, . 
 
t$ 
 
 ^earning anb |Tubor. 
 
 LIBRARY 
 
 University of Illinois. | 
 
 A CLASS. BOOK. VOLUN 
 
 t 530,5 El .: \ 
 
 # 
 
 - 
 
 < V~ 
 
 * 
 
 ,. 
 
or' 
 
 .^ 
 
 Latest Date stamped below. A 
 charge is made on all overdue 
 books. 
 
 U. of I. Library 
 
 I 
 
 (79 
 
 22 19 
 FEB 5' 
 
 
 1984 
 
 1762S-S 
 
FIELD COLUMBIAN MUSEUM 
 
 PUBLICATION 5. 
 ZOOLOGICAL SERIES. VOL. i, No. i. 
 
 ON THE STRUCTURE AND 
 DEVELOPMENT 
 
 OF THE 
 
 VERTEBRAL COLUMN OF AMIA 
 
 WV. 
 
 BY 
 
 O. P. HAY, PH. D., 
 
 Assistant Curator of Ichthyology. 
 
 D. G. ELLIOT, F. R. S. E., Curator of Department. 
 
 CHICAGO, U. S. A. 
 October, 1895. 
 
VII' 
 
CONTENTS. 
 
 1. Description of the gross structure of the vertebral column. ... 5 
 
 2. The minute structure of the vertebral column 15 
 
 3. Discussion of the vertebral column of fossil fishes 20 
 
 4. Discussion of the vertebral column of the higher vertebrata. . .24 
 
 5. The origin and development of the vertebrae of Amia dur- 
 
 ing its earlier stages 25 
 
 6. Further consideration of the vertebral column of fossil fishes.. 39 
 
 7. Further discussion of the vertebral structure of Teleostei 42 
 
 8. Further consideration of the vertebral column of living and 
 
 extinct Amphibia 45 
 
ON THE STRUCTURE AND DEVELOPMENT OF THE VERTE- 
 BRAL COLUMN OF AMIA. O. P. HAY, PH. D. 
 
 I. DESCRIPTION OF THE GROSS STRUCTURE OF THE VERTEBRAL 
 
 COLUMN. 
 
 For our knowledge of the vertebral column of the fish Amia we 
 are indebted to a considerable number of writers, whose works will 
 be found in the list at the end of this paper. Their views, so far as 
 they concern our present purpose, will receive consideration as our 
 discussion proceeds. 
 
 In again calling attention to the spinal column of this fish, we 
 must first of all consider a peculiarity which has struck all investiga- 
 tors, and which distinguishes this species from all other living osse- 
 ous fishes. This peculiarity consists in the possession, throughout the 
 greater portion of the tail region, of apparently twice as many verte- 
 bral bodies as there are myomeres and neural and haemal arches. 
 Franque (26) appears to have been the first to remark on this struc- 
 tural feature of this fish. He regarded those vertebral centra which 
 are devoid of upper and lower arches as intercalated vertebrae 
 similar to those which are found in certain sharks and rays. He says: 
 ' Sunt igitur corpora vertebrarum inter vertebras intercalata." 
 He refers to the fact that among the Rays Rhinobatus has intercalated 
 vertebrae, while among the Squali Sphyrna malleus has intercalated 
 superior arches. Almost all other writers who have dealt with the 
 subject have adopted the same interpretation, while it is the express 
 purpose of a recent paper by Ludwig Schmidt (56) to establish this 
 view. On the other hand, Dr. G. Baur (9) and Dr. Carl Zittel (60) 
 hold that the two segments of the vertebral column which are found 
 in each of the myomeres in question are "centra" and " intercen- 
 tra," (pleurocentra* and hypocentra.), corresponding to those ele- 
 ments which together make up a vertebral body in some of the Stego- 
 cephali. 
 
 I signify my acceptance of the opinion that the whole vertebral 
 column of the ancestors of Amia was composed, in each myomere, of 
 
 *The term centrum has long been in, use to distinguish the principal portion of the vertebra 
 independently ot any theory concerning its origin and composition. To apply now this term to dis. 
 tinguishone of the elements that may enter into the construction of a vertebral body would intro- 
 duce confusion. I prefer therefore to employ in this paper pleurocentrum and fiypocentrum to des- 
 ignate the elements of the body. 
 
 5 
 
't 
 
 6 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 the two elements named, and that in the caudal region of this fish 
 we have both of these elements present, and still distinct from each 
 other. I find it impossible to ignore the results of palaeontological 
 researches when they are exhibited in so clear a light. These results, 
 which reveal an extraordinary similarity between the stegocephalous- 
 vertebral column and that of so many of the earlier osseous fishes, 
 cannot be set aside lightly, merely because there is in the tail of cer- 
 tain Elasmobranchs an evident duplication ; at -least not until it has 
 been shown by embryology that the vertebral bodies in the two cases 
 arise in the same way ; and this has certainly not yet been done. 
 
 If we shall regard a pleurocentrum and a hypocentrum in the tail 
 as together constituting a vertebral body, the total number of these 
 in the whole vertebral column will, of course, be considerably reduced 
 below the number usually given by authors. Franque says that the 
 number, including the intercalated ones, is commonly ninety-two, 
 although the number may vary somewhat. Of this number thirty- 
 seven or thirty -eight belong to the trunk. L. Schmidt found in one 
 skeleton thirty-nine vertebrae in the trunk and forty-seven in the 
 tail. In his figure of another specimen he represents fifty segments in 
 the tail, of which about fourteen are the so-called intercalated pieces. 
 The anterior four to six caudal vertebrae are simple, and resemble in 
 all respects those of the hinder dorsal region, except that the lower 
 arches are closed below to form the haemal canal. Near the end of 
 the tail, again, the intercalated pieces are not developed. Therefore, 
 counting pleurocentrum and hypocentrum as one, we find about sev- 
 enty-five vertebra? in the whole column, although the number may be 
 from one to three less. 
 
 Externally the vertebrae of Amia differ from those of most other 
 osseous fishes in the almost total lack of excavations, bony ridges and 
 subsidiary processes. Those of the trunk especially may be regarded 
 as approximately circular disks, each with a conical excavation at each 
 end. These disks become gradually shorter as we move from the 
 hinder dorsal vertebrae towards the head. In the tail the two ele- 
 ments, pleurocentrum and hypocentrum, taken separately, are con- 
 siderably shorter than a hinder dorsal vertebra ; but, if we consider 
 the two elements as constituting a single vertebral body, we find that 
 their combined lengths usually exceed somewhat that of a dorsal cen- 
 trum, while the length of an anterior caudal vertebra is only about 
 seventy-seven per cent of that of a hinder dorsal vertebra. As far for- 
 ward as the twenty-third dorsal vertebra from the head the length of 
 each vertebral body remains nearly the same as that of the hinder- 
 most. In front of this they generally grow shorter, so that the most 
 
OCT. 1695. VERTEBRAL COLUMN OF AMIA HAY 7 
 
 anterior is but little more than half the length of the twenty-third. 
 Schmidt's description, but not his figure, would lead one to suppose 
 that the first two vertebrae behind the head are shorter, when com- 
 pared with the succeeding ones, than they really are. 
 
 In transverse section the dorsal vertebrae are somewhat broader 
 than high, the perpendicular axis of the first dorsal being about 
 eighty-five per cent of the transverse axis. As we move backward, 
 the form changes somewhat, so that sections become more and more 
 nearly circular, the hindermost dorsal body being nearly perfectly so. 
 The caudal centra, on the other hand, become more and more com- 
 pressed towards the tip of the tail. Thoroughly macerated verte- 
 brae, as well as sections through decalcified vertebrae in various 
 regions, show that the notochord has not suffered complete constric- 
 tion, but runs continuously through the vertebral column. 
 
 It is no unusual thing to find the pleurocentrum and hypocen- 
 trum of some of the myomeres of the middle portion of the tail con- 
 solidated. Stannius (58, p. 21) had observed this union of elements, 
 and that it occurred at different points in different individuals. 
 Schmidt has called attention to the same phenomenon, and has 
 figured such a resulting vertebra. In such cases there is the closest 
 possible resemblance between the vertebra so resulting and one of 
 the anterior caudal vertebrae. At the same time there can be no 
 doubt concerning the complete homology of an anterior caudal verte- 
 bra and any of those of the dorsal region. 
 
 On the other hand there may occur a union of the elements of 
 different vertebrae. A specimen in my possession shows an evident 
 case of the consolidation of the hypocentrum with both the pleuro- 
 centrum in front of it and the one behind it. 
 
 Even if we should not be able to find in the middle tail region a 
 vertebra formed by such apparently abnormal union of pleurocen- 
 trum and hypocentrum, we might find it instructive to compare these 
 elements of any myomere with a posterior dorsal and an anterior 
 caudal vertebra. The general form of the two elements taken to- 
 gether is the same as that of the simple vertebral body. In the case 
 of the middle tail segments, the upper and lower arches rest 
 on the hinder element, the intercentrum. In the dorsal region the 
 arches repose on the hinder half of the vertebral body. In 
 "both the middle and the anterior tail-regions the upper and the lower 
 arches are separated by suture from the corresponding bodies. If 
 we should form our conclusions regarding the composition of the dor. 
 sal and the anterior vertebrae from what a macroscopic view of 
 the parts affords, we would, I think, conclude that they have been 
 
8 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 formed through a union of distinct parts, whether these be regarded 
 as pleurocentra and hypocentra or as principal and intercalated ver- 
 tebrae. 
 
 With reference to the relations of the upper arches to the verte- 
 bral bodies, authors have not been wholly accurate. Franque figures 
 three vertebral bodies and two superior arches ; the bases of the lat- 
 ter reposing equally on the upper surfaces of two contiguous ver- 
 tebral bodies, so that the upper arches are intervertebral in posi- 
 tion. In his figure of the complete skeleton, he represents the bases 
 of all the upper arches of the dorsal region as resting in a similar 
 way intervertebrally. The bases of the succeeding simple vertebrae 
 are represented as resting almost wholly on the bodies of their 
 respective vertebrae, while the upper arches of the remainder of the 
 tail are borne by their respective intercentra. The following quota- 
 tion from Franque is produced: 
 
 ' ' Inter apophysin superiorem et corpus utrimque in parva f os- 
 sula pauxillum cartilaginis inclusum est, quod in spuriis vertebris rep- 
 eris, cui cartilaginis apophysis affixa est, et eo quidem modo ut usque 
 ad locum quendam unaquaeque apophysium duobus vertebrarum cor- 
 poribus addicenda sit." 
 
 Shufeldt (57) reproduces Franque's figures, and states that the 
 bases of the neural arches " articulate between each consecutive pair 
 of vertebrae, these latter having a form to accommodate themselves 
 to this unique condition." Furthermore, in Franque's figures the 
 hinder border of the base of each arch is represented as coming into 
 contact with the anterior border of the next base behind. 
 
 Schmidt refers to Franque's and Shufeldt's descriptions and fig- 
 ures. He affirms that he has not been able to find, in his specimens, 
 the upper arches either to be placed between the vertebral bodies, or 
 to come into contact by their adjacent basal borders. In both his 
 text and his figures he represents the upper arches of the whole dor- 
 sal region as sitting on the hinder half of their respective bodies, and 
 extending over the next body behind by only a little process; also, 
 as having between the successive bases, even in the region close to 
 the head, a considerable interspace. In so doing he has fallen into 
 as great an error as that of the authors whom he attempts to correct, 
 but of an opposite kind. 
 
 When we come to examine these parts accurately, we find that 
 in the hinder portion of the tail, where pleurocentrum and hypocen- 
 trum are both developed, the neural arches have their bases ex- 
 panded anterio-posteriorly, and rest almost wholly on the hypo- 
 centra. Nevertheless, the anterior process of the base projects 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 9 
 
 forward somewhat over the hinder border of the pleurocentrum in 
 front, while the hinder border of its own hypocentrum is left 
 somewhat uncovered. As we move forward we find the neural bases 
 shifted gradually backward, so that in the case of the most anterior 
 free intercentrum the upper arch reposes on its upper surface, extend- 
 ing neither on the pleurocentrum in front nor on that behind. In 
 the hinder dorsal region we find that the neural bases are set still fur- 
 ther backward, so that their hinder angles begin each to overlap 
 slightly the anterior border of the vertebra next behind. This 
 backward displacement of the upper arches goes on until, in the 
 anterior end of the vertebral column, the bases are placed be- 
 tween two vertebrae and rest equally on both. That is, if we 
 regard each dorsal vertebra as consisting of a pleurocentrum and a 
 hypocentrum united, we may observe that on passing from the tail 
 to the head the pairs of neural bases change from a position of rest- 
 ing each on its own hypocentrum and partly on its pleurocentrum to 
 that of resting partly on- its own hypocentrum and partly on the 
 pleurocentrum of the verteibra next behind. 
 
 As regards the distance of the bases of the successive arches 
 from each other, we find that just behind the head they approach very 
 closely, if they do not come into actual contact. Soon there 
 is a space developed between them, and this increases to near the 
 end of the tail. As regards the relations of the arches to the verte- 
 brae and to neighboring arches, it would almost seem as if Franque 
 had drawn his conclusions wholly from an examination of the ante- 
 rior end of the vertebral column, and Schmidt from the other ex- 
 tremity. 
 
 At the hinder end of the head are two vertebrae which are 
 strongly united to the skull. Their presence is indicated by lines 
 running across this portion of the base of the skull, and by their 
 movable neural arches. The halves of each arch meet above the 
 myelon, but do not unite, nor are they prolonged upward. To the 
 upper ends of each of these arches, or at least the hindermost pair, 
 is articulated a compressed inter-spinous bone (axonost). Sagemehl 
 has already recorded the occurrence of these vertebral bodies, the 
 arches and the axonosts. The lateral halves of the first arch behind 
 the head are prolonged above the neural canal but a short distance. 
 To their upper ends is articulated by a freely movable joint a knife- 
 shaped bone, which must be regarded as homologous with the axo- 
 nosts which support the rays of the dorsal fin. The succeeding 
 arches increase gradually in length. To the second is articulated an 
 axonost similar to, but longer than, the first one. In two specimens 
 
io FIELD COLUMBIAN 'MusEU>M ZOOLOGY, VOL. i. 
 
 examined, the third arch has no corresponding axonost. The fourth 
 arch in one specimen has one lateral half articulated to the next axo- 
 nost, while the other half lies free behind the axonost. The fourth 
 axonost is loosely attached to the upper extremity of the fifth arch. 
 I find only four of these axonosts, but Franque figures seven, only one 
 of which appears to be closely connected with a neural arch. Others 
 lie between successive arches, as do two of them in one of my speci- 
 mens. Schmidt very incorrectly figures all the anterior neural 
 arches as being extended oiit into lateral halves longer than those 
 which succeed them. 
 
 Between the lateral halves of each of the neural arches and 
 overlying the neural canal is found a pair of intercalated cartilages 
 (Fig. i, i. c.). These are vertebral in position, as is usually the 
 case with the bony fishes. In Lepisostcus Balfour found similar 
 masses of cartilages to be intervertebrally placed. Above these car- 
 tilages runs the superior longitudinal ligament. In the adults these 
 cartilages become more or less ossified, and the bony layer which 
 inve'sts each cartilage is continuous with that of the corresponding 
 half of the neural arch. 
 
 As regards the lower arches, we find that those of the middle of 
 the tail are supported by the hypocentra alone, and the bases of these 
 arches lack a little of reaching the anterior border of the hypo- 
 centrum. In the posterior portion of the dorsal region the 
 transverse processes arise from the very hinder border of the verte- 
 bral bodies (Fig. 2). As we advance towards the head these trans- 
 verse processes are moved a little forward, until at about the middle 
 of the dorsal region they occupy a position near the middle of the 
 length of the vertebral body (Fig. 3). Near the head again 
 the processes appear again to be slightly nearer the hinder border of 
 the body. With respect to the level of origin of the pro- 
 cesses, it may be observed that at the anterior end of the vertebral 
 column they spring from the sides of the vertebrae halfway up. Fur- 
 ther back the processes slowly descend, so that those of the last dor- 
 sal vertebra arise from its lower side. In length these processes 
 increase from the first to the middle of the trunk, and then gradually 
 grow shorter to the last dorsal vertebral body. In the tail the lower 
 arches are joined to the centra by suture. Below the haemal canal 
 the lateral halves unite to form a spine. The spines of the first four 
 to six caudal vertebrae are articulated to their respective arches 
 just below the point where the lateral halves have united. The 
 elevated origin of the transverse processes in the dorsal region is no 
 doubt connected with the enlargement which that portion of the body 
 cavity has suffered. 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY n 
 
 Franque states that the first vertebra, omitting the two which 
 are consolidated with the skull, has no transverse processes, while 
 the next has these, and likewise sometimes bears ribs. In a specimen 
 before me the processes of the first vertebra are certainly not con- 
 spicuous, but they can hardly be said to be absent. They stand out 
 each as a bony ring surrounding a shallow pit, formed by a shrinking 
 of cartilage. In this specimen, too, these processes support a pair 
 of ribs, each of which is as long as the first five vertebrae taken to- 
 gether. Shufeldt could find no ribs on the first vertebra. In 
 another specimen I find no ribs on the first vertebra, but there is a 
 pair on the second. Schmidt figures the first pair of ribs on the 
 third vertebra. In like manner my specimen has ribs on the last dor- 
 sal vertebra, as also did the specimen figured by Schmidt. Neither 
 Franque nor Shufeldt found such ribs in the specimens investigated 
 by them. 
 
