HARVARD UNIVERSITY ^ Library of the Museum of Comparative Zoology BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, TN CAMBRIDGE VOL. 114 CAMBRIDGE, MASS., U. S. A. 1955 - 1956 The Cosmos Press, Inc. Cambridge, Mass., U. S. A. /'^ CONTENTS PAGE No. 1. — The Size of Parietal Foramen and Organ in Reptiles. A Rectification. By Tilly Edinger. Oc- tober, 1955 . 1 No. 2. — A Classification of Therapsid Reptiles. By D. M. S. Watson and A. S. Romer. February, 1956 . 35 No. 3. — The Placenta of the Wolverine {Gulo Gulo Luscus (Linnaeus)). By George B. Wislocki and Emmanuel C. Amoroso. (8 plates.) February, 1956 89 No. 4. — The Functional Anatomy of the Delphinid Nose. By Barbara Lawrence and AYilliam E. Sche- vill. (30 figures.) February, 1956 . . . .101 No. 5. — Silicified Middle Ordovician Trilobites : The Odontopleuridae. By H. B. Whittington. (24 plates.) May, 1956 153 No. 6. — A Cat Skeleton with an Anomalous Third Hind Leg and Abnormal Vertebrae. By Thomas S. Parsons and John M. Stein. (4 plates.) May, 1956 291 No. 7. — Post-Triassic Nautiloid Genera. By Bernhard Kummel. (28 plates.) July, 1956 .... 319 Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol-. 114, No. 1 THE SIZE OF PARIETAL FORAMEN AND ORGAN IN REPTILES. A RECTIFICATION By Tilly Edinger CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM October, 19.").") Publications Issued by or in Connection WITH THE MUSEUM or COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 114. Breviora (octavo) 1952 — No. 48 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 5.") JoHNSONiA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 3, no. 34 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin oi the Museum of Comparative Zoology AT H A R V A E D COLLEGE Vol. 114, No. 1 THE SIZE OF PARIETAL FORAMEN AND ORGAN IN REPTILES. A RECTIFICATION By Tilly E dinger CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM October, 195r> No. 1 — The Size of Parietal Foramen and Organ in Reptiles. A Rectification. By Tilly Edinger^ CONTENTS Page Introduction 3 Opinions 6 Foramen-eye relations in lizards 11 Literary data 11 Figured data 16 Foramen and skull .......... 19 The size question 19 An answer ........... 20 Application to fossils 24 Summary 29 Literature cited 30 INTRODUCTION The pineal apparatus consists of one or two organs evaginated from the roof of the diencephalon and their nerve fiber connec- tions with and inside the brain ( Fig, 1 ) . Study of such an organ complex would seem a subject altogether inaccessible to students of extinct vertebrates, of which only the hard parts are preserved. This is in great part true of the pineal organ proper (glandula pinealis, epiphysis cerebri), the pineal body which develops in almost all vertebrates. In the extant forms, at least, the epiphysis has no relation to osseous elements of the head, and it is only in some very ancient fishes that the skull bones clearly indicate the presence, above the brain, of two organs in the pineal complex (Edinger, in press). The parapineal organ fails to develop in the vast majority of living vertebrates. In most Lacertilia, how- ever, it is a stalked, dorsomedian, photoreceptive sense organ whose presence is reflected in the structure of the skull roof. Whenever, in a lizard, this parietal organ is well developed as a 1 This study Is one of a series aided by a grant from the Penrose Fund of the American Philosophical Society. BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY parietal eye, its influence causes and maintains an opening in the bony shield over the brain. This opening is the parietal foramen — always median but variable in position even within species (El- Toubi, 1945, p. 5), lying either between the two parietal bones, between the two frontal bones, or in the fronto-parietal suture (Fig. 4). A corresponding foramen is of frequent occurrence in Paleozoic Agnatha and fishes, and in Paleozoic and Mesozoic amphibians and reptiles. BICTAL OdCAN y PINEAL OWCAN |). EDINQER: PARIETAL FORAMEN 9 Even Spencer, the first to compare parietal eye preparations from many (28) lacertilian species, observed that in every speci- men there is at least some one feature in which the eye or the parietal skin spot is less perfect than in other specimens. He concluded "that the pineal eye in Lacertilia is a rudimentary structure — that at the present time it is not so highly developed as it must have been at some previous period when fully func- tional" (1886, p. 216). This idea found much support, authors citing very large parietal foramina in some reptiles of the Past. One must realize that, in contrast to the foramen, the structure of the parietal organ is revealed only in microscopic sections. Students of such preparations have not in general been concerned with the size of the organ. I found only the following data in the literature. Spencer was amazed to find the axis of the eye 0.4 mm. long both in a 2-foot long Sphenodo7i and in a 6-foot long Varanus (1886, p. 183). Longitudinal and transverse diameters varied from 0.18 to 0.20 mm. and from 0.06 to 0.07 mm. in the Anniella specimens of Coe and Kunkel (1906, p. 393) ; medio- dorsal skull length is about 8 mm., and length of the animals 105 to 152 mm. from snout to cloaca, plus a post-cloaeal portion varying from 16 to 75 mm. {ibid., p. 351). Slightly larger parietal eye diameters in Phrynosoma have been expressed in microns : 258pi and ni[K (Ritter, 1891, p. 212). In short, the diameters of measured lizard parietal eyes were fractions of millimeters. In the vast majority of lizards the foramen — correspondingly, it appears — ranges from pinpoint size to diameters of about 1 mm. Diameters "of considerable size", namely 2.2 and 2.6 mm., were noted by Schmidt (1909, p. 373) in two agamid in- dividuals. Also among the dozens of lizard skulls I studied, diameters of more than 1 mm. are rare. I found 4 Agamidae and 3 Iguanidae with diameters of 1.1 to 2.0 mm. ; and the largest, by far, in the MCZ Department of Herpetology is the parietal fora- men of the lizard giant, Varanus komodoensis, which is 3 mm. long and 2 mm. broad (the skull, 190 mm. long). On the other hand, among extinct Reptilia one can encounter a parietal foramen as much as 50 mm. in length and 20 mm. in breadth — in a Jurassic plesiosaur, Pliosaurus, whose skull, lacking tip of muz- zle, is 1100 mm. long (Linder, 1913, p. 394). Certainly one must conclude that an extinct reptile such as 10 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY this plesiosaur had a larger parietal eye than have the Laeertilia. On the other hand, there are no data of any kind to support the idea that in the Past parietal eyes were different, were more elaborate in structure, than they are in the Present. If there had existed, for example, auxiliary structures such as muscles and extrinsic nerves to move the parietal eye, their traces should be found in the walls of parietal foramina of at least the type shown Fig. 3. Tapinocephaloid. Meaiau section through back part of skull. F.m., foramen magnum, P.p., parietal foramen. (After Haughton 1918; xl/6). in Figure 3. Further, the ontogenesis of the lacertilian pineal complex has been thoroughly investigated, over the past 70 years, by a great number of students, in innumerable embryos of many species. Although the process is by no means always the same in every detail, never has anything been observed which could be interpreted as indicating an originally more complicated structure of the parietal eye. In the concept that in fossil reptiles with a large parietal foramen the organ had great photoreeeptive power, it need not be assumed that the structure was different from that in lizards. Size obviously has no necessary relation to function in the lateral eyes of vertebrates ; the big eyes of a cow see less well, the tiny eyes of an eagle see far better than human eyes. The parietal photoreceptor, however, in contrast to the highly differentiated lateral eye, has no provisions to control the access of light rays. The amount of light received by its retina must depend only on the size of the lens, and thus of the vesicle. The tacit assumption EDINGER : PARIETAL FORAMEN 11 is reasonable that the larger a parietal organ is, the more light stimuli it will absorb and transmit. Although the volume of the parietal vesicle must in some lizards be several times that in others, one cannot well imagine that among the Recent Lacertilia photosensitive capacity is very- much greater in larger than in smaller parietal eyes. It can hardly matter much whether in these minor aids to the major photoreceptors (later eyes and skin) the lens diameter is a greater or lesser fraction or, possibly, all of 1 mm. Far greater lenses, however, must be assumed to kave existed in several extinct branches of the Reptilia. It is with regard to these that the questions of parietal eye size become interesting. In consequence, great importance must be attributed to the widespread opinion according to which a large parietal foramen does not necessarily reflect the presence of a large parietal organ. It was formulated as follows by Camp (1923, p. 395). "The size of the foramen has no relation to the size of the eye in the latter group [Lacertidae]. Those who would relate fiLnctional signifi- cance of the eye with size of the foramen in fossil forms should note this." This challenge to paleontology has never been met. It has not even been discussed. Yet it can be successfully combatted — with, of course, the material of neontology. Camp's studies of actual specimens, like my own, were not of the organ but of skulls. His conclusion as to the absence of size relations between organ and foramen was drawn from zoological papers that he quotes. During my more extensive perusal of the literature I came to recognize this opinion as one of the many generalizations about the pineal apparatus that have acquired seeming overall validity merely by being repeated for decades. FORAMEN-EYE RELATIONS IN LIZARDS Literary data. The papers dealing specially with parietal eyes contain in their texts very little information about size relations between eye and foramen. With the exception mainly of Stud- nicka (1905) and Schmidt (1909), those who described the organ did not mention the foramen — although bony foramen margins occasionally appear in one or both upper corners of their figured microscopic preparations. 12 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The consequences are evident whenever a general statement has seemed necessary, such as in the big monographs mentioned in the introduction. In one of them the authors agree with Camp and others : ' ' The size of the foramen differs and bears no direct relation to the size of the parietal eye" (Tilney and Warren, 1919, p. 131). Another survey, in a chapter entitled "The Parietal Foramen in Reptiles", says of the parietal eye that "as a rule ... its diameter is about one-third the diameter of the foramen" (Gladstone and Wakeley, 1940, p. 270). No references are given. That 1 :3 proportion has indeed been re- ported— namely, in one Phrynosoma individual. "While not stressed by the describer of the specimen (Ritter, 1891), the condition was mentioned by Studnicka in his monograph — but mentioned as a case of quite exceptional discrepancy between organ and foramen (1905, p. 160). In Phrynosoma, Studnicka 's own specimen also showed "an extraordinarily broad parietal foramen. It is a big gap in the bony skull roof, about four times as broad as the parietal eye" (1905, p. 181). A foramen with an abnormally large transverse diameter seems quite likely to occur when the opening lies in the crossing of the frontoparietal and mediosagittal sutures in a thin skull, particularly in a young lizard. It can then even have the shape of a T. In such a case, the transverse bar of the T is an unclosed middle section of the frontoparietal suture, the longi- tudinal bar a gap either between the two frontal bones (Engel- brecht, 1951, fig. 10) or between the two parietal bones: a parietal foramen not fully framed. In none of the fossil reptiles I know is there a particularly broad parietal foramen ; it is either a longitudinal oval or a circle, very rarely a little broader than long. Should a T-shaped opening be found in a fossil, it would certainly be recognized as a fontanelle rather than a foramen. There could be no misunderstanding the fact (Engelbrecht, 1951, p. 21) that only the middle section is related to the parietal organ. The concept of independent size variation of the parietal organ and its foramen is, admittedly, a generalization, based on data reported in special studies of the organ. The basis is, I have found, psychological rather than factual. When Spencer (1886, p. 21.")) wrote that in his speeiiiiens of Laccrta ocellata and Cyclodus "the eye is placed within the EDINGER: PARIETAL FORAMEN 13 parietal foramen, fitting it closel}'", this was an incidental ob- servation in a discussion of positions of the parietal eye. Schmidt was the only author who, with a large amount of material, paid attention to and discussed size proportions of eye and foramen. Understandably, he noted and reported the conspicuous and the exceptional conditions — as did Studnicka in his survey. Only the following data are mentioned in Schmidt's text (1909). Much as in a previously figured Anolis, in which "the parietal eye occupies the entire foramen space in thick bone" (p. 374; our Fig. 2A), the foramen was only slightly wider than the eye in one of Schmidt's Anolis cristellatiis specimens, but "approxi- mately double the size of the eye" in another specimen (p. 381). In one of Schmidt's other Iguanidae, a specimen of Hoplurus, the diameter of the foramen was approximately 6 times that of the eye (ibid.). In some but not all specimens of Agama the diameter of the foramen was "a multiple" of the eye diameter. It was because the foramen was almost filled by the eye in the Lacerta ocellata specimens of other authors that Schmidt men- tioned one case of a foramen "more than five times as large as the parietal eye" (p. 392). And it was in connection with this case that Schmidt quoted, and gave enduring character to the opinion laid down, without any example, by Carriere (1889, p. 146) : "There is no relation at all between the size of the parietal foramen and the size and development [Ausbildung] of the parietal organ, nor between the size of the foramen and the position of the organ. . . . Bone structure in this case allow^s no conclusions as to the soft organ we regard as related. The common conclusion, based primarily upon the larger size particu- larly of a fossil foramen, that the organ was on a higher level of development, is inadmissible." While this statement, being carried on and on in the literature, must have discouraged paleontologists, certain observations of Schmidt himself teach the student dealing solely with bones how he can recognize cases in which the foramen may be considerably larger than was the organ. In the Hoplurus specimen, Schmidt reports, the parietal bone tapers to a very thin lamina as it approaches the foramen. In all the Agama specimens with a wide foramen, the parietal bones were thin plates, thinning further towards the foramen (Fig. 14 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 2B). The Laeerta ocellata specimens which Schmidt found described in the literature had parietal bones far thicker than the specimen in which he found the wider gap. The numerous specimens of 40 lizard species which he studied, taken together, convinced Schmidt (p. 374) that the parietal foramen is small, more or less fitting the organ, in thick bone. A foramen much wider than the organ occurs in bone which is thin and which, further, tapers toward the foramen. "It seems plausible that in a thick bone the tendency and capacity to lessen or close an opening are greater than in a thin bone." One wonders whether the latter condition (thin bone : wide foramen) is a phase preceding the former (thicker bone: close- fitting foramen) in ontogeny. Apparently the question has never been raised as to whether in the Lacertilia a wide parietal fora- men is a character of very young individuals. In descriptions of post-embryonic specimens, body length is hardly ever given, and individual age never. Probably, as little was known of the age of these Kecent specimens as is the case of all fossil indi- viduals. Hanitsch (1889) did compare two age stages; but his specimens were, besides one adult, two newborn Anguis whose parietal bones were still unossified. Significantly, however, in a nearly mature embryo of Laeerta the parietal foramen was described as "a very large pentagonal fontanelle" (Parker, 1880, p. 630). Among the specimens cited above as examples of absolute si^e of the parietal organ is a Phrynosoma in which the organ diam- eters were 0.26 and 0.17 mm. Two other Phrynosoma specimens were cited among those with foramina much wider — 3 x and 4 X wider — than the organ. However, in one of the two MCZ Phrynosoma skulls, both 15 mm. long dorsomedially, I could not find the foramen with a magnifying glass, and in the other the foramen is a minute circle, just visible to the naked eye : it is not larger than the organ whose diameters were measured. The conclusion forcibly presents itself that in the two individuals with great size discrepancies between organ and foramen the parietal bone was not fully ossified. Presumably they were younger than either the individual in which the organ was meas- ured or those represented by the MCZ skulls. (Perhaps very young heads were purposely chosen for sectioning in preference EDINGER: PARIETAL FORAMEN 15 to heads containing such thickened, knobby, and strongly spined skulls.) Studying the skulls alone can not, of course, answer the question as to whether the relation between the soft organ and the skull foramen is different in the very young from that in the adult. Study of the MCZ collection of lizard skulls, all of them well ossified, showed only that there is much individual variation in foramen size, and that fortification of the foramen C - -I Fig. 4. Iguana iguana {A, C) and Iguana fhinolopha {B, D). D2, Ventral view of cranial roof, showing remnant of dorsally closed parietal foramen. X about 2/3. by a ring of dense bone (Fig. 4) develops at different ages in different individuals; and it suggested that closure of the fora- men (Fig. 4Z>) occurs late in individual life. It may well be because specimens of fossil reptiles are not, in general, those of very young individuals that I know of no fossil parietal foramen with tapering margins. Many types of 16 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY extinct reptiles have, on the contrary, the foramen margins raised and buttressed, as have some lizards. In several kinds of mam- mal-like reptiles, such fortification developed into "a sort of Cyclopean periscope" (Gregory, 1926, p. 186) rising above the general level of the parietal region — a large perforated boss (e.g., Moschops; ibid., pi. XV) or a chimney-like structure {Hippo saur us ; Haughton, 1929, fig. 20). A strikingly large parietal opening is often found in very thick bone — indeed, a canal rather than a foramen (Fig. 3). Thick bone, be it noted, is a condition in which, according to Schmidt's considerable ma- terial of Lacertilia, major size discrepancies between parietal foramen and organ do not occur. While occasionally highlighting cases in Avhich the parietal foramen was found strikingly larger than the organ, the descrip- tions of the parietal eye in Lacertilia have nowhere answered the question of what, if any, are its normal size relations to the foramen. Figured data. Not the texts of the parietal eye studies but their illustrations constitute a record of relations between the organ and the parietal foramen. In all suitable figures that I could find, I have measured the maximum diameter of the eye (without the surrounding "sclera") and the minimum diameter of the foramen. The ratios of eye diameter to foramen diameter are listed in the table (p. 00). It is unfortunate that these ratios must be interpreted without regard to age factors, as the latter are unknown. Moreover, the list is rather short. The big surveys by their rich illustration suggest that the literature they summarize must contain many figures well-suited to the purpose, showing in section not only the eye but also both sides of the foramen. Actually, a 1940 figure for example is likely to be a copy, from Studnicka 1905, of a 19th century figure. The data I brought together were found in figures published by the following authors: Boveri (B), de Graaf (G), Hanitsch (H), Leydig (L), Owsiannikow (0), Schmidt (S), Spencer (Sp), and Studnicka (St). EDINGER: PARIETAL FORAMEN 17 TABLE OF RATIOS Between PARIETAL EYE (=1) and parietal foramen Seps (L) 0.4 Lacerta (L) 1.4 Chamaeleo (S) 0.5 Lacerta (St) 1.4 Chamaeleo (S) 0.7 Lacerta (Sp) 1.5 Anguis (G) 0.8 Varanus (Sp) 1.5 Seps (Sp) 0.8 Chamaeleo (Sp) 1.5 Lacerta (L) 0.8 Lacerta (L) 1.6 Lacerta (L) 0.8 Chamaeleo (Sp) 1.7 Far anus (Sp) 1.0 Varanus (Sp) 1.7 Maiuia (S) 1.0 Pseudopus (St) 1.8 Lacerta (L) 1.1 Varanus (B) 1.8 Lacerta (L) 1.1 Lacerta (L) 1.1 Anolis (Sp) 1.1 Lacerta (Sp) 2.1 Eumeces (S) 1.1 Caiofes (Sp) 2.1 Lacerta (0) 1.2 Leiodera (Sp) 2.2 Chalcides (S) 1.2 Moloch (Sp) 2.3 Pseudopus (St) 1.3 Lacerta (L) 1.3 Agama (B) 1.3 Chamaeleo (B) 1.3 Agama (S) 3.2 Anguis (H) 1.4 PZica (Sp) 5.8 0-1,2 I.3-L5 1.6-1.8 19-2.1 2.2-24 2.5-2.7 2B-30 3.l-i3 5.8 Fig. 5. Distribution of diameter ratios between parietal eye and foramen; data from table. The illustrations in which I measured diameters of parietal eye and foramen represent 37 individuals. The table, and the graph illustrating the table (Fig. 5), show that the ratios fall 18 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY into three groups. The small gap between the first and second group is, however, probably not a genuine separation but merely due to the paucity of specimens. By far the largest group is the first. Thirty-one ratios form a continuous series in which the foramen diameter varies from about one-half^ to one and four-fifths times the diameter of the eye. One or both the Anolis specimens mentioned by Schmidt also belong in this group. The majority of the specimens (19) are in the 1 :1 to 1 :1.5 range. This is a span of size relations to be expected between an organ and an orifice formed by the more or less remote influence of the organ to keep it open and by the tendency of bone growth to close it. One must remember that for capacity function of the parietal eye the foramen need not be much larger than the "lens"; that there is great variation in the thickness of the "scleral" tissue outside the eye proper, as also in the amount of its blood capillaries ; and that the eye is never directly in touch with the foramen margins. One also has to consider a feature of the material on which the table is based. "When the direction of a figured section is mentioned at all in the caption, it is identified as medio-sagittal, i.e., longi- tudinal. As the foramen is either circular or a longitudinal oval or slit, I must have, in general, measured the longest diameter of the foramen, while transversely the eye : foramen ratio was presumably smaller. In the second group, in 4 specimens, the foramen diameter is slightly more than twice the diameter of the eye. This brings to 35, out of 37, the cases which, to me at least, demonstrate that, although the two related structures are not in contact, there does exist a definite if variable size relationship between foramen and eye. Only 2 specimens are in the third group, that with foramina 3 to 6 times as long as the eyes. This, however, is the group to which must be added cases described in the literature, quoted above : 2 Phrynosoma (x3, x4), 1 Hoplurus (x6), 1 Lacerta (more than 5 x) and possibly some of Schmidt's Agama specimens ("a multiple"). As such great differences in size were observed in particularly thin skulls, they probably occur in particularly young individuals. 1 The Chamaeleo cases are abnormal : the degenerated eye lies above the foramen. EDINGEB: PARIETAL FORAMEN 19 Taking together the evidence of published figures and the data reported in texts, we see that there are 35 cases of what may be taken as normal size relations, against the 6 definite cases of considerable discrepancy which have been so highlighted in the literature as to put in doubt all paleontological research on the pineal complex. From the structure of the lacertilian parietal organ we have above concluded that it must be a type of photoreceptor whose capacity is proportionate to its size. In my opinion, the size relations between foramen and eye in the extant Lacertilia entitle the paleontologist to conclude that the larger a parietal foramen is in a fossil skull, the larger was the organ it served. FORAMEN AND SKULL The size question. What do the authors mean when they write about "small" and about "large" parietal foramina ? It is odd that this question should have to be asked, but I have not found it answered in either the neontological or the paleonto- logical literature. Wettstein, zoologist, stated that Sphenodon, "representative of the lowest evolutionary stage among the living reptiles . . . has a relatively large parietal foramen" (1931, pp. 2, 33). This, presumably, is meant to say that a tuatara skull has a larger parietal foramen than a similar-sized skull of other living rep- tiles. This is indeed the ease ; the foramen is 3 x 1 mm. both in an MCZ tuatara skull (mediosagittal length 56 mm.) and in Wettstein 's specimen (ibid., fig. 37; skull 51 mm.). Among the comparable lizard skulls I measured, the largest foramen is 1x1 mm. Huene, paleontologist, made a survey as to "which stego- cephalians and reptiles have a relatively large-sized parietal foramen" (1933, pp. 372-373). He concluded from the distribu- tion of ' ' large ' ' foramina that the parietal eye was a more potent organ in certain underwater conditions than it was, and is, in terrestrial habitats. From this it would seem that by size Huene meant absolute size, and by ' ' relatively large ' ' he meant that the foramen is absolutely larger in one group than in another. 20 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY I realized, however, that this interpretation was wrong when I checked on one of Huene's examples — the suborders of the Sauropterygia, an extinct order on which Huene is an expert, and of which I have measured the foramen parietale in numerous skulls (Edinger, MS, in preparation). "In the placodonts that foramen is large, in the nothosaurs smaller, and similar in the plesiosaurs" (Huene, loc cit.). But, in e.g. the Middle Triassic, Placodns skulls (median length 158 to 200 mm.) have the parietal foramen 10 to 19 mm. long, with breadths from 10 to 14 mm. ; and in Nothosaurus skulls (125 to about 800 mm. long), length of the foramen ranges from 4 to as much as 23 mm., breadth from 2.5 to 23 mm. Parietal foramina still larger developed later in at least one infraorder of the Plesiosauria. A 50 mm.-long foramen was mentioned above ; there have further long been on record another Jurassic Pliosaurus (maximum skull length 1160 mm.) with a 40 x 15 mm. foramen (Linder, 1913, p. 394), and a Cretaceous Brachauchenius (skull about 800 mm.) vrith a fora- men approximately 40 mm. in length (Williston, 1907, pp. 479, 484) . It thus appears that Huene's statements regarding the size of the opening did not imply absolute size but, rather, size rela- tive to skull size. The fact is that the parietal foramen is a far more conspicuous feature in the center of a short, broad Placodus head than in the back part of the slender, elongated head of Nothosaurus or a plesiosaur. I believe that this is what most authors have in mind when they call a parietal foramen large or small ; they mean the size of the foramen relative to the size of the whole skull. Others apparently intend to say that the foramen they describe is larger or smaller than in some other form whose head, or perhaps whose body is of comparable size. Those few who paid attention to the foramen: organ ratio called "small" a parietal foramen approximately the size of or smaller than the organ, while "large" in such a study means that a foramen is larger than the organ. There has, further, been occasional comment on small size of the parietal foramen as compared with the lateral eye orbits. Still further, some authors must have been thinking merely of absolute size of the parietal foramen. A7i answer. Great differences in absolute size of the parietal EDINGER: PARIETAL FORAMEN 21 organ and hence of the parietal foramen presumably are, we have seen, significant physiologically. But in the comparative anatomy of a group as vast as the Class Reptilia the meaning of absolute size of the foramen differs enormously. "We have to compare tiny animals with animals whose body is a hundred or a thousand times larger. Obviously, the size of the parietal organ relative to the size of the body would be not only difficult to calculate but also of dubious value. The same is true of the size of the foramen relative to the skull as a whole. We find the foramen in skulls compactly constructed, and in skulls drawn out into short or long muzzles; but the parietal foramen is a part of the cranium, the ''Hirnschadel" of German terminology which, like the brain, is never involved in the process of skull elongation as much as the facial portion ("Kieferschadel"). A meaningful expression of parietal eye size can be found through the fact that this organ, in contrast to the lateral eyes, is entirely a part of the brain. Skulls do not build a separate orbit to lodge the parietal eye. This lies either within the brain chamber or, if in the parietal foramen, in an extension of the brain chamber. Seeing the parietal eye as one of the brain organs, one is not surprised by the discovery of Spencer quoted above — namely, that this organ can be of the same size in a two-foot long and in a six-foot long reptile. The entire brain, including the extensive olfactory tracts whose length necessarily varies with skull length, was found to be 70 mm. long in a Fython of about 6.5 m. length, and 40 mm. in a PytJion of about 2.5 m. length (Edinger, 1929, p. 100). The difference in brain-to-body length ratios (1:93 and 1:63, respectively) would be far greater if only the contents of the braincase proper had been measured. It follows that calling the parietal foramen small in large animals or in large skulls may be as unfortunate as the common usage of calling the brains of elongated heads smaller than the brains in short heads. It is even assumed that the former are less efficient than brains which constitute a large portion of the head. Efficiency of the various brain functions depends, of course, much more on the composition than on the size of the brain, relative or absolute. Greater or lesser relative size, within the brain, of one or more of its components, is in general the 22 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY anatomical correlate of greater or lesser importance of the func- tions they serve. In consequence, a parietal organ should be regarded as large when it is a relatively large appendage of the brain. Its osteologi- cal reflection, the parietal foramen, is large when it is a consider- able interruption of the braincase roof, otherwise closed. In many reptiles an approximate size relation of the parietal organ to the rest of the brain can be shown three-dimensionally by endocasts of the braincase (Fig. 6). This method, however, is not practicable in the Lacertilia and the many other groups whose braincase is not a complete bony capsule. But even in species whose braincase is incompletely ossified, an index of brain size is available. The occipital area, which lodges the hindbrain, is usually well ossified. Consequently, macerated or fossil skulls in which the foramen parietale is also present include the foramen occipitale magnum. This surrounds the posterior end of the brain. The lumen of the foramen mag- num is somewhat larger than but corresponds approximately to a cross section of the medulla oblongata. The relation between bone and medulla is here similar to that between parietal foramen and parietal eye. In both these cranial openings, size differences roughly represent size differences of the related organs. Com- paring the diameters of foramen magnum and foramen parietale is a method not as exact as paleoneurology can use in other in- vestigations but preferable, I believe, to the vague methods of comparison hitherto used. I regard as a fortunate circumstance the fact that the brain area whose osseous counterpart we can compare with that of the parietal eye is in the most conservative portion of the brain. The medulla, in contrast to the brain as a whole, is of similar shape in all reptiles ; and its thickness varies mainly, or exclusively, with body size. For the appraisal of parietal organ size by measuring and comparing the diameters of foramen parietale and foramen magnum, an immense amount of material is available in extinct and extant reptilian skulls. However, for large-scale comparisons of the two foramina, paleontology is under more than its usual handicaps. Crushing of course can practically eliminate even the foramen magnum (for example, in all plesiosaurs preserved in slates). A fragment may contain one but not both foramina. The good fossil skulls are held in collections all over the world; de- EDINGER: PARIETAL FORAMEN 23 scriptions of them rarely list foramen diameters, and figures show but indistinctly that small feature, the parietal foramen. A few examples will be given here of this means of comparison. We will first reconsider from this viewpoint animals mentioned above in one or another context. We can then examine with this method two of the general evolutionary concepts mentioned which, incidentally, are contradictory (as is typical of the pineal complex literature). Is the size of the parietal organ correlated with the habitat of a reptile? Do large parietal foramina occur only in very ancient and/or very primitive reptiles ? The parietal foramen of Sphenodon is larger than in lizards relative to the brain, as expressed by the approximate breadth and height of the medulla oblongata as well as by other standards of comparison. In the two skulls mentioned, in which the parietal foramen is 3 x 1 mm., the foramen magnum is 5 x 5 (MCZ) and 5x4 mm. (Wettstein, 1931, fig. 32). The relatively smaller size of the parietal foramen in Lacertilia may be illustrated by two series of skulls in the MCZ Department of Herpetology. One series consists of 8 Iguana skulls with dorsomedian skull lengths from 38 to 100 mm. Nos. I and IV- VII are I. igua^ia; II, III and VIII are I. rhinolopha (see Fig. 4). The foramen maguum is 5 x 4 mm. in the smallest, 5x5 mm. in the next four, 6x6 mm. in a 64 mm.-long skull, but 7x7 mm. in a 60 mm.-long and in the largest skull. The parietal foramen meas- ures fractions of 1 mm. in I, IV and VI, and reaches diameters of 1 mm. in II (length even 1.2 mm.). Ill, V, VII. The foramen is dorsally closed in the 100 mm. skull ; its remnant is a ventral circular pit with 1 mm. diameters.^ The other series is one of six skulls of Ctenosaura pectinata, 55 to 84 mm. long. Breadth of the foramen magnum rises from 5 to 6 mm. and so does height except that, while it is 6 mm. in specimens IV and VI, height is only 5 mm. in V. The parietal foramen is 1 x 1 mm. large only in IV and in VI, the largest skull. Length is 1 mm. also in the smallest specimen, but breadth is less than 1 mm., as are the diameters in the other three speci- mens. 1 As the closure phenomenon is apparently not previously noted in lizards, I mav add here that it is also shown in a series of eight Phifsignathiis cocinciiius skulls ."G to 52 mm. long. The parietal foramen, invisible to the naked eye in the smallest, tiny in all others except in one of the five 45 to 47 mm.-long skulls (1 xO.5 mm.), is closed in the largest siieeimen. 24 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The parietal foramen of the giant, Varanus komodoensis, so conspicuously large with diameters of 2 and 3 mm., is similar in relative size to that of some of the smaller lizards ; the foramen magnum is 14 mm. broad and 12 mm. high. If an attempt is made to compare the areas rather than diameters of the two foramina — by the method of multiply- ing with each other the diameters of the openings — the difference between the rhynchocephalian and the lizards becomes quite striking. The parietal foramen then compares to the foramen magnum as 1 -.7 and 1 .-8 in Sphenodon, but as 1 :28 in Varanus komodoensis, and the largest parietal foramina in the Iguana and Ctenosaura series have ratios of 1 :21 and 1 :36, respectively. Briefly, in the lizards I measured, the relatively largest diam- eters of the parietal foramen are only about one-fifth of the diameters of the foramen magnum ; in most specimens they are far less. This is the osteological expression of a parietal eye so small that its existence remained unknown for many decades of research on lizard brains. Application to fossils. Much larger than in any of the Kecent Reptilia was the parietal foramen in specimens of the two sauropterj^gian genera discussed above. There are late Middle Triassic (Upper Musehelkalk) skulls both of Placodus and of Nothosaurus in which the foramen parietale is almost as large as the foramen magnum. Before this phase, however, parietal foramen diameters in Lower Musehelkalk Nothosaurus averaged only about half those of the foramen magnum ; and they seem to have shrunk to about one-third in an Upper Triassic Notho- saurus. In the placodonts, the foramen ultimately closed {Hen- odus, Upper Triassic; Edinger, MS. in preparation). Similarly, the foramen had disappeared in the last of the plesiosaurs (at least Hydrotherosaurus, late Upper Cretaceous; Welles, 1943, p. 133). The Placodontia and Nothosaurus were semi-aquatic seashore dwellers at the times when their parietal eyes were small and when they were large. It is really not possible to correlate evolu- tion of a large parietal eye with the environment of the animals in which it developed. The history of this structure in placodonts and Nothosaurus constitutes but two examples of a phenomenon EDINGER: PARIETAL FORAMEN 25 which occurred repeatedly in the history of the reptiles. The parietal foramen was enlarged for some time in several branches of the Reptilia, in different habitats, in different periods of earth history. The story of Nothosaurus may. further, serve as one of the actual records which do not fit into the theory that the parietal foramen is larger in primitive than in progressive forms. In this long-lived genus, neither the evolutionary enlargement nor the reduction of the parietal eye was paralleled by any notice- able change in other structures, except that some of the Upper Muschelkalk individuals were very large animals. When it is said that the genuine foramen parietale occurs only in primitive vertebrates, this is true in so far as neither foramen nor eye are left in birds and mammals. On the other hand, the statements are unacceptable which assert that among those vertebrate groups in which the foramen occurs it is present, or it is large, only in primitive forms. Such assertions do not agree with the record. The vertebrates with parietal foramina are extinct Agnatha, Pisces, and Amphibia, and both extinct and extant Reptilia. I have not yet found a large parietal foramen in a fish; among amphibians, the foramen appears to be large in those labyrinthodont skulls which represent young individuals — but my studies are far from concluded. At present it seems to me that parietal foramina of strikingly large relative size occurred, in gnathostomes, only among Reptilia. Apparently the related ancient photoreceptor came to play a larger role in this class than in the other classes. This view finds support, I believe, in differences between the dorsal eyes of living verte- brates, that is, of lampreys and of reptiles (Fig. 1 A and C). Only in the reptiles is the ocular structure perfected ; only the reptilian parietal eye has a distinct, clear lens. The various extinct Reptilia in which the parietal eye became large cannot all be regarded as primitive. Permian Therapsida, mammal-like reptiles of the suborders Theriodontia and Dino- cephalia were mentioned above as having the parietal opening as a wide canal through a thick skull roof (Fig. 3), or having it strongly buttressed by a superstructure of the parietal bone. In the parietal cone of Delphinorinathus conoccphalus, Seeley found a circular parietal foramen "nearly 2 cm. in diameter"; the foramen magnum "appears to have been not more than 2.7 26 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY cm. high by 1.3 cm. wide" (1892, pp. 469, 471). A certain im- portance of the parietal eye in some kinds of therapsids is also demonstrated by the endocranial cast of a later, early Triassic representative of still another suborder, Dicynodontia. In a 141 mm. -long skull of Lystrosaurus (Fig. 6^), the parietal foramen does not appear to be particularly large. In the brain (Fig. 6B), the parietal organ must have been an appendage of considerable size. The channel it produced in the skull is 18 mm. high caudally where it joins the midbrain chamber. On the Fig. 6. Lystrosaurus sp. A, Skull in dorsal view, from Huene 1931; x 1/3. B, Its endocranial cast in left side vieWj M.C.Z. 2124 (duplicate of specimen in Geol. -paliiont. Inst. d. Univ. TUbingen; x 1). EDINGER: PARIETAL FORAMEN 27 outer skull surface the diameters of the opening (9x7 mm.) are about half those of the foramen magnum (18 x 13 mm.) ; but below that, the east has revealed a 10 x 10 mm. swelling which presumably indicates the position of the parietal eye. As ap- parently everywhere among the Reptilia when the parietal fora- men is known in more than one individual, its variability can be seen in Lystrosaurus. In a 171 mm. -long skull, the parietal canal is 17 mm. deep, and its diameters on the surface are 10 and 7 mm., much as in our figured specimen; but the diameters of the foramen magnum, 17 and 23 mm., are more than twice those of the foramen parietale in this larger skull (van Hoepen, 1913, pp. 25, 34). To judge from Brink's figures (1951, 3D and E), the parietal foramen was still smaller, relative to the foramen magnum, in the largest of complete Lystrosaurus skulls.^ The Therapsida are a reptilian order greatly specialized in the direction of mammals. It contains many forms in which the parietal foramen was very small, and the foramen was lost before the therapsids gave rise to the new class. But in this order there also developed at least some individuals in which the parietal eye was half to fully as large in section as the medulla oblongata. The most primitive reptiles are included in the order Cotylo- sauria. According to the general concept of parietal eye devolu- tion, the early cotylosaurs should have the largest parietal fora- mina among the Reptilia. A paragon, in the literature, of parietal foramina ' ' exception- ally large" (Gregory, 1951, p. 338), is that of the early Permian cotylosaur Diadectes. It is indeed huge as seen in the short and massive skulls, a perpendicular tunnel through an extremely thick skull table. Less impressive is its endocast, seen as a dorsal extension of an endocast of a cranial cavity (Gregory, 1951, figs. 12. 13. Al). As I have reason to suspect that this endocranial cast is not that of one specimen but a composite which Cope took from several cranial fragments, the size of the parietal organ is better appraised by the method used above. Among the Diadectes specimens in the American Museum of Natural History, the often-figured, undistorted skuU No. 4839 contains 1 Postscript. Six Lystrosaurus skulls were measured in the paleontological col- lection of Bavaria and in the British Museum (Nat. Hist.). The foramen magnum diameters are in one case more than three times, in three cases more than twice, and in two cases less than twice those of the foramen parietale. 28 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the foramen magnum as well as the foramen parietale. The former is 17 mm. high and 19 mm. broad. The latter is prac- tically the same size : 16 mm. long and 20 mm. broad (the respec- tive "areas" being 323 and 320 mm.^). A parietal foramen far larger (at least in an absolute sense) is seen in the Diadectes fragment AMNH 4352, the partial roof of a skull which ap- parently was but little larger than No. 4839, and also in the excellent natural (stone-) mould of its ventral side (Gregory, 1951, figs. 12.13. A 2 and A3). While in other specimens the parietal foramen is more or less circular, in this specimen it is Fig. 7. Right halves of sectioned cotylosaur skulls. Price; x 3-1/3. B, Diadectes, after Gregory; x 1/3. A, Captorhinus, after 15 mm. broad and as much as 28 mm. long. Thus some Diadectes may have had a parietal organ larger, relatively, than e.g. Placodus. To judge from figures in the literature, some later diadectomorphs indeed had a parietal foramen larger than the foramen magnum (e.g. Pareiasaiirus, Boonstra, 1930, figs. 1-2). These later Diadectomorpha, nota bene, were among the types more advanced than Diadectes; and Diadectes itself was not a truly primitive but a rather large and in several respects a specialized cotylosaur. EDINGER: PARIETAL FORAMEN 29 It is the other, generally more primitive branch of the Cotylo- sauria, the Captorhinomorpha, which "may lie close to the stem of the reptiles, if, indeed ,the stem reptile itself was not a member of it" (Romer, 1945, p. 173). Figure 7 serves to compare L. I. Price's (1935) reconstruction of a medio-sagitally sectioned Captorhinus braincase, a careful combination of his findings in six specimens, and W. K. Gregory's (1951) similar reconstruction of Diadectes, based on the AMNH material. The figure shows that the length of the foramen parie- tale was much less in the primitive than in the specialized coty- losaur, relative to the height of the foramen magnum. It is true that the foramen magnum is so high in Captorliinus that surely not its entire height was occupied by the medulla; its breadth, however, is also greater than the diameters of the parietal fora- men, which is circular. In Protorothyris, Romeria and other earlier and more primitive captorhinoids, the relative sizes of foramen parietale and foramen magnum appear to be essentially the same as in Captorhinus, according to Mr. John B. Clark, who is studying this group. He further informs me that the occiput is generally distorted so that his conclusion is in part based on observations of the braincase as a whole and its relationship to the whole skull, where there is a very strong resemblance to Captoi'hinus. In Seyynouria, primitive in combining amphibian and reptilian characters, AYhite (1939, p. 333) described the parietal foramen as shaped like an inverted tunnel, its diameter on the ventral surface of the bone about one and one-half times that on the dorsal surface ; reconstructing the braincase on the basis of sev- eral specimens, White shows (his fig. 10) the external parietal orifice about one-third as long as the foramen magnum is high. Thus in at least some of the truly primitive Reptilia the parietal foramen was relatively smaller than in various progres- sive and in later forms. It was, however, far larger than it is in any surviving reptile. SUMMARY It has been claimed that paleontology is unable to contribute reliable data to the evolutionary history of the pineal complex, 30 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY although it was in the Paleozoic and Mesozoic eras that numerous types of vertebrates had the parietal foramen. In modern times, this foramen occurs only in Lacertilia and the tuatara, and it is invariably associated with a parapineal photoreceptive organ, the parietal eye. The idea that the size of the foramen can not reveal the size of the organ is here rejected. The size proportion between organ and foramen does not vary as much in the Lacertilia as was assumed because a few (possibly juvenile) cases of exceptional discrepancies were repeatedly highlighted in the literature. The conditions found in lizards, rather, give confidence in paleonto- logical study — not, of course, of the entire pineal apparatus, but of the parietal eye. The large parietal foramina in some fossil reptilian skulls are valid evidence as to the one-time presence of parietal eyes far larger than have survived into the Cenozoic era. The concepts that large parietal foramina and organs de- veloped in a particular habitat, or occurred only in very ancient or primitive Reptilia, are put to a test. They appear to be mis- conceptions due in great part to the lack of a definition as to how the size of the foramen should be judged. As the organ is a part of the brain, it is suggested that the parietal foramen be com- pared with another cranial opening, the foramen magnum, cor- responding approximately to the circumference of another brain part, the medulla. Whereas the diameters of the foramen mag- num are, in lizards, at least five times those of the parietal fora- men, parietal foramina relatively far larger were evolved in various other branches of the Reptilia. Examples here given of extinct reptiles with parietal foramina close to the foramen magnum in size include terrestrial Permian, and semi-aquatic Triassic forms. LITERATURE CITED Bargmann, W. 1943. Die Epiphysis cerebri. Mollendorff 's Haudb. mikrosk. Anat. d. Menschen, 6 (IV) : 309-502. BOONSTBA, L. D. 1930. A contribution to the cranial osteology of Fareiasawrus serridens (Owen). Ann. Univ. Stellenbosch, 8 (A): 1-19. EDINGEB: PARIETAL FORAMEN 31 BOVERI, V. 1925. Untersuchungen iiber das Parietalauge der Eeptilien. Acta Zool., 6: 1-57. Brink, A. S. 1951. On the genus " Lystrosauru^ " Cope. Trans. Eoy. Soc. S. Africa, 33: 107-120. Burton, M. 1954. Misleading ' ' Third Eye. ' ' 111. London News, 1954 : 270. Camp, C. L. 1923. Classification of the lizards. Bull. Amer. Mua. Nat. Hist., 48: 289-481. Carriere, J. 1889. Neuere Untersuchungen iiber das Parietalorgan. Biol. Cbl., 9 (1890) : 136-149. Clausen, H. J. and B. Mofshin 1939. The pineal eye of the lizard (Anolis carolinensis) , a photorecep- tor revealed by oxygen consumption studies. Jour. Cell. Comp. Physiol., 14: 29-41. CoE, W. R. and B. W. Kunkel 1906. Studies of the California limbless lizard, Anniella. Trans. Conn. Acad., 12: 349-403. Dean, B. 1891. Pineal fontanelle of placoderm and catfish. Rep. Conun. Fisheries New York, 19: 306-363. 1895. Fishes, living and fossil. Columbia Univ. Biol. Sex., 3. New York and London. Doderlein, L. 1890. Vertebrata. In Steinmann, G. und L. Doderlein, Elemente der Palaontologie, pp. 516-827. Leipzig. Edinger, T. 1929. Die fossilen Gehirne. Erg. Anat. Entwicklungsgesch., 28: 1-249. (In press). Paired pineal organs. Folia Psychiatr. Neurol. Neurochir. Neerl., Suppl. El-Touibi, M. R. 1945. Notes on the cranial osteology of TJromastyx aegyptiam (For- skal). Bull. Fac. Sci. Fouad I Univ., 25: 1-12. Engelbrecht, D. v. Z. 1951. Contributions to the cranial morphology of the chamaeleon Miorosaura pumila Daudin. Ann. Uuiv. Stellenbosch, 27 (A) : 3-31. 32 BULLETIX: MUSEUM OF COMPARATIVE ZOOLOGY Gladstone, R. J. and C. P. G. Wakelet 1940. The pineal organ. The comparative anatomy of median and lateral eyes, with special reference to the origin of the pineal body; and a description of the human pineal organ considered from clinical and surgical standpoints. XVI + 528 pp. London and Baltimore. Graaf, W. H. de 1886. Bijdrage tot de kennis van den bouw en de ontwikkeling der epiphyse bij amphibian en reptilien. Leiden. Gregory, W. K. 1926. The skeleton of Moschops capensis Broom, a dtnocephalian rep- tile from the Permian of South Africa. Bull. Amer. Mus. Nat. Hist., 56: 179-251. 19.")1. Evolution emerging. 2 vols. New York. Griffiths, I. 1954. On the nature of the fronto-parietal in Amphibia, Salientia. Proe. Zool. Soc. London, 123: 781-792. Hanitsch, E. 1889. On the pineal eye of the young and adult Anguis fragilis. Proc. Liveipool Biol. Soc, 3: 87-95. Haughton, S. H. 1918. Some new carnivorous Therapsida, with notes upon the brain- case of certain species. Ann. S. Afr. Mus., 12: 175-216. 1929. On some new therapsid genera. Ann. S. Afr. Mus., 28: 55-78. HoEPEisr, E. C. N. v. 1913. Bijdragen tot de Kennis der Eeptielen van de Karrooformatie. Ann. Transvaal Mus., 4: 1-46. HUENE, F. V. 1931. Beitrag zur Kenntnis der Fauna der siidafrikanischen Karroofor- mation. Geol. palaont. Abh., 22 (N.S. 18) : 157-228. 1933. Zur Lebensweise und Yerwandtschaft von Placodus. Abh. senckenb. naturf. Ges., 38: 365-382. KiTAY, J. I. and M. D. Altschule 1954. The pineal gland. A review of the physiologic literature. Cam- bridge, Mass. Leydig, F. 1890. Das Parietalorgan der Amphibien und Eeptilien. Anatomisch- histologische Untersuchung. Abh. senckenb. naturf. Ges., 16: 441-551. LiNDER, H. 1913. Beitrage zur Kenntnis der Plesiosaurier-Gattungen Peloneustes und Pliosaurus. Geol. palaont. Abh., 15 (N.S. 11) : 337-409. Neal, H. Y. and H. W. Rand 1936. Comparative Anatomy. Philadelphia. EDINGER: PARIETAL FORAMEN 33 OWSIANNIKOW, P. 1888. Ueber das dritte Auge bei Petromyzon fiuviatilis nebst eiaigen Bemerkungen iiber dasselbe Organ bei anderen Thieren, M6m. Acad. Sci. St. Petersbourg, (7) 36 (9) : 1-26. Parker, W. K. 1880. On the structure and development of the skull in the Laeertilia. I. On the skull of the common lizards (Lacerta agilis, L. viridis and Zootoca vivipara). Phil. Trans. Roy. Soc. London, 170: 595-640. Price, L. I. 1935. Notes on the brain cai5e of Captor'hinus. Proc. Boston Soc. Nat. Hist., 40: 377-386. RiTTER, W. E. 1891. The parietal eye in some lizards from the western United States. Bull. Mus. Comp. Zool., 20: 209-228. ROMER, A. S. 1945. Vertebrate Paleontology. Chicago. Schmidt, W. J. 1909. Beitrage zur Kenntnis der Parietalorgane der Saurier. Zeitschr. wiss. Zool., 92: 359-426. Seeley, H. G. 1892. On Delphinognathns conoccphalus (Seeley) from the Middle Karoo Beds, Cape Colony, preserved in the South African Mu- seum, Capetown. Quart. Jour. Geol. Soc. London, 48: 469-475. Spencer, W. B. 1886. On the presence and structure of the pineal eye in Laeertilia. Quart. Jour. Microsc. Sci., (N.S.) 27: 165-238. StadtmItller, F. 1936. Kranium und Visceralskelett der Stegocephalen und Amphibien. Bolk's Handb. vergl. Anat. Wirbelt., vol. 4, pp. 501-698. Stigler, R. 1950. Versuche iiber das Parietalauge von Agama colonorum. Zool. Anz., 145: 316-318. Studnicka, F. K, 1905. Die Parietalorgane. Oppel's Lehrb. vergl. mikrosk. Anat. Wir- belt., vol. 5, VIII + 254 pp. TiLNEY, F. and L. F. Waiiren 1919. The morphologj' and evolutional significance of the pineal body. Amer. Anat. Mem., 9: 1-257. Trost, E. 1933. Die Histogenese und Ilistologie des Parietalauges von Anguis fragilis und Chulcides oceUatus. Zeitschr. Zellforsch., 38: 185-217. 34 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Welles, S. P, 1943. Elasmosaurid plesiosaurs, with a description of new material from California and Colorado. Mem. Univ. Cal., 13: 125-254. Wettstein, 0. V. 1931. Ehynchocephalia. Kiikenthal's Handb. Zool., 7 (I) : 1-235. White, T. E. 1939. Osteologj' of Seymouria haylorensis Broili. Bull. Mus. Comp. Zool., 85: 325-410. Welliston, S. W. 1907. The skull of Brachauchenius, with observations on the relation- ships of the plesiosaurs. Proc. U.S. Nat. Mus., 32: 477-489. WOLTERSTORFF, M. 1886. Ueber fossile Frosche, insbesondere das Genus Palaeotatrachus. I. Jahresber. Abh. naturw. Ver. Magdeburg, 1 (1885) : 1-93. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 114, No. 2 A CLASSIFICATION OF THERAPSID REPTILES by D. M. S. Watson and A. S. Romer CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM February, 1956 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 114. Breviora (octavo) 1952 — No. 51 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JoHNSONiA (cjuarto) 1941 — A publication of the Department of ]\Iollusks. V^ol. 3, no. 34 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, ( ambridge 3S, Massachusetts. . Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 114, No. 2 A CLASSIFICATION OF THERAPSID REPTILES by D. M. S. Watson and A. S. Romer CA^iIBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM February, 1956 No. 2 — A Classification of Therapsid Reptiles By D. M. S. Watson and A. S. Romer CONTENTS Page Introduction 38 Order Therapsida 42 Suborder Theriodontia 46 Infraorder Titanosuchia 47 Family Brithopodidae 48 Anteosauridae 49 Jonkeriidae 49 Infraorder Gorgouopsia 51 Family Aelurosauridae 52 Aelurosauropsidae 52 Arctognathidae 53 Arctognathoididae 53 Broomisauridae 54 Burnetiidae 54 Cynariopsidae 54 Galerhinidae 55 Galesuchidae 55 Gorgonognathidae 55 Gorgonopsidae 56 Hipposauridae 56 Inostranceviidae 56 Pachyrhinidae 57 Phthinosuchidae , 57 Eubidgeidae 57 Scylacocephalidae 58 Scylaeopsidae 59 Scymnognathidae 59 Sycosauridae 60 Gorgonopsians lacking family characteristics ........ 60 Gorgonopsians represented by inadequate material ... . . 61 Infraorder Cynodontia 61 Family Procynosuchidae 62 Thrinaxodontidae 63 Cynognathidae 63 Diademodontidae 64 Gomphodontosucliidae 65 Cynodontia incertae sedis 65 38 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Infraorder Ictidosauria 66 Family (unnamed) 66 Tritheledontidae 66 Tritylodontidae 67 Microcleptidae 67 Possible mammals from the ' ' Ehaetic. " 67 Infraorder Therocephalia 68 Family Pristerognathidae 68 Trochosuchidae 69 Whaitsiidae 69 Euchambersiidae 71 Incertae sedis 72 Infraorder Bauriamorpha 72 Family Lycideopsidae 73 Ictidosuchidae 73 Nanictidopsidae 74 Silpholestidae 74 Scaloposauridae 74 Ericiolacertidae 75 Bauriidae 75 Incertae sedis 76 Suborder Anomodontia 76 Infraorder Dinocephalia 77 Family Deuterosauridae 77 Mosehopidae 78 Tapinocephalidae 78 Infraorder Venyukovioidea 79 Family Venyukoviidae 79 Infraorder Dromasauria 80 Infraorder Dicynodontia 85 Therapsida int;ertae sedis 86 INTRODUCTION In 1838 a Scottish engineer, Andrew Geddes Bain, who con- structed military roads in South Africa, and was the first to determine the general geology of that region, discovered fossil reptiles in rocks north of Fort Beaufort in the eastern part of the colony of the Cape of Good Hope. These were the first therapsids to be found (unless some of those from the Ural mountains were collected earlier) and were later described by Richard Owen. Bain and his son Thomas, Dr. Atherstone and some of his friends. WATSON AND ROMER : THERAPSID CLASSIFICATION 39 and some others made large private collections, most of which ultimately reached the British Museum and were also described by Owen, In 1889 H. G. Seeley, aided by a grant from the Royal Society, made a special expedition in order to collect fossil rep- tiles in the Karroo. This was the first expedition of such a kind ever to leave England, and the first to go to the Karroo. Since then local collectors and visitors from many universities have sought reptiles there. The vision and enthusiasm of Robert Broom, then in medical practice, led to the collection by him, and by men inspired by him, of immense numbers of fossil skulls, few of which are adequately prepared. Broom described these materials, founding on them several hundred species, each with a short .statement of matters wherein it differed from others, and a figure, or more recently several figures. Other men added — and are still adding — to the collec- tions and to the list of conventionally diagnosed species. Thus there are now some 688 species described from the Karroo. Of these the great bulk are therapsids. The rocks which have yielded this immense number of terrestrial vertebrates cover a period from some point in the Middle Permian to the end of Triassic time, about 45 million years. The therapsids fall essentially into two groups, plant-eating and animal-eating, the first in effect serving as the food supply of the second series. The first group comprises in fact the only plant- eating terrestrial vertebrates of their time, and with a few ex- ceptions (essentially their ancestors) all fall into the group "Anomodontia," typified by Dicynodon. The range in size of anomodonts is very great ; adult skulls vary from less than 3 cm. to more than 100 cm. in length. There exist very large num- bers, certainly several thousand, of complete but usually unpre- pared skulls, but skeletons are of the greatest rarity; not many more than a dozen are to be found in the museums of the world. The attempts made up to the present to divide this great mass into genera have not reached far, and evidently their understand- ing will depend not only on much further preparation, but on a greatly improved precision of stratigraphical knowledge. Nothing is so important at the present time as a detailed study of the stratigraphy. The remaining therapsids, Gorgonopsia, Theriodontia, etc., are more susceptible to classification, and we felt that for our own 40 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY purposes some attempt should be made to draw up a scheme, with short diagnoses, isolating groups of species which could thus be handled in the literatui'e. The great bulk of the species were described by Robert Broom, a man of genius with a unique visual memory and a critical mind. He recognised some hundreds of therapsids which were not anomodonts. Such a Avealth of species is in no way improbable ; it should be compared with the number of mammalian species found in the Tertiary of the United States from Upper Eocene to Pleistocene times. What this number may be we do not know, but it must vasth^ exceed the number of Karroo reptiles. Thus we are inclined to believe that the majority of Broom's species may be well founded. Both the present authors have handled many Karroo fossils, including a considerable part of the type materials. Thus it seemed that direct comparison of figures would allow us to draw up a classification, which could make no pretence to complete- ness or even probability, but would at least pro\nde some sort of framework for further studies. The material to which the following classification applies is comparable in variety and complexity to the Class Mammalia. The classification is therefore designed on the broadest lines, the families, apart from the Gorgonopsia, being wide groups includ- ing a range of often very varied forms, for no useful purpose could be served by establishing small, tightly defined ones. Our experience of considering systematically the whole litera- ture of the therapsids has shown us that the continued descrip- tion of new species of these reptiles on the level which is now customary is most undesirable. What is wanted is a complete preparation, by the acetic and formic acid techniques, of all the existing type skulls. These should then be fully described, and illustrated by figures representing dorsal and palatal views of the skull, lateral views of skull and lower jaw, an occipital view, and probably a detailed figure of the otic region. The figures should distinguish fractures from genuine margin; restorations of broken outlines (which are often useful) should be made in dotted lines outside the damaged edges. The locality and horizon of the specimen must be stated. The classification of such very varied groups can only progress when the rocks from which they are derived have been adequately WATSON AND ROMER : THERAPSTD rLASSIFTOATION 41 subdivided into stages, and these mapped. The good exposures usual in the Karroo make such mapping easy (in a technical sense) though it involves time and much walking. Such work is by far the most urgent requirement of students of these reptiles, and should have complete priority over the description of any more new species. Miss Townend made for us tracings of all published figures, with the lateral views of skulls drawn facing the same way. Spreading these out side by side, sorting them into groups and making use of the published descriptions, and of our personal knowledge of the types, we grouped them into "clumps" and tried to draw up a workable definition of each. Recently we have had the very great advantage of a publication by S. H. Haughton and A. S. Brink, "A bibliographical list of Reptilia from the Karroo Beds of Africa," Palaeontologia Af- ricana, II, 1954. This is an admirable work, listing each species with a full set of bibliographic references, and giving the horizon and place of discovery of the type specimen. It provides also a series of families, and gives a diagnosis for each genus. The families we have set up were made independently, and chiefly before the publication of Haughton and Brinlv's list. We were able to collaborate in this way because the senior author held an Alexander Agassiz Professorship in the Museum of Comparative Zoology in Harvard University in 1952, and was thus able to discuss the drawings Miss Townend (with help from the Royal Society) made for us. Thanks are due to these two bodies for their generosity. In general we have carried the classification down to the genus only. In the case of the gorgonopsians, however, generic distinc- tions are none too certain in many cases and hence we have listed all species. We have not attempted subdivision of the dicynodonts, since much greater knowledge of skull structure than we possess today is needed before the systematics and evolution of this group can be unravelled. The vast majority of the therapsids listed are from South Africa ; we have noted localities only for forms found in other regions. Stratigraphic horizons for types from South Africa and Russia are given in terms of the zones currently accepted for the Permian and Triassic of those regions. 42 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ORDER THERAPSIDA Advanced theropsid reptiles derived from the Pelycosauria. Characterised by the relatively great size and downtiirning of the paroccipital process so that it abuts directly against the inner surface of the squamosal opposite the inner side of the quadrate, not far above the articulation for the lower jaw. The temporal fossa is widened by an outpushing of the zygomatic arch, which is associated with a lateral extension of the squamo- sal. The posterior surface of this bone faces backward lateral to the ridge which bounds the area of attachment of neck muscles and lies at the outer end of the paroccipital process. The quadrate is always attached to the quadratojugal in a characteristic theropsid manner, the latter bone resting upon the upper surface of the projecting outer condyle of the quadrate, being then separated from it by a quadratojugal foramen, and finally rising as a thin sheet closely applied to the posterior surface of the quadrate. It thus, to a greater or lesser extent, separates the quadrate from that depression on the front face of the lateral extension of the squamosal in which it lies. The suture between the epicondyle of the quadrate and the quadratojugal closes early, and has very rarely been seen. The quadrate-quadratojugal complex becomes much reduced, finally forming little more than a flattened nodule providing the articu- lar surface for the lower jaw. In early forms the distal end of the stapes lies in a depression, the stapedial recess, in the inner edge of the quadrate, and in some later forms actually articu- lates directly with it, the stapedial recess vanishing. The braincase is a massive structure, its component bones usually fusing early in life. The condyle always includes contri- butions from the exoccipitals, and in very advanced forms these alone remain functional, the basioccipital withdrawing from con- tact with the atlantal intereentrum. The prootic (which tends to be more expanded than in pelycosaurs), opisthotic and supra- occipital usually fuse early in life. The paroccipital process is massive and distally has a large abutment with the squamosal. Proximally the opisthotic joins with the prootic to form a fenes- tra ovalis completed by the posterolateral corner of the basi- sphenoid. The fenestra is placed very ventrally so that it is usually completely below the base of the brain. The floor of the WATSON AND ROMER : THERAPSID CLASSIFICATION 43 braincasp, orioinally thick, becomes much thinner with time in the history of some groups, and may eventually be extremely reduced. The basipterygoid process in the most primitive Russian form still articulates with the pterygoid by a smooth cartilage-covered surface. In all later forms the articulation is immobile and usually sutural. In general the process is converted into a flattened sheet of bone projecting horizontally and laterally well in front of the i)ituitary fossa. To it the pterygoid is attached below, and the epipterygoid rests on it. The anterior part of the brain usually lies in a sphenethmoid, a mesial bone resting on the upper border of a deep sheet of bone which is part of the basisphenoid extending forward below the pituitary fossa. Ex- (•ej)tionally the sphenethmoid resolves itself into a pair of orbito- sphenoids. Superficial membrane bones may, as in Dinocephalia, be rough- ened by an ornament, but in the later forms they are usually smooth (with such exceptions as the snout of Cynognathus) . The membrane bones of the temporal region in early forms lie superficially, but later come to be closely attached to the brain- case and largely buried by muscles. There is no supratemporal, and the dermosupraoccipital and tabulars, which long survive, lose all superficial exposure and become thin films "plastered" onto the posterior surface of the widened supraoccipital and parietals and the squamosals which border them. They never transgress the flange which marks the limits of attachment of the neck muscles on the squamosal. The parietals in early forms are widely expanded on the roof of the head, the brain cavity being limited by ridges on their ventral surfaces which may lie much mesial of the free border of the bone. They rest on the supraoc('ii)ital and are tied to it by the dermosupraocciptals and tabulars. The lateral border of the parietal is primitively, and always to some extent, attached to a process from the postorbital which extends backward mesial to the temporal fossa, towards and in more primitive forms to the squamosal. In most forms the parietals separate to surround a pineal foramen, and in some groups are separated by a median preparietal bone. The parietals may form a deep and very narrow sagittal crest. The frontals are relatively small bones usually forming some part of the orbital margin. They are bordered by a postfrontal and, on 44 BULLETIN: MUSEUM OP COMPARATIVE ZOOLOGY the final loss, b}^ reduction in width, of that bone, by the post- orbital which may itself be lost. Anteriorly they join the nasals and are bordered by the prefrontals. The lacrimal, short but always present, retains its superficial position. The face may be very long, or very short, in specialised forms. It is seldom, if ever, pointed and usually is parallel sided, or even bulbous, so that the premaxillae are wide. In consequence the nostrils, in many groups, face as much forward as outward, and are often overhung by a projection of the nasal. There is a septomaxilla, with a superficial exposure in most early forms, withdrawn within the nasal cavity in some advanced animals. Sclerotic plates are occasionally present. There is some evidence which suggests that the latest therapsids possessed mammal-like turbinal bones in their very large muzzles, implying that they may have been warm blooded. The presence of large foramina on the face of late forms suggests that the skin was muscular and sensory and possibly had hairs, wdtli the glands associated with them. Primitively the quadrate ramus of the pterygoid, arising from the attachment to the basipterygoid process, pa.sses directly to the quadrate, but in some advanced forms it may be cut short, scarcely extending l^ehind the basipterygoid process, its place being taken by an epipterygoid process which may itself fail to reach the quadrate. The rest of the pterygoid stretches outward to make a transverse flange against the lower jaw, and forward to unite with the vomers and palatines to roof the mouth. The interpterygoid vacuity is reduced, often to a very small hole between the transverse flanges, or a narrow slit in advance of the basipterygoid processes. The palate so formed is greatly varied : it may be essentially flat, it may be largely vaulted, or it may bear a deep but narrow gulley passing backward from the posterior nostrils. In some forms, in differing ways, a secondary palate is formed so that the point at which inspired air enters the mouth becomes placed far back, hi primitive forms the palate is continuous l)elow the eye but in some a suborbital fenestra, between the pterygoid, ectopterygoid and palatine, makes its appearance. And this suborbital fenestra may in turn vanish with a re-expansion of the bones which surround it. A hyoid bone is not uncommonly seen ; it stretches forward and inward from the neighborhood of the jaw articulation and pre- WATSON AND ROMER : THERAPSID CLASSIFICATION 45 snraably implies the presence of a muscular tongue. The lower jaw is always compouud, but the dentary may be- come very large and a place of muscle attachments. There is always a reflected lamina of the angular. There is usually a marginal dentition, and often palatal teeth, but all these may be lost and functionally replaced by horny sheets. The jaw articulation is opposite or anterior to the level of the occipital condyle. In some cases the lower jaw is capable of very free movements so as to grind up food in the mouth. The vertel)ral column usually lacks intercentra, except an- teriorly; the centra are usually not notochordal. To an in- creasing extent it is diversified, cervicals tend to he sharph- distinguished from the dorsals which follow them, and a lumbar series with short fused ribs becomes established in later forms. There are usually at least three sacrals, and there may be as many as seven. The tail varies immensely in length, being sometimes as long as the rest of the body, sometimes a mere conical hinder end. The limb girdles steadily become modernised, so that although a precoracoid and coracoid are always present the screw-shaped glenoid cavity of pelycosaurs is lost, and the precoracoid ceases to contribute to its surface at all. The clavicles are little ex- panded ; short, broad interclavicles are universal. The cleithrum finally disappears. An ossified .sternum is often present shewing the points of articulation of sternal ribs. The pelvis, at first pely- cosaur-like, gains a thyroid fenestra; the iliac blade lengthens and approaches the mammalian position. The short pubis is still deflected anteriorly, and the ischium shortened. Both fore and hind limbs become more advanced by bringing the elbow and knee in toward the body ; thus the humerus becomes narrower and less twisted than in pelycosaurs and the femur gains an inturned head. The humerus always has an entep- icondylaj-. and sometimes an ectepicondylai- foramen. The del- topectoral crest is elongate. The number of digits is five in both hand and foot of all known thera'psids. The numbers of phal- anges, at first 2.3.4.5.4(3), is reduced to 2.3.3.3.3 in both fore and hind foot, an intermediate stage being known in which the primi- tive number of phalanges is retained but the third phalanx of the third toe, and the third and fourth of the fourth toe are reduced to verv short discs. There is one centrale in the tarsus. 