph condition of this muscle is probably diffuse, and lies in the lateral wall of the opercular such as is found among certain perciforms and cavity between and above the branchiostegal rays. triacanthodids. Here the posteromedial fibers of As such, it is found in triacanthoids and balistoids. the sternohyoideus tend to pass dorsally and be- It is fairly well developed in the ostracioids, and come attached (triacanthodids) to the median there is a ventromedial slip to the anterior face septum separating the opercular cavities from each of the cleithrum from the opercular valve. It is other and from the heart. The fibers separate out well developed in triodontids. In the other farther in triacanthids, balistoids, and ostracioids. gymnodonts, it becomes even better developed, Insertion is still broad and somewhat diffuse in and acquires attachment to the pterotic and triacanthids. In balistoids, the anterior aponeurosis hyomandibular dorsally (although it is poorly attaches to the urohyal, while the posterior one developed laterally in the molid Ranzania). In inserts on hypobranchial 3. Two subdivisions are diodontids and tetraodontoids, it is an extremely also present in the ostracioids, the first including powerful muscle, with some fibers passing from hypobranchial 3 and the second including cerato- the fascia between epibranchials 3 and 4 to the branchial 4 in sites of insertion (these subdivisions posteromedial wall of the cavity. In tetraodontoids, in ostracioids are completely separate from each the posterolateral fibers form a semicircle around other and the sternohyoideus). Both conditions ap- the opercular valve, connecting the dorsal cleithrum pear to be synapomorph, apparently arising from to the dorsolateral coracoid. Just below this, a different directions of consolidation of the broad bundle passes anteriorly from the coracoid to the aponeurotic sheet probably present in the common cleithrum. The large development of the muscle ancestor. The muscle is absent in all gymnodonts. is gymnodonts (particularly in the last three fam- In a broad context, this is plesiomorph, for the ilies) is presumably associated with the increased muscle has seldom been reported in perciforms rigidity of the palatal arch, and the increased role (although Dietz, 1914:133, reports its presence in of the opercular cavity in respiration. gobioids, and the orientation of the posteromedial 18. STERNOHYOIDEUS (Figure 22a).—In most fibers of the sternohyoideus in many acanthoptery- teleosts, this muscle is posteriorly continuous with gians). If, however, one accepts that the gym- the obliquus inferioris body musculature, and in- nodonts are tetraodontiforms (and the evidence serts on the urohyal. As thus defined, it is found is impressive), then it is true to say that the subse- in triacanthoids. In balistoids, a medial subdivision quent loss of the muscle is apomorph. develops which inserts on the ventral tip of the Since no tetraodontiform possesses well-developed urohyal. There may also be a horizontal subdivi- toothed, lower pharyngeals, and since the sterno- sion as in some balistids (e.g., Rhinecanthus, branchialis attaches anteriorly onto the bran- Balistapus) or the medial subdivision may be absent chial arches, it seems that its presence must be (e.g., the monacanthid Oxymonacanthus). Ostra- correlated with respiration (by aiding the depres- cioids also possess this medial subdivision, but sion of the branchial and hence hyoid arches), and here the muscle arises primarily from the fascia its absence may once again be referrable to the between the cleithrum and the coracoid (syna- branchiostegal ray pumping mechanism of the dio- pomorph). In gymnodonts, the sternohyoideus is dontids and tetraodontoids. This latter may not small, and is not laterally continuous with the be wholly correct, since the muscle is also absent obliquus inferioris. In diodontids and tetraodon- in triodontids and molids. Possibly it was originally tids, the posteromedial fibers tend to sweep more lost by a common ancestor similar to Triodon, dorsally than anteriorly. where the cleithrum closely underlies the branchial 80 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY basket (and hence there is little room for vertical ing the two families. Triodontids have lost the movement of the branchial and hyoid arches). muscle of the first arch, but the other two are It is also possible that the similarity of the well developed. All other gymnodonts have lost triacanthid and balistoid-ostracioid condition is the arch-shaped ligament between the third hypo- due to parallel evolution. If the triacanthodid con- branchials. Molids and diodontids have all three dition is plesiomorph for tetraodontiforms, then obliqui ventrales present, but only the third is the muscle may have become consolidated and present in tetraodontoids. Although this could be separated in those two lines independently. Thus, interpreted as a link between the tetraodontoids the common balistoid + ostracioid-gymnodont an- and the ostracioids, I feel that the loss of the cestor may well have retained the triacanthodid arch-shaped ligament and associated muscle fibers condition, and the gymnodonts themselves may is far more subject to the laws of parsimony than never have developed a sternobranchialis. This the postulated loss, in more than one phyletic line, view is expressed as an alternative in Figure 22c. of muscles whose absence or presence is more 20. PHARYNGOCLAVICULARIS EXTERNUS (Figure labile. 23a).—The muscle is consistently present in perci- 23. TRANSVERSI VENTRALES (Figure 24a).—Most forms, tetraodontiforms, and many other fishes. fishes possess two of these muscles, which connect Variations found in tetraodontiforms were the the ventromedial faces of the fourth and fifth large development of the muscle in ostracioids, its ceratobranchials across the ventral midline. There bifid nature in many monacanthids, the triple in- is little variation in them in tetraodontiforms. sertion in canthigasterids, and the medial position Transversus ventralis V consists of separate an- relative to the pharyngoclavicularis internus in terior and posterior sections in balistids. In mona- diodontids. Only the first has phylogenetic signifi- canthid, both muscles consist of two sections, which cance, the other variations being autapomorphic. cross over each other in the midline (the relative 21. PHARYNGOCLAVICULARIS INTERNUS (Figure position of one to the other being variable). In 2$b).—Normally, this muscle passes horizontally addition, there is a third transversus ventralis be- forward from the medial face of the deithrum to in- tween the ventral tips of the third hypobran- sert on the anteroventral surface of ceratobranchial chials—the muscle apparently being formed from 5. It is present in this form in triacanthodids, fibers contributed by obliquus ventralis III and ostracioids, and triodontids. In triacanthids, the rectus ventralis IV. Transversus ventralis IV is insertion is broad and by way of numerous tendons. single in the monacanthid Paraluteres. Canthi- In molids, the whole muscle is broad and diffuse. gasterids possess only the fifth muscle—it seems Balistoids are linked by the insertion of the muscle probable that the fourth has fused with that part on ceratobranchials 4 and 5. Diodontids and tetrao- of the pharyngoclavicularis internus that crosses dontoids are specialized in that some fibers insert the midline just ventral to the fourth ceratobran- on basibranchial 3. In addition, tetraodontoids chials. possess a section of the muscle which crosses the 24. RECTI VENTRALES (Figure 246).—These mus- midline. cles, which interconnect the basal regions of the 22. OBLIQUI VENTRALES (Figure 23c).—In most arches to one another, are of mosaic occurrence fishes, there are three pairs of these muscles, which in teleosts. The most frequent one (often called span the ventral articulations between the cerato- the interarcuales IH/IV) connects the anterior tip branchials and hypobranchials of the first three of ceratobranchial 4 to the ventral process of hypo- arches. Fibers of obliquus ventralis HI often at- branchial 3, usually with some fibers attaching to tach to the arch-shaped ligament between the ven- the arch-shaped ligament between the hypobran- tral processes of the third hypobranchials. This chial processes. Triacanthodids and triodontids condition is met with in triacanthoids and balis- have this muscle, as well as one between cerato- toids, with minor variations. In monacanthids, the branchial 1 and the dorsohyal. Among the molids, muscle of the second arch is absent; except in Mola has only the first rectus, while Ranzania also Oxymonacanthus; the third muscle is also absent possesses the second muscle. However, neither of in Anacanthus. In ostracioids, only obliquus ven- them possesses a rectus IV. In the triacanthids, tralis III is present—a synapomorph condition link- balistoids, and ostracioids, recti I, II, and IV are NUMBER 155 81 present. Rectus II attaches primarily to the urohyal linked by the fact that the origin of the muscles or to its antimere where the urohyal is small or is from the ventral surface of the prootic shelf, absent (ostracioids). Balistoids are apomorph in which appears to have evolved for this purpose. the possession of an extra section to the fourth Ostracioids are apomorph in that the enlargement rectus (the original section being absent in the of the shelf has allowed the anterior migration monacanthids Paraluteres, Oxymonacanthus, Pseu- of the branchial arches and the associated muscula- dalutarius, and Anacanthus, and poorly developed ture to a position beneath the orbit. Gymnodonts in most other monacanthid genera). Diodontids other than triodontids have the fourth levator and tetraodontoids possess four rectus muscles, the much enlarged. In molids, the entire origin of tetraodontoids being linked by the presence of a these muscles is from the dorsomedial face of the lateral division of the first rectus, which attaches hyomandibular. Diodontids and tetraodontoids to the lateral face of ceratobranchial 1 (canthi- have in common the even greater development of gasterids) or to the dorsomedial face of the hyo- levator IV, and the insertion of its posterior fibers mandibular (tetraodontids). Since the arch-shaped on the cleithrum. Origin here includes the pterotic ligament between the ventral processes of the third as well as the prootic (some fibers arising from hypobranchials is absent in the last three families the hyomandibular in diodontids). mentioned, the fourth rectus attaches to the pos- 27. LEVATORES INTERNI (Figure 25c).—The mus- terior face of the ventral process of hypobranchial cles generally originate from the prootic, and, in 3. perciforms, insert on infrapharyngobranchials 2 25. RECTUS COMMUNIS (Figure 25a ).—The mus- and 3. They are found as defined above in triacan- cle is found throughout the acanthopterygian fishes, thoids and balistoids. Levator III usually lies some- and usually passes from the urohyal to the antero- what lateral to levator II, and originates more ventral face of ceratobranchial 5. It is present in posteriorly. Ostracioids show an enormous devel- this condition in the triacanthoids, balistoids, opment of these muscles, whose origin has ex- ostracioids, and triodontids. In certain monacan- panded to include not only the prootic, but the thid genera, two small bundles of muscle fibers pterotic, epiotic, exoccipital, and (ostraciids) arise from the posterior face of transversus ven- parasphenoid as well. Much of the origin is thus tralis III and attach to the dorsal faces of the from the posterior face of the skull. In triodontids, tendons of this muscle. They apparently represent levator II inserts on infrapharyngobranchials 2 the remains of the ventral sections of rectus IV, and 3, while in molids levator III inserts on these whose posterior sites of attachment have shifted bones. Diodontids and tetraodontoids are linked from the anteroventral faces of the fourth cerato- by the posterior migration of the origin of levator branchials onto these tendons. The remaining fam- II to the pterotic, the fibers crossing medial to ilies of gymnodonts have lost the urohyal, and those of levator III at an angle of about 45 degrees. the muscle originates from the ventrohyal (except In the insertion of these muscles, tetraodontoids diodontids, where the origin is from the posterior are similar to molids, and diodontids more like face of basibranchial 1). Tetraodontoids are spe- triodontids, but the character state at this level cialized in that both the origin and insertion are is difficult to determine. Tetraodontoids possess a tendinous, the muscle fibers being confined to a medial connection of levator II to the lateral face belly in the middle of the length of the muscle. of transversus dorsalis III. 26. LEVATORES EXTERN I (Figure 256).—There 28. OBLIQUUS DORSALIS (Figure 26a).—Perci- are generally four of these muscles, which originate forms usually possess a single such muscle between from the prootic region of the skull beneath the the infrapharyngobranchial and epibranchial of hyomandibular fossa and insert on the dorsolateral the third arch. This condition is found in triacan- faces of epibranchials 1-4. Often, as in a number thoids, balistoids (except the monacanthid Alu- of tetraodontiform families, the manner of inser- terus), aracanids, triodontids, and diodontids. tion varies—e.g., the muscle of the first arch may Balistoids have developed an additional anterior send a slip of fibers to the second arch. Triacan- section which connects infrapharyngobranchial 2 thoids and triodontids exhibit the presumed plesio- to epibranchial 3. Molids show some familial morph condition. Balistoids and ostracioids are variation. In Mola, most of the fibers attach to 82 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY infrapharyngobranchial 2, while in Ranzania an sity, phylogenetic interpretation does not seem obliquus dorsalis II has developed between the warranted, except to state that the tetraodontids infrapharyngobranchial and epibranchial of the and canthigasterids are sister groups, since they second arch (possibly being derived from trans- share the unique anastomosis of fibers with those versus dorsalis II). Tetraodontoids are linked by of levator internus II. the fact that epibranchial 4 contributes partially 31. RETRACTOR DORSALIS (Figure 27a).—This (tetraodontids) or wholly (canthigasterids) to the muscle, which, in the plesiomorph condition, con- lateral site of the muscle's attachment. In both nects infrapharyngobranchial 3 to the anterior por- these families, fibers pass anteromedial to infra- tion of the vertebral column, is found as such in pharyngobranchial 2. triacanthoids and ostracioids. All other families have 29. OBLIQUUS POSTERIOR.—The muscle is present become specialized in the site of insertion, origin, in all the tetraodontiform families except the or both, and all (with one exception) in different molids (it is also absent in the monacanthid Ana- ways. In both balistoid families, the origin of the canthus). Frequently, fibers of this muscle inter- muscle has shifted anteriorly onto the lateral face mingle with those of the sphincter oesophagi or of the basioccipital, the insertion remaining on adductor V. Nothing could be deduced from it infrapharyngobranchial 3. phylogenetically. 32. ADDUCTORES (Figure 27b).—In most tele- 30. TRANSVERSI DORSALES (Figure 266).—In the osts, there are two of these muscles, adductor IV plesiomorph condition, there are two such muscles connecting the medial faces of the fourth epibran- which connect the epibranchial, infrapharyngo- chials and ceratobranchials and adductor V con- branchial, or both elements across the midline necting ceratobranchial 5 to epibranchial 5 or, between the second and third arches. Triacanthoids when that bone is absent (as in most fishes), to and balistoids exhibit this condition, except that epibranchial 4. This latter condition is met with most of the fibers of transversus II arise from the in triacanthoids, balistoids, triodontids, and molids. lateral face of the parasphenoid instead of from Ostracioids are linked by the loss of adductor V the midline. In the latter group, some fibers of (ostraciids having lost adductor IV as well). In transversus III may attach to epibranchial 4. Ostra- diodontids and tetraodontoids, each arch possesses ciids, on the other hand, have four such muscles, an adductor (except the tetraodontid Arothron, inserting on epibranchials 1-4. The fibers serving where adductores III and IV are absent). epibranchials 3 and 4 are poorly separated from 33. SPHINCTER OESOPHAGI (Figure 27c).—The nor- each other. Gymnodonts appear to show almost all mal condition of this muscle, where it simply en- possible sites of insertion. In triodontids, the an- circles the esophagus behind the gill arches, is terior muscle inserts on epibranchial 4 and infra- found in the triacanthoids. Balistoids are linked pharyngobranchial 2, the posterior one attaching by the possession of a bundle of fibers passing to epibranchial 3. Molids have an anterior mus- dorsomedially from epibranchial 4 to infrapharyn- cle passing posterolaterally from the parasphenoid gobranchial 3, while ostracioids possess a pair of to infrapharyngobranchial 2 (and epibranchial 3 ventral tendons attaching to the medial faces of in Ranzania). Diodontids possess the anterior mus- the cleithra, and lack a section of the muscle pass- cle to the third epibranchial and infrapharyngo- ing anterodorsal to the retractor dorsalis. Gym- branchial, while the fibers of the posterior muscle nodonts all have a longitudinal bundle of fibers insert on epibranchial 4. In tetraodontids, trans- passing posteriorly from at least one infrapharyn- versus II arises partially from the parasphenoid, gobranchial, and all of them except the triodon- inserting on infrapharyngobranchial 1, epibran- tids have a similar orientation of the anteroventral chials 1 and 2, the medial face of levator internus fibers. Diodontids, canthigasterids, and some tetra- II, with some fibers passing to the wall of the odontids possess a lateral bundle passing anterior opercular cavity. Transversus II inserts on epibran- to (and sometimes attached to) the epibranchial- chials 3 and 4. Canthigasterids have the anterior ceratobranchial region of the fourth arch. In addi- muscle to infrapharyngobranchial 2 and the medial tion, the ventral longitudinal bundle in these three face of levator internus II, and the posterior fibers families attaches to the basibranchials (to the to epibranchial 4. In view of this incredible diver- third in diodontids and tetraodontids, where it NUMBER 155 83 seems to be plesiomorph, and to the first in can- rather than on the shaft of the rays distal to their thigasterids). Tetraodontids and canthigasterids bases. The possible explanation is the same as have also developed a partial separation between that advanced for the analogous flanges on the the more dorsal longitudinal bundle and the ven- lateral halves of the rays, as discussed under the tral transverse component of the ventral region abductor superficialis of the muscle. 38. ADDUCTOR PROFUNDUS (Figure 29a).—Once 34. ABDUCTOR SUPERFICIALIS (Figure 28a).—The again, little can be said except that the diodontids fibers originate from the posterior face of the ala and tetraodontoids apparently share a common laminaris externus and the lateral cleithrum be- ancestor, the ventral six to nine tendons inserting hind it. Tendons insert on the anterolateral bases at the base of the ray instead of on the now non- of the principal fin rays. The only variation of existent ventral flanges. phylogenetic significance in tetraodontiforms is 39. ARRECTOR DORSALIS (Figure 296).—Origin found in diodontids and tetraodontoids. Here, of the muscle is primarily from the medial face some ventral (diodontids, canthigasterids) or all of the cleithrum, the insertion being on the medial the rays (tetraodontids) develop anterodorsal face of the vestigial fin ray. All gymnodonts show flanges from their bases, at the same time losing a considerable reduction of the muscle, which, in the posteroventral flange. Consequently, the ten- addition, becomes covered medially by the adduc- don of the abductor superficialis inserts on the tor muscles (the latter also occurs in aracanids, flange rather than on the anterodorsal base of the but here the arrector dorsalis is well developed). ray. Molids lack flanges on the fin rays, both lat- 40. CORACORADIALIS (Figure 29c,d).—The fibers erally and medially. of this muscle generally connect the posterodorsal The development of the anterodorsal flange ap- region of the coracoid to the posteroventral face pears to be related to the more vertical axis of of the fourth radial, in the apparently plesiomorph the fin, and would seem to allow better control condition. As such, it is found in triacanthodids, of the fin rays. monacanthids, aracanids (small), triodontids, and 35. ABDUCTOR PROFUNDUS (Figure 28&).—The tetraodontids. In balistids and triacanthids it is muscle arises from the coracoid and posterior clei- a well-developed muscle, originating mainly from thrum beneath the abductor superficialis, and in- the fascia over the medial face of the adductor serts on the tips of the posteroventral flanges of profundus; in ostraciids and molids it is absent; the principal and vestigial fin rays. In diodontids and in diodontids some of the dorsal fibers are and tetraodontoids it inserts on the posteroventral continuous with fibers of the obliquus inferioris. base of the shaft of some ventral or all the rays, The two former apparently synapomorph condi- for the reasons discussed above. Both monacanthids tions give rise to the phylogeny in Figure 29c. and diodontids lack a section of the muscle to the In the above analysis, the muscle's form was vestigial fin ray, but the former do not possess interpreted as linking the triacanthids and balis- the anterodorsal flanges of the ventral fin rays. tids as sister groups, and also linking the ostraciids 36. ARRECTOR VENTRALIS.—Origin is from the and molids together. To take the latter first, there lateral face of the cleithrum above and beneath is a reductive trend in the ostracioids—the muscle the abductor superficialis, the fibers inserting tendi- is small in aracanids and absent in ostraciids. nously on the anterolateral face of the medial Molids are unique in possessing an anterior prong half of the vestigial fin ray. No significant varia- of the postcleithrum which articulates with the tions were noticed in the tetraodontiforms exam- lateral surfaces of the radials. It is doubtful ined. whether the radials are very mobile under these 37. ADDUCTOR SUPERFICIALIS (Figure 28c).— circumstances (the apparent function of the cora- There is little variation in the form of this mus- coradialis), and it is not surprising that the muscle cle among tetraodontiforms. However, diodontids is absent in molids. and tetraodontoids have developed dorsally di- The large form of the muscle and its position rected flanges on the ventral six to nine rays, with medial to the adductor series in triacanthids and a concomitant loss of the ventral flanges. The balistids is difficult to explain in terms of the tendons serving these rays insert on the flanges proposed phylogeny. It should theoretically either 84 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY be present in this form in all the families (in ing in the decrease or loss of movement of the which case it has reverted to the ancestral condi- girdle. However, this seems to have taken place tion in four separate lineages—triacanthodids, for different reasons. In triacanthids, it is associated monacanthids, ostracioids, and gymnodonts), or with increased support for the pelvis and the pelvic else it has reached the same modified, enlarged spines. In ostracioids, there is no pelvis, and the condition separately in triacanthids and balistids. increased rigidity results from the close association The former seems unlikely. The latter is possible, of the pectoral girdle with the bony cuirass of the especially if one could offer an explanation as to exoskeleton, (Figure 30d). why this should occur, and what advantage (s) it 44. ARRECTOR DORSALIS PELVICUS (R.T.) (Figure confers. However, this cannot be done until the 31a).—This muscle, which is found in most perci- function of the muscle has been accurately deter- forms and a number of other fishes, originates mined, and its influence on the mechanism of the from the dorsolateral face of the pelvis and inserts pectoral fin defined. on the pelvic spine. It is present in triacanthoids, 41. ADDUCTOR RADIALIS (Figure 30a).—This but absent in all other tetraodontiform families. muscle is present in many perciform groups, and The great development of the muscle relative to is found in the triacanthoids, where it inserts on the presumed plesiomorph condition in the tri- a maximum of three rays. It is absent in all other acanthoids seems synapomorph. However, the tetraodontiform families. loss of the muscle may be synapomorph, or it may 42. PROTRACTOR PECTORALIS (Figure 30&).—The have occurred in more than one lineage. Although muscle is consistently present in most euteleosts the former view is tentatively followed here, the including tetraodontiforms (except the ostracioids). presence of a rather normal-looking pelvis and It arises from the pterotic and inserts on the anter- pelvic fin in the Eocene Eoplectus may mean that odorsal face of the cleithrum. In molids, it is it has been lost independently in the gymnodom reduced to a thin strip of fibers. In diodontids and and the balistoid-ostracioid lineages. tetraodontoids, the muscle inserts on the antero- 45. ARRECTOR VENTRALIS PELVICUS (R.T.) (Figure dorsolateral face of the cleithrum, and, in tetrao- 316/r).—The fibers of this muscle arise from the dontoids, part of the muscle attaches to the ventral face of the pelvis and insert on the pelvic supracleithrum. This seems to be associated with spine. It is extremely well developed in the tri- the operation of the postcleithral apparatus (see acanthoids (more so than in the presumed plesio- Winterbottom, 1971a, for details). morph condition exemplified by perciforms), much 43. LEVATOR PECTORALIS (Figures $0c,d).—In reduced (or absent—five genera of monacanthids) most perciforms and triacanthodids, the muscle in balistoids, and absent in all the other families. passes from the pterotic to the supracleithrum, The absence of the muscle in ostracioids and cleithrum, or both. Triacanthids possess a clei- gymnodonts needs reinterpreting, since it is ap- thrum which abuts firmly against the skull, and parently the result of parallel evolution. Ostraci- the muscle is absent. It is also absent (apparently oids have lost the pelvic girdle as a result of the for a similar reason) in ostracioids. In balistoids, development of the exoskeleton; in gymnodonts, it attaches to the anterodorsomedial rim of the its reduction and loss is either associated with the cleithrum, mainly in front of the supracleithrum. body form (molids) or with the inflatory abilities. In all gymnodonts, the fibers pass posteromedially It is not certain whether the muscle followed the to the supracleithrum to insert on the tip of the same reductive processes in the balistoids and posterodorsal extension of the cleithrum. In ostracioids (i.e., the balistoid condition may be molids, the muscle is much reduced, but still passes plesiomorph for the two groups), and therefore medial to the supracleithrum to attach to the clei- its presence cannot be used to link the balistids thrum (Figure 30c). and monacanthids. In the above analysis, the point of disagreement 46. ADDUCTOR SUPERFICIALIS PELVICUS (R.T.) is the linking of the triacanthids and ostracioids (Figure 32a,&).—In percifonns, the muscle inserts by virtue of the absence of the muscle. In both both on the pelvic spine and on the pelvic fin cases, the reason appears to be the very firm at- rays. A single slip to a pelvic fin ray was found tachment of the pectoral girdle to the skull, result- only in the triacanthodid Parahollardia. The sec- NUMBER 155 85 tion to the pelvic spine is very well developed in odontoids, possibly because the fibers become as- triacanthoids, much reduced and passing through sociated with the transversus cutaneous system. In a bony canal in balistoids (absent in five monacan- addition, there may well be an elusive, functional thid genera), and absent in all the other families. reason. In all cases where the muscles are reduced In the reinterpretation of this muscle, the rea- or absent, the fin is short based, and, when used in soning used for the arrector ventralis pelvicus ap- locomotion, it flaps synchronously from side to side plies in the same way, except that the passage of with the anal fin (Figure 336). the tendon through a narrow, bony canal in 49. ERECTORES DORSALES (R.T.pars) (Figure 34a, balistoids is considered to be synapomorph, and 6).—Generally, these muscles are fairly small, and thus links the two families as sister groups. insert on the anterolateral base of each spine and 47. ADDUCTOR PROFUNDUS PELVICUS (R.T.) (Fig- fin ray. In the spiny dorsal fin, triacanthodids pos- ure 32c).—Among the perciforms, this muscle in- sess up to six, triacanthids up to five, while there serts on the bases of the pelvic fin rays. It is found are two in balistoids and triodontids. Since there only among members of the Triacanthodidae (but is overlapping variation in the first two families, is absent in seven of the genera—see Winterbot- these muscles cannot be used to differentiate be- tom, 1970). tween them. They are absent in all other families Since this muscle is very much reduced, and (which do not possess a spiny dorsal fin). Those to frequently absent, in the triacanthodids, the trend the fin rays are much enlarged in ostracioids, and in its reduction seems obvious. It is, therefore, are enormous, being partially to wholly fused to better to assume that no phylogenetic information the depressors in molids (Figure 34a). can be deduced from it at present. In the initial analysis, the ostracioids were linked 48. INCLINATORES DORSALES (R.T.pars) (Figure with the gymnodonts (minus triodontids). This $5a,b).—In the apparently plesiomorph con- is apparently the result of a reductive trend, since dition, these muscles originate from the fascia the former group would have little use for a spiny overlying the epaxialis and insert on the lateral dorsal fin (being encased in thick, bony armor) bases of the spines and rays of the dorsal fin. They and, as in the gymnodonts mentioned, loss of the are present to the spines only in the triacanthoids, spiny dorsal fin may well be associated with the where triacanthodids may have up to six (inserting posterior migration and small size of the soft dorsal on the spines and the basal pterygiophores) and fin (Figure 346). triacanthids possess two to four (inserting only on 50. DEPRESSORES DORSALES (R.T.pars) (Figure the pterygiophores). The muscles attach to all the 54c,d).—In the most plesiomorph condition, each dorsal fin rays in triacanthoids, balistoids, and spine and fin ray receives a depressor muscle. There triodontids. They are much reduced or absent in are up to six in the spiny dorsal fin of triacantho- all the other families (0-1 in aracanids and dio- dids, up to five in triacanthids, three in balistids, dontids, absent in molids and ostraciids, 2 to the two in most monacanthids, and one in triodontids. posterior rays in canthigasterids, and continuous The muscles to the fin rays are enlarged in os- with the transverse cutaneous attaching to the tracioids, and fused to the erectors in molids. pterygiophores in tetraodontids—see Figure 33a). Once again, a reductive trend is apparent. The In the initial interpretation, the ostracioids were postulated ancestor of the balistoids + ostracioids linked with the gymnodonts minus triodontids. + gymnodonts must have possessed a spiny dorsal This was based on the loss or reduction of the fin. This is retained symplesiomorphly in the muscles in these families. In ostracioids, the loss balistoids, but lost in the ostracioids (the same seems to be associated with the large development reasoning used for the erectores dorsales applies of the erector and depressor dorsales muscles, and here). Triodontids are plesiomorph in the reten- the reduction and freeing of the dorsal moiety of tion of a single depressor to the spiny dorsal fin, the epaxialis (their normal site of origin). This is the fin being absent in all the other gymnodonts. taken a step further in molids, where the epaxialis 51. INCLINATORES AN ALES (Figure 35a).—These almost entirely disappears, and the erector and de- muscles generally insert on the lateral bases of all pressor muscles are enormously enlarged. Inclinators the anal fin rays. In all families (except most are also reduced or absent in diodontids and tetra- ostracioids) there is a residual inclinator attaching 86 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY to the distal face of the first basal pterygiophore. where these are separate. In the tetraodontiforms, It is well developed in balistoids, and attaches to its presence is apomorph, and it is found only in the tip of the postcleithrum in most tetraodontoids. the balistoids. These two families are the only There are 0-2 normal inclinators in diodontids, ones where the anterior migration of the spiny and four in canthigasterids. Molids show two ex- dorsal fin has separated it from the soft dorsal fin. tremes in the form of the residual inclinator. In In all other forms (apart from the triacanthoids), Ranzania, the muscle attaches to the base of the the spiny dorsal fin has apparently migrated pterygiophore and fades into the subdermal con- posteriorly. nective tissue, while in Mola it attaches to the tip 56. SUPRACARINALIS POSTERIOR.—In the normal of the postcleithrum but does not reach the teleostean condition, this muscle connects the last pterygiophore. It is therefore described as part of pterygiophore of the soft dorsal fin with the neural the obliquus inferioris, for it seems as likely to be spine of the last complete vertebra, attaching to part of this muscle as to be an inclinator. the tips of the neural spines of the intervening Inclinators to the fin rays are reduced (tetrao- vertebrae. The variations found in tetraodonti- dontoids) or absent (most ostracioids, some dio- forms are both autapomorphs—the muscle is absent dontids and molids). In view of the intrafamilial in molids, and attaches to the dorsomedial faces variation, no interpretation will be placed on this of the bifid neural spines in diodontids. facet of the distribution of the muscle. 57. INFRACARINALIS ANTERIOR (R.T.) (Figure 52. ERECTORES ANALES (Figure 35&).—The only 37a,b).—In the apparent plesiomorph condition variation found in this muscle complex is in ostra- (such as is found in most teleosts), the fibers cioids (where the origin is more lateral owing to originate from the posteroventral regions of the the divergent nature of the pterygiophores) and cleithrum and coracoid and pass posteriorly to in molids (where the muscles are enormously de- insert on the anteroventrolateral face of the pelvis. veloped, and partially to wholly fused with the Triacanthoids are apparently synapomorph in that depressores anales). the muscle has partially fused with both the ar- 53. DEPRESSORES ANALES (Figure 35

