UN HARVARD UNIVERSITY Library of the Museum of Comparative Zoology MUS. COMP. ZOOL LIBRARY OCCASIONAL PAPERS JUN g £ of the MUSEUM OF NATURAL HISTORY The University of Kansas Lawrence, Kansas NUMBER 99, PAGES 1-17 JUNE 8, 1982 DISCOVERY OF INTRACRANIAL OSSICLES IN A CARBONIFEROUS NORTH AMERICAN PALAEONISCID (PISCES: ACTINOPTERYGII) Cecile M. Poplin1 During the preparation of serial sections of an endocranium of an Upper Carboniferous palaeoniscid from the United States2, I noticed two small bony elements whose shape and dimensions were almost identical and which were tightly applied to the walls of the endocranial cavity in the diencephalic and mesencephalic regions. These small bones are complete, being entirely surrounded by a periostic lamella. Their symmetry indicates that they are in their natural position and, thus, that they are not indeterminable bony fragments like those which are sometimes found scattered in matrix near endocrania. Subsequently, mechanical preparation of many specimens of the same species yielded these ossicles in normal or inverse relief under different angles according to the breaks of the nodules. It became obvious that these elements are present in most specimens, and that their dimensions vary within precise limits. This paper deals with the origin of these ossicles, the reasons for their variability, and why they have never been observed until now in other vertebrates besides this North-American palaeoniscid. I will first describe these enigmatic bones and their variations, 1 La 12 du CXRS, Institut de Paleontologie, 8 rue de Buffon, 75005 Paris, France. 2 Designated throughout this paper as Palaeoniscid B (as referred to by- Watson, 1925) which comes from the Virgil Series (Upper Carboniferous, Pennsylvanian) in the vicinity of Lawrence. Kansas (Poplin, 1974). The large number of specimens and their excellent state of preservation allows an exhaustive study which will be published in the future; on that occasion this new ta-xon will be named. 2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY and will then discuss the origin of the postero-superior part of the basisphenoid, and of the endocranial walls supporting the ossicles. DESCRIPTION The ossicles are always symmetrically placed in the diencephalic and mesencephalic regions of the endocranial cavity which are often difficult to approach for the observer. In order to eliminate the uncertainties which result from incomplete observations. I used only those 18 specimens of Palaeoniscid B in which this part of the endocranium is completely visible. All of these individuals are probably adults judging from the degree of ossification. When entirely ossified the ossicles comprise two parts, anterior (diencephalic) and posterior (mesencephalic), which are con- nected by a bony isthmus. The posterior part is placed vertically between the corresponding anterior bulge of the optic lobe and the internal face of the endocranium. The anterior part, closer to the median plane, lies on the ventral, horizontal face of the diencephalon3 ( Fig. 1 ) . Direct observation and the study of the serial sections show that the internal structure of the intracranial ossicles is identical to that of the bone which forms the endocranium itself (Nielsen, 1942; 1949). They are surrounded by a continuous periostic lamella (Poplin, 1974:31) which is adjacent to that of the endocranial wall Fir.. 1. Schematic reconstruction of intracranial ossicles of Palaeoniscid B ( Pennsylvanian of Kansas ) on a natural cast of the endocranial cavity ( KUVP 56371, Museum of Natural History, University of Kansas, Lawrence, Kansas). Stippled areas: ossicles; A. right lateral view; B. ventral view; C. left lateral view, c.i, foramen for the internal carotid; di, diencephalon; isth, bony isthmus; l.t>l>f, optic lobe; p.di, p.mes, anterior (diencephalic) and posterior (mesen- cephalic) parts of ossicles; II, optic foramen; IV, foramen for the trochlear nerve. :; As a simplification, I call "optic lobe" or "diencephalon" the natural casts of the regions oi the endocranial cavity which contained these organs of the nervous system in the living animal. PALAEONISCID INTRACRANIAL OSSICLES 3 on their lateral face. These two characters demonstrate the in- dividuality of the ossicles and the absence of cartilage. The number of specimens available permits an examination of ossicle variation. Four major ways in which the ossicles vary are as follow: 1. Outlines-: The right and left ossicles are always symmetrical in size, thickness and general shape, and only the minute sinuosities of their outlines are variable (Fig. 1). 2. Size: The size of the anterior portion is almost the same in all specimens where it can be observed, but the posterior part differs in that respect. Figure 1 shows the most common case where the posterior part extends horizontally from the proximal foramen of the canal for the trochlear nerve (IV) to the anterior limit of the optic foramen (II), and where it lies above the optic foramen to half the height of the optic lobe. In some specimens, however, it is approximately half the size. In one specimen it extends downward to the foramen through which the internal carotid reaches the endocranial cavity. 3. Thickness: The average thickness of the ossicles, estimated from serial sections and from direct observation, is 0.3 to 0.4 mm. In contrast, the posterior part of one specimen is thicker and ex- tends 1 mm into the space of the endocranial cavity that houses the optic lobe (Figs. 3, 4D). 