HARVARD UNIVERSITY Library of the Museum of Comparative Zoology UNIVERSITY OF KANSAS miscellaneous MUSEUM OF NATURAL HISTORY publication No. 64 IVIUS. COMP. ZOOU. UIBRARY DEC 1 b ',9 I HARVARO The Phylogeny and Biogeography of Fossil and Recent Gars (Actinopterygii: Liepisosteidae) By E. O. Wiley UNIVERSITY OF KANSAS LAWRENCE 1976 November 12, 1976 UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly published in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publications, Museum of Natural History, began with nimiber 1 in 1946. Longer research papers are pub- lished in that series. Monographs of the Museiun of Natural History were initiated in 1970. All manuscripts are subjected to critical review by intra- and extramural specialists; final acceptance is at the discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain the Occa- sional Papers and Miscellaneous Publications by addressing the Exchange Librarian, University of Kansas Library, Lawrence, Kansas 66045. Individuals may purchase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, University of Kansas, Law- rence, Kansas 66045. The University of Kansas Museum of Natural History Miscellaneous Publication No. 64 November 12, 1976 The Phylogeny and Biogeography o£ Fossil and Recent Gars (Actinopterygii: Lepisosteidae) By E. O. Wiley A dissertation submitted in partial ftdfilhnent of the re- quirements for the degree of Doctor of Philosophy, The City University of Neiv York, 1976. The University of Kansas Law^rence 1976 University of Kansas Publications, Museum of Natural History Editor: Richard F. Johnston Miscellaneous Pubhcation No. 64 pp. 1-111; 72 figures; 4 tables Published November 12, 1976 Museum of NAxtmAL History The University of Kansas Lawrence, Kansas 66045 U.S.A. Printed by University of Kansas Printing Service Lawrence, Kansas CONTENTS ABSTRACT 1 INTRODUCTION 1 MATERIALS AND METHODS 5 Systematic Methodology 7 Biogeograpliic Method 13 RELATIONSHIPS OF GARS TO OTHER ACTINOPTERYGIAN FISHES 13 The Skull 14 The Hyoid and Visceral Arches 25 Postcranial Skeleton 34 Summary Hypotheses of Actinopterygian Relationships 37 Summary 37 SYSTEMATIC ACCOUNTS 39 Division Ginglymodi 39 Family Lepisosteidae 4 1 Genus Lepisosteus 41 Genus Atractosteus 61 PHYLOGENETIC RELATIONSHIPS AMONG GARS 83 Monophyly of the Genera 84 Relationships Among Lepisosteus Gars 85 Relationships Among Atractosteus Gars 87 A CLASSIFICATION OF GARS 92 GAR BIOGEOGRAPHY 93 ACKNOWLEDGEMENTS 97 SUMMARY 97 LITERATURE CITED 99 APPENDIX A— MATERIAL EXAMINED 110 L Abstract The relationships, taxonomy, and biogeog- raphy of gars are the focus of this study. The phylogenetic method of Hennig (1966) is used to analyze current hypotheses concerning the relationships among gars and of gars to other actinopten,'gian groups. Hennig's method is discussed and se\eral points taken up in detail. Croizat's ( 1958 ) method of biogeographic anal- ysis is used to describe the major features of gar biogeography. Gars comprise a monophyletic group, the sister-group of the Halecostomi (Amiidae plus Teleostei). These three ta.xa comprise another monophyletic group, the Xeopter>'gii, the sister- group of the Chondrostei. These conclusions corroborate certain previous hypotheses and re- fute others. Sixteen species of gars are recognized. They are split equally among two genera, Lepisosteus and Atractostetis. The genus Lepi- sosteus includes a newly described fossil species from the Cretaceous of Montana and the fol- lowing seven species in approximate phyloge- netic order: L. cuneatus (Eocene, North Amer- ica); L. platostomus (Recent, North America); L. indicus (Cretaceous, India); L. osseus (Re- cent, North America); L. fimbriatus (Eocene and Oligocene, Europe); L. oculatus (Recent, North America); L. platyrhincus (Recent, North America). The interrelationships of these spe- cies are discussed. The genus Atractosteus includes, in approx- imate phylogenetic order: A. strausi (Eocene, Europe); A. tropicus (Recent, Middle Amer- ica); A. simplex (Eocene, North .America); A. africanus (Cretaceous, Africa); A. occidentalis (Cretaceous, North .\merica); A. atrox (Eocene, North America); A. spatula (Recent, North and Middle America); and A. tristoechus (Recent, Cuba and the Isle of Pines). Track analysis of the biogeographic distri- butions of both genera indicate that both may have had a Pangean distribution and the min- imum age for both genera is hypothesized to be 180 million years before present. Various tracks within each genus are discussed and relative levels of v-icariance are hypothesized. Introductiox The lepisosteids, or gars, are carniv- orous fishes of sluggish habits and are now restricted to the ^^^estem Hemi- sphere from Costa Rica to southern Can- ada. Fossil gars are known from North America (Cretaceous to Recent), Europe (Cretaceous to Oligocene), .Africa (Cre- taceous), and India (Cretaceous). There are seven currently recognized Recent species (Suttkus, 1963) and nine diag- nosable fossil species. The living gars inhabit the larger ri\'ers, streams, and lakes of their range. Some species also frequent brackish and marine coastal waters (see Suttkus, 1963, for a summar\' of occurrence in these habitats). Anatomically, they combine various primitive ( plesiomorphous ) and derived (apomoqihous) characters. No- tably primitive characters include inter- locking ganoid scales, skull roofing bones with enameloid tubercles, a semi-hetero- cercal tail, and fulcral scales on the medial fins (Suttkus, 1963; Patterson, 1973). Derived characters not found in any other group of actinopter\'gians in- clude an attenuated snout produced by ethmoid elongation, opisthocoelous ver- tebrae, and plicidentine teeth. The structure and development of gars have been extensively studied since the first works of Louis Agassiz (1834, 1843; anatomy) and Alexander Agassiz (1878; general aspects of early develop- ment). These studies include: Earl>' development: Wright, 1879; Balfour and Parker, 1882 (many aspects of both development and structure, com- parisons \\Tth other fishes); Beard, 1889; Mark, 1890; Dean, 1895a, 189.5b, 1896a, 1896b (comparisons \^-ith Amia); Ziegler, 1900; Reighard and Phelps, 1908 (ad- hesive organ); Lindahl, 1944 (adhesive organ and hvpophvsis); and Virchow, 1894. Skull development: Parker, 1882; \'eit, 1907, 1911. 1924 (chondrocranium); Hammarberg, 1937 (chondrocranium and dermal bones); Aumonier, 1941 (dermal bones). MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Vertebrae and ribs: Gegenbaur, 1867; Baur, 1887 (ribs); Schaeffer, 1967a (vertebrae ) . Miscellaneous developmental studies: Wilder, 1876, 1877 (fins and brain); Nickerson, 1893 (scales); Beard, 1895 (pronephrons ) , 1896 (yolk sac and mer- ocytes); Allen, 1911 (origin of sex cells); Landarce and Conger, 1913 (lateral line primordia); Brookover, 1914 (olfactory nerve); Hammett and Hammett, 1939 (proportional snout length); Garrett, 1942 (corpuscles of Stannius); Kullin, 1950 (forebrain); Bodemer, 1957 (ex- trinsic ocular muscles); Kerr, 1967 (teeth); Jessen, 1972 (pectoral girdle). Skull stmcture: Veit, loc. cit; Baur, 1889a (comparison of occipital region with Amia); Allis, 1919 (comparison of otic region with other fishes); De Beer, 1926 (comparison of orbito-temporal re- gion with other fishes); Rayner, 1948 (neurocranial ossifications compared to other fishes); Patterson, 1973 (compari- sons with other neopteiygians ) , 1975 (comparison with other actinoptery- gians); Reagan, 1923 (skeleton, with comparisons); Mayhew, 1924 (skull ossi- fications); Gregory, 1933 (comparisons with other fishes); Holmgreen and Sten- sio, 1936 (skull and visceral arches); Westoll, 1937 (cheek bones); Stensio, 1947 (relationship of lateral line system to skull bones); Parrington, 1956 (pat- terns of dennal bone ossification); Gos- line, 1965 ( circumorbital bones); Gardi- ner 1963 (snout), 1967 (preopercular). Sensory canals: Collinge, 1892, 1895; Allis, 1905, 1934; Stensio, loc. cit. Hyoid arch: Tatarke, 1939; Bertmar, 1967; McAllister, 1968; Nelson, 1969a. Visceral skeleton: Wijhe, 1880, 1882; Allis, 1911; Edgeworth, 1911, 1935 (mus- cles); Hohngren and Stensio, loc. cit.; Nelson, 1969a. Pectoral girdle: Sewertzoff, 1934; Quertermus, 1967 (cleithral shape); Jes- sen, 1972, 1973 (with comments on course of spinal nerves in vertebral re- gion ) . Vertebrae and ribs: Baur, 1887, 1889b; Haines, 1942; SchaeflFer, 1967a. Scale morphology: Agassiz, 1843; Williamson, 1849, 1851; Jackson, 1856, Reissner, 1859; Nickerson, 1893; Scupin, 1896; Goodrich, 1909; Kurr, 1952; Sutt- kus, 1963. Swimbladder: Valentine, 1840; Hoe- ven, 1841; Hyrtl, 1852a; Parkard, 1859. Respiratory function of the swim- bladder: Weidersheim, 1904; Potter, 1927; Suttkus, 1963; McGonnack, 1967; Renfo and Hill, 1970; Rahn, Rahn, How- ell, Gans, and Tenney, 1971; Hill, Schnell, and Echelle, 1973. Other anatomical studies include: Wyman, 1844 (tooth stmcture); Miiller, 1844 (gut); Wilder, 1877 (brain), 1878, (gut); Hyrtl, 1851, 1852b (arteries); Macallum, 1886 (gut and pancreas); Kingsbury, 1897 (encephahc invagina- tions); Miiller, 1897 (pseudobranchs); Alhs, 1908 (pseudobranchs); Allen, 1907, 1908 ( subcutaneous blood vessels of the head and tail respectively); Brookover, 1908 (Pinkus's nerve); Tlieunissen, 1914 (motor nerve arrangement); Danforth, 1916 (coronaiy and hepatic nerves); Gasto, 1966 (liver); Landolt and Hill, 1975 (gill area and respiration). Finally, Goodrich (1930) and Jollie (1962) pro- vide good summary infonnation on gar anatomy. Although traditionally considered primitive actinopterygian fishes, the re- lationships of gars to other major groups have been controversial. Various authors aligned them with the polypterids ( Miil- ler, 1844), the amiids (Huxley, 1861; Goodrich, 1909, 1930; Berg, 1940, 1965; Rayner, 1941, 1948; Nelson, 1969a; Jes- sen, 1972, 1973, and others), and a group composed of amiids and teleosts among Recent fishes (Westoll, 1944; Gardiner, 1960, 1963, 1967; Patterson, 1973). Among fossil fonns, gars have been aligned with aspidorhynchids (Reis, 1887), or semionotids (Rayner, 1941, 1948; Gardiner, 1963, 1967; Romer, 1966), or have been set apart from other known fossil groups (Patterson, 1973). Regard- less of the alignment proposed, no author THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS has doubted their monophyly and gars usually are put in an order or division of their own (Ginglymodi of Cope, 1872; Patterson, 1973; Lepidosteifomies of Berg, 1940; Lepisostei of Suttkus, 1963). The name Ginglymodi is used here for reasons of priority. One of the major aims of this study is to evaluate the various hypotheses of ginglymod relationships to other acti- nopterygian groups using the phyloge- netic methodology of Hennig ( 1966 ) . Various characters of the skull, hyoid and visceral arches, pectoral girdle, and axial skeleton of gars are compared to chondrosteans, semionotids, haleco- morphs (Amia, etc.), and teleosts. Anal- yses of characters are extended to other groups where needed. Characters not previously hypothesized as plesiomor- phous or apomorphous are interpreted while characters previously inteipreted as plesiomorphous or apomorphous (for example, those studied by Nelson, 1969a, and Patterson, 1973) are re-evaluated. Particular attention is placed on synapo- morphies which corroborate the mono- phyletic nature of the Ginglymodi and the synapomoiphic characters shared among groups of actinopteiygians which indicate ginglymod relationships. Recent gars are a relatively well known component of the North and Middle American fish fauna. Except for the work of Suttkus (1963), there has been little recent systematic work on the group, and no comprehensive study of their interrelationships. Gar nomencla- ture began with Linnaeus' (1758) de- scription of Lepisosteus osseus, which he placed with the pikes in the genus Esox. Bloch and Schneider (1801) followed Linnaeus' generic placement when they described Atractosteus tristoechus. La- cepede (1803) placed the gars in their own genus, Lepisosteus. Rafinesque (1818a, 1818b, 1820) added four genera, Litholepis (a mythical fish drawn by Audubon), Sarchirtis, Cylindrosteus, and Atractosteus. Cuvier (1825) erected the family Lepisosteidae. Throughout the nineteenth centuiy many nominal Re- cent species were described by such workers as Agassiz (1843), De Kay (1842), Girard (1858), Gill (1863), Winchell (1864), and Dumeril (1870). The majority of these names since have been placed in synonomy. The prolifer- ation of synonyms was brought about primarily by misunderstanding of onto- genetic changes and geographic vari- ation, and a tendency on the part of some workers to describe specimens from newly sampled areas as new species. The nomenclature and relationships of Recent gars was in confusion through- out the nineteenth centuiy and the first part of the twentieth centuiy because of a poor understanding of the nominal gen- era and the large number of synonyms thought valid. Agassiz (1848) divided gars into sharjo-nosed and flat-nosed spe- cies and (1850) commented on the spe- cies he thought valid. Cope (1865) and Fowler (1910) recognized two genera, Lepisosteus and Cylindrosteus. Dumeril (1870) recognized these genera plus Atractosteus. Jordan and his colleagues attempted to deal with the nomenclature of the entire family, while accepting the manv genera and species then current (Jordan, 1885; Jordan and Gilbert, 1883; Jordan and Evemiann, 1896; Jordan, Evemiann, and Clark, 1930). All these efforts to comprehend gar relationships resulted in a complicated nomenclature, since simplified by recognition of a sin- gle genus, Lepisosteus ( Hubbs and Lag- ler, 1943; Eddy, 1957; Moore, 1957, 1968; Suttkus, 1963), and seven recent species (Suttkus, 1963). Suttkus (1963) provided a key to these species, rede- scribed four of them, and split the genus into two subgenera, Lepisosteus and Atractosteus. For reasons discussed be- low, I recognize these subgenera as genera. The genus Atractosteus, as defined here, includes three extant named spe- cies; A. spatula Lacepede (North and Middle America), A. tristoechus (Bloch and Schneider) (Cuba and the Isle of MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Pines), and A. tropicus Gill (Middle America). In addition, there is a still undescribed narrow-snouted morphotype from the coastal plain of Texas. The status of this moiphotype cannot be de- cided until more material is available for study. The genus Lepisosteus, as defined here, includes four living species, L. platostomus (Rafinesque), L. osseus (Linnaeus), L. oculatus Winchell, and L. platyrhincus (DeKay), all found in the eastern half of North America. Fossil gars are found in North Amer- ica, Europe, Africa, and India. Most names are based on fragmentary mate- rial and the number of valid morpho- types and their interrelationships are essentially unknown. Nevertheless, as discussed below, all of the fossil material can be assigned to one or the other Recent genus. Wood (1846) described the first fos- sil gar, Lepisosteus fimhriatus, a Euro- pean species from the Eocene and Oligo- cene commonly known as L. suession- ensis Gervais ( 1853 ) . Leidy described three of the five North American species recognized here; A. occidentalis (Leidy, 1856a; Cretaceous), A. otrox (Leidy, 1873a, Eocene), and A. simplex (Leidy, 1873a, Eocene). The specimens Leidy used for his descriptions were undiag- nosable, but later workers (Eastman, 1900a; Estes, 1964) described these mor- photypes and associated the names with more complete and diagnostic material. Cope (1884) described the fourth rec- ognized morphotype from the North American Eocene, L. cuneatus. He placed this morphotype in the genus Clastes. These and other workers also named a number of fomis not consid- ered valid here. The last North Amer- ican morphotype recognized here, L. opertus, is described as a new species from the Cretaceous. Kinkelin (1884) added the second European species rec- ognized here, A. strausi. Woodward (1908) described L. indicus from the Cretaceous of India. The last form rec- ognized here is A. africanus, described by Arambourg and Joleaud (1943) under the generic name Paralepidosteus. Two workers. Woodward (1895) and Hay (1902, 1929), attempted to deal with fossil gars as a group. Woodward (1895) justifiably doubted the validity of many of the described fonns. Hay (1902, 1929) summarized the literatiu'e on North American fossil gars. The species listed above are treated in a foiTnal systematic account. This is organized according to the phylogenetic relationships among the species. The di- vision, family and genera are diagnosed by synapomoi-phies (shared derived characters). Each species is placed in phylogenetic order within its respective genus by a listing convention that was previously discussed by Nelson ( 1972a ) and applied to neopterygian fishes by Patterson and Rosen (in press). The species are diagnosed and synonymies provided. The synonymies of Recent species include only name changes or significant systematic references, whereas those of fossil foiTns include all refer- ences found. Fossil material not diag- nosable to species is listed by its as- signed name at the level at which its affinities can be assessed (for example, Atractosteus sp. indet., incertae sedis Atractosteus africanus, etc.). The ob- jective of the systematic accounts is to elucidate relationships among species and not to evaluate intraspecific vari- ation. Thus, the descriptive comments are largely confined to characters im- portant for relationships of named taxa and only secondarily to intraspecific vari- ation that might cause problems in iden- tification of those taxa. In the analysis of phylogenetic re- lationships among gars, each character is analyzed to detemiine its relative apomoq^hous (derived) or pleisomor- phous (primitive) nature. After char- acter analysis, the relationships among taxa are summarized in a series of phylo- genetic hypotheses for each genus. There have been few discussions of THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS gar biogeography outside of descriptions of ranges. Casier ( 1961 ) discussed the biogeography of fossil gars. Rosen (1975) has discussed the distribution of Atrac- tosteus in relation to generalized fea- tures of the Middle American and Antil- lean biotas. Rosen's discussion followed a "vicariance" methodology proposed by Croizat (1958, 1962) and summarized by Croizat, Nelson, and Rosen (1974). His views are evaluated in light of cur- rent knowledge of gar relationships. Dis- tiibution of the other Recent and fossil species are also analyzed by the same "vicariance" methodology. The purposes of this study, stated above, can be summarized in a series of questions: 1. Are the gars a monophyletic group? If so, what characters do they exhibit that corroborate this hypothesis? 2. What characters shared by gars with other actinopterygians permit the evaluation of various hypotheses of acti- nopterygian relationships? Which hy- pothesis is to be preferred? 3. What Recent and fossil species can be recognized within the Family Lepisosteidae? What are the relation- ships among these species as evidenced by shared derived ( synapomorphous ) characters? 4. What are the distributional pat- terns of the family and the monophyletic groups within the family? Can these patterns be tied to generalized patterns of distribution reported by other work- ers? Materials and Methods Fossil and Recent specimens from a number of institutions were used. The material examined is listed for each spe- cies in the systematic account by State or Countiy and District, and locality (if a fossil) in Appendix A, Material Exam- ined. Catalogue Numbers of important specimens of other actinopterygian groups are referred to in the text. StiTic- tural abbreviations used in figures are identified in the figure caption. Institu- tional abbreviations are defined in Ap- pendix A. Fossil preparations. — Fossil prepara- tions include acid, mechanical, and lye preparations. The lye technique was ap- parently originated by Herr Otto Feist of Neider-Ramstadt, West GeiTnany who uses the technique for preparing speci- mens from the Messel fonnation, includ- ing the specimens of Atractosteus stransi described here. Herr Feist imbeds the specimens in fiberglass (apparently in the field), soaks them in a lye solution and scrubs them with brushes, begin- ning with wire and ending with a tooth- brush. I have used a strong KOH solu- tion with the same results. Specimens need little additional preparation. The lye apparently breaks down the hydro- carbons in the Messel matrix, which is a kind of compressed peat. Recent preparations. — Recent speci- mens examined included articulated and disarticulated Recent osteological mate- rial, alizarin preparations, and alcohol preserved whole material. Gill arch stiTicture was studied in alcohol pre- served specimens by excising the arch, soaking it in distilled water overnight, staining it with Alizarin red S, destain- ing in distilled water, and returning it to alcohol (this procedure from G. Nel- son, pers. comm.). Muscle, bone, and cartilage patterns were detemiined by visual inspection. Cartilage was stained with Methylene blue and destained in alcohol during inspection. General methods. — Osteological and myological patterns and meristic counts were detemiined by visual inspection with and without optical aids. Bausch and Lomb and Wild dissecting micro- scopes were used. Meristic counts follow those of Suttkus ( 1963). Enameloid pat- terns were detemiined by visual inspec- tion of unprepared bone and checked in representative specimens of each species by either staining the bone with alizarin (which does not stain the enameloid) or by "smoking" the bone with ammonium chloride (which brings out the enam- 6 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY eloid and bony ridges in relief). The ammonium chloride technique was used for bone photography. Drawings were made with the use of several optical de- vices, including the camera lucida, copy stand, and tracing from photographs. Approximate dimensions of drawings or photographs are provided by a 10 mm bar below the figure or by stating the greatest length of the figured specimen. Color patterns were detenu ined by visual inspection of preserved specimens. The following terms are used to describe these color patterns : Blotch: a concentration of melano- phores fomiing a large pigment patch with more or less definite borders. Flank stripe: a pigment stripe run- ning anteriorly along the side from the base of the caudal fin to or through the eye. Dorsal stripe: a pigment stripe run- ning anteriorly along the dorsum from the base of the caudal fin to the nape of the head. Belly stripe: a pigment stripe run- ning posteriorly from the base of the pectoral fin to the base of the anal fin along the lateral edge of the belly on each side of the midline. Belly stripes on each side usually join at the anal fin base to continue posteriorly as a medial ventral stripe on the caudal peduncle. Preopercular stripe: a pigment stripe running along the lateral ami of the preopercular from its junction with the subopercular anteriorly to the lower jaw. Retroarticular stripe: a small vertical pigment stripe on the back of the lower jaw. Lower jaw stripe: usually a contin- uation of the flank stripe on the lower jaw, and occasionally a separate pigment stripe on the coronoid process. Measurements were taken with either a dial micrometer or dividers measured against a metric rule. Reported measure- ments were selected because they were obtainable or partly obtainable from some fossil, as well as Recent specimens. Neither the measurements nor counts presented here are meant to represent statistical samples, nor to characterize adequately the variation expected in Re- cent North American species. Dr. Royal Suttkus (Tulane University) has taken and continues to take extensive meristic and morphometric data on Recent gars and my efforts in this regard would sim- ply produce duplication. The purpose of measurements in this account is to pre- sent a small number of proportions of dorsal head length for Recent species to be compared with similar proportions obtained from fossils. Dried, articulated material was used exclusively with two exceptions — if the Recent species is not well known (A. tristoechus, A. tropicus), or if less than 8 articulated skulls were available. The measurements used in this study, their abbreviations, and their definitions are listed below. Measure- ments were taken to the nearest 0.1mm. DHL (Dorsal head length): from the most anterior end of enameloid de- velopment of the premaxillary process to posterior junction of the parietals. MHL (Medial head length): from the anterior tip of the rostrum to pos- terior junction of the operculum and suboperculum. SL (Snout length): from the anterior tip of the rostnmi to orbital edge of the anterior medial circumorbital. PS ( Post snout length ) : from orbital edge of the anterior medial circumorbital to posterior junction of the opercular and subopercular. MSW ( Maximum snout width) : dis- tance across the snout at junction of the posterior infraorbitals and the anterior lacrimals. LSW ( Least snout width ) : distance across the snout at junction of the an- terior infraorbitals and the antorbitals. PL (Parietal length): distance be- tween anterior and posterior junctions of the parietals. FL (Frontal length): distance be- tween posterior and anterior junctions of the frontals. PmL (Length of premaxillary proc- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS ess ) : distance from posterior junction of the processes to anterior end of surface ornamentation. LLJ (Length lower jaw) : from sym- physis to end of the angular. OrW ( Orbit width ) : distance be- tween the orbits across top of the skull. OrD (Orbit diameter): maximum distance across the orbit. OpW ( Opercular width ) : from pos- terior junction of the opercular and sub- opercular anterior to the end of surface ornamentation. Systematic Methodology Gar relationships are evaluated using Hennig's ( 1966 ) method of phylogenetic analysis under the philosophy of deduc- tive hypothesis testing advocated by Popper (1968a, 1968b). Popper's ap- proach dictates adoption of an empirical methodology and the attitude on the part of the investigator that he attempt to falsify rather than confimi his hypoth- eses. Hypotheses which have not been refuted are said to be corroborated, and the degree of corroboration is directly related to the number and severit)^ of valid tests applied to them. Where con- flicting hypotheses compete because none is totally falsified by the evidence at hand, the hypothesis that has been rejected the least number of times is preferred (Wiley, 1975). Regardless of the number of times a hypothesis is cor- roborated, it is never considered con- firmed. Rather, a hypothesis must al- ways remain falsifiable and thus subject to refutation if it is to remain a part of science. The phylogenetic methodology of Hennig (1966) has been extensively dis- cussed by its proponents (Hennig, 1950, 1966, 1975; Bmndin, 1966, 1968; Schlee, 1969, 1971; Nelson, 1969a, 1969b, 1970, 1971, 1972a, 1972b, 1973a, 1974; Crow- son, 1970; Farris et al., 1970; Griffiths, 1972, 1973; Cracraft, 1974, 1975; Wiley, 1975) as well as its critics (Colless, 1967, 1969a, 1969b; Mayr, 1969, 1974; Darling- ton, 1970; Bock, 1973; Ashlock, 1971, 1973, 1974; Sneath and Sokal, 1973; Sokal, 1975; and others ) . I do not intend to discuss all the merits of the argu- ments presented by these authors. How- ever, where criticisms of the phyloge- netic method have proven valid, that part of the methodology has been de- emphasized or dropped; this does not invalidate the methodology (for example, the "deviation rule," Schlee, 1971; the "biographic" or "progression" nde, Croi- zat et al., 1974, and Nelson, 1974). Other objections have proved insubstan- tial or have been successfully answered by proponents of the methodology (for example, Mayr's 1974 objections are an- swered by Rosen, 1974b, and Hennig, 1975). There are two alternative meth- odologies. The evolutionary taxonomic method of Simpson ( 1961), Mayr (1969), and others is not vised here because in the only instances in which it differs from Hennig's methods the evolutionary taxonomic methods result in untestable hypotheses (Wiley, 1975). The other possible alternate, numerical taxonomy (Sneath and Sokal, 1973), is not used here because it is concerned with phe- netic, not phylogenetic, relationships. Hennig's major phylogenetic princi- ples may be summarized briefly. Species or groups of species are related by rela- tive recency of common ancestry. Other criteria, such as overall resemblance or occupation of similar "adaptive zones," are rejected as grouping criteria. Taxa that share an immediate common an- cestor are tenned sister groups, and, since they originate from a splitting of the common ancestral species, they have the same time of origin. Only those fea- tures (characters) of taxa that indicate immediate common ancestry are used to elucidate relationships between sister groups. Characters purporting to dem- onstrate sister group relationships (i.e., immediate common ancestory) are termed apomorphous characters, and taxa sharing these characters are said to have synapomorphous characters in com- mon. Characters indicating a phyloge- 8 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY netic relationship but not a sister group relationship are tenned plesiomorphic characters. Characters that do not dem- onstrate common ancestry at any level are tenned nonhomologous characters. Neither plesiomorphous nor nonhomolo- gous characters are pemiitted in the analysis because they do not pertain to the problem at hand: the elucidation of immediate common ancestry between taxa. Taxa must be monophyletic in the strict sense, that is, descended from a single ancestral species and including all descendents of that species. Ancestors are not identified but remain hypothet- ical. Finally, classifications derived from phylogenetic hypotheses must reflect fully the relationships of the phyloge- netic hypotheses. This principle dictates that sister groups have coordinate posi- tions, and therefore coordinate ranks, in the classification, because they have the same time of origin. I will consider three aspects of this methodology in depth. The relationship between a phylogenetic concept of ho- mology and the tenns apomoi-phous, plesiomorphous, and nonhomologous are discussed. The uses of characters for testing phylogenetic hypotheses under Popper's philosophy of deductive hy- pothesis testing are outlined. Finally, the integration of fossil and Recent taxa into a single analysis and classification is discussed. Homology. — A concept of homology is basic to any methodology that makes comparisons between two or more or- ganisms or taxa. In a general sense, characters that can be validly compared in studying relationships among organ- isms are considered homologues, whereas invalid comparisons involve nonhomolo- gous characters. This is not to say that non-homologues cannot be validly com- pared in problems not concerned with relationships. For example, comparisons of bat and bird wings may be perfectly valid to an investigator interested in comparing two kinds of vertebrate flight dynamics. The definition and use of the term homology depends largely on the aims and interests of the investigators em- ploying the tenn. If one wishes to study only the phenetic relationships among taxa, then one will use a phenetic defi- nition of homology (Sneath and Sokal, 1973). Those interested in phylogenetic relationships will adopt a phylogenetic definition (Simpson, 1961; Bock, 1969; Hennig, 1966). I have argued that a phylogenetic definition of homology is preferable to a phenetic definition be- cause it leads to hypotheses of homology that contain all the potential falsifying observations of phenetic similarity and dissimilarity and all the potential falsi- fiers provided by rejection or corrobora- tion of the phylogenetic hypotheses with which the homologies are associated (Wiley, 1975). In this study, two or more characters are said to be homologous if they are transfonnation stages of the same orig- inal character present in the ancestor of the taxa displaying the character (Wiley, 1975, modified from Hennig, 1966). There are two logical derivations of this definition. First, characters that are de- rived in the immediate common ancestor of the taxa compared, and retained in these taxa, may be termed synapo- morphous characters. Such characters are hypothesized as having originated in the immediate ancestral species as unique, or autapomorphous, characters. The presence of these characters in the descendent taxa is evidence of immedi- ate common ancestry of the descendent taxa. Second, characters that are derived in an ancestor more genealogically dis- tant than the immediate common ances- tor, and retained in all subsequent com- mon ancestors of the taxa in question, may be termed symplesiomorphic char- acters. Symplesiomorphies are not evi- dence of immediate common ancestry of the taxa considered because they are not unique to the immediate common ances- tor of the taxa under consideration. Sim- ilar characters hypothesized' not to be THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS present in the immediate common ances- tor of the descendent taxa but present in two or more of the descendent taxa may be termed nonhomologous characters. Convergent and parallel characters are here considered nonhomologous. Whether a particular homologous character is considered a plesiomorphy or an apomorphy depends on the level of universality of the phylogenetic hy- potheses with which the character is associated as a subset. A phylogenetic hypothesis at the highest level of uni- versality would incorporate all species of organisms known or recognized. All characters associated with this hypoth- esis as subsets would be synapomorphies, because all hypothetical immediate com- mon ancestors would be present in the analysis. Thus, characters would be in- corporated into the hypothesis at the point where they originated and all ho- mologous characters would be synapo- morphic. Plesiomoi-phies would not be proper subsets in such an hypothesis be- cause all homologous characters would already be incoiporated into the hypoth- esis where they exist as synapomoiphies. Incorporation of a plesiomorphy would mean that a single homologous character had been used twice in the same analy- sis. But, no one has attempted to pro- duce a hypothesis of the highest univer- sality. Rather, subsets of this phylogeny are evaluated. For example, the lowest level of universality ( if species are taken as the minimal taxonomic units) would be a phylogenetic hypothesis concerning the relationships of three species. Be- tween the lowest and highest levels of universality are phylogenetic hypotheses of varying levels of universality. The level of universality a given hypothesis occupies depends on the number of taxa levels it incorporates; an hypothesis in- corporating four species exists on a higher level of universality than one in- corporating three species. At any level of phylogenetic univer- sality other than the highest both plesio- morphous and apomorphous characters must be considered, because the investi- gator must sort out those characters in the organisms that demonstrate immedi- ate common ancestry from those that do not. This may be framed as a question: which characters are uniquely derived in ancestors included in the hypothesis and which characters are uniquely derived in ancestors not included in the hypothesis, but retained by one or more ancestors included in the hypothesis? The synapo- morphies of a phylogenetic hypothesis are the characters demonstrating im- mediate common ancestory at the level of universality considered. Symplesio- moi-phies differ in that they supposedly demonstrate immediate common ances- tory at a level of universality higher than the hypothesis under consideration. Thus, symplesiomorphies are relevant to a larger phylogeny ( as synapomorphies ) containing the phylogeny under consid- eration as a proper subset. For example, the character "tetrapod limb present" belongs, as a synapomorphy, to the level of universality of the sarcopterygian am- phibians and their sister group, the am- niote sarcopterygians ("reptiles", etc.). This character would not provide a valid corroboration of the hypothesis that liz- ards were more closely related to rhyn- chocephalians than they were to snakes. This is because the character has al- ready been used to test the proposition that the tetrapod sarcopterygians are all more closely related to each other than is any to the rhipidistian sarcopterygians. The use of this character to test the snake-lizard-rhynchocephalian hypothe- sis is an example of the use of a homol- ogy to test a phylogenetic proposition at the wrong level of universality. Homol- ogies applied at the wrong level of uni- versality provide invalid tests of phylo- genetic hypotheses. AutapomoqDhous characters properly belong only to species. (This statement is made under the assumption that spe- cies can be considered individuals as Ghiselin, 1974, has suggested. If how- ever, species must be thought of as 10 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY groups of individuals then I think that autapomoi-pliies would properly belong only to individual organisms and that the "autapomorphy" of a species is only the synapomorphy uniting its component individuals.) When the tenn is applied to a character of a higher taxon, it is only because that higher taxon is consid- ered a single entity for the purpose of testing an hypothesis of relationship and is really being applied to the hypothet- ical ancestral species of the members of that taxon. When used in this way, it has some of the same properties as a plesiomoq^hy in that it cannot elucidate problems of immediate ancestry between the taxa considered, and it is an un- testable homology ( because the ancestor is not observed ) . Two further character- istics are apparent — its nature does not change as the level of universality is raised (like a synapomorphy), and its nature does change with a lowering of the level of universality (like a synapo- morphy), but it does not immediately change to a symplesiomoiphy, but to a synapomorphy. At a level at which it is synapomorphous, it is both testable as an homology and pertinent to elucida- tion of common ancestry. When the term autapomorphy is ap- plied at the species level, it is simply a unique character. There seems to be no phylogenetic argument that can be ap- plied to test a unique character as de- rived other than by showing it to be a member of a synapomorphous pair. Thus, unique characters are accepted as apomorphous only by parsimony. Be- cause parsimony does not constitute a test of a scientific hypothesis, I conclude that autapomorphies, like symplesio- morphies, are not testable propositions and cannot themselves be applied as tests of phylogenetic hypotheses. Each hypothesis of synapomorphy is tested in a two-step process. First, it may be tested by its own set of potential falsifiers, without reference to the phylo- genetic hypothesis with which it is asso- ciated as a proper subset. Most potential falsifiers in this round of testing are sim- ilarities and differences, morphological or otherwise (Wiley, 1975). In the sec- ond stage of testing, the hypothesis of synapomorphy is associated with an hy- pothesis of phylogeny, and the phylog- eny and synapomorphy are tested with other hypotheses of synapomorphy. If these other hypotheses of synapomorphy refute the phylogeny, they also refute the supposed synapomorphy unless one of two conditions is found: (1) the re- futing "synapomorphy" is actually a symplesiomorphy, or (2) that the refut- ing "synapomorphy" is actually a non- homology. Neither of these types of characters represent valid tests (for rea- sons discussed above) and thus neither can refute the phylogenetic hypothesis (Wiley, 1975). If other hypotheses of synapomorphy are congruent with the phylogeny and its associated synapomor- phy, then both hypotheses ( phylogenetic and homologous) are said to be corrob- orated. The refutation of a character as a synapomorphy and corroboration of that character as a symplesiomoiphy can only be accomplished by raising the level of universality of the problem at hand either by finding the character in the sister group of the entire system tested, by finding the character so commonly outside the group that it is considered a symplesiomoiphy, or by applying a de- velopmental or ontogenetic rule of char- acter transformation. Such a laile auto- matically raises the level of phylogenetic inquiry, for such rules are held only because they are common to large num- bers of organisms outside the group of immediate interest. Testing phylogenetic hijpotJieses. — All statements of phylogenetic relation- ships involve a minimum of three taxa at the lowest level at which they can be tested. Such hypotheses usually take the fomi that two taxa are said to share a common ancestor not shared by the third taxon. This relationship can be ex- pressed by a phylogram. The distiibu- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 11 tion of characters among the three taxa provides deductive tests of the relation- ship. Without reference to Hennig's spe- cific method but with reference to any test of relationship, we may say the only tests that can be considered valid at- tempts to refute a given three-taxon problem are those involving a single character present in two and only two of the three taxa. This is because only such a character is capable of refuting a given hypothesis of relationship. Characters shared by all three taxa or characters unique only to one are congnaent with the four possible testable hypotheses for any three taxa. In a purely phenetic system, all shared characters, whether primitive or derived, have equal status in discussions about which branching diagram should be adopted as most par- simonious. Hennig's (1966) method dif- fers fundamentally from a purely phe- netic method in that all the shared characters are not used to refute a given relationship; only synapomorphous char- acters are used. Such testing can be accomplished only in an open system, that is, by considering taxa outside the three (or more) taxon system. Such con- siderations may be termed outgroup comparisons. The one condition placed on this procedure is that the three (or more) taxa must foiTn a monophyletic group. The designation of outgroups for comparison pemiits an investigator to sort out which of the observed charac- ters are unique to the three taxon system and which characters have a more gen- eral distribution. The outgroup compar- ison automatically raises the level of universality of the phylogenetic hypoth- esis to a new level. And, it allows the investigator to put his three-taxon prob- lem in context with an hypothesis of a higher level of universality. The most parsimonious phenetic solution without reference to the outgroup may not nec- essarily be the most parsimonious phylo- genetic solution within the context of the higher level phylogeny. So, to achieve overall parsimony, the phylogenetic in- vestigator will have to reject certain characters as valid indicators of relation- ship (G. F. Engelmann, pers. comm.). And, it is the characters analyzed as symplesiomorphies or nonhomologies that can be objectively rejected as valid indicators of relationship within the three-taxon problem. This is because they are not pertinent to the elucidation of immediate common ancestry (as dis- cussed above), and that acceptance of plesiomorphies as valid indicators within the context of the higher level phylogeny would lead to rejection of the tliree- taxon unit as a monophyletic group; tliis would violate the basic condition of the validity of the investigation itself. Syn- apomorphies, then, are the only valid tests of a phylogenetic hypothesis, and tliis testing is carried out as discussed above in the paragraph about testing synapomorphies. It should be pointed out here that a synapomorphy which is used to produce a phylogenetic hypoth- esis via "induction" does not test that phylogeny. Only after the phylogeny is proposed do synapomorphies provide deductive tests. Finally, when a situ- ation exists where all hypotheses have been rejected, then that one wliich has been rejected the least number of times is preferred on the basis of parsimony. Integrating Fossil and Recent taxa. — Two problems are relevant here. First, can fossil and Recent taxa be integrated into the same analysis and classification? Second, should ancestor recognition be attempted if fossil and Recent taxa are integrated into the same analysis? Patterson and Rosen (in press) have concluded that fossil and Recent taxa can be integrated into the same analysis and classification and that the objections of Hennig (1966) and Crowson (1970) can be overcome by considering fossil taxa as terminal branches. Further, they have outlined two conventions permit- ting a classification of both fossil and Recent taxa that exactly reflects the phy- logenetic relationships of all taxa in the classification. These conventions are out- 12 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY lined in the classification of gars pre- sented below, I conclude from their discussion that the integration of fossil and Recent taxa is possible with the phylogenetic methods of Hennig ( 1966). Regarding recognition vs. non-recog- nition of ancestors, two questions are relevant. First, can ancestral taxa be supraspecific taxa or must they be spe- cies or populations? Second, can any jDarticular hypothesis of ancestor-de- scendent relationship be tested in an objective manner? The question of whether supraspecific ancestors can be recognized is closely tied to the definition of monophyly adopted. The concept of monophyly sensu Hemiig (1966, or if one prefers, Hennig modified sensu Ashlock, 1973) leads to the logical corollary that supra- specific taxa cannot be considered an- cestors because some of the members of the "ancestral" taxon would actually be more closely related to taxa outside the group than to taxa inside the group. On this basis I would agree with Patterson and Rosen (in press) that ancestral units, if recognized, must be species or popu- lations and not supraspecific groups. One might argue, however, that Hen- nig's definition of monophyly is not pre- ferred. Hennig's concept of monophyly has at least three points which, in my opin- ion, make it superior to concepts of monophyly outlined by Simpson ( 1961 ) , Ashlock (1971), and Mayr (1974). The first point concerns the evolutionaiy proc- ess. What little we know of evolutionary process indicates that populations are the highest level taxa that evolve and spe- cies are the highest level taxa that can be considered populations. Since higher taxa do not evolve, they cannot be con- sidered ancestral units. The reality of higher taxa is based solely upon whether they are accurate reflections of past spe- ciation events and thus higher taxa are historical constructs and have no reality as active units of evolution. The second point concerns the concept of group membership. Hennig's (1966) definition of monophyly dictates that all the de- scendents of an ancestor be placed in the same group as the ancestor. This confonns to basic set theory and is meth- odologically precise. The removal of a species or group of species from a taxon that includes its ancestor and sister group and then calling the original group monophyletic and ancestral to the species or group of species is untenable because, firstly, it breaks up a logical set of two logical subsets into one logical set (the species or group of species, the "s" group of Ball, 1975:413) and one illogical set (the "ancestral" taxon, whose general components can no longer be de- fined, the "s-bar" group of Ball, 1975:413), and, secondly, it makes the tenn mo- nophyly an open term, that is, any taxon could be teimed monophyletic ( Hennig, 1966). Thirdly, Rosen (1975) has pointed out that use of non-monophyletic groups sensu Hennig (1966) in biogeographic analysis leads to apparently incomplete distribution patterns because some spe- cies, or groups of species, are not classi- fied with their nearest relatives. This leads to an underestimation of the an- cestral range of the hypothetical common ancestor of the non-monophyletic group. Thus, adoption of an evolutionary taxo- nomic definition, such as that of Simp- son (1961), obfuscates biogeographic analysis. Two of Hennig's original premises are that ancestors (=ancestral species) remain hypothetical and that the search for ancestors is futile. Hennig specifi- cally rejected the idea that ancestor- descendent relationships between two or more species could be scientifically dem- onstrated in either the phyletic sense (A gives rise to B through time without a geographic isolation event) or the cla- distic sense (A gives rise to B and C by splitting). Nelson (1970a, 1973a) and Cracraft ( 1974 ) state that ancestors are empirically unknowable. Whether this position or Bock's (1973) assertion to the contrary is correct THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 13 will not be taken up here for three rea- sons : ( 1 ) my thinking on these ques- tions has been influenced by my col- league George Engelmann and thus is better left for a future joint paper; (2) those species of gars diagnosed without autapomorphies and thus which might in some circumstances be considered possible ancestors are so fragmentary that hypotheses of ancestry can be put aside for lack of information; (3) all other fossil gars studied are derived, and questions of specific ancestors never came up in the analysis. I will state, as an unsupported conjecture, that hypoth- eses of ancestry relating to this study of gars are trivial hypotheses and were not pursued. BlOGEOGRAPHIC METHOD The panbiogeographic or "vicariance" method is used to describe certain dis- tributional patterns among gars. This methodology has been extensively dis- cussed by its originator, Croizat ( 1958, 1962), and summarized by Croizat, Nel- son, and Rosen (1974) and Rosen (1974a). The vicariance method attempts to find general patterns of distribution to provide a general solution to explain the individual species pattern in the most economical manner. This is accom- plished by track analysis. A track is a line around the range of a species or enclosing the various ranges of the taxa of a monophyletic group. When more than one species is enclosed within the track, then the track is an estimate of the range of the ancestor of the monophy- letic group. Tracks enclosing non- monophyletic groups, such as paraphy- letic groups, are not as infonnative as those enclosing monophyletic groups be- cause they enclose only part of a mono- phyletic group and therefore underesti- mate the range of the common ancestor of the taxa included within the track. Analysis begins by plotting as many as possible of the tracks observed and looking for general distributional pat- terns, that is, those comprising more than a single track. Such general distri- butional patterns are termed generalized tracks. The more individual tracks mak- ing up the generalized track, the more corroborated the generalized track. Ad- ditionally, generalized tracks made up of a number of distantly related taxa are more highly corroborated than general- ized tracks with the same number of individual tracks made up of closely re- lated taxa. The generalized track esti- mates an ancestral biota in the same way an individual track estimates the ancestral range of the ancestral species of the monophyletic group. The presence of a generalized track rejects the hypothesis that the individual tracks comprising it could have origi- nated by the independent chance dis- persal of, or migration of, each individ- ual species comprising the track. This is because it is assumed that the chances of a general distributional pattern emerging from the independent dispersal of many component species which have diflFerent biological needs and dispersal capabilities is nil. Instead, a generalized track calls for explanation on a general level. That is, it calls for an explanation that involves the biota as a whole to produce the observed pattern, rather than separate explanations involving each member of the biota individually. The vicariance method recognizes the reality of dispersal. Dispersal is identified in one of four ways : ( 1 ) by the overlap of generalized tracks, ( 2 ) by the observation that a species or group of species does not confonu to any gen- eralized track, ( 3 ) by direct observation of dispersal or migration and subsequent settlement, and (4) by sympatry be- tween sister groups. All of these meth- ods of identifying dispersal except (3) are founded on the assumption that the process of speciation is usually allopatric. The Relationship of Gars to Other actinopterygians Gars traditionally have been consid- ered primitive actinopterygian fishes, but 14 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY their exact sister group relationships have been subject to dispute. Miiller (1844) included the gars and Amia with the polypterids in his order Holostei. Huxley (1861) removed the polypterids from the Holostei and placed them in the Crossopterygii. This alignment of Amia and gars was accepted by most workers of the nineteenth century and has most recently been advocated by Nelson (1969a) and Jessen (1972, 1973). Others have argued that Amia is more closely related to teleosts than to gars and that Holostei is a grade or para- phyletic group ( Westoll, 1944; Gardiner, 1960, 1963, 1967; JolHe, 1962; Patterson, 1973 ) . When fossil fishes are considered, gars have been considered by most work- ers as being most closely related to se- mionotifonns (Goodrich, 1909, 1930; Rayner, 1941, 1948; Gardiner, 1960, 1963, 1967; Romer, 1966) or aspidorhynchids (Reis, 1887; Goodrich, 1904). These con- clusions were challenged by Patterson (1973) who argued that semionotifonns are more closely related to halecomorphs (Amia) and teleosts than to gars, and that aspidorhynchids are teleosts. In assessing the relationships of gars to other actinopterygians, I attempted to establish which of the characteristics of gars are autapomorphic, which are syn- apomorphic with one or more non-gar actinopterygian groups, and which are symplesiomorphic or nonhomologous and thus of no value in assessing phylo- genetic affinities. Three major anatomi- cal areas are covered: the skull, the visceral arches, and the post-cranial skeleton. Four major hypotheses of gar rela- tionships are tested: ( 1 ) Gars and Amia are sister groups and are more closely related to teleosts than are chondrosteans (Nelson, 1969a; Fig. la). (2) Gars and Amia are sister groups and chondrosteans are the sister group of teleosts (Jessen, 1972; Fig. lb). (3) Amia is the sister group of tele- osts and gars are more closely related to this group (haleco- stomes) than are chondrosteans (Patterson, 1973; Fig. Ic). (4) Semionotids are the sister group of gars and this group the sister group of amiids and teleosts (Westoll, 1944 and others cited above; Fig. Id). Fig. 1. — Four hypotheses of the relationships between chondrosteans (C), gars (G), amiids ( A ) , teleosts ( T ) , and semionotids ( S ) . a. after Nelson (1969a); b. after Jessen (1972); c. af- ter Patterson (1973); d. after Rayner (1948). The Skull The pattern of dermal and endo- chondral ossifications is basically similar in gars, other neopterygians, and in many chonodrosteans, making hypoth- eses of homology possible (Patterson, 1973:239). The pattern of dernial ossi- fication of gars is illustrated in Fig. 2. The etlimoid region of gars differs from those of other actinopterygians in being elongated to produce the charac- teristic lepisosteid snout. There are no etlimoid ossifications. Patterson ( 1975: 499) reviewed the occurrence of ethmoid ossifications in various actinopterygian groups and concluded that etlimoid ossi- fications are primitive for osteichthyans. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 15 Fig. 2. — Lateral (upper) and dorsal (lower) views of the skull of Lepisosteus oculatus (LACM 33916-3). a, angular; Ant, antorbital; Co, circumorbital; d, dentary; d Co, dorsal circumorbital; 'Dhy', "dermohyal"; Dpt, dermopterotic; Dsp, demosphenotic; Fr, frontal; lo, infraorbital; Lac, lacrimal; Na, nasal; Op, opercular; Pop, preopercular; Q, quadrate; Qj, quadra tojugal; Ro, rostral; s, surangular; So, suborbital; Sop, subopercular; St, supratemporal. Lateral ethmoids are found in teleosts, Amia, large Acipenser, Pohjpterus, and Latimeria (Patterson, 1975). They are also found in palaeoniscids (Bergeria, Nielsen, 1949; Perleidtis, Patterson, 1975), fossil halecomoiphs (Catiirus, Rayner, 1948; Sinamia, Stensio, 1935), "holosteans" ( Macrepistius, Schaeffer, 1971; Lepidotes, Patterson, 1975). Fully ossified forms, such as some paleoni- scids, all parasemionotids, and the se- mionotifonns Heterolepidotes and Da- pedium, probably had internal ossified ethmoids dvning development (Patter- son, 1975:499). I conclude that the elon- gation of the etlimoid cartilage and the lack of ossifications in this cartilage are synapomorphous characters shared by gars. Gars and palaeoniscids lack an en- doskeletal rostrum of the type seen in Recent chondrosteans, saurichthyids, amiids, and teleosts. Independent ossi- fications of this cartilage are known only in Amia, pachycomiids, and Recent tele- osts, but could have been present during the ontogeny of solidly ossified forms such as saurichthyids, pholidophorids, Dapedium, and Heterolepidotes (Patter- son, 1975:502). Patterson (1975) con- cluded that (1) an endoskeletal rostmm with a well defined nasal septum and laterally or dorsally oriented nasal pits was independently derived in several lineages and that lack of this cartilage is plesiomoiphic for actinopterygians, and (2) an endochondral rostral bone in front of the lateral etlimoids is synapomorphic for halecostomes (the pre-ethmoids of Amia and P achy cor mus, the supra- ethmoids and ventral ethmoids of tel- eosts ) . The dermal components of the lepi- sosteid snout consist of a medial rostral, paired nasals and antorbitals, underlain by the premaxilla and paired vomers. 16 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY The number and position of demial snout elements have been discussed by many authors. Gardiner (1963) con- ckided that the primitive pattern of ossi- fication consisted of a postrostral sepa- rating the nasals and a pair of compound rostro-premaxillo-antorbitals. Schaeffer (1973) suggested that there was a me- dial rostal as well as postrostral. Wenz (1968) concluded that evidence did not peimit a primitive hauplan for the acti- nopteiygian snout to be established, and Patterson ( 1975 ) apparently agreed. Al- though the snout ossifications of chon- drosteans are variable, the pattern within the Neopterygii is more stable and per- mits unambiguous comparisons. Gars are similar to other neopterygians in having a medial dermal rostral which carries the etlimoid commissure of the infraorbital canals, paired nasals which contact each other and carry the supra- orbital canals, and separated antorbitals which carry the infraorbital canals. These are underlain by paired premaxil- laries and vomers. The rest of the snout is overlain by the ascending processes of the premaxillaries and the frontals. Both pairs of bones carry the supraorbital sensory canal. The snout is bordered by a series of toothed infraorbitals that carry the infraorbital sensory canal. Gars differ from other actinoptery- gians in having two commissures be- tween the supraorbital and infraorbital sensory canals (Fig. 3a). The first is produced by a backward bending of the supraorbital canal on the nasal. This commissure joins the two sensory canals between the nares and traverses the na- sals and antorbitals. The second com- missure is a branch of the infraorbital canal that runs from the antorbital to the premaxillary arm posterior to the nares (Jollie, 1969). Jollie (1969) stated that this condition is unique among ac- tinopterygians and has suggested that the internarial commissure is homolo- gous with that in Polypterus and Aci- penser but not homologous with the in- ternarial commissure in Amia (which P com Sor Pm Fig. 3. — Lateral view of the snout of five osteichthyan fishes. A. Lepisosteus after Jollie, 1969; B. Amia after Jollie, 1969; C. Elops after Forey, 1972; D. Eusthenopteron after Stensio, 1947; E. Pteroniscuhis after Stensio, 1947. Ant, antorbital; E. com, etlimoid commissure; Fr, frontal; I. com, internarial commissure; lo, in- fraorbital; loc, infraorbital canal; Na, nasal; P. com, postnasal commissure; Pmx, premaxilla; Ro, rostral; Ro-Deth, rostro-dermethmoid; Sor, supraorbital sensory canal; Vo, vomer. loins between the nares but does not involve the nasal and develops late in ontogeny. Fig. 3b). The teleostean snout has a single commissure, which is a branch of the infraorbital canal travers- ing the antorbital (as in Amia) and either connects with the supraorbital ca- nal posterior to the nares ( osteoglossids, Nelson, pers. comm.) or ends before meeting the supraorbital canal {Elops, Nelson, 1969c; Fig. 3c). These differences are best assessed by surveying patterns of sensory canals and commissures among other gnatho- stomes. In the snout of sharks, the prim- itive pattern may be that of Chlamy- doselachus, in which the infraorbital and supraorbital canals meet behind the nares (JolHe, 1969). In acanthodians the supraorbital canal is joined to the infra- orbital canal anterior to the orbit, but the relationship of the commissure to the nares is apparently unknown (see Wat- ton, 1937). Among sarcopterygians there THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 17 are several snout patterns. The dipnoans Protopterus (Panchen, 1967) and Neo- ceratodus ( Stensio, 1947 ) have no com- missures between the supraorbital and infraorbital canals but each is connected via a commissure to its counterpart across the snout. In rhipidistians (Fig. 3d) there is a single commissure between the supraorbital and infraorbital canals in front of the snout, and the infraorbital canals are connected to each other across the snout (Holoptychhis, Jarvik, 1947; Etisthenopteron, Jarvik, 1944; see Sten- sio, 1947, for summary but note that fig- ure 29b, p. 106 is probably incorrect). In Latimeria chalumnae there are commis- sures between the supraorbital and in- fraorbital canals both anterior to the nares and between the nares ( Millot and Anthony, 1958; Jollie, 1969). In at least one fossil coelacanth (Nesides schmidti) there is no indication of an internarial commissure. I note that Pohjpterus dis- plays the sarcopterygian pattern (see Jollie, 1969, and Stensio, 1947, for illus- trations ) . Within the Actinopteiygii the prim- itive condition (Fig. 3c) seems to be that in which the infraorbital and supra- orbital canals are connected by an inter- narial commissure and the infraorbital canals on each side of the head are con- nected via the rostral commissure (Gar- diner, 1963; Jollie, 1969). This condition is found in Pteronisculus (Nielsen, 1942), Bergeria and Austral osomus (Nielsen, 1949), Moythomasia (Jessen, 1968), Ategotrachelus, Kentuckia and others, but not in Canobius and Babastrania (Gardiner, 1963). This snout sensory ca- nal pattern is probably apomorphic for the Actinopterygii and thus plesio- morphic within the group. What the plesiomoiphous pattern is for teleos- tomes in general is not clear. The pat- tern seen in the coelacanth Latimeria incorporates both the rhipidistian and actinopterygian patterns. Dipnoans have a different, and presumably apomor- phous, pattern. But whatever the basic teleostome pattern may be, it is logical to assume that the internarial commis- sure between the supraorbital and infra- orbital canals seen in gars is homologous with the same commissure seen in most chondrosteans. And, the presence of a postnarial commissure between these sensoiy canals is hypothesized to be apo- moiphous for taxa within the Actino- pterygii. Two questions remain: (1) is the internarial commissure of Amia ho- mologous to the internarial commissure of gars and chondrosteans?; and (2) what are the homologies of the post- narial commissures of gars and teleosts? Jollie (1969) stated that the internarial commissure of Amia calva is not fully developed until relatively late in devel- opment (200 mm). Alhs (1889) figured the early development of the commis- sure. In 10 mm specimens of Amia calva the commissure has the same orientation as adult Elops, that is, the branch of the infraorbital canal is directed upward and is behind the nares. It then contacts the edge of the posterior nares (11.5 mm), and, as the posterior nares move farther back during development, the infra- orbital branch comes to lay between the nares. These ontogenetic changes cor- roborate a hypothesis that the internarial commissure of gars and Amia are non- homologous. Further, it corroborates a hypothesis of synapomorphy between the internarial commissure of Amia and the postnarial commissure of teleosts and gars. Thus it would seem that a postnarial commissure or an infraorbital branch on the antorbital is apomorphous for neopteiygians. Gars are unique in having both the plesiomorphous and apomorphous commissures. Amia and teleosts lack the internarial commissure of gars and chondrosteans, an apomor- phous condition. Further, Amia calva is autapomorphous in having the postnarial commissure between the nares. The premaxilla of gars includes a toothed anterior part, the premaxilla proper, and the premaxillary or nasal process (pmx. Fig. 2, and Patterson, 1973) which makes up as much as 50% of the 18 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY length of the snout and carries the supra- orbital canal. Hammarberg (1937) con- tended that the supraorbital canal was found on the nasal process because two nasal ossification centers and their asso- ciated sensory placodes were incorpo- rated into the nasal process. Thus, he concluded that the premaxilla is a com- pound bone composed on one anasmatic bone and parts of one sensory canal bone. He temied this bone the pre- maxillo-nasals (other names applied to the premaxilla of gars include the ethmo- nasals, Allis, 1905, and the naso-premax- illaries. May hew, 1924). Aumonier (1941) studied the premaxillary process and concluded that no nasal ossification centers were involved and that the asso- ciation of the premaxillary process with the supraorbital canal was produced by simple posterior growth of the premax- illary arm. Further, he could find no evidence that the premaxilla arose from two ossification centers. Patterson ( 1973 ) compared the pre- maxillary process of gars with that of Amia and he pointed out three basic similarities between the two: (1) both line the nasal pits, (2) both suture with the frontal, and (3) both are perforated by the olfactory nerve. He surveyed the distribution of nasal processes and found them in parasemionotids, semionotids, caturids, and amiids. The olfactory nerve perforated the nasal process in Semiono- ttis, Lepidotes, and Eunjcarmus (loc. cit.: 510). Patterson also found that the lat- eral dermethmoids of philodophorids oc- cupy the same topographic position as the nasal processes of amiids and the fossils named above, and that these lat- eral dermethmoids differ from the nasal process and are primitively toothed. Patterson ( 1975) concluded that the pre- maxillary processes of all neopterygians are basically homologous and that pre- maxillary processes arose by backgrowth of the small process such as that seen in parasemionotids. There are, however, differences be- tween the premaxillary processes of gars and Amia. Pehrson (1940) studied the ontogeny of the premaxilla of Amia and concluded that it was a compound bone composed of two ossification centers, the premaxilla proper and a posterior ossi- fication center, the rhinal bone of Bjer- ring ( 1972 ) . In contrast, the gar pre- maxilla has either a single ossification center (Aumonier, 1941:20), or is de- rived from an anasmatic bone and parts of a sensory bone (Hammarburg, 1937). Apparently, the developmental se- quences of the premaxillae of Amia and gars differ. In addition, the premaxillary process of gars makes up a significant portion of the dorsal surface of the snout, whereas that of Amia and the semiono- tids I have examined lies beneath the nasals. Finally, parasemionotids, the pre- sumed sister group of amiids, lack a well developed premaxillary process, and, unless parasemionotids are apo- morphic in lacking a well developed process, we must assume that the com- mon ancestor of all halecomor]Dhs also lacked a well developed premaxillary process. Regardless of the significance placed on the topographic dissimilarity or developmental dissimilarity, the phy- logenetic argument seems to refute the conjecture that the premaxillaiy process of gars and Amia is homologous. Whether the premaxilla of Amia is ho- mologous to that of teleosts (Patterson, 1973) is also open to question, but I have no additional observations to add to those of Patterson (1973, 1975). The vomers of gars are paired. This is apparently the plesiomorphous condi- tion and is found in a variety of other actinopterygians (see Patterson, 1973). Lepidotes, some other semionotids, ha- lecomorphs, and teleosts have a single vomer (Rayner, 1948), a character that refutes the ancestor-descendent or sister group relationship proposed by Rayner (1948) for gars and Lepidotes (Patter- son, 1973). The vomers overlie the para- sphenoid and the ectopterygoids. These bones will be discussed under the palatal section below. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 19 The remainder of the dorsal snout is made up of the anterior half of the fron- tals bordered by a series of 5 to 10 toothed infraorbital bones and the atro- phied maxilla. Both the long series of toothed infraorbitals and the atrophied maxilla are unique for gars and are con- sidered synapomorphic for the group. In early growth stages the maxilla is much larger and occupies the correct topo- graphic position of the nonnal actino- pterygian maxilla. It begins to atrophy relative to other bones when the young reach about 26 mm total length (Ham- marburg, 1937; Aumonier, 1941). Lepi- dotes has a small series of infraorbitals running onto the snout (Westoll, 1937), but these are neither toothed, nor as numerous as those of gars, nor do they border the snout margin. I conclude that a hypothesis of synapomoqihy be- tween the infraorbitals of gars and Lepi- dotes is a weak hypothesis and that the presence of non-toothed infraorbitals in Lepidotes is a plesiomorphous condition. The dermal bones posterior to the snout of gars are the usual frontals, pa- rietals, demiopterotics, supratemporals, and post-temporals dorsally, and the lac- rimals, circumorbitals, suborbitals, and the opercular series laterally (see Fig. 2). Gars differ from other actinoptery- gians in the relationships of the denuo- pterotic and dennosphenotic to their en- dochondral counterparts, the sphenotic and pterotic. Gars have a demiopterotic- sphenotic articulation with the distal end of the sphenotic frequently being seen exernally under the lateral wing of the demiopterotic. If Patterson's (1975) hypothesis that the pterotic is present in gars and missing in amiids is correct (see discussion of neurocranium below ) , then gars have both the sphenotic and pter- otic articulating with the dermopterotic and they have a dennosphenotic without an endochondral articulation. Amiids and teleosts differ in having an epi- occipital-dei-mopterotic articulation and the usual sphenotic-dermosphenotic ar- ticulation primitive for actinopteiygians. And, amiids are unique in having lost the pterotic and thus in lacking a pter- otic-demiopterotic articulation. Behind the parietals and demiopter- otics of actinopterygians is a supra- temporal series that carries the extra- scapular or supratemporal commissure between the supraorbital canals. Gars have two to six supratemporals on each side of the midline (Fig. 4b) whereas amiids and most teleosts have a single supratemporal on each side of the mid- line (Fig. 4c, d) (or, in some teleosts, a complete loss of the supratemporals). Palaeoniscoids such as Moijthomasia (Jessen, 1968) and Pteronisctdus (Fig. 4a, from Nielsen, 1942), branchoptery- gians (Polypterus, Daget, 1950), and a variety of sarcopterygians have two or more supratemporals on each side of the midline (Fig. 4a). I hypothesize that this condition is plesiomorphous. Within the Neopterygii, semionotfforms have two per side. A single supratemporal per side is found in the fossil haleco- morphs ( as figured from various sources by Lehman, 1966) Parasimionotus (Fig. 118), Promeco.somia (Fig. 122), Eoeu- gnathus (Fig. 126), Meter olepidofes (Fig. 129), Fiiro (Fig. 128), Caturus (Fig. 133), Oneoscopus (Fig. 135), Urocles (Fig. 136), Microsemius (Fig. 143), Ophiopsis (Fig. 141). A single supratemporal per side is also found in Pachycormus (Lehman, 1966, Fig. 147); and in the philodophorids and lepto- lepids (but not in Sinamia, Lehman, 1966, Fig. 138). I hypothesize that a single supratemporal on each side of the midline is a synapomorphy uniting ha- lecomorphs and teleosts. Gars have a complete circumorbital ring ending anteriorly in three lacrimals, a condition regarded by Gosline (1965) as primitive (Fig. 2). Patterson (1973) found this conclusion premature. Some palaeoniscoids do not have a complete circumorbital ring. Rather, the nasals form the anterior borders of the orbits (Pteronisculus, Boreosomus, Canobius, and platysomids, Stensio, 1947; Poly- 20 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY pterus, Jarvik, 1947). Others have a complete circumorbital series (Cheiro- lepis, Watson, 1925; Discellopyge, Brough, 1931). The most ancient se- mionotifonn, Acentrophorus, lias a com- plete circumorbital ring (Gill, 1923), as do other semionotiforms such as Lepi- dotes (Westoll, 1937), Semionotus (Leh- man, 1966), Dapedium (Wenz, 1968). Thus, it appears that a complete circum- orbital ring is primitive within the Neo- pterygii, regardless of its condition within Actinopterygii. I conclude that the secondaiy loss of some components of the circumorbital series is a synapo- morphy uniting halecostomes. The dermosphenotic carries the junc- tion of the otic and postorbital portions of the infraorbital canal. In some gars the dermosphenotic is incoiporated into the circumorbital ring, a plesiomoq^hous character shared with some palaeoni- scoids, Amia, and teleosts. The relationships of the supraorbital and infraorbital canals relative to the dennal roofing bones are variable within the Actinopterygii. In palaeoniscids there is, primitively, no commissure be- tween the two canals (Fig. 4a). This condition is also seen in young Amia calva (Allis, 1889) and adult Pholido- phorus macrocephalus (Patterson, 1975). In gars there is a commissure between the otic branch of the infraorbital canal and the supraorbital canal on the dermo- pterotic ( Fig. 4b ) . In Amia the commis- sure is further forward and involves the frontals (Fig. 4c). Teleosts primitively lack a commissure ( pholidophorids, lep- tolepids, Elops, Fig. 4d), but more de- rived groups have a commissure (Gos- line, 1965). The condition in gars (and perhaps Amia and some Recent teleosts) is apparently derived, and such commis- sures apparently have been derived in- dependently in several lineages. Gars have a mosaic of suborbitals between the circumorbitals and the oper- cular series (SO, Fig. 2). The presence of suborbitals is considered plesiomor- phous by Schaeffer (1973) and Patterson Fig. 4. — Dorsal view of the posterior skull roof of four actinopterygians. A, Pteronisculus, after Nielsen, 1942; B, Lepisosteus, after Patterson, 1973; C, Amia after Patterson, 1973; D, Elops, after Forey, 1972. Dpt, dermopterotic; Dsp, dermosphenotic; Fr, frontal; loc, infraorbital canal; Pt, post-temporal; Soc, supraoccipital; Sor, supraorbital sensory canal; Sor conim, supraorbital commissure; St, suprateniporal; St. com, suprateniporal commissure. (1973). They are found in a large num- ber of palaeoniscids and fossil haleco- stomes (see Patterson, 1973, for discus- sion) but are missing in Amia calva and recent teleosts. Patterson (1973:245) in- teq^reted this absence of suborbitals as evidence for relationship because, al- though some primitive teleosts (i.e. pho- lidophorids) have a few suborbitals, the parallelism is an "almost unique condi- tion." I cannot accept this conclusion. A character that can demonstrate rela- tionships is one found in the common ancestor of the taxa related. But most evidence leads to the conclusion that the common ancestor of amiids and teleosts had suborbitals and that primitive tele- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 21 art. Dpi Fig. 5. — Pterygoid bones of Atractostetis tristoechus (USNM 111309). Upper, dorsal view; Lower, lateral view. Dermopalatine omitted, art. Dpi, articular surface of dermopalatine; art. lo, articular surface of infraorbitals; Ecpt, ectopterygoid; Enpt, endopterygoid; Mpt, metapterygoid; Q, quadrate. osts ( pholidophorids and leptolepids) retained this character. The absence of suborl)itals is a parallehsm and not a synapomorphous condition in spite of the uniqueness of the parallehsm. Thus, it cannot be used to indicate a sister group relationship between Amia and teleosts. Incorporated in the suborbital mosaic of gars is a bone identified by Jollie (1962) as the demiohyal (Dhy, Fig. 2). Patterson (1973:245) concluded that the evidence for identification of this bone is weak, and that, while it may be the demiohyal, it could be the homologue of the suprapreoperculum. I agree with Patterson and have been unable to de- termine if the bone is a demiohyal. The question has some phylogenetic signif- icance, for if it is the demiohyal, then aniiids and teleosts would share the syn- apomorphy of loss of the bone. If it is not the demiohyal, then gars share the loss with other neopteiygians, making the character a synapomorphy of neo- pterygians. The opercular series of gars includes the opercular, subopercular, and pre- opercular (Fig. 2). Gars lack an inter- opercular, as do palaeoniscids. Haleco- nioiphs, semionotifomis, and teleosts have an interopercular, a synapomorphus condition relating them and excluding gars. Conjectures concerning the sec- ondary loss of the interopercular of gars (Rayner, 1948) depend on the assump- tion that gars are the direct descendants of semionotifomis (McAllister, 1968). This assumption is rejected on the basis of other evidence ( Patterson, 1973 ) . The palate of gars consists of elon- gate ectopterygoids overlain by toothed dermopterygoids (demiopalatines), en- dopterygoids, and metapterygoids (Fig. 5). Gars lack an autopalatine, an endo- chondral bone found in Aniia, teleosts (Patterson, 1975), "Gogo palaeoniscoids" (Gardiner, 1973), Pteronisculus (Nielson, 1942), semionotids (Gardiner, 1960; Wenz, 1968), and acanthodians (Miles, 1973). Gars differ from other actino- pterygians in that the metapterygoid is connected to the hyomandibular and preopercular only by cartilage and con- nective tissue. The suspensorium of gars is unique in several respects ( Fig. 6 ) . The quad- rate is supported entirely by the ecto- pterygoid medially and the quadratojugal posteriorly. The quadrate is situated in 22 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Hyo Sop Sym Pop Fig. 6. — Medial view of the post-orbital region of Atractosteus tropicus (AMNH 27939). Sus- pensorial ossifications unstippled, other bones stippled. Co, circumorbital; Dsp, dermosphe- notic; Hyo, hyomandibular; La, lacrimal; Op, opercular; Pop, preopercular; Qj, quadrato- jugal; So, suborbitals; Sop, subopercular; Sym, symplectic. front of the orbit rather than behind. The symplectic is not in close association with the quadrate but is found far pos- teriorly on the preopercular. Patterson ( 1973 ) reviewed the sus- pensorium of neopterygians and con- cluded that ( 1 ) the placement of the quadrate in front of the orbit is an auta- pomorphy of gars, (2) the loss of the quadratojugal and the double articula- tion of the lower jaw via both the sym- plectic and the quadrate was an apo- morphy of amiids (including Caftirus and Furo), and (3) the fusion of the quadrate and the quadratojugal is an apomorphy of teleosts. The separate quadratojugal of gars is a primitive char- acter shared with Lepidotes (figured by Patterson, 1973). In many specimens of gars the quadratojugal makes up an ex- ternal component of the skull and with the ectopteiygoid provides the only ossi- fied support of the quadrate. The sym- plectic of gars is "L" shaped and articu- lates only with the quadratojugal and the preopercular. The shape and topo- graphic position of the symplectic of gars is unique among actinopterygians. In regard to the teleostean condition of the quadrate and quadratojugal, Allis (1909) and Holmgren and Stensio (1936), with Patterson (1973), identify the splint-like process of the teleost quadrate as the (quadratojugal. Fred Cochocki and I observed this process in the leptocephalus larva of Elops saurus (UMMZ 165213, 25 mm) before it fused to the quadrate. The splint-like bone was not associated with any cartilage and occupied the exact topographic posi- tion of the quadratojugal of L. ossetis of the same length. In later stages of Elops saurus the process fuses with the quad- rate. These observations corroborate the hypotheses of Patterson and other work- ers. The quadrate of teleosts, then, is a compound bone composed of one endo- chondral (quadrate) and one deraial (quadratojugal ) element. The hyomandibular of gars articu- lates with the auditoiy capsule of the neurocranium above the foramen for the lateral head vein and the ramus hyo- mandibularis VII. Like other actino- pterygians there is an articulation with the opercular, and the hyomandibular fits into a groove of the demiopterotic. The position of the hyomandibular in relation to the lateral head vein and ramus hyomandibularis VH of actino- pterygians agrees with that of Pohjpterus (Goodrich, 1930), acanthodians (Miles, 1968), and Paleozoic sharks (Schaefl^er, 1967b). In dipnoans and Recent sharks the hyomandibular articulates below the foramen of the lateral head vein and ramus hyomandibularis VII, and this lower articulation was considered by Goodrich (1930) to be the plesiomorphous gnathostome condition. Schaeffer (1967b) and Gardiner (1973) disagreed, and stated that the plesiomorphous condition is a high articulation. If this is true, then the condition in Recent sharks may be a synapomorphy for that group. The ramus hyomandibularis VII of actinopterygians passes medially to the hyomandibular and either penetrates the bone (palaeoniscids, neopterygians) or continues to pass laterally (Acipenser). In either case it does not branch into the mandibular and hyoid branches until it has either passed or penetrated the hyo- mandibular. In dipnoans (insofar as a THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 23 msc Fig. 7. — Lower jaw of Atmctosieus spatula (AMNH uncat. ). Upper, lateral view; lower, medial view, a, angular; ar, articular; c, coronoid; d, dentary; Iwp, lateral wing of preartic- ular; mec, Meckel's cartilage; msc, mandibular sensory canal; p, prearticular; r, retroarticular; s, surangular. hyomandibular can be identified) and Recent sharks, the ramus hyomandibu- laris VII passes the hyomandibular later- ally and then branches. In rhipidistian sarcopterygians (Eusflienopteron, Mega- lichthyes, and other osteolepifomis and porolepiforms where the condition is known; Jarvik, 1954) and in coelacanths {Nesides, Jarvik, 1954; Latimeria, Millot and Anthony, 1958, 1965) the orientation of the ramus hyomandibularis VII to the hyomandibular is similar to actinoptery- gians, but the mandibular and hyoid branches apparently fork before pene- trating the hyomandibular. The signif- icance of the similarity between dip- noans and Recent sharks is obscure. Polypterus has the branching of the ra- mus hyomandibularis VII before the hyomandibular, like rhipidistians, but only the hyoid branch penetrates the hyomandibular whereas the mandibular branch curves in front of the hyoman- dibular (Goodrich, 1930, Fig. 446). The structure of the lower jaw of actinopterygians has recently been stud- ied by Nelson (1973b). He considered the presence of discrete articular, retro- articular and mentomeckelian endochon- dral ossifications and discrete prearticular and surangular dermal ossifications un- fused with endochondral elements or themselves to be plesiomorphous char- acteristics of actinopterygians. The ossi- fications of the lower jaw of gars is shown in Fig. 7. Nelson ( 1973b ) stated that a mentomeckelian is present as a separate ossification in Amia and is pres- ent and fused to the dentary in teleosts. Citing Starks (1916), Nelson mentioned the bone in large sturgeons. It is also present in Pteronisculus (Nielson, 1942), Latimeria, and Polypterus (Nelson 1973b). The absence of a mentomeckel- ian may be a synapomorphy of gars. Chondrosteans, gars, and Amia have unfused, discrete articular and retro- articular ossifications. In chondrosteans and Amia the surangular and retroartic- ular are separated by an unrestricted mass of Meckel's cartilage, whereas in gars these bones are in contact, Teleosts 24 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY lack a siirangular. Gars differ from other actinopterygians in having a lateral wing (Iwp, Fig. 7) on the prearticular which restricts Meckel's cartilage and articu- lates with the surangular. The ventral surface of the lateral wing of the pre- articular fonus the dorsal roof of the cavity that houses the insertion of the adductor mandibular muscle. The pre- articular and surangular meet above to produce a coronoid process. The den- tary is long, reflecting the general length- ening of the skull. In dried preparations, the Meckelian groove runs one-third to one-half of the length of the dentary. Coronoids cap the dentary medially and support an outer row of small teeth and an inner row of large teeth. The man- dibular sensory canal penetrates the an- gular and runs the length of the dentary (msc. Fig. 7; Allis, 1905). The development and structure of the gar neurocranium has been studied by Veit (1907, 1911, 1927), Mayhew (1924), Hammarberg (1937), and De Beer (1937). Rayner (1948) compared the neurocranial ossifications of gars with those of semionotids and pointed out the similarities between the neuro- cranium of gars and Lepidotes. Her conclusions have formed the major basis for considering gars to be semionotids. An extensive study of the neurocranium of actinopterygians including gars by Patterson (1975) summarized earlier pertinent data and added new infomia- tion on the structure of the neurocranium of phylogenetically important taxa. The ossifications of the neurocranium of Lep- isosteus oculatus are shown in Fig. 8. Patterson (1975:566) concluded that a hypothesis of loss of neurocranial ossi- fications is preferable to one of fragmen- tation or gain of neurocranial ossification centers. Further, he concluded that the palaeoniscoid Perleichis displays the primitive actinopterygian pattern. Table 1 is a summary of the ossification centers present in major groups of actinoptery- gians. The primitive pattern of Perlei- dus is most closely approximated by those of parasemionotids and philodo- phorids. Following this hypothesis gars and Lepidotes are more apomorphic than halecomorphs and teleosts. Both have lost the endochondral intercalar, the opisthotic, and (following Patterson, 1975) the "epioccipitals." The gars dif- fer from Lepidotes in that gars have also lost the basisphenoid. The endochondral intercalar has been lost within both the Halecomorphii (aniiids) and the Tele- ostei (pholidophorids, leptolepids and more derived groups) but the denual intercalar is found in all halecostomes (Patterson, 1975). The opisthotic has been lost within the Teleostei and Ami- idae. Further, the amiids have lost the pterotic (following Patterson, 1975). The loss of both the opisthotic and en- dochondral intercalar is, therefore, not unique to gars and Lepidotes within the Actinopterygii; both have been lost at least twice within other groups. To consider these losses as synapo- morphies uniting gars and Lepidotes into a single group, is, at the least, a weak hypothesis. Such a hypothesis is refuted by several other characters uniting Lepidotes to halecomoiphs and teleosts. To consider these losses as plesiomorphous of Neopteiygii would be unparsimonious in view of the similar- ities between parasemionotids, pholido- phorids, and Perleidus. The adoption of this hypothesis would require the inde- pendent acfjuisition of several ossifica- tion centers in halecostomes and pale- oniscoids, or the loss of ossification cen- ters in the ancestor of neopterygians followed by a re-acquisition in two line- ages and a loss within each lineage in- dependently. While such hypotheses are possil)le, it is more parsimonious to con- sider the losses by gars and Lepidotes to be independent losses and therefore nonhomologous. Gars have no anterior myodomes. The anterior myodomes are found in a number of chondrosteans, parasemiono- tids, Dci))ediiini, caturids, and amiids (Patterson, 1975:515-516). Patterson THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 25 Exo Pto Spo Psp Fig. 8. — Neurocranial ossifications of Lepisosteus ociilatus ( LACM 33914-2). Top, posterior view; middle, ventral \ie\v; bottom, lateral view. Boc, basioccipital; Exo, exoccipital; Ors, or- bitosphenoid; Pro, prootic; Psp, parasphenoid; Pto, pterotic; Pts, pterosphenoid; Spo, sphenotic. considered the gar condition secondary, and thus a synapomorphous character. Gars share with Polypterus and most primitive palaeoniscoids a transverse ca- nal in the otic region, a characteristic Patterson (1973:254) thought precluded gars having secondarily lost a posterior myodome. Thus, Patterson reasoned that the large posterior myodome of Amia and teleosts is a synapomorphy of ha- lecostomes. Patterson (1973:253-254) stated that the basipterygoid process of gars is in- termediate between the condition in chondrosteans, where the process is en- tirely endochondral (i.e., via the prootic), and the derived condition in teleosts, where it is entirely dermal ( lateral wings of the parasphenoid). The Hyoid and Visceral Arches Hyoid arch. — The hyoid arch of gars consists of paired endodermal hypohyals, anterior and posterior ceratohyals, inter- hyals, and dermal paired primary basi- 26 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Table 1. — Neurocranial Ossifications in seven groups of actinopterygian fishes. ^ Bone T axon -^ tn •l-l ■3 a 0 en es cc o J2 n-) 1 W5 ^ 1 •i-H s 0) in •S 0 '0 sa to CS ♦* rS a a: M •~J n <. n Basioccipital - X X X X X X X Exoccipital - X X X X X X X Epioccipital .. X X X X X Supraoccipital .. ? X X Demi-intercalar „ X X X X Endo-intercalar — _ .. X X X Opisthotic - X X X Pterotic X X X X X X Prootic _. X X X X X X X Basisphenoid .. X X X X X X Sphenotic .. X X X X X X X Orbitosphenoid — - - X X X '? X X X 1 Patterson (1973). hyal toothplates and branchiostegals (Fig. 9). Gars, like amiids and chondio- steans, lack the basihyal of teleosts (Wijhe, 1882; Veit, 1911; Hammarberg, 1937; Nelson, 1969a). A basihyal is hy- pothesized here to be a synapomoqihy of teleosts, and may be derived from anterior prolongation of the first basi- branchial copula (Nelson, 1969a). The hypohyals of gars are rectangu- lar and connected by connective tissue to the basihyal toothplates of the spatu- late "tongue." The anterior ceratohyal is long and uncompressed, whereas that of halecostomes is laterally compressed. The posterior ceratohyal is bent upward at a right angle to the vertical plane of the arch, articulating with the anterior ceratohyal, the interhyal, and the pre- opercular. The interhyal is usually un- ossified and connects the hyoid arch with the skull at the junction of the sym- plectic and hyomandibular cartilages. Three branchiostegals are present ( Mc- Allister, 1968). Two branchiostegals ar- ticulate with the posterior ceratohyal and one articulates with the anterior cera- tohyal, or occasionally at the junction of the two ceratohyals. That the most pos- terior branchiostegal is the homologue of the interopercular of halecostomes (De Beer, 1937) has been rejected by Mc- AlHster (1968). What, if any, phyloge- netic significance can be attached to the small number of branchiostegals in gars is not clear because many groups have apparently reduced the number of branchiostegals independently (see Mc- Allister, 1968, for discussion). Gars are the only known actinoptery- gian group having a long series of paired primary basihyal toothplates associated with a long spatulate "tongue" (Nelson, 1969a). These toothplates make up the abd hyo ro hyo st hyo Fig. 9. — Diagrammatic lateral view of the hyoid apparatus of a gar. ACH, anterior ceratohyal; abd hyo, abductor hyoides; b, branchiostegal; HH, hypohyal; PCH, posterior ceratohyal; pro hyo, protractor hyoides; st hyo, sternohyoides; tend, tendon. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 27 pn^QDCOffl Tongue Fig. 10. — "Tongue" of Lepisosteits osseus (AMNH 509). Dorsal view. ACH, anterior ceratohyal; BHTP, basihyal tooth plates; HH, hypohyal. dorsal surface of the tongue, articulate with each other and are supported by the dense connective tissue underlying them (Fig. 10). Whereas the presence of basihyal toothplates is probably plesi- omorphous for actinopterygians, the number and arrangement of the basihyal toothplates in gars is hypothesized as synapomoi-phous. The hyoid arch is connected to the cleithiaim by the sternohyoids, which originate on the medial wing of the cleithrum and insert on the hypohyals (st hyo. Fig. 9) . The protractor hyoideus originates on the posterior half of the anterior ceratohyal and inserts along the lower edge of the posterior half of the lower jaw (pro hyo. Fig. 9). Gars have a ligament, not reported in other fishes, which connects the posterior ceratohyal with the retroarticular of the lower jaw (tend, Fig. 9). This ligament does not seem to be the functional analogue of the tendon of halecostomes that con- nects the interopercular with the lower jaw and which has a respiratory function (Schaeffer and Rosen, 1961). In gars the ligament may provide the functional equivalent of the halecostome protractor hyoideus. In halecostomes this muscle is well developed and inserts on the lower jaw symphysis, whereas in gars it is weak, not inserted at the symphysis, and is probably not capable of coordinating the movement of the lower jaw and hy- oid arch. The ligament of gars, however, seems capable of depressing the lower jaw upon contraction of the sternohyo- ideus. It could also function in keep- ing the lower jaws in line and working in concert with the hyoid arch. Thus, I hypothesize that the ceratohyal-retro- articular ligament of gars has a primary feeding function in contrast to the in- teropercular-retroarticular tendon of ha- lecostomes which has a primarily respira- tory function. The hyoid arch is connected to the opercular series via the abductor hy- oideus. Tliis muscle consists of a series of muscle sheets running between the branchiostegals and inserting on the opercular bones. Finally, the hyoid arch is connected with the anterior end of the basibranchial copulae by articulation via the hypobranchials and via a tendon running from the anterior ceratohyal to the first hypobrancliial. Visceral arches. — Various morpholog- ical structiues of the visceral arches of gars have been cited as evidence for a monophyletic Holostei (Nelson, 1969a) and against a monophyletic Holostei (Patterson, 1973). Gars, like other ac- tinopterygians, have five paired arches united ventrally by a series of basi- branchial copulae. The basibranchial copulae of gars have been described and compared with those of other fishes by Nelson (1969a). Gars are similar to halecostomes in hav- ing four copulae (Fig. 11a, b). The "Gogo palaeoniscoids" have a single basibranchial, a condition Gardiner (1973) considered plesiomorphic for the Actinopterygii. Gardiner (1973) stated that this plesiomorphic condition is re- tained in Polijptenis (Fig. lie). In Aci- penser there are three ossifications in a single basibranchial copula (Wijhe, 1882; Gardiner, 1973), whereas in Pohjodon (Fig. lid) there are three copulae (Nel- son, 1969a), a condition similar in the palaeoniscoids Birgeria (Stensio, 1921) and Pteronisculus (Nielsen, 1942). Nel- son (1969a) and Gardiner (1973) re- viewed the number of copula in other teleostomes. Four copulae, then, seem to be synapomorphous for the Neo- pterygii. 28 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY BB1 BB2 BB3 a.CB5 BB4 a.HBI a.HB2 aHB3 a.HB4 C. Fig. II. — Lateral view of the basibranchial copulae of four actinopterygians, after Nelson, 1969a. a. Lepisosteus platostomiis; b. Amia calva; c. Pohjodon spatitJa; d. Pohjpterus sp. Cartilage stippled, hypobranchial articulatory surfaces black, bone unstippled. a. CBS. artic- ulatory surface of ceratobranchial 5; a. HB 1-4, articulatory surfaces of hypobranchials 1-4; BBl-4, basibranchials 1-4. The position of the articulation of the hypobranchials to the basibranchial copulae also varies between actinoptery- gian groups. In gars (Fig. 11a) hypo- branchial 1 articulates at the junction of the basibranchials 1 and 2, hypobran- chial 2 articulates at the junction of basi- branchials 2 and 3, hypobranchials 3 and 4 and ceratobranchial 5 articulate with basibranchial 3, and basibranchial 4 sup- ports no paired arch elements (Nelson, 1969a). In Pteroniscultis, basibranchial 1 supports no paired arch elements and was termed the basihyal by Nielsen (1942). Pohjodon and Amia (Figs, lid, b) are similar to each other in having hypobranchials 1 and 2 articulating with basibranchial 1. In Pohjodon, hypo- branchials 1 to 3 all articulate with basi- branchial 1 and if this basibranchial is homologous to basibranchials 1 and 2 in Ptcroniscidus, then they have a similarity not present in neopterygians. Amia, teleosts, and chondrosteans (where the condition is known) are sim- ilar in having the articular ends of hypo- branchial 4 penetrated by the ventral aorta. Gars have hypobranchials that are not penetrated by the ventral aorta, a condition hypothesized here as apo- morphous. Amia and gars have a perichondral ossification on the copula between arches 2 and 3 (Fig. 11a, b). Gars occa- sionally have two other ossifications (Patterson, 1973). The significance of this similarity is lessened by the observa- tions that ( 1 ) the ossifications are not found on the same copula, being found on basibranchial 2 in Amia and basi- branchial 3 in gars, and ( 2 ) perichondral ossifications of basibranchials are found in palaeoniscids (Stensio, 1921; Nielsen, 1942; Gardiner, 1973) and may be prim- itive for actinopterygians or derived several times independently. These ob- servations weaken Nelson's (1969a) hy- pothesis that this ossification is a syn- apomorphy uniting Amia and gars. Gars are similar to chondrosteans and Amia in having a separate fourth hypobranchial, which teleosts lack ( Nel- son, 1969a). The condition in teleosts is hypothesized to be apomorphous rela- tive to those of other actinopterygians. In gars the hypobranchials and cera- tobranchials are simple rod-like stiiic- tures. There are four epibranchials. The first two epibranchials have uncinate processes (UP, Fig. 12b). Chondrosteans lack uncinate processes (Fig. 12a), whereas Amia and teleosts have uncinate processes on the third epibranchial, as well as on the first two epibranchials (Fig. 12c). In Atnia the third uncinate process is not prominent and is carti- laginous, whereas in teleosts it is well developed and ossified. I hypothesize that the lack of uncinate processes, the condition of chondrosteans, is plesio- morphous for actinopterygians. Gars, Amia, and teleosts share the synapo- morphy of having uncinate processes. Gars have the relatively plesiomoqohous condition of two uncinate processes rela- tive to the three possessed by halec- ostomes. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 29 IB 1-3 C. Fig. 12. — Dorsal view of the upper gill arch elements of: (A) "Gogo palaeoniscoid," after Gardiner, 1973, (B) Atractosteus tropicus; and (C) Amia caJva. Cartilage stippled, bone unstippled. Suprapharyngobrancliials omitted in B and C. EBl-3 or 4, epibranchials 1-3, 4; IBl-3, infrapharyngobranchials 1-3; SBl-2, suprapharyngobranchials 1-2; UP, uncinate process. Gars, Amia, and chondrosteans lack a fourth infrapharyngobranchial (Fig. 12), whereas teleosts have this structure. Presence of a fourth infrapharyngo- branchial was considered plesiomorphous by Nelson ( 1969a ) but apomorphous by Patterson (1973). Nelson's (1969a) con- jecture was based on the observation that elasmobranchs and acanthodians have a fourth infrapharyngobranchial. Patterson's (1973) conjecture was based on the observation that all actinoptery- gians except teleosts lack a fourth infra- pharyngobranchial. Acceptance of Nel- son's ( 1969a ) hypothesis would require acceptance of a monophyletic group composed of chondrosteans, gars, and Amia. This hypothesis is incongruent with hypotheses based on other charac- ters. I accept Patterson's ( 1973 ) hy- pothesis and suggest that the fourth infrapharyngobranchial of teleosts origi- nated as a subdivision of the third infra- pharyngobranchial. Thus, the fourth infrapharyngobranchial of teleosts is hy- pothesized to be not homologous with the fourth infrapharyngobranchial of elasmobranchs and acanthodians. Chondrosteans have infrapharyngo- branchials that are posteriorly supported by the epibranchials (Fig. 12a; Gardiner, 1973; Nielsen, 1942). In gars, Amia, and teleosts, the infrapharyngobranchials are laterally supported by the epibranchials (Fig. 12b, c). This change in orientation for support of the infrapharyngobran- chials is hypothesized an apomorphy for neopterygians. Neither gars nor chondrosteans have a fifth epibranchial. Alhs (1897) figured a fifth epibranchial in Amia but Bertmar (1959) failed to find this stioicture. I have been unable to find a fifth epi- branchial on 37-40 mm cleared and stained specimens of Amia. A "fifth epi- branchial" is found in primitive teleosts as a small ball of cartilage, and Nelson (1969a) interpreted this character as a retained plesiomorphy. Tliis hypothesis is weakened by two observations — it is absent in other actinopterygian groups, and additional balls of cartilage are found in higher teleosts (Rosen, pers. comm.). It is possible that the "fifth epibranchial" is simply another of these balls of cartilage and thus an apomorphy found in certain teleosts. Dermal Components of the Arches. — The demial ossifications associated with the endoskeletal visceral skeleton may be divided conveniently into dermal toothplates and gill rakers, these rakers being modified toothplates. Nelson (1969a, 1970b) hypothesized that the 30 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY primitive condition for gnathostomes consists of a well developed dermal arch skeleton composed of separate tooth- plates covering the buccal cavity from the jaw margin to the phaiyngo-esopha- geal border. Apomoq^hous conditions usually involve reductions of demial ele- ments and have occurred in both elasmo- branchs (Nelson, 1970b) and osteichthy- ans (Nelson, 1969a). Nelson (1969a, 1970b) hypothesized the plesiomorphous condition for osteichthyans to consist of six rows of dermal elements: (1) lateral plates, (2) lateral gill rakers, (3) two rows of medial toothplates, (4) medial gill rakers, and (5) inner plates. Lati- jneria has retained the primitive osteich- thyan features (Nelson, 1969a). Among chondrosteans, Pteronisculus has several rows of toothplates associated with each arch (Nielsen, 1942). Living chondro- steans have reduced their demial plates to a single row of medial toothplates on the first hypobranchials (Nelson, 1969a). Gars have reduced their medial tooth- plates to a greater degree than Amia or teleosts. Gars have a single row of medial toothplates on hypobranchials 1, 2, and 3, ceratobranchials 2, 3, and 4, and on epibranchial 1. Gars have lost medial toothplates on ceratobranchial 1. Two rows of medial toothplates (the plesiomorphous condition) are found on hypobranchial 4 and infraphaiyngo- branchial 2, while several rows of tooth- plates are found on ceratobranchial 5 and infrapharyngobranchial 3. Amia and primitive teleosts retain the plesiomor- phous number of rows of medial tooth- plates on each of the lower arches. Amia differs from gars and teleosts in lacking the inner and lateral plates. Nelson ( 1969a ) concluded that neo- pterygians are the only osteichthyan fishes with significantly developed upper pharyngeal dentition. He hypothesized that the organization of the toothplates on the third infrapharyngobranchial and the lack of a fourth infraf)haryngo- branchial were synapomorphies uniting Amia and gars. This hypothesis rests on two conjectures: (1) that the fourth in- frapharyngobrancliial is plesiomorphous, a conjecture refuted above, and (2) that the several rows of teeth on the third infrapharyngobranchial of holosteans are derived from two or more arches (Nel- son, 1969a, suggested that deimal ele- ments of arches 3, 4, and possibly 5 are involved). Conjecture two seems to rest on the assumption that the toothplates on the third infrapharyngobranchial are homologous with only the medial tooth- plates of the lower arch elements. But there are only five or six rows of tooth- plates on infrapharyngobranchial 3 of gars, and it is possible that these are homologous with the entire six rows of dermal elements present primitively on the lower arches. I offer tliis conjecture as an alternative hypothesis and suggest that the two upper patches of teleosts arose from the simple subdivision of the primitively single third infrapharyngo- brancliial. Visceral arch muscles. — The visceral arch musculature of actinopterygians has been used to corroborate a hypoth- esis of a monophyletic Holostei ( Nelson, 1969a). Other groups in which this char- acter complex has been used successfully to demonstrate relationships include the eels (Nelson, 1966), the neoteleosts (Rosen, 1973), and the tetradontifomis (Winterbottom, 1972). Dorsal arch musculature. — The prim- itive dorsal arch musculature of actino- pterygians is hypothesized here to be similar to that seen in living Polyodon and Acipenser. In these genera there is a series of levator muscles running from the epibranchials to the skull roof and two medial muscle layers; a longitudinal layer lining the buccal cavity overlain by a circular layer. Both layers are un- differentiated from the same muscle lay- ers lining the esophagus. Gars, Amia and teleosts differ from chondrosteans in having discrete muscles derived from the circular muscle layer. These muscles in- clude the transverse dorsalis and oblique dorsalis (Fig. 13). I conclude that the THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 31 Fig. 13. — Dorsal view of the upper gill archs of (A) Amia calva; (B) Lepisosteus oculatus; and (C) Albula viilpes. Semidiagrammatic; bone stippled, muscles lined, dorsal aorta un- stippled. EB 3, 4, epibranchial 3, 4; LABE 3, 4, levator arcus branchialis exterior 3 and 4; LABI a and p, levator arcus branchialis interior, anterior and posterior; OD, oblique dorsalis; RA, retractor arcus branchialis; TD, a, p, transverse dorsalis, anterior, posterior. presence of such muscles is an apomor- phy of neopterygians. Retractor muscles are found in all three neopterygian groups. Retractors move the posterior infrapharyngobran- chial(s) and originate on the vertebrae. Teleost retractors are not found in all lower teleosts (Nelson, 1966, 1967; Rosen, 1973) and are derived from the inner longitudinal muscle layer ( Rosen, 1973 ) . The retractors of gars and Amia are de- rived from the outer circular muscular layer (pers. observ. ). Dietz (1912, 1914, 1921), Nelson (1966, 1967), and Rosen (1973) concluded that the retractors of teleosts are not homologous with the hol- ostean retractors. Rosen (1973) pointed out that teleost retractors have arisen a number of times within the Teleostei. Nelson (1969a) concluded that the retractors of amiids and gars are synapo- morphous. Patterson (1973) suggested that multiple independent development of retractors in teleosts weakens Nelson's 32 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY (1969a) hypothesis. In support of Nel- son's argument are the observations that the retractors of both gars and Ainia originate on the vertebral cohimn and insert on the third infrapharyngobran- chial, and that they have a similar em- bryological development, at least to the point that both are derived from the same muscle layer. Embryological stud- ies indicate that the retractors of both groups are derived from either the esophageal sphincter (Edgeworth, 1911, 1928) or from a posterior muscle mass independent of the esophageal sphincter (Edgeworth, 1935). In refutation of Nelson's ( 1969a ) hy- pothesis are the following observations. The retractors of gars (Fig. 13a) orig- inate separately, one on each side of the vertebral column on the third, fourth, and occasionally fifth vertebrae. They remain separate muscles and insert sepa- rately on the posterior edge of the third infrapharyngobranchial. They do not share muscle fibers with the transverse dorsalis but do share muscle fibers with the underlying circular muscle layer of the esophagus. Each transverse dorsalis of gars originates separately along the midline connective tissue and dorsal sur- face of the third infrapharyngobranchial and inserts in conjunction with the oblique dorsalis on the anterolateral car- tilaginous tip of the ossified lateral arm of the third infrapharyngobranchial. In Amia calva (Fig. 13b), the retractors originate as a single muscle on the verte- bral column, separating anteriorly into the two masses. They insert on the dor- sal surface of the third infrapharyngo- branchial and are confluent with the transverse dorsali. The transverse dor- sali do not insert in conjunction with the oblique dorsali; rather, the retractors in- sert with the oblique dorsali, and the transverse dorsali are found anterior to this insertion. Finally, the retractors of Amia do not share muscle fibers with the circular muscle layer of the esoph- agus. Lower teleosts, for example Al- bula ( Fig. 13c ) , have no retractors. But, the posterior transverse dorsali have the same origin and insertion and the same topographic relationship with the oblique dorsali of the fourth epibranchial as gars. And, the insertion of the posterior trans- verse dorsalis and the oblique dorsalis of the fourth epibranchial of Albula is the same as the insertion of the retractor and the oblique dorsalis of the fourth epibranchial of Amia calva. I interpret the retractor of Amia as a derivative of the transverse dorsalis and not as the homologue of the retrac- tor of gars. This interpretation would explain three observations : ( 1 ) the asso- ciation of the oblique dorsalis of the fourth epibranchial and retractor in Amia, (2) the sharing of fibers of the circular muscle layer and retractor of gars and the non-sharing of circular fi- bers and the retractors of Amia, and ( 3 ) the confluence of the transverse dorsali and retractors in Amia and the sepa- ration of these muscles in gars. I con- clude that the retractors of Amia and gars are independently derived and therefore not evidence for a relationship between the two groups. Ventral arch musculature. — The ven- tral gill arch muscles of osteichthyans have been studied by Edgeworth ( 1928, 1935) and Nelson (1967). Nelson (1967) hypothesized that oblique ventralis mus- cles are associated with each hypobran- chial and ceratobranchial as a primitive condition for osteichthyans and that loss of a hypobranchial is correlated with the presence of a transverse ventralis muscle on the arch. Chondrosteans lack trans- verse ventralis muscles but have a pair of oblique dorsali on each arch. Gars and Amia differ from chondrosteans and are similar to teleosts in having a trans- verse ventralis muscle on the fourth arch (Fig. 14a, b). Gars and Amia differ from teleosts in retaining a fourth hypobran- chial and a fourth oblique ventralis which shares fibers with the transverse ventralis. I hypothesize that the pres- ence of a transverse ventralis is a syn- apomorphy of neopterygians, that the THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 33 Fig. 14. — Ventral view of the posterior gill arches of Arnia calva (left) and Atractosteiis spatula (right). Seniidiagrammatic; bone and cartilage stippled, muscles lined, tendons white. BB 3-4, basibranchials 3-4; CB 3-5, ceratobranchials 3-5; HB3, hypobranchial 3; OV3-OV4, oblique ventralis 3 and 4; RC, rectus communis; t. OVp, tendon of oblique ventralis posterior. retention of a fourth pair of oblique ventralis muscles and the fourth hypo- branchials are plesiomorphies, and well developed fourth oblique ventrali are synapomorphies of teleosts. Aniia and teleosts differ from gars in having a pair of rectus communis mus- cles (Fig. 14b, c). These muscles orig- inate at the base of the third hypo- branchials and insert via a tendon on either the fourth ceratobrancliial (tele- osts) or the fifth ceratobrancliial (Amia). In both these groups of halecostomes the rectus communis is innervated by fourth arch nerve fibers and shares muscle fi- bers with the fourth transverse ventralis. Nelson (1967) suggested that the muscle was derived from the fourth transverse ventralis by forward growth. The pres- ence of a rectus communis is hypothe- sized an apomorphy of halecostomes. The ventral muscles of the fifth arch in actinopterygians are, like the dorsal muscles, derived from the outer circular muscle layer of the buccal and esopha- geal cavities. In the chondrosteans Pohjodon and Acipenser, there are two muscles, the transverse ventralis running between the fifth ceratobranchials and the coracobranchialis which originate on the midline connective tissue and insert on the coracoid of the pectoral girdle. In PoIyodo7i the coracobranchialis is composed of many separate muscle bun- dles and is not attached to the basi- branchial copula. This muscle in Pohj- odon is little difi^erentiated from the transverse ventralis, which runs between the ceratobranchials immediately poste- rior to the coracobranchialis, and it is undifferentiated from the circular mus- cle layer of the esophagus. In Acipenser the coracobranchialis is better defined, originating at the midline as a pair of discrete muscles connected by a tendon to the basibranchial copulae and insert- ing on the coracoid of the pectoral gir- dle. The transverse ventralis, like that of Pohjodon, originates on the midline connective tissue and inserts along the fifth ceratobranchials. In neopterygians the coracobranchialis does not originate on the midline connective tissue but on the lateral surfaces of the fifth cerato- 34 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY branchials. In Amia and gars the coraco- branchialis inserts on the cleithrum whereas in Flops and Sahno it inserts on both the cleithram and the coracoid (Jes- son, 1972). The transverse ventrahs of gars and Amia is essentially like that of chondrosteans except that it is better dif- ferentiated from the circular muscle layer of the esophagus (but it still shares fi- bers extensively). Teleosts are different. In Elops and AJbula, the transverse ven- trahs no longer inserts on the fifth cera- tobranchials but on the cleithrum via a tendon. The distribution of this charac- ter will have to be investigated in other teleosts before its significance can be assessed. POSTCRANIAL SKELETON Dermal components of the pectoral girdle. — The dermal pectoral girdle of gars includes the post-temporal, supra- cleithiami, and cleithrum (Fig. 15). The post-temporal articulates with the supra- temporals above and the pterotic medi- ally and ventrally. The articular surface of the gar post-temporal is convex, al- lowing for a ball-and-socket-like articu- lation of the supracleithrum via the con- cave supracleithral articular surface. The post-temporal of gars is similar to those of chondrosteans and Polijpterus in lacking a post-temporal process. Amia and teleosts have a post-temporal process that articulates with the dermo-inter- calar (see figures in Jessen, 1972; pers. observ. of Elops, Megalops, Alhula, Clu- pea, Osteoglossum, and Salmo). Patter- son (pers. comm.) reports the process in at least some semionotids. The presence of a post-temporal process in haleco- stomes is hypothesized apomorphous relative to its absence in gars and chon- drosteans The mode of articulation between the post-temporal and supra- cleithrum in gars is unique and hypothe- sized apomorphic for that group. The supracleithrum of gars has a supracleithral process that is connected to the basicranium via a ligament. Poly- pterus also has a supracleithral process Sclm P rad. m. w. CIm Fig. 15. — Medial view of the pectoral girdle and post-temporal of Lepisosteus osseus. Der- mal bones unstippled, endoskeletal mesocoro- coid stippled. A. P rad, articulatory surface of pectoral radials; Clm, cleithrum; Mc, meso- corocoid; m.w. Clm, medial wing of cleithrum; pro. Sclm, supracleithral process; Pt, post-tem- poral; Sclm, supracleithrum. (Daget, 1950; Jessen, 1972), but Jessen (1972) reported that the process is con- nected via a ligament to the epaxial body muscles, not the basicranium. Such a process has not been reported in chon- drosteans and there is little reason to conclude that the supracleithral proc- esses of Polijpterus and gars are homol- ogous. I conclude that the process is apomorphous and independently derived in both groups. The cleithiTim of gars differs from that of other actinopterygians in having a well-developed medial wing (M.W. CI, Fig. 15). This sti-ucture serves as the attacliment area for the pectoral ad- ductors, the ventral body musculature, and the sternohyoideus (Jessen, 1972). This structure is hypothesized apomor- phous for gars. The cleithrum is con- nected to the supracleithrum via con- nective tissue and articulates with its opposite medially along the ventral edges of the cleithral wing. The cleithriun is connected to the vertebral column via Baudelot's ligament and to the visceral THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 35 arches via the coracobranchiaHs and sternohyoideus. Gars, hke Amia and teleosts (and all known fossil neopterygians ) , lack a clav- icle. Attempts to homologize the small ossicles found on some gars (Jarvik, 1944) and the small flagellae of Amia with the clavicle of chondrosteans are not convincing and I hypothesize that the lack of a clavicle is a neopterygian synapomorphy. Endoskeletal pectoral girdle and Jes- sen's hypothesis of Actinopterygian rela- tionships.— The endoskeletal pectoral girdle of chondrosteans, gars, Amia, and teleosts has been studied in detail by Jessen ( 1972 ) . His inteq^retation of the shoulder girdle of actinopterygians led him to hypothesize that holosteans have a fundamentally different endoskeletal structure than teleosts and chondro- steans. Thus, Jessen (1972, 1973) con- cluded that chondrosteans are the sister group of teleosts and that Amia, gars, and Bergeria are the sister group of all other actinopterygians. Jessen's hypothesis rests on three characters hypothesized by him (Jessen, 1973) to be synapomoqDhies of gars and Amia. Two are shoulder girdle charac- ters and one concerns the course of the spinal nerves (and will be discussed here for completeness). 1 — The medial process of the scapu- lar region of chondrosteans and teleosts is not homologous with a similar process in Amia and gars that Jessen tenned the pontiform process. 2 — Gars and Amia lack a coracoid region that teleosts and chondro- steans have. 3 — ^The spinal nerves of gars and Amia penetrate the body muscu- lature ventral to the transversely oriented pleural ribs, whereas in chondrosteans, teleosts, and all other gnathostomes the spinal nerves follow the inner side of the body musculature mesial to the pleural ribs. Patterson (1973:259-260) outlined and discussed these points. He con- cluded that: character (1) is refuted by his suggestion that the differences in the upper endoskeletal girdle of gars and Amia could be accounted for by changes in the orientation and relative size of the parts, character (2) is refuted by demonstrating that gars have a cora- coid canal of the same type as other actinopteiygians and which marks the boundary between the scapular and cor- acoid regions of the endoskeletal girdle. Thus, the "anterior process of the middle region" (Jessen, 1972) is the homologue of the "anterior process of the coracoid region" (Patterson, 1973). Patterson (1973) concedes that the course of the spinal nerves is a significant similarity. I conclude from this that a hypothesis of synapomoq^hy concerning the course of the spinal nerves of Amia and gars cannot be refuted on morphological grounds, and therefore must be refuted on phylogenetic grounds, that is, by its incongruence with other hypothesized synapomorphies. Vertebrae. — The structure and de- velopment of actinopterygian vertebrae have been reviewed by Schaeffer (1967a) and his analysis will serve for the de- velopmental statements made below. The axial skeleton of all gnathostomes has, primitively, neural and haemal arches derived from sclerotomic mesen- chyme at the position of the myosepta. Intercalaries are present in sarcoptery- gians (except Neoceratodus) , chondro- steans, Amia, and perhaps gars (in a modified form ) , but intercalaries are ab- sent in teleosts and Polypterus. Whether centra are a primitive neopterygian feature or not is not clear. All three neopterygian groups have centra, but in at least one primitive haleco- morph centra are missing (Patterson, 1973 ) . The centra of amiids and teleosts are amphicoelous whereas those of gars are opisthocoelous (Fig. 16), a character unique among all actinopterygians ex- cept the blenny Andamia (Schaeffer, 36 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 16. — Vertebra of Atractosteus africanus (MHNP N.29, length of ventral ridge 13mm). 1967a). The amphicoelous vertebrae of teleosts differ in development from those of Amia. In teleosts the centrum forms as a double ring, an inner ring calcifying in the fibrous sheath of the notochord (the chordacentrum ) and an outer ring which ossifies later in the perichordal tube (the autocentrum) and later re- places the chordacentRuu to form the definitive adult centrum. In Amia a thin layer of bone forms in the perichordal tube and this is rapidly overlain by can- cellous bone derived from the sclero- tomic mesenchyme. Gars show a third pattern. Cartilaginous rings develop intrasegmentally in the perichordal tube. These rings increase in size and detach from the arches. Before ossification the rings constrict the notochord and each is split by a transverse canal which forms the opisthocoelous joint. This split is an intrasegmental rather than intersegmen- tal subdivision of the perichordal tube and the notochord. Because of the dif- ferences in development and adult struc- ture, I hypothesize that the opisthocoe- lous vertebrae of gars are apomorphous for that group. Caudal skeleton. — Gars have numer- ous hypurals, ural centra, and epurals. The neural arches of the ural centra are not modified into uroneurals and remain paired. The first ural centrum is usually fused to the last preural centrum and thus a compound centrum supports the parhypural and first hypural. Occasion- ally a second and even third hypural will be associated with the fused centnim. Patterson ( 1973 ) reviewed the cau- dal fin structure of neopterygians and concluded that: ( 1) the presence of uro- neurals in teleosts is apomorphous rela- tive to the undifferentiated neural spines of Amia and gars; (2) the presence of median neural spines in Amia is inter- mediate between teleosts (with median uroneurals =neural spines) and gars; (3) numerous ural centra, hypurals and epurals are a primitive feature of actino- pterygians; (4) the fused hypural-ural centra in Amia is apomorphous relative to the unfused condition in teleosts and gars; (5) the one-to-one correspondence between the middle hypurals and fin rays of gars and all of the fin rays and hypurals of Amia may have been de- rived independently; (6) epaxial fin rays developed independently in recent tele- osts and Amia; (7) the absence of ra- dials at the tips of the last few haemal spines of Amia is apomorphous relative to their presence in gars and teleosts, and (8) that two hypurals articulating with a first ural centrum is found in gars as an individual variation. I can find no reason to reject Patterson's ( 1973 ) first seven points, but the condition of the THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 37 fused first ural and last preural centra and their support of the parhypural and first hypural may be an apomorphy for gars and not a matter of individual variation. Summary Hypotheses of Actinoptery- GiAN Relationships The synapomorphies discussed above are organized below in five phylogenetic hypotheses (Figs. 17-19). In each case, the least rejected hypothesis is presented first and is followed by the alternative hypothesis which has corroborating in- stances based on characters that cannot be refuted as synapomorphies on mor- phological grounds. Alternate hypoth- eses for which no corroborating instances were found are not presented (for ex- ample, Jessen's, 1972, hypothesis). Char- acters discussed by Patterson ( 1973 ) as ahnost unique parallelisms among Amia and recent teleosts are not included be- cause Patterson's (1975) analysis of fossil halecostomes indicates that these char- acters are nonhomologous, thus elimi- nating them from considerations of re- lationship. Fig. 17a is the least rejected hypoth- esis of recent actinopterygian relation- ships based on the assumption followed throughout the analysis that chondro- steans are a monophyletic group. This is the same hypothesis accepted by Pat- terson (1973) and, eliminating fossil groups, by Westoll (1944), Gardiner (1960, 1963, 1967), and others. The monophyly of halecostomes is corrobo- rated by 13 synapomorphies (characters 1-13, Fig. 17a). The monophyly of neo- pterygians is corroborated by 7 synapo- morphies (characters 14-20, Fig. 17a). This hypothesis is not compatible with two characters, similarities of the endo- skeletal pectoral girdle and the course of the spinal nerves of A?nio and gars (Jessen 1972, and 1973, respectively). Both of these characters corroborate a hypothesis of monophyly of the Recent Holostei and are incorporated into the alternate hypothesis of relationship shown in Fig. 17b. This hypothesis was forwarded by Nelson (1969a). Nelson (1969a) based the hypothesis on gill arch characters, which, as discussed above, are open to alternate interpretations in- corporated in the first hypothesis (Fig. 17a). Fig. 18a and b summarizes alternate hypotheses of relationships of the fossil semionotid genus Lepidotes. Fig. 18a summarizes the corroborating observa- tions forwarded by Patterson (1973) that Lepidotes is a halecostome. Fig. 18b summarizes the corroborating observa- tions discussed by Patterson ( 1975 ) for Rayner's (1941, 1948) hypothesis that Lepidotes is more closely related to the Ginglymodi than to halecostomes. Lepi- dotes shares five synapomorphies (char- acters 1-5, Fig. 18a) with halecostomes and only two (characters 6, 7, Fig. 18b) with ginglymods. Both of these charac- ters are reduction characters that have arisen independently within both the Halecomorphi (amiids) and the Tele- ostei ( see Patterson, 1975 and discussion above ) . Fig. 19 summarizes the synapomor- phies uniting the ginglymod genera Lepisosteus and Atractosteus (characters 1-27). No alternative hypothesis has been forwarded which refutes this hy- pothesis. Summary Gars fomi a monophyletic group, the Ginglymodi. Gars are the sister group of the Halecostomi, a group composed of "semionotids" (Lepidotes, etc., a para- phyletic or polyphyletic assemblage; Patterson, 1973) the Halecomorphi (amiids and their fossil relatives the par- asemionotids, Patterson, 1973), and the Teleostei. Gars and Halecostomes form the monophyletic group Neopterygii and this group is the sister group of the Chondrostei. 38 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY G AT C G A Fig. 17. — Alternate hypotheses of the relationships of chondrosteans (C), gars (G), amiids (A), and teleosts (T). (a) Least refuted hypothesis; (b) alternate hypotliesis with corrob- orating characters. Synapomorphies (black rectangles) connecting taxa are: (1) a dermal intercalar; (2) a posterior myodome; (3) a supraniaxilla; (4) a post-temporal process; (5) loss of intemarial commissure; (6) an endochondral rostral; (7) maxilla with internal articu- latory head; (8) a single supratemporal on each side of the midline; (9) circumorbital ring incomplete; (10) an interopercular; (11) uncinate process on third infrapharyngobranchial; (12) a rectus communis muscle; (13) median neural spines in caudal region; (14) basiptery- goid process entirely composed of parasphenoid; ( 15 ) a postnarial commissure between the supra- and infraorbital canals; (16) no clavicle; (17) uncinate processes on first arid second infrapharyngobranchials; (18) infrapharyngobranchials laterally supported; (19) differentiated dorsal gill arch musculature (i.e. presence of OD, TD, etc.); (20) four basibranchial copulae; (21) course of spinal nerves; (22) general similarities in pectoral girdle anatomy. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 39 Fig. 18. — Alternate hypotheses of the phyloge- netic position of the semionotifomi genus Lepi- dotes (L) to gars (G) and halecostomes (H). (a) Least rejected hypothesis, (b) alternate hypothesis with corroborating characters. Syn- apomorphies (black rectangles) connecting taxa are: (1) medial neural spines in caudal region; (2) a posterior myodome; (3) a supra- maxilla; (4) an interopercular; (5) a maxilla with internal articulatory head; (6) loss of opisthotic; (7) loss of epioccipital. Systematic Accounts Division Ginglymodi Cope, 1872 Holostei Miiller, 1845:119 (in part). Ginglymodi Cope, 1872:328. Rhomboganoidei Jordan and Ever- mann, 1896:108. Aetheospondylii Goodrich, 1904:495; 1930:xvii (in part). Lepidosteoidei Goodrich, 1909:340. Orthoganoidei Zittle and Koken, 1911:105. Lepisosteifomies Hay, 1929:701 (in part). Lepidosteifomies Berg, 1940:211. Semionotidea Romer, 1966:353 (in part). Lepisosteida Matsubara, 1955:170. Lepisostei Suttkus, 1963:61. Diagnosis. — The ginglymods differ from all other actinopteiygian fishes in the synapomorphous characters shown in Fig. 19. The more obvious charac- ters for identification are: opisthocoelous vertebrae, plicidentine teeth, ethmoid elongation with snout bordered by toothed infraorbitals, premaxillary with nasal process carrying the supraorbital canal, an atrophied maxillary, quadrate in front of orbit, cleithrum with medial wing, retractor muscles in upper gill arches not associated with the transverse dorsalis. Description and remarks. — Body and head elongate. Body with interlocking ganoid scales covered with enameloid but lacking a dentine layer. Caudal fin semiheterocercal, without epaxial fin rays, with numerous haemal spines sup- porting fin rays, and numerous hypurals, epineurals, and ural centra; fulcral scales bordering the upper fin margin. Anal and dorsal fins far back on the body, caudal peduncle short. Anal and dorsal fins with fulcral scales on their anterior edges. Pelvic fins abdominal, internally supported by simple pectoral plates and without fulcral scales. Pectoral girdle consisting of dermal cleithrum, supraclei- thi-um, and postcleithrum, without clav- icle; endodeiTnal mesocoracoid support- ing radials and 11 to 14 fin rays. Verte- bral column consisting of a series of opisthocoelous vertebrae with paired neural spines. Pleural ribs articulating with epipleural ribs that reach the outer body wall. Post-temporal an integral part of skull roof, without post-temporal process. Two to five small, rectangular supratemporals on each side of the mid- hne. Parietals and dennopterotics equal in size and not conspicuously elongate. Frontals elongate. Premaxillary with a long process, carrying the supraorbital canal from the frontals to the nasals. Antorbitals and nasals crescent-shaped and small. Rostral U-shaped and carry- ing the infraorbital commissure. Snout bounded by three to ten toothed infra- orbitals. Maxillary small and attached to the posterior infraorbital via tendon. Two or three lacrimals. A complete cir- cumorbital series and a large mosaic of suborbitals. Opercular series consisting of an opercular, subopercular, and pre- opercular; no interopercular. Neuro- cranial ossifications including a medial basioccipital, paired exoccipitals, pro- otics, sphenotics, sphenoids, and a me- 40 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY H A. Fig. 19. — Phylogenetic analysis of the gar gen- era Lepisosteus (L) and Atractosteits (A) to the Halecostomi (H) and Chondrostei (C). Synapomoiphies are black rectangles whereas plesiomorphies are white rectangles. Snyapo- niorphous characters are: (1) no ethmoid os- sifications; (2) elongate ethmoid region; (3) a series of toothed infraorbitals bordering the snout; (4) supraorbital canal on premaxillary process; ( 5 ) premaxillary process present, forming an external dermal component of the snout; (6) atrophied maxilla; (7) dermopter- dial oibitosphenoid. Neurocraniiim lack- ing a dennal or endochondral intercalar, epiotics, or basisphenoid. Interorbital septum complete. Ethmoids cartilagi- nous. Parasphenoid long, flattened, grooved posteriorly for passage of the dorsal aorta, and extending from the posterior end of the basisphenoid to mid-snout. Basipterygoid process incor- porating portions of both the parasphe- noid and the prootics. Vomers overlying the parasphenoid anteriorly. Vomers elongate, paired, toothed, and lance- shaped. Quadrate in front of orbit, sup- ported medially by the ectopterygoid and posteriorly by the quadratojugal. Hyomandibular and L-shaped symplec- tic not supporting the quadrate. Endo- pterygoids and metapterygoids articu- lating on the ectopterygoid and not contacting the quadrate or hyomandib- ular. Elongate ectopterygoid with teeth, overlain by dennopalatines anteriorly. Elongate, toothed dentary overlain by coronoids medially and with conspicuous meckelian groove on medial side extend- ing 1/3 to 1/2 the length of the bone. Surangular and angular making up a coronoid process. Articulation of lower jaw via separate articular and retro- articular. Lateral posterior end of lower otic-sphenotic articulation; (8) epioccipital lost; ( 9 ) postorbital commissure of supraorbital and infraorbital sensory canals on the dermo- pterotic; (10) ectopterygoid elongate; (11) autopalatine missing; ( 12 ) endopterygoid and metapterygoid not supporting the quadrate; (13) quadrate position in front of orbit; (14) symplectic "L" shaped and not contacting quadrate; ( 15 ) prearticular with lateral wing that restricts Meckel's cartilage; (16) nienlo- meckelian missing; ( 17 ) opisthotic and basi- sphenoid lost; (18) no anterior myodome; (19) a series of paired primary basihyal toothplates supported by a spatulate tongue; (20) a retro- articular-posterior ceratohyal ligament; (21) hypobranchials not penetrated by the ventral aorta; (22) medial toothplates reduced on vis- ceral arches; (23) a retractor not associated with the transverse dorsalis; (24) a supra- cleithral process; (25) a medial wing on the cleithrum; (26) opisthocoelous vertebrae; (27) plicidentine teeth; (28) synapomgrphies unit- ing gars to other neopterygians: see figure 18a. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 41 jaw overlain by the deniial angular. Hy- oid arch without a basihyal and visceral arch without a fourth infraphaiyngo- branchial. A long series of paired pri- mary basihyal tooth-plates. A single highly vascularized swimbladder. Conus arteriosis with seven tiers of eight valves, no bulbous arteriosis. Gut with spiral valve remnant. Females with oviduct directlv connected to ovaries, males with staggered testes. Color pattern variable but primitively with dorsal flank and belly stripes. Lepisosteus sinensis Bleeker (1873: 154) from China is a belonid, not a gar as reported by Wagner (1912:734). ? Lepisosteus alessandrii Ameghino (189(S) from Argentina is listed by Pascual (1970) as an improbable member of the genus. Litholepis adamantinus Rafi- nesque (1818a:447) is a mythical fish drawn by J. J. Audubon (Suttkus, 1963: 69). Etijmologij. — From the Greek ginghj- nios, a hinge joint, referring to the mode of articulation of the opisthocoelous ver- tebrae. Family Lepisosteidae Cuvier, 1825 Lepisosteidae Cuvier, 1825:2, 307. Lepidostei Agassiz, 1832:140. Fitzinger, 1873: 52. Sauroides Agassiz, 1843:2. Lepidosteinii Miiller, 1844:208. Cams, 1875: 590. Diagnosis and descriptive remarks. — Those of the division Ginglymodi. Other remarks. — The following nom- inal species are Lepisosteidae, genus and species indeterminate: fNaisia apicalis Miinster, 1846:34 (Upper Eocene, Gennany). \Trichiuridea sagittidens Winkler, 1876:31 (Middle Eocene, Bel- gium ) . ]Fneumatosteus nahunticus Cope, 1869:242; 1875:31. Eastman, 1900a: 68 (Miocene, North Caro- lina ) . \Lepisosteus knieskerni Fowler, 1911: 150 ( PCretaceous, New Jersey). ^ Atractosteus emmonsi Hay, 1929: 709 (based on Emmons, 1858: 244; PMiocene, North Carolina). \Faralepidosteus praecurser Casier, 1961:42 (Early Cretaceous, Af- rica ) . The following references refer to fos- sil gars for which no specific determina- tion was attempted by the authors. Gid- ley, 1915:539 (Upper Cretaceous, Fort Union Fm., Montana); 1927:274 (Pleis- tocene, Florida). Gilmore, 1916:302 (Upper Cretaceous, New Mexico); 1920: 8, 68 ( Upper Cretaceous, Kirtland Fm., Wyoming). Hay, 1903:120; 1927:274 (Pleistocene, Florida). Reeside, 1924:21, 23, 31, 38, 42 ( Upper Cretaceous, Fruit- land, McDermott, Ojo Alama, and Naci- miento Ems., New Mexico). Russel, 1935:118 (Cretaceous, Milk River Beds, Alberta). Bjork, 1967:229 (Eocene, Slim Buttes Fm., South Dakota). Lepisosteus Lacepede Lepisosteus Lacepede, 1803:331 (type species L. gavialis by subsequent designation, Jor- dan and Evermann, 1896:109). Sarchinis Rafinesque, 1818a:418 (type species S. vittatus by subsequent designation, Jor- dan, 1877:9). CijUndrostcus Rafinesque, 1820:72 (type spe- cies C. platostomtis by subsequent desig- nation, Jordan, 1877:11). Lepidosteus (Lacepede): Koenig, 1825:12; Agassiz, 1843:2. Diagnosis. — Gars with an ectoptery- goid-prem axillary articulation on the premaxillary process, without enlarged dermopalatine fangs, with projecting ridges above and below the articular socket of the supracleithrum. The fron- tal bone elongate anteriorly and postero- laterally, extending past the dennopter- otic laterally to produce the character- istic shape of the frontal (Fig. 20). Lepisosteus also differs from Atractosteus in that Lepisosteus gars retain, primi- tively, small pear-shaped gill rakers (Fig. 21) and medial toothplates on the first infrapharyngobranchial (Fig. 22) and the first three hypobranchials and cera- tobranchials. 42 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 21. — Gill rakers in seven Recent lepiso- steids. (a.) L. platostomus; (b. ) L. osseus; (c. ) L. oculatus; (d. ) L. platyrhincus; (e.) A. tristoechus; (f. ) A. spatula; (g. ) A. tropicus. Fig. 20. — Frontal shape of eight Lepisosteus. (a.) L. opertus (MCZ P.13392); (b.) L. cuiie- a*«* (AMNH P.4622); (c.) L. platostomus (UMMZ 190846); (d.) L. indicus (BMNH P.12178); (e.) L. osseus (LACM 33917-4); (f.) L. fimbriatus (composite, BMNH P.1700 and P.13330); (g.) L. oculatus (LACM 33915-1); (h.) L. platyrhincus (LACM 33912-2). Etymology. — A compound masculine nominative derived from the Greek lepis (:=scale) and Latin osteus (^bone). The following nominal species and name combinations are Lepisosteus spe- cies indetemiinate: \Clastes cycliferus Cope, 1873:634; 1877b: 40; 1884:54. Woodward, 1895:445. Merrill, 1907:8 (Eocene, Wyoming). \Lepidosteus cycliferus (Cope): Eastman, 1900a:68. Cockrell, 1908:163. \ Lepisosteus cycliferus (Cope): Hay, 1902: 337; 1929:377. \Lepidosteus longus Lambe, 1908:13 (Oli- gocene, Saskatchewan, Canada). The following references refer to Pleistocene and Pliocene records of Lep- isosteus, sp. indet. : Hay ( 1927, Pleisto- cene of Florida); C. L. Smith (1954, Pleistocene of Oklahoma; 1958, Pleisto- cene of Oklahoma and Kansas; 1962, Lower Pliocene of Oklahoma); Uyeno and Miller (1962, Pleistocene of Texas; 1963, summary of North American rec- ords); Dalquest (1962, Pleistocene of Texas ) ; Uyeno ( 1963, Pleistocene of Texas); Hibbard and Dalquest (1966, Pleistocene of Texas); Lundberg (1967, Pleistocene of Texas); Swift (1968, Pleis- tocene of Texas); and Wilson (1968, Pliocene of Kansas). fLepisosteus opertus, new species Figures 20a, 23, 24a, 25a, 27a. Lepisosteus occidentalis: Estes, 1964:43; 1969: 11 (in part, bones of A. occidentalis mixed with bones of L. opertus). Diagnosis. — Lepisosteus opertus dif- fers from all other Lepisosteus in retain- ing the primitive enameloid pattern of the family on the deirnopterotics and parietals (Figs. 25a, 27a). Types.— The holotype (MCZ P.13392) is an incomplete frontal (Figs. 20a, 23a). Paratypes include two fragmehtaiy fron- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 43 IB1-3 EB1-4 '^- B. Fig. 22. — Ventral views of the upper gill arch elements of (A) Lepisosteus oculatus and (B) Atractosteus tropicus. Semidiagramniatic; cartilage stippled. EB 1-4, epibranchials 1-4; IB 1-3, infrapharvTigobranchials 1-3. tals (MCZ P. 13396), two dermopterotics (MCZ P. 13397, Fig. 23b), one preopercu- lar (MCZ P.13393, Fig. 23c), 17 infra- orbitals (MCZ P.13394-95), three pari- etals (MCZ P.9374, Fig. 27a). Type locality. — Bug Creek Anthills, section map locality SW 1/4-9-T22N- R43E (or approximately 25 mi. SSE of Ft. Peck), McCone County, Montana. Formation and age. — Hell Creek For- mation, Upper Cretaceous, probably in other formations such as the Lance, and Belly River series. Description and comparisons. — Lepi- sosteus opertus is a small Cretaceous gar known only from fragmentary remains. No counts or measurements were obtain- able. All premaxillaries from the Hell Creek formation have a double row of premaxillary teeth, but I cannot assign the smaller preserved premaxillaries to either A. occidentalis or L. opertus. I assume that L. opertus had two com- plete rows of premaxillary teeth like L. platostomus. The deiTnopalatines are unknown. Infraorbitals are narrow and elongate, with rounded enameloid tuber- cles. Circumorbitals have enameloid, but the condition of the dorsal circum- orbital and relationship of the dermo- sphenotic to the orbital margin is un- known. The number of supratemporals per side is unknown. Supracleithrum has a long bony process on the ventral side of the articular facet (Fig. 24a). Frontal shape is like that of other Lepisosteus (Fig. 20a), and it has rows of rounded enameloid tubercles along the bony ridges. Shape and enameloid pattern of the dennopterotic and parietal can be compared to other Lepisosteus shown in Figs. 25a and 27a. Lepisosteus opertus differs from Atractosteus gars in the shape of the posterior end of the frontal bone (Fig. 20a). Lepisosteus opertus differs from Atractosteus occidentalis in that L. oper- tus has enameloid on the frontals, infra- orbitals, circumorbitals, and preopercu- lar, whereas A. occidentalis lacks enam- eloid or has only occasional, minute, rounded, enameloid tubercles on the 44 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 23. — Various bones of Lepiaosteus opertua. (a.) Frontal ( MCZ P.I3392, 21mm, holotype); (b.) infraorbital (MCZ P.I3395, 13 mm, para- type); (c.) preopercular (MCZ P.13393, 23- mm, paratype ) . preopercular. In addition, L. opertus has sheets of enameloid on the parietals and dermosphenotics and elongate in- fraorbitals and a Lepisosteus snpra- cleithrum, whereas A. occidentalis has rounded enameloid tubercles on the dennosphenotic, square infraorbitals, and a supracleithrum more like that of A. spatula, A. tristoechus, and A. atrox. Etymology. — From the Latin, oper- tus (=hidden), referring to the obser- vation that it has remained undescribed among the material assigned to another species, A. occidentalis. f Lepisosteus cuneatus (Cope) Figs. 20b, 25b, 27b, 29b, 31, 32 Clastes cuneatus Cope, 1878:9; 1880:303; 1884:55. Lepisosteus cuneatus: Eastman, 1900a: 68; 1900b:57. Hay, 1902:377; 1929:708. Hussakof, 1908:78. Cockerell, 1909:796. Hussakof and Bryant, 1919:195. Stromer, 1925:. 360. Diagnosis. — Differs from all other Lepisosteus in that the width of the opercular and subopercular is greater than the distance from the margin of the opercular apparatus and the suborbitals anterior to the medio-posterior orbital margin, whereas in all other Lepisosteus the width of the opercular and suboper- cular is less than this distance. Type.—AMNH P.2517. A complete fish with crushed skull displaying a large opercular typical of the species. Type locality. — Manti Beds, Central Utah. No exact locality given. Formation and Age. — Green River Formation, Lower Eocene. Descriptive comments. — Lepisosteus cuneatus is a small, short-snouted gar from the Eocene of North America. De- tails of the posterior half of the skull of AMNH P.4622 in dorsal and lateral view are shown in Fig. 31. Meristic data are shown in Table 2. Presence or absence of second pre- maxillary tooth row not detemiined. Dennopalatine without demiopalatine fangs. Five to six infraorbitals. Number of circumorbitals not detennined, but deiTnosphenotic included in the orbital margin and there are three circum- orbitals forming the posterior orbital margin. Three lacrimals. Few suborbit- als (AMNH P.4623, Fig. 32). Opercu- lar and subopercular large compared to postorbital head length. Two supratem- porals on each side of the midline. Su- pracleithrum only partly observable, condition of articular facet not ob- servable. Frontal shape, as far as determinable, characteristic of the genus (Fig. 25b). Shape and enameloid pattern of denno- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 45 Fig. 24. — Supracleithra of six Lepisosteus species, (a.) L. opertus (AMNH P.9323); (b.) L. platostomus (AMNH 7147, 27mm); (c.) L. osseiis (LACM 33916-4, 16mm); (d.) L. fim- briatus (BMNH P.1300, 24mm); (e.) L. oculatus (LACM 33914-3, 20mm); (f.) L. platyrhincus (LACM 33912-4, 17mm). pterotic, parietal, and opercular shown in Figs. 25b, 23b, and 29b, respectively. All skull bones with large amounts of enameloid in the fonn of broad, elongate continuous tubercles or large, rounded tubercles. Differs from L. opertus in having a reduced number of convoluted and in- terconnected enameloid ridges on the dennopterotics and parietals, and in other details of enameloid pattern ( Figs. 25b, 27b). Differs from L. fimhriatus and all Recent Lepisosteus in having wide interconnecting enameloid ridges on the parietals and dermopterotics, and dorsal half of the opercular, and in hav- ing large, rounded enameloid tubercles on other parts of these bones, whereas in fimbriatus and Recent Lepisosteus the enameloid is thinner, and usually occurs as disconnected tubercles (reduced in many species to series of small, rounded tubercles). Differs from L. indicus in that cuneatus has enameloid and a short snout, whereas indicus lacks enameloid and has a long snout (Fig. 20). Etymology. — From the Latin cune- atus (m wedge-shaped), referring to the skull of the type. 46 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY rmM Fig. 25. — Dermopterotics of three Lepisosteus. ( a. ) L. opertus ( MCZ P. 13397, paratype, 27mm ) ; (b.) L. cuneatus (AMNH P.4622, 23mm); (c.) L. platostomus (UMMZ 190846R, 25mm). Lepisosteus platostomus Rafinesque Shortnose Gar Figs. 20c, 21a, 24b, 25c, 27c, 29b, 33, 34a Lepisosteus platostomus Rafinesque, 1820:72. Jordan and Evermann, 1896:110. Suttkus, 1963:71. Lepisosteus albus Rafinesque, 1820:73. Lepisosteus platystomus Giinther, 1870:329. Jordan and Gilbert, 1883:91. Jordan, 1885: 13. Cylindrosteus scabiceps Fowler, 1910:607. Jor- dan, Evermann and Clark, 1930:37. Cijlindrosteus platostomus: Jordan, Evermann, and Clark, 1930:37. Diagnosis. — Differs from other Lepi- sosteus except L. opertus in having two complete rows of prem axillary teeth (up to 8 in L. platostomus, 2-4 in L. osseus, 1-2 in L. oculatus, L. platijrliinchus, and L. fimhriatus) . Differs from L. opertus in having thin enameloid ridges usually disconnected into oblong tubercles, whereas L. opertus has wide, intercon- nected tubercles. Type. — No type material was col- lected by Rafinesque. Descriptiojis and comparisons. — A small gar without flank stripes or belly pigmentation in adults. A typical skull is shown in Fig. 33. Meristic counts are shown in Table 2 and various moipho- metric measurements expressed as ra- tios or dorsal head length are shown in Tables 3 and 4. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 47 Fig. 26. — Dermopterotics of four Lcpisosteus. (a.) L. osseiis (LACM 33917-4, 30mm); (b.) L. fimbnatus (BMNH P.1700, 27mm); (c.) L. ocidatus (LACM 33915-1, 32mm); (d.) L. platijrhincus (LACM 33912, 30mm). Head and dorsum darker than flank and belly. Flank stripe variable, usually indistinct or same Intensity as dorsum. Flank usually with a series of vertical pigment bars between some scale rows. Usually two distinct pigment blotches on caudal peduncle immediately in front of caudal fin. Flank stripe more distinct on cheek and may be broken into two stripes on the opercular as an individual variation, continuing through eye and ending on coronoid process of lower jaw. Juveniles with faint pre- opercular and retroarticular stripes, these ab- sent or diffuse in adults. Mid-dorsal stripe ab- sent or faint. Gular region with some pigment bordering medial margin of lower jaw. Belly without belly stripes or pigment blotches. Pec- toral and pelvic fins without transverse pigment bars. Anal and dorsal fins with two or three transverse rows of pigment blotches. Caudal fin with variable pigment blotches on fin rays and fin membranes. Premaxillary process with ridges for articu- lation of ectopterygoid and with two complete rows of premaxillary teeth. Dermopalatine with two tooth rows, the irmer row of adults some- what enlarged but not as large as infraorbital fangs. Seven to nine infraorbitals. Eight to ten circumorbitals, dorsal circumorbital not en- larged, dermosphenotic included in the orbital margin, three circumorbitals including the der- 48 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 27. — Parietals of three Lepisosteus. (a.) L. opertus (MCZ P.9374, 24nim); (b.) L. cuneatus (AMNH P.4622, 21mm); (c.) L. platostomus (UMMZ 190846, 21mm). Enameloid pattern shown in black on outline drawings. mosphenotic fomiing posterior orbital margin. Three lacrimals. Suborbitals not numerous, tire ventro-posterior marginal suborbitals distinctly larger than those of the internal mosaic. Two or four supratemporals on each side of midline. Supracleithrum with projections above and be- low articular facet, its shape shown in Fig. 24b. Enameloid on all dermal roofing bones of skull. These bones with low-lying bony ridges capped by more or less continuous elongate enameloid tubercles. Shape and enameloid pat- tern of dermopterotic, parietal, and opercular shown in Figs. 25c, 27c, and 29c, respectively. First basihyal toothplate paired. Gill rakers small, pearshaped (Fig. 21a), and not nimier- ous (Table 2). A single row of medial tooth- plates on the first arch. A single complete row of medial toothplates on second and tliird lower arch elements and two complete rows on the fourth arch (Fig. 34a), a single row of medial toothplates on first infrapharyngobranchial, two rows on the second infrapharyngobranchial. Diff^ers from L. osseiis and L. indiciis in that L. platostomus has a shorter snout. Differs from L. oculatus and L. plattjrhincus in num- ber of lateral line scales (59-65 in platostomus, 53-59 in oculatus and platijrhincus) and head color pattern (blotched in oculatus and platij- rhincus plain in platostomus). THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 49 Fig. 28. — Paiietals of four Lepisostetis. (a.) L. ossetw (LACM 33917-4, 38mm); (b.) L. fimhriatus (BMNH P.33525, 19mm); (c.) L. oculatus (LACM 33915-1, 36mm); (d.) L. platijrhincus (LACM 33912-2, 33mm). Enameloid patterns shown in black on outline drawings. Etymology. — From the Greek plator (r=broad or flat) and stomus (= mouth). Range. — Northeastern Texas north to Montana, east to southern Ohio and south to Mississippi (Schultz, 1965). Other comments. — Schultz (1965) and G. R. Smith (1964) report L. platostomus from the Pleistocene of Kansas. A report from Florida by Hay (1917) is consid- ered by Uyeno and Miller ( 1963 ) to be L. platyrhincus. Lepisosteus osseus (Linnaeus) LoNGNOSE Gar Figs. 10, 15, 20d, 21b, 24c, 26a, 28a, 29c, 34b, 35 Esox osseus Linnaeus, 1758:313. Esox viridis Gmelin, 1789:1389 (In: Linnaeus, 1789; not Lepidosteus viridis Giinther, which is A. spatula). Lepisosteus gavialis Lacepede, 1803:333. Sarchirus vittatus Rafinesque, 1818a:419. ?Lepisosteus stenorhynchus Rafinesque, 1818b: 50 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 29. — Operculars of three Lcpisosteus. (a.) L. cuneatus (AMNH P.4622, 20mm); (b.) L. platostomus (UMMZ 190846, 11mm); (c.) L. osseus (LACM 33917-4, 21mm). Enameloid patterns shown in black on outHne drawings. 447 (Hsted by Suttkus, 1963, as a doubtful synonym ) . Lepisosteus oxyiirus Rafinesque, 1820:73. Lepisosteus longirostris Rafinesque, 1820:74. ?SaTchirus argenteus Rafinesque, 1820:86 (listed by Suttkus 1963, as a doubtful synonym). Lepidosteus osseus: Agassiz, 1843:2. Giinther, 1870:330. Lepidosteus semiradiatus Agassiz, 1843:2. Lepidosteus gracilis Agassiz, 1843:2. Lepisosteus huronensis Richardson, 1836:237. Lepidosteus rostratus Cuvier, 1836:238 (In: Richardson, 1836). Lepidosteus bison De Kay, 1842:271. Lepisosteus lineatus Thompson, 1842:145. Macrognathus loricatus Gronow, 1854:148. Lepidosieus leptorhynchus Girard, 1858:351. Lepidosteus otarius Cope, 1865:86. Lepidosteus crassus Cope, 1865:86. Lepidosteus tieculi Dumeril, 1870:327. Lepidosteus ynidberti Dumeril, 1870:228. Lepidosteus harlani Dumeril, 1870:329. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 51 V. Fig. 30. — Operculars of three Lepisosteus. (a.) L. fimbriattis (BMNH P. 1529, 14mm); (b.) L. oculatus (LACM 33915-1, 24mm); (c.) L. platijrhincus (LACM 33912-2, 20mm). Enameloid patterns shown in black on outline drawings. Lepidosteus smithi Dumeril, 1870:330. Lepidosteus ayresii Dumeril, 1870:331. Lepidosteus copei Dumeril, 1870:332. Lepidosteus lesueurii Dumeril, 1870:335. Lepidosteus elisabeth Dumeril, 1870:336. Lepidosieus lamarii Dumeril, 1870:337. Lepidosteus clintonii Dumeril, 1870:338. Lepidosteus twostii Dumeril, 1870:339. Lepidosteus piquotiamus Dumeril, 1870:340. Lepidosteus horatii Dumeril, 1870:341. Lepidosteus thompsoni Dumeril, 1870:342. Lepidosteus louisianensis Dumeril, 1870:343. Lepisosteus osseus: Suttkus, 1963:79. Diagnosis. — Lepisosteus osseus dif- fers from all other Recent species of the genus in having an extremely attenu- ated snout (snout length 79% to 83% of dorsal head length in osseus, less than 75% in other species). Type. — A dried specimen with bro- 52 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 31. — Dorsal (upper) and lateral (lower) views of the posterior skull of L. cuneatus (AMNH P.4622, 38inm from back of skull to anterior end in dorsal view). ken snout on deposit in the Linnean Col- lection of the Linnean Society of Lon- don. Description and comparisons. — Lepi- sosteus osseus is a medium sized (to 1500 mm, Suttkus, 1963) gar with a nar- row attenuated snout. The skull of a typical individual is shown in Fig. 35. Various count data shown in Table 2. Moi*phometrics expressed as ratios of dorsal head length shown in Tables 3 and 4. Color pattern of adults varying with en- vironmental conditions (Suttkus, 1963:77). Juveniles with continuous mid-dorsal stripe missing in adults, where dorsum darkened by general melanophore development. Flank stripe of juveniles continuous from base of caudal fin through eye onto lower jaw. Flank stripe of juveniles serrate on upper margin. Adult flank stripe either missing or reduced to a series of THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND REGENT GARS 53 Sop Pop Fig. 32. — Medial view of the post-orbital region of Lepisosteus cuneatus (AMNH P.4623, 60mm). Co, circumorbital; Dsp, dermosphe- notic; Op, opercular; Pop, preopercular; So, suborbital, Sop, subopercular. pigment blotches. Preopercular stripe incom- plete. Head witli or without small random pig- ment blotches. Posterior half of lower jaw with scattered melanophores. Gular region with pig- ment lining lower jaw, such pigment extending from symphysis to branchiostegals. Isthmus with or without scattered pigment of variable intensity. One belly stripe per side in juveniles, extending from pectoral base to anal fin, where stripes fuse on the ventrum of caudal pe- duncle. Belly between stripes usually unpig- mented; belly stripes usually missing in adults. Pectoral fin base pigmented in juveniles, usually unpigmented in adults. Pectoral fins with 2-5 transverse rows of pigment blotches. Pelvic fins with two rows of transverse pigment blotches. Anal and dorsal fins usually with three trans- verse rows of pigment blotches. Caudal fin with numerous pigment blotches, pattern in- dividually variable. Urostyle of juveniles un- pigmented. Premaxillary with outer tooth row of two to four teeth. Dennopalatine of adults without fangs, both rows of teetli small in adults. Nine to ten circumorbitals, dorsal circumorbital not enlarged, deiTnosphenotic included in orbital margin, three circumorjjitals comprising pos- terior orbital border including the dennosphe- notics. Eight to ten long, thin infraorbitals. Suborbitals not numerous, marginal suborbitals larger tlian internal mosaic. Three lacrimals. Two to three supratemporals on each side of midline. Supracleithrum with bony projections above and below articular facet, its shape dis- tinctive compared with other Lepisosteus ( Fig. 24c). Frontal shape shown in Fig. 20d, frontal attenuated, frontal enameloid reduced. Shape and enameloid pattern of the dermopterotic, parietal, and opercular shown in Figs. 26a, 28a, and 29c respectively. Differs from L. octiJotiis, L. pJatijrhincus, and L. fimhriatus in that L. osseus has two to ^TSiWAgip*-^-'' Fig. 33. — Lateral (upper) and dorsal (lower) views of the skull of Lepisosteus platostomus (UMMZ 190846, DHL-82mm). 54 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 34. — Semidiagiammatic dorsal view of the left lower gill arches of tliree Lepisosteus. (a.) L. platostomus; (b.) L. osseus; (c. ) L. ocidatus (L. platyiliinciis is similar to c). Gill rakers omitted; cartilage stippled. BBC, basibranchial copula; HA, hyoid arch; MTP, median tooth plate; 1-5, gill arches 1-5. Fig. 35. — Lateral (upper) and dorsal (lower) views of the skull of Lepisosteus osseus (AMNH uncat., DHL-224mm). THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS oo four teeth in outer row of premaxilla, others with a single medial tooth in outer row (rarely two in plattjrhincus). Also differs from L. ocii- latus and L. platyrhincus in that osseus lacks large pigment blotches on head (and other details of juvenile and adult color patterns) and has paired first basihyal toothplates, whereas octilatus and platyrhincus have large pigment blotches on head and fused first basihyal tooth- plates. Differs from L. platostomus and L. opertus in tliat L. osseus has incomplete outer row of premaxillaiy teeth, whereas the above species have complete outer row of teeth. Also differs from L. platosiomtis in juvenile and adult color pattern and from L. opertus in that the latter has wide interconnected enameloid ridges on parietals and dennopterotics, whereas osseus has thin continuous tubercles or series of round or oblong tubercles on these bones that are not interconnected. Differs from L. indicus in thaf osseus has enameloid on the dennal skull bones, whereas indicus lacks enameloid. Etymology. — From the Latin osseus (rzrbone). Range. — Quebec to Florida along the Atlantic coast; westward to the Great Lakes region and south to northern Mex- ico and western Texas. Other comments. — Lepisosteus osseus has been reported from the Pleistocene of North Carolina (Hay, 1923) and Kan- sas (Schultz, 1965; Neff, 1975). Arche- ological reports include Keenlyside et al. (1974). f Lepisosteus indicus (Woodward) Figs. 20e, 36 Belonostomtis (?) indicus Woodward, 1890:23; 1895:439. Lepidosteus indicus Woodward, 1908:2. Diagnosis. — Differs from all other Lepisosteus in having dermal bones without ridges or enameloid. Type.— BMNR P.12178, a crushed skull, also vertebrae, and scales (BMNH P.12185, P.12186) figured by Woodward, 1908. Type Locality. — Lameta Beds, Don- gargoan, Madhya Pradesh, Central Prov- ince, India. Formation and Age. — Lameta For- mation, Upper Cretaceous. Description and comments. — The skull (BMNH P.12178) is crushed with the dorsal side showing (Fig. 36). There is no counterpart. The vertebrae are fig- ured by Woodward (1908) and are either on deposit in India or lost. Lepisosteus indicus is a long-snouted gar witli no dermal skull ornamentation. Both fron- tals are partly presei"ved, elongate, and most similar to L. osseus (Fig. 20e). Right frontal 160 mm in length, left frontal is too fragmen- tary to measure. Both premaxillaries are partly preserved. Ascending process of premaxillary a significant part of snout roofing bones, typically Table 2. — Ranges of various meristic counts for seven species of gars.i Character Species L. L. L. L. A. A. A. platostomus osseus oculatus platyrhincus tropicus spattda tristoechus No. predorsal scales 50-55 47-55 45-54 47-51 42-48 49-54 49-51 No. lateral line scales . _. 59-65 57-63 53-59 54-59 41-56 58-62 56-63 No. transverse scale rows — 21-24 19-24 18-24 21-25 20-24 23-32 21-24 No. dorsal fin rays 8-9 6-9 6-9 7-8 7-8 7-10 7-8 No. anal fin rays 8-9 8-10 7-9 7-8 7-8 7-10 7-8 No. caudal fin rays 11-12 11-14 12-13 12-13 11-12 12-14 11-12 No. pectoral fin rays 11-12 10-13 9-13 9-11 11-12 11-15 11-12 No. gill rakers. 27-33 14-31 15-24 19-33 59-62 59-66 67-81 "^ Some data from Suttkus (1963). 56 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 36. — Type skull of Lcpisostetis indicus. d, dentary; Ecpt, ectnpterygoid; Fr, frontal; lo, infraorbital; Pmx, premaxilla; Pop, preopercular; Psp, parasphenoid. gar-like, carrying part of supraorbital canal. Left lower jaw is exposed with its inner surface showing; 181 mm from tip of retroarticular to broken anterior end. Posterior articular series is fragmentary. Articular-retroarticular articu- lating surface is concave, similar in shape to that in other gars. Inner surface of dentary broken, exposing meckelian groove. Both large and small plicidentine teeth are preserved. Part of right infraorbital chain is preserved. Wood- ward (1908:4) incorrectly identified this series of tliree infraorbitals as "palato-pterygoid ar- cade." Additional preparation revealed articu- lations between infraorbitals and infraorbital sensory canal pores. Infraorbitals lack bony ridges or enameloid. One circumorbital was identified. Hyomandibular fragmentary, similar to that of other gars. Parasphenoid (partly pre- served) lies between the frontal s. Left ecto- pterygoid preserved, plus fragments of vomers and dermopalatines. No indication of dermo- palatine dentition. One bone, which fits into position of supraorbital on frontal concavity, does not carry infraorbital canal (like that of Lepisosteus, but not Atractostcus). One canal- bearing bone, tentatively identified as circum- orbital, is located left of the partly preserved hyomandibular. Etymology. — The trivial name indi- cus refers to the country of occurrence. fLepisosteus fimbriatus Wood Figs. 20f, 24d, 26b, 28b, 30a, 37, 38a Lcpidotus fimbriatus Wood, 1846:6, 122. Lcpidosteiis stiessioncnsis Gervais, 1852:4; 1859:517; 1874a:846; 1874b:459. Dollo, 1893:193. Leriche, 1900:188; 1902:44; 1907:243; 1923:183, 186; 1932:369: Priem, 1901:489; 1908:81, 86, 90, 98. White, 1931:80. easier, 1943:2. Lepidosteus maximiliani Vasseur, 1876:295. Lepisosteus sp.: Rutot, 1884: XV. Lepidosteus fimbriatus: Woodward, 1895:442. Diagnosis. — Differs from all Lepiso- steus except L. octdatus and L. pJoty- rhincus in that L. fimhriatus and the two other species have a single medial tooth THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 57 Fig. 37. — Various bones of Lepisosieus fimtjriatus. (a.) dentary (Bi\l.\H l^.lSOO, 4ymin); (b.) partial infraorbital series (BMNH P.33522, 39mm); (c.) frontal (BMNH P.1300, 36mm). on the outer tooth row of the premaxil- lary (Fig. 37a). Differs from L. ociilatiis and L. platyrhincus in details of the shape and enameloid pattern of the deiTnopterotic, parietals, and siipra- cleithrum. Differs from L. indicus in frontal bone shape ( Fig. 20 ) , and in that L. fiinbriatus has enameloid, whereas L. indicus lacks enameloid. Differs from A. strausi in that A. strausi has sheets of enameloid on the demiopterotics, pa- rietals, frontals, and supratemporals, whereas L. fimbriatus has elongate to oval enameloid tubercles on these bones. Types.— BMNH P.25252 (partial den- tary) and P.25254 (opercular). Formation and age. — Known from most freshwater Eocene and lower Oli- gocene deposits of Belgium, France, and England. Description and comments. — All spec- imens of L. fimhriatus studied were small. Various skull elements not shown elsewhere are found in Fig. 37. No me- ristics or morphometries were obtainable. Premaxillary with single medial tooth on outer tootli row (Fig. 37a). Dermopalatines of BMNH P.1300 with small fangs comparable to fangs on small L. platostomiis. Six infraorbitals (BMNH P.33522). Number of circumorbitals indeterminate, dermosphenotic included in or- bital margin. Number of lacrimals indetermi- nate. Preserved suborbitals could not be counted. Two to four supratemporals on each side of midline. Supracleithrum with small bony projections above and below articular facet, its shape compared to other Lepisosteus shown in Fig. 24, 58 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Frontal shape like other relatively short- nosed Lepisosteus ( Fig. 20 ) . Shape and enam- eloid pattern of dermopterotic, parietal, and opercular compared to other Lepisosteus shown in Figs. 26b, 28b, 30a. Enameloid pattern and shape of other bones shown in Fig. 37. Verte- bral centrum ovoid, ventral ridges not wide, transverse process not wide. Scales smooth. Etijmologij . — From the Latin fim- briatus (r^bordered with hairs or fi- brous), referring to the enameloid pat- tern of the opercular. Other comments. — Some v^ertebrae from the Blackheath Beds are subtri- angular, have wide transverse processes and widely spaced ventral ridges similar to A. strausi. No demial skull bones as- signable to A. strausi have been collected in these beds, but collection of additional specimens may show its presence in the Upper Eocene and Lower Oligocene of England. Casier's ( 1943 ) figures may include a lower jaw of Atractosteus. Specimens referred to by Gervais ( 1859, 1874a, 1874b) as L. suessionensis include bones identical to British L. ftmbriatus, as well as bones identified here as Amia sp. indet. Lepisosteus oculatus Winchell Spotted Gar Figs. 2, 8, 20g, 21c, 22a, 24e, 26c, 28c, 30b, 34c, 38b ^Lepisosteus latirostris Girard, 1858:352 (placed in synonymy witli C. castelnaucli=L. ocu- latus by Jordan, Evermann, and Clark, 1930:37). Lepisosteus oculatus Winchell, 1864:183. Sutt- kus, 1963:71. Cylindrosteus productus Cope, 1865:86. Cylindrosieus agassiz Dumeril, 1870:347. Jor- dan, Evermann, and Clark, 1930:37. Cylindrosteus hartonii Dumeril, 1870:347. ?Cylindrosteus zadocki Dumeril, 1870:353. Lepisosteus platostomus Jordan and Evermann, 1896:110 (in part; L. oculatus listed as synonym of L. platostomus) . Cylindrosteus castelnaudii: Jordan, Evermann, and Clark, 1930:37 (in part). Lepisosteus productus: Hubbs and Lagler, 1943:76. Eddy, 1957:40. Diagnosis. — Differs from all extant Lepisosteus except L. platyrhincus in having large pigment spots on the head, having a single medial tooth in the outer premaxillary tooth row, and no medial toothplates on the first visceral arch. Differs from L. platyrhincus in that adults have bony ossicles on the ventral isthmus surface under the gill mem- brane, whereas L. platyrhincus lacks these ossicles (Suttkus, 1963). Type.— UMMZ 55062, a whole dry preserved speciiuen. Description and comparisons. — This species is a gar of medium size and pro- portion, reaching a reported length of 819 mm total length (Suttkus, 1963). The skull of a typical specimen is shown in Fig. 2. Meristic counts are shown in Table 2. Various skull measurements expressed as ratios of dorsal skull length are shown in Tables 3 and 4. Females have proportionately longer snouts than males ( Hubbs and Lagler, 1943; Suttkus, 1963). Lepisosteus oculatus shows much variation in color pattern intensity, especially belly pig- mentation. Generally dark above, lighter on flank. Flank stripe continuous in young, usually broken into series of blotches in adults, running from caudal peduncle anteriorly through eye, onto lower jaw, there fusing with the preoper- cular stripe. Preopercular stripe usually con- tinuous, occasionally broken. Retroarticular spot present; may not contact flank stripe in front of orbit. Head with numerous large pig- ment blotches on dorsal surface. Edges of the infraorbitals with distinct pigment dashes. Gu- lar region motded to solid over its entire area, extending onto istlnnus. One belly stripe per side, usually interconnected by reticulate pig- ment blotches between stripes. Some adults with uniformly dark belly pigment obliterating belly stripes, others widiout belly pigmentation. Belly stripes unite at anal fin and continue posteriorly as a single stripe. Pectoral and pel- vic fins of adults with as many as five transverse pigment bars, juveniles with as few as two. Lower half of pectoral finbase dark. Dorsal, anal, and caudal fins with varying numbers of pigment blotches. Premaxilla with ridges on process for articu- lation of the ectopterygoid and with a single medial tooth in the outer tooth row. Dermo- palatine with two tooth rows, inner row in adults not enlarged as fangs. Six to eight infra- orbitals. Eight to nine circumorbitals with the dorsal circumorbital not enlarged and the der- mosphenotic incoiporated into orbital margin. Three circumorbitals forming posterior margin THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 59 Fig. 38. — Prema>dllas of three gar species, (a.) Lepisosteus oculatus (LACM 33915-1, tol. length eOiiim); (b.) L. fimbriatus (BMNH P.1300, 17mm); (c.) Atractosteus occidentatis (AMNH P.9323, 19mm). of the orbit. Three lacrimals. Suborbitals not numerous. The ventral-posterior marginal sub- orbitals larger than internal mosaic. Two to four supratemporals on each side of midline. Supracleithrum with projecting ridges above and below articular facet, its shape most similar to tliat of L. platyrhincus (Fig. 24e). Dermal skull bones with elongate enameloid tubercles and rounded tubercles. Frontal shape shown in Fig. 20g. Shape and enameloid pat- terns of the dermopterotic, parietals, and oper- cular shown in Figs. 26c, 28c, and 30b, re- spectively. First basihyal toothplate fused. Gill rakers small and pear-shaped (Fig. 21c) numbering 15 to 24 on die first arch outside row (Table 2). Medial toothplate pattern shown in Fig. 34c. Medial toothplates missing on first arch, reduced on second arch to a single row on the hypobranchial, reduced to a single row on the tliird and fourth arches. Medial toothplates missing on the first infrapharyngobranchial. Lepisosteus oculatus differs from L. platostomus in number of lateral line scales (53-59 in L. oculatus, 59-65 in L. platostomus, Suttkus, 1963) and from L. 60 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 39. — Lateral (upper) and dorsal (lower) views of the skull of L. plattjrhinciis (UMMZ 159805-1, DHL-126 mm)- osseus in snout length. Lepisosteus ocu- latus differs from L. fimbriatus in that the medial wing of the cleithrum of L. oculatus is large, whereas in L. fimbri- atus it is small. L. oculatus differs from L. indicus in frontal bone proportions and in that L. oculatus has enameloid on dermal bones of the skull, whereas L, indicus lacks enameloid. Etymology. — From the Latin oculus (=eye) and atus (^provided with). Range. — From the Great Lakes south to the Gulf of Mexico, eastward along the Guff Coast to Western Florida and westward to Central Texas (Suttkus, 1963). Lepisosteus platyrhincus De Kay FLORffiA Spotted Gar Figs. 20h, 21d, 24f, 26d, 28d, 30c, 39 Lepisosteus platyrhincus De Kay, 1870:355. Suttkus, 1963:83. Ctjlindrosteus castelnaudi Dumeril, 1870:355. Jordan, Evermann, and Clark, 1930:37. Lepisosteus platostomus: Jordan and Ever- mann, 1896:111 (in part). Ctjlindrosteus megalops Fowler, 1910:609. Jor- dan, Evermann, and Clark, 1930:37 (men- tioned as doubtful ) . Diagnosis. — Differs from all Lepiso- steus, where the character is known, ex- cept L. oculatus and L. fimbriatus in having one (occasionally two) medial teeth on the outer prem axillary tooth row as opposed to having uniformly two or more teeth on the outer premaxillary tooth row. Type. — Not examined. Supposedly on deposit at the Lyceum, Alexandria, Virginia. Description and comparisons. — Lepi- sosteus platyrhificus is a medium sized gar ( to 1,330 mm total length; Hammett and Hammett, 1936) and usually is darkly pigmented. A skull is shown in Fig. 39. Meristics taken in this study and reported by Suttkus (1963) are shown in Table 2. Morphometries taken from a small series and expressed as ra- tios of dorsal head length are shown in Tables 3 and 4. Color patterns of adults showing polymor- phic variation in color intensity (from "im- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 61 maculate" to very dark individuals from same localities, Suttkus, 1963). Dorsal stripe con- tinuous in juveniles, adults with uniform dark or mottled dorsum. Flank stripe in young from base of caudal fin through the eye, and to tip of the lower jaw. Flank stripe of adults usually broken into series of darker and lighter blotches. Juveniles with stripe between dorsal and flank stripes from nape to one-half the distance be- tween pectoral and pelvic fins. Belly with one or two belly stripes per side, fusing at anal fin to form one stripe on caudal peduncle. Most individuals with belly mottled between belly stripes. Head with large pigment blotches on dorsum. Preopercular and retroarticular stripes fused, and usually fused with flank stripe just anterior to eye. Gular region usually solidly pigmented, isthmus mottled. Pectoral fin base pigmented. Three to five transverse rows of pigment blotches on pectoral and pelvic fins. Dorsal and anal fins with three or four trans- verse rows of pigment blotches. Caudal fin witli many randomly placed pigment blotches. Other color notes given by Suttkus ( 1963 ) . Premaxilla with single medial tooth on outer tooth row (occasional individuals have two teeth on one side, one tooth on other). Dermo- palatine of adults without enlarged fangs. Eight to ten circumorbitals, dorsal circuniorbital not enlarged, dermosphenotic included in or- bital margin. Three circumorbitals comprising posterior orbital margin, including dermo- sphenotic. Six to seven infraorbitals. Three lacrimals. Suborbitals not numerous, marginal suborbitals larger than internal mosaic. Supra- cleithrum with bony projections above and be- low articular facet; shape of supracleithrum shown in Fig. 24f. Shape and enameloid pat- tern of dermopterotic, parietal, and opercular shown in Figs. 26d, 28d, and 30c, respectively. Lepisosteus platyrhinctis differs from L. platostomus in that platyrhincus has large pigment blotches on the head (none in platostomus) and fewer lateral line scales (54-59 in platyrhincus, usually 59-65 in platostomus). L. platyrhincus differs from L. osseus in that osseus either has small pigment blotches on the head or lacks them and has a narrow snout, whereas platyrhincus has large pigment blotches on the head and a wider snout (least width of snout 12.9- 25.7 times in snout length in osseus, 4.8-8.2 in platyrhincus; Suttkus, 1963). Differs from L. indicus in shape of the frontal bone (relatively shorter in pla- tyrhincus. Fig. 20). Differs from L. cu- neatus in that cuneatus has wide enam- eloid tubercles on the derm op tero tics and parietals whereas L. platyrhincus has series of rounded tubercles or rela- tively thin elongate tubercles on these bones (Figs. 26, 28, and 30 respectively). Etymology. — From the Greek platys (=rflat) and rhynchos (= snout). Range. — From the southern tip of Florida northward to the lowlands of Georgia (Suttkus, 1963). Other comments. — Fowler's ( 1917 ) report of L. platostomus from the Pleis- tocene of Florida is considered by Uyeno and Miller ( 1963 ) to be L. platyrhincus. Suttkus ( 1963 ) gives saltwater and eco- logical references. Atractosteus Rafinesque Atractosteus Rafinesque, 1820:75 (type species A. ferox by subsequent designation, Jordan, 1877:11). Clastes Cope, 1873:633 (type species herein designated C. atrox Cope, 1873:633). Clastichthijs Whitley, 1940:243 (a substitute name for Clastes Cope, preoccupied in Arachnida ) . Paralepidosteus Arambourg and Joleaud, 1943: 42 (type species P. africanus Arambourg and Joleaud, 1943, by monotypy). Diagnosis. — Gars with large laterally- compressed gill rakers that are convo- luted on their dorsal edges (Fig. 21), lacking medial toothplates on the first three hypobranchials and ceratobran- chials, and reduced numbers of tooth plates on other gill arch elements. Atractosteus also differs from Lepiso- steus in retaining, primitively, an ecto- pterygoid-premaxillary articulation be- hind the premaxilla proper, in having the articular socket of the supracleithrum without projecting ridges, and in having enlarged demiopalatine fangs as adults. Etymology. — A compound mascuHne nominative derived from the Greek atractos fr^spindle) and Latin osseus (=bony). The following names are considered here as Atractosteus, species indeter- minate: \Clastes pastulosus Sauvage, 1897:94 (Cre- taceous, Portugal). 62 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY ^Atractosteus lapidosus Hay, 1919:110; 1923:375 (? A. spatula. Pleistocene, Florida). f Atractosteus strausi (Kinkelin) Figs. 40-46, 47a, 49a, 51a, 53b Lepidosteus strausi Kinkelin, 1884:244; 1903: 25-26. Andreae, 1893:7; 1894:359. Wood- ward, 1895:444. Diagnosis. — Differs from all Atracto- steus except A. tropicus in that A. strausi has sheets of enameloid on the frontals, parietals, demiopterotics, and supra- temporals. Differs from A. tropicus in having long, continuous enameloid tu- bercles on the opercular bones, whereas A. tropicus has reduced amounts of enameloid on these bones and in having larger scales (lateral line scale count of approximately 45 in strausi, 51-56 in tropicus). Type. — A series of scales deposited at the Senckenberg Museum, not exam- ined in this study. Formation and age. — Messel Local- ity, Dannstadt, Gemiany, Eocene. Description and comparisons. — Atrac- tosteus strausi is an Eocene gar attain- ing large size. A whole specimen (AMNH P.4626) is shown in Fig. 40. A closeup of this skull, two other skulls (AMNH 33839, 33856), and accompany- ing outline drawings of these skulls are shown in Figs. 41-46. Meristics are shown in Table 2. Morphometries ex- pressed as proportions of dorsal head length of AMNH P.4626 are shown in Tables 3 and 4. Premaxilla with two complete rows of teeth. Dennopalatine with an inner row of enlarged fangs. Six infraorbitals, three lacrimals, seven to eight circumorbitals, dorsal circumorbital thin, dermosphenotic included in the orbital margin. Suborbitals large, not as numerous as those of the spatula group. Two supratemporals on each side of midline. Frontal shape typical of tlie genus, most sunilar to A. simplex (probably a result of sim- ilar skull compression, Fig. 53). Shape and enameloid patterns of dermopterotic, parietal, and opercular shown in Figs. 47a, 49a, and 51a, respectively. Enameloid usually in broad, inter- connecting sheets on dorsal dermal bones, and in long, narrow tubercles on opercular series. Infraorbitals, circumorbitals, and suborbitals with enameloid tubercles. Vertebrae subtrian- gular, with broad ventral ridges. Scales smooth. Fig. 40. — Dorsal view of Atractosteus strausi (AMNH P.4626, DHL-52.6mrti). THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 63 fe <: 64 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY L:7 / THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 65 CD n3 ^ • »t C/3 ■73 B c a 03 "a in Tj* 0} p. 66 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 67 Atractosteus sfratisi differs from A. simplex in that A. strausi has three lacri- mals and three circumorbitals on the posterior orbital border, whereas A. sim- plex has two lacrimals and two circum- orbitals bordering the posterior orbital margin. Also differs from A. occidentalis, A. atrox, A. spatula, and A. tristoechus in that A. strausi has a narrow dorsal circumorbital that does not exclude the demiosphenotic from the orbital margin and a large orbit, whereas the spatula group has a broad dorsal circumorbital which excludes the demiosphenotic from the orbital margin and a small orbit. Etymology. — The trivial name in honor of Herrn Banquier Caesar Strauss. Other comments. — See other com- ments for L. fimbriatus. Atractosteus tropicus Gill Tropical Gar Figs. 6, 21e, 22b, 47b, 49b, 51a, 54, 55a, 56c Atractosteus tropicus Gill, 1863:172. Dumeiil, 1870:367. Jordan, Evermann, and Clark, 1930:38. Atractosteus helandieri: Dumeril, 1870:368 (Listed as a synonym of A. tropicus). Lepisosteus tropicus: Jordan and Evermann, 1896a: 111. Miller, 1954:230. Suttkus, 1963:70. Diagnosis. — Differs from all other Atractosteus except A. strausi in having convoluted sheets of enameloid on the dermal skull roofing bones. Differs from A. strausi in having reduced amounts of enameloid on the opercular series (A. strausi has numerous elongate tvibercles on these bones) and more lateral line scales (see diagnosis of A. strausi). Description and Comparisons. — Atrac- tosteus tropicus is a small Middle Amer- ican gar with more body pigmentation than other Atractosteus. A skull of t}'p- ical specimen is shown in Fig. 54. Juvenile color pattern (based on AMNH 33851, 155 mm specimen): dorsum dark, flank light. Flank stripe as dark as dorsum, not ser- rate, running from caudal peduncle to eye, there fusing with general head pigmentation. A thin, less pigmented area between flank stripe and dark dorsum extends from opercular ante- rior to orbital margin. Paired ventral belly stripes on each side extend from opercular membrane to anal fin, merge, and continue as single stripe to caudal fin. Belly between beUy stripes wltli irregular pigment blotches. Ven- tral base of pectoral fin and first half of fin rays and membrane dark brown. One additional transverse pigment stripe on pectoral approxi- mately two-thirds the distance from base to tip. Head with distinct retroarticular and preoper- cular stripes, preopercular stripe extending onto lacrimals. Lower jaw dark except for lightly pigmented area between retroarticular stripe and other darker pigment. Head uniformly dark, without blotches. Some lighter areas along junction of infraorbitals. Gular region with dark stripes along medial jaw margin, irregular pig- ment blotches between stripes. Pelvic fins v^dth three transverse pigment bars. Dorsal, anal, caudal fins with rays alternately light and dark, but fin membranes with little or no pigment. In adults, no distinct dorsal stripe; dorsum either uniformly dark or light with numerous brovm pigment blotches. Flank stripe contin- uous only on caudal peduncle, broken into browTi pigment blotches anteriorly if specimen has a light dorsum with blotches, or solid brown pigment blotches anteriorly if speci- men has a light dorsum with blotches, or solid brown pigment of dorsum extending ventrally to level of ventral margin of flank stripe on caudal peduncle. Flank stripe consolidated on opercular (dark) and sub- operculars (light), ending at orbital margin. Occasional individuals with thin reddish-brown vertical stripes on posterior edge of body scales extending vertically from dorsum to belly. No ventral belly stripes or pigment blotches on belly. Base of pectoral fins without pigment. Dorsum of head uniformly dark. Preopercular stripe missing or restricted to posterior half of preoperculum. Retroarticular stripe missing or indistinct. Brown pigment on lower jaw re- stricted to coronoid process. Pectoral fin with- out distinct pigment blotches. Pelvic fins with- out distinct transverse bars, occasionally with diffuse blotch on posterior half of fin. Dorsal, anal, caudal fins with rays dark and membrane light. Occasional membrane pigmentation at base of caudal fin. Premaxilla without ridges on arm, ecto- pterygoid articulating with premaxilla proper. Premaxilla with two complete tooth rows. Der- mopalatine with tvvo tooth rows, inner row of adults enlarged as dermopalatine fangs. Six to eight infraorbitals. Nine to ten circumorbitals; single dorsal circumorbital not enlarged, dermo- sphenotic incorporated into orbital margin. Three circumorbitals on posterior orbital border including dermosphenotic. Three lacrimals. Suborbitals not numerous, those of ventral- posterior border larger than internal mosaic. 68 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 47. — Dermopterotics of three Atractosteus. (a.) A. strausi (AMNH P.4626, 13mm); (b. ) A. tropicus (AMNH 27939, 29mm); (c.) A. simplex (AMNH P.4203, 31mm). Enameloid pattern shown in black on outline drawings; enameloid of c removed from most of the bone. Two or three supratemporals on each side of midline. Articulatory facet of supracleithrum without ridges, its shape shown in Fig. 55a. Dorsal dermal roofing bones with sheets of enameloid. Enameloid reduced on infraorbitals and opercular series. Shapes and enameloid patterns of the dermopterotics, parietals, and opercular shown in Figs. 47a, 49b, and 51b, respectively. First basihyal toothplate paired. Gill rakers laterally compressed, convoluted on dorsal edge and sitting on basal plates (Fig. 21e). Gill rakers numbering 57-62 on outside of first arch (Table 2). Medial toothplate missing on first three hypobranchials and ceratobranchials and first infrapharyngobranchials, reduced to a single incomplete row on fourth hypobranchials and ceratobranchials and second infrapharyngobran- chials (see Fig. 56c). Range. — Rio San Juan in Costa Rica, Lake Nicaragua, and Rio Usumacinta Drainage of Guatemala and Mexico; Chiapas, Mexico (Miller, 1954). Etymology. — The trivial name tropi- cus refers to the species' occurrence in Middle America. Other comments. — See comments for A. spatula from the Rio Sapoa, Nica- ragua and Costa Rica, and from Lake Nicaragua. t Atractosteus simplex (Leidy) Figs. 47c, 49c, 51c, 53c, 55b, 57, 58 Lepidosteus glaber Marsh, 1871:105 (name only). Cope, 1874:441 (listed not recog- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 69 /* — • • • . ^^^^ I .*.♦*.*••.*•'•••''; •• ^ O. '.•.'»•.*.. .. \«, • . . • . , * • • . ;^> Fig. 48. — Dermopterotics of four Atractosteus. (a.) A. occidentalis (AMNH P.4304, 37mm); (b.) A. atrox (USNM P.4755); (c.) A. spatula (TU 388); (d.) A. tristoechus (USNM 11309). Enameloid pattern shown in black on outline drawings. 70 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 49. — Parietals of three Atractosteus. (a.) A. strausi (AMNH P.4626, 13mm); (b.) A. tropicus ( AMNH 27939, 25mm ) ; ( c. ) A. simplex ( AMNH P.4302, 35mm ) . Enameloid pattern shown in black on outHne drawings; enameloid of c removed from much of the bone. nizable). Woodward, 1895:444. Eastman, 1900a:67 (nomen dubium). Cockerell, 1908:163. Lepidosteus simplex Leidy, 1873a: 73; 1873b: 189. Woodward, 1895:444. Eastman, 1900a: 74. Cockerell, 1908:163; 1909:796. Clastes glaber: Cope, 1873:634; King, 1878: 376, 405. Lepidosteus aganus Cope, 1877b :40. Cockerell, 1908:163. Clastes integer Cope, 1877b :41. Merrill, 1925: 361. Lepisosteus glaber: Hay, 1902:377; 1929:708. Lepisosteus simplex: Hay, 1902:377; 1929:708. Hussarkof and Bryant, 1919:195. Lepisosteus aganus: Hay, 1902:377; 1929:708. Stromer, 1925:361. Lepisosteus integer: Hay, 1902:377; 1929:708. Clastes aganus: Merrill, 1907:7. Lepidosteus integer: Loomis, 1907:358. Cock- erell, 1908:163. Stromer, 1925:361. Diagnosis. — Differs from A. atrox in having a low, bony ridge on the dermal skull roofing bones overlain by round or slightly elongate enameloid tubercles, whereas A. atrox has high bony ridges that are transversely grooved. Atracto- steus simplex differs from all other Atractosteus (for which the characters are known) in having only two lacrimals and two circumorbitals bordering the posterior orbital margin, including the dermosphenotic. Atractosteus simplex differs from A. strausi and A. tropicus in having the dennal skull bone patterns THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 71 Fig. 50. — Parietals of three Atractosteus. (a.) A. atrox (USNM P.4755, 39mm); (b.) A. spatula (TU 388, 65mm); (c.) A. tristoechtis (USNM 111308, 50mm). Enameloid patem shown in black on outline drawings. described above, whereas A. tropicus and A. strausi have sheets of enameloid. Atractosteus simplex differs from A. occi- dentalis, A. atrox, A. spatula, and A. tristoechus in having the demiosphenotic incorporated into the orbital margin, whereas the other species have an en- larged dorsal circumorbital excluding the dermosphenotic from the orbital margin and in that A. spatula has minute, round enameloid tubercles and A. tristoechus lacks enameloid and A. simplex has larger rounded tubercles. Types. — USNM P.2174, one basioccip- ital and three vertebrae. USNM P.21173, basioccipital-parasphenoid, three verte- brae, a tooth, an "antorbital," and three scales designated cotypes. Type locality. — Junction of the Green and Big Sandy rivers, Sweetwater County, Wyoming. Formation and Age. — Bridger, Wa- satch, and Green River Formations, Lower Eocene. Description and comparisons. — The type and cotype resemble the bones as- sociated with more complete AMNH and USNM specimens. The skull of AMNH P.4302 is shown in Figs. 57 and 58. Me- ristic data of Eastman (1900a) are shown in Table 2. Morphometries from this skull: DHL-139.6 mm; PS-72.4 mm; POr-54.2 mm; OpW-18.6 mm; LLJ-86.6 72 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 51. — Operculars of three Atractosteus. (a.) A. strausi (AMNH 33839, 12mm); (b.) A. tropicus (AMNH 27939, 20mm); (c.) A. siinplex (AMNH P.4302, 19mm). Enameloid pattern shown in black on outline drawings; enameloid of c removed from most of the bone. mm; OrD-17.8 mm. Comparisons of these measurements with other gars are shown in Tables 3 and 4. Two complete rows of premaxillary teeth (USNM P.16712). Dermopalatine has inner tooth row of enlarged dermopalatine fangs (USNM P.16712, AMNH P.4302). Six infra- orbitals (AMNH P.4302, P.4308), two lacrimals (AMNH P.4302). Eight circumorbitals (AMNH P.4302), dorsal circumorbital tliin; dermosphe- notic included in orbital margin. Two circum- orbitals make up posterior border of orbit, including dermosphenotic. Relatively few sub- orbitals; marginal suborbitals bordering pre- opercular distinctly larger than internal mosaic. Two supratemporals on each side of midline (AMNH P.4302, P.4308). Supracleithmm with- out distinct projecting ridges above and below articular facet. Supracleithrum shown in Fig. 55b. Frontal typical of genus ( Fig. 53c ) , most similar to A. strausi. Dermopterotic, parietal, and opercular shown in Figs. 47c, 49c, and 51c. Infraorbital enameloid reduced relative to A. strausi and A. tropicus. Enameloid on other skull roofing bones missing or reduced to small, round or slightly oblong tubercles. Vertebrae THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 73 Fig. 52. — Operculars of three Atractosteiis. (a.) A. atrox (USNM P.4755, 43mm); (b.) A. spatula (TU 388, 50mm); (c.) A. tristoechits (USNM 11309, 30mm). Enameloid pattern shown in black on outline drawing. of large specimens distinctly subtriangular. Scales smooth. Etymologij. — From the Latin, sim- plex (= simple) referring to the smooth scales. fAtractosteus africanus (Arambourg and Joleaud) Figs. 16, 59 Paralepidosteus africanus Arambourg and Jo- leaud, 1943:42. Casier, 1961:41. Patterson, 1973:295. Diagnosis. — Differs from all other Atractosteus in the bony ridge pattern of the one preserved infraorbital ( Fig. 16 ) . Vertebrae are larger than other Atracto- steus but are not diagnosable. Type material— MimV N.27-30, N. 155 (vertebrae), N.37 (infraorbital). Type locality. — From the vicinity of Damergou, Niger. Formation and age. — Damergou Beds, Senonian, Upper Cretaceous. 74 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 53. — Frontal shape of six Atractosteus. ( a. ) A. tropicus (AMNH 27870); (b.) A. strausi (AMNH P.4626); (c.) A. simplex (AMNH P.4302); (d.) A. atrox (USNM P.4755); (e.) A. spatula (USNM 26166R); (f.) A. tristoechus (USNM 111309). A lOinm bar is below each drawing. Descriptive comments. — N.155 a large ab- dominal vertebra, 41 mm along ventrum of centrum. Transverse processes not preserved, ventral ridges wide, incomplete neural arches. N.27 a large abdominal vertebra (33 mm cen- trum length, 39.6 mm centrum width, 35.8 mm maximum centrum width), subtriangular, with ventrally directed transverse processes and two widely spread ventral ridges. N.29 a small ( 13 mm centrum length), almost complete, abdom- inal vertebra with ventral ridges like N.27. N.28 a posterior abdominal vertebra, with strongly downtumed transverse processes, meas- uring 19 mm in centrum length. N.30 a caudal vertebra, 25 mm in centrum length, with rec- tangular articulating surface. Infraorbital, N.37, with a random pitted pattern of dermal orna- mentation, not organized into radiating ridges like otlier Atractosteus. This bone interpreted by Arambourg and Joleaud ( 1943 ) as part of dentary. Here interpreted as infraorbital, be- cause: (1) in all other Atractosteus the last in- fraorbital has two closely situated rows of out- wardly directed teeth, and outer row of teeth not much larger than inner row, a characteristic of N.37; and (2) the infraorbital is flattened Uke typical posterior Atractosteus infraorbitals. Etijtnology. — The trivial name afri- canus refers to the continent of occur- rence. f Atractosteus occidentalis (Leidy) Figs. 38c, 48a, 55c, 60 Lepidosteus occidentalis Leidy, 1856a: 120; 1856b:73. Cope, 1877a:574. Woodward, 1895:126. Lambe, 1902:29; 1904:21, 36, 43. Cockerell, 1908:163. Lepidotes haydeni Leidy, 1856a: 120; 1856b:73. Woodward, 1895:125. Hatcher, 1905:67. Clastes occidentalis: Cope, 1884:52. Lepisosteus occidentalis: - Cross, 1896:227. Hay, 1902:337; 1903:119; 1910:296; 1929: 708. Williston, 1902:953; Matthew, 1916: 485. Gilmore, 1924:27. Sternberg, 1924:68. Stomer, 1925:360. Estes, 1964:43; 1969:11. Lepisosteus haydeni: Cockerell, 1908:136. Lepidotes occidentalis: Osbom, 1902:11. Hatcher, 1905:67. Lambe, 1907:179. Peale, 1912:746, 754. Diagnosis. — Atractosteus occidentalis differs from A. strausi, A. tropicus, and Lepisosteus opertus in that occidentalis lacks enameloid on the infraorbitals, cir- cumorbitals, frontals, and preopercular, and has reduced amounts of enameloid on the dermopterotic, whereas the above three species have enameloid on the in- fraorbitals, circumorbitals, frontals, and preopercular, and have broad sheets of enameloid on the dermopterotic. Differs from A. simplex in that simplex has a relatively narrow dorsal circumorbital and a dennosphenotic incorporated into the orbital margin, and low, round or oblong enameloid tubercles on the der- mopterotic, whereas occidentalis has a thick dorsal circumorbital (which could presumably restrict the dermosphenotic from the orbital margin), and high, rounded enameloid tubercles on the THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 75 Fig. 54. — Lateral (upper) and dorsal (lower) views of the skull of Atmctosteus tropicus (AMNH 29075, DHL-123mm). dennopterotic. Differs from A. atrox in that atrox has bony ridges that are trans- versely striated on the frontals and der- mopterotics, whereas occidentalis has non-striated bony ridges devoid of enam- eloid on the frontals and low ridges with rounded enameloid tubercles on the der- mopterotics. Differs from A. spatula in having larger, rounded tubercles, whereas spatula has minute, round tu- bercles. Differs from A. tristoechus in that tristoechus lacks enameloid, and oc- cidentalis has enameloid. Type. — Five scales supposedly de- posited at ANSP. I failed to locate the types. Type locality. — Bad Lands of the Ju- dith River, Nebraska Territory. No other locality given. Formation and Age. — Hell Creek, Judith River, Belly River, Lance, and Arapahoe — Denver Formations, Upper Cretaceous. Description and comparisons. — Atrac- tosteus occidentalis is known only from disarticulated material. Large premax- illas from the Hell Creek formation are presumably A. occidentalis and have two complete rows of prem axillary teeth (Fig. 38c). Dermopalatines with enlarged demiopala- tine fangs. Infraorbitals almost square, without enameloid. Number of infraorbitals undetermi- nable but dorsal circumorbital large; presum- ably excluded dermosphenotic from orbital margin. Those suborbitals that can be distin- guished from L. opertus small, with few scat- tered and rounded enameloid tubercles on the bony ridges. Number of supratemporals per side undetenninable. Supracleithrum with sim- ple articular facet, axis of the bone straight. Supracleithrum most similar in shape to A. atrox (Fig. 55). Overall shape of frontal unknown; frontal without enameloid and with low bony ridges. Parietal shape unknown. Shape and enameloid pattern of dermopterotic shown in Fig. 48a. Preoperculum with high bony ridges on the lateral arm, usually devoid of enameloid ( P"ig. 60 ) . When present, enameloid tubercles on preoperculum minute, widely scattered ( 1 or 2 per ridge. Fig. 60). Etymology. — From the Latin occi- dentalis ( = western ) . Other comments. — A. occidentalis oc- MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY curs in the same localities as L. opertus. Only those bones that can be definitely assigned to A. occideritalis are discussed here, and further studies on such bones as the opercular, subopercular, and cleithi-um will have to wait the collec- tion of more and better preserved mate- rial. The name A. occidentalis is applied to the above material because Estes (1964) implied that A. occidentalis was close to A. spatula. f Atractosteus atrox (Leidy) Figs. 48b, 50a, 52a, 53d, 55d, 61, 62 Lepidosteus atrox Leidy, 1873a:73; 1873b: 189. Woodward, 1895:445. Eastman, 1900a:57; 1900b: 69. Sauvage, 1901:80. Schlosser, 1901:408. Stromer, 1925:360. Lepidosteus notabilis Leidy, 1873a: 98; 1873b: 192. Woodward, 1895:444. Clastes anax Cope, 1873:634; 1884:53. Merrill, 1907:7. Clastes notabilis: Cope, 1877b: 40. Clastes atrox: Cope, 1884:54. Merrill, 1907:7. Lepisosteus notabilis: Hay, 1902:377; 1929: 708. Lepisosteus atrox Hay, 1902:377; 1929:708. Jordan, 1905:32. Merrill, 1907:12. Hus- sakof, 1908:78. Hussakof and Bryant, 1919: 195. Diagnosis. — Differs from all other Atractosteus in having very thick skull roofing bones that have high bony ridges with transverse striations capped with minute enameloid tubercles. Type.— VSNM P.2145, an anterior vertebrae, indistinguishable from other large Atractosteus vertebrae. Type locality. — Junction of the Big Sandy and Green Rivers, Sweetwater County, Wyoming. Formation and age. — Bridger Forma- tion, Lower Eocene. Description and comparisons. — A large Eocene Atractosteus, distinguished by the enameloid pattern described above. Descriptive comments are based on USNM P.4755 (Figs. 61, 62) and MCZ P. 5168 (a complete specimen with crushed skull). Meristics of Eastman (1900a) are shown in Table 2. Morpho- metries taken from the USNM P.4755 in- clude: DHL-309 mm; PL-46.6 mm; FL-139.9 mm; PmxL-126.8 mm; OpW- 41.5 mm. These measurements, ex- pressed as proportions of dorsal head length are shown in Tables 3 and 4. Two complete rows of premaxillary teeth. Dermopalatine not observable. Five to six in- fraorbitals. Number of circmnorbitals not de- terminable. Dorsal circumorbital enlarged, pre- sumably excluded dermosphenotic from orbital margin. Preserved suborbitals small, relative size of marginal suborbitals and internal mosaic not determinable. Three supratemporals on each side of midline. Supracleithrum with sim- ple articular facet, without distinct projecting ridges. Shape of supracleithrum most similar to A. spatula and A. tristoechus (Fig. 55). Frontal shape similar to others in genus (Fig. 53d). Shape and enameloid pattern of die dermopterotic, parietal, and opercular shown in Figs. 48b, 50a and 52a. Differs from A. strausi, A. tropicus, and A. simplex in having enlarged dorsal circumorbital excluding dermo- sphenotic from orbital margin. Differs from A. spatula in that atrox has enameloid pattern as above, whereas spatula has less numerous, larger, rounded tubercles on non-striated ridges. Differs from A. tristoechus in that atrox has enameloid, whereas tristoechus lacks enameloid. Etymology. — From the Latin atrox (n: savage or hideous). THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS Fig. 55. — Supracleithra of six species of Atractosteus. (a.) A. tropicus (AMNH 27939, 25mm); (b.) A. simplex (AMNH P.4302, 28mm); (c.) A. occidentalis (AMNH P.9323); (d.) A. atrox (articulatory head only, USNM P.4755, 25mm); (e.) A. spatula (USNM 172205R, 55mm); (f.) A. tristoechus (USNM 11309, 31mm). Fig. 56. — Semidiagrammatic dorsal view of die left lower gill arches of three Atractosteus gars. Left: A. tropicus; middle: A. spatula; right: A. tristoechus. Gill rakers omitted. BBC, basi- branchial copula; HA, hyoid arch; MTP, median toothplates; 1-5, gill arches 1-5. 78 MISCFT.T.WKOrS PURIJC \TTOX MrSKUM OF NATURAL HISTORY Dpt elm d _ La a s , , ^ So Pop Q\ Dsp (^/--'^ ""Sop Figs. 57 and 58. — Dorsolateially compressed skull (top) and outline drawing (bottom) of Atractosteus simplex (AMNH P.4302, DHL-144mm). a, angular; Ant, antorbital; Chn, cleithrum; d, dentary; Dpt, dermopterygoid; Dsp, derniosphenotic; Fr, frontal; lo, infraorbital; La, lacrimal; Na, nasal; Op, opercular; Pa, parietal; Pmx, premaxillary; Pop, preopercular; Qj, quadratojugal; s, surangular; Sclm, supracleithrum; So, suborbital; Sop, subopercular; St, supratemporal. Table 3. — Range, mean, and standard deviation of four measurements expressed as ratios of dorsal head length for eight species of gars. Species N SL/DHL PS/DHL SW/DHL LSW/DHL L. platostomus _ ... 12 .72-.75(.74,.01)* .43-.52( .47,.03) .14-.22( .17,.02) .09-.12( .10,.0] ) L. osseus 12 .79-.83( .81,.01 ) .25-.36( .33,.05) .09-.12( .11,.01) .04-.06( .05,.01) L. oculatus 8 .70-.77( .74,.02) .33-.52( .42,.07) .12-.22( .18,.03) .08-.15( .1L,02) L. platyrhincus ..... 8 .69-.83( .73,.04) .33-.53( .47,.06) .17-.22( .19,.02) .09-.16( .12,.02) A. tropicus 32 .71-.79( .74,.03) .38-.52( .48,.03) .18-.26( .22,.02) .11-.17(.14.,02) A. simplex 1 .52 .17 A. spatula 11 .69-.75( .72,.02) .48-.62( .55,.04) .20-.30( .27,.04) .16-.21( .18,.02) A. tristoechus 8 .71-.74(.73,.01) .39-.53( .47,.07) .26-.32( .29,.03) .17-.20( .18,.01) * smallest ratio — largest ratio ( mean, standard deviation ) . THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 79 Atractosteus spatula (Lacepede) Alligator Gar Figs. 7, 21f, 48c, 50b, 52b, 53e, 55e, 56a, 63 Lepisosteus spatula Lacepede, 1803:333. Sutt- kus, 1963:83. Lepisosteus ferox Rafinesque, 1820:73. Kiit- land, 1844:18. Lepisosteus belandieri Girard, 1858:353. Atractosteus hicius Dumeril, 1870:366. Lepisosteus viridis Giintlier, 1870:329 (not Esox viridis Gmelin, which is Lepisosteus osseus). Woodward, 1895:441. Goodrich, 1909:341. Litholepis adamantimus Rafinesque: Jordan, 1877:16. Litholepis spatula: Jordan, 1877:16. Lepisosteus tristoechus: Jordan, 1885:13 (in part. North American populations only). Jordan and Evemiann, 1896:111. Forbes and Richardson, 1920:35 (in part. North American populations only). Atractosteus spatula: Jordan, Evemiann, and Clark, 1930:38. Diagnosis. — Atractosteus spatula dif- fers from A. tropicus, A. strausi, A. sim- plex, and A. atrox, in demial roofing bone enameloid patterns (Figs. 47-52), from A. tropicus in lateral line scale count (58-62 in A. spatula, 51-56 in A. tropicus), and number of predorsal scales (49-54 in A. spatula, 43-48 in A. tropi- cus). Atractosteus spatula differs from A. tristoechus in having enameloid on the dermal roofing bones and in gill raker count (59-66 rakers on the first arch outside row in A. spatula, 67-81 in A. tristoechus). Type. — A mounted specimen depos- ited at the Museum National d'Histoire Naturelle, Paris, France. Description and comparisons. — Atrac- tosteus spatula is the largest of living gars reaching a maximum reported length of over 3 meters and weight of 137 kilos (Suttkus, 1963). Meristic counts are shown in Table 2. Skull meas- urements, as expressions of ratios of dor- sal head length, are shown in Tables 3 and 4. Color pattern descriptions have been given by Suttkus ( 1963 ) for adults and juveniles and by Moore et al. (1973) for a small juvenile specimen. Fig. 59.— Infraorbital ( MHNP N.37 ) of Atrac- tosteus africanus. Top, dorsal view; bottom, ventral view. Fig. 60. — Various bones of Atractosteus occidcntaUs. Right, Infraorbital (AMNH P.9323, 12mm); Left, preopercular (MCZ P.9379, 51mm). 80 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 81 Fig. 63. — Lateral ;top) and dorsal (bottom) views of the skull of Atractosteus spatula (AMNH 220946, DHL-186mm). Juveniles witli flank stripe, fading with age. Some adults with flank blotches on caudal pe- duncle. Juveniles and adults without belly stripes, belly without pigment blotches. Head of juveniles dark, no head blotches. Juveniles with pair of dark dorsal stripes bordering a medial light stripe, these stripes running from back of head to base of caudal fin, fading with age. Fins of juveniles with pigment blotches; some adults with blotches. Fin rays usually brown. Adults with dark brown to tan dorsum, fading ventrally to white or yellowish belly. Throat region speckled. Premaxilla with two complete rows of teeth. Dermopalatine fangs as large as infraorbital fangs in adults. Vomer with single, enlarged fang. Five to seven infraorbitals, with enam- eloid lost or reduced to a few minute blisters on the bony ridges. Remaining skull roofing bones with minute rounded enameloid tuber- cles. Opercular series with little or no enam- eloid. Approximately eight circumorbitals. Der- mosphenotic excluded from orbital margin. Many suborbitals; approximately 10 or more along posterior border, including the "demio- hyals." Three supratemporals on each side of midline (a total of six). Supracleithrum with- out projecting ridges above and below the ar- ticular socket, most similar to that of A. tristoe- chus in shape (Fig. 55e). Vertebrae of large adults subtriangular in shape, with wide ventral ridges. Anterior scales with ridges, posterior scales smooth. Gill rakers ornate, sitting on basal plate and laterally compressed (Fig. 21f). Medial tooth- plates missing on first three arches, reduced to single incomplete row on fourth arch and 3-4 rows on fifth ceratobranchial (Fig. 56a); first basihyal toothplates fused. Etymology. — From the Latin spatula (=broad piece or spoon), referring to the broad snout. Range. — From Veracruz, Mexico, northward to the Mississippi River drain- age, including the lower reaches of the Ohio and Missouri rivers, and eastward along the Gulf coast to Choctawhatchee Bay, Florida. Also known from the mouth of the Rio Sapoa, Rivas Province, Nicaragua (TU 388) and, based on iden- tifications made by Dr. W. A. Bussing (Universidad de Costa Rica), from Lake Nicaragua and the Rio Sapoa at La 82 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 64. — Lateral (top) and dorsal (bottom) views of tlie skull of Atractosteus tristoechus (AMNH 3097, DHL-145nini). Criz, Guanacaste Province, Costa Rica. Occurrence of this species between the two disjunct populations is unknown. Other comments. — Atractosteus spa- tula has been reported from the lower Pliocene of Kansas (C. L. Smith, 1962) and the Pleistocene of Texas ( Hay, 1926; Uyeno and Miller, 1962; Swift, 1968) and Florida (Hay, 1919). Atractosteus tristoechus (Bloch and Schneider) Cuban Gar or Manjurai Figs. 5, 48d, 50c, 52c, 53f, 55f, 56b, 64 Esox tristoechus Bloch and Schneider, 1801:395. Lepidosteus maniuari Poey, 1854:273. Litholepis tristoechus: Jordan and Gilbert, 1883:92 (in part, Cuban populations only). Lepisosteus tristoechus: Jordan and Evennann, 1896:111 (in part, Cuban populations only). Atractosteus tristoechus: Jordan, Evemiann and Clark, 1930:38. Alayo, 1973:11. Lepisosteus tristoechus: Suttkus, 1963:70. Diagnosis. — Differs from all other species of Atractosteus in lacking enam- eloid on the dermal roofing bones of the skull. Differs from all Atractosteus except A. occidentalis, A. atrox and A. spatula in having an enlarged dorsal circum- orbital that excludes the dennosphenotic from the orbital margin, from A. spatula in gill raker count (67-81 in A. tristo- echus; 59-66 in A. spatula), and from A. spatula, A. occidentalis and A. atrox in lacking enameloid on the skull bones. Type. — Not examined, presumed on deposit with Bloch and Schneider (1801) types at the Humbolt University Mu- seum, East Berlin. Description and comparisons. — A medium-sized Atractosteus gar of pre- sumably plain coloration. The skull of a typical specimen is shown in Fig. 64. Meristic data are shown in Table 2. Var- ious morphometric measurements, ex- pressed as ratios of dorsal head length, THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 83 are shown in Tables 3 and 4. Specimens examined in this study were old and bleached, but what color pattern re- mained seemed similar to A. spatula. Premaxilla without ridges on its process, ectopterygoid articulating with premaxilla proper. Two complete rows of premaxillary teeth. Derniopalatine widi two tooth rows, in- ner row enlarged as fangs. Three to five infra- orbitals. Eight to nine circumorbitals; dorsal circumorbital enlarged, excluding dennosphe- notic from orbital margin. Three circumorbit- als lining posterior orbital margin, including dorsal circumorbital. Three lacrimals. Sub- orbitals numerous, ventral-posterior marginal suborbitals only slightly larger than internal mosaic. Three or occasionally five supratem- porals on each side of midline. Articular facet of the supracleithrum simple, without projecting ridges, devoid of enameloid. Shape of supra- cleithrum most similar to A. spatula ( Fig. 55f ) . All dermal roofing bones of tlie skull lacking enameloid. Shape of the dermopterotic, pari- etals, and opercular showii in Figs. 48d, 50c, and 52c. Frontal shape shown in Fig. 53f. First basihyal toothplate paired. Gill rakers laterally compressed, convoluted on dorsal edge, sitting on a basal plate (Fig. 21g). Gill rakers numbering 67-81 on first outside arch ( Suttkus, 1963; 67-77 on specimens examined in this study). Medial toothplates absent on first three hypobranchials and ceratobranchials and first infrapharyngobranchials (Fig. 56b), reduced to single incomplete row on fourth arch and sec- ond infrapharyngobranchials. Etymology. — From the Latin tri ( = three) and Greek steachos ('=:rows) re- ferring to the rows of teeth of the lower jaw. Range. — Western Cuba and the Isle of Pines. Phylogenetic Relationships Among Gars The characters used in the foregoing descriptions of the genera and species of gars are discussed below, along with conclusions concering their relative apo- moiphous or plesiomorphous nature. After this discussion, four phylogenetic hypotheses are presented as a summary, two for each genus. The first for each incoiporates only the Recent species, whereas the latter two place the fossil morphotypes at the level to which their preserved characters allows. The ration- ale of this approach rests primarily on the inherent incompleteness of the fossil specimens I examined. I have concluded that the relationships of the fossil fonns are best investigated within the context of an understanding of their Recent rela- tives (following Greenwood et al., 1966; and Nelson, 1969b), for two reasons. First, an analysis based primarily on Re- cent species is especially desirable (and essential) for gars because many of the major synapomorphies of the two genera are either not preserved in fossils or have not been observed to date in fos- sils. Other characters, taken in the con- text of a well corroborated hypothesis of Recent species relationships, and pre- served in fossil gars, pennit elucidation of relationships of the fossil morpho- types. Thus, much of the analysis of fossil gars concerns the discovery of pre- servable characters that are correlated Table 4. — Range, mean, and standard deviation of four measurements expressed as ratios of dorsal head length in ten species of gars. Species N PmxL/DHL L. platostomus 12 .41-.53( .48,.03)' L. osseus 12 .52-.64( .58,.04) L. oculatus ____ 8 .35-.52( .44,.06) L. plaUjrhincus .___ 8 .41-.47( .44,.02) A. strausi 1 .36 A. tropicus 32 .35-.45( .39,.03) A. simplex 1 .33 A. atrox 1 .41 A. spatula 11 .36-.45( .40,.03) A. tristoechus 8 .31-.38( .34,.02) FL/DHL PL/DHL LLJ/DHL .27-.34(.31,.03) .26-.36(.29,.03) .29-.38(.34,.03) .34-.40(.36,.02) .34-.45(.41,.02) .44 .45 .34-.42(.38,.02) .40-.48(.44,.03) .20-.23(.22,.01) .12-.18(.15,.02) .16-.23(.20,.03) .19-.24(.21,.02) .15-.27(.21,.02) .20 .15 .20-.25(.23,.02) .18-.21(.19,.01) .63-.67(.65,.01) .73-.78(.75,.02) .64-.70(.67,.02) .61-.67(.64,.02) .66 .60-.68(.65,.02) .62 .59-.66(.63,.02) .63-.72(.73,.01) " smallest ratio — largest ratio (mean, standard deviation). 84 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY with synapomoiphies of minimal am- biguity. Beyond these considerations, analyza- ble fossil gars are rare. In some species, such as L. opertus and A. occidentolis, the moiphotype is represented only by disarticulated remains; many of these remains are not definitely assignable to one or the other species. In A. afri- canus the entire morphotype is repre- sented by one infraorbital and a few vertebrae. In other species, such as L. incUcus and A. atrox, the morphotype is represented by one or two skulls. Be- cause of the small number of specimens available for analysis, intraspecific vari- ation cannot be assessed. MONOPHYLY OF THE GeNERA Frontal bones. — The shapes of the frontal bones of all gars reflect the lengthening of the etlimoid and otic re- gions of the gar snout. As such, the elongate frontals of gars are apomor- phous relative to the short snouts of most other actinopterygians. Within the family, the Atractosteus gars have shorter frontals ( Fig. 53 ) than the Lepi- sosteus gars (Fig. 20). This is reflected not only in the length of the bone, but also in its articulation with both the pre- maxillary process and the parietals and dennopterotics. The premaxillary-frontal articulation in Lepisostens is narrow lat- erally and the posterior arm of the pre- maxillary process is very thin. In Atractosteus the lateral boundary of ar- ticulation is broad and slopes gradually toward the posterior of the premaxillary process. In Lepisostens gars the lateral edge of the frontal grows back along the lateral border of the dermopterotic, whereas in Atractosteus the dermopter- otic and frontal meet at more or less right angles to the long axis of the skull. The narrow and more elongate snout of Lepisostens which is reflected in the shape of the frontal and its articulation with other skull roofing bones is hypoth- esized to be more apomoi-phous than the snout and frontal bone shape of Atracto- steus. Gill Rakers. — There are two types of gill rakers found among Recent gars. The first, found in Lepisostens gars, is not numerous, is pear-shaped and stud- ded with many relatively large teeth. These gill rakers are similar to those of Amia calva, Alhula, and semionotids (Fig. 21a-d). They are hypothesized to be primitive relative to the gill rakers of Atractosteus. The Atractosteus gill rak- ers are numerous, laterally compressed, convoluted on their dorsal edge, and sit on basal plates (Fig. 21e-g). No other actinopterygian is known to have such gill rakers. The Atractosteus raker is considered a synapomorphy uniting Atractosteus into a monophyletic group. Medial toothplates of the visceral arches. — As discussed above, the demial arch elements primitively consist of se- ries of lateral plates, gill rakers, and medial toothplates. There are no medial toothplates in Atractosteus gars on the hypo- and ceratobranchials of the first three arches and these toothplates are reduced to a single incomplete row on these bones on the fourth arch ( Fig. 56 ) . Further, Atractosteus gars lack tooth- plates on the first infrapharyngobran- chial and have a reduced number on the second and third infraphaiyngobran- chials (Fig. 22). Lepisostens gars have these toothplates, albeit reduced as com- pared with those of other neopterygians such as Amia and teleosts ( Fig. 34 ) . I conclude that the absence of medial toothplates on the first three arches and a reduction of the fourth arch is a svn- •I apomorphy uniting Atractosteus gars. Ectopter [I goid-pre maxilla articulation. — The two genera differ in the way the ectopterygoid and premaxilla articulate with each other. In Atractosteus, the ectopterygoid articulates immediately posterior to the nasal foramen on the premaxilla proper. This is essentially the same type of articulation found in other actinopterygians (although, of course, the autapomorphous premaxillary proc- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 85 ess of gars produces a ectopterygoid- premaxillaiy association not similar to that of other actinopterygians ) . The ar- ticulation in Lepisosteus differs in that the ectopterygoid articulates with the premaxillary process and not with the premaxilla proper. In L. pJatostomns, L. oculatus, and L. pkityrhincus, this is re- flected by a series of ridges on the pre- maxillaiy process (Fig. 38a). In L. os- seus, the ridges have presumably been obliterated by increased snout length (see discussion below). Unfortunately, no fossil gar has this character pre- served. I conclude that Recent Lepiso- steus gars share a synapomoq^hic ecto- pteiygoid-premaxilla articulation. Dermopalatine fangs. — During on- togeny the demiopalatines of both Atractosteus and Lepisosteus juveniles have dermopalatine fangs as large as those on the infraorbitals (Suttkus, 1963). As growth proceeds the dennopalatine teeth of Lepisosteus do not grow as large as those of Atractosteus. Fully grown Atractosteus gars have an inner row of dermopalatine fangs as large as the infraorbital fangs and an outer row of teeth corresponding in size to the outer row of smaller infra-orbital teeth. Adult Lepisosteus gars lack the dermo- palatine fangs, having only two rows of the smaller teeth. Enlarged demiopala- tine teeth are typical of many semiono- tiforms and of Amia. The presence of dennopalatine fangs in outgroups and the ontogenetic evidence leads to a con- clusion that the lack of dennopalatine fangs in adult Lepisosteus is a synapo- morphy of the genus. Enatneloid. — Some species of both Lepisosteus and Atractosteus have broad, flattened enameloid tubercles which in- terconnect to form convoluted patterns on the parietals and dermopterotics as well as more or less extensive amounts of enameloid on the other skull roofing bones (i.e. L. opertus, Figs. 25a, 27a, and L. cuneatus. Figs. 25b, 27b; A. strausi, Figs. 47a, 49a and A. tropicus, Figs. 47b, 49b). In all Lepisosteus gars except L. indicus there are also large amounts of enameloid on the infraorbit- als and preoperculars (in L. opertus, Fig. 23, it is difficult to decide because of weathering). All Atractosteus gars show some reduction (or loss) of enam- eloid on the infraorbitals and the pre- operculars or both ( Fig. 60 ) . The initial loss, or reduction, of enameloid on the infraorbitals and the tendency toward continued reduction of enameloid on other skull roofing bones is hypothesized here to be apomorphous for the genus Atractosteus, whereas the retention of this enameloid in primitive Lepisosteus gars is considered plesiomorphous. Supracleithrum. — Gars differ from other actinopterygians in having a supra- cleithrum with a concave articular facet for articulation with the post-temporal. All Atractosteus gars have simple artic- ular facets without bony projections above and below the facets (Fig. 55) while Lepisosteus gars have these pro- jections (Fig. 24). The absence of these projections in Atractosteus or other ac- tinopterygians leads to a conclusion that the projections are synapomor]^)hic for Lepisosteus. The shape of the supra- cleithrum also differs in the two genera (Figs. 24, 55), but the characterization of either as plesiomorphous or apomor- phous is not justified. Relationships Among Lepisosteus Gars PremaxiUary tooth pattern. — The pre- maxillary teeth are known in six of the eight species of Lepisosteus. Lepisosteus platostomus and probably L. opertus differ from all other species in having two complete rows of teeth on the pre- maxilla. This is similar to the pattern in Atractosteus (Fig. 38c) and is consid- ered plesiomorphous. Other Lepisosteus display a reduction trend in the number of teeth on the outer row. Lepisosteus osseus has two to four teeth, whereas L. oculatus, L. platyrhincus, and L. fim- briatus (Fig. 38b) have a single tooth on the outer row (occasionally two in platyrhincus). The condition in L. os- 86 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY seus is considered plesiomorphous rela- tive to those in the latter three species. Ectopterygoid-premaxillo articulation. — The ectopterygoid-prem axilla articula- tion is known only from Recent species. The articulation of the ectopterygoid and the nasal process of the prem axilla is considered a synapomorphy of Lepi- sosteus and this articulation is mani- fested by a series of ridges on the side of the premaxilla in L. platostomus. L. oculatus, and L. pkityrhinciis (Fig. 38a). Lepisosteus osseus does not have these ridges but maintains the usual Lepiso- steus articulation. It is logical to assume that the loss of ridges on the nasal proc- ess in L. osseus is an apomoiphic condi- tion related to greatly increased snout length in which the ectopterygoid has elongated to such an extent that the bone is very thin at its anterior end. Medial toothplates of the gill arches. — The distribution of medial toothplates is known only in Recent forms. As dis- cussed above, Lepisosteus differs from Atractosteus in having some medial toothplates on the first three hypobran- chials. Both L. platostomus and L. 0.9- seus have medial toothplates on the first hypobranchials, whereas L. oculatus and L. platyrhincus lack medial toothplates on the first hypobranchials (Fig. 34). This condition in the latter species is hypothesized to be derived relative to the foraier. Color pattern characteristics among Lepisosteus. — Color patterns vaiy in adults between veiy plain (L. platosto- mus) to very spotted (L. platyrhincus). Three species, L. osseus, L. platyrhincus, and L. oculatus retain, as adults, more of the juvenile color pattern character- istics common to all Recent species of Lepisosteus and to juvenile A. tropicus, whereas L. platostomus loses or greatly reduces the intensity of these color pat- terns. The flank stripes, dorsal stripe and belly pigmentation of L. oculatus, L. platyrhincus, and L. osseus are hy- pothesized to be plesiomorphous relative to the condition found in large juvenile and adult L. platostomus on the basis of ontogenetic and outgroup criteria. The further elucidation of synapomorphies between the other Recent species of Lepisosteus is complicated by the range of intraspecific variation found in L. osseus and L. oculatus (see descriptions above and color notes by Suttkus, 1963 ) . One character seems relatively unambig- uous— the large blotches on the head of L. oculatus and L. platyrhincus are hy- pothesized to be apomorphous. This hypothesis is weakened by the observa- tion that some L. osseus have head blotches (albeit small) while some lack head blotches entirely. Enameloid patterns on skull roofing hones. — As discussed above, the plesio- morphous enameloid pattern for gars is hypothesized to consist of sheets of enameloid on the supratemporals, pari- etals, and demiopterotics. This pattern is found in L. opertus (Figs. 25a, 27a). Lepi- sosteus cuneatus (Figs. 25b, 27b) retains most of the plesiomoq^hous pattern but shows some reduction of enameloid com- pared to L. opertus. Lepisosteus ocu- latus, L. platyrhincus, L. fimhriatus, and L. platostomus tend to have elongate enameloid tubercles or enameloid tuber- cles present as series of oblong tubercles on these bones (Figs. 25-28). Lepisosteus osseus (Figs. 26a, 28a) has less enam- eloid on the parietals and demiopterotics than other species, except L. indicus (which lacks enameloid completely). All Lepisosteus show a tendency to have small rounded enameloid tubercles on the frontal bones (but not L. opertus. Fig. 23a, or L. cuneatus. Fig. 31 ) . Lepi- sosteus osseus has reduced the amount of enameloid on the frontal bones to a greater extent than other Lepisosteus which have enameloid, and the condition in L. osseus is hypothesized to be inter- mediate between the usual condition in the genus and the condition seen in L. indicus, in which the enameloid is miss- ing and the bony ridges are reduced in height. The complete loss of enameloid THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 87 in L. indicus and A. tristoechus is con- sidered here as a nonhomology. Ethmoid elongation. — Lepisosteiis os- seus is autapomoiphic among Recent spe- cies of Lepisosteus in the length of its snout. The long snout of L. indicus is hypothesized to be a synapomorphy with L. osseus when fossil and Recent species are considered (Figs. 20d, e). Relationships Among Atractosteus Gars Dermosphenotic and the circum- orbital series. — The position of the der- mosphenotic, the relative size of the other circumorbitals, and the relative size of the orbit are known in all Recent species of Atractosteus, A. strausi, A. simplex, and can be inferred from the remains of A. atrox and A. occidentalis. In all of these species except A. tropicus (Fig. 54), A. strausi (Figs. 41, 42), and A. simplex (Figs. 57, 58) the dermo- sphenotic is excluded from the orbit by an enlargement of the dorsal circum- orbital (the supraorbital) and the orbit is relatively small (Figs. 61, 62, 63, 64). In all Recent and fossil Lepisosteus where the condition is known the dermo- sphenotic is included in the orbital mar- gin, the dorsal circumorbital is thin, and the orbit is large, just as seen in A. tropicus, A. strausi, and A. simplex. I conclude that the conditions seen in A. tropicus, A. strausi, and A. simplex are plesiomorphic relative to that seen in the spatula species group. Infraorbital enameloid. — The only two species of Atractosteus retaining sig- nificant amounts of enameloid on the in- fraorbitals are A. tropicus and A. .strausi. Other species either completely lack enameloid on the infraorbital or have minute amounts as an individual vari- ation ( c.f . A. spatula ) . All disarticulated infraorbitals of A. occidentalis examined lacked infraorbital enameloid (Fig. 60a), as do the infraorbitals of the articulated skulls of A. atrox and A. simplex. The single preserved infraorbital of A. afri- canus lacks enameloid (Fig. 59). Enam- eloid is present on the infraorbitals of all Lepisosteus, and since large amounts of enameloid on dernial bones seems to be a primitive actinopterygian charac- teristic, I interpret the presence of infra- orbital enameloid as plesiomorphous. Thus, the lack ( or virtual lack ) of enam- eloid in A. africanus, A. occidentalis, A. simplex, A. atrox, A. spatula and A. tristoechus is considered to be derived relative to the condition of A. tropicus and A. strausi. Other enameloid patterns. — Enam- eloid patterns within Atractosteus in- clude both proliferation and reduction of enameloid. The primitive pattern for the family is hypothesized to be similar to A. tropicus and A. .strausi (as well as L. opertus), which have broad, flat, in- terconnected enameloid tubercles run- ning on top of the bony ridges of the dermopterotics, parietals and other roof- ing bones (Figs. 47a, b, 49a, b). This hy- pothesis is not refuted by the obser- vation that many chondrosteans have essentially the same type of enameloid. A. atrox has a unique bony ridge pattern and minute enameloid blisters (Figs. 48b, 50a). Atractosteus occidentalis and A. simplex have reduced numbers of round or oblong tubercles, which are considered to be plesiomorphic relative to the minute enameloid tubercles of A. atrox and A. spatula. Atractosteus tristo- echus lacks enameloid completely and is hypothesized to have the most apomor- phous condition within this group. Color pattern. — Atractosteus tropicus juveniles have all the color pattern char- acters considered to be plesiomorphous for the family, and adults retain some of these conditions. Atractosteus spatula and A. tristoechus lack many of these pigment patterns and their color pattern similarities are hypothesized to be apo- morphous relative to the patterns of A. tropicus. Summary of relationships and discus- sion.— The four phylograms presented below in Figs. 65-68 represent the least rejected hypotheses of relationships among the species of each genus. 88 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 65. — Least rejected hypothesis of phylogenetic relationship among recent Lepisosteiis gars. Synapomorphies (black rectangles) connecting taxa are: (I) large pigment blotches on head; (2) no medial toothplates on first infrapharyngobranchial; (3) single tootli on outer premaxillary tooth row; (4) long narrow snout; (5) ridges on premaxillary ami missing; (6) number of teeth on outer premaxillary tooth row four or less; (7) adults showing a reduction of the primitive color pattern, i.e., no belly stripes, reduced flank stripe, no transverse pig- ment blotches on paired fins; (8) no enlarged dermopalatine teeth; (9) ecterygoid articu- lation on premaxillary arm; (10) frontal narrow anteriorly and growing backward posteriorly past the lateral anterior end of the dermopterotic; (11) enameloid on skull roofing bones in narrow elongate tubercles; (12) synapomorphies of the family Lepisosteidae: see Fig. 19. Fig. 65 is a summary of the least rejected hypothesis of relationships of Recent Lepisosteus gars. The monophyly of the genus is corroborated by 4 char- acters (8-11). That L. osseus is more closely related to L. occulatus and L. platijrhincus is corroborated by one character (6), while the monophyly of the oculatus-platyrhincus species pair is corroborated by 3 characters (1-3). Fig. 66 summarizes the least rejected hypothesis of relationship among fossil and Recent Lepisosteus. The only de- rived character that L. opertus shares with other Lepisosteus is the correct frontal-dermopterotic articulation. Lepi- sosteus cuneatus has this character, plus a lack of demiopalatine fangs (a condi- tion not observable in L. opertus). Lepi- sosteus cuneatus is hypothesized to be THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 89 Fig. 66. — Least rejected hypothesis of phylogenetic relationship among fossil and Recent Lepisosteus. Synapomorphies (black rectangles) connecting taxa are: (1) syTiapomorphies 1-3, Fig. 65; (2) single medial tooth on outer premaxillary tooth row; (3) frontal long and narrow anteriorly; (4) enameloid reduced or missing on frontal; (5) number of teeth on outer premaxillary tootli row four or less; (6) enameloid reduced to thin, elongate and/or dis- connected tubercles, not in sheets; (7) no dermopalatine fangs; (8) enameloid on skull roofing bones in wide, continuous tubercles, not in sheets; (9) frontal shape of the genus. more derived than L. opertus based on its more derived enameloid pattern. Lepisosteus indicus is hypothesized to be the sister species of L. osseus, based on the relatively more elongate frontals of both and on the hypothesis that L. osseus is inteiTnediate in enameloid pat- tern and bony ridge height between L. indicus and other Lepisosteus. Lepiso- steus jimhriatus is hypothesized to be the sister species of the oculatus- platyrhincus species pair, based on the shared single, medial tooth in the outer tooth row of the premaxillaiy. Fig. 67 summarizes the least rejected hypothesis of relationship among Recent Atractosteus gars. The monophyly of the genus is corroborated by three synapo- morphies (7-9). That A. spatula is the Recent sister species of A. tristoechus is corroborated by five characters (2-6). One autapomorphy is also presented. Fig. 68 summarizes the relationship of fossil and Recent Atractosteus. Atrac- 90 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 67. — Least rejected hypothesis of phylogenetic relationships among Recent Atractosteus gars. Synapomorphies (black rectangles) connecting taxa are: (1) No enameloid on skull bones; (2) loss of flank, belly and dorsal stripes in adults; (3) dorsal circumorbital enlarged, dermosphenotic excluded from the orbit; (4) suborbitals numerous compared to otlier gars; (5) enameloid eitlier missing or reduced to minute, round tubercles on skull roofing bones; (6) no enameloid on infraorbitals; ( 7 ) no medial toothplates on first three gill arches ventrally; (8) gill rakers large, laterally compressed, and with basal plates; (9) infraorbital enameloid reduced compared to Recent Lepisosteus; ( 10 ) synapomorphies of the family Lepisosteidae, see Fig. 19. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 91 Fig. 68. — Least rejected hypotliesis of phylogenetic relationship among fossil and Recent Atractosteus gars. Synapomorphies (black rectangles) connecting taxa are: (1) enameloid tubercles missing or minute and not numerous compared to other Atractosteus; (2) enameloid tubercles minute but numerous; ( 3 ) dorsal circumorbital enlarged ( condition of A. africanus unknown); (4) no enameloid or virtually no enameloid on infraorbitals; (5) enameloid on dermal skull roofing bones reduced to small, rounded or oblong tubercles, not in sheets; (6) enameloid on opercular bones reduced; (7) enameloid on infraorbitals reduced. tosteus strausi is considered primitive to other Atractosteus and its inclusion in the genus is based on reduction of enam- eloid on the infraorbitals. Atractosteus tropicus is hypothesized to be more de- rived than A. strausi based on the reduc- tion of the enameloid tubercles on the preopercular of A. tropicus and other Atractosteus relative to the retention of these long enameloid tubercles in A. strausi. Other Atractosteus either lack enameloid on the infraorbitals or have less enameloid on the infraorbitals than either A. strausi or A. tropicus. Atracto- steus simplex is placed above A. strausi and A. tropicus. Atractosteus simplex has a reduction of enameloid on the skull roofing bones but retains the primitive 02 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY conditions of a thin dorsal circumorbital and a large orbit ( at least in small speci- mens). A. africamis shares with A. occi- denfolis and the other members of the spatula species group a severe reduction in infraorbital enameloid (on the one specimen available) and it has subtrian- gular vertebrae (a similarity produced by growth, i.e., small A. spatula have oval centra in cross section, whereas large specimens have triangular centra in cross section). The remaining species (A. atrox, A. occidentalis, A. spatula and A. tristoechus) of the spatula species group have a large dorsal circumorbital which excludes the dennosphenotic from the orbital margin and have a small orbit relative to the large orbit found in other Atractosteus and in all Lepisosteus. Within this group A. atrox is autapo- morphic in having a unique enameloid pattern on the parietals, denuopterotics, and frontals. Atractosteus occidentalis is plesiomorphous relative to other mem- bers of the group in having large, rounded enameloid tubercles on the skull roofing bones relative to the minute enameloid tubercles found in other spe- cies. Atractosteus tristoechus is the most derived species based on a complete lack of enameloid in this species. A Classification of Gars The classification presented here summarizes the relationship of gars to other actinopterygians and the interrela- tionships of gars among themselves. This classification adopts two conventions used for combining recent and fossil groups in a single classification, that of Nelson's (1972a, 1973c) use of the tenn "incertae sedis", and that of Patterson's and Rosen's (in press) use of the tenn "plesion." Nelson (1972a, 1973c) sug- gested that the term incertae sedis be reserved for fossil species or fossil groups of uncertain relationship. These species or groups of species of uncertain posi- tion are listed here at the level that their preserved characters allow phylogenetic placement. Patterson and Rosen (in press ) feel the advantage of Nelson's use of the tenn is that it separates the un- certainties about placement of Recent groups })rought about because of defi- cient theories of relationship from the uncertainties about placement of fossil groups brought about by poor specimen preservation (for example, A. africanus). Patterson and Rosen ( in press ) also used incertae sedis for interchangeable taxa and for the inclusion of non-monophy- letic groups. Neither of these problems arise in gar classification, and neither connotation is implied by the use of the tenn here. The "plesion" (Patterson and Rosen, in press ) is a fossil group or a fossil spe- cies that is sequenced by a listing con- vention in a classification and which is the primitive (plesiomorph) sister spe- cies (group) of all species listed below it in the classification. Plesions are not given fonnal rank in the classification presented below. When fonnal rank is given a plesion, this name is applied onlv within the context of the classifi- cation of the Recent group of which it is a part, and the plesion does not affect the hierarchic position of Recent groups with which it is associated. Thus, if later information is gathered concerning re- lationships of the plesion, then its posi- tion in the classification can be changed without affecting the formal hierarchic ranks of the Recent groups with which it is associated in the classification. The advantages of this system of conventions can be summarized as ( 1 ) uncertainties arising from the incom- pleteness of fossil specimens are clearly identified in the classification by the tenn incertae sedis, (2) fossil species or groups of species can be incorporated into a classification of Recent organisms without changing the fonnal ranks of these Recent organisms (thus increasing the information content of the classifi- cation over one that included only Re- cent organisms), (3) the plesion groups can be changed as new infonnation about them is discovered without affect- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 93 Fig. 69. — Track analysis of the distributional patterns of the genera Lepisosteus (dots) and Atractosteus (dashes). Black circle is the locality of Lepisosteidae species indeterminant (Casier, 1961), which defines the known southern limit of the family in Africa. ing the ranks of Recent groups, and (4) the phylogenetic relationships among both the fossil and Recent groups can be expressed exactly. Infraclass CHONDROSTEI Infraclass NEOPTERYGII Division HALECOSTOMI Subdivision HALECOMORPHI Subdivision TELEOSTEI Division GINGLYMODI Family LEPISOSTEIDAE Genus Lepisosteus plesion f L. opertus plesion \L. cuneatus L. platostomus L. osseus species pair plesion \L. indicus L. osseus L. oculatus species group plesion f L. fimbriatus L. oculatus L. platyrhincus Genus Atractosteus plesion f A. strausi A. tropicus plesion f A. simplex A. spatula species group f A. africanus and fA. occi- dentalis incertae sedis in A. spatula species group. plesion f A. atrox A. spatula A. tristoechus Gar Biogeoraphy The distribution of gars is analyzed below using the "vicariance" method of analysis discussed in the methods sec- tion. Although Recent gars are known only from the northern part of the West- ern Hemisphere, fossil gars are known from other continental areas (see ranges for the various fossil species above ) . Of interest here are the distributional pat- terns, or tracks, of the two genera Lepiso- steus and Atractosteus, and various monophyletic groups of species within each of the genera. These tracks are shown in Figs. 69-72 and are discussed below. Distributional patterns of the genera. — Fig. 69 shows the distributional pat- terns of Lepisosteus and Atractosteus projected on a map of present continen- tal positions. The generalized track of Lepisosteus includes the ranges of a 94 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Pig. 70. — Track analysis of the distributional patterns of the osseus-indicus species pair ( dots ) and tlie oculatiis species group (dashes). Black areas denote the known range of the oculatus species group, black circle denotes the known range of L. indicus. number of species from North America (Upper Cretaceous to Recent), one Euro- pean species (L. fimbriatus, Eocene to Oligocene), and one Indian species (L. indicus, Cretaceous ) . Both northern and southern land masses are inchided within the track. The Lepisosteus gen- erahzed track includes three species group tracks within it, and will be exam- ined in a separate section beyond. The Atractosteus generalized track (Fig. 69) includes the ranges of several fossil and one living North American species (Cretaceous to Recent), two nu- clear Middle American species (A. trop- icus and a disjunct population of A. spatula), one Caribbean species (A. tristoechus), one African species (A. afri- canus), and one European species (A. strausi). Like the generalized track of Lepisosteus, that of Atractosteus in- cludes both northern and southern land masses within its area. The Atractosteus generalized track is composed of several individual tracks that will be discussed below in a separate section. Two observations can be made from the tracks of the two genera: firstly, both genera are found on parts of what once were Laurasian and Gondwanian land masses; secondly, the genera show a large amount of sympatry, both on Recent and fossil distributions. Three conclusions can be drawn: ( 1) both gen- era may have had ancestral Pangean distributions; (2) because sympatry im- plies dispersal, one or both of the genera must have dispersed (possibly over a Pangean landscape rather than a land- scape of present continental positions); (3) because an allopatric speciation event (a vicariance event) must come before dispersal producing sympatry, the vicariance event producing the two gen- era may have occurred before the break- up of Pangea. Thus, the minimum age of the genera of gars is hypothesized here to be 180 million years before the present. Lepisosteus species cUstrihutional pat- terns.— Lepisosteus contains a number of individual tracks. The more primitive members of the genus, L. opertus and L. cuneatus, are western North American forms, whereas the other members of the genus are either eastern North Ameri- THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 95 Fig. 71. — Track analysis of the distributional patterns of the A. spatula species group, open circles are fossil localities, black areas are Recent species ranges. Black circle is tlie disjunct population of A. spatula from Nicaragua and Costa Rica. can, European, or Indian. Whether the western ranges of the primitive members of the genus represent an early vicari- ance event or two early vicariance events is problematical, because it is difficult to determine whether their ranges are the result of sample error in collecting, availability of suitable formations, or whether their observed ranges are repre- sentative of their natural ranges. This problem could be resolved by tying these two species to fossil biotas and showing that they form part of a gener- alized track among Cretaceous and Eocene faunas in general (the same problem exists in determining the bio- geographic importance of the western North American fossil Atractostetis) . The track describing the range of the group made up of L. platostomus and more derived species is identical with that describing the range of the osseus species pair plus the oculatus species group. This is because L. platostomus is found within the range of L. osseus. However, it is interesting to note that L. platostomus is a relatively westerly form. The osseus species pair track (Fig. 70) contains two species, L. osseus of North America and L. indicus of India (Cretaceous). This track is hypothesized to be older than the oculatus species group track because it indicates Pangean distribution. The oculatus species group track (Fig. 70) contains two North American species (L. oculatus and L. platyrhincus, both Recent) and one Eu- ropean species ( L. fimbriatus. Eocene to Oligocene). This track conforms to a more generalized track between Europe and eastern North America composed of a number of fossil and Recent groups (McKenna, 1975), and the vicariance event that split the common ancestor of L. fimbriatus and oculatus-platyrhincus is hypothesized to be Early Eocene in age. Atractosteus species group distribu- tional patterns and Rosens hypothesis of Caribbean biogeography. — Rosen (1975) reviewed the biotic composition of the Caribbean region and concluded that several generalized tracks composed of many diverse organisms were involved in producing the distributional patterns appearing in the region today. Two of 96 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Fig. 72. — Track analysis of the spatnla-tristoechus species pair. Dashes enclose the range of A. spatula, dots connect the range of A. tristoechiis with A. spatula. these generalized tracks involve species of the A. spatula species group — the North American-Caribbean generalized track and the Caribbean-West African generalized track. The A. spatula species group fonns a track (Fig. 71) composed of species from North America, nuclear Middle America, Africa, and the Caribbean. Within this track is another (Fig. 72) composed of the two disjunct popula- tions of A. spatula and the Caribbean A. tristoechus. Rosen (1975) concluded that the Recent Atractosteus distribu- tions in the Caribbean region were prob- ably a part of an older North American- Caribbean generalized track of Pangean or Laurasian affinities, as opposed to a younger North American-Caribbean track with primarily Gondwanian affini- ties (a component of which would be Gamhusia, for example). Additionally, Rosen concluded that the final determi- nation of phylogenetic affinities among Recent Atractosteus would help deter- mine the relative age of the vicariance event that produced A. tristoechus. Within the A. spatula species group there are at least two levels of vicari- ance. First, the species group as a whole confonns to a track drawn between North America and Africa, and this track predates the fomiation of the Caribbean Region (to a time when, following Ro- sen's conclusions concerning the origin of the Caribbean, the region would still have been part of the Pacific seafloor). The minimum age of this track can be inferred from the current estimate of time of separation of western Africa from North America, which supposedly occurred during the Jurassic. Second, the spatula-tristoechus species pair con- fonns to the generalized North Amer- ican-Caribbean track discussed by Rosen (1975), and is not necessarily older than similar distributions confonning to this track (i.e. Gamhusia, etc.). Third, Ro- sen's conclusion that Atractosteus as a whole is Pangean is corroborated, since the most primitive species in the region, A. tropicus, has Pangean affinities and a relative age of vicariance older than the A. spatula group. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 97 Discussion. — This brief statement of gar distributions can be summarized in a relative hierarchy of vicariance events: (1) The vicariance of the two genera probably occurred on Pangea and later dispersal events have obscured the orig- inal vicariance pattern. (2) The vicari- ance event that produced the indicus- osseus species pair is older than the event that produced the present distri- bution of the L. oculatus species group. (3) The L. oculatus species group is a component of a generalized Early Eo- cene track and thus has a minimum age of Early Eocene. (4) The distribution of the A. spatula species group was prob- ably produced by a West Africa-North American vicariance event of Jurassic age, and this track is relatively older than the spatula-tristoecJius track that is a component of a younger North Amer- ican-Caribbean generalized track. The model of gar biogeography pre- sented above predicts that Lepisosteus gars will be found as a component of the African biota. Furthermore, the lack of gars from South America is interesting and it is possible that the vicariant sister group of gars may be found in the South American fossil biota. Acknowledgements To my colleagues and friends at the American Museum of Natural History I express my thanks for many hours of discussion about gars and systematic theory. Special thanks go to Dr. Donn E. Rosen, whose help and encourage- ment throughout the project are grate- fully acknowledged. Thanks also go to Dr. Gareth Nelson, who spent many hours discussing problems of fish phy- logeny and system atics with me and who taught me many techniques used during the project. Dr. Bobb Schaeffer gave me valuable advice and guidance on the fos- sils. Mr. Walter Sorensen helped and guided me in fossil preparation. The section concerning ancestor recognition grew out of a dialogue with my col- league George Engelmann. Ms. Louise LoPresti typed the rough draft, Ms. Lynne Judge and Ms. Joan Mey typed the final draft. Their skills are appre- ciated. My thanks also go to Dr. Colin Pat- terson (British Museum, Natural His- tory) and to M. Daniel Goujet (Museum National d'Histoire Naturelle) for their help during my visit abroad, and to Dr. Darrell D. Hall, Sam Houston State Uni- versity, and his students for assisting with field work in Texas. The following persons and institu- tions lent specimens or accommodated me during visits: Dr. Reeve Bailey ( Mu- seum of Zoology of the University of Michigan); Dr. J. E. Boehlke (Academy of Natural Sciences, Philadelphia); M. Daniel Goujet ( Museum National d'His- toire Natiu-elle); Dr. Brian Gardiner (Queen Ehzabeth College); Dr. Colin Patterson (British Museum, Natural His- tory); Dr. Clayton Ray (U.S. Museum of Natural History); Dr. Gerald Smith (Museum of Paleontology of the Univer- sity of Michigan ) ; Dr. Camm Swift ( Los Angeles County Museum); Dr. Royal Suttkus (Tulane University); and Dr. Stanley Weitzman (U.S. National Mu- seum of Natural History). My studies at the British Museum (Natural History) and the Museum Na- tional d'Histoire Naturelle were made possible through a grant by the Theo- dore Roosevelt Fund, American Museum of Natural History. The American Mu- seum of Natiu-al History is gratefully acknowledged for providing financial and logistical support during my tenure as a graduate student. The City Univer- sity of New York provided partial fi- nancial assistance by a University Fel- lowship and by equipment grants. Summary ( 1 ) The objectives of this study were fourfold: to detennine if the lepiso- steids, or gars, are a monophyletic group, to detennine which of the current the- ories of the relationships of gars to other actinopterygian groups is most highly 98 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY corroborated, to detemiine which nom- inal fossil and recent gar species could be recognized as valid species and what the interrelationships of these species are, and to describe the major features of gar biogeography. (2) Hennig's (1966) phylogenetic method is used to study the first three objectives. Hennig's method is briefly summarized and several points taken up in detail. Croizat's (1958, 1962) bio- geographic method is used for the bio- geographic study. (3) Anatomical features of gars are compared to those of other actinoptery- gian and teleostome groups. The analy- sis concentrates on the skull, visceral arches, pectoral girdle, and postcranial skeleton. At least twenty-seven charac- ters are unique for gars among actino- pterygian fishes and these characters corroborate a hypothesis of monophyly for gars as a group. Gars share seven derived (synapomorphic) characters with halecomorphs (Amia, etc.) and teleosts that they do not share with chondrosteans. These characters corrob- orate a monophyletic Neopterygii. Amia shares thirteen derived characters with teleosts that neither group share with gars, while Amia shares only two char- acters with gars not shared with teleosts that can not be refuted as derived, based on moiphological criteria. Parsimony fa- vors a monophyletic Halecostomi (Amia plus teleosts) rather than a monophyletic Holostei (Arnia plus gars). It is also more parsimonious to consider the se- mionotids as halecostomes rather than as the sister group of gars, because the semionotid Lepidotes shares five derived characters with halecostomes while shar- ing only two loss characters with gars which can not be refuted as synapomor- phies based on morphological criteria. (4) The Division Ginglymodi, the family Lepisosteidae, and the two gen- era Lepisosteus and Atractosteus are diagnosed, and synonymies are provided. Each of the recognized species is diag- nosed, described or redescribed, and briefly compared to other species in its genus. Synonymies of Recent species in- clude only name changes; the synony- mies of fossil species include all liter- ature citations found. Described fossil fonns that cannot be diagnosed to spe- cies are placed at the level their pre- served characters allow (i.e. Lepisosteus sp. indet., etc.). One species, L. opertus, is described as a new species from the Hell Creek Fomiation, Cretaceous, Mon- tana. Other Lepisosteus gars recognized are: L. cuneatus (Eocene, North Amer- ica); L. platostomus (Recent, North America); L. indicus (Cretaceous, In- dia); L. osseus (Recent, North Amer- icaj; L. fimhriatus (Eocene to Oligocene, Europe); L. oculatus (Recent, North America), and L. platyrJiincus (Recent, North America). Eight Atractosteus gars are recognized: A. strausi (Eocene, Eu- rope); A. tropicus (Recent, Middle America); A. simplex (Eocene, North America); A. africanus (Cretaceous, Af- rica); A. occidentalis (Cretaceous, North America); A. atrox (Eocene, North America); A. spatula (Recent, North and Middle America); and A. tristoechus (Cuba and the Isle of Pines). (5) Each genus is monophyletic based on derived characters. (6) Within the genus Lepisosteus, L. opertus is the primitive sister species of all other Lepisosteus gars. Lepiso- steus cuneatus is more primitive than L. platostomus. Lepisosteus platostomus is the sister species of a group composed of the L. osseus-indicus pair and the L. oculatus species group. Within the L. oculatus species group, L. fimhriatus is the sister species of the oculatus-platy- rhincus species pair. (7) Within Atractosteus, A. strausi is the primitive sister species of other Atractosteus gars. Atractosteus tropicus and A. simplex are relatively more prim- itive than the A. spatula species group. Within the A. spatula species group, A. africanus and A. occidentalis are incertae sedis while A. atrox is the sister species of the spatula-tristoechus species pair. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 99 (8) A classification of Recent and fossil gars is presented. (9) Track analysis of both Lepiso- steus and Atractosteus indicate that both genera may have had Pangean distribu- tion and thus may be as old as 180 mil- lion years before present. Within Lepi- sosteus, the track of the osseus-indicus species pair may be older than that of the oculatus species group. The oculatus species group track (Eastern North America to Europe) seems correlated with a generalized Eocene track based on mammalian distributions. Witliin Atractosteus the spatula species group track connects North America and West Africa. This track may predate the for- mation of the Caribbean region. The track of the spatula-tristoechus species pair is correlated with a generalized North American-Caribbean track. The biogeographic model presented here pre- dicts that Lepisosteus gars will be found in the fossil fauna of Africa and that the vicariant sister group of gars may be found among the fossil fauna of South America. Literature Cited Agassiz, a. 1878. The development of Lepi- dosteus. Proc. Amer. Acad. Arts and Sci., 13:65-76. Agassiz, L. 1832. Untersuchungen iiber die fossilen Fische der Lais-Fonnation. Jahrb. Min. Geogn. Geol., 1832:139-149. Agassiz, L. 1834. Remarks on certain points in the anatomy of Lepidosteus. Proc. Zool. Soc. London, 2:119-120. Agassiz, L. 1843. Recherches sur les poissons fossiles. Tome 2, pt. 2. Paris. Agassiz, L. 1848. Monograph of garpikes (abst.). Proc. Amer. Assoc. Adv. Sci., 1848:70-71. Agassiz, L. 1850. Remarks on species belong- ing to die genus Lepidosteus. Proc. Bos- ton Soc. Nat. Hist., 3:182. Allen, B. M. 1911. The origin of the sex cells of Amia and Lepidosteus. J. Morph., 22:1-36. Allen, W. F. 1907. Distribution of the sub- cutaneous vessels in the head region of the ganoids, Polyodon and Lepisdostcus. Proc. Washington Acad. Sci., 9:79-158. Allen, W. F., 1908. Distribution of the sub- cutaneous vessels in tlie tail region of Lepidosteus. Amer. J. Anat., 8:49-88. Allis, E. p. 1889. The anatomy and develop- ment of the lateral line system in Amia calva. J. Morph., 2:463-467. Allis, E. P. 1897. The cranial muscles and cranial and first spinal nerves in Amia calva. J. Morph., 12:487-772. Allis, E. P. 1905. The latero-sensory canals and related bones in fishes. Intemat. Monatsschrift Anat. Physiol., 21:401- 500. Allis, E. P. 1909. The cranial anatomy of the mail-cheeked fishes. Zoologica, Stuttg., 22:1-219. Allis, E. P. 1908. The pseudobranchial and carotid arteries on the gnathostome fishes. Zool. Jahrb. (Abtli. Anat.), 27: 103-134. Allis, E. P. 1911. The homologies of the muscles related to the visceral arches of the gnatliostome fishes. Quart. J. Micro. Sci., 63:303-406. Allis, E. P. 1919. On certain features of the otic region of the chondrocranium of Lepidosteus, and comparison with other fishes and higher vertebrates. Proc. Zool. Soc. London, 1919:245-266. Allis, E. P. 1934. Concerning the course of the latero-sensory canals in recent fishes, prefishes, and Necturus. J. Anat., 68: 361-415. Ameghino, F. 1898. Sinopsis geologico-pale- ontologica. Segundo censo de la Re- publica Argentina. Cap. I. Territorio, Tercera Parte, vol. 1. Buenos Ares. Andreae, a. 1893. Vorliiufige Mittheilung iiber die Ganoiden (Lepidosteus und Amia) des Mainzer Beckens. Verh. Nat. Medic. Ver. Heidelberg., (2)2:1-15. Andreae, A. 1894. Beitrage zur Kenntniss der fossilen Fische des Mainzer Beckens. Abh. Senck. Nat. Ges., 18:351-364. Arambourg, C, Joleaud, L. 1943. Vertebres fossiles du bassin du Niger. Bull. Direc- tion des Mines A.O.F., 7:1-74. AsHLOCK, p. H. 1971. Monophyly and asso- ciated terms. Syst. Zool., 20:63-69. AsHLOCK, P. H. 1973. Monophyly again. Syst. Zool., 21:430-437. AsHLOCK, P. H. 1974. The uses of cladistics. Ann. Rev. Syst. Ecol., 5:81-99. AuMONiER, F. J. 1941. Development of the dermal bones of the skull in Lepidosteus osseus. Quart. J. Micro. Sci., 83:1-33. Balfour, F. N., Parker, W. K. 1882. On the structure and development of Lepido- steus. Phil. Trans. Roy. Soc. London., 173:359-442. 100 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY Ball, I. R. 1975. Nature and formulation of biogeographic hypotheses. Syst. Zool., 24:407-430. Baub, G. 1887. On tlie morphology of ribs. Amer. Nat., 1887:942-945. Baur, G. 1889a. On the moiphology of the vertebrate skull. J. Morph., 3:467-474. Baur, G. 1889b. On the morphology of ribs and tlie fate of the actinosts of the me- dial fins in fishes. J. Morph., 3:463-466. Beard, J. 1889. On the early development of Lepidosteus osseiis (preliminary report). Proc. Royal Soc., 46:108-118. Beard, J . 1895. The pronephros of Lepidosteus osseus. Anat. Anzeiger., 10:198-201. Beard, J. 1896. The yolk-sac, yolk and me- rocytes, in SctjUium and Lepidosteus. Ibid., 12:334-347. Berg, L. S. 1940. Classification of fishes botii recent and fossil. Trav. Inst. Zool. Acad. Sci. URSS, 5:85-517. Berg, L. S. 1955. Classification of fishes and fish-like vertebrates, living and fossil. Second addition, corrected and enlarged. Trudy Zool. Inst. Leningr. 20:1-286. Bertmar, G. 1959. On the ontogeny of the chondral skull in Characidae, with a dis- cussion on the chondrocranial base and the visceral chondrocranium in fishes. Acta Zool., 40:203-364. Bertmar, G. 1967. On the hyoid arch skeleton in fishes. Zool. Rev., 23:45-60. Bjerring, H. 1972. The rhinal bone and its evolutionary significance. Zool. Scripta, 1:193-201. Bjork, p. R. 1967. Latest Eocene vertebrates from Northwestern South Dakota. J. Paleont., 41:227-236. Bleeker, p. 1873. Memoire sur la faune ichthyologique de Chine. Nederlandsch. Tijdschrift f. Dierkunde, 4:113-154. Block, M. E., Schneider, J. B. 1801. Sys- tema ichthyologiae iconibus ex illustra- tum. Post obitum auctoris opus incho- atum absolvit, correxit, interpolavit. Jo. Gottlob Schneider, Soxo. Belolini. Bock, W. J. 1969. The concept of homology. Ann. N.Y. Acad. Sci., 167:71-73. Bock, W. J. 1973. Philosophical foundations of classical evolutionary taxonomy. Syst. Zool., 22:375-392. Bodermer, C. W. 1957. The origin and de- velopment of the extrinsic ocular mus- cles in the gar pike (Lepisosteus osseus). J. Morph., 100:83-111. Brookover, C. 1908. Pinkus's nerve in Amia and Lepidosteus. Science, 27:913-914. Brookover, C. 1914. The development of the olfactory nerve and its associated gan- glion in Lepidosteus. J. Comp. Neurol., 24:113-130. Brough, J. 1931. On the fossil fishes from the Karroo system, with some general con- siderations of tlie bony fishes of the Tri- assic period. Proc. Zool. Soc. London, 1931:235-269. Brundin, L. 1966. Transantarctic relationships and tlieir significance, as evidenced by chironomid midges. Kungl. Svenska Vetenskap. Handl., 11:1-472. Brundin, L. 1968. Application of phylogenetic principles in systematics and evolution- ary theory. In: Nobel Symposium 4, Current problems in lower vertebrate phylogeny (T. 0rvig, ed.). Alguist and Wiksell, Stockholm: 473-495. Carus, J. V. 1875. Handbuch de zoologie. Lster Bd. Wirbeltliiere, Mollusken und Molluscoiden. Leipzig, 1868-1875. Casier, E. 1943. Contributions a I'etude des poissons fossiles la Belgique — Restes de genre Lepidosteus du Landenien conti- nental de la Hesbaye. Bull. Mus. Hist. Nat. Belg., 19:1-12. Casier, E. 1961. Materiaux pour la faune ichthyologique eocretacique du Congo. Annzs. Mus. r. Afr. Cent. (8vo, sci. geol.), 39:1-96. Cockerell, T. D. a. 1908. The fishes of the Rocky Mountain region. Univ. Colorado Stud., 5:159-178. Cockerell, T. D. A. 1909. A fossil gar pike from Utah. Science, 29:796. CoLLESs, D. H. 1967. The phylogenetic fal- lacy. Syst. Zool., 16:289-295. CoLLESs, D. H. 1969a. The phylogenetic fal- lacy revisited. Syst. Zool., 18:115-126. CoLLESS, D. H. 1969b. The interpretation of Hennig's "Phylogenetic Systematics" — a reply to Dr. Schlee. Syst. Zool., 18:134- 144. CoLLiNGE, W. E. 1892. The lateral canal sys- tem of Lepidosteus osseus. Proc. Bir- mingham Phil. Soc, 8:263-272. CoLLiNGE, W. E. 1895. The sensoiy canal of fishes, I. Ganoids. Quart. J. Micro. Sci., 36:499-537. Cope, E. D. 1865. Note on the fishes brought from the Platte River near Ft. Riley, by Dr. W. A. Hammond. Proc. Acad. Nat. Sci. Phil., 1865:85-87. Cope, E. D. 1869. Second addition to the his- tory of the fishes of the Cretaceous of the United States. Proc. Amer. Philos. Soc, 9:240-244. Cope, E. D. 1872. Observations on the sys- tematic relationships of the fishes. Proc. Amer. Assoc. Adv. Sci., 20th Meeting, Indianapolis, 1871:317-343. Cope, E. D. 1873. On the extinct Vertebrata of the Eocene of Wyoming, observed by the expedition of 1872, with notes on the geology. U.S. Geol. Surv. Terr. 1872, Washington, D.C.: 565-649. Cope, E. D. 1874. Report on the vertebrate palaeontology of Coloradb. Ann. Rept. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 101 Gecl. Surv. Terr. Washington, D.G., 1874:427-533. Cope, E. D. 1875. Survey of the Vertebrata whose remains have been preserved in the formations of North Carolina. Rept. Geol. Sui-v. North Carohna 1, appendix B:29-52. Cope, E. D. 1877a. Report on the geology of the region of the Judith River, Montana, and on vertebrate fossils obtained on or near the Missouri River. Bull. U.S. Geol. and Geog. Surv., 3:565-597. Cope, E. D. 1877b. Report on the extinct Vertebrata obtained in New Mexico by parties of the expedition of 1874. Geog. Surv. West one hundredth Meridian, 4, Paleontology. Washington, D.C. Cope, E. D. 1884. The Vertebrata of the Ter- tiary fomiations of the West, Book I. Rept. U.S. Geol. Surv. Terr., 3:1-1099. Cracraft, J. 1974. Phylogenetic models and classification. Syst. Zool., 23:71-90. Cracraft, J. 1975. Paleontology and phyloge- netics: a response to Bretsky. Syst. Zool., 24:119-120. Croizat, L. 1958. Panbiogeography. Pub- lished by tlie author, Caracas. Croizat, L. 1962. Space, time, and form: the biological synthesis. Published by the author, Caracas. Croizat, L., Nelson, G. J., Rosen, D. E. 1974. Centers of origin and related con- cepts. Syst. Zool., 23:265-287. Cross, W. 1896. Post Laramie and Tertiary Geology. Arapahoe formation, Denver formation. Age of Arapahoe and Denver formations. Mon. U.S. Geol. Surv., 27: 151-252. Crowson, R. a. 1970. Classification and Biol- ogy. Heinemann Educational Books, Ltd., London. CuviER, C. L. C. F. D. 1825. Recherches sur les ossemens fossilles, ou Ton rerablit les characters de plusieurs animaux dont les revolutions du globe ont detruit les especes. 3rd ed., Paris. CuviER, C. L. C. F. D. 1836. Ibid., 4th Edi- tion. Daget, J. 1950. Revision des affinities phy- logenetiques de Polypterides. Mem. Inst. fr.Afr. noire, 11:1-178. Dalquest, W. W. 1962. The Good Creek Formation, Pleistocene of Texas, and its fauna. J. Paleont., 36:568-582. Danforth, C. H. 1916. The relation of coro- nary and hepatic arteries in the common ganoids. Amer. J. Anat., 19:391-400. Darlington, P. J. 1970. A practical criticism of Hennig-Brundin "Phylognetic (sic) Systematics" and antarctic biogeography. Syst. Zool., 19:1-18. Dean, B. 1895a. The early development of gai-pike and sturgeon. J. Morph., 11:1- 62. Dean, B. 1895b. Notes on the early develop- ment of the ganoids, Lepidosteus, Aci- penser, and A77iia. Rept. British Assoc. Adv. Sci., 65:734. Dean, B. 1896a. On the larval development of A7nia calva. Zool. Jahrbiicheren, 1895: 639-672. Dean, B. 1896b. The early development of Amia. Quart. J. Micro. Sci., 38:413-444. De Beer, G. B. 1926. Studies on the verte- brate head II: the orbitotemporal region of the skull. Quart. J. Micro. Sci., 20: 263-370. De Beer, G. B. 1937. The development of the vertebrate skull. Clarendon Press, Ox- ford. DiETZ, P. A. 1912. Vergelijkende anatomic van de kaak-en kieuw-boogspieren der Tele- ostei. Leiden. DiETZ, P. A. 1914. Beitrage zur kenntnis der Kierfer- und Kiemenbogenmuskulatur der Teleostiere I. Mitt. Zool. Stat. Neapel, 22:99-162. DiETZ, P. A. 1921. t)ber die systematische Stellungder Gadidae. Zugleich Nr. 2 der "Beitrage zur kenntnis der Kiefer- und Kiemenbogenmuskulatur der Teleo- stierre." Mitt. Zool. Stat. Neapel, 22: 433-457. De Kay, J. E. 1842. Zoology of New York or the New York Fauna. Part 4, Fishes. White and Visscher, Albany, N.Y. DoLLO, L. 1893. Sur le Lepidosteus suesso- niensis. Bull. Scient. France Belgique, 25:193-197. DuMERiL, A. 1870. Histoire Naturelle des pois- sons ou ichthyologie generale. Tome II: Ganoides, Dipnes, Lophob ranches. Paris. Eastman, C. R. 1900a. Fossil lepisosteids from the Green River shales of Wyoming. Bull. Mus. Comp. Zool., 36:67-75. Eastman, C. R. 1900b. New fossil bird and fish remains from the Middle Eocene of Wyoming. Geol. Mag., 4:54-58. Eddy, S. 1957. How to Know the Freshwater Fishes. Wm. C. Brown Co., Dubuque. Edgeworth, F. H. 1911. On the morphology of the cranial muscles in some verte- brates. Quart. J. Micro Sci., 56:167-316. Edgeworth, F. H. 1928. The development of some of the cranial muscles of ganoid fishes. Phil. Trans. Roy. Soc. London, (B)217:l-89. Edgeworth, F. H. 1935. The Cranial Muscles of Vertebrates. Cambridge. Emmons, E. 1858. Report of the North Caro- lina Geological Survey. Agriculture of the eastern counties, together with de- scriptions of the fossils of the marl beds. Raleigh. 102 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY EsTEs, R. 1964. Fossil vertebrates from the Late Cretaceous Lance fonnation, East- em Wyoming. Univ. California Publ. Geol. Sci., 49:1-180. EsTES, R. 1969. Lower vertebrates from the Late Cretaceous Hell Creek formation, McCone County, Montana. Breviora, 337:1-37. Farris, J. S., Kluge, a. G., Eckhart, M. J. 1970. A numerical approach to phylo- genetic systematics. Syst. Zool., 19:172- 189. FiTziNGER, L. J. F. J. 1873. Versuch einer natiirlichen Classification der Fische. Sitzber. Akad. Wiss. Wien., Math.-Nat. CI., 67:5-58. Forbes, S. A., Richardson, R. E. 1920. The fishes of Illinois, 2nd edition. Illinois Nat. Hist. Surv. Circ. 3:1-357. Fowler, H. W. 1910. Notes on chimaeroid and ganoid fishes. Proc. Acad. Nat. Sci. Phila., 62:602-612. Fowler, H. W. 1911. A description of the fossil fish remains of the Cretaceous, Eo- cene, and Miocene formations of New Jersey. Bull. Geol. Surv. N.J., 4:22-182. Gardiner, G. B. 1960. A revision of certain actinopterygian and coelacanth fishes, chiefly from the lower Lias. Bull. Br. Mus. Nat. Hist. (Geol.), 4:239-384. Gardiner, G. B. 1963. Certain palaeoniscoid fishes and the evolution of the snout in actinopterygians. Bull. Br. Mus. Nat. Hist. (Geol.), 8:255-325. Gardiner, G. B. 1967. The significance of the preopercular in actinopterygian evolu- tion. J. Linn. Soc. (Zool.), 47:197-209. Gardiner, G. B. 1973. Interrelationships of teleostomes. In: Interrelationships of fishes (Greenwood, Miles, and Patterson, eds.). J. Linn. Soc. (Zool.), 53, suppl. 1:105-1.35. Garrett, F. S. 1942. The development and phylogeny of the capsule of Stannius in ganoid and teleostean fishes. J. Morph., 70:41-67. Gegnebaur, C. 1867. Ueber die Entwickelung er Wirbelsaule des Lepidosteus mit ver- gleichend-anatomischen Bemerkungen. Jena Zeitschr. Naturev., 3:359-419. Gervais, P. 1852. Zoologie et paleontologie francais (animaux vertebres), 2 (first edition). Paris. Gervais, P. 1859. Ihid. (2nd edition.) Gervais, P. 1874a. Presence du genre Lepi- dostee parmi les fossiles du bassin de Paris. C. R. Acad. Sci. Paris, 79:844- 846. Gervais, P. 1874b. Ihid. J. Zool. Paris, 3: 457-461. Ghiselin, M. T. 1974. A radical solution to the species problem. Syst. Zool., 23: 536-544. Gidley, J. W. 1915. An extinct marsupial from the Fort Union, with notes on the Myr- mecobidae and other families of this group. Proc. U.S. Nat. Mus., 48:359- 402. Gidley, J. W. 1927. Preliminary list of species of Pleistocene vertebrates found at Mel- bourne, Florida. Publ. Carnegie Inst. Wash., No. 322B:274. Gill, E. L. 1923. The Permian fishes of the genus Acentrophonis. Proc. Zool. Soc. London, 1923:19-40. Gill, T. 1863. Descriptive enumeration of a collection of fishes from the western coast of Central America, presented to the Smithsonian Institution by Capt. John M. Dow. Proc. Acad. Nat. Sci. Phila., 1863:162-180. GiLMORE, C. W. 1916. Contributions to the geology and paleontology of San Juan County, New Mexico. 2: Vertebrate faunas of the Ojo, Alamo, Kirtland and Fruitland formations. U.S. Geol. Surv. Prof. Pap., 98:279-308. GiLMORE, C. W. 1920. Reptilian faunas of the Torrejon, Puerco, and underlying Upper Cretaceous formations of San Juan County, New Mexico. U.S. Geol. Surv. Prof. Pap., 119:1-68. GiLMORE, C. W. 1924. Report on a collection of vertebrates from Wood Mountain, southern Saskatchewan. Canad. Geol. Surv. Dept. Mines, Bull. 38 (Geol. Se- ries 43): 27-28. GiRARD, C. F. 1858. Fishes of North America, observed on a survey for a railroad route from the Mississippi River to the Pacific Ocean. Washington, D.C. Gmelin, J. F. 1789. In: Linnaeus, 1789. Sys- tema Naturae, Edito decime tertia, To- mus I(3):1126-1516. Goodrich, E. S. 1904. On the dermal fin-rays of fishes — living and extinct. Quart. J. Micro. Sci., 47:465-522. Goodrich, E. S. 1909. Part IX, Vertebrata Craniata (First Fascicle: Cyclostomes and Fishes). In: A Treatise on Zoology (Lankester, ed.). Adam and Charles Black, London. Goodrich, E. S. 1930. Studies on the Struc- ture and Development of Vertebrates. Macmillan, London. GosLiNE, W. A. 1965. Teleostean phylogeny. Copeia, 1965:186-194. Greenwood, P. H., Rosen, D. E., Weitzman, S. H., Myers, G. S. 1966. Phyletic studies of teleostean fishes, with a pro- visional classification of living foims. Bull. Amer. Mus. Nat. Hist., 131:339- 456. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 103 Gregory, W. K. 1933. Fish skulls: a study of the evolution of natural mechanisms. Trans. Amer. Phil. Soc, 23:1-481. Griffiths, G. C. D. 1972. The phylogenetic classification of Diptera Cyclorrhapha. Ser. EntomoL, 8:1-340. Griffiths, G. C. D. 1973. Some fundamental problems in biological classification. Syst. Zool., 22:338-343. Gronow, L. T. 1854. Catalogue of fish col- lected and described by Lawrence Theo- dore Gronow, new in the British Mu- seum. Edited from the manuscript by Albert Giinther. London. Gunther, a. 1870. Catalogue of the fishes in the British Museum, (vol. 8). Taylor and Francis, London. Haines, R. W. 1942. Eudiarthrodial joints in fishes. J. Anat., London, 77:12-19. Hammarberg, F. 1937. Zur Kenntnis der on- togenetischen Entwicklung des Schadels von Lepidosteus platijstomus. Acta Zool., 18:210-337. Hammett, F. S., Hammett, D. W. 1939. Proportional length growth of gar {Lepi- dosteus platyrhincus De Kay). Growth, 3:197-209. Hatcher, J. B. 1905. Vertebrate fauna of the Judith River Beds. Bull. U.S. Geol. Surv., 257:67-103. Hay, O. p. 1898. Classification of the amiid and lepisosteid fishes. Amer. Nat., 32: 341-349. Hay, O. p. 1902. Bibliography and catalogue of the fossil Vertebrata of North Amer- ica. Bull. U.S. Geol. Surv., 179:1-868. Hay, O. p. 1903. On some recent literature bearing on the Laramie formation. Amer. Geol., 32:115-120. Hay, O. p. 1910. Where do the Lance Creek ("ceratops") beds belong, the Cretaceous or in the Tertian/? Proc. Indiana Acad. Sci., 25:277-303.' Hay, O. p. 1917. Vertebrata mostly from stratum No. 3, at Vero, Florida, together with descriptions of new species. Kept. Florida Geol. Surv., 9:43-68. Hay, O. p. 1919. Descriptions of some mam- malian and fish remains from Florida of probably Pleistocene Age. Proc. U.S. Nat. Mus., 56:103-112. Hay, O. p. 1923. The Pleistocene of North America and its vertebrate animals from the states east of the Mississippi River and from the Canadian Provinces east of longitude 95°. Carnegie Inst. Wash., Publ. No. 322:1-499. Hay, O. p. 1926. A collection of Pleistocene vertebrates from southwestern Texas. Proc. U.S. Nat. Mus. 68:1-18. Hay, O. p. 1927. The Pleistocene of the west- em region of North America and its vertebrate animals. Carnegie Inst. Wash, Publ. No. 322B: 1-346. Hay, O. p. 1929. Second bibliography and catalogue of the fossil vertebrates of North America, vol. I. Carnegie Inst. Wash. Hennig, W. 1950. Grundziinge einer Theorie der phylogenetischen Systematik. Deut- scher Zentralverlag, Berlin. Hennig, W. 1966. Phylogenetic Systematics. Univ. Illinois Press, Urbana. Hennig, W. 1975. "Cladistic analysis or cla- distic classification?": a reply to Ernst Mayr. Syst. Zool., 24:244-256. HiBBARD, C. W., Dalquest, W. W. 1966. Fos- sils from the Seymour formation of Knox and Baylor counties, Texas and their bearing on the late Kansan climate of that region. Contrib. Mus. Paleont. Univ. Mich., 21:1-66. Hill, L. G., Schnell, G. D., Echelle, A. A. 1973. Effect of dissolved oxygen con- centration on locomotory reactions of the spotted gar, Lepisosteus oculatus (Pisces, Lepisosteidae). Copeia, 1973:119-124. Hoeven, J. VAN DER. 1841. Uebcr die zellige Schwimmblase des Lepidosteus. Arch. Anat., 1841:221-223. Holmgren, N., Stensio, E. 1936. Kranium und Visceralskelett der Akranier, Cyclo- stomen und Fische. In: Handb. der vergleichenden Anat. der Wirbeltiere (Bolk, Coppert, Kallius, Labosch, eds.), 4:233-500. HuBBS, C. L., Lagler, K. F. 1943. Annotated checklist of the fishes of Foots Pond, Gibson Co., Indiana. Indiana Lakes and Streams, 2:73-83. HussAKOF, L. 1908. Catalogue of the types and figured specimens of fossil verte- brates in the American Museum of Nat- ural History. Part I: Fishes. Bull. Amer. Mus. Nat. Hist., 25:1-103. HussAKOF, L., Bryant, W. L. 1919. Catalogue of the fossil fishes in the museum of the Buffalo Society of Natural Sciences. Bull. Buffalo Soc. Nat. Sci., 12:1-346. Huxley, T. H. 1861. Preliminary essay upon the systematic arrangement of the fishes of the Devonian epoch. Mem. Geol. Surv. UK, Decade, 10:1-40. Hyrtl, C. J. 1851. Ueber das Arterien-System des Lepidosteus. Sitzber. Akad. Wiss. Wien, 8:234-241. Hyrtl, C. J- 1825a. Ueber die Schwimmlase des Lepidosteus osseus. Sitzber. Akad. Wiss. Wien, 8:71. Hyrtl, C. J. 1852b. Uber die Pori abdomi- nales die Kiemenarterien, und die Glan- dula thyroidea der Ganoiden. Sitzber. Akad. Wiss. Wien, 8:179-185. Jackson, C. T. 1856. On the composition of 104 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY the scales of the garpike. Proc. Boston Soc. Nat. Hist., 5:92. Jarvik, E. 1944. On the dermal bones, sensory canals and pit-lines of the skull of Eusthenopteron foordi. K. Vet. Akad. Handl. Stockholm, (3)21:1-48. Jarvik, E. 1948. On the morphology and tax- onomy of the Middle Devonian osteo- lepid fishes of Scotland. K. Vet. Akad. Handl. Stockholm, (3)25:1-301. Jarvik, E. 1954. On tlie visceral skeleton in Eusthenopteron with a discussion of the parasphenoid and palatoquadrate in fishes. Kunge. Sven. Vet. Handl., (4)5: 1-104. Jessen, H. 1968. Moythomasia nitida Gross und M. cf. striata Gross Devonische Palaeonisciden aus dem Oberen Platten- balk Mulde Bergish-Gladback-PafFrather (Rlieinisches Schiefergebiege ) . Palaeon- tographica, 128A:87-121. Jessen, H. 1972. Schultergiirtel und Pectoral- flosse bei Actinopterygiem. Fossils and Strata, 1:1-101. Jessen, H. 1973. Interrelationships of actino- pterygians and brachiopterygians: evi- dence from pectoral anatomy. In: In- terrelationships of fishes ( Greenwood, Miles, and Patterson, eds.). J. Linn. Soc. (Zool.), 53, supp. 1:227-232. JOLLiE, M. 1962. Chordate morphology. Rlieinhold, New York. JoLLiE, M. 1969. Sensory canals of tire snout of actinopterygian fishes. Trans. Illinois St. Acad. Sci'., 62:61-69. Jordan, D. S. 1877. Review of Rafinesque's memoirs on North American fishes. Con- tribution to North American ichthyology I. Washington, D.C.:l-53. Jordan, D. S. 1885. A catalogue of the fishes known to inhabit the waters of North America, north of the Tropic of Cancer, with notes on the species discovered in 1883 and 1884. Ann. Rept. Comm. Fish and Fisheries, 1884:1-185. Jordan, D. S. 1905. A Guide to the Study of Fishes. Henry Holt and Co., N.Y. Jordan, D. S., Evermann, B. W. 1896. The fishes of North and Middle America: a descriptive catalogue of the species of fish-like vertebrates found in the waters of North America, north of tlie Isthmus of Panama. Bull, U.S. Nat. Mus., 147: 1-3313. Jordan, D. S., Evermann, B. W., Clark, H. W. 1930. Checklist of tlie fishes and fish-like vertebrates of North and Middle America nortli of the northern boundary of Venezuela and Columbia. Rept. U.S. Comm. Fish., 1928, part 2: 1-670. Jordan, D. S., Gilbert, C. H. 1883. A synop- sis of the fishes of North America. Bull. U.S. Nat. Mus., 16:1-1018. Keenlyside, D. L., Stewart, F. L., Boucher- White, N. 1947. Notes on the pre- historic distribution of long nose gar, Lepisosteus ossetis, in Lake Erie. Ca- nadian Field Naturalist, 88:239. Kerr, T. 1952. The scales of primitive living actinopterygians. Proc. Zool. Soc. Lon- don, 122:55-78. Kerr, T. 1967. Development and structure of some actinopterygian and urodele teetli. Proc. Zool. Soc. London, 133:401-422. King, C. 1878. Systematic geology. U.S. Geol. Explor. 40th Parallel, Washington, D.C. Kingsbury, B. F. 1897. The endocephalic in- vaginations in ganoids. J. Comp. Neurol., 7:37-44. Kinkelin, G. F. 1884. Die Schleusen Kammer von Frankfurt-Niederrad und ilire Fauna. Ber. Senckenberg. Naturforsch. Ges., 1884:219-257. Kinkelin, G. F. 1903. Die Originale der palaontologischen Sammlung im Senck- enbergischen Museum und die auf die- selben bezungliche Literatur. Ber Senck- enberg. Naturforsch. Ges., 1902-1903: 3-88. Koenig, C. 1825. Icones Fossilium Sectiles. London. Kullin, B. 1950. Contributions to the ontog- eny of the everted forebrain. K. Fusiogr. Siillsk. Lund. Forh., 20:87-110. Lacepede, B. G. E. 1803. Historic naturelle des poissons, vol. .5. Paris. Lambe, L. M. 1902. On Vertebrata of the Mid-Cretaceous of the Northwest Terri- tory. 2. New genera and species from the Belly River series (Mid-Cretaceous). Contrib. Canad. Pal., 3:25-81. Lambe, L. M. 1904. Progress of vertebrate paleontology in Canada. Trans. Roy. Soc. Canada, 2:13-56. Lambe, L. M. 1907. (Letter to R. S. Lull on fossils collected on the Red Deer River, Alberta.) J. Mon. U.S. Geol. Survey., 49: 178-179. Lambe, L. M. 1908. The Vertebrata of the Oligocene of tlie Cypress Hills, Sas- katchewan. Contrib. Canad. Pal., 3:1- 65. Landcare, F. L., Conger, A. C. 1913. The origin of the lateral hne primordia in Lepisosteus osseus. J. Comp. Neurol., 23:575-633. Landodt, J. C, Hill, L. G. 1975. Observa- tions of the gross structure and dimen- sions of the gills of three species of gars. Copeia, 1975:470-475. Lehman, J. P. 1966. Actinopterygii. In: Traite de Falcontologie 4, \'ol. 3 (J. Piveteau, ed. ). Masson, Paris: 1-242. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 105 Leidy, J. 1856a. Notice of remains of extinct reptiles and fishes, discovered by Dr. F. V. Hayden in the Bad Lands of the Judith River, Nebraska Territory. Amer. J. Sci., 2:118-120. Leidy, J. 1856b. Notice of remains of extinct reptiles and fishes, discovered by Dr. F. y. Hayden in the Bad Lands of the Judith River, Nebraska Territory. Proc. Acad. Nat. Sci. Phil., 8:72-73. Leidy, J. 1873a. Notice of remains of fishes in the Bridger Tertiary formation of Wy- oming. Proc. Acad. Nat. Sci. Phila., 1873:97-99. Leidy, J. 1873b. Contributions to the extinct fauna of the Western Territories. Rept. U.S. Geol. Sur. Terr., 1:14-358. Leriche, M. 1900. Fauna iclithyologique des sables a Unios et Teredines des environs d'Epernay (Mame). Ann. Soc. Geol. du Nord, 29:173-196. Leriche, M. 1902. Les poissons tertiaires de la Belgique, I. Les poissons paleocenes de la Belgique. Mem. Mus. Roy. Hist. Nat. Belgique, 2:1-48. Leriche, M. 1907. Sur la faune iclithyologique et sur I'age des faluns de Pourcy (Mame). C. R. Acad. Sci., Paris, 145: 442-444. Leriche, M. 1923. Les poissons paleocenes et eocenes du Bassin de Paris (note addi- tional). Bull. Soc. Geol. France, (4)22: 177-200. Leriche, M. 1932. Les poissons eocenes du Bassin de Paris ( Deuxieme note addi- tionnelle). Bull. Soc. Geol. France, (5)2: 357-374. LiNDAHL, E. 1944. Zur Kenntnis der Entwick- lung von Haftorgan und Hypophyse bei Lepidosteus. Acta. Zool., 25:97-133. Linnaeus, C. 1758. Systema naturae sive regna tria naturae, systematice proposita per classes, ordines, genera et species, cum characteribus, differentiis, synony- mis, locis, etc. Tomi i. Holmiae. LooMis, F. B. 1907. Origin of the Wasatch deposits. Amer. J. Sci., 23:356-364. Lundberg, J. G. 1967. Pleistocene fishes of the Good Creek fomiation, Texas. Copeia, 1967:453-455. McAllister, D. E. 1968. Evolution of bran- chiostegals of teleostome fishes. Nat. Mus. Canada Bull., 121:1-239. Macallum, a. B. 1886. The alimentary canal and pancreas of Acipenser, Amia, and Lepisosteus. J. Anat. Physiol., 20:604- 636. McCormack, B. 1967. Aerial respiration in the Florida spotted gar. Quart. J. Flor- ida Acad. Sci., 30:68-72. McDaniel, B., Castro, S. 1966. The structure of tlie liver of the Alligator Gar Lepi- sosteus spatula (Lacepede). J. West Nat., 11:303-305. McKenna, M. C. 1975. Fossil mammals and Early Eocene North Atlantic land con- tinuity. Ann. Missouri Bot. Gardens, 62:. 335-353. Mark, E. L. 1890. Studies on Lepidosteus. Part 1. Bull. Mus. Comp. Zool., 19:1- 128. Marsh, O. C. 1871. (Communication on some new reptiles and fishes from the Cre- taceous and Tertiary. ) Proc. Acad. Nat. Sci. Phila., 1871:103-105. Matsubara, K. 1955. Fish Morphology and Heirarchy. Pts. I-III. Ishizaki-Shoten, Tokyo. Matthew, W. D. 1916. A marsupial from the Belly River Cretaceous. With critical observations upon the affinities of the Cretaceous Mammals. Bull. Am. Mus. Nat. Hist., .38:315-341. Mayhew, R. L. 1924. The skull of Lepisosteus platostomus. J. Morph., 38:315-341. Mayr, E. 1969. Principles of Systematic Zool- ogy. McGraw Hill Book Co., New York. Mayr, E. 1974. Cladistic analysis or cladistic classification? Z. Zool. Syst. Evolut.- forsch., 12:94-128. Merpull, G. p. 1907. Catalogue of types, co- types and figured specimens of fossil vertebrates in the Department of Geol- ogy, U.S. National Museum. Bull. U.S. Nat. Mus., 53:1-81. Miles, R. S. 1968. Jaw articulation and sus- pensorium in Acanthodes and their sig- nificance. Nobel Sj'mposium 4, current problems in lower vertebrate phylogeny (T. 0rvig, ed. ). Alguist and Wikell, Stockholm: 109-127. Miles, R. S. 1973. Relationships of acan- thodians. In: Interrelationships of fishes (Greenwood, Miles, and Patterson, eds.). J. Linn. Soc. (Zool.), 53, supp. 1:63- i03. Miller, R. R. 1954. The southern limit of the garfishes, family Lepisosteidae. Copeia, 1954:230-231. MiLLOT, J., Anthony, J. 1958. Anatomic de Latimeria chalumnae. Vol. I, squelette et muscles et formations de soutien. Centre Nat. Rech. Sci., Paris. Millot, J., Anthony, J. 1965. Anatomic de Latimeria chalumnae 2. Systeme ner- veux et organes de sens. Centre Nat. Rech. Sci., Paris. Moore, G. A. 1957. Fishes. In: Vertebrates of the United States (Blair, Blair, Brod- korb, Cagle, and Moore, eds.). McGraw Hill Book Co., New York. Moore, G. A. 1968. Ibid, 2nd Edition. Moore, G. A., Trautmax, M. B., Curd, M. R., 1973. A description of post-larval gar 106 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY (Lepisosteus spatula Lacepede, Lepiso- steidae) with a list of associated species from tlie Red River, Choctaw County, Oklahoma. Southwest. Nat., 18:343-344. MiJLLER, F. W. 1897. Ueber die Entwicklung und morphologische Bedeutung der Pseudobranchie und ihrer Umgebung bei Lepisosteus osseus. Arch. Mirk. Anat., 49:463-502. MiJLLER, J. 1844. Ueber den Bau und die Grenzen der Ganoiden und iiber des naturliche System der Fische. Phys. Math. Abn. Akad. Wiss. Berlin, 1844: 117-216. MiJLLER, J. 1844. IbicL Ber. Akad. Wiss. Ber- lin, 1844:416-422. MiJLLER, J. 1845. Ibid. Arch. Naturges, 11:91- 141. MiJNSTER, G. VON. 1846. Ueber einen in den tertiaren Ablagerungen der Gegend von Magdeburg gefundenen neuen Fisch- zahn. Beitr. Petrefacten-kunde, 7:34-35. Neff, N. a. 1975. Fishes of the Kanopolis Local Fauna (Pleistocene) of Ellsworth Couny, Kansas. Univ. Michigan Pap. Paleont. No. 12. Studies on Cenozoic Paleontology and Stratigraphy in honor of Claude W. Hibbard: 39-48. Nelson, G. J. 1966. Gill arches of teleostean fishes of tlie order Anguilliformes. Pa- cific Sci., 20:391-408. Nelson, G. J. 1967. Branchial muscles in some generalized teleostean fishes. Acta. Zool., 48:277-288. Nelson, G. J. 1969a. Gill arches and the phy- logeny of fishes with notes on the classi- fication of vertebrates. Bull. Amer. Mus. Nat. Hist, 141:475-552. Nelson, G. J. 1969b. Origin and diversifica- tion of teleostean fishes. Ann. New York Acad. Sci., 167:18-30. Nelson, G. J. 1969c. Infraorbital bones and their bearing on the phylogeny and ge- ography of the osteoglossomorph fishes. Amer. Mus. Novit., 2394:1-37. Nelson, G. J. 1970a. Outline of a theory of comparative biology. Syst. Zool., 19: 373-384. Nelson, G. J. 1970b. Pharyngeal denticles (placoid scales) of sharks with notes on the dermal skeleton of vertebrates. Amer. Mus. Nov., 2415:1-26. Nelson, G. J. 1971. "Cladism" as a philos- ophy of classification. Syst. Zool., 20: 471-472. Nelson, G. J. 1972a. Phylogenetic relationship and classification. Syst. Zool., 21:227- 231. Nelson, G. J. 1972b. Comments on Hennig's "Phylogenetic Systematics" and its in- fluence on ichtliyology. Syst. Zool., 21: 364-374. Nelson, G. J. 1973a. Classification as an ex- pression of phylogenetic relationships. Syst. Zool, 22:344-359. Nelson, G. J. 1973b. Relationships of clupe- omorphs, with remarks on the structure of the lower jaw in fishes. In: Inter- relationships of fishes (Greenwood, Miles, and Patterson, eds.). J. Linn. Soc. (Zool.), 53, supp. 1:333-349. Nelson, G. J. 1973c. Notes on the structure and relationships of certain Cretaceous and Eocene teleostean fishes. Amer. Mus. Nov. 2524:1-31. Nelson, G. J. 1974. Darwin-Hennig Classi- fication: a reply to Ernst Mayr. Syst. Zool., 23:452-458. Nickerson, W. S. 1893. The development of the scales of Lepisosteus. Bull. Mus. Comp. Zool., 24:1-139. Nielsen, E. 1942. Studies on Triassic fishes from East Greenland. I. Glaucolepis and Boreosomus. Palaeozool. Greenlandica, i. K0benhaven: 1-394. Nielsen, E. 1949. Studies on Triassic fishes from East Greenland. II. Australosomus and Birgeria. Middr. Gr0nland, 146:1- 309. OsBORN, H. F. 1902. On vertebrata of the Mid-Cretaceous of the Northwest Terri- tory. I: Distinctive characters of the Mid-Cretaceous fauna. Contrib. Canad. Pal., 3:1-21. Packard, A. S. 1859. The structure of tlie swimming bladder of the gar pike Lepi- dosteus. Proc. Acad. Nat. Sci. Phil., 11. Panchen, a. L. 1967. The nostrils of choanate fishes and early tetrapods. Bio. Rev. Cambridge, 42:374-420. Parker, W. K. 1882. On the development of the skull in Lepidosteus osseus. Proc. Royal. Soc. London, 33:107-112. Parrington, F. R. 1956. The patterns of der- mal bones in primitive vertebrates. Proc. Zool. Soc. London, 127:389-411. Pascaul, R. 1970. Evolution de communi- dades, cambios faunisticos y intergra- ciones biocinoticas de Argentina. Actas IV Congreso Latinoamericano de Zo- ologia. Vol. II. Caracas 10-16 de No- vembre de 1968:991-1088. Patterson, C. 1973. Interrelations of holo- steans. In: Interrelationships of fishes (Greenwood, Miles, and Patterson, eds.). J. Linn. Soc. (Zool.), 53, supp. 1:233- 305. Patterson, C. 1975. The braincase of pholi- dophorid and leptolepid fishes, with a review of the actinopterygian braincase. Philo. Trans. Royal Soc. London, 269: 275-579 (No. 899). Patterson, C, Rosen, D. E. (In Press.) Re- view of ichthyodectiform and other THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 107 Mesozoic fishes and the theory and prac- tice of classifying fossils. Bull. Amer. Mus. Nat. Hist. Peale, A. C. 1912. On the stratigraphic posi- tion and age of the Judith Ri\'er fomia- tion. J. Geol., 10:530-549, 640-652, 738- 757. Pehrson, T. 1940. Development of dermal bones in the skull of Ariiia calva. Act. Zool., 21:1-50. PoEY Y Aloy, F. 1851. Memorias Sobre la Historia Natural de la Isla dc Cuba, Acompanadas de Sumarios Latinos tj Ex- tractos en Frances. Tome I, Habana. Popper, K. R. 1968a. The Logic of Scientific Discovery. Harper Torchbooks, Inc., New York. Popper, K. R. 1968b. Conjectures and Refu- tations: The Growth of Scientific Knowledge. Harper Torchbooks, Inc., New York. Potter, G. E. 1927. Respiratory function of the swim bladder in Lepisosteus. J. Exp. Zool., 49:45-67. Priem, F. 1901. Sur les poissons de I'Eocene inferier des environs de Reims. Bull. Soc. Geol. France, (4)1:477-504. Priem, F. 1908. fitude des poissons fossiles du basin parisien. Me. Ann. Paleont., 1908: 1-144. Querternus, C. 1967. A key to the North American species of Lepisosteus (Class Pisces) based on the cleithrum. Trans. Illinois St. Acad. Sci., 60:45-48. Rafinesque, C. S. 1818a. Museum of natural history. Further discoveries in natural history made during a journey through the western region of the United States. Am. Monthly Mag. Grit. Rev., 3:445- 447. Rafinesque, C. S. 1818b. Description of three new genera of fluviatile fish, Pomoxis, Sarchirus and Exoglossum. J. Acad. Nat. Sci. Phila. (N. S.), 1:417-422. Rafinesque, C. S. 1820. Ichthyolgia ohien- sis, or natural history of the fishes in- habiting the river Ohio and its tributary streams, proceeded by a physical de- scription of the Ohio and its branches. Lexington, Kentucky. Rahn, K., Howell, K. B., Rahn, B. J., Gans, C., Tenney, S. M. 1971. Air breathing of the garfish ( Lepisosteus osseus ) : Resp. Physiol., 11:285-307. Rayner, D. H. 1941. The structure and evo- lution of tlie holostean fishes. Biol. Rev., 16:218-237. Rayner, D. H. 1948. The structure of certain Jurassic holostean fishes with special ref- erence to their neurocrania. Phil. Trans. Roy. Soc. London, 233B:287-345. Reeside, J. B. 1924. Upper Cretaceous and Tertiary formations of the western part of the San Juan basin, Colorado and New Mexico. U.S. Geol. Surv. Prof. Pap., 134:1-70. Regan, C. T. 1923. The skeleton of Lepido- steus with remarks on tlie origin and evolution of the lower neopterygian fishes. Proc. Zool. Soc. London, 1923: 445-461. Reinghard, J., Phelphs, J. 1908. The devel- opment of the adhesive organ and head of A7nia. J. Morph., 19:469-496. Reis, O. M. 1887. Ueber Belonostomus, Aspi- dorhijnchus und ihre Beziehungen zum lebenden Lepidosteus. Sitzbr. K. Bazer. Akad. Wiss., Math. Phys., 17:151-177. Reissner, E. 1859. Ueber die Schuppen von Pohjpterus und Lepidosteus. Arch. Anat. Physiol. (Reichert), 1859:254-269. Renfro, J. L., Hill, L. G. 1970. Factors in- fluencing aerial breathing and metabo- lism of gars (Lepisosteus). Southwest. Nat., 15:45-54. Richardson, J. 1836. Fauna Boreali-Ameri- cana; or, the zoology of the northern parts of British America, containing de- scriptions of the objects of natural his- tory collected on the later northern land expedition, under command of Sir John Franklin, R.N. vol. 3, Fauna Boreal, Amer. London and Norwich. RoMER, A. S. 1966. Vertebrate Paleontology. Univ. Chicago Press, Chicago. Rosen, D. E. 1973. Interrelationships of eu- teleosts. In: Interrelationships of fishes (Greenwood, Miles, and Patterson, eds.). J. Linn. Soc. (Zool.), 53, supp. 1:397- 513. Rosen, D. E. 1974a. Space, time, form: the biological synthesis. [A review of Leon Croizat.] Syst Zool., 23:288-290. Rosen, D. E. 1974b. Cladism or gradism?: a reply to Ernst Mayr. Syst. Zool., 23: 446-451. Rosen, D. E. 1975. A vicariance model of Caribbean biogeography. Syst. Zool., 24:431-464. Russel, L. S. 1935. Fauna of die upper Milk River beds, southern Alberta. Trans. Roy. Soc. Canada, 29 (sect. 4): 115-127. RuTOT, A. 1884. Quelques mots sur les nou- velles decouvertes d'Erquelinnes. Ann. Soc. malae. Belgique, 19:xv. Sauvage, H. E. 1897. Note sur les lepidos- teides du terrain garumnien du Portugal. Bull. Soc. Geol. France, 25:92-96. Sauvage, H. E. 1901. Abstract of Eastman (1900). Rev. Grit. Paleozool., 5:80. Schaeffer, B. 1967a. Osteichthyan vertebrae. J. Linn. Soc. (Zool.), 47:185-195. Schaeffer, B. 1967b. Comments on elasmo- branch evolution. In: Sharks, Skates, 108 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY and Rays ( Gilbert, Mathewson, and Rail, eds. ), Johns Hopkins Press, Baltimore: 3-35. ScHAEFFER, B. 1971. The braincase of the holostean fish Macrepisiius, with com- ments on neurocranial ossification in the Actinopterygii. Amer. Mus. Novit., 2459:1-34. ScHAEFFER, B. 1973. Interrelationships of chondrosteans. In: Interrelationships of fishes (Greenwood, Miles, and Patterson, eds.). T- Linn. Soc. (Zool.), 53, supp. 1:207-226. ScHAEFFER, B., RosEN, D. E. 1961. Major adaptive levels in the evolution of the actinopterygian feeding mechanism. Amer. Zool., 1:187-204. ScHLEE, D. 1969. Hennig's principle of phylo- genetic systematics, an "Intuitive Sta- tistico-phenetic Taxonomy"? Syst. Zool., 18:127-134. ScHLEE, D. 1971. Die Rekonstrucktion der Phylogenese mit Hennig's Prinzip. Aufs. Reden Senckenberg naturf. Ges., 20:1- 62. ScHLOssER, M. 1901. Review of Eastman. C. R. Neues Jahrb. Min. Geol. Pal., 2:480. ScHULTz, G. E. 1965. Pleistocene vertebrates from the Butler Spring local fauna, Meade Count v, Kansas. Pap. Michigan Acad. Sci., 50': 235-265. ScupiN, H. 1896. Verleichende Studien iiber die Histologic der Ganoidschuppen. Arch. Naturgesch., 1896, 62, Tahrg:145- 186. Sewertzoff, a. N. 1934. Evolution der Bauchflossen der Fische. Zool. Jahbr. Abt. Anat., 58:415-500. Simpson, C. G. 1961. Principles of Animal Taxonomy. Columbia Univ. Press, New York. Smith, C. L. 1954. Pleistocene fishes of the Berends Fauna of Beaver County, Okla- homa. Copeia, 1954:282-289. Smith, C. L. 1958. Additional Pleistocene fishes from Kansas and Oklahoma. Copeia, 1958:176-180. Smith, C. L. 1962. Some Pliocene fishes from Kansas, Oklahoma, and Nebraska. Copeia, 1962:505-520. Smith, G. R. 1964. A late Illinoian fish fauna from southwestern Kansas and its cli- matic significance. Copeia, 1964:278- 285. Sneath, p. H. a., Sokal, R. R. 1973. Nu- merical Taxonomy. W. H. Freeman and Co., San Francisco. Sokal, R. R. 1975. Mayr on cladism — and his critics. Syst. Zool., 24:257-262. Starks, E. C. 1916. The sesamoid articular. Leiand Stanford Univ. Pubis. Univ. Ser., 1916:1-40. Stensio, E. 1921. Triassic fishes from Spitz- bergen. Part 1. Vienna. Stensio, E. 1935. Sinamia zdanskyi, a new amiid from the Lower Cretaceous of Shantung, China. Palaeont. Sin., (3c)]: 1-48. Stensio, E. 1947. The sensory lines and der- mal bones of the cheek in fishes and amphibians. Kungl. svensk. Vetensk. Akad. Handl., 3 (24): 1-195. Sternberg, C. M. 1942. Notes on the Lance formation of Southern Saskatchewan. Canad. Field-Natur., 33:66-70. Stromer, E. 1925. Der Rickgang der Ganoid fische von der Kreidezeiten. Zeitschr. deutsch. Geol. Ges., 73:348-371. Suttkus, R. D. 1963. Order Lepisostei. Mem. Sears Found. Marine Res., 1:61-88. Swift, C. 1968. Pleistocene freshwater fishes from Ingleside Pit, San Patricio County, Texas. Copeia, 1968:62-69. Tatarke, K. 1939. The mandibular and hyoid arches and gill cover in Polypterini, Lepisosteidei, and Aminoidei. Pap. Anim. Morph. Kiev: 17-71. (Russian with English summary. ) Teller, S., Bardack, D. 1975. New records of Late Pleistocene vertebrates from the southern end of Lake Michigan. Amer. Midi. Nat., 94:179-189. Theunissen, F. 1914. Over der rangschikking der motorische cellen in de hersenen van Acipenser ruthenus en Lepidosteus ossetis. Versl. Wis. Nat. Afd. Akad. Wet. Amsterdam, 22:963-971. Thompson, Z. 1842. History of Vermont, Nat- ural, Civil, and Statistical. Burlington, 1842. Uyeno, T. 1963. Late Pleistocene Fishes of the Clear Creek and Ben Franklin local faunas of Texas. J. Grad. Res. Centr., S. Methodist Univ., 31:168-173. Uyeno, T., Miller, R. R. 1962. Late Pleisto- cene fishes from a Trinity River terrace, Texas. Copeia, 1962:338-345. Uyeno, T., Miller, R. R. 1963. Summary of Late Cenozoic freshwater fish records of North America. Occ. Pap. Mus. Zool. Univ. Michigan, 631:1-34. Valentine, G. G. 1840. Ueber die Organi- sation der Trabeculae carneae in der schwimblase von Lepisosteus spattda Lacepede. Repert. Anat. Physiol. (Va- lentin), 5:392-397. Vasseur, G. 1876. Sur la couche a lepidostees de I'argile de NeuflesSaint-Martin, pres Gisors. Bull. Soc. Geol. France, 4:295- 304. Veit, O. 1907. Ueber einige Besonderheiten am Primordialcranium von Lepidosteus osseiis. Anat. Hefte. 1. Abth., 1907:155- 203. THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 109 Veit, O. 1911. Beitriige zur Kenntnis de Kopfes der Wirbeltiere I. Die Entwick- elung des Primordialcranium von Lepi- dosteus osseus. Anat. Hefte. 1. Abth., 1911:93-225. Veit, O. 1924. Beitrage zur Kenntnis der Kopfes der Wirbeltiere II. Fruhstadien der Entvvickelung des Kopfes von Lepi- dosteus osseus und ihre prinzipielle Be- deutung fur die Kephalogeneses des Wirbeltiere. Morpholog. Zuhrbuch., 53: 319-390. ViRCHOW, H. 1894. Einige embryologische und angiologische Erfahrungen iiber nordamerikanische Wirbeltiere. Sitzber. Naturf. Freunde Berlin, 1894:33-44. Wagner, G. 1912. On the present status of Lepisosteus sineiisis Bleeker. Zool. An- zeiger Bd. XXXIX Nr. 25/26, 2:738- 741. Watson, D. M. S. 1925. The structure of cer- tain palaeoniscids and the relationships of that group with other bony fish. Proc. Zool. Soc. London, 1925:815-870. Watson, D. M. S. 1937. The acanthodian fishes. Phil. Trans. Roy. Soc. London, 228B:49-146. Weidersheim, R. E. E. 1904. Machtragliche Bemerkungen zu meinen Aufsatz iiber Kehlkopf der Ganoiden und Dipnoer. Anat. Anz., 24:651-652. Wenz, S. 1961. Reduction du preopercle chez Dapedium. Comptes rendus des seances de I'academie de Sciences, 252:1371- 1373. Wenz, S. 1968. Complements a I'etude des poissons actinopterygiens du Jurassique Frangais. Cent. Nat. Res. Sci., Paris, 1967:1-276. Westoll, T. S. 1937. On the cheekbones in teleostome fishes. J. Anat., 71:362-382. Westoll, T. S. 1944. The Haplolepidae, a new family of Late Carboniferous bony fishes. Bull. Amer. Mus. Nat. Hist., 85: 1-122. White, E. F. 1931. The Vertebrate Faunas of the English Eocene. Vol. I. Oxford Univ. Press, London. Whitley, G. P. 1940. The nomenclator zo- ologicus and some new fish names. Aust. Nat. Sydney Mag., 1940:241-243. Wijhe, J. W. von. 1880. Over het Visceral- skelett in de zenuwen van den kop der Ganoiden. Doesburg, Leiden. Wijhe, J. W. von. 1882. Ueber das Visceral- skelett und die Nerven des Kopfes der Ganoiden und von Ceratodus. Nieder- landisches Arch. Zool., 5:207-320. Wilder, B. G. 1876. Notes on the North American ganoids, Amia, Lepidosteus, Acipenser, and Pohjodon. Proc. Amer. Assoc. Adv. Sci., 24:151-196. Wilder, B. G. 1877. On the brains of some fish-like vertebrates. Abst. in Proc. Amer. Assoc. Adv. Sci., 25:259-263. Wilder, B. G. 1878. On a remnant of the spiracle in Ainia and Lepidosteus. Amer. Natur., 19:192. Wiley, E. O. 1975. Karl R. Popper, system- atics, and classification: a reply to Wal- ter Bock and otlier evolutionary tax(3n- omists. Syst. Zool., 24:233-243. Williamson, W. C. 1849. On the microscopic structure of the scales and dermal teeth of some ganoid and placoid fishes. Phil. Trans. Roy. Soc. London, 1849:435-475. Williamson, W. C. 1851. Investigations into the structure and development of the scales and bones of fishes. Phil. Trans. Roy. Soc. London, 1851:643-702. WiLLisTON, S. W. 1902. The Laramie Cre- taceous of Wyoming. Science, 16:952- 953. Wilson, R. L. 1969. Systematics and faunal analysis of a Lower Pliocene vertebrate assemblage from Trego County, Kansas. Contr. Mus. Pal. Univ. Michigan, 22: 75-126. WiNCHELL, A. 1864. Description of a gar-pike, supposed to be new Lepisosteus (Cylin- drosteus) oculatus. Proc. Acad. Nat. Sci. Phila., 16:183-185. Winkler, T. C. 1876. Deuxieme memoire sur les dents de poissons fossiles du terrain bruxellian. Arch. Mus. Teyler, 4:16-48. WiNTERBOTTOM, R. 1972. The familial phy- logeny of the Tetradontiformes (Actino- pterygii: Pisces) as evidenced by their comparative myology. Smithsonian Contrib. Zool., 155:1-201. Wood, S. V. 1846. On the discovery of an alligator and of several new Mammalia in the Hordwell Cliff, with observations upon the geological phenomena of that locality. London Geol. J., 1:1-7, 117- 122. Woodward, A. S. 1890. Description of a fish- skull (from the Lameta beds, Nagpur, Indian). Rec. Geol. Sur. India, 23:23- 24. Woodward, A. S. 1895. Catalogue of the fos- sil fishes in the British Museum. Part III. London. Woodward, A. S. 1908. On some fish-remains from the Lameta beds at Dongargaon, Central Provences. Pal. Indica (N.S.), 3:1-6. Wright, E. P. 1879. On the development of the garpike (Lepidosteus osseus). Na- ture, 19:100-101. Wyman, J. 1844. On the microscopic structure of the teeth of the Lepidostei, and tlieir analogues with those of the labryrintho- 110 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY donts. Proc. Boston. Nat. Hist. Soc, 1: 131-132. ZiEGLER, H. E. 1900. Die neuren Forschungen in der Embryologie der Ganoiden. Zool. Centralblatt, 7:227-244. ZiTTEL, K. A., KoKEN, E. 1911. GnindzUge der Palaontologie (Palciontologie) von Karl A. van Zitfel. Berlin. Appendix A — Material Examined Specimens of gars used in this study are listed by Country or State and museum number for recent species, and Country, State (or re- gion) and formation for fossil specimens. Frag- mentary material not identifiable to species is not listed. The use of a "P." in front of the catalogue number indicates that tlie specimen is on deposit in the Paleontology Department of the institution cited (except the British Mu- seum, Natiual History), whereas no prefix in front of the number indicates that the specimen is on deposit in the Ichthyology Department of the institution cited. British Museum specimens are entirely from tlie Paleontology collections but may be labeled with or without the "P." Specimens catalogued in a separate osteological collection are prefixed by "ost." The following institutional abbreviations are used: AMNH, The American Museum of Natural History ANSP, Academy of Natural Sciences, Phila- delphia BMNH, British Museum (Natural History) LACM, Los Angeles County Museum MCZ, Museum of Comparative Zoology, Harvard University MHNP, Museum National d'Histoire Natu- relle, Paris SMC, Sedgwick Museum, Cambridge Uni- versity TU, Tulane University Collection of Fishes UMMP, University of Michigan Museum of Paleontology UMMZ, University of Michigan Museum of Zoology USNM, United States National Museum of Natural History Lepisosteus opertus Montana. Hell Creek Formation: MCZ P. 13392 (holotype); MCZ P.13393-P.13397 (Paratypes); MCZ P.9374 (paratype). AMNH P.9323 (catalogued with specimens of A. occi- dentalis ) . Lepisosteus cuneatus Utah. Manti Formation: AMNH P.2517 (type). AMNH P.4622-P.4625; MCZ P.13325. Lepisosteus platostomus Arkansas: TU 447722 (2 spec); TU 59709 (1 spec). Kansas: ANSP 621 (1 spec, type Cylindrosteus scabriceps Cope). Illinois: UMMZ 14705 (4 spec). Louisiana: TU ost. 297-299 (1 spec, each); TU ost.397-398 (1 spec, each); TU 47529 (3 spec); TU 47546 (4 spec); TU 47579 (2 spec); TU 47657 (1 spec); TU 70151 (2 spec); TU 87357 (7 spec); USNM 172825 (6 spec); USNM 173088 (2 spec). Iowa: UMMZ 10115 (9 spec). Mississippi: USNM 129249 (6 spec); USNM 129334 (2 spec); USNM 129342 (8 spec); USNM 129459 (5 spec). Missouri: TU 53817 (8 spec); UMMZ 147916 (2 spec); UMMZ 190842 (1 spec); UMMZ 190846 (2 spec). Nebraska: UMMZ 134778 (1 spec). Tennessee: USNM 32373 (1 spec). Lepisosteus indicus India. Madliya Pradesh, Dongargoan, La- meta Beds: BMNH P.12178 (type): 40 mi. WNW of Nagpur; BMNH P.12186 near Takli; BMNH P.12185. Lepisosteus osseus Florida: LACM 33915-33921 (22 spec, to- tal). Indiana: USNM 64917 (1 spec). Iowa: UMMZ 173448 (3 spec). Kentucky: USNM 89440 (1 spec). Michigan: UMMZ 173340 (1 spec); UMMZ 174554 (4 spec); UMMZ 174558 (3 spec); UMMZ 180463 (1 spec); UMMZ 182051 (1 spec); UMMZ 189179 (4 spec); UMMZ 56017 (100 spec); UMMZ 60648 (4 spec); UMMZ 82335 (1 spec). Mississippi: USNM 129231 (1 spec). Mis- souri: UMMZ 147917 (4 spec); UMMZ 148093 (1 spec); UMMZ 148825 (3 spec); UMMZ 150213 (2 spec); UMMZ 150762 (1 spec). New York: AMNH 599 (1 spec); AMNH 28657 (2 spec); USNM 69947 (1 spec). Ohio: ANSP 77987 (1 spec). Texas: USNM 89430 (1 spec); AMNH uncat. (4 spec). Lepisosteus fimbriatus England, U.K. London. Dulwic, Wool- wich Beds: BMNH P.5504; P. 15488; P.33531; P.37201; P.39001; P.40338; P.41080. Hants., Barton Beds: BMNH P.12625; P.13057; P.39185-39189. Headon Beds: BMNH P.1529 and P.1529a; P.1700; P.13330; P.21058; 25252 and 25254 (types); 25258; P.27603; P.27607; P.27706; 28540; P.30295; P.33522-33526; P.33530; P.38104; P.46388-46389; P.47568- 47569; SMC P.31389-31405; SMC P.31414. Kent, Oldhaven Beds: BMNH P.15280; P.16695- 16696; P.31181. Blackheath Beds: BMNH P.14610; P.14613; P. 14635-14637; P.14640; THE PHYLOGENY AND BIOGEOGRAPHY OF FOSSIL AND RECENT GARS 111 P.14680; P.14753; P. 16333- 16334; P.19884; P.19908-19933; P.20010-20018; P.28066-28067; P.28575-28576; P.39313; P.1611a-i; P.46074- 46075; P.51284; P.51297; P.51648; P.55510; P.55512-55517; P.55742-55748; P.55941-55947. Suffolk Pebble Beds: BMNH 29017. Sussex, Worthing: BMNH P.20127. Thornton Beds: P.38602. France. Paris Basin: MHNP P.8959; MHNP P.1874-638; MHNP P.4-1876; MHNP Lemoine collections, lots 1-3. Lepisosteus oculatus Ahhama: UMMZ 103506 (1 spec). Ar- kansas: UMMZ 123149 (1 spec). Florida: LACM 33914-33916 (22 spec, total); LACM 33912 (1 spec); TU 23157 (1 spec); TU 23795 (1 spec); TU 23837 (1 spec); TU 40572 (1 spec); UMMZ 165168 (1 spec). Louisiana: TU 268 (1 spec); TU ost.300 (1 spec); TU 6376 ( 1 spec); TU 6506 ( 1 spec); TU 11447 (5 spec); TU 11618 (2 spec); TU 13877 (3 spec); TU 16842 (1 spec); TU 17115 (2 spec); TU 17680 (1 spec); TU 41453 (1 spec); UMMZ 170787 (1 spec); USNM 172093 (1 spec). Michigan: UMMZ 55062 (holotype); UMMZ 166511 (1 spec); UMMZ 178806 (1 spec). Mississippi: TU 85996 (3 spec); TU 86199 (1 spec); TU 86464 (1 spec). Texas: TU 22289 (5 spec); TU 24597 (2 spec); TU 66629 (1 spec); TU 85567 (1 spec); UMMZ 165203 (2 spec); UMMZ 165210 (1 spec); UMMZ 166184 (1 spec). Lepisosteus platyrhinctis Florida: LACM 33912-33913 (22 spec, to- tal); UMMZ 158596 (1 spec); UMMZ 158624 (7 spec); UMMZ 159805 (5 spec); UMMZ 166536 (1 spec); USNM 26214 (1 spec); USNM 92832 (1 spec); USNM 133399 (1 spec); USNM 133429 (4 spec). Georgia: UMMZ 158093 (1 spec). Atractosteus strausi Germany. Vicinity of Darmstadt, Messel Locality: AMNH P.4626; AMNH 33839; AMNH 33856; also 16 uncatalogued casts of privately held specimens at AMNH; BMNH P.33506-33519. AMNH 25649 (4 spec); AMNH 25790 (1 spec); AMNH 27937 (1 spec); AMNH 28075- 28076 (1 spec, each); AMNH 33851 (1 spec); UMMZ 144241 (1 spec); UMMZ 144244 (1 spec); UMMZ 144247-144249 (1 spec each); UMMZ 144251-144252 (1 spec, each); UMMZ 144254 (1 spec). Mexico: TU 84923 (10 spec); UMMZ 184612 (2 spec); UMMZ 184631 (1 spec); UMMZ 187727 (1 spec): UMMZ 187747 (1 spec); UMMZ 187793 (1 spec). Nicaragua: TU 24277 (2 spec); USNM 44175 (1 spec); USNM 120715 (1 spec). Atractosteus simplex Wtjoming. Bridger Formation: USNM P.2174 (type); USNM P.21173 (cotype). Green River Formation: AMNH P.4302; AMNH P.4305; MCZ P.5318; USNM P.22752. New Mexico. Wasatch Formation: USNM P.2582 (type of C. aganus Cope); USNM P.2584 (type of C. interger Cope). Atractosteus occidetitalis Montana. Hell Creek Formation: AMNH P.4304; BMNH P.56533-56537; MCZ P.9377- 9379; MCZ P.9385. Lance Formation: BMNH P.48140-48153. Laramie Fomiation: BMNH P. 10738-10739. Canada. Belly River Forma- tion, Alberta: BMNH P.11906, BMNH P.12222- 12223. Atractosteus atrox Wyoming. Bridger Formation: USNM P.2145 (type); USNM P.4755; MCZ P.5168. Green River Fomiation: USNM P.3974 (type Clastes anax Cope). Atractosteus spatula Louisiana: TU ost.119-123; TU ost.131; TU ost.315; TU ost.347-348; TU ost.351-353; TU ost.364-365 (1 spec each); TU 17115 (2 spec). Texas: UMMP 55462-55463; UMMZ 111049 (1 spec, each); UMMZ 1131001 (12 spec); UMMZ 131165 (1 spec). Mexico: TU ost.415 (parts of 5 spec); TU ost.477 (parts of 8 spec); USNM 1003 (1 spec, type A. belanderi Girard). Nicaragua: TU ost.388. Atractosteus tropicus Costa Rica: UMMZ 175920 (1 spec); USNM 6806 (1 spec, holotype). Guatemala: AMNH 22092-22099 (1 spec each); AMNH 25192 (3 spec); AMNH 25622 (1 spec); Atractosteus tristoechus Cuba: AMNH 3097 (9 spec); UMMZ 30745 (1 spec); USNM 12496 (1 spec, type L. manfuari Poey); USNM 24794 (1 spec); USNM 11309 (parts of 5 spec). RECENT MISCELLANEOUS PUBLICATIONS UNIVERSITY OF KANSAS MUSEUM OF NATURAL HISTORY 52. Reproductive cycles in lizards and snakes. By Henry S. Fitch. Pp. 1-247, 16 fig- ures in text June 19, 1970. Paper bound, $5.00 postpaid. 53. Evolutionary relationships, osteology, and zoogeography of leptodactyloid frogs. By John D. Lynch. Pp. 1-238, 131 figures in text, June 30, 1971, Paper bound, $7.00 postpaid. 54. The dentition of glossophagine bats: development, morphological characteristics, variation, pathology, and evolution. By Carleton J. Phillips. Pp. 1-138, 49 figures in text. September 24, 1971. Paper bound, $5.00 postaid. 55. Middle American lizards of the genus Ameiva (Teiidae) with emphasis on geo- graphic variation. By Arthur C. Echtemacht. Pp. 1-86, 28 figures in text De- cember 14, 1971. Paper bound, $3.00 postpaid. 56. Systematics of the chiropteran family Mormoopidae. By James Dale Smith. Pp. 1-132, 40 figures in text. March 10, 1972. Paper bound, $5.00 postpaid. 57. A systematic review of the Teiid lizards, genus Bachia, with remarks on Heter- odactylus and Anotosaura. By James R. Dixon. Pp. 1-47, 15 figures in text. February 2, 1973. Paper bound, $1.50 postpaid. 58. Systematics of three species of woodrats (genus Neotoma) in central North Amer- ica. By Elmer C. Bimey. Pp. 1-173, 44 figures in text. April 13, 1973. Paper bound, $7.50 postpaid. 59. Systematics and evolution of the Andean lizard genus Pholidobolus (Sauria: Teiidae). By Richard R. Montanucci. Pp. 1-52, 8 figures in text. May 14, 1973. Paper bound, $1.75 postpaid. 60. Mammals of the Black Hills of South Dakota and Wyoming. By Ronald W. Turner. Pp. 1-178, 14 figures in text. April 3, 1974. Paper bound, $7.50 postpaid. 61. Reproductive strategies in a tropical anuran community. By Martha L. Crump. Pp. 1-68, 13 figuures in text. November 15, 1974. Paper bound, $4.50 postpaid. 62. A demographic study of the ringneck snake (Diadophis punctatus) in Kansas. By Henry S. Fitch. Pp. 1-53, 19 figures in text. April 3, 1975. Paper bound, $3.00 postpaid. 63. Morphology of the bony stapes (columella) in the Passeriformes and related groups: evolutionary implications. By Alan Feduccia. Pp. 1-34, 16 plates, 7 figures in text. May 30, 1975. Paper bound, $3.00 postpaid. CT* erne Bookbinding Co.. Inc. 300 Summer Street B«8ton, Mass. 0??r) 3 2044 093 361 632 Date Due