46 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Distal carpal and tarsal five are absent. The effect of these changes is to raise the body of the animal off the ground, so that with the development of a digitigrade foot and the inturning of the elbow and knee, the animal's gait becomes very mammal- like, and its speed presumably largely increased. There are apparently never any epiphyses in therapsids, so that residual cartilages of some substance seem to have remained (throughout life ?) in most members of the group. The therapsids are customarily divided directly into a number of groups such as dinocephalians, dromasaurs, dicynodonts, theri- odonts, etc. It appears, however, highly probable that at a very early stage in tlierapsid evolution there occurred a dichotomy resulting in two main lines, one including mainly carnivorous types and leading toward and to the mammals, the other con- sisting of herbivores, with the dicynodonts as its most conspicuous component. We therefore consider the order Therapsida as consisting of two suborders, here termed Theriodontia and Anomodontia. The former group is an expansion of the Therio- dontia, as usually defined, to include, on the one hand, the car- nivorous dinocephalians and, on the other, the ictidosaurs. In the Anomodontia are here included the herbivorous dinocephal- ians and uroiuasaurs as well as the dicynodonts. The present classification necessitates the division of the Dinocephalia as currently conceived. This last term is here restricted to the herbivorous forms (tapinocephaloids). The carnivorous forms (titanosuchids) fall into two groups, one of which may be called the Titanosuchia whilst the other includes certain very primitive members of the Gorgonopsia, the Phthinosuchidae. SITBORDER TI1P]RI0D()NT1A Therapsids generally wdth a carnivorous dentition in which there is a series of incisor teeth in the premaxilla, while the maxillary series contains towards its anterior end one or more enlarged teeth, the canines. The latter are usually oval in sec- tion and deep rooted. In some late forms this dentition may be converted (as in many mammals) into an herbivorous den- tition by loss or reduction of the canine and modifications of the cheek teeth. Palatal teeth, usually restricted to thickened re- gions of the palatine and pterygoid, are commonly present in WATSON AND ROMER : THERAPSID CLASSIFICATION 4:7 members of the group. A tranverse flange, composed maiiily of pterygoid but with a modest contribution from the ectopterygoid bone, always occurs ; the quadrate ramus of the pterj^goid arises abruptly from the flange, usually considerably dorsal to the admesial part of its lower border, and extends outwards toward and generally to the quadrate. This ramus, for a greater or less part of its length, touches the basisphenoid (in which the para- sphenoid is included). At some point between the pterygoid flange and the posterior end of the basisphenoid (at the tubera which support the front part of the border — usually cartilagi- nous— of the fenestra ovalis) the pterygoid passes ventral to and is attached to the lower surface of a horizontal sheet of bone projecting outwards from the side of the basisphenoid. This projection is the basipterj-goid process, whose upper sur- face is in contact with the base of the epipterygoid. The oc- cipital aspect of the skull has a central area composed of the supraoccipital, the exoccipitals, and the usually powerful opis- thotics, largely coated by the fused dermosupraoccipitals and tabulars. It is bounded by a baclvwardly projecting ridge, made jointly by the tal)ulars and squamosals with contributions from the parietals and interparietal in advanced forms. The supra- temporal has completely vanished, and the posterior surface of the squamosal (lateral to the ridge) is usually of considerable width. The face is never greatly shortened; the lacrimal never reaches the septomaxilla. In contrast to anomodonts, the face is not notably bent down- ward on occiput and suspensorium. The epipterygoid is usually broad. The suspensorium is never of great height, the jaw articulation being essentially in line with the tooth row and about at the level of the occipital condyle. The under surface of the skull tends to be flattened. The fenestra ovalis lies close to the ear cavity. A coronoid bone is present and a "coronoid process" frequently developed in the jaw. Infraorder Titanosuchia The parietal region broad, the temporal fossa lying largely on the outer side of the skull in early forms, its margin carried backwards on to a backwardly turned border of the squamosal. The face is long. There is no preparietal bone. Typically four 48 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY stout incisors, a prominent canine, usually circular in section, and a variable number — typically 8 to 16 — of small cheek teeth in both upper and lower jaws. The quadrate is a bone inserted in front of the lower border of the squamosal, but largely visible from behind. The external nostrils are narrow slits, floored largely by the septomaxilla, just touching the nasal, but largely surrounded by the premaxillae, whose internarial processes slope steeply backwards, separating the anterior ends of the nasals for a considerable distance. The lower jaw is shal- low, with the articular surface directed largely backward and below the line of insertion of cheek teeth into the dentary. The reflected lamina is quite deep and its region of insertion lies far forward on the angular. Postfrontal present. No secondary palate. Paired vomers. No suborbital vacuities. Quadrate ramus of pterygoid well developed and of some height. Quadratojugal relatively large with laterally exposed area. Narrow ascending ramus of epipterygoid. Occipital condyle single. No projecting coronoid process. Cleithrum present. Little iliac expansion ; no thyroid fenestra. The titanosuchians have generally been included in the Dino- cephalia, of which the typical members are large herbivores. Both are unquestionably primitive in many regards and in their most generalised members close to the base of the therapsid stock. But apart from primitive features the two groups have little in common except a clumsy build and a tendency toward thicken- ing of the skull roof, both of which characters appear to be accom- paniments of the usual large size. Family Brithopodidae. Primitive titanosuchians, characterised by a face which is rather deep in proportion to its width and re- tains something of the triangular section found in Dimetrodon. The zygomatic arch is shallow and the temporal fossa large. The parietal foramen is very far back and the plane of the occiput nearly vertical. The quadrate is relatively short and not appreciably carried forward ventrally ; the condyle is separated into two by an oblique spiral depression. Thickening and rough- ening of the bones of the skull is incipiently deA'eloped. Teeth are developed on the palate. The precoracoid enters the glenoid cavity, which retains a trace of its original screw shape. Phalan- geal formula 2.3.3.3.3 in both limbs. The tail very long. WATSON AND ROMER : THERAPSID CLASSIFICATION 49 Very primitive therapsids from the Russian Permian. They were long known only from fragmentary remains, but Efremov has recently described a good skull of Titanophoneus and Orlov has given a general description of the skulls of Admetophonens and Syodon. Admetophoneus Efremov 1954, Russian zones 1 and 11 ; Briiho- pus Kutorga 1838 {Dinomurus Fischer 1847, Eurosaurus Fischer 1842, Orthopus Kutorga 1838, fUhopalodon Fischer 1841), Rus- sian zone I; Syodon Kutorga 1838 (Cliorhizodon Twelvetrees 1880), Russian zones I and 11 ; Titanophoneus Efremov 1938, Russian zone II. Limb bones found with brithopids are brigaded together as Phreatosuchidae by Efremov, and are remarkable because their structure includes many features which characterise pelycosaurs. They are described as Phreatophasma Efremov 1954, Phreato- saurus Efremov 1954, and Phreatosuchus Efremov 1954. Family Anteosauridae. Titanosuchians of large size with skulls usually greatly thickened in aged individuals. The pineal fora- men lies relatively far forward, but the intertemporal surface is wide, the postorbital strips being largely covered by muscle. The face is short, deep, and triangular in section (with a rounded median ridge). The frontal area is swollen and usually forms a l)row over the snout. The facial processes of the premaxillae have a nearly straight dorsal border and extend far backward as narroAV points betAveen the nasals. Their dentigerous border slopes upward in front. The incisors are simple conical teeth which, with the canines, point somewhat forward ; the numerous small maxillary teeth lie on a festooned border. The quadrate and attached quadratojugal are largely exposed below the squamosal, but continue the general direction of the hinder sur- face of that bone. The basis crauii is thin ; the paroccipital proc- ess, though wide from back to front, is shallow. Anfeosauriis Watson 1921 (Titanognathus Broili and Schroe- der 1935, Dinosuchus Broom 1936, Broomosuchus Camp, Taylor and Welles 1942), Micranteosanrus Boonstra 1954, Paranteosau- rus Boonstra 1954, Pseudanteosaurus Boonstra 1954 ; all from the Tapinocephalus zone. Family Jonkeriidae. Large titanosuchians with a very elon- gated face, wider than high and of nearly the same depth at the 50 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY nostril and the front of the orbit. The face is rounded in section and in no way triangular. The temporal region has a wide bar between the fenestrae, composed very largely of parietals, with the pineal foramen far forward. The temporal fossa is much smaller than in the Anteosauridae, and the quadrate very much larger, the quadratojugal attached to it being largely visible below the zygoma in a side view of the skull. The quadrate con- dyle faces largely forward and lies so far in front of the occiput that the temporal fossa appears floored by bone in direct dorsal view. The palate lacks the tooth-covered thickenings found on the palate of the anteosaurs, and has enormously long and wide pala- tal nares, but is generally of similar type. The teeth carried by the widened premaxilla differ from those of Anteosaurus in that they possess basin-shaped crushing surfaces (like those of the tapinocephaloids) behind the high labial spike. The upper sur- face of the parietal and hinder part of the frontal are raised as a boss, a forehead standing out above the nearly straight mid-dorsal line of the face. The jonkeriids as above described have a resemblance, which may be of importance, to the tapinocephaloids. Each group has an expanded cranial region from which a relatively small face projects forwards, and in each group certain members, at any rate, have the highly characteristic crushing premaxillary teeth best explained by Efremov in Deuterosaurus. These unique crushing incisors, taken with the development in each of a similar association of a swollen posterior part of the skull with a relatively slender face, suggest a community of origin, the two groups having parted from one another not very long before their first recognition at the base of the fossiliferous Karroo. lArchaeosuclius Broom 1905 (not identifiable) ; Dinartamus Broom 1923; Dinocifnodon Broom 1929; Dinophoriens Broom 1923; Enohiiis Broom 1923; Glaridodon Seeley 1888 (not identi- fiable) ; Jonkeria van Hoepen 1916 {Dinospliageiis Broom 1929, Dinopolus Broom 1936); ILamiasaurus Watson 1914; Phoneo- suchus Broom 1929; Scapanodon Broom 1904 (not identifiable) ; SdiHiki. Broom 1929; TitanosHchus Owen 1879, all from the Tapinocephalus Zone. watson and romer : therapsid classification 51 Infraorder Gorgonopsia Muzzle typically long and heavj'. Temporal region relatively short, low and moderately expanded. Greatest skull width some- what anterior to the quadrates. Dorsal processes of premaxillae relatively short here and in remaining theriodonts (in contrast to the titanosuchians). The roof of the skull in the temporal region is wide, the postorbital meeting the squamosal and excluding the parietal from the border of the temporal fossa. There is in all later forms a preparietal bone, not usually in contact with the parietal fora- men. The postfrontal is always present. The squamosal is at- tached medialty to the parietal, tabular, and the end of the paroccipital process and laterally forms a thick transversely placed plate which turns forward to form the upper border of the zygomatic arch. Immediately lateral to the end of the paroc- cipital process the anterior face of the lower part of the squamosal is excavated into a recess, with a sharply defined upper border which may actually overhang the recess. Into this the quadrate and the adherent quadratojugal are received in such a way that much of the height of both bones is hidden by the squamosal in occipital view; in more advanced forms only a small proportion of the height of the quadrate can be so seen. There is no sec- ondary palate, the choanae being very extensive openings which extend far back and in all later forms are enlarged on their outer sides anteriorly to form a hole through which the lower canine passes when the mouth is closed. In this and the following theriodont groups the vomers are fused. No suborbital vacuities. In Gorgonopsia, in contrast to the Therocephalia, the lateral border of the quadrate ramus of the pterygoid instead of follow- ing a straight line from its origin at the transverse flange, passes back parallel to the principal plane of the skull for a long way before turning out to join the quadrate, often with a curved border. This condition, though less well developed, anticipates the "girder" existing in higher cynodonts. The strap-shaped epipterygoid often has a quadrate ramus which may very nearly reach the quadrate. In the lower jaw the hinder end of the dentary projects freely above the surangular. Incisor teeth, usually 5 in number, are of large size ; a precanine tooth in the maxilla very rarely occurs; there is a single canine and the 52 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY number of cheek teeth rarely exceeds 5. Palatal teeth occur. The digital formula is 2.3.4.5.4. The preeoracoid is excluded from the glenoid cavity in this and the remaining theriodont groups. Little acromial development. Ilium little expanded; thj'roid fenestra present. The gorgonopsians are relatively primitive theriodonts, sug- gestive of ancestry to the cynodonts; however some features, such as the reduced cheek dentition indicate that typical gor- gonopsians are aberrant to at least some extent. Most gorgonop- sian skulls exhibit but minor variations on a fairly uniform pattern. For systematic purposes most might either be "lumped" in a single large family or divided into a considerable number of small families separated by relatively slight distinctions. The latter course is followed here, although in pursuing it there is found to be a number of genera in which proof is lacking of the distinctive qualities of any of the families erected. Family Aelurosaurldae. Gorgonopsians in which the inter- temporal region is wide and the zygomatic arches probably do not spread widely. The snout is long but rather narrow in proportion to its depth, and the transverse section across it has a rounded upper margin. The dentition is 5.1.4-5, the dentiger- ous surface of the premaxilla sloping downward from the first incisor tooth, and the molar series inserted in a gently convex border of maxilla. The palate possesses teeth in irregular rows along the transverse flanges of the pterygoids, and there are sharply limited patches of teeth on the pterygoids and palatines posterior to the hinder ends of the internal nares. This is a good recognizable group, but difficult to define. ""Aelinoynathus" serratidens Haughton 1915, L. Cistecephalus zone; " Aelurognatlius" cf. serratidens Broili and Schroeder 1934, Cistecephalus zone; '* Aelurogtiathus" sollasi Broili and Schroeder 1935, U. Endothiodon zone; Aelurosauroides watsoni Boonstra 1934, Endothiodon zone; Aelurosaurus felirius Owen 1881 (Aelurosaurva hreviceps Broom 1931, Aelurosaurus striati- dens Broom 1912, Aelurosaurus tenuirosiris Broom 1911, Aeluro- saurus whaitsi Broom 1911), Endothiodon zone; "Gorgonopsier" Broili and Schroeder 1936, E. Africa, "Lower Bonebed." Family Aelurosauropsidae. Small gorgonopsians character- ised by the flattened face, which is nearly twice as wide as it is WATSON AND ROMER : THERAPSID CLASSIFICATION 53 high. The premaxilla below the nostril is deep, the face very little less than half the total length. The intertemporal width exceeds the interorbital. The pineal foramen touches the pre- parietal and is well behind the postorbital bar. The occipital condyle is the most posterior point of the skull. Aelurosauropsis {Aelurusauroides) ivilmanae Broom 1940, M.- U. Cistecephalus zone. Family Arctognathidae. Advanced gorgonopsians with a short muzzle, rounded in section and with its tooth rows widely separated. The tooth-bearing lower border of the maxilla forms a large segment of a circle, the point of origin of the Z3'goma lying high up. Upper teeth 4.1.6. The intertemporal surface is narrow ; the occipital surface meets it far forward ; the pre- parietal lost or very small. The palate has a deep median trough extending back from the posterior nares, overhung by small tooth-bearing projections on the palatines. Behind the ptery- goid flanges the pterygoid and parasphenoid form a narrow girder, the quadrate rami of the pterygoids taking off from it very far back and not reaching the quadrates. The lower sur- face of the basisphenoid forms a gently concave triangular area recalling that of cynodonts. The lower jaw is deep anteriorly, with a pronounced chin, and the dentary extends upward, free from the surangular, for a considerable distance. \'ery recognizable forms. ArctognafJnis curvimola Owen 1876, Cistecephalus zone; Arc- toytwthus ivhaitsi Ilaughton 1924, M. Cistecephalus zone; Lycae- nodontoides haihijvhinus Ilaughton 1929, Cistecephalus zone. Family Arctognaflioididae. Large gorgonopsians, character- ised by a broad face which is about half the total skull length. The orbit is placed high up with a deep jugal below it. The temporal fossae are of moderate size and the occiput stretches forward between the fossae to their anterior ends. The pineal foramen, when present, lies just behind the parieto-frontal su- ture. There is usually no preparietal. Upper dentition 4-5.1.4-6. There are patches of teeth on the pterygoid and palatine. "Aelurognathus" haughtoni Huene 1950, Lower Cistecephalus zone, E. Af r. ; Arctognatiwides hreviceps Boonstra 1934, Cistece- phalus zone; Leontocephalus cadlei Broom 1940, V. Cistecephalus 54 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY zone; Leontosaurus vanderhorsti Broom and George 1950, Cis- tecephalus zone. Family Broomisauridae. Gorgonopsians with a relatively short, -wide, and high face with a rounded upper surface. The continuous tooth-bearing border of the upper jaw is bowed dowuAvard, there being scarcely any trace of a step. The frontal enters the orbital margin by a narrow, parallel-sided process. The temporal fossa is short and wide, the temporal roof forming a third, or little more, of the extreme width. The occipital con- dyle lies far back. Broomisaurus planiceps Broom 1913, Tapinocephalus or Ciste- cephalus zone ; Broomisaurus ruhidgei Broom 1940, Cistecephalus zone; " Lycaenops" pricei Broom and Robinson 1948, L. Ciste- cephalus zone. IFamily Burnetiidae {Burnetiamorpha). Small theriodonts in which the roof of the skull is thickened, shows in general no sutures, and is composed of fine cancellar bone, with essentially no superficial table of hard bone. There is a wide temporal roof, with laterally placed temporal fossae, presumably of gorgonop- sian pattern. The palate lacks suborbital vacuities, has teeth on the pterj^goids and palatines, and apparently a median groove ; there may be 11 post-canine maxillary teeth. The two animals reported to this family differ greatly in age. They may have been produced by parallel evolution from differ- ent parent stocks. Burnetia miraMlis Broom 1923, U. Cistecephalus zone ; Styra- cocephalus platyrhynchus Haughton 1929, U. Tapinocephalus zone. Family Cynariopsidae. Gorgonopsians with a face which is never greatly elongated, rounded on its upper surface and with rather short temporal fossae. The most characteristic feature is the smooth curve made by the tooth-bearing margins of the pre- maxilla and maxilla which bring the anterior incisor so high up that the nose aljove it is extraordinarily shallow. Dentition 5.1.5 or thereabouts. There is a marked groove in the palate behind the posterior nares. Cynarioides gracilis Broom, 1930, U. Endothiodon zone; Cynarioides laticeps Broom 1935, ?L. Cistecephalus zone ; Cyiia- WATSON AND EOMER : THERAPSID CLASSIFICATION 55 rioides teyuns Broom 1925, V. Endotliiodon zone; Cynariops ro- hustus Broom 1925, U. Endothiodon zone. Family Galerhinidae. Small gorgonopsians with a rather flat skull in which the temporal fossa is markedly short, the occiput extends far forward, and the occipital condyle is the most pos- terior part of the whole skull. The family may be distinguished by the lack of a special, narrow, parallel-sided, orbital process of the frontal from some Tapinocephalus zone forms with other- wise comparable features. The zygomatic arch is noticeably shallow. The dentition is ?.1.5 in the upper jaw. Galerhinus ruMdgei Broom 1936, L.-M. Cistecephalus zone ; Galerhinus ruhidgei of Huene 1950, East Africa, ?Cisteeephalus zone; Galerhinns polyodon Broom 1935, ?Cistecephalus zone. Family Galesuchidae. Gorgonopsians in which the frontal enters the orbital margin as a more or less parallel-sided process between the prefrontal and very wide postfrontal. The temporal fossa is large and is not carried back as far as the level of the occipital condyle. In Galesiichus the basicranial region forms a broad expansion ventral to the condyle, unknown in other gorgonopsian groups. In Eoarctops, which may belong here, there are 4 or 5 upper incisors, a canine and 3 molars arising from a continuous jaw margin which seems to show no sign of a step ; in the lower jaw the coronoid process of the dentary projects above the surangu- lar, the hinder part of the jaw being unusually deep. Cerdodon tenuidens Broom 1915, Tapinocephalus zone; Eoarctops vanderlyli Haughton 1929, Tapinocephalus zone; Galesuchus gracilis Haughton 1915, Tapinocephalus zone; Scyla- cognathus major Broom 1935, L. Cistecephalus zone ; Scylaco- gnathus parvus Broom 1913, Tapinocephalus zone. Family Gorgonognafhidac. Large gorgonopsids, with a very long face exceptionally wide at the premaxilla-maxilla suture. The orbit is small (because of the animal's size), the temporal fossa rather small, Avith a very deep zygoma below it, the ventral border of the skull having a deep rounded notch below the orbit. The frontal reaches the orbit by a narrow process, as in Dimetro- don. Distinguished from Rubidgeidae by little except the en- trance of the frontal into the orbit. Gorgonognaihus longifrons Haughton 1915, Cistecephalus 56 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY zone; Gorgonognathus maximus Huene 1950, E. Africa, L. Ciste- cephlaus zone; Gorgonorhinus luckhoffi Broom 1937, ?L. Ciste- cephalus zone; Gorgonorhinus minor Broom 1948, ?L. Cistece- phalus zone ; Tigricephalus kingwilli Broom 1948, Cistecephalus zone. Family Gorgonopsidae. Gorgonopsids in which the intertem- poral region is comparatively wide and the zygomatic arches spread posteriorly. The snout is long and heavy, being both wide and high and the section across it a little in front of the eyes is noticeably square-cut. The squamosals are considerably laterally extended at the sides of the occiput and are shallow. The denti- tion is 5.1.1-5. A very marked step in the maxilla, raising the incisors above the canines and cheek teeth. A small sharply de- fined area of the pterygoid, near the midline, and on the level of the transverse flanges, bears small, sharp-pointed teeth. The basis eranii is very shallow, so that the condyle and paroccipital processes have very little height, and the posttemporal fossae lie very near the lower border of the occiput. Gorgonops forvia^ Owen 1876, Endothiodon zone; Leptofra- chelus eupachygnafhus Watson 1912, Endothiodon zone; "Lycae- nops" kingoriensis Huene 1950, E. Africa, L. Cistecephalus zone. Family Hipposauridae. Gorgonopsians of small to middle size which are characterized by possessing a markedly curved dorsal contour; a very wide parietal region w^ith much excavated occi- put. The interorbital width is great and the orbits, always large, may become relatively enormous. The quadrate region is deep, produced forward as well as downward, and the quadrate itself projects very far below the ventral border of the squamosal and of the ridge made by the tabular. The articular border of the quadrate lies far below the basis eranii, the stapes passing as much downward as outward. Hipposauroides ruhidgei Broom 1 940, Cistecephalus zone ; Hipposaurus hoonstrad Haughton 1929, U. Tapinocephalus zone ; Hipposaurns major Boonstra 1952, Tapinocephalus zone; Ictido- rhiv})s warfiiisi Broom 1913, U. (Usteeephalus zone; LeDiiirosaii- rus pricei Broom 1949, L. Cistecephalus zone ; Pseud ohipposaur us kifcliingi Broom 1948, ?Endothiodon zone. Family Inostrancfviidae. Gorgonopsians which reach a very large size. They have a very wide parietal region, with relatively WATSON AND ROMER : THERAPSID CLASSIFICATION 57 small temporal fossae. The occiput is very wide, stands nearly vertical, and the pineal foramen is immediately in front of its dorsal border. The paroccipital processes are deep. The face is long, its tooth-bearing border without a noticeable step. The dentary has a well marked "chin." The reflected lamina may bear a deep pocket on its lateral surface. Distinguished from Rubidgeidae by lacking the extreme localised widening of the temporal region in later members of that family and the nearly circular widely separated temporal fossae. Amalitzkia and Inostrancevia may be distinct genera. Amalitzkia annae Pravoslavlev 1927, U, Permian, Russia; Amalitzkia wladimiri Pravoslavlev 1927, U. Permian, Russia ; Inostrancevia alexandri Pravoslavlev 1927, U. Permian, Russia ; Inostrancevia latifrons Pravoslavlev 1927, U. Permian, Russia ; Inostrancevia parva Pravoslavlev 1927, U. Permian, Russia; Inostrancevia procJivis Pravoslavlev 1927, U. Permian, Russia. Family Pachyrhinidae. Large gorgonopsids, in which the face is about as high as widp ; the orbits small (no doubt owing to the overall size of the animal) ; the temporal fossae are relatively large, the bony roof which separates them being relatively nar- row. The occipital condyle lies far in front of the posterior borders of the squamosals. The frontal reaches the orbit by a definite process between the pre- and large postfrontals. The dentition is 5.1.4 and there is no sign of a step. There is a small median groove behind the posterior nares. Differs from Galesuchidae in the anterior position of occipital condyle. This may be largely a size effect. The frontal shape seems to be common in the Tapinocephalus zone. Pachyrhinos kaiseri Broili and Schroeder 1934, U. Tapinoce- phalus zone. Family Phthinosuchidae. The number of teeth is higher than in typical Gorgonopsia. The lower jaw .shallow. The step in the dentition is little developed. The exposure of the squamosal lateral to the occipital surface is relatively very small. There is no preparietal? The pineal foramen is as posterior as possible. Phthinosaiirus Efremov 1938, Russian zone II (or III?); Phthinosnchus Efremov 1954, Russian zone I. Family Rubidgeidae. A family of gorgonopsians, the mem- 58 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY bers of which attain very large size. The skull has a relatively large and narrow snout; the upper dentition being composed of a series of very powerful incisors, usually 5 in number, a canine, and a variable number of cheek teeth, ranging from 5 to 0. The maxilla has a rounded tooth-bearing border and is very deep. With increase in size, the whole temporal region increases in width ultimately to a very great extent, and this increase in general width goes with a corresponding and disproportionately enlarged broadening of the intertemporal region. The occipital region inclines forward and in large individuals reaches the pineal foramen, which lies immediately in front of the temporal openings. The occiput is necessarily very wide and the quadrates and squamosals do not seem to extend ventrally to the level of the occipital condyle. The palate lacks a suborbital vacuity and is of standard gorgonopsian pattern. A feature which may be rather characteristic is that the immensely deep anterior end of the mandible sweeps into the lower border of the jaw. The re- flected lamina of the angular often bears a deep ridge, behind which are one, or two, deep pockets. Dinogorgon {Broomicephalus) laficeps Brink and Kitching 1953, Cistecephalus zone; Dinogorgon ondehergensis Brink and Kitching 1953, Cistecephalus zone ; Dinogorgon quinquemolaris Huene 1950, Cistecephalus zone, E. Africa; Dinogorgon ruhidgei Broom 1936, M. Cistecephalus zone ; Dinogorgon (Dracocephalus) scheepersi Brink and Kitching 1953, Cistecephalus zone; Pro- ruhidgea maccahei (pugnax) Broom 1940, ? Cistecephalus zone; Prornhidgea rohusta Brink and Kitching 1953, Cistecephalus zone ; Buhidgea atrox Broom 1938, Cistecephalus zone ; Ruiidgea hitchingi Broom 1938, Cistecephalus zone; Buhidgea laticeps Broom 1940, II. Cistecephalus zone ; Buhidgea majora Brink and Kitching 1953, Cistecephalus zone ; Buhidgea platyrhina Brink and Kitching 1953, Cistecephalus zone; fSmilesaurus ferox Broom 1948, Endothiodou zone; ?S!}ni]esanrus maccahei Broom 1948, Endothiodon zone; fTangagorgon tenuirostris Boonstra 1953, E. Africa, Lower Bone Bed ; Tigrisaurus pricei Broom and George 1950, fCistecephalus zone. Family Scylacocephalidae. Small gorgonopsians characterised by a long temporal fossa, the temporal roof being about twice as wide as the interorbital width. The snout is less than half the WATSON AND ROMER : THERAPSID CLASSIFICATION 59 total skull length, and is high and narrow. The frontal enters widely into the orbit. The lower jaw has a raised symphysial region, the canine socket being above the line of insertion of the cheek teeth. The hinder part of the jaw is unusually long and the reflected lamina deep and far forward. Scylacocephalus wafermeyeri Broom 1940, Cistecephalus zone. Family Scylacopsidae. Small gorgonopsians in which the face forms about half the skull length. The face is rounded in cross section and widens gently, passing smoothly into the orbital re- gion and so to the widest point of the skull about half way along the zygomatic arch. The interorbital and intertemporal regions are of nearly equal width. The postorbital bones are of some width and pass down to rest on the zygoma by a widened arc. The temporal fossa is longer than is usual in gorgonopsids. The occiput slopes forward; the squamosals are turned back- ward at their temporal border. There is a basioccipital with large tubera which join with those from the basisphenoid to support the cartilage surrounding the fenestra ovalis. The basisphenoidal tubera are separated by a pit but anteriorly are confluent and continued by a deep narrow ridge nearly to a very small interpterygoid vacuity. Dental formula 4-5. 1. 4 (-7?). Cyniscopoides hroomi Brink and Kitching 1953, Cistecephalus zone; fCyniscops hroomtanus Huene 1950, Cistecephalus zone, E. Afr. ; fCyniscops cookei Broom 1948, Cistecephalus zone; fCyniscops kitching i Broom 1948, Cistecephalus zone ; fCyniscops longiceps Broom 1941, Cistecephalus zone; fCyniscops ruhldgei Broom 1937, M. Cistec'e])halus zone; Galerhynchus ruhldgei Broom 1937, Cistecephalus zone; " Gorgonopsid " Watson 1913, Cistecephalus zone; Scylacops capensis Broom 1913, Cistece- phalus zone ; Sycocephalus })igendens Brink and Kitching 1953, L. Cistecephalus zone. Family Scymnognaihidae. Gorgonopsids with a comparatively wide intertemporal region, and spreading zygomata. The snout is relatively long, square-cut in transverse section, and there is a marked step from the premaxilla to the maxillary tooth-bearing border. The paroccipital process is robust, deep below the post- temporal fossa, and the condyle and floor of the braincase thick. 60 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Arcfops willistoni Watson 1914, ?Endothiodon zone ; Chiweta- saurus dixeyi Haughton 1926, Cistecephalus or V. Endothiodon zone, E. Africa; Dixeya nasuta Huene 1950, Cistecephalus zone, E. Africa; Dixeya quadrata Haughton 1926, Cistecephalus zone, E. Africa; Lycaenoides anyusticeps (minor) Broom lf)13, Ciste- cephalus zone ; Scymnognathus hohnesi Broom 1948, Endothiodon zone ; Scymnognathus major Olson and Broom 1937, Endothio- don zone; Scymnognathus parringtoni Huene 1950, Endothiodon zone, E. Africa; Scymnognathus whaitsi Broom 1912, Endothio- don zone. Family Sycosauridae. Gorgonopsians of medium size, with a face as long as the rest of the skull and rather heavily built, with a semicircular anterior end. Nostril very far above lower border of premaxilla. The postorbital bar is broad (cf. Rubid- geidae), the zygoma rather deep, the temporal fossa very short. The occiput is steeply inclined and makes a large bay with dorsal surface. The frontal is excluded from the orbital margin by a meeting of the pre- and postfrontals. There is no preparietal. Dentition 5. 1. 5. Sycosaurus hrodiei Broom 1941, Cistecephalus zone; Syco- saurus laticeps Haughton 1924, ?L. Cistecephalus zone. Gorgonopsians represented hy adequate skulls hut lacking definite family characteristics: — Aelurognathus microdon Boon- stra 1934, Cistecephalus zone; Aelvrognathus nyasaensis Haugh- ton 1926, Cistecephalus zone ; Aelurognathus tigriceps Broom and Haughton 1913, Cistecephalus zone; "Aelurognathus" minor Brink and Kitching 1953, Cistecephalus zone ; Alopecorhynchus rubidgei Brink and Kitching 1953, Cistecephalus zone ; " Arctops ' ' watsoni Brink and Kitching 1953, Cistecei)halus zone ; Cerdor- hinus parvidens Broom 1936, U. Tapinocephalus zone ; Clelandina major Broom 1948, Endothiodon zone; Clelandina rubidgei Broom 1948, U. Endothiodon or L. Cistecephalus zone ; Cyono- saurus longiceps Olson 1937, L. Cistecephalus zone; Lycaenops ornatus Broom 1925, U. Endothiodon zone; '' Lycaenops" alt i- ceps Brink and Kitching 1953, Cistecephalus zone ; Nanogor- gon gracilis Broom and Robinson 1948, ?Cistecephalus zone ; Sauroctonus progressus Hartmann-Weinberg 1938, Cistecephalus zone, Russia. WATSON ANT) ROMER : THERAPSID CLASSIFICATION 61 Gorgonopsians represented hy material inadequate for proper assignment : — " Aelurosaurus" brevirostris Broom 1948, Endothi- odon zone; Aloposauroides tenuis Brink and Kitching 1953, Cis- tecephalus zone ; Aloposaurus gracilis Broom 1910, U. Endothio- don zone; Arctusuchus tigrinus Owen 1876, ?Endothiodon zone; Ccrdognathus greyi Broom 1915, fCistecephalus zone; Cgna- rioides grimbeeki Broom 1935, U. Endothiodon zone; Cy7iiscodo7i lydekkeri Broom 1915, Tapinoeephalus zone; Cynodraco major Owen 1876, h. Endothiodon zone; Cynodraco serridens Owen 1876, fEndothiodon zone; Delpliaciognathus {AstJienognathus) paucidens Broom 1915, Cisteeephalus zone; Eriphostoma micro- do)) Broom 1911, ■Tapinoeeplialus zone; Genovum hroilii Huene 1950, Cisteeephalus zone, E. Africa; Lycaenodon longiceps Broom 1925, L. Cisteeephalus zone ; Lycosaurus pardalis Owen 1876, ? Cisteeephalus zone; Pardocephalus waUacei Broom 1948, Cisteeephalus zone; " Scy)}})w • • PLATE 2 PLATE 3 Fig. 5. A low power, mk-roscopic view of the tyincal appeaianee of the zonarv placenta of the wolverine. A maternal artery is visil)le (arrows) jnirsuino- its way toward the fetal surface of the placental hJiyrinth where it liranclies. J~, junctional zone; gs, deep glandular zone. X 9. Fig. 6. The border of the wolverine's placenta, at a point where there is a small, round hematoma as well as hemotrophe and histotrophe between the nu'mbranous chorion on the one hand, and the placental labyrinth and tlie parajdacental junctional zone on the otlier hand. The field enclosed in a rectangle is shown at higher magnification in Figure 17. The uterine lumen, mucosa and musculature are visible at the very bottom of the picture. X 11.25. PLATE 3 PLATE 4 Fig. 7. The placental labyrinth shown at high magnification, illnstrating the maternal sinusoidal capillaries lined by hypertrophied endothelial cells which rest upon a liasement membrane which is intensely stained by the periodic acid-Schiff reagent. Surrounding the basement membrane the trophoblast is visible; it is syncytial in chaiacter and contains nunun'ous ■■intraepithelial" fetal capillaries. Pale, unstained, relatively acellular stroma without any basement meml)rane, abuts the trophoblast. X 200. Fig. 8. A portion of the laliyrinth stained by Masson's trichrome mix- ture. With this staining method, the endothelium lining the maternal vessels and the syncytial tropholilast is well differentiated, but the intervening basement membrane which is selectively stained by the ]ierio(lic acid-Schiff reagent (Fig. 7) is entirely unstained. X '200. Fig. 9. A portion of the laljyrinth close to the fetal surface, illustrating the spurious presence of giant cells or syncytial tags, resulting from the tangential planes in which the endothelium of the recurrent maternal ves- sels in this region is cut. X 152. PLATE 4 PLATE 5 Fig. 10. Higher power niicroseoijit- view of the junctional and glandular zones of the wolverine's placenta. Compare with Figure 5 (./~, gz). Two entire glandular compartments and parts of two others are visible; their lumens are filled witli histotro]ih(' and tlie i)artitions between them carry tlie afferent and efferent i)lacental maternal blood vessels. X 32. Fig. 11. Illustrates a single glandular compartnieut with a simple tioor and elaborate side Avails; the walls consist of folds or leaves of stroma, cov- ered by uterine epithelium some of which is syncytially transformed and contains vesicles. X 144. Fig. 12. Another compartment, showing a tongue of fetal stronui covered l)y columnar trophoblastic cells projecting into a mass of amorphous histo- trophe in the chamber. The fetal projection is flanked on its sides by folds of vesiculated nuiternal mucosa, some which is visibly disintegrating to form histotrophe. X 144. PLATE 5 PLATE 6 Fig. 13. A section through the larger of the two minute hematomas on the internal surface of the placenta near the insertion of the umbilical cord in Figure 1. In the upper corner of the ijhotograph there is a similar lieniatoma of microscopic dimensions. X 40. Fig. 14. A section through the small, pale nodule on the internal surface of the placenta near the attachment of the large central hematoma (cf. Fig. I). This hematoma consists primarily of elaborate infolded masses of the chorion with very little actual extravasated blood in its interior. It is interpreted as being an older hematoma in which resorption is nearly completed. X 32. Fig. 1.5. A higher jjower view of part of the wall of the hematoma seen in Figure 13, showing folds of chorion dipping into the interior of the extrava- sated maternal l)lood. The low columnar trophoblastic cells covering the clKjriduic folds are phagocytic and are actively engaged in resorbing the blood. X 144. Fig. l(i. A view of an extremely minute, very recent hematoma in con- tact with the placental labyrinth at a site where the extravasation of blood occuired probably fiom neighboring maternal sinusoidal capillaries. X 100. ^^l ^v?!?<^. !?»^ 7 '^ ■« i " ■ ■' ^ " , ■■' --■ ■"■'.» ■ • ' -v' ■' ■ "^^*r ^^ *t:"';>^;-^- ';'' 14 \ Sfo^^^' 16: PLATE 6 PLATE 7 Fig. 17. A field at the border of the placenta, magnified from the area shown in the rectangle in Figure 6. The area in this figure delimited by a rectangle is similar to the one shown at higher magnification in Figure 18. The two arrows in this pliotograph indicate parts of the paraplaeental mem- branous chorion similar to that shown at higher magnification in Figure 19. X -iO. Fig. 18. A field, similar to the one contained in the rectangle in the pre- ceding figure, still further magnified, to show a fold of chorion dipping into a mass of extravasated maternal lilood. The chorionic epithelial cells are variously laden with disintegrating blood cells, pigment, granules and vacuoles. X 144. Fig. 19. Paraplaeental membranous chorion from a site similar to the ones indicated by arrows in Figure 17, again showing the columnar chorionic epithelial cells engaged in the resorption of extravasated maternal blood. X 144. Fig. 20. Glands in the paraplaeental uterine mucosa. The low surface epitheliimi lining the uterine cavity is apparent above the glands. The mucosa is visil)le at nuu-h lower magnification in Figure 6. X 144. *^%V'''*^ «>*^*^ r.- ■ ' PLATE 7 PLATE 8 Fig. 21. Umbilical cord of the wolverine, hnlf n centimeter from the umbilicus, showing paii'ed umbilical arteries and veins with the slit like lumen of the allantois between them and. below, the vitelline artery and vein. X 16. Fig. 2'2. A portion of the denser connective tissue in tlie neighborhood of the umbilical arteries, showing a rich supply of small, interstitial blood vessels which vascularize the dense, central stroma of the cord. X 160. Fig. 23. A longitudinal section through the tip of the yolk sac, the chorionic membrane and the wall of the allantois. Abbreviations: ys, yolk sac; e.r, exocoelome; cli, membranous chorion; al, allantoic cavity. X 40. Fig. 24. Th? detailed appearance of a portion of the wall of the yolk sac, at high magnification. X 260. Fig. 25. A portion of the membranous (du)rion t)verlying the yolk sac. The chorionic epithelium is hiwei- and more inactive here than in the para- placental region. X 2(')0. Fig. 26. A ptirtion of the membranous chorion at a distance from the liorder of the i)lacenta. The cells, lu'i-e, are (piite low and inactive. X 260. \ ■ V 71 21 *; - — ' ■ ',/ ■''/' /' , •.'.*»>.? #^--*^ 23 PLATE 8 Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 1 14, No. 4 THE FUNCTIONAL ANATOMY OF THE DELPHINID NOSE by Barbara Lawrence and William E. Schevill With Thirty Figttres CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM February, 1956 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 114. Breviora (octavo) 1952 — No. 51 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 5'). JoHNSONiA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 3, no. 35 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 — Published in connection with the ^Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, ( 'ambridge 38. Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volume? will be published under Museum auspices. BiiUetin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 114, No. 4 THE FUNCTIONAL ANATOMY OF THE DELPHINID NOSE by Barbara Lawrence and William E. Schevill With Thirty Figures CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM February, 1956 No. 4 — The Funcfional Anafomxi of the Delphinid Nose'^ By Barbara Lawrence and William E. Schevill TABLE OF CONTENTS Page Introduction 103 Materials and methods . 106 General description of the nasal passage 108 Detailed description of the nasal passage . . Ill Outer part: blowhole and vestibular sacs Ill Constriction: tubular and connecting sacs 113 Inner part: nasal plugs and premaxillary sacs 116 Musculature of the blowhole region in Tursiops, Stenella, and Belphinus 118 Introduction ........ 118 Pars nasalis of the maxillonasolabialis 119 Posteroexteruus layer .... 119 Intermedins layer 121 Anteroexternus layer 122 Posterointernus layer .... 123 Anterointernus layer . . 124 Profundus layer ... 126 Nasal plug muscle .... . 126 Pars labialis of the maxillonasolabialis 127 Major intrinsic muscle of the tubular sac 128 Ligament muscle 129 Minor intrinsic muscle of the tubular sac 129 The blowhole mechanism ...... 130 Observations on its functioning . . . . 130 Anatomical explanation of function . . . . 136 Contrast with mysticetes 138 Discussion of previous work 140 Effect of hydrostatic pressure on the blowhole mechanism 147 Summary 148 Acknowledgments . 149 References . . 150 INTRODUCTION No description or anatomical examination can possibly give an adequate picture of a free-swimming porpoise's breathing cycle. Nor can they convey a sense of the urgent purpose behind the iContribution uo. 79.3 from the Woods Hole Oceanographic Institution 104 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY grace and power with which it rolls to the surface and the split- second timing when it goes down. The strong rush of air as it is expelled is distinctly audible, the slight gasp of the far briefer inhalation much less so. There is a definite feeling of force be- hind this swift exchange of air that is never apparent in land mammals even during hurried or labored breathing. A thin puff of moisture more often than not accompanies a blow, especially when the porpoises are swimming fast and so speeding up the breathing cycle. In Tur slops these puffs are sometimes visible for quite a distance and frequently betray the presence of in- dividuals that are rolling low and otherwise quite difficult to see. Occasionally the spouts are larger, spreading rapidly and mush- rooming at the top, suggesting those of a humpback whale in miniature. To a casual observer all this seems something to admire rather than to marvel at, and yet here is a mammal which is keeping its ancestral dependence on air while living its entire life in water. No wonder zoologists have long been fascinated by the adapta- tions arising from this most successful compromise. Of these, one of the most obvious and yet the most baffling is the breathing of the porpoise. How does it keep the air in or the water out? Investigation of this problem was begun a long time ago ; as early as 1787 Plunter considered it in his "Observations on the struc- ture and oeconomy of whales." Since then numerous workers have reported on different aspects of the blowhole region, but none of them with sufficient accuracy of detail about the structure and interrelationship of the parts involved to make possible a correct interpretation of their function. More particularly, de- scriptions that deal with muscle layers do not give sufficient information about the thickness of the different layers, the way in which they merge at origin or insertion, or the actual eft'ect of their contraction. Furthermore, descriptions of the nasal sacs lack adequate detail on their position and that of their openings relative to the rest of the nasal passage. This is no reflection on the diligence of these early workers ; rather it is an expression of the difficulty of distinguishing these details in any l)ut fresh (or recently thawed) material, and of the importance of having many specimens available. It should also be borne in mind that the focus of much of this LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 105 earlier work has been on the possible homology of structures of the blowhole region with the nasal cartilages, bones of the nose, and facial muscles of land mammals. As a result, function has, in a sense, been speculated on rather than investigated. Such speculation, not checked on living animals, has produced a con- flicting mixture of fact and fancy, and even the most careful workers have failed to grasp some of the most important points. The purpose of our work has been to unravel some of this confusion. In this paper we have attempted to explain the move- ment of air in the outer nasal passages, how the blowhole is opened, how it closes, and how the seal is tightened. Since such explanation will be easier to follow if the myology of the blow- hole region and the interrelationships of the nasal sacs and main air passage are understood, these will be described first. We have accepted Huber's (1934, p. 131) opinion that the rostral muscles are the pa7's JabiaUs of the maxillonasolahialis and that all the layers of blowhole muscle are derived from the pars nasalis of the maxillonasolabialis. We recognize six such layers, all functionally distinct, which we have called : posteroexternus ipe), intermedins {i), ant eroext emus {ae), poster oint emus (pi). anterointcrnus (ai), and profundus (pr) (Figs. 10-17). This recognition implies a more clearcut anatomical separation than actually exists. In our cross sections the only obvious division is between pi and ai (Figs. 7-9) ; also various layers, notably pr, in- clude more than one section, but where there is no sharp distinc- tion either in origin, insertion, or direction of fibers, such subdivisions within a layer have not been considered separately. In studying the nasal sacs we have found that only the pre- maxillary sac has been consistently so identified. This term we have continued to use; otherwise we have selected appropriate descriptive names. For the sake of simplicity the descriptions are usually of the right side only ; hence the sacs, although paired, are usually referred to in the singular. In keeping with the characteristic asymmetry of the skull (Fig. 1), the passage and associated sacs on the left are all less developed than those on the right. The part of the nasal passage here considered is the distal part which lies superficial to the openings of the paired bony nares, in other words that part which lies in the soft tissues of the head between the bones of the skull and the protuberance of dense, 106 BL'LLETIX : MUSEUM OF COMPARATIVE ZOOLOGY fatty, connective tissue above the rostrum, commonly called the melon. MATERIALS AND METHODS This paper is based on a stud}^ of both dead and living ma- terial. Detailed dissections were made of two Tur slops truncatus, two Stenella plagiodon, and one Delphinus delphis, all preserved by freezing'. This kept the material in excellent condition even when partially thawed for dissection and then refrozen repeat- edly. Cross and sagittal sections were made of two fresh adult Tursiops, and sagittal sections and dissections of a fetal one preserved in formalin. Dissections and comparative observations were also made on a number of Glohicephala melaena and a single specimen of Lagenorhynchus acutus. Latex casts of the passage of Stenella and Delphinus were generously supplied by Dr. Henrj^ Kritzler. In addition, the movement of air in the passage was studied l)y pumping air up the trachea of specimens of Tursiops, Delphinus, and Stenella. Most of the observations on live animals were made during the course of four visits to Marine Studios. Marineland, Florida. Here there was ample opportunity to watch, from both above and below the surface of the water, the breathing cycles of captive Tursiops, and less frequently of Stenella. Motion pictures taken at 64 frames per second show the relation of closing of the blowhole to submerging, both in groups of animals swimming at various speeds and in two different individuals idling near the surface. Other close-up pictures taken at the same speed show, in detail that the unaided eye cannot catch, how various parts of the passage move at such times. Finally, and most important, investigations were made inside the blowhole of a live bull Tursiop.s which was kept under such observation for three days. By holding a finger in various parts of the passage (on occasion as far down as the bony nares) for extended periods between and during blows it was possible to feel both muscular and pneumatic motion. It was also possible to introduce a rubber tube and take pressure readings with a mercury manometer in different parts of the passage. The behavior of our live animal during the three days when we had him under ol)servation is of considerable interest in view LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 107 of a rather widely held opinion that porpoises will not tolerate tampering with the blowhole. This is probably based in part on Triie's observation (1891, p. 202) that stranded Tursiops at Hatteras "invariably showed signs of discomfort by lashing the tail violently" when he touched the blowhole. In contrast, our bull never thrashed around until in his death flurry, when he was apparently unconscious. The rest of the time he was remarkably passive, whether afloat or aground. This does not mean that he submitted with complete tranquillity to our investigations. His efforts to hold his nasal passage closed will be described later. In addition, as he lay on his side, he occasionally bent his head down away from the investigator, at the same time lifting it slightly from the deck, and sometimes opened his mouth a little as if gagging. This occurred particularly w^hen the plugs were pulled open by hand, allowing air to escape, or when he was probed near the entrances to the tubular and connecting sacs. These occasional efforts seemed to be to aA^oid rather than to attack the investigator. The only time he snapped was once when we attempted to muzzle him with a rope, a precaution we gave up as unnecessary. Irving, Scholander, and Grinnell (1941, p. 158), when working on the respiration of Tursiops, found them so sensitive to inter- ference Avith their breathing that on occasion individuals as- phyxiated themselves. These authors stated that "if a little water enters the nostril during inspiration, the animal often will not breathe again." This certainly was not the case with our animal, even when we ourselves opened his blowhole under water. Apparently neither the presence of water in the outer nasal j^assage nor interference with the mechanics of l)lowhole action are by themselves sufficiently disturbing to inhibit l)reath- ing. Some other factors must also be involved. This individual (length 262 cm., weight about 260 kg.) was not in perfect health when caught, as shown hj an extraordinary amount of flatulence and a rather bloated appearance. For this reason the mild behavior described above raav not be entirelv typical ; it is, however, in keeping with our previous observations on these creatures. 108 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY GENERAL DESCRIPTION OF THE NASAL PASSAGE A brief general description of the nasal passage, together with the accompanying diagrams, will help to orient the reader and enable him to understand the subsequent detailed discussion of the structure and functions of the various parts. The 'inner or bony nares pass upward through the skull as a pair of simple, somewhat rounded tubes between the cranium and the rostrum (Fig. 5). Being surrounded by bone, they are not collapsible, and they lack diverticula. Paired plugs occlude their upper ends, shutting them off from the collapsible and far more complicated superficial portion of the passage which slopes more anteriorly (Fig. 2). In this paper the term "plug" refers to the mass of tissue which occludes the upper end of the bony nares, is moved by the nasal plug muscle, and has a lateral ex- tension on the premaxillary bone. For most of its extent the distal nasal passage is paired, being divided by a cartilaginous extension of the bony septum which separates the internal nares (Figs. 1, 5, 19, 23). The blowhole, or opening through Avhieh the passage communicates with the exterior, is single, smaller than the rest of the passage, approxi- mately crescentic in outline when closed, and nearly round when wide open (Figs. 26-28). Essentially, the main air passage may be considered as made up of (1) a relatively fixed posterior wall sloping downward and backward from the posterior lip of the blowhole, and (2) a highly movable anterior wall comprised of the valve and the plugs, which are closely coapted to the posterior wall (Fig. 2). Bisecting the nasal passage in the midline a short distance within the blowhole is the cartilaginous septum men- tioned above. Anteriorly this attaches to the wall of the passage between the plugs (Figs. 5, 19) ; posteriorly it attaches to the bony wall of the passage and thence curves upward and forward beneath the posterior lip (Fig. 28). The plugs are freely movable on each side of this septum, which itself limits the extent to which the passage can he drawn open anteroposteriorly. That the apparent simplicity of a view down the open blow^- hole is very misleading is immediately discovered when the closed passage is in^•estigated. Then one finds that the overlapping posterior wall fits so snugly on top of, as well as behind the anterior wall (Figs. 2, 3) that a probe thrust down must follow LAWRENCE AND SCIIEVILL : THE DELPHINID NOSE 109 a tortuous course to reach the bony nares. As a further compli- cation there are four pairs of diverticula (Figs. 1, 20), two of which, the premaxillary and the vestibular, are broadly con- tinuous with the passage and so widen it to a diameter far greater than that of the nares or blowhole. The paired premaxillary sacs are the largest (Fig. 1) and lie immediately on top of the smooth portion of the premaxillary bone around the opening of the bony nares. They communicate with the nares internal to the plugs along the entire narial border of the premaxillae, and their superficial wall appears as an extension of the integument covering the nasal plugs (Figs. 2- 4). For the most part, the premaxillary sacs are internal to the plugs, but each has a small opening at its posterolateral corner external to the plug (Fig. 22) through which air may pass into the outer passage even when the plugs are seated. The outermost pair of sacs, the vestibular, are posterolateral to, and lie immediately within the lips of the blowhole (Figs. 10, 12) . Thej^ are highly distensible and, when uninflated, roughly half the size of the premaxillary sacs. Like the latter they are continuous with the main passage, so that it is difficult to deter- mine where the one begins and the other ends. Together, passage and sacs form a widened vestibular area whose inner limit is the transverse slit where the plugs pass beneath the overhanging posterior wall (Figs. 21, 22). This slit or constriction is of par- ticular importance, as it is here that the porpoise can volun- tarily tighten the closure of the passage. There is a considerable mass of tissue lying between the vestibular area and the dorsal surface of the plugs (that is to say, beneath the ventral surface of the vestibular sac) and it is here that the last two pairs of sacs, the iuhular and the connecting, lie embedded. These are both true sacs in the sense that they are perfectly distinct from the main air passage and connect with it only via rather small, well-marked openings. The tubular sacs are U-shaped, lie almost horizontally in the head, and on each side surround the slitlike opening between the top of the plugs and the posterior wall (Figs. 1-3). A transverse ligament beneath the posterior limb of the tubular sac helps to stiffen the posterior wall of the passage here (Fig. 20). The connecting sacs appear as small appendices lateral and internal to the tubular sacs, into which they open (Figs. 1, 20, 22). They also lie directly over the external opening 110 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of the premaxillary sac described above (Fig. 22), so that there is a direct passage for air from the outer margin of the premaxil- lary sac via the connecting sac into the tubular sac. When the tubular sac is thus inflated it pneumatically tightens the seal of the inner passage. Surrounding the passage and sacs are six layers of blowhole muscle. Omitting the small and often poorly developed inter- medins layer, the other five take origin over a wide area in the fronto-maxillary region (Fig. 23) and are found to be in two groups which insert respectively posterior and anterior to the passage, but alternate, and overlap broadly at their origin. Thus the anteroexternus partially separates the two posterior layers, while the anterior portion of the posterointernus lies between the anteroexternus and the anterointernus (Fig. 11). Where the two posterior layers are not separated by an anterior layer they tend to merge. Thus while these four layers are functionally distinct, they are by no means anatomically so throughout their entire extent. The remaining, or deepest layer, the profundus, is dis- tinguished from the anterointernus by its more intimate connec- tion with the rostral muscles and its insertion, chiefly in the melon (Fig. 16). The effect of these muscles on the sacs is inci- dental to their control of the passage ; the vestibular and pre- maxillary sacs lack intrinsic muscles. The tubular sac is unique in being provided with a weak l)ut rather extensive intrinsic musculature and the connecting sac is partially affected by a portion of this. The nasal plugs have their own distinct muscles. In summary, then, the distal passage maj' be considered to be divided into two rather wide spaces separated by a slitlike con- striction. For a more detailed description it will be convenient to deal with these three regions separately, the blowhole and vestibular sacs forming the outer part, the tubular and con- necting sacs and associated parts forming the constriction, and the nasal plugs and premaxillary sacs forming the inner part. In fundamental plan the distal nasal passage is very similar in all three forms studied, so our description of Tnrsiops trunca- tus is equally applicable to Stenella plagiodon and Delphinus delphis. The same is true of the rather general remarks on the musculature of this region. A more thorough description of the muscle layers, as well as an account of certain slight differences between the three forms, is given below (pp. 118-129). LAWRENCE AND SCIIEVILL : THE DELPIIINID NOSE 111 DETAILED DESCRIPTION OF THE NASAL PASSAGE Outer Part : Blowhole and Vestibular Sacs The blowhole is itself somewhat asymmetrically curved with its convexity directed backward. The anterior lip, or valve, slopes back to pass under the posterior, thus forming a slight de- pression which often holds a little water when the animal sur- faces. The lips are thick and non-muscular, being composed of what appears to be very fibrous, less oily blubber than that w'hich covers the rest of the animal. The posterior lip in particular is stiff and densely fibrous ; in a live animal only the margin along the middle third was found to l)e of softer, freely movable tissue. Since the lips have no muscle attachments in them, they can move only as the underlying tissue is moved. Thus the thick, fibrous mass of the posterior lip, projecting as it does well l)eyond the subjacent muscles, moves relatively little, while the anterior lij) or valve moves freely as part of the mass of connec- tive tissue on which it lies. The valve in the living animal is soft and pliable to the touch, and has a well-marked margin where it passes beneath the posterior lip. It slopes rather steeply in the midline, and is continuous below and laterally with the two j)lugs, which are separated by the septum. On either side the passage widens into a large sac, the vestibular, which when un- inflated lies lateral and somewhat posterior to the blowhole (Fig. 10). At the lips of the blowhole the smooth wall of the passage passes gradually into the more wrinkled lining of the sac with no real division between sac and passage (Fig. 13). On the under or forward side of the passage the valve passes with no sharp boundary into the anterior fold of the vestibular sac (Figs. 2, 13, 21). This fold divides the anterior part of the sac into two unequal pockets, an inner one which is shallower anteroposteriorly, and an outer one which is much deeper and communicates directly with the blowhole opening. The fold is thickest at its medial end, and here takes origin midway between th(^ margin of the valve and the jdug. Laterally it tapers to nothing just beyond the end of the slitlike opening ])etween the plug and the posterior w^all. When the passage is closed and the sac uninflated, this fold covers the slit, thrusting against the less well developed posterior fold which forms a soft yielding mass Ix^tweeu the stiff overhanging posterior lip and the almost 112 BILLETIX : MUSEUM OF COMPARATIVE ZOOLOGY equally stiff convex inner mass that interlocks with the plug (Figs. 13, 19). This posterior fold has no real pocket beneath it and appears merely as a puckering of the wall of the uniu- flated sac. Various degrees of inflation change the shape of this sac in various ways, as is shown by latex casts. When it was slightly inflated, the latex reached only that part of the sac beneath the anterior fold; when strongly inflated both folds are obliterated. The vestibular sacs lie close beneath the blubber and are separated externally from the thick integument around the blowhole by the thin aponeuroses which form part of the insertion of the outermost layers of blowhole muscle, pe, i, and ae (Figs. 10, 11). Posteriorly these sacs are tightly bound to the insertion of the underlying muscles, so that here they appear thick walled, while anteriorly they are thinner walled and more loosely held in the investing tissue. As well as enclosing the sac, the outermost layers of blowhole muscle insert internal to the dense collagenous connective tissue mass comprising the outer lips. As with the rest of the blowhole muscles, these layers are not entirely distinct from the surround- ing layers, but usually merge with them at either origin or in- sertion so that it is extremely difficult to separate them on dissection. For this reason the pull that they exert is on an area, not a particular spot, and their contraction has a general rather than a local effect. The four outermost layers, pe, i, ae, and pi (Figs. 10-13), are chiefly concerned with opening the blowhole by drawing the valve down and forward, and by pulling upward and backward on the walls of the nasal passage which lie immediately inside the rigid posterior lip. These layers fur- ther act to hold the vestibular sac in place. Pulling on them does not stretch this sac, neither does it keep it from expanding. They do however fix this sac, determining the direction in which it expands when air is forced into it from below. Furthermore, when contracting to open the outer part of the blowhole, the outermost layers of muscle exert pressure which tends to flatten the sac, thus preventing air from being impounded there as the l)lowhole closes. In opening the passage, pe and pi act chiefly on the posterior part. Pe draws up and back the outer wall of the vestibular sac and that portion of the passage which lies just inside the posterior LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 113 lip. The tendinous insertion of pi (Fig. 14) curves over the nasal bones to a deeper attachment, chiefly in the slightly de- veloped posterior fold of the vestibular sac and the tissue ventral to this, including the outer side of the tubular sac. Posteriorly pe and pi are not separated by other layers, and at their insertion are not truly distinct. Here their thick tendinous mass fills the space between the nasal bones and the nasal passage, and, where the vestibular sac lies on top of this mass, it is closely attached to it. While pe and pi have their main mass posteriorly and control the posterior part of the nasal passage, ae has its main mass anteriorly and acts on the anterior part of the passage (Figs. 11, 13). It is largely responsible for depressing the valve by pulling the tissue beneath it anteroventrally, and further helps to open the nasal passage by pulling the well-developed anterior fold of the vestibular sac anterolaterally as well as ventrally, away from its position on the top of the slitlike opening of the passage above the plugs. In addition, ae anchors the underside of the vestibular sac, so that when inflated it has to expand on top of ae. Pe, ae, and pi are all separated from the melon and the pars labialis of the maxillonasolabialis by the collagenous connective tissue in which the posterior, external fibers of this latter muscle end (Figs. 11, 13). Layer i is very thin with no bony attach- ments, and makes a connection, otherwise lacking, between pe and ae and the melon (Fig. 11). Constriction : Tubular and Connecting Sacs Although, for convenience of description, the nasal passage is here divided into three, there is no real separation between these regions. As noted above the innermost portion of pi attaches on the tubular sac, and the slitlike opening at the inner end of the vestibular area also marks the outer end of the constricted area. Surrounding the slitlike opening on each side and more or less embedded in fatty connective tissue are the tubular sacs, which are paired, asymmetrical, and U-shaped (Figs. 15, 22). The posterior portion of each of these lies between the nasal bones and this opening. Together with its surrounding muscles, the blowhole ligament, and connective tissue, this portion of the sac forms a transverse, convex band in the posterior wall of the 114 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY passage, which, when the passage is closed, fits into a concavity on the superficial surface of each plug (Fig. 3). The smaller anterior portion of the sac lies beneath the valve. The tubular sacs each open downwards into the nasal passage through a more or less horizontal slot in the posterior part near the bend of the U. This slot slopes postero ventral ly for a very short distance to open immediately dorsal to each plug and slightly, anterior to where the plug lies against the posterior wall of the bony nares (Figs. 1, 3, 17, 20, 22). Thus the only entrance to the sac is from below, and when the passage is closed this opening is external to the plugs. Mediallj^ the opening lies very close to the bony wall of the naris; laterally, owing to the slope of the skull, the opening lies a slight distance in front of this wall. Lateral to the tubular sac and deeper, lies a small diverticulum, the connecting sac (Figs. 1, 16, 20), which is separated from the former by a deep section of the major intrinsic muscle of the tubular sac, which is here poorly developed and liberally streaked with collagenous connective tissue ; the connecting sac opens downward into the nasal passage near the posterolateral corner of the na-^al plug, and upward into the end of the ventral opening of the tubular sac (Figs. 20, 22). The opening of the connecting sac into the nasal passage lies over the lateral opening of the premaxillary sac. It is Ijounded anterior^ by the blowhole liga- ment (Fig. 20), to which its ventral surface is closely attached, so that its entrance does not collapse when the passage is closed. This makes it easily possible, even when the passage is closed, for air passing out of the posterolateral corner of the premaxil- lary sac external to the plug to enter the connecting sac, and difficult for it to go elsewhere. The blowhole ligament passes from the premaxillary bone to the lateral extremity of the slit- like opening above the plugs, where in some specimens it is stiffened by a small band of cartilage ; from here a few fibers form a transverse band immediately posterior to this opening and end in the tissue beside the septum. This ligament is a very important structure that strongly anchors the commissure of the opening and makes taut the posterior wall above the plugs. Moreover it is slightly elastic, so that contraction of pi can lift it away from the dorsal surface of the plug, but when this layer is relaxed the ligament snaps back into place. Immediately beneath it and with similar orientation is a poorly developed slip of the major in- LAWRENCE AND SCHEVILL : THE DELPIIINID NOSE 115 trinsic muscle, the lio-araent muscle ; together they separate the connecting sac from the lateral edge (Figs. ]6, 20) of the pre- maxillary sac. Of the layers of blowhole muscle, three {j)i, ai, and pr) are more or less associated with this middle region of the distal nasal passage (Figs. 13-16). Pi, as described earlier, acts to some de- gree on the outer part of the passage and has, externally, a strongly tendinous insertion with anteriorly directed fibers (Fig. 13). Deeper dissection shows the direction of the fibers to be progressively less horizontal, more anteroventral, and where the innermost layers attach on the tubular sac there is no intervening mass of tendon (Fig. 14). Ai and pr have in common a rather coarse and tendinous struc- ture, attachments in front of the nasal passage, and an intimate connection with the pars labialis. Ai has as its primary function to help ae depress the valve by drawing the tis.sue beneath it downward and forward. The main mass of ai is anterior, and the thin aponeuroses which form the insertion of the outermost layers lie in part on top of the anterior part of the tubular sac (Fig. 14). Insertion of the deeper portion is more via slender strands than aponeuroses, and is largely anterior to the tiibular sac, which is firmly attached to it (Fig. 15). As the posterodorsal border of this muscle tends to follow the edge of the premaxilla, and as in this region its fibers course anteroventrally as well as medially, its contraction would not put pressure on the lateral corner of the nasal plug, although it would do so on the anterior part of the connecting sac. Pr is made up of many subdivisions which arc impossible to separate fully, but Avliich show a progressive change in the direction of the fibers (Figs. 16, 17) from the outermost, which take about the same direction as those of ai, to the innermost, which are nearly horizontal. At their origin and. except for the innermost layers, at their insertion, the fibers of pr are continuous with those of the pars labialis of the rostrum. The innermost layers curve inward anterior to the nasal plug muscle and be- tween it and the melon (Fig. 18). Functionally also, the two muscles work together to draw the mass of the melon down to- wards the skull and back against the premaxillary sacs, thus providing a relatively i-igid backing for the nasal ])liig muscles, and putting pressure on these sacs. 116 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY lu addition to these layers of blowhole muscle, there are also associated \^dth the tubular sac, and to a certain extent also with the connecting sac, two pale, ill-defined muscles, the major and minor intrinsic muscles of the tubular sac (Pigs. 15, 16) and a small slip of the former here called the ligament muscle ; these are exceedingly hard to separate from the surrounding connective tissue. The major intrinsic muscle attaches to the posterior and underside of the posterior part of the tubular sac, and runs posterolaterally to end in the fatty connective tissue between this sac and the skull ; near the bend of the tubular sac the fibers change direction as shown (Fig. 15), and pass to the underside of the sac. Anteriorly the muscle fibers become much more sparse ; the outer ones insert along the anterior margin of the tubular sac and at their origin are hard to distinguish from the long slender threads of the tendinous insertion of ai (Fig. 15). This combi- nation of scattered muscle fibers and fatty collagenous connective tissue is rather thick in this region, and, as well as attaching to the underside of the tubular sac, attaches also to the top of the connecting sac. Closely connected with the anterior part of the major intrinsic muscle is the ligament muscle, which has its origin in the fascia beneath the blowhole ligament and inserts in part in the tissue at the commissure of the passage and in part in the wall of the premaxillary sac where this attaches external to the plugs. Wrapping around the anterior part of the posterior half of the tubular sac is a similarly pale and even less well developed muscle, the minor intrinsic (Fig. 15), which arises in the fatty collagenous tissue between the sac and the nasal passage, and passes posterodorsally to insert beneath, on the anterior edge, and on top of this part of the sac. Apparently the function of both the major and minor intrinsic muscles is to help anchor and hold open the tubular sac. The ligament muscle probabh" helps also to hold open the external opening of the premaxillary sac. Inner Part : Nasal Plugs and Premaxillary Sacs The above described, non-rigid parts of the passage are shut off from the permanently open Ijony nares by paired plugs attached in the midline to a cartilaginous septum which divides the passage and limits the extent to which it can be pulled open (Figs. 19, 27, 28). The openings of the nares face anterosuperiorly so that LAWRENCE AXD SCHEVILL : THE DELPHIXID XOSE 117 the plugs lie obliquely across the passage on each side of the septum (Figs. 2, 3). This septum is roughly triangular, with the hind margin extending along the upper end of the bony septum, then up the bony wall of the nares and finally curving forward under the posterior lip; and with the front margin rising more nearly vertically between the two plugs (Figs. 5, 19). At the anterior edge of the bony nares the lower margin of each plug is closely attached to the septum; this then is the pivot on which the rest of the mass works as it is drawn anteroventrally to open the passage. At this point the plugs meet the septum almost at right angles and form a very small posterior projection on each side of it. As they slope back across the passage they project in- creasingly far, till at their upper edge septum and plugs meet at the apex of a Y (Figs. 5, 22). This allows the posterior part of the plugs to be freely movable along each side of the septum. The position and extent of the premaxillary sacs is marked on the skull by a characteristic smooth area on the premaxillae surrounding the bony nares. Lying as they do beneath the plugs, these sacs communicate with the inner passage along the narial borders of the premaxillae (Figs. 1, 23). Their anterior portion forms a well-defined pocket with no other openings. Along its peripheral margin the sac attaches to the premaxilla and then along the maxillo-premaxillary suture where its black wall is visible when pr is cut away (Figs. 16, 18. 19). Medially the sac wall is continuous with the integument of the under surface of the plugs, except posteriorly where it attaches on top of the liplike ex- tension of the plug instead of being continuous with its margin. This forms an opening to the sac external to the plugs (Fig. 22) where air which has passed into the premaxillary sac along the hony margins of the nares can pass outward into the upper nasal passage without unseating the plugs. This opening lies close beneath the opening of the connecting sac, so that when the main passage is closed, air from here reaches first the connecting sac and then the tubular sac (Fig. 20). The dorsal surface of each plug is not horizontal, but slopes somewhat ventrolaterally and is characteristically smooth-skinned and concave. Fitting into this concavity is the above-described convexity foi"med by the posterior limb of the tubular sac, the blowhole ligament, and the mass of tissue associated with them (Fig. 3). On their under surface, posteriorly, the plugs are 118 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY roughly oval and somewhat protuberant. The mass of collagenous connective tissue of which they are formed is thickest here where the plugs lie snugly against the margins of the opening of the bony nares. Anteroventrally this cushion becomes thinner and passes without a well-defined boundary into the membranous external wall of the premaxillary sac (Fig. 19). Laterally the cushion of the plugs is more sharply set off from the liplike extension which lies on top of the premaxillary bone beneath the connecting sac (Figs. 18, 19). The paired openings of the bony nares are made oval by a tough membrane (Fig. 19) stretched across the angle where the maxilla and mesethmoid meet, that is to say, the posterolateral corner. This membrane is so sitviated that air from below cannot lift the liplike extension of the plug up against the opening of the tubular sac, while it can press the medial part of the plug more tightly against the interlocking mass of the posterior lip. The paired nasal plug muscles themselves are distinct from the layers of the pars nasalis and from the pars labialis. They take origin in part lateral but chiefly anterior to the premaxillary sacs and course posteromedially to a diffuse insertion in the col- lagenous connective tissue of the cushion of the plug, and are themselves heavily streaked with collagenous connective tissue (Figs. 3, 19). Their contraction draws the plugs down and forward. MUSCULATURE OF THE BLOAVHOLE REGION IN TURSJOPS, STENELLA, AND DELPHINUS Introduction Before describing in detail the various layers of blowhole muscle (Iluber's pars nasalis of the maxillonasolabialis), a few general remarks may be helpful. The layers are in thin sheets. Those controlling the posterior part of the passage are thickest posteriorly, hard to distinguish from each other in this region, and, at their origin, extend less far forward. Obviously, also, because of the shape of the passage they are the more external layers. Those controlling the anterior part of the passage are correspondingly thickest anteriorly Avhere they tend to merge with one another, and extend less far posteriorly. Thin layers of LAWREXCE AND SCIIEVILL : THE .DELPIIIXID NOSE 119 loose connective tissue separate more layers of muscle fibers than seem to be functionally distinct. Of the distinct layers the outer- most are inserted via aponeuroses of varj'ing thickness while the inner ones have a more tendinous insertion, the individual strands often being discrete instead of in bundles. It will also be seen from the figures that while the actual muscle fibers of the super- ficial layers extend at least as far as the vestibular sac before being replaced by aponeuroses, the muscle fibers of the deep layers become increasingly shorter as the passage widens. Other interesting features are the way in which the anterior and pos- terior layers alternate with each other, as well as the fanning out of certain layers. The effect of this rather intricate arrangement is to balance the forces pulling in various directions, so that while the passage is being opened it is also anchored against too strong a pull in any one direction. The three forms studied resembled each other closely. Such differences as were found were chiefly in the position and degree of development of minor subdivisions of the various layers. For this reason Tursiops is discussed in detail and only the ways in which the others differ are noted. Some of these differences may be merely individual variation; the number of specimens dis- sected is not sufficient to show how great this is ; and the bilateral asymmetry is considerable. It should also be noted that the way in which the bones in this region overlap varies within a species so that, while the region of attachment of each layer remains fairly constant, the actual bones involved may vary. For instance, in the specimen figured the frontal does not intervene broadly between the occipital and the nasals, while in others it does. Pars Nasalis of the Maxillonasolabialis Posteroexternus layer of the pars nasalis of the maxillonasolabialis Tursiops {pe; Figs. 10-12, 23) This is the outermost of the blowhole muscles and takes origin in a broadly continuous sheet over a wide area. Posteriorly it is thick, and at its posteromedial edge along the raised margin of the naso-frontal bones it is hard to distinguish from pi which lies 120 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY beneath it. Anteriorly it tapers to a very thin sheet which is diffi- cult to separate from i. At its insertion the converging fibers overlap in a way that makes the muscle appear double. A very thin inner section is separable onty with difficulty from the main mass of the muscle and intimately connected with ae where the two meet. Origin : On the maxilla, frontal, and occipital ; where the occipi- tal forms a crest along the occipitofrontal suture, pe attaches beneath it. Anteriorly pe is difficult to separate from ae. Insertion: Posterior to the blowhole opening. The anterior part of the muscle has dorsally directed fibers which pass in a very thin aponeurosis across the vestibular sac and then continue su- perficial to the posterior part of the muscle to attach in the denser collagenous connective tissue of the wall of the nasal passage just internal to the stiffened, crescentic posterior lip. The pos- terior part of the muscle has fibers that run anteromedially, passing deep to the anterior part to insert in the collagenous connective tissue of the posterior Avail of the nasal passage. It also inserts on the vestibular sac. When this is flattened and uninflated these attachments are along the posteromedial mai'gin and merge with similar attachments of the posterior part of ae. The anterior part of pe is intimately connected with the loose connective tissue investing the vestibular sac, chiefly over its superficial surface. These attachments of pe on the vestibular sac merge anteriorly with similar attachments of i which here lies internal to it. Funciion: The posterior part when contracted pulls back the posterior wall of the blowhole and the movable, medial part of the posterior lip ; more laterally it draws up and back the pos- terior wall of the vestibular sac, including its posterior fold. The attachments on top of and along the posteromedial margin of the vestibular sac are such that contraction of this layer does not stretch the sac, while the aponeuroses which lie on top of the sac help to flatten it when the blowhole opens. Delphinus Resembles Tursiops closely except that the anterior portion of pe has its origin slightly more posteriorly and the insertion of this part forms a more conspicuous transverse band across the insertion of the posterior part. Also the anterior margin of pe LAWRENCE AXD SCHEVILL : THE DELPPIIXID XOSE 121 and posterior margin of i are more distinct. Stenella The fibers of the anterior and posterior sections do not cross at their insertions as in Tursiops and Delpliinns, bnt converge evenly. Tn StcneUa, pe differs further in having a more distinct inner layer which has, however, the same origin and insertion as the outer layer and so is here considered as part of pe. Intermedins layer of the pars nasalis of the raaxillonasolabialis Tursiops (i; Figs. 10, 11) This is a small, not very distinct, anterodorsally directed muscle which lies between pe and ae. Origin: Beneath pe from the connective tissue sheath of ae. Insertion: In the fatty, fibrous mass of the melon and on the an- terior part of the superior surface of the vestibular sac (Figs. 10, 11). Its insertion, via a thin aponeurosis, on the vestibular sac lies beneath and merges with that of the most anterior part of pe. In front of the vestibular sac its fibers merge with those of ae to attach via strong aponeuroses in the connective tissue of the valve at the commissure of the blowhole, as well as beneath the valve more medially. Those anteriormo.st fibers which insert in the melon are quite distinct from the underlying layers, and it is this distinctness which is the principal reason for recognizing this layer as a separate one. Function: Forms the main connection between the outer layers of blowhole muscle and the melon. Delphinus Layer / is broader and better developed than in the other two genera, with fibers directed more anteriorlv. Posteriorlv a few fibers originate directly from the maxilla, and the attach- ments of this layer on the superficial aspect of the A'estibular sac extend further posteriorh' and nearer the ventral margin of the uninflated sac. Stevella Differs chiefly in having the fibers directed slightly more dor- sally and less anteriorly than in Tursiops, so that their insertion in the melon is nearer the bloAvhole ; further, part of the origin is from the maxilla via a very thin aponeurosis. 