(65a); epaxialis (68b); and obliquus superioris BRANCH POINT 4 (70a). Of the 24 muscles listed, 11 have become syn- The hypothesis presented here is that the Trio- apomorph in the balistoids + ostracioids and 15 in don tidae form the sister group of the other gymno- the gymnodonts, which indicates that the former dont families. The evidence is provided by the are slightly more plesiomorph than the latter. following synapomorphs: section A 1 (lb), section Given, then, that the ancestral species at Branch A 2 (2b), and section A Oligocene. formation, a discussion of such forms as have been The Protacanthodinae (Tyler, 1968:237-238) ap- discovered is essential to any study purporting to parently represent an offshoot of the triacanthodid- be phylogenetic. Much of the information comes like stock leading toward the triacanthids. This from Tyler's (1968) excellent summary of fossil agrees basically with Tyler's (1968) concept, except triacanthoids. that I would derive the balistoids (and the other Tyler (1968) placed the Spinacanthinae as a nontriacanthoid groups) from a slightly more basal subfamily of the Triacanthodidae, and considered position, before the separation of the protacan- them to be the most basal group of the order (i.e., thodins. The reduction of all but the first dorsal the sister group of all other tetraodontiforms), spine and the elongate soft dorsal and anal fins mainly on the basis of the extremely well-developed, of protacanthodins are apparently synapomorphic six-spined dorsal fin. The two monotypic genera, with the triacanthid condition, while they share the Spinacanthus [S. cuneiformes (Blainville)] and form and configuration of the pelvis with tria- Protobalistum [P. imperiale (Massolonga)], are canthoids generally. Tyler's placement of Acan- from the upper Eocene of Monte Bolca. Tyler's thopleurus and Marosichthys in close proximity objections to these fishes having balistoid affinities to the triacanthids seems undeniable. are the six-spined dorsal fin and the short-based The last fossil which seems pertinent to this soft dorsal and anal fins. The former objection discussion is Cryptobalistes brevis (Rath), placed falls away, since Eoplectus has at least five (and in its own subfamily by Tyler (1968:243-249). Of probably six) dorsal spines, and this must have this extremely confusing Oligocene fossil, Tyler been the condition of the common ancestor of (1968:247) has said, "Cryptobalistes simply is not the gymnodonts and balistoids + ostracioids. The on a lineage which led to any Recent balistids or fact that the spines are better developed than in to any other Recent fishes, even though it has any other tetraodontiform seems irrelevant, for paralleled the fundamental trends within the it could just as easily be an apomorph condition Triacanthidae-Balistidae line of evolution." The as a plesiomorph one, and besides, dorsal spine reader is referred to Tyler for the reasoning behind reduction is apparent independently in the tri- this statement. Concerning the phylogenetic posi- acanthoid, gymnodont, and balistoid-ostracioid tion of the fossil, the apparently synapomorph lines, the spines being better developed in the fossil nature of the basin-like pelvic shaft with that of examples. Short-based soft dorsal and anal fins, Recent triacanthodins seems to be the only per- while not found in Recent balistoids, are present in tinent character, and the subfamily is tentatively ostracioids and gymnodonts (with a few secondary placed as the extinct sister group of the triacantho- exceptions), and the character is, in the Spinacan- dins (although this must imply that the hollardiins thinae, plesiomorph (i.e., the state of the fins is diverged somewhat before the Oligocene from the as one would postulate for the ancestor at Branch triacanthodins). Point 2), the longer based balistoid fin thus being Vom Rath's specimens (one in counterpart, the synapomorphic. Evidence that the Spinacanthinae other single) are no longer at Bonn (pers. comm. belong to the balistoid-ostracioid lineage rather H. K. Erben, Institute fur Palaontologie, Bonn) than the gymnodont line is provided by the and every effort to relocate them for reexamination elongate ethmoid region between the upper jaw has failed. The description given by vom Rath NUMBER 155 97