 On the upper and lower surfaces of each of the free pleuro- 
 centra of the tail, in a line on each side with the bases of the 
 arches, are found, in fresh specimens, slightly projecting masses of 
 cartilage. Schmidt describes these and gives figures of them viewed 
 externally and in microscopic section. He regards them as rudiment- 
 ary arches, upper and lower, belonging to the intercalated vertebral 
 bodies. Stannius (58, p. 21) refers to these, and compares them to 
 the masses of cartilage which are seen between the true vertebrae 
 and the arches resting on them. But Franque had observed them 
 still earlier, as is shown in the paragraph already quoted from him. 
 
 Schmidt has described another set of cartilages, which, he claims, 
 have not been mentioned by other writers. These are said to be 
 found in front of the bases of the upper arches of the dorsal ver- 
 tebrae, and they fill up partially the space between the successive 
 arches. Schmidt regards these cartilages as homologous with those 
 just described as occurring on the upper side of the " intercalated " 
 bodies of the tail, and therefore as rudimentary upper arches. I will 
 say here that I have examined these masses by means of transverse 
 and longitiidin^.1 sections, and find that they are not distinct carti- 
 lages, but the anterior portions of the masses which occur between 
 the several vertebral bodies and their neural arches. 
 
 There is another system of cartilages which is of much interest, 
 and which has received little attention. These are found on the under 
 side of all the dorsal vertebrae, there being two to each vertebra. On 
 most of the vertebrae these two cartilages project somewhat beyond 
 the surface of the bone in a linear form, and are placed one on each 
 side of the tract occupied by the dorsal aorta. In the dried skeleton they 
 
i2 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 show themselves as two linear pits running parallel with each other 
 and with the axis of the vertebra. In the more anterior vertebrae 
 they are nearly circular. Here they lie nearer the front end of the 
 vertebral body (Fig. 4). In the succeeding vertebrae the anterior 
 ends of the cartilages come close to the anterior ends of their respec- 
 tive centra, but become gradually more extended backward. Finally, 
 in the hinder two-thirds of the trunk they are so lengthened as to 
 occupy the whole of the length of the centra (Figs. 2 and 3). 
 These cartilages are found even on the vertebral bodies which are so 
 closely united with the head. They appear to have been noticed 
 only by Dr. M. Sagemehl (54, 57) who observed them on the verte- 
 brae at the hinder end of the skull and followed them backward along 
 the vertebral column. He states that in young individuals these car- 
 tilages penetrate deeply into the substance of the centra, while in 
 older specimens only thin plates of cartilage can be recognized rest- 
 ing superficially on the vertebrae. He enters into no explanation of 
 these structures. They will engage our attention later. We may, 
 however, note here, that in the case of three or four species of Amia 
 described by Dr. Joseph Leidy (44, p. 185, pi. xxxii.) from the 
 Bridger Beds of the Eocene of Utah, there occur on the lower side 
 of the vertebral body a pair of fossae, which occupy exactly the posi- 
 tion of the cartilages in Amia. The same fossae have been observed 
 and figured by Prof. E. D. Cope in species of Pappichthys, also occur- 
 ring in the Bridger Beds (17. p. 56, PI. iii.). They are spoken of 
 as two parallel fissures which further backward become oblong 
 fossae, and again near the caudal vertebrae are narrowed. Prof. Cope 
 further says that in the caudal vertebrae these fossae are as well devel- 
 oped as are the neurapophysial pits, and are much like them. He 
 evidently had in mind here the articulatory surfaces of the haemal 
 arches, which are very different things. 
 
 The structure of the terminal portion of the vertebral column 
 has been more or less accurately described by Franque (26), Kolliker 
 (43), Wilder (59), and Shufeldt (57). Of these accounts it seems to 
 me that that of Kolliker is the most complete and his figures the 
 most accurate. Shufeldt presents an original figure of this region, but 
 he omits a representation of the cartilaginous elements, and, as it 
 appears to me, of some of the bones. Dr. Wilder had not seen Kol- 
 liker' s paper when he wrote his notes, and, through what appears to 
 have been a misunderstanding of Aug. Dumeril's language (25, ii. 
 401) he ascribed to Kolliker views not held by the latter author. So 
 far as they bear on the same points, Wilder' s results coincide with 
 those of Kolliker. 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 13 
 
 Kolliker's figures and descriptions cover twelve vertebral bodies, 
 the prolongation of the vertebral axis behind these, and the arches 
 and other parts connected with these two regions. Of the centra 
 described and figured, the anterior seven are called "complete," in- 
 asmuch as they are provided with fully developed neural and haemal 
 arches. The last five are "incomplete," that is, they have haemal 
 arches but not ossified neural arches. Belonging to the seven com- 
 plete vertebras there are, however, only six neural arches, since the 
 second arch, counting from before, is situated over the articulation 
 between the second and third centra, and belongs equally to both. 
 As was first shown by Franque and later by Kolliker, the so-called 
 intercalated vertebral bodies are no longer developed in this portion 
 of the vertebral column. Shufeldt's figure represents the alternation 
 of pleurocentra and intercentra as ending in front of the tenth ver- 
 tebra from the tip of the ossified part of the vertebral axis. This 
 failure of the "intercalated" centra to develop is only a part of the 
 degeneration which has affected this region and been associated with 
 the production of the heterocercal tail of Amia. 
 
 In the case of three specimens examined by myself I find that 
 the vertebra called by Kolliker the last complete one, being the last 
 one furnished with an osseous neural spine, shows a different state of 
 affairs. There is no osseous neural arch, but there are two cartilag- 
 inous plates present, which meet, or nearly meet, above the 
 spinal cord. These are also segmented off from the cartilage 
 which represents the next segment behind. Shufeldt represents 
 this vertebra as having a neural spine, and hence we may conclude 
 that the spine is sometimes developed, sometimes not. 
 
 As Kolliker has shown, the haemal arch of the vertebra just men- 
 tioned is different from those which precede it. The latter all have 
 conspicuous cartilaginous bases, one on each side of the haemal canal, 
 and, as shown by sections, penetrating to the center of the vertebral 
 body. The former arch, represented by a haemal spine, or hypural 
 bone, is co-ossified to the lower side of the corresponding centrum. 
 
 The five terminal vertebrae have ossified centra, and each has its 
 own haemal arch, but there are no distinct neural arches. In the 
 place of these there is a continuous band of cartilage which roofs over 
 the spinal canal, and is prolonged posteriorly to the end of the noto- 
 chord, some distance beyond the distal ends of the hypural bones. 
 The last five vertebrae are more or less reduced in size as we proceed 
 backward, while the portion of the vertebral axis represented by 
 them is turned strongly upward. That portion of the continuous 
 cartilage which overroofs the last five vertebrae represents, as Kolli- 
 
14 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 ker indicates, the neural arches of these vertebrae. These arches have 
 become expanded anterio-posteriorly and have coalesced. Belonging 
 lo each of these arches is a haemal arch in the form of a hypural 
 bone. Each is co-ossified with the under side of its proper vertebra 
 just as is the last hypural belonging to the last so-called complete 
 vertebra. The last three of the haemal arches are directed nearly 
 horizontally backward. These are followed by two other similar 
 bones, which Kolliker says belong to the last vertebra, the hinder- 
 most looking like a continuation of the vertebral column. On this 
 point I shall have something to say when I come to speak about the 
 microscopic structure of these parts. 
 
 Lying between the bone last mentioned and the cartilage which 
 is prolonged behind the ossified centra, Kolliker describes another 
 similar bone which does not reach the last centrum. It functions as 
 .a ray-bearer, and doubtless is the lower arch of a vertebra which has 
 ceased to attain development. I find this bone in two specimens. 
 
 Kolliker describes that portion of the cartilage which projects 
 beyond the last vertebral body as consisting of a tube which incloses 
 the spinal cord, or its representative. On the underside of the carti- 
 laginous rod is a furrow, at first shallow, and in this lies the noto- 
 chord. Posteriorly the furrow deepens, and at length, near the ter- 
 mination of the cartilage, the notochord is enveloped in the substance 
 of the cartilage. Everywhere there is, according to Kolliker, a 
 partition of cartilage between the notochord and the spinal cord. 
 This matter will come up for consideration later. Kolliker regards 
 the extension of cartilage behind the last vertebra as representing 
 not only the upper arches but also the continuation of the vertebra] 
 bodies. It thus represents a cartilaginous vertebral column ("eine 
 ganze knorpelige Wirbelsaule darstellt"). 
 
 The tail of Amia has not attained the stage of complete hetero- 
 cercality, that is, it*s not hypocercal. There are still a few fin-rays 
 lying above the vertebral axis. Of these epural fin-rays, Kolliker 
 figures five, all unsegmented. Shufeldt figures apparently three, 
 all much segmented. The specimens which I have examined show 
 three rays, only the most posterior being segmented sparingly. Of 
 epural interspinous bones, Kolliker figures four, the most posterior 
 being one which lies closely against the upper side of the prolonged 
 cartilage. Shufeldt figures six interspinous bones', the hindermost of 
 which is apparently the one which lies close to the rod of cartilage. 
 This series of bones, as shown by the author last named, completely 
 fills up the space between the last dorsal ray and the upper border of 
 the caudal fin. He is of course in error when he states that these 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 15 
 
 bones have not been previously observed. Shufeldt remarks (57, ex- 
 planation of plate ix.) that these bones, taken in connection with the 
 free spines (interspinous bones) found over the vertebrae immediately 
 behind the head, suggest that in the early ancestors of Amia the fin 
 was continuous from the base of the cranium to the tail. In Oligo- 
 plciirus, of the Upper Jura, which has a far shorter dorsal fin, there 
 is a series of such bones occupying the whole space between the head 
 and the dorsal fin. (Zittel, 60; iii. 231.) 
 
 Of these epural interspinous bones I have found, in two speci- 
 mens examined, also four. The third of these bones is in close rela- 
 tion with the base of the next to the last epural fin-ray, the second of 
 those found by me. The fourth bone is closely bound to the upper 
 surface of the prolonged cartilage. Its proximal end reaches for* 
 ward so as to overlap somewhat the last but one of the ossified verte- 
 bral centra. The distal end appears to be embraced by the lateral 
 halves of the hindermost epural fin-ray. 
 
 2. THE MINUTE STRUCTURE OF THE VERTEBRAL COLUMN. 
 
 Kolliker has dealt with the microscopical structure of the termi- 
 nal portion of the tail. Schmidt has also described and illustrated 
 with a number of figures some points in the finer structure of the 
 vertebral column. 
 
 In all the vertebras of this fish the bases of the arches, upper 
 -and lower, come into contact with the sheath of the notochord. \ In 
 my specimen, 12. 5 cm. long, these bases consist of clear cartilage, but 
 at the inner ends of the bases small portions of the cartilage are ab- 
 sorbed. This process of absorption continues as the individual 
 increases in size, until probably the greater part of the cartilage is 
 removed. While this is going on, trabeculae of bone are depos- 
 ited, which look exactly like the bone deposited elsewhere in the cen- 
 trum; but the crust of bone which first spread over the base of the 
 arches serves to define their boundaries. See figures i, 5, 6, 7. 
 
 It is interesting to observe that in the case of the simple verte- 
 brae, as those, of the dorsal region, the cartilage contained in the 
 upper arch is segmented off at the surface of the centrum from the 
 cartilage enclosed within the centrum; while the cartilage of the 
 arches in the middle of the tail is continuous with that of the cen- 
 trum. 
 
 The cartilages which have already been referred to as appearing 
 in pairs on the underside of each dorsal vertebra, may be now more 
 fully described. They are shown in figures i, 5, 6. As seen in 
 longitudinal section, they may be regarded as thin trigtilanar plates, 
 
16 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 whose bases are seen on the surface of each of the vertebrae, 
 while their apices reach usually to the notochordal sheath. The car- 
 tilage appears to persist even in adult life. It is ensheathed in a thin 
 layer of bone, and in cross sections this is seen to extend outward as 
 far as does the cartilage itself. Hence when the cartilage shrinks 
 through drying, there is left a little rim of bone surrounding the pit. 
 Inmost of the trunk region, the bases of these cartilages do not come 
 into contact with the intravertebral bases of the transverse processes 
 (Fig. i). In the posterior dorsal region, however, where the pro- 
 jesses have descended considerably, the cartilages in question come 
 into contact with the intravertebral bases of the lower arches, 
 and soon have the appearance of being suspended from their lower, 
 or inner, sides (Fig. 5). Toward the last dorsal vertebra these 
 cartilages, which, from their relation to the aorta, may be called the 
 aortal supports, become shorter. In the last dorsal vertebra (Fig. 6) 
 the cartilage is short, and is attached to the intravertebral basis of the 
 haemal arch half-way from the notochord to the outer surface of the 
 vertebra. In the vertebras of the tail the cartilages are missing. 
 There is, however, in my younger specimen, what seems to be ves- 
 tiges of them in the first caudal vertebra. Nothing, however, can be 
 more certain than that the lower arches of the trunk are bent down 
 to form the arches of the tail, and that the aortal supports have 
 nothing to do with the formation of the caudal haemal arches. In the 
 last dorsal vertebra we find, at the anterior end, the aortal supports 
 looking somewhat like the lower arch-bases. Further back, 
 however, the latter come into view, in a series of sections, and 
 the aortal supports cling to their inner surfaces (Fig. 6). Figure 
 7 shows a section through the first caudal. The bases of the lower 
 arches are seen in same position as in the last dorsal. 
 
 The origin and the homology of these aortal supports will be fur- 
 ther considered when we come to examine younger specimens. In 
 many, if not most, of the osseous fishes, there is to be found on the 
 under side of each of the dorsal vertebrae, a pair of bony ridges, one 
 on each side of the aorta. To these ridges the aorta appears to be 
 suspended. In the anterior portion of the vertebral axis of Acipenser 
 a plate of cartilage grows downward from the lower edge of the base 
 of each half of the haemal arch, and gives support and protection to 
 a portion of the aorta. Further back, these cartilaginous plates bend 
 inward towards each other, and finally unite below the aorta, thus 
 enclosing it in a canal. It appears to me that the aortal supports of 
 Amia may be regarded as the equivalents of those cartilaginous pro- 
 cesses which protect the aorta in Acipenser. In Amia the cartilages 
 

 
 
 m .0 f I'l nsmiosqa lo Bidaliav iBeiob lohaJnf: ns rfgooirii noiloaS .1 .giH 
 , siv .tnorfooJon aril ol gniiEiJsnsq a^" 1 auonigfifiliBO 5o eiir^q aaiiil adi gniwoila ,not 
 iswol erto lo saasd ariJ ,89iloi laqqu aril lo esafid aril noqqoa rioidw 
 
 .elioqqua Ijslios srfl bns 
 
 jjDiworia .namiosqa ifubfi lo Bidslisv Ifiaiob dl()8,,3ffi Jo waiv IsiinaV .S . 
 -qua I^liojs gooniBlinO sdl ^d baiquooo SIB rioiriw ancd aril ni zqsg aril 
 
 .aiioq 
 
 .UtibB io asidaliav Ifiaiob rlli-1 hni; di<:I in v/siv iBilnaV .F, .giH 
 .llubfi io Srtdsiiav iBaiob bni bnB laf lo wsiv IsilnaV .i- .gi"*! 
 .nol .m .D c.Sf ns^paqg lo jsidaliav [saiob slBroillunaq lo noiloaS .c .i'i 
 .Jrr.1 erf. 1 gjs nerniosqg smBa lo jndaliav Ifiatob Jafil lo noiloaS .0 .gi'i 
 
 'Hi lo atbbirn aril 
 
 AMI A 
 
PLATE I. 
 
 Fig. 1. Section through an anterior dorsal vertebra of a specimen 12.5 c. ra 
 long, showing the three pairs of cartilaginous rays penetrating to the notochord, viz., 
 the cartilages which support the bases of the upper arches, the bases of the lower 
 arches, and the aortal supports. 
 
 Fig. 2. Ventral view of the 3(5th dorsal vertebra of adult specimen, showing 
 especially the gaps in the bone which are occupied by the cartilaginous aortal sup- 
 ports. 
 
 Fig. 3. Ventral view of 13th and 14th dorsal vertebrae of adult. 
 
 Fig. 4. Ventral view of 1st and 2nd dorsal vertebrae of adult. 
 
 Fig. 5. Section of penultimate dorsal vertebra of specimen 12.5 c. m. long. 
 
 Fig 6. Section of last dorsal vertebra of same specimen as the last. Taken 
 near the middle of the vertebral centrum. 
 
FIELD COLUMBIAN MUSEUM 
 
 ZOOLOGY. PL. 
 
 ..sp.c 
 
 O.P.H it:. 
 
 BMeisel lith.Bostm 
 
 AMIA 
 
FIELD COU 
 
 .nsmiosqa amis. io id9Ji;> v lbjjf> 
 
 .gaol .fnm cl ,nmk gnuoy; lo aoi^nVH^^BBI^BE 
 isqqu adJ io esasd sriJ awodS 
 
 sdl no Jesi aavlasoiadi doid 
 .mm ? ftiwl. lo BidsJisv leaiob dJ8 dgiioiriJ noiJosa saia 
 
 .8 .g 
 ri sdi bniriad-iaui 
 
 .G 
 
 AMI A 
 
PLATE II. 
 
 Fig. 7. Section through the first caudal vertebra of same specimen. 
 
 Fig. 8. Sagittal section through the axial region of a young Amia, 15 mm. long. 
 Taken just behind the head. Shows the bases of the upper arches placed between 
 the intercalated cartilages, which themselves rest on the elastica externa. 
 
 Fig' 9. Transverse section through 8th dorsal vertebra of Amia 27 mm. long. 
 
FIELD COLUMBIAN MUSEUM 
 
 ZOOLOGY. PL. II 
 
 A.a, 
 
 0.?H iJe! 
 
 B-MeisellitUwM. 
 
 AMIA 
 
FIELD COLUVHAK MUSEUM ZOOLOGY. PL 
 
 .III 3TAJ C I 
 
 lo list adJ lo 9gd sdl JK noigai IBIXB aril riguoidl noiiosa iKrubuJi^ooJ .01 .iH 
 isqqu adl .bnsd Jrfgn -)ri) no .vlfBiD^qHS f.-MoriZ ^nol .mm 08 WtwK lo aarniDaqa 
 isqqu stii )o noiJieoqisic; ;l*BS^9JfilE3i9Jni isv/of sdl briB 
 
 ;biorioolon sril bn esdoiK i^qn 31 
 no asrfoiE iswol srf; Jo ?.s?.sd orfj nsswldd gniinsv/ STJ; aagEliJifi:) baJfilBDiaJni sdJ 
 
 .bnsri rfsl 
 
 lo am LI loo IjaidtiJisv erfl aX^v/aiv sbiS .It .^i'i 
 
 bsJjsIfioisJni esgfilbiBD adi bo fid eworiS .-\MWI\ 
 
 sdJ lo noilioq bsievoonrj HE esriou, > ii nasvtftfad oals ; msdi 
 
 .biodooloa 
 .mrnf;S; trnwk ne lo^M^^Wl^MDrohsJeoq E dyooidJ noilosa-asoiD .SI .1"? 
 
 lo aJnioq }zaih3 aril lo noilEUlia srii ^HEiosqaa ewodg 
 
 AND ACiPENSER. 
 
PLATE III. 
 
 Fig. 10. Longitudinal section through the axial region at the base of the tail of 
 a specimen of Amia 30 mm. long. Shows especially, on the right hand, the upper 
 and the lower intercalated cartilages; on the left hand, the interposition of the upper 
 intercalated cartilages between the bases of the upper arches and the notochord; 
 while the intercalated cartilages are wanting between the bases of the lower arches on 
 the left hand. 
 
 Fig. 11. Side view of a portion of the dorsal region of the vertebral column of 
 Adpenser. Shows bases of upper and lower arches and the cartilages intercalated 
 between them ; also between the two series of arches an uncovered portion of the 
 notochord. 
 
 Fig. 12. Cross-section through a posterior dorsal vertebra of an Amia 23mm. 
 Shows especially the situation of the earliest points of ossification. 
 
FIELD COLUMBIAN MUSEUM 
 
 ZOOLOGY. PL. III. 
 
 
 O.P.H del. 
 
 
 not. 
 
 B Meisel Mh Bos'w 
 
 AMIA AND ACIPENSER. 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 17 
 
 are apparently somewhat more independent of the haemal arches, but 
 there seems to be no reason why this condition should not be the re- 
 sult of specialization. 
 
 A microscopical examination of the caudal region of a specimen 
 about 1 2 smm. in length, has given me results which may be of some 
 value. 
 
 The last but one of the " complete" vertebrae is in all respects 
 similar to the vertebrae preceding it. Both arches, the upper and the 
 lower, sit upon cartilaginous bases, and these latter penetrate close 
 to the notochord. The next vertebra, like the corresponding verte- 
 bra in each of the two specimens microscopically examined, lacks the 
 neural spine, while also the lateral halves of the neural arch appear 
 not to meet over the spinal canal. The cartilages on which the neu- 
 ral arch is based reach nearly to the center of the vertebra, so that 
 on the upper side the notochord is constricted, just as in the case of 
 the more anterior vertebrae. The haemal arch, on the contrary, has 
 no such broad cartilaginous bases, although its proximal end contains a 
 nodule of cartilage. The lower arms of the cross seen in cross-sections 
 of the more anterior vertebras are here wanting, and the lower side 
 of the notochord is extremely little constricted. The ossified hypural 
 spine is consolidated with a thin shell of bone which forms the lower 
 side of the vertebral centrum. The appearance presented suggests 
 that the base of the haemal spine has originated at a relatively late 
 period in the development of the individual, so that it has not stood 
 in the way of the growth of the notochord. 
 
 Giving our attention to the five vertebrae which follow the one 
 just described, we find that each possesses the upper arms of the car- 
 tilaginous cross and that upon the ends of these arms, rest the edges 
 of the strip of cartilage which roofs over this part of the spinal canal ; 
 just as the neural arches of the more anterior vertebrae rest on the 
 upper ends of the cartilaginous crosses of their respective vertebrae. 
 Moreover, that part of this roofing cartilage which overlies the first two 
 of these last five vertebrae is almost wholly segmented off from that 
 behind it by an incision which begins in the lower border and runs 
 upward. This, taken in connection with the spineless condition of 
 the vertebra just preceding, presents a gradual transition between 
 the normal neural arches and those which have become concrescent ; 
 so that it is quite certain that this anterior portion of the prolonged 
 cartilage is made up alone of the neural arches. It is also to be ob- 
 served here that the upper side of the notochord has suffered con- 
 striction at the points touched by the lower ends of the intravertebral 
 cartilages. 
 
i8 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 The five hypural bones of this section are coossified to their re- 
 spective vertebrae in the same way as has been described in the case 
 of the vertebra immediately preceding. No cartilage is present, but 
 the lower surface of the vertebral centrum is composed of a thin 
 crust of bone, possibly of cartilaginous origin, and to this is coossi- 
 fied the corresponding hypural. The notochord enclosed by these cen- 
 tra has, on the underside, undergone little constriction. 
 
 My longitudinal sections do not confirm Kolliker's view that the 
 next two hypurals are attached to the terminal centrum of the tail. 
 While the base of the first of these hypurals does come close up to the 
 hinder border of the last centrum, it is entirely distinct from the lat- 
 ter. At the same time its proximal end reaches backward and sup- 
 ports its own portion of the notochord. The next hypural is in its 
 turn brought into contact with the notochordal sheath still further 
 back, and is not at all coossified with the last developed centrum. 
 An examination of the next to the last hypural bone shows that its 
 base closely resembles that of the vertebrae in front; since there is, 
 next to the notochordal sheath, a layer of ossified cartilage. We 
 have therefore some reason to believe that the base of this hypural 
 has become coossified with the lower portion (hypocentrum) of an 
 imperfectly developed vertebral centrum. The base of the last 
 hypural is also somewhat expanded, and this expansion possibly rep- 
 resents another rudimentary centrum. It is not at all improbable that 
 at a later period of life these hypurals coossify by their proximal ends 
 jto the hindermost vertebral centrum, so as to produce the condition 
 described by Kolliker. 
 
 It is interesting to note that the notochord over these two hypu- 
 rals just described, has suffered slight constriction, as if in sympathy 
 with the effort to develop additional centra. 
 
 In my sagittal sections I find, in place of the staff-like bone 
 which is sometimes seen above the last described hypural, and which 
 is itself undoubtedly but another hypural, a row of three cartilages. 
 The two most anterior of these nodules of cartilage are covered with 
 a layer of bone. 
 
 Turning our attention now to the band or rod of cartilage which 
 is prolonged beyond the ossified vertebral column and ends between 
 the lateral halves of the twenty-first caudal ray, we find that Kolli- 
 ker, as already stated, regards it as representing not only the upper 
 arches, but also the continuation of the vertebral centra. According 
 to Kolliker's description and figures, the cartilage under considera- 
 tion forms, for some distance beyond the last vertebral body, a com- 
 plete tube, which encloses the spinal cord, or its representative. On 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 19 
 
 the underside of this tube is a furrow, at first shallow, but growing 
 deeper towards the hinder end. In this furrow runs the notochord. 
 Near the tip of the cartilage the edges of the furrow meet and co- 
 alesce below the notochord. Kolliker's figures show a relatively 
 thick wall of cartilage everywhere between the spinal cord and the 
 notochord. In my sections there is, on the contrary, from the last 
 ossified centrum to the tip of the notochord, no cartilage between 
 the latter and the spinal cord. The narrow space between them is 
 occupied by loose connective tissue and blood-vessels. The lower 
 borders of the cartilage incline inward between the two structures, 
 but do not meet. The condition described by Kolliker is probably 
 .attained at a more advanced age. Farther back, at a point some dis- 
 tance behind the distal ends of the hypurals, the cartilage gradu- 
 ally spreads downward over the sides of the notochord, and finally 
 completely encloses it. Under these circumstances it appears to me 
 probable that this whole strip of cartilage has resulted from the 
 downward extension of coalesced neural arches, and none of it from 
 the lower arches. It is also interesting to observe that there are in 
 this portion of the terminal cartilage seven shallow nicks along its 
 lower border. These suggest a partial separation into distinct seg- 
 ments. 
 
 The notochord in this region has a thick cuticular sheath, outside 
 of which is to be distinctly seen the elastica externa. In my sections 
 -of the hinder tail region of the specimen 125111111. long, the noto- 
 chordal sheath is far from having a structureless appearance. Espec- 
 ially along the portion where the incomplete vertebrae are developed, 
 but also further backward, the sheath is traversed from elastica to 
 notochord by lines, or fibers, of some substance which stains deeply 
 with carmine. Where the notochord is constricted by the upper por- 
 tion of the incomplete vertebrae, these fibres appear to start from the 
 lower surface of the elastica. Here they are so close-set as to resem- 
 ble the hairs of a brush, but as they approach the inner surface of the 
 sheath they become finer and less numerous. 
 
 In this same region the inner cuticular sheath has undergone 
 another modification, to which I will call attention. Where the last six 
 vertebrae of the upturned portion of the tail have constricted the 
 notochord, especially on the upper side, we find that the outer half, 
 or somewhat less, of the cuticular sheath stains deeply. One is 
 reminded of the modification suffered by the sheath of the notochord 
 of Lepisosteus, as described by Balf our and Parker (4, vol. i, 781; 
 vol. iv., pi. 41 (Fig. 69). InAmia, however, the pulley-like band so 
 modified does not extend far forward and backward beyond the 
 
20 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 most constricted portion of the centrum. In Lepisosteus it extends- 
 in each direction to the end of the centrum. In Amia the fibres- 
 described in the preceding paragraph appear to pass right through 
 the modified band of the sheath. 
 
 3. DISCUSSION OF THE VERTEBRAL COLUMN OF FOSSIL FISHES. 
 
 Having now considered the structure of the vertebral column of 
 the living Amia, it is proper to determine what light we may obtain 
 regarding its peculiarities by a consideration of such fossil forms as 
 may be related to it. For a presentation of the facts relating to the 
 Amioid fishes the reader is referred to Dr. Zittel's " Handbuch der 
 Palseontologie, " vol. iii. Through Amia, Eurycornms, Callopterus, 
 Caturus, and related genera, we are apparently led back to more and 
 more primitive arrangements. From a study of these forms attempts 
 have been made to reach an explanation of the mode of development 
 of the vertebrae of these fishes. That adopted by Dr. Zittel may be 
 thus stated. The hypocentrum (originally of two lateral halves) is 
 developed on the lower side of the notochord in each myomere, and 
 supports the lower arch. On the upper side of the notochord are devel- 
 oped two pleurocentral plates, which may also become consolidated 
 into one piece. Such an arrangement may be found in Caturus. In 
 a more advanced stage of development the hypocentrum grows fur- 
 ther upward on the sides of the notochord, while the extremities of 
 the pleurocentrum grow downward. The borders of the two ele- 
 ments may at length come into contact, and thus cover in the whole 
 surface of the notochord. Such a condition may be seen in some por- 
 tions of the vertebral column of Caturus and in that of Callopterus. 
 Further growth of the plates results finally in the upper ends of the 
 hypocentrum meeting above the notochord, and thus forming a com- 
 plete ring. In a similar way, the pleurocentrum of each myomere 
 forms another ring. We find such rings in their complete develop- 
 ment in the middle region of the tail of Amia. 
 
 On the other hand, Ludwig Schmidt appears to regard the two 
 rings in each myomere as the primitive condition, and each of these 
 as a distinct vertebral centrum, the one with, the other without, 
 neural arches. As to the manner of development of the vertebral 
 centra of the dorsal region, Schmidt presents two contradictor} 7 views. 
 He at first states that these vertebrae must be regarded as resulting 
 from the coalescence of two such vertebras as are found in each myo- 
 mere of the tail (56, pp. 755 and 760). And he presents as evi- 
 dences and illustration of this fusion the varying number of com- 
 plete vertebrae in the anterior tail region and the occasional fusion. 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 21 
 
 of the two centra in the middle tail region. On another page of his 
 paper, when Schmidt comes to consider the fossil forms with rhachi- 
 tomous vertebrae, such as Eurycormus and Euthynotus, he explains 
 these structures on the hypothesis that, starting with the two rings in 
 each myomere throughout the body, the hypocentrum has grown at 
 the expense of the pleurocentrum. The latter becomes reduced in 
 size and is confined to the upper side of the notochord. Eurycormus 
 furnishes one stage ; Euthynotus a more advanced one. At length 
 the pleurocentrum becomes a mere rudiment ; and in Amia, the last 
 of the series, Schmidt finds the pleurocentrum represented by only 
 the small masses of cartilage, "the rudimentary upper arches,," alleged 
 to lie in front of the bases of the developed upper arches. Of course, 
 one or the other, or both, of these explanations must be wrong ; and 
 neither of them explains the arrangement found in Caturus, as is 
 admitted by Schmidt. On the other hand, the theory that the rings 
 have resulted from the growth of the pleuro-and hypocentral plates, 
 seems capable of explaining all the known facts. 
 
 Having considered the mode of development of the vertebrae of 
 the Amioidei, it may be profitable to inquire whether or not any 
 other groups of fishes have possessed similar vertebras. The so- 
 called Ganoids first claim our attention. 
 
 Our modern species of Lepisosteus show no indications in their 
 adult condition of the presence of pleuro-and hypocentra. But Aspi- 
 dorhynchus, a member of the Rhynchodontidae, had dorsal vertebrae 
 composed evidently of two portions. These formed rings, which 
 were divided along the sides by sutures. This being true of the dorsal 
 region, we can hardly floubt that the caudal vertebrae also originated 
 from once distinct pleuro-and hypocentra. The tubular vertebrae of 
 Belonorhynchus, too, must have had a similar origin. 
 
 In the family Macrosemiidae, Cope ^Saurodontidae, Zittel) some 
 genera have simple tubular vertebrae. Others, as Eugnathus; have 
 the vertebrae composed each of two distinct pieces. The species of 
 Pholidophorus sometimes possess undivided vertebrae, but usually the 
 centra are made up of pleurocentra and hypocentra (Zittel, 60, p. 
 315). In the tail these two pieces stand nearly opposite each other 
 and form a bony sheath on which rest the upper and the lower 
 arches. It seems quite improbable that the closely related genera 
 fropterus and Histionotus should have developed vertebrae in an en- 
 tirely different way. 
 
 Little is known concerning the state of the vertebral column of 
 the Sphaerodontidae and the Stylodontidae. At any rate, the ossifica- 
 tions were feeble. Tctragonolepis is said to have had ring- like vertebrae. 
 
22 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 Dapedius is regarded as having possessed composite vertebrae. These 
 meager accounts of the vertebral structures of these fishes make it 
 probable that all the species of the families, so far as they possessed 
 anything representing vertebral centers, had the two elements, pleu- 
 rocentra and hypocentra. And the same having been shown to be 
 true, or highly probable, of every family of the order Lepisostei, 
 we are, I think, justified in reaching the conclusion that whenever 
 ossification of the vertebral column has begun in members of this 
 group, it has been through the formation in each somite of a lower 
 piece or pieces, on which rest the lateral halves of the haemal arch ; 
 while on the upper side of the notochord upper pieces of bone have 
 been deposited in the region of the bases of the upper arch ; that is, 
 pleurocentra and hypocentra are normally constituents of the ver- 
 tebral centrum. 
 
 Of the condition of the ancient Crossopterygia, as regards the 
 vertebral column, we have little knowledge. Prof. Cope (15, p. 19} 
 states that in Ectosteorhachis (Megalichthys) the vertebrae are repre- 
 sented by annular ossifications resembling those of Cricotus. There 
 appears, however, to be only a single ring in each myomere. 
 
 From the foregoing survey of the condition of the vertebral 
 column in the different families of the so-called Ganoids, one may 
 easily become convinced that originally the ossified vertebral centra 
 of all the species that possessed such were composite in their struc- 
 ture; that is, each centrum included in its composition a pleurocen- 
 trum and a hypocentrum. Each of these two elements was doubtless 
 itself, at an earlier period of its history, double. If the centrum is 
 in any given case simple in structure, this has probably resulted from 
 the coalescence at some time in the animal's life of the two elements, 
 or possibly sometimes from the suppression of one of them. If we 
 find in one fish the notochord surrounded in one portion of the body 
 by a solid vertebral body, and in another portion by two complete, 
 but distinct, rings; or in another fish the notochord in one region 
 protected by pleurocentra and hypocentra, while in another region 
 there are mere rings, it appears quite improbable that these different 
 structures have no genetic connection. It is equally improbable that 
 in species of the same family, or of closely related families, the ver- 
 tebral centers originated sometimes as mere tubular incrustations of 
 the notochord, sometimes as double rings, and sometimes as pleuro- 
 and hypocentra. Some one of these structures must have been 
 the fundamental one ; and if so, then the elements which have been 
 called pleuro-and hypocentra must be regarded as the most primitive. 
 
 To what extent do the conclusions we have reached apply to the 
 fishes which we call the Teleostei ? 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 23 
 
 By the consent of perhaps all modern ichthyologists, the relation- 
 ships of the Ganoids and the Teleosts are very close. Dr. Gill (31) 
 regards the two groups as distinct subclasses, but admits that in some 
 of their members they approach closely. Liitken (46) excludes from 
 the Ganoids large groups which have been held to belong there. 
 Among these excluded forms are the Amioidei. Others abandon the 
 conception of the Ganoids as forming a distinct group. Cope forms 
 of the bony fishes, and some of the Ganoids, his subclass Actinopteri. 
 It is hardly open to doubt that the Teleosts have descended from more 
 or fewer of the families of ancient Ganoids. The Halecomorphi are 
 intimately related to the Isospondyli, and have probably furnished 
 the ancestors of the latter. Hence we might reasonably expect to 
 find in some of the recent or fossil Teleosts, in the stricter sense, 
 traces of the composite nature of the vertebral centra. And in fact 
 Prof. Cope (22) expresses the opinion that the fossil genera which 
 possess annular vertebrae, as Aspidorhynchus, should be assigned to 
 Isospondyli. According to the views of Liitken, most of the genera 
 which display pleuro-and hypocentra would be thrown among the 
 Teleosts. The Hoplopleuridae, whose members range through the 
 Mesozoic, are now usually assigned to the Teleost series. In some of 
 these the vertebral column is imperfectly or not at all ossified. Be- 
 lonorhynchus has feebly ossified vertebrae. The vertebral elements 
 appear as small ossifications at the bases of the arches, and are prob- 
 ably to be regarded as pleuro-and hypocentra. Deecke (24, p. 131) 
 points out the close resemblance of this fish to Be/one belone, living 
 now in the Atlantic. He states that if the vertebral column-of Bel- 
 onorhynchus were ossified it would be difficult to distinguish the two 
 genera. We are therefore justified in expecting that our living 
 Teleosts will in some way display in their vertebral structure the 
 existence of the elements which we have found in so many of the 
 Ganoids. Nevertheless, we are not yet in a condition to demonstrate 
 the presence of such elements. The vertebrae of our fishes have not 
 yet been investigated sufficiently from this point of view to enable 
 us to speak with definiteness. 
 
 It is, of course, entirely possible that in the lapse of ages one or 
 the other of the elements of the vertebral centrum of the ancient 
 fishes has entirely disappeared, while the other remains as its sole 
 constituent. And this is what Prof. Cope appears to believe has hap- 
 pened. He says (22, 1019) : " The so-called centra of fishes are inter- 
 centra [hypocentra], as in the Batrachia." This opinion may be cor- 
 rect, but the grounds on which it appears to have been based seem to 
 afford it insufficient support. " The descriptions and figures of Prof. 
 
24 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 Zittel render it perfectly clear that the fishes rarely develop complete 
 vertebral centra [pleurocentra], the order Halecomorphi being the 
 only example." On the contrary many genera belonging to a 
 mimber of families are mentioned by Prof. Zittel as possessing pleu- 
 centra and hypocentra. 
 
 4. DISCUSSION OF THE VERTEBRAL COLUMN OF THE HIGHER 
 VERTEBRATA. 
 
 But the Ganoid and Teleost fishes are not the only vertebrates 
 which manifest in their vertebral centra the possession of the ele- 
 ments which we have found so well developed in the vertebrae of the 
 Amioid fishes. Hermann von Meyer was the first to point out that 
 the vertebrae of the fossil Archegosaurus, among the Amphibians, 
 were each made up of a numher of distinc t pieces. This genus, and 
 others closely related to it, show indeed a remarkable resemblance in 
 the structure of their vertebrae to the fishes related to Amia, espec 
 iaily to Caturus. Others again, as Cricotus and Diplovertebron, have 
 the vertebrae of at least the tail constructed exactly like the vertebrae 
 of the middle portion of the tail of Amia. Some other fossil genera, 
 as Mastodonsaurus, usually have solidly ossified vertebrae, but Von 
 Meyer has shown that such vertebrae may, during the youth of the 
 animal, have the rhachitomous structure of Archegosaurus. Some spe- 
 cies, again, as those classified as Lepospondyli, had simple tubular 
 vertebrae ; but in closely related genera, such vertebrae consisted each 
 of a right and left half, the two meeting by a suture along the dorsal 
 and ventral sides of the notochord. Whether in such cases we have 
 two pleurocentra or two hypocentra, or the two elements coalesced, 
 it may yet be impossible to say. Cope holds the view that the verte- 
 brae of modern Amphibians consist only of hypocentra. This view 
 will be discussed further on. 
 
 On the other hand it has been shown by Cope, Albrecht, Dollo 
 and Baur, that the vertebral centra of the reptiles, birds and ani- 
 jnals are really pleurocentra, while the hypocentra appear only 
 occasionally in reptiles, especially as " subvertebral wedge-bones," 
 and in all Amniota as the " body " of the atlas. 
 
 The finding of the elements pleurocentra and hypocentra in the 
 vertebral column of so many of the osseous fishes, the amphibians, 
 and the Amniota may well lead us to suspect that future investigations 
 will reveal a still more general participation by them in the structure 
 of the vertebral column. Such investigations may further prove to 
 what extent these elements enter into the constitution of the primi- 
 tive vertebra from which all others have been derived. 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 25 
 
 5. THE ORIGIN AND DEVELOPMENT OF THE VERTEBRAE OF AMIA DURING 
 
 ITS EARLY STAGES. 
 
 Since my results obtained from the study of adult specimens 
 have been put in practically the form presented above to the reader, 
 I have had the good fortune to acquire a considerable amount of lar- 
 val and young material. Some of this, consisting of larvae from 10 
 to i5mm. in length, was received from Dr.J. E. Reighard, of the Uni- 
 versity of Michigan, through the kind offices of Dr. Eycleshymer, of 
 the University of Chicago. For the possession of the young of larger 
 growth, varying from 23 to 44mm., I am indebted to Dr. S. A. 
 Forbes, of the University of Illinois, and his assistant, Prof. Frank 
 Smith. The University of Illinois supports a biological station on 
 the Illinois river at Havana; and it was here that Prof. Smith secured 
 the materials sent me. As a result of a study of the young of Amia, 
 I am able to throw some light on the development of its vertebral 
 column, and that of fishes in general. 
 
 I have prepared transverse and sagittal sections of specimens 
 lomm. in length. An enlarged figure of a specimen of this length 
 may be found on PI. xxx. of Mr. Allis's paper on Amia (i.) In 
 specimens of this size the notochord is already vacuolated. The 
 walls of the vacuoles stain deeply with carmine, but not, at least for 
 me, with hsematoxylin. Peripherally they pass into a layer resem- 
 bling in every way themselves, but thicker. The thickness of this 
 external layer I make to be about .oo66mm. but it is thicker where 
 the vacuole walls enter it. It apparently represents the epithelial 
 layer, which is seen at a later stage, but I find in it no traces of nu- 
 clei, or little else to suggest the presence of cells. Outside of it lies 
 a highly refractive, extremely thin layer, one of the sheaths of the 
 notoehord. 
 
 Anteriorly the pointed end of the notochord is buried in the' con- 
 nective tissue closing in the pituitary space, but the greater portion 
 of the cranial division is enclosed in the well developed cartilage at 
 the base of the skull. This cartilage may be followed backward, in 
 longitudinal sections, into the two ridges of the skeletogenous tissues 
 which give origin to the upper arches. The specimens of the length 
 being considered do not all appear to be in the same stage of devel- 
 opment. In one examined the most anterior upper arches consist of 
 [hyaline cartilage. In another, the cartilage at the base of the skull 
 passes insensibly into procartilage in the region of the neural arches. 
 In the intervals between the nerve roots may be seen the bases of the 
 upper arches, distinguished apparently only by being slightly more 
 
26 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 condensed than the tissues which lie between the arches. On the 
 lower side of the notochord there may be seen at intervals, in sagittal 
 sections of the more advanced lomm. specimen, little masses of more 
 dense, more deeply staining tissues, the beginning of the lower 
 arches. In cross sections these may be traced from the lower outer 
 side of the notochord downward to near the aorta, their lower ends 
 staining a little more deeply in haematoxylin. In the less advanced 
 specimens the indications of the arches, upper and lower, fade out 
 posteriorly ; but in other specimens they may be followed into the 
 tail, where also the cartilage becomes better differentiated. Here 
 the lower arches may be seen to extend down the sides of the aorta 
 and pass into the indifferent tissue which surrounds the caudal vein. 
 The upper arches, however, have become very short, mere cartilagi- 
 nous papillae at the sides of the spinal marrow. 
 
 When we come to examine specimens having a length of i5mm. 
 we are able to observe a number of important changes in the axial 
 structures. The notochord has attained a diameter of about .3mm. 
 There is present a distinct cuticular sheath, the thickness of which I 
 make .0033111111. Outside of this is an extremely thin, but distinct, 
 membrana elastica. Lying within the cuticular sheath, and surround- 
 ing the notochord there is present a very distinct epithelial layer. 
 The nuclei of the cells stain deeply, and the whole layer has a thick- 
 ness of about .oo66mm. How this layer of cells originates I am tin- 
 able to say. I find no traces of it in a specimen i2mm. in length, 
 except it be the already described cell-less layer just inside the cuti- 
 cular sheath. Scheele describes and figures a distinct epithelial 
 layer as existing in the Trout (55). Goette has denied the exist- 
 ence of such a structure, but it has been demonstrated by a number 
 of competent observers, as Gegenbaur, Grassi, etc. 
 
 In the anterior dorsal region the upper arches rise two-thirds the 
 height of the spinal cord, passing above into the delicate connective 
 tissue which closes in the latter organ. In the posterior region they 
 are very short. When sagittal sections of this length are examined, 
 we find that in the anterior region some new elements are beginning 
 to show themselves. These are small masses of incipient cartilage, 
 which are placed immediately in front of the bases of the upper 
 arches (Fig. 8). Whether or not these masses are ever at any 
 time united with the bases of the arches by continuous cartilage, 
 incipient or well developed, I find myself unable to say definitely. In 
 nearly all cases there may be seen at least a narrow line of connective 
 tissue cells surrounding each little mass and cutting it off from the 
 bases immediately in front and behind. But occasionally, both while 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 27 
 
 the cartilage of the mass is in the undifferentiated condition and 
 when it has become quite hyaline, a section may be found in which 
 the bases and the intermediate masses appear to be quite continuous. 
 These masses are, as is to be expected, best developed in front, and 
 become, in this stage, very obscure about the middle of the back. 
 They rest immediately on the elastica, while the bases of the arches- 
 seem to have been crowded upward out of contact with it. If these 
 little cartilages just described are ever actually continuous with the 
 bases of the arches immediately before and behind them, they cer- 
 tainly become first cut off from the bases just in front, and later from 
 the bases immediately behind them. 
 
 In the region of the tail, masses of cells suggesting the early 
 stage of cartilage are seen to lie between the bases of the successive 
 half-arches, and often they cannot be sharply separated from the ad- 
 joining arch bases. This applies both to the upper and the lower sides. 
 of the notochord. While we may be somewhat uncertain whether 
 or not these newly appearing masses of cartilage have a direct con- 
 nection at any time with the bases of the arches, this much is very 
 certain : there is never at any stage a contimiation of the cartilage 
 of the bases, or of the intermediate masses of one side, across the noto- 
 chord to those of the other side. Nor have I found anywhere that 
 the cartilage of the upper arches is continuous with that of the lower 
 arches. Hence there is nothing resembling a tube of cartilage en~ 
 closing the notochord. 
 
 In the anterior portion of the body the cartilages which repre- 
 sent the lower arches have not changed greatly from their earlier 
 condition. The cartilage is better differentiated, showing now cejls 
 lying in a distinct ground-mass. The masses appear in the form of 
 flat bands lying against the outer sheath of the notochord. The 
 bands come down quite close to the aorta, and it is at their lower end 
 that we find the cartilage best defined. In the hinder dorsal region 
 the bands appear to descend to a lower level on the sides of the noto- 
 chord, the lower ends coming into quite close contact with the aorta. 
 Furthermore, each band, or arch-base, sends outward from near its 
 upper end a short process into the region where later the proximal 
 end of a rib will appear. This may be seen throughout the whole 
 trunk region. In the tail the halves of each lower arch have united 
 at their distal ends, so as to enclose the blood vessels. 
 
 My specimen next in length to those just described is 23mm. 
 long. In this the notochord has attained a diameter of about .4mm, 
 exclusive of the two sheaths. The epithelial layer has no longer the 
 distinct appearance that it had in the preceding stage. Here and 
 
28 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 there nuclei may be distinguished in the deeply staining layer which 
 surrounds the notochord ; but the latter has in many places an appear- 
 ance as if it also had become vacuolated. If so, the mesh work is very 
 fine. I have not in any stage of development seen any such columnar 
 epithelium as that figured by Hasse. (39, Taf. vi., Fig. 14.) 
 
 It may be as well to state here that I have measured the diame- 
 ter of the notochord in specimens up to 44mm. in length, and find 
 that in the vertebral regions, where it is of course most constricted, 
 the diameter has not appreciably increased beyond about .4mm. And 
 yet in a specimen about 12.5011. the opening through the vertebral 
 centrum has a diameter of about .6mm., and this is filled up by the 
 .notochord, which retains the appearance it has in younger specimens. 
 This shows that this organ, instead of being suppressed by the devel- 
 opment of the cartilaginous and bony structures that develop around 
 it, continues to grow slowly until the animal reaches a considerable 
 size. In respect of its notochord, Amia displays a transitional stage 
 between the lower "Ganoids" and the " Teleosts." 
 
 Surrounding the notochord of the specimen 23mm. in length is 
 found the cuticular sheath, whose thickness has increased greatly 
 since the earlier stage. This thickness is now about .O2mm. There 
 .are seen to be indications of the fibrous structure, which has been 
 observed by several writers. Outside of the just described cuticu- 
 lar sheath is the elastica. It forms a thin, highly refractive lin- in 
 cross-sections, and often presents a wavy course. This I attribute 
 to distortion due to postmortem changes. I have never found in it 
 the slightest indication of cells, nuclei, or any other structure. 
 
 The axial structures outside of the notochord and its coverings 
 are of great interest. Throughout the length of the body the lateral 
 halves of the neural arches have met above the neural canal, and 
 their backwardly directed distal ends lie alongside of each other. As 
 is known, they never unite to form neural spines. Interspinous car- 
 tilages are also now developed, both above and below. The proxi- 
 mal member (axonost) is long and slender, its proximal extremity 
 reaching nearly to the tips of the arches, upper and lower. Each 
 is followed distally by two short nodules of cartilage. 
 
 In a cross-section through the dorsal region, the upper arches 
 may be seen to rest on short cartilages which themselves come in con- 
 tact by their lower ends with the elastica (Fig. 9) . Against the right 
 .and left sides of the notochordal rod we find the bases of the lower 
 arches. These, as in the case of the upper arches, are of hyaline 
 cartilage, and their bases are in direct contact with the notochordal 
 sheaths. Distally there is attached to the outstanding process of the 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 29 
 
 half-arch on each side a bar of cartilage, the rib. This rib, as in the 
 case of the ribs of osseous fishes, passes outward immediately between 
 the peritoneum and the lateral muscles. 
 
 Anteriorly, at the third vertebra, the bases of the lower arch 
 stand at the sides of the notochord close to the bases of the upper 
 arch, but I do not think that they at any time coalesce. As a series 
 of sections are. run through toward the rear, the bases of the lower 
 arches descend to a lower level and a greater space intervenes between 
 them and the upper arches. 
 
 It has already been noted that the bases of the lower arches, 
 especially along the anterior portion of the vertebral column, send 
 downward towards the aorta on each side a process of cartilage. 
 As may be seen from Fig. 9, the lower end of this thickens and 
 grows downward, so as to project a slight distance below the noto- 
 chord. These lower swellings of the cartilage are, as may be sur- 
 mised, the beginnings of the lowest cartilaginous rays which have 
 been described as occurring in the adult vertebrae, and which have 
 been called the aortal supports. My supposition, derived from the 
 examination of these structures in the adult, that they originate from 
 the lower arches is fully confirmed. As in the case of the lower 
 arches themselves these aortal supports develop from the front to the 
 rear. When a specimen 3omm. long is examined by means of cross- 
 sections, it is seen that, when about the twenty-fourth vertebra is 
 reached, the aortal support is cut off from the main cartilage by a 
 narrow spicule of bone. Behind this region the support is apparently 
 developed independently of the main mass of the lower arch. The 
 whole appearance suggests that this terminal portion of the arch- 
 base, becomes developed into true cartilage just a little too late to 
 escape being separated from its parent mass by the now developing 
 bone. 
 