122 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Anteroexternus layer of the pars nasalis of the maxilloiiasolabialis Tursiops {ae; Figs. 10-13, 23) While the main mass of this muscle is anterior, it is entirely distinct from the melon and the pars labialis of the maxillonaso- labialis, from which it is separated by a thin fibrous sheet. Origin: Chiefly on maxilla, with a thin extension onto the frontal. While at its origin it is largely distinct from pe, it tapers pos- teriorly to a thin sheet of connective tissue which is difficult to separate from this layer. Anteriorly its connections with the two layers internal to it are equally close. Here at its insertion on the maxilla it merges with pi and anterior to this a few fibers are inseparable from ai. Insertion: Via strong aponeuroses chiefly in the dense, colla- genous connective tissue beneath the valve from the commissure of the blowhole almost to the midline, and continuously from this onto the anterior fold of the vestibular sac. In addition, thin aponeuroses attach over the vestibular sac. Anteriorly these are distinct; posteriorly they merge with similar layers of pe. In effect, pe, i, and ae enclose the vestibular sac in thin aponeu- roses, pe and i attaching more to the margin with only a few sheets passing to the outer and inner sides. As described above, the insertion of i merges in part with that of ae; the insertions of ae and pi are, however, entirely distinct. Function: This layer controls quite a large area. Contraction depresses the valve by pulling the tissue beneath it and as far internally as the beginning of the separation of the plugs ; it also draws the anterior fold of the vestibular sac forward, laterallv, and down. Counteracting this forward pull on the vestibular sac, the posterior, less elastic part of the layer helps to hold the sac in place. The internal attachments of this muscle are such that the sac must expand on top of, rather than underneath it. Dclphinns In general ae resembles that of Tursiops. It differs in having a very thin l)ut distinct inner layer attaching to the underside of the vestibular sac and merging with pi posteriorly. This in Tursiops is probably represented liy a tenuous layer of con- nective tissue. LAWRENCE AND SCHEVILL : THE DELPIIIXID NOSE 123 Stenella Here there seems to be a more obvious separation of ae into two parts, the second being posterior as well as internal to the first. Together they perform the function of ae in Tursiops and their joint origins and insertions differ only in a more extensive attachment of the posterior part on the posterior margin of the vestibular sac and in a closer connection with pi at the maxil- lary origin. Posterointernus layer of the pars nasalis of the maxillonasolabialis Tursiops (pi; Figs. 11-14, 23) Although this layer has a very wide origin, its greatest de- velopment is posterior. The actual muscle fibers are peripheral; beneath the vestibular sac and superficial to the plugs these fibers end in and are replaced by thin aponeuroses. Origin: On the maxilla and frontal with a small area on the occipital. At its origin it is thickest posteriorly, covering a much wider area than the outer layers. In the region where the maxillo- frontal suture bends upward it tapers to a thin layer without muscle fibers. Anteriorly, it l)ecomes more muscular and again somewhat thicker. At its anteriormost attachment on the maxilla it is hard to separate from ai: it also merges with ae. From this margin thin layers of tissue merge with the strands at the end of the pars labialis of the maxillonasolabialis. Posteriorly where it arises beneath pe it is distinct from this layer. Insertion: Its principal insertion is internal to pe in the posterior wall of the nasal passage beneath the vestibular sac, as well as in the shallow posterior fold of the sac. This insertion is via a thick, shiny, tendinous mass, whose fibers trend anteroventrally to fill the space between the nasal bones and the passage, and which is closely attached to the underside of the vestibular sac, particularly posterior to the slitlike opening of the main passage between the dorsal surface of the plugs and the posterior wall. Where the tendinous insertion of pe lies on top of that of pi, the two are ditfieult to separate. Deeper dissection shows the fibers of pi to slope progressively more ventrally, less anteriorly, to a less tendinous insertion on the tubular sac. Additional lesser attachments are via aponeuroses to the underside of the vestibular 124 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY sac in the region lateral and slightly anterior to the slitlike open- ing between the dorsal surface of the plugs and the posterior wall ; other similar sheets extend forward between ae and ai to merge in the melon with the insertion of the pars labialis of the maxil- lonasolabialis. These attachments are of very secondary im- portance, and most of the muscle, even the anterior part, converges to insert behind the nasal passage, as described above. Function: Contraction of this muscle helps pe to draw back the posterior fold of the vestibular sac; it also (and this is very important) helps to open the slitlike passage between the pos- terior wall and the dorsal surface of the plugs. This is accom- plished by pulling up and back the convex mass including the blowhole ligament and the tubular sac which lies in the con- cavity of the plugs. The anterior insertions of this layer help to counteract this powerful backward pull and, together with ae, prevent the contraction of the very strong posterior section from distorting the entire passage. A further effect of the contraction of pi is to put pressure on the posterior part of the tubular sac, which would force out air trapped here. Delphimis Differs from Tursiops and SteneUa in having the anterior attachments reduced to a very few poorly differentiated strands which pass to the posterolateral commissure of the nasal passage beneath the vestibular sac. Further, at its origin it tapers gradually from a thick posterior portion to a thin anterior, but lacks a pronounced thinning midway. At its origin anteriorly it is also distinct from ae, though it merges with ai. Stenella Two very distinct parts of this layer were found, the external of which differs from Tursiops and Delphinus in having the anterior fibers directed more dorsally and with better developed attachments anterior to the nasal passage. The thicker internal layer controls the posterior wall of the passage beneath the ves- tibular sac. Antei-ointeruus layer of the pars nasalis of the maxillonasolabialis Tursiops {ai; Figs. 11, 13-15, 23) This is a very thin tendinous layer, slightly thicker anteriorly, LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 125 which could possibly be considered as merely a very distinct part of pr. Along its anterior and posterior margins it is closely connected with pr, but otherwise the two layers of muscle fibers are quite distinct, the boundary being marked by strong shining tendons connected by thinner sheets of connective tissue. The actual muscle fibers are chiefly lateral to the maxillo-premaxil- lary suture, the posterodorsal ones following around this rim. Origin: On the maxilla and frontal internal to pi and anteriorly internal to ae as well. A few of the anteriormost fibers merge with pi. At its origin anteriorly it merges with the pars labialis of the maxillonasolabialis so that the two are impossible to separate. Insertion: Anterior to the nasal passage in the dense collagenous connective tissue beneath and lateral to the valve. The strong transverse sheet of its insertion shows many slender tendons which branch and vanish and are hard to separate from similar insertions of pr. In relation to the tubular sac, to which it is firmly attached, the insertion is in tissue which lies in part on top of this sac, in part anterior to it. The fibers converge towards this area from their rather wide origin, the most posterior ones assuming an anteroventral direction to do so. As the main mass passes dorsally anterior to the slitlike opening between the dorsal surface of the plugs and the posterior wall, it lies on top of the tubular sac; laterally it only partly covers this sac. At its in- sertion anterior to the tubular sac, some of its fibers are hard to distinguish from those of the major intrinsic muscle of the tubular sac. Function: Helps depi-ess the valve by pulling downward and forward on the tissue beneath it, and puts pressure on the an- terior part of the tubular sac. Delphinus Differs in having a better developed posterior section, with origin extending as far as the nasal bones, and some of its in- sertion via thin aponeuroses in the area posterolateral to the slitlike opening of the passage between the plugs and the posterior wall. Anteriorly it also differs in being more closely connected with ae at its origin. Stenella This layer does not differ significantly from that of Tursiops. 1 26 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Profundus layer of the pars nasalis of the maxillonasolabialis Tursiops {pr; Figs. 16, 17, 18, 23) This muscle is composed of a number of sections which are separated with difficulty. The fibers of the outermost ones have approximately the same direction as those of ai. The deeper ones are directed progressively more anteriorly (less dorsally). The outer layers of pr are inseparable both at their origin and their insertion from the pars labialis of the maxillonasolabialis. The innermost layers are distinct at their insertions only. This muscle is the most tendinous of all the blowhole muscles with numerous small bipinnate sections. The tendons of these sections converge to form strong strands which divide again to insert in the melon. Origin: Beneath ai. The inner layers arise on the maxilla and from the strong, thick, lateral margin of the premaxillary sac. Insertion: In the fibrous fatty mass of the melon anterior to the tubular sac and its major intrinsic muscle. The innermost layer forms a more or less transverse band across the attachment of the nasal plug muscle, separating it from the pars labialis of the maxillonasolabialis. Function: Draws the melon stronglj' down towards the premax- illa. This flattens the premaxillary sac and also makes a firm mass towards which the nasal plug muscles can pull. The band across the base of the nasal plug muscle helps to limit its motion here. Delphinus Does not differ significantly from Tursiops. Stenella Does not differ significantly from Tursiops. Nasal Plug Muscle Tursiops (Figs. 2-4, 8, 9, 19, 23) This muscle is entirely independent of the six layers of blow- hole muscle. It is abundantly streaked with collagenous con- nective tissue which becomes more and more concentrated until it finally replaces the muscle entirely in the pad of the plugs. From its origin, it takes a posteromedial direction. LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 127 Origin: Chiefly on the premaxilla anterior to the premaxillary sae, with a few fibers arising in the connective tissue band along the margin of the premaxilla lateral to the sac. Insertion: In the dense collagenous connective tissue of the plugs. Function: To withdraw the plugs from the openings of the bony nares. The diagonal direction of the fibers and their strong medial insertion makes the dorsomedial parts of the plugs the most movable. This motion can be seen in photographs taken at 64 frames per second. Delphinus As in Tursiops. Stenella As in Tursiops. Pars Labialis of the Maxillonasolabialis Tursiops (Figs. 11-13, 16-17, 18, 23, 25) This large muscle lies anterior to the layers of blowhole muscle and the nasal plug muscle. It is very coarse and mixed with numerous heavy collagenous connective tissue fibers, espe- cially externally. Here the muscle fibers are directed laterally and posterolaterally ; deeper dissection shows a gradual shift in direction, first to more vertical and then to anteromedial. The muscles on each side, although for the most part separate, are connected by a few fibers anteriorly, and behind this by a poorly developed transverse band of connective tissue beneath the melon. Origin: From the maxilla ; on the premaxilla it merges gradually with the tissue beneath the melon. The origin of the deeper por- tions is not distinct posteriorly from that of the deeper layers of blowhole muscle. Insertion: Externally the more ventral fibers end in the fibrous connective tissue between the upper lip and the melon, the more dorsal in the melon itself and in the fibrous sheet of tissue separating the melon from the outer layers of blowhole muscle. Internally the fibers insert in the melon. Function: Helps keep the melon under tension and makes a strong connection between it and the rostrum. Delphinus As in Tursiops. 128 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Stenella As in Tur slops. Major Intrinsic Muscle of the Tubular Sac Tursiops (Figs. 15, 16) This is a poorly developed muscle with a strong mixture of connective tissue, particularly in its anterior part. It merges with and vanishes in the adjacent tissue to such a degree that it is extremely difficult to determine its limits. Anteriorly it is thick, with the muscle fibers becoming sparser as one dissects deeper. The direction of the fibers is difficult to determine ; pos- teriorly they run anteromedially and gradually change direction to run posteromedially in their anterior part. Origin: In the fatty, collagenous connective tissue between the tubular sac and the skull posteriorly and laterally. It also at- taches in a thick mass of this same tissue on top of the connecting sac. Anteriorly it intermingles with the connective tissue in- sertion of ai. Insertion: On the margin and underside of the tubular sac, the anterior part being attached more on the underside and less on the margin. d' Funciion : The posterior part helps to anchor and hold open the tubular sac ; the part lying between the tubular and connecting sacs may also help to separate the walls of the latter. Belphinus The structure of this muscle was exceedingly hard to deter- mine here. The anterior part in particular was not clearly separable from the surrounding layers. S^tenella As in Delphinus, this muscle seems to be largely posterior. It diff'ers from that in Tursiops in having the fibers run antero- medially to insert on the tubular sac. Thus while the anchoring effect is apparently the same, the direction of the fibers is dif- ferent. It should be emphasized again that this muscle is poorly developed and hard to trace. LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 129 Ligament Muscle (not figured) Tursiops This poorly developed band of muscle fibers is apparently an internal slip of the major intrinsic muscle, with which it is closely connected. It lies beneath and in front of the blowhole ligament, on top of the connecting sac. Origin: Beneath the blowhole ligament. Insertion: In the tissue near the commissure of the slitlike open- ing above the plugs. Comments: Failure to identify this slip in Stenella and Del- phinus could easily have been due to the great difficulty of separating it from the rather formless mass of connective tissue and muscle fibers in this area. A special effort to determine its extent in Lagenorhynchus showed it to be distinguishable from the major intrinsic muscle only by the direction of the fibers. In Lagenorhynchus the ligament muscle, as well as inserting near the commissure, has a few fibers that insert in the outer wall of the premaxillary sac where this latter overlaps or lies on top of the lip of the plug. It is possible that this helps open the passage here. Minor Intrinsic Muscle op the Tubular Sac Tursiojjs (Fig. 15) A very poorly developed muscle wrapping around the anterior part of the posterior half of the tubular sac. Origin: In the fatty collagenous connective tissue between the tubular sac and the nasal passage both internal and anterior to the tubular sac, so that the fibers trend dorsoposteriorty. Inserfion: Weakly around the anterior part of the posterior half of the tubular sac. Function: As with the major intrinsic muscle of the tubular sac, this helps to anchor and hold open the tubular sac. Delphinus Not found here, possibly due to state of preservation. Stenella Relatively better developed here than in Tursiops, and passing across the tubular sac near the bend to merge with the major intrinsic muscle of the tubular sac. 130 BTTLLETIN : MUSEUM OF COMPARATTV^E ZOOLOGY THE BLOWHOLE MECHANISM Structure must be understood in order to understand func- tion, but the detailed dissections described above do not by themselves suffice to explain the blowhole mechanism of por- poises. The additional information obtained from experiments on living and freshly killed animals has solved a number of controversial problems which have baffled zoologists in the past. Observations on its Functioning All of the evidence agrees in showing that the passage is opened l)y muscle action, closed by the natural elasticity of the tissues, and the seal is tightened not by muscle action but by forcing the air back up the passage into the tubular sac. Inflation of this sac makes a very efficient seal where the passage narrows to a slit beneath the vestibular sac so efficient that internal to this the porpoise can keep its bubble, or internal air, at greater than ambient pressure. It makes a less efficient seal at the blow- hole entrance, where it is possible with small effort to separate the lips. When the passage opens, the plugs, the anterior fold of the vestibular sac, and the valve move so nearly simultaneously that pictures at 64 frames per second fail to show any differential sequence of events (Fig. 26), and we could not detect any per- ceptible lag between the first movement and the outward rush of air. When the passage closes it is possible, again in slow motion pictures, to see the plugs falling into place first, next the anterior fold of the vestibular sac, and finally the valve (Figs. 27, 28). This opening of the passage is in no sense the separating of the lips of an orifice ; it is the drawing apart of masses of tissue which, when undisturbed, lie pressed closely together. Most of the motion is in the anterior part of the passage, the great wedge of tissue formed by the valve, the folds of the vestibular sac, and the plugs being drawn down and forward away from their ])osition beneath and against the posterior wall. Although the posterior part moves relatively little, it is possible to see the mid-part of the crescent drawn back to widen the blowhole aperture (Fig. 26) and to feel the wall inside lifted up and back from its position on top of valve and plugs. LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 131 Within the vestibular sac and well behind the opening of the main passage, it is possible to feel a very strong contraction of muscle layer pe between the outer wall of the vestibular sac and the blubber as the blowhole opens. Deeper than this, individual action of the muscle layers cannot be felt, although it is easy to feel the swift separating of the actual walls of the passage. It is, however, apparent from their arrangement and the relative strength of their parts as well as from pulling on them by hand, that pe and pi control the posterior wall of the passage, drawing the lip slightly back and lifting the wall of the passage, including the convex mass, away from the valve and dorsal surface of the plugs. Layers ae and ai act on the anterior mass external to the l)lugs, drawing the valve and anterior fold of the vestibular sac forward and down. Finally, the nasal plug muscle withdraws the plugs from their position across the bony nares. The nasal plug muscle attaches in the cushion of the plugs, and of necessity it is that part of each plug which lies against the posterior wall beside the septum which must move the most to open the passage. The liplike, lateral extension of the plugs is moved only indirectly, and very little, as is shoAvn in motion pictures. This shows less clearly in the photographs, but in Figure 26 this part of the plug shows near the back corner of the otherwise widely opened passage. There is also very little motion at the lower end of the plugs where they attach on the cartilaginous septum close to the anterior margin of the bony nares (Fig. 19). For this reason, withdrawing them does not open the premaxillary sac, nor can the plugs slide forward into the sac. Instead, the plugs hinge here and their forward move- ment makes a distinct and perceptible midline bulge in the melon, which at the same time seems to flatten on the sides. Apart from this bulging of the melon when the blowhole opened, no other muscle action was seen or felt through the thick layer of blubber which coA'ered the Tursiops head to a depth of about two-thirds of an inch. All of this happens very quickly, and during a normal blow the main column of air passes directly and with considerable force to the outside. There is no preliminary inflation of the vestibular sac, as determined by manometer readings, though there is here a verv slight rise from 2 mm. of mercurv to about 132 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 8 mm. during a blow, with readings in the bony nares going up to at least 30 mm. at such times. This rise is part of the general rise in pressure at exhalation, and one sequence of pictures showed a slight bulge over the vestibular sac at this time. On inhalation the pressure drops, and the same sequence then showed a flattening of this area. The pattern of a blow is for the exhala- tion to take longer than the inhalation. This can best be told by ear, as there is no visible change in the widely opened passage when exhalation stops and inhalation begins. The timing of events during the period between the first opening of the blow- hole and its complete closing varies. In eight blows of a single individual loafing near the surface, it was found that the per- centage of time for opening the blowhole averaged about the same as that for closing, though sometimes the one was swifter, sometimes the other, and on two occasions the blowhole seemed to widen slightly after it was apparently fully open. These blows were analyzed as follows : Total iloic .63 sees. .77 .69 .77 .75 .74 .78 .47 That the opening and closing took so nearly the same time is especially interesting in view of the fact that only in closing is there a clearly visible sequence of events (Figs. 27, 28). At this time the plugs fall in place first, then the folds of the vestibular sac, and last the valve. More often than not the closing appears asymmetrical with, in two porpoises at least, a lag on the left side. As is often the case with aquatic mammals (HoAvell 1930, p. 87), the closed position is the involuntary one. In the forms studied this is dependent on the great elasticity of the tissues. The posterior lip presses on top of as well as behind the anterior, so that AA'hen it is pulled up by hand and released it falls back quickly in spite of the absence of any muscles to draw it in this Opening Closing 17.5% 12.5% 16.3 20.0 19.0 18.2 14.2 14.3 16.6 14.6 14.9 19.2 14.0 20.0 20.0 16.6 LAWREXCE AND SCHEVILL : THE DELPHIXID XOSE 133 direction. Similarly, the valve when depressed by finger im- mediately lifts into place when the finger is removed. A finger inside the vestibular sac during repeated blows detected no muscle action between the sac and the blubber to pull the lips of the blowhole together, nor does anatomical investigation show any muscle capable of this action. Air is not retained in the vestibular sac at the end of a blow, as repeated investigation of the live animal showed. ^Manometer readings showed no pressure here at such times, and opening only the vestibular sac by sepa- rating the lips of the blowhole right after an inhalation never resulted in the escape of air, even when this was done under water. Beneath the vestibular sac, the passage narrows, as pre- viously described, to a transverse slit between the top of the plugs and the posterior wall of the passage. The soft mass of the anterior fold of the vestibular sac covers this slit so com- pletely that considerable probing is needed to get into the passage beneath it. In the posterior wall internal to this slit, the elastic blowhole ligament and associated tissues form a taut convex mass which is difficult for the finger to lift from the top of the plug, and which immediately snaps back into place when released. Deeper than this it is possible further to open the passage by unseating the plugs manually. This does not neces- sarily result in the escape of air, indicating that at such times, anyhow, the animal's internal air is at no greater pressure than his surroundings. Attempts to open forcibly different parts of the nasal passage showed that when the Tursiops resisted it did so by inflating the tubular sac. This resistance could be felt strongly internal to the vestibular sac and weakly at the lips of the blowhole. Within the vestibular sac over the posterior part of the tubular sac, an obvious swelling could be felt at such times, and when a finger was forced down the slit between the plugs and the posterior wall a tightening could be felt on each side of the finger. Further, when the pOrpoise was thus resisting, manometer readings in the lateral margin of the premaxillary sac showed sharp rises in pressure, measured in millimeters of mercury above ambient (^ 0 mm.). At these times the pressure fluctuated irregularly, the usual pattern being for the animal to hold pressures of about 6-10 mm. and sometimes as high as 20 mm. between blows, Avith 134 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY occasional rapid excursions higher. The greatest pressure re- corded at such a time was 124 mm., though peaks were more often between 40 and 60. On the inhalation the pressure dropped to about 0. Also at times of resistance there was often a slight but appreciable bulging of the melon somewhat in front of the valve, as might be caused by inflation of the premaxillary sac. With the manometer tube at the entrance to the tubular sac, similar changes in pressure were recorded. These were somewhat more erratic, probably because of leaks around the rubber tube, and reached as high as 86 mm. ; as with the premaxillary sac, the pressure dropped at the end of each blow, seemingly on the inhalation. Clearly, efforts to hold the blowhole closed involved forcing air back up the passage under considerable pressure ; unless the plugs were unseated by hand, or the walls of the pas- sage near the entrance to the connecting sac were separated, this air did not escape but was retained internal to the vestibular sac. More often than not the porpoise made no perceptible re- sistance to our explorations of the vestibular sac, as he almost always did when we tried to push further down the passage. On the not very frequent occasions when the porpoise resisted separating the lips of the blowhole, a lifting of the cushion of the valve as by inflation of the tissue beneath it could be felt, sometimes accompanied by what seemed to be an inflation of the premaxillary sac. The lifting was immediately external to the anterior part of the tubular sac and made a visible depression along the hinge line of the valve. This, however, was not strong enough to prevent the valve from being forced do\^'n by hand with very little effort. When this was done with the blowhole submerged, Avater ran into the vestibular sac. At the next blow, the water was ejected. That this use of the vestibular sac as a water trap is not acci- dental was suggested by many pictures of swimming porpoises in which one frame shows the blowhole well open with the water at the margin, and the next frame shows the blowhole submerged. This evidence was later confirmed bv a long series of similar observations of another porpoise floating at the surface in a confined space. Here sequence after sequence showed the animal submerging with the blowhole open (Fig. 29). As it went under, the water trickled in at the forward corner of the blowhole, on either side ; if it delayed closing for .05 to .06 second, the trickle LAWREXCE AXD SCHEVILL : THE DELPHIXID XO^E 135 spread to a stream extending around the rim of the opening to the midline in back. At such times there was often a visible spurt of air and water as the animal submerged while closing the blowhole (Fig. 29), suggesting that if there is any danger of the animal flooding itself, it forces air back up the passage at the end of an inhalation. The spray was particularly conspicuous if water had l)een allowed to trickle in along the midline, that is to say, posteromedial to the entrance to the vestibular sac. Water leaking in here could not be trapped in this sac, and the only way to get rid of it would be to blow it out. Only part of the water was ejected in this way. On the subsequent blow, a thick spray was apparent before pictures showed a visible opening of the passage (Fig. 30). When the spray was fairly well dissi- pated the blowhole began to open visibly ; in seven sequences this delay varied from .05 to .12 second, depending apparently on the amount of water blown out. This seems to indicate that the animal may force air up the passage, as explained below, Avithout retracting the plugs, while there is water in the vestibule. This would eliminate the danger of water making its way down into the bony nares past the plugs when the animal pulls them. If this should happen it could endanger the larjmx or the .Eustachian tube, as the porpoise has no mechanism for getting rid of water once it is in the inner nares. The larynx projects into the back of the bony nares and is held in place by a sphincter muscle in such a way that a shallow trough is left around the two, near the bottom of which is the opening of the Eustachian tube. We do not see how water accumulating here, ventral to the opening at the end of the larynx, could either be blo^vn out or ejected by muscle action. The mechanism for closing the blowhole is not strong enough to retain any volume of air in the vestibular sac under much greater than ambient pressure. Small (|uantities of air can, how- ever, be trapped in this sac beneath the folds. This was demon- strated in two Avays. A small amount of latex blown into tho vestibular sac from the inner passage will fil] the space beneath the folds, but not reach the outer part. Experiments in pumping air up the nasal passage of a Delphl)U!s soon after death show- that after the tuliular sac has been inflated, slightly more pressure will inflate the vestibular sac with scarcely anj^ air escaping from the blowhole. AiDparently, undei- moderate pressin-(\ air from 136 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY below will lift the anterior fold up against the posterior lip, thus tightening the seal, which helps to trap air in the inner part of the vestibular sac. "When internal pressure is released, the sac deflates slowlv, and it is difficult to force out the residual air bv hand pressure. Similarly in a dead Lagenorhxjnclius where air was found in the tubular sac, this air could be forced into the vestibular sac when the anterior part of the tubular sac was squeezed by pulling ai. When this tension was relaxed the air moved back into the tubular sac. It did not escape at the blow- hole nor did it find its way past the plugs and out at the cut end of the trachea. These two experiments seem to show that once the tubular sac is inflated, as in the first case, it is difficult to force air past the slitlike opening which it constricts. If, on the other hand, the tubular sac is squeezed more or less empty, as in the second experiment, air passes more easily through this slit in either direction. The fact that air returned to the tubular sac when pressure was released would seem to indicate that this sac ordi- narily remains partly open. Anatomically, particularly at the bend, this is possible. Of further interest is the presence of observable amounts of air in the outer passage of this individual only, among the various ones dissected. Since the animal was harpooned and lanced, and died after considerable flurry, it seems very likely that this tightening of the seal of its air passage was part of its death struggle. Anatomical Explanation of Function The observations on the various changes that take place in the outer nasal passage during and betAveen blows are both con- firmed and explained by the accumulated anatomical data. The absence of muscles for closing the blowhole was striking. No muscles were found on top of the vestibular sac with attachments near the corner of the crescent which could possibly draw the posterior lip forward. Instead, the two large layers, pe and pi. attaching posterior to the passage, have medial attachments and exert a strong posterior pull. Similarly the wide spi-ead of the vestibular sac and the position of the plugs make it impossible for muscle action to lift the valve up and back against the pos- terior lip. The two layers, ae and ai, which attach in front of the LAWRENCE AND SCHEVILL : THE DELPHIXID NOSE 137 passage beneath the valve, pull strongly downwards and for- wards. In each instance the outermost layers attach well beneath the thick layer of blubber and connective tissue which surrounds the orifice, far too deep to draw the edges together in any way. While elasticity of the tissues closes the passage, the normal watertight fitting together of its parts is ensured by the texture of the surfaces and consistency of the masses involved. In the posterior wall there are two stiffened and very smooth surfaces. The more external is the underside of the crescentic posterior lip, which closes against the equally smooth anterior wall where it slopes posteroventrally from the valve. The more internal is the convexity formed by the posterior part of the tubular sac and the associated muscles and connective tissue, and stiffened by the blowhole ligament. This, too, has a very smooth surface which lies in the no less smooth concavity of the external surface of the plugs. Between these two stiffened surfaces is the soft wrinkled mass of the two folds of the vestibular sac. Thus be- tween two pairs of beautifully conforming surfaces there is a gasket-like cushion that can be distorted to fill any crack. The pneumatic closing action of the tubular sac is also easily understood if the structure of this I'egion is examined in detail. As is most readily seen in sagittal section, the plugs lie diagonally across the upper end of the bony nares in such a waj' that air under pressure from below will push the posteromedial part of the plugs more tightly against the surface dorsal to them, thus hindering the direct upward passage of air while at the same time lifting the lateroventral part of the plugs to open the entrance to the premaxillary sac so that air can pass in this direction. While the cushion of the plug is thus lifted by air forced up the bony nares, the lateral extension is not so affected, lying as it does on top of the protecting diagonal membrane and the premaxillary bone. This means that this part of the plug is not pushed up against the entrance to the tubular sac, which thus remains open while the main passage closes more tightly. From the premaxillary sac air passes into the connecting sac, whose opening lies immediately above the lateral external opening of the premaxillary sac. From here the direct upward passage of air takes it into the tubular sac, whose intrinsic muscles are suited to open it, facilitating the passage of air in this direction. As described above, this inflation of the tubular sac tightens the 138 BULLETIN' : MUSEUM OF COMPARATIVE ZOOLOGY seal of the passage. When latex was blown up the passage of a Delphinus, a concentration of it was found in the lateral edge of the premaxillary sac continuous with the latex in the bony nares and that lying external to the lip of the plug. This shows how air from the premaxillar}' sac thus passes externally to the plugs into the connecting sac. There is no evidence that hydrostatic pressure is important in sealing the blowhole. Animal tissue differs little from water in its compressibility; so, once the outer passages are emptied of air, sea pressure could be used in this way only if the air in the bony nares were at less than ambient pressure, and the por- poise's voluntary efforts to tighten closure of the passage involve, as we have seen (pp. 133-134), a rise to greater than ambient. This is surely the easiest arrangement, for achieving an intrapul- raonary pressure less than ambient would involve expanding the thorax, which would be difficult at any but the shallowest sub- mergence. To put it differently, this postulates the animal working against its medium (the water) to tighten the closure of its nasal passage when it is organized to do this much more easily by working with the medium. It is further significant that in relying on an increase in internal pressure to tighten its seal, the porpoise has perfected an arrangement which allows it to retain its internal air under greater than ambient pressure. Contrast with Mysticetes The reasons for these complications in the porpoise's outer nasal passage are not completely understood. Aquatic mammals in general develop some mechanism for closing the nostrils. In delphinids this is particularly elaborate, far more so, for in- stance, than in their relatives the baleen whales. Both groups are faced with many of the same problems, and the different ways in which they have solved these must inevitably depend on the variation possible in each group. In the mysticetes. or baleen whales, the nasal passage is much less modified than in the odon- tocetes. The nostrils are still paired, the passage is more nearly horizontal, and part]y roofed over by the nasal bones. All this has made possible a comparatively simple method of closing. In one form dissected. Balaenoptera acuto-rostrata. which is prob- ably fundamentally similar to the others, the nares are closed LAWRENCE AND SCHE^^LL : THE DELPHINID NOSE 139 by two large plugs moA'ing on each side of a stiff median septum. These pass beneath the posterior wall which is stiffened by carti- lage. Posterolateral to this opening of the inner passage and extending anteriorly are shallow diverticula which probably form the basis for Kernan's somewhat misleading and oft-quoted reference to "spiral folds" as a closing device (Kernan in Schulte, 1916, p. 435). The muscles opening the nostrils and moving the plugs are simple, and the elasticity of the tissue as felt when one thrusts an arm down the nasal passage of a dead whale is surely great enough to close it. There has been no development of an elaborate pneumatic mechanism to tighten closure. Both groups of cetaceans have in common their need for some way to keep water out of the inner passage. Stormy weather and fast swimming with a minimum of time at the surface make this a real problem for each, but the solution is different. The odon- tocetes. as we have discovered, have a water trap inside the blowhole. The mysticetes have a conspicuous V-shaped pair of ridges which elevate in front of the nostrils as these open. Ob- servations at sea of the humpback, Megaptera novae-angliae, and the little piked whale, Balaenopfera acuto-rostrata, show these to be very efficient water deflectors, at least in moderate weather. The use of buoyancy changes may also be common to both groups, although we have not observed it in the mysticetes. We have frequently seen our captive Tursiops sink verticalh^ This sudden sinking without exhaling and without any horizontal motion (thus excluding the possibility of planing forces driving the porpoise down) seems best explained by a buoyancy reduc- tion from making the body smaller. This could be done only by compressing the internal air, which requires a tight seal in the i-espiratory passage ; since human swimmers can do this in slight degree, it is not surprising to find a cetacean adept at it. Such a maneuver differs from the sinking after exhaling often observed by us in Stenella plagiodon and reported by "Woodcock and Mc- Bride (1951, p. 215). One is reminded of the old whalers' ac- counts of "settling" by the sperm whale, supposedly as an emergency maneuver when surprised by a whale boat (e.g., Davis, 1874. p. 187). Different solutions to similar problems based on fundamental 140 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY dissimilarities in structure only partly explain the elaborate pneumatic arrangement of the odontocete passage. Other dif- ferences based on different habits and activities must also con- tribute. One of the most conspicuous features of the odontocetes, and one in which they differ widel}^ from the mysticetes, is their habitual production of a wide variety of underwater sounds. Experiments have led us to believe that these sounds are laryn- geal, and we have found that thev all can be made without necessarily allowing air to escape from the blowhole. Undoubt- edly the ability to subject the internal air to considerably greater than ambient pressure is important in this connection. Regula- tion of the air pressure in the sinuses connected with the middle ear may also be affected by this arrangement, much as a noseclip aids a human diver, especially if he goes deep. DISCUSSION OF PREVIOUS WORK Much of the early work on the blowhole region does not deal with function, and since this is our main interest an historical account of these investigations is not attempted here. The reader who wishes to pursue this can find fairly detailed references in Murie (1873) and Gruhl (1911). It is, however, opportune to summarize the varying opinions on the manner in which the outer nasal passage is opened and closed and the anatomical evidence on which these opinions rest. Since the early 19th century there have been two general theories ; the one held by such workers as von Baer, Kiikenthal, Boenninghaus, and Gruhl, that the blowhole closes when its muscles relax and is sealed tighter by increasing hydrostatic pressure, and the other held by Cuvier, Sibson, Murie, and more recently by Huber and others, that certain of the layers of blow- hole muscle pull the blowhole closed while others open it. Fur- ther, there are differences of opinion as to the number of layers involved. The first group of workers tends to emphasize the essential singleness of the blowhole muscle in structure as well as in function. The second group, in accordance with the double function they assign to this musculature, defines numerous layers. Thus, those workers who believed the blowhole to be opened, not closed, by muscle action felt it unnecessary to recog- nize more than a very few laj^ers, while those whose approach is LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 141 through dissections (accurately showing more numerous muscle layers) have felt it necessary to assign a double function to these. Neither group has adequately described or shown the use of the nasal sacs. A considerable part of these investigations was carried out on Phocoena. Piecing together the accounts of phocoenids by von Baer, Kiikenthal, Sibson, Boenninghaus, Gruhl, and Howell, one finds a certain similarity in structure between these and the delphinids, but the differences are such that more anatomical detail and an actual examination of live specimens is necessary before function can be compared in the two groups. However, we see no reason to doubt von Baer's assertion of 1826 (col. 820) that the lips of the blowhole are closed except when opened by muscles. Turning to the delphinids, we find that the earliest good descriptions and illustrations of the blowhole region and asso- ciated muscles in this group are those of Murie, who published on Grampus rissoanus {= g risen s) (1870), Lagenorhynchus al- hirostris (1871), and Glohicephalus nielas (^ Glohicephala melaena) (1873), stating in the latter paper (p. 247) that the function of the layers is "nearly identical in the several forms." Of these genera the second is the most closely related to the three discussed in this paper, and dissection by us of Lagenorhyn- chus acutus shows this species, at least, to resemble our three closely. In this it disagrees with Murie 's figures of L. alhirostris. which differ chiefly in the wide areas of insertion of the outer layers of blowhole muscle, structure and musculature of the tubular sac, and the arrangement of the rostral muscles. In spite of this, a comparison of his text and figures leads us to believe that his layers (1871, pp. 147-148 and pi. 5) may be homologized with ours as follows : first layer, occipito-frontalis, with pe, and probably also i; second layer, levator labii superioris alaeque nasi, with ae (the posterior part is probably homologous with our inner layer of pe which we found merging with ae) ; third layer, levator superioris proprius, with pi: fourth layer, zygo- maticus, and the posterior part of the fifth layer, depressor alae nasi, jointly with ae and pr; the anterior part of the fifth layer, which he calls naso-labialis, with the pars labialis of the maxil- Icmasolabialis. The nasal plug muscle is described (1871, p. 149) l)ut not named or figured. Finally pyramidalis, in spite of its difference in size and orientation, is probably equivalent to our 142 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY major intrinsic muscle of the tubular sac. Murie's sacs also maj^ be homologized as follows : the maxillary with our vestibular sac, the naso-frontal with the posterior part of our tubular sac, and the anterior or premaxillary with our premaxillary sac. While our examination of Lagenorhynchus was not as thorough as that of the other three forms, we found nothing to suggest that it functions differently. In spite of this strong similarity, we also found that as a group Tursiops, Stenella, and Delphinus resemble each other more closely than any of them does Lagenorhynclius. In some features this latter, a shorter beaked porpoise, tends to approach Glohicephala. Although the evidence is not conclusive, the resemblance (in such detail as Murie published) of Grampus and Globicephala to our material suggests very persuasively that here also function will be found to be very similar. Our own superficial examination of the passage and sacs in Glohicephala tends to confirm this. Murie himself felt this to be true of his forms and so stated (see below) in the last of his three accounts. His work on Glohicejjhala, although completed first, was not published till after the other papers had come out. His work on Grampus was completed next and published first, while that on Lagenorliyncliiis refers to both of the previous dissections. Homologous outer layers are given the same names in all three although they are sometimes numbered differently. Identification of the inner layers varies, and one more layer is recognized in Lagenorhynchus than in the other two. This inconsistency is based not so much on an apparent difference in the layers as on attempts to homologize them with facial muscles of other mam- mals, and so does not concern us here. What is of interest is Murie's inability to come to any single, final conclusion as to the functions of these layers in spite of his statement (1873, p. 247) : " As regards the action of the diff'erent layers in Glohiceps etc., they are nearly identical in the several forms." In the paper on Grampus he says that the three outer layers close the blowhole while the fourth layer opens it, as does also the nasolabialis. In the paper on Lagenorhynchus he claims the opposite function for these same three, saying they "dilate'' the blowhole; the use of the fourth is more obscure : he appears to feel that primarily it helps close the passage by compressing the nasal blubber, while it could also have a tendency to open the blowhole ; the fifth layer he feels has a divided function: the anterior part (1871, p. LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 143 149) "drags forward and everts the anterior lip," the posterior part retracts it. Evidently in Glohicepliala the various layers function as in Lagenorhynchus. This is apparent when on com- paring text and figures (1873, pp. 247-248, figs. 63-65), it be- comes clear that the "superior layer" of Glohicepliala equals oe- eipito-frontalis and levator labii superioris alaeque nasi (the first and second layers of Lagenorhynchus) , while the "second sheet" equals levator superioris proprius (or the "third layer" of Lagenorhynchus) . Clearly, he frequently had in mind that what applied to the blowhole mechanism of one genus was equally true of another, as witness his numerous cross references in lieu of restatement. Nevertheless, he never summarized or clarified these conflicting remarks, and his sequence of dissections and publication is such that it is not possible to decide what his final conclusions were. Although this might make his work seem confused, one has only to consider how difficult his dissections must have been to realize how remarkable it is that he was able to find out so much. As regards the anterior part of the naso- labialis, he seems to understand that this has an opening func- tion, though, as in the case of the other layers, the actual region of the nasal passage affected is not made clear. Kiikenthal's (1893) discussion of the odontocete nose in his monumental whale monograph is chiefly concerned with homolo- gies and phylogeny. He adds little to what von Baer had long before said of function, and his anatomical work bearing on these matters is adequately built on by Gruhl (1911). Gruhl's important comparative account of the nasal passage in the Ceta- cea deals primarily with odontocetes, of which he dissected four genera, and includes a detailed discussion of previous work. His comments on function and such anatomical details as he gives for Tursiops and Delphinus need to be considered here. Very properly he realizes that elasticity of the walls closes the passage, but he dismisses the musculature as essentially a large paired muscle which embraces the outer sacs and is suited to widen them. As with previous workers who have understood the basic single function of the blowhole muscle layers, there is no apparent grasping of the fact that the valve moves in one way. the pos- terior wall of the blowhole in another, and the folds of the vestibular sac still differently. It is this difference in the way particular parts of the nasal passage have to move in order for 144 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY it to open which has made it necessary for the musculature to be organized in alternating layers with differently directed fibers attaching in specific places, not all around the blowhole. In describing the sacs in Tursiops and Delphinus he calls attention to considerable variation in them, particularly in the connecting sac, which he found to be smaller in Delphinus than in Tursiops. This is in agreement with our findings. However, his lack of live or fresh material has led him to believe that the small size of the tubular sac means that it is a mere vestige of the "Regio olfactoria, " and hence insignificant. He also thinks that the function of the premaxillary sacs is to facilitate the action of what we call the nasal plugs. His detailed and careful descrip- tion of what we have called the blowhole ligament is of particular interest in that he refers to it as a cartilage. This suggests that chondrification may be more usual than we found it. Apparently disregarding Gruhl's work, Kellogg (1928, pp. 194-195; 1938, pp. 659) refers to a muscular closing of the nose which he describes (1928) as follows: "... the narial cavity is divided into a series of pockets into which the narial passages open, and between these pockets are a series of plugs controlled by muscles that apparently contract when pressure is exerted on the external flap of fibro-elastic cartilage that covers the external nostril." There are no further anatomical details nor explanations, and since this description does not agree with what we have seen, we are at a loss to reconcile it with our findings. The most recent investigations of the blowhole mechanism in odontocetes are those begun by Ernst Huber as part of his careful series of papers on the evolution of the facial musculature of vertebrates. Through studying the innervation in both Tursiops and Monodon he came to the, conclusion that only one muscle, the maxillonasolabialis, is involved. The blowhole laj^ers and nasal plug muscle are derived from the pars nasalis of this muscle and the rostral muscles from the pars labialis. In addition to studying homologies, he also considers the mechanism of the blowhole, returning to the earlier theory that muscle action helps to close it. A posthumously published paper, assembled from iiis notes and sketches, is very explicit as to how this might be done. T"'^nfortunately he himself never lived to complete this work, so the anatomica] basis for this theorv is not sufficientlv detailed ; LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 145 indeed the publiished schematic diagrams of muscle action do not agree with the figures of dissections. The latter show in more detail than had previously been done the layering of the blowhole muscle, how the outermost layers are more posterior, the inner ones more anterior, and the increasing horizontality of the deeper layers. The direction of the fibers of the pars labialis of the maxillonasolabialis and their continuity with the inner layers of the blowhole muscle are also clearly shown. What are lacking are complete descriptions of the individual layers and their subsidiary parts, their relative thickness, fiber direction, and the minutiae of their insertion, as well as detailed descrip- tions of the sacs and their relation to the main passage. Probably it is in part lack of this evidence which has led him to con- clusions so at variance with what our studies have demonstrated. The assumption that the outer layers of muscle close the blow- hole is based largely on their attachment in the tissue on each side of the passage. This is easily seen, and it is interesting to note that while Huber interprets this to mean that they pull the blowhole closed, earlier workers interpreted this as evidence of a dilating action. Actually, if thickness of the layers and insertion of their strongest parts are known, it is perfectly clear that the layers which attach around the passage in this way have a powerful portion which moves one part of the passage so as to open it. and a weaker, thin portion which helps anchor an opposite part against distortion. The deeper layers he be- lieved assisted in closing the blowhole by pulling the "elastic cushion" beneath the valve backward, thus squeezing the valve back and up against the posterior lip. We found no layers which could do this. Either they attach posterior to the passage or too deep for their contraction to have this effect. Though he thus postulates a closing action for layers which we find open the passage, he also refers (1934, p. 124) to "the dilator musculature of the blowhole orifice. ' ' This is not further described or figured so it is difficult to know Avhat muscles or layers he had in mind. As well as ])ostulating muscle action to close the blowhole, Huber also said that air trapped in the vestibular sac helps to keep the plugs seated ; but after a blow we always found this sac empty of air. The pneumatic action which he felt around the outside of the blowhole was probably deeper than he suspected. 146 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY His forcing the animal to breathe by pressing here might have been due, as he suggests, to unseating the plugs. We ourselves were only once able to make a porpoise breathe by squeezing lateral to the blowhole, though we tried repeatedly. Unseating the plugs by means of a finger inserted down the passage often made it blow. Sometimes, also, when it was resisting our opening the blowhole it was possible to make it blow by pressing down on the valve. In respect of these remarks about cetaceans, it is interesting to note that of seals Howell (1930, p. 88) says that two muscles control the nose : the maxillonasolabialis opens it and the naso- labialis closes it. Now we find in cetaceans, where the closing is not by muscle action, that the nasolabialis is rudimentary. Huber's paper was prepared for publication by A. B. Howell, who gives no indication of disagreeing with the theories therein. However, in an earlier work while speculating on the origins of the sacs, he says (1930, pp. 98-99), "The intricate and laminated muscle layers which converge to operate its opening cannot func- tion as dilators with precisely equal force at each and everv' point, however, but their pull must be very unequal." This ex- cellent observation is subsequently confused by a statement that the nasal valve (which etiuals our anterior lip and may also be used here to include the plugs) "is opened by contraction of certain of the rostral muscles, and that in closure there is tension from laterad of the blowhole, but whether [it] is voluntary . . . is not clear." As the rostral muscles, those beneath the melon, are obviously not the same as the intricate muscle layers converg- ing on the blowhole and as he refers the reader to Huber for further details, one gathers that he is in fundamental agreement with the latter. Howell's published diagram (1930, p. 98, fig. 15) is further misleading in that the interlocking arrangements of the supracranial part of the passage are not shown, nor is the attachment of the nasal plug (valvular plug of Huber) to the membranous septum at the margin of the premaxillary sacs. This is what makes it impossible for a forward movement of the plug to open the premaxillary sac as Howell's diagram shows. "The handbook of R. H. Burne's cetacean dissections" (Fraser 1952), although it does not deal with function, is of interest in that the sagittal sections of Delphiniis resemble closely ours of Tiirsiops. This is to be expected in view of the close resemblance LAWRENCE AXD SCHEVILL : THE DELPHINID NOSE 147 we found between the two forms in the arrangement of sacs and muscle layers. Comparing the terms used for the various struc- tures which he found, his "boss" matches well our nasal plugs, while his vestibular and premaxillary sacs are the same as ours. The intermediate areas as shown in his Figure 19 are harder to identity with the structures found in our dissection of Del- phi nus. Apparently the "tongue-like" process equals both our cartilaginous septum and the concave mass of the posterior wall ; the "subsidiary sac" seems closest to our connecting sac, while the space above it could easily be the tubular sac. The dissection shown is exceedingly difficult to make without distorting the region, and this probably accounts for the discrepancies between our findings and his. EFFECT OF HYDROSTATIC PRESSURE OX THE BLOWHOLE MECHANISM The failure of earlier workers properly to explain the blow- hole mechanism has been partly due to lack of observations on living or recently-killed animals. Further, the anatomical basis for these theories has been incomplete. The arrangement of the muscles had been only partly worked out and too little emphasis put on the actual shape of the air spaces. In addition, the relation of pressure to closing the passage had not been properly analyzed. The effect of increasing sea (hydrostatic) pressure on a closed, collapsible container full of air is to reduce the volume of that air so that pressure inside the container is the same as pressure outside. This means that two very common, essentially opposite theories about pressure in relation to closing the passage are false. The first holds that pressure tends to expel air from the animal, thus necessitating a very strong closing mechanism. The second supposes that pressure itself is the closing mechanism — on the assumption that increasing hydrostatic pressure squeezes the lips of the blowhole ever more tightly together. As far as the first is concerned, the OA'erlapping and close- fitting-together of the lips and walls of the passage form a series of check valves which trap the air so that its relative density cannot make it bubble out. Further, the cetacean diaphragm slopes in such a way as to allow far more compression of the lungs than in land mammals. Thus the porpoise as it submerges 148 BULLETIN : MITSEUM OF COMPARATIVE ZOOLOGY is effectively a compressible closed container of gas, and subject as such to Boyle's law. The second assumption is equally wrong, since the compressi- bility of the air bubble results in the animal maintaining an internal pressure equivalent to the hydrostatic pressure around it. If the pressure on both sides of his valves remains the same, any increase in outside pressure squeezing the passage shut would be acted against by an equivalent increase in internal pressure forcing the walls apart. Outside pressure could tighten the seal only if it acted against a lesser inside pressure. This would be possible only if the porpoise could keep its pulmonary air at less than sea pressure, which, as we have said (p. 138), would be a difficult feat involving holding the ribs lifted against very considerable outside pressure. It is far easier for the porpoise to increase than to decrease its internal pressure, and it is just such an increase, as we haA'e shown above, which pneumatically tightens the seal of the passage. SUMMARY Investigations have been made on the blowhole and related nasal structures in the delphinid genera Tur slops, Stenella, Del- phinus, Lagcnorhynchus, and Glohicephala. These have included dissections and pneumatic studies of fresh and frozen carcasses, slow-motion moving pictures, and observations on live animals, both in captivity and at large. We have demonstrated that the nasal passage is opened by muscle action and closed by the elasticity of the tissues, and that this closure may be voluntarily tightened by a pneumatic seal. The blowhole musculature which opens the passage is incom- pletely divided into six different layers, of which five are func- tionally important. Of these, two insert posterior to the passage and three anterior, alternating with each other to do so and overlapping broadly at their origin. Those layers which lift the posterior part of the passage up and back are in general more external and are thickest posterior to the passage ; those which draw the anterior part down and forward are more internal and have their main mass anterior and lateral to the passage. The passage is anchored against distortion by thinner portions of these various lavers attaching in such a wav as to counteract the LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 149 pull of each major portion. As the walls of the passage are drawn apart, the plugs covering the opening into the bony nares are withdrawn by an independent muscle. When the passage is closed the posterior wall presses closely on top of and behind the anterior. The outer j^art of the passage, including the ves- tibular sac, may be considered as a vestibule. This is used as a water trap when the animal submerges with the blowhole open. Beneath this is a transverse slit which marks the entrance to the deeper part of the passage. Lying around this slit on each side is the U-shaped tubular sac, which can be inflated without withdrawing the plugs and whose inflation tightens the seal at this slit. This control allows the porpoise to keep its internal air at greater than ambient pressure. Beyond a certain limit an increase in pulmonary pressure forces the air past this point and into the vestibular sac, where a small amount can be trapped beneath the anterior fold, any excess escaping at the blowhole, where there are no constricting muscles. The blowhole mechanism is efficient in protecting the inner passages from flooding and also in allowing the porpoise a certain independence from its medium in controlling its pulmonary pressure. The odontocete nose is so much more complicated than that of the mysticetes that it suggests other non-re.spiratory functions. Of these, buoyancy reduction, vocalization, and pres- sure equalization in the sinuses connected with the middle ear are probably all important. ACKNOAVLEDGMENTS This work was supported chiefly by the Office of Naval Re- search through a contract with the Woods Hole Oceanographic Tn.stitution. As u.sual, we are grateful to many people for help. In par- ticular we wish to thank Marine Studios, Marineland, Florida, for their hospitality over a number of years. Not only have they repeatedly granted us the use of their unparalleled facilities, but they have also provided and maintained experimental animals for us. and their staff has helped enormously with unflagging co- operation and interest. AVe would like to mention especially Messrs. W. P. Rolleston, F. G. Wood, Jr., and Frank S. Essapian. as Mell as Captain R. Capo and INIr. L. Andreu. 150 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY We are indebted also to Dr. Henry Kritzler for his help in the early stages of this work and for the two latex injected specimens used in this study. The opportunity to study Globicephala in Trinity Bay, New- foundland, was due to the kindness of Arctic Fisheries Products, Ltd., Mr. Bernard Andrews, Mgr., and the help of Dr. David E. Sergeant, Fisheries Board of Canada. We wish also to thank Drs. H. B. Bigelow and R. H. Backus for reading the manuscript, and especially Drs. C. P. LjTuan and D. C. Fawcett for giving us a great deal of time and for their constructive criticism. The drawings were made by Miss Jessie H. Sawj'er. Photo- graphs of transverse sections were taken by Mr. Frank S. Essa- pian, and the motion pictures by the authors. EEFERENCES VON Baee, K. E. 1826. Die Nase der Cetaceen, erlautert dnrch Untersuchung der Nase des Braunfisches (Delphinus Phocaena) . Tsis von Oken, Jg. 1826, Heft 8, columns 811-847. pis. 5-6. BOENNIXGHAUS, GeORG 1902. Der Eachen von Phocaena coininunis Less. Zoologische Jahrbii- cher, Abt. Anat., 17 (1-2) : 1-98, 20 text figs., 1 pi. Davis, William M. 1874. Nimrod of the sea. New York, Harper & Brothers, 403 pp.. 30 text figures. [Eraser, Francis C] 1952. Handbook of R. H. Burne's cetacean dissections. London, British -Museum (Natural History), 70 pp., 41 figs. Gruhl, Kurt 1911. Beitrage zur Anatomie und Physiologie der Cetaceennase. Jena. Zeitschr. Xaturw., 47(3) : 367-414, 21 text figs. Howell, A. Brazier 1927. Contribution to the anatomy of the Chinese finless porpoise, Xeomeris phocaenoides. Proc. U. S. Nat. Mus., 70 (art. 13") : 1-43. 14 text figs., 1 pi. 1930. Aquatic mammals. C. C. Thomas, Springfield (Dlinois) and Baltimore (Maryland), xii -|- 338 pp., 1 pi., 53 text figs. LAWRENCE AND SCHEVILL : THE DELPHINID NOSE 151 HuBER, Ernst 1934. Anatomical notes ou Pinnipedia and Cetacea. Carnegie Instit. Washington, publ. no. 447: 105-1.36, 12 text figs. Hunter, John 1787. Observations on the structure and oeconomy of whales. Phil. Trans. Eoy. Soe. London, 77 (pt. 2): 371-450, pis. 16-23. Irving, Laurence, P. F. Scholander, and S. W. Grinnell 1941. The respiration of the porpoise, Tursiops truncal us. Jour. Cell. Comp. PhysioL, 17(2) : 145-168, 15 text figs. Kellogg, Remington 1928. The history of whales — their adaptation to life in the water (concluded). Quart. Rev. Biol., 3(2): 174-208, figs. 12-24. 1938. Adaptation of structure to function in whales. Cooperation in Research. Carnegie Instit. Washington, publ. no. 501: 649-682. Kukenthal, Willy 1893. Vergleichend-anatomische uud entwickelungsgeschichtliche Unter- suchungen an Walthieren: Bau und Entwickelung der Cetaceen- nase. Denkschr. Med. Naturw. Ges. Jena. 3(2): 322-349, text fig. 51, pis. 21-24. MuRiE, James 1870. On Risso's grampus, G. rissoanus (Desm.). Jour. Auat. Physiol., 5: 118 138, 2 text figs., pi. 5. 1871. Notes on the white-beaked bottleuose, Lagenorhynclius albirostris, Gray. Jour. Linn. Soc, Zool., 11: 141-153, pi. 5. 1873. On the organization of the eaaing whale, Globicephalus melas. Trans. Zool. Soc. London, 8(4): 235-301. 6 text figs., pis. 30-38. Schulte, H. von W. 1916. Anatomy of a foetus of Balaenoptera borealis. Mem. Am. Mus. Nat. Hist., n. s., 1(6) : 389-502, 10 text figs., pis. 43-57. SiBSON, Francis 1848. On the blowhole of the porpoise. Phil. Trans. Roy. Soe. London, 1848: 117123, pi. 12. True, Frederick W. 1891. Observations on the life history of the bottlenose porpoise. Proc. U. S. Xat. Mus., 13: 197-203, 1 text fig. Woodcock, A. H., and A. F. McBride 1951. Wave-riding dolphins. Jour. Exp. Biol., 28(2'): 215-217. EXPLANATION OF FIGURES All of the figures are of Tursiops truncatus (Montagu). The dissections are of a young female (skull length about 150 mm.), the sagittal sections of an old female, and the transverse sections of an old bull. In Figures 1, 21-23, and 26-30 the anterior end is at the bottom of the picture; in all others except the transverse sections it is at the right. Orien- tation of the head is the same in Figures 10, 11, 13-17, and 19. Abbreviations used: Layers of the pars nasalis of the maxillonasolabialis muscle pe, posteroexternus ae, anteroexternus t, intermedins pi, posterointernus ai, anterointernus pr, profundus MAJ. I\T., major intrinsic muscle of tubular sac MIN. IXT., minor intrinsic muscle of tubular sac VEST. S., vestibular sac TUB. S., tubular sac CONX. S., connecting sac PBEMAX. S., premaxillary sac Fig. 1. Plan of nasal sacs in dorsal view, and positions of sagittal sec- tions, Figs. 2-4. In vertical sequence the vestibular sac is the outermost, then the tubular, the connecting, and the premaxillary. The posterior wall of the bony nares is stippled lightly; the margin of the premaxillary sac is shown by a broken line, the posterior margin of the plug by a solid line. The left vestibular sac is omitted. BULL. MUS. COMP. ZOOL OFL'NING- BEriAiELN DORSAL' SURFACE OF PLU&S L PCST ^MALL Pkfmax. S. Fig. 2. Sagittal section through blowhole made at a, Fig. 1. BULL. MUS. COMP. ZOOL. Fig. 2 to CD Qi < Q_ ^ ^ 2 ° ^ < ^ < ^ > o 1 CO CO o n ^ t— " LLJ > O Q O / / < O H- CO I- -^ > , CO CD > O L._ 1 ^^ CO . 1 q: h- ' 1 O \ cb \ _ \o \ ZD , 2 CQ ^ \ < < 5 Q_ \ err 1 r ~i ^ ^ \ < o CQ ' CO c!^S ^< <; u_j y ^c:> Fig. 3. Sagittal section lateral to blowhole made at 6, Fig. 1. BULL. MUS. COMP. ZOOL. Fig. 3 CO cn an ^^ o 3 $,-^ Q_ a_ '-^ ,^ 1 ^9 < _j < Fig. 4. Sagittal section near lateral edge of passage made at c, Fig. 1. BULL. MUS. COMP. ZOOL. Pig. 4 cr I — CO Fig. 5. Diagram of sagittal section through right side of nasal passage of fetus with valve and right plug drawn forward to show attachment of cartilaginous septum between plugs. In older animals the larynx lies farther lip in the posterior bony nares. BULL. MUS. COMP. ZOOL. Fig. 5 Tub. s. Left plug- ^^^. behind septum Post lip Valve "f\l&HT PLL'(j DKAIMN rOR- HARD PrEMAvX. Sac • Tip of larynx Trache/A Fig. 6. Oblique side view of skull showing positions of transverse sec- tions, Figs. 7-9, 24, 25. The nares are shaded. Fig. 7. Rostral face of approximately transverse section, posterior to bony nares, made at a, Fig. 6, showing relative position of sacs and passage as well as principal separation Itetween layers of blowhole muscle. See Fig. 24 for photograph of this same section with passage cut open along lines shown here. Maxilla Nasal BULL. MUS. COMP. ZOOL. M&StTHMOlD I ; ,B LOIN HOLE I / *.'^ I Lacrymal Occipital-. Frontal- Parietal-'' Temporal'' Palatine / Pterygoid Bony nares Fig. 6 HEAD MAXILLA Fig. 7 Separation Blubber Tu B^ s. Passa&e between PLU& L POST I^ALL -Ant fold vests -'^Cranial CAVITY .' In Fig. 24 passa&e cut here f^assaoe beneath dia&onal membrane Passage above diagonal membrane Fig. 8. Eostral face of approximately transverse section through bony nares made at b, Fig. 6. The heavy black line shows the extent of the paired premaxillary sacs lateral and anterior to the nares. On the left of the figure the lining of the naris has been removed. Fig. 9. Eostral face of transverse section anterior to bony nares, made at c, Fig. 6. The premaxillary sacs are here separated bj' a thin partition, and the separation l)etween the anterior and posterior layers of the blowhole muscle is well marked. BULL. MUS. COMP. ZOOL. Nasal plu'j muscle Bony septum^ Fig. 8 pRfMA XILLA B^lUBBER Frontal Premaxillary s. .-Separation be- ^ ^n\.i.H pLiol 'Naris lntranceto eustachian TUBE Sphincter MUSCLE Melon Nasal plu& muscle pREMAXILLARY ""--------. SAG Mesethmoid cartila&e ^' Vomer'' ]_ Fnd of ptery&oid sinus Premaxilla -Sinus con- nected lAllTH eustachian tube """"""Jaiai Fig. 9 RATION BE- vi'im pu Lac Palatine ""Jaiaj Maxillary sinus Fig. 10. Outermost layers of blowhole muscle; vestibular sac beneath posterior lip dotted in. Fig. 11. Posteroexternus layer partly cut away to show attachments of outer layers of blowhole muscle on and around vestibular sac, and orienta- tion of deeper layers. The fibrous sheet separating anteroexternus and pars labialis has been reflected. BULL. MUS. COMP. ZOOL. Fig. 10 pe POST. PART Vest, s ^.Blo/m hole .pe ANT PART -^?C -OOLLA&ENOUS CONNECTIVE TISSUE INSERTION OF PARTS ^^f 6LCLe /Melon, LESS DENSELY fibrous ,MelON, MORE DENSELY FIBROUS m^MemMirgc O'^l ps Fig. 11 /7^,THIN INNER LAYER ^DOE Ofae MERGING lAllTH /?(? 'ED&E OF 6 MERGING lAIITH /7C ^ y' ^'^""-'^.^--Vest. s. ^,6, REFLECTED -Pars labialis PL Layer of fat between , blubber l muscle aaS^^^MlS? Fig. 12. Dorsal view of outer layers of blowhole muscle Fig. 13. Posterior lip of blowhole partly cut away and folded back, and vestibular sac cut open to show its inner boundary where the plugs pass beneath the posterior wall. See Figs. 21 and 22 for transverse extent of this opening. BULL. MUS. COMP. ZOOL. Fig. 12 R' Nasal BONE /Vest. s. BL0^^HOLE ^-Pars labialis ^(f PARTLY CUT AWAr TO SHOlAj^/, Blubber pe P' Ve Fig. 18 ST. S- Cut ed&e: of vest s. \\post fold of vest s. Smooth unin& cr post lip passing into wrinkled linins ofvesi s. ,,'Mar&in OF valve against i^hich post, lip closes 'Fold markino edge or tougher external tissue -'-Ant fold of vest s. pulled forward •Opening or passage betvneen PLUa & post lAJALL nsektion 0f'^(? -Deeper fibers of paks labiaus MERGING lAllTh aL .-''"Pars labialis, post part reflected Fig. 14. Vestibular sac and folds cut away to show insertion of an- terointernus and posterointernus. Fig. 15. Posterointernus and outer part of anterointernus cut away to show tubular sac and its intrinsic muscles. BULL. MUS. COMP. ZOOL. Insertion cf pc Post, lip .Valve Fig. 14 'Opening of passa&e be- ^NEEN FLU G-&. POST. l^ALL I _U-CUT MARGINS OF VESTS. ~ ~ at Post lip or blcai hole Fig. 15 /Valve Openin(j of passag-e be- TIAIEEN PLUG- h POST l^ALL Inslrtiom [)^aL >-J.