(1859) is sparse, but the figure looks much like of the balistoids + ostracioids. Tyler hinted a balistid with a triacanthoid pelvis and pelvic (1962c:800) that the gymnodonts share a common fin. Should Cryptobalistes ultimately prove to be ancestor with the triacanthoids + balistoids + (in spite of Tyler's statement quoted above) a ostracioids (= his Sclerodermi), but in a later paper balistoid sister group, it may be that the pos- (1970a: 29) he states that ". . . the Triodontidae terodorsal flange of the balistoid pelvis represents have on other grounds been considered as the the posterior elongation in the process of resorb- most generalized family of the suborder Gymno- tion. However, such a postulate raises certain dontes, the latter probably derived from the basal difficulties. It seems fairly clear that Cryptobalistes Sclerodermi." By "basal Sclerodermi," Tyler pre- would be on the balistoid line rather than the ostra- sumably means the triacanthoid fishes, but he does cioid or ostracioid + balistoid one if it should not commit himself as to whether the gymnodonts prove in any way related to these forms. The 11 syn- are derived from a pretriacanthodid, pretriacanthid, apomorphs listed under "Branch Point 2" which or prebalistoid + ostracioid stock. Since the latter link the ostracioids and balistoids seem to be case is the one presented here, and since no formu- powerful ones, and the sister group status of these lated arguments or alternatives have been proposed two groups is extremely likely. If so, then to this hypothesis as yet, further discussion awaits balistoids and ostracioids separated from their new evidence or reinterpretation of the old. common ancestor well before the Eocene, for One last point may be mentioned in connection ostracioid fossils, apparently fully differentiated, are with the postulated sister group status of the known from Monte Bolca. Until more specimens balistoids + ostracioids and the gymnodonts— of Cryptobalistes are found, or the original speci- their normal method of locomotion involves the mens rediscovered, it seems safer not to form any synchronous, side-to-side flapping of the dorsal and firm ideas on the position of the genus, and I have anal fins. The possible exception to this is Triodon, accordingly followed one of Tyler's (1968) inter- which has small dorsal and anal fins and a some- pretations. what lunate caudal fin. This is probably a sec- There are thus two main likely points of dif- ondary condition (particularly since the caudal ference from the interpretation of tetraodontiform fin of ^Eoplectus and nearly all other gymnodonts phylogeny postulated here. One of these involves is rounded). It therefore seems likely that this the sister group status of the triacanthodids and method of locomotion was developed by the postu- triacanthids, which has been extensively discussed, lated common ancestor of these fishes. It also and, although I believe the triacanthoid pelvis appears likely that an in-depth study of the to be a significant synapomorph, it is quite possible anatomical systems associated with these fins will that the triacanthodids separated off first from the reveal further synapomorphs linking the two ancestral stock (thus forming the sister group of superfamilies, particularly in the fields of neurology, all the other tetraodontiform families) and that osteology, and myology. they were followed soon after by the triacanthids (which thus would form the sister group of all the My own views on the phylogenetic positions of nontriacanthoid families). The other point is my the fossil groups discussed above are incorporated placement of the gymnodonts as the sister group in Figure 47.

A New Classification of the Tetraodontiformes

The importance of the distinction between doing this. In a hierarchical phylogenetic classifica- phylogeny and systematics has already been em- tion, no taxon (be it species, genus, or family) phasized, and the ideal system would appear to can contain more than two subtaxa, since the be one of phylogenetic systematics, i.e., a classifica- moment a third subtaxon is added, a new rank tion of animals in a hierarchical manner such that has to be imposed in order to preserve the phy- the phylogeny of the components is accurately logenetic information. The use of the classical, represented. There is, at present, no method of hierarchical system results in a number of, say, 98 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY families within a superfamily, and it is impossible Taxon A (1) Taxon E (3.2) Taxon I (5.1) to deduce the phylogeny of the contained units. Taxon B (2) Taxon F (3.2) Taxon J (5) Taxon C (3.1) Taxon G (4) Taxon K (6) The establishment of a classification based on Taxon D (3.1) Taxon H (5.1) "coordinate" units poses a number of apparent The main difficulty with this system is that one problems at the present time. Since no taxon can is forced to make arbitrary decisions as to what be the evolutionary equivalent (in the strict sense constitutes any given hierarchical level. It is of the word) of any other, it seems that one can- obvious that definitions based on morphological not imply equivalency in the hierarchical ranking dissimilarity will only lead to confusion if used of taxa. Logical extension of the Hennigian in a phylogenetic classification, since they ignore methods in these directions appears to lead ulti- Hennig's "Rule of Deviation" (1966:207). Hen- mately to a ranking system of classification without nig's suggestion that hierarchical levels be de- the imposition of arbitrary levels of implied termined by their geological age is more logical, equivalency (such as family and genus). Although even if still arbitrary and sometimes clouded by the naming of all branch points may be the final our lack of knowledge. By suggesting only broad result, it appears to be unnecessary for the present stages, Hennig (1966:185) "makes it possible for (unless an immediate purpose is served by doing the time being to reserve the well known category so). The hierarchical classification presented here designation ('class' and 'subclass') for the more is, therefore, provisional, pending the outcome of important and morphologically isolated groups. the attention currently directed to the discussion All possibilities of future refinement of the method of these problems. The relative time of branching remain open." This view can apply equally well (i.e., the phylogeny) has been indicated by the to lower categories, and there seems to be no reason development of the following method, which allows why the phylogeny should not be indicated (where for the inclusion of more than two subgroups in necessary) by the numerical suffix method outlined any group, the phylogeny of the contained units above. Where it has not proved possible to being indicated by a simple system of numbers elucidate the phylogeny, there is no cause to (derived from the phylogenetic tree) behind each reject the system prevailing prior to the in- taxon. The first taxon to diverge is suffixed by 1, vestigation. That changes in systematic rank may the second by 2, and so on. Should any of these result in applying phylogenetic findings to the lines subsequently radiate, the resulting units could classification is irrelevant, since hierarchical levels be labeled (if the radiation took place in the are completely arbitrary, and have, in the past, lineage labeled, say, 2) 2 in the case of sister been based solely on the amount of similarity- groups, or, if there were three resulting lines, 2, dissimilarity, without regard for the state of the 2.1, 2.1, where the units labeled 2.1 are sister characters. groups, and together form the sister group of 2. The above would only arise, of course, if the The following phylogenetic tree is drawn up taxa were to be treated as equivalent hierarchical from the myological evidence presented in this units. It can be seen from the following example paper, to which has been added my views on the phylogenetic position of the various fossil groups. (which contains more units than would normally Where alternative branch points are shown, the be encountered) that each prime number con- dashed line represents the less likely possibility. tains the taxon (or taxa) whose sister group is The numbers (1 and 2, 3 and 4) are used to in- composed of all taxa with a larger prime number dicate the alternatives in the hierarchical classifica- (Figure 46). tion presented after Figure 47. Using the method of interpretation put forward The hierarchical arrangement of these groups above, we get: thus becomes: NUMBER 155 99

1. Suborder Triacanthoidei OR 2. Suborder Triacanthodoidei (1) Family Triacanthodidae Family Triacanthodidae Subfamily Triacanthodinae (1) Subfamily Triacanthodinae (1) Subfamily? tCryptobalistinae (1) Subfamily? tCryptobalistinae (1) Subfamily Hollardiinae (2) Subfamily Hollardiinae (2) Family Triacanthidae Suborder Triacanthoidei (2) Subfamily Triacanthinae Family Triacanthidae Subfamily tProtacanthinae Subfamily Triacanthinae Subfamily tProtacanthinae Suborder Tetraodontoidei Suborder Tetraodontoidei (2) Superfamily Balistoidea Family tSpinacanthidae (1) Family Balistidae (1) Subfamily Balistinae Subfamily Monacanthinae Family Ostraciidae (2) Subfamily Aracaninae Subfamily Ostraciinae Superfamily Tetraodontoidea 3. Family fEoplectidae (1) OR 4. Family tEoplectidae (1) Family fZignoichthyidae (2) Family Triodontidae (2) Family Triodontidae (3) Family Tetraodontidae (S) Family Tetraodontidae (4) Subfamily Tetraodontinae Subfamily Tetraodontinae Subfamily Canthigasterinae Subfamily Canthigasterinae Family Diodontidae (3) Family Diodontidae (4) Family Molidae 4) Family Molidae (5) The following systematic and hierarchical changes Tetraodontidae,1 and that the fEoplectinae of are suggested in order to bring tetraodontiform Tyler (1973) be transferred from the Triacan- classification into line with the phylogenetic rela- thodidae to the superfamily Tetraodontoidae, and tionships postulated in this paper: raised to familial rank, in keeping with its 1. That two suborders be provisionally recog- phylogenetic position. nized (Triacanthoidei and Tetraodontoidei), the It seems unlikely, at this stage of our knowledge, first containing two families, the second consisting that the Tetraodontiformes is phylogenetically of two superfamilies. equivalent to the Perciformes, and the sister group 2. That a superfamily Balistoidea be recognized, of the former order (as alluded to in the text) containing the families f Spinacanthidae, Balistidae, probably lies among the acanthurid-siganid-chaeto- and Ostraciidae. dontid fishes. Since these latter groups are 3. That the families Monacanthidae and Ara- presently classified in the order Perciformes, it canidae be treated as subfamilies of the Balistidae may well prove necessary to place them in the and Ostraciidae, respectively, and that the fSpin- tetraodontiforms when sufficient evidence to war- acanthinae of Tyler (1968) be removed from the rant such a change has been assembled. This, Triacanthodidae and placed as the sister group however, must await further investigations. of the Balistidae, being given equal rank (as fSpinacanthidae). 1 In this paper, the conventional view of the canthigasterids 4. That a superfamily Tetraodontoidea be forming the sister group of the tetraodontids has been fol- lowed. Recently, Tyler (ms) has demonstrated that the sister formed to contain the families fEoplectidae, Tri- group of Canthigaster consists of at least Carinotetrodon, and odontidae, Tetraodontidae, Diodontidae, Molidae, quite possibly Monotreta and Cltelonodon as well. No rep- and—should fZignoichthys prove to be farther resentatives of the latter three genera were dissected for the along the gymnodont line than -fEoplectus, as purposes of this study. Should Tyler's view prove to be cor- hinted by Tyler (1973) and tentatively preferred rect, the subfamilial rank afforded to Canthigaster .above would have to be rescinded, and the genus would become here—the fZignoichthyidae. completely incorporated within the Tetraodontidae. A de- 5. That the family Canthigasteridae be re- tailed study of the generic interrelationships of this family duced to subfamilial rank within the family is badly needed. 100 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

