AMERICAN M[USEUM[ NOVITATES Published by Number 1110 the AMERICAN MUSEUM of NATURAL HISTORY May 16, 1941 New York City

Home , Diplurus, Libys, Rhabdoderma, Undina (fish)

AMERICAN M[USEUM[ NOVITATES Published by Number 1110 THE AMERICAN MUSEUM OF NATURAL HISTORY May 16, 1941 New York City

A REVISION OF COELACANTHUS NEWARKI AND NOTES ON THE EVOLUTION OF THE GIRDLES AND BASAL PLATES OF THE MEDIAN FINS IN THE COELACANTHINI BY BOBB SCHAEFFER' INTRODUCTION The existence of coelacanth fishes in the University, for the discovery and presenta- Upper Triassic of North America has long tion of the first specimens from this latter been known from the description by New- site, to Mr. V. E. Shainin for the presenta- berry (1878, 1888) of the very poorly pre- tion of the very fine specimens in his pos- served Diplurus longicaudatus. Since then, session, and particularly to Professor A. W. Lambe (1916) has described Coelacanthus Pollister of the Department of Zoology banifensis from a single fin and more re- for permission to study his private collec- cently Bryant (1934) has reported Coela- tion and for much helpful advice. He also canthus newarki from the Newark Series. wishes to thank Dr. G. L. Jepsen for per- From a new locality much more abundant mission to study the type specimen, Pro- material of a form that is exceedingly close fessor W. K. Gregory and Mr. H. C. Raven to C. newarki has become available; and in for critically reading the manuscript and view of the scarcity of coelacanth remains finally Dr. A. J. Ramsay for his assistance from the Triassic of North America, it is with certain of the photographs. All the desirable to report this material in some specimens collected in the Granton Quarry detail. have been presented to The American It has been pointed out by Stensio (1921) Museum of Natural History and catalogued and Moy-Thomas (1937) that a great num- as A.M.N.H. No. 15222. ber of the coelacanths referred to the genus Coelacanthus possess characters that are OSTEOPLEURUS, NEW GENUS quite distinct from those of the genotype, GENOTYPE.-Coelacanth?is newarki Biyant. which is the Permian Coelacanthus granula- DIAGNOSIs.-Long, ossified, pleuiral ribs. Pelvic fin midway between first and second dor- tus of Agassiz (1839). In view of this fact, sal fins. Plates of pelvic girdle quadrilateral in and for reasons to be mentioned later, it shape. Basal plate of anal fin bifurcated. seems a(lvisable to establish a new genus Basal plate of first dorsal, triangular. Basal for the reception of the species from the plate of second dorsal, deeply bifurcated. Scales possessing about five prominent parallel Upper Triassic of Pennsylvania, described ridges. as Coelacanthus newarki by Bryant, and also for the material to be described in this HORIZONS AND LOCALITIES paper, which is referred to the same genus The Pennsylvania material (one com- and species. plete specimen and several fragments) was The name Osteopleurus (Gr. osteon, bone, collected by Dr. Jepsen in the Upper and Gr. pleuron, rib) is herewith proposed Triassic shales of the Newark Series in an for this genus in allusion to the long, ossi- exposure of the Lockatong formation, near fied ribs. It will tentatively contain just the town of North Wales. The New Jersey one species, newarki. The type is Prin. specimens were found in the Upper Triassic Univ. Pal. Mus. No. 13695. black shales of the Stockton formation of The author is indebted to Dr. G. M. Kay the Newark Series in the Granton Quarry of the Department of Geology, Columbia at North Bergen, New Jersey. The char- 1 Department of Zoology, Columbia University. acter of the sediments and the presence of 2 AMERICAN MUSEUM NOVITATES [No. 1110 the phyllopod Estheria ovata Lea indicate five prominent longitudinal ridges, the that both formations are quite probably middle one being the longest and largest. fresh water and less likely brackish water There are also fine striae which converge in deposits. the covered portion. Annular rings are present which tend to break the continuity MEASUREMENTS of the striae. The details of ornamentation There are no observable qualitative are well preserved in the scale illustrated in characters which can be used to separate Fig. 2,A. the forms from the Lockatong and the Stockton. There are, however, quantita- SKULL tive differences which at first were considered to be of specific significance. An at- Due to crushing it is impossible to make tempt was made to use quantitative meth- out the detailed structure of the skull in ods outlined by Simpson and Roe (1939) any of the specimens. Some features of for comparing small samples and single importance can be determined in one of specimens in an effort to determine the specimens from the Stockton (Fig. 1,A), whether we are dealing with a single species and, also, from numerous isolated plates. or two separate species. This was not The opercular is a large bone, as it is in successful, however, since nothing is known most coelacanths. An isolated example concerning relative ontogenetic growth in shows that it is oval in outline and lacks coelacanths in general, and since the num- ornamentation. The outline of the post- ber of measurable specimens of Osteopleurus orbital cannot be determined in the pre- is very small. As Dr. Simpson pointed served skull. An isolated plate (Fig. 4,B), out, it is impossible to determine in this which has a different shape than the oper- case whether the quantitative differences cular, has been tentatively identified as the are specific differences or merely different postorbital, although it may be relatively stages in the growth of a single species. too large to be this bone. The lack of a Therefore, until more well-preserved mate- lateral line canal would seem to indicate rial is obtained, all of the specimens are re- that the medial surface is exposed. The to one newarki. postorbital lateral line canal is, however, ferred the species, very prominent in the preserved skull. MEASUREMENTS IN MILLIMETERS That the squamosal, opercular, preopercu- Pennsyl- lar, and subopercular are missing from New vania this skull is indicated the presence of a Jersey Speci- by Speci- mens depressed region (Fig. 1,A,x) behind the mens (approx.) orbit. Furthermore, the squamosal branch Body length (minus supple- of the postorbital lateral line canal is not mentary caudal) 110 71 evident. Head length 30 19 Length post. border of head The angular bone can be readily identi- to ant. border of caudal 45 26 fied, as the mandibular branch of the lat- Length first caudal 35 26 eral line runs over it. The Length first plus second dentary, splen- caudal 45 ? ial, and coronoid are not distinct. The Fin locations, mm. from angular plates possess the long, oval shape terior end to: characteristic of the coelacanthids. They First dorsal 35-40 22 are ornamented with fine which fol- Second dorsal 62 36 ridges Pelvic 45 25-26 low the outline of the bone. One end has Anal 68 40 a wide, shallow groove which tapers to a Caudal 80 45 point in the middle of the plate. They are Body depth at ant. end of 11 mm. in length and 2.7 mm. in width first dorsal (approx.) 24 15 (Fig. 2,C). SQUAMATION The suture between the frontals and in- Each scale appears to be an elongated tertemporals is very evident, the bones oval. The exposed portion possesses about having been abnormally separated during Fig. 1. Osteopleurus newarki. A. Lateral view of skull. X2.8. ang., angular; ant., antorbital; clav., clavicle; cl., cleithrum; dn., dentary; excl., extracleithrum; fr., frontal; int., intertemporal; laju., lacrimo-jugal; pas., parasphenoid; po., postorbital; st., supratemporal; x., opercular region. B. Lateral view of almost complete specimen showing position of all the fins, scale imprints, and ossified ribs. X .1.

3 Fig. 2. Osteopleuerus newarki. A. Isolated scale showing details of sculpturing. X 18.0. B. Unidentified plate. X9.6. C. Gular plate. X7.4. D. Lateiral vieew of a second almost complete specimen showing well-preserved neural arches. X1.4. 4 19411 REVISION OF COELACANTHUS NEWARKI 5 preservation. The supratemporal and ex- which meets the ventral border of the trascapulars cannot be distinguished from cleithrum. the intertemporal, although the supratem- A pelvic fin is preserved in place in two poral is undoubtedly present as a poorly of the specimens, but the pelvic girdle is defined plate lying just dorsal to the oper- covered with a layer of matrix which cannot cular region. What appears to be an ant- be removed. It is preserved in the type, orbital is present as a small triangular bone being well ossified, quadrilateral in outline, on the anterior rim of the orbit. The post- and with concave margins. The fin has rostrals and premaxillary are crushed and about twelve lepidotrichia and is relatively their outlines cannot be determined. The further behind the pectoral girdle than the circumorbital plates have not been pre- pelvic fin of Coelacanthus granulatus. served. The lacrimojugal characteristi- cally frames the ventral rim of the orbit. UNPAIRED FINS A bar of bone running lengthwise through The anterior dorsal fin including the the orbit may be the somewhat dorsally basal plate is well preserved in several speci- displaced parasphenoid. It resembles that mens (Fig. 3,A). This plate has a more or part of the bone between the more posterior less triangular shape, with the apex pointed expanded portion and the ethmoid ossifica- anteriorly. The lepidotrichia, about nine in tion, as figured by Moy-Thomas (1937) number, are attached directly to the plate, for Rhabdoderma. the proximal ends being forked to fit over The ossified part of the neurocranium the plate. The former are grooved and (Fig. 4,C) is poorly preserved in dorsal as- flattened laterally and are not broken up pect on one of the slabs in the Princeton into segments, except at their distal ends. collection (P.U.P.M. No. 13826). The The basal plate of the posterior dorsal parasphenoid has its edges upturned to a fin is preserved in the complete Pennsyl- greater degree than in Wimania. On each vania specimen. As Bryant states, the side of the channel thus formed a bony wing plate is bifurcated. The lower process, projects sidewards, as in the Undina para- however, is about the same length as the sphenoid (Aldinger, 1930). Very frag- upper, as can be seen in the example figured mentary bits of bone anterior to the para- in PI. vIII, fig. 1, of Bryant's paper. The sphenoid have been identified as the eth- fin contains about twelve lepidotrichia. moid ossifications. The location of the The basal plate of the anal fin (Fig. 9,E) basisphenoid, pro6tic, and basioccipital is is also only known from the type specimen. indicated, but it is impossible to make out It is deeply bifurcated, the forks being more any of the structural details. There are slender than those on the basal plate of the several examples of a well-preserved iso- posterior dorsal. The process with which lated plate of constant shape which cannot the fin rays articulated is quite narrow but be positively identified (Fig. 2,B). It is somewhat expanded posteriorly. possibly a part of the ossified neurocranium, Both the epichordal lobe and the hypo- or even of the pterygoid complex. chordal lobe of the caudal fin contain about ten to twelve lepidotrichia. The supple- PAIRED FINS AND GIRDLES mentary caudal fin is beautifully preserved The pectoral fin is barely indicated on in negative in one of the New Jersey speci- Mr. Shainin's specimen, lying on what mens (Fig. 3,B). It is relatively longer may be the remains of the ossified swim than the one found in Undina and consists bladder (Fig. 2,D). One specimen has a of dermal rays in the usual tuft. well-preserved pectoral girdle (Fig. 1,A). The cleithrum is apparently quite similar AXIAL SKELETON to that found in Rhabdoderma. Externally, The neural arches are robust forked there is a ridge which runs from the upper structures with fairly long, spinous proc- posterior border to the lower anterior esses which are apparently rather well ossi- border. A curved clavicle is indicated fied (Fig. 2,D). There are about thirty Fig. 3. Osteopleurus newarki. A. Trunk region showing first dorsal fin, with its basal plate, neural arches, and ossified ribs. X2.9. B. Specimen showing caudal fin and well-preserved supplementary caudal fin. X 2.3. REVISION OF COELACANTHUS NEWARKI 7 pairs of long, ossified pleural (subperi- scales have a great number of fine longitu- toneal) ribs (Fig. 2,D and Fig. 3,A).. This dinal ridges. The only. observable resem- is an important character which, above all blance between the two is the triangular others, separates Osteopleurus from the shape of the basal plate of the first dorsal, genotype of the genus Coelacanthus, which which, as will be shown, is a character com- has very short, pleural ribs. Haemal arches mon to most members of the order. are evident posterior to the last rib (Fig. By establishing the new genus Osteo- 2,D). pleurus for the identifiable remains of With regard to the Upper Triassic coela- coelacanths other than those of Diplurus canths from North America, Coelacanthus from the Triassic of Pennsylvania and banffensis (Lambe, 1916), from Alberta, New Jersey, it is hoped that a further con- is, as Moy-Thomas points out, far too tribution has been made toward a more fragmentary for generic determination. natural classification of the Coelacanthini. The evidence indicates that Osteopleurus is Diplurus longicaudatus (Newberry, 1878) probably most closely related to Coelacan- from the Brunswick Formation of the Up- thus, although, as pointed out above, the per Triassic of New Jersey is quite distinct pelvic plates have a very different shape. from Osteopleurus. It reaches a length of Until better preserved skull material is three feet and lacks ossified ribs. The found, it is unsafe to speculate further.

-CU .B Fig. 4. Osteopleurus newarki. A. Outline of opercular plate. X 3.5. B. Outline of plate tentatively identified as postorbital. X3.8. C. Dorsal view of neurocranium. X8.1. eth., ethmoid; ps., parasphenoid; bs.,basisphenoid; pro., prootic; bo., basioccipital.

THE EVOLUTION OF THE GIRDLES AND BASAL PLATES OF THE MEDIAN FINS IN THE COELACANTHINI In an attempt to determine the possible (1932, p. 46), however, that the group affinities of Osteopleurus a comparative shows a much greater range of variation study of the paired and unpaired fin sup- than formerly suspected and that it is pos- ports has been undertaken. It is well sible to distinguish several natural groups known that in most respects the coelacanths of genera. There are some striking varia- are a very homogeneous and rather con- tions in the pelvis and in the basal plates of servative order. Stensio has pointed out several of the median fins. These varia- 8 AMERICAN MUSEUM NOVITATES [No. 1110 tions must be considered along with the Moy-Thomas has suggested (1937) that it skull in dividing the Coelacanthini (Ac- may be fused with the cleithrum in many tinistia) into smaller categories. Up to cases, i.e., Rhabdoderma. the present time, no attempt has been made The supracleithrum is evidently loosely to subdivide the order or to construct at united by ligaments to the cleithrum in all least a tentative phylogeny. Such an at- but a few genera, such as Coelacanthus and tempt is made in this paper. possibly Osteopleurus, in which it appears

F F 6 Fig. 5. Series of pectoral girdles. Scales various. A. Eusthenopteron foordi. Upper Devonian. Medial view (after Bryant). B. Diplocercide8 kay8eri. Upper Devonian. Lateral view (after Stensi6). C. Rhabdoderma elegans. Upper Carboniferous (composite, after Huxley and Moy-Thomas). D. Coelacanthu8 granulatus. Permian. Lateral view (after Moy-Thomas and Westoll). E. Laugia groenlandica. Lower Triassic (after Stensib). F. Osteopleurus newarki. Upper Triassic (original). G. Macropoma speciosum. Cretaceous (after Reis). H. Latimeria chalumnae. Recent. Lateral view and cross-section at level of arrows (after Smith).

THE PECTORAL GIRDLE to be suturally united. It is interesting to The pectoral girdle (Fig. 5) in all coela- note in this connection that there is consid- canths consists of the following dermal ele- erable variation in the intimacy of the ments: a somewhat triangular supraclei- union between the supracleithrum and the thrum, a dominant cleithrum, and a cla- cleithrum in the teleosts. vicle which is never as long or as wide as The cleithrum is quite constant in shape, the cleithrum. The endoskeletal portion the variations being of minor importance. of the girdle is but rarely distinguishable. It is, in general, the longest bone of the 1941] REVISION OF COELACANTHUS NEWARKI 9 girdle and is in most cases rather narrow. clavicle and is very slightly overlapped by It is relatively much narrower than the them on its anterior margin. same bone in the rhipidistians. In Lati- The earliest coelacanthid pectoral girdle meria (Smith, 1940) the cleithrum is curved known, that of Diplocercides, possesses a about its long axis as can be seen in the separately ossified scapulo-coracoid ele- cross-section in Fig. 5,H. This is appar- ment of relatively large size. As men- ently not a general feature, for in Rhabdo- tioned above, the endoskeletal portion of derma it is a strongly ridged but essentially the girdle is not as independent of the der- a flat plate. mal as it is in the rhipidistians. In Diplo- The coelacanth clavicle has its greatest cercies, it consists of a flat, lenticular plate extent in the same vertical axis with the very intimately associated with the clei- cleithrum. It lacks the prominent hori- thrum and clavicle. Besides the sugges- zontal portion which is present in the tion of Moy-Thomas that it may be fused rhipidistian clavicle. The clavicle of the with the cleithrum, the possibility that it Upper Devonian Diplocercides does possess was often but poorly ossified must be con- a fairly wide horizontal ramus, approaching sidered. In Undina and Macropoma, the the condition in Eusthenopteron. In all endoskeleton has a posterior projection genera the clavicles meet in the midventral which extends beyond the cleithrum. This line. In Latimeria each clavicle has a dis- process is quite characteristic of these tinct facet for articulation with its fellow. Mesozoic genera. To it was attached the There is no evidence of the presence of an skeleton of the pectoral fin. In no case is it interclavicle as recently described by possible to distinguish the scapular element Gross (1936) in certain rhipidistians. from the coracoid, as they are completely One of the most interesting features of co6ssifled. the pectoral girdle is the presence of a der- There is considerable variation in the lo- mal bone apparently not found in any other cation of the scapulo-coracoid plate, and vertebrate. This bone, named by Moy- hence, in the pectoral fin in relation to the Thomas the extracleithrum, has so far been cleithrum and the pectoral girdle as a whole. observed only in Coelacanthus granulatus In Diplocercides the plate is located on the (Fig. 5,D) and in Latimeria chalumnae lower extremity of the cleithrum, even (Fig. 5,H), although it may have been overlapping the clavicle to a considerable present in Osteopleurus. It is located on degree. In nearly all other coelacanths the lateral surface of the girdle behind the the endoskeleton has apparently migrated, cleithrum, contacting the clavicle ventrally. at least relatively, dorsalward to the The phylogenetic history of the extra- middle of the cleithrum. In Macropoma cleithrum is at present unknown. It does and Undina, the same positional result is seem odd that it should appear as a neo- obtained by a relative increase in the morph in only a few genera. Moy-Thomas length of the clavicle. It would appear (1935) does not believe it to be homologous that the various positions of the pectoral with the postcleithrum of actinopterygi- fin were specializations associated with the ans since it is ventral to the endoskeletal different modes of living in the Coelacan- portion of the girdle. In any case the very thini. For instance, the pectoral fin of distant affinities of the coelacanths and Laugia could only have been used for teleosts seem to preclude any possibility maneuvering movements while swimming, of determining its origin and homologies while that of Diplocercides, like Eusthenop- in that manner. It appears extremely un- teron, was probably also used for propelling likely that the extracleithrum is a new ele- the body over the sea floor. Unfortu- ment in the coelacanths. Very probably nately, the endoskeletal girdle of Latimeria it is indistinguishably fused with the was not recovered, but Smith (1940) is of cleithrum in most crossopts at an early the opinion that it is situated just inside stage in ontogeny. In Latimeria, it is situ- the dorsal apex of the extracleithrum, as in ated superficially to the cleithrum and the Coelacanthus granulatus, or at about the tk o ~0

00 o L A .Q, L4

.0 0 ~ §13 s>ta 0-

KU .-j

QZi

- 1.% llm::::. 'It.. A. I- I Q.C.z QI REVISION OF COELACANTHUS NEWARKI 11 middle of the cleithrum. The pectoral horizontal, as in the Rhipidistians, or fin, however, has a rather long, fleshy por- slightly V-shaped. In such forms as tion, thus carrying the fin well down below Wimania and Coelacanthus the space be- the ventral surface of the body. This sug- tween the adjoining pelvic plates would be gests that it can be used for pulling the a mere slit, while in Rhabdoderma and Un- body over the ocean floor. dina it would be relatively quite large. The anterior and posterior divisions (so- THE PELVIC GIRDLE called by Stensi6, 1932) were expanded to various degrees for the attachment of the The variation in the shape of the pelvic body and fin musculature. The proximal plates results in a natural division into radial(s) of the pelvic fin articulated with three types represented by the plates of the posteroventral region of the posterior Rhabdoderma, Laugia, and Osteopleurus division. (Fig. 6). Unfortunately the pelvis has been described or preserved in but few of The pelvis of Laugia is situated so far the known genera and the differences be- forward that it evidently made a ligamen- tween these known types are extreme tous connection with the clavicle of the enough to indicate that the range of varia- pectoral girdle. This very percomorph- tion must have been rather great. Al- like position plus the great size of the pelvic though the pelvis is unknown in the Upper fins were probably responsible for the very Devonian forms, it is well preserved in the different shape of the plates of this girdle. Carboniferous Rhabdoderma. This prob- In spite of the specializations, it is possible ably represents a very close approach to to recognize homologous parts in the pelvic the primitive type. The resemblance to plates of Laugia and Rhabdoderma. A the pelvis of Eusthenopteron is not too great plate of the former has triangular anterior but is close enough to indicate a derivation and posterior divisions with a prominent from the rhipidistian type. Through the but not denticulated medial process spring- kindness of Professor Gregory and Mr. ing from the latter. As in the percomorphs, Raven I am permitted to include a figure the pelvic girdle is hence not only anchored of the second known specimen of a pelvic by the hypaxial musculature but also by plate of Eusthenopteron which will be de- attachment to the pectoral girdle. This scribed by them in detail at a subsequent condition undoubtedly evolved along with time. It suffices here to note that this the enlargement of the pelvic fin which, plate does not resemble very closely the incidentally, extends posteriorly almost to one figured by Goodrich (1901) but does the base of the anal fin. Such a large fin agree more closely with the Rhabdoderma would offer great resistance to the water type. during movement, necessitating the coinci- The Rhabdoderma type of pelvic plate is dental development of such a bracing also found in Coelacanthus granulatus, Un- mechanism. A pelvic plate of Laugia is dina, Macropoma, and Wimania. The further specialized in having the anterior part of the plate showing the greatest division twisted medially and the nmedial amount of variation in this group is the and lateral borders of the posterior divi- medial process. This process may (Rhab- sion turned more or less dorsally, thus doderma) or may not (Undina) possess a offering the most advantageous positions denticulated medial border. It varies in for the attachment of the fin and body size from little more than a serrated region muscles. on the medial border between the anterior Stensi6 considers the pelvis of Laugia to and posterior divisions in Wimania (Sten- be suggestive of the Rhabdoderma type, par- si6, 1921, P1. VIII, fig. 7) to a very promi- ticularly of Rhabdoderma elegans. It was nent process in Rhabdoderma. This proc- probably derived from a form with a rather ess articulated with its counterpart of the well-developed medial process. opposite side to produce a girdle which in The pelvis of the Osteopleurus is at pres- transverse section was either perfectly ent quite baffling. A study of Bryant's 12 AMERICAN MUSEUM NOVITATES [No. 1110 original material has not revealed with cer- In the latter, the two processes are more tainty which is the lateral and which the nearly the same size. The anterodorsal medial borders of the plates. Until this process is in all cases on the same longitu- point can be definitely settled it is unwise dinal axis with the posterior expanded por- to make detailed comparisons. In any tion of the plate, while the anteroventral event, the pelvic plates are very unlike any- process meets the posterior division at an thing previously described and if found angle of about thirty-five degrees. The alone would certainly not be considered as posterior portion reaches its greatest extent coelacanthid. A plate is quadrilateral in in Rhabdoderma, in which the anterior outline, rather than pentagonal as Bryant processes are relatively quite short. In states, with a rather short anterior projec- Osteopleurus, on the other hand, the proc- tion. The border of each side is slightly esses are relatively long and narrow and the concave. posterior division is reduced. Laugia has a well-developed posteroventral process. BASAL PLATE OF THE FIRST DORSAL This extension is indicated in Coelacanthus FIN granulatus, Undina penicillata, and in Maw- In most genera the basal plate of the sonia minor (Woodward, 1908). The basal first dorsal fin (Fig. 7) is triangular with plate of the second dorsal fin is certainly one the apex directed forward. One observed of the diagnostic structures, along with the exception to this is found in Rhabdoderma, caudal fin, of the postcranial skeleton. in which the plate is oval or kidney-shaped. The ventral border is notched in some spe- BASAL PLATE OF THE ANAL FIN cies of this genus for articulation with the The variation in the shape of this plate neural spines. The other exception is found (Fig. 9) is, indeed, very extreme. In in Scleracanthus in which the plate, while be- Rhabdoderma it resembles that of Eusthe- ing essentially triangular, is slightly forked nopteron rather closely and for that reason at its anterior end. In the rhipidistians, the former may approach the primitive the basal plate of this fin is simply a rod type. The simple quadrilateral plate of with no anteroposterior expansion (Moy- Coelacanthus granulatus may represent the Thomas, 1939, Fig. 22,A). A single radial posterior expanded portion of the Rhabdo- is situated between this plate and the lepi- derma type. In Osteopleurus, Undina, and dotrichia of the fin. In the coelacanths, Macropoma, the plate is deeply bifurcated on the other hand, the lepidotrichia articu- like the basal plate of the second dorsal fin. late directly with the dorsal border of the This type must have arisen by the develop- fin. ment of an anteroventral process from the posterior expanded division. The basal BASAL PLATE OF THE SECOND DORSAL plate of this fin in Laugia is so utterly un- FIN like any of the others that its origin is not The basal plate of the second dorsal fin ascertainable. In the ventral view, it is (Fig. 8) is without exception a bifurcated cross-shaped. A large median keel pro- structure. The open part of the V is al- jects dorsally at right angles to the hori- ways directed forward. In Eusthenopteron zontal portion. These various extensions the anterodorsal process is very short and certainly anchored the plate firmly in the rounded, while the anteroventral process ventral musculature, and their develop- is much longer and quite robust. Such a ment must have been- associated with some plate could very well be the structural an- sort of torsion acting on the plate during cestor of the same plate in the coelacanths. fin movement. 19411 REVISION OF COELACANTHUS NEWARKI 13

24 PNYLOqEJVY OF r7E COELA V II ' |Recent Latirneria

o Cretaceous I Mawsonia| Coccodernia Jlurassic|Jurassic| ) V ~/,_Libq ena/

7l a.i A|eiiaL ' \Scleracanthius

Pnermian Coelacantthus 5permatodus7 Carbon- ife'rousRhabdodermfa Dplocercides tipper esroides J)evonian RhipidistiaIn Stock Figure 10.

CLASSIFICATION AND PHYLOGENY The accompanying phylogeny of the within the Crossopterygii have been re- Coelacanthini (Fig. 10) is simply a very cently summarized by Moy-Thomas (1939) tentative attempt to indicate the relation- and will only be incidentally mentioned ships between certain of the better known here. genera. As Stensi6 (1932) has pointed out, The Upper Devonian forms from Ger- certain of the genera do fall together in many, Diplocercides, Nesides, and Euporo- natural groups, possibly into families or steus, are certainly a natural and primitive superfamilies. The characters which es- group. The neurocranium is ossified in tablish the coelacanths as a separate order two separate parts and a true basipterygoid 14 AMERICAN MUSEUM NOVITATES [No. 1110 process is present on the basisphenoid. coelacanths flowered in the Triassic. The Both of these characters are present in the number of known distinct Triassic genera is rhipidistians. Furthermore, the presence certainly greater than in any other period of a horizontal limb on the clavicle is a (Fig. 11). The forms from Spitzbergen primitive rhipidistian feature not found in Axelia, Sassenia, Scleracanthus, and Myla- later coelacanths. canthus are apparently a natural group Moy-Thomas (1937) has done a great branching off the line leading to the later service by revising the Carboniferous and Mesozoic forms. There are numerous de- Permian coelacanths of the British Isles tails of the neurocranium which demon- and thus eliminating a vast amount of strate this relationship (Stensi6, 1921, pp. synonymy. Although all the Carboniferous 120-129), and also the close affinities of forms have been referred to the genus this entire group to Coelacanthus on the Rhabdoderma, Moy-Thomas himself ad- one hand and the later Mesozoic genera on mits that in the future this genus may have the other. to be subdivided into several genera. This Scleracanthus has a dentition adapted group may then constitute a family. Rhab- for crushing. It is closely related to Axelia doderma illustrates a further step in the but has a more robust pterygoid, more evolution of the order in the employment specialized dentition and differently shaped of the antotic process (extending laterally scales. Mylacanthus is closely related to from the dorsal border of the basisphenoid) Axelia in having a less specialized denti- as the articular contact of the neurocranium tion. Heptanema is very difficult to distin- with the metapterygoid. There is some re- guish (Moy-Thomas, 1935) from Scleracan- duction in the size of the basipterygoid thus, while the latter is also considered by process. The neurocranium is subdivided the same author to be very close to Whitea. into a series of ossifications instead of two A basic character uniting this Axelia group major ones. All later coelacanths have a is the shape and configuration of the ptery- similarly constructed neurocranium. The goquadrate complex. In Axelia, the ptery- pelvis of Rhabdoderma is the most ancient goid is large and thick with a very broad known and appears to represent the primi- posterior vertical limb and a long, thin tive coelacanthid type, showing very dis- anterior horizontal limb. In Mylacanthus tinct but derivable differences from rhipi- the posterior limb is narrow but longer; in distians. The same is true for the endo- Scleracanthus it is also long but wider. skeletal supports of the other fins, paired The latter is also true of Whitea and Hep- and unpaired (except possibly the basal tanema. The metapterygoid and quad- plate of the first dorsal). rate are strongly and independently ossi- Now that the genus Coelacanthus has fied in this group. been properly defined (Moy-Thomas and Westoll, 1935), it may, in the future, form Wimania, while it may belong to the the nucleus of a Permian family. It is fur- Axelia group, shows more definite affinities ther evolved from the Rhabdoderma stage to Coelacanthus and Rhabdoderma and may by the loss of the basipterygoid process and have to be placed in a separate family. the presence of only the antotic process. This conclusion is based on the shape of the This condition holds for all later coela- parasphenoid, the pterygoid, the cheek canths. Coelacanthus, although more pro- plates, and the large opercular. gressive, differs markedly from Rhabdo- Stensi6 considers Laugia to be very sug- derma only in the presence of an extraclei- gestive of Coelacanthus, the neurocranium thrum. Spermatodus appears to be most exhibiting about the same degree of ossi- closely related to Coelacanthus (Moy- fication. He points out that the neuro- Thomas, 1939). Besides other similarities cranium is certainly intermediate between in the construction of brain case, both show the condition found in Diplocercides and a reduction in the ossification of the orbital such later forms as Axetia and Macropoma. region. In a number of the characters of the post- Our present knowledge indicates that the cranial skeleton, previously enumerated, ; as eE L. 1; r.-4 0, m a -, .z 16 AMERICAN MUSEUM NOVITATES [No. 1110

Laugia is quite specialized and must repre- Latimeria is worthy of further comment, as sent an offshoot from the Coelacanthus type it not only articulates in the usual manner in the Permian. with the quadrate, but also separately with Osteopleurus resembles Coelacanthus with the symplectic. Smith points out that two respect to the postcranial skeleton except distinct sockets can be observed in the for the apparent very great difference in lower jaw of Diplocercides, and he has tenta- the shape of the pelvic plates. The former tively identified the symplectic in Wimania must represent a line isolated from the and Axelia, indicating a similar type of jaw more central types in the Permian. suspension in these forms. That this con- Coccoderma, Libys, and Undina appear dition is common to all coelacanths is to be rather closely related. For instance, doubtful as there is no evidence of its exist- in each the dermopalatine is an independ- ence in the well-known Macropoma (Wat- ent bone along the anterior limb of the son, 1921) or in Undina. The symplectic pterygoid. In other forms it is fused with bone likewise articulates with the lower the pterygoid. Other details of the skull jaw in Amia, making a joint which is func- also bear out this conclusion. The direct tionally more important than that formed derivation of this group from the Coelacan- with the quadrate. There is no evidence, thus type is indicated not only by the con- however, to support Smith's suggestion struction of the skull, but also by the de- that this relationship between the mandibu- tails of the postcranial skeleton. lar and hyoid arches in Latimeria is "the There are a number of minor details in typically primitive one." It appears quite the construction of the dermal cranial roof certain that this double suspension of the which indicate relationship between Ma- lower jaw has been independently acquired cropoma and the Axelia group. For in- in the coelacanths and the amioids. stance, Macropoma and Whitea have a very The hyomandibular, which is very well similar supraorbital series. The shape of developed in Latimeria, has been generally the pelvic plates is another example. considered to be reduced in all coelacanths. Mawsonia seems to be most closely re- Smith considers an element in Wimania, lated to Undina, more distantly to Macro- identified by Stensi6 as an epihyal, to be poma. Both have the metapterygoid par- an ossified portion of the hyomandibular. tially fused with the pterygoid. Other de- In any case, there is now good evidence for tails of the skull are very similar. believing that the coelacanth hyomandibu- Latimeria is considered by Smith to be lar was much larger than has been formerly more closely related to a Rhabdoderma- supposed, although it may have been Wimania type than to Macropoma, al- mostly cartilaginous as in Latimeria. though it bears a strong superficial resem- The hyomandibular, representing the more blance to the latter. This conclusion ap- dorsal part of the same embryonic mass pears to be supported by the pattern of the which gives rise to the symplectic, neces- dermal cheek bones, the method of lower sarily supports the latter and hence could jaw suspension, and the presence of an ex- not have been reduced, at least in the tracleithrum. forms having the double suspension of the The method of lower jaw suspension in lower jaw.

REFERENCES AGASSIZ, L. BROOM, R. 1833-44. Recherches sur les Poissons Fos- 1905. On a Species of Coelacanthus from the siles, II, pp. 168-180. Upper Beaufort Beds of Aliwal North: ALDINGER3.H. Rec. Albany Mus. Grahamstown, So. 1930. Uber das Kopfskelett von Undina Africa, I, pp. 338-339. acutidens Reis und den kinetischen BRYANT, W. L. Schiadel der Coelacanthiden: Cen- 1919. On the Structure of Eusthenopteron: tralb. f. Min., Geol., Psa.. Abt. B Bull., Buffalo Soc. of Nat. Sciences, No. 1, pp. 22-48. XIII, No. 1, pp. 1-23. 19411 REVISION OF COELACANTHUS NEWARKI 17

1934. New Fishes from the Triassic of REIs, 0. M. Pennsylvania: Proc. Amer. Phil. 1888. Die Coelacanthinen, Palaeontograph- Soc., LXXIII, No. 5, pp. 319-326. ica, XXXV, pp. 1-94. COPE, E. D. SIMPSON, G. G., AND ROE, ANNE 1873. On some new Batrachia and fishes 1939. Quantitative Zoology, McGraw-Hill, from the Coal-measures of Linton, New York, 414 pp. Ohio: Proc. Acad. Nat. Sci. Phila., SMITH, J. L. B. pp. 340-343. 1940. A Living Coelacanthid Fish from GOODRICH, E. S. South Africa: Trans. Roy. Soc. of 1901. On the Pelvic Girdle and Fin of So. Africa, XXVIII, pt. 1, pp. 1-106. Eusthenopteron: Quart. Journ. Micr. STENSI1, E. A. Sci., (N. S.) XLV, pp. 311-324. 1921. Triassic Fishes from Spitzbergen: GROSS, W. Adolph Holzhausen, Vienna, pt. 1, 1936. Beitrage zur Osteologie baltischer und pp. 1-307. rheinischer Devon-Crossopterygier: 1932. Triassic Fishes from East Greenland: Pal. Zeit., Bd. 18, Nr. 3-4, pp. 129- Medd. Om. Gronland, XXXVIII, 155. No. 3, pp. 1-305. HUSsAKOF, L. 1937. On the Devonian Coelacanths of 1911. The Permian Fishes of North Amer- Germany with Special Reference to the ica: Carnegie Instn. Wash., No. 146, Dermal Skeleton: Kung. Svenska pp. 155-175. Veterskaps Handling, XVI, No. 4, pp. HUXLEY, T. H. 1-56. 1866. Figures and Descriptions illustrative 1940. Uber die Fische des Devons von of British Organic Remains: Mem. Ostgr6nland: Mitt. Der Naturf. Geol. Surv., dec. 12, pp. 1-45. Gesell., Bd. 16, Nr. 6, pp. 132-137. LAMBE, L. M. TRAQUAIR, R. H. 1916. Ganoid Fishes from near Banff, Al- 1905. Notes on the Lower Carboniferous berta: Trans. Roy. Soc. Canada, Fish of Eastern Fifeshire: Proc. (3) VIII, pp. 35-44. Roy. Phys. Soc. Edinburgh, XVI, pp. Moy-THOMAs, J. A. 80-86. 1935. The Coelacanth Fishes of Madagascar: WATSON, D. M. S. Geol. Mag., LXXII, No. 851, pp. 1921. On the Coelacanth Fish: Ann. and 213-227. Mag. Nat. Hist., (9) VIII, pp. 320-337. 1937. The Carboniferous Coelacanth Fishes WESTOLL, T. S. of Great Britain and Ireland: Proc. 1939. On Spermatodus pustulosus Cope, A Zool. Soc. London, Ser. B, pt. 3, pp. Coelacanth from the Permian of 383-415. Texas: Amer. Mus. Novit., No. 1017, 1939. Palaeozoic Fishes, Chem. Pub. Co., pp. 1-23. Inc., New York, 149 pp. WOODWARD, A. S. MoY-THOMAS, J. A., AND WESTOLL, T. S. 1891. Catalogue of the Fossil Fishes in the 1935. On the Permian Coelacanth, Coela- British Museum, part 3, 567 pp. canthus granulatus, Ag.: Geo. Mag., 1895. The Fossil Fishes of the Talbragar LXXII, No. 856, pp. 446-457. Beds (Jurassic?): Mem. Geo. Surv. NEWBERRY, J. S. of N. S. W., Paleontology, No. 9, pp. 1878. Descriptions of New Fossil Fishes from 1-27. the Trias: Ann. N. Y. Acad. Sci., I, 1908. On Some Fossil Fishes from the Cre- pp. 127-128. taceous Formation of Bahia, Brazil. 1888. Fossil Fishes of the Triassic Rocks of Quart. Jour. Geo. Soc., LXIV, pp. New Jersey and the Connecticut 358-362. Valley: Monographs of the U. S. ZITTEL, K. A. Geological Survey, XIV, part 2, pp. 1932. Text-Book of Palaeontology. II, 17-76. Macmillan, London, 464 pp.