- MaJ. INT Fatty connective tissue ~ ~ ~~~MlN. INT Fig. 16. Tubular sac partly cut away to show connecting sac, edge of premaxillary sac, and inner part of major intrinsic muscle. Fig. 17. Connecting sac and blowhole ligament cut away. The tubular sac is partly cut away, and its entrance pulled open to show the plug in- ternal to it. The premaxillary sac is cut along the margin of the premaxilla and pulled open to show its relation to the tubular sac and the main passage. Two probes show respectively the opening of the tubular sac into the passage (solid line) and the lateral extent of the passage between the plug and the posterior wall (dotted line). BULL. MUS. COMP. ZOOL. Fiff. 16 ■'^ffwwwi^jyay Tub. S. DRAt^N BACK TO SHOl/^ CONN. S. Opening or tub. s. into nasal PASSAO-f WALL LaT IDQl OF PREMAK. S P^OBt IN OPENING OrTUB.S. INTO PASSAGE Post part or tub. s. i/nith entrance pulled open Fig. 17 w lateral cor- PASSAGE BETl^EEN PLUO «oPOST/yALL vmiwr ,tiltf»"'''^'''^^ Fig. 18. Eight half of melon cut away to show inner edge of insertion of pars labialis and its relation to profundus. The premaxillary sac lateral to the naris has been pulled open to show the diagonal membrane. BULL. MUS. COMP. ZOOL. Fig. 18 D_ o K- =3 CJ c/j GO QQ ~ — 1 vC \— < u— U-J -z. O > —1 < z < > o o_ \ \ CD V— \ u_ CO O O n ct^ U_J \ -z. \ ^ / ?^ o / \ C3 / oo / C/O A o ^ cr:^ c^ < Li_l -ZL < (>i CQ Fig. 19. Valve and plug drawn forward to show anterior and posterior limits of attachment of plug on septum and position of anterior and posterior folds of vestibular sac on left side. < CO BULL. MUS. COMP. ZOOL. -^ (±3 j4^ Q Fig. 19 1 1 o < > < ^— < 2 Li_ Li- U_J ce: 1- 00 < — en U_I < < :z Q^ >- CQ 2: ^ 0 Ll_) CO 2 CD 1 < or U-J •z. Z C3 a:: CD \ O) Fig. 20. Exploded diagram of the sacs cut open and distorted, to show their relation to each other and to the main passage. The outer part, con- striction, and inner part have been separated as shown by arrows. Jagged lines show where the sacs have been cut open. Details of the plug and main passage are omitted, and the folds of the vestibular sae are not shown. Outer part, A. This shows the vestibular sac (vs) as an expanded part of the passage between the valve (v) and the slit (s), Avhere the plug (p) passes beneath the posterior wall (piv)- This slit is at once part of the main passage and the inner opening of the vestibular sac. The outer open- ing (o) of the vestibular sac is at the lips of the blowhole. • Constriction, B. The tubular sac (ts) lies on each side of the slit (s), and its opening (e) is downward into the passage external to the plug. The entrance to the connecting sae (cs), which has been stretched open (ef. Figs. 1, 22), lies external to the liplike extension (I) of the plug and passes upward (u) to join the entrance of the tubular sac. Inner part, C. The plug has been cut away medially to show the naris (n) and the inner opening of the premaxillary sac (ps) around the mar- gin (m) of the premaxillarj- l)one. Laterally the premaxillary sac has been cut, and part of the blowhole ligament (t) removed to show the second opening (h) of this sae which is into the passage external to the liplike extension (/) of the itlug. The opening of the connecting sae (cs) downward into the imssage in the region marked i has been mm-li widened. BULL. MUS. COMP. ZOOL / ^ Y (cut ends) Fig. 21. Dorsal view of passage cut open at commissure of blowhole. Anterior fold of vestibular sac drawn forward to show transverse extent of passage at point where plugs pass beneath posterior wall. Fig. 22. Diagram of dorsal view of passage with outer part including vestibular sac cut away. On the left side of the figure the opening where the plug passes beneath the posterior wall marks the inner limit of the vestibular area; the position of the tubular sac around this slit and the outline of the plug beneath this sac are indicated by broken lines; the con- necting sac is omitted. On the right side of the figure, the posterior wall of the passage has been cut away from the skull (along irregular line) and folded forward to show the lateral entrance to the premaxillary sac. The dotted line shows the attachment of the superficial wall of the premaxillary sac external to the liplike extension of the plug. The broken line shows the margin of the plug continuous with that of the premaxillary sac. The entrance to the connecting sac (as well as the sac, which is not shown) is very much smaller on this side than on the right (cf. Fig. 1). BULL. MUS. COMP. ZOOL. Fig. 21 Vesi s.. Ofenin(j of ^-^ PASSAGE BETl/^EEN DORSAL SURFACE or PLUCj h posi nMi Ant fold of rio-ht vest s. /Post fold of VESI s. /Post, lip ^'DRAIAIN BACK/ Valve Opening of passage \ BETlAjEEN dorsal /\ / v / \ SURFACE OF PLUG / \ \ Sb POST l^ALL /^ //Right \ // plug \ Fig. 22 Transverse extent of passage beneath vests. Bony naris PUKE EXTENSION OF PLUG ?0 ST PART- - ~%y' OF TUB.S Premaa. s. / Concave dorsal sur- face DELEFT PLUG .lAT ENTRANCE TO / PREMAX. S. EX- TERNAL TO PLUGS Entrance TO CONN. s. Entrance to TUB. s Ventral surface of post imall Fig. 23. Dorsal view of skull showing muscle attachments (on left of figure), and arrangement of bones around nares. BULL. MUS. COMP. ZOOL. Occipital Fig. 23 ^^/^ WW&;. ae TM/AP^ lligg] at REMAXILLA pr pars lab/cbJis nasa) p)ua m Fig. '24:. Photograph of rostral face of section at a, Fig. (i. The cuts iiidi eated in Fig. 7 have been made and pulled open to show liplike extension of plug and entrance to connecting sac as well as the interior of the posterior part of the tubular sac. The string following the passage outward is hidden as it passes between the dorsal surface of the plug and the concave mass of the posterior wall, visible as it passes around the anteiior fold of the vestiljular sac, and hidden again till it emerges from lilowhole. A1)brevia- tions: a, pajier marker at entrance to connecting sac; //, lateral end of opening of tuliulai' sac into passage; c, paper marker in bend leading to anterior jiart of tubular sac; d, interior of posterior i^art of tubular sac; e, valve, or anterior lip of blowhole; /, diagonal membrane in left side of passage; r/. anterior fold of vestibular sac; //, passage beneath plug; i, string entering passage superficial to plug; ,/, anterior nutrgin of opening of tubular sac; 7;, liplike extension of plug. Fig. l^.l. Rostral face of section through melon made at ^7, Fig. fi. showing insertions of pars labialis and the gradual transition from the more oilv melon to the more fibrous blubber. Abbreviations: ))t, melon; n, blubber; o, rostrum ; /), pars labialis. BULL. MTS. COMP. ZOOL. Fi-. 24 BULL. MIS. (■0.\li\ ZOOL. Fig'. 25 Fig. 26. Beginning of blow. Alternate frames from pictures taken at 64 frames per second looking into left side of passage as it opens. In this and Figures 27 and 28 the following alibreviations are used: b, bony wall of naris; r, convex mass of posterior wall; (1, diagonal membrane; /, anterior fold of vestibular sac; /, liplike extension of plug ; p, plug ; s, septum ; r, valve. Here as well as in the following sequences (Figs. 27-.30), anterior is at the bottom of the picture, posterior at the top. In this sequence in the fifth frame the edge of the diagonal memlirane shows internal to the convex mass; as the blowhole aperture widens, the liplike extension of the plug appears as a pale mass external to the diagonal membrane in the lower left hand corner of the open passage (frames 11, 13, and 15). BULL. MUS. COMP. ZOOL. Fiff. 26 Fig. 27. End of blow l)egun in Figure 26. Uninterrupted sequence of frames from pictures taken at ci- second slio\vin«- water running into open blowhole as porpoise sulinierjics, ami low spray as some water is forced liack out. BULL. MUS. COMP. ZOOL. Fig. 29 Fig. 30. Uninterrupteil sequeiu-e of frames from pit-tures taken at 64 frames per second showing wet Idow following submergence with blow- hole open. BULL. MLTS. COMP. ZOOL. Fig. 30 Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 114, No. 5 STLICIFTED MIDDLE ORDOVICTAN TRILOBTTES THE ODONTOPLEURTDAE By H. B. Whittington With Twenty-pour Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM May, 1956 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 1 14. Breviora (octavo) 1952 — No. 53 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 3, no. 35 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 114, No. 5 SILICIFIED MIDDLE ORDOVICIAN TRILOBITES THE ODONTOPLEURIDAE By H. B. Whittington With Twenty-four Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM May, 1956 No. 5 — Silicificd Middle Ordovician trilohites. The Odontopleuridae By H, B. Whittington CONTENTS Introduction and acknowledgements . Terminology Stratigraphical occurrence and localities Part I : Morphology, Ontogeny and Evolution Morphology of holaspid exoskeleton Cephalon .... Thorax ..... Pygidium .... External surface Abnormal specimens Ontogeny ..... Locus of segmental divisions Mode of life .... Origin and evolution . Part II : Systematic Descriptions Famity Odontopleuridae Burmeister Discussion of svstematics of Odontopleuridae Subfamily Odontopleurinae Burmeister Odoniopleura Emmrich .... Primaspis K. and E. Richter . P. ascitus n. sp. . . . . . Leonaspis R. and E. Richter . Diacanthaspis AYhittingtou D. coo peri Whittington D. lepidus n. sp D. secretus n. sp. ..... Discussion of ontogeny of Diacanthaspis D. ulrichi n. sp D. aff. ulrichi n. sp. . Z>. orandensis n. sp. . D. scitidus n. sp. . Page 159 160 162 166 166 166 173 175 175 178 178 182 185 186 193 193 194 195 196 198 199 205 207 211 216 220 224 225 228 228 230 156 BULLETIN : MUSEUM OF COMPARATm: ZOOLOGY Acidaspis Murchison .... Dudleyaspis Prantl and Pfibyl Radiaspis R. and E. Richter . Subfamily Miraspinae R. and E. Richter . Miraspis R. and E. Richter . fM. sp. ind Ceratocephala Warder .... C. laciniata Whittington and Evitt C. rarispina n. sp. C. (Ceratocephalina) n. subgen. . C. (C.) tridens n. sp. .... Proceratocephala Prantl and Pfibyl Whittingtonia Prantl and Pfibyl . Dicranurus Conrad .... Notes on other miraspinid genera . ?Subfamily Miraspinae R. and E. Richter . Odontopleurid protaspis .... Subfamily Apianurinae n. subfara. . Apianurus n. gen. ..... A. harhatus n. sp. . Specimens at localities 2 and 3 differing A. 'barhatus n. gen., n. sp., A. glaber n. sp. . A. sp. ind. ...... A. aff. fiircata Linnarsson . Calipernurus n. gen. .... C. insoUtns n. sp. Subfamily Selenopeltinae Corda Selenopeltis Corda ..... Other genera, subgenera and species sometimes ferred to Odontopleuridae References Explanation of Plates 1-24 from re Page im 234 235 235 236 238 238 239 242 243 244 247 248 248 248 250 250 252 252 254 205 268 269 270 271 273 279 279 282 283 289 ILLUSTRATIONS Plates Plate 1. Primaspis ascitus n. sp. WHITTINGTON : SILICIFIED ODONTOPLEUBIDAE 157 Plate 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20 21. 22. 23. 24. Primaspis ascitus n. sp. . Diacanthaspis cooperi Whittington . Diacanthaspis lepidus n. sp. . Diacanthaspis lepidus n. sp. Diacanthaspis secretus n. sp. . Diacanthaspis secretus n. sp. and D. lepidus n. sp Diacanthaspis ulrichi n. sp. Diacanthaspis aff. ulrichi n. sp. and D. cooperi Whittington Diacanthaspis orandensis n. sp. Diacanthaspis orandensis n. sp. and D. cooperi Whittington Diacanthaspis scitulus n. sp. Diacanthaspis scitulus n. sp. . Miraspis sp. ind. and Ceratocephala laciniata Whittington and Evitt .... Ceratocephala rarispina n. sp. and C. {Cerato cephalina) n. subgen., sp. ind. Ceratocephala (Ceratocephalina) tridens n subgen., n. sp. Apianurus harhatus n. gen., n. sp. . Apianurus harhatus n. gen., n. sp. . Apianurus harhatus n. gen., n. sp. . Apianurus harhatus n. gen., n. sp., and A. aff furcata Linnarsson .... Apianurus glaher n. gen., n. sp., Apianurus n gen., sp. ind., and miraspinid ? protaspis Calipernurus insolitus n. gen., n. sp. Calipernurus insolitus n. gen., n. sp. Calipernurus insolitus n. gen., n. sp. . Figure 1. 2. 3. Figures Notation for spines of Stage 0 cranidium . Ilypostomes characteristic of subfamilies . Classification, relationships, range in time of Odontopleuridae ...... Page 161 171 : 88-1 89 158 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Figure Page 4. Odontopleura ovata Emmrich .... 197 5. Primaspis primordialis (Barrande) . . . 198 6. Development of Primaspis ascitus n. sp. . . 204 7. Leonaspis n. sp. . . . • • ■ 206 8. Age and comparison of species of Diacanthaspis 208-209 9. Development of Diacanthaspis cooperi Whit- tington ........ 213 10. Keconstruction of Diacanthaspis lepicius n. sp. g^g 11. Development of Diacanthaspis secretus n. sp. 222 12. Development of Diacanthaspis ulrichi n. sp. . 227 13. Acidas2:)is hrightii Murchison .... 233 14. Dudley aspis quinquespinosa (Lake) . . 234 15. Miraspis mira (Barrande) .... 237 16. Ceratocephala laciniata Whittington and Evitt 239 17. Whittingtonia Mspinosa (M'Coy) . . . 247 18. Dicranurus monstrosus (Barrande) . . . 249 19. Reconstruction of Apianurus barhatus n. gen., n. sp 253 20. Cephalon of Apianurus barbatas n. gen., u. sp. 256 21. Triangular graph of dimensions of Apianurus harbatus n. gen., n. sp. . . . . . 259 22. Development of Apianurus barbatus n. gen., n. sp 261 23. Triangular graph of dimensions of Apianurus barbatus n. gen., n. sp., and Calipernurus insolitus n. gen., n. sp. . . . . . 272 24. Reconstruction of Calipernurus insolitus n. gen., n. sp 274 25. Selenopeltis buchi (Barrande) .... 281 Tables Table Page 1. Numbers of exoskeletal parts of two species of Diacanthaspis at localities 2-4 . . . 217 2. Numbers of exoskeletal parts of three species of Diacanthaspis at locality 8 . . . 229 3. Numbers of odontopleurid cranidia at local- ity 8 242 4. Numbers of exoskeletal parts of all sizes of Apianurus n. gen., and Calipernurus n. gen. 267 WHITTINGTON : SILICIFTED ODONTOPLEURIDAE 159 INTRODUCTION AND ACKNOWLEDGEMENTS The photographs show the remarkable nature of the silicified material from Virginia — superbly preserved, free from the matrix, undistorted, and including adults and growth stages. These specimens afford a detailed picture of an early part of the history of odontopleurid trilobites, and thus throw new light on the morphology and evolution of the family. The best odon- topleurid material previously known was that described by Barrande (1852, 1872) and recently re-studied by Prantl and Pi^'ibyl (1949). Although these Bohemian odontopleurids range in age from Ordovician to Devonian, they are not as well pre- served, nor are growth stages known. It is my good fortune that the Museum of Comparative Zoology houses the Schary Collec- tion (to which Barrande originally had access), for this circum- stance enabled me to study the Bohemian species at first hand. A visit to Britain in 1953, and generous loans from American museums, have permitted me to see additional specimens. The type and certain other species have been described elsewhere (Whittington, 1956b), and two silicified species from Virginia were treated of by Whittington and Evitt (1954). In Part II of this paper the remainder of the silicified odontopleurids are described in detail. In Part I, the morphology, ontogeny, and evolution of the family are discussed in general terms. Part II embodies a classification arising out of the new data, and my views on evolution, and includes notes on all the genera which have been referred to the Odontopleuridae. The amount of morphological detail which the silicified exo- skeletons reveal enables fine distinctions to be drawn between genera and between species, i.e. it encourages "splitting." Ex- amples are the new subgenus Ceratocephala (Ceratocephalina), the differences between Apianurus n.gen. and Calipernurus n.gen., and between some of the species of Diacanthaspls. On the other hand, the seven species of Diacanthaspis seem to form part of a natural, related group, and I have widened the original diagnosis of Diacanthaspis rather than split it into subgeneric groups. In dealing with the less well-preserved and often in- complete material on which some Ordovician and all later genera have been established, I have tended to "lump" rather than 160 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY "split." Thus, widely varying degrees of morphological dif- ference separate the genera shown in Text-figure 3, and these differences are not of equal rank. Dr. G. Arthur Cooper and his colleagues at the U. S. National Museum first discovered silicified trilobites in the Middle Ordo- vician limestones of Virginia about 1935. In succeeding years a large collection was prepared, and in 1946 Dr. Cooper invited me to study it, under project grant 491-46 from the Penrose Bequest of the Geological Society of America. Meanwhile, Dr. William R. Evitt, University of Rochester, independently dis- covered these fossils, and since 1947 he and I haA^e collaborated in the preparation of additional material and in studies of it. The silicified trilobites here described are thus derived from our own as well as IJ. S. National Museum collections. A special debt of gratitude is owed to Mrs. Evitt for her painstaking sort- ing of the finest residues, which brought to light many of the tiny growth stages. I am also grateful to Mrs. Stanley J. Olsen, Mrs. Robert E. Kay, and Mr. Ira B. Laby for preparing the enlargements from my negatives and for aiding in mounting the plates. Text-figures 10 and 19 were drawn from my sketches by Mr. F. Y. Cheng, the remainder by Miss Pat Washer. Professor L. Stormer, Paleontological Institute, Oslo, kindly permitted me to describe a Norwegian species here. TERMINOLOGY The terminology used herein is the same as that employed by Whittingtou and Evitt, 1954, pp. 11-14, with certain additions and emendations : — Cephalic spines: to simplify description of cephalic spines a system of letters and numbers has been adopted (Text-fig. 1). Median or paired axial spines of the glabella have been num- bered 1-5, commencing with the median occipital and numbering forwards; 2a is an additional pair appearing later in the on- togeny than 1 and 2. These numbers do not correspond w4th those used by Whittington and Evitt (1954, text-fig. 16), where 2a is numbered 2, 2 is 3, 3 is 4, etc. Certain paired cheek spines have been lettered as shown, in the protaspis and Stage 0 cephala of Diacanthaspis (Text -figs. 9, 11, 12) and Apianurus n. gen. (Text-fig. 22) there is a long, backwardly-directed spine WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 161 at the extremity of the posterior border of the fixed cheek, as well as a long spine on the border of the free cheek. These spines have been called fixigenal and lihrigenal respectively (terms proposed by Richter, 1932), in preference to the older terms metacranidial and parial of Raw (1925). Antennular notch: a notch in the margin of the free cheek immediately outside the anterior branch of the facial suture (a in Text-fig. 20; see also PI. 1, figs. 2, 5; PI. 17, figs. 7, 10). As Figure 1. Cranidium of Stage 0 odontopleurid based on that of Apianurus harbatus n.gen., n.sp., with paired spines numbered and lettered as follows : 1, 2a, 2, 3, 4, 5, six pairs of axial glabellar spines numbered forward from the occipital ring; Ai, A2, A3, on fixed cheek; B, C, on posterior border; D, on fixed cheek behind palpebral lobe; Er, on eye ride; Fx, fixigenal; Op, paired occipital spine; PI, on summit of palpebral lobe. explained under "Cephalon: hypostome" and "Mode of Life," this notch may permit the antennule to protrude forward when the animal is resting on the sea bottom. Pleural spines: The pleura may be divided by a pleural furrow into a convex anterior and posterior band, and each of these bands may be prolonged distally into a spine (the anterior into more than one spine), here called anterior pleural (not terminal 162 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of Prantl and Pfibyl, 1949, p. 128) and posterior pleural spines respectively. Pleural region is used for that part of the pygidium outside the axis, rather than pleural lobe. Pygidial spines: Commonly one paired spine is longer and stouter than the others, and is here called the major spine; it may arise from the border or the surface of the pleural region, and the base is always connected to the first axial ring by a low ridge, here termed the pleural ridge. Sagittal (sag.), Exsagittal (exs.), and Transverse (tr.) refer respectively to the median line, a line parallel to, but outside of the median, and a direction at right angles to the median. The abbreviations used in the text are given in parentheses. STRATiaRAPHICAL OCCURRENCE AND LOCALITIES The specimens described here were obtained from limestones of the Edinburg, Oranda and Martinsburg formations of the Shenandoah Valley, northern Virginia. The stratigraphy of these formations has been described by Cooper and Cooper (1946), and the classification and correlation is summarized in Twenhofel et al. (1954). Trilobites from the Edinburg and underlying limestones have been described by Evitt (1951), Cooper (1953), and Whittington and Evitt (1954), and the latter work also contains an account of the mode of occurrence, preservation, and method of extraction of the silicified specimens. Trilobites from the lower Martinsburg formation have been de- scribed by Whittington (1941), Evitt (1953) and Evitt and Whittington (1953). Odontopleuridae first appear in the lower Lincolnshire limestone (Ceratocephala triacantheis Whitting- ton and Evitt and fragments of Apianurus n. gen.), below the Edinburg formation. In the lowest part of the latter formation (localities 2-4) they are abundant and exhibit the greatest variety. In the slightly higher horizon at locality 6 odonto- pleurids are rare, but they become numerous and varied again some 400 feet higher in the section in the Oranda formation. The latter is not more than 50 ft. thick, and in the lowest part of the succeeding Martinsburg formation silicified odontopleurids again occur in fair abundance and varietv. Not all the blocks of WHITTINGTOlSr : SILICIFIED ODONTOPLEURIDAE 163 limestone collected yield good material, and the same species do not appear at different localities in the same formation — per- haps because of slight differences in horizon. Thus Diacanthaspis ulrichi n. sp. is the only odontopleurid at the lower Edinburg locality 7, is rare at locality 4, and unknown at 2 and 3. Pri- maspis ascifus n. sp. is only known at locality 10 in the lower Martinsburg, and does not accompany t). cooperi at localities 9, 11 and 12. The localities (cf. Whittington and Evitt, 1954, pp. 5-6) and the odontopleurids obtained at each are listed below. Tables 1-4 give numbers of specimens recovered of particular species. Edinburg limestone Locality 2 — Lower part of Edinburg limestone (bed 18 of Cooper and Cooper, 1946, geologic section 19, pp. 94-95), yellow- ish-weathering argillaceous limestone forming edge of quarry and along strike of same bed, in field between quarry and rail- road ; just north of railroad tracks at switch a quarter of a mile east of Strasburg Junction, just west of Strasburg, Shenandoah County, Virginia. Some of the finest specimens illustrated here came from blocks collected at this locality and prepared by Dr. G. Arthur Cooper. Diacanthaspis lepidus n. sp. Diacanthaspis secretus n. sp. Ceratocephala laciniata Whittington and Evitt, 1954 Apianiirus barhatns n. gen., n. sp. Apianiirus glaher n. gen., n. sp. Apianurus sp. ind. Calipernurus insoUtus n. gen., n. sp. Locality 3 — Lower part of Edinburg limestone, section in field on south side of road, 0.2 mile east of Strasburg Junction, just west of Strasburg, Shenandoah County, Virginia. The sec- tion dips about 38° SE. The lowest beds were 6 feet of granular limestones with GirvaneUa sp., about 91 feet from the east edge of the quarry dump at the top of the field. These may be upper Lincolnshire limestone. About 20 feet stratigraphically higher, alternations of dark granular limestone and dark fine-grained limestone with sponge spicules were seen. These beds seem to be 164 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the basal Botetourt limestone member of the Edinburg formation (Cooper and Cooper, 1946, p. 80), and blocks from them were collected and prepared by Dr. G. Arthur Cooper, Dr. and Mrs. W. R. Evitt, and Whittington. These yielded all the species known from locality 2 except Apianurus sp. ind. In addition Miraspis sp. ind. and Ceratocephala {Cerafocephalina) tridens n. subgen., n. sp., occur. Locality 4 — Botetourt member, lower part of Edinburg lime- stone, in upper part of field northeast of Virginia State Highway 639, at a point 0.25 mile from its junction with U.S. Highway 11. This junction is 0.7 mile southwest of Strasburg, Shenandoah County, Virginia. The outcrop is approximately half a mile southwest of locality 3 along the strike of the beds. Collected and prepared by Dr. and Mrs. W. R. Evitt and by Whittington. and notable for yielding some of the finest tiny specimens. Diacanthaspis lepidus n. sp. Diacaniliaspis sccretus n. sp. Diacanthaspis ulrichi n. sp. Ceratocephala laciniata Whittington and Evitt, 1954 Ceratocephala {Ceratocephalina) tridens, n. subgen., n. sp. Apianurus harhatus n. gen., n. sp. Apianurus glaher n. gen., n. sp. Locality 6 — Edinburg limestone, lower part. Hupp Hill, at entrance to Battlefield Crystal Caverns, and in field on opposite (east) side of U. S. Highway 11, about ly^ miles north of Stras- burg, Shenandoah County, Virginia. Discovered by Dr. G. Arthur Cooper, later collections by Whittington. Odontopleurids are rare, only one or two parts of exoskeletons of the following being known: — ^te Diacanthaspis secretus n. sp. Ceratocephala laciniata Whittington and Evitt, 1954 Apianurus harhatus n. gen., n. sp. Locality 7 — Lower part of Edinburg limestone, 300 feet ± south 40° east of bridge, li/4 miles east of Edinburg, Shenandoah County, Virginia. Originally collected by E. 0. Ulrich, later (1931) by Charles Butts, and yielding only one odontopleurid, Diacanthaspis ulrichi n. sp. WHITTTNGTON" : SILTCIFIED ODONTOPLEURIDAE 165 Oranda formation Locality 8 — Lower 5 feet of formation, eobbly limestone, in bank and pasture on north side of Virginia secondary highway 777, just west of its junction with Virginia secondary highway 910, and circa 300 yards north of Greenmount church, five miles north of Harrisonburg, Rockingham County, Virginia. Discovered by G. Arthur Cooper, later collections by Cooper and A. R. Loeblich, Jr., W. R. Evitt, and Whittington. Diacanthaspis orandensis n. sp. Diacanthaspis scitulus n. sp. Diacanthaspis aff. ulrichi n. sp. Ceratocephala rarispina n. sp. Apianurus harhatus n. gen., n. sp. Martinsburg shale Locality 9 — Road cut, gutter, and loose blocks in pasture on west side of Virginia secondary highway 910, about half a mile north of Greenmount church, five miles north of Harrisonburg, Rockingham County, Virginia. Same as locality 1 of Evitt and Whittington, 1953, p. 55. Collected and prepared by Dr. and Mrs. W. R. Evitt, Dr. G. Arthur Cooper, and Whittington. DiacantJiaspis cooptri Whittington, 1941 Locality JO — Pasture on north side of Virginia secondary highway 772, about 1 mile east of Greenmount church, five miles north of Harrisonburg, Rockingham County, Virginia. Collected and prepared by Dr. and Mrs. W. R. Evitt. Primaspis ascitus n. sp. . Diacanthaspis cooperi Whittington, 1941. Locality 11 — Loose blocks in pasture on north side of Virginia secondary highway 616, Vj mile east of intersection with Virginia secondary highway 699, and 2i/2 miles north-northeast of Spring Hill, 7 miles north of Staunton, Augusta County, Virginia. The Oranda formation immediately underlies the lower Martinsburg formation and outcrops a short distance to the west. The locality is that from which all the material described by me (including Diacanthaspis cooperi) in 19.41 came. At that time (Whitting- ton, 1941, p. 492) tAvo localities were given, distant respectively 2^2 miles north-northeast, and 3 miles north-northeast, of Spring 166 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Hill and Long Glade. Spring Hill is a new name for the settle- ment formerly called Long Glade, and the confusion probably arose because the blocks of limestone were collected at different times (cf. Evitt, 1953, p. 34). The first blocks were collected by Dr. G. Arthur Cooper, later ones by Cooper and "Whittington. Locality 12 — In field on south side of Virginia secondary highway 753, 1 mile west of intersection with Virginia secondary highway 732, and 31/2 miles north-northeast of Spring Hill, Augusta County, Virginia. This locality, visited by Dr. G. Arthur Cooper and Whittington, is one mile northeast of locality 11, and yields Diacanthaspis cooperi. PART I: MORPHOLOGY, ONTOGENY A^^) EVOLUTION Morphology' of Holaspid Exoskeleton The family diagnosis (p. 193) epitomises the morphology of an odontopleurid, and in this section certain aspects are com- mented on more fully, these being chiefly features on which I have information additional to that available to Prantl and Pfibyl (1949). Cephalon Attitude: The characteristic attitude of the strongly convex cephalon is discussed below under "Mode of Life." The convex- ity of the cephalon and its position relative to the thorax and pygidium were recognized by some earlier authors (e.g. Weller, 1907, PL 23, figs. 1-4; Warburg, 1933, p. 9, footnote; see also Whittington and Evitt, 1954, p. 54), but not, apparently, by Barrande. Many of Barraude's drawings were made from speci- mens flattened in shale or calcareous mudstones, and show an exterior view of the cephalon combined with a dorsal view of thorax and pygidium (for definitions of these terms see Whit- tington and Evitt, 1954, p. 11). The present reconstructions attempt to remedy this situation (compare Text-figs. 4, 15, 25 with Barrande, 1852, PI. 39, figs. 33, 1; PL 37, fig. 25, respec- tively). Occipital ring: Conspicuous in odontopleurids is the way in WHITTINGTON- : SILICIFIED ODONTOPLEURIDAE 167 which the antero-lateral portion of the occipital ring merges with the inner posterior corner of the cheek. Behind this region the posterior border is distinctly separated from the remainder of the occipital ring, and the axial furrow more or less well developed, as it is beside the lateral glabellar lobes. The occipital ring in most odontopleurid genera is strongly convex and com- monly elongated so that it projects behind adjacent parts of the posterior border. As Prantl and Pfibyl pointed out (1949, p. 127), a median occipital tubercle is always present, and a median spine or paired spines may also be developed. The species of Diacanthaspis here described (Text-fig. 8) show that the median tubercle, or paired spines, or both may be greatly enlarged, or none may be especially prominent, in a group of closely related species. In genera such as Ceratocephala and Miraspis (Text- fig. 15), where long, stout, paired occipital spines are developed, a posterior occipital band, below and behind the bases of these spines, is developed (cf. Warburg, 1933). Such a band is, how- ever, not present in Acidaspis (Text-fig. 13) or Bicranurus (Text-fig. 18), though the median (in the former) and paired (in the latter) occipital spines are large. The view of Reed (1925. p. 420) that the posterior occipital band represents the occipital ring, and that the part in front belongs to the glabella, has been adequately refuted bv Warburg (1933) and Opik (1937. pp. 45-47). Where I have been able to observe both surfaces of the exo- skeleton, as in Primaspis ascitus, n. sp. (PI. 1, figs. 1, 6), the occipital lobes are small, gently convex, subcircular in outline, situated immediately behind the deep outer part of the occipital furrow. Their convexity, rather than any distinct furrow, sep- arates them from the rest of the occipital ring inside and behind them, and they are indistinctly bounded posteriorly. On the inner surface they are seen to be in front of the outer part of the doublure of the occipital ring. In structure and position they are analagous to the lateral glabellar lobes, though they do not extend back to the posterior margin of the segment, and are much less distinct than the first and second glabellar lobes. In Ceratocephala laciniata and C. triacantheis (Whittingtou and Evitt, 1954, PL 6, fig. 4; PI. 8, figs. 1, 2; PI. 25, figs. IQ, 11) occipital lobes are scarcely distinguishable, whereas in certain 168 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY younger species of the same genus (e.g. Warburg, 1933, text- figs. 2, 3) the outer part of the occipital ring, between the occipi- tal furrow and posterior band, is inflated. Similarly, in Dicra- mirus hamatus (Whittington, 1956b, PI. 60, figs. 11, 12) the inflation of the antero-lateral part of the occipital ring, though vaguely defined, especially postero-laterally, is greater than in D. mo7istrosus (Text-fig. 18). Species of other genera show the same features, and indicate that, when occipital lobes are de- veloped (they are absent altogether in some genera and absent in certain species only of other genera), they are situated in the antero-lateral corner of the occipital ring, probably outside the margin of the doublure, and ill-defined on the posterior and inner sides. Their variability suggests that presence or absence of such lobes can scarcely be used as a generic character. Glahellar lohes and furrows: Two pairs of lateral glabellar lobes are always present, the first (basal) pair the larger, and, except in a few genera and certain species of other genera, small, variably-developed, third lateral lobes are present. Particularly in Ceratocephala and Miraspis the third lobes are small and de- pressed, and have been regarded as not developed by Prantl and Pfibyl (1949, pp. 180, 194).- The convexity of the' lateral lobes and degree of their separation from the median glabellar lob(^ are variable (contrast, for example, Leonaspis, Text-figure 7, with Whittingtonia, Text-figure 17), and in Apian urns n.gen. and Primaspis keyserlingi (Barrande, 1852, PI. 36, figs. 10, 12) the lateral lobes are fused. The peculiarities of the glabellar lobation of Selenopeltis (Text-fig. 25) set it apart. The frontal lobe varies in width, sometimes being only as wide as the median lobe {Acidaspis. Text-fig. 13) but usually it is expanded to extend in front of the second or third lobes (Leonaspis n sp., Text-fig. 7 ; Dicranurus liamatiis, Whittington, 1956b, PI. 60, fig. 15; Miraspis mira, Text-fig. 15). The eye ridge terminates opposite the lateral extremity of the frontal lobe, and may merge into it {Ceratocephala, see Whittington and Evitt, 1954, PI. 9. fig. 2) or be separated from it by a low depression, the axial furrow (Primaspis asciius n. sp., PL 1, figs. 2, 5). Hupe's dia- gram (1953, p. 102, text-fig. 62; repeated in 1955, p. 236, text- fig. 205) conveys a misleading impression, in that figure 62 (1), .'howing 3 pairs of lateral glabellar lobes, is labelled "primitif" WHITTINGTON : SILICIFIED ODOXTOPLEURIDAE 169 and ''normale," and figure 62 (3) with 2 pairs of lateral lobes, is labelled as showing ''regression progressive et complete" of the fourth segment (from the posterior margin). In fact no such progressive reduction occurs, for genera with 2 and 3 pairs of lateral lobes appear at about the same time {Diacanthaspis, Primaspis, and Ceratocephala), and Ceratocepliala and Dicra- nurus, both having 3 pairs of lateral lobes, are known in the Middle Devonian. The outer part of the occipital furrow, and each glabellar furrow (especially at the inner end), is deepened and forms a strong ridge on the inner surface (PI. 1, fig. 6; PI. 5, fig. 4; PI. 7, fig. 5; PI. 12, fig. 7"; PI. 15, fig. 5; PI. 20, fig. 11; PI. 21, fig. 6; PI. 22, fig. 6). The quartz along the crests of these ridges may appear darker, strengthening the impression that these ridges are points of muscle attachment. So far as is known, however, these thick ridges are not extended as ventral processes. The outer part of articulating and ring furrows is similarly deepened (e.g. PI. 12, figs. 6, 18) but does not form so strong a ridge on the inner surface. Eye lohe, eye ridge, facial sutures: The external covering of the eye of Ceratocephala has been described ("Whittington and Evitt, 1954, p. 16) as having the outer surface smooth or faintly divided into numerous closely spaced, slightly raised facets, the inner surface showing shallow pits similarly arranged. The ex- ternal covering of the eye of the silieified specimens of other genera described here appears to be similar, the facets on the outer surface sometimes clearly delineated (PL 2, figs. 18, 22; PI. 3, fig. 21; PI. 16, fig. 23; PI. 17, fig. 21; PI. 23, figs. 9, 10). This type of eye surface seems to be typical of odontopleurids. The presence of the eye ridge, and of sutural ridges ("Whit- tington and Evitt, 1954, pp. 13, 17-19), notably where the suture approaches either border, is characteristic. Fusion of the facial sutures, presumably a secondary phe- nomenon, has been suggested as a criterion for distinguishing genera of odontopleurids. It is true that in some specimens in which the exoskeleton is preserved it is almost impossible to detect the course of the suture, but whether or not this means that secondary fusion has taken place is difficult to determine. In any event this single character is not regarded as a reliable 170 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY criterion upon which to base a genus (Whittington and Evitt, 1954, p. 53). Librigenal spine: Prantl and Pfibyl (1949, pp. 126, 131, etc.) stated that, in Selenopeltis and miraspinids. the librigenal spine arose from the "surface of the cheeks above the genal angle," rather than being a backward and outward extension of the rolled borders, at the genal angle, as it is in odontopleurinids. In Ceratocephala (AVhittington and Evitt, 1954, PI. 6, fig. 17; PI. 25, figs. 12-14), the broad base of the librigenal spine merges into the upper surface of the posterior and lateral borders and the sutural ridge. In Ceratocephala {C cratocephalina n. subgen.) (PI. 16, figs. 23, 24) it likewise arises from the upper surface of the border, and a line of spines on the margin of the border is continuous below it. The fact that the doublure extends beneath the base of the librigenal spine (PI. 16, fig. 0) shows that the spine arises from the border and not the surface of the cheek (cf. Whittington and Evitt, 1954, p. 52). Thus I do not agree with Prantl and Pfibyl (1949, p. 131, etc.) that the miraspinid librigenal spine arises in a way that is fundamentally different from, and originates from other cephalic segments than, that of odontopleurinids. I l)elieve rather that the difference between the librigenal spines in the two subfamilies is one of direction and not place of origin. In apianurinids the librigenal spine also arises from the cephalic border, but from a point farther for- ward. The ontogeny of Diacanthaspis, of Apianurus n. gen., and of Ceratocephala shows that the librigenal spine develops in much the same way in each group, but does not show to which cephalic segment it may belong. Hypostome: The hypostomes of a species of Primaspis and those of Apiayiurus n. gen. and Calipernurus n. gen. are described here, and others in Whittington, 1956b. Thus the hypostomes of more than half of the odontopleurid genera are known, and seem to fall into four types, exemplified in Text-figure 2. The "family resemblance" between them is shown particularly by the small anterior wing (lacking the wing process), lateral notch and pointed shoulder, middle body with large depression in antero-lateral corner, from which the middle furrow runs inward and backward, and tiny posterior wing. The types distinguished form one of the main bases of the four subfamilies recognized. WHITTINGTOX : SILICIFIED ODONTOPLEURIDAE 171 Prantl and Pfibyl (1949, p. 134, etc.) stressed the ditfereuce between the odontopleurinid and the S elenopeltis-C eratocephala types of hj'postome. The latter two present a considerable like- ness, and in both the crescentic tip of the posterior lobe of the middle body is inflated and extends outside the depression in the antero-lateral corner of the middle body. In the odontopleurinid (and apianurinid) type the tip of the posterior lobe of the mid- Figure 2. HjTDOstomes charaeteristie of each of the four subfamilies of Odontopleuridae. A, Odontopleurinae, Primaspis, based on that of P. asffitus n.sp. B, Apianurinae, Apianuras, based on that of A. harbatus n.geu., n.sp. C, Miraspinae, C eratocephala, based on that of C. laciniata Whitting- ton and Evitt. T), Selenopeltinae, Srlptiopeltis, based on that of S. hvchi (Barrande). die body does not extend as far forward. I hardly think this difference is as great as Prantl and Pfibyl tend to make it by saying that in odontopleurinids the middle furrow starts from the lateral border furrow, whereas in Selenopeltis and Cerafo- cephala it starts from the anterior border furrow. The silicified specimens of hj^postomes have shown the exist- 172 BULLETIN : MUSEUM OF COMPABATn^E ZOOLOGY ence of a cirealar bole through the doublure (PI. 3, figs. 9, 10; PI. 10, figs. 21, 24; PI. 13, figs. 9, 11; PL 14, fig. 12; PI. 23, fig. 15), situated ou or near the sharp flexure associated with the shoulder. This hole is seen in species belonging to several genera, but is not exhibited by all species of a single genus — for ex- ample, Diacanthaspis. A similar-appearing opening in the ce- phalic and thoracic doublure of other trilobites is the Pander- ian opening, but whether or not this similarity implies that they are analagous structures is uncertain. The macula is not discernible in the odontopleurid hypostomes here studied, except perhaps as the smooth area at the inner end of the middle furrow of Pnmaspis ascitus, n. sp. (PI. 1, fig. 20). The depression in the antero-lateral corner of the middle body is conspicuous, however, in that it is smooth on the external surface, and the quartz is often differently coloured, and on the inner surface there may be some exfoliation. Thus it resembles the similarly-situated area in Sphaerexochus (Whittington and Evitt, 1954, p. 25) and may be an area of muscle attachment. The problem of the manner in which the hypostome is attached in odontopleurids has been discussed previously (Whittington, 1941, p. 516; Whittington and Evitt, 1954, pp. 20, 55-56). Iso- lated rostra of the silicified species have not been found, but the unique specimen of Acidaspis cincinnafievsis recently described ("Whittington, 1956b) has the rostrum in place. The posterior edge of this rostrum is straight (except distally where it curves to meet the inner edge of the cheek doublure), and obscurely bevelled in this specimen. Measurement shows that the straight edge fits exactly against the straight, bevelled, anterior edge of the hypostome. The fit of the bevelled edges suggests that the hypostome lies approximately in the horizontal plane (on the orientation adopted here), or slopes slightly downward and backward. The tip of the small anterior wing lies beneath the intersection of the axial and preglabellar furrows, just in front of the eye ridge. This intersection, and the border furrow just outside it, is usually deepened in odontopleurids, though it does not form a large projection on the inner surface. Thus the attachment of the hypostome seems to be as previously suggested — at the suture and vnth a muscular link between anterior wing and/or pit at anterolateral corner of tbe middle bodv to WHITTINGTON : SILICIFIED ODOXTOPLEUEIDAE 173 unspecified point or points on cranidiuni. One of the latter is perhaps the preglabellar furroAv and border furrow outside the inner end of the eye ridge (i.e. about where the anterior boss is in other trilobites). The hypostome of Apianurus harhafus n. gen., n. sp., is restored in approximately this position in Text- figure 20. This attachment appears not to have been as rigid as that in cheirurids and allied families (Whittington and Evitt, 1954, pp. 19-21), and the evidence cited bj' Prantl and Pfibyl (1949, p. 196) of many specimens of Miraspis mira in which the hypostome is preserved Init rotated through 180° and lying just in front of the cephalic margin, is interpreted by them as indieating an ''easily movable connection." However, the flat surfaces that are apposed along the hypostomal suture are not suggestive of movement at the suture having been possible during life. Antennular notch: Figures 5, 7, of Plate 17 shoAv that there is a -conspicuous notch in the anterior cephalic border immedi- ately outside of where the connective suture crosses the border. It has previously' been argued (Whittington and Evitt, 1954, p. 20) that the antenuule is attached to the surface of the anterior boss and passes througli the lateral hypostomal notch. It may also have passed through the notch (here called the antennular notch) in the anterior cephalic border, W'hich would have per- mitted the antennule to be forwardly extended, when the cephalic margin w^as resting on the sea bottom, in the manner suggested under "Mode of Life." The antennular notch is developed to a greater or lesser extent in most odontopleurids (e.g. PI. 1. figs. 2, 5; PI. 12, fig. 2; PI. 22, figs. 3, 6). Thorax The eight thoracic segments of Leonaspis n. sp. (Whitting- ton, 1956b ) are unusual, nine or ten being most commonly seen in the family. The backwardly-couvex curve of the posterior margin of the prominent axial ring (and the occipital ring) is complemented by the curve of the articulating furrow. Medially, the ring adjacent to this furrow is less elevated. The articulating half ring- is long (sag.), particularly that of the first segment. In some species and genera the antero-lateral part of the axial ring is inflated, e.g. SeJenopeltis (Text-fig. 25), Miraspis (Text- 174 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY fig. 15), and Proceratocephala (Whittington, 1956b, Pi. 60, figs. 2, 10). A posterior band of the type seen on the occipital ring in some genera has not been seen on axial rings. The horizontal pleurae may be nnfurrowed as in Apianurus n. gen. (PI. 18, figs. 6, 9) and Ceratocephala (Whittington and Evitt, 1954, PI. 8, fig. 10), but are generally divided by the straight, transverse pleural furrow into a narrow anterior and wider (exs.) posterior band. The narrow, raised ridge of the posterior pleural band of Selenopeltis, with its strong curve convex forward, is distinctive. A posterior flange is present in many species. At the fulcrum the pleural bands are continued by spines: a stout, more or less horizontal, posterior spine, and one or more slimmer, shorter, anterior spines, downwardly di- rected. The doublure is curled under at the base of these spines, and forms the fulcral articulating socket and process. In odonto- pleurinid genera the broad-based librigenal spine is directed horizontally outward and backward beside the thorax, and con- sequently the first two or three thoracic pleurae are facetted so that they fit beneath the posterior cephalic border and libri- genal spine. Thus the anterior pleural spine is absent, and the posterior pleural spine short (e.g. PI. 3, fig. 11; PI. 8, fig. 12). The anterior pleural spines of the third or fourth, and successive, segments are present and the posterior pleural spine is long and backwardly directed. In miraspinids (Text-figs. 15, 18) and apianurinids (Text-fig. 19), however, the librigenal spines are respectively directed upwardly, and situated far forward on the cephalon, and no facetting and tnnication of the thoracic pleural terminations is necessary to permit fitting behind the cephalon. Thus the posterior pleural spines are successiveh' directed obliquely forward, outward, and obliquely backward. This difference in development and direction of pleural spines has been regarded by Prantl and Pfibyl (1949) as indicating a distinction of family rank between odontopleurinids and mira- spinids, but hardly seems to be of this magnitude (cf. Whit- tington and Evitt, 1954, p. 52). Likewise the "stunting" of the first two or three thoracic segments hardly seems an adequate basis for the subgenus Leonaspis {Keftneraspis) Prantl and Pfibyl (1949, p. 165; see Whittington, 1956b). An inflation of the posterior pleural haiul at the fulcrum WIIITTINGTON : SILICIFIED ODONTOPLET-RTDAE 175 characterises certain odontopleurinids, e.g. Priniaspis (PI. 1, figs. 11-14), some species of Leonaspis, and Acidaspis (Whitting- ton, 19561)) and may indicate relationship. Pygidium Major border spines, i.e. one notably larger pair among the several arising from the pleural regions, are characteristic, but may not be developed at all {Diacanthaspis secrefus n.sp., PI. 7, fig. 10; Badiaspis radiata, see R. and E. Richter, 1917, text-fig. 10) or only slightly developed, as in some species of Cerafoce- phoJa (Barrande, 1852, PL 38, figs. 5, 18). When present, the major spine is connected at its base to the first axial ring by the pleural ridge which runs across the pleural region. Commonly, the major spine arises from the upper surface of the pygidial border, but may take its origin from inside the border of the pleural region (e.g. Diacanthaspis orandensis n.sp., PI. 10, figs. 17-19; Apianurus harbatns n.gen., n.sp., PL 18, fig. 12; Caliper- nuriis insolitus n.gen., n.sp., PL 24, figs 27, 28). In this latter case it seems quite independent of the border spines, though the pleural ridge is present. When the major spine arises from the border it is commonly the third or fourth spine, and in Miraspis is unusually far forward on the border, though still the third spine (cf. Prantl and Pfibyl, 1949, pp. 128-129). As remarked under "Locus of segmental divisions" the major spine and pleural ridge are homologous with the thoracic posterior pleural band and spine. The doublure of the pygidium is curled under, not wide, and there is a small projection posteriorly, directed up toward the axis. External Surface Short, thorn-like spines, tubercles, and granules are character- istic; their form and arrangement are shown by the plates (cf. Whittington and Evitt, 1954, Pis. 6-9, 25, 26). Rarely is the external surface everywhere smooth, as it seems to be in Primaspis kcyserlingi, but commonly the deeper parts of furrows, and the doublure, are smooth. A symmetrical arrangement of the prominent spines, making transverse rows across the exoskeleton, 176 BULLETIX : MUSEUM OF COMPARATIVE ZOOLOGY is typical of early developmental stages (see below) and may persist into the adult. In most adults this symmetry is less obvious, and may be lost altogether (e.g. Primaspis ascitus, PI. 1, fig. 1). Conspicuous paired spines on axial rings are common, though rarely are they as large on the thoracic rings as in Proceratocephala (Whittington, 1956b, PI. 60, figs. 2, 5, 10). It was shown previously (Whittington and Evitt, 1954, pp. 56-59, text-fig. 1, etc.) that small openings occur on the distal side of tubercles scattered along occipital and librigenal spines, and at and near the tips of spines on the cephalic and pygidial border. Such openings are characteristic of the specimens de- scribed here (e.g. PL 7, figs. 12-11; PI. H, fig. 16; PI. 13, fig. 14; PL 23, fig. 6; PL 24, figs. 21, 24). It was further suggested that in Ceratocephala laciniata there may have been a single opening, possibly occupied by a hair, at the tip of the thorn-like spines (Whittington and Evitt, 1954, p. 59; PL 8, fig. 11; PL 9, figs. 2, 4; text-fig. 1) scattered over the external surface of the cranidium. Among the few available cranidia of 3-. 'rarispina n.sp., however, is one (PL 15, fig. 28) in which the thorn-like spines, where not obviously broken, are closed at the truifcated tip by a plate in which there are minute depressions (appearing as dark spots in the photograph). The diameter of these de- pressions is less than that of some of the quartz grains replacing the exoskeleton, so that it is difficult to be sure of their nature. It seems probable that they are the orifices of canals through the exoskeleton. The thorn-like spines on the cranidium of Diacan- thaspis are similar, sometimes showing an evenly eilt off tip with a relatively large central opening (PL 7, fig. 15). A well- preserved specimen of Z>. scitulus n.sp. (PL 13, figs. 16, 17), hoAv- ever, has the truncated tip covered by a convex plate in which there are tiny depressions. These may show a regular arrange- ment of a central larger pit surrounded l)y smaller pits. A speci- men of D. lepidus n.sp. (PL 4, fig. 22) shows a similar arrange- ment. In Z>. cooperi and D. orandensis n.sp. (PL 9, fig. 6; PJ. 11, fig. 19) the tip is slightly expanded and subspherical, with a large central depression or opening surrounded by smaller ones. In B. aff". ulrichi n.sp. (PL 9, figs. 7, 9) some spines show only a central opening, and others appear to show several openings. One may conclude that in DiacantJiaspis, as in C. rarkpina n.sp.. WHITTIXGTON : SILJCIFIED ODOXTOPLEUEIDAE 177 the tip of the thorn-like spines was pierced by one or more minute openings, but whether a hair issued from each opening is uncer- tain. The tubercles of Primaspis ascitus n.sp. (PI. 2, fig. 21) appear to be closed at the summit, and in Apionurus harhatus n.gen., n.sp. (PI. 18, figs. 19. 22) the characteristically curved tips of the thorn-like spines appear closed. In Diacanthaspis lepidus n.sp. (PI. 5, fig. 1), D. secretus n.sp. (PI. 6, fig. 22), and D. ulrichi n.sp. (PI. 8, fig. 27) the tip of the librigenal spine is hooked. The hooked portion appears to be smooth, without openings, but proximally to the tip the usual type of opening is present. Such openings are numerous at the tip of the libris'enal spines in species which do not have the hook (PI. 13, fig. 14). Tubercles on the external surface, as in Primaspis asciins n.sp. or Calipernurus insolitus n.gen., n.sp.. seem never to have had openings at the summit. The median occipital tubercle in ])oth these species, however (PI. 2, figs. 21, 23; PI. 23- figs. 5, 7, 8), is larger than any other, and has four pits, arranged to outline a square (and rarely a faint, small, median pit) in its low, domed surface. It appears early in ontogeny (PI. 2, figs. 6. 8; PI. 24, fig. 1) and is retained in the largest holaspis known. The pits, appearing as dark spots, do not seem to be the openings of canals through the exoskeleton. The occipital doublure extends close underneath the tubercle and well in front of it in P. ascitus n.sp. (PI. 1, fig. 6). No similar tubercle is seen on the axial rings of thorax, or pygidium. The same type of median occipital tu- bercle has4jeen observed in other species of Primaspis, in Dia- canthaspis hjoperi (PI. 11, fig. 18), D. ulrichi n.sp. (PI. 8, figs. 24, 30), D. aff. ulrichi n.sp. (PI. 9, figs. 8, 9), Z). orandcnsis n.sp. (PL 11, fig. 20), and Leonaspis leonhardi Barrandc This tubercle may be compared with the median glabellar tubercle of Tretaspis seticornis (Stormer, 1930. pp. 85-87, figs. 36, 37), which is siudlar but displays more clearly a fifth, centrally situated pit, as does the four-celled sense-organ of Anaspides (Hanstrom, 1934). Temple (1952, pp. 254, 258) has described the median occipital tubercle of protaspid and early meraspid stages of Dalmanifina olini as bearing five tiny tubercles, one central and the others forming a square. This may be an anala- 178 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Q Odontopleuro I I I 52 a. V) o c o u o Q D a q: Apianurus ii'H o o I WHITTINGTON : SILICIFIED ODONTOPLBUBIDAE 189 Miraspinae Sele n 0 pelt inae Number of Genera w o tn 3 =3 C Q. O ) o o D Q> c O, ^ ^^ o I hi 190 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY four groups have much in common, each has its distinctive characters, one of the most important being the type of hypo- stome (Text-fig. 2). There is no comparable diA^ersification after this, for only two subfamilies survive the Ordovician and no new- ones arise. In later Middle Ordovician times three (and possibly a fourth) new genera and one new subgenus arise (making a total of at least eight in the Middle Ordovician), and two new ones appear in Upper Ordovician times. During the Silurian period five new genera arise, and in the Devonian four. In the Middle and earliest Upper Devonian the nine surviving genera become extinct, this extinction being almost as rapid as the early diversi- fication. Text-figure 3 shows the variation in time of number of genera and reflects the abrupt appearance and extinction of the group, though not the maximum morphological diversity of Ordovician time. A striking feature of odontopleurid evolution is the long range in time of certain genera — Ccratocepliala endures for some 100 million j'ears, and each of four other genera, Primaspis, Leonaspis, Acidaspis and 3Iiraspis, for some 50-70 million years (on present estimates of absolute time). Other genera appear to have a relatively much shorter range. It is suggested here (Text-fig. 3) that certain of these long- ranging genera provided a root-stock from which the short- ranging genera were derived. Ceratocephala affords the best example, the other miraspinid genera differing from it in one or more morphological characters (e.g. the big paired occipital spines of Dicranurus, or the inflated median glabellar lobe of Whittingtonia) but retaining the main structural plan. Not all the seemingly derived genera are short-ranging, e.g. MirasjDis. though this genus may prove to be rather a closely related but distinct root-stock. Acidaspis and Primaspis-Leonaspis seem to be the root-stocks of the Odontopleurinae. The other two sub- families are relatively short-lived, not so diverse, and the Seleno- peltinae especially seem an aberrant oft'-shoot. If any general picture is afforded by the pattern of odonto- pleurid evolution, it seems to be one of persistent main themes and relatively brief appearances of variations on these themes. Some of these variations, such as large, paired or single, occipital spines, major pygidial spines, marked inflation of certain glabel- lar lobes, appear more than once during the history of the family, and in different combinations, suggesting a process of shuffling WHITTTXGTOX : SILICIFIED ODONTOPLET'RIDAE 191 of inorphologrical characters. The stratigraphical position of each of the species of Diacan- fhaspis is shown in Text -figure 8. There appear to be three main strands within the evolving plexus of species, viz. oran- densis-cooperi, Icpidus-secretus-scitulus, and ulrichi-afl. idriclii. The species are similar to each other, the distinctions being size and length of median and/or paired occipital spines, outline of glabella and degree of inflation of lobes, presence or absence of, and position of origin of, major pygidial spine, size and distri- bution of spines on external surface, etc. The illustrations show that the.se characters are relatively minor distinctions within a major framework, and that they appear and disappear, forming (.litterent combinations. Thus the species of Diacanthaspis show the same evolutionary pattern in a part of the family as in the whole — the presence of several main themes, occurrence of minor variations on these themes,* and shuffling and recombina- tions of characters within the plexus. No other genus is repre- sented in the collections by as many species as is Diacanthaspis. In other genera represented by more than one species no particu- lar trends of morphological change are evident. Turning to consider the rate of morphological change among the silieified trilobites, it is seen to be variable in different lines during lower Edinburg to Oranda time — Apianunts harhafus n.gen., n.sp., shows no change, and there is little in the Dia- cantliaspis idricin — aff. nlrichi line, while the D. lepidus-secre- tus-scitulns line shows considerably more. Cerafocephala is present in the pre-Edinburg Lincolnshire limestone (C triacan- theis Whittington and Evitt, 1954), a second species C. laciniata is common in the lower Edinburg, and C. rarispiiia n.sp. is rare in the Oranda. Thus there is change in this presumed line of descent, though no particular trend is evinced. An apparent oft'-shoot from this line in the lower Edinburg is C. {Ceratocepha- lina) iridens n.subgen., n.sp. h\ Oranda to lower Martinsburg time there is slight morphological change in the Diacanthaspis orandensis-coopen line. The ontogeny of certain fossil animals has been interpreted as showing recapitulation of ancestral adult morphology, but no examples of this process have been adduced from trilobites. Following Stiibblefield's .suggestion (1936), Stflrmer (1942), 192 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and most vigorously Hupe (1954), have urged that paedogenesis or neotony is an important process in the evolution of trilobites. In the case of Diacanthaspis we know the development in detail of the oldest and youngest species (Text-figs. 9, 11; Whittington, 1941, text-figs. 2-6), and something of that of other species (Text-fig. 12). However, not more than one complete ontoge- netic series is known from any one of the three lines of descent in DiacantJiasph. Thus we do not have sufficient evidence to reveal whether or not recapitulation or paedogenesis are im- portant processes in the evolution, but there are no indications of either being operative. What is clear is that the earliest developmental stages of the various species are remarkably similar, though even at this stage specific differences may be observed. As development proceeds through the meraspid stages the peculiar specific characters of each (long, paired, occipital spines, stout, median, occipital spine, shape of eye lobe, glabellar lobation, etc.) become increasingly evident. Thus the develop- ment of species of Diacanthaspis affords an excellent example of von Baer's laws (De Beer, 1951, p. 8). Comparison of the ontog- eny of Diacayiihaspis with that of Apianurus (Text-fig. 22) and of CeraiocepJiaJa (Text-fig. 16; Whittington and Evitt, 1954, text-fig. 16) shows that the early meraspid stages are remarkably alike, though generic and subfamily differences are quite evi- dent. As development proceeds the divergence between genera increases. The protaspides of the Middle Ordovician odonto- pleurids suggest that, if paedogenesis were an important evolu- tionary process, we might expect, for example, holaspid Devonian odontopleurids to bear fixigenal spines, or to have only basal lateral glabellar lobes, or extremely short, paired, occipital spines. These expectations are not fulfilled. A clearer understanding of the evolutionary processes may be possible when ontogenetic series of later Ordovician, Silurian, or Devonian odontopleurids are known, but present evidence seems not to favour recapitula- tion or paedogenesis, and seems to suggest that the importance of this latter process in the evolution of trilobites in general may have been overestimated by Ilupe (1954) and others. WHITTINGTON : SILICIFIED ODONTOPLBITBIDAE 193 PART II: SYSTEMATIC DESCRIPTIONS Family ODONTOPLEURIDAE BURMEISTER, 1843 Diagnosis: Cephalon convex, so that postero-mediau rpfjion stands high above antero-lateral marjiin. Glabella with maximum width generally at occipital ring, sub-parallel sided or tapering forward, occipital ring may be elongated (sag.) and prominent, lateral lobes may be present, median and/or paired spines or tubercles characteristic; 2-3 pairs of lateral glabellar lobes. Cheek convex, inner posterior corner merges with antero-lateral corner of occipital ring; eye lobe prominent, situated centrally on cheek or inside and behind this point, eye ridge present, anterior branch suture runs forward and inward, posterior branch outward and backward to cross posterior margin, sutural ridges characteristic. Librigenal spine usually stout, broad base merging with postero-lateral borders; commonly row of shorter spines arising from oiTter edge of border of free cheek, y>vo- gressively shorter anteriorly ; antennular notch in border of free cheek adjacent to anterior branch of suture. Rostrum short (sag. and exs.) and wide (tr.L llypostome of width greater than, or equal to, length, postero-lateral margins rounded, convex middle body, faint middle furrows running backward and slightly inward from depression at antero-lateral corner of middle body, lateral and posterior border of moderate and similar width, small anterior Aving, no wing process, posterior wing extremely small. Thorax of 8-10 segments, convex axial ring and long articulat- ing half -ring; pleura horizontal, gently or moderately convex, undivided or divided into two bauds, long pleural spine on posterior band, directed outward and progressively more strongly backward posteriorly, short pleural spine or spines on anterior band, directed outward and downward. Either or both pleural spines maj* be missing from the tirst (and sometimes seccnid ) segment because of the size of the facet. Rygidiuin short, sub- triangular in outline, 2 axial rings, may be faint third ring ; row of horizontal border spines, one pair often larger than remainder, may be upwardly directed, and with base connected by low pleural ridge across pleural region to first axial ring; or major spine may arise from pleural region. External surface of exoskeleton rarely smooth, generally with 194 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY thorn-like spines or tubercles, granules between them, arrange- ment of larger spines or tubercles may be symmetrical about mid- line. Scattered openiiitis occur near tips of large spines (libri- genal, posterior pleural, major pygidial), and sometimes also at tips of thorn-like spines or tubercles. Latter may, however, be closed at tip, or may exhibit several tiny openings. Four tiny depressions, arranged at corners of square, maj^ occur at summit of median occipital tubercle. Doublure narrow, curled under. Appendifers not developed. Geological Range: Late Lower Ordovician (late Canadian or Arenig) to early Upper Devonian. Discussion of Systematics of Odontopleuridae Since Prantl and Pfibyl published (1949j their new arrange- ment of the Odontopleuridae, it has been followed, with some emendation, by Erben (1952a, 1952b), and adopted with reserva- tions by Hupe (1953). Whittington and Evitt (1954, pp. 52-53) voiced some criticisms, and these and others are explained in detail here. Study of the silicified material, and of type speci- mens of most genera (Whittington, 1956b), has provided a wealth of new data, and my taxonomic scheme is summarized in Text-fig. 3. This adds a new group, but retains the three main groups of Prantl and Pfibyl, assigning to them subfamily rather than familv rank. 1 consider the homogeneitv of the odonto- pleurids, as well as the relatively small size of the group, sug- gestive rather of family than superfamily rank (cf. Hupe, 1953, p. 230). The similarity of early larval stages of different genera may be interpreted as indicating a monophyletic origin for the family and not "at least diphyletic" as maintained by Prantl and Pfibyl (1949, p. 131 ). in subdividing odontopleurids those authors laid emphasis on the form of the cephalon, and particu- larly the hypostome, the '■different origin of the genal spines," and the direction of iIk^ i)osterior pleural thoracic spines. With some of these 1 agree, while others seem unimportant. 1 have based the .subfamilies here used principally on the following : — a) Form of the hypo.stomes, which, while all are of a type peculiar to the family, seem to fall into four sub-types corres- ponding with the subfamilies (Text-fig. 2). b) Form and inflation of the glabellar lobes and occipital ring. The characters of this region are of prime impni-tance in sub- WHITTIXQTO>r : SILICIFIED ODOXTOPLEt'RIDAE 195 family divisions, and their nature is summarized in the diagnoses. c) Form of the cheek and position of eye lobe, which together help to determine the course of the dorsal facial suture and the angle between the two branches. Such features as type of occipital spine or spines, form of thoracic segments, direction of posterior pleural spines, sliapc of pygidium and development of major border spines seem to me to be of less importance. Whittington and Evitt (1954, p. 32) suggested that the librigenal spine of Ceratoccphnla did not arise in a fundamentally different way from that of Diacan- thaspis, and further evidence for this view is given here. Thus 1 do not accept the suggestion that a major systematic division may be based on the supposedly different origin of this spine in different genera (Prantl and Pi-ibyl, 19-19, p. 131). A forward direction of the posterior pleural spines of the first two or three thoracic segments is observed in some miraspinid and in apianurinid genera. In these forms the librigenal spines are well outside the distal parts of the anterior thoracic segments, and so this direction of the pleural spines is possible. In odontopleurinid genera the librigenal spine is sometimes broad- based, and is directed back just outside the distal parts of the anterior thoracic segments. Hence these segments are facetted, and may lack pleural spines, or they are short. This difference in direction of the pleural spines of the anterior thoracic seg- ments was regarded as important taxonomically by Prantl and Pfibyl, but it is really dependent on the form of the cephalon. and in my view of little taxonomie value (cf. Whittington and Evitt, 1954, p. 52). Subfamily ODOXTOPLEURIXAE BURMEISTER, 1843 (=Odontopleuridae of Prantl and Pfibyl, 1949) Diagnosis: Greatest width of glabella at occipital ring, taper- ing forward slightly or moderately : occipital ring may be elongated, median or paired tubercles or spines. Eye lobe far back, variable distance out across cheek, angle between two branches of suture near eye lobe 90-120". Librigenal spine slim to stout, row of border spines on cheek Avell-developed. Hypo- stome (when known) slightly wider than long, middle furrow commences in front of mid-length and runs inward at a low angle ; shoulder sharp. Lateral notch shallow. Pleurae of thorax 196 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY with broad (exs.), convex, posterior band, continuous with stout posterior pleural spine. Pygidium usually with major pair of border spines, 1-2 pairs small border spines between them. Geological Range: Middle Ordovician to early Upper Devonian. Discussion: Prantl and Pribyl (1949, pp. 135, 151, etc.) di- vided this group into two parts, the first (their Odontopleurinae) including Odontopleura, Acidaspis, rrimaspis, Radiaspis and DiacantJiaspis. In this group the anterior branch of the facial .suture is said to run just outside the eye ridge, so that no sub- triangular area is enclosed between suture, eye ridge and the anterior cephalic margin. In the second group (their Acanthalo- minae) which included Leonaspis and Dudley aspis, a striking subtriangular area was claimed to be developed. Text-figures 4, 5, 7, 10, 13, 14 suggest that such an area is variously developed in all these genera, and that it scarcely aft'ords a basis for a sub- family division. An additional distinction between the two groups was the "configuration of the occipital ring and the origin of the occipital spines" (Prantl and Pribyl, 1949, p. 135). 1 do not regard such differences as of this high a value. An example of our different points of view is oui- treatment of Dudleyaspis. Prantl and Pribyl (1949, pp. 171-172) regarded this genus as allied to Leonaspis because of the presence of the subtriangular area enclosed between the anterior branch of the suture and the eye ridge. This area seems to me to be no larger than that in Acidaspis, and the glabella lobation and cephalic form of these two genera are extremely similar, and I therefore regard them as allied, and Dudleyaspis as derived from Acid- aspis. Cienus Odontopleura Ennnrieh, 1839 Text-figure 4. Type Species: Odontopleura ovata Emmrich, 1839 {=0. prc- vosti Barrande, 1846). See Prantl and Pribyl, 1949, p. 135. Discussion: Only the type and one very similar species, from the \^'e^lockian of Europe, are known. Odontopleura is dis- tinguished from Leonaspis (Text-fig. 7) by the larger, more elevated, lateral glabellar lobes (particularly the basal pair), the elevation and elongation of the median part of the occipital ring, which bears paired spines, the faintly defined occipital lobes. WHITTIXGTON' : SILICIFIED ODOXTOPLEURIDAE 197 the smaller eve lobes, situated farther outward and forward, and consequent different direction of the anterior branch of the suture, the slim librigenal spine and more numerous lateral border spines, the long anterior pleural spine and slimmer, less backwardly directed posterior pleural spine of the thorax, and the relatively wider (tr.) pygidium. In glabellar lobatiou. convexity of cheeks, position of eye lobes, course of facial sutures, etc. Odontopleura approaches Figure 4. Odontopleura ovata Emmrich, Middle Silurian, Motol Beds, ea2, Bohemia. Eeeonstruetion, dorsal view, X IVs- Based on MCZ 4170, Butovice; MCZ 4164, Lodenice. Primaspis (Text-fig. 5), but is distinguished by the form of the occipital ring and stouter occipital spines. Prantl and Pfibyl (1949) reproduced some of Barrande's drawings and also new photographs, but all these illustrations are of specimens flattened in shale, as is that of Hupe (1953, fig. 134). In my drawing I have endeavoured to portray the convex cephalon in its true relation to the thorax, using uncom- pressed specimens from Butovice, Bohemia. 198 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Genus PkimasPIS K. and E. Kichter, 1917 Text-figure 5. Type Species : Odontopleura primordialis Barrande, 1846. See l^rantl and Pfibyl, 1949, p. 144. Diagnosis: Glabella with small third lateral lobes, greatest width across occipital ring and basal lobes, occipital ring not Figure 5. Frimaspis priinordialis (Barrande), Middle Ordovician, DraboT quartzites, d5, Bohemia. A, B. C, cranidiuni, anterior, right lateral, and dorsal views respectively, X 3. D, pygidium, dorsal view, X 3. Based on MCZ 4139, Drabov. greatly inflated or lengthened, bearing median tubercle or paired spines, small occipital lobes. Eye lobe far back and at about half width of cheek. Lateral border spines short, librigenal spine broad at base. Thorax of ten segments, posterior pleural band inflated at fulcrum and continued into stout posterior pleural spine ; anterior pleural spine small. WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 199 Discussion: This genus is represented in Bohemia by species ranging: in age from Upper Llandeilo to Ashgill, and perhaps Silurian. Primaspis heyserlingi (Barrande, 1846) shows the characteristic features of the genus including the anterior pleural spines, which are not lacking (as stated by Prantl and Pfibyl. 