Literature Cited

Alexander, R. McN. Synopsis of the Genera. Annals and Magazine of 1967. The Functions and Mechanisms of the Protrusile Natural History, series 11; 10:1-18. Upper Jaw of Some Acanthopterygian Fish. Journal 1951. The Ocean Sunfishes (Family Molidae). Bulletin of Zoology, London, 151 (1) :43-64. of the British Museum (Natural History), Zoology, 1968. Animal Mechanics, x + 346 pages. Seattle: Uni- 1 (6) :89-121. versity of Washington Press. Gabriel. M. L. Berry, F. H.. and W. J. Baldwin 1940. The Inflation Mechanism of Spheroides maculatus. 1966. Triggerfishes (Balistidae) of the Eastern Pacific. Biological Bulletin, 79(2) :372. Proceedings of the California Academy of Sciences, Gans, C, and W. J. Bock series 4, 34 (9): 429-474. 1965. The Functional Significance of Muscle Architec- Borcea, I. ture: A Theoretical Analysis. Ergebnisse Anatomic 1906. Observations sur la musculature branchiostegale des und Entwicklungsgeschicte, 38:116-142. Teleosteens. Annales scientifiques Universitatea Geoffroy Saint-Hilaire, E. Jassy (Roumania), 4:203-225. 1818. Philosophie Anatomique. xxxix + 517 pages. Paris. Breder, C. M., and E. Clark Geoffroy Saint-Hilaire, I. 1947. A Contribution to the Visceral Anatomy, Develop- 1827. Histoire naturelle des poissons du Nil. Volume ment and Relationships of the Plectognathi. Bulle- 1 (part 1): 141-579 in Description de I'Egypte, . . . tin of the American Museum of Natural History, Histoire Naturelle. Paris. 88 (5) :287-32O. Goodsir, J. Brundin, L. 1841. On Certain Peculiarities in the Structure of the 1966. Transantarctic Relationships and Their Significance Short Sunfish (Orthagoriscus mola). Edinburgh as Evidenced by Chironomid Midges. Kunglia New Philosophical Journal, 30:188-194. Svenska vetenskapsakademien Handlingar, 11 (1): Greene, C. W. 1-472. 1913. An Undescribed Longitudinal Differentiation of the 1968. Application of Phylogenetic Principles in System- Great Lateral Muscle of the King Salmon. Anatom- atics and Evolutionary Theory. Pages 473-495 in T. ical Record, 7:99-101. 0rivg, editor, Current Problems of Lower Vertebrate Greenwood, P. H., G. S. Myers, D. E. Rosen, and S. H. Weitz- Phytogeny, Stockholm: Almqvist and Wiksell. man Cleland, J. 1967. Named Main Divisions of Teleostean Fishes. Pro- 1862. On the Anatomy of the Short Sun-fish (Orthagoris- ceedings of the Biological Society of Washington, cus mola). Natural History Review (London), new 80:227-228. series, 2:170-185. Greenwood, P. H., D. E. Rosen, S. H. Wekzman, and G. S. Cuvier, G. L. C. F. D. Myers 1817. La regne animal. Volume 2, 532 pages. Paris. 1966. Phyletic Studies of Teleostean Fishes, with a Pro- Dietz, P. A. visional Classification of Living Forms. Bulletin of 1914. Beitrage zur Kenntnis des Kiefer- und Kiemenbog- the American Museum of Natural History, 131 (4): enmuskulatur der Teleostier, I: Die Kiefer- und 339-456. Kiemenbogenmuskeln der Acanthopterygier. Mit- Gregory, W. K., and H. C. Raven teilungen Statzione zoologica Neapel, 22 (4) :99-162. 1934. Notes on the Anatomy and Relationships of the Edgeworth, F. H. Ocean Sunfish (Mola mola). Copeia, 4:145-151. 1935. The Cranial Muscles of the Vertebrates, viii + 493 Grenholm, A. pages. England: Cambridge University Press. 1923. Studien iiber die Flossenmuskulatur der TeleostieT. Fraser-Brunner, A. 296 pages. Doctoral dissertation, University of Upp- 1935. Notes on the Plectognath Fishes, II: A Synopsis of sala. the Genera of the Family Ostraciontidae. Annals Hennig, W. and Magazine of Natural History, series 10, 16:313- 1966. Phylogenetic Systematics. 263 pages. Urbana: Uni- 320. versity of Illinois Press. 1941a. Notes on the Plectognath Fishes, VI: A Synopsis of Kesteven, H. L. the Genera of the Family Aluteridae, and Descrip- 1943. The Evolution of the Skull and Cephalic Muscles, tions of Seven New Species. Annals and Magazine I: The Fishes. Memoirs of the Australian Museum, of Natural History, series 11, 8:176-199. 8 (2) :65-132. 1941b. Notes on the Plectognath Fishes, VII: The Aracan- Le Danois, Y. idae, a Distinct Family of Ostraciontoid Fishes, 1955. Sur le remaniement du sous-ordre des poissons with Descriptions of Two New Species. Annals and Plectognathes et la definition d'un nouveau sous- Magazine of Natural History, series 11, 8:306-313. ordre: Les Orbiculates. Comptes Rendus de V 1943. Notes on the Plectognath Fishes, VIII: The Classi- Academic Sdentifique, Paris, 240:1933-1934. fication of the Suborder Tetraodontoidea, with a 1958. Systeme musculaire. Volume 13 (1) :78S-813 in NUMBER 155 101

Traite de Zoologie. P.-P. Grassc, editor. Paris: American Museum of Natural History, 141 (4) :475- Masson et Cie. 552. 1959. Etude ost£ologique, myologique et syste'matique Nishi, S. des poissons du sous-ordre dcs Orbiculatcs. An- 1938. Muskeln des Rumpfes. Volume 5:351-446 in Bolk, nales de I'lnstitut Oce"anographique, nouvelle serie, Goppert, Kallius, and Lubosch, editors, Handbuch 36(1) :1-27S. der Vergleichenden Anatomie der Wirbeltiere. 1961. Remarques sur les poissons orbiculates du sous- Berlin and Vienna: Urban and Schwarzenberg. ordre des Ostracioniformes. Memoirs du Museum Owen, R. National d'Histoire Naturelle, Serie A. 29 (2) :207- 1866. On the Anatomy of Vertebrates, Volume 1: Fishes 358. and Reptiles, xlii + 650 pages. London. Lubosch, W. Parr, A. E. 1917. Vergleichende Anatomie der Kaumuskeln der Wir- 1927. On the Functions and Morphology of the Post- beltiere, in fiinf Teilen. Zweiter Teil. Die Kaumus- clavicular Apparatus in Spheroides and Chilo- keln der Knochenfische, Ligamentum maxillo- mycterus. Zoologica, 9(5): 245-269: mandibulare, Wesen der Streptognathie und Genese Randall, J. E. des Squamoso-dentalgelenkes. Jenaische Zeitschrift 1967. Food Habits of Reef Fishes of the West Indies. fUr Naturwissenschaft, 54:277-332. Studies in Tropical Oceanography, 5:665-847. 1929. Vergleichende Anatomie der Kaumuskeln der Raven, H. C. Wirbeltiere, in fQnf Teilen. Zweiter Teil (Fortset- 1939a. Notes on the Anatomy of Ranzania truncata, a zung): Die Kaumuskeln der Teleosteer. Gegenbaurs Ptectognath Fish. American Museum Novitates, Morphologisches Jahrbuch, 61:49-220. 1038:1-7. 1938. Muskeln der Kopfes: Viscerate Muskulatur, B: 1939b. On the Anatomy and Evolution of the Locomotory Teleosteer. Volume 5: 1011-1024, in Bolk, Goppert, Apparatus of the Nipple-tailed Ocean Sunfish Kallius, and Lubosch, editors. Handbuch der (Masturus lanceolatus). Bulletin of the American Vergleichenden Anatomie der Wirbeltiere, Berlin Museum of Natural History, 76(4): 143-150. and Wien: Urban and Schwarzenberg. Regan, C. T. Maurer, F. 1903. On the Classification of the Fishes of the Suborder 1912. Die ventrale Rumpfmuskulatur der Fische. Jenais- Plectognathi, with Notes and Descriptions of New che Zeitschrift fUr Naturwissenschaft, 49(1) :I—118. Species from Specimens in the British Museum Col- Mayr, E. lection. Proceedings of the Zoological Society of 1969. Principles of Systematic Zoology, xi + 428 pages. London, 1902 (2) :284-303. New York: McGraw-Hill Inc. Rensch, B. Meckel. J. F. 1959. Evolution Above the Species Level, xvii + 419 1828. System der Vergleichenden Anatomie. Third part. pages. John Wiley and Sons Inc. xii + 670 pages. Halle. Rosen, D. E., and C. Patterson Monod, T. 1969. The Structure and Relationships of the Paracan- 1950. Sur trois dispositifs de "verrouillage" osseux chez thopterygian Fishes. Bulletin of the American Mu- des poissons. Comptes Rendus Premiere Conference seum of Natural History, 141 (3) :357-474. Internationale Africanistes de I'Ouest, Dakar, In- Rosen, N. stitut Francois d'Afrique Noire, 1:211-215. 1913. Studies on the Plectognathi, 4: The Body Muscles. 1959a. Sur le syste'me de "verouillage" de la premiere Arkiv for Zoologi (Stockholm), 8(18): 1-14. epinl dorsale dcs Balisddae. Proceedings of the Sarkar, S. P. XVth International Congress of Zoology, section 5, 1960. A Study of Functional Morphology of the Head of paper 25:409-412. an Indian Puffer Fish, Sphaeroides oblongus 1959b. La nageoires ventrales du Balistes forcipaius Gmelin (Bloch). 119 pages. Doctoral dissertation, Univer- 1789. Bulletin de VInstitut Francois d'Afrique sity of Leiden. Noire, series A, 21 (2) :695-709. S0rensen, W. E. 1959c. Notes sur l'lpine latlro-caudale et la queue de 1883. Om oppustningssaeken hos Tetrodon og aanded- YAcanthurus monroviae. Bulletin de I'lnstitut raettet hos Ctarias. Naturhistorisk Tidsskrift Francois d'Afrique Noire, series A, 21 (2) :71O-734. (Copenhagen), 13 (3) :374-414. 1960. La premiere nageoire dorsale du Balistes forcipatus 1884. Om lydorganer hos fiske. En physiologisk og com- Gmelin. Bulletin de I'lnstitut Francais d'Afrique parativanatomisk unders0gelse. 245 pages. Doctoral Noire, series A, 22(3): 1008-1012. dissertation, University of Copenhagen. Myers, G. S. Souch£, G. 1958. Trends in the Evolution of Teleostean Fishes. 1932a. Morphologie comparative des muscles ellvateurs de Stanford Ichthyological Bulletin, 7 (3) :27-30. la mandibule chez les poissons. Memoires de la Nelson, G. J. Socie"ti des Sciences physiques et naturelles de 1969. Gill Arches and the Phytogeny of Fishes, With Notes Bordeaux, series 7, 3 (2): 1-292. on the Classification of Vertebrates. Bulletin of the 1932b. Contribution a l'ltude anatomique et comparative 102 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