 Scheele (55) has observed in certain Cyprinidae (Rhodeus) struc- 
 tures which, consisting of fibrous tissue, arise from the lower and 
 hinder portions of the lower arches, and descending on each side 
 unite below the aorta. This vessel is therefore surrounded below 
 each vertebral body by this structure and supported and protected 
 by it. Scheele regarded it as an outgrowth from the base of the 
 lower arch, and I think that he was correct in his conclusion. The 
 structures which he describes are undoubtedly homologous with the 
 aortal supports of Amia. But in the Cyprinidae there has occurred 
 an enormous reduction of the cartilage which usually enters into the 
 formation of the axial structures, and these supports have felt the 
 effect of it. I think that in most fishes there will be found on the 
 
.30 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 lower side of the vertebras some rudiment of these supports, either 
 in the form of bone or dense connective tissue. 
 
 But Scheele is not content merely to note the occurrence of these 
 connective tissue aortal supports. He regards them as being of great 
 morphological importance, the representatives of a series of arches, 
 .absent, so far as yet known, from all other Teleosts. They are called 
 by him " hsemapophyses, " not indeed homologous with what have 
 usually been called such in the tail of osseous fishes, but with the 
 lower arches in the tail of Selachians, Urodeles and Ganoids. Scheele 
 would have us believe that in Rhodeus there are present representa- 
 tives of three series of arches, viz.: (i) the neural arches, (2) the 
 ribs and their homologues in the tail which enclose the haemal canal, 
 .all of which he calls parapophyses, and (3) the vestigial ' ' haemapo- 
 physes." The latter have been inherited from the Selachians and 
 Ganoids, but have, with the alleged exception of those of Rhodeus, 
 "been lost in the Teleosts. The parapophyses, therefore, belong to a 
 .series of arches which are placed at a higher level in the body of the 
 iish, derived originally, indeed, Scheele claims, from the upper arches, 
 " Theile der obern Bogen." 
 
 Now, there is not an argument advanced by Scheele for the pur- 
 pose of establishing his ideas regarding the aortal supports as repre- 
 sentatives of a hitherto unrecognized series of arches in Teleost 
 fishes, and as homologous with the lower arches of sharks, and Gan- 
 oids, and Urodeles, that will not apply with greater force to the car- 
 tilages which support the aorta in Amia. And yet it is perfectly ob- 
 vious that these processes from the parapophyses do not become mod- 
 ified in this Ganoid's tail into the lower arches. As we pass back- 
 ward in a series of cross-sections of an Amia 27mm. long, we find 
 that near the last dorsal vertebra, the cartilaginous bases of the lower 
 .arches assume in section a triangular form. The parapophyses arise 
 from the upper outward angle, the one pointing outward beneath 
 the peritoneum, the aorta supports from the lower inner angle. The 
 outer angle is plainly prolonged in the last dorsal vertebra ; \vhile in 
 the next vertebra, the first caudal, it is united with its fellow beneath 
 the caudal vessels. In short, the parapophyses are developed from 
 the upper portion of the bases of the lower arches, the aortal sup- 
 ports from the lower portion. The latter are simply differentiated 
 parts of the bases of the lower arches. 
 
 In Fig. i o is shown a sagittal section through the axial region of 
 a specimen of Amia somewhat larger than the one just studied; it is 
 3omm. long. The section includes the anterior and a portion of the 
 middle tail region. The notochord and its two sheaths are repre- 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 31 
 
 sented, and the cartilages that rest on them above and below. Wher- 
 ever these cartilages come in contact with the notochord, the latter 
 structure is seen to be slightly constricted, more in the case of the 
 larger cartilages than of the smaller ones. 
 
 But the cartilaginous structures are of the greatest interest to us. 
 We are now in the region in which, in the adult, there occurs a change 
 from the single to the double vertebrae. The arches which are seen 
 on the extreme left of the figure are those of the second caudal ver- 
 tebra Between the long haemapophyses toward the right are seen 
 three small nodules of cartilage. On the upper side of the notochord 
 are seen other neural arches, and between the bases of those on the 
 right hand two small cartilages like those found between the lower 
 arches. But it is evident from the section that a third little mass of 
 cartilage has fused with the base of the next half-arch toward the 
 right. The section being taken a little to one side of the middle line, 
 it is evident that for each little nodule of cartilage seen, there is an- 
 other on the opposite side. Hence there would be four such masses 
 in front of the upper and lower arches belonging to each myomere. 
 
 It will be further observed that the upper arches on the left side 
 of the figure appear to be segmented at short distance below their 
 bases, while the upper arches standing behind the small nodules of 
 cartilage are not thus segmented. This suggests that those basal 
 pieces on which the larger portion of each half-arch appears to rest 
 are really the equivalents of those little masses which lie between the 
 undivided arches. This is shown, too, I think, not only from the 
 fact that the little cartilages of the middle tail region and the basal 
 pieces of the region in front of them occupy corresponding positions, 
 but also because of the similarity in the manner of their development. 
 Both are evidently formed in the intervals between the bases of the 
 arches immediately in front and behind. Those of the dorsal and an- 
 terior tail region secondarily push themselves backward under the 
 bases of the neural arches. The upper arches become differentiated 
 first, and their bases come down to the membrana elastica. At the 
 same time incipient cartilage cells may be made out lying between 
 the bases, possibly continuous with them. Later, these also are trans- 
 formed into clear cartilage, and increasing in size so as almost to 
 touch one another, they push the bases of the neural arches away 
 from contact with the notochord. 
 
 In the case of the lower arches of the dorsal region, we do not 
 find little cartilages lying between them. On the contrary, we find 
 that the anterior end of the base of each half-arch has grown forward, 
 vso as to occupy the place of the missing nodule. Or, it is posssible 
 
32 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 that at sometime in the history of the larva, or during its ancestral 
 history, the base of the lower half-arch has fused with the little inter- 
 mediate cartilage which lay in front of it ; but I have seen no signs 
 of its presence. It has likely been suppressed. Now, as will be 
 readily conjectured, and as will be further demonstrted, the ''interca- 
 lated vertebrae" of the adult, or as we have called tjiem, tJie pleurocentra, 
 have originated from the four little masses of cartilage lying in front of 
 the arches belonging to each myomere; while the hypocentra have had their 
 origin from the union of the upper and lower arches. Furthermore, the 
 simple vertebra of the region in front of this are developed from the bases 
 of the lower arches and the basal cartilages on which the upper arches 
 rest. 
 
 Again, there are none of the extra cartilages placed between the 
 bases of the upper and lower arches in the upturned region of the 
 tail. It is possible that, on account of the crowding together of 
 structures in this region, these nodules never made their appearance 
 there. Or, being possibly younger productions, they were the first to 
 feel the effects of influences leading to degeneration, and at length 
 disappeared. 
 
 Before we investigate the later history of the arches and the 
 accompanying cartilages, it will be well to inquire whether anything 
 corresponding to the latter occur elsewhere, especially in related 
 lower forms. These are found in a position entirety similar to what 
 are called intercalated cartilages in sharks. Whether or not they are 
 homologous structures in the two forms I do not here decide. But 
 similar intercalated cartilages are found also in the cartilaginous Gan- 
 oids. Fig. it has been drawn from the dorsal region of Acipen- 
 3er. Here we have upper and lower arches resting by their bases- 
 on the sheath of the large and partly exposed notochord. Between 
 the bases of the upper arches are seen intercalated masses of cartilage, 
 each mass consisting of from one to four pieces. Between the ex- 
 panded bases of each two of the lower arches is also a group of car- 
 tilages varying in number, size and form. These cartilages were 
 long ago observed by V. Baer and J. Muller. (V. Baer, 2; J. Muller, 
 49, p. 87.) While these intercalated cartilages are thus broken up in 
 Acipenser into a number of distinct pieces, we find that in Polyodon 
 (Spatularia) each, "both above and below, consists of a single piece. 
 See figure of a portion of the vertebral column of Spatularia in 
 Wiedersheim's " Vergleichende Anatomic der Wirbelthiere." 
 
 If now, in either Acipenser or Polyodon, the upper intercalated 
 masses were to increase in size to any considerable extent in an ante- 
 rio-posterior direction, the bases of the upper cartilages would be 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 33. 
 
 either lifted tip out of contact with the notochord or be suppressed. If 
 in addition to this, the lower intercalated cartilages were to be sup- 
 pressed or to become fused with the bases of the lower arches, we 
 should have exactly the same condition that we find in the greater 
 part of the vertebral axis of the young Amia. 
 
 ' It has already been shown that in Amia the notochord is not at 
 any stage surrounded by a tube of cartilage. At most there are to 
 be found only four bands of more or less differentiated cartilage, two 
 above and two below the notochord, and it is not certain that these 
 bands are continuous. In Lcpisosteus, however, according to Balfour 
 and Parker (4) the notochord is at one stage almost completely en- 
 closed by a tube of well defined cartilage. The tube is interrupted 
 only by gaps at the sides of the notochord between the bases of the 
 upper and the lower arches. Nothing like distinct and independent 
 intercalated cartilages were observed" by these authors; but "inter- 
 vertebral rings of cartilage " were observed and figured by them. 
 These were placed in the intervals between successive pairs of arches, 
 and completely surround the section of the notochord where they 
 occur. Balfour and Parker found that these rings do not have a uni- 
 form thickness, but that each showed two dorsal and two ventral 
 thickenings. These thickenings are in line with the bases of the 
 arches, and were regarded by the observers as presumptive evidence 
 that the rings had been produced through the fusion of the bases. 
 However, there were missing the very stages in which the rings took 
 their origin, and it is by no means certain that they are not produced 
 from the fusion of four intercalated cartilages for each body segment. 
 If the four intercalated cartilages of each segment in the tail of Amia 
 should expand and then fuse with one another and with the adjacent 
 bases of arches, we should have the condition found in Lepisosteus. 
 
 Ossification of the vertebral column begins probably soon after 
 the young Amia has reached a length of i5mm. At least, no deposit 
 of bony matter has been observed in specimens of this length, while 
 such deposits are abundant in a specimen 23mm. long. 
 
 The place of origin of the earliest formed bone of the vertebra? 
 of fishes has been much discussed and has received from investiga- 
 tors the most varied and often contradictory answers. In the settle- 
 ment of this question is further involved the view that is to be held 
 concerning the composition of the vertebra, especially whether or 
 not any part of it is derived from the envelops of the notochord. In 
 discussing this subject we shall, for the sake of brevity, consider that 
 any part of the notochord or of its sheaths which does not change to 
 either cartilage or bone does not form a portion of the vertebra. 
 
34 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 Von Baer (2 p. 36), basing his conclusions on his studies especially 
 of the Cyprinidae, thought that no part of the vertebra is derived from 
 the notochord or its sheaths. For him the vertebra was formed 
 through the simple union of the bone arising from the upper and the 
 lower arches and extending around the notochord so as to produce 
 a ring. 
 
 From the study of Blennius viviparus Rathke (51) concluded that 
 each vertebral centrum arises from ossifications which are located in 
 the bases of the upper and the lower arches, which four points by 
 their union form a ring. Later Rathke (52) appears to have accepted 
 J. M tiller's view that the vertebral ring is distinct in its origin from 
 the arches. 
 
 Johann Miiller at first (49, Osteology, pp. 83 and 143) held the 
 opinion that ossification of the vertebral column begins outside of the 
 notochordal sheaths, and that the vertebra is composed wholly of the 
 arches and the ossifications arising in them. Later, from studies on 
 the Elasmobranchs, he concluded (49, Neurology, p. 69) that in the 
 case even of the osseous fishes, the notochordal sheath ossifies, and 
 that an undivided ring is formed independently of the arches. The 
 vertebra then consists of the upper and the lower arches and the cen- 
 tral ring. 
 
 August Miiller's views are essentially those of Johann Miiller (48, 
 p. 260.) 
 
 Gegenbaur has expressed his conclusions on the origin of the ver- 
 tebral body in his paper on Lepisosteus (28.) He believes that the 
 centrum arises outside of the elastica externa, and that it proceeds 
 from the arches themselves. 
 
 Theophil Lotz (45) made investigations on Salmo salar. For him 
 the vertebral centrum arises outside of the elastica, its earliest ap- 
 pearance being at the points of contact of the lower arches with the 
 notochordal sheaths. It then grows upward on each side of the no- 
 tochord, and its extremities at length pass under the bases of the 
 upper arches and meet above the notochord to form a complete ring. 
 
 Oscar Cartier (13) agrees with Lotz, except that he finds that the 
 ossification which produces the ring arises in the elastica. The latter 
 disappears as the ring is developed. 
 
 Goette (32 and 33) contends that ossification begins in what is 
 usually called the elastic sheath. He denies the elastic nature of 
 this sheath, and holds that it is cellular in structure. The centrum 
 is developed from below upward, as taught by Lotz and Cartier. 
 
 Balfour, in 1881, in his Comparative Embryology appears to 
 accept in general the views of Goette. 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 35 
 
 In 1882 Balfour and Parker published their extremely important 
 paper on the structure and development of Lepisosteus (4). They 
 showed in this work that there is surrounding the notochorda cuticular 
 sheath and a very distinct elastica externa. Cartilaginous arches 
 arise both above and below the notochord, their bases coming into 
 contact with the elastica. At a somewhat later stage, there is found 
 to be an almost complete tube of cartilage surrounding the notochord 
 outside of the elastica. This has already been described on another 
 page. At this stage a thin crust of bone envelops the upper and the 
 lower arches and the whole vertebral body. But the bone has not 
 arisen in the elastic layer of the notochord, neither does it extend into 
 this. While its exact place and manner of origin is so far unknown, 
 it apparently does not even arise close to the elastic sheath, but is 
 separated from the latter by the layer of cartilage. Here, so far 
 as we can see and as was believed by Balfour and Parker, the verte 
 brae originate from the union simply of the two sets of arches and 
 their subsequent ossification. 
 
 Klaatsch (41 and 42) holds that the vertebrae of the Teleostomi 
 arise wholly outside of the notochordal sheath. 
 
 Scheele (55) who has investigated Rhodeus among the Cyprinidae 
 and the Trout among the Salmonidae, regards the skeletogenous layer 
 outside of the elastica as giving origin to all the ossified portions of 
 the vertebrae. 
 
 As a result of my investigations I conclude that in Arnia, ossifi- 
 cation of the vertebral centrum does not begin in either the cuticular 
 sheath or in the elastica. On the contrary, it takes its origin wholly 
 in the layer of embryonic connective tissue which lies immediately 
 outside of the elastica. The latter notochordal sheath may be seen 
 in young, just ossifying Amia, and even in adults with a length of 
 i2.5cm., as a bright, highly refractive line immediately between the 
 layer of developing bone and the cuticular sheath. Indeed it is n'ot 
 infrequent that one may distinguish cells resembling bone-cells lying 
 between the elastica and the lamella of growing bone. This is most 
 likely to be observed near the ends of the centra. 
 
 The earliest osseous rudiments of the vertebral body do not form an 
 undivided ring. On the contrary, the bone which is to lay the foun- 
 dation of each centrum starts from a number of distinct points. 
 These are located in the angles between the right and left sides of 
 each half -arch and the adjacent portion of the membrana elastica. 
 There are thus two of these centers of ossification for each half-arch 
 .and two for each of the intercalated cartilages of the middle tail 
 region (Fig. 12). I have thoroughly satisfied myself by the exam- 
 
2,6 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 ination of cross and longitudinal sections that the bony deposits found 
 at the base of each half arch do not at first extend all around the 
 base. The bone is wanting both in front and behind the base; neither 
 does it pass between the base and the elastica. 
 
 While it is possible to regard these centers of bone growth as 
 'common to the arches an,d to the tissues in which the primitive ring 
 is developed, as Rathke thought, they appear to appertain more 
 properly to the arches. On their first appearance they extend fur- 
 ther up on the arch than they do along the circumference of the no- 
 tochord, and in many instances they are seen hardly to touch the lat- 
 ter. In other words, the centers belong to the arches rather than to 
 any tissue in closer relations with the notochord. However, the little 
 deposits of bone soon- begin to extend themselves over the notochord 
 more rapidly than along the arches, and then they appear to belong 
 rather to the former. 
 
 The bone is perichondrial in its origin where it is in connection 
 with the cartilage of the arches. It forms at first an extremely thin 
 layer wholly devoid of any bone-cells, but as the layer thickens these 
 cells are seen to be included. When, however, the bone lies against 
 the membrana elastica, the bone cells are included from the start. 
 
 Through the extension of the various centers of each vertebra 
 around the notochord and their union with one another, a continuous 
 ring of bone is formed. Meanwhile, the two centers at the base of 
 each arch have grown forward and backward and met, and the deposit 
 continues to advance until it covers the whole centrum with a bony 
 crust. In a larva 3omm. long, ossification has so far advanced in the 
 dorsal region as to join all the centers into one band, which has 
 attained nearly the full length of the centrum ; while in the middle 
 of the tail only a tiny flake of bone is to be seen on each side of the 
 base of each half-arch. 
 
 Soon after the vertebral band has closed around the notochord 
 the bone begins to push itself out into the connective tissue. The 
 new deposit is in the form of narrow bars which shoot out at right 
 angles from the primitive ring, and these again become united by 
 other bars more or less parallel with the ring. A loose meshwork 
 of bone is thus finally produced, such as is to be seen in the adult 
 vertebra. This network of bony spicules is first seen between the 
 bases of the lower arches of the dorsal region in the loose connective 
 tissue just above the aorta. 
 
 In the case of the centers of ossification belonging to the bases 
 of the lower arches in the trunk region, I find the bone appearing on 
 the lower surface of the cartilage, both mesiad and laterad of the 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 37 
 
 aortal support (Fig. 12). The latter structure is, of course, even 
 in adult life, not covered with bone at its distal end. Whether the 
 bone found on the two sides of this originates on one side and then ex- 
 tends around it to the other, or whether there are two distinct centers, 
 I am unable now to say. As already remarked, it is found that along 
 the posterior dorsal region the crust of bone is developed a little 
 sooner than the cartilaginous aortal support, so that the latter never 
 lias a direct connection with the main portion of the cartilage of the 
 lower arch. The hindermost of the supports are rather tardy in 
 making their appearance. As already stated, all the intercalated car- 
 tilages, both of the dorsal and of the caudal regions, become ossified, 
 ach from the two centers, which soon unite into one. The upper 
 half-arches of the dorsal and anterior tail regions are incrusted each 
 with a sheath of bone, which arises at an earlier period than do the 
 ossifications closer to the notochord. This sheath covers each half- 
 arch except at the proximal and distal ends. The upper half-arches 
 of the middle tail region and all of the lower half-arches have ossifica- 
 tions at their bases, next to the notochord, each arising from two cen- 
 ters, and have also a sheath of bone covering the greater extent of 
 their length. This sheath, as in the case of the dorsal half-arches, 
 leaves the distal extremity of the cartilage free, and proximally comes 
 down close to the basal ossification, without, however, ever uniting 
 with it at any time of life. We may therefore say that each half- 
 arch has two ossifications, that contiguous to the notochord and the 
 distal sheath, except the upper half-arches of the dorsal and anterior 
 tail regions. These have only the distal sheath, and lack the basal 
 ossification. My judgment is that the latter has been suppressed by 
 the ossification arising from the neighboring intercalated cartilage. 
 The cartilage found in each half-aich is unbroken from the notochord 
 to the tip, except those of the dorsal and anterior tail regions. 
 
 We are now to learn how the vertebral bodies of the different 
 regions are formed. In the trunk and anterior tail regions each cen- 
 trum is produced by the fusion of the bones which have arisen from the 
 upper intercalated cartilages with those at the bases of the correspond- 
 ing lower arches. The upper arch has nothing whatever to do with the 
 formation of the centrum. Its sheath simply rests on the upper sur- 
 face of the centrum, and is separated from it by a permanent suture. 
 A parapophysis of less or greater length forms on each side, a part 
 of each vertebral centrum, being originally a portion of the lower 
 arch. In the anterior tail region there is no parapophysis, the ossifi- 
 cation not extending beyond the surface of the vertebra. 
 
38 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 In the middle tail region we find apparently two kinds of centra r 
 those with and those without arches. The latter, the so-called pleu- 
 rocentra, are formed wholly through the union of the ossifications 
 arising from the upper with those arising from the lower intercalated 
 cartilages. Of course the cartilages themselves continue to consti- 
 tute a part of such a centrum. On the other hand, the centra of this 
 region which are furnished with upper and lower arches are produced 
 through the fusion of the ossifications which arise at the bases of 
 these two arches. They contain nothing more ; unless, to be exact,, 
 we expressly mention the persisting cartilage and the enclosed por- 
 tion of the notochord and its sheaths. 
 
 The posterior tail vertebrae are mostly like the vertebrae last de- 
 scribed. They are produced wholly through the union of the bases 
 of the upper and lower arches. We find, therefore, in the vertebral 
 column of Amia no support for the doctrine that the vertebra is com- 
 posed partly of a central portion which is distinct in origin from the 
 cortical portions, and which has its origin in one of the notochordal 
 sheaths. In Amia the whole vertebra originates from the ossifications- 
 which spread out from the bases of the arches or from the intercal- 
 ated cartilages, or from both. Clear evidences to the same effect are 
 given by Balfour and Parker's work on Lepisosteus. In that fish, too, 
 neither of the notochordal coverings furnished any portion of the 
 vertebral centrum ; all the bone developed from the cartilages sur- 
 rounding the notochord. Indeed, it is only where there are gaps in 
 the cartilaginous tube surrounding the notochord that the bone comes 
 at all in contact with the elastica, and evidently the bone did not orig- 
 inate where these gaps are. The notochord can therefore hardly be 
 truly said to take the active part in the development of the bony ver- 
 tebral axis that some authorities have attributed to it. 
 
 Nevertheless, I am not prepared to say that the coverings of the 
 notochord never undergo ossification. In some of the more advanced 
 larvae of Amia itself, I find that the inner sheath, the cuticula chordae, 
 has every optical appearance of having become calcified, and it stains- 
 exactly like the basis of the bone outside of the sheaths. But I am 
 not sure that this has happened. Goette agrees with Kolliker that in 
 Anguilla the cuticula chordae undergoes partial ossification (33, p. 
 125). It is, however, the outer sheath, the elastica, which is regarded 
 by some writers as calcifying and furnishing the primitive ring of the 
 vertebra. It is quite certain that this elastica is a secretion product 
 of the innermost layer of the skeletogenous tissue. It would not 
 then be strange if the earliest film of bony substance poured out by 
 this skeletogenous tissue were to occupy the apposed elastica. But 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 39 
 
 the more recent investigations are rather opposed to the doctrine that 
 the elastica does become calcined. It is quite evident that primitively 
 it does not. 
 
 6. FURTHER CONSIDERATION OF THE VERTEBRAL COLUMN OF FOSSIL 
 
 FISHES. 
 
 It becomes now proper to consider the composition of the verte- 
 brae of the larval Amia in connection with the fossil members of the 
 group and those of the stegocephalous Amphibians. 
 
 In doing so we cannot fail, it seems to me, to recognize in the 
 upper intercalated cartilages and the ossifications arising from them, 
 the pleurocentra, the bones which cover in the upper surface of the 
 notochord in both the ancient Amioid fishes and the Stegocephali. In 
 a form like Caturus (Zittel, 60, p. 138, Fig. 146) we find the pleuro- 
 centrum occupying a position between the bases of the upper arches 
 just as do the intercalated cartilages in the tail of Amia. In Gallop- 
 terus the upper arches repose on the pleurocentra, out of contact with 
 the hypocentrum, as they do in the dorsal region of Amia. If this con- 
 clusion is correct, and the homologising of the intercalated cartilages 
 with those of Acipenser and Poly o don is justified, we reach the impor- 
 tant result that the pleurocentrum, which plays such an important 
 part in some fishes, amphibians, and all Amniota, has developed from 
 the upper intercalated cartilages of the lower fishes. 
 
 In the tail of our young Amia these pleurocentra furnish only 
 the upper half of the bony material which enters into those vertebral 
 centra which have no arches. The explanation which has hitherto 
 been given of these centra is, that the lateral extremities of the pleu- 
 rocentra have continued to grow downward around the notochord 
 until they have met and coalesced on its under side. From Amia we 
 learn that this is not, at least always, the case. Two other elements 
 in each muscle-segment contribute to the formation of this vertebral 
 ring; these are the two lower intercalated cartilages and their ossifi- 
 cations. In fishes no distinct bones which correspond to those lower 
 elements appear yet to have been discovered ; but in the stegocepha- 
 lous genus Arehegosaurus, Hermann von Meyer (47, p. 104) has de- 
 scribed as being found in each vertebra a distinct ossification which 
 occupies the space between the bases of adjacent lower arches, and 
 on the lower side of the notochord. This ossification Fritsch (Fauna 
 der Gaskohle) calls the hypocentrum pleurale. In the tail of Archc- 
 gosaurus, according to Hermann von Meyer, this bone is divided into 
 two lateral halves. In this condition it corresponds in every respect 
 to the two ossifications in each segment, which result, in an early 
 
40 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 stage, from the two lower intercalated cartilages of Amia. It seems 
 to me, therefore, that we must recognize these two lower intercalated 
 cartilages and their ossifications as representing the hypocentrum pleu- 
 rale. For the sake of brevity I propose in this paper to employ for this 
 bone the name hcemacentrum, although the term is not wholly satisfac- 
 tory. The vertebra which has hitherto been called the pleurocen- 
 trum, is therefore a pleuro-haemacentrum. 
 
 In the Aimioid fishes, and in the stegocephalous Amphibians, as 
 has been described, the lower surface of the notochord is often em- 
 braced, in each segment, by a curved plate of bone which rises on 
 the sides of the notochord and to which the lower arch is closely 
 attached. This plate, sometimes divided into two lateral halves, has 
 received the name hypocentrum. In Amia, the ossifications which 
 produce the lower portion of each centrum that is provided with 
 upper and lower arches, spread from the bases of the lower half-arches 
 of that vertebra. These ossifications belong to the arches from which 
 they spring as truly as the ossifications connected with the interca- 
 lated cartilages belong to the latter. As already stated they meet in 
 the middle line below and grow upward on the sides of the notochord. 
 These horse-shoe-shaped pieces of bone must therefore correspond 
 to the hypocentra of the ancient representatives of Amia, and doubt- 
 less to the hypocentra of the Stegocephali. I would therefore define 
 the hypocentrum as the vertebral element which results from the 
 union of the ossifications arising from the bases of the lower arches. 
 To the hypocentrum may be attached by suture or coossification the 
 bones springing from the remainder of the lower arch. In the trunk 
 region of Amia the parapophyses must be regarded as portions of the 
 hypocentrum ; In the tail region, the lower arches, which enclose the 
 blood-vessels, are morphologically distinct bones, although they be- 
 long to the same cartilaginous arches as the hypocentra. 
 
 In Amia the hypocentra of the middle tail region grows upward 
 on the sides of the notochord and unite with the ossifications which 
 grow downward from the corresponding upper arches. The verte- 
 bral centra thus formed are those which have borne the name of 
 hypocentra, or intercentra. They are not hypocentra simply. We 
 might call the bone resulting from the union of the ossifications of 
 the bases of the upper arches, epicentra. The vertebrae now under 
 discussion would then be epi-hypo-centra. 
 
 In the dorsal and anterior tail regions the vertebrae are produced 
 each by the union of the hypocentrum and the pleurocentrum. They 
 are therefore pleuro-hypocentra. 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 41 
 
 A proper consideration of the facts deduced from the study of 
 the development and composition of the vertebrae of Amia suffice to 
 give us notions regarding the manner in which the vertebrae of this 
 tish and its ancestors have been produced quite different from those 
 which have hitherto been held. The two bony rings belonging to each 
 muscle-segment of the tail of Amia, and of some other fossil fishes, in- 
 stead of being in one of the last stages of vertebra-formation, needing 
 only to unite in order to form a definitive vertebra, are really in one 
 of the earlier stages. The transformation of the condition found in 
 the tail of Amia to that in the dorsal region, has apparently been 
 effected by the great increase in size of the hypocentral portion of the 
 epi-hypocentrum, the resulting suppression of the haemacentral por- 
 tion ot the pleuro-haemacentrum, the enlargement of the pleurocen- 
 tral portion .of the pleuro-haemacentrum and the concomitant sup- 
 pression of the epicentral portion of the epi-hypocentrum. In 
 other words, the lower portion of the anterior of the two rings was 
 suppressed by the expansion of the lower portion of the hinder ring, 
 while the uppermost portion of the anterior ring grew at the expense 
 of the uppermost portion of the hinder ring. The dorsal vertebrae 
 of Eurycormus, (Zittel, 60, p. 230, Fig. 242) it seems to me, are in a 
 stage of transition from the one condition to the other. That the 
 epicentrum, the ossification at the bases of the upper arch, was at 
 length suppressed, is evident from its not appearing in either fossil 
 or in larval Amias. In the latter the upper arch is not simply lifted 
 up out of contact with the notochord ; its base is aborted, the ossifi- 
 cations which ought to appear there never developing. 
 
 In case this course of transformation has been pursued, there 
 must have been stages in which the sutures between the two elements 
 were oblique or parallel with the axis of the fish, instead of perpen- 
 dicular. And such, indeed, appears to have been the condition in 
 some of the Lepisostei (Pholidopkoruf). Dr. Zittel says of the genus 
 named : "In der hinteren Caudal-region stehendie Hypo-und Pteu- 
 rocentren einander fast gegeniiber und bilden dadurch eine knb'cherne 
 Scheide, auf welcher die untern und oberen Bogen aufsitzen." 
 
 While the course of modification which has just been described 
 has been pursued by many, perhaps most, of the relatives of Amia, 
 and possibly by other fishes, it appears probable that some, such as 
 Cat urns and Callopterus (Zittel, 60, pp. 221 and 230, Figs. 241, 243) 
 followed, so far as they went, a different course. In Caturus the ver- 
 tebral elements often lack much of covering the notochord. Never- 
 theless the lower intercalated elements, as an ossification at least, is 
 absent, and possibly also the base of the upper arch. Evidently no 
 
42 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 
 
 ossified rings have ever existed. The pleurocentrum and hypocen- 
 trum push themselves over the notochord until they come into con- 
 tact and form triangles on the sides of the notochord. Whether such 
 elements ever continue to grow around the notochord and produce 
 complete rings I do not know. In view of what happens in Amia, it 
 would appear to be most reasonable to deny this until the evidence is 
 made unquestionable. It is interesting to note that in both Caturus and 
 Callopterus, according to Zittel's account, no rings appear in any por- 
 tion of the body. Euthynotus possibly furnishes us with a case in 
 which the hypocentrum has grown around a considerable portion of 
 the notochord, while the pleurocentrum is reduced to small propor- 
 tions, but the epicentrum is probably present and the genus therefore 
 belongs nearer Eurycormus. With the same elements to start with, 
 the various genera have developed a considerable variety of struc- 
 ture in their vertebral structure, and it will require much investigation 
 to discover all the relationships. 
 
 Attention ought to be called to the fact that the pleurocentrum 
 and the ring containing it belong to the anterior end of the vertebra, 
 or the vertebral complex, into which it enters. In the middle 
 tail region of Amia, whenever, with rare exceptions, which may be 
 regarded as abnormal, the two vertebral rings unite, the one contain- 
 ing the pleurocentrum is anterior. In the dorsal region, it is the 
 intercalated cartilages placed in front of each arch which push them- 
 selves backward beneath the latter and unites with the hypocentrum 
 to form the centrum. It is necessary to keep this in mind in exam- 
 ining the text and figures of Dr. Zittel's Handbuch, in which the 
 statement is sometimes made that the pleurocentrum is the hinder 
 element. 
 
 I have already, on page no, called attention to Schmidt's expla- 
 nation of the development of the dorsal vertebrae of Amia. My invest- 
 igations on the young of Amia have demonstrated that neither of his 
 hypotheses is true. There is no fusion of two such simple rings for 
 each myomere as are found in the tail. Furthermore, the pleurocen- 
 trum is not suppressed by the growth of the hypocentrum. Such a 
 process as he describes would make of the centrum simply a hypo- 
 centrum. On the other hand, in Amia, the pleurocentral element 
 forms the whole upper half of each dorsal vertebra. 
 
 7. FURTHER DISCUSSION OF THE VERTEBRAL STRUCTURE OF THE 
 
 TELEOSTEI. 
 
 We may return for a moment to the consideration of the verte- 
 bras of our living Teleosts. Does the pleurocentrum enter into the 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 43. 
 
 composition of any of them ? I am not able to answer definitely. I 
 am not aware that anything like intercalated cartilages has been found 
 in any of our bony fishes belonging to the group of Teleosts. The 
 Clupeidce, as being most closely related to the Ami idee, might be ex- 
 pected to throw light on the question. In case that the cartilages 
 forming the upper arms of the cross found in the vertebrae of some 
 fishes, as Lucius (sox), were segmented off from that contained in 
 the remainder of the upper arch, I should regard it as representing 
 the pleurocentrum ; but I find mention of only one fish showing such 
 a condition. This is the genus Monacanthus, referred to by Goette 
 (33, p. 129). Usually the cartilage of the upper arch comes down on 
 the notochord, and is unbroken from end to end, as it seems to be in 
 Lucius and Salmo. It may be that in these the pleurocentrum is 
 wholly missing. 
 
 I do not think that there can be any question that the lower por- 
 tion of the osseous ring which encircles the notochord and forms the 
 earliest rudiment of the centrum is the hypocentrum, Does this 
 give origin to all the bone which enters into the centrum ? If so, the- 
 opinion of Prof. Cope is correct. The statements of Lotz, Cartier, 
 and Goette that the primitive vertebral ring grows from the points 
 of contact with the notochord iipward and meet above the notochord, 
 are in favor of this idea. If, on the other hand, bony centers, how- 
 ever small, should be found at the bases of the upper arches, the ver- 
 tebra would not be wholly a hypocentrum. And this, I suspect, will 
 be found to be the case. If those observations are correct which go 
 to prove that the ring starts at four points, the fishes in which this 
 happens would have something more than hypocentra. If no pleu- 
 rocentra are developed, the centrum would be formed from the bases 
 of the upper and the lower arches, and be similar to those of the 
 middle tail region of Amia which are furnished with arches. It is 
 extremely probable that our modern bony fishes are derived from 
 many distinct sources among the ancient Ganoids, and it is possible 
 that among the latter were some which never developed bony inter- 
 vertebral elements. And even among the relatives of Amia there were 
 some, like Hypsocormus, which seem not to have the intervertebral 
 ossifications, possibly not the cartilages. We must therefore be pre- 
 pared to find that the composition of the vertebral centra of our fishes 
 is various. 
 
 The ribs of Amia deserve a portion of our attention. None are 
 present in larvae i5mm. long; all have been laid down in cartilage 
 in specimens 23mm. in length. Hence I cannot, as I hoped to do, 
 make observations on the method of their origin, whether in direct 
 connection with the transverse processes or independent of them. 
 
44 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 Do the ribs pass backward and form a part of the lower arches 
 -of the tail ? In case it should be demonstrated that ribs chondrify 
 independently of the transverse processes, there are no ribs in the 
 tail of Amia. Otherwise it may be a question of definition. While 
 there is a suture separating the bone of the lower arch from the 
 vertebral centrum in the tail, there is no segmentation of the carti- 
 lage. The lateral halves of the most anterior six or seven arches 
 unite below the great blood vessels to form the haemal canal, and 
 immediately below the point of union there is a joint followed by a 
 long spine. This spine might possibly be regarded as being a pair of 
 fused ribs. 
 
 The ribs and the transverse processes pass outward immediately 
 beneath the peritoneum, as do those of most, but not all, Teleostomi. 
 Anteriorly the transverse processes rise high up on the sides of the noto- 
 -chord ; but, as already stated, I have nowhere found them coalescing 
 with the bases of the upper arches. Scheele has found the bases of 
 the lower arches in the most anterior vertebras of Rhodeus to spring 
 from a common mass of cartilage with the bases of the upper arches. 
 Proceeding from this fact he concludes that the Cyprinidae are a more 
 primitive form than the Salmonidse, in which upper and lower arches 
 .arise independently of each other; and that the ribs of Teleosts are 
 homologous with those of Amphibians. In the latter conclusion he 
 wholly ignores the fact, first demonstrated by Aug. Miiller and later 
 emphasized by Rabl (50), Hatschek, and Baur (12), that the ribs of 
 these two groups of animals occupy very different positions with ref- 
 erence to the lateral muscles, and the further fact that some fishes 
 possess both kinds of ribs. As to the Cyprinidae, it has appeared to 
 me that they are in many respects much more highly differentiated 
 fishes than the Salmonidae. This is shown in the universal absence 
 of teeth on the bones of the mouth, the great modification of the 
 lower pharyngeal bones and the development of teeth on them, the 
 modification of the anterior vertebrae to the service of hearing, and 
 the enormous reduction of the cartilage found in the vertebral col- 
 umn of the lower fishes. Moreover, instead of the Cyprinidae being 
 more ancient than the Salmonidae, they appeared, so far as our knowl- 
 edge goes, at a considerably later period, the beginning of the Mio- 
 cene ; the Salmonidae near the beginning of the Cretaceous. 
 
 Hence, we have every reason for concluding that the condition 
 of the upper and the lower arches and the cartilages connected with 
 them is in a more primitive stage in the Salmonidae than in the Cyp- 
 Tinidae. We have now the testimony of Amia to the same effect. Upper 
 .and lower arches arise almost universally independently of each 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 45. 
 
 other, and their union is a secondary matter. In Rhodeus the coales- 
 cence of the two sets of arches has probably occurred as a result of 
 their being crowded close to each other, the result of the lifting- up of 
 the lower arches to give room for the contents of the abdomen. 
 
 8. FURTHER CONSIDERATION OF THE VERTEBRAL COLUMN OF LIVING 
 AND EXTINCT AMPHIBIA. 
 
 It may now prove profitable to examine a little more closely the 
 condition of the vertebral column of the Amphibia, living and extinct,, 
 in the light furnished by the embryology of Anita. 
 
 In the tail of Archegosaurus, and also in the two posterior presa- 
 cral vertebrae of Chelydosaurus, there occur, according to H.von Meyer 
 (47, p. 104) and Fritsch (Fauna der Gaskohle), basal pieces which 
 Fritsch has called hypocentra pleuralia. These have already been 
 referred to, and the name haemacentra proposed for them. We find 
 then in the Stegocephali the same four vertebral elements that are 
 found in Amia during its larval condition. Furthermore, Fritsch tells 
 iis that in the tail of old specimens of Arctygosaurus there are indica- 
 tions of embolomerous structure of the vertebrae. This might very 
 naturally be brought about by the coossification of the bases of the 
 upper and lower arches to form one centrum, and the union of the 
 pleurocentra with the haemacentra to form the other. The identity 
 of such a condition with that found in the middle tail region of Amia 
 would be complete. But should such double vertebrae not occur in 
 Archegosaurus, they are found in the genera , Cricotus and Diploverte- 
 bron. Prof. Cope tells us that such vertebrae are found throughout 
 the vertebral column of Cricotus. It seems to me that this arrange- 
 ment must have arisen here as it does in Amia, and that the ancestors 
 of the amphibians possessed both upper and lower intercalated car- 
 tilages throughout the length of the vertebral axis. 
 
 On the composition of the vertebrae of the Lepospondyli, with 
 their tubular vertebrae, our studies so far throw but little light. Pos- 
 sibly we may understand them better after an examination of the 
 vertebral structures of the living amphibia. In the Branchisauridae 
 the thin tubular centra consist, as already stated, each of two lateral 
 halves, separated dorsally and ventrally by sutures. We have here 
 evidence to the effect that the vertebral rings are primitively not 
 simple. 
 
 It is a matter of the highest interest to explain, if possible, the 
 morphogeny of the vertebrae of the higher fishes, the amphibians- 
 and the Amniota. Probably in the case of none of these groups shall 
 we be able to do this satisfactorily. In the attempt to elucidate the 
 
.46 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 ^composition of the vertebrae of our living amphibians, I am aware, 
 that I tread on uncertain ground, and that my conclusions may be 
 erroneous. On the other hand, some of the suggestions may have 
 some value. 
 
 The origin and structure of the vertebrae of the amphibia have 
 been quite as much discussed as have the origin and structure of 
 those of the fishes; and there has been the same lack of unanimity 
 of opinion. Johannes Muller concluded (49 Neur., p. 73) that in the 
 majority of the members of the class the vertebral centrum arises as 
 .an ossified integral ring in the outer sheath of the notochord. In the 
 case of certain Anura (Pelobates, Pseudis, etc.), however, the centra 
 were found to arise wholly from the upper arches, and the notochord 
 took no part in the process. 
 
 Among the Urodela Gegenbaur (29) recognized the existence 
 'Of two sheaths surrounding the notochord, an inner thicker, and an 
 outer thinner. XDutside of the latter is the skeletogenous layer, in 
 which, at intervals, are developed the cartilaginous upper arches. 
 The bases of the latter rest on the outer sheath of the notochord. 
 In the tail, lower arches possess similar relations. Midway between 
 the bases of the successive arches there is formed a ring of cells, which 
 later change to cartilage. There is no continuous cartilage surround- 
 ing the notochord. Ossification begins by the formation of a delicate 
 ring of bone immediately outside the outer sheath, and in contact 
 with the bases of the arches. Gegenbaur seems to imply that the ring 
 is complete from the beginning. Its edges extend forward and back- 
 ward until they approach those of the neighboring centra. While 
 this is progressing, the cartilaginous intervertebral rings have devel- 
 oped and have become enclosed each within the approximating ends 
 of two centra. In some cases the cartilaginous ring remains short 
 anterio-posteriorly ; in other cases it grows forward and backward 
 so as to extend from near the center of one vertebra to that of the 
 next behind. The arches at an early stage undergo ossification and 
 become continuous with the corresponding centra. We have then a 
 series of biconcave vertebras alternating with a series of cartilaginous 
 rings, which are enclosed in the concave ends of the vertebral centra. 
 Finally, each cartilaginous ring may become more or less segmented 
 transversely, one portion going to each of the two adjoining centra. 
 In the case of the frogs, Gegenbaur regarded the notochord as being 
 .surrounded by a continuous tube of cartilage from which spring the 
 upper arches and with which the intervertebral rings are also con- 
 nected. Ossification is introduced by the deposit of calcific matter in 
 the cartilage at the base of the upper arches. There is a center on 
 
OCT. r&95. VERTEBRAL COLUMN OF AMIA HAY 47 
 
 each side of the notochord, and from these the bone extends around 
 it, so as finally to form a ring. From the ring the bone spreads up- 
 ward on the upper arches. 
 
 Goette (32) found in Bombinator a continuous cartilaginous band 
 along the upper side of the notochord, distinct from the bases of the 
 arches. Ossification begins in this band by the appearance, sometimes 
 at least, of lateral centers beneath the bases of the arches. These 
 at length form bony rings. Between them the cartilages develop 
 into intervertebral cartilages, which become cross-segmented and 
 form the epiphyses of the vertebral centra. 
 
 Hasse (36) regards the vertebral centra of the Anura as being 
 composed principally of the bases of the upper arches. In the case 
 of the Urodela, Hasse (35) holds that the intervertebral cartilages lie 
 ^between the two sheaths of the notochord; while the ring of bone 
 "which forms the earliest rudiment of the centrum arises in the elas- 
 tic a, first of all on the lower side. Spreading fore and aft these ver- 
 tebral centra come to enclose the intervertebral cartilages in their 
 concave ends. Later, the cartilages divide, and a portion of each is 
 distributed to each of the two enclosing vertebrae. 
 
 On the side of the palaeontologists, von Meyer (47, p. 95, seq.) y 
 accepting the view that the vertebral centrum originates as a ring 
 independently of the arches, concludes that in such forms as Archego- 
 saurus the vertebral column is in an embryonic condition, and that 
 there are no centra present. 
 
 As already stated, Prof. Cope regards the vertebrae of the mod- 
 ern amphibians, like that of the Teleostomi, as being derived wholly 
 from the hypocentra. 
 
 On page 123 of this paper I have called attention to the inter- 
 vertebral rings of cartilage which, in a certain stage of development, 
 surround the notochord of Lepisosteus. and I have there suggested 
 that these rings occupy exactly the positions occupied by the interca- 
 lated cartilages of related fishes. The fusion of these cartilages 
 with one another to form a ring, and their further fusion with the 
 bases of the neighboring arches, is a process which might take place 
 secondarily. If this supposition is allowable in the case of Lepisosteus, 
 less objection can be urged in the case of the Urodeles; since 
 here the cartilages in question at no time coalesce with the bases of 
 the arches. We might easily explain the case by supposing that the 
 four intercalated cartilages belonging to each segment have fused 
 and formed a ring ; that this has remained, at least for a long time, 
 in a state of cartilage ; and that the vertebral bands of bone in front 
 ..and behind it have expanded over it and enclosed it between them. 
 
48 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 One objection against this hypothesis may be urged in the case 
 of Lepisosteus and all the amphibians. In all of these the interver- 
 tebral ring becomes cross-segmented, and the parts are distributed 
 to different vertebrae, the anterior portion becoming the hinder end 
 of the vertebra in front, the hinder portion the front end of the ver- 
 tebra behind it. In answer to this objection it may be said that orig- 
 inally the intervertebral ring, whatever its origin, is an integral mass 
 of cartilage ; that is, after the fusion of the circle of constituent 
 pieces ; and that the transverse segmentation is also a secondary 
 matter. It certainly would not be difficult to find examples of this 
 secondary segmentation of cartilage in many animals. 
 
 A further objection against homologizing the invertebral carti- 
 lages of the Urodeles with the intercalated cartilages of Amia, may 
 be found in the apparent fact that the intervertebral ring of the Uro- 
 deles is located between the inner and the outer sheaths of the noto- 
 chord. Hasse's observations (35) appear to prove that such is the 
 fact ; nevertheless, it is desirable that they shall be confirmed. Geg- 
 enbaur did not reach the same conclusion. Hasse, in the paper just 
 referred to, and elsewhere, regards the intercuticular position of the 
 cartilages as indicating a relationship between the Elasmobranchs 
 and the Dipnoi, on the one hand, and the Urodeles on the other. 
 Later (39, p. 91) he concludes that, as regards the Dipnoi, he is prob- 
 ably wrong, and that the just named group does not stand in close 
 relation with the Urodeles. The so-called elastica externa of Dipnoi 
 and the cartilaginous Ganoids may, after all, be the homologue of the 
 inner sheath of Elasmobranchs. 
 
 In view of these difficulties and doubts it appears to be easier to 
 believe that the intercalated cartilages have, through pressure or some 
 other influence, caused the elastica to be dissolved, and have thus 
 come to lie in contact with the inner sheath. It is to be noted that 
 the groups of cells which, in the salamanders, give origin to the inter- 
 vertebral rings, are placed exactly where we should expect the inter- 
 calated cartilages to appear. In T-jito:* tceniatus, investigated by 
 Hasse, these appear in the intervals between the successive arches 
 above ; while on the lower side of the notochord, they form one cen- 
 ter in the middle line, as though from a coalescence of the two lateral 
 masses. 
 
 Gegenbaur attached much importance to the existence of the 
 intervertebral cartilages in the vertebral column of the amphibians. 
 He says, "Allen aber ist mit Auftreten des Intervertebralknorpels 
 ein eigenthiimliches Unterscheidungsmoment gegen die Fische 
 hingeworden. und in dieser neuen Bildung is zugleich eine fur die 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 49 
 
 ganze kiinftige Gestaltung des Wirbelkorpers maasgebende Einrich- 
 tung zu Stande gekommen." It is not impossible, however, that we 
 have here, after all, no new structures, but, as in so many other 
 cases, only new adaptations and new combinations of structures 
 already well known in lower forms, 
 
 To what extent the bases of the lower arches, the hypocentra, 
 may be represented in the trunk region of the Urodeles is questiona- 
 ble. Possibly with the disappearance of the lower series of ribs the 
 bases of the arches also have totally disappeared. This seems to be 
 the condition of the Anura, certainly in the case of those with epi- 
 chordal vertebrae. But evidences of hypocentral ossifications should 
 be looked for in the Urodeles in the trunk, and such ossifications are 
 certainly present in the tail. 
 
 The opinion of Prof. Cope that the vertebral centra of the Am- 
 phibians are represented by only the hypocentra was doubtless the 
 outcome of the idea that only pleurocentra and hypocentra enter into 
 the composition of centra. That the bases of the upper arches and 
 haemacentral bones may also participate in the composition, there can 
 now be no doubt. 
 
 In the " Biologisches Centrallblatt," Band vi., 1888, Dr. G. Baur 
 has discussed the morphogeny of the vertebral column and given the 
 history of the question and the literature bearing on the subject. The 
 view proposed first by Cope, and accepted by Albrecht, Dollo and 
 Baur, that the vertebral centrum of Amniota has been derived from 
 the pleurocentrum is, I believe, correct. The evidences in favor of 
 this hypothesis have been presented by Baur in the paper just cited. 
 A consideration of the situations in which the various elements con- 
 nected with and forming the vertebra confirms the hypothesis. 
 
 The arches, upper and lower, of the Amniota are unquestion- 
 ably homologous with those of the lower vertebrates, and, as in the 
 latter, are developed in the intersection of two sets of membranous 
 septa, viz. : those between the myomeres and those between the two 
 sides of the body, including in the latter the subperitoneal membrane. 
 The hypocentra, therefore, are cut by the transverse septa. The 
 intercalated cartilages, or pleurocentra, on the other hand, fall be- 
 tween the septa and in the myorneres. When the hypocentrum enters 
 into the composition of the vertebral centrum, the lower arches will 
 naturally be found intimately connected with the centrum. And such 
 is its position in all the living tailed amphibians (Stannius, Am- 
 phibia, page n) being attached to the middle or hinder half of the 
 centrum. In the fossil Hylonomus fritschi, Credner, however, the 
 lower arches are represented as being attached between two contig. 
 
50 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 uous vertebrae. This condition I am not at present able to explain. 
 
 If now the hypocentrum should vanish and its place be taken by 
 the enlarged pleurocentrum, as is believed to be the case among 
 the Amniota, the bases of the lower arches, if present, would 
 primarily fall between the vertebral centra. And it is between 
 the vertebrae, or close to their articulation, that we find the attach- 
 ments of the lower arches, or chevron-bones, of most of the higher 
 vertebrates. The departures from the rule may be easily explained 
 as secondary modifications. 
 
 The upper arches, like the lower, are developed in the transverse 
 septa. This being the case, the connection of the arches with the 
 pleurocentra must be secondary. Hasse, in his " Beitrage zur Allge- 
 meinen Stammesgeschichte der Wirbelthiere," appears to have op- 
 posed this view, regarding, as other writers have done, the arches as 
 developing in direct union with the centra. On the other hand, Fro- 
 riep (27) holds that the body of the vertebra has an origin independ- 
 ent of the arches. The arch is the fundamental and earliest struc- 
 ture, the centrum a secondary one. He rejects the hypothesis that 
 the vertebral body of the higher vertebrates proceeds from the union 
 of the basal portions of the arches. According to Froriep, the prim- 
 itive membranous arches grow around the notochord and unite below 
 the latter to form his " hypochordale Spange." The latter is evi- 
 dently the representative of the hypocentrum. The upper arch at 
 length chondrifies and unites with the centrum next behind. The 
 latter is the pleurocentrum. The shifting of the arch backward to 
 become attached to the pleurocentrum reminds us of the movement 
 of the arches of Amia to join partly the pleurocentrum next behind. 
 Hoffman (40) tells us that in the tortoises the ribs and the upper 
 arches, during the early stages of development, rest intervertebrally 
 on the cartilaginous tube which surrounds the notochord. According 
 to Gegenbaur, there is, in Lacertilia, a stage in which the upper arch 
 is placed intervertebrally. (29, p. 44). The ribs of the amphibia 
 being developed in the transverse septa, must have their connection 
 primitively with the hypocentra, rather than with the pleurocentra. 
 It has been shown by Cope (14, p. 518) that in some of the Pelyco- 
 sauria (Theriodontia) the head of the rib is attached to the intercen- 
 trum ; and Dr. Baur has pointed out (6, 8) that this is true also in the 
 case of the cervical vertebrae of the very primitive reptile Spheno- 
 don, of those of the Crocodilia, and of those of some of the Dino- 
 sauria. When now the hypocentrum becomes rudimentary or disap- 
 pears, the head of the rib will, at first at least, be attached between 
 the pleurocentra, now become the centra. Such is its attachment in 
 
OCT. 1895. VERTEBRAL COLUMN OF AMIA HAY 51 
 
 a large number of the Amniota. As a result of subsequent modifi- 
 cations, the rib-head may, of course, form an attachment with the 
 centrum, usually with the one next behind ; occasionally to the hinder 
 end of the one in front. 
 
 In case the rib develops a second point of attachment to the ver- 
 tebral elements, this will naturally be with the upper arch, since rib 
 and arch lie in the same transverse septum ; the tubercular portion of 
 the rib will also be carried backward, and this may have some in- 
 fluence in causing also the head of the rib to take a more posterior 
 articulation. As a matter of fact, the tubercular process of the ver- 
 tebra usually springs from the upper arch. The Ichthyosauria fur- 
 nish us with a case in which the tubercular attachment has descended 
 low down on the centrum. In the dorsal region of the Crocodilia the 
 rib-heads also desert the centra and adhere to the transverse pro- 
 cesses. 
 
 We have seen that in Amia the pleurocentrum, both in the dor- 
 sal and in the middle portion of the tail, goes to form the anterior 
 portion of the vertebra to which it belongs. On the contrary, Fro- 
 riep holds that the upper arch, rib, and "hypochordale Spange" belong 
 primarily to the centrum immediately behind them, thus making the 
 hypocentrum anterior. The true explanation may be that the hypo- 
 centrum and the elements connected with it belong no more to the 
 pleurocentrum behind it than to the one in front. It is then free to 
 form connection with either, and this connection is different in the 
 different groups of vertebrates. 
 
 In concluding this paper I wish to acknowledge my indebtedness 
 to Dr. George Baur for the opportunity to conduct my investigations 
 in his laboratory in the University of Chicago, and for many val- 
 uable suggestions offered during the progress of my work. 
 
52 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 < 
 
 BIBLIOGRAPHY. 
 
 1. Allis, Edw. P., 1888. The Anatomy and Development of the Lateral Line 
 System in Amia calva. Journal of Morphology, Vol. II. pp. 463 566. Pis., 
 XXX-XLII. 
 
 2. Baer, Ernest v. 1835. Untersuchungen iiber die Entwicklungsgeschichte 
 der Fische. 
 
 3. Balfour, P.M. 1881. A Treatise on Comparative Embryology. Vol. II. 
 
 4. Balfour, F M and Parker, W. K. 1883. The Structure and Development of 
 Lepidosteus. Transactions of the Philosophical Society; also in Memorial edition 
 of works of F. M. Balfour. 1885, Vol. I, pp 738- 847. Pis. 3442. 
 
 5. Baur, G. 1886. Ueber die Morphogenie der Wirbelsaule der Amnioten. 
 Biologisches Centralblatt, Band VI. S. 332342. 353303. 
 
 6. 1886. The Ribs of Sphenodon (Hatteria). American Naturalist, Vol. XX., 
 
 pp. 979981. 
 
 7. 1886. Proatlas, Atlas, and Axis of the Crocodilia. Amer. Nat., Vol. XX., 
 
 pp. 288293. 
 
 8. Osteologische Notizen iiber Reptilien. Zoologischer Anzeiger, Bd. IX., S. 
 
 733743. 
 
 10. 1887. On the Morphology of Ribs. Amer. Nat., Vol. XXI., pp. 942-945. 
 
 11. 1887. Arrangement of the Sauropsida. Journal of Morphology, Vol. I., pp. 
 
 93104. 
 
 12. 1893. Ueber Rippen und ahnliche Gebilde und deren Nomenclatur. Anat- 
 
 omischer Anzeiger, Jahrgang IX,, S, 116 120. 
 
 13. Cartier, O. 1875. Beitrage zur Entwicklungsgeschichte der Wirbelsaule. 
 Zeitschrift fur wissenschaftliche Zoologie, Bd. XXV.. Sup., S. 65-70. 
 
 14 Cope, Edward D. 1878. Descriptions of Extinct Batrachia and Reptilia 
 from the Permian Formation of Texas. Palaeontological Bulletin No. 29, pp. 505- 
 530 (Extracted from the Proceedings of the Amer. Phil. Soc., 1878.) 
 
 15. 1880. Second Contribution to the History of the Vertebrata of the Permian 
 
 Formation of Texas. Palaeontological Bulletin No. 32, pp. 1 22. (From Proc. Amer. 
 Phil. Soc., 1880.) 
 
 16. 1882. The Rhachitomous Stegocephali. Amer. Nat., Vol. XVI., pp. 
 
 334335. 
 
 17. 1883. The Vertebrata of the Tertiary Formations of the West. Washing- 
 ton, D. C. 
 
 18. 1884. The Batrachia of the Permian Period of North America. American 
 
 Nat, Vol. XVII , pp. 2639. 
 
 19, 1886. The Batrachian Intercentrum. Amer. Nat., Vol. XX., pp. 76 ; 175. 
 
 20. 1886. On the Intercentrum of the Terrestrial Vertebrata. Trans, Amer. 
 
 Phil Soo., Vol. XVI., pp. 243253. 
 
 21. 1887. The Homology of the Chevron Bones. Amer. Nat., Vol. XII., 
 
 p. 319. 
 
 22. Zittel's Manual of Palaeontology. Amer. Nat., Vol. XXI,, pp. 10141019. 
 
 23. Credner, Hermann. 18811886. Die Stegocephalen und Saurier aus dem 
 Rothliegenden des Plauen'schen Grundes bei Dresden. Zeitschrift der deutschen 
 geologischen Gesellschaft, Bd. XXXIII., 1881; XLV., 1893; Naturwissenschaftliche 
 Wochenschrift, Bd. V., Berlin, 1890. 
 
 24. Deecke, W. Ueber Fische aus verschiedenen Horizonten der Trias. Palaeon- 
 tographica, Bd. XXXV., pp. 97138. Taf. VI, VII. 
 
VERTEBRAL COLUMN OF AIMIA HAY. 53 
 
 25. Dumeril, Aug. 1870. Histoire naturelle des Poissons ou Ichthyologie 
 Generalle. Tome II. 
 
 26. Franque, Henricus, 1847. Nonnulla ad Amiam calvam accuratius cognos- 
 cendam. With one plate. 
 
 27. Froriep, Aug. 1886. Zur Entwicklungsgeschichte der Wirbelsaule. II. 
 
 /Beobachtung an Saugethieren. 
 28. Gegenbaur, C. 1867. Ueber die Entwicklung der Wirbelsaule des Lepid- 
 osteus mit vergleichenden anatomischen Bemerkungen. Jenaische Zeitschrift, Bd. 
 III. 
 
 29. 1862. Untersuchungen zur vergleichenden Anatomie der Wirbelsaule bei 
 
 Amphibien und Reptilien. Vier Tafeln. 
 
 30. 1876. Einige Bemerkungen zu Goette's "Entwicklungsgeschichte der 
 
 Unke." Morphologisches Jahrbuch, I., 288 345. 
 
 31. Gill, Theodore. 1871. Arrangement of the Families of Fishes. Washing- 
 ton, D. C. 
 
 32, Goette, Alex. 1875. Die Entwicklungsgeschichte der Unke. With Atlas 
 of 22 plates, 
 
 33. 1879. Beitrage zur Morphologic des Skeletsystems der Wirbelthiere. 
 
 Archiv fur mikroskopische Anatomie, Bd. XVI, 117152; Taf. VII IX. 
 
 34. Grassi, B. 1883. Developpement de la Colonne vertebrale chez les Pois- 
 sons osseux. Archives Itallienne de Biologic. Tome IV., pp. 236244; 251 268. 
 J 35. Hasse, C. 1890. Die Entwicklung der Wirbelsaule von Triton t<enialus. 
 Zeitschrift fur wissenschaftliche Zoologie, Bd. LIII., Sup. S., 1-20. Taf. I. -III. 
 
 36. 1892. Die Entwicklung der Wirbelsaule der ungeschwanzten Amphibien. 
 
 Zweiter Abhandl. Mit Tafel XII. Zeitschr. f. wiss. Zoologie. Bd. LV., S. 252-264. 
 
 37. 1892. Die Entwicklung der Wirbelsaule der Elasmobranchier. Dritter 
 
 Abhandl. Mit Tafel XXL. Zeitschr. f. wiss. Zoologie, Bd, .LV., S. 519-531. 
 
 38. 1892. Die Entwicklung der Wirbelsaule der Dipnoi. Vierte Abhandl. Mit 
 
 Tafel XXII. Zeitschr. f. wiss. Zoologie, Bd. LV., S. 533-542. 
 
 39. 1893. Die Entwicklung und der Bau der Wirbelsaule der Ganoiden. 
 
 Fiinfte Abhandl. Mit Tafeln V VII. Zeitschr. f. wiss. Zoologie, Bd. LVII. S. 
 76 !>(i. 
 
 40. Hoffman, C. K. 1878. Beitrage zur vergleichenden Anatomie der Wirbel- 
 theire. VIII. Ueber die Chorda dorsalis bei den Schildkroten. Niederlandisches Archiv 
 fur Zoologie, Bd. IV, pp. 185-199; Taf. XII. 
 
 41. Klaatsch, H. 1893. Beitrage zur vergleichenden Anatomie der Wirbel- 
 saule. I. Ueber den Urzustand der Fischwirbelsaule, (Mit Taf. XXVI. und eine Fig- 
 urimText.) Morphologisches Jahrbuch, Bd. XIX., pp. 649680. 
 
 42. 1893. Beitrage zur vergleichenden Anatomie der Wirbelsaule. II. Ueber 
 
 die Bildung knorpeliger Wirbelkorper bei Fischen. (Mit Taf. VII, und 6 Figuren im 
 Text.) Morphologisches Jahrbuch, Bd. XX., S. 143186. 
 
 43. Kolliker, Albert. 1860. Ueber das Ende der Wirbelsaule der Ganoiden 
 und einige Teliostier. Mit 4 Tafeln. 
 
 44. Leidy, Joseph. 1873. Contributions to the Extinct Vertebrate Fauna of 
 the Western Territories. Report U. S. Geological Survey of the Territories, Vol. I., 
 Pt. I., pp. 358, Pis. I-XXXVII. Washington, D. C. 
 
 45. Lotz, Theophil. 1864. Ueber den Bau der Schwanzwirbelsaule der Sal- 
 moniden, etc. Zeitschrift fur wissenschaftliche Zoologie, Bd. XIV., pp. 81 106; Taf. 
 X.-XIII. 
 
54 FIELD COLUMBIAN MUSEUM ZOOLOGY, VOL. i. 
 
 46. Liitken, Chr. von. 1873. Ueber die Begrenzung und Eintheilung der 
 Ganoiden. Palaeontographica, Bd. XXII., S, 1-54. 
 
 47. Meyer, Hermann von. 1857. Reptilien ausder Steinkohlen-Formation in 
 Deutschland. Zweite Lieferung. Palaeontographica, Bd. VI. S. 59-218; Taf. VIII 
 XXIII. 
 
 48. Miiller, August. 1853. Beobachtungen zur vergleichenden Anatomic der- 
 Wirbelsaule. Archiv fiir Anatomic und Physiologic. 
 
 49. Miiller, Johann. 1835 1845. Vergleichende Anatomie der Myxinoiden. 
 
 50. Rabl, Carl. 1893. Theorie des Mesoderms, II. (Mil Taf. IV VII. ; Fig. 
 10-13 imText.) Morphologisches Jahrbuch, Bd, XIX., S. G5 144. 
 
 51. Rathke, Heinrich. 1839. Entwicklungsgeschichte der Natter. 
 
 52. 1866. Untersuchungen iiber dieEntwicklung und den Korperbau der Krok- 
 odilen. 
 
 53. Reis, Otto M. 1888. Die Coelacanthinen, mit besonderer Beriicksichtgung 
 der im weissen Jura Bayern vorkommenden Arten. Palaeontographica, Bd, XXV., 
 S. 196., Taf. I V. 
 
 54. Sagemehl, M. 1883. Beitrage zur vergleichenden Anatomie der Fische. 
 I. Das Cranium von Amia calva. Morphologisches Jahrbuch, Bd. IX., S. 177. 
 
 55. Scheele, C. 1893. Beitrage zur Entwicklungsgeschichte der Teliostier- 
 Wirbelsaule. Morphologisches Jahrbuch, Bd. XX., S. 147, Taf I. III. 
 
 56. Schmidt, Ludwig, 1892. Untersuchungen zur Kenntniss des Wirbelbaues 
 von Amia calva. Zeitschrift fiir wissenschaftliche Zoologie, Bd. LIV., S 748 764' 
 Taf. XXXIV. 
 
 57. Shufeldt, R. W. 1885. The Osteology of Amia calva, etc. From Report 
 of U. S. Commission of Fish aud Fisheries for 1883. PI. I. XIV. Contains transla 
 tion of Sagemehl's paper. 
 
 58. Stannius, Hermann. 1854. Handbuch der Anatomie der Wirbelthiere. 
 
 59. Wilder, B. G. 1877. On the Tail of Amia. Proceedings of the Ameri- 
 can Association for the Advancement of Science for 1876. Pp. 264 266. 
 
 60. Zittel, Carl von. 18871890. Handbuch der Palaeontologie. Bd. Ill; 
 Pisces, Amphibia, Reptilia, und Aves. 
 
PLATES. 
 
 EXPLANATION OF FIGURES, 
 ao. Aorta, 
 as. Aortal support, 
 cu. ch. Inner notochordal sheath. 
 e. ex. External elastica. 
 h. a. Haemal arch. 
 
 i. c. Cartilage intercalated between halves of upper arch above the spinal cord. 
 1. 1. s. Superior longitudinal ligament, 
 n. Nerve. 
 
 n. a. Half of neural arch, 
 not. Notochord. 
 
 n, s. Distal extremity of lateral half of neural arch, 
 par. Parapophysis. 
 
 p. c Cartilage supporting lateral half of neural arch (pleurocentrum.) 
 r. Rib. 
 
 s. n. Spinal nerve, 
 sp. c. Spinal cord, 
 sp. s. Spinal ganglion, 
 v. Vein. 
 In the colored figures cartilage is represented by blue, bone by red. 
 
.V! 
 
 \ lo sbiz i\al adt io noilsBlqoq^ri bns noitefilqo^rf adJ zin9?.3iq?>i atfilq eidT 
 -o)odq moiT .snod Isdoun adJ }o n-?vi oats ai wsiv A v/ofad moi^ nf? 
 
PLATE IV. 
 
 This plate represents the hyoplastron and hypoplastron of the left side of Proto- 
 stega gigas seen from below A visw is also given of the nuchal bone. From photo- 
 graph. 
 
FIELD COLUMBIAN MUSEUM. 
 
 ZOOLOGY, PL. IV 
 
 PROTOSTEGA GIGAS. 
 

 
 .V 
 
 
 sno'uisnq 
 }o bns 
 
 aril wodf? ol 
 Blqtqa adJ lo 
 
 ad} 
 
 io noitsBlq aril lo noilio)a9i Ir.h'icq A 
 .allanKJnol adl ioasis aril bn xanod aril Jo 
 la SIB Bilafllqidqix adt 
 
PLATE V. 
 
 A partial restoration of the plastron of Protostega, to show the relative positions 
 of the bones and the size of the fontanelle. Only the bases of the epiplastra and of 
 the xiphiplastra are shown. 
 
FIELD COLUMBIAN MUSEUM. 
 
 ZOOLOGY, PL. V. 
 
 PROTOSTEGA GIGAS. 
 
-' 
 
 H 
 
W ''. A >^V 
 
 
 r