1949, p. 14), but is unusual in that the lateral glabellar lobes are fused, there are no occipital spines, and the external surface is smooth. In North America, Primaspis is represented by P. ascitus n.sp. (described below), /'. trentonensis (Hall, 1847) (Whittington, 1941, p. 502, PI. 74, figs. 31-37; Prantl and Pribyl, 1949, p. 149), and probably by P. crosotus (Locke, 1843) from the Eden of the Cincinnati district. All these species have a median occipital tubercle, but not paired spines. In the size and convexity of the basal and median lateral glabellar lobes, position of the eye lobe, course of the facial sutures, form of the pygidium, etc., such species as P. crosotus approach typical Leonaspis, and Prantl and Pribyl (1949, p. 146) have remarked on the Primaspis features displayed by Leonaspis coronata and L. deflexa. This seems to me to point to the derivation of Leonaspis from Primaspis, as suggested in Text-figui'e 3. The close relation of Odontopleura to these genera has been discussed. Text-figure 3 also suggests that Acidaspis (Text-tig. 13; Whit- tington, 1956b) is related to Primaspis. Such a species as P. trentonensis, for example, has a strikingly similar thorax and pygidium, with the posterior pleural bands inflated at the ful- crum, and while the cephalon does not display the special fea- tures of Acidaspis, it resemliles it in having the median lateral glabellar lobe considerably smaller than the basal, in convexitj" of cheeks, position of eye lobe, course of facial sutures, etc. i'Ki]MA«pis ASCITUS Whittiugton, n.sp. Plates 1 and 2 ; Text-figure 6 Uolotype: USNM 116515 (Plate 1, figures 1, 2, 5, 6) ; locality 10. Other Material: Paratypes USNM 116516a-h; all figured speci- mens in USNM. Geological Horizo)i and Localitij: Lower Martinsburg .shale, locality 10. Description : Cephalon crescentic in outline, moderately convex 200 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY longitudinally, strongly convex transversely. Glabella widest across basal lobes, slightly narrower at occipital ring, narrowing rapidly- anteriorly, w^here it overhangs the border slightly. Occipital ring longest (sag.) medially, where it slopes gently forward to shallow occipital furrow ; laterally, behind basal glabellar lobe, a low occipital lobe is developed, defined most clearly on the inner side by a shallow depression. Low median occipital tubercle. Occipital furroAv deeper laterally. Three pairs of lateral glabellar lobes, basal (first) largest, sub-oval in outline, inflated and standing above part of median lobe separat- ing them; second lobes inflated, defined by basal and second furrows directed inward and backward at about 45°, the basal furrow deepest at the inner end ; inner ends of second and basal furrows joined to each other and to occipital furrow by shallow longitudinal furrow; anterior (third) lateral lobes with faint independent convexit}', defined by a .short, shallow furrow run- ning inward from the antero-lateral corner of glabella. Axial furrows faint beside glabella, where fixed cheek merges with antero-lateral corner of occipital ring, clearly defined beside lateral lobes, dying out anteriorly where eye ridge merges with glabella. Narrow anterior border separated from glabella by shallow furrow. Cheek most strongly convex in inner corner inside eye lobe, which is situated at about half width (tr.) and close to posterior margin, opposite posterior portion of basal glabellar lobe. Palpebral lobe short, low, defined on inner side by deep palpebral furrow which curves around behind eye lobe and dies out. Anteriorly palpebral lobe and furrow merge with broad eye ridge and furroAV on inner side. Eye surface \\ith facets faintly discernible on inner side (PI. 2, fig. 22). Anterior branch of suture runs forward and then curves inward a short distance outside eye ridge, crosses border furrow on low sutural ridge and meets rostral suture at edge of border. Posterior branch runs in "S" curve back from eye lobe, out across cheek, crosses border furrow on sutural ridge, and runs over border to doublure beneath base of librigenal spine. Anterolateral border of cheek moderately convex, margin curves forward just outside anterior branch of suture. Row of short border spines, increasing in length posteriorly, directed down and slightly out from lower .surface of border. Posterior border narrow (exs.) beside occip- WHITTIXGTON : SILICIFIED ODONTOPLEURIDAE 201 ital ring, widening and swelling outward to merge with the broad base of the librigenal spine. Doublure of cephalon of same width as border. Low projections formed by inwardly projecting outer part of occipital furrow and the three glabellar furrows, particularly the deeper, inner part of the basal furrow. Small depression, with sharply raised anterior edge, in doublure in front of librigenal spine (PI. 1, fig. 4), is the socket for fulcral process of anterior segment of thorax. Rostrum unknown, but evidently a transverse plate on outer edge of anterior border, perhaps curled under at hypostomal edge. Hypostome slightly wider than long, gently convex middle body, middle furrow running backward and inward a short distance from circular depression at antero-lateral corner, macula faintly defined by convexity, tips of crescentic posterior lobe slightly inflated. Nar- row anterior border not defined by furrow ; lateral border with small, triangular anterior wing directed dorsally, small lateral notch and sharp, swollen shoulder projecting outward; postero- lateral border widest, gently convex, small median projection in faint, wide median notch. Doublure extending between shoul- ders, of same width as borders, posterior wing a small, twisted projection (PI. 1, fig. 18), inner edge of doublure curled in in medial portion. External surface of exoskeleton tuberculate, ex- cept in deeper parts of furrows, anterior border (which is gran- ulate), inner part of posterior border, median area of posterior lobe of middle body of hypostome and adjacent maculae, and doublure (except median part of hypostomal doublure). Tuber- cles vary in size, larger ones may display symmetrical arrange- ment, including 5 or 6 pairs on fronto-median lobe, some on lateral lobes, and fixed cheek inside eye lobe. Toward tip of genal spine (PI. 2, fig. 10) tubercles become longer and distally di- rected, and there appear to be tiny openings at the base on the distal side of some of them. Other tubercles closed at top. Median occipital tubercle with 4 tiny depressions (not the openings of canals through the exoskeleton) arranged in a square, and some- times a small central pit (PI. 2, figs. 21, 23). Number of thoracic segments unknown. Convex axial ring narrowest in mid-line, where it slopes gently forward to deep articulating furrow. Latter with steep anterior slope, articulat- ing half -ring long (sag.). Horizontal pleurae divided by slightly 202 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY diagonal pleural furrow into gently convex anterior band and strongly convex posterior band. Anterior pleural spine blade- like, shorter on anterior segments, facet at antero-lateral corner of segment. Posterior pleural band becomes more swollen distally at fulcrum, and is extended into spine, which is short on anterior segments, progressively longer and more backwardly directed on succeeding segments. Ring process and socket well developed, but axial socket and process inconspicuous. Anterior surface of pleura is flattened, with a projecting rim along the upper edge, and this surface fits against the flattened surface of the posterior flange, the projecting rim fitting against and above the curved upper edge of the posterior flange (PI. 1, figs. 12, 13). Doublure extends in to fulcrum, and has an anterior projection, the ful- cral process, which fits into the notch (with raised anterior edge to act as a "stop") in posterior part of doublure (PI. 1, fig. 15). Axis of pygidium with strongly convex first ring, artic- ulating furrow and half-ring as in thorax, second axial ring lower and less convex, tip of axis merging into border. Pleural regions flat, bounded by a low border anteriorly, strongest antero- laterally, and a broader, more convex postero-lateral border; crossed by a strong ridge connecting first axial ring to the base of major border spine. Five smaller spines, longest posteriorly, project horizontally from postero-lateral border and one major spine (with a swollen base which merges with proximal part of fourth horizontal spine) arises from upper, outer part of pos- terior border and projects upward and backward. Doublure curled under, as wide as border, at tip of axis most strongly curled and with a slight inward projection, inner, anterior corner forming fulcral articulating process. External surface of thorax and pygidium tuberculate, tubercles densely packed on axial rings, posterior pleural spines and pygidial border spines, less closely packed in articulating furrow and on pleurae and pleural regions, and absent from upper surface of posterior pleural band, articulating half ring, and inner part of doublure. On under side of proximal parts of spines tubercles smaller. Con- spicuously larger tubercles as pair on posterior slope of axial ring, including two rings of pygidium, and a third pair (or row of 3) close together at the tip, suggesting a third segment. Large tubercles also in row along anterior pleural band, and one at WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 203 about half ^Yidth (tr.) of posterior band, on posterior slope. Towards tips of thoracic posterior pleural and pygidial border spines tubercles elongated and distally directed, and some open- ings occur at bases of these tubercles. Elsewhere tubercles closed, and particularly the larger ones appear to expand slightly distally to give a bulbous tip. Discussion: Primasjyis asritus n.sp. is extremely similar to P. trentonensis (Hall; see Whittington, 1941, pp. 501-502, PI. 74, figs. 31-37) from the Trenton, Sherman Fall formation, of New York and Ontario. It differs from P. trentonensis in various minor ways, of which the moi'e obvious are: 1) the greater con- vexity of the posterior part of the cephalon, so that the posterior surface of the basal glabellar lobes and fixed cheeks, particularly, is longer and descends vertically; 2) the inflation at the fulcrum of the posterior pleural band of the thorax is less; 3) the major border spines of the pygidium of P. trentonensis are farther apart, and the base does not merge with the distal part of the minor border spine inside it, but is separate. Also in P. trento- nensis there are two minor border spines in front of the major spine, not three. Tubercles on the external surface are similarly developed ^nd distributed in the two species, and the median occipital tubercle of P. tre7itonensis exhibits the four tiny de- pressions. Paired tubercles can scarcely be distinguished on the axis of P. trentonensis, except on the posterior edge of the two pygidial axial rings. Development: Smallest cranidium of length (sag.) 0.6 mm. Text-figure 6A (cf. PI. 2, figs. 1, 2) summarizes the characters: glabella narrow (tr.), almost parallel sided, basal lobes outlined, faint second lateral glabellar furrows, median occipital tubercle, large median tubercle (with a suggestion of a subdivision indi- cating that it will later divide and form axial spines 2a) opposite basal glabellar lobes, 3 distinct pairs of tubercles in front (axial spines 2-4) ; palpebral lobes about opposite axial spines 3, strong eye ridge running forward to antero-lateral corner of glabella, posterior border well-defined, widening (exs.) outward; tuber- cles on fixed cheeks and border in sub-symmetrical arrangement, including Ai, B, and C. Next largest cephalon known (Text- fig. 6B; PI. 2. figs. 5, 6), length (sag.) 0.85 mm., shows consider- able change. Both the cephalon as a whole, and individual parts, 204 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY are more convex. Fronto-median glabellar lobe iiov steep-sided, standing highest, basal and median lateral lobes well defined and inflated, width of glabella across basal lobes now 35 per cent width of cranidium, rather than 22 per cent as at the smaller stage. Eye lobe opposite anterior part of basal glabellar lobe, fixed ^O O W% n\ O).. o i ol ' B Figure 6. Prima-^pis uitcitK.s n.sp. A, small i-ranidiuiii, e.xterior view, ilrawn from orisiiial of Plate 2, fitruie 2. B, cranidium ami right free ilieek, exterior view, drawn from original of Plate 2, figure 6, and free cheek of appropriate size. X 38. Spines numbered and lettered as in Text- figure 1. cheek inflated at inner corner, which merges with occipital ring. Median occipital tubercle large, low; on posterior margin of occipital ring prominent pair of tubercles, other pairs of smaller tubercles in front of these. Five pairs of tubercles on median glabellar lobe — axial 2a, 2-5 — the most anterior not quite sym- WHITTINGTON : SILICIPIED ODONTOPLBURIDAE 205 metrical, paired tubercles on lateral lobes, larger tubercles on fixed cheeks and border sub-symmetrically situated. A larger eranidium of length (sag.) 3.2 mm. (PI. 2, figs. 11, 12) shows that these trends of change have continued — convexitv of lat- I'ral glabellar lobes and inner part of fixed cheek haA'e increased, and fixed cheek and basal glabellar lobes are level with median glabellar lobe in transverse line. Width across basal lobes now -14 per cent width of eranidium. Eye lobe farther back, paired tubercles, including four on median glabellar lobe, may still be distinguished. In this size of eranidium, third lateral glabellar lobes and occipital lobes are distinct, as are 4 pits in median occipital tubercle. Further increase in size shows slight change continuing, so that width across basal glabellar lobes may become almost 50 per cent of that of eranidium ; basal glabellar lobes are inflated sufficiently to stand above part of median lobe be- tween them (PI. 1, fig. 5) ; posterior slope of fixed cheek becomes steeper. The conspicuous larger tubercles of earlier stages lose their prominence and cannot be readily picked out in larger specimens. The smallest hypostome (PL 2, figs. 8, 4) differs little from larger ones. Only transitory pygidium known (PI. 2, fig. 9) is probably stage holaspid-1, since the portion behind the first segment is like the true pygidium, lacking only the first two lateral border spines. The axis is convex, each ring with a pair of prominent tubercles on the posterior edge — recalling the pair on the posterior margin of the occipital ring. Small true pygidia are like larger ones, but may liave four border spines in front of the major spines, the anterior extremely short. In larger speci- mens this small anterior border spine is reduced, presumably with enlargement of the facet, and disappears. Genus LeoXASPIS R. and E. liiehter, 1917 Text-figure 7. i K. and E. Richter (1952) regard as undesirable the revival of Conrad's (1840) name Acanthaloma by Prantl and Pribyl (1949, p. 151). I am in agreement with them, and have sub- mitted a proposal for the suppression of this name to the Inter- national Commission on Zoological N'omenclature. 206 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Tj'pe Species: Odontopleiiraleonhardil^arvande, 1846. Diagnosis: Glabella as AAide across large basal glabellar lobes as across occipital ring, two pairs glabellar lobes, the anterior Figure 7. Leonaspis new species, Lower Devonian, Haragan shale, Arbuckle Mta., Oklahoma. A, complete exoskeleton, dorsal view. B, eephalon, anterior view. C, hypostome, exterior view. X 2. Drawn from originals of Whittington 1956b, Plate 57, figures 10-16. WHITTIXGTOX : SILICIPIED ODOXTOPLEURIDAE 207 smaller than the basal, occipital ring not greatly lengthened or inflated, median tubercle or short median spine. Eye lobe oppo- site posterior part of basal glabellar lobe or occipital furrow, large to medium size ; two branches of suture almost at right angles adjacent to eye lobe, cross border furrows on sutural ridges. Librigenal spines broad-based. Thorax of 8-10 segments. pleurae with anterior pleural spines short, may be bifid distally, posterior band continuous with stout posterior pleural spine; anterior two or three segments may be facetted and pleural spines reduced. Pygidium usually with prominent pair major border spines, one or two pairs small border spines between them. Geological Range: Lower Silurian to Middle Devonian. Discussion: Elsewhere (Whittington, 1956b) I have described two American species of Leonaspis, and one is illustrated here (Text-fig. 7). Both are very like the type species. I have also discussed the genus, and the unsatisfactory nature of the basis for the two subgenera of Leonaspis recently proposed by Prantl and Pfibyl. Genus DiACAXTHASPIS Whittington, 1941 Type Species: Diacanthaspis cooperi Whittington, 1941. Diagnosis: Glabella widest (tr.) across occipital ring and basal lobes, length (sag.) greater than maximum width; two pairs of lateral lobes, of which the anterior pair is the smaller, sometimes faint, tiny third lateral lobes; occipital ring with paired and median spines, of which one median and/or one pair may be conspicuously longer and thicker than the remainder; small lateral occipital lobes may be present. Eye lobe situated well inwards opposite basal glabellar lobe, sutural and eye ridges distinct. Row of spines on anterolateral cephalic border. Libri- genal spine may be hooked at tip. Plypostome shield-shaped, width (tr.) across anterior wings and shoulders about equal; may be circular hole through doublure of shoulder. Number of thoracic segments unknown; pleurae with narrow anterior and broad posterior bands, latter not inflated at fulcrum. Pygidium 208 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY t/1 "5 re iNJ 9unasNiiyviAi ubmoi UIJ VQNVyO WHITTINQTON : SILICIFIED ODONTOPLEURIDAE 209 to Q, to < z: z < < c o -4-' ^ n 0^ C 1— 1 o J* «f-i > fcc ^ s^ en C5 o ,c T) c a M <» -M •t-t 6 o tn 13 o li g3 S o iNj ounaNias ubmoi > .1— o t^ en /^\ s,*- O P, o o r/1 «> P< *5 CQ • rH P .s a g a 09 o O a qi o 00 r-^ > 0) C) Eh n U be c3 ^ "3 -i^ .9 X 210 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of three segments, two axial rings distinct. Border with 6 or 7 pairs of horizontally directed spines, of which the 4th may be elongated; upwardly and backwardly directed major spine may be present on upper surface of border postero-laterally ; major spines linked b}' ridge to first axial ring. External surface of exoskeleton with median and symmetri- cally situated thorn-like spines on axial and pleural regions, granulation between these spines. Geological Range: Middle Ordovician. Discussion: The diagnosis has been emended to embrace the type and the new species shown in Text-figure 8. A recently described Scottish species (Tripp, 1954, pp. 663-664, PI. 1, figs. 24-29) also lacks large paired occipital spines and major spines on the pygidial border, but appears to belong here. It also has faint third glabellar lobes, and such faint lobes are seen in D. scitulus n.sp. On the other hand, three Esthonian and Swedish species placed in this genus by Prantl and Pfibyl (1949, p. 150) are here placed in Apianurus n.gen.. and I consider it unlikel}' that "Acidaspis" tremenda Barrande, 1852, belongs in Diacanthaspis, as .suggested by Prantl and Pfibyl (1949, p. 150 ). The similarities between DiacaniJiaspis and Primaspis have previously been commented on (Whittington, 1941, p. 502; Prantl and Pfibyl, 1949, p. 150), and the similarities and con- trasts are well illustrated by the Virginia species described here. The third glabellar lobes are clearly developed in Primaspis, and species of this genus have stout major pygidial spines. The infla- tion of the posterior pleural band at the fulcrum and of the outer part of the posterior cephalic border and base of the librigenal spine, are also chai-acteristic of Primaspis, but not of Diacanihaspis. The silicified species, though distinct from each other, have much in common, and they also fall into groups related in time — lepidus — secretus — scitulus, orandensis — cooperi, and ulrichi — aff. ulrichi, as discussed in Part I. These groups give a glimpse of the evolving plexus of Diacanthaspis, and might be regarded as subgenera, though this step is not taken here. The basic unity of all the species is shown by the similarity betAveen the develop- WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 211 mental stages of each, a similarity which is most marked in the earliest stages. The basic unity is also shown by the presence of many common characters which are differently developed in dif- ferent species, e.g. a relatively large median occipital spine is present in the tiny cranidia of D. secret us and Z>. ulrichi, but in the former it loses its prominence in later stages, while in D. ulrichi it is retained to the adult. The development of the paired occipital spines, present in all the young stages, offers another such example. In all the species of Diacanthaspis there is a deep depression separating the basal glabellar lobe from the outer part of the occipital ring, the slope down to the depression from the occipital ring being gentle, that from the glaljellar lobe steep. In D. orandensis and D. cooperi the forward-sloping antero-lateral part of the occipital ring develops a low, subcircular rise, bearing one or two thorn-like spines. This is the occipital lobe, a faintly developed feature peculiar to these two species. The spines and granules on the external surface of the different species are described and illustrated in detail, and have been discussed in general terms in Part I. It should be emphasized that only in the best preserved specimens is the tip of the thorn- like spines seen to be truncated by a disk in which are tiny pits or openings. Their arrangement is not as regular as that of the four pits at the tip of the median occipital spine, and it seems unlikelv that thev are different-sized versions of the same structure. DIACANTHASPIS COOPERI Whittiugtoii. 1941 Plate 3; Plate 9, fig. 6; Plate 11, figs. 16-18; Text-figures 8, 9. Geological Horizon and Localities: Lower Martinsburg forma- tion, localities 9-12. Description: The new material enables the following additions to be made to the original description (Whittington. 1941. pj). O02-508, PI. 74, figs. 1-30, text-figs. 2-6). The eye surface is preserved in some specimens (PL 3. fig. 21; PI. 11, fig. 17), the outer surface with the tiny facets faintly convex, corresponding pits on inner surface deeper. Librigenal spine ends in a sharp point, not a hook (PL 3, fig. 20). 212 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Slightly larger hypostomes (PI. 3, figs. 7, 9, 10, 13, 14) than those described previou-sly have a length (sag.) little more than two-thirds the maximum width across the anterior wings. There is a distinct lateral notch, and the shoulders are sharply pointed, the tip directed ventrally. The doublure is widest near the shoulder, the inner edge extended in a small ventrally-directed posterior wing. There is a circular hole through the doublure at the shoulder. Posteriorly the doublure is narrowest. Orna- ment of granules, but coarse tubercles on tips of crescent -shaped posterior lobe of middle body, and along posterior margin. Segments from all parts of thorax now known (PI. 3, figs. 11, 12, 15, 16) but not complete number. The anterior segment (PI. 3, fig. 11), since it fits behind the cephalon, lacks the anterior, and has an extremely short po.sterior, pleural spine. The articulating half ring is, however, larger and more convex than that of posterior segments. The photographs show the progressive increase in size of both pleural spines in successive segments, as well as the changes in direction. Axis of the pygidium (PI. 3, fig. 19) with two rings and a pair of spines at the tip, suggestive of a third segment. The base of the large, upwardly directed spine on the margin of the pleural region is situated between the bases of the 3rd and 4th pleural spines, and is connected by a low ridge to the first axial ring. Photographs at high magnifications (PI. 9, fig. 6j show the dorsal external surface of the exoskeleton covered with sym- metrically situated thorn-like spines with granules between. The granulation is absent only from the posterior edge of the figure 9. Diacanthnspis coopcri Whittingtoii. A, B, Protaspis, dorsal and ventral views, drawn from the original of Plate 3, figures 1, 2. In A, the right free cheek has been drawn in its correct position rather than displaced as in B and the original. In B, the outline of the jjoorly preserA'ed hypostonie is indicated by a broken line. C, Stage 0 exoskeleton with only right free cheek restored in position. Drawn from original of Plate 3, figure 3, and free cheek of appropriate size. D, small eranidiuni and left free cheek drawn from originals of Plate 3, figure 8. Approximately X 62. Paired spines numbered and lettered as in Text-figure 1; Lb is librigenal. WHITTIXGTON : SILICIFIED ODOXTOPLEl'RIDAE 213 Figure 9 214 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGV occipital ring, the upper part of the posterior cephalic border, corresponding edge of the thoracic axial rings and the posterior pleural bands, the border of the pygidium and the raised ridg(^ on the pleural regions, and the paired occipital, librigenal, and posterior pleural spines. Distally directed openings may be seen on the occipital and librigenal spines, and between and on the granules at the tips of the lateral cephalic and pygiclial border spines openings were present (PI. 11, fig. 16). The thorn-like spines, when not obviously broken, appear bluntly rounded at the tip, with a distinct, tiny central opening surrounded by an irregular ring of tiny depressions or openings (PI. 9, fig. 6). The median occipital tubercle is short and bluntly rounded at the tip, and some specimens show four depressions, arranged in a square, on the slopes of the tip (PI. 11, fig. 18). These appear to be depressions, not openings through the exoskeleton, and resemble those seen in, for example, D. aff. ulrkhi n.sp. and D. orandensis n.sp. (see below). Development: Protaspis (PI. 3, figs. 1, 2, 5, 6 ; Text-fig. 9 A, B) of length (sag.) 0.42 mm., maximum width (tr.) at base of fixigenal spines 0.51 mm. Shield subcircular in outline, gently convex, most strongly so anteriorly ; divided by faint, curving transverse furrow into larger cephalic and smaller pygiclial por- tion ; doulilure narrowest posteriorly. Glabella outlined by broad, shallow furrows, gently convex, a median occipital and three pairs (2, 8 and 4 of Text-fig. 9A) of axial spines; occipital furrow faint, in front of outer parts pair of extremely faintly defined, subcircular, basal lateral lobes. Cheeks slope quite steeply antero-laterally, connected by narrow anterior border; low eye lobe far forward and outward, almost in line with axial spines 4. Eye surface not preserved. Free cheek narrow, the two branches of the suture forming a straight line on the dorsal side, curving across the doublure as shown in Text-figure 9B. Posterior border not clearly defined, extended into long, broad-based fixigenal spine. Short spines or tubercles arranged symmetrically on cheeks and anterior border include Ai, A2, A3, B, C, D, Er, two pairs on the anterior border, and additional small tubercles on the fixed cheek. Five spines on vertical margin of free cheek, posterior (librigenal) the longest. Protopygidium with strongly convex axis, especially posteriorly, bearing two pairs of tubercles. Pleural regions depressed near WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 215 axis, convex toward rolled border ; two stout pairs of border spines, outer pair with spine at base, inner pair directed back- ward and inward. Rostrum and hypostome poorly preserved (PI. 3, fig. 2), general outline as suggested in Text-figure 9B. Exoskeleton at Stage 0 (PI. 3, figs. 3, 4; Text-fig. 9C) of length (sag.) 0.57 mm., width (tr.) at base fixigenal spines 0.6 mm., length (sag.) of cephalon 0.34 mm. A dissociated free cheek of appropriate size has been drawn in position in Text- figure 9C. Glabella parallel-sided, more convex, defined by deep axial furrows ; deeper occipital furrow and ring relatively wider, with large median tubercle and slimmer but longer paired spines ; low median tubercle in front of axial spines 4. Tiny basal lateral glabellar lobe faint, almost concealed by increased inflation. Cheek more convex, eye lobe larger and farther back, course of cephalic suture similar. Posterior border defined distally, fixi- genal spines as long as in protaspis. Tubercles (or short spines) as on protaspis, on posterior border additional spine between Ai and B. Free cheek with librigenal spine stouter and longer than fixigenal, five spines on edge of border, decreasing in size forward. A new row of spines has appeared on the border, outside and below the other, directed outward and downward, extending to base of librigenal spine and increasing in length posteriorly. Transitory pygidium (width (tr.) at anterior mar- gin 0.5 mm.) with well-defined axis, pleural regions fiat with rolled border. Three pairs of spines on axis, and three on borders, inner directed slightly inward. First interpleural fur- row faint, two spines on pleura, outer at base of border spine, third small spine on outer side border spine near base. Upright spine at base second border spine. Smallest cranidium previously described (Whittington, 1941, PI. 74, fig. 1; text-fig. 2) slightly larger than those of Stage 0, and without trace of fixigenal spines. A cephalon of length (sag.) 0.51 mm. is here restored (Text-fig. 9D ; PL 3, fig. 8) and is at same stage of development as original of my 1941, Plate 74, fig. 2 and text-figure 3. Basal part of glabella wider, basal glabellar lobes much larger, axial spines 2a have appeared, as have axial spines 5 (in place of the median tubercle of Stage 0). Eye lobe larger and farther back, but most striking is absence of fixigenal spine, so that librigenal spine is at genal angle. o 216 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY There are no specimens showing any transition (e.g. a reduction of the genal spines) between Stage 0 and this stage. The smallest transitory pygidium previously described (Whit- tington, 1941, p. 508; PI. 74, fig. 11) is probably that of Stage 1, while the larger one (op. cit., p. 508; PI. 74, fig. 12, 13) may belong to a stage between 1 and 8. A new specimen (PL 3, figs. 17, 18) appears to belong to Stage 8 (assuming that the holaspid thorax is composed of 10 segments). Anterior segment defined by interpleural groove, and right posterior pleural spine curves backward and slightly inward. Proximal portions of the posterior pleural spines of second segment may be seen, and inside them there are four pairs of small marginal spines. On the right side (the left is broken off), just inside the margin, is a long upwardly and backwardly directed spine, its base in line with the fourth pair of spines on the axis. The inclination of this spine is different from that of the posterior pleural spines, and it appears to be the upwardly directed major pleural spine so characteristic of the true pygidium. On the posterior pleural bands of the first two segments are two short spines, upwardly directed, the outer at the base of the posterior pleural spine, and forming a curved line between the axial ring and posterior })leural spine. A spine corresponding to this outer spine is pres- ent on the pleural portion of the third segment, just in front of the upwardly directed major pleural spine. The curve be- tween the axis and the base of the major spine corresponds with that of the first two segments, and suggests that the major pleural spine is the modified posterior pleural spine of the first segment of the true pygidium. DIACANTHASPIS LEPIDUS WhittiugtOn, ll.sp. Plates 4, 5 ; Plate 7, figure 15 ; Text-figures 8, 10 Holotype: USN^I 116517 (PI. 5, figs. 1, 4, 7, 9), locality 2. Other Material: Paratypes USNM 116518 a-c; all figured speci- mens in USNM. Geological Horizon and Localities: lower Edinburg limestone, localities 2, 3, 4. Description: At the localities given above, two species of Diacanthaspis occur, D. lepidus n.sp., and D. secretus n.sp., the WHITTINGTOX : SILICIFIED ODONTOPLELTIIDAE 217 latter described beloAv. Material from these localities, collected at different times, has been prepared, and the relatii-e frequency of occurrence of parts of the adult exoskeleton is as follows (counts of complete or nearly complete parts only) : Table 1 Xumbers of Exoskeletal Parts of Two Species of Diacanthaspis at Localities 2-4. Locality 2 Locality 3 Locality 4 Cooper and Evitt and Cooper Evitt Whittington Whittington Collection Collection Collection Collection U. Ii pidu.s h.sp. Cranidia 41 112 45 1^ Free ohe; ks •48 185 20 8 Pygidia 78 1122 12 14 D. secret Its n.sp. Cranidia 57 24 4 14 Free cheeks 75 30 — 15 Pygidia 98 14 — 12 The rarity of D. secretns at locality 3, and the fit of the free cheeks, has led to the associations made here (Text -fig. 8), and support is afforded particularly by the nature of the ornament. No specimens are intermediate in morphology between those here called D. hijidus and D. secretns. and there is little variation in even minor features of the ornament between individuals of either species. It might be suggested that these two morphologi- cal types, occurring together, are sexual dimorphs of one species, rather than two distinct species. The fact that D. secretns is rare at locality 3, but occurs in almost equal numbers to those of D. lepielus at localities 2 and 4, seems to argue against this suggestion. Diacanthaspis lepielus is considerably older than the type spe- cies, D. coo peri, and is distinguished from it at once in lacking the 218 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY WHITTINGTOK : SILICIFIED ODONTOPLEURIDAE 219 stout occipital spines. The cephalon is less convex transversely, and both the width (tr.) between the palpebral lobes, and of the occipital ring, are lesser fractions of the cranidial width. Occipi- tal lobes are absent. Tip of librigenal spine hooked. The pygid- inm of D. lepidus is relatively shorter than that of D. cooperi, and has the elongated fourth marginal spines. That of D. cooperi has a long upwardly-directed spine situated on the border inside the base of the fourth marginal spine, which latter is not elongated. One hypostome from locality 3 is referred to D. lepidus, since this is the common species at this locality. It is of the same general type as that of D. cooperi (compare PI. 3, figs. 7, 9, 10 with PI. 5, figs. 15-17), but different in detail — e.g. the outline of the lateral and posterior margins, depth of lateral notch, projection of shoulder directed laterally rather than ventrally, greater convexity of middle body, and shorter middle furrows. The doublure is incomplete postero-laterally, but at the shoulders it shows no circular opening. The long thorn-like spines on the external surface are hollow, and the tips of even the longest and best preserved are not pointed, but cut off (PI. 7, fig. 15). The end is usually open, but in some is covered by a plate having several tiny holes or depressions in it (PI. 4, fig. 22). Openings, directed distally, occur on the librigenal spines, tips of the cephalic outer border spines, on the posterior pleural spines of the thorax, and near the tips of the pygidial border spines. These openings may have been occupied by sensory hairs. On the cephalon, between the thorn-like spines, the external surface is granulate, except on the convex surface of the borders and occipital ring. Similarly, granulation is absent from the convex surface of the axial rings, the posterior pleural bands, and the border of the pygidium. Development : A series of cranidia down to the length (sag.) U.48 mm. is shown in Plate 4, figures 8-10, 12-14, 16-18. The smallest (PI. 4, fig. 6) is distinguished from that of a smaller size of D. secretus n.sp. (PI. 6, fig. 12; Text-fig. IIC) by the presence of a median, as well as small paired, occipital spines, faint basal glabellar lobes, a less marked indentation in the anterior margin, a slightly different arrano-ement of the spines on the fixed cheek on and inside the palpebral lobe (spines Ao, A3, D, and Er), and a tiny additional spine at the extremity 220 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of the posterior border near the posterior edge, between spines B and C. In cranidia of 0.5-0.66 mm. in length (compare PI. 4. fig. 7, with PI. 6, figs. 14, 15), some of these differences are more prononnced — -the smaller indentation in the anterior margin in D. lepidus, the different arrangement of the spines referred to on the fixed cheek and the additional spine on the posterior border. The first two of these differences persist into larger sizes, and aid in differentiating the species. Another specific character already evident at this size is the different shape of the palpebral lobe. In larger cranidia the differences between D. Icpichis and D. secretiis become more prononnced. All the tiny cranidia with fixigenal spines (probably Stage 0) known from localities 3 and 4 are of the type shown in Plate 6, fignres 6. 7 (cf. Text -fig. IIB), i.e. all the spines are long, with a thick base and blnntly terminated tip, there is a single occipital and axial pairs 2, 3, and 4, and the anterior margin is indented. This tjT)e is so like the cranidial portion of the protaspis from locality 4 shown in Plate 6, fignres 1-5 that the conclusion that they form part of a series is inevitable, and this series is traced into the holaspid D. secrcfus. However, a second pi'otaspis known from locality 4 (PL 4, figs. 1-5) is clearly odontopleurid, and differs from that of D. secretus only in the following characters : (a) the spines on the dorsal surface are low, appearing rather as high tubercles ; ( b ) the occipital ring bears a median spine and a suggestion of small, paired spines just behind the median spine; (c) small basal glabellar lo])es are present; (d) axial spines 5 are represented by tiny tubercles. The characters (a), (b), and (c) are those which distinguish the smallest known cranidium of D. lepidus, the spines on the external surface being less massive than those of D. secretus. Axial paired spines 5 are not known in this size of D. lepidus cranidium, but there is a median axial 5. This second protaspis from locality 4 is thus tentatively regarded as that of D. lepidus, for comparison be- tween it and smallest known cranidia of other species of Dia- canthaspis, and of Ceratocephala, Apianurus n.gen., and Ccdi- pernurus n.gen., reveals far greater differences. DiACANTHASPis SECRETUS Whittington, n.sp. Plate 6 ; Plate 7, figures 1-14 ; Text-figures 8, 11. Holotype: USNM 116519 (PI. 7, figs. 1, 3, 5, 11), locality 2. WHITTINQTOX : SILICIFIED ODONTOPLEt'RTDAE 221 Other Material: Paratypes USNM 116520 a-d; all figured speci- mens in USNM. Geological Horizon and Localities: lower Edinburg limestone, localities 2, 3, 4. Description : The occurrence of this species is discussed under Z>. Icpidus n.sp. The seemingly minor but persistent differences that distinguish D. secretus from D. lepidus, summarized in Text- figure 8, are: 1) the longer, lower palpebral lobe, more strongly curved course of the anterior branch of the suture, and better- defined eye-ridge (compare PI. 5, figs. 1, 4, 7, with PL 7, figs. 1, 3, 5) ; 2) the different direction and arrangement of the four long spines on the fixed cheek between the palpebral and lateral glabellar lobes (compare PI. 5, fig. 1 with PL 7, fig. 1) ; 3) the l)ackward curve of the tips of the outer spines on the border of tlie free cheek, and the lesser number of spines on the upper surface of this border (compare PL 5, fig. 9 with PL 7, fig. 7) ; 4) the relatively longer pygidium, in which the fourth marginal spines are not elongated (compare PL 5, fig. 12 with PL 7, fig. 10). A single small ornamental spine, upwardly directed, is present at the base of the 4th, 5th, and 6th marginal spines; in D. lepidus two such spines are present on the 3rd, 4th, 5th. and 6th marginal spines. Other ways in which the exoskeletal parts of the two species can be discriminated may be seen from a study of Plates 5 and 7. It will be seen that the minor orna- mental spines of D. lepidus tend to be longer and sharper, as well as more numerous in certain regions (compare PL 5, figs. 1, 7, 9 with PL 7, figs. 1, 3). The thorax of neither species is eompletety known, but isolated segments of D. secretns have shorter, blunter ornamental spines, and the lateral barbs on the anterior pleural spine arc fewer (^compare PL 5, figs. 2, 3, lO, 11 with PL 7, figs. 4, 6). The librigenal, cephalic, thoracic pleural and pygidial boi'der spines of the exoskeleton show openings (PI. 7, figs. 12-14) like those described in D. lepidus. The thorn-like spines have the tips cut off and closed, with tiny openings or pits in the end. The granulation is distributed as in Z>. lepidus. Intermediate between the thorn-like spines and granules are high tubercles, and these apparently are closed at the tip. Development: Metaprotaspis (PL 6, figs. 1-5; Text-fig. IIA) 222 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY B >=*• / '^o' ^ % \ '■? ^1 N so o\ Figure 1 I WHITTINGTON : SILICIFIED ODONTOPLEl^RIDAE 223 subcircular in outline, gently convex ; free cheeks, rostrum and h3^postome unknown. Cranidial portion divided by shallow, broad, poorly-defined axial furrows into glabella, with very gentle convexity, and gently convex fixed cheeks. Glabella not divided into rings, but with stout median occipital spine and three pairs in front, a fourth pair on the anterior border, these four pairs forming tAvo lines diverging slightly forward. Pos- terior area of cheek with prominent spines on dorsal surface symmetrically arranged, lettered A,, Ao, A3, B, C, D, in Text- figure 11 A. At margin, at end of curving row formed by median occipital and spines A| and B, is posteriorly directed fixigenal spine. On anterior area of cheek, spines included are those labelled PI, A3, and Ei- respectively in Text-figure 11 A, and one on anterior border. x\nterior margin faintly indented at midline. Only a gentle slope divides cranidium from proto- P3'gidium. Axis prominent, two pairs of spines, tAvo horizontal pairs directed back from margin of pleural regions, and one pair on dorsal surface at base anterior marginal spine. Stage 0 specimen reconstructed (Text-fig. IIB) from originals of Plate 6, figures 6, 10, and similar specimens. It is larger in size (length (sag.) about 0.5 mm., length cranidium 0.32 mm.) but otherwise exceedingly similar, except for the presence of a third pair of border spines on the pygidium. Free cheek ex- tremely narrow (tr.), small eye lobe far forAvard, roAv of five spines on upper surface of convex border, roAv of seven on outer surface of border, long librigenal spine Avith hooked tip. Rostrum and hypostome unknoAvn. Text-figures IIC-F (cf. PL 6, figs. 8, 12, l-Jr, 15) shoAv larger cranidia and associated free cheeks. Not until the cephalon is about 0.5 mm. in length (sag.) (Text-fig. HE, PI. 6, fig. 14) do small paired occipital spines appear, faint Figure 11. Diacanthaspis secretus n.sp. A, Pi'otaspis, dorsal view, drawn from the original of Plate (5, figures 1-4. Approximately X .52. B, Stage 0, dorsal vIbav, left free cheek separated from cranidium, draAvn from originals of Plate 6, figures 6, 10. Approximately X 32. C, D, E, F, cranidium and right free cheek of individuals of increasingly larger size, exterior A'iews. Drawn from originals of Plate 6, figures 8, 9, 12, 14, 15 respectively. Ap proximately X 32. Paired spines numbered and lettered as in Text-figure 1; Lb is librigenal. 224 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY basal glabellar lobes, and a median axial spine in front of axial spines 3. The eye has moved back at this stage, and considerably farther back at the next stage shown, when the second lateral glabellar lobes and axial spines 5 appear (Text-fig. IIF; PI. 6, fig. 15). The indentation in the anterior margin is marked in small cranidia but becomes much less obvious, though not quite obliterated, as size increases. Additional paired spines appear on cranidia of length (sag.) about 1 mm. (PL 6, fig. 16) upwards. A slightly larger transitory pygidium than Stage 0 (PI. 6, fig. 13) has three pairs of spines on the axis, the border with three pairs of spines directed horizontally (posterior pleural spines), progressively shorter backward. At the antero-lateral corner is the broken base of a tiny anterior pleural spine. On the pleural region a short spine is situated at the base of the first two posterior pleural spines, remainder of region granulated, and showing the first pleural furrow and interpleural groove. A considerably larger transitory pygidium (PI. 6, fig. 20) shows 4 pairs of axial spines and 4 posterior pleural spines. Short, blade-like anterior pleural spines are present on the first three segments (mostly concealed in dorsal view) ; there are two spines on the posterior pleural band of the first segment, the outer one longest. Discussion of Ontogeny of Diacanthaspis If these ontogenetic series of D. lepidus and D. secret us be accepted, it is remarkable that the protaspis of the latter (PI. 6, figs. 1-5) has far the more massive and prominent spines on the external surface, while the reverse is the case in the holaspis, where the spines, though massive, are fewer and shorter and thus less prominent than the long, thin, thorn-like spines of D. lepidus (compare PI. 7, figs. 3, 15). Comparison of the metapro- taspis and Stage 0 specimens of D. lepidus and D. secretus with corresponding specimens of D. cooperi (PI. 3, figs. 1-6; Text-figs. 9A-C) reveals the close similarity — in outline and form, shape of glabella, presence of fixigenal spines, num])er and arrange- ment of spines on the external surface, and axial and border spines of the protopygidium. Distinctive of D. secretus is the size of the spines on the external surface, while in both the others these spines are short and rounded. In D. lepidus and WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 225 D. cooperi basal glabellar lobes are present in the metaprotaspis. Comparison between later developmental stages (PI. 4, figs. 6-10, 12-14, 16-18; Text-figs. 9D, IIC-F; Whittington, 1941, text-figs. 2-6) reveals both the parallelisms and specific divergence. Fixi- genal spines are lost abruptly in all. In D. secretus lateral glabel- lar lobes and axial spines 2a and 5 appear at a later stage than in either of the other species. While paired spines are present on the occipital ring of the older species, they never become bigger than the median occipital tubercle, whereas in D. cooperi their growth is rapid. DiACANTHASPis ULRiCHi Whittington, n.sp. Plate 8 ; Text-figures 8, 12. Holotijpe: USNM 116521 (PI. 8, figs. 1-5), locality 7. Other Material: Paratypes USNM 116522a-c : all figured speci- mens in USNM. Geological Horizon and Localities: lower Edinburg limestone, localities 3, 4, 7. This species has not been found at locality 2, and only 2 pygidia are known from locality 3. From locality 4, 8 cranidia and 14 pygidia, together with 26 immature cranidia, have been recovered. At locality 7 D. ulrichi is not accompanied by D. lepidus and D. secret us, whereas at locality 4 these latter two species are rather more abundant (see Table 1, p. 217) than is D. ulrichi. Description: D. ulrichi is readily distinguished from con- temporary and later species (except D. aff. ulrichi) by: (1) the convex fronto-median glabellar lobe and narrow (tr.), low lateral lobes, wbich latter are separated from the fixed cheeks by ex- tremely faint axial furrows; (2) the straight course of each branch of the suture adjacent to the eye lobe, which gives a distinctive outline to both cranidium and free cheek; (3) the short, thick, bluntly terminated median and paired main occipi- tal spines; (4) the convex axis and outwardly -sloping pleural regions of the pygidium; (5) the short spines on the outer edge of the cheek border and border of the pygidium. The external surface of the cephalon (PI. 8, figs. 24, 30) bears, besides the main occipital and outer border spines referred to, smaller, thick, blunt-tipped spines. Inside the cephalic 226 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY borders the surface between the spines is covered with fine gran- ules. The blunt spines are symmetrically arranged on the lateral glabellar lobes and cheeks, but on the fronto-median lobe, while the number and general disposition is similar in different speci- mens, it is not exactly the same, nor is the arrangement sym- metrical (PI. 8, figs. 1, 30). On the posterior edge of the occipital ring (PI. 8, fig. 24) there is a median and a pair of these spines, and in some specimens a second pair is situated in front of the main pair. Hypostome unknown. Pleurae of thoracic segments (PI. 8, figs. 12, 13, 17, 21) with narrow (exs.) anterior and wide (exs.) posterior band, former bearing short anterior pleural spine, latter bearing larger spine, progressively longer and more backwardly curved on successive segments. Axis with 2 pairs spines, 3 on posterior band, area between smooth ; articulating groove, anterior band, and pos- terior flange tuberculate. Axis of pygidium with two rings, each with pair of spines. Six pairs of spines on outer edge of border, the first very small, the fourth elongated and connected by a low ridge to the first axial ring. Prominent spine on upper surface of border at base of elongated fourth spine. One specimen (PI. 8, fig. 9) has only one, rather than two, pairs of spines between the major border pair. The distal portions of the main median and paired occipital spines (PI. 8, figs. 24, 30) display an irregular, hummocky sur- face at high magnifications. There may have been openings at the base of the tubercles on the paired spines, and the tip of the median spine shows the four tiny depressions arranged in a square. The distal tips of the posterior outer border spines of the free cheek, and the longer border spines of the pygidium, may have been like those of the paired occipital spines. Other, shorter, spines are rounded at the tip and may be closed, and the distal part of the librigenal spine does not show any openings. Development: Protaspis unknown. Development of cephalon from Stage 0 shown in Text-figure 12 A-D (compare PL 8, figs. 10. 14-16). The smallest cranidium has the characteristic trapezoidal outline, low, parallel-sided glabella divided only by the occipital furrow, low palpebral lobe far forward, faint eye ridge, and fixigenal spine. The presence of the latter suggests WHITTIXGTON : SILICIFIED ODONTOPLEURIDAE 227 that it belongs to Stage 0. The median occipital spine is repre- sented by a large, low tubercle, and the paired occipital spines, present in some specimens of this size, are tiny. Remainder of glabella without paired spines. On fixed cheeks paired spines Aj, Ao, A3, B, and C have been recognized. Surface of cranidium between spines covered with fine granules. In the next size of cranidium found, fixigenal spines are absent, the palpebral lobe is farther back, and paired spines 2. 3, 4 appear on the glabella B Figure 12. Diacanihaspis ulrichi n.sp. A, B, C, D, crauidia of increasing size, exterior views, right free c-heek shown in D. Approximately ^ 38. Drawn from the originals of Plate 8, figures 10, 14, 15, 16, 28 respectively. Paired spines numbered and lettered as in Text-figure 1. in front of the occipital furrow. At this stage the cranidium is quite like that of D. cooperi (AYhittington, 1941, text-fig. 2) and D. secret us (PI. 6, fig. 12; Text-fig. IID^, but the large median occipital tubercle and fewer and shorter paired spines on fixed cheeks and borders are distinctive. The later develop- ment parallels that of other species, except that beyond a length (sag.) of about 0.6 mm. the sj'uimetrical arrangement of spines on the glabplla in front of the occipital ring is lost. 228 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY On Plate 8, figures 25, 26 show a transitory pygidium of an unknown stage. Axis with three pairs of spines ; two long border spines are posterior pleural spines of two segments that will be released into thorax; anterior pleural spine of first is short, curved. Behind axis, two pairs of tiny border spines. A small holaspid pygidium is shown in Plate 8, figures 22, 23. DiACANTHASPTs afi:. ULRiCHi Whittiugton, n.sp. Plate 9, figures 1-5, 7-9 ; Text-figure 8. Material: all figured specimens in USNM. Geological Horizon and Localities: Oranda formation, locality 8. Description: a few fragments only of this extremely rare species are known, and deemed insufficient as the basis for a new specific name. It is like D. ulrichi n.sp., but ditfers (compare PI. 9, figs. 1, 5, with PI. 8, figs. 1, 6), for example, in the more convex fronto-median glabellar lobe, the longer median occipital spine, the deeper axial furrows of the cranidium, particularly outside the lateral glabellar lobes, and the more triangular outline of the p^^gidium. The spines and granules on the external surface of the two species are similar both in appearance and distribution (compare PL 9. figs. 7-9 with PI. 8, figs. 24, 30). The smooth tip of the median occipital spine shows clearly the four depressions arranged in a square. The smooth, blunt tips of the spines on the glabella and free cheeks show a tiny central opening. DiACANTHASPis OBANDENsis Whittiugton, n.sp. Plate 10 ; PI. 11, figures 1-15, 19, 20 ; Text-figure 8. Holotype: USNM 116523 (PI. 10, figs. 1, 3, 5-7), locality 8. Other Material: Paratypes USNM 116524a-f ; all figured speci- mens in USNM. Geological Horizon and Locality: Oranda formation, locality 8. Three species occur at this locality, and approximate numbers of complete or fairly complete parts of pxoskeletons obtained are as foUoAvs : WHITTINGTON" : SILICIFIED ODONTOPLEURIDAE 229 Table 2 Numbers of Exoskeletal Parts of Three Species of Diacanthaspis at Locality 8. D. orandcnsin n.sp. I), scitiilas u.sp. D. aff. idrichi n.sp. Cranidia 120 25 2 Free cheeks 150 38 1 Pygidia 100 31 2 The most common hypostome, 130 specimens, is regarded as belonging to D. orandensls, the less common, 16 specimens, to D. scitulus. Description: Diacanthaspis orandensis is exceedingly similar to D. cooperi, differing only in minor but persistent characters, of which the more obvious are : (1) relatively longer outer border spines on free cheek and pygidium (compare PI. 10, figs. 2. 17. 18, with PL 3, fig. 19, and ^Yhittington, 1941, PL 74, fig. 23) ; (2) more prominent axis of pygidium; (3) thorn-like spines on external surface tend to be longer and sharper; (4) particularly axial, but also longitudinal, furrows bounding first and second lateral glabellar lobes tend to be deeper and lobes to be more prominent (compare PL 10, figs. 1, 3, 5; PL 11, fig. 19 with PL 9, fig. 6, and Whittington, 1941, PL 74, figs. 24. 25, 29). As might be expected, the early developmental stages of the two species are indistinguishable (compare PL 11, fig. 1 with PL 3, fig. 8) ; only later do the specific differences become clear. No complete specimens of the librigeual spines are known, i.e. the extreme tip is always broken, but almost certainly it is not hooked. The dorsal external surface of D. cooperi has been described in detail, and most of the remarks apply to D. oran- densis. Specimens of the latter, however, are better preserved, and there is no doubt of the presence of distally-directed open- ings, with raised rims, on the paired occipital, librigenal, posterior pleural, and pygidial border spines. On the rounded tip of the median occipital spine of D. cooperi tiny depressions, arranged in a square, were observed. Similar depressions are 230 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY present on the occipital spine of D. orandensis (PI. 11, fig. 19), and further, the tips of the thorn-like spines are likewise bluntly rounded -with a group of tiny depressions, of which the central may be larger (PI. 11, fig. 19). The depressions are extremely small, and whether or not they are the openings of canals through the exoskeleton is uncertain. DiACANTHASPis sciTULUs "Whittington, n. sp. Plates 12, 13 ; Text-figure 8. Holotype: USNM 116525 (PL 12, figs. 1-3), locality 8. Other Material: Paratypes USNM 116526 a-c ; all figured speci- mens in USNM. Geological Horizon and Localities: Oranda formation, locality 8. Description: As Table 2 (p. 229) shows, this species is much less common than D. orandensis at locality 8. The general plan of the exoskeleton, and particularly the outline of the glabella, absence of prominent paired occipital spines, and arrangement of thorn-like spines on the external surface of the exoskeleton (and particularly of the fixed cheek), ally it with D. lepidiis n.sp. and D. secretus n.sp. rather than with D. orandensis and D. cooperi. The lack of large median or paired occipital spines, and the form of pygidium, distinguish it from D. ulrichi. D. scitulus differs from D. lepidus and D. secretus in that: (1) the cephalon is more convex transversely between the eye lobes, the eye ridges and sutural ridges are stronger and more clearly defined, and the row of border spines on the free cheek extends back onto the base of the librigenal spine (compare PI. 12, figs. 1-3 with PL 5, figs. 1, 7, 9 and PL 7. figs. 1. 3) ; (2) there are seven pairs of horizontally directed border spines on the pygid- ium, approximately equal in length, and a stouter, longer spine arises from the upper surface of the border and is directed back- ward and slightly upward (compare PL 12, fig. 14 with PL 5, fig. 12 and PL 7, fig. 10). In the type of major pygidial spine, convexity of the cranidium, strength of sutural ridges, and extension of lateral border spines on to the base of the librigenal spines, D. scitulus approaches the D. orandensis-cooperi group (compare PL 12, figs. 1-3, 14 with PL 10, figs. 1, 3, 5, 27). Thus the species of Diacanthaspis, while they fall into groups, never- WHTTTINGTON : SILICIFIED ODONTOPLBURIDAE 231 theless betray their close relationship. Large specimens of B. scitidus show, particnlarly on the inner snrface (PI. 12, fig. 7), a faintly impressed furrow directed transversely on the anterior glabellar lobe Avhere it slopes down to the second glabellar fur- row. This furrow is not well enough developed to be termed a third glabellar furrow, but may be the incipient stage of such a furroAv. The hypostome is of Diacanthaspis type, and dis- tinguished from that of D. orandensis (with which it occurs, but in lesser numbers) by the more curved outline of the anterior margin, the less angulate postero-lateral outline, and the nar- rower (sag.) posterior border (compare PI. 13, fig. 8 with PL 10, fig. 22). The doublure of the shoulder shows the large circu- lar opening. The one known specimen of the hypostome of D. lepidus is of similar type, but does not show the opening through the doublure. The dorsal external surface of the exoskeleton is shown in Plate 13, figures 12, 14-17. The symmetrically arranged thorn- like spines are truncated and closed at the tip, and there are tiny depressions (which may be openings) in the tip. Granula- tion extends between the thorn-like spines, but is absent from the cephalic borders, central part of occipital ring and inner corner of fixed cheek around large spine, upper surface of lateral glabellar lobes and eye ridge, axial rings, posterior pleural l:)ands. borders of pygidium and pleural ridge, and border and librigenal spines. Tiny tubercles are present particularly near the tips of the librigenal and border spines. Distally directed openings with a raised rim also occur on the distal parts of these spines, including the tip of the librigenal spine (PI. 13, fig. 14). which is not hooked as in D. lepidus, I), secretus and D. idrichi. Development : The smallest known cranidium of D. sciiidus is 0.93 mm. in length (sag.) (PL 13, figs. 3-5), and is very like cranidia of about the same size of D. lepidus (PL 4, figs. 12-14 j and D. secretus (PL 6, figs. 16-18), even in the size and position of the main thorn-like spines. On the fronto-median glabellar lobe axial spine pairs 2, 3, 4 are present, a median spine only in the position of pairs 2a and 5. On the fixed cheeks spines A^, Ao, A3, B, C, and PI may be seen. The shape of the palpebral lobe, and the strength of the eye ridge and sutural ridges, are espe- cially distinctive of D. scitulus, while examples of smaller dif- 232 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ferences are the greater number of spines on the cranidium of D. lepidus, and the fewer and differently situated spines of D. secretus. The small free cheeks of D. scitulus (PI. 13, fig. 13) have the distinctive border spines, as also do the small pygidia (PI. 12, fig. 16). Genus AciDASPIS Murchison, 1839 Text-figure 13. Synonym (objective) : Pseudomonaspis R. and E. Richter, 1917. Type Species : by monotj^py, Acidaspis brightii Murchison, 1839 (Wliittington, 1956b). Diagnosis: Two pairs of lateral glabellar lobes well developed, separated from median lobe by deep longitudinal furrows, in type species frontal glabellar lobe rounded, projecting well in front of anterior lateral glabellar lobes. In Ordovician species small third lateral glabellar lobes are present. Median part of occipital ring inflated and prolonged backward as thick median spine, separated from rest of glabella by shallow median part of occipital furrow; lateral part of occipital ring with low, gently convex occipital lobe in inner corner, behind deeper outer part of occipital furrow. Ej^e lobe elevated, situated opposite most posterior part of basal glabellar lobe ; anterior branch of the facial suture runs straight forward and inward, diverging from the course of the eye ridge, crosses border furrow on sutural ridge and curves over the anterior border. Posterior branch curves downward and outward across the fixed cheek in front of the posterior border furrow, and on to the inner side of sutural ridge, against which this furrow ends. It then curves over posterior border, inside base of librigenal spine. Convex anterolateral border with row of stout spines directed down- ward, longest posteriorly and diminishing forward ; antennal notch present, anterior border between sutures projects slightly. Librigenal spines long, curved. Thorax of 10 segments, posterior pleural band convex and inflated at fulcrum. Pygidium with 7 pairs border spines, 5th the major. External surface tuber- eulate. Geological Range: Middle Ordovician to Middle Devonian. WHITTINGTON : SILICrFIED ODONTOPLEURIDAE 233 Figure 13. Acidaspis brightii Murchison, Wenlock limestone, Middle Silu- rian, England. A, B, C, cephalon, dorsal, anterior, and right lateral views respectively, approximately X 3. (After Whittington 1956b. text-figure 1.) 234 BULLETIN : MUSEUM OF COMPARATI\^ ZOOLOGY Genus DUDLEYASPIS Prantl and Pfibyl 1949 Text-figure 14. Type Species: Acidaspis quinquespinosa Lake, 1896 (Whittington, 1956b). Diagnosis: Occipital ring with short median spine and two further pairs spines on posterior margin; not prolonged back- Figure 14. Dudleyaspis quinquespinosa (Lake), Wenlock limestone, Mid- dle Silurian, England. A, B, cephalon, dorsal and right lateral views re- spectively, approximately X 6. After Whittington 1956b, text-figure 2. ward and lacking lateral lobes. In front of large basal lateral lobes glabella narrows rapidly; small third lobes, third lateral furrows short, transversely directed, frontal glabellar lobe pro- jecting in front of third lobes. Eye lobe situated opposite mid- WHITTINGTON : SILICIFIED ODON'TOPLEURIDAE 235 part of basal lateral glabellar lobes, anterior branch suture runs on sutural ridge straight for^Yard and slightly inward to anterior border; suture then runs more directly inward and over outer edge of border to meet rostral suture about where anterior border projects forward. Sutural ridge also connects eye lobe and swollen base librigenal spine, and posterior branch suture runs along inner side of this ridge and over border inside base librigenal spine. Convex anterolateral cephalic border with row vertical spines as in Acidaspis; librigenal spines slim, curved; two pairs spines on posterior border backwardly directed. Thorax of 10 segments, pygidium with 2 pairs spines between major pair. Geological Range: Silurian (mainly Middle). GTenus RadIASPIS R. and E. Richter, 1917 Type Species: A7^ges radiatus Goldfuss, 1843 (Prantl and Pi-ibyl, 1949, p. 142). Discussion: The type species has been described by R. and E. Richter (1917, pp. 468-472, text-figs 9, 10; 1926, pp. 109-110; 1930, text-fig. 2). The form of the cephalon is like that of Acidaspis — in convexity, lobation of glabella, elongation of occipital ring, presence of lateral border spines, etc., but is dis- tinguished by the paired occipital spines. The thorax is of 9 segments, and the pygidium is notable for the absence of major border spines (there being 8 pairs of equal length) and the bilobed form of the posterior part of the axis. I regard Badiaspis as most closely related to Acidaspis, not Odontopleura, contrary to Prantl and Pfibyl's opinion. The genus is knoAvn from Lower to earliest Upper Devonian of Germany and Bohemia. Subfamily MIRASPINAE R. and E. Richter, 1917 (=Ceratocephalidae of Prantl and Pfibyl, 1949 ; of Erben, 1952) Diagnosis: Glabella wide; occipital ring long (sag. and exs.), convex, with prominent paired spines arising from swollen base ; well-defined, sub-parallel sided median lobe, gently to strongly convex, 2 pairs lateral glabellar lobes, small third pair usually present. Convex cheek of characteristic subrectangular outline, 236 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY antero-lateral portion projecting, librigenal spine arising from upper surface of border, directed upward and outward, may or may not be spines on cheek border ; eye lobe about centrally situ- ated, may be pedunculate ; two branches of suture usually inclined to each other at an obtuse angle. Hypostome subrectangular in outline with median posterior notch; middle furrow in form of triangular depression in anterolateral corner of middle body. Thorax of 9-10 segments; pleurae lacking pleural furrow or di- vided by it into narrow convex anterior and broader convex posterior bands, anterior pleural spine characteristically blade- like, with lateral barbs, and downwardly directed; posterior pleural spine much larger, 'horizontal, first 2 or 3 directed out- ward and in some eases slightly forward, successive spines di- rected more strongly backward. Pygidium with border spines, may or may not be major pair, may or may not be unpaired median posterior border spine. Geological Rayige: Middle Ordovician to Middle Devonian. Discussion: Since this group includes Miraspis, the oldest available name Miraspinae is used for it. Prantl and Pribyl and Erben include here the three best-known genera — Miraspis, Ceratocephala and Dicranurus. As a result of the present study I include Proceratocephala, Whitti7igtonia, and Ceratocephala (Ceratocephalina) n. sub. gen. I have no new information on Ceratonurus, Koneprusia, Orphanaspis, or Selenopeltoides, but presume that they may also belong here. It appears to me to be an overestimate of the value of morphological differences to divide this group into three subfamilies, and some of the criteria used by Prantl and Pribyl — e.g. supposed fusion of facial sutures in Ceratocephala, lack of anterior pleural spines of thorax in Dicranurus — are either of doubtful value or errone- ous. Genus MiRASPIS R. and B. Richter, 1917 Text-figure 15. Type Species: Odontopleura mira Barrande, 1846. (Whittington, 1956b.) Diagnosis: Occipital ring with wide (sag.) posterior band. Small third lateral glabellar lobes. Eye lobe situated opposite WHITTINGTON : SILICIFIED ODONTOPLELTJIDAE 237 mid-part of basal glabellar lobe, pedunculate. Row long, slim spines, diminishing in length forwards, on antero-lateral cephalic border. Nine thoracic segments, well-marked pleural furrow ; Figure 15. Miraspifi mira (Bariande). Motol Beds, ea2, upper middle Silurian, Bohemia. A, B, cephalon, anterior and dorsal views, respectively ; C, thoracic segment, dorsal view; D, pygidium, dorsal view. Approximately X IV2. After Whittington 1956b, text-figure 3. slim additional pleural spine between anterior and posterior. Pygidium with pleural ridge running out at first transversely from first axial ring, then turning abruptly to join base of curved 238 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY major spine. Four to eight pairs small border spines between major pair, 2 in front. Geological Range: Middle ( ?) Ordo^ician to LoAver Devonian. Discussion: In addition to the material described elsewhere is the single silicified pygidium discussed below, and the incomplete thorax and Miraspis-type pygidium from the Middle Ordovician of Wales (Whittington and Williams, 1955, p. 425, PL 40, fig. 119). Complete exoskeletons of Ordovician odontopleurids with Miraspis-type pygidium are as yet undescribed, so the range of the genus is uncertain. ?MiRASPis sp.ind. Plate 14, figures 1, 7. Material: one incomplete pygidium and one free cheek, both from locality 3, lower Edinburg limestone. Discussion: The pygidium is of characteristic Miraspis form, the short, convex axis with a prominent first ring, bearing a pair of spines. Pleural region crossed by a ridge connected to the axial ring, this ridge directed at first outward and then turning sharply to merge with the base of the major border spine. In front of ridge is first pleural furrow, separating it from low convex band bearing row of three spines. Major border spine long, directed backward and slightly upward, two small spines outside the major spine, 4 pairs inside, upper surface of border with small spines, most prominent pair behind axis. Small spines scattered on pleural regions. The free cheek has the librigenal spine arising from the upper surface of the border, latter bearing a row of short spines, and the eye lobe is pedunculate. These features recall Miraspis, and suggest that possibly this cheek belongs to the same species as the pygidium. Genus CeeATOCEPHALA Warder, 1838 Synonyms (subjective) : Onchaspis (Onychaspis) Raymond, 1925; see Whittington and Evitt, 1954, p. 53. Trapelocera Corda, 1847 ; see Prantl and Pribyl, 1949, pp. 180-181, for sum- nmvy of argument. WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 239 Ceratocephala laciniata Whittington and Evitt, 1954 Plate 14, figures 2-6, 8-15 ; Text-figure 16. Discussion: Two cranidia from locality 3 (PI. 14, figs. 2, 3) are smaller (length (sag.) 0.46 and 0.57 mm.; maximum width 0.75 and 0.85 mm. respectively) than those previously described, though not as small as that of C. triacantheis (Whittington and Evitt, 1954, p. 60). They do not have fixigenal spines, but are larger than the cranidia with fixigenal spines of Apianurus n.gen. and Diacanthaspis described here, and thus seem to represent the immediately succeeding stage when these spines are lost. Notable is the convexity of the glabella of the smallest, on which there are axial pairs of spines 2. 3, and 4 (Text-fig. 16 ; equivalent Figure 16. Ceratocephala laciniata Whittington and Evitt. Cranidium, exterior view, X 38. Drawn from the original of Plate 14, figure 3. Paired spines numbered and lettered as in Text-figure 1. to 3, 4, and 5 of Whittington and Evitt, 1954, p. 61, fig. 16), the glabella being highest where spines 3 are situated. Lateral glabellar lobes cannot be distinguished. The palpebral lolje is far forward, bearing a long, curved spine (PI in Text-fig. 16) and there are three spines on the eye ridge. On the convex fixed cheeks spines Aj, A2, A3, B, C, and D may be recognized, and between the spines on both glabella and cheeks there are tiny close-spaced spines. The larger of these two cranidia (PI. 14, fig. 2) shows characters intermediate between the .smaller and the originals of Whittington and Evitt (1954, PI. 7, figs. 1-8). Pairs of spines 2a, 2, 3, 4, and 5 (equivalent to 2-6 of Whittington and Evitt, 1954, text-fig. 16) are now present on the fronto-median glabellar lobe, as well as a pair on the flanks between 2 and 3. 240 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Basal lateral lobes are faintly developed, and bear a stout median spine, and with this inflation the fixed cheek no longer slopes inward to the axial furrow. Additional and better-preserved hypostomes (PI. 14, figs. 11, 12) have shown that the inner edge of the posterolateral doub- lure is flexed upward, and just behind the lateral notch it is also thickened. This slight projection may be regarded as the posterior wing. The doublure is flexed (the flexure running almost transversely) at the shoulder, and on the anterior side of the flexure, near the outer edge, a hole pierces the doublure. This hole is present in small hypostomes, is variable in size and usually elongated transversely. From localities 2, 3, and rarely 4, come the objects shown in Plate 14, figures 5, 6, 10, 13-15. They show a range in size and are both right and left-handed. The two divergent blade-like spines, with a row of thorn-like barbs, curved distally, along each side, are like the anterior thoracic pleural spines and the pygidial border spines of C. laciniaia. As in these latter spines, there are openings at the tip of the barbs (PI. 14, fig. 14). The fused base of the two spines is attached to a flat strip which resembles the outer part of a thoracic pleura of C. laciniata; it is without a "pleural" furrow, the transverse edge is flattened and bears a row of tiny tubercles along the "upper" edge. The anter- ior edge of the strip is also very like the anterior edge of the pyg- idial pleural region, notably in the way the flattened edge slopes forward at the "fulcral articulating process" (PI. 14, figs. 8, 10). On the opposite ("ventral") side the end of the "pleura" is rolled under and there is an antero-lateral projection ("fulcral articulating process") and a "fulcral" socket beneath the fused base of the spines. AVhen placed as in Plate 14, figures 4, 8, 9, between the outer part of a posterior thoracic segment and a pyg- idium, the resemblance of this object to the outer part of a pleura is extremely close. Further, the angle made by the lateral margin, and parallel inner margin, of the "doublure," with the sagittal line is a large one, larger than that of an obviously posterior (with backwardly directed posterior pleural spine) thoracic segment. This angle is such as to make the line of the inner edge of the "doublure" fit between that of the p.ygidium and of a posterior segment, as shown in Plate 14, figure 4. Thus this WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 241 object, if it is part of a thoracic pleura, belongs to the posterior thoracic segment. It is not crossed by a pleural farrow, but such furrows, though present in some posterior segments (AVhitting- ton and Evitt, 1954, PI. 6, fig. 6), are extremely shallow and not seen in posterior segments in which the posterior pleural spine points inward and backward. When this object is arranged as in Plate 14, figures 4, 8, 9, it is thus difficult to think of it as anything but the outer part of the posterior thoracic segment. There are difficulties to be surmounted before accepting this con- clusion, viz. : (1) Most serious is the fact that the ''pleura" of the object is of constant length and the inner termination is a V-shaped edge, the anterior limb of the "V" the longer, the edge bevelled (PI. 14, figs. 13, 15). If this object is the distal portion of a segment, why is it terminated in this way. when all other distal parts of segments found are obviously broken (PI. 14, fig. 4; see also Whittington and Evitt, 1954, PI. 6, fig. 6) ? The flattened edge is like that of the anterior and posterior edges, which would be the sutures between segments. Is the inner edge of this object also a suture? If so these sutures would run symmetrically on either side of an axial ring and innermost pleural parts, a peculiar situation seemingly unique among trilobites. Xo such median part of this supposed segment has been found. (2) Why are the two "pleural spines" of the object similar to each other, and not of the distinct anterior and posterior types of all other known segments ? The border spines of the pygidium are also of the thoracic anterior pleural spine type, and so, if the object is the outer part of the last segment, it also represents this posterior simplification of spine type. I have examined two entire specimens of C verneuili in the Museum of Comparative Zoology collections, but they show no peculiar last thoracic .segment, and no such segment is illustrated by Barrande (1852, PI. 38, figvs. 5, 6). The objects discussed above occur in material in which C. laciniata is fairly abundant, and they also occur in the lower Lincolnshire limestone with C. triacantheis Whittington and Evitt, 1954. C. rarispina n.sp. is not abundant in the Oranda formation, and these objects have not been found at locality 8. Tentatively I regard these objects as part of a thoracic segment of Ceraiocephaln. but the various 242 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY problems mentioned above remain at present unsolved. If they are truly distal parts of the last thoracic segment, the reconstruc- tion given by Whittington and Evitt (1954, fig. 13) requires modification on the lines suggested by Plate 14, figure 8. Ceratocephala rarispina Whittington, n.sp. Plate 15, figures 1-25, 28, 29. Holotype: USNM 116527 (PI. 15, figs. 1, 4, 5, 7), locality 8. Other Material: Paratypes USNM 124698 a-d ; all figured speci- mens in USNM. Geological Horizon and Locality: Oranda formation, locality 8. This species is rare at locality 8, as shown by the following table : — Table 3 Numbers of Odontopleurid Cranidia at Locality 8. Diacanthaspis orandensis n.sp 120 25 2 22 7 Diacanthaspis scitulus n.sp. . . . Diacanthaspis aff. ulrichi n.sp. . Apianurus harhaius n.gen., n.sp. Ceratocephala rarispina n.sp. . . Description: This species differs from both Ceratocephala triacantheis and C. laciniata (Whittington and Evitt, 1954, pp. 54-60, Pis. 6-9, 25, 26, figs. 1-17), bearing perhaps more re- semblance to the former, older, species than the latter. Points of discrimination are :— 1) Occipital ring lacks posterior band, and paired spines are more strongly curved outward. 2) Third lateral glabellar lobes not developed. 3) First and second lateral glabellar lobes more inflated and separated from both median lobe and cheeks by well marked furrows. 4) Eye lobe situated farther back than in C. laciniata. 5) Border of free cheek subdivided by groove into two convex bands of about same width, both of which merge into base of librigenal spine. WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 243 6) Few, larger spines on external surface of cephalon. Six pairs on median glabellar lobe include 2a, 2, 3, 4, 5, the latter being far apart, and an additional pair far apart between 2 and 3. All have counterparts on other two species (e.g. compare PI. 15, fig. 1, with Whittington and Evitt, 1954, PI. 8, fig. 1 ; PL 25, fig. 10), as has swollen base of pair 3. Two spines on eye ridge rather than three. Spines along margin of free cheek longer. Conspicuous on C. rarispma are the tiny spines that line the edges of occipital furrow and furrows on cheeks (PL 15, figs. 1, 29). 7) Hypostome with shallower lateral notch and narrower postero-lateral border. 8) Pygidium with much shorter median border spine. In Ceratocephala laciniaia some of the thorn-like spines scat- tered over the glabella and cheeks appear to have a single open- ing at the tip (Whittington and Evitt, 1954, p. 59). In at least one specimen of C. rarispina (PL 15, fig. 28), however, some of these spines are covered over at the truncated tip, there being several minute depressions or openings in the plate cover- ing the tip (appearing as darker spots in the photograph). This structure is like that in Diacanthaspis described above, and whether or not hairs emerged from the tips of these spines is uncertain. Subgenus CerATOCEPHALINA Whittington, n.subgen. Type Species: Ceratocephala {Ceratocephalina) tridens Whittington, n.subgen., n.sp. Discussion: C. {Ceratocephalina) tridens displays many of the exoskeletal characters of Ceratocephala — the subtrapezoidal outline of the convex cephalon, broad, long (sag.) occipital ring, large basal, smaller second, and extremely small third lateral glabellar lobes, prominent eye ridge and eye lobe, and librigenal spine originating on the postero-lateral border of the free cheek ; the thoracic pleurae are not furrowed and bear two pleural spines, the anterior shorter, with lateral barbs; the pygidium is short, triangular; paired axial spines are prominent on each segment, and the ornament is of large, well-spaced spines. Yet it is distinguished from any of the three species of Ceratocephala 244 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY in the Middle Ordovieian of Virginia (and other species of this genus) by a well-marked group of characters: (1) the lesser inflation of the lobes (fronto-median and lateral) of the glabella : (2) the position of the large eye lobe, far back opposite the midpoint of the basal glabellar lobe, and the consequent align- ment of the branches of the dorsal suture; (3) the prominent median occipital spine, as large as the paired spines; (4) the relatively longer pygidium, axis well-marked, and lacking the median border spine. The differences between C. laciniata, C. iriacantheis ("SVhittington and Evitt, 1954) and C. rarispina n.sp., seem to be less than those between any one of them and C. {€.) tridens, and I have recognized these differences as of subgeneric rank. If the hypostome associated with C. (C.) tridens is correctly placed, the subsquare outline and well-defined subtriangular anterior lobe of the middle body afford further distinguishing characters of the subgenus. Only the cephalon is known of Whittingtonia hispinosa, from the Upper Ordovieian of Eire and Sweden (Text-fig. 17). It is distinguished from Ceratocephala (Ceratocephalina) by the far more prominent median glabellar lobe, the relatively narrow basal glabellar lobe, the more anterior position of the eye lobe, lack of median occipital spine, and less prominent ornament. Proceratoeephala (Whittington, in56b), also an Upper Ordo- vieian genus, has more inflated median and lateral glabellar lobes, lacks the median occipital spine and has the paired spines dift'erently situated, has well-marked pleural farrows in the thoracic segments, and a pygidium with more prominent border spines. Ceratocephala (Ceratocephalina) tridens Whittington, n.subgen., n.sp. Plate 16. Holotype: USXM 124699 (PI. 16, tigs. 1-3, 18), locality 4. Other Material: Paratypes USXM 124700 a-c, all figured speci- mens in USXM. Geological Horizon and Locality: lower Edinburg limestone, localities 3 and 4. Description: Outline and form of cephalon shown in Plate 16, figures 1-7. Occipital ring set off' from posterior border by WHITTIXGTOX : SILICIFIED ODOXTOPLEURIDAE 245 change in slope and backward curve of posterior margin, longest (sag.) medially; occipital furrow deep only behind basal lateral glabellar lobes. Lobes of glabella gently convex, separated from each other by broad, shallow furrows, only the inner parts of the first and second glabellar furrows relatively deeper, and the third furrow represented by a subcircular pit behind the eye ridge (PL 16, figs. 3, 6) ; third glabellar lobe extremely small. Change of slope only separates glabellar from anterior border, axial furrows shallow beside first glabellar lobes, elsewhere glabella merges into cheeks. Eye lobe large, subspherical, situ- ated on highest jDoint of cheek, and opposite a point on midline just in front of occipital furrow. Palpebral lobe vertical, anteriorlj- merging with eye ridge, which curves forward to join with most anterior part of glabella. Eye surface (PI. 16. figs. 23, 24), with each facet convex externally, concave in- ternally. Anterior branch of suture curves forward, diverging slightly from eye ridge, crosses anterior border furrow on low sutural ridge, and turns abruptly to run inward and downward across anterior slope of border to meet rostral suture at obtuse angle. Rostral suture runs along outer edge of narrow (sag. and exs.) anterior border. Posterior branch of suture (PI. 16, fig. 18) runs down the vertical cheek, in a curve convex outwards, and across posterior border furrow and l^order. Two branches of suture are approximately aligned and thus do not make an angle at eye lobe. Free cheek (PI. 16, figs. 23, 24) with broad border occupying almost half the width at the librigenal spine, shallow border furrow, midpart of border, between rolled margin and inner part, concave upward. Librigenal spine arising inside edge of border, base merging into border, the low swelling ex- tending forward especially conspicuous. Interior view (Pi. 16. figs. 6, 9) shows doublure widest laterally, and ridges formed by deepest parts of occipital and glabellar furrows. Rostrum must be short (sag. and exs.) but broad (tr.). Hypostome tentatively placed here (PI. 16, figs. 19-22) sub-square in outline, middle body convex, divided by shallow furrows into sub-triangular anterior lobe and crescentic posterior lobe, well-marked depres- sion in antero-lateral corner. Lateral notch extremely shallow, shoulder small, pointed, posterior border wide (exs.) with median notch. Major cephalic spines are librigenal and paired 246 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY and median occipital spines. The next size, thorn-like, include those on the major spines, those projecting from edge of lateral border, paired spines of glabellar lobes, palpebral lobe, eye ridge, free cheek, etc. Fine granular ornament extends over dorsal surface between these spines. Number of thoracic segments unknown. Axis broad, pleura relatively narrow (tr.). Construction of axial rings and pleurae like Ceratocephala (cf. Whittingtoii and Evitt, 1954, PL 8, figs. 8, 10), but appendiferal pits shallower and anterior pleural spines curved, directed outward and downward, not steeply downward as in C. laciniata and C. triacantheis. The posterior pleural spines are directed slightly upward and progressively more strongly backward, so that on a posterior thoracic segment (PI. 16, fig. 12) they point almost directly back. The pair of spines on the axial ring are long and curve outwards. Thorn- like spines occur on them and on the pleural spines. Pygidium (PI. 16, figs. 15-17) with axis undivided, but the two pairs of axial spines suggest it may be composed of at least two segments. Pleural regions without border, one major pair of spines arising behind axis on upper surface, directed backward and upward ; doublure narrow with broader (sag.) "tongue" projecting toward axis. Many short spines on edge of border, smaller spines on pleural lobes. Discussion: From locality 4, where most of the specimens of C. (C.) tridens have been found, comes a single small cranid- ium of the same general type, but displaying certain differences (PI. 15, figs. 26, 27, 30). Most striking are the long curving pair of occipital spines, the bases closer together, and the small median occipital spine. The external surface is covered with tiny tubercles, and the spines in the row on the fixed cheek are longer than those of C. (C.) tridens. The different appearance of the external surface of this cranidium may be the result of preservation, but the difference in the occipital spines is sugges- tive of a specific difference. Perhaps this cranidium represents a second species of Ceratocephala (Ceratocephalina) . WHITTINGTON : SILICIFLED ODONTOPLEURIDAE 247 Gemis PkOCERATOCEPHALA Prantl and Pfibyl, 1949 Synonym : Drummuckaspis Prantl and Pfibyl, 1949 Type Species : Acidaspis terribilis Reed, 1914. Discussion: The type species has been described in detail elsewhere (Whittington, 1956b), and is like Ceratocephala, but differs in the presence of long paired spines and lateral lobes on axial and occipital rings, possessing 9 thoracic segments with deep pleural furrows, and a pygitlium with a long major and short median border spine. The cephalon displays some features recalling Miraspis, but is too poorly known to allow detailed distinctions to be made. P. terrihilis is known only from the Upper Ordo^dcian of Scotland. / / 0 / y / Figure 17. Whittingtonia bispinosa (M'Coy), Upper Ordovician, Chair of Kildare, Eire. A, B, C, incomplete cephalon, exterior, anterior, and left lateral views, approximately X 3. Dra'VNii from originals of Whittington 1956b, Plate 59, figures 4, 5, 7, 8. 