du muscle ptirygo-maxillare chez les poissons 122 (3) :87-124. osseux. Bulletin de la Station Biologique d'Arca- 1973. A New Species of Triacanthodid Fish (Plectog- chon, 29:41-82. nathi) from the Eocene of Monte Bolca, Repre- 1932c. Contribution a l'etude du muscle adducteur de l'arc senting a New Subfamily Ancestral to the Triodon- palatin chez Lophobranches et les Plectognathes. tidae and the Other Gymnodonts. Memoria di Bulletin de la Station Biologique d'Arcachon, 29 Museo Civico di Storia Naturale di Verona (in (2) :267-269. press), 1935. Nouvelles reserches sur l'adducteur de Tare palatin nis. The Osteology, Phytogeny, and Higher Classification chez les poissons. Bulletin de la Station Biologique of the Fishes of the Order Plectognathi. d'Arcachon, 32:53-72. van Dobben, W. H. Takahasi, N. 1935. Uber die Kiefermechanismus der Knochenfische. 1917. On the Homology of the Median Longitudinal Archives nierlandaises de Zoologie, 2(1): 1-72. Muscles—Supracarinalis and Infracarinalis—with van Roon, J. M. the Fin Muscles of the Dorsal and Anal Fins, and 1942. Some Additional Notes on External Features and Their Function. Tokyo Journal of the College of on the Jaw Muscles of Orthagoriscus mola (L.). Agriculture, 6 (3): 199-213. Zoologische Mededeelingen, 23:313-317. Thilo, O. van Roon, J. M., and J. J. ter Pelkwijk 1940. Mechanism of Jaw and Body Muscles of Orthagoris- 1879. Die Sperrgelenke an den Stacheln einiger Welse, cus mola. Zoologische Mededeelingen, 22:65-75. des Stichlinges und des Einhornes. 21 pages. Doc- vom Rath, J. J. G. toral dissertation. University of Dorpat [Tartu], 1859. Bcitrag zur Kenntniss der fossilen Fische des Estonia. Plattenberges im Canton Glarus. Zeitschrift der 1896. Die Umbildungen an den Gliedmasscn des Fische. Deutschen geologischen Gesellschaft, 11:108-132. Gegenbaurs Morphologische Jahrbuch, 22:287-355. Wahler, G. v.. and H. v. Wahler 1899. Die Entstehung der Liiftsacke bei den Kugdfischen. 1964. Eine bemerkenswerte Anordnung von Schuppen- Anatomische Anzeiger, 16:73-87. fortsatzen beim Knochenfisch Batistes bursa, und 1914. Die Vorfahren der Kugelfische. Biologische Central- ihre hydrodynamische Erklaring. Naturwissen- blatt, 34:523-545. schaften, 51:45. 1920. Das Maulspitzen der Fische. Das Enstehen und Weitzman, S. H. Vergehen seiner Mechanik. Biologische Centralblatl, 1964. Osteology and Relationships of South American 40:216-238. Characid Fishes of the Subfamilies Lebiasininae Tyler. J. C. and Erythrininae With Special Reference to Sub- 1962a. The General Osteology of Representative Fishes of tribe Nannostomina. Proceedings of the United the Order Plectognathi. xxiii + 388 pages. Doctoral States National Museum, 116(3499): 127-170. dissertation, Stanford University. Wickler. W. 1962b. The Pelvis and Pelvic Fins of Plectognath Fish: A Study in Reduction. Proceedings of the Acade- 1962. Pseudalutarius nasicornis (Monacanthidae, Balisti- my of Natural Sciences of Philadelphia, 114(7) :207- formes). Schwimmenbewegungen. Encyclopaedia 250. Cinematographica (Institut Wissenschaft, Gottin- 1962c. Triodon bursarius, a Plectognath Fish Connecting gen). E61/1962. the Scleroderms and Gymnodonts. Copeia, 4:793- 1963. Triacanthus biaculeatus (Triacanthidae, Balisti- 801. formes). Schwimmen. Encyclopaedia Cinemato- 1963a. A Critique of Y. le Danois' Work on the Classifi- graphica (Institut Wissenschaft, Gottingen). E67/ cation of the Fishes of the Order Plectognathi. 1962:1-5. Copeia, 1:203-206. Wiedersheim, R. E. E. 1963b. The Apparent Reduction in the Number of Pre- 1887. Das Geruchsorgen der Tetrodonten nebst Bemerken caudal Vertebrae in Trunkfishes (Ostraciontidae, uber die Hautmuskulatur derselben. Pages 73-84 Plectognathi). Proceedings of the Academy of Nat- in Festschrift A. von Kolliker zur Feier seines ural Sciences of Philadelphia, 115(7): 153-190. siebenzigstcn Geburtstages gewidmet von seiner 1968. A Monograph on Plectognath Fishes of the Super- Schulern. Leipzig. family Triacanthoidea. Monographs of the Acade- Willem, V. my of Natural Sciences of Philadelphia, 16: 364 1941. Le mode respiratoire de Batistes et d'Ostracion. pages. Bulletin de I'Academie royale de Belgique, serief 1970a. The Progressive Reduction in Number of Elements 5, 27:228-294. Supporting the Caudal Fin of Fishes of the Order 1942. Contribution a l'etude des organes respiratoire chez Plectognathi. Proceedings of the Academy of Nat- les teleosteens Plectognathes, Ire Partie: Les Balis- ural Sciences of Philadelphia, 122(1): 1-85. tides. Bulletin du Musie royal d'Histoire naturelle 1970b. Osteological Aspects of Interrelationships of Sur- de Belgique, 18(35): 1-24. geon Fish Genera (Acanthuridae). Proceedings of 1944. Contribution a l'etude des organes respiratoire chez the Academy of Natural Sciences of Philadelphia, les teleosteens Plectognathes, 2* Partie: Les Chae- NUMBER 155 103

todontiformes. Bulletin du Musee royal d'Histoire Bulletin du Musie royal d'Histoire naturelle de naturelle de Belgigue, 20(6): 1-21. Belgique, 25 (24): 1-4. 1945. Contribution a l'ltude des organes respiratoire Winterbottom, R. chez les t£l£osteens Plectognathes, 3* Partie: Les 1970. Evolution of the Pelvic Fin Ray Musculature of Osttaciomides. 4e Partie: Le role de l'interoper- Some Triacanthoid Fishes (Plectognathi). Copeia, culaire dans la manoeuvre respiratoire chez les 3:453-456. t£leost£ens. Bulletin du Musee royal d'Histoire 1971a. A Reinterpretation of the Mechanism of Extension naturelle de Belgique, 21 (23) :1-16. of the Tetraodontid Pufferfish Postcleithral Appa- 1947a. Contribution a l'etude des organes respiratoire chez ratus, with Notes on the Oblique Eye Muscles of les tele'osteens Plectognathes, 5* Partie: Te'trodon- the Diodontidae (Plectognathi). Notulae Naturae, tes et Diodon. Bulletin du Music royal d'Histoire 440:1-7. naturelle de Belgique, 23 (17) :1—17. 1971b. Movement of the Caudal Spine of Some Surgeon- fishes (Acanthuridae, Perciformes). Copeia, 3:562- 1947b. Le manoeuvres respiratoires chez les poissons 566. telt-ostecns. Bulletin du Musie royal d'Histoire Winther, G. P. naturelle de Belgique, 23 (29) :1-14. 1877. Fiskenes ansigt: en comparativ-anatomisk under- 1949. Contribution a l'ctude des organes respiratoire chez s0gelse. Tredie afsnit. Naturhistorisk Tidsskrift les tclcostccns Plectognathes. 6* Partie: Mola (Copenhagen), 11(3:267-478. List of Abbreviations

In figures where both muscles and bones are illustrated, muscle abbreviations are underscored

A 1. Adductor mandibulae A 1, a, p, and y Di Diodontidae A 2. Adductor mandibulae A 2, a, P. and y D.L. Dorsal lobe (of pelvis) A 3. Adductor mandibulae A 3 D.O. Dilatator operculi A (a. Adductor mandibulae A to D.RETR. Retractor dorsalis A.A.P Adductor arcus palatini D.S. Dorsal spine ABD.P. Abductor profundus EBR. Epibranchial ABD.S. Abductor superficial ECTO. Ectopterygoid ACT. Radial EHY. Posterohyal (=epihyal) AD. Adductores EOC. Exoccipital ADD.P. Adductor profundus EP. Epural ADD.R. Adductor radialis EPAX. Epaxialis ADD.S. Adductor supcrficialis EPIP. Epipleural AD.OP. Adductor operculi E.P.L. Epihyal-pterygoid ligament A.D.P. Arrector dorsalis pelvicus EPOT. Epiotic A.H. Adductor hyomandibulae EREC. Erectores dorsales or anales A.M.L. Articulo-maxilla ligament ETH. Ethmoid ANG. Angular F. Frontal A.P.P. adductor profundus pelvicus F.D. Flexor dorsalis Ar Aracanidae F.D.S. Flexor dorsalis superior ARR.D. Arrector dorsalis FOR.B. Blood vessel foramen ARR.V. Arrector ventralis F.R. Fin ray ART. Articular F.R.E. Fin ray element ASC.P. Ascending process F.V. Flexor ventralis A.S.P. Adductor superficialis pelvicus F.V.E. Flexor ventralis externus A.V. Abdominal vertebra F.V.I. Flexor ventralis inferior A.V.P. Arrector ventralis pelvicus H.AB. Hyohyoidei abductores Ba Balistidae H.AD. Hyohyoidei adductores BBR. Basibi anchial HBR. Hypobranchial BHY. Basihyal H.L. Hypochordal longitudinalis B.L. Baudelot's ligament H.S. Haemal spine BOC. Basioccipital HY.IN. Hyohyoideus inferioris B.PT. Basal pterygiophore HYOM. Hyomandibular B.R. Branchiostegal ray HYP. Hypural B.V. Buccal valve IBR. Infrapharyngobranchial C. Coelom I.E.L. Interopercle epihyal ligament Ca Canthigasteridae IHY. Interhyal CART. Cartilage IMD. Intermandibularis CBR. Ceratobranchial INC. Inclinatores dorsales or anales CHY. Anterohyal (=ceratohyal) INF A. Infracarinalis anterior CL. Clei thrum INF.M. Infracarinalis medius COR. Coracoid INF.P. Infracarinalis posterior COR.R. Coracoradialis INT. Interradialis C.P. Cartilagenous plug I.O.L. Interoperde-opercle ligament C.S. Cut surface IOP. Interopercle C.T. Connective tissue I.S.L. Interopercle-subopercle ligament C.V. Caudal vertebra L.A.P. Levator arcus palatini DEN. Dentary L.CUT. Longitudinal cutaneous DEPR. Depressores dorsales or anales L.EXT. Levatores externi D.HHY. Dorsohyal (=dorsal hypohyal) LIG. Ligament 104 NUMBER 155 105

LINT. Levatores interni R.CAR. Rostral cartilage L.O. Leva tor operculi R.COMM. Rectus coramunis MAX. Maxilla RECT. V. Rectus ventralis Me Monacanthidae R.I. Retractor interoperculi M.E.L. Maxillo-ethmoid ligament R.MD. Ramus mandibularis V MESO. Mesopterygoid R.MX. Ramus maxillaris V META. Metapterygoid S.ANT. Supracarinalis anterior Ml Molidae S.B. Swim bladder M.P.L. Maxillo premaxilla ligament S.BL.a. Swim bladder "a" N.S. Neural spine SCAP. Scapula OBL.D. Obliquus dorsalis SCL. Supraclei thrum OBL.I. Obliquus inferioris S.M. Posterior swim bladder muscle OBL.P. Obliquus posterior S.MED. Supracarinalis medius OBL.S. Obliquus superioris S.N. Supraneural OBL.V. Obliquus ventralis s.o. Sphincter oesophagi O.FOR. Foramen for olfactory nerve soc. Supraotcipital O.N. Olfactory nerve SOP. Suboperde OP. Opercle SPIN. Spinalis OR. Olfactory rosette S.POST. Supracarinalis posterior Os Ostraciidae SPOT. Sphenotic PAL. Palatine STB. Sternobranchialis PAS. Parasphenoid STH. Sternohyoideus PCL. Postcleithra SYM. Symplectic PEL. Pelvis T. Tendon of ... PF. Prefrontal Te Tetraodontidae PHC.E. Pharyngoclavicularis externus Ti Triacanthidae PHC.I. Pharyngoclavicularis internus To Triacanthodidae PM. Premaxilla Tr Triodontidae POP. Preopercle TR. Levator pectoralis P.P. Protractor pectoralis TR.C. Transversus caudalis P.P.L. Palatino-palatine ligament TR.CUT. Transversus cutaneous P.R. Pelvic ray TR.D. Transversus dorsalis PR.HY. Protractor hyoidei TR.V. Transversus ventralis PROT. Prootic U.C. Ural centrum PROX. Proximalis U.HB.L. Urohyal-hypobranchial ligament P.S. Pelvic spine UHY. Urohyal PT. Posttemporal U.HYX. Urohyal-hypohyal ligament PTS. Pterosphenoid VA. Valvulus PTOT. Pterotic V.F.R. Vestigial fin ray Q Quadrate V.HHY. Ventrohyal (=ventral hypohyal) R.A.P. Retractor arcus palatini VOM. Vomer R.B. Ramus buccalis VII FIGURE 1 Body outlines of representative triacanthodids: a, Parahollardia (Hollardiinae); b, Halimochirurgns (Triacan- thodinae). FIGURE 4. Body outlines of representative monacanthids: a, Pervagor; b, Oxymonacanthus; c, Anacanthus.

FIGURE 5.—Body outline of a representative aracanid, Aracanostracion.

FIGURE 2.—Body outline of a representative triacanthid, Trixiphichthys.

FICITKE 6.—Body outlines of representative ostraciids: FIGURE 3.—Body outline of a representative balistid, Batistes. a, Acanthostracion; b, Lactoria. NUMBER 155 107

FIGURE 7.—Body outline of Triodon macropterus.

FIGURE 10.—Body outlines of representative diodontids: a, Diodon; b, Chilomycterus.

a

FIGURE 8.—Body outlines of representative itetraodontids: a, Lagocephalus; b, Arothron.

FIGURE 9.—Body outline of a representative canthigasterid, FIGURE 11.—Body outlines of representative molids: a, Mola; Canthigaster. b, Rumania. 108 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

Tr Te Ca Di Ml To Ti Ba Me Ar 0s Tr Te Ca Di Ml To Ti B)ia WIe ArV Qs L

1 J L FICURE 12. Phylogenetic tree for section A 1 of the adductor FIGURE 13.—Phylogenetic tree for section A 2 of the adductor mandibulae. (See "List of Abbreviations".) mandibulae.

To Ti Ba Me Ar 0s Tr Te Ca Oi ml To Ti Ba Me Ar 0s Tr Te Ca Di Ml

j _J r 1

To Ti Ba Me Ar 0s Tr Te Ca Di Ml To Ti Ba Me Ar 0s Tr Ta Ca Di Ml

FICURE 15.—Phylogenetic trees: a, the levator arcus palatini; To Ti Ba Me Ar Oa Trj T.Tee Ca Dii Ml b, the dilatator operculi. To Ti 8ya Me Ar Os Tr Te Ca Di Ml J L FIGURE 14.—Phylogenetic trees: a, section A 3 of the ad- ductor mandibulae; b, A oo of the adductor mandibulae; r e, reinterpreted for section A to. a To Ti Ba Me Ar 0s Tr Tve Ca Di Ml

FIGURE 16.—Phylogenetic trees: a, the levator operculi; b, the adductor arcus palatini. NUMBER 155 109 To Ti Ba We Ar Qs Tr Te Ca Di Ml To Ti Ba Me Ayr 0s Tr Te Ca Di ml Y x 1 C

To Ti Ba We AVr Qs Tr Te Ca Di Ml To Ti Ba We Ar Qs Tr Te Ca Di ml I . Z—I

-I C

FICURE 17.—Phylogenetic trees: a, the retractor arcus pala- FICURE 18.—Phylogenetic trees: a, the adductor hyomandib- tini; b, the adductor operculi. ulae; b, reinterpreted for the adductor hyomandibulae.

To Ti Ba We Ar Os Tr Te Ca Di ml To Ti Biea MIe ArY Qs Tr TeM Ca lD:i Ml

To Ti Ba We Ar Os Tr Te Ca Di Ml To Ti Ba Me Ar 0s Tr TTe cCai TDi Ml

J L

b FICURE 19. Phylogenetic trees: a, the intennandibularis; b, FICITRE 20.—Phylogenetic trees: a, the hyohyoideus mfenons; the valvulus. b, reinterpreted for the hyohyoideus inferiorit. 110 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY To Ti 8va Me Ar 0s Tr Te Ca Qi Ml To Ti Ba TM,c AVj Q1s Tr T.re C,Ta D.Ti M.1 J L

To Ti Ba Me Ar 0s Tr Te Ca Oi Ml To Ti Ba Me A; Oi Ml

FIGURE 21. Phylogenetic trees: a, the hyohyoidei abductorcs; To Ti Ba We Ar 0s Tr Te Ca Oi Ml b, the hyohyoidei adductores. To Ti Ba Me AVr 0s Tr Te Ca Di Ml

FIGURE 22.—Phylogenetic trees: a, the sternohyoideus; b, the sternobranchialis; c, alternative for the sternobranchialis.

