American Museum Published by the American Museum of Natural History Central Park West at 79Th Street New York, N.Y

NovitatesAMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET NEW YORK, N.Y. 10024 U.S.A. NUMBER 2658 OCTOBER 10, 1978

BOBB SCHAEFFER AND WALTER W. DALQUEST A Palaeonisciform Braincase from the Permian of Texas, With Comments on Cranial Fissures and the Posterior Myodome

ovitcttesAMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2658 pp. 1-15, figs. 1-9 October 10, 1978

A Palaeonisciform Braincase from the Permian of Texas, With Comments on Cranial Fissures and the Posterior Myodome

BOBB SCHAEFFER1 AND WALTER W. DALQUES-F

ABSTRACT A single, large palaeonisciforn braincase from This specimen has also occasioned a review of the the Lower Permian Lueders Formation, Baylor actinopterygian cranial fissures and the posterior County, Texas, is described as a new genus and myodome. species within the Palaeoniscifonnes, incertae sedis. INTRODUCTION The completely ossified, sutureless braincase was presented by Emsoff and Walter Dalquest characteristic of most paleopterygians (the to the American Museum of Natural History. Chondrostei of most classifications) and of a Preparation of the braincase, by Walter Soren- few neopterygians has been described in vary- sen, proved difficult and tedious. Various tech- ing detail in about 15 genera. Although this niques were used to expose the specimen, braincase type is generally regarded as primi- which is complete and little damaged except for tive for the Actinopterygii, there are frequently dorsoventral compression that has caved in the differences in general proportions and in the fossa bridgei and produced an up-warping of development of certain characters such as the the ethmoid region. As acid preparation proved basipterygoid process, the fossa bridgei, the to be inadvisable, the braincase was cut mid- posterior myodome, the ventral otic and otico- sagittally to reveal some of the internal struc- occipital fissures and the vestibular fontanelle tures. that give the braincase of a particular taxon a In addition to describing the braincase and distinctive appearance. associated rostral bone (fig. 1), we have briefly In 1973 Ted Emsoff, a student at Mid- discussed two problems involving the paleo- western University, discovered a large (150 pterygian neurocranium that have recently re- mm. long), massively ossified neurocranium, ceived considerable attention. One relates to the apparently palaeoniscoid, and an associated me- position and significance of the ventral otic fis- dian rostral element, in the Lower Permian sure and the other to the origin and history of Lueders Formation near the Lake Kemp Dam the posterior myodome. The systematic position in Baylor County, Texas. This unique specimen of the Lueders palaeoniscoid remains obscure,

'Curator, Department of Vertebrate Paleontology, the American Museum of Natural History. 2Professor of Biology, Midwestern State University.

art sfbr, articular facet for 1st suprapharyngobranchial asp, ascending process of parasphenoid bhc, bucco-hypophysial canal bls, bony lamina overhanging saccular recess bocc, basioccipital ossification bpt, basipterygoid process bsph, basisphenoid ossification cro, occipital crest (ridge) crsp, cranio-spinal process eff art, opening for efferent branchial arteries erm, external rectus muscle fhym, hyomandibular facet fm, foramen magnum font, dorsal fontanelle (covered by dermal bone) fontv, vestibular fontanelle fos, otico-sphenoid fissure fotc, otico-occipital fissure fotv, ventral otic fissure goa, groove for orbital artery hv, head vein ic, foramen for internal carotid artery ifv, inferior orbital vein ios, interorbital septum juc, jugular canal and groove not, notochordal canal or pit ?np, ?nasal pit occa, foramen for occipital artery opa, ophthalmic artery groove ora, foramen for orbital artery pas, parasphenoid pf, pituitary fossa pmy, posterior myodome prb, prootic bridge popr, postorbital process ?ptr, ?postrostral pv, pituitary vein FIG. 1. Luederia kempi, new genus and species, sef, sub-epiotic fossa AMNH 8998. Radiogram of neurocranium and asso- spic, opening for spiracular canal ciated rostral element in dorsal aspect prior to prepa- sv, sacculus vasculosis ration. Xl. utr, utricular recess I, canal for olfactory nerve but on the basis of present evidence it has II, foramen for optic nerve several unique derived characters in both the III, foramen for oculomotor nerve braincase and the rostral element that justify IV, foramen for trochlear nerve formal taxonomic treatment. V and VII op sup, fossa with foramina for superficial ophthalmic nerves ABBREVIATIONS VII hym, foramen for hyomandibular branch of facial nerve ANATOMICAL IX, foramen for glossopharyngeal nerve aort, aortic canal X, foramen for vagus nerve aps, foramen for pseudobranchial artery art ifbr, articular facet for 1st infrapharyngobranchial INSTITUTION art pq, articular facet for autopalatine AMNH, the American Museum of Natural History 1978 SCHAEFFER AND DALQUEST: PALAEONISCIFORM BRAINCASE

ACKNOWLEDGMENTS branch of the facial nerve and associated blood vessels. The interorbital septum is complete We are indebted to Dr. Cecile Poplin and from the ethmoid area to the basisphenoid pillar Dr. Colin Patterson for helpful comments and the posterior orbital wall. The canal for the provided during the preparation of the present olfactory nerves thus passes through the entire paper. The drawings were made by Ms. Lor- length of the septum. The post-mortem com- raine Meeker and the photographs were taken pression has obscured and distorted much of by Mr. Chester S. Tarka. the detail on the posterior face of the ethmoid DESCRIPTION and in the orbitotemporal region including the posterior wall of the orbit. There is no sure In its general topography and proportions, evidence of anterior myodomes or of canals for including the relative size of the basipterygoid the ethmoid ramifications of the trigeminal and processes and the size of the ascending pro- facial nerves, as identified by Nielsen (1942, cesses of the parasphenoid, the Lueders brain- fig. 65) in Boreosomus and by Poplin (1974, case (figs. 2-7) resembles the neurocranium of figs. 18, 19) in Kansasiella. Pteronisculus (Glaucolepis) as described and The basipterygoid process is relatively long figured by Nielsen (1942), Lehman (1952), and deep as in Pteronisculus, and has both Beltan (1968), and Patterson (1975, p. 461). It dermal and endochondral components. The fo- differs from the Pteronisculus neurocranium in ramen for the efferent pseudobranchial artery is the complete ossification of the interorbital sep- situated on the anterior face of the process tum, in the near absence of a vestibular fon- laterally to the groove for the ophthalmic tanelle, in the partial closure of the lateral otic artery. As the interorbital septum is completely fissure, and in the relations of the ventral otic united to the basisphenoid pillar, the optic fora- fissure to certain neighboring structures and fo- men (which is vaguely evident) is divided ramina (see p. 5). where it exits from the braincase. It is obvious The dorsal surface of the neurocranium is from the midsagittal section that the posterior covered with thick, rugose or tuberculated der- myodome is unpaired; its opening into the orbit mal bones. Sutures are not visible in spite of is also well defined. The position of the jugular careful cleaning, and it is possible that some of canal can be accurately located but the details the dermal elements are co-ossified. The sepa- of the trigeminofacialis chamber are obscure. A rate rostral bone is finely tuberculated and there depression in the postorbital wall, dorsal and is evidence of similar ornamentation above the somewhat medial to the trigeminofacialis cham- orbits. A dorsal fontanelle in the neurocranium ber, presumably includes the foramina for the is indicated by an additional depression in the superior ophthalmic and trochlear nerves. In a dermal roof between and behind the orbits (for photograph of the postorbital wall in Ptero- comparison see Nielsen, 1942, fig. 21). The nisculus magnus (Nielsen, 1942, pl. 6, fig. 2) anterodorsal surface of the ethmoid area is there is an unidentified depression just behind partly exposed but the paired depressions iden- the trochlear foramen that resembles the one in tified by Nielsen (1942, fig. 7) as nasal pits the Lueders braincase. (fossae) are not evident. The anterior part of the optic region is domi- The ethmoid articular processes for the auto- nated by the broad and high ascending process palatines are unusually well developed for a of the parasphenoid, which extends upward to paleopterygian. Grooves situated anterior and the ventral exit of the spiracular canal. The last dorsal to these processes presumably housed typically opens into the anterior part of the the maxillary branch of the trigeminal nerve fossa bridgei, in this specimen almost elimi- and related structures (Nielsen, 1942, fig. 6). nated by the post-mortem compression. The ar- On the ventral surface of the ethmoid region ticular facet of the hyomandibular has a nearly paired grooves separated by an elevation of the vertical orientation, indicating that the suspen- nasal septum may have carried the palatine sorium was oblique as in many other palaeo- 4 AMERICAN MUSEUM NOVITATES NO. 2658

FIG. 2. Luederia kempi, new genus and species, AMNH 8998. Neurocranium in dorsal, ventral, and lateral views. x 1.25.

niscoids. The posterior opening of the jugular of the articular protuberance for the first in- canal and the jugular groove behind the lower frapharyngobranchial (see Nielsen, 1942, fig. 6; part of the hyomandibular facet are also typ- Gardiner and Bartram, 1977, fig. 2), and which ically situtated. Midway along the posterior may have continued anteriorly as the internal border of the ascending process there is a fora- carotid. The orbital artery (the external carotid men that presumably served as the entrance to of Nielsen, but see Poplin, 1974, p. 92) has a the elongated orbital artery, which occupied the separate opening into the jugular canal from well-defined groove bordering the medial side which it presumably passed into the trigemi- 1978 SCHAEFFER AND DALQUEST: PALAEONISCIFORM BRAINCASE 5

FIG. 3. Luederia kempi, new genus and species, AMNH 8998. Key diagrams for figure 2. nofacialis chamber and the orbit. A small fora- more or less covers the "triangular area" be- men below and somewhat behind the orbital tween the grooves for the lateral aortae, as in foramen may have been the exit for the hyo- Pteronisculus (Patterson, 1975, p. 538), the mandibular branch of the facial nerve. ventral fissure (as revealed in sagittal section) is The parasphenoid slightly overlaps the ven- in the same relative position as in other palaeo- tral otic fissure and has a relatively greater niscoids. Because of this extension, however, extension behind the basipterygoid processes the articulations for the first infrapharyngo- than in, for instance, Pteronisculus. Although it branchials are lateral to the parasphenoid rather 6 AMERICAN MUSEUM NOVITATES NO. 2658

efferent branchial arteries, as in Pteronisculus stensioi (Nielsen, 1942, fig. 6). Apparently there was considerable variation among the palaeoniscoids in the number and arrangement of the openings for the efferent branchials (Poplin, 1975). In lateral aspect, the otic region has two opa prominent topographic features-the tuberosity that carries the articulation for the first suprapharyngobranchial and the robust craniospinal process that includes the articulation for the second suprapharyngobranchial. Below the tuberosity, on the right side, the infravagal por- FIG. 4. Luederia kempi, new genus and species, tion of the otico-occipital fissure is open and AMNH 8998. Cross section of neurocranium at level lined with perichondral bone. It terminates ven- of postorbital wall. trally in the much-reduced vestibular fontanelle. Above the tuberosity, and on the left side, the otico-occipital fissure appears to be closed, sug- gesting that this is an aging phenomenon. The location of the glossopharyngeal foramen is not evident, but it must be nearly in line with the jugular depression and the vagal foramen. In midsagittal section (fig. 6), the posterior myodome, the prootic bridge and the basisphenoid ossification are well defined. The ventral otic fissure, which is typically situated behind the posterior myodome, is not lined with perichondral bone and, as preserved, is filled with cancellous bone. The significance of the filling is obscure; in any case, the fissure does not join the vestigal vestibular fontanelle. Various sef cro other features visible in the midsagittal plane have been distorted by the post-mortem compression. fhymK In posterior aspect it is evident that the cranial roof has been forced downward to the thick, projecting rim of the foramen magnum. This rim is more or less confluent with that of art ifbr not the deep notochordal pit. On both lateral sur- faces of the rim there is an oblique groove and cort bpt related foramen for an occipital artery. The most unusual aspect of the posterior occipital surface is the presence of paired depressions on either side of the foramen magnum that may be FIG. 5. Luederia kempi, new genus and species, AMNH 8998. Neurocranium in view. regarded as sub-epiotic fossae.1 They are bor- posterior dered dorsally by the horizontal occipital pro-

than behind it. Paired openings for efferent 'According to Patterson (personal cornmun.) these depres- branchial arteries are situated near the posterior sions are the sub-epiotic fossae of Phillips (1942), not the border of the parasphenoid, and behind them post-temporal fossae, which lie anterior and lateral to the there is a median, slotlike aperture, also for otico-occipital fissure. 1978 SCHAEFFER AND DALQUEST: PALAEONISCIFORM BRAINCASE 7 tuberances (see Poplin, 1974, figs. 11 and 13) ated immediately in front of the ethmoid (fig. and ventrolaterally by strongly developed cra- 1). Notches on both lateral borders lead into a niospinal processes. The occipital crest separat- transverse canal-presumably the ethmoid com- ing the two fossae above the foramen magnum missure. The outer, anterior surface is finely has been mostly eliminated by the depression denticulated like the anterior part of the skull of the cranial roof. It is probable, however, roof. The median (or inner) surface is divided that the distance from the dorsal rim of the by a strongly developed horizontal lamina. foramen magnum to the posterior edge of the There is some indication that this shelf repre- roof approached that in Pteronisculus (Nielsen, sents fused, paired elements, although there is 1942, p. 32) and in Boreosomus (Nielsen, no indication of such fusion elsewhere (see also 1942, fig. 61). Patterson, 1975, fig. 138). Above the shelf The associated noncranial rostral element there is a shallow concavity that may have (fig. 7) is a tooth-bearing bone that was situ- formed the anterior wall of the nasal capsules.

fm bls

~~~~~~~~~~~/~ -zs~

not

,~Ir -~ -

I - / vi I1 pmy fotv aort

aort fotv pmy ic pas

FIG. 6. Luederia kempi, new genus and species, AMNH 8998. Sagittal section of neurocranium. Most of the interorbital septum is included in the left half. X2. 8 AMERICAN MUSEUM NOVITATES NO. 2658

FIG. 7. Luederia kempi, new genus and species, AMNH 8998. Rostral element in anterior, lateral and posterior (intemal) views. x3.4.

Below the shelf, and fused to it, are a pair of SYSTEMATICS robust, tusklike teeth that are separated from each other by a deep, dorsally directed con- On the basis of available evidence it is not cavity. The spacing of these enlarged teeth also possible to formulate a hypothesis of relation- favors the possibility that the shelf, at least, ship for the fish represented by the Lueders represents a fusion of paired elements. The braincase and rostral element. It is probably a smaller teeth along the ventral margin of the palaeoniscoid, which means only that it has denticulated outer surface are entirely anterior numerous primitive actinopterygian neu- to the much larger paired tusks. Because of rocranial characters, including a sutureless crushing on the upper part of the ethmoid re- braincase, a notochordal canal that extends for- gion, the exact relationships of the rostral bone ward to the ventral otic fissure, and an un- to other dermal elements or to the neu- divided, oblique or nearly vertical hyoman- rocranium cannot be determined. dibular facet. However, comparison at a The homologies of this element are indeed broader level, with other paleopterygian neu- problematical, particularly in the absence of the rocrania, supports the opinion that it has certain other dermal snout bones. Because it is me- unique derived characters that favor formal tax- dian, tooth-supporting, canal-bearing and shows onomic treatment, as follows: no evidence of an orbital margin, identification in terms of Gardiner's (1963) criteria is difficult. As the bone to a CLASS OSTEICHTHYES belonged mature indi- SUBCLASS ACTINOPTERYGII vidual, fusion of premaxillae is possible. ORDER PALAEONISCIFORMES (PALAEONISCIDA), Patterson (1975, fig. 138) has noted this condi- INCERTAE SEDIS tion in an East Greenland specimen of Per- LUEDERIA,1 NEW GENUS leidus. An additional "fusion," with the canal- bearing rostral, could produce a single, median TYPE SPECIES: Luederia kempi. rostropremaxilla. The Lueders element also re- DISTRIBUTION: Baylor County, Texas. sembles the median rostrodermethmoid of Lower Permian (Leonardian) Lueders Forma- Pachycormus curtus (Patterson, 1975, fig. 139), tion, Maybelle Aa2 Limestone Member. but as the presence of separate premaxillae can- DIAGNOSIS: Large palaeonisciform approx- not be demonstrated in the former, further speculation about its identity seems pointless. 'For the Lueders Formation. 1978 SCHAEFFER AND DALQUEST: PALAEONISCIFORM BRAINCASE 9 imating size of Scanilepis dubia (Woodward) ingly, we are following Miles (1973, p. 74) in on basis of braincase dimensions. Neu- grouping the acanthodians and osteichthyans in rocranium differs from that of other palaeo- the category Teleostomi. In this same paper, nisciforms in having well-defined ethmoid Miles (pp. 85-89) discussed the components of articular processes for the autopalatines, aortic the Acanthodes basicranium (also Moy-Thomas grooves, and articulations for the first in- and Miles, 1971, p. 66) and compared the fis- frapharyngobranchials situated lateral to the sures between these components with the lateral posterior extension of the parasphenoid, well- occipital fissure and the ventral otic fissure of defined sub-epiotic fossae that are bordered primitive actinopterygians and crossopterygians. ventrolaterally by strong craniospinal processes, The possibility of inferring some embryonic and in having an associated median rostral ele- chondrocranial features from the braincase of ment that is tooth-supporting, canal bearing, an extinct, adult osteichthyan depends on the and without orbital margins. presence of morphological landmarks that can also be identified in the developing and mature Luederia kempi,1 new species neurocrania of related living forms. The Figures 1-7 basitrabecular (basipterygoid) process, the lat- HOLOTYPE: AMNH 8998, complete brain- eral commissure, the ventral otic and otico- case and associated median rostral element. occipital (metotic) fissures, the vestibular fon- Only known specimen. tanelle and various canals and foramina (along HORIZON AND LOCALITY: Lueders Forma- with their topographic relationships) represent tion, Maybelle Aa2 Limestone Member, Lower such landmarks. The developmental basis for Permian (Leonardian), about 200 m. south of differences in the relationships of these fissures south end of the Lake Kemp Dam, Baylor with each other and with the vestibular fon- County, Texas (locality 2 of Berman, 1970, fig. tanelle in adult extinct paleopterygians and 2). early neopterygians remains problematical, but DIAGNOSIS: Same as for genus. some inferences can be derived from living forms. A perusal of DeBeer (1937) indicates COMMENTS ON THE CRANIAL that the lateral auditory wall in the vicinity of FISSURES AND THE POSTERIOR the foramina for the facial, glossopharyngeal, MYODOME and vagus nerves has a fairly uniform development. The cartilaginous wall becomes per- In order to distinguish between the primitive ichondrally and sometimes endochondrally and the derived aspects of the palaeoniscoid ossified without any significant changes in to- braincase, outgroup comparison at various lev- pography except for those related to increase in els of universality is obviously necessary. In size. this regard Patterson (1975) has provided a de- Romer (1937, pp. 39-43) was apparently the tailed comparison of paleopterygian and neop- first to deduce the embryonic components of an terygian neurocrania with emphasis on the extinct osteichthyan neurocranium, in this in- origin and fate of the ventral otic fissure, the stance for the rhipidistian Ectosteorhachis otico-occipital fissure, the vestibular fontanelle (non-Megalichthys). He concluded that the tra- and the posterior myodome. More recently Gar- beculae formed the basal part of the eth- diner and Bartram (1977) have discussed the mosphenoid moiety and the parachordals the homologies of the ventral otic fissure in pal- same for the otico-occipital portion. Romer also aeoniscoids and in the rhipidistian Eusthenop- compared the Ectosteorhachis braincase with teron and Glyptolepis. Although the relation- Watson's (1925) "Palaeoniscoid A" and "Pa- ships of the acanthodians are still somewhat laeoniscoid B," which have been subsequently controversial (Jarvik, 1977), we believe that the redescribed by Poplin (1974). Although Watson hypothesis of a sister group relationship with (1925, p. 837) figured the ventral otic fissure of the Osteichthyes has not been falsified. Accord- "B," this fissure was not discussed by him, and Romer (1937, p. 57) made no comparison 1For Lake Kemp, Baylor County, Texas. with the rhipidistian condition. Jarvik (1954, p. 10 AMERICAN MUSEUM NOVITATES NO. 2658

7), as pointed out by Gardiner and Bartram Beer, 1937, p. 391), the persistent ventral otic (1977), apparently confused the ventral otic fis- fissure in adult teleostomes may be behind sure in Pteronisculus (Nielsen, 1942) with a rather than at the actual embryonic chordal- vaguely defined, inconstant furrow in Eus- prechordal junction. As noted by Patterson, the thenopteron that runs transversely from the ves- fissure cartilage may ossify in some specimens tibular fontanelles to the U-shaped gap in the of Boreosomus and this may be the case in floor of the notochordal canal. Schaeffer (1968) Luederia. The fissure may also disappear in and Gardiner and Bartram (1977) have sug- Saurichthys (Stensi6, 1925), in Perleidus (Leh- gested that the lower part of the intracranial man, 1952), in the parasemionotids, in some joint in crossopterygians is the developmental caturids, in Dapedium and in "early" pho- homologue of the ventral otic fissure in the lidophorids and leptolepids (Patterson, 1975). primitive actinopterygians for reasons that are In some forms the closure of the ventral otic discussed in their respective papers. fissure may be age-related, but out-group com- In Polypterus, Acipenser, Lepisosteus, parison suggests that the presence of cartilage Amia, and various teleosts the anteroventral in the adult fissure is primitive. In advanced part of the metotic fissure (which housed the actinopterygians the prootics have grown down- vagus nerve) becomes incorporated into the ward between the basisphenoid and the basioc- basicapsular fenestra through the orientation of cipital, but the fissure persists as a synchondro- the basicapsular commissure as it fuses with the sis between the basioccipital and the prootic parachordal (DeBeer, 1937, p. 399). During according to Patterson (1975). In regard to the subsequent growth of the auditory capsule, of acanthodians, Miles (1973, p. 89; 1977, p. 49) the basal plate and of the anterior and posterior has noted that the cleft between the posterior basicapsular commissures, the foramina for the margin of the basisphenoid and the middle ven- facial, glossopharyngeal and vagus nerves ac- tral ossification in Acanthodes is top- quire their adult spatial relationships. Although ographically in the position of the actino- these foramina are usually well separated, the pterygian ventral otic fissure (see also Patter- ninth and tenth cranial nerves in, for example, son, 1975, pp. 537-538). Although this deduc- Clupea (?harengus) emerge through the same tion is based only on external molds of the foramen, and the posterior basicapsular com- Acanthodes braincase, it is the most par- missure is apparently absent (DeBeer, 1937, p. simonious explanation. As preserved in the 132). The anterior and posterior commissures negative, the several components of the brain- obviously play an important role in the devel- case in Acanthodes were presumably separated opment of the lateral auditory wall. Although by areas of cartilage; there is no evidence of they cannot be observed per se in the adult fusion between them. It is also possible that the chondrified or ossified neurocranium, their con- cartilage-filled ventral otic fissure opened into dition in the developing chondrocranium (in- the rear of the orbit in Acanthodes, as in cluding their absence) can frequently be Mimia. At least there is no ossification that inferred by the relationships of the relevant fo- would prevent this in the adult Acanthodes ramina and fissures. (Miles, 1973, fig. 4). The ventral otic fissure in adult palaeo- The otico-occipital fissure requires only brief niscoids (fig. 6) is usually interpreted as a per- comment as we are not concerned here with sistently cartilaginous zone between the ways in which it is eliminated in higher ac- basisphenoid and basioccipital ossifications (the tinopterygians. It represents the embryonic fissure is thus a synchondrosis), which may be metotic fissure between the auditory capsule near or at the junction of the prechordal and and the occipital arch, and its position is read- chordal components of the basal plate ily determined by the location of the vagal (Schaeffer, 1968, p. 13; Gardiner, 1973, p. 106; foramen, which in living immature actino- Patterson, 1975, p. 416, 466; Gardiner and Bar- pterygians is always situated in the fissure. As tram, 1977, pp. 239-240, 241-242). However, in actinopterygians that retain the otico-occipi- if the basipterygoid processes mark the pos- tal fissure in the adult stage, the apparently terior limit of the prechordal basicranium (De- homologous fissure in Acanthodes may be lined 1978 SCHAEFFER AND DALQUEST: PALAEONISCIFORM BRAINCASE I I with perichondral bone (Miles, 1973, p. 89; ossified auditory wall and the related foramina. 1977, p. 49) which indicates the absence of In Patterson's (1975, p. 416) opinion, the cartilage. In crossopterygians, the auditory cap- primitive state for the actinopterygian ventral sule and the occipital arch are generally fused otic fissure includes; (1) a bony separation be- in the mature neurocranium, although Jarvik tween it and the otico-occipital fissure; (2) con- (1954, p. 7) has found a more or less separate tinuation of the fissure cartilage to the orbit, supravagal portion in some specimens of Eu- and (3) a position definitely anterior to the sthenopteron, and there is indication of a fis- vestibular fontanelle. This hypothesis is diffi- sure around the vagal foramen in cult to test by comparison with the rhipidistian Ectosterorhachis (Romer, 1937, fig. 21). The condition because of the marked differences in otico-occipital fissure and the ventral otic fis- the cranial wall around the trigeminal nerve. In sure are confluent and cartilage-filled in the the rhipidistians, the ventral otic and otico-oc- Devonian dipnoans from Australia (Miles, cipital fissures are well separated, and the ven- 1977, p. 49), but the vestibular fontanelle is tral otic fissure is some distance in front of the absent. vestibular fontanelle (Jarvik, 1954, fig. 11). The vestibular fontanelle, derived on- These modifications are presumably related to togenetically from the basicapsular fenestra, is the intracranial joint rather than only to os- present in several groups of adult living teleosts sification of the anterior basicapsular com- as a membrane- or cartilage-filled space be- missure and related areas. The otico-occipital tween the otic and occipital ossifications (Pat- fissure and the fontanelle are absent in the ac- terson, 1975, p. 428). Closure of the fontanelle tinistians. The presence of a continuous cranial with cartilage must mean that this hiatus is fissure, without an intervening fontanelle in eliminated before ossification begins, and in some Devonian dipnoans (Miles, 1977) neither this case, that the fontanelle represents a part of falsifies nor corroborates the above hypothesis. the chondrocranial wall that has failed to os- Patterson (1975) has rightly emphasized that sify. The absence of perichondral bone around the ventral otic fissure and the otico-occipital the edge of the vestibular fontanelle, as in fissure have separate origins and different fates, Pachycormus (Patterson, 1975, fig. 106), im- which is generally true for the teleostomes. The plies that this area was filled with cartilage. In ultimate reason for the persistence or disap- Perleidus the vestibular fontanelle is variously pearance of a particular fissure, or part thereof, developed in the adult; it was presumably car- and of the vestibular fontanelle, may represent tilage filled when not closed by bone (Lehman, a compromise between the constraints of phy- 1952). Patterson has noted that the differences logeny, the demands of function, and the in the expression of the fissure and the fon- effects of aging. tanelle may be, in part, age related, which A particularly interesting aspect of the prim- seems to be the case in Luederia. Incidentally, itive actinopterygian basicranium involves the the term "fontanelle" is usually defined as a relative position of the ventral otic fissure. Gar- membrane-covered (but not cartilage filled) diner (1973, p. 109), Patterson (1975, p. 543), space in the chondrocranium or between neigh- and Gardiner and Bartram (1977, p. 242) have boring dermal or endoskeletal ossifications. suggested that an increase in the size of the As implied above, "confluence" of the ven- posterior myodome is responsible for a pos- tral otic fissure or of the otico-occipital fissure terior migration of this fissure. In a succinct with the vestibular fontanelle means only that summary of posterior myodome history, Patter- they are not separated by areas or bridges of son (1975, p. 543) noted that it is absent in bone. Embryological evidence indicates that os- some Devonian actinopterygians (e.g., Mimia). sification in the lateral wall of the auditory In Kentuckia (Rayner, 1951), a Mississippian region follows closely the final pattern of the genus, the paired myodomes are separated by a chondrocranium. Absence of the anterior or the bony partition of the basisphenoid but are posterior basicapsular commissures, or failure joined by the canal for the pituitary vein, which of either to fuse with the parachordal, will passes transversely through the partition. Bjer- apparently influence the configuration of the ring (1977, p. 169) claimed that the posterior 12 AMERICAN MUSEUM NOVITATES NO. 2658 myodomes of Kentuckia are only parts of the fissure to the various landmarks discussed by respective trigeminofacialis chambers, but aside Gardiner and Bartram (1977) plus additional from being paired, they resemble the myo- features such as the basipterygoid process, the domes of Pteronisculus in most respects (figs. foramina for the oculomotor and the glos- 8, 9). In Luederia, as in Pteronisculus, and sopharyngeal nerves, the utricular recess and most other actinopterygians, there is a single the canal for the internal carotid artery. As median posterior myodome that is joined to the noted above, Luederia has a large posterior pituitary space dorsally and includes the path of myodome that is anteroventral to the prootic the pituitary vein. The recess for the sacculus bridge and to the partition in front of the ven- vasculosus is also dorsal to the median myo- tral otic fissure. In most ways it agrees with the dome, and the bucco-hypophyseal canal passes situation in Pteronisculus (Nielsen, 1942, fig. through the anterior part of the myodome to its 9). opening in the basicranium. As a result of this comparison, we can find Gardiner (1973), Patterson (1975), and Gar- little evidence that the ventral otic fissure be- diner and Bartram (1977) have concluded that, tween the basisphenoid anteriorly and the along with enlargement of the myodome and basioccipital posteriorly has migrated or been consequent atrophy of the front of the pushed backward in the palaeopterygians. If the notochord, the ventral otic fissure migrated pos- occipital segment of the basicranium is drawn teriorly from its primitive position anterior to to unit size in Mimia, Kentuckia, Kansasiella, the vestibular fontanelle to a location within the Pteronisculus, and Luederia (fig. 8), it is evi- chordal part of the basal plate. As interpreted dent that the landmarks mentioned above have by Gardiner and Bartram (1977, pp. 241-242), about the same spatial relationships with each the position of the ventral otic fissure in pal- other and with the ventral otic fissure. In the aeoniscoids varies from an anterior one, where neopterygians, however, the prootics have en- the ventral fissure is confluent with the orbit, to larged and separated the basisphenoid from the a posterior location where it "passes through basioccipital. The basisphenoid has become rel- the base of the otic region and runs into the atively smaller (C. Patterson, personal com- vestibular fontanelles." In spite of this pre- mun.) and most of the posterior myodome floor sumed migration, these authors list certain con- is formed by the prootics. The ventral otic stant topographical relationships between the fissure is now bordered anteriorly by the pro- ventral fissure and certain other structures, in- otics, which, through their increase in size, cluding (a) the anterior tip of the notochord; (b) have apparently moved the fissure into a rela- the partition between the auditory and cranial tively more posterior position. cavities; (c) the level of the origin of the orbital In order to appreciate the meaning of these arteries, and (d) the place where the lateral changes it should be recalled that the external commissures join the basicranium. In addition, recti muscles differentiate before the posterior they found constant relationships between the myodome is actually formed and while the lat- position of the fissure in Mimia and in embryos eral chondrocranial wall is still membraneous of Acipenser, Lepisosteus, and Exocoetus in (see DeBeer, 1937, p. 428; pl. 35, fig. 1; pl. relation to the transverse pituitary vein, the fo- 47, fig. 2; Bjerring, 1967, fig. 18). The position ramina for the palatine branch of the facial and length of the posterior myodome (and, in nerve and the openings for the orbito-nasal ar- fact, of the other myodomes) is thus related to teries as they pass into the orbit (see DeBeer, the position and the length of the external recti 1937, pl. 30, fig. 1; pl. 38, fig. 1; p. 50, fig. (or other eye muscles). Increase in the length 3). of the obliquely oriented external recti muscles In order to interpret some of the structures (fig. 9) is a derived condition in the actino- visible in the midsagittal section of the pterygians that initially resulted in shallow, Luederia braincase we have compared it with paired myodomes and eventually in a single sagittal sections of other palaeoniscoid neu- posterior myodome that finally extended into rocrania (fig. 8) and, where possible, with the the basioccipital. It thus appears to us that the topographical relationships of the ventral otic enlargement of the myodome in the advanced 1978 SCHAEFFER AND DALQUEST: PALAEONISCIFORM BRAINCASE 13

prb utr

pmy-

ic - B E

C FIG. 8. Postorbital portion of the neurocranium in sagittal section in various palaeonisciforms and in a pholidophorid. The basicranial part of the occipital segment is the same length in all figures. A. Mimia toombsi, after Gardiner and Bartram, 1977. B. Kentuckia deani, after Rayner, 1951. C. Kansasiella eatoni, after Poplin, 1974. D. Pteronisculus stensioi, after Nielsen, 1942. E. Luederia kempi. F. Pholidophorus bechei, after Patterson, 1975. actinopterygians involved reduction in or the well as some backward displacement of the absence of chondrification and ossification fissure. As noted by Patterson (1975) enlarge- around the elongating external recti muscles as ment of the pituitary vein canal was not in- 14 AMERICAN MUSEUM NOVITATES NO. 2658

fotv pv pmy pv pf fotv hv pmy pv A B C FIG. 9. Schematic reconstructions of palaeonisciform neurocrania in dorsal aspect to show the development of the posterior myodome in relation to the elongation of the external recti muscles. A. Posterior myodomes not developed as in Mimia toombsi (Gardiner and Bartram, 1977). B. Separate posterior myodome for each external rectus muscle, as in Kentuckia deani (Rayner, 1951). C. Single posterior myodome for both external recti muscles, as in Pteronisculus stensioi (Nielsen, 1942) and most other actinopterygians. volved, at least initially. Nevertheless, the Cahiers de Paleont., C.N.R.S., Paris, 135 canal and the incipient myodomes were in close PP. association, and the median myodome incorpo- Berman, David S. rates the pituitary vein. 1970. Vertebrate fossils from the Lueders For- According to our scenario reconstruction of mation, lower Permian of north-central posterior which in Texas. Univ. Calif. Publ. Geol. Sci., vol. myodome history, agrees 86, pp. 1-39. most respects with that of Patterson (1975, pp. Bjerring, Hans C. 540-544), the external recti muscles primitively 1967. Does a homology exist between the bas- were attached to the surface of the basisphenoid icranial muscles and the polar cartilage? portion of the postorbital wall at the level of Colloq. Internatl., C.N.R.S., no. 163, pp. the pituitary fossa (fig. 9). As some Devonian 223-267. palaeoniscoids had no myodomes, it follows DeBeer, G. R. that their external recti muscles were short and 1937. The development of the vertebrate skull. contained within the orbit. This was apparently Oxford, Clarendon Press, xxiv + 552 pp. the primitive gnathostome condition. Elonga- Gardiner, Brian G. tion of the external recti muscles obviously oc- 1963. Certain palaeoniscoid fishes and the evolution of the snout in actinopterygians. Bull. curred among the palaeoniscoids, resulting in Brit. Mus. (Nat. Hist.), Geol., vol. 8, no. the paired myodomes of Kentuckia and in the 6, pp. 257-325. single posterior myodome of most actino- 1973. Interrelationships of teleostomes. In pterygians that is also confluent with the pitui- Greenwood, P. H., R. S. Miles, C. Pat- tary fossa. terson (eds.), Interrelationships of fishes. (Suppl. no. 1, Zool. Jour. Linnean Soc., LITERATURE CITED vol. 53). Academic Press, New York, pp. 105-135. Beltan, Laurence Gardiner, Brian G., and A. W. H. Bartram 1968. La faune ichthyologique de l'Eotrias du 1977. The homologies of ventral cranial fissures N.W. de Madagascar: Le Neurocrane. in osteichthyans. In Andrews, S. Mahala, 1978 SCHAEFFER AND DALQUEST: PALAEONISCIFORM BRAINCASE 15

R. S. Miles, A. D. Walker (eds.), Prob- tolepid fishes, with a review of the ac- lems in vertebrate evolution. (Linnean tinopterygian braincase. Phil. Trans. Roy. Soc. Symp. Ser. no. 4), Academic Press, Soc. London (Biol. Sci.), vol. 269, pp. New York, pp. 227-245. 275-579. Jarvik, Erik Phillips, J. B. 1954. On the visceral skeleton in Eusthenopteron 1942. Osteology of the sardine (Sardinops with a discussion of the parasphenoid and caerulea). Jour. Morph., vol. 70, pp. palatoquadrate in fishes. Kungl. Svenska 463-500. Vetenskapsakad. Handl., vol. 4, no. 1, Poplin, Cecile pp. 1-104. 1974. Etude de quelques paleoniscides pennsyl- 1977. The systematic position of acanthodian vaniens du Kansas. Cahiers de Paleont., fishes. In Andrews, S. M., R. S. Miles, C.N.R.S., Paris, 151 pp. A. D. Walker (eds.), Problems in verte- 1975. Remarques sur le systeme arteriel epi- brate evolution. (Linnean Soc. Symp. Ser. branchial chez les actinopterygiens primi- no. 4), Academic Press, New York, pp. tife fossiles. Colloq. Internatl. C.N.R.S., 199-225. no. 218, pp. 265-271. Lehman, Jean-Pierre Rayner, Dorothy 1952. Etude complementaire des poissons de 1951. On the cranial structure of an early pal- l'Eotrias de Madagascar. Kungl. Svenska aeoniscid, Kentuckia, gen. nov. Trans. Vetenskapsakad. Handl., vol. 4, no. 6, Roy. Soc. Edinburgh, vol. 62, pp. 53-83. pp. 1-201. Romer, Alfred S. Miles, Roger S. 1937. The braincase of the Carboniferous 1973. Relationships of acanthodians. In Green- crossopterygian Megalichthys nitidus. wood, P. H., R. S. Miles, C. Patterson Bull. Mus. Comp. Zool., vol. 82, pp. (eds.), Interrelationship of fishes. (Suppl. 1-73. no. 1, Zool. Jour. Linnean Soc., vol. 53), Schaeffer, Bobb Academic Press, New York, pp. 63-103. 1968. The origin and basic radiation of the Os- 1977. Dipnoan (lungfish) skulls and the relation- teichthyes. In Orvig, T. (ed.), Current ships of the group: a study based on new problems of lower vertebrate phylogeny. species from the Devonian of Australia. (Nobel symposium no. 4). New York, In- Zool. Jour. Linnean Soc., vol. 61, pp. terscience Publishers, John Wiley and 1-328. Sons, Inc., pp. 207-222. Moy-Thomas, J. A., and Roger S. Miles Stensi6, Erik 1971. Palaeozoic fishes. (2nd edition). Phila- 1925. Triassic fishes from Spitzbergen. Part II. delphia, W. B. Saunders, Co., 259 pp. Kungl. Svenska Vetenskapsakad. Handl., Nielsen, Eigil vol. 2, pp. 1-261. 1942. Studies on Triassic fishes from East Watson, D. M. S. Greenland. I. Glaucolepis and Boreo- 1925. The structure of certain palaeoniscids and somus. Medd. Gronland 138, pp. 1-403. the relationships of that group with other Patterson, Colin bony fishes. Proc. Zool. Soc. London, no. 1975. The braincase of pholidophorid and lep- 53, pp. 815-870.