248 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Genus WhITTIXGTONIA Prantl and Pfibyl, 1949 Text-figure 17. Type Species : Acidaspis bispmosus M'Coy, 1846. Discussion : Only the cephalon is known, and a well-preserved specimen has recently been described (Whittington, 1956b). It is distinguished from Ceratocephala by the three narrow (tr.) pairs of lateral lobes and wide (tr.), strongly convex fronto- median lobe which overhangs the anterior border and has a convex band across the base, and the short, thick spines on the lateral cephalic border. Specimens are known from the Upper Ordovician of Eire and Sweden. Genus DiCRANUEUS Conrad, 1841 Text-figure 18. Type Species : Acidaspis hamata Hall, 1859 (Whittington, 1956b). Diagnosis: Occipital ring without posterior band, paired spines thick and long, recurved in a hook extending over the thorax. Small third lateral glabellar lobes. Eye lobe situated opposite basal glabellar lobe on highest part of convex cheek. No spines on lateral cephalic border. Nine thoracic segments; anterior pleural spine blade-like, with lateral barbs, curved downward and backward ; convex posterior pleural band con- tinued into stout posterior pleural spine, anterior directed out- ward, remainder outward and backward. Pygidium with pair major border spines only, connected by strong pleural ridge to first axial ring. ■Geological Range: Lower to Middle Devonian. Notes on Other Miraspinid Genera Selenopeltoides Prantl and Pfibyl, 1949, type species by orig- inal designation Acidaspis hawlei Barrande, 1852, from the Figure 18. Dicranurus monstrosus (Barrande), Prokop limestones, goo, Middle Devonian, Lochkov, Bohemia. A, anterior view of cephalon, B, C, dorsal and right lateral views of entire exoskeleton, approximately X 1%. After Whittington 1956b, text-figure 4. WHITTINQTON : SILICIFIED ODONTOPLEURIDAE 249 Figure 18 250 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY upper Middle Silurian of Bohemia. No material other than the original incomplete holotype (and an isolated pygidium, now missing) is known, so that the basis for a genus is most unsatis- factory. Erben (1952b) suggested that the type species was most closely related to Dicranurus, which appears probable. Orphanaspis Prantl and Pribyl, 1949, type species by original designation Trilohites orphanus Barrande, 1852, Middle Silurian. Beside the type species, two others have been recognized in the Lower and early Middle Devonian of Germany (Erben, 1952a. pp. 306-308, 314-316, text-figs. 52, 54, PI. 20, figs. 7, 13; 1952b), but all are based on pygidia only, of miraspinid type. Koneprusia Prantl and Pfibyl, 1949, type species by original designation Acidaspis fuscina Novak, 1883, lower Middle De- vonian. Three species originally described by Barrande were placed in this genus by Prantl and Pribyl (1949, pp. 199-202), but of neither the type species nor these other species was new material described or new figures given. The pygidium of the type species is distinctive, the border spines being only one stout pair and a median posterior, but the cephalon is poorly known. Tentatively this material may be recognized as repre- senting a separate genus, presumably of miraspinid type, but until it is better known its position will remain uncertain. Ceratonurus Prantl and Pfibyl, 1949, type species by original designation Acidaspis krejcii Novak, 1883. I have no new obser- vations to add to those of recent authors (Prantl and Pfibyl, 1949, pp. 189-192; Erben, 1952a, pp. 308-313 ; Erben 1952b), and all are agreed that this Lower and Middle Devonian genus is of Ceratocephala type. It may be derived from this latter genus or from Miraspis. ?Subfamily MIRASPINAE R. and E. Richter Odontopleurid Protaspis Plate 21, figures 23, 24. Material: USNM, one incomplete specimen from locality 4. Discussion: This sjDecimen is smaller than either of the pro- taspides of Diacanthaspis lepidus n.sp. (PI. 4, figs. 1-5) and D. secretus n.sp. (PI. 6, figs. 1-5) from the same locality, but is of extremely similar form. It difi:'ers principally in that the fixi- WHITTINGTON: SILICIFIED ODONTOPLEURIDAE 251 genal spiue and those on the external surface are slimmer and less swollen at the base. On the parallel-sided glabella only the thick median occipital and axial spine pair 3 are present. On the cheek, which is most convex in the inner part, the fixigenal and spines Ai, A2, A3, B, C (directed almost straight outward and somewhat upward), D (about midway between C and palpe- bral lobe), PI and Er may be seen, and there is an additional spine about midway between Ao and B. The protopygidium has a markedly triangular outline, convex axis bearing two pairs of spines, two pairs of spines on borders, at base of anterior of which is short upwardly-directed spine. The presence of only spine pair 3 on the glabella and the outline of the protopygidium, as well as the slimmer spines on the external surface, make it unlikely that this protaspis is that of B. lepidus or D. secretus. The Stage 0 cranidium of D. ulrichi n.sp. (PI. 8, fig. 10), has stout fixigenal and median occipital spines, but others are faint, and the outline also is unlike that of the cranidium of this protaspis. The Stage 0 cranidium of Apianurus harhatus n.gen., n.sp. (PI. 19, fig. 2), has stouter spines, including paired occipital, pairs 2 and 4, and those of the anterior border, all of which distinguish it. Thus 1 am driven to suggest that this protaspis might be that of Ceratocephala laciniafa, which is fairly abundant at this locality, but cannot exclude the possibility that it is that of the rare C. (CeratocephaUna) tridens, n.subgen., n.sp. No evidence can be offered for or against this latter possibility, since tiny cranidia of C. (C.) tridens are not known. The smallest known cranidia of C. laciniata (PI. 14, figs. 2, 3) have an outline like that of the protaspis, notably the curve of the lateral sutural margin, anterior border without spines, eye lobe far forward, fixed cheeks lacking fixigenal spine but having all lettered spines and some additional. The glabella of these small C. laciniafa cranidia, however, is stronglj^ convex, spine pair 3 most prominent, but long, curved paired occipital spines are present as well as pairs 2 and 4, and the median occipital spine is small. It is chiefly this difference in the glabellae that makes me hesitate to identify this protaspis as that of C. laciniata, though there are also dif- ferences in the disposition of the lettered spines on the fixed cheek. Only the discovery of cranidia intermediate between this 252 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY protaspis and the smallest of C. laciniata would show how. if they belong to the same species, the transition takes place. Subfamily APIANURIXAE Whittington, n.subfam. Diagnosis: Glabella narrows forward, occipital ring long (sag.), convex, long paired spines and median tubercle; well defined, parallel-sided median lobe, two pairs of lateral lobes fused. Eye lobe situated far back and about midway across cheek, two branches of suture forming a straight line inclined inward and forward to the sagittal line, free cheek narrow, librigenal spine arising about midway along lateral border and curving back. Hypostome shield-shaped, middle furrow arises at antero-lateral corner of middle body and runs inward and backward; small, pointed shoulders and shallow lateral notch. Thorax of unknown number of segments, pleurae convex (exs.), single large pleural spine. Pygidium with paired border spines, unpaired median border spine may be present, long major spine upwardly directed. Geological Range: Middle to Upper Ordovician. Genus ApiAXURUS Whittington, n.gen. Type Species: Apianurus harhaius Whittington, n.gen., n.sp. Diagnosis: Long occipital spines diverge at 60-80°; fused lateral lobes kidney-shaped. Large eye lobe opposite basal glabel- lar lobe. Hypostome widest anteriorly, convex middle body divided into triangular anterior and crescentic posterior lobe ; small anterior, tiny posterior, wings. Pygidium with six or seven pairs border spines, flat pleural region bearing centrally-situated upright major spine. Long spines (except occipital) with thorn- like lateral spines, remainder of exoskeleton tuberculate or spinose. Geological Range: Middle and Upper Ordovician. An incom- plete cranidium and fragmentary free cheeks and thoracic seg- ments testify to the presence of a species of this genus in the Lincolnshire limestone, below the Edinburg limestone. WHITTINGTON : SILICIFIED ODOXTOPLEURIDAE 253 254 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Apianurus barbatus Whittington, n.gen., n.sp. Plates 17-19 ; 20, figures 1-17, 19 ; Text-figures 19-22 Holotype: USNM 124701 (PI. 17, figs. 3, 4, 6), locality 2. Other Material: Paratypes, USNM 124702 a-e; all figured speci- mens in USNM. Geological Horizon and Localitiei^: Edinburg limestone, localities 2, 3, 4, 6, 8. Description: Cephalon moderately convex, outline in dorsal and anterior aspect elliptical. Glabella moderately convex trans- versely, gently convex longitudinallj', maximum width across occipital ring, narrowing forward to half this width at anterior margin ; length about two-thirds maximum width. Occipital ring of length (sag.) in dorsal aspect half that of rest of glabella, becoming narrower (exs.) laterally, outline of posterior margin a curve strongly convex posteriorly, outline of anterior margin a curve more gently convex anteriorly. Stout occipital spines diverge at about 60° and curve upward and backward, extend- ing to a length three times the sagittal length of the cephalon. Small median occipital spine just behind occipital furrow. Lat- ter shallow medially, deep behind basal lateral lobes. First and second glabellar lobes fused to give a kidney shape, the second the smaller, the first glabellar furrow represented by a subcircu- lar pit adjacent to the median lobe. A pit anterior to the second glabellar lobe represents the second glabellar furrow, and is bounded anteriorly by the eye ridge as it fuses with the frontal glabellar lobe. Outline of anterior margin of latter curve con- vex forward, and separated by change of slope from narrow anterior border. Fronto-median lobe gently convex, change of slope separating it from lateral lobes, faint additional swelling running transversely across it between first glabellar lobes. Axial furrow not deep, but a narrow, unornamented band, mark- ing reversal of slope between cheek and glabella. Cheek semi- circular in outline, maximum width opposite second glabellar lobe, rising steeply to large eye lobe. Transverse line through mid-point of eye lobe passes through mid-point of first glabel- lar lobe and runs just in front of occipital furrow. Anterior branch of facial suture runs straight forward and inward at about 45° to the midline, curving inward a little more as it WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 255 crosses the border on a sutural ridge. Posterior branch runs out and back in line with anterior branch for a short way, then curves in over posterior border. Latter longest (exs.) at suture, convex, and becomes shorter inward. Thus the occipital ring merges with inner corner of cheek rather than with posterior border. Fixed cheek .slopes vertically behind eye lobe. Palpebral lobe with rim becoming well-defined anteriorly, and passing into broad, convex eye-ridge. Latter runs with a slight curve inward and forward to merge with frontal glabellar lobe. Eye surface almost hemispherical, external surface (PI. 17, fig. 21) showing faintly the tiny facets. Lateral cephalic border rolled, shallow border furrow interrupted at about mid-length by swelling at the base of librigenal spine. Latter directed outward and slightly forward at first, then curving and slimming to point backward, and reach to a length about three times that of the cephalon (sag.). Just outside anterior branch of suture is antennular notch (PI. 17, fig. 5), with a sharp projection at outer margin. Anterior border between sutures narrow (sag. and exs.), rostral suture running along outer edge. Rostrum unknown, but pre- sumably wide (tr.) and short, sloping downward and inward. Doublure narrow on free cheek, absent on posterior border of cranidium, articulating half-ring long (sag.). On inner surface of cranidium (PI. 17, fig. 12) outer part of occipital ring, first and second glabellar furrows make rounded projections but are not extended as appendifers. Width of hypostome (PI. 18, figs. 1-5) at anterior margin greater than length, rounded postero- lateral outline, faint median posterior notch. Anterior wing small, tip rounded, no wing process. Lateral border narrow, gently convex, deep border furrow, shallow lateral notch and sharply-pointed shoulder. Po.sterior border wider and in median portion separated by shallow depression from median body. Outermost part of posterior border bent up sharply near mid- line. From the broad depression at antero-lateral corner the middle furrow runs straight inward and backward to define tri- angular, convex anterior lobe of the middle body. Posterior lobe crescentic in outline, inflated. Doublure of hypostome widest laterally, disappearing at anterior wing and pressed close to median part of posterior border so that edge is sharp. Ridge crosses doublure from point of shoulder to inner edge, immedi- ately behind ridge doublure extended inward to form small tri- 2r>6 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY r abf Figure 20. Apianurus barbatm n.gen., n.sp. Outline reeonstruetioiis of cephalon, approximately X 10. A, ventral view, right half of rostrum and left half of hypostome shown, anterior portion of left free cheek incomplete. B, anterior view, rostrum not shown and anterior portion of left free cheek incomplete. C, left lateral view, rostrum and left free cheek not shown, a, antennular notch; abf, ridge on inner surface corresponding to anterior liorder furrow; d, doublure of right free cheek; 2gf, ridge on inner surface corresponding to second glabellar furrow; h, hypostome; m, boss corre- sponding to depression at outer end of middle furrow; r, rostrum. WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 257 angular posterior wing. Perforation through doublure posterior to ridge (PI. 17, fig. 19). By mounting a cranidium and free cheek of appropriate size iu juxtaposition, as shown in Plate 17, figures 5, 7, 9-11, the approximate position and attitude of the hypostome has been determined (Text-fig. 20). The gently-curved hypostomal suture (PI. 18, figs. 1, 5) fits against the rostrum, and the anterior wing is directed upward and outward, the tip lying beneath the deep- ^i-ned outer part of the anterior border furrow where it meets the furrow bounding the outer side of the eye ridge. There may have been muscles linking the wing and inner side of the furrow, but there is no wing process or socket. The depression at the outer end of the middle furrow forms a considerable projection on the inner surface (PI. 18, fig. 2), and may have been a point of muscle attachment. Perhaps it was linked by muscles to the outer end of the anterior border furrow. The shape of the rostrum is suggested, and is consistent with the form of the few odontopleurid rostra knoAvn (Whittington, 1956b, PI. 58, fig. 7; PI. 59, fig. 3). If the cephalon rests on a flat surface on the antero-lateral cephalic border and librigenal spines, as portrayed in Text-figure 20 B, C, the middle body of the horizontal hypo- stome would lie on this surface. Ornament of cephalon of close-spaced spines which curve characteristically toward the closed tip (PI. 17, fig. 21; PI. 18. fig. 22). The spines A'ary in size, the exoskeleton between their bases smooth, as are also the axial, occipital, and glabellar fur- rows. Arrangement of spines not perfectly regular, but sym- metrical pairs of larger spines ma}^ be distinguished even on largest cranidia, on the occipital ring, median and lateral glabel- lar lobes, eye ridge, etc. (PI. 17, fig. 2). Single row of s.ymmetri- eal spines on anterior border of cranidium. Large, curved spines are also regularly arranged on the border of the free cheek and the proximal portion of the librigenal spine (PI. 17, figs. 1, 7, 14, 15) — e.g. four long and one short spine on the border behind the librigenal spine ; stout spine at edge of autennular notch, one of same size midway to base of librigenal spine ; 6 spines along outer side of proximal part of librigenal spine, etc. The character- istic curved spines are not present on the distal part of the librigenal spine, nor on the occipital spines. Much shorter, blunt, 258 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY spines are present, directed outwards on the proximal portion of the spine, but distally inclining more and more in the direction of the main spine, so that at the tip they lie at a low angle to the surface (PI. 17, fig. 3). Toward the tip of the main spine, on the distal side of the base of the small spine, a minute opening is sometimes seen — apparently the locus of a sensory hair. Other cephalic spines do not seem to show such openings. Proximally, in the region where there are the large lateral spines, the under .side of the librigenal spine is smooth, as is the doublure of the free cheek. The characteristic curved spines occur on the lateral and posterolateral borders of the hypostome (PL 18, fig. 1), but are reduced to granules on the posterior border. Similar spines are scattered on the lateral areas of the middle body, but the central portion is smooth. Number of thoracic segments unknown. A sufficient number of the delicate segments are preserved to permit the reconstruc- tion (Text-fig. 19 A, B), which is based on the assumption of a total of ten. Axis broad, two-thirds total width (excluding pleural spines) at anterior segment (PI. 18, fig. 6), narrowing back to half its width at posterior segment (PI. 18, fig. 9). Axial ring moderately convex, posterior margin forms a curve, convex posteriorly. Articulating furrow more strongly curved in same sense, outermost part deepened. Articulating half-ring longer (sag.) than axial ring. Pleura narrowest (tr.) in anterior seg- ments, horizontally extended, without pleural furrow. Narrow (exs.) flanges developed, outer surface of each flat and inward and downward sloping. Ring and axial articulating sockets and processes scarcely at all developed, but lateral margin of pleura rolled under and fulcral socket and process prominent. Base of pleural spine swells out from upper, outer surface of pleura. Length, curvature and direction of pleural spines indicated in reconstruction. The characteristic curved ornamental spines are close together on the axial ring, varying in size, some larger ones paired. Similar spines more widely scattered on pleurae. Pleural spines ornamented like librigenal, with row of lateral, curving hooks on each side of proximal part, and the tiny in- clined spines distally (PI. 18, fig. 9). Similar openings to those on the tips of the occipital and librigenal spines seem to be present. Pygidium (PI. 18, figs. 11-13) semi-oval in outline, length WHITTINGTOX : SILICIFIED ODOXTOPLEURIDAE 259 (sag.) a little more than one-third width at anterior margin. Axis moderately convex, a little wider than long, rounded pos- teriorly, rising from the flat pleural regions and not defined by deep furrows. Articulating furrow and half-ring like those of segments. First ring furrow broad and deep, curving forward to midline, extremities widened and deepened. Second ring Figure 21. Triangular graph comparing relative proportions of dimensions of cranidia of Apianttrus barhatus n.gen., n.sp. Those from the lower Edin- luirg formation, localities :!, 3, sliown in dots enclosed by solid line. Cranidia from the Oranda formation, locality 8, shown by open circles enclosed by broken line. Two tiny cranidia from locality 3 shown by X (originals of Piate 19. figures 2, 9). L, sagittal length of cranidium. "W, width between spines B on posterior border. E. length (exs.) from anterior margin to midpoint of palpebral lobe. Measurements were made in ex- terior view, i.e. with isolated cranidium resting on a horizontal surface. 260 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY furrow represented only by pair of oval, unornamented areas on terminal part of axis, corresponding in position to wider outer part of first furrow. In many specimens these two pairs of oval areas are translucent, the quartz much thinner over them, or broken, indicating- that they may be areas of muscle attachment (cf. Whittington and Evitt, 1954, pp. 24-25). Long spine rises vertically from flat pleural region, opposite first ring furrow, and distall}^ curves a little backward. Border of pleural regions with 6 pairs of spines directed outward and slightly upward, first pair short, second longer, third longest, and remaining pairs shorter inwards. Doublure of pygidium narrow, rolled under at antero-lateral corner forming articulating process. Ornament of curved spines on pygidium (PI. 18, figs. 12, 13, 19) disposed as on thorax — dense, varying in size, and with some conspicuous pairs on the axis, more widely spaced on pleural regions. Upright and border spines with lateral barbs, and towards tips the tiny inclined spines are present, and the open- ings may be present. Discussion : In addition to the material from the lower Edin- burg formation (localities 2, 3, 4; Pis. 17-19), that from the Oranda formation (locality 8; Pi. 20, figs. 1-17, 19) also repre- sents Apianurus harbatus. Comparison of parts of the exoskele- tons from all localities reveals no consistent differences between them. Measurements of length and width of eranidia and of posi- tion of palpebral lobe in specimens from both horizons fall within almost the same area of a triangular graph (Text-fig. 21). Local- ity 8 in the Oranda formation is some 450 ±: feet higher in the section than localities 2-4 in the lower Edinburg, so that A. harbatus evidently has a long range in time. A few specimens (see Table 4) are knoAvn from locality 6, at an intermediate point in the section. Figure 22. Ainannrus harhatus, u.geu., ii. sp. A, Stage 0 exoskeleton, dorsal view, drawn from originals of Plate 19, figures 1, 2, 5. Approximately X 32. B, t-ranidium, exterior view, drawn from original of Plate 19, figure 9. Approximately X 32. C, cranidium and right free cheek, exterior view, drawn from originals of Plate 19, figures 7, 8. Approximately X 32. D, cranidium and right free cheek, exterior view, free cheek from original of Plate 19, figure 11. Approximately X 32. E, cranidium, exterior view. Ap- liroxiniately X 32. F, cranidium, exterior view. Approximately X 16. WniTTIXGTOX : SILICIFIED ODOXTOPLEURIDAE 261 262 BULLETIN : MUSEUIM OF COMPARATIVE ZOOLOGY Development. Stage 0: Protaspis unknown, but the smallest known cranidium, free cheek and pygidium from localities 3 and 4 (PI. 19, figs. 1, 2, 5) have been associated (by analogy with Diaca7ithaspis) to show the probable nature of the Stage 0 exo- skeleton (Text-fig. 22A). Outline of cranidium (length (sag.) 0.41 mm.) trapezoidal, gently convex longitudinally, more gently transversely. Glabella outlined by broad, straight, axial furrows, short (sag. and exs.) occipital ring with prominent median and short curved pair spines, occipital furrow defined only distally, remainder of glabella without furrows, four prominent pairs short, curved, axial spines (2-5 in Text-fig. 22A). Narrow an- terior border, fixed cheeks gently convex, palpebral lobes at two- fifths cranidial length (exs.) from anterior margin. Posterior margin of fixed cheek runs directly outward from the axiai furrow, and just beyond half the width bends abruptly to run outward and forward to the base of the long fixigenal spine. On the outer, postero-lateral area of the fixed cheek there are 4 paired spines (A, B, C, D in Text-fig. 22A), and tiny spines along the inner side of the fixigenal spine and adjacent outer part of the fixed cheek. Certain specimens (PI. 19, figs. 4, 6) have the fixigenal spine retained on one side only. Their significance is discussed under "Ontogeny" in Part I. Free cheek of ap- propriate size to fit this smallest cranidium (PI. 19, figs. 1, 3; Text-fig. 22 A) narrow, large eye surface situated at half length, big librigenal spine curving outward and backward. Anterior branch of the suture straight, posterior runs in curve convex out- ward, doublure projecting inward and fitting against doublure at base of fixigenal spine. Notch in anterior border outside suture very small. Long spines arranged in a constant pattern along outer edge of border and proximal edges of librigenal spine — conspicuous is variation in size of those along antero-lateral edge, curved spine at anterior basal edge of librigenal spine, numbers, curvature and/or inclination of remainder. Transitory pygidium of width at anterior margin 0.51 mm., length (sag.) 0.30 mm. Prominent axis bearing three pairs of curved spines, first two rings distinct, tip indistinct where posterior part of pygidium is bent sharply down. Pleural regions flat, borders bent down, 8 border spines arise on dorsal surface inside border, curving out- ward and backward. Doublure rolled under. Small barbs along sides of these border spines. WHITTINGTON* : SILICIFIED ODONTOPLEURIDAE 263 Further development of cephalon: Next largest cranidia (PI. 19, figs. 9, 10; Text-fig. 22B) from localities 3 and 4 only. Par- allel-sided glabella more convex, defined by deep axial furrows, tiny fifth pair of axial spines (2a in Text-fig. 22B) present just in front of occipital furrow Small lateral pairs of spines also present outside pairs 2, 3, and 4. Extremity of occipital furrow a shallow pit, in front of which is small swelling representing the beginning of basal glabellar lobe. Large palpebral lobe situated at about the same position. Most striking is the absence of fixigenal spine and rounded outline of genal angle of fixed cheek. Paired spines A, B, C and D (Text-fig. 22B) are de- veloped as in the smaller cranidium, and there are one or two additional spines at the genal angle between B and C, as well as a small one on the fixed cheek between A and B. Free cheek associated with this size of cranidium very like that of Stage 0, but has eye surface situated behind mid-length, posterior branch of suture running in curve convex outward and seemingly little or not at all modified despite absence of fixigenal spine. Notch in anterior border is wider, spines along margins and proximal part of librigenal spine similar in size and arrangement to those on the smaller cheek. The next largest cephala (PI. 19, figs. 7, 8 ; Text-fig. 22C) show a marked increase in convexity of glabella and cheeks and in height of eye lobes. Occipital ring projects back farther behind cheeks, and paired spines now reach back to a length equal to that of cranidium. Midpoint of eye lobes is at about half the length, cheeks slope steeply behind them to border furrow, which runs out to margin between spines A and B. Eye ridge faintly de- fined. Median occipital and five paired spines of glabella prominent, additional lateral pairs developed, including 2 pairs (one central and one just in front) on convex basal glabellar lobe. On cheek and borders main spines developed as liefore but additional pairs also present. Next largest cephalon (PI. 19, figs. 11, 14-16; Text-fig. 22D) similar, eye lobes at about .same position, eye ridge more clearly defined. Occipital ring with longer paired spines, and relatively reduced median spine. In front of pits at extremity of occipital furrow are convex, ovate, basal glabellar lobes. In front of these, in line with anterior part of eye lobe and at base of side of glabella, a prominent pair of 264 BULLETIN' : MUSEUM OF COMPARATIVE ZOOLOGY spines. In slightly larger cranidia (PL 19, figs. 18-20, 22, 23; Text-fig. 22E ) the area around the base of these spines is slightly swollen, and represents the second glabellar lobes. A deep pit represents the first glabellar furrow, and a faint swelling outside it connects the two lobes. Thus the bean-shaped fused lateral lobes are present at this stage, and become more prominent as the cranidium increases in size, swelling up to partly fill the de- pression between eye lobes and glabella. Other changes as size increases include appearance of swelling on median glabellar lobe just in front of occipital furrow, backward movement of eye lobes with consequent steeper slope behind them, and increased prominence of eye ridges. Many more ornamental spines appear, mostly paired, and a median row on the median glabellar lobe. The relative size of these spines diminishes, but even in the larg- est cranidia (PI. 17, figs. 2, 3) the numbered and lettered pairs may be recognized by their slightly larger size, and most spines retain the characteristic curve so prominent in tiny specimens. The chief change in the free cheek from the size shown in Text- figure 22A is the backward movement of the eye lobe, so that the postero-lateral slope behind it is steep or overhanging. The arrangement of the spines on the border and proximal part of the librigenal spine remains basically the same, though the curved spine at the anterior basal edge of the librigenal spine is reduced, and an additional large spine (making six in all) appears along the anterior proximal margin of the librigenal spine. The margin of the antennular notch remains smooth, but the outer edge bears a large spine. Because the broad, swollen base of the librigenal spine merges into the cheek a border becomes defined by a shallow border furrow only near the branches of the suture. Smallest known hypostome (PI. 19. figs. 12, 13) 0.4 mm. in length (sag.), 0.55 mm. in width across anterior wings, propor- tions of various parts similar to adult. There is little change with increasing size. Thorax and pygidium. Next largest transitory pygidium (PI. 19, fig. 24) to that assigned to Stage 0 (PI. 19. fig. 5) has three pairs of spines on the axis, 3 pairs long border spines, with lateral barbs, similarly directed to those of the earlier stage. An upright spine at the base of the anterior two border spines. Larger transitory pygidia (PI. 19, figs. 25, 27-29) have four WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 265 pairs axial spines and four pairs border spines, the fourth border pair short, blunt, situated at tip. Latter less steeply bent down. First three border spines with lateral barbs and upright spine at base, direction and curvature shows slight variation, perhaps associated with stage. Latest transitory pygidium (PL 19, fig. 26) of length (sag.) 0.36 mm., width 0.7 mm. with 4 pairs of axial spines, posterior two close together at termination. Long, curved spines arising from the antero-lateral portions of pleural regions are pleural spines of what is to become last thoracic seg- ment. Bases of broken-oif pair of upright spines on pleural regions are opposite 3rd and 4th axial spines. Four pairs of border spines, third pair larger than others. Posterior tip of this pygidium not bent down. Small true pygidium (PI. 18, figs. 23, 24) has first axial ring defined, with its pair of axial spines, and posterior part of axis bears two pairs of axial spines (and pre- sumably therefore includes at least 2 segments). Swollen base of upright spine on the pleural region is connected with first axial ring. Margin with tiny anterior pair of spines and 4 larger pairs, third of these distinctly longer. As size increases (PI. 18, figs. 16-18, 20, 21) the adult appearance is attained, except that there are still only 5 pairs of marginal spines, the anterior short, the others becoming equal in size. There are three pairs of axial spines, and a median spine on the tip of the axis. Large pygidia (PI. 18, figs. 11-13) have six pairs of mar- ginal spines, the first much shorter and slimmer than the rest. Behind the ring furrow the axis includes two pairs of axial spines, and a median spine at the tip. Specimens from Localities 2 and 3 Differing from Apianurus barhatus n.gen., n.sp. Under this heading may be grouped the following : (a) Cephala and pygidia from both localities which are so strikingly different as to be regarded as representing a species of a distinct genus, described below as Calipernuriis insolitus n. gen., n. sp. (b) Cranidia from locality 2 which are like A. harhatus but are smoother — i.e. the spines are short, reduced to tubercles, or absent. These are described below as A. glaber n.sp. 266 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (c) Pygidia occurring at both localities, distinguished from A. harhatus b}' being smoother and having shorter, inwardly- curved upright spines. These are here regarded as belonging with cranidia (b). (d) A second type of pygidiuni occurring at locality 2, hav- ing the typical curved spine ornament of A. larhatus, but the upright spines are very short, and there are 7 pairs of spines on the border. This pygidium is described as Apianurus sp.ind. (e) Three types of hypostome from locality 2, each different from that of A. harhatus. The one differing most has been assigned to Calipernurus insolitus n.gen., n.sp., the smooth one to A. glaher n.sp., the third described under '' Apianurus sp. ind." Each of the groups of specimens included under (a) to (e) is distinct, and there is no gradation between groups or between any one of them and A. harhatus. Within each group except (e) there are specimens of different sizes, showing that the dis- tinctive characters of the group are not those of one particular size (i.e. groAvth stage). Table 4 shows the relative abundance of the different species at the different localities. Only A. harhatus is found in strata above the lower Edinburg. If the specimens grouped under (b), (c), and (d) were expressions of the morphological range of variation of A. harhatus one might expect them to be represented in the sample from locality 8. The fact that they are not suggests that they represent distinct species. This same pattern of distribution is displayed by species of other genera from localities 2 and 3, e.g. among the hundreds of specimens of Dimeropyge virginiensis Whittington and Evitt (1954, pp. 37-42, PI. 2, PI. 3, figs. 1-30) are a small number that are distinctly different and do not grade morphologically into the common form, and which I regard as a separate species. One possible explanation of the ''common" and "rare" species of the same genus at the same locality is that they represent sexually dimorphic forms of the same species, but this explanation can hardly be removed from the realm of speculation. In this con- nection Hintze (1953, p. 150) has discussed the occurrence of two pliomerid species at the same horizons of the Pogonip group, but in this case the two species are apparently equally abundant. WHITTINGTON : SILICIFIED ODONTOPLEURIDAE 267 id lique interior view showing hole in doublure, X 30. Locality 10. 15-16. Incomplete posterior thoracic segment: dorsal, left lateral views, X 6. Locality 10. 17-18. Stage 8 transitory pygidium : exterior, interior views, X 30. Locality 12. 19. Ilolaspid pygidium: dorsal stereograph, X 4%. Locality 12. 20. Tip of genal spine : oblique exterior view, X 30. Locality 9. 21. Eye surface: oblique exterior view, X 30. Locality 9. PLATE 3 PLATE 4 •''iin^ »»*'• Plate 4 Diacnnthaspis hpidu.s n.sp. Lower Edinburg limestone Figtire 1,2,3,4,5. Incoiuplete protaspis : exterior stereograph, anterior view, left lateral view, postero-dorsal stereograph, antero-dorsal stereo- graph, X 33.6 Photograph by W. E. Evitt. Locality 4. 6. Small eranidimu: exterior stereograph, X 30. Loeality 3. 7. Small cranidium: exterior stereograph, X 30. Locality 3. 8. 9, 10. Small cranidium : dorsal, anterior, right lateral views, X 15. Locality 3. Small pygidium : dorsal \'iew, X 20. Locality 2. Small cranidium: dorsal, anterior, left lateral views, X 15. Locality 2. Transitory pygidium: duisal vicAV, X 30. Locality 4. Small cranidium: dorsal, anterior, left lateral views, X 15. Locality 2. Cranidium: exterior stereograph, anterior view, X G. Locality 2. Cranidium: exterior stereograph, X 6. Locality 2. Posterior part of cranidium: exterior view, X 30. Locality 2. Arrow points to spine witli truncated tip and tiny depressions or openings. 11. 12, 13, 14. 15. 10. 19, 17, 21. 18. 20. Figure Plate 5 Diacanthaspis lepidus n.sp. Lower Edinburg limestone 1,7,9. Holotype cranidium and paratype free cheek: dorsal stereo- graph; anterior, oblique exterior views, X 10. USNM 116517, 116518a. Locality 2. 2.3,0,6. Incomplete thoracic segment: exterior, anterior, posterior, in- terior views, X 12.5. Locality 4. i. Holotype cranidium: interior view, X 15. USXM 116517. Locality 2. 8,13,1-4. Pygidium: interior view, X 6; posterior and right lateral views, X 4. Locality 2. 10,11. Part of thorax and pygidium, paratype: dorsal and ventral stereographs, X 6. USNM 116518c. Locality 3. 12. Pygidiiun : dorsal stereograph, X 9. Locality 2. 15, 16, 17, 18. Paratype hj^ostome: interior stereograph, X 20; exterior stereograph posterior and anterior views, X 15. USNM 116518b. Locality 3. PLATE 5 PLATE 6 Plate 6 BiacantlMspis secretus n.sp. Lower Edinburg limestone Figure 1,2,3,4,5. Incomplete protaspis: exterior stereograph; anterior, left lateral, antero-dorsal, posteero-dorsal views, X 33.6. Photo- graphs by W. E. Evitt. Locality 4. Original of Text-figure llA. 6. Stage 0 eranidium and free cheek: exterior stereograph, X 30. Locality 4. Original of Text figure IIB. 7, Same eranidium as figure 6.: interior view, X 30. Locality 4. 8,11. Small eranidium: exterior stereograph; interior view, X 30. Locality 4. Original of Text-figure IIC. 9. Small free cheek: oblique exterior view, X 30. Locality 4. Original of Text-figure IIC. 10. Stage 0 pygidium : dorsal ^•iew, X 30. Locality 4. Original of Text-figure IIB. 12. Small eranidium and free cheek: exterior stereograph, X 30. Locality 3. Original of Text-figure IID. 13. Transitory pygidium: dorsal view, X 30. Locality 4. 14. Small eranidium and free cheek: exterior stereograph, X 25. Locality 2. Original of Text-figure HE. 15. Small eranidium and free cheek: exterior stereograph, X 15. Locality 2. Original of Text-figure IIF. 16,17,18. Small eranidium: dorsal, right lateral, anterior views, X 9. Locality 2. 19. Same eranidium as figure 14: interior view. X 20. Locality 2. 20. Transitory pygidium : dorsal view, X 30. Locality 4. 21,22. ParatjT)e free cheek: oblique interior views, X 15. L^SX^^M 116520a.. Locality 2. Figure PlxiTE 7 Diacanthaspis secretus n.sp. Lower Edinburg limestone 1,3,5,11. Holotype cranidium : dorsal and anterior views, X 9; interior view, X 6; left lateral view, X 9. USNM 116519. Locality 2. 2, 7. Paratype free eheek : dorsal and oblique exterior views, X 10. USNM 116520a. Locality 2. 4. Paratype thoracic segment: doi'sal view, X 6. USNM 116520b. Locality 2. 6. Paratype, incomplete thoracic segment : dorsal view, X 6. USNM 116520b. Locality 2. 8. Pygidium : interior view, X 9. Loe^ity 2. 9. Pygidium: dorsal view, X 9. Locality 4. 10. Paratype pygidium: dorsal stereograph, X 9. USNM 116520c. Locality 2. 12. Pygidium, exterior view, tips of border spines shoAving open ings, X 30. Locality 2. 13, 14. Free cheek: oblique exterior views, border spines and librigenal spine showing openings, X 30. Locality 2. Diacanthaspis lepidus n.sp. Lower Edinburg limestone 15. Cranidium : anteiior view to show spines X 30. Locality 2. PLATE 7 PLATE 8 Plate 8 Diacatnthatipis iilriclii n.sp. Lower Edinburg limestone Figure 1,2,3,4,5. Holotype cranidium: dorsal stereograph, X 15; right lateral, interior, anterior, posterior views, X 10. USNM 116521. Locality 7. (5, 7, 8. Paratype pygidium : dorsal, posterior, right lateral views, X 15. USNM 116522c. Locality 7. Pygidium: dorsal view, X 30. Locality 7. Stage 0 cranidium: exterior stereograph, X 30. Locality 4. Original of Text-figure 12A. Pygidium: ventral view, X 20. Locality 7. Two paratype segments from anterior part of thorax: exterior view, X 15.USXM 116522b. Locality 7. Paratype thoracic segment: dorsal view, X 15. USNM 116522b. Locality 7. Small cranidium: extei ior view, X 30. Locality 4. Original of Text-figure 12B. Small cranidium: exterior view, X 30. Locality 4. Original of Text-figure 12C. Small cranidium: exterior view, X 30. Locality 4. Original of Text-figure 12D. Paratype thoracic segment: dorsal view, X 15. USNM 116522e. Locality 7. Small cranidium: anterior, dorsal, right lateral views, X 15. Locality 4. Paratype thoracic segment: dorsal view, X 15. USNM 116522b. Locality 7. Small pygidium: exterior and interior views, X 20. Locality 7. Cranidium: two exterior views to show spines and depressions at tip of occipital spine, X 30. Locality 7. Transitory pygidium: exterior and interior views, X 20. Lo- cality 7. 27. Free cheek: oblique exterior view, X 9. Locality 7. 28. Small free cheek: oblique exterior view, X 30. Locality 7. Original of Text-figure 12D. 29. Free cheek: oblique exterior view, X 30. Locality 7. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18, 19, 20. 21. 22 23. 24, 30. 25, 26. Figure Plate 9 Diacanthaspis aff. ulrichi n.sp. Oranda formation, locality 8 1-3, 7, 9. Cranidium : 1, 2, 3, exterior stereograph, dorsal and left lateral views, X 9 ; 7, oblique exterior view showing spiiies and granu- lation on external surface, X 30; 9, oblique exterior view show- ing pits in smooth tip of median occipital spine, spines and granulation on external surface, X 30. 4, 8. Incomplete cranidium : 4, exterior stereograph, X 9 ; 8, ex- terior view showing pits in the smooth tip of median occipital spine, X 30. Pygidium : dorsal stereograph, X 15. IJiacantliaspis cooperi Whittingt Lower Martinsburg shale, localit on itv 10 6. Incomplete cranidium: exterior view showing openings in tips of spines and granulation, X 30. PLATE 9 PLATE 10 Figure Plate 10 Diacanthaspis orandensis n.sp. Oranda formation, localitv 8 1,3,0-7. Holotype cranidium: exterior stereograph, dorsal, anterior, left lateral, posterior views, X 9. USNM 116523. 2,4,20. Paratype free cheek: oblique exterior, dorsal, and oblique interior views, X 9. USNM 116524a. 8-12. Paratype segment from anterior part of thorax: dorsal, ven- tral, posterior, anterior, right lateral views, X 10. TJSX^M 116524c. 13, 14. Cranidium : interior and dorsal views, X 9. 15. Paratype segment from posterior part of thorax: dorsal view, X 15. USXM 116524d. 16. Paratype, right pleura of segment from posterior part of thorax, dorsal view, X 15. USNM 116524e. 17. Incomplete pygidium, dorsal view, X 9. 18. Incomplete pygidium, dorsal view, X 9. 19. Incomplete pygidium, dorsal view, X 9. 21-24. Paratype hypostome: interior, exterior, posterior, right lateral views, X 15. USNM 116524b. 25. Incomplete pygidium, ventral view, X 6. 26,27. Paratype pygidium: 26, posterior view, X 6; 27, dorsal stereo- graph, X 9. USNM 1165241 Plate 11 Diacanthaspis orandensi^ n.sp. Oranda formation, locality 8 Figure I. Smallest cranidiuui ; exterior view, X 30. 2,3. Small hypostome: interior, exterior views, X 20. 4. Paratype free cheek (original of Plate 10, figures 2, 4, and 20) : oblique exterior view, X 9. USNM 116.524a. 5. Small cranidium: exterior view, X 30. C. Small free cheek: oblique exterior view, X 10. 7. Small cranidium : exterior view, X 30. 8. Pygidium (original of Plate 10, figure 18) : jjosterior view, X 6. 9. Small pygidium : exterior view, X 20. 10. Small cranidium: exterior view, X 15. II. Small cranidimu : interior vieAV, X 15. 12-15. Small cranidium: exterior stereograph; riglit lateral, dorsal, anterior views, X 15. 19,20. Incomplete cranidium: 19, exterior view of part of external surface ; on right fixed cheek rounded tips of spines show minute dark spots, the locii of depressions or openings, X 30; 20, exterior view of part of ()cci])itai ring, showing pits in tip of median spine, X 30. Diacanthaspis cooperi Whittington Lower Martinsburg shale 1(5,17. Free cheek: 16, oblique exterior view of base of librigenal spine and adjacent border showing openings in tips of border spines, X 30; 17, oblique interior view of eye surface showing facets, X 30. Locality 9. 18. Postero-median part of incomplete cranidium showing pits in tip of median occipital spine, X 30. Locality 10. PLATE 11 PLATE 12 Figure Plate 12 Diacnnthaspis scitulus n.sp. Oranda formation, locality 8 1-3. Holotype cranidium and left free cheek: dorsal stereograph, anterior and anterolateral views, X 9. USXM 116525. 4-6. Paratype segment from anterior part of thorax: posterior, dorsal, interior views, X 10. USXM 116526b. 7. Cranidium: interior view showing short third glabellar fur- lows, X 20. 8-10. Paratype segment from posterior i^art of thorax : anterior, dorsal, right lateral views, X 10. USX^M 116526c. 11. Small pygidiuni : dorsal view, X 15. 12-1-t. Paratj-pe pygidiiun: posterior, right lateral views, dorsal stereograph, X 9. USNM 116526e. 15. Pygidium: interior view, X 9. 16. Small pygidiuni : dorsal view, X 30. 17-19. Paratype, last two segments of thorax and pygidium: 17, left lateral view, X 15; 18, ventral view, X 10; 19, dorsal view, X 15. USXM 116526d. 20. Free elieek: oblique exterior view showing broad flattened edge along suture, X 15. Plate 13 Biacanthaspis scitulus n.sp. Oranda formation, locality 8 Figure 1, 2. Small eranidium: dorsal, anterior views, X 9. 3-5. Smallest knovm eranidium: dorsal, anterior, right lateral views, X 15. 6,7. Free cheek showing injury: oltlique interior and oblique ex- terior vieAvs, X 9. 8,10,11. Paratype hypostome: exterior stereograph, posterior and left lateral views, X 15. USXM 116526a. 9. Hypostome : interior stereograph, X 15. 112.1"). Paratype pygidium (original of Plate 12, figures 12-14): enlarged views of external surface to show tips of spines, X 30. USXM 116526e. 13. Small free cheek : oblique exterior view, X 30. 14. Free cheek, tip of librigenal spine showing openings in spines on external surface, X 30. 16,17. Cranidium: 16, part of glabella, right fixed cheek and an- terior border showing spines and granules, X 30; 17, part of frouto-niedian glabellar lobe showing minute openings in tips of paired axial spines, X 50. PLATE 13 PLATE 14 Plate 14 ■'Miraspis sp.ind. Lower Edinburg formation, locality 3 Figtire 1. Pygidium: dorsal stereograph, X 9. 7. Free cheek, oblique exterior view, X 9. Ceratoceplmla lacinmta Whittington and Evitt 2. Small eranidium : exterior stereogi'aph, X 30. 3. Small eranidium: exterior stereograph, X 30. Original of Text figure 16. 4. Outer parts of pleurae of two posterior thoracic segments, pygidium, and objects of uncertain affinities (pin supporting this object not blacked out) arranged in approximate relative positions: ventral view, X 6. 5. (i. Objects of uncertain affinities, smallest size known: exterior and interior views, X 10. 8, 9. Left- and right-hand objects of uncertain affinities, arranged in juxtaposition with a iiygidinm : dorsal and posterior views, X 6. 10. Objects of imcertain affinities: "dorsal" view, X 9. 11. Small liypostome: interior view showing tiny opening through doublure at shoulder, X 10. 12. Paratype hypostome (original of Whittington and Evitt, 1954, PI. 6, figs. 10-13) : enlarged view of part of interior, showing openings through doublure at shoulder, X 20. USXM 116503e. 13. Object of uncertain affinities: "ventral" view showing pos- terior edge of "pleura" X 10. 14,15. Objects of uncertain affinities: 14, tips of "pleural" spines showing openings at tips of barbs, X 30; 15, oblique interior view showing anterior and inner edges of "pleura" X 15. Plate 15 Ceratocephala rarispina u.sp. Oranda formation, locality 8 Figure 1, 4, 5, 7. Holotype cranidium : exterior stereograph, anterior, interior and dorsal views, X 9. USNM 116527. 2,12. Paratype free cheek: exterior and right lateral views, X 9. USNM 124698a. 3, 6. 8. Fragment of left side of cranidium with iacomplete free cheek attached, showing course of suture and poorly preserved eye lobe: oblique exterior, exterior, and anterior views, X 9. 9. Pleural portion of segment from anterior part of thorax: dorsal view, X 9 10,13. Paratype segment from anterior part of thorax: posterior and dorsal views, X 9. USNM 124698c. 11,14-16. Cranidium: left lateral, anterior, posterior, exterior views, X 10. 17. Incomplete segment from posterior part of thorax: dorsal view, X 9. 18,20,21. Paratype pygidium: dorsal, posterior and left lateral views, X 15. USNM 124698d 19. Pygidium: dorsal view, X 30 22, 23. Paratype hypostome: interior and exterior views, X 30. USNM 12469Sh. 24. Small oranidiuni and left free cheek: exterior stereograph, X 20. 25, 28. Oranidiuni, incomplete but showing entire anterior border (compare with Figure 1) : 25, exterior view, X 10; 28, part of axial region showmg depressions (appearing as dark spots) in tips of spines, X 30. 29. Free cheek: oblique exterior view, X 20. Ceratoceplmla (Ceratocephalina n.subgen.) sp.iml. Lower Edinburg formation, locality 4. 26,27,30. Cranidium: anterior, right lateral, dorsal views. X 15. PLATE 15 PLATE 16 Plate 16 Ccratocephahi (Ceratoceplialina) tridens u.subgeu., n.sp. Lower Edinburg limestone Figure 1,3. Holotype eranidiuni and yaratype free cheek: exterior and anterior stereographs, X 9. Locality 4. USNM 124699, 124700a. 2,18. Holotype cranidium: right lateral and dorsal views, X 9. Locality 4. USNM 124699. Ostracod carapace adhering to base of glabella on right side (figure 18) was removed before photographs forming figures 1-3 were made. 4, o, 7. Cranidium : right lateral view, anterior stereograph, dorsal view, X 9. Locality 4. 6. Incomplete cranidium : interior view, X 10. Locality 4. 8. Small cranidium and free cheek: exterior stereograph, X 20. Locality 4. 9. Paratype free cheek: oblique interior view, X 9. Locality 4. USNM 124700a. 10,11,13. Paratype, 2 thoracic segments: dorsal, ventral and anterior views, X 9. Locality 4, USNM 124700b. 12. Paratype thoracic segment : dorsal view, X 9. Locality 4. USNM 124700b. 14. Small free cheek : oblique exterior view, X 30. Locality 4. 15-17. Paratype pygidiuui: dorsal stereograph, right lateral, pos- terior views, X 15. Locality 4. USNM 124700c. 19,20. Hypostome: interior and exterior views, X 30. Locality 4. 21,22. Hypostome: interior and exterior views, X 30. Locality 4. 23. Free cheek: oblique exterior view showing eye surface, X 20. Locality 3. 24. Free cheek: oblique exterior view, X 20. Locality 4. Figure Plate 17 Apianurus barbatus n.gen., n.sp. Lower Edinburg limestone 1,16. Paratype free cheek: exterior and left lateral views, X 4. Locality 2, USNM 124702a. 2. Cranidium : exterior stereograph, X 4. Locality 3. 3, 4, 6. Holotype cranidium : exterior stereograph, dorsal and left lateral views, X 4. Locality 2, USNM 124701. 5,7,9-11. Cranidium from Locality 2, free cheek from locality 3: an- terior view, anterolateral stereograph, right lateral view, left antero-lateral view, dorsal stereograph, X 4. 8,212. Cranidium: posterior and interior views, X 4. Locality 3. 13. Eight free cheek: lateral view, X 9. Locality 2. 14. Left free cheek: exterior view, X 9. Locality 2. 15. Left free cheek: ventral view, X 4. Locality 3. 17,19. Hypostome: exterior view, oblique interior view showing opening through doublure at shoulder and posterior wing, X 9. Locality 2. 18. Free cheek: interior view, X 6. Locality 3. 20. Free cheek with complete librigcual spine: dorsal view, X 6. Locality 2. 21. Right free cheek: exterior view showing eye surface, X 30. Locality 2. PLATE 17 PLATE 18 Figure Plate 18 Apianurus barbatus n.gen., n.sp. Lower Edinburg limestone 1-5. Paratype hypostome: exterior stereograph, interior, left lat- eral, posterior, anterior views, X 4. Locality 2. USNM 124702b. 6. Paratype anterior thoracic segment: dorsal view, X 4. Local- ity 2. USXM 124702d. 7, 8. Anterior thoracic segment : dorsal and posterior views, X 10. Locality 2. 9. Paratype thoracic segment : dorsal stereograph, X 4. Locality 2. USNM 124702c. 10,14,15. Paratype incomplete thoracic segment: left lateral, dorsal, anterior views, X 4. Locality 2. USXM 124702d. 11-13. Paratype pygidium: left lateral view, dorsal stereograph, pos- terior view, X 4. Locality 2. USNM 124702e. 16-18. Small pygidium: posterior, left lateral, dorsal views, X 10. Locality 2. 19. Pygidium: enlarged view of anterior part of axis showing spines on external surface, X 30. Locality 2. 20,21. Small pygidium: dorsal and posterior views, X 10. Locality 2. 22. Cranidium : enlarged view of median part of occipital ring between occipital spines showing external surface, X 30. Lo- cality 2. 23,24. Small holaspid pygidium: dorsal and ventral views, X 30. Locality 4. Figure 12, 13. 14 IG. 17, .21. 18 2U. Plate 19 Apianurus harbatus n.gen., n.sp. Lower Edinburg limestone 1,3. Stage 0 free c-hcek: exterior and interior views, X 30. Local ity 4. Original of Text-figure 22A. 2. Stage 0 cranidium : exterior stereograph, X 30. Locality 3. Original of Text-figure 22A. 4,6. Stage 0 (?) cranidium lacking right fixigenal spine, exterior and interior views, X 30. Locality 4. 5. Stage 0 transitory pygidiuni: exterior stereograph, X 30. Locality 4. Original of Text-figure 22A. 7. Small cranidium: exterior view, X 30. Locality 2. Original of Text figure 22C. 8. Small free cheek: exterior view. X 30. Locality 4. Original of Text-figure 22C. 9,10. Small cranidium: exterior and interior views, X 30. Locality 3. Original of Text-figure 22B. 11. Small free cheek: exterior view, X 15. Locality 4. Original of Text-figure 22D. Small hypostome: exterior and interior views, X 30. Locality 3. Small cranidium: dorsal, anterior, right lateral views, X 15. Locality 2. Small hypostome: exterior and interior views, X 30. Locality 3. Small cranidium: dorsal, anterior and left lateral views, X 10. Locality 2. 22, 23. Small cranidium: exterior and left lateral views, X 15. Local- ity 2. 24. Transitory pygidium : dorsal view, X 30. Locality 4. 25. Transitory pygidium: dorsal view, X 30. Locality 4. 26. Transitory pygidium, Stage holaspis-1 ; dorsal view, X 30. Locality 2. 27. 28. Transitory pygidium: dorsal and ventral views, X 30. Locality 4. 29. Transitory pygidium: dorsal view, X 30. Locality 4. PLATE 19 PLATE 20 Figure Plate 20 Apicunurus harbatus n.gen., ii.sp. Oranda formation, locality 8 1-3,11. Cranidium: exterior stereograph, X 4; anterior, right lateral, and interior views, X 6. 4. Cranidium: exterior stereograph, X 6. 5,7. Free cheek: dorsal and right lateral views, X 4. 6, 8, 10. Hypostome : right lateral view, exterior stereograph, posterior view, X i. 9. Small hypostome: exterior view, X 30. 12.16. Segment from anterior part of thorax: dorsal and left lateral views, X 4. 13,14,19. Pygidium: dorsal stereograph; ventral and posterior views, vieys, X 4. 15.17. Segment from posterior part of thorax; right lateral and dorsal views, X 4. Apianurus aff. furcata (Linnarsson) Middle Ordovician, Chasmops limestone, southern Norway 18,21. Incomplete cranidium: anterior view, exterior stereograph, X 2. Guttormsberget. PMO 66690. 20. Pygidium: dorsal view, X 10. Bygd£?y. PMO 5647. 22,24,25. Incomplete cranidium: dorsal stereograph, anterior and ex- terior views, X 4. Guttormsberget. PMO 66691a. 23. Cranidium: exterior view, X 2.2 Fernholmen. PMO 3673. Figure Plate 21 Apianurus glaber ii.geii., u.sp. Lower Edinburg limestone, locality 2 1,3. Holotype eranidiiim: exterior stereograph, right lateral, dorsal views, X 4. USXM 124703. 4,5. Incomplete cranidlum: dorsal and exterior views, X 4. 6,7. Incomplete cranidium: interior and posterolateral views, X 6. 8, 10. Paratype pygidium: posterior, right lateral and interior views, X 4. USXM 124704a. 11,12. Incomplete cranidium: exterior and left lateral A'iews, X 4. 13,14. Paratype hypostome: interior view, exterior stereograph, X 6. ITSXM 124704b. 13. Pygidium: dorsal stereograph, X 4. Apianurus sp.ind. Lower Edinburg limestone, locality 2 16,17,19. Pygidium; dorsal stereograph, X 6; posterior and left lateral views, X 4. USNM 124705a. 18. Pygidium: dorsal stereograph, X 6. TSXM 124705b. 20, 25. Hypostome : exterior stereograph, interior view, X 9. USXM 124705d. 21. Pygidium: dorsal view, X 15. USXM 124705c. 22,26. Hypostome: interior and exterior views, X 9. USNM 124705e. Odoutopleurid protaspis, ? subfamily Miraspiuae Lower Edinburg limestone, locality 4 23, 24. - Exterior stereograph, right lateral view, X 33.6. Photograph by W. R. Evitt. PLATE 21 PLATE 22 Figure Plate 22 Calipernurus bisoUtus n.gen., u.sp. Lower Edinburg limestone 1-3, H. Holotyye, incomplete ceplialon: exterior stereograph, left lateral, anterior, posterior views, X 9. Locality 3. USNM 124711. 4,5,7. Cranidiiun (Locality 3) and free cheek (Locality 2): dorsal and antero-lateral stereographs, exterior view, X 6. Fi(/iire Plate 23 Calipernurus insolitus ii.gen., ii.sp. Lower Edinljurg limestone 1,3. Holotype, incomplete cepbalon: ventral, dorsal views, X 9. Lo- cality 3. USNM 124711. 2. 4. Crauidium (original of Plate 22, figures -4, 5, 7) : posterior, interior views, X 4. Locality 3. 5.7. C'ranidiiim: dorsal and exterior views, X 9. Locality 3. (i. Tip of libiigenal .si)ine slio\vin<;' openings at base of barbs, X Md. Locality 3. 8. Cianidiuin ; exterior view showing pits in median occipital tul)ercle, X 15. Locality 3. 9, lU. Free cheek (original of Plate 22, figures 4, 5 and 7) ; oblique interior and oldique exterior views of eye surface, X 30. Local- ity 2. 11, 12,15. I'aratype hypostome: exterior stereograph, X G; interior view, X 6; oblique interior view, X 15 (arrow points to tiny open- ing in doublure). Locality 2. USNM 124712a. 13,14. Hypostome: exterior stereograph, interior view, X 9. Locality PLATE 23 PLATE 24 Plate 24 Calipernurus insolitus n.gen., n.sp. Lower Edinburg limestone Figure I. Small cranidiuiu : exterior stereograph, X 30. Locality 3. 2-4. Small craiiidiuin: right lateral, anterior, dorsal views, X 10. Locality 2. 5-7. Small erauidiimi (Locality 2) and tree cheek (Locality 3) : left lateral view, exterior stereograph, anterior view, X 10. 8.13,14. Small pygidium : left lateral, posterior, dorsal views, X 1,. Locality 3. 9. Cranidium : exterior stereograph, X 10. Locality 2. 10. Small segment from posterior part of thorax: dorsal view, X 10. Locality 3. II. Small pygidinm lacking median border spine: dorsal view, X 12. Locality 3. 12. Transitory pygidium : dorsal view, X 15. Locality 2. 15,16,19. Paratype segment: anterior, dorsal, posterior views, X 4.5. Locality 3. USNM 124712b. 17,20. Pygidimn with median border spine distally bifurcated: dorsal and ventral views, X 10. Locality 3. 18. Pygidium lacking median border spine : dorsal view, X 9. 21,25,27. Paratype pygidium: 21, oblique posterior view showing tips of major border spines and openings at base of barbs, X 30; 25, posterior view, X 6 ; 27, dorsal stereograph. X 6. Locality 2, USNM 124712d. 22,23. Incomplete segment from posterior part of thorax: dorsal and left lateral views, X 6. Locality 2. 24. Tip of major border spine of pygidium showing openings at base of barbs, X 30. Locality 2. 26. Paratype, 2 incomplete segments from posterior part of thorax: dorsal view, X 9. Locality 3. USXM 124712c. 28. Pygidium: dorsal stereograph, X 6. Locality 2. Bulletin of the Museum of Comparative Zoology AT H A R V A E T) COLLEGE Vol. 114, No. 6 A CAT SKELETON WITH AN ANOMALOUS THIRD HIND LEG AND ABNORMAL VERTEBRAE By Thomas S. Parsons and John M. Stein WiTH Four Plates CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE :\r U S E U M i\l.\Y, 1956 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 114. Breviora (octavo) 1952 — No. 53 is cui-rent. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. Johnsonia (quarto) 1941 — A publication of the Department of Mollusks. Vol. 3, no. 35 is current. Occasional Papers of the Department of ^Iollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo "i 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, ( 'ambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 114, No. 6 A CAT SKELETON WITH AN ANOMALOUS THIRD HIND LEG AND ABNORMAL VERTEBRAE By Thomas S. Parsons and John M. Stein With Four Plates CAMBRIDGE, MASS., U.S.A. printed for the imuseum May, 1956 No. 6 — A Cat Skeleton with an Anomalous Third Hind Leg and Abnormal Vertebrae By Thomas S. Parsons and John M. Stein Harvard Biological Laboratories Cambridge 38, Massachusetts INTRODUCTION The cat skeleton here described (see Plate 1) displays on the right side a highly abnormal innominate bone articulating with two legs, the more anterior of which has two tibiae and is evi- dently the product of a fusion of two legs. The posterior verte- brae show various fusions and abnormalities. The skeleton was kindly lent to us by its owner, Mr. Elling 0. Eide, of Sarasota, Florida. The cat was a family pet which lived over six and a half years and could run and climb trees despite its extreme pelvic abnormalities. Its appearance in life is shown in Plate 3A. It was buried shortly after its death and the skeleton dug up fifteen months later. Although a few of the smaller bones were not recovered, the completeness of the recovery is shown by the finding of twenty-eight of the sesamoids. The mother of the cat was apparently normal and produced many litters of kittens, probably many by the same father whose only abnormality was Polydactyly. The one described here had two normal littermates. The mother was slightly over seven and a half years old when this litter was born. She was very sick for several weeks after their birth, but recovered and had normal litters afterwards. Most of her kittens were normal except for numerous cases of Polydactyly, but one born approximately three years before the one here described had two symmetrically fused heads. It was found dead, and whether or not it was alive at birth is not known. Due to the burial, only the skeleton of the animal was re- covered, and no attempt has been made to reconstruct any of the soft parts. The skeleton has, however, been compared with five normal ones, one articulated and four disarticulated, as well as with the extensive descriptions and illustrations in Jayne (1898). The only definite evidence on the articulation of this skeleton in life comes from the nature of the articulatory surfaces and the excellent x-ray shown in Plate 2. The skeleton was mounted as shown in Plate 1 ; no attempt was made to include most of 294 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the tail, the ribs, sternum, phalanges, or hyoid apparatus in this reconstruction. Although the work of making the original descriptions was divided between the authors, both of us have gone over all the parts and are equally responsible for this account. In order to give morphological descriptions, it is necessary for us to give names to several of the anomalous elements before the reasons for our interpretations of their nature can be stated. However, we hope that the reader can separate our theories from the more objective accounts of the various bones. We would like to thank several people for their kind assistance in the writing of this paper : first, Mr. Elling 0. Eide, who not only loaned us the skeleton, but also gave us its history, provided the photograph of the live cat, and made the drawings ; Dr. C. K. Newton, D.V.M., of Bradenton, Florida, who provided the x-ray of the hind legs; and especially Dr. Ernest E. Williams of Har- vard University, who offered much helpful advice throughout the course of the study. DESCRIPTION General Much of the skeleton, such as the pectoral girdles and limbs, is normal and needs no detailed description. The skull and jaws are also normal, but have a very stocky appearance. Both are short and very wide ; the areas of attachment of the jaw muscles are noticeably larger than average. The canine teeth are very large. The third right lower premolar, the third right upper incisor, and the third left upper premolar are missing and the sockets filled with l^one, but these probably were lost after ma- turity was reached. The two second upper premolars are also missing, but they seem never to have been present, a relatively common occurrence, according to Jayne (1898). The first and second left upper incisors were broken oft', apparently before death. Only the tAvo stylohyals of the hyoid apparatus were re- covered. They appear normal. The ribs seem normal except that the twelfth right one is lack- ing due to the malformation of the vertebra. The recovery of the sternum was incomplete, but the manubrium and five sternebrae (two of them fused) are present. All appear slightly thicker PARSONS AND STEIN : ANOMALOUS CAT SKELETON 295 than usual for their length, especially the fused pair. Their ends appear rather asymmetrical, but in the absence of the cartilages between them this indicates little. The fused ones each send out a short prong of bone, on either side, at the fused ends which must liave articulated with the costal cartilages. Vertebral column The anterior lialf of the vertebral column appears normal; the posterior part shows abnormalities. There are seven eervicals, thirteen thoracics, .seA'en lumbars, four sacrals (one more than usual), and nineteen caudals. The last number, although low, is within the range given by Jayue (1898), and some of the terminal ones ma}' not have been recovered. Only those vertebrae which show marked peculiarities Avill be described here. The first such is the tenth thoracic. The neural spine and post- zygapophyses are twisted approximately 15 degrees to the left. The posterior articular surfaces are more nearly horizontal and the anapophyses shorter than usual, but this vertebra is still relatively normal in general appearance. The last three thoracic vertebrae are fused into one mass (Fig. 1). The eleventh is joined to the tAvelfth by the dorsal part of A B Fig. 1. Eleventh, twelfth, and thirteenth thoracic vertebrae. Anterior end to right. A, dorsal view. B, ventral view. xl. the centrum and a small flange of bone just ventral to the left anapophysis. In the former the two sides of the neural arch meet dorsally in the midline, but are not fused, and there is no trace of a neural spine. All of the articular processes appear quite normal in shape, although the postzj-gapophyses are deflected slightly to the left and the right metapophysis is more anterior 296 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY than lateral to the prezygapophyseal facet. Well developed tuber- cular facets are present and the capitula of the ribs articulate intervertebrally as is usual in the more anterior thoracic verte- brae, but not common in the eleventh (see Jayne, 1898, p. 84; however, his fig. 55, p. 82, shows a situation similar to that in this specimen). The twelfth thoracic vertebra is represented ventrally only by the wedge-shaped left half of the centrum (Fig. IB). On the right side the eleventh and thirteenth thoracics come together without any gap between them. Dorsally (Fig. lA), both sides of the neural arch of the twelfth are present and tightly fused to the last thoracic, but the right side is far smaller than the left. Only a left rib articulates with this vertebra, and that to a large convex capitular facet. The prezygapophyses are quite normal, but the right metapophysis is much reduced. The postzygapo- physes are fused with the prezygapophyses of the following verte- bra forming dorsal mounds. An anapophysis is present only on the left. The small neural spine points straight dorsally. This vertebra is thus anticlinal, using Jayne 's (1898) definition; he gives the eleventh as the normal anticlinal vertebra, while Gott- lieb (1914) states that it can be either the tenth or the eleventh. The intervertebral notch on the right appears double, so pre- sumably the normal number of spinal nerves was present. The thirteenth thoracic vertebra (Fig. 1) is the most normal looking of the three fused ones. The prezygapophyses lack dis- tinct metapophyses due to their fusion with the postzygapophyses of the preceding vertebra, but there are raised roughened areas which probably correspond to them. The neural spine is quite small and directed anteriorly. The postzygapophyses and ana- pophyses are displaced to the left with the right ones dorsal to the left. The latter are normal, but the right postzygapophysis is elongate, while the right anapophysis is higher than long, ar- ticulating with the entire height of the following prezygapo- physis. The first lumbar vertebra (Fig. 2A) is markedly asymmetrical with its centrum about 9 mm. long on the right side and 12 mm. on the left. The neural arch appears slanted to the left so that the left pedicle is abnormally thick while the right, although in contact with the centrum, is not fused to it. The neural canal is thus displaced to the right. The right prezygapophysis is vertical PARSONS AND STEIN : ANOMALOUS CAT SKELETON 297 and massive while the left is thinner and projects dorsolaterally. Only the latter bears a normal metapophysis. The postzygapo- physes are both displaced to the left as are the anapophyses. The Fig. 2. Lumbar vertebrae. Anterior end to right in B, D, and E and to left in C. A, first in posterior view. B, second in dorsal view. C, third in posterolateral view. D, fourth and fifth in ventral view. E, sixth in dorsal view. xl. latter are quite short and round in section with the right one well dorsal to the left. The rather small neural spine arises from the normal position on the neural arch and slopes towards the right to the midline. The pleurapophyses are rudimentary prongs about 2 mm. long on the right, and 4 mm. on the left. The former is somewhat dorsal to its usual position while the latter is curved slightl}' posteriorly. The second lumbar vertebra (Fig. 2B) is also highly asym- metrical. The prezj'gapoph^'ses, massive structures with nearly vertical articular facets, are displaced to the left with only the right bearing a normal metapophysis. The postzygapophyses and anapophyses appear quite normal, the only irregularity being a slight elongation of the right postzygapophysis. As in the first lumbar, the neural spine leans to the right, almost meeting the 298 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY right prezygapophysis. The pleurapophyses are very small and do not extend ventral to the centrum. The right one points lat- erally and the left anteriorly. The third lumbar vertebra (Fig. 2C) displays spina bifida with both sides of the neural arch forming a separate neural spine. The middorsal cleft is exceedingly narrow, but the two sides do not actually touch at any point. The left neural spine is directed anteriorly and appears much like the anterior half of a normal neural spine. The more posterior right one is directed straight dorsally so that a V-shaped notch is seen between them in lateral view. All the zygapophyses and metapophyses are normal with the left prezygapophysis and right postzygapophysis slightly larger than the others. The anapophyses, especially the left, are reduced in size and are spinous rather than blade-like. The pleurapophyses while normal in shape are also reduced, being intermediate in size between those of the first and second lumbar vertebrae of the four normal cats used for comparison. The fourth and fifth lumbars (Fig. 2D) are fused by a large rounded mass of bone ventral to the two centra. The normal parts of the two, which are held in the articulated position, show no fusion. The ventral mass is fused to the sides of the centrum of the fourth lumbar leaving a small midventral canal between it and the centrum ; no such canal is present in the following vertebra. The anterior face of the fourth lumbar centrum faces anteroventrally producing a distinct angle in the articulated spine at this joint. The pleurapophyses are small and more ventrallj' directed than usual. The right one points slightly more laterally than tlie left. The metapophyses are not distinctly de- veloped and the left anapophysis is rudimentary; otherwise the processes of this vertebra are normal. The fifth lumbar vertebra also has quite normal dorsal proc- esses. The right anapophysis is almost completely absent, a con- dition which was also seen in the normal skeletons used for com- parison. The left postzygapophysis is slightly elongated, and the posterior end of the neural spine is deflected to the right. The other processes show no modifications. The pleurapophyses are somewhat reduced (Fig. 2D), and the right one especially does not extend laterally as far as usual. The complex of the fourth and fifth lumbars shows a slight curvature with the length of the two centra slightly greater on the left than on the right. It PARSONS AXD STEIN : ANOMALOUS CAT SKELETON 299 is impossible to say exactly how this curvature is brought about because of the mass of bone ventral to the two centra obscuring their articulation. The sixth lumbar vertebra (Fig. 2E) also has greatly reduced pleurapophyses with the right one considerably smaller than the left. Both extend anteroventrally but not laterally, as would be the normal condition. The posterior face of the centrum is some- what dorsally directed, thus making another sharp angle in the articulated spine. The prezygapophyses and metapophyses are normal with those on the left slightly larger. The neural spine leans towards the left. On the right, the postzygapophysis is much reduced and the anapophysis completely absent, while on the left the former is greatly enlarged and the latter present although very small. The seventh lumbar vertebra is tightly fused to the sacrum dorsally, although ventrally the centra remain separate (Fig. 3). It does not enter into the sacroiliac joint. The left pleurapophy- sis, while short, is fairly normal, but the right one is absent except for a broad truncated base extending just anterior to the side of the centrum. The left prezygapophysis is a simple dorsal projection lacking a metapophysis completely (as in one of the four disarticulated normal cats used for comparison). On the right the entire process is a low mound reaching posteriorly as far as the right neural spine. The left postzygapophysis is com- pletely fused to the prezygapophysis of the first sacral vertebra, the two being represented by a small mound. There is no trace of any corresponding articulation on the right. The neural arch apparently did not fuse middorsally until very late in develop- ment since there is a distinct suture curving to the left poste- riorly. Two neural spines are present, their arrangement being similar to that in the third lumbar except that in the seventh the right is anterior. The latter expands anteroventrally where it has a concave anterior face into which the prezygapophysis ex- tends. The neural canal is triangular rather than quadrilateral in end view. The sacrum is shown in Figure 3. It is, as a unit, somewhat twisted. The posterior part is slightly curved to the left, mainly by unequal growth of the sides of the second sacral vertebra, and accompanying this there is a counterclockwise rotation about the long axis of the body when seen in posterior view. The left side 300 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of the sacrum is quite normal despite the presence of a fourth sacral vertebra. There are three small mounds representing: fused zygapophyses and three sacral foramina both dorsally and ventrally. The sacroiliac joint on this side is apparently normal Fig. 3. Seventh lumbar vertebra and sacrum. Anterior end to right, dorsal view. B, ventral view. C, lateral view. x%. A, although the auricular facet is slightly larger than figured by Jayne (1898, fig. 75, p. 103) or in the four normal cats exam- ined. The neural spines are fused into a low longitudinal ridge with a definite spine for the first, and small gentle rises for the middle two, but no distinct spine on the fourth sacral. The left pleurapophysis of the fourth sacral is slightly smaller and more laterally directed than the right. The right side of the sacrum is highly abnormal. There are no signs of the three pairs of fused zygapophyses and only the pos- terior two pairs of sacral foramina are normal. The surface articulating with the ilium is split into two separate parts. The larger and more posterior is an irregular, posterolaterally di- rected, oval facet on the expanded transverse process of the sec- ond sacral. The other is a small circular facet at the posterior end of the first sacral about midwav dorsal. It is laterallv di- rected. The first sacral lacks a large expanded transverse process on the right, liaving only this articular facet on a small process which does not extend as far laterally as the second sacral (from which it is separated by a narrow groove), and a posteriorly directed ventral process somewhat resembling a caudal rib. This PARSONS AND STEIN : ANOMALOUS CAT SKELETON 301 ends just ventral to, but not fused with, the transverse process of the second sacral. There is thus neither a dorsal nor a ventral sacral foramen l)etween these two vertebrae, but a large lateral intervertebral foramen. Of the nineteen caudals, only the first six show any abnormali- ties. The first three show slight asymmetries in the zygapophyses, but nothing else unusual. Numbers four through six are fused together by their neural arches, with their centra remaining quite distinct. As in the last three thoracics, the middle one, the fifth caudal, is represented only by one side, in this case the right. Apparently the left side is completely absent even dorsally, the postzygapophysis of the fourth articulating directly with the prezygapophysis of the sixth on that side. On the sixth there is a small prong of bone on the centrum representing part of a haemal arch ; a similar condition was observed in one of the nor- mal cats used for comparison. Pelvic girdles and appendages The left innominate bone seems quite normal except that the auricular articulation is unusually large and the spine of the pubis for the attachment of the rectus abdominis is absent. One of the normal cats u.sed for comparison also lacked the latter structure. The left femur, tibia, and fibula all appear quite normal. The right innominate bone (Fig. 4) is a very deformed struc- ture, seemingly divided into two parts : one part consists of the right half of the more normal or left pelvis; the other is a very disfigured complete second pelvis. This second or teratological pelvis and the more normal right innominate are solidly fused in two places : one between the spine and horizontal ramus of the more normal pubis and the medial portions of the teratological right ischium, the other between the anteromedial portion of the more normal ilium and the anterolateral portion of the tera- tological right ilium. A distinct opening occurs between the two l)oints of fusion (Fig. 4A, oi). Thus the teratological pelvis is suspended by the more normal right innominate in a parasagittal plane of the animal. The part of the teratological pelvis normally ventral faces laterally to the right. The ilium of thp more normal right innominate is about 1 cm. 302 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY shorter than the left. The suture (Figs. 4A and B, sut) between the two ilia is solid yet clearly defined. Approximately 7 mm. posterior to the suture on the medial side is an oval articulation with the sacrum (4x7 mm. in size; Fig. 4A, al). .Starting about 10 mm. from the anterior end of the ilium and extending for 40 mm. on the lateral side and even overlapping for al 0' so sp Fig. 4. More normal right innominate and entire teratological pelvis. A, medial view. B, lateral view. C, ventromedial view. Abbrevia- tions: al, regular auricular facet; a2, right auricular facet of tera- tological pelvis ; act, acetabulum of teratological pelvis ; ft, ischial tuberosities of teratological pelvis; Iti, left teratological ilium; oi, gap l)etween regiilMi- right innominate and teratological pelvis; ot, obturator f enestrae of teratological pelvis ; sp, spongy bone mass; sitt, suture between regular right innominate and teratological pelvis ; nyin, regular pubic symphysis ; xs, bony ridge ventral to pubic sjTnphysis. x^ cy2. 