To Ti Ba Me Ar 0s Tr Te Ca Qi Ml Y To Ti Ba Me Ar 0s Tr Te Ca Oi Ml

To Ti Ba Me Ar 0s Tr Te Ca Di Ml V Y To Ti •

FIGURE 23—Phylogenetic trees: a, the pharyngoclavicularis O externus; b, the pharyngoclavicularis interims; c, the obliqui _ ventrales. FIGURE 24.—Phylogenetic trees: m, the transversi ventrales; b, the recti ventrales. To Ti Ba l«lc Ar Os Tr Te Ca Di Ml To Ti Ba Me Ar Os Tr Te Ca Di Ml Y

To Ti Ba Me Ar Os Tr Te Ca Di Ml V ftr Qg Tf V

FIGURE 26.—Phylogenetic trees: a, the obliquu* dorsalis; b, the transversi dorsales. To Ti Ba Me AVr Os Tr Te Ca Oi Ml To Ti Ba Me Ar Os Tr Te Ca Di Ml

FIGURE 25.—Phylogenetic trees: a, the rectus communis; b, the levatores externi. e, the levatores interni.

To Ti Ba Me Ar Os Tr Te Ca Di Ml

To Ti Ba Me Ar Os Tr Te Ca Di Ml

To Ti Ba Me Ar Os Tr Te Ca Di Ml

To Ti Ba Me Ar Os Tr Te Ca Di Ml

To Ti Ba Me Ar Os Tr Te Ca Di Ml FIGURE 28.—Phylogenetic trees: a, the abductor superfidalis; y y b, the abductor profundus; c, the adductor superfidalis.

-FIGURE 27.—Phylogenetic trees: a, the retractor dorsalis; b, the adductores; c, the sphincter oesophagi. 112 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

To Ti 8a We Ar O.s Tr Te Ca Di ml T,° Ti Ba Me Ar Os Tr Te Ca Di mi

To TI 8a Me Ar 0,8 Tr T« Ca 01 To Ti Ba Me Ar Os Tr Te Ca 01 Ml

To T.i Ba Me Ar Os Tr Te Ca Di ml V To 1Ti Ba me Ar Os Tr Ti Ca Di Ml JL

Tp Ti Ba Me A\rp Qs Tr Te Ca Di ml To Ti Ba Me Ar QOas Tr Te Ca Di Ml

FIGURE 29.—Phylogenetic trees: a, the adductor profundus; FIGURE 30.—Phylogenetic trees: a, the adductor radialis; b, the arrector dorsalis; c, the coracoradialis; d, reinterpreted b, the protractor pectoralis; c, the levator pectoralis; d, rein- for the coracoradialis. terpreted for the levator pectoralis. NUMBER 155 113

T T 1 TVo Ti 31a M1e AIr OIs Tr TIe CIa DIi MIl vV° A 8»V "I Ar Os Tr Te Ca Di MJ

3 C

T:o Ti B8a MWie Ar Oa Tr Ta Ca Di Ml To TTii Ba fflpPfc Ar Os Tr Te Ca DOii mMll III i izz:

b

To Ti Ba Me Ar 9s Tr Te Ca Di Ml To Ti Ba We Ar Os Tr Te Ca Di HI

J L

FIGURE 31.—Phylogenetic trees: a, the arrector dorsalis pel- FIGURE 32.—Phylogenetic trees: a, the adductor superfidalis vicus; b, the arrector ventralis pelvicus; c, reinterpreted for pelvicus; b, reinterpreted for the adductor superfidalis pel- the arrector ventralis pelvicus. vicus; c, the adductor profundus pelvicus.

To Ti Ba He _Alr IOs Tr Te Ca Di Ml

To Ti Ba He Ar Os Tr Te Ca Di Hi

t——1 '

FIGURE S3.—Phylogenetic trees: a, the indinatores donates; b, reinterpreted for the indinatores dorsales. To Ti Ba Me Ar Qs Tr Te Ca Di Ml To Ti 8Va Me AVr Qs Tr Te Ca Oi Ml

J L

To Ti Ha flic Ar Os Tr Te Ca Oi ml To Ti Ba We Ar Qs Tr Te Ca Di mil I I I I

To Ti Ba me AT OS Tr Te Ca Di ml To Ti Ba Me Ar Qs Tr Te Ca Di ml V

FICURF. 35.—Phylogcnctic trees: a, the indinatores anales; b, the erectores anales; c, the depressores anales. To Ti Ba me Ar Qs Tr Te Cra TDi To Ti Ba Re Ar Os Tr Te Ca Di Ml i J,

FIGURE 34.—Phylogenetic trees: a, the erectores dorsales; b, reinterpreted for the erectores dorsales; c, the depressores dorsales; d, reinterpreted for the depressores dorsales.

To Ti Ba IDc Ar Os Tr Te Ca Di Ml

To Ti 8a me Ar Os Tr Te Ca Di Ml

FIGURE 36.—Phylogenetic trees: a, the supracarinalis anterior; b, reinterpreted for the supracarinalis anterior; c, the supra- carinalis medius. NUMBER 155 115 TVo TT i 8Va We Ar Os Tr TVe Ca Di Ml To Ti Ba Me Ar Os Tr Te Ca Oi Ml L

To Tl Ba Me Ar 09 Tr Te Ca Oi Ml To Ti Ba Me Ar Os Tr Ca 04 Ml V VV TT I V J I

To Ti Ba me Ar Os Tr Te Ca Di 1*11 To Ti Ba Me Ar Os Tr Tre Ca Oi Ml il

FIGURE 38.—Phylogenetic trees: a, the interradialis; b, the hypochordal longitudinalis; c, the flexor dorsal is. To Ti 3a We Ar Os Tr Te Cn Oi ml V

FIGURE 37.—Phylogenetic trees: a, the infracarinalis anterior; To Ti b, reinterpreted for the infracarinalis anterior; c, the infra- c Ar Qs Tr Ta Ca 0 carinalis medius; d, reinterpreted for the infracarinalis medius.

To Ti B3a M15ec Ar QOs Tr Te Ca Di ml •I 1—L, i

FIGURE 39.—Phylogenetic trees: a, the flexor dorsalis supe- rior; b, reinterpreted for the flexor dorsalis superior. 116 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

To Ti Ba Me AVr Os Tr Te Ca Di Ml To Ti Ba Me Avr Os Tr TVs Ca Di Ml

J L

To Ti Ba Me Ar Os Tr Te Ca Oi Ml To Ti Ba Me Ap Q« Tr Tve Ca Oi Ml I I I

To Ti Ba Me AT OS Tr Te Ca Oi M. To Ti Ba Me Ar Os Tr Te Ca Di Ml

FIGURE 41.—Phylogenctic trees: a, the transversus caudaliv. "YY111 WW

FIGURE 40.—Phylogenetic trees: a, the flexor ventralis; b, the flexor ventralis externus; c, the flexor ventralis inferior; d, reinterpreted for the flexor ventralis inferior. NUMBER 155 117

To Ti B,a Me Ar Oa Tr Te Ca Di Ml To Ti Ba We Ar Os Tr Te Ca 01 Ml

To Ti Ba Me Ar Os Tr Te Ca Di Ml To Ti Ba Me AVr Os Tr Te Ca Di Ml

FIGURE 42.—Phylogenetic trees: a, the obliquus superioris; To Ti Ba Me Ar Os Tr Te Ca Di Ml 6, the obliquus inferioris.

» . I

To Ti Ba Me Ar Os Tr Te Ca Di Ml

FIGURE 43.—Phylogenetic trees: a, the spinalis; b, reinter- preted for the spinalis; c, the transversus cutaneous; d, the longitudinal cutaneous. 118 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

FIGURE 44.—A hypothesized phytogeny of the Tetraodontiformes, based on the analysis of those muscles yielding phylogenetic information. Solid black: The phylogeny included in the shaded area is as indicated in the phylogenetic tree. If only two families are linked, the two families are sister groups. Wavy lines: The contained families share a common ancestor, but further subdivision into sister groups is not possible. Oblique dashes: The common ancestor of the families linked by the wavy lines share* a common ancestor with the families covered by this symbol. Dots: As above, but the next lowest level of application. Arrow: Used to link supposed sister groups where the hypothesized phylogeny is incorrect, if confusion is otherwise likely. A: An autapomorph condition, where this would otherwise affect the interpretation. General: Where symbols extend halfway across the gap separating families and then cease or change, it indicates that the family (ics) thus implicated should be linked with the family (-ies) which is (are) elsewhere represented by the same half-symbol. It only occurs where the hypothesized phylogeny is incorrect. The numerals in the right-hand margin correspond to the text number used for that muscle. Where a muscle has been reinterpreted, "a" and "b" are appended after the number. The numerals within the tree itself will be used to designate branch points in the discussion. The letters "a" and "b" are used at points of branching to clarify the lines in which the evidence applies. 40* 40k

4t 41 • 43 k 44

4tk 49*

60* 80k

TO 120 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

CRYPTOBALISTINAE CDEFGHJJK

TRIACANTHIMAE .TRIACANTHODINAE TO BALISTOIDS*, SPINACANTHINAE PROTACANTHODINAE FIGURE 46.—Cladogram for taxa A-K.

FIGURE 45.—The phylogenetic tree of the basal tetraodonti- forms. (Drawn from Tyler, 1968, fig. 6.)

FIGURE 47.—Phylogenetic tree of living and fossil tetraodontiforms. Daggers indicate fossil groups; question mark emphasizes the tentative placement of the Cryptobalistinae as the sister group of the Triacanthodinae; the digits 1 through 4 correspond to the hierarchical alternatives in the classification on p. 99; outline of fossils from Tyler (ms). NUMBER 155 121

M\X

P\L PPL

MEL. PM.

PPl MAX

ASC.P. DEN. ran

c d

FIGURE 48.—Triacanthodes anomalus, ANSP 117493, 87 mm SL: a, medial view of right half of jaw region; b, dorsal view of jaws (left palatine and maxilla removed); c, anterior view of left maxilla; d, medial view of left maxilla. Triacanthus biaculeatus, ANSP 111536, 111 mm SL.: e-f, left lateral view of snout region, -howing degree of upper jaw protrusion. (See abbreviations list for explanation.) 122 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

FIGURE 49.—Triacanthodes anomalus, ANSP 11749S, 87 mm SL, left lateral view of superficial anterior musculature. NUMBER 155 123

SPOT

UHY.' HY. IN. '.R.COMM. >STH. MNF.A.

FIGURE 50.—Triacanthodes anomalus, ANSP 117493, 87 mm SL, left lateral view of the branchial arch region. (Palatal arch and parts of the hyoid arch removed.) RECTV. TRY. PHC.I.

L1G.

V.HHY.

BHY

UHY.

D.HHY.' UHR

CUR PHCE. 'S.U

CM

FIGURE 51.—Triacanthodes anomalus, ANSP 117493, 87 mm SL. ventral view of branchial arches, urohyal displaced to left. (Left rectus communis and pharyngoclaviculares removed.)

FIGURE 52.—Triacanthodes anomalus, ANSP 117493, 87 mm SL, left lateral view of dorsal branchial arch region. (Levatores externi removed.)

tPlP

N S

D.RETR.

CBR

1CM. NUMBER 155 125

CL. F.R

ACT.

COR.

PEL

C.S.

1 CM

FIGURE 53.—Triacanthodes anotnalus, ANSP 117493, 87 mm SL, medial view of left pectoral and pelvic girdles and the associated musculature.

FIGURE 54.—Triacanthodes anomalus, ANSP 117493, 87 mm SL, left lateral view of the spiny dorsal fin. (Indinatores donates and the epaxialis removed.) SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

IV

IKK

FIGURE 55.—Triacanthodes anomalus, ANSP 117495. 87 mm SL, left lateral view of the sixth to ninth fin rays of: a, soft dorsal fin; b, anal fin. (Body muscles and first two inclinatores removed.)

b nix iU- •SPOSI

INF.P,

F \ 1 EPAX,

2BLJS.-

SPOSI

INFP

INFP IR.t FICURE 56.—Triacanthodes anomalus, ANSP 117493, 87 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, after removal of body musculature; c, deep layer of muscle. NUMBER 155 127

PR. S. ANT

FIGURE 57.—Tydemania navigator is, ANSP 117532, 73 mm SL, left lateral view of superficial anterior musculature.

a

FIGURE 58.—Macrorhamphosodes platycheilus, ANSP 117533, 88 mm SL: a, left lateral view of superficial anterior musculature; b, mirror image of right lateral view of anterior snout region. 128 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

SPOT

L.EXT.

mm.

FICURE 59.—Macrorhamphosodes platycheilus, ANSP 117533, 88 mm SL, left lateral view of branchial arch muscles (superficial).

a

OBL.l.

ABO.S

HY IN' H.AB.' HAD.1 'STH N1M£JL Rj.

FICURE m.—Halimochirurgus centriscoides, ANSP 117494, 114 mm SL: a, left lateral view (slightly dorsal) of superficial anterior musculature; b, left lateral view of anterior snout, with mouth fully open. NUMBER 155 129

FIGURE 61.—Parahollardia lineata, ANSP 101031, 68 mm SL, left lateral view of superficial musculature. 130 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

•\DD.R.

IOR

F.R ll>K K

1CM.

FIGURE 62.—Parahollardia lineata, ANSP 101031, 68 mm SL, medial view of left pectoral girdle, in the region of the radials. (Adductor superficial!* and profundus removed.)

1NF.A

FICURE 63.—Parahollardia lineata, ANSP 101031, 68 mm SL, dorsolateral view of left side of pelvis. (Postdeithrum, obliquus inferioris, and adductor superficial pelvicus section serving the pelvic spine removed.) NUMBER 155 131

FIGURE 64.—Hollardia hollardi, ANSP 117491, 92 mm SL, left lateral view of superficial anterior musculature. 132 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

FIGURE 65.—Hollardia hollardi, ANSP 117491, 92 mm SL, ventral view of superficial anterior musculature.

FIGURE 66.—Trixiphichthys weberi, ANSP 117492, 122 mm SL, left lateral view of superficial anterior musculature.

HYOM. DEPR.

POP

PR.HVf NUMBER 155 133

IT

A. A. P.

IMS

1 CM

FIGURE 67.—Trixiphichthys weberi, ANSP 117492, 122 mm SL, posterolateral view of anterior orbital region. (Adductor mandibulae complex removed.)

FIGURE 68.—Transverse sections progressing posteriorly of the left adductor mandibulae complex, viewed from anterior: a, Hollardia hollardi, ANSP 117491, 53 mm SL; b, Triacanthus biaculea- tus. ANSP 111556. HI mm SL. a \ I. A 1

A2*.

R.MD.

CM. b 134 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

TR.U.y L.EXT., PROTv PTS SPOT iPTOT

PJ.

FIGURE 69^—Trixiphichthys weberi, ANSP 117492, 122 mm SL, left lateral view of branchial arch region. (The palatal and hyoid arches, and the abductor superfidalis removed.) OBL.V. iRECT.V RECT.V. TR.V

UHY

PHC.I.

FIGURE 70.—Trixiphichthys weberi, ANSP 117492, 122 mm SL, ventral view of branchial arches, urohyal displaced to left. (Pharyngoclavicularis externus, right rectus communis, and part of sternobranchialis removed.)

FIGURE 71.—Trixiphichthys weberi, ANSP 117492, 122 mm SL, left lateral view of dorsal branchial arch region. (The levatores externi, the first arch, and ceratobranchials 3 and 4 removed.)

SPOT.

PTOT-

PROT.

LINT.-

PAS.

EBR.

IBR-

CBR.

1CM. I'll'K

Lit.

•I RU

UPT.

SI'IS

FIGURE 72.—Trixiphichthys weberi, ANSP 117492, 122 mm SL, left lateral view of dorsal spine region. (Epaxialis and inclinatores dorsales removed.)

FIGURE 73.—Trixiphichthys weberi, ANSP 117492, 122 mm FICURE 74.—Trixiphichthys weberi, ANSP 117492, 122 mm SL, medial view of left pectoral girdle. (Pelvis removed.) SL, left lateral view of the sixth to eighth dorsal fin rays. (Epaxialis and the sixth and seventh inclinatores dorsales removed.)

INC

LKH NUMBER 155 137 b SPOST

OBL.S. INF.P. FV.E PROX/

HCURE 75.—Trixiphiehthys weberi, ANSP 117492, 122 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, deep musculature.

EREC

L.O HI)

LAP INt

EPAX

OBL.S.

ARR.V. R.MD.

OBL.I.

INF.M.

w. HAD STH. ABD.S/ ABD.P 1CM.

FIGURE 76.—Triacanthw biaculeatus, ANSP 111536, HI mm SL, left lateral view of superficial anterior musculature. 138 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

LIG.» A.S.P> INF.M. INC.

FICURE 77.—Tripodichthys oxycephalus, ANSP 89397, 82 mm SL, left lateral view of superficial anterior musculature. NUMBER 155 159

DEFR.

FIGURE IB^-Balisics vetula, ANSP 117535, 1S7 mm SL, left lateral view of superficial anterior musculature. FIGURE 79—Batistes vetula, ANSP 117535, 137 mm SL, left lateral view of snout region. (Most of the adductor mandibulae complex removed.)

FIGURE 80.—Batistes vetula, ANSP 117535, 137 mm SL, left lateral view of head region. (Body muscles, abductor superficial, lateral portion of the sternohyoideus, and the palatal arch removed.)

S.MEI).

COR.' ARR.V.' 'ABDP '1NF.A. '1NF.M. NUMBER 155 141

REcry TRY.

R.COMM/

OBL.V.1

Fir.iiRF 81.—Balistes vetula, ANSP 117535, 1S7 mm SL, ventral view of branchial arches. (Left rectus communis, pharyngoclaviculares and right stemobranchialis removed^

SPOT

FIGURE 82.—Balistes vetula, ANSP 117535, 137 mm SL, left lateral view of dorsal branchial arch region. (First branchial arch, palatal arch, pectoral girdle, and levatores externi removed.) 142 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

C.S

FIGURE 83.—Batistes vetula, ANSP 117555, 137 mm SL, medial view of left pectoral girdle and associated muscles.

FIGURE 84.—Batistes vetula, ANSP 117535, 137 mm SL. left lateral view of pelvis, cut longitudi- nally slightly to the left of the midline, showing the muscles of the right side. NUMBER 155 143

n. a F.D.S.

S.I'OSl

H.L.

F.R.

INF.P. INT.

1 CM.

SPOST. b B.PT

N.S.

B.PT. INF.P.

FIGURE 85.—Batistes vetula, ANSP 117535, 137 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, deep musculature. 144 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

FIGURE 86^-Balistapus undulatus, ANSP 117598, 83 mm SL, left lateral view of superficial anterior musculature. NUMBER 155 145

A Id..

Kit.

I CM.

Kit-

1CM.

FIGURE 87.—Transverse sections progressing posteriorly of the left adductor mandibulae complex, viewed from anterior: a, Balistapus undulatus, ANSP 117598, 83 mm SL: b, Pervagor melano- cephalus, ANSP 117500, 85 mm SL. 146 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

FIGURE 88.—Melichthys niger, ANSP 117515, 153 mm SL, left lateral view of superficial anterior musculature. NUMBER 155 147

FIGURE 89.—Rhinecanthus aculeatus, ANSP 117546, 159 mm SL, left lateral view of superficial anterior musculature. 148 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

FIGURE 90—Odonus niger, ANSP 102779, 257 mm SL, left lateral view o( superficial anterior musculature. NUMBER 155 149

SfOT \ D S

TIL' rtL 'INFM

FIGURE 91.—Pervagor melanocephalus, ANSP 117512, 87 mm SL, left lateral view of superficial anterior musculature. 150 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

BPT.

EREC.

ETH

Q. IOP.' PRHYf FIGURE 92.—Pervagor mehmocephalus, ANSP 117512, 87 mm SL, left lateral view of the deep muscles of the cheek. (Most of the adductor mandibulae complex removed.)

FIGURE 93.—Pervagor melanocephalus, ANSP 117512, 87 mm SL, medial view of right palatal arch (posteroventral region). (Hyoid arch, except interhyal, removed.)

HYOM.

R.I.

META

POP

l.E.L NUMBER 155 151

•ML

TR.V.

R.COMM/ STB.' BBR

FIGURE 94.—Pervagor melanocephalus, ANSP 117512, 87 mm SL, ventral view of branchial arches. (Left rectus communis, sternobranchialis, and pharyngoclaviculares removed.)

- EPAX.

INF A.

FIGURE 95.—Pervagor melanocephalus, ANSP 117512. 87 mm SL, left lateral view of superficial branchial arch musculature. (The palatal and hyoid arches removed.) 152 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

PRO I

P\S

SCO I

LINT DRtTR

OBL.I). IK I)

1 UK

oiu r IliK

_Ui

FIGURE %.—Pervagor melancephalus, ANSP 117512, 87 mm SL, dorsolateral view of the dorsal branchial arch region. (The levatores exteri, first branchial arch, and ceratobranchial 2 removed.) NUMBER 155 153

ADDS.

ADD.P

FICURE 97.—Pervagor melanoccphalus, ANSP 117512, 87 mm SL, medial view of left pectoral girdle. (Postdeithrum, removed.)

C.S.

A.V.P.' A.S.P. • FOR.B.'

1 CM.

FIGURE 98.—Pervagor melanocephalus, ANSP 117512, 87 mm SL, left lateral view of pelvis, cut longitudinally slightly to the left of the midline. 154 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

FD.S.

H.L

OBL.S.

OBL.I

1NF.P.-

1 CM

F.DS.

FD.

C.V

II.S

B.PT

IM.IV

FICURE 99.—Pervagor melanocephalus, ANSP 117512, 87 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, deep musculature. TRCUT

FIGURE 100—Stephanolepis auratus, ANSP 117526, 64 mm SL, left lateral view of superficial anterior musculature.

FIGURE 101.—Stephanolepis auratus, ANSP 117526, 64 mm SL, left lateral view of deep cheek musculature. (A la. A la', and A 2 removed.)

H M). 156 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

EPAX

A !«t?

FIGURE 102.—Paramonacanthus barnardi, ANSP 117506, 78 mm SL, left lateral view of super- ficial anterior musculature. NUMBER 155 157

S.MED.

EREC.

STH INF.A TR.CUT

1CM

FICURE 103—Chaetoderma spinosissimus, ANSP 117514, 42 mm SL, left lateral view of super- ficial anterior musculature. 158 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

EREC.

FIGURE 104.—Paraluteres prionurus, ANSP 106644, 57 mm SL, left lateral view of superficial anterior musculature. NUMBER 155 159

DEPR.

I ()

R.MD

INF.M.

INF.A.

FIGURE 105.—Cantherhines pardalis, ANSP 117530, 126 mm SL, left lateral view of superficial anterior musculature. FIGURE 106.—Oxymonacanthus lonprostris, ANSP 117496, 72 mm SL, left lateral view of superficial anterior musculature.

Sill.1 < INF.A. INF.M.' FICURE 107.—Pseudalutarius nasicornis, ANSP 100830, 45 mm SL, left lateral view of superficial anterior musculature. EREC

A I* EPAX

R.MI)

FIGURE 108.—Aluterus monoceros, ANSP 117554, 345 mm SL, left lateral view of superficial anterior musculature.

wmac.

RJMX

RJHD

STH

INF.A

,1CM.

FIGURE 109.—Aluterus monoceros, ANSP 117554, S45 mm SL, left lateral view of the deeper muscles in the head region. 162 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

RAP,

FIGURE 110.—Aluterus monoceros, ANSP 117554, 345 mm SL, left lateral view of the cheek. (The adductor mandibulae complex removed.)

HBR. R.COMM.v STBj OBL.V. RECT.V. VCBR.

OBL.V.

RECT.V.

FICURE 111—Aluterus monoceros, ANSP 117554, 545 mm SL, ventral view of branchial arch musculature. (Most of the muscles of the left side and the right rectus communis removed.) NUMBER 155 163

D.RETR.

L.

TR.D.

1CM.

FIGURE 112.—Aluterus monoceros, ANSP 117554, 345 mm SL, left lateral view of dorsal branchial arch musculature. (Palatal arch, pectoral girdle, and first infrapharyngobranchial removed.)

PCL

F.R

ACT.

FIGURE 113.—Aluterus monoceros, ANSP 117554, 345 mm SL, medial view of the left pectoral girdle and associated muscles. 164 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

SMBD. ,ETAX.

FIGURE 114.—Anacanthus barbatus, ANSP 117516, 166 mm SL, left lateral view of superficial anterior musculature.

PROT F .PTOT. UNI S.ANT. PTS / i y CART. B.PT.

OBL.1

(SOP. 'OP. ' ARR.V, x ABD.P.' ABD.S.^ ADD.P.^OR.^ PCL. FIGURE 115.—Kentrocapros aculeatus, ANSP 117531, 91 mm SL, left lateral view of superficial anterior musculature. NUMBER 155 165

A 2f.

A2J.

RJ.

FIGURE 116.—Transverse sections progressing posteriorly of the left adductor mandibulae com- plex, viewed from anterior: a, Kentrocapros aculeatus, ANSP 117531, 91 mm SL; b, Ostracion tuberculatus, ANSP 117542, 102 mm SL. FIGURE 117. Kentrocapros aculratus, ANSP 117531, 91 mm SL, dorsolateral view of the deep muscles of the cheek. FIGURE 118.—Kentrocapros aculeatus, ANSP 117531, 91 mm SL, ventral view of superficial anterior musculature.

H.AD STB ' CL.' ABD.S NUMBER 155 167

FICURE 119.—Kentrocapros aculeatus, ANSP 117531, 91 mm SL, left lateral view of superficial branchial arch musculature. 168 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

CM.

RECT.V.

PHC£.

UHY

U.HY.L.

SSL

R.COMM. 9J.

STB.' OBLV.

1CM.

FIGURE 120.—Kentrocapros acuteatus, ANSP 117531, 91 mm SL, ventral view of branchial arch musculature.

PAS PROT. < AD.OP. yA.H.

FIGURE 121.—Kentrocapros aculeaius, ANSP 117531, 91 mm SL, left lateral view of dorsal branchial arch musculature. (The levatores externi, interni, and the first arch removed.) NUMBER 155 169

SJOST

FICIRE 122.—Kentrocapros aculeatus, ANSP 117531, 91 mm SL, left lateral view of dorsal and anal fin musculature.

ADDS.

•ARR.D.

COftJt.

COR. CL.

R/.

FICURE 123.—Kentrocapros aculeatus, ANSP 117531, 91 mm SL, medial view of left pectoral girdle and associated musculature, from somewhat posterodorsally. 170 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

UtX

IBX

F.VJ.

FIGURE 124—Kentrocapros aculeatus, ANSP 117531, 91 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, deep musculature.

OBLJ.v S. ANT.t EPAX\ INC

UINT.

INT

OBLS.

PR.HY.

I)EPR

X10. FIGURE 125.—Capropygia unistriata, ANSP 109574, 67 mm SL, left lateral view of superficial musculature. StLL

FU.

FICUKE 126.—Ostracion tuberculatus, ANSP 117542, 102 mm SL, left lateral view of superficial anterior musculature.

FIGURE 127.—Ostracion tuberculatus, ANSP 117542, 102 mm SL, left lateral view (slightly dorsally) of the deep check musculature.

ETH.

VOM.

ART.

ANG/ IOP POP

I CM. 172 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

IHY.

FIGURE 128.—Ostracion tuberculatus, ANSP 117542, 102 mm SL, dorsolatcral view of snout region after removal of the adductor mandibulae complex.

C.S.

PROT

FIGURE 129.—Ostracion tuberculatus, ANSP 117542, 102 mm SL, left lateral view of superficial branchial arch musculature. NUMBER 155 173

CBR.

UHY.

U.HHY.

U.HY.L

K.tOMM.

HBR OBL.V.

1CM.

FICURE ISO—Ostracion tuberculatus, ANSP 117542. 102 mm SL, ventral view of the branchial arch musculature. (Left rectus communis removed.)

PAS

LINT.

OBLP CBR.

1CM.

FICURE 131.—Octradon tuberculatus, ANSP 117542, 102 mm SL, dorsolateral view of dorsal branchial arch musculature. (Levatores externi removed.) 174 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

KL

FR

ARRD.

CL

COR.

FIGURE 132.—Ostracion tuberculatus, ANSP 117542, 102 mm SL, posteromedial view of left pectoral girdle and associated musculature.

FR

S.POST.

c.v.

INFP. B.PT.

SJjJL*.

ERECT DEPR/ >F.R. FICURE 133.—Ostracion tuberculatus, ANSP 117542, 10Z mm SL, lett lateral view oi aeep Doay musculature. NUMBER 15S 175 b

r.yj. r.vx

FICURE 134.—Ostracion tuberculatus, ANSP 117542, 102 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, deep musculature.

FICURE 135.—Lac tor ia cornuta, ANSP 117539, 176 mm SL, left lateral view of superficial anterior musculature.

L.A.T. EPAX.

DO.

L.O

A2«C

A2P. ARR.V.

ABD.S.

ABU.P. A3.

OBL.I.

R.MD

PR.HY. H.AD/ 'HY.IN. H.AB. STH. FICURE 156.—Tetrosomus concatenate, ANSP 117555. 196 mm SL, left EPAX. lateral view of superficial anterior musculature. LA*

DO

LO

\i

R.MD

AX.

A3

FIGURE 1S7<—Aeanthostra- cion quadricornis, ANSP 98848, 126 mm SL, left lateral view of superficial OBL.1. anterior musculature.

H.AB. NUMBER 155 177

FIGURE 138.—Triodon macropterus, ANSP 103902, 270 mm SL, left lateral view of superficial anterior musculature. 178 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

R.MD.

A 1.

A3.

A3.

I CM. b

113639, 124 mm SL. NUMBER 155 179

FICURE 140—Triodon macropterus, ANSP 10S902, 270 mm SL, left lateral view of the deep muscles of the cheek.

MR. HBR

C.S.

UHY.

U.HY.L.

RCOMV RECTV

FIGURE 141.—Triodon macropterus, ANSP 103902, 270 mm SL, ventral view of branchial arch musculature. (Left rectus communis and pharyngoclaviculares removed.) 180 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

PROT

LEXT

LJNT.

(UK

I CM.

FIGURE 142.—Triodon macropterus, ANSP 103902, 270 mm SL, left lateral view of dorsal branchial arch region. (Most of the levatores extern! removed.)

B.L. ADD.S. ARR.P.

CL.

PCL.

COR. FIGURE 143—Triodon macropterus, ANSP 103902, 270 mm SL, medial view of left pectoral girdle and associated musculature. JNC

FICURE 144.—Triodon macropterus, ANSP 103902, 270 mm SL, left lateral view of dorsal and anal fin musculature.

FIGURE 145.—Triodon macropterus, ANSP 103902, 270 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, aha removal of the aponeurosis; c, deep musculature. a

F.VJ./ 'r.v. i. ANT.

DEN

mru

FIGURE 146.—Lagocephalus lunaris, ANSP 113639, 124 mm SL, left lateral view of superficial anterior musculature.

FIGURE 147.—Lagocephalus lunaris, ANSP 113639, 124 mm SL, left lateral view of cheek musculature after the removal of A 1.

-PTOT.

ETH

I CM NUMBER 155 183

PAL

SYM.

1CM.

FICURE 148.—Lagocephalus lunaris, ANSP 113639, 124 mm SL, left lateral view of the deep muscles of the cheek.

I CM.

POP.

FIGURE 149.—Lagocephalus lunaris, ANSP 113639, 124 mm SL, medial view of articulation region of left lower jaw. 184 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

SfOT.-

PROT

RECT.V.

c.s.

CHY

H.Afc JUL I CM

FIGURE 150.—Lagocephalus lunaris, ANSP 113639, 124 mm SL, left lateral view of superficial branchial and hyoid arch musculature.

FIGURE 151.—Lagocephalus lunaris, ANSP 113639, 124 mm SL, ventral view of branchial arch musculature.

MLV. PHC.E.

LIG

mm PHC.I.

TRY.

R.COMM.

CBR. PROT DRLTR.

TR.D

II1K

FIGURE 152.—Lagocephalus lunaris, ANSP 113639, 124 mm SL, left lateral view of dorsal branchial arch musculature.

FIGURE 153 Lagocephalus lunaris, ANSP 113639, 124 mm SL, medial view of left pectoral girdle and associated musculature.

ADOS. CL

COR.R. 186 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

1 EREC> Dl I B.PT INC.

1NF.M- INC.

1CM.

FICURE 154.—Lagocephaltis lunaris, ANSP 113639, 124 mm SL, left lateral view of dorsal and anal fin musculature.

a b

S.POST.v N.S.'

FICURE 155—Lagocephalus lunaris, ANSP 113639, 124 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, deep musculature. NUMBER 155 187

D.O.V L.O.V C.S PP. TR.

K\p. EPAX.

A l«t OBL.l

A2W.

INF.M. PR.HY

H.AB. FICURE 156.—Amblyrhynchotes honckenii, ANSP 117547, 106 mm SL, left lateral view of superficial anterior musculature.

L-O.v C

A2i.'

PR.HY. VAB. 'HAD. STH. ^INF.A. NL.CUT .P. WBD.S. XARR.V.

FICURE 157. Colomesus psittacus, ANSP 98887, 117 mm SL, left lateral view of superficial anterior musculature. 188 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

EREC.v DEPR.

LO

LA*.

EPAX.

A Id.

FR.HY H.AB.

I CM.

FKUU 158.—XtnopUnu tutritus, ANSP 117495. 102 mm SL, left lateral view of superficial anterior musculature.

LO DO

A If •AX.

AU.

ABOT. ABD.S. XAR».V.

1= FIGURE 159^-ilrortron irnmacufato, ANSP 117528, 68 mm SL, left lateral view of superficial musculature. ("?INCH implies that the identification of these fibers as part of the incUnator (analis) system is tentative.) NUMBER 155 189

C.S» P.P.

\:»

MM. I' Vlilis N \KK\

FIGURE 160.—Tetraodon schoutedeni, ANSP 117501, 62 mm SL, left lateral view of superficial anterior musculature. ("?INC", as for Figure 159.)

0-

r PR.HY H.AD. 'STH. 'ARR.V. ^ ABDA V OBL.I. VlNFM. FICUIUE 161.—Ctmthigaster margaritatus, ANSP 117517, 52 mm SL, left lateral view of superficial anterior musculature. 190 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

A.A.P.. PF ETH

FKun 182.—Cmthiputer mmrgmritatm, ANSP 117517, 52 mm SL, left lateral view of deep muscles of the cheek.

FKUKB lte—C*nthig*sUr mmrgaritatus, ANSP 117517, 52 mm SL, left lateral view of superficial branchial arch musculature.

SPOT.'

PTS

PROT

c.s. HBK

R.COM.M. STH. I CM. NUMBER 155 191

DRETR

I CM.

FICURE 164.—Canthigaster margaritatus, ANSP 117517. 52 mm FICUKE 165.—Canthigaster nutrgaritatus, ANSP 117517, 52 mm SL, ventral view of branchial arch musculature. SL, left lateral view of dorsal branchial arch region. (The first infrapharyngobranchial, epibranchials 1 and 2, ceratobran- chials 1-5, and the levatores extern! removed.)

BPT S.ANT., EREC, i DEPR

soc

INC INF.M.

Fir.iRF 166.—Canthigaster margaritatus, ANSP 117517, 52 mm SL, left lateral view of dorsal and anal fin musculature. 192 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY b

N.S/

EPAX.

F.R. C.V

OBI..S.

.INT. H.S

OBL.I. 1NF.P. 'F.V.I.

I CM.

FIGURE 167.—Canthigaster margaritatus, ANSP 117517, 52 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, deep musculature.

FIGURE 168.—Canthigaster valentini, ANSP 117534, 45 mm SL, left lateral view of superficial anterior musculature.

s. ANT.

EREC.

A IP.

OBL.S.

INC

STH. IABD.P. ^ABD.S. XARR.V. ^ OBL.I. MNF.M.

1 CM. NUMBER 155 19S

PTOT. v OP SCL. ,CL. ARR.V. PCL. yETAX.

INC

•JNC

PR.HY/ H.AD.' B.R.' HAR' 'ABD.S. FICURE 169—Diodon hystrix, NMC 68-997, 77 mm SL, left lateral view of superficial anterior musculature.

HYOM. PTOT. LAP. SPOT.

PROT

DEN.

FIGURE 170.—Diodon hystrix, NMC 68-997, 77 mm SL, left lateral view of the deep muscles of the cheek. 194 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

CL.

PHC.I. V.HHY.

STH.

I CM.

FICURE 171.—Diodon hystrix, NMC 6S-997, 77 mm SL, left lateral view of superficial branchial arch musculature.

CBR.

HBR

BBR

RECT.V c.s.

1 CM.

FICURE 172.—Diodon hystrix, NMC 6&-997, 77 mm SL, ventral view of branchial arch musculature. NUMBER 155 195

soc. EOC N.S. ,A.V.

EffOT.

tBR'

FIGURE 173.—Diodon hystrix, NMC 68-997, 77 mm SL, left lateral view of dorsal branchial arch musculature. (Epibranchials 1 and 2, ceratobranchials 1-3, and the levatores externi removed.)

PCL.

F.R.

ACT.

FIGURE 174.—Diodon hystrix, NMC 68-997, 77 mm SL, medial view of left pectoral girdle and associated musculature. 196 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

a

EEA&

F.VJ.

FIGURE 175.—Diodon hystrix, NMC 68-997, 77 mm SL, left lateral view of caudal peduncle region: a, superficial musculature; b, deep musculature.

P.P., TR

A \JL

H.AD/ H.AB/ 'ABDS N ABD P **OBLI. INF.M

FIGURE 176.—Chilomyctertu orbicularis, ANSP 117510, 80 mm SL, left lateral view of superficial anterior musculature. A2«C

A 1*.

HAD \ABDP \\BDS

FIGURE 177.—Chilomycterus schoepfi, ANSP 117518, 138 mm SL, left lateral view of superficial anterior musculature.

FIGURE 178.—Mola mola, BM(NH) 1938.623.38, 287 mm SL, left lateral view of superficial anterior musculature.

EPOT C.S. (EREC.+DEPR.)

H.AR> HAD. SIHf PHC.E. \=i PAL PAS

R.MD

IOP

FICURE 179 Mola mola, BM (NH) 19S8.6.2S.S8, 287 mm SL, left lateral view of the deep muscles of the cheek. FICURE 180 Mola mola, BM (NH) 1938.62S.S8, 287 mm SL, left lateral view of superficial branchial arch musculature.

SPOT

V.HHY

CL.

SDL NUMBER 155 199

CART

HBK CBR.

IJISK

cs S.0,

TR.V.

D.HHY.

PMCJ. PHC.E. OBL.V. RECT.V.

R.COMM.

FIGURE 181.—Mola mola, BM(NH) 19S8.623.S8, 287 mm SL, ventral view of branchial arch musculature.

PAS

IBR.

EBR.'

CBR

FIGURE 182 Mola mola, BM (NH) 1938.62S.38, 287 mm SL, left lateral view of dorsal branchial arch musculature. (Epibranchials 1 and 2, ceratobranchials 1-3, and the levatores externi and interni removed.) ADD.P.

CL.

COR.

FR

AC I

FIGURE 183—Afofo mola, BM (NH) 1938.6.23.38, 287 mm SL, medial view of left pectoral girdle and associated musculature.

BPT

EPAX. EREC. + DEPJL

CART.

FRF.r +DEPR NUMBER 155 201

EREC+DEPR

EREC+DEPR.

A2«L

nunr

OBL.I.' AMDS. ^AMKF. »AWLV.

FIGURE 185.—Ranzania laevis, ANSP 106723, 93 mm SL, left lateral view of superficial anterior musculature.

FICURE 184-—Mola mola, BM(NH) 1938.6.23.38, 287 mm SL, left lateral view of dorsal, anal, and caudal fin musculature.

•ft U.S. GOVERNMENT PRINTING OFFICEl 1974 1974 516^223/23

Publication in Smithsonian Contributions to Zoology

Manuscripts for serial publications are accepted by the Smithsonian Institution Press, sub- ject to substantive review, only through departments of the various Smithsonian museums. Non- Smithsonian authors should address inquiries to the appropriate department. If submission is invited, the following format requirements of the Press will govern the preparation of copy. Copy must be typewritten, double-spaced, on one side of standard white bond paper, with IY2" top and left margins, submitted in ribbon copy with a carbon or duplicate, and accompa- nied by the original artwork. Duplicate copies of all material, including illustrations, should be retained by the author. There may be several paragraphs to a page, but each page should begin with a new paragraph. Number consecutively all pages, including title page, abstract, text, litera- ture cited, legends, and tables. The minimum length is 30 pages, including typescript and illus- trations. The title should be complete and clear for easy indexing by abstracting services. Taxonomic titles will carry a final line indicating the higher categories to which the taxon is referable: "(Hymenoptera: Sphecidae)." Include an abstract as an introductory part of the text. Identify the author on the first page of text with an unnumbered footnote that includes his professional mailing address. A table of contents is optional. An index, if required, may be supplied by the author when he returns page proof. Two headings are used: (1) text heads (boldface in print) for major sections and chap- ters and (2) paragraph sideheads (caps and small caps in print) for subdivisions. Further headings may be worked out with the editor. In taxonomic keys, number only the first item of each couplet; if there is only one couplet, omit the number. For easy reference, number also the taxa and their corresponding headings throughout the text; do not incorporate page references in the key. In synonymy, use the short form (taxon, author, date:page) with a full reference at the end of the paper under "Literature Cited." Begin each taxon at the left margin with subse- quent lines indented about three spaces. Within an entry, use a period-dash (.—) to separate each reference. Enclose with square brackets any annotation in, or at the end of, the entry. For references within the text, use the author-date system: "(Jones, 1910)" and "Jones (1910)." If the reference is expanded, abbreviate the data: "Jones (1910:122, pi. 20: fig. 1)." Simple tabulations in the text (e.g., columns of data) may carry headings or not, but they should not contain rules. Formal tables must be submitted as pages separate from the text, and each table, no matter how large, should be pasted up as a single sheet of copy. Use the metric system instead of, or in addition to, the English system. Illustrations (line drawings, maps, photographs, shaded drawings) can be intermixed throughout the printed text. They will be termed Figures and should be numbered consecu- tively; however, if a group of figures is treated as a single figure, the components should be indicated by lowercase italic letters on the illustration, in the legend, and in text references: "Figure 9b." If illustrations (usually tone photographs) are printed separately from the text as full pages on a different stock of paper, they will be termed Plates, and individual components should be lettered (Plate 9b) but may be numbered (Plate 9: figure 2). Never combine the numbering system of text illustrations with that of plate illustrations. Submit all legends on pages separate from the text and not attached to the artwork. An instruction booklet for the preparation of illustrations is available from the Press on request. In the bibliography (usually called "Literature Cited"), spell out book, journal, and article titles, using initial caps with all words except minor terms such as "and, of, the." For capitalization of titles in foreign languages, follow the national practice of each language. Underscore (for italics) book and journal titles. Use the rolon-parcnthcses system for volume, number, and page citations: "10(2) :5-9." Spell out such words as ''figures," "plates," "pages." For free copies of his own paper, a Smithsonian author should indicate his requirements on "Form 36" (submitted to the Press with the manuscript). A non-Smithsonian author will receive 50 free copies; order forms for quantities above this amount with instructions for pay- ment will be supplied when page proof is forwarded. K 1

•

"iHr ^ ^MBI