4. Nature of ossification: The two parts of the ossicles are Table 1. Ossification of ossicles observed in 18 specimens of Palaeoniscid B. Anterior part Ossified Imprint Imprint Absent Absent Posterior part Ossified Ossified Absent Ossified Absent No. of Observations 3 3 2 8 2 ABC Fig. 2. Schematic transverse sections of intracranial ossicles of Palaeoniscid B at the level of the diencephalic region showing different aspects of the fossils with regard to the anterior part of the ossicles. A. ossicles missing; B. ossicles present; C. the anterior part of ossicles is missing, leaving its imprint on the inferior face of the endocranial cavity. Hatching: endocranial walls; black areas: ossicles; stippled areas: endocranial cavity. 4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY equally ossified in only 3 of the 18 specimens where this region could be observed. Two of the 18 endocrania have no trace of an ossicle. In the other specimens the anterior and/or posterior parts are always symmetrically ossified, or absent, or represented by their imprint on the endocranial wall which forms a corresponding bulge on the natural cast of the endocranial cavity when the latter is preserved (Figs. 2, 3, 4B). Table 1 summarizes the observed combinations of these different aspects. Three conclusions can be drawn from Table 1: a) the posterior mesencephalic part of the ossicles is by far the most observable since it occurs as a distinct bone in 14 endocrania out of 18. In contrast, the anterior dience- phalic part is preserved more rarely. It was observed ossified in 3 specimens, appears as a cast in 5 specimens, and is totally missing in the others; b) combinations other than those in Table 1 are possible even though they were not observed; and c) the morphol- ogy of the ossicles and their variability leads to the conclusion that they are formed from two centers of ossification, an anterior and a posterior, which are partly independent from each other. HYPOTHESES AND DISCUSSIONS The frequency with which these ossicles are found, their sym- metrical position, and the fact that they are always tightly applied to the internal face of the endocranium refute the hypothesis that they are artifacts of fossilization. Their frequency and symmetry also refute the hypothesis that the ossicles can be interpreted as pathological. Their bony structure eliminates also the hypothesis that they are otoliths which would have been, by chance, sym- metrically displaced in front of the hypophyseal region after death. I conclude that the intracranial ossicles are normal elements of the endocranium of Palaeoniscid B. Two possible hypotheses concerning the origin of ossicles are from soft tissues (the meninges) or skeletal tissues. Meningeal Origin of Ossicles It can be hypothesized that the intracranial ossicles of Palae- oniscid B represent a more or less ossified state of a structure, or of an organ, which is usually made of soft tissue in other verte- brates. This hypothesis could explain the size variation of the ossicles, the fact that the anterior part may show only its imprint, and the inconstant occurrence of one or the other part. It is well known that the only soft organs able to ossify in a non-pathological way are tendons, ligaments and meninges. We can eliminate tendons and ligaments, which evidently do not occur in this region of the skull, and focus on the meninges which are in immediate proximity of the ossicles. PALAEONISCID INTRACRANIAL OSSICLES 5 a) The dura mater, the most external meninx and thus in contact with the cranium, in higher vertebrates (including amphibians, Pirlot, 1969) comprises an internal sheet and an external or periost sheet. The latter acts as an internal periost for the bones of the E E m E E ro B E E ro E E ro Fig. 3. Palaeoniscid B ( Pennsylvanian of Kansas). Stereoscopic views of the natural casts of the endocranial cavity. A. ( KUVP 56374) ventral view, ossicles completely ossified, the posterior part is partly destroyed and mainly visible as a cast; B. (KUVP 56371) ventral view, cast of the anterior part of ossicles as a double bulge on the cast of the endocranial cavity; the periostic lamella of the posterior part has been destroyed during preparation of the specimen; C. (KUVP 56373) dorsal view, posterior part of ossicles visible as a horizontal section; D. (KUVP 56372) backwards oblique view, specimen upside down; bone of the posterior part of the left ossicle partly broken showing its alveolar structure identical to that of the endocranial walls, also broken; anterior part of ossicle totally missing. 6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY p mes \ d ant Fig. 4. Palaeoniseid B ( Pennsylvanian of Kansas). Diagrams corresponding to Fig. 3ABCD. Stippled areas: ossicles (as either bone or casts); hatching: broken parts (matrix or bone) of the endocranium. c.i, foramen for the internal carotid; f.cl.ant, anterior dorsal fontanelle; l.opt, optic lobe; p.di, p.mes, anterior and posterior parts of ossicles; //, optic foramen; ///, foramen for the common oculomotor nerve; TV, foramen for the trochlear nerve. cranium. Its osteogenetic activity is particularly well known in humans (Paturet, 1964). The periost sheet is able to regenerate bone after trepanation; it sometimes shows areas of ossification which are most often observed at the level of the tentorium (that part of the dura mater which is attached to the petrosal and separates the brain from the cerebellum), and of the falx cerebri (which separates the two hemispheres of the brain). Ossification of the tentorium is rare in humans, and occurs mainly in older individuals. It is much more frequent and normal in other mammals, particularly in recent non-human primates where it has different forms (Saban, 1963:23, 311). A comparable- ossification has been observed in some cephalaspids (Janvier, pers. comm.). PALAEOMSCID INTRACRANIAL OSSICLES 7 The presence of an unidentified bone in a recent anuran, Eleutherodactylus rugulosus (Tanzer and Baldauf, 1971), may be attributed to a comparable phenomenon. This unpaired median bone was located behind the prootic level and between the carti- laginous roof of the neurocranium and the dura mater. Because the authors had no knowledge of any comparable ossification in other vertebrates, they inferred it was a malformation. There is uncertainty as to whether the bone was built by the dura mater since connective tissue separated these two organs; but the observa- tion was unique and made on only one adult individual. The unidentified bone of Eleutherodactylus is apparently the only re- ported case which can be directly compared to the ossicles of Palaeoniscid B. b) The problem of the basisphenoid: this bone has a variable morphology (Daget, 1964) in recent teleosts, and is of dermal origin according to Allis (1897) or of cnchondral origin according to De Beer (1937). Conversely, Chabanaud (1936) thinks that it is a membranous bone made of two distinct elements: a postero-dorsal part formed by one or two fused ossifications of the dura mater (therefore the name "meningoste") and an antero- ventral part (the "belophragme") formed by the conjunctive interorbital septum. It is known that the endocranium of palaeoniscids is made of two bony blocks in which it is impossible to distinguish the differ- ent parts which are found in more recent actinopterygians. The term "pars basisphenoidea" indicates the median and two lateral pillars which separate the posterior myodome and the hypophysial fossa from the orbital cavities in front, and which seem to be the topological homologues of the belophragme of teleosts. The os- sicles of Palaeoniscid B, which are exactly above and on both sides of the pars basisphenoidea, would then represent the prefiguration of the meningoste which was not yet fused to the belophragme. A homology with the meningoste can be suggested for the bony spur of the pterosphenoid of Pholidophorus and Leptolepis, two primitive Jurassic teleosts (Patterson, 1975). This spur, of mem- branous origin according to Patterson, is located immediately below the optic lobes and probably serves as a connection to the mem- branous veil which closes the anterior part of the optic foramen in front of the optic nerve.4 Moreover, the comparable locations of this spur and the posterior part of the ossicles might indicate a homology. 4 The pterosphenoid of Hiodon tergisus (a Recent osteoglossomorph teleost) has on its anterior edge a vertical, inwardly directed lamina, dorsally developed, which forms the anterior limit of the cerebral cavity. According to Taveme (1977), it represents the ossified remnant of an embryonic cartilaginous epiphysial bar. This lamina does not seem to be related to the bony spur cited here, nor with the intracranial ossicle. 8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY This hypothesis would support Chabanaud's (1936) views, and would provide an explanation of the origin of the ossicles, but it presents a major difficulty. The postero-superior part of the basis- phenoid (meningoste) forms the lower and lateral edges of the optic foramen (Gregory, 1959: Fig. 2, 119; Patterson, 1975: Fig. 85, 101) in Recent aetinopterygians, but the ossicles of Palaeoniscid B are located above the optic foramen and do not form part of the posterior wall of the orbit. They would have had to undergo a downward migration during evolution in order to fuse with the belophragme of teleosts. It is more reasonable to postulate a homology between the meningoste and the region of the posterior face of the orbit which borders the optic foramen laterally and above the pillars of the pars hasisphcnoidea. Conversely, the hypothesis of a meningeal origin of part of the basisphenoid is far from proven, despite acceptance of this view by many authors such as Devillers (195S) and Daget (1964). Chabanaud supports his assertion of a membranous origin with an observation by Schleip (1904) who published a transverse section of a salmon fry where this bone appears within the membranous septum which separates the cerebral cavity from the rectus muscles of the eyes. Examination of Schleip's original text and figure re- veals that the author alludes neither to the dura mater nor to any meningeal tissue, even though he recognized that the basisphenoid shows many distinct centers of ossifications. Thus, it seems that Chabanaud has extrapolated Schleip's results. c) Finally, it is known that the protective membranes of the nervous system in fishes are far more simple than those of higher vertebrates (Scharrer, 1944; Bertin, 1958; Pirlot, 1969) despite the statements of Allis (1897:9) and De Beer (1935:35). In fact, osteichthyans have a primary meninx made of conjunctive, some- times fibrous, tissue rich in blood capillaries and adipose matter which has neither the structure nor the function of a dura mater. From this it may be inferred that the hypothesis of a meningeal origin of the ossicles, the postero-superior part of the basisphenoid, and the bony spur of the pterosphenoid of Pholidoplwriis and LeptoJepis probably should be rejected. Skeletal Origin of Ossicles A second hypothesis for the origin of the ossicles is that they derive from parts of the cranial skeleton. There are two possible elements, the endocranial walls and the anazygal. A. Hypotheses of derivation from the endocranial walls 1. First, it is necessary to determine the endocranial walls of Palaeoniscid B against which the ossicles are placed. Their anterior diencephalic part is applied to the wall which is formed either by PALAEONISCID INTRACRANIAL OSSICLES 9 the anterior part of the pterosphenoid or by the posterior part of the orbitosphenoid in adult actinopterygians where the neurocranium is built of distinct ossifications. The posterior mesencephalic part of the ossicles is located in an area which corresponds to the median part of the pterosphenoid (Gregory, 1959: Fig. 25, 119; Patterson, 1957: Fig. 63, 65, 90 etc. . . ). Not all authors agree on the em- bryological origin of the pterosphenoid and orbitosphenoid bones in osteiehthyans. The orbitosphenoid, which is frequently unpaired in Pholido- phoridae, Leptolepidae (Patterson, 1975) and Tcleosts (Daget, 1964), despite its double origin, is formed from the pila prcoptica according to Daget ( 1964). The pila preoptica, which could be also called pila antoptica, is the serial homologue of the pila metoptica, pila antotica and following pilae (De Beer, 1937; Kuhn, 1971), and it derives from the somitic sclerotome, precisely from dorsal arcual elements. Another hypothesis can be proposed in which the orbit- osphenoid proceeds also from the sclerotome, but through the taenia marginalis. The pterosphenoid (called also pleurosphenoid) is a derivative of the pila lateralis ("alisphenoid pedicle") according to Devillers (1958). The visceral origin of this pila is supported in many works (De Beer, 1937; Jarvik, 1972; and Bjerring, 1977). Goodrich (1930) considered the pila lateralis as originating from the spread- ing orbital cartilage. De Beer (1937), Goodrich (1930) and Daget ( 1964 ) stated that the pterosphenoid developed from the taenia marginalis. A third possibility is that this bone derives from the pila metoptica since it covers the same area [between the foramina for the optic nerve and the common oculomotor nerve (III), De Beer, 1937]; this has been suggested by Bjerring (1971:209) for Ettsthenopteron foordi. Despite the variety of proposed origins for these two bones, they are formed by the somitic sclerotome since they are most certainly derivatives of either the pila preoptica, the pila metoptica, the orbital cartilage, or the taenia marginalis. In Palaeoniscid B it is possible to eliminate immediately a hypothesis of derivation of the ossicles from the pila lateralis, which has been observed in part or completely in Birgeria (Stensio, 1921), Pteronisculus (Nielsen, 1942), Kansasiella (Poplin, 1974, 1975), Macrepistius (Schaeffer, 1971), Pholidophoridae, Leptolepidae and Amia (Patterson, 1975). The pila lateralis lies on the posterior wall of the orbit, divides the anterior meatus of the trigemino-facialis chamber in two parts, and has no connection with the endocranial wall in any of the fossil actinopterygians just noted. Concerning the embryological origin of the area which supports the anterior diencephalic part of the ossicles, I can envisage the same hypothesis of origin as that proposed for the orbitosphenoid 10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY and pterosphenoid. The pila preoption connects the orbital carti- lage and the corresponding trabecle in front of the optic foramen in embryos; its fusion with the symmetric pila leads to the formation of the bony interorbital septum of the tropibasic skull of telosts (De Beer, 1937). According to this definition, it appears that the pilae prcopticac cannot have formed the endoeranial floor which supports the diencephalon of Palaeoniscid B. Moreover in palaeon- iscids the interorbital septum is not ossified, and the orbits were separated by a membranous partition (Poplin, 1974). It seems much more likely that this endoeranial area is derived directly from the taenia marginalis. Concerning the embryological origin of the endoeranial wall against which the posterior mesencephalic part of the ossicles lie, the hypothesis of the taenia marginalis and of the pila mctoptica (proposed above for the pterosphenoid) are not as yet contra- dicted, in my opinion, and therefore are retained. Thus it appears that the endoeranial area which supports the ossicles derives from the embryological somitic sclerotome through the taenia marginalis, and perhaps the pila mctoptica, for topo- graphical reasons. 2. With regard to the intracranial ossicles, it appears that the hypothesis of a visceral origin can be eliminated in light of the above. Indeed, it is well known that if elements of the primitive visceral skeleton are incorporated in the neurocranium (these views are supported by Jarvik, 1972 and Bjerring, 1977, following authors such as Huxley, Allis, and Holmgren), they build the more external walls which do not directly surround the brain (as for instance the trabecles ventrally, and the pila lateralis and the lateral commissure laterally). Further, the "notohylic mesoderm" (Bjerring, 1968) (="ehordo- mesoblaste" or "chordoblaste" of Daget, 1964 and Pirlot, 196S) is unpaired, median and situated below the floor which supports the brain, as are its derivatives (for example, the notochord and the exchordal cartilages). Therefore, it cannot have given rise to the intracranial ossicles. The most plausible and logical hypothesis is the derivation of the ossicles from somitic sclerotomes which are known to have originated in the bony cover protecting directly the nervous struc- tures at the level of the skull as well as that of the vertebral column (like the orbitosphenoid and the pleurosphenoid according to most authors) in gnathostomes. One can content oneself with this hypothesis because embryo- logical considerations on fossils are speculative, particularly when the authors do not agree on the origin of the bones which form the endoeranial wall in the region dealt with in recent fishes. 3. Anyway one can go further with this hypothesis by attempting PALAEOMSCID INTRACRANIAL OSSICLES 11 to determine, as for the endocranial wall, from which exact sclero- tome area the ossicles originated. The morphology of the ossicles permits the supposition that they develop from two points of ossifi- cation, anterior and posterior, and this would account for the varia- tions of the two parts owing to their different osteogenetic activities. The diencephalic part possibly derives from an anterior point of ossification, and may have two origins. The first one is the taenia marginalis. This presumes a duplication of the taenia marginalis in order to form both the ossicle and the adjacent endocranial wall. Such a process is not unknown as shown by the embryology of Polypterus senegalus. Daget, Bauchot and Amoult (1964), in their detailed study of the development of this brachiopterygian, de- scribed how the taenia marginalis splits into two parts and forms a secondary taenia which is lateral to the first one and which fuses to it in a later stage. Such a duplication may have occurred as a superimposition in Palaeoniscid B, and not as a juxtaposition as in Polypterus. Anyway the ventral location of the ossicles in the dien- cephalic region is somewhat inconsistent with the major function of the taenia marginalis to build the lateral part of the endocranium roof in the orbital region. Thus, I would abandon this hypothesis. A second possible origin for the diencephalic part of the ossicles is the pila preoptica. I have shown above that it does not form a bony interorbital septum in palaeoniscids, but it is possible that in Palaeoniscid B the anterior part of the ossicles represent the re- duced pila preoptica. The mesencephalic part of the ossicles possibly derives from a posterior ossification point of which the pila metoptica could be the origin as for the adjacent endocranial wall. This presumes a duplication by superimposition of that embryonic pillar, like that of the taenia marginalis cited above. In that hypothesis, the spur of the pterosphenoid of Pholidophorus and Leptolepis could be due to a similar phenomenon. It should be noted that it is not possible to suggest such a duplication by superimposition of the pila metoptica without using "Holmgren's Principle of Delamination," which was defined by Holmgren (1940) and named by Jarvik (1959). The terms of this principle are frequently quoted, but are sometimes considered imprecise (Schaeffer, 1977; Patterson, 1979). In my opinion, delamination can be used to explain the formation of the intra- cranial ossicles only in an ontogenetic sense and by means of real successive cellular generations. B. Hypothesis of derivation from the anazygal In another paper (Poplin, 1981), I discussed in some detail the problems dealing with the homologies of the anazygal and the 12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY prootic bridge in osteichthyans. I will summarize here the principal conclusions. 1. The anazygal is a bony, median, single element which over- Lays the cephalic notochordal canal in its anterior region, and on which the brain lies partly. It has been described in Latimcria chalumnae ("piece preoccipitale suschordale" of Millot and An- thony, 1958), and later in fossil coelacanthiforms and in "rhipidis- tians" such as Eusthenopteron foordi which have two successive anazygals. In actinopterygians, it appears that the anazygal is a normal element in the endocranium of the more primitive fossil repre- sentatives, mainly palaeonisciforms, and that its disappearance in more recent forms has resulted from the backward extension of the posterior myodome. The anazygal is most often single, but it is made of two sym- metrical ossifications in Mimia toombsi (the oldest palaeonisciform for which the endocranium is known, Gardiner and Bartram, 1977), the osteolepiform Eusthenopteron foordi, and the eoelacanthiform Nesides schmidti (Bjerring, 1971), all Devonian. This suggests that the anazygal probably has a paired origin. According to Jarvik (1972) and Bjerring (1971, 1975, 1978), the anazygal is formed from the somitic sclerotome by the two ventral arcual elements of a cephalic vertebra persisting in the endocranium. I favor the hypothesis that the anazygal originates from dorsal arcual elements ( interdorsal or basidorsal) owing to its position with regard to the notochord. 2. The prootic bridge (also called dorsum selloc or crista sellaris) covers, postero-dorsally, the hypophysial fossa, the saccus vasculosus, and, in teleosts, the middle part of the posterior myo- dome. Most hypotheses on its origin suggest that it originates from arcual elements. Its location on the median plane above and in front of the anterior extremity of the notochord supports a hypoth- esis that the prootic bridge may result from fusion of two sym- metric dorsal arcuals immediately anterior to the anazygals. This indicates that the prootic bridge and the anazygal could be serial homologues. 3. The intracranial ossicles have the following characteristics: a) the brain rests partly upon them (on its ventral face in the diencephalic area and on its anterior face at the level of the optic lobes) as well as on the prootic bridge and on the anazygal; b) each ossicle is made of two parts which have their symmetry on the other ossicle, as in the double anazygals of Mim:a, Eusthenop- teron and Nesides cited above; and c) the anterior part of each ossicle may derive from the pila prcoptica and the posterior part from the pila metoptica. These two pilae, formed by somitic sclero- tome, correspond to dorsal arcual elements. PALAEONTSCID INTRACRANIAL OSSICLES 13 Based on the above statements about the origin of the anazygal and of the prootic bridge, I propose a hypothesis in which the anterior symmetrical parts of the two ossicles and the posterior symmetrical parts constitute two pairs of bony elements which represent two double anazygals. This "double diencephalic anazy- gal," the "double mesencephalic anazygal," the prootic bridge, and the anazygal which overlies the notochord represent four successive serial homologues. CONCLUSIONS The origin of intracranial ossicles Two hypotheses have been proposed for the origin of intra- cranial ossicles. The first is based on the osteogenetic power of the dura mater in many vertebrates. Such a meningeal origin has been proposed by Chabanaud ( 1936 ) for the postero-superior part of the basisphenoid (the "meningoste" ) whose location in the skull reminds that of the intracranial ossicles. Patterson (1975) sug- gests that the two bony symmetric spurs attached to the medial face of the pterosphenoid in two primitive Jurassic teleosts (PlioJidoplwrus and Leptolepis) have a membranous origin. Does a homology exist between the intracranial ossicles, the "meningoste" and these bony spurs? Despite its simplicity, I do not favor this hypothesis since it has two difficulties: a) although the dura mater of higher vertebrates is frequently shown to have the function of a periost, such an osteogenetic function has not been clearly demonstrated for the primary meninx of osteichthyans; and b) the location of the ossicles above the optic foramen, where they do not form the posterior wall of the orbital cavity, is not consistent with that of the "meningoste" of the basisphenoid. Thus, the evidence suggests only a possible homology between the intracranial ossicles and the bony spurs of the pterosphenoid of Pholidophorus and Leptolepis. The second hypothesis points to a skeletal origin of the ossicles. Evidence from classical embryological works on the skull of Recent actinopterygians shows that the endocranial walls which support the ossicles of Palaeoniscid B probably originated from the somitic sclerotome. It seems probable that the intracranial ossicles have the same origin. More precisely, they may be derived from dorsal arcual elements. This hypothesis has two different conclusions, each of which seems equally possible: a) the anterior part of each ossicle may correspond to the reduced pila preoptica, and the posterior part may have been formed from the pila metoptica after a process of duplication by superimposition ( Holmgren's "Principle of Delamination") which may also account for the origin of the 14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY bony spurs of the pterosphenoids of Pholidophorus and LeptoJepis; and b) the anterior and posterior parts of the two ossicles, which are symmetrical, constitute two successive double anazygals which could be the serial homologues of the prootic bridge and the anazygal overlying the notochord in Palaeoniscid B. I favor this hypothesis, recognizing that my conclusions on the origin of the ossicles are based mainly on embryological considera- tions made by comparisons with Recent actinopterygians, and are speculative in nature. Following Jarvik (1972: Fig. 99B), Bjerring (1977) presented a synthesis of his views on the fundamental metameric characters of the cephalic region of craniota. According to Bjerring (1973), the terminal metamere (which is identified as immediately anterior to the premandibular metamere) comprises dorsal and ventral arcual elements which have formed the sclerotic coating of the eyes. He does not, however, mention the pila preoptica, which represents the dorsal arcual element of the terminal metamere since it precedes immediately the pila metoptica which belongs to the premandibular metamere. Since the pila preoptica perhaps forms the anterior part of the intracranial ossicles and the orbit- osphenoid of teleosts ( Daget, 1964), the ventral arcual elements of the terminal metamere, by themselves, would have formed the sclerotic coating of the eyes according to Bjerring. Variation in intracranial ossicles I have shown previously that variation in intracranial ossicles occurs mainly in their size and the thickness of their posterior part, and also in the nature of their appearance (i.e. presence as bone or cast, or total absence). It follows from the hypothesis of a skeletal origin that each ossicle was probably formed by two centers of ossification, anterior and posterior, the osteogenetic ac- tivity of which might vary from one individual to another. The ossicles are very thin (0.3 to 0.4 mm) and easily broken. A part of one might have been detached before fossilization so that it ap- peared only as an imprint on the endocranial wall or on the natural cast of the endocranial cavity. In addition, it is also possible that the ossicles of Palaeoniscid B are evidence of structures which may have been more developed in their ancestors. I have shown that the anterior part of the ossicles is perhaps the reduced pila preoptica, and it is well known that remnant organs often exhibit great intraspecific or individual variations. Lack of previous observations of intracranial ossicles in vertebrates The number of observations and the location of the ossicles PALAEONISCID INTRACRANIAL OSSICLES 15 reported herein clearly demonstrate that they are normal elements of Palaeoniscid B. Nevertheless, my inquiries to many colleagues did not reveal similar ossicles in any other Recent or fossil verte- brate, and none are mentioned in the literature. Three explanations for this lack of information can be proposed: a) It is well known that one finds mainly that for which one is looking. It is possible that intracranial ossicles did not attract the attention of paleontol- ogists and anatomists in the past. Re-examination of previously investigated taxa might reveal the presence of intracranial ossicles; b) The ossicles are small, fragile, and applied against the en- docranial wall without being connected to it by a suture, factors not favorable to fossilization. Due to the excellent state of preser- vation and abundance of the Pennsylvanian material from Kansas, the ossicles of Palaeoniscid B were clearly evident. There probably exist known examples of fossil species in which the endocranium has lost its ossicles before fossilization, but natural casts or prints might reveal them. This is not, however, the case for specimens of Kansasiella eatoni, which were obtained from the same localities as Palaeoniscid B and were in the same excellent state of preserva- tion. An intensive study of that species ( Poplin, 1974 ) did not reveal any trace of ossicles, and leads to a third explanation; c) The fact that the ossicles might be remnants, as stated above, may explain their absence in fossil species as old as Palaeoniscid B and in more recent actinopterygians. The preceding paragraphs represent a great deal of speculation and raise many questions. Because intracranial ossicles are now well documented in the Pennsylvanian Palaeoniscid B, it would be useful to investigate this anatomical phenomenon in more fossils. ACKNOWLEDGEMENTS The problems dealt with in this paper have been discussed with many colleagues. Some of them helped with suggestions, by ob- taining references, or by criticizing the manuscript. I am very grateful to P. Geistdoerfer, D. Goujet, P. Janvier, F. Poplin, J. C. Rage, R. Saban, B. Schaeffer, and L. Taverne. I particularly want to thank E. O. Wiley who veiy kindly turned my bad English into good American. RESUME Description de deux ossicules symetriquement places contre la face interne de la paroi endocranienne dans les regions diencephal- ique et mesencephalique de nombreux endocranes appartenant a un Palaeoniscide ( Aetinopterygien) pennsylvanien du Kansas. La discussion porte sur leur origine (probablement des elements arcuaux dorsaux du sclerotome somitique) et sur celle des parois adjacentes de l'endocrane des Palaeonisciformes et des Actinop- 16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY terygiens plus recents. L'hypothese dune possible homologie serielle entre ces ossicules, le pout prootique et 1'anazygal sur- montant la notochorde est proposee. Les raisons des variations morphologiques de ces ossicules et du fait qu'ils n'ont ete observes jusqu'a present que chez ce seul Palaeoniscide sont egalement discutees. LITERATURE CITED Allis, E. P. 1897. The morphology of the petrosal bone and of the sphenoidal region of the skull of Amia clava. Zool. Bull., 1(1): 1—26. Bertin, L. 1958. Systeme nerveux. In Grasse P. P. Traite de Zoologie, 13(l):854-922. Bjerrixg, H. C. 1968. The second somite with special reference to the evolution of its myotomic derivatives. In Current problems of lower vertebrate phylogeny (T. 0rvig, editor). Proc. 4th Nobel Symp., pp. 341-357, Almqvist and Wiksell, Stockholm. Bjerrixg, H. C. 1971. The nerve supply to the second metamere basicranial muscle in osteolepiform vertebrates, with some remarks on the basic composition of the endocranium. Acta Zool., 52, 189-225. Bjerrixg, H. C. 1973. Relationships of Coelacanthiforms. In Interrelation- ships of fishes. (P. H. Greenwood, R. S. Miles and C. Patterson, editors). Zool. J. Linn. Soc, 53:179-205. Bjerrixg, H. C. 1975. Contribution a la connaissance de la neuro-epiphyse chez les Urodeles et leurs ancetres Porolepiformes, avec quelques remarques sur la signification evolutive des muscles stries parfois presents dans la region neuro-epiphysaire des Mammiferes. In Prob- lemes actuels de Paleontologie-Evolution des Vertebres. Coll. Intern CNRS, 218:231-256. Bjerrixg, H. C. 1977. A contribution to structural analysis of the head of craniate animals. The orbit and its content in 20-22 mm embryos of die north american actinopterygian Amia calva L. with particular reference to the evolutionary significance of an nonocular orbital muscle innervated by the oculomotor nerve and notes on the metameric char- acter of the head in craniates. Zool. Scripta, 6:127-183. Bjerrixg, H. C. 1978. The "intracranial joint" versus the "ventral otic fissure". Acta Zool., 59:203-214. Chabaxaud, P. 1936. Le neurocrane osseux des Teleosteens dissymetriques apres la metamorphose. Ann. Inst. Oceanogr., 16(3) : 22.3-298. Daget, J. 1964. Le crane des Teleosteens. Mem. Mus. Nat. Hist. Nat., A. Zool., 31 (2): 163-341. Daget, J., Bauchot, M. L. and R. Arxoult J. 1964. Developpement du chondrocrane et des arcs aortiques chez Polypterus senegalus Cuvier. Acta Zool., 46:201-244. De Beer, G. R. 1937. The development of the vertebrate skull. Clarendon press, Oxford, 552 pp. Devillers, C. 1958. Le crane des Poissons. In Grasse P. P. Traite de Zoologie, 13(l):551-687. Gardixer, B. G. and A. W. II. Bartram. 1977. The homologies of ventral cranial fissures in Osteichthyans. In Problems in vertebrate evolution. Linnaean Soc. Symp., 4:227-245. Goodrich, E. S. 1930. Studies on the structure and development of Verte- brates. Macmillan and Co., London. 837 pp. Gregory, W. 1959. Fish skulls. A study of the evolution of natural mech- anisms. E. Lundberg, Laurel, Florida. 481 pp. Holmgrex, N. 1940. Studies on the head of fishes. Part I. Development of the skull in sharks and rays. Acta Zool., 21:51-121. PALAEONISCID INTRACRANIAL OSSICLES 17 Kuh.v, II. J. 1971. Die Entwicklung und Morphologie des Schiidels von Tachyglossus aculeatus. Abh. Senckenb. Naturforsch. Ges., 528:1-224. Jarvik, E. 1959. Dermal fin rays and Holmgren's principle of delamination. Kungl. Svensk. Vetensk. Handl., 6(1):1-51. Jarvik, E. 1972. Middle and upper devonian Porelepiformes from east Green- land with special reference to Glyptolepis groenlandica n. sp. and a discussion on the structure of the head in the Porolepiformes. Medd. om. Gr0nland, 187(2):307 pp. Millot, J. and J. Anthony. 1958. Anatomie de Latimcria chalumnae. T. I. Squelette et muscles, (ed. CNRS). Nielsen, E. 1942. Studies on triassic fishes from East Greenland I. Glaucolepis and Borcosomus. ( Palaeozoologica Groenlandica). Medd. om Gr0nland, 138:1-103. Nielsen, E. 1949. Studies on triassic fishes from East Greenland. II. Australosomus and Birgeria ( Paleozoologica Groenlandica). Medd. om Gr0nland, 146(1 ): 1-309. Patterson, C. 1975. The brainca.se of pholidophorid and leptolepid fishes, with a review of the actinopterygian braincase. Phil. Trans. Roy. Soc. London, 269(899) : 275-579. Patterson, C. 1979. Cartilage bones, dermal bones and membrane bones, or the exoskeleton versus the endoskeleton. In Problems in vertebrate evolution. Linnaean Soc. Symp., 4:77-121. Paturet, G. 1964. Traite d'anatomie humaine. T. IV. Systeme nerveux. Masson ed., Paris. Pirlot, P. 1969. Morphologic, evolution des Chordes. Presses de l'Univ. de Montreal. Poplin, C. 1974. Etude de quelques Paleoniscides pennsylvaniens du Kansas. Cahiers de Paleont. 151 pp. Poplin, C. 1975. Kansasiclla nomen novum remplacant Kansasia Poplin 1974 (Poissons Paleonisciformes). Bull. Soc, Geol. de France, Supp. T. XVII: 26. Poplin, C. 1981. Les homologies du pont prootique chez les Osteichthyens. Cybium, 5(1):3-17. Saran, R. 1963. Contribution a l'etude de l'os temporal des Primates. Description chez 1'homme et les Prosimiens. Anatomie comparee et phylogenie. Mem. M.N.H.M., A Zool., 29:1-377. Schaeffer, B. 1971. The braincase of the holostean fish Macrepistius with comments on neurocranial ossification in the Actinopterygii. American Mus. Nov., 2459:1-34. Schaeffer, B. 1977. The dermal skeleton in fishes. In Problems in Vertebrate evolution. Linn. Soc. Symp. Ser., 4:2.5-52. Scharber, E. 1944. The histology of the meningeal myeloid tissue in the Ganoids Amia and Lepisosteus. Zeits. Wiss. Zool., 72:467-524. Schleip, W. 1904. Die Entwickelung des Kopfknochen bei dem Lachs und der Forelle. Anat. Hefte., 13(72) : 331-427. Stensio, E. 1921. Triassic fishes from Spitzbergen, Vienna. 307 pp. Tanzer, E. C. and R. J. Baldauf. 1971. Contribution to the cranial mor- phology of Eleutherodactylus rugulosus (Cope) (Anura: Leptodactyli- dae). Texas Journ. Sci., 23(l):29-55. Taverne, L. 1977. Osteologie, phylogenese et systematique des Teleosteens fossiles et actuels du super ordre des Osteoglossomorphes. Pt. I. Osteologie des genres Hiodon, Eohiodon, Lycoptera, Osteoglossum, Scleropages, Heterotis et Arapaima. Acad. Roy. Belgique. Mem. CI. Sc, XLII(3): 1-235. Watson, D. M. S. 1925. The structure of certain Palaeoniscid and the re- lationship of that group with other bony fishes. Proc. Zool. Soc. London, 54(33): 815-870. 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 number 1 in 1946. Longer research papers are published in that series. Monographs of the Museum of Natural History were initiated in 1970. All manuscripts are subject 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 Occasional Papers and Miscellaneous Publications by addressing the Exchange Librarian, The University of Kansas Li- brary, 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, The Univer- sity of Kansas, Lawrence, Kansas 66045. Editor: E. O. Wiley Managing Editor: Joseph T. Collins printed by university of kansas printing service lawrence, kansas 1 0 7 I U \» 3 2044 093 361 699 DATE DUE ADD 1 <\ onrtA HHv-1