12 mm. on the medial surface lies a spongy mass of bone (Figs. 4A and B, up). Part of this spongy mass forms the articulation for the more normal right femur. The acetabulum is irregular in shape and has a very rough surface as does the head of the articulating fenuir. This spongy bone seems to have replaced arthritic cartilage and to have been applied to the more normal bone underlying it. Some rough bone also appears on the medial side of the ischium, dorsal to the obturator fenestra. PARSONS AND STEIN : ANOMALOUS CAT SKELETON 303 The iscliiiim of the more normal innominate is bent ventrally and laterally to the right more strongly than in any of the four cats used for comparison. The bone on the ventral side of the ischium is rough and the tuberosity is unusually large. At the symphysis, the fit between the right and left halves of the left pelvis is good, but a small ridge of bone extends ventrally along the entire right part of the symphysis (Fig. 4C, xs) . The descend- ing ramus of the pubis of its right innominate is much thicker dorsoventrally than that of the four normal specimens. The hori- zontal ramus is fused with the teratological right ischium for about 17 mm. This suture is very solid, but the line of fusion can be seen quite clearly by the grain of the bone. The teratological pelvis looks like a fat, symmetrical "Y" from a lateral view. There is a small fenestra near the base of the Y, this presumably being the fused obturator fenestrae from both sides (Fig. 4A, of). The right ilium (the posterodorsal arm of the Y) is shorter than the left and has an oval auricular facet (Fig. 4A, a3) about 3x5 mm., located 9 mm. from the right upper end of the Y, and 4 mm. from the line of fusion with the more normal right ilium. The teratological left ilium is located on the right side of the cat 's central axis. The left teratological ilium looks slightly more normal in shape than the right teratological one does. The right pleurapophj'sis of the seventh lumbar verte- bra is fused to the left teratological ilium. It will be remembered that the rest of the seventh lumbar is fused to the sacrum. From ventrolateral aspect, the teratological innominate has one fairly smooth acetabular articulation (Fig. 4B, act). This articu- lation has two distinct sides, right and left. Posterior to the fused obturator fenestrae are the fused tuberosities of the isehia (Fig. 4A, ft). The ilia on both sides are of smooth solid bone. The articulation of the femur with the more normal right in- nominate is a knotty, gnarled mass of bone (Fig. 5 A, ar). The articular surface itself seems worn enough to have allowed some slight movements. The abnormal area seems to have been formed on top of smoother bone and is probably ossified cartilage. Only the posterolateral part of the head is present and that is formed in rough bone. An irregular pit occurs in place of the rest of the head (Fig. 5A, pt). The lesser trochanter (Fig. 5B, It), also covered with rough bone, seems also to have articulated with the innominate slightly posterior and medial to the head articu- 304 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY oc Figure 5 PARSONS AND STEIN : ANOMALOUS CAT SKELETON 305 latiou. This places the femur in a position approximately per- pendicular to the central axis of the cat. The shaft is almost C3'lindrical, i.e., not graded in thickness as were those of the cats used for comparison. Although slightly rough in appearance, the greater trochanter has a normal shape. Distally, the condyles are bent about 15 degrees medially, considering the bone in the normal position (Fig. 5B). The articulatory surfaces of the condyles are smaller than those of the four normal cats ; they are also slanted inward towards the axis of the bone and do not extend outward as they should. The medial condyle has a long curved articulation (Fig. 5B, ic) ; the lateral condyle has a short oval articular surface (Fig. 5B, oc). Proximally the more normal right tibia has unevenly developed articular surfaces. The lateral articular surface (Fig. 5C, oa) is less extended than in the normal eats, whereas the medial surface (Fig. 5C, ia) extends in an arc with the lateral articulation as a center. Considering the shape of this articulation and that of the distal end of the femur, it seems as though the knee joint had a swivel action as well as a hinge in order to compensate for the very slight motility of the acetabular articulation. The rest of the bone appears normal, as does the fibula of this leg. The fusion of two femora, one right and one left, apparently formed the double femur which articulates at act, Figure 4B. with the fused teratological pelvis. The articulation is smooth and allowed hip movement. The distal portions of the femora are fused at an angle of about 60 degrees (Fig. 5D). Proximally the two component bones have turned so that their heads and tlu- greater and lesser trochanters all have fused with their counter- parts from the other side, the double bone having one head, one greater trochanter, and one lesser trochanter all located in the Fig. 5. A and B, more normal right femur in anterior and posteroventral views. C, more normal right tibia in posterodorsal view. D, E, and F, fused teratological femur in actual ventromedial (morphological posterior), actual anterior (morphological left lateral), and actual lateral (morphological anterior) views, respectively. Abbreviations: ar, irregularly formed head; c, canal; gt, greater trochanter; h, head; ia, medial tibial condyle; ic, medial femoral condyle; It. lesser trochanter; oa, lateral tibial condyle; oc, lateral femoral condyle; pt, rough pit; sr, sharp ridge. B, C, and D foreshortened to show articular surfaces. x%. 306 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY medial plane of the bone {Fig. 5F). On one side at the proximal end in the medial plane there is a sharp thin ridge (Fig. 5F, sr) running for about 25 mm. In this region the bone is thin latero- medially and thick anteroposteriorly. Moving distally, the bone gradually widens lateromedially and at the distal end is as thick that way as it was anteroposteriorly before (Figs. 5E and F). A canal (Fig. 5F, c) perforates the bone starting anteriorly (here and for the rest of the description of the long bones of this leg, anterior, posterior, medial, and lateral refer to the normal position, not that in this specimen) about 25 mm. from the distal extremity and ending 10 mm. from that end on the pos- terior side. This canal is directly in the midline of the bone. A large articulation for a sesamoid is present on the lateral surface of the right component. All four condyles are well formed (Fig. 5D). The x-ray shown in Plate 2 gives the relationships at this joint clearly. The two separate tibiae which articulate with the double femur are almost exact mirror images of each other. They lie so that their anterior surfaces are pointed outward from their mutual articulations. These articulations between the two tibiae lie one at the posterior side of the head (Plate 2) and the other at the distal extremity. There are two proximal articular surfaces on each of the bones ; they lie on the posterolateral and anteromedial parts of the head. It appears that the joint between the double femur and the two tibiae must have been almost completely im- mobile. The shafts are unusually thin and the bones are dispro- portionately long in comparison to the other bones in the cat. The tubercle is prominent and the crest is short. Distally all the parts are recognizable, although modified. The overall organization of this double leg is rather baffling to the authors. The anterior halves of the teratological pelvis and of the double femur appear to represent the morphological left side, an interpretation which fits not only the shape of the elements, but also the most probable pattern of fusion. Nonethe- less, certainly in the foot and very probably in the lower leg, the morphological left is posterior. It is as though the knee were mistakenly articulated backwards in the mounting of the skele- ton ; however, the nature of the articular surfaces and the photo- graphs of the living cat seem to preclude this latter possibility. The double fibula is a symmetrical structure formed by the PARSONS AND STEIN : ANOMALOUS CAT SKELETON 307 complete fusion of two fibulae, one right and one left, along their medial sides. At the proximal end, the fusion has caused the loss of the medial parts of both bones, leaving only the lateral parts recognizable. The shaft is thin and bent concavely from the posterior aspect. Distally the median parts of the bones are not lost. There is a fossa with a central ridge in the line of fusion and indications of the lateral maleoli are present on both sides. The astragalar articulations form one single articular surface which is quite rough. The feet of both the regular hind legs appear completely normal, but the tarsus of the abnormal limb is much modified by being in the flexed rather than the normal extended position so as to form an angle of about 120 degrees with the tibiae. This tarsus is a double structure bearing metatarsals 1, 2, 3, 4, 3, 2, and 1, hence with two sides corresponding to the normal median and without any normal lateral. In the description, dorsal, ventral, lateral, and median will be used for the condition in this foot ; all are the reverse of the normal morphological pattern. Although only the disarticulated bones are available, the articular surfaces can be matched very precisely in all cases so that the restoration is almost certainly correct although two elements are missing (Plates 3B, C, and 4). The calcanea, astragali, and scaphoids are fused into one sym- metrical mass, but in most places the sutures can be distinguished. The first form a single midventral bar which expands posteriorly in a rough and immovable articulation with the single fibula and both tibiae. Each astragalus is an L-shaped element with the two arms of approximately equal length. It lies along the side of the calcaneum and has one arm extending dorsally from its proximal end to form the main articulation with the tibiae. The dorsal surfaces are fused into one triangular facet with the apex anterior. The scaphoids are fused together and cover the distal ends of the astragali, but not the more ventral calcanea. The tuberosities are thus dorsal and lateral. The cuboids are represented by a single midventral element which articulates proximally with the calcanea and dorsally with the scaphoids and ectocuneiforms. It is more or less cylindrical with the length about twice the diameter. The distal end is con- cave to receive the proximal end of the single fourth metatarsal. On the left, the ectocuneiform and mesocuneiform are fused 308 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY into a single element which covers almost all of the distal surface of the left scaphoid. The part representing the mesocuneiform is far larger than usual being almost equal to the ectocuneiform in size. However, the former is, as usual, shorter in the proximal- distal direction. The articular surfaces have the same general pattern as in normal specimens, but those of the mesocuneiform are much expanded, especially in height. The left entocuneiform and probably also the vestigial first metatarsal are represented by a single fairly normal-appearing element which lies in the expected position dorsal and lateral to the mesocuneiform and proximal end of the second metatarsal. It also articulates with the scaphoid. On the right, only the ectocuneiform is present. It is, like most of the elements of this tarsus, smaller than usual, compared to normal cats of the same overall size. It lies distal to the ventro- median half of the right scaphoid. The hook on the plantar (here dorsal) surface is much reduced and the ventrolateral corner is drawn out into a small prong, but the element is immediately recognizable as an ectocuneiform. The mesocuneiform was lost but, judging fi'om the articular surfaces of the scaphoid and sec- ond metatarsal, must have been quite large and similar to the one on the left. The right entocuneiform and first metatarsal are also lost and were probably similar to the left, although there is no smooth articular surface for them on the right scaphoid. It is possible that these elements were never present. The overall symmetry of the foot and the notch on the dorsolateral surface of the proximal end of the second metatarsal strongly suggest their presence however. The five metatarsals require little description except for the middle one (number 4, probably representing a fusion of that element from both sides). The others, numbers 2 and 3 of each side, bear normal proximal articular surfaces although those of the third metatarsals are in contact with each other dorsally and both have a small facet for this articulation. The distal ends of these third metatarsals are normal while the seconds both end in rather amorphous rounded swellings. These four are all curved with the concave surface dorsal ; that is, they are not modified for the inverted position of the foot. The outer pair are also curved laterally at their distal ends. The right second metatarsal is somewhat shortened, but the other three are all normal length PARSONS AND STEIN : ANOMALOUS CAT SKELETON 309 although considerably thinner than is usual. The central, fourth, metatarsal is a symmetrical element with both ends much modified. Proximally, it bears a small convex head articulating with the cuboid, while distally it has a larger convex swelling with a small dorsal projection. The curvature is the reverse of the other metatarsals with the concave surface ventral. It lies ventral as well as median to the third metatarsals and is in contact with them proximally. This gives it a triangular shape in section although it is rounded distally. Its length is equal to that of the third metatarsals. There are also three phalanges presumably associated with this foot. However, their position cannot be told from the dis- articulated bones and it is possible that some were lost. Two of the three have fairly normal proximal concave faces, and distally small swellings showing none of the usual features. The third is a short thin cylindrical bone without any distinctive features. DISCUSSION General The main significance of highly abnormal specimens is the in- formation they may give concerning developmental processes or inheritance; however, detailed consideration of these points is not possible in this case although we will mention possible genetic mechanisms. It is hoped that the description is detailed and accurate enough to give future workers another example which will be useful in reaching am' general conclusions. The literature of teratology and the genetics of abnormal de- velopment is very extensive and no attempt at compiling a com- plete bibliography of the described abnormalities was made. The papers cited include only those available ones which seem to bear most directly on the problems discussed. More extensive bibliographies of teratology can be found in such general works as Bateson (1894), Schwalbe (1906-1913), Schlegel (1921), and Stroer (1937). Griineberg (1947) includes many references to the genetical literature, especially concerning rodents. 310 bulletin : museum of comparative zoology Vertebral column The anomalies shown by the vertebrae of this cat are all well known and have been much studied, especially in humans. How- ever, single specimens with so many deformities are rare and seldom reach maturity. While some of the peculiarities may be a direct result of the pelvic malformation, others, though quite possibly caused by the same factors, are not. Possibly some, such as the fusion of the fourth and fifth lumbars, were caused by disease after maturity was reached and are thus outside the prov- ince of this discussion. Spina bifida (or rachischisis) is a well known defect, but in most studies on teratology only extreme cases are considered (as in Kermauner, 1909). Examples such as the eleventh thoracic and third lumbar of the present specimen should have little or no adverse effect upon the animal and hence are of small interest medically. Probably such cases are of more frequent occurrence than the literature would indicate since Jayne (1898) apparently found at least one in the cats he used (see his fig. 83, p. 125). Unfortunately he does not saj' how many were examined, but he had at least part of fifty or more spinal columns. In mice there are many mutations which can cause spina bifida (Griineberg, 1952). The two half-vertebrae, the twelfth thoracic and fifth caudal, also represent a well known anomaly of great theoretical interest. Baur (1891), on the basis of several reptilian skeletons, put forth a theory of increase in the number of segments bj* intercalation with such hemivertebrae representing incomplete cases. Brash (1915) studied a human spine with this defect in a cervical, and reviewed the literature on all cases reported in man. He con- cluded that the hemivertebrae could arise from either the addi- tion or loss of half of a normal vertebra and that in his specimen the condition was secondary, with the absence of the spinal nerve on one side as the primary cause. Similar cases have been re- ported in many forms including the cat (Jaj'ue, 1898, fig. 84. p. 125). The twelfth thoracic in our specimen is presumably an example of the failure of half of the centrum to form, since counting it there are the usual thirteen thoracies. Judging from the foramina, the normal number of spinal nerves was present, so here the condition is primary. Concerning the fifth caudal, no PARSONS AND STEIN : ANOMALOUS CAT SKELETON 311 definite statements can be made. Even commoner and better known are variations in the num- ber of sacral vertebrae. Bateson (1894) gives many cases of this and Dwight (1901) discusses its occurrence in man. Jayne (1898) mentions its occurrence in the cat. The incorporation of the first caudal into the sacrum is not very common according to the last author, but no figures are given. On the other hand, partial fusion of the seventh lumbar and sacrum, often with a tendency for the lumbar to become involved in the sacroiliac joint, is exceedingly common. Klatt figures such a condition in a dog and a cat (1939, figs. 7 and 8, pp. 230 and 231). In the present case there is no actual participation of the last lumbar in the joint, but the right side shows modifications of the pleura- pophysis. This seems to approach sacral articulation with the extra pelvis. More recent workers such as Dauforth (1930b) and Backman (1934) consider the boundaries between the various regions of the vertebrae to be determined by gradients in differentiation. Thus the homology of individual vertebrae cannot be determined by simple counting, whether from the anterior end or any other selected point. This opinion seems w^ell confirmed by the work of Sawin and his co-workers, especially Green, on the lumbar region in rabbits and mice (see Griineberg, 1947, for references). The twenty-seventh vertebra in the cat here described is, accord- ing to this conception, partly sacral as w' ell as lumbar ; the name seventh lumbar is used merely as a convention. This viewpoint also casts grave doubts on the earlier ideas concerning hemi- vertjbrae mentioned above. The spinae bifidae and hemivertebrae could be independent of tHe leg anomalies with the sacral malformations directly de- pendent on the latter, although similar cases of deformed sacra cire known where there are no pelvic abnormalities. However ir is also possible that all the spinal abnormalities are a result of the pelvic duplication since they are all consecutive, running from the tenth thoracic through the sixth caudal. The general asymmetry of the posterior part of the vertebral column may also be partly or wholly dependent on the other anomalies since certainl}" the muscles must have been very abnormal in this area. However such asymmetry is well known especially in the tail where it is very common in some forms such as certain dogs which 312 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY are otherwise quite normal (see Klatt, 1939, and Stoekard et al., 1941, for a discussion of this). Berg (1912) describes a series of cats in which the tail is greatly deformed. However, in almost all these cases the tail is markedh' shortened and the malforma- tion usually is greatest distally, neither of which is true of the cat under considei-ation. The genetics of such deformed tails has been studied by a large series of workers, but is not well understood and need not be considered here. Bamber (1927) reviews the genetieal literature on eats. In mice and other rodents there are many mutants listed by Griineberg (1947 and 1952) which show similar deformation of the caudal vertebrae. The morphological expression and genetic action of these various mutations are very diverse. While most of them tend to shorten the tail, some such as the mutation ' flexed ' (Hunt, Mixter, and Perniar, 1933) have no such effect. In rats, 'stubs' (Ratcliffe and King, 1941) affects not only the tail but also causes abnormalities in the hind limbs. Both of these are recessive genes. There are also several dominants such as 'Brachyury' or 'T' (Chesley, 1935) and 'Sd' (Gluecksohn- Schoenheimer, 1943) which are lethal when homozygous. It is of interest that in a mouse of genotype T/+ Sd/+ there is a cumu- lative effect producing a more highly aberrant vertebral column than either gene does alone, although both normally produce a similar effect ( Gluecksohu-Schoenheimer, 1943). Most of these mutations cause not only asymmetry of the vertebrae, but various other defects, such as spina bifida, are common in mice bearing these mutant genes. It therefore seems possible, though in no wise proven, that the vertebral column of this cat developed abnormally under the control of a mutation similar to some of those known for mice, and that this abnormality was accentuated by the occurrence of another developmental anomaly, genetically controlled or not, resulting in the formation of a second and much modified pelvis and pair of limbs. Another possibility is that all the anomalies are the result of a single event, either a mutated gene or some environmental factor operating during the very early develop- ment. If it is a single mutation, then it would probably be dom- inant although there is the possibility of a recessive or the ex- treme expression of a homozygote in which the heterozygote is relatively normal as in Wright's (1935) polydactylous guinea PARSONS AND STEIN : ANOMALOUS CAT SKELETON 313 pigs. Any sucli theory postulating homozygosity for a mutation in this cat Avould ahnost certainly require that the father was not the one suspected but a half-brother of the cat being studied, since the chance of the same mutation appearing in two unrelated cats is very small. Another problem which arises in considering possible genetic mechanisms is that any genes involved may show a very low penetrance, as is the case in Danforth's (1925 and 1930a) strain of mice to be discussed below. Thus, for example, the mother and ai^proximately half her offspring could be carry- ing a gene responsible for the abnormalities described in this specimen without their showing any phenotypic effects. Pelvic girdles and appendages Like most described polymelous specimens, this cat exhibits posterior duplication — however, in a much modified form. The commonest pattern of such abnormalities appears to be one in which two approximately equal pelvic girdles are present, pro- ducing a Y-shaped main axis of the animal. However, the spinal column is usually single and much deformed posteriorly. In such cases there is a marked tendency for the reduction and fusion of the inner pelvic elements of the two sides. Cases in point have been described by Conrow (1917) and Mainland (1929) where six legs were clearly present, and by Lesbre (1900) and Horsley (1920) where fusion has reduced the number to five. Reports of cases of asymmetry like the present one are some- what rarer and far less complete. Murie (1934) reports a five- legged rabbit, but very little is told of the nature of the extra leg. Three other papers describe animals in which the doubling is confined to the appendicular skeleton on one side with the other side essentially normal: Bateson's (1894) macaque (his no. 504) shows this doubling extending into the pelvic girdle while in Weimer's (1942) sheep only the limb, and in Dwight's (1892) human, only the forearm and hand are affected. The cat under consideration conforms to neither of these patterns. There is posterior duplication resulting in four pelvic limbs and two girdles. Reduction and fusion characterize the right girdle (the teratological pelvis of the descriptive part) and limbs, while the left side remains essentially normal except for the right innominate, thus forming the more normal pelvic girdle of the cat. It is impossible to tell whether any duplication oc- 314 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY curred in the soft parts. However, the viscera probably did not show much duplication since the right pelvic canal is completely closed by extensive fusion. As mentioned previously, fusion in the limb is a common fea- ture of polymelous specimens, there being great variation in de- gree. Here the femora, fibulae, and feet are very closely fused but the two tibiae are completely separate. The macaque of Bateson (1894) is very similar in this respect except that there the positions are reversed and the fibulae are separate and the tibiae joined together. There is very little known about possible genetic mechanisms causing such posterior duplication. Danforth (1925 and 1930a) has described a mutant strain of mice with this character, but the exact mode of inheritance is not known. By selecting for it, he was able to develop only a line with an incidence of 12 per cent, so it is considered probably recessive. The expression was very variable with extensive fusion normally occurring. The spinal column was also abnormal with the tail generally kinky, and occasional cases of spina bifida. In the case of the cat de- scribed in this paper, there is no evidence that the cause of the duplication is genetic as opposed to environmental. The functional problems arising from the extra leg are of some interest, but little can be definitely said about them. It seems most unlikely that the much deformed double foot could have borne any weight, so that leg was presumably of little use despite the very smooth hip articulation. The more normal right leg, on the other hand, has much irregular and apparently arthritic ossification at the hip resulting in what must have been an almost immovable joint. Despite the presumably compensa- tory freedom at the knee evidenced by the articular surfaces there, this leg too would seem greatly inferior to the normal condition. Thus it seems probable that the cat depended on the three normal limbs, making little use of the two right hind legs. SUMMARY The cat skeleton described in this paper shows an unusual type of posterior duplication in which the left pelvic girdle and limbs are relatively normal and the right ones much modified. The right girdle is a single Y-shaped element fused to the antero- ventral border of the right innominate of the left pelvis, and bears PARSONS AND STEIN : ANOMALOUS CAT SKELETON 315 one limb. This consists of a single element representing two femora, two tibiae, one fibula, and a highly abnormal foot in the flexed position bearing metatarsals 1, 2, 3, 4, 3, 2, and 1. The vertebrae from the tenth thoracic through the sixth caudal are all abnormal showing various degrees of asymmetry and fusion and cases of spinae bifidae and hemivertebrae. Possible genetic causes of such a condition are discussed, but no definite conclu- sions can be reached concerning this case. BIBLIOGRAPHY Backman, G. 1934. Die Abliangigkeit morphologischer Variationeu von Differenzie- rungs- und Waehstunisgradienten. Anat. Anz., vol. 79, pp. 78-87. Bamber, E. C. 1927. Genetics of domestic cats. Bihliog. Genetica, vol. 3, pp. 1-86. Bateson, W. 1894. Materials for the study of variation. London. Baur, G. 1891. On intercalation of witchrae. Jour. Morph., \'ol. 4, pp. 331 33G. Berg, W. 1912. Ueber stummelsehwanzige Katzen und Hunde. £in Beitrag zu der Lehre von der Eeduktion des Schwanzes. Zeitschr. Morph. Anthrop., Sonderheft 2, pp. 227-267. Brash, J. C. 1915. Vertebral column witli six and a half cervical and thirteen true thoracic vertebrae, with associated abnormalities of the cervical spinal cord and nerves. Jour. Anat. Physiol., vol. 49, pp. 243-273. Cheslet, p. 1935. Development of the short-tailed mutant in the house mouse. Jour. Exp. Zool., vol. 70, pp. 429-459. CONROW, S. B. 1917. A six-legged rat. Anat. Rec, vol. 12, pp. 365-370. Danforth, C. U. 1925. Hereditary doubling suggesting anomalous chromatin distribu tion in the mouse. Proc. Soc. Exp. Biol. Med., vol. 23, pp. 145 147. lySOa. Developmental anomalies in a special strain of mice. Anier. Jour. Anat., vol. 45, pp. 275-287. yl6 BULLETIN : MUSEHJM OF COMPARATIVE ZOOLOGY 1930b. Numerical variation and homologies in vertebrae. Amer. Jour. Phys. Anthrop., vol. 14, pp. 463-481. DWIGHT, T. 1892. Fusion of hands. Mem. Boston Soc. Nat. Hist., vol. 4, pp. 473- 486. 1901. Description of the human spines showing numerical variation in the Warren Museum of the Harvard Medical School. Mem. Bos- ton Soc. Nat. Hist., vol. 5, pp. 237-312. Gluecksohn-Schoenheimer, S. 1943. The morphological manifestations of a dominant mutation in mice affecting tail and urogenital system. Genetics, vol. 28, pp. 341-348. Gottlieb, H. 1914. Die Antikliuie der Wirbelsiiule der Saugetiere. Morph. Jahrb., vol. 49, pp. 179-220. Gruneberg, H. 1947. Animal genetics and medicine. London. 1952. The genetics of the mouse. 2nd ed. The Hague. HORSLEY, J. S., Jr. 1920. A description of a six-legged dog. Anat. Rec, vol. 19, pp. 1-27. lIi'NT, H. R., R. MixTER, and D. Permar 1933. Flexed tail in the mouse, Mus musculii^s. Genetics, vol. 18, pp. 335-366. Jayne, H. 1898. Mammalian anatomy. Part I. The skeleton of the cat. Phila- delphia. Kermauner, F. 1909. Missbildungen des Rumpfes (Fortsetzung). In Schwalbe, E., Die Morphologie der Missbildungen des Menschen und der Tiere. Jena. vol. 3, Lief. 1, pp. 86-112. Klatt, B. 1939. Erbliche Missbildungen der Wirbelsiiule beim Hund. Zool. Anz., vol. 128, pp. 225-235. Lesbre, F. X. 1900. Observations de deux vaches et d'une cane pygomeles, avec con- siderations generales sur la pygomelie. Jour. yied. Vet., ser. 5, vol. 4, pp. 625-652. PARSONS AND STEIN! ANOMALOUS CAT SKELETON '517 Mainland, D. 1929. Posterior duplicity in a dog, with reference to mammalian tera- tology in general. Jour. Anat., vol. 63, pp. 473-495. MURIE, A. 1934. A five legged rabbit. Jour. Mammal., vol. 15, p. 162. Ratcliffe, H. L., and H. D. King 1941. Developmental abnormalities and spontaneous diseases found in rats of the mutant strain, stubs. Anat. Rec, vol. 81, pp. 283-305. SCHLEGEL, M. 1921. Die Missbildungen der Tiere. Ergebn. AUgem. Path. u. Path. Anat., vol. 19, pp. 650-732. Schwalbe, E. 1906-1913. Die Morphologie der Missbildungen des Menschen und der Tiere. 3 vols. Jena. Stockakd, C. E., et al. 1941. The genetic and endoerinic basis for differences in form and behavior as elucidated by studies of contrasted pure-line dog breeds and their hybrids. Amer. Anat. Mem., no. 19. Stroer, \V. F. H. 1937. Die Extremitatenmissbildungen und ihre Beziehungen zum Bauplan der Extremitiit. Zeitschr. Anat. Entw.-Gesch., vol. 103. pp. 136-160. Weimer, B. E. 1942. A case of hind-limb duplication in the sheep. Proc. West Virginia Acad. Sci., vol. 15, pp. 82-84. Wright, S. 1935. A mutation of the guinea pig, tending to restore the peutadactyl foot when heterozygous, producing a monstrosity when homo zygous. Genetics, vol. 20, pp. 84-107. PLATE 1 Plate 1. Mounted skeleton of the eat. The greater part of the tail, the ribs, sternum, phalanges, clavicles, patellae, and hyoid apparatus are not included. PLATE ■:? I'LATK 1^. X-i-,iy of the (losteiidr p;nt of tlu' cat sliow'.iig tlic regular right anil teratological h'lul linilis in lato'al view. x%. ii PLATE 2 PLATE 3 Plate ?>. A. snapslicit of the cat in life. B and C, teratological foot in ventral (normal dorsal) and dorsal (normal i)lantar) views. 7-5 ••nid (" x4/.T. *•, , I- B PLATE 3 PLATE 4 Plate -i. Teiatological foot in laternl view compared to ii iKirmal left foot in dorsal view. x6/5. PLATE 4 Bulletin of the Museum of Comparative Zoology AT HARVARD C O L L t: G E Vol. 114, No. 7 POST-TRIASSIC NAUTILOID GENERA By Berniiard Kummel With Twenty -eight Plates CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM August, 1956 Publications Issued by or in Connection WITH THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 — The current volume is Vol. 114. Breviora (octavo) 1952 — No. 57 is current. Memoirs (quarto) 1864-1938 — Publication was terminated with Vol. 55. JoHNSONiA (quarto) 1941 — A publication of the Department of Mollusks. Vol. 3, no. 35 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 18 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. The continuing publications are issued at irregular intervals in numbers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts. Of the Peters "Check List of Birds of the World," volumes 1-3 are out of print; volumes 4 and 6 may be obtained from the Harvard University Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices. Bulletin of the Museiun of Comparative Zoology AT HARVARD COLLEGE Vol. 114, No. 7 POST-TRIASSIC NAUTILOID GENERA By Bernhabd Kummel With Twenty-eight Plates CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM August, 1956 No. 7 — Post-Triassic Nautiloid Genera By Bernhard Kummel TABLE OF CONTENTS Page Introdiu'tion 324 Classification of Post-Triassic Nautiloids 325 Evolution 344 Systematic Paleontology 360 Family Xautilidae d 'Orbigny, 1840 3f50 Subfamily Nautilinae d 'Orbigny, 1840 361 Genus Cenoceras Hyatt, 1883 361 Genus Eutrephoceras Hyatt, 1894 377 Genus Pseudocenoccras Spath, 1927 384 Genus Carinonautilus Spengler, 1910 385 Genus Ohinautilus Kobayashi, 1954 386 Genus Nautilus Linne, 1758 387 Subfamily Pseudaganidinae nov 389 Genus Pseudaganides Spath, 1927 390 Genus Pseudonautilus Meek, 187fj 397 Subfamily Paracenoceratinae Spath, 1927 399 Genus Paraccnoceras Spath, 1927 402 Genus Aulaconautilus Spath, 1927 409 Genus Tithonoceras Eetowski, 1894 410 Genus SomalinautiJns Spath, 1927 411 Subfamily Cymatoceratinae Spath, 1927 412 Genus Cymatoceras Hyatt, 1884 417 Genus Paracymatoceras Spath, 1927 427 Genus Procymatoceras Spath, 1927 428 Genus Cymatonautilus Spath, 1927 430 Genus Anplonautilus Spath, 1927 430 Genus Eucymatoceras Spath, 1927 431 Genus Syrionautilus Spath, 1927 433 Genus Eeminautilus Spath, 1927 434 Genus Deltocymatocera.s n.gcn 438 Genus Epicymatoceras n.gen 439 Subfamily Hercoglossiuae Spath, 1927 440 Genus Cimomia Conrad, 1866 445 Genus AnguUthes Montfort, 1808 452 Genus Hercoglossa Conrad, 1866 457 Page Genus Aturoidea Vredenburg, 1925 462 Subfamily Aturinae Hyatt, 1894 464 Genus Aturia Bronn, 1838 464 Addendum 468 Bibliography 469 ILLUSTRATIONS PLATES Plate 1. Cenoceras orbignyi 2. Cenoceras trechmanni 3. Cenoceras striatus 4. Cenoceras simillimus 5. Cenoceras inornatus 6. Cenoceras hurtonenais 7. Cenoceras and Syrionautilus 8. Paracenoceras and Cenoceras 9. Ohinautilus and Cenoceras 10. Heminautilus and Pseudocenoceras 11. Carinonautilus ariyalurensis 12. Pseudonaiitilus geinitzi 13. Paracenoceras hexagonnm 14. Aulaconautilus and Somalinautilus 15. Tithonoceras zitteli 16. Cymatoceras pseudoelegans 17. Procymaioceras suhtrwicatus 18. Cymatonautilits julii 19. Paracymatoceras asper 20. Anglonautilus undulatus 21. Eucymatoceras plicatus 22. Deltocymatoceras leiotropis 23. Epicymatoceras vaelsensis 24. Cimomia burtini 25. Angulithes and " Bisiphytes' ' 26. Angulithes triangularis 27. Hercoglossa orhiculata 28. Aturoidea and Pseudaganides FIGURES ^ Figure Page 1. Bar chart of nautiloid genera per period 345 2. Phylogenetie relationships of late Triassic and early Juras- sic nautiloids • • • • 347 3. Phylogeny of post- Triassic nautiloids 348 4. Bar chart of number of genera of post-Triassic nautiloids per series 353 5. Bar chart of number of species of post-Triassic nautiloids per series 357 6. Distribution map of Cenoceras 358 7. Distribution map of Eutrephoceras 359 8. Cenoceras 363 9. Cenoceras 365 10. Cenoceras 367 11. Cenoceras excavatus 369 12. Cenoceras burto?ievsis 371 13. EiitrepJioceras 378 14. Pseudaganides 391 15. Ps.eudaganides and Pseitdonautilu-s 393 16. Pertnoceras and Pseudonautilus 399 17. Distribution map of Paracenoceratinae 401 18. Paracenoceras hexagonum • 403 19. Paracenoceras and Aulaconautilus 404 20. Paracenoceras and Reminautilus 406 21. Distribution map of Lower Cretaceous Cymatoceratinae . . . 414 22. Distribution map of Upper Cretaceous Cymatoceratinae . . 415 23. Cymatoceras, Paracymatoceras, Deltocymatoceras, Procyma- toceras, Epicymatoceras 419 24. Cymatoceras and Paracymatoceras 422 25. Procymatoccras suhtruncatus 429 26. Anglonautilus undidatus 431 27. Eucymatoceras pUcatus 432 28. Reminautilus saxbii 436 29. Distribution map of Cretaceous Hercoglossinae 442 30. Distribution map of Tertiary Hercoglossinae 443 31. Hercoglossa, Cimomia, and Atiimidcs 446 32. Cimomia and Angidithes 448 33. Angidithes 454 34. Hercoglossa and Aturoidea 458 35. Distribution map of Aturinae 465 INTRODUCTION Studies on tetrabranchiate cephalopods have added much data to our understanding of the patterns and modes of evolution as displayed by fossil animals. Few invertebrate phyla show such marked periods of expansion in evolutionary intensity con- trasted with periods of extreme contraction to the point of near extinction. The generalized evolutionary history of the ammon- oids with periods of near extinction at the end of the Paleozoic and again at the end of the Triassic, and their final extinction at the end of the Cretaceous is well known. Among the nautiloid cephalopods a similar evolutionary history exists but with some important differences. The nautiloid cephalopods had their greatest radiation soon after their first appearance in the Upper Cambrian, that is in the Ordovician. In this early phase of their history numerous distinct phyletic trends are well established. However, by Silurian time there is already a marked decline in the number of genetic groups. This decline in the viability of the nautiloids continued, until today we have only a single genus (Nautilus) living as a relic form in the southwest Pacific. In this long, gradually diminishing diversity the nautiloids nearly became completely extinct in the late Triassic, a period of extinction coinciding with a similar phase in the evolution of the ammon- oids. A single stock derived from the Triassic Syringonautilidae survived into the Jurassic. From this surviving stock all post- Triassic nautiloids are directly or indirectly derived. The purpose of this study is to analyze the evolutionary his- tory of the family Nautilidae which includes the surviving stock from the Triassic and all post-Triassic nautiloid genera. This phyletic group is of particular interest since it includes an evo- lutionar}^ complex which has survived a period of near extinc- tion of the class, went through a minor eruptive phase in the development of new adaptive types, and is now represented by a single genus. No explanation is given as to the cause of the KUMMEL : POST-TRIASSIC NAUTILOID GENERA 325 decreasingr viability of the nautiloids, but the patterns anrl modes of evolution of this declining group are instructive. Post-Triassic nautiloids have received but scant attention, most workers having merely referred all species to the genus Nautilus. In 1927, L. F. Spath presented a very comprehensive, but short, review of post-Triassic nautiloids and proposed a taxonomic scheme including many new genera. In 1951, I spent eight months at the British Museum of Natural History, and at that time had the opportunity of studying the very large and excellent collections of that institution including the types of many of the genera proposed by Spath. At the time Spath (1927a) proposed his taxonomic scheme for post-Triassic nau- tiloids he did not give any diagnosis of his genera or illustrations of his type species, many based on Sowerby species not previously illustrated except in woodcuts. In the present report all genera of post-Triassic nautiloids are diagnosed and illustrated. An intensive search has been made of the literature to bring together, as far as possible, all species names that have been proposed over the past hundred odd years. In this way data was obtained on the variability, geologic range and geographic distribution of each generic group. This study is an outgrowth of the preparation of a chapter on Mesozoic nautiloids for the Treatise on Invertebrate Paleon- tology, edited by R. C. Moore. A similar study on Carboniferous to Triassic nautiloids has already been published (Kummel, 1953e). I wish to express my very sincere appreciation to Dr. L. F. Spath for many long hours of stimulating conversations and for facilitating in every way my work at the British Mu- seum. Fruitful discussions with Dr. Curt Teichert and Dr. Ernest Williams on taxonomic and evolutionary problems were most helpful. CLASSIFICATION OF POST-TRIASSIC NAUTILOIDS The only previous comprehensive discussion of post-Triassic nautiloids is that of L. F. Spath (1927a). In recent years several large works on Tertiary nautiloids have been produced by Miller which have included descriptions of many new Tertiary faunas. Spath 's discussion of post-Triassic nautiloids was only a small and incidental part of his brilliant work on the famous Jurassic 326 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY fauna of Kachh (Cutch). Even so, he laid the framework for future work on these animals. Nautiloids of Mesozoic age and, until recently, Tertiary age have received only brief incidental treatment in paleontologieal publications. For one thing post-Triassic nautiloids are not common fossils compared to the contemporaneous ammonoid, pelecypod, or gastropod faunas. They likewise are seemingly not nearly so diverse. The evolutionary decline of the nautiloids, already very marked in the Silurian, progresses steadily towards the Recent. Thus in studying Mesozoic and Tertiary nautiloids we are closely examining the decline and near extinction of a large animal group. Until the contribution by Spath (1927a) there had been few genera proposed for post-Triassic nautiloids and most of these were not generally accepted or used. The most common practice was to place all post-Triassic species in the genus Nautilus. Those genera that had been proposed before Spath 's 1927 re- vision include Nautilus Linne, 1758, Bisiphytes Montfort, 1808, Cenoceras Hyatt, 1883, Eutrephoceras Hyatt, 1894, Digonioceras Hyatt, 1894, Angulithes Montfort, 1808, Hercoglossa Conrad, 1866, Pseudonautilus Meek, 1876, Cimomia Conrad, 1866, Cyma- toceras Hyatt, 1883, Tithonoceras Retowski, 1893, Carinonautilus Spengler, 1910, and Aturia Bronn, 1838. As mentioned above, most of these genera were never used after their original pro- posal ; however, such generic names as Eutrephoceras, Herco- glossa, Cymatoceras, and Aturia are fairly common in the early literature. Most of these genera were considered as members of the Nautilidae. In his revision of post-Triassic nautiloids Spath (1927a) pro- posed 16 new genera and adopted 5 families, namely the Nautili- dae d'Orbigny, 1840, the Hercoglossidae Spath, 1927, the Cyma- toceratidae Spath, 1927, the Paracenoceratidae Spath, 1927, and the Aturidae Hyatt, 1894. On a straight morphological basis, the Nautilidae include those genera directly or indirectly in the main stock giving rise to Nautilus. The Hercoglossidae include those genera with '"goniatitic'* or very sinuous sutures. The Cymatoceratidae include those stocks characterized by ribbing — the only stock of po.st-Triassic nautiloids with anj' ornamen- tation. The Paracenoceratidae are characterized by differentia- tion of the periphery, and the Aturidae are characterized by KUMMEL : POST-TRIASSIC NAUTILOID GENERA 327 the unique and peculiar dorsal siphuncle. There have been no significant changes made to this taxonomic arrangement since its proposal in 1927. Restudy of the large nautiloid collections in the British Museum and the Museum of Comparative Zoology plus a comprehensive review of the litera- ture have brought forth further data on the range of variation within the various genera, and also new data on the geographic and stratigraphic range of most of the species described. The nautiloids reflect no significant change in their evolution- ary development from the late Paleozoic into the Triassic. In fact, the evolutionary pattern of Triassic nautiloids is merely a culmination of trends begun back in the Carboniferous ( Kum- mel, 1953c). Towards the end of the Triassic most of the long- lived stocks became extinct. A single persisting stock repre- sented by Cenoceras survived the Triassic, and it represents the root form from which, directly or indirectly, all post-Triassic nautiloids are derived. On examination of the available data on post-Triassic nautiloids, one is impressed by the essential homo- geneity of the group. Distinctive radiations are recognizable, each representing an elaboration of particular morphological characters. Homeomorphous developments within the group and with pre-Jurassic genera are common. The time-space relation- ships of the post-Triassic nautiloid "species" and their mor- phological modifications, interpreted in terms of adaptive radia- tion, make possible a constructive phylogenetic interpretation of the group. The interpretation presented in this report gives a much clearer understanding of the various generic groups and the kinds and ranges of variation within the groups. The classifica- tion adopted reflects the phylogeny of post-Triassic to Recent nautiloids. The classification used in this report is as follows: Family Nautilidae d'Orbigny, 1840 Subfamily Nautilinae d 'Orbigny Genus Nmitilus Linne, 1758 Type species: XctidiUis ponipiHns Limn' Genus Cenoceras Hyatt, 1883 Type species: Nautilus orbignyi Prinz Genus Eutrephoceras Hyatt, 1894 Type species: Nautilus delayi Morton Genus Pseudocenoceras Spath, 1927 Type species: Nautilus largilliertianus d'Orbigny 328 , BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Genus Carinonautilus Spengler, 1910 Type species: Carinonautilus ariyalurensis Spengler Genus Obinautilus Kobayashi, 1954 Type species: Obinautilus pulclira Kobayashi Subfamily Pseudaganidinae nov. Genus Pseudaganides Spath, 1927 Type species : Nautilus Tcutchensis Waagen Genus Pseudonautilus Meek, 1876 Type species: Nautilus geinitzi Oppel Subfamily Paracenoceratinae Spath, 1927 Genus Paracenoceras Spath, 1927 Type species: Nautilus hexagonus J, de C. Sowerby Genus Aulaconautilus Spath, 1927 Type species: Tautilus sexcarinatus Pictet Genus Somalinautilus Spath, 1927 Type species: Nautilus antiquum Daeque Genus Tithonoceras Retowski, 1894 Type species : Tithonoceras zitteli Retowski Subfamily Cymatoceratinae Spath, 1927 Genus Cymatoceras Ilyatt, 1883 Type species: Nautilus pseudoelegans d'Orbigny Genus Procymatoceras Spath, 1927 Type species: Nautilus subtruncatus Morris and Lycett Genus Cymatonautilus Spath, 1927 Type species: Nautilus julii d'Orbigny Genus Paracymatoceras Spath, 1927 Type species : Nautilus asper Oppel Genus Syrionautilus Spath, 1927 Type species: Nautilus libanoticus Foord and Crick Genus Anglonautilus Spath, 1927 Type species : Nautilus undulatus J. Sowerby Genus Eucymatoceras Spath, 1927 Type species: Nautilus plicatus Fitton Genus Heminautilus Spath, 1927 Type species : Nautilus saxbii Morris Genus D.eltocymatoceras n.gen. Type species: Nautilus leiotropis Schliiter Genus Epicymatoceras n.gen. Type species: Nautilus vaelsensis Binckhorst Subfamily Hercoglossinae Spath, 1927 Genus Hercoglossa Conrad, 1866 Type species : Nautilus orbiculatus Tuomey Genus Cimomia Conrad, 1866 Type species: Nautilus burtini Galeotti KUMMEL : POST-TRIASSIC NAUTILOID GENERA 329 Genus Angulithes Montfort, 1808 Type species: Nautilites triangularis Montfort Genus Aturoidea Tredenburg, 1925 Type species: Nautilus parl-in,'nides .V. ahltcnensis Schliiter, 1876 — Eutr.ephoceras y. alabamensis Morton, 1834 — Aturia 332 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Aturia alaskensis Schenck, 1931 — Aturia N. alhensi.s d 'Orbigny, 1850 — • Cijmatoceras Eutrephoccras nJcescnse Eeeside, 1927 — Eutrephoceras N. allani Fleming, 1945 — Eutrephoceras N. allioni Miehelotti, 1840 — Eutrephoceras K. altavensis Pomel, 1889 — Cymatoceras N. altifrovti Chapman, 1915 — - Eutrephoceras N. altisiphites Prinz, 1906 — Cenoc.eras N. amasianus Gugenberger, 1928 — Cenoceras N. ammoni Loesch, 1912 — Pseudaganides N. amorettii Parona, 1897 — Cenoceras Cymatoceras andranofotsyense CoUignon, 1951 — Cymatoceras N. anguliferous Schliiter, 1876 — Cymatoceras N. angustata Conrad, 1849 {in Dana) — Aturia N. angustus Blanford, 1861 — Cimomia X. anomplialus Pia, 1914 — Cenoceras X. (iiifiqifus Daeque, 1910 — Somalinautilus X. (ipplanal i(s Wanner, 1902 — Fseudocenoceras X . arariforin'm Pia, 1914 — Cenoceras X . araris Dunioitier, 1869 — Cenoceras X. aratus Quenstedt, 1846 — Cenoceras X. o 0 o 11 ;o a: 2 r^ 1 0 ■^— O c o M ' ' 12 1 1 "i V"\ o 0 z 0 CO CO < c .0 'c 0 0 c i — c 0 "c < ^4 0 m 0 D 0 ■0 0 <3J D 1 C 0 / 0: 0 0 a. c 0 cr c >• Cenoceras trechmanni 0) 0 0 :3 0 c e / c ^%^ / 0 ^^^r^'^'. / - !c / *^-. / >t ^ ^ •:_ f-„'- Fig. '2. Phylogenetic diagram illustrating the relationships of the Liassio Xautilidae with the Triassie nautiloid families. The columns representing the Triassie families have been terminated at an arbitrary even line in the Xorian. The exact relative extinction dates of these families in the Xorian are not known. specialized oxycones" (See Holzapfel, 1899). A similar situa- tion is found with the early Triassie ophiceratids of the Hima- layas (Diener, 1897) and Greenland (Spath, 1930a, 1935a). More examples can he found within the Jurassic and Cretaceous am- monoids. 348 BULLETIX : ML7SEUM OF COMPARATIVE ZOOLOGY Recent ^^ A'^^ ££. >- DC < \- (T. H "-»« "^z,;. (9 •O ^. \ ^b-^"- jF 3 % CO O UJ O ? UJ 1 \ \ o E O O O "^ ^ ^ ^ ^ ^«„ >>' x^ 3 O c o 3 0) Q. o. 3 < q: 3 -3 ?S*^ c^^ ^*^ ^ Cenoceras ipv .e^ ,i'^<* A^' Fig. 3. Diagram showing inferred phylogeny and geologic distribution of post-Triassic nautiloids. KUMMEL : POST-TRIASSIC NAUTILOLD GENERA 349 Most early Jurassic nautiloids are here considered species of a single genus, Cenoceras. The earliest species of Cenoceras. C. trechmanni. is from Carnian strata of Nbav Zealand and is derived from the Syringonautilidae (Kiimmel, 1953b). Xo species of any other Triassic stock survived into the Jurassic. In the Lias there is thus a Avorld-wide fauna of involute to evolute, strigate to smooth forms with a wide range in whorl shapes — a genetic com])lex of great plasticity. From this homo- geneous but plastic stock there arose from diverse parts of the complex several distinct evolutionary lines built on specializa- tion of one or more morphological features. From this Cenoceras complex arose the persisting stable stock (Eutrephoceras) which gave rise to other members of the Nautilinae. Likewise the Pseudaganidinae, Paracenoceratinae and most probably the Cymatoceratinae arose directly out of the Cenoceras complex (Fig. 3). The Pseudaganidinae are characterized bj' their sinuous su- tures. In this respect they are adaptive types like the Triassic Clydonautilidae, Gonionautilidae, and Siberionautilidae. They are likewise similar to these Triassic groups in that they show stability in the general form and patterns of the suture and great varialulity in conch shape. The Pseudaganidinae range from the Liassic into the Lower Cretaceous. There are only two genera involved in this radiation, of which Pseudaganides is the main line and PsciidonauHliis is a specialized offshoot of Upper Jurassic and Lower Cretaceous age. Tn the evolutionary history of the uautiloids from the Carboniferous to the Recent, adaptive trends expressed in terms of sinuous — ''goniatitic" — sutures occurred at four separate times. The first is known through the genus Permoccras Miller and Collinson based on Aganides hitauniensc Haniel from Middle Permian strata of Timor. In luy opinion. Pcrmoceras is related to the Grypoceras-Doma- toceras evolving .stock and is an aberrant development of this •Stock. Grgpoceras iPlummeroceras) Kummel (1953c) likewise had a highly sinuous suture with a deep ventral lobe and a deep lateral lobe but it has a very evolute conch. The second radiation repeating this adaptive trend is tliat in the Upper Triassic in- cluding the Clydonautilidae, Gonionautilidae and Siberionautili- dae. All of the genera of these families include species with very involute conchs but show great variability in conch shape. The 350 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY only other Triassic species with a "'gouiatitic'' suture is Cly- menonautUus ehrlichi Mojsisovics which in all features but the suture is allied to the Syringonautilidae. The suture however has a deep, tongue-shaped, narrow-, lateral lobe. The Upper Triassic families listed above which include those involute species with highly sinuous sutures, evolved from involute, smooth forms with nearly straight sutures (Paranautilidae). The third radiation of this series is that of the Pseudaganidi- nae of Lower Jurassic to Lower Cretaceous age. This group evolved out of some part of the Cenoceras complex in the Lias. There is almost perfect homeomorphy between Permoceras hitaimiense Ilaniel of the Middle Permian and Pseudonaidilus geinitzi Oppel of the Upper Jurassic (Miller and Collinson, 1953). Each however represents similar adaptive types from different root stocks and they are not related. The fourth and last radiation centered on the sinuosity of the suture is that of the Hercoglossinae and Aturinae. The Aturinae with its single genus Aturia is a very specialized de- velopment out of the Hercoglossinae. The specialization in this case is not so much on a further elaboration of the suture but in the development of the peculiar dorsal siphuncle. Within the Hercoglossinae there is a beautiful developmental series (usually treated as genera) showing gradual increase in sutural complexity. The subfamily has its origin within the stable evolving stoclc of the Xautilinae represented by Eutrephoceras which is characterized by an involute smootli conch with straight or slightly sinuous sutures. Cimoinia of the Hercoglossinae has a slightly more individualized suture and is gradational with En- trephoceras, as it is also with Hercoglossa which has very distinct lobes and saddles. Angulithes is merely a sagittate dei-elopment of these forms. Jn Aturoidea the sutural individuality is carried slightly farther. The Paracenoceratinae is a small subfamily of mainly Jurassic age with four genera characterized by specialization of the venter. The main stock of this subfamily, Paracenoceras, is nearly world- wide in distribution and has by far the largest number of species. The other genera, namely SomaUnautilus, Aulaconaufilus, and Tithonoceras are more highly specialized forms with few known species. Tn Paracenoceras the whorl section is subtrapezoid, generally with a broad, sulcate venter. SomaUnautilus has angu- KUMMEL : POST-TRIASSIC NAUTILOID GENERA 351 lar ventral shoulders but a broad arched venter. Aulaconautihis has four or more longitudinal ribs. In this respect it is homeo- morphous to Aulametacoceras of the Permian and Triassic. Tithonoceras has prominent rounded ventrolateral keels with a sulcate venter. Paracenoceras and Somalinautilus are undoubt- edly derived from distinct elements of the Cenoceras complex but Aulaconautilus and Tithonoceras appear to be specialized developments of Paracenoceras. One of the must successful and diverse groups to stem directly or indirectly from the Cenoceras complex is the Cymatoceratinae. This subfamily comprising 10 genera ranging from the Juras- sic to the mid-Tertiary is characterized by conchs bearing ribs. This is the only group of post-Triassie nautiloids to have orna- mentation, aside from Aulaconautilus of the Paracenoceratinae. Radiation within the Cymatoceratinae is reflected in the shape of the concn, suture and ornamentation. The main evolving stock, Cinnatoccras, has an involute, rounded conch with only a slightly sinuous suture. ]\Iost of the other genera of this sub- family are thought to represent various specialized groups de- rived from Cymatoceras. Paracyniatoceras has a more sinuous suture, in fact much like that of Hercoglossa, and on even this feature it is gradational with Ci/niatoceras. Those forms dif- ferentiated on the basis of conch shape include Honinautilus, Deltocymatoceras (n. gen. p. 438), Epicymatoceras (n. gen. p. 439) and Cymatonautilus. Herminautilus has a compressed in- volute conch with a highly sinuous suture, that has deep ventral and lateral lobes. Deltocymatoceras is a homeomorph of Afigu- lithes with a sagittate whorl section. Epicymatoceras has an evolute, highly compressed conch with a subrectangular whorl section. Cymatonautilus has a concave venter and concave lat- eral areas. Those genera differentiated on the basis of modifica- tion of the ribbing pattern include Eucymatoceras, Anglonau- tilus, Procymatoceras, and Syrionautilus. In Eucymatoceras the ribs form prominent V-shaped salients on the venter and flanks. In Ancjlonautilus there are coarse folds on the venter whereas Procymatoceras has a rapidly expanding robust conch with ribs mainly on the lateral areas. Syrionautilus has peculiar .shaped ribs with wide interspaces like that of Proclyclonautilus spirolobus of the Triassic. Cymatonautilus and Procymatoceras are Jurassic developments, the remaining genera mostly Cre- 352 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY taceous. In fact the most widespread and characteristic nautil- oids of the Cretaceous are genera of the Cymatoceratinae. Thinking of tlie Cymatoceratinae as representing a single genetic unit with its origin within the Cenoceras complex, and having as the basic common denominator the ribbing pattern, all of the diverse morphologic types represented by the various genera reflect a broad adaptive radiation which produced numer- ous homeomorphs of other genera of the Nautilidae. The most persistent stock evolved from the early Jurassic Cenoceras complex is that of Eutreplioceras. This genus has a world-wide distribution and ranges in time from the Upper Jurassic to mid-Tertiary. It and Cymafoceras are the most com- mon post-Triassic nautiloids. Some 90 so-called species, assigned to this genus, have been described. The genotype has a tightly involute, subglobular, smooth conch with a nearly straight suture. This basic, very simplified, conch pattern became modi- fied either by greater compression, depression or by slightly looser coiling. The variants, however, on the basis of the record available appear to be completelj" random in their chronologic or geographic distribution. Each such variant has usually been treated as a distinct species. The geologic record thus shows the occurrence of nautiloids with the basic, simplified, globose conch persisting from Upper Jurassic to mid-Tertiary time and with numerous variants of this basic pattern. Whereas the suture is generally straight or nearly so there are gradational forms to Ciniomia. Data are not available as to whether there is any particular chronologic or geographic rela- tionship in regard to the sutural variants,' which appear to be completely random as is the case with conch form. The eutrephoceratids are thus a persisting generalized stock which arose from the Cenoceras complex, and they show no par- ticular adaptive trends throughout their history. They do show, however, a certain variability in conch shape and suture. The group has the longest range of any of the post-Triassic nau- tiloids. It is from this persisting generalized form that the re- maining more specialized genera and subfamilies arose. The most important offshoots are the genus Nautilus and the whole subfamily Hercoglossinae. The nature of the evolutionary pat- tern from Enfrephoceras to Cimomia is not easy to decipher. The only significant difference is the degree of sinuosity of the KUMMEL : POST-TRIASSIC NAUTILOID GENERA 353 suture, that in Eutrephoceras being straight, or nearly so, and that in Cimomia having a distinct lobe and saddle on the lateral areas. There are gradational forms between these two genera. The time range of the two genera is approximately the same, Cimomia being slightly younger in origin. Tt seems quite clear that Cimomia arose from Eutrephoceras but whether or not there was a single time and point of origin in the late Jurassic, or if -13- -12- « 1 -||- ""^ -10- -9- -8- -7- ill — b— -5- IpPplHiHp ill ill-iiiliiliii -4- ■ /" -3- _ o ■. i ■'■ -i — t :r':\ "■ ^l - ■" 1 ■ . . 1 . - "■■•■■•L-l-ilil !;:i:k;:::::: ^ -■■—■ ::,;;-i:i::i;j;:-::i::i!iiiil L. Jur. M. Jur. U. Jur. L. Cret. U. Cret. Paleo Eo. Oligo. Mio. Plio. Ptest R. Fig. 4. Bar chart showing total number of genera of nautiloids present in each series of the Mesozoic and Tertiary (light stippling) and number of new genera appearing for the first time in each series (dense stippling). 354 BULLETIN : JIUSEUM OF COMPARATIVE ZOOLOGY there were numerous times of origin, throughout the Cretaceous and Tertiary cannot be told from the kind of data available. The alternatives are a parallel evolutionary pattern of the two genera or a consideration of these transitional forms of suture (Cimomia) as a form genus including numerous distinct radia- tions from the eutrephoceratid evolving stock. Iterative patterns of evolution are recognized within the Car- boniferous through Triassie nautiloids. With these nautiloids three main, slowly-evolving, persisting stocks are recognized, nameh- the Domafoceras-Grypoceras line, the Metacoceras-Mojs- varoceras line, and the Liroceras-Paranautilus line. Each of these stocks has a relatively large number of species (which is merely an indication of relative abundance) and is widespread geographically. Throughout their history they show only rela- tively minor evolutionary changes. Each of these lines repre- sents a stable, slowly-evolving, l)ut persisting parent stock of their particular family groups. These lines are the evolutionary reservoirs from which other, generally more specialized, groups evolved. The offshoots most often are adaptations accentuating one or more generalized characters of the parent stock. Most of the offshoots are forms with few species and limited geographic range. However, some of these offshoots became in themselves virile, evolving stocks, each with a Avide adaptive range and many species (e.g. Pleuronautilidae from the Metacoceras-Mojs- varoceras line and the Clydouautilidae from the Liroceras-Para- nautilus line ) . The E ntrephoceras stock gave rise to four other small and aberrant groups, Pseudocenoceras, Ohinautiliis, Carinonautilus, and Nautilus. Pseudocenoceras is a compressed form somewhat evolute and with an arched venter, flattened sides and steep umbilical shoulders. Carinonautilus is a much compressed form with a prominent rounded keel. It is a monotypic form from Upper Cretaceous strata of South India. Ohinautilus is another monotypic form recently described l)y Kobayashi from Oligocene strata of Japan. It is a much compressed involute form with a shallow but distinct furrow on the venter. This genus is still very incompletely known since on the type and only available specimen neither the suture nor position of the siphuncle is preserved. No fossil species are assigned to the genus Nautilus. No Plio- KUMMEL : POST-TRIASSIC NAUTILOID GENERA 355 oene or Pleistoeenp nautiloids are kno-\vn. Several species of Eocene, Oligoeene, and Miocene age have been at times assigned to Nouiiliis but all of these are here considered as having closer affinities to the persisting eutrephoceratid stock than to modern species of Nautilus. TABLE 1 ^ ~ ;~ ?. T c a rt ?- ^ C ^ :ij £i Cenoceias x x Eutrephoceras x x x x x x x Pseudofonocc'.a : x x Carinonautilus x Obinantihis x Xaiitihis X Pseudfi^anides x x x I'soiidonautilus x x Paracenoeeias x x x Aulaeouautilus x Tithonoceras x Somalinautilus • x x Cyniatoi-eras x x x Paraeyinatoeeras x x x Pioeymatoceras x x Cyniatonautilus x x Anglonautilus x x Eueyniatoeeras x Syrioiiaiitilus x Heminautilus x Deltocymatoceras x Epieymatoceras x Cimomia x x x x x x Angulithes x x x x x Hercoglossa x x x x Aturoidea x x x Aturia x x x x 356 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The geologic range of the geuei-a of post-Triassic nautiloids is summarized on Table 1, and the bar chart of Figure 4 shows the number of genera and number of new genera per series division of the Jurassic, Cretaceous, and Tertiary. This bar chart clearly brings out the increasing adaptive diversity of the nautiloids in their resurgence after their near extinction in the late Triassic. There is a gradual increase in numbers of new types through the Jurassic. However, of the total of 11 genera in the Upper Juras- sic, only 3 {Eutrephoceras, Paracymatoeeras and Cimomia) are still destined for a long history. More than half of the total Upper Jurassic fauna (6 genera) do not survive into the Cretaceous and 2 genera {PscudonautUus and Paracenoccras) have their main evolutionary plaj^ in the Upper Jurassic with only a few species known from Lower Cretaceous. Thus the Jurassic faunas are replaced in the Cretaceous by new phyletic lines, mainly of the C^^llatoceratinae, with a much smaller portion from the Nautilinae and Hercoglossinae. By the early Tertiary the eyma- toceratid radiation is almost at an end, with a single surviving species in the middle Tertiary. Only one new genetic unit, the Aturinae, comes into the picture in the Paleocene ; this became by mid-Tertiary time the most conspicuous element of the fauna. Early Tertiary time was the heyday of the Hercoglossinae ; how- ever, the origin of all members of this radiation is well down in the Cretaceous. Wherever early Tertiary faunas are found, members of the Hercoglossinae are the principal elements; the great number of species of this subfamily for the early Tertiary merely reflects the large amount of study concentrated on them in recent years, especially by Miller. Thus from this summary picture, post-Triassic nautiloid evo- lution took place in three successive waves, one in the Jurassic, another in the Cretaceous, and the last in the early Tertiary. Each wave introduced new genetic lines which gradually re- placed existing or previous lines. The Jurassic and Cretaceous radiations are of approximately equal magnitude ; that of the early Tertiary is greatly reduced. Whether or not Nauiilus should be considered an additional radiation is a moot question. Of the derivations from the eutrephoceratid surviving stock only the Hercoglossinae had an extensive evolutionary history. The other offshoots including Nautilus are either monotypic or repre- sented b.y few species of no great geographic or stratigraphic KUMMEL : POST-TRIASSIC NAUTILOID GENERA 357 range. It seems higlily likely that Xautihis will follow this same fate. In most of the principal evolutionary trends of the jiost-Trias- sic nautiloids the main evolving stock of each snhfamily is i-ejn-e- -100- -90- -80- -70- -60- -50- -40- -30- -20- -10- L. Jur. M. Jur. U. Jur. L. Gret. U. Gret. Paleo, Eo. Cligo. Mio. Plio. Pleist. Fig. 5. Bar chart showing total munl)er of species of nautiloids for each series of Mesozoic and Tertiary. sented by the largest number of species, with the other genera having only a fraction of that amount. Thus of the Xautilinae, Cennceras and Enircphocevas have by far the largest numl)er of species. In the Paracenoceratinae, it is Paracenoccras; in the 856 BL7LLETIX : MUSEUM OF COMPARATIVE ZOOLOGY KUMMEL : POST-TRIASSIC NAUTILOID GENERA 359 o II !~ tn li ci; «1 o O ?a. O J-. (/, ■W cr, tx h- 3 11 A crt o flfl ■^^ — fii o H r ^^ j; ■4-1 « _r^ o « w; r^ ^ tt zj n • r— o o •— o »J -^^ T r-i Oi o ^T* ^ >— • O o .i -*— [/) ^ aj — ^ zt W II M 3 ■Jl H; '" O a & r^. ►— ' ^_ :j; 5i- — ix ^ P r*' 360 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Cymatoceratinae, it is Cyinatoceras; in the Pseudaganidinae it is Pscudaganides, and in the Hereoglossinae it is Cimomia that has the greatest number of species. The Aturinae has only the genus Aturia. \\\ spite of the unsatisfactory nature of our understand- ing of most nautiloid species, these named units do express morphologic diversity, geographic range, and general abundance in nautiloid faunas of these periods. These data likewise tend to substantiate the thesis presented here for persisting, evolving stocks of wide geographic distribution from which more re- stricted adapted types evolved. On Figure 5 are summarized the nninl)er of species per series of the Jurassic, Cretaceous, and Tertiary. Except lor the large number of early Tertiary species, interpreted here as a monographic high, this bar chart reflects veiy well tlie liar chart on di .ti-iI)ution of genera during this time span. syste:^[atic paleoxtology Family NAFTILTDAE d'Orbigny, 1840 This family is interprete;! here, to include all post-Triassic nautiloid genera, which are placed in six subfamilies: Nautilinae (I'Orbigny, Pseudagauidinae nov., Paracenoceratinae Spath. Cymatoceratinae Siiath. Hereoglossinae Spath, and Aturi- nae Hyatt. The overall evolutionary pattern of post-Trias- sic nautiloids with the plastic Cenoveras complex of the early Jurassic from which developed directly or indirectly several distinct jihyletic trends produced a very homogeneous evo- lutionary unit. Tlie subfamily units are adaptive trends inter- preted for the most jjart on the basis of single characters — in the Cymatoceratinae it is the presence of ribbing; in the Para- cenoceratinae it is elaboration of the periphery ; in the Hereo- glossinae and Pseudagauidinae it is elaboration of the suture; and in the x\turinae it is the unique dorsal siphuncle and its structure. None of these trends is really a major shift in the evolutionary complex but each is more in the nature of elabora- tion of genetic potentials in the evolving parent stocks. It thus seems that the taxonomic ranks used here reflect the phylo- genetic picture. KUMMEL : P08T-TRIASSIC XAUTILOID GENERA 361 Subfamily NAUTTLIXAE d'Orbigny, 1840 The Nautilinae includes among its genera the earliest forms and the latest and livinu- species of the Nautilidae. Thus in the history of this subfamily we have the origin of the Xautilidae. the persisting stock which gave rise to the remaining subfamily units, and the few living species represented by the relic genus Nautilus. The genera included in the Nautilinae are: Nautilus, Cenoceras, Etitrcphoceras, Pseudocenoceras, Carino nautilus, and Obinautilus. Cenoceras is the plastic evolving complex which survived the great period of extinction at the end of the Triassic and experienced a very intense radiation in the early Jurassic. From the Cenoceras complex arose the Pseudaganidinae, Para- eenoceratinae, and probably the Cymatoeeratinae. Eutrepho- ceras is interpreted as a slowly evolving, long persisting, general- ized stock also derived from the Cenoceras com])lex and from which the remaining members of the Nautilinae were derived ])lus the Ilercoglossinae. Carino nautilus and Ohinautilus are monotypic ; Pseudoceno- ceras has 8 species, and Nautilus only 5 species (all Recent). However, Cenoceras has 97 species and Eutrephoceras 90. This great number of species (or named units) reflects the wide geo- graphic range, morphologic diversity and relative abundance of these two genera. Cenoceras is mainly confined to the Lower and Middle Jurassic (with one Upper Triassic [Caruian] species), and Eutrephoceras ranges from the Upper Jurassic into the Miocene. Roth o-enera have world-wide distributions (Figs. 6, 7). Genus CenOCEEAS Hyatt. 1883 Cenoceras Hyatt, 1883, pp. 300 301. Cenoceras Hyatt, 189-1, p. aoO. Digonioceras Hyatt, 1894, p]'. 548.549. Nautilites Priiiz, 1906, p. 201. Cenoceras Spatli, 1927a. pp. 20 24. OiJhionautihis Spatli, 1927a, pp. 21, 24. Sphaeronautilus Spath, 1927a, pp. 21, 24. Cenoceras Flower and Kiimmel, 1950, p. 615. Bisiphytes (Cenoceras) Kummel, 1954, p. 322. T]ipe species. Xautihis intermedins d'Orbigny non Sowerbyr=X. nrbignyi Prinz (1906, p. 213). By original designation. Type illustration repro- duced on Plate 1, tigs. 1, 2. 362 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The genus Cenoceras is here interpreted to include those nau- tiloid species of the Liassic and Inferior Oolite that are part of the evolving complex which survived the Triassic and which in the great j)lasticity of the gi'oup reflect an extensive adaptive radiation, it is from this very plastic adaptive unit that the more stable, more clearly defined evolutionary lines evolved. The nautiloids nearly became extinct at the end of the Triassic. A single surviving line derived from the Syringonautilidae gave rise to the sole surviving stock (Spath, 1927a, p. 28; Kummel, l!)5;3b). This surviving stock entered an environmental vacuum (insofar as nautiloids are concerned) in the earliest Jurassic. The re-occupation of the diverse ecological niches available for a homogeneous unit stock gave rise to many morphological types, especially in regard to conch form. It is not possible at this time to decipher the individual evolutionary lines within the mam complex. There is a complete range of variation and gradation in couch form, suture, position of siphuncle, ornamentation, etc. To bring the taxonomy of these nautiloids to reflect the phylo- genetic interpretation, all of the genera previously proposed for aberrant types in the Liassic are placed in synonymy of Ceno- ceras. There is no particular reason for retaining these genera since their included species fit well within the range of variabil- ity and evolutionary pattern of the genus Cenoceras. Fig. 8. Cross sections of the couch of species of Cenoceras. Eedrawn from Pia (1914, pi. 8 [5]). A, C. arari.s (Duniortier) after Dumortier, diameter 210 mm.; B, C. arariformi.'i (Pia) after Pia, diameter 190 ram.; C, C. adneticu.H (Pia) after Pia, diameter 137 mm.; J), C. dumeyhns (d'Orbigny) after d'Orbigny, diameter 150 mm.; E, C. stoppani (Parona) after Parona, diameter 142 mm.; F, C. haoonicua (Vadtisz) after Vadasz, diameter 50 mm.; G, C. altisipJtitcN (Prinz) after Prinz, diameter 60 mm.; n, C. Julianas (Fueini) after Fueini, diameter 20 mm.; /, C. siiiuHimu.'< (Foord and Crick) after Pia, diameter 107 mm.; J, C. secernendu.^ (Pia) after Pia, diameter 135 mm.; K, C. ornatus (Foord and Crick) var. atanaten.sis (Pia) after Pia, diameter 158 mm.; L, C. parctoi (Gemmellaro) after M. Gemmellaro, diameter 17.5 mm.; M, C. sp. ind. No. 2, after Prinz, diameter 65 mm.; X, C. striatm (Sowerby) after d'Orbigny, diameter 110 mm.; 0, C. striatits (Sowerby) after Hauer, diameter SO mm.; P, C. .ituri (Hauer) after Hauer, diameter 65 mm.; Q, C. semseyi (Prinz) after Prinz, diameter 88 mm.; B, C. rohustus (Foord and Crick) after Foord, diameter 200 mm.; S, C. geyeri (Prinz) after Geyer, diameter 80 mm.; T, C. fourneti (Dumortier) after Dumortier and Pia, diameter 80 mm. KUMMEL: POST-TRIASSIC NAUTILOID GENERA 368 B ^rx X H M N 0 Fiajure 8 .S64 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY N^oineiic'latuiv of early .lurassic uautiloids has an involved and ambignons history. As pointed ont by Spath (1927a, p. 20), Hyatt's genus Ccnoceras (for the group of Nautilus intermedins d'Orbigny non So-vverl)y") has been universally rejected, since Hyatt left in Nautilus s.s. other members of the same group, such as Nautilus striatus. However, Spath in his 1027 revision of post- Triassie nautiioids accepted Cenoceras as a distinct unit on the basis of Hyatt's (1894, p. 550) revised definition limiting Ccno- ceras to forms with a trigonal ananepionic, a subquadragonal metanepionic, and a dor,>al]y suleate nepionic stage. At the same time Spatli readily recognized that the groups of Nautilus stri- atus Sowerby and Nautilus intermedius Sowerby could not be differentiated on the shape of the couch and they "agree in ornamentation, position of the siphuncle, presence of an annular lobe, and chiefiy in the course of the septal suture" (Spath, ]9:27a, p. 20). Now for the group of Nautilus si riatus Sowerby, Spath revived the genus Bsiplnjtes Montfoi't ( 1808). This genus has been dis- cussed to a varied degree by Spath (1927a, pp. 19-24), Miller, Dunlmr, and Condra (19;}3,"pp. 42, 43), Teichert (1940, p. 591) and Kummel (1954, pj). o22-32')). Montfort listed as type of his genus Bisipln/hs ry figures 1, 2, Plate 6, X 1. located. The group of Nautilus hurtonensis includes the most evolute adaptation of the Cenoceras complex but there is within this complex a nearly complete gradational series in degree of involution. 372 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Whereas the group of Nautilus burtonensis includes the most evolute elements of the Cenoceras complex, the most involute form is Nautilus pisanus Fucini (1895, p. 828, pi. 13, fig. 5) from lower Liassic strata of Italy for which Spath (1927a, p. 21) established the genus Sphaeronautilus (Fig. lOf/). In its arcestid-like shape Nautilns pisanus is indeed an extreme aber- rant type. The conch is small and smooth. The suture has only a slight lobe on the ventral and lateral areas. There is not known at present a perfect gradational series from "typical" Ceno- ceras to the extremely involute Nautilus pisanus. However, its stratigraphic position and general morphological characters make it logical to consider Nautilus pisanus as the extreme in- volute development of the Cenoceras complex and the genus Sphaeronautilus is placed in synonymy of Cenoceras. More species have been described for Cenoceras than for any other genus of post-Triassie nautiloids. Of the 96 species, two-, thirds are Lias in age and the remaining third Middle Jurassic in age. Cenoceras is a truly cosmopolitan form being known from nearly all continental areas (Fig. 6). The species belonging to Cenoceras with their age and geographic distribution are listed below. DISTEIBUTION OF SPECIES OF THE GENUS CENOCEBAS Stratigraphic Geographic Species Distribution Distriliution C. (Khu'iicus (Pia) 1914 ('. afflnis (CMiaiJuis aiul DewaUjue) 185li C. altisiphites (Prinz) 190(3 C. amasianus (Gugeiiberger) 1928 C. amorettii (Parona) 1897 C. anompJialus (Pia) 1914 C. arariformis (Pia) 1914 C. araris (Dumortier) 1869 Jurassic Austria (Lias) Jurassic Luxembourg (Lias) Jurassic Hungary (Lias) Jurassic Turkey (Lias) Jurassic Italy (Lias) Jurassic France (Lias) •Turassic Austria (Lias) Jurassic Europe (Lias) KUMMEL : POST-TBIASSIC NAUTILOID GENERA 373 Species Stratigraphic Geographic Distribution Distribution C. aratus (Quenstedt) 1846 C. arthaberi (Gugenberger) 1928 C. aatacoides (Young and Bird') 1828 C. austriacus (Hauer) 1856 C. baconicvf^ (Vadasz) 1911 C. hnlsamoorivellii (Parona) 1897 ('. brudfordcnsis (Crick) 1898 r. brancui (Gemmellaro) 1884 C. hreislacJ.-i (Parona) 1897 C. burto7ieniiis (Foord and Crick) 1890 C. catonis (Gemmellaro) 1886 C. clvlensis (Huppe) 1854 C. demonensis (M. Gemmellaro) 1911 ?6'. dispansus (Morris and Lycett) 1850 i . distefanoi (Gemmellaro) 1884 C. domeyJcus (d'Orbigny) 1842 C. egregius (Pia) 1914 C. ercycinus (Tagliarini) 1901 C. excavatus (J. de C. Sowerby) 1826 C. exiguus (Crick) 1898 Jurassic Germany (Lias) Jurassic Turkey (Lias) Jurassic Europe (Lias) Jurassic Alps (Lias) Jurassic Himgary (Lias) Jurassic Italy (Lias) Jurassic England (Bajocian) Jurassic Italy (Lias) Jurassic Italy (Lias) Jurassic England (Bajocian) Jurassic Sicily (L. Dogger) Jurassic Chile (Lias) Jurassic Sicily (Lias) Jurassic England (Bathonian) Jurassic Italy (Lias) Jurassic Chile (Lias) Jurassic Austria (Lias) Jurassic Sicily (L. Dogger) Jurassic England (Bajocian) Jurassic England (Bajocian) 374 BULLETIN : JrUSEUM OF COMPARATIVE ZOOLOGY Species Stratigraphic Geographic Distribution Distribution C. exterehratus (Criek) 1898 C. fi-'icheramts (Foord and Crick) 1890 C. fourneti (Dumortier) 1874 C. geyeri (Prinz) 1906 C. hnllstattensis (Spengler) 193 9 C. imlayi (Kuniinel) 1954 C. impendens (Crick) 1898 C. inoniatu.s (d'Orbigny) 1842 E. intermeduis (J. Sowerby) 181 fi C. jourdani (Dumortier) 1874 C. juUanvs (Fucini) 1895 C. jurensis (Quenstedt) 1858 C. Uneatus (J. Sowerby) 1813 C. lineolatus (Foord and Crick) 1890 C. Iiipheri (Kummel; 1954 C. lutatii (Gemmellaro) 1886 C. maUi.erhii (Terquem) 1855 C. manissad jinni (Gugenberger) 1928 C. mariani (M. Gemmellaro) 1911 C. marii (Gemmellaro) 1886 Jurassic England (Bajocian) Jurassic France (Lias) Jurassic Europe (Lias) Jurassic Austria (Lias) Jurassic Austria (Lias) Jurassic Alaska (Bajocian) Jurassic England (Bajocian) Jurassic Europe (Lias) Jurassic Europe (Lias) Jurassic Europe (Lias) Jurassic Italy (Lias) Jurassic Europe (Lias) Jurassic England (Bajociauj Jurassic England (Lias and France Bajocian) .Jurassic Oregon (Bajocian) (U.S.A.) Jurassic Sicily (L. Dogger) Jurassic Luxembourg (Lias) Jurassic Turkey (Lias) Jurassic Sicily (Lias) Jurassic Sicily (L. Dogger) KUMMEL : POST-TRIASSIC XAUTILOID GENERA 375 Species Stratigraphie Geographic Distril lilt ion Distribution C. mazzarensis (Tagliarini") 1901 C. meyrati (Ooster) 1858 C. multiseptaius (Foord and Oick) 1890 C. obesus (J. Sowerby) 1816 C. ohstructus (Deslongebanips) 1878 C. orbignyi (Prinz) 1906 C. nrnatus (Foord ;md Crick) 1890 C. nrnatus var. atanatcnsls (Pia) 1914 C. paretoi (M. Gemmellaro) 1911 C. Tperinflatus (Foord and Criok') 1890 C. pfirornaius (Crick) 1894 C. pe.rtexiii>i ( J)uni(ii'tli'r) 18(>7 C. pimniis (Fucini) 189o C. po!>i(/'>nali.<: (.7. d.' C. Sowerby) 1820 C. postriutuK (Prinz) 1900 C. prof andisiphytes (Prinz) 1900 C. pseudolincalus (Foord and Crick) 1890 C. pseu(lori((/osus (Pia'^ 1914 C. psfiudotruncatuft (Crick) 1921 C. quadrangularis (Pia) 1914 C. robu.ifvs (Foord and Crick) 1890 Jurassic Sicily (L. Dogger) Jurassic Alps (Lias) Jurassic England (Bajocian) Jurassi* England (Bajocian) Jurassic France, (Bajocian) England Jurassic Europe (Lias) Jurassic England (Bajocian) Jurassic Austria (Lias) Jurassic Sicily (Lias) Jurassic England (Bajocian) Jurassic Western (Bajocian) Australia Jurassic France (Lias) Austria Jurassic Italy (Lias) Jurassic England (Bajocian) Jurassic Europe (Lias) J urassie Austria (Lias) Jurassic England (Bajocian) Jurassic Alpine (Lias) Region Jurassic England (Lias) Jurassic Austria (Lias) Jurassic France, (Lias) Austria 376 BL LLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Species Stratigraphie Geographic Distribution Distribution C. rotundum (Hyatt) 189-4 C. rotwndum (Crick) 1898 C. rugosu^ (Buvignier) 1832 C. schlumbergeri (Terqueiii) 1855 C. schmidti (Giebel) 1852 C. schwalmi (Prinz) 190ti C. secernendus (Pia) 1914 C. semiornatus (Crick) 1898 C. .lemistriatus (d'Orbigny) 1843 C. semseyi (Prinz) 1904 C. simillimiis (Foord and Crick) 1890 C. smithi (Foord and Crick) 1890 C. spreaficoi (Parona) 1897 C. steiiimanni (Moricke) 1894 C. stoppanii (Parona) 1897 C. striutiis (J. Sowerby) 1817 C. 6-tu,ri (Hauer) 1856 C. subrotundus (Crick) 1898 C. subtruncatus (Prinz) 1906 C. terebraUis (Dumortier) 1874 J urassic England (Bajocian?) Jurassic England (Bajocian) J urassic Europe (Lias) Jurassic Luxembourg (Lias) Jurassic Germany (Lias) J ura;-sic Hungary (Lias) •J urassic Austria ( Lias ) Jurassic England (Bajocian) Jurassic France (Lias) Jurassic Hungary (Lias and L. Dogger) Jurassic England, (Lias) Austria J urassic England (Bajocian) Jurassic Italy (Lias) Jurassic Chile (Middle) J urassic Italy (Lias) Jurassic Europe (Lias) Jurassic Alpine (Lias) Region Jurassic England (Bajocian) Jurassic France (Lias) Jurassic France, (Lias) England KUMMEL : POST-TRIASSIC XAUTILOID GENERA 377 Species Stratigraphie Geographic Distribution Distribution Jurassic Sicily (L. Dogger) Jurassic France, (Lias) Germany, Jurassic Austria (Lias) Jurassic Europe (Lias) Jurassic Sicily (L. Dogger) Jurassic Sicily (L. Dogger) Jurassic Sicily (Middle) C. thyrrenus (Tagliarini) 1901 r. ioarcensis (d'Orbigny) 1849 C. tricarinatus (Vadasz) 1911 C. truncatus (Sowerby) 1816 C. icaageni (Gemmellaro) 1886 C. zignoi (Gemmellaro) 1886 C.zitteli (Gemmellaro) 1886 Genus ErTREPHOCKKAS Hyatt, 1894 Eutrephoceras Hyatt, 1894, p. 555. Eutrephoceras Eeeside, 1924, p. 2. Eutrephoceras Eeeside, 1927b, pp. 6, 7. Eutrephoceras Miller and Thompson, 1933, pp. 300-304. Eutrephoceras Spath, 1927a, pp. 21, 24. Eutrephoceras Spath, 1927b, pp. 425, 426, 428. Eutrephoceras Stenzel, 1940, pp. 738-742. Eutrephoceras Miller, 1947, pp. 26-39. Eutrephoceras Miller, 1951, pp. 33-36. The genus Eutrephoceras represents the long-lived and per- sistent stock arising from the lower Jurassic Cenoceras complex and is considered to be the root from which most of the remain- ing Cretaceous and Tertiary uautiloids, excepting the Cymato- oeratinae, are derived. It is most similar to Cimomia and Xau- tilus. Cimomia arose from Eutrephoceras in the Upper Jurassic or Cretaceous and is in itself the root stock of the Hercoglossinae. Cimomia differs from Eutrephoceras in the greater sinuosity of the suture, but many intermediate forms are known. Thus the interpretation of Eutrephoceras presented here vis- ualizes a persisting, slowly evolving stock consisting of generally involute, smooth forms with straight or nearly straight sutures. and a conch that is generally subglobular but can be quite com- pressed or depressed. The rather wide variation in the shape of the conch is illustrated iu Text Figure 13. 378 BULLETIN: MrSEr:\[ OF COMPARATIVE ZOOLOGY Figure 1 3 KUMMEL : POST-TRIASSIC XAUTILOID GENERA 379 Probably the eoiumouest and most widespread genus of post- Triassic nautiloids is Eutrephoceras. To date, approximately 89 species are recognized ranging in age from the Jurassic to the JMiocene ; of these, 6 species are Jurassic in age, 39 Cretaceous and 44 from the Tertiary. Miller has recently described the American Tertiary species of Eutreplioccras and his comprehensive diagnosis of the genus is quoted here (1947, p. 27) : "Conch nautilicouic and typically subglobular ; whorls reniform in cross section, broadly rounded veutrally and laterally, and moderately deeply impressed dor- sally. Aperture marked ventrally hy a broad shallow rounded liyponomie sinus. Umbilicus small and inconspicuous; umbilical shoulders low and rounded. Surface of conch smooth or essen- t ially so. Septa moderately convex apicad ; sutures of t^-pical forms slightly sinuous, but at least the external sutures of some forms are essentially straight. An annular lobe is present in some forms but not in the genotype. Siphuucle small, circular in cro.ss section, and orthochoanitic in structure ; its position varies considerably in the different species, but in no case is it marginal.'' The large number of species of Eutrephoceras reflects their general abundance and widespread distribution geographically and stratigrajihically. The wide variability in conch shape, degree of involution, and inflection of the suture suggest a wide adaptive range for this genetic stock. Probably in no other Fig. 13. Cross sections of the eoneh of A, Eutrephoceras cyclotus (Op- l)el) from Zittel 1868, pi. 3, fig. 2a, diameter 117 mm.; B, E. doissieri (Pictet) 1867, pi. 8, fig. 4b, diameter 63 mm.; C, E. nuhinjlatus (d'Orbigny) 1850, from d'Orbigny 1840, pi. 37, fig. 2, diameter 88 mm.; D, E. jonesi Miller and Thompson, from Stenzel 1940, fig. 115, diameter 225- mm.; E, E. piermntii (Sergio) 1933, pi. 2, fig. 4b, diameter 144 mm.; F, E. Justus (Blanford), from Stoliczka 1866, pi. 93, fig. 2a, diameter 64 mm.; G, E. dekayi (Morton), from Stenzel 1940, fig. 115, diameter 46 mm.; H, E. fictormnum Teichert 1947, fig. 5, diameter 102 mm. ; /, E. reesidei Stenzel 1940, fig. 115, diameter 14.6 mm.; J, E. yieitiergicus (Eedtenbacher) from •Schliiter 1876, pi. 48, fig. 4, diameter 90 mm.; K, E. dekayi (Morton^ from Stenzel 1940, fig. 115, diameter 85 nmi.; L, E. sloani Eeeside from Stenzel 1940, fig. 115, diameter 120 mm.; M, E. carolinense Kellum, from Stenzel 1940, fig. 115, diameter 20.7 mm.; :N', E. lentiformis (Stoliczka) 1866, pi. 93, fig. la, diameter 74 mm.; 0, E. cooTcanum (Whitfield), Stenzel 1940, fig. 115, diameter 220 mm.; P, E. laverdei Durham 1946, pi. 63, fig. 4, diameter 61 mm. 380 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY generic group are the recognized species so gradational in charac- ter and generally lacking in objectivity. Variability studies of large populations would be extremely helpful in assessing the many species groups. Since EufrepJioceras is interpreted as the basic persisting stock derived from the Cenoceras complex, its relationships to the other genetic groups are discussed in detail under each of these derived groups and need not be repeated here. xl similar long-persisting, smooth, involute, generalized stock is represented in the Permo-Carboniferous and Triassic by the Liroccras-Poranautilus trend. From this particular trend was derived the Upper Triassic Clydonautilidae which is a stock adapted for sinuosity of the suture ; a homeomorphous de- velopment of the Ilercoglossinae. EutrepJwceras is world-wide in distribution and ranges in age from the Upper Jurassic into the ^Miocene. DISTRIBUTION OF SPECIES OF THE GENUS EUTREPHOCERAS Species Stratigraphic Geographic Distribution Distribution K. ahltenensu'i (Sehliiter) 187fi E. aleetienne Reeside, 1927 E. allani (Fleming) 1945 E.alUoni (Michelotti) 1840 E. altifrons (Cliapnian) 1915 E. archiacianus d'Oibigny 1840 E. 'balcomben-sis (Cliapnian) 1915 E. hrllcropho)! (Lundgren') 1867 E. benyl Miller, 1947 E. Uanfordi (Douville) 1929 E. hoissieri (Pictet) 1867 E.houchardianus (d'Orbignyj 1840 E. bryani (Gabb) 1877 Cretaceous Germany (Upper) Cretaceous Montana (Upper) (U.S.A.) Oligocene f New Zealand Miocene Italy, Malta Miocene Australia Cretaceoua France Miocene Australia Cretaceous Sweden (Danian) Eocene North Carolina (U.S.A.) Eocene Pakistan Cretaceous Switzerland, (Lower) Algeria Cretaceous France (Albian) Eocene New Jersey (U.S.A.) KUMMEL : POST-TRIASSIC XAUTILOID GENERA 381 Species Stratigraphic Geographic Distribution Distribution E. burl:arti (Castillo and Aguilera) 1895 E. butonensis (Martin) 1933 E. carolinense Kellum, 1926 E. centraHs (J. Sowerby) 1812 E. charpentieri (Leymerie) 1851 E. clemientinus (d'Orbigny) 1840 E.cooJcanum (Whitfield) 1892 E. crassus (Sehafhautl) 1863 E. cydotus (Oppel) 1865 E. dartevellei Miller 1951 E. darupensis (Schliiter) 1876 E. decipiens (Michelotti) 1861 E. dehayi (Morton) 1834 E. depres.suft (Binckhorst) 1861 E. dersertorum (Quaas) 1902 E. dietrichi (Zwierzyeki) 1914 E. douvillei Spath, 1927 E. dubaleni (Peyrot) 1932 E. euthymi (Pictet) 1867 E. expansum (J. de C. Sowerby) 1824 E. faxoense Hyatt, 1S94 E. felix (Chapman) 1915 Jurassic Mexico (Tithonian) Oligocene East Indies Eocene Xorth Carolina (U.S.A.) Eocene England, Germany Cretaceous France (Upper) Cretaceous England (Albian) France Eocene Xew Jersey (U.S.A.) Eocene Bavaria Jurassic Moravia (Tithonian) Eocene Landana, West Africa Cretaceous (4ermany (Senonian) Miocene Italy Cretaceous U.S.A. (Upper) Cretaceous Belgium (Maestrichtian) Cretaceous Libya (Danian) Cretaceous East Africa (Neocomian) Jurassic Eg>-pt (Oxfordian) Miocene France Cretaceous France, (Lower) Switzerland Cretaceous England (Cenomanian ) Cretaceous Denmark (Danian) Miocene Australia 382 Bl'LLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Species Stratigraphic Geographic Distribution Distribution E. ftamweus (Bom-lietti) 19-17 E. francomontaniis (Kului) 1939 E. geelo7i(/ensis (Foord) 1891 E. gosavicus (Eedtenbacher) 1873 E. hannai Vokes, 1937 EJ hallidayi (Waring) 1914 E. hendersoni (Etheridge) 1901 E.indicum (SpenglcM) 1910 E. izmnoensis Yokoyania, 1913 E. japonicus (Shimizu) 1926 E. javanus (Martin) 1879 E. johnsoni Miller, 1947 E. 'jonesi Miller and Thompson, 1933 E. lahcchi (d'Arehiac and Haime) 1854 E. laverdei Durham, 1946 E. leonei (Negri) 1934 E. mayJcsi Miller, 1947 E. marucoensis (Giovine) 1950 E. meta-fiettriuiitii (Douville) 1929 E. montanensi Kummel, 1954 E. montmoUim (Pictet and Campiche) 1859 E. neuhergicus (Eedtenbacher) 1873 E. oregonense Miller, 1947 Cretaceous Tripolitania (Maestrichtian) Jurassic Germany (Upper) Miocene Australia Cretaceous Austria (Upper) Eocene California (U.S.A.) Paleocene California (U.S.A.) Cretaceous Queensland, (Lower) Australia Cretaceous India (Upper) Tertiary Japan Eocene Japan Tertiary Java Paleocene ? Alabama (U.S.A.) Paleocene Alabama (U.S.A.) EoL-eno Pakistan Cretaceous Colombia (Aptian) Eocene Libya Eocene California (U.S.A.) Cretaceous Argentina (Hauterivian) Cretaceous N. Africa (Upper) L. Tertiary Pakistan Jurassic Montana (Callovian) (U.S.A.) Cretaceous Switzerland (Albian) Cretaceous (Upper) Eocene France Austria Oregon (U.S.A.) KUMMEL : POST-TRIASSIC NAUTILOID GENERA 383 Species Stratigvaphic Disti-il:)iition Geograpliic Distribution E. ovoid.eum Crick, 1907 E. i)armensis (Deshayes) 1866 E. perlatus (IMorton) 1834 E. piersantii (Sergio) 1933 E. planoventer Stephenson, 1941 E. quadrilincatus (Favre) 1869 E. reesidei Stengel, 1940 E. regalis (J. de C. Sowcrby) 1843 E. regvJaris (Schafhuutl) 1863 E. restrict us (Griepenr-kerl') 1889 E. rrsupinaiits (Kedtenbacher) 1873 E. scDnfiJippoi (Sorrentino) 1932 E. simile Spath, 1953 E. sloani Eeeside, 1924 E. sphaericum (Forbes) 1846 E.staadti (Cossniann) 1902 E. stephensoni (Dickerson) 1914 E. stricteumtilicatus (Stchepinsky) 1943 E. subinflatvs (d'Orbigny) 1850 E. s tibia evigatum (d'Orbigny) 1840 E. subplicatum (Philippi) 1895 E.szoniaghi (Yogi) 1910 E. thomi Eeeside, 1927 E. tumescens (Franscher) 1895 Cretaceous Zululand Eocene France Cretaceous Alabama (Upper) (U.S.A.) Oligocene Italy Cretaceous Tennessee, (Upper) Texas (U.S.A.) Cretaceous Germany (Upper) Eocene Texas (U.S.A.) Eocene England Eocene Bavaria Cretaceous Germany (Senonian) Cretaceous Austria (Upper) Cretaceous Tripolitania (Maestrichtian) Cretaceous Graham (Upper) Land Eocene South Carolina (U.S.A.) Cretaceous India Eocene France Paleocene California (U.S.A.) Oligocene Turkey Jurassic France (Kimmeridgian) Cretaceous England, (Upper) India, Madagascar, Libya Cretaceous Argentina, (Upper) Chile Eocene Hungary Cretaceous Montana (Upper) (U.S.A.) Eocene Austria 884 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Species Stratigraphic Geographic Distribution Distribution Cretaceous South Africa (Yalanginian; Eocene France Eocene England Cretaceous Poland (Upper) Tertiary Italy ( Lower) Eocene Australia Oligocenc Italy E. uitenhageiise Spath, 1930 E. iimhilicaris (Deshayes; 1835 E. urhanus (Sowerby) 1843 E. vastus (Kner) 1850 E. vicentinvs (Oppenheim; 1901 E. Victorian um (Teichert) 1943 E. vina-ssai (Yenzo) 1937 Genus PsEUDOCENOCEEAS Spath, 1927 Typt .species. Xai'.tiius largiUiertianvs d'Orbiguy, 1840, pi. 18 (by original designation). Picsiotypp illusti'ated on Piate 10, figs. H, 4. This genus can be diagnosed as follows : Conch involute, com- pressed, smooth. AVhorl section sul)rectaugular, venter broad, flattened, ventral shoulders rounded. Whorl sides flattened, sub- parallel. Umbilical shoulders rounded, umbilical wall nearly vertical. The suture is onlv slightlv sinuous, essentiallv straight across the venter and with a broad, generally shallow, lateral lobe. The siphuncle is subeentral in position, that is, closer to the dorsum. As stated by Spath (1927a, p. 24), "Pseudocenoceras is easily distinguished from its Cretaceous contemporaries by its steep umbilical edge, truncated whorl-shape, entomarginal siphuncle and reclined septal edges." The Cretaceous contemporaries of I'seiidocenoceras include members of the Cymatoeeratinae, the Hereoglossinae, EutrepJw- ceras, and Carinonauiilns. The smooth conch and simple suture would ally Euirephocera^ and Pficudoccnoceras; all of the other generic groups are specialized forms trending on quite different patterns. There are only 8 species of Cretaceous nautiloids that can be assigned to Pseudocenoceras. and all are for the most part very similar. The type species is a common form in Cretaceous strata of England and France. Nautilus heryiasensis Pietet (1867) from Lower Cretaceous strata of France appears to be closely allied to Nautilus largilliertianus but has a more central siph- uncle. Nautilus lehardyi Binckhorst (1861) is a quite aberrant KUMMEL ; POST-TRIASSIC XAUTILOID GENERA 38o form with subaugular ventral shoLilders aud a concave venter. It is placed in Pseudocenoceras with question. Pseudoceiioceras is a fairly common form in Cretaceous rocks of Europe from England to the Crimea. One species, Nautilus applanaius Wanner (1902), is from the Upper Cretaceous of Libya. A list of the species of Pseudocenoceras, their age, and geographic distribution is given below. DISTEIBL'TIOX OF SPECIES OF THE GEXrS PSEUDOCEXOCEEAS Stratigraphic Geographic Species Distribution Distribution P. applanaius (Wanner) 1902 P. herriasensis (Pictet) 1867 P. campichei (Karakasch) 1WU7 P. ntoni (Sharpe) 1853 P. fialicianu^ (Alth) 1850 P. largilliertianus (d 'Orbignj- j 1840 PJ lehardyi (Binekhorst) 18()1 P.pirteti (Karakasch) 1907 Genus CaEIXOXAL'TILI'S Spengler, 1910 Type upccies. Carinonautihis ariyalurensis Spengler, 1910, p. 149, pi. 14. .figs. 1 a (' (monotypic). Illustrated here on Plate 11. The only data available on this monotypic genus are the description and illustration of the type species. However, from this source the following diagnosis is made. The conch is very involute, compressed, and much higher than wide. The umbilicus is small anci shallow. The whorl sides are broadly rounded and converge toward the venter. A distinct furrow aligns the ventral shoulder. The venter has a prominent rounded keel that on the adoral part of the living chamber is much broader and is divided by a median furrow. The umbilical shoulders are broadly rounded. The widest part of the whorl section is just ventral of the umbilical shoulders. The suture is onlv slightlv convex. Cretaceous Libya (Upper) Cretaceous France (Lower) Cretaceous Crimea (Lower) Cretaceous Europe (Upper) Cretaceous Po'.and Cretaceous Europe (Cenomanian) Cretaceous Belgium (Maestriehtian) Cretaceous Crimea (Lower) ■ 186 BULLETIN : Ml'SEUil OF COMPARATH'E ZOOLOGY The ouh- markings on the shell are growth lines -which form a deep sinus over the venter. The position of the siphuncle is not known. At a diameter of 1.5 cm. the venter is not yet keeled but merely sharply rounded. At 3 cm. the single prominent keel is present, and at 6 cm. diameter the median furrow is well developed on the widened keel. The measurements of the type species are as follows (Spengler, 11)10, p. 149) : Diameter 93 mm. Height of last whorl 54 mm. AYidth of umbilicus 5.5 mm. Width of last whorl 32 mm. Spengler (1910, p. 149) recognized the uniqueness of this species and suggested that it was derived from Nautilus angustus Blanford {=Ci))iumia) by sharpening of the ventral area. Con- temi^oraneous forms include members of the Cymatoceratinae, Hercogiossinae, Eutrephoceras and Pseudocenoceras. I- am in essential agreement with Spengler on the phylogenetic position of Carifw7iauiihis but would derive them from Eutrephoceras, which is here considered the persistent evolving stock. There is slight difference between the eutrephoceratid stock and the cimomids. The eutrephoceratids show great variability in their smooth, involute conchs and appear to be the only evolutionary reservoir available to give rise to Carinonautilus. This monoty])ic genus is known only from the Ariyalur group (Campanian) of the Trichinopoly district of southeast India. Genus ObixaUTILUS Kobayashi, 1954 Tijpe .specie^. Obiiiaiitiluti pulchra Koliayashi, 19:')41). Type illustiatloii ro produced on Plato 9, figs. 1, 2. Kobayashi 's description of his new monotypic species and genus is quoted below. The specimen is 78 mm. in diameter ; last whorl 25 mm. in width. 46 nnn. in height and 35 mm. in median height. ' ' Shell discoidal, compressed laterally, rapidly coiling and quite involute ; umbilicus almost closed, if not completely ; flanks a little convex, but nearl.v parallel in the inner half and inclined abruptly near the umbilicus ; ventral sinus shallow but fairly KUMMEL : POST-TRIASSIC NAUTILOID GEXERA 387 broad and provided with an obtuse peripheral carina on each side which is siibangulated inside but outwardly merges with the flank. "Radial ribs narroAv, flattopped, separated by narrow and shallow grooves and frequently branching distally by insertion of fine grooves. These ribs and grooves distinctly bent forward from the umbilicus, nearly straight or even slightly concave backward in the broad flank, swing back a little near the venter and moderately sinuated behind on the venter. The curvature of these ribs is very similar to the growth lines of living Nautilus." Kobayashi based the above description on a single well- preserved specimen. The shell is very well preserved and he (Kohayashi) notes "its deep and light bro^ai shades must be the original colour pattern."' Unfortunately, neither the suture nor position of the siphuncle was observed. The ribbing de- scribed by Kobayashi appears to be accentuated growth lines rather than the type of ribbing that is characteristic of the Cymatoceratinae. The most characteristic feature 0/ Otiyiauiilus is the shallow I)ut distinct ventral furrow. Because of this modification of the periphery, Kobayashi rightly pointed to possible affinities with the Paracenoceratidae. However, I agree completely with Kobayashi that OhinaiifUus should be considered a development of the Xautilinae rather than derived from any member of the Paracenoceratinae. Since the suture and position of the .siphuncle are not known, this interpretation will have to remain tentative until more data become available. Ohinautil lis pulchra is known from one specimen collected at a cutting between Aburatsu and Obi towns in the Province of Hyuga (Miya^^aki Pref.), Japan, on the occasion of the high- way operation. The locality is situated in the pre-Mioceue ter- laiii where the Nichinan formation containing the Oligocene Ashiya fauna lies unconformably beneath the ^Miocene Aburatsu formation (Kobayashi, 1954b, p. 183). Genus NAUTILUS Linne, 1758 7/7)' .sprcicK. Xdiitihis pompilius Linne. The interpretation of the scope of the genus Nautilus has changed considerably- since its formal introduction by Linne. 388 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY In the early period of paleoutological research on this group of tetrabranchiate cephalopods nearly all species were placed in the single genus Nautilus. Gradually as more and more genera were proposed, Nautilus became a waste-basket genus for unassigned forms. The greater amount of attention paid to Paleozoic nautil- oids soon pre-empted the use of the genus Nautilus for species of that age. The genus has persisted for most species of Mesozoic and Tertiary nautiloids until rather recently. At the present, for instance. Miller (1951, p. 32) restricts Nautilus to only the Recent species and all of the Tertiary forms previously assigned to Nautilus are placed in Eutrephoceras or Cimomia. In this interpretation I am in full agreement. Eutrephoceras is the persisting stock out of the Cenoceras complex in Upper Jurassic time that shows no particular evolutionary trends from its time of origin to the time of its apparent extinction in the late Tertiary. It is the longest-lived single genetic stock of post-Tri- assic nautiloids, and was ancestral to several other aberrant groups, namely Pseudocenoceras, Carinonautilus, and Ohinau- tilus. It is also ancestral to Cimomia of the Hercoglossinae which developed into an extremely abundant and diverse evolutionary complex. There appear to be no records of any Pliocene or Pleistocene nautiloids. Aside from Aturia, all other species of nautiloids seem to be more closely related to the eutrephoceratid evolving stock than to modern Nautilus. The number of species and specimens of Miocene nautiloids is still very few. However, with the above interpretation Nautilus is thought to arise from the eutrephoceratid stock at some time in the mid or late Ter- tiary. The biology and ecology of modern Nautilus has been very ably summarized by Stenzel (1948, 1952). Similar comprehen- sive discussions of the morphology, etc., can be found in Miller (1947). The extensive literature on modern Nautilus is covered in the bibliographies of Stenzel 's and Miller's papers. Three species of living NautiJiis are generally recognized, namely N. poynpilius (with the umbilicus practically obliterated by the deposition of callus) N. macromphalus (with an open im- perforate umbilicus), and N. umhilicatus (with an open finely perforate umbilicus) (Miller 1947, p. 13). Several "varieties" of N. pompilius are recognized but their significance and rela- tionships are poorly known. To the above three species, Iredale KUMMEL: POST-TRIASSIC NAUTILOID GENERA 389 (1944) has added two new species, namely N. alumnus, from off Queensland and New South Wales, and N. repertus from off Western Australia. Living NautUus is restricted to the southwest Pacific from South Australia to the southern Philippines and eastward to the Fiji Islands. Stenzel (1948, p. 84) gives a good summary of the known records for live specimens of Nautilus. Subfamily PSEUD AGANIDINAE nov. This new subfamily is established for two genera of mainly Jurassic age that include adaptive types characterized by great sinuosity of the suture. These genera are Pseudaganides Spath, and Pseudonautihis Meek which Spath (1927) had previously included in his family Ilercoglossidae. As is discussed in more detail below, Pseudaganides is considered to be a direct radiation from the Cenoceras complex in the Lower Jurassic ; by the Upper Jurassic it was a widely distributed form in the European geosynclinal areas and show^ed a wide variation in conch form and suture. The group did not survive into the Cretaceous. Sometime during the Upper Jurassic an offshoot from Pseuda- ganides developed, characterized mainly by greater elaboration of the suture, that is, the formation of a deeper ventral lobe and angular lateral lobes. Species of this genus range into the Lower Cretaceous. In the Carboniferous to Recent history of the nautiloids there were four distinct adaptive trends built around elaboration of the suture. As would be expected, there are many homeomorph- ous forms between these adaptive trends. Each of these trends had its origin in long-persisting, smooth, generalized stocks ^yitll very simple, nearly straight sutures. The duration of each of these trends is likewise somewhat limited. The Permian trend is known only through a single species of the genus Permoceras from Middle Permian strata of Timor. The Triassic radiatiou is mainl}^ Upper Triassic (Carnian-Norian) in age. The Jurassic- trend, represented by Pseudaganidinae, is Lower Jurassic to Lower Cretaceous in age with the principal radiation in the Upper Jurassic. The last such trend, that of the Hercogiossinae and Aturinae, is Upper Cretaceous to mid-Tertiary in range with the principal radiation in the Lower Tertiary. 390 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Genus PsEUDAGANIDES Spath, 1927 Pseudaganides Spath, 1927a, pp. 22, 25. Hercoglossoceras Spath, 1927a, pp. 22, 26. Nautilus (Pseudaganides) Jeannet, 1951, pp. 17-21. Type species. Nautilus hutcliensis Waagen, 1873, pi. 3, fig. 4 (Plate 28, figures 3, 4 of this report). Type by original designation. When Spath (1927a, p. 22) established this genus the only remark he made as to its characters was the following- statement : "Pseudaganides, which comprises the early 'aganitici', like Somalinautilus, has its origin in 'Cenoceras' and includes, for example, the large suhsinuatus group of the Inferior Oolite, as already mentioned." In his discussion of the type species, Pseudaganides ktitchensis (Waagen), Spath (1927a, pp. 34-85) mentions the following species as belonging to this genus : A'. aganiticiis Schlotheim, N. scMotheimi Loesch {=Naatilus {Her- coglossa) aganiticus Foord and Crick, 1890b, p. 394, fig. 1), and N. ammoni Loesch; he also included with question N. girardoti de Loriol and N. siihhiangulafus d'Orbigny {=Nautilus hiangu- latus d'Orbigny, 1845, pi. 34, figs. 1-3). There are approximately 40 species of Jurassic nautiloids with "goniatitie" sutures. Besides the "aganitici" included by Spath in Pseudaganides, he established the genus Hercoglossoceras [type species — Hercoglossa {'gravesiana (d'Orbigny) var.') kochi (Prinz)], a highly compressed form. Pseudonautilus was established by Meek for Nautilus gcinitzi Oppel {in Zittel, 1868), and is characterized by a pointed lateral lobe and a very deep ventral lobe. A number of species from the Upper Jurassic (recently discussed by Loesch, 1914), Spath tentatively assigns to Hercoglossa (principally an Upper Cretaceous-Eocene de- velopment) though he does mention that they probably repre- sent an independent development. My own studies lead me to conclude that all of the Jurassic nautiloids with "goniatitic" sutures are part of a single evolu- tionary line developed from some compressed elements of "Ceno- ceras" and this line is quite independent of the later hercoglos- sid development. The type species of Pseudaganides — N. kutchcnsis — comes from the Middle Chari group {anceps zone) of Kachh (Cutch) Pakistan. The type measures 47 mm. in diameter, 28 mm. for KUMMEL : POST-TRIASSIC NAUTILOID GENERA 391 length of whorl, 27 mm. as width of the whorl, and the umbilicus is 5 mm. across. AVaagen's illustration is reproduced here on Plate 28, figures 3, 4. The form of the shell is involute with a somewhat rectangular whorl section with flattened sides and venter. The greatest width of the whorl is near the umbilicus. The suture has a prominent lateral lobe and a saddle next to the umbilicus. The suture is projected toward the venter au'l has a shallow lobe on the venter. The shell has fine longitudinal strigation and growth lines. The strigations disappear towards the aperture. Fig. li. Diagrammatic representation of sutures of A, Pseudaganidc.s Jjrunhuheri (Loeseli) 1914, tig. oc; B, P. franconicus (Oppel) from Loescli 1914, fig. la; C, P. subMangidatus (d'Orbiguy) 1843, from d'Orbignv 1850. pi. 34, fig. 3; 1\ P. rogeri (Loescli) 1914, fig. 3a; E, P. pellerensis (Vialli) 1937, fig. la; F, P. sinuafu.^ (Sowprby). from Vialli, 1937, fig. la. All figures greatly reduced. From the Carl)oniferous to the Tertiary, evolutionary trends, formulated on a cfjmplication of the suture, have occurred re- 392 BULLETIN : MUSEUM OF (OMPARATIVK ZOOLOGY peatedly. In most cases it can be established that such evolu- tionary lines arose from ancestors with essentially uncomplicated (nearly straight) sutures. Many of these evolutionary lines are represented by a single or very few species. However, some groups as the Triassic Clydonautilidac have large numbers of species. In genera such as Clydonantdus and Procly do nautilus there is a constancy in the pattern of the suture with, of course, a degree of modification in shape and depth of the lobes, etc. ; there is, however, a very striking difference or range of variation in the shape of the conch (see Kummel, ll)53c, figs. 38, 40, 41). This suggests that the genetic complex governing the nature of the suture is less adaptive than the shape of the conch. Among the Jurassic "agauitici" there is a certain homogeneity in the suture, mainly in the pronounced lateral lobe, and in the pro- jected ventral saddle with its shallow lobe. The differences en- countered are more a matter of degree, with some element of the suture being larger, deeper, or of slightly different shape (Fig. 14). The scope of variation is similar to that found in species of the Clydonautilidac. The early "aganitici" such as ''Hercoglossa" kochi (Prinz) of the Lias, Nautilus subsinuatus d'Orbignj' of the Inferior Oolite, N. crassisinuatus Crick also of the Inferior Oolite, N. pel- lerensis Vialli of the Bajocian of north Italy, and N. (Inilii Gem- Fig. 15. Cross sections of the conch of A, Pseudaganides pulchellus (Jeannet) 1951, tig. 45, diameter -14 mm.; B, P. hrenlceVi (Jeamict) 1951, fig. 37, diameter 50 mm.; C, P. oppeli (Zittel) 1868, pL 4, fig. lb, diameter 100 mm.; D, P. suhliungulatus (d'Orbigny) 1850, from d 'Oibigii.v, 1843, pi. 34, fig. 2, diameter 92 mm.; E, P. rugerl (Loesch) 1914, fig. 31), height 50 mm.; F, P. .s('h}iekli (LuesL-h) 1914, fig. 2a, height 35 mm.; c;, P. agdniticus (Selilotheim) from Jeannet, 1951, fig. 42, height 25 mm.; //, /'. lodii (Prinz) 1906, fig. 5, height 77 mm.; /, P. frlchensis (Jeannet) 1951, fig. 36, diameter 46 mm.; J, P. stramhergensis (Oppel) 1865, pi. 2, fig. 8a, diameter 98 mm.; K, P. biunhuhiri (Loesch) 1914, fig. 51j, height 50 mm.; living chamber; L, P. bninhiihoi (Loesch) 1914, fig. 5a, height 39 mm.; phragmocone; M, P. ainuatus (Sovverby) from d'Orbigny, 1843, pi. 32. diameter 188 mm.; N, P. gravesianus (d'Orbigny) 1843, pi. 38, diametei 270 mm.; 0, Pseudonautiius aturioidts (Pictet) 1867, pi. 11, fig. Ic, diametei 60 mm.; P, P. malhosl (Pictet) 1867, pi. 9, fig. 2b, diameter 100 mm.; Q. Pseudaganides dumasi (Pictet) 1867, \)\. 10, fig. 2Ii, diameter (iO mm. KUMMEL : rO.-T-TRlASSIC NAUTILOID GENERA 393 H K Figure 1 5 894 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY mellaro are all compressed, involute forms with the whorls higher than wide. The sutures have pronounced lateral lobes and are strongly projected towards the venter which they cross in nearly a straight line. In nearly all other species of Fseudagan- ides there is a slight lobe on the peripheral area. The lateral saddle in the above early "aganitici" is quite variable. It is very pronounced in iV. subsinuatus and N. crassisitmatus but very low in N. pellerensis and N. diiilii. In each of these latter two species the umbilical seam passes through about the middle of the lateral saddle. The sutures in most of the remaining ' ' aganitici ' ' of the Juras- sic have a shallow lobe on the venter. This small ventral lobe appears to persist in forms like N. ammoni Loesch, N. rogeri Loesch, N. schncidi Loesch, etc., but in ^Y. schlosseri, after the ventral lobe having been present, it disappears in the last few sutures and is nearly straight or only slightly arched across the venter (see Loesch, 1914, p. 115, fig. 8). There is a rather large degree of variation in the shape of the conch in the species here assigned to Pseiidaganides (Fig. 15). All of the species have nautiliconic conchs but vary in the character of the ventral shoulders and in the width to height proportions of the whorl section. In the type species, N. kut- chensis, the whorl section is subrectangular with subangular ventral shoulders, an arched venter, and flattened convergent flanks. Subangular to acutely rounded ventral shoulders are present in N. argoviensis Loesch, N. aiiimoni Loesch, N. klehels- bergi Loesch, N. aganiticus Schlotheim, N. inbingensis Loesch, N. helveticus Loesch, N. brunhubcri Loesch, N. frickensis Jean- net, and N. snbbknigidaftts d'Orbigny. Well rounded ventral shoulders and arched venters are found in X. pidcheUus Jeannet, N. schneidi Loesch, A'. st)'anibergriisis ()i)pel, N. franconicus Oppel and N. oppeli Zittel. Some species have slightly concave venters, as in N. snbbiangulatus d'Orbigm- and X. rogeri Loesch. In N. ammoni an:l N. schlosseri the venter becomes slightly con- cave on the living chamber, but is arched on the phragmocone. The width and height measurements of the whorl section in N. kutchensis are approximately equal. This same ratio is present in many other species of Pseiidaganides (e.g. N. pulchellus Jean- net, N. rogeri Loesch, N. bninhnberi Loesch, .V. schneidi Loesch, N. ammoni Loesch, etc.). More compressed species include N. Kl'.M.AIEl. : I'OST-TKIASSIC XAUTILQID GENERA 39; subbiaitgiilatns d'Orbigny, N. frickensis Jeannet, N. stramher- (joisis Oppel. etc. Xautilus kochi has a highly compressed conch Avith couvei-gent whorl sides and a rather narrow rounded venter. It appeai-s to be merely an extreme form of Pseudaganides antl does not Avarrant separation as a distinct genus. Hercoglos- soccras established by Spath for N. kochi is here considered a ■synonym of Pseudaganides. Whereas the degree of convergence of the whorl sides is greatest in N. kochi the degree of compres- sion of the conch is found in several other species. Loesch (1914) has described several species of Upper Juras- sic nautiloids which are very similar in conch shape and suture to typical HercogJossa of the late Cretaceous and early Tertiary. That is, these species have well rounded A^entral shoulders and arched venters. Spath (1927a) included them tentatively in Hercoglossa l)ut also expressed the opinion that they "probably foi'in an independent development." These Upper Jurassic • ' h(n-coglossids " include among others: N. franconicus Oppel, r\'. schneldi Loesch, N. rogcri Loesch, .V. hrnnhuheri Loesch, I\ . roemeri Loesch, X. schwertschlageri Loesch, N. schlosscri Loesch, N. stromeri Loesch, N. oppeli Zittel, and N. stramhergensis Op- ])(']. In all of these species, however, there is a shallow ventral lobe which is not fouihl in the Upper Cretaceous-Tertiary Herco- glossa. The great similarity of these two groups is thought to be a case of homeomor|)hy. The true Hercoglossa develop from Cretaceous CUmojiiia whereas the Jurassic "aganitici" are de- rived from Liassic Cenoceras. Pseudaganides is Avidely distributcnl in Jurassic strata through- out Europe, and in Pakistan. Of a total of 37 species listed below, 1 is from the Lias, 9 are from ]\Iiddle Jurassic strata, and 28 are known from Upper Jurassic formations. DISTBIBUTIOX OF SPECIEK OF THE Gr:xr,S PSEUDAGANIDES Species Stratigi;ii)lii(' Geographic Distribution Distriljutioii P. aganiticus (Schlotheini ) IS'20 P. aniDioiii (Loesch) 1914: P. argoviensis (Loesch) 191:^ Jurassic France (Oxfordian) Jurassic Europe (Upper) Jurassic Switzerlan (Oxfordian 396 ni'LLETIX: ilUSEUM OF COMPARATIVE ZOOLOGY Species Stratigraphic Geographic Distribution Distribution P. bodeni (Loesch) 1914 P. bnnihvberi (Loesch; 1914 r. (•/«l/.s/^v (d'Orbigny) 184:^ r. crass itiiiiua tug (Crick) 1898 P.drepanensi^ (Tagliarinij 1901 P.duilii (Genimellaroj 188ti P. franconicus (Oppel) 1865 P. frickensis (Jeannet) 1951 P. girardotl (de Loriol) 1903 P. glaber (Foord and Crick ) 1890 P. graves ianu,s (d'Orbigny) LS4;) P.helveiicus (Loesch) 1914 P.klebelsbergi (Loesch) 1914 P. kocht (Prinz) 19u0 F.krenTceli (Jeannet) 1951 P. kutchensis (Waagen) 1873 P. ledonicus (de Loriol) 1903 P.oppeli (Zittel) 1868 P. pellereiisi^ (Yialli) 1937 P. portlandicus (Foord and Crick) 1890 P. pulchellus (Jeannet) 1951 Jurassic (Oxfordian ) Jurassic (Upper) Jurassic (Bajocian) J urassic (Bajocian) Jurassic (Middle) Jurassic (Middle) Jurassic (Tithouiau ) Jurassic (Callovian) Jurassic (Oxfordian) Jiirassic .1 urassic Jurassic (Upper) Jurassic (Tithonian) Jurassic (Lias) J urassic (Upper) Jurassic (Upper) Jurassic (Oxfordian ) Jurassic (Upper) Jurassic (:\nddle) Jurassic (Upper) Jurassic (Oxfordian) France Europe France England Sicily Sicily Europe Switzerland Smtzerland England, France France Switzerland Czecho- slovakia Austria, Hungary Switzerland Pakistan France Europe North Italy England Switzerland KUMMEL : POST-TRIASSIC XAUTILOID GENERA 397 Species Stratigiaphir (Jeograpliic Distribution Distribution /'. roemeri (Loeseh) 1914 P. rogcri (Loeseh) 1914 P. r oyer I (de Loriolj 1872 P.schlosfteri (Loeseh) 1914 r. schlothcimi (Loeseh; 1914 P. schneiili (Loeseh) 1914 P. schutrtsclilugeri (Loeseh) 19lL' P. strdrnbergpiisis (Oppel) 186.5 P. sironwri (Loeseh) 1914 P. suhbwnr/tilatim (d'Orbigny) 1850 P. nubshiuntus (d'Orbigny) ]8o0 P. tubingnusi.s (Loeseh) 1914 P.turl-criii (Gemmellaro) 188() Jurassic East Europe (Upper) Jurassic East Europe (Upper) Jurassic France (Upper) Jurassic East Europe (Tithouiau'i Jurassic Lorraine (Middle) Jurassic Europe (Tithonian) Jurassic Europe (Upper) Jurassic Czecho- ( Tithonian) Slovakia Jurassic Czeeho- ( Tithonian; Slovakia Jurassic France (Bathonian) Jurassic England (Middle) Jurassic Germany (Upper) Jurassic Sicily (Middle) Geuus PSECDOXAI I'lLi ts Meek, 1876 P/ieadunaaUiu.s Meek, 1876, p. 491. Paeudonautilus Spath, 1927a, pp. 20, 22, 26. Pseudonautilus Miller aud Collinson, 1953, pp. 293-295. T\ipc species. Nautilus geiiiitzi Oppel, in Zittel, 1868, p. 45, pi. 2, tigs. 17 (Plate 12, figures 1, 2 and Text-Figure 16 of this report). This genus is well characterized by the extreme " goniatitic ' ' development of its suture. Only four species are assignable to the genus and the following diagnosis has been compiled from speci- mens in the M.C.Z. and the description and illustrations of these species. Conch involute, compressed, and smooth. Whorl sides flat- tened, venter low and broadly arched. The ventral and umbilical shoulders are rounded. The whorl section is higher than wide. 398 BULLETIN: MUSEUM OP^ COMPARATIVE ZOOLOGY The umbilicus is very small and deep. The suture has a deep narrow ventral lobe, a large V-shaped lateral lobe followed by a rounded saddle with a small second lateral lobe just above the umbilical seam. There is a deep pointed dorsal lobe. The siphuncle is in a subventral position. The most diagnostic character of Pseudoyiautilus is the sinuous suture with the deep V-shaped ventral lobe. The four species assignable to Pseudonautilus are : P. geinitzi Oppel, the type species; N. aturioides Pictet (1867) ; N. dumasi Pictet (1867) ; and N. malhosi Pictet (1867). Loesch (1914, p. 130-131) dis- cussed in some detail various aspects of N. gemitzi and N. aturi- oides. Nautilus geinitzi is from the Upper Jurassic (Tithonian) of Moravia, Silesia, and France, and has also been recorded from the Berriasian of Tunisia by Arnould-Saget (1953). Nautilus aturioides is from the zone of Hoplites hoissieri, lowest Valan- ginian (Loesch, 1914, p. 131). Nautilus malhosi Pictet has more angular ventral shoulders than in the type species and likewise the ventral and lateral lobes are rounded rather than pointed. In respect to the suture, N . malhosi and N. dumasi are more or less intermediate forms between typical Pseudaganides and N. geinitzi, the type species of Pseudonautilus. The assignment of these two species to Pseudonautilus is open to question. Nautilus malhosi has been recorded from Lower Cretaceous formations in central Europe (Pictet, 1867), Algeria (Pomel, 1889), and from the Crimea (Karakasch, 1907). One of the most remarkable cases of homeomorphy among post-Carboniferous nautiloids is the similarity of Nautilus (Aga- nides) hitauniensis Ilaniel from the Permian Bitauni beds (Leonardian) of Timor and Pseudonautilus gei7iitzi Oppel (Fig. 16.) Miller and Collinson (1953) have recently discussed this problem and erected the genus Pcrmoceras and the family Permoceratidae for Haniel's si)ecies. They likewise recognized the homeomorphous relationship of these two nautiloids. Among Triassic nautiloids sinuous sutures are found in the Clydonautili- dae, Gonionautilidae Siberionautilidae, and in the genus Clyme- nonautilus of the Syringonautilidae. Sinuous sutures are, how- ever, not common in Permian nautiloids. Among the forms comprising the Grypoeeratidae, Grypoceras (Plummeroceras) plummeri (Kunimel, 1953c) — a compressed evolute, quadrate whorled form — has a suture with a deep ventral lobe and a broad KUMMEL : POST-TRIASSIC NAUTILOID GENERA 399 deep lateral lobe. The general shape of the conch except for the degree of involution is rather similar in G. (Plummeroceras) and Pennoceras. Within the complex of Permian nautiloids now known it seems more logical to consider Permoceras an offshoot of the Domatoceras-Grypoceras evolving stock and thus a mem- ber of the Grypoceratidae. Fig. 16. Diagrammatic representation of sutures and cross sections of Permoceras hitauniense (Haniel) A, B, and Pseudonautilus geinitzi (Oppel) C, B, all X 1. From Miller, and Collinson, 1953, fig. 1. Subfamily PARACENOCERATINAE Spath, 1927 When Spath established this taxonomic unit (originally as a family) he included in it Paracenoceras Spath, Tithonoceras RetoAvski, ^omalinautilus Spath, Aulaconaiotilus Spath, Carino- nautilus Spengler and with question Heminaiitilus Spath. Ex- 400 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY amination oi" topotypes of Heminautilus etheringtoni Durham and the descriptions and illustrations of H. lalUerianus d'Or- bignj^ (especially in Douville, 1916, pi. 17) lead me to conclude that these species belong more properly in the Cymatoceratinae. They have the characteristic ribbing", even though in some speci- mens the ribbing is very weak or absent, and a hercoglossid suture, more sinuous than the suture in most species of the Paraceno- ceratinae. However, on the basis of the ornament alone this genus should not be left in the Paracenoceratinae. Carinonautilus is a monotypic form known only from the Upper Cretaceous of south India. The paracenoceratids are primarily a Jurassic radiation from the Cenoceras complex and it is difficult to visualize a derivation of Carinonautilus from this source. It is here allied to the Nautilinae and considered a derivative of Eutrephoceras which is the long-persisting, slowly- evolving stock ranging from the Upper Jurassic to the middle Tertiary. Of the remaining genera, only Paracenoceras is wide- spread and fairly common as far as Mesozoic nautiloids go. It has by far the largest numlier of species ; of a total of 42 species in this family 35 belong in Farace^ioceras. Tithonoceras is mono- typic, Somalinautilus and Anlaconautilus have three or less species each. Whereas Spath did not diagnose this family at length, he did state that it was "characterized by differentiation of the peri- phery', generally associated with increase of the sinuosity of the suture line, and a ventral lobe" (Spath, 1927a, p. 25). Para- cenoceras has a broad, generally sulcate periphery. Tithono- ceras has a broad flattened periphery with a median furrow and the ventrolateral area projected into a prominent blunt keel aligned by furrows. Somalinautilus has an arched venter with angular ventral shoulders. Aulaconautilus has longitudinal ribs on the peripheral area. Both Paracenoceras and Somalinautilus are interpreted as off- shoots from the Cenoceras complex. Both of these genera first appear in the Middle Jurassic and it seems likely that they stem from separate sources within Cenoceras. Aulaconautilus and Tithonoceras are more specialized developments stemming from Paracenoceras. The subfamily is largely confined to the eastern hemisphere, being known from England, Europe, east and north Africa, Arabia, and Pakistan. The only occurence of a member KUMMEL: POST-TKIASSIC NAUTILOID GENERA 401 402 BULLETIN : :\IUSEUM OF COMPARATIVE ZOOLOGY of this subfamily in the western hemisphere is Parace7ioceras in Cuba. As is so characteristic of post-Triassic nautiloids the main stock of a subfamily is generallj^ a highly variable form with many described species and a wide geographic distribution. This is well illustrated in the distribution of species of this subfamilv (Fig. 17). Genus PakaCENOCERAS Spath, 1927 Paracenoceras Spath, 1927a, pp. 22, 25. Paraoenoeeras Si^ath, 1935b, pp. 224-225. Paracenoceras Scott, 1943, pi). 88, 89. Nautilus {Paracenoceras) Jeannet, 1951, pp. 10-17. Paracenoceras Miller and Collinson, 1952, pp. 626-628. Type species. Nautilus hexagonus J. de C. Sowerby, 1826, pi. 529, fig. 2 (illustrated here on Plate 8, fig. 1, and Plate 13, figs. 1, 2; Text Fig. 18). Type by original designation. Paracenoceras is the only genus of the Paracenoceratinae that is nearly world wide in distribution and includes a relatively large number of species. Of the remaining genera of this sub- family one is monotypic and the remaining two genera contain no more than three species apiece. The type species — Nautilus hexagonus J. de C. SoAverby — was previously only known through the rather unsatisfactory description and woodcut illus- tration (J. de C. Sowerby, 1826, pi. 529, fig. 2). Sowerby 's type is illustrated on Plate 13, figures 1, 2, and Plate 8, figure 1 and Text Figure 18. The genus can be diagnosed as follows : Couch robust, involute, smooth. Whorl section subtrapezoidal in outline. Whorl sides flattened, convergent towards a broad venter that may be flat- tened or truncate to distinctly sulcate. Ventral shoulders well rounded. Umbilicus generally small, umbilical shoulders rounded, umbilical walls steep. The shell is smooth except for sinuous growth lines which form a deep sinus on the broad venter. The suture is only slightly sinuous but with a distinct ventral and lateral lobe and a small saddle at the umbilical shoulder and wall. The position of the siphuncle is quite vari- able but never in an extreme ventral or dorsal position. The polished median section of Sowerby 's tj'pe specimen shows that the siphuncle is beaded and with short funnels (Plate 8, fig. 1). Similar beaded siphuncles have been also noted by Spath (1927a, KUMMEL : POST-TRIASSIC NAUTILOID GENERA 4()o p. 23) in species of '' Bisiphytes," " Digonioceras/' and Cyma- foceras. Approximately 35 species of Paracenoceras have been recorded most of which are from Upper Jurassic strata. Spath (1935b, p. 224) has described Paracenoceras prohexagonum from strata o. Fig. 18. Paracenoceras hexagonuia \,ttowerby). Diagrammatic cross sec- tion of holotype represented by figures 1, 2, Plate 13, and figure 1, Plate 8, X 1. Bathonian? age. This is the oldest species known, and the youno'cst species of Paracenoceras appears to be P. rJiodani (Roux) from the Albian of Switzerland. Sowerby's type of NaKfilus liexagonns, refignred here on Plate 13, figures 1, 2 and Text Figure 18, is a large phragmocone meas- uring 95 mm. in diameter, 54 mm. in height of last whorl, and 67 mm. in width of the last whorl. The whorl sides are only slightlv convex and converge towards a broad sulcate venter. 404 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY F'igure 19 KUMMEL: POST-TRIASSIC NAUTILOID GENERA 405 The umbilical and ventral shoulders are well rounded. The umbilicus is fairly large, deep, and funnel shaped. The umbili- cal Avails are very steep. The large siphunele is subcentral in position. There is much A'ariation in shape of the whorl section among the various species of Paracenoceras. The cross sections of 12 species are illustrated in Figure 19 to show the degree of varia- tion. Sulcation of the venter of the adoral volution as on the type species, is not present on several of the species which have merely truncated venters or ventral areas slightly arched. Para- cenoceras maroccnse Miller and CoUinson (1952) has a broad sulcate venter at earh^ maturity but the venter becomes arched with increasing size (Text Figure 19 G-I). The whorl sides also show some variation from being slightly concave as in P. liexa- (jonuni and P. marocense, to flattened or slightly convex in most of the remaining species. Variations in suture are mostly a mat- ter of degree. Sutures of several species of Paracenoceras are illustrated in Figure 20 for comparison. Paracenoceras costatum Scott (1943, p. 88, pi. 24, figs. 1, 3), with prominent sigmoidal costae on the whorl sides, belongs in the genus Procymatoceras. Scott (1943, p. 89) compared "Para- cenoceras" costatum with P. juniarense (Waagen, 1873, p. 21, pi. 4, figs, la, b) which has sigmoidal ribs only on the whorl sides, that terminate in large tubercles at the ventrolateral shoulder. However, Spath (1927a, p. 32) questions the accuracy of Fig. 19. Cross sections of the coneh of A, Paracenoceras calloviensis (Oppel), from Waagen 1873, pi. 3, fig. 2b, diameter 65 mm.; B. P. suhhexa- gonum Jeannet 1951, fig. 19, height i2 mm.; C, P. roUieri (Loesch) from Jeannet 1951, fig. 29, height 27 mm.; D, Aulaeonautilus Mcarinntus Jeannet 1951, fig. 32, diameter 41 mm.; E, Paracenoceras icandaen,sis (Waagen) 1873, pi. 4, fig. 3b, diameter 85 mm.; F, Aulaeonautilus sexcarinatus (Pietet) 1867, pi. 10, fig. lb, diameter 107 mm., the genotype of Aulaeo- nautilus; G, H, I, Paracenoceras marocense Miller and Collinson, 1952, fig. 1 {G, height 28 mm.; H, height 48 mm.; I, height, 72 mm.); J, P. prohexagonum Spath 1935, fig. 46, height 44 mm.; K, P. hexagonum (Sowerby) from Spath 1935, fig. 4d, height 56 nun., the genotype of Para- cenoceras; L, P. dilatatus Jeannet 1951, fig. 8, height 140 mm.; J/, P. Icumagunensi^ (Waagen) 1873, pi. 3, fig. lb, diameter 72 mm.; N, P. hersnachcnsis Jeannet 1951. fig. 14, height 100 mm.; 0, P. enniunum (Dacque), from Scott 1940, pi. 24, fig. 2, height 70 mm.; P, P. arduennensis (Loesch, 1914) from Jeannet 1951, fig. 26, height 19 mm.; Q, Aulaeo- nautilus picteti (Oppel), from Zittel 1868, pi. 3, fig. 36, diameter 95 unn. 406 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Waagen's illustration and points out that the specimens are badly weathered. Spath quite rightly maintains that the affini- ties of P. jumarense are very doubtful and the species can be retained in Paracenoceras with question. A list of the species assigned to Paracenoceras with their gen- eralized geologic and geographic data is given below. i\Iost of the species are from the Jurassic Tethyian geosynclinal area and adjoining epicontinental seas of Europe. Several occurrences lig. 20. Diagrammatic representation oi suiures oi A, Famcenoceras dilatatus Jeannet 1951, fig. 8; B, P. hexagnnum (Sowerby), from Spath 1935, fig. 4e, the genotype of Paracenoceras; C, P. acMini Jeannet 1951, fig. -25; D, P. marocense Miller and Collinson 1952, fig. lA ; E, P. rollieri (Loesch, 1914) from Jeaunet 1951, fig. 29; F, P. hersnachoisis Jeannet 1951, fig. 14; G, Eeminaiitilus stantoni Scott 1940, fig. 179; H, Paraceno ceras prohexagonum Spath 1935, fig. 4e; /, Heminautilus etheringtoni Dur- ham 1946, fig. 3C; J, Heminautilus lallierianus (d'Orbigny) from Durham. 1946, fig. 3A. All figures greatly reduced. KUMMEL : POST-TRIASSIC NAUTILOID GENERA 407 in eastern Africa, Madagascar, Arabia, and in Pakistan form an eastern extension of the European faunal distribution. The onh* occurrence known in the Western Hemisphere is in Cuba (Sanchez Roig, 1951). As noted above, most of the species of Paracenoceras are Upper Jurassic in age. DISTRIBUTION OF SPECIES OF THE GENUS PARACENOCERAS Species Stratigraphic Geographic Distribution Distribution P. acklini ( Jeannet) 1951 P . arduennensis (Loesch) 191i P. bnintrutanae (Kuhn) 1936 P. calloviensis (Oppel) 1858 P. dilatatus (Jeannet) 1951 P. divesianus (Kuhn) 1936 P. dorsatus (Roemer) 1836 P. dorsoexcavatum (Parona and Bonarelli) 1897 P. ennianwm (Dacque) 1905 P. giganteus (d'Orbigny) 1825 P . gramulosum (d'Orbigny) 1843 P. herznachensis Jeannet, 1951 P. hexagonoides Spath, 1927 P. heragoninn (J. de Sowerby) 1826 P. jeannet i Sanchez Eoig, 1951 PJ jumarense (Waagen) 1873 Jurassic SAvitzerland (Callovian) Jurassic France (Oxfordian) Jurassic Germany (Upper) Jurassic Europe (Callovian) Jurassic Switzerland (Callovian) Jurassic France (Upper) Jurassic Germany (Upper) Jurassic France (Callovian) Jurassic Somaliland (Upper) Jurassic France (Upper) Jurassic France (Upper) Jurassic Switzerland (Callovian) Jurassic Pakistan (Tithonian) Jurassic Europe (Upper) Jurassic Cuba (Oxfordian) Jurassic Pakistan (Bathonian) 408 BULLETIN : MUSEUM OF COMPAEATIVE ZOOLOGY Species Stiatigrapliic (geographic Distribution Distribution P. Icumagunense (Waagen) 1873 P. hitifrons (Zwierzycki) 1914 P. marncense Miller and Collinson, 1952 P. mjatsclil:uwa7iits (Loescb) 1914 P. miyreani (d'Orbigny) 1842 P. mullerriedi Sanchez Eoig, 1951 P. prohexagonum Spath, 1935 P.rhodani (Eoux) 1848 P. roUieri (Loesch) 1914 P.sattleri (Krenkel) 1910 P.schattenhcrfii (Kuhn) 1936 P. schusteri (Loesch) 1912 P. siruJus (Gemmellaro) 1868 P. sinuosn^ (Roemer) 1836 P. staffelbergensis (Kuhn) 1936 P. suhhexagorms ( Jeannet) 1951 P.voJgcnsis (Nikitin) 1888 P.u-andaense (Waagen) 1873 P. wilviae (Jeannet) 1951 Jurassic Pakistan (Callovian) Jurassic Tanganyika (Upper) Jurassic Morocco (Upper) ? Jui'assic Russia (Upper) Jurassic France (Kimmeridgian) Jurassic Cuba (Oxfordian) Jurassic British (Bathonian?) Somaliland Cretaceous Switzerland (Albian) Jurassic Switzerland (Oxfordian) Jurassic Tanganyika (Kimmeridgian) Jurassic Germany (Upper) Jurassic France (Kimmeridgian) Jurassic Sicily (Upper) Jurassic (Upper) Jurassic (Upper) Jurassic (Callovian) Jurassic Russia Jurassic Pakistan (Upper) Jurassic Switzerland (Callovian) Germany Germanj- Switzerland KUMMEL : POST-TRIASSIC NAUTILOID GENERA 409 Genus AULACONAUTILUS Spath, 1927 Aulaconautilus Spath, 1927a, pp. 22, 25. Nautilus {Aulaconautilus) Jeannet, 1951, p. 17. Type species. Xaidilus sexcaririatus Pictet, 1867, PI. 10, figs, la-c (repro- duced here as Plate 14, figs. 1, 2). Type by original designation. The following diagnosis is made from study of the descrip- tions and illustrations of the three species that can be assigned to this genus. No specimens were available for study. Conch involute, generally compressed, with greatest width of the whorls just below the umbilical shoulders. AYhorl sides smooth, convex and converging. Ventral shoulders rounded, venter broad, low, and with longitudinal ribs. Umbilicus very small, umbilical shoulders rounded. Suture sinuous with a shallow ventral lobe, a broad deep lateral lobe with a saddle at the umbilical shoulder. Position of the siphuncle not known. Only three species can be placed in Aulaconautilus; these are A. sexcarinatus (Pictet) the type species, A. picteti (Oppel), and A. hicarinatus (Jeannet). There is variability in the conch shape, suture, and ornamentation. The type species, A. sexcari- natus, has eight longitudinal ribs on the venter (Plate 14, figs. 1, 2 and Text Figure 19i^), divided by shallow furrows. The median furrow is the most prominent. The ribs at the ventro- lateral edge are the weakest of the ribs on the venter. Aulaco- nautilus picteti has only four longitudinal ribs all of equal prominence ; the median furrow is the widest and deepest of the furrows between these ribs. This species likewise is much more inflated than the type species (Text Fig. 19^). Aulaconautilus hicarinatus has a more depressed, subquadratic whorl section with a broad flattened venter. There are four widely spaced longitudinal ribs on the venter (Text Fig. 19D). The orna- mentational pattern of longitudinal ribs on the venter, which characterizes Aulaconautilus, is unique for post-Triassic nautil- oids. A homeomorphous development is the Permian and Trias- sic genus Aulametacoceras Miller and Unklesbay. The type of this genus is Middle Permian in age (from Arizona). The only other species assigned to it is Nautilus rectangularis Hauer of Upper Triassic age. Aulametacoceras is considered a radiation from the stable evolving Metacoceras complex (Kummel, 1953c). The sutures of species of this genus, especially those of the 4:10 BULLETIN : MUSEUiM OF COMPARATIVE ZOOLOGY type species, are very similar in plan and degree of sinuosity to those of Pseudaganides. The other groups of the Paraceno- ceratinae have sutures with A-ery shallow lol)es and saddles. The three known species of Aidaconaniihis come from Upper Jurassic strata of Europe. Genus TiTHONOCERAS Retowski, 1894 TithoHOceras Ectowski, 189-4, p. 223. Tithonoceras Crick, 1908, p. 14. Tithanoceras Spath, 1927a, pp. 20, 22, 25. Type species. Tithonoceras zitteli Retowski, 1894, p. 223, pi. 13, figs. 2a, b, c (refigured here on Plate 15, figs. 1, 2). This is another of the numerous mouotypic Mesozoic nautiloid genera. The type species is from Tithonian (Upper Jurassic) strata of the Crimea. The conch is evolute, compressed, and smooth. The whorls are subrectangular in outline, being higher than wide. The venter is broad, flattened, and with a shallow median furrow. The ventral shoidders are inflated and form a smooth keel-like ridge that is aligned on the venter and the whorl sides by a furrow. The ventral furrow aligning the ventrolateral keel is rather narrow but very distinct. The fur- roAV on the whorl sides adjoining the ventrolateral keel is con- cave and broad occupying nearly half of the whorl side. The dorsal half of the whorl side is convex passing onto a broadly rounded umbilical shoulder. The umbilicus is large, measuring one-eighth the diameter of the conch. The shell is smooth except for sinuous growth lines which are deeply concave on the venter. The suture is very sinuous with ventral and lateral lobes. The position of the siphunele is not known. Proper evaluation of monot.ypic foi'ms is not entirely satis- factory. Little can be said of TitJionoccras exce])t that it appears to be a specialized offshoot of Paracenuceras. The general conch form is unique among Carboniferous to Recent nautiloids. There are no homeomorphous types of other ages known to me. Tn its derivation from Paracenoccras there was merely greater compression of the conch and further elaboration of the ventral area. The type and only known species of TiiJionoceras came from Tithonian strata (Upper Jurassic) of the Crimea, Russia. KUMMEL : POST-TRIASSIC NAUTILOID GEXERA 411 Genus SOMALIXAUTILUS Spath, 1927 Tiipe species. Xautilus mitiquus Dacque, 1910 i=X. hisulcutus Dacque, 1905, p. 144, pi. 16, figs. 3a, b). Type by original designation. Type figure reproduced here on Plate 14, figs. 3, 4. The Paracenoceratinae are derived from Liassie cenoceratids and are characterized by modified venters. The earliest species of this subfamily are included in the genus Somalinautilns. At this date only three species of Somalinautilns have been re- corded. The type species is from Upper Jurassic (Kimmerid- iiian) strata of Somaliland. The other species definitely assigned to this genus are Xantilus fuscus Crick (1898, p. 122) and N. clausus in Foord and Crick (1890b, p. 284) Jion d'Orbigny (1842, p. 158, pi. 33). From the description and illustration of the type species and study of several specimens of Nautilus fuscus in the British ]\[uscum (Natural History), the following diagnosis is made. Conch involute, depressed, rapidly expanding, whorls wider than high. Venter slightly arched with subangular ventral shoulders, with a distinct furrow on venter adjoining the ventral .■-lioulder. The whorl sides are concave adjacent to the ventral shoulders and convex towards the umbilicus. Umbilical should- ers are sharply rounded and the umbilical wall very steep. Umbilicus measures less than one-quarter the diameter of the conch. Surface of the conch with sinuous growth lines and may or may not have longitudinal striae. Suture forms shallow ventral lolie. generally an angular saddle at the ventral shoulder followed by a broad, shallow, lateral lobe, with a low saddle on the umbilical wall. The position of the siphuncle in the type species is not known but in Nautilus fuscus it is slightly below the center. Each of the generic groups of the Paracenoceratinae is very distinct and Somalinautilus with its sul^angular ventral shoulders and low arched venter is readily differentiated from the other genera. Somalinautilus most probably arose directly out of the Cenoceras complex and not from Paracenoceras. There are several quadrate whorled forms of Cenoceras which could have given rise to those Middle and Upper Jurassic species placed in Somalinautilus. Species of this genus are known from Eng- land, France, and Somaliland. 412 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Subfamily CYMATOCERATINAE Spath, 1927 This subfamily now comprises 10 genera of post-Triassic nautiloids which have variable conch shapes and degree of sinuosity of the suture but have in common an ornamentational pattern consisting of ribs. Ornamentation is relatively rare in post-Triassic nautiloids, excepting those forms within the Cyma- toceratinae. There are no species with nodose ornamentation known. Within this subfamily there is displayed a wide range of adaptive radiation as seen in the extent of variation in conch shape and suture. The genus Cymatoceras is the basic adaptive type ; it has an involute, rounded conch with only a slightly sinuous suture. Paracymatoceras is like Cymatoceras in conch form but has a more sinuous suture with distinct lobes and saddles. Syrionaiitilus, Anglonautilus, and Eucymatoceras are differentiated on modifications of the ribbing plan. Cymato- nautilus, Heminautiliis, Epicymatoceras (n. gen., p. 439), and Delfocynmtoceras (n. gen. p. 438) are differentiated by their modified conch shapes. Heminautilus and Epicymatoceras have highly compressed conchs and flattened venters. The former is very involute and the whorl section more convergent in plan. The latter is more evolute and with a simple, only slightly sinuous, suture. Deltocy matoceras has a subtriangular whorl section and CymatonautUvs has an evolute conch with a concave venter and wiiorl sides. The only common and geographically-widely -distributed genus is Cymatoceras. This fact is clearly expressed in the number of species belonging to each group. Cymatoceras 64 species Parari/tytatoceras 6 species Heminautilus 5 species Procymatoceras 4 species Eucymoioceras 3 species Anglonautilus 3 species Cymatonautilus 2 species Syrionautilus 1 species Epicymatoceras 1 species Deltocymatoceras 1 species Cymatoceras has a truly world-wide distribution. The other genera are largely confined to Europe in the epicontinental seas KUMMEL: POST-TRIASSIC NAUTBLOID GENEKA 413 adjoiuiiig the Tethyiaii geosyneline (Text-Figs. 21, 22). Eucy- matoceras, Anglonautilus, Syrionaiitilus, Cymatonautilus, Epi- c y mat oc eras, aud Dcltocymatoccras are known only from a belt extending from England to the Crimea or Caucasus Mountains in Kussia in sediments of the Jurassic and Cretaceous epicon- tinental seas of and adjoining the Tethys proper. Procymato- ccras is known from England, Pakistan, and Ethiopia; Hcmi- naufilus and Parucymafoccras are known outside the European Tethyian belt from North and South America, Ethiopia, and Japan. The known distribution of fossil nautiloids is so dependent on factors of preservation and concentration of field collecting that it is probably dangerous to make any too-sweeping conclu- sions. However, the number of species does at least reflect the relative abundance of the various genera. If we think of the Tethys as an ancient circum-equatorial sea the Cymatoceratinae are found to be largely confined to the borders of this seaway. Many genera (e.g. Eucyniatoceras, Syrionnutilus, Cymatona\i- tilus, Epicymafoceras, and Deltocyniatoceras) are confined en- tirely to the European-Middle East portions of this circum- equatorial sea. Analysis of the distribution of the Cymatoceratinae on a time basis throws some light on the general problem of their evolu- tion. Procymatocenis and Cymaionautilus are known only from Middle and Upper Jurassic strata. The species of these genera are confined to the marginal seaways of the Tethys from Eng- land to India. Whereas, to the best of my knowledge, the species of these two genera have never been found in a single stratum and locality, their geographic and time ranges do over- lap. In the Cretaceous, the genus Cyuiatoceras is widely distributed throughout the period. Anglonautilus, Eucyniatoceras, and Heminautilus are confined to the Lower Cretaceous. Anglo- nautilus and Eiicymatoceras have approximately the same geographic range from England to the Black Sea. Whereas Heminautilus also occurs in this European-Mediterranean-belt it is likewise known from Ethiopia, Arkansas (U.S.A.), and Colombia. The Upper Cretaceous genera Epicymatoceras and Deltocymatoceras occur together in the type Maestrichtian area of Belgium, and Syrionautilus is confined to the Middle East. 414 BULIjETIN : MUSEUM OF COMPARATIVE ZOOLOGY KlTMiMEL: POST-TRIASSIC NAUTILOID GENERA 415 o 5 i* ai O o ^ CD -5 a 13 =w C a o be ;: O r <^ 3 «4-l rK M .5 416 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY The wide range in variation of conch shape and suture within the Cymatoceratinae produced many homeomorphs to other members of the Nautilidae. The presence of the characteristic ribbing- makes differentiation easy. The conch shape and suture of Deltocymatoccras are closely similar to that of Angulithes of the Hercoglossinae. The conch shape and suture of many species of Cymatoceras and Eucymatoceras are nearly the same as that found in many species of Eutrcphoceras and Cimomia. The suture of Paracymatoceras is identical in plan with that found in Hercoglossa. This great variability in conch shape and suture leads me to believe that the Cymatoceratinae form a phyletic group which has as its main genetic base the ribbing character. Once estab- lished, this genetic line ■ had its own adaptive radiation which is expressed in the many conch shapes, rib patterns, and sutures as it is in any other group of Mesozoic nautiloids. This subfamily is the dominant group of Cretaceous nautiloids. Ribbing, on a pattern somewhat like that in the Cymatoceratinae, occurred only once before, and that was in a single species of Upper Triassic nautiloid-Proclydonautilvs spirolohus (Dittmar). The ribbing in this species is very fine and like that in Syrionautilus, whereas in most other genera and species of the Cymatoceratinae the ribbing is coarser. It is important to note that Proclydonau- tilus spirolohus belongs to a family that characteristically has smooth eonchs except for a few forms with nodes developed during some stage of their ontogeny {C osmonautilus Hyatt and Smith, 1905, and CaUaionautilus Kieslinger, 1924). The ribbing patterns found in species of Pleuronanttlus are quite different from those of the Cymatoceratinae. Data are insufficient to decipher precisely the evolutionary lines within the Cymatoceratinae. The evidence available seems to point to Cymatoceras as the parent stock which gave rise to most of the remaining adaptive types. The origin of the subfamily must be directly or indirectly among the cenoceratid complex of the Lower and Middle Jurassic. Even though Cyma- tonautilus and Procymatoceras are the oldest forms known, they are both rather specialized and probably not ancestral to Cyma- toceras, which as far as we now know first appears in the Cre- taceous. Neocymatoceras Kobayashi (1954) was established for a single KUMMEL: POST-TRIASSIC NAUTILOID GENERA 417 specimen from the Oligoeene of Kjoisliu, Japan. Neocymatoceras tsukushiense, the type species, is a typical eymatoceratid in conch shape and rib pattern but has a strongly projected suture which forms a rounded ventral saddle, a broad sweeping lobe occupying nearly the whole whorl side, and with what appears to be the beginning of a saddle on the umbilical area. Cymatoceratids with sinuous sutures characterize Paracymatoceras. However, in that genus the saddles and lobes are well defined ; that is, there is always a well-defined saddle on or near the umbilical shoulder. It is also true though that Paracymatoceras is grada- tional with Cyynatoceras in regard to sutures. Those cymato- ceratids with more sinuous sutures show the sutures to be more strongly projected on the ventral area and in this way give the sinuosity ; the lateral saddle at the umbilical area is usually very weak. Cymatoceras pseudonegama Spengler (Fig. 24 C) and C. semilobatus Spengler (Fig. 24 D) have such strongly projected sutures on the umbilical area but not quite to the extent found in the Japanese Oligoeene species. Similar strongly projected sutures in very small juvenile specimens of Cyma- toceras have been described and figured by Kummel (1953a) from Lower Cretaceous formations of Texas. However, the suture of Neocymatoceras tsiikushiense appears to me to lie within the range of variation of the suture of Cymatoceras, and the lack of any lateral saddle in the umbilical area disallows any affinity with Paracymatoceras. The distinctness of the suture in Neocymatoceras tsukushiense is only in the marked projection on the peripheral area, a modification present in several Cre- taceous species. This interpretation places the Japanese species in the genus Cymatoceras, and Neocymatoceras is considered a synonym of Cymatoceras. It should be emphasized that C. tsukushiense is the only Tertiary Cymatoceratinae known. Genus CYMATOCERAS Hyatt, 1884 Cymatoceras Hyatt, 1884, p. 301. Cymatoceras Hyatt, 1894, p. 553. Nautilus (Cyynatoceras) Spengler, 1910, p. 127. Cymatoceras Spath, 1927a, pp. 21, 25. Cymatoceras Miller and Harris, 1945, p. 2. Cymatoceras Durham, 1946, p. 429. Neocymatoceras Kobayashi, 1954a, pp. 18-20. 418 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Type species. Nautilus pseudoelegans d'Orbiguy, 1840, p. 70, pi. 8 (by original designation). Species of Cymatoceras are the most common and widely dis- tributed nautiloids ol' the Cretaceous. The characteristic pat- tern of ribbing and the slightly sinuous septa make identification of the genus easy. A cast of the type specimen (which is in the Museum of Natural History, Paris) is illustrated on Plate 16, figures 1, 2. This same specimen has previously been illustrated by Foord and Crick (1890a, p. 551, figs. 3a, b), and this same figure was reproduced in Foord 's second Catalogue of the Fossil Cephalopoda in the British Museum (Foord, 1891, p. 255, fig. 59 a, b). Foord and Crick (1890a, p. 552) quote a letter from Dr. Paul Fischer expressing doubt as to the identity' of d'Orbigny's type specimen. Fischer states: "The type of Nau- iilus yseudoelegans is difficult to recognize. According to the dimensions given in the original diagnosis (diameter 240 milli- metres, thickness 160 mm. Paleont. Franc, p. 60) our specimens, No. 4834D, which were sent to you are typical. [Foord and Crick, page 552, state that two specimens were sent, both num- bered 4834 D; the larger one is represented by their figure 3.] Moreover, they come from the neighborhood of Vandeuvre Fig. 23. Cross sections of the conch of A, Cynmtoceras clegans (,T Sowerby) from d'Orbigny, 1840, pi. 19, fig. 2, diameter 190 mm.; B, C. sliarpei (SchlUter) 1876; pi. 46, fig. 6, diameter 168 mm.; C, C. tourtiae (Schliiter) 1876, pi. 46, fig. 2, diameter 88 mm.; 1), ('. rvnomunensis (SchlUter), 1876, pi. 45, fig. 2, diameter 115 mm.; E, Pnicj/iiKilnci'rds siih^ truncatus (Morris and Lycett) 1850, pi. 1, fig. 22, diameter 72 mm., the genotype of Procyviatocerns ; F, Frocymatoccras baheri (Morris and Lycett) 1850, pi. 1, fig. la, diameter 82 mm.; G, Faracymafoceras asper (Oppel) from Zittel, 1868, pi. 3, fig. la, diameter 83 mm., the genotype of Faraeyiitatoceias ; II, Frocymnioceras.' intumescens (Waagx-n) 1873, ]il. 3, fig. 3, diameter 90 mm.; I, C. radiatus (Sowerby) from d'Orbigny, 1840, pi. 14, fig. 2, diameter 188 mm.; J, C. neooomiensis (d'Orbigny) 1840, pi. 11, fig. 2, diameter 192 mm.; K, C. deslongchampsianum (d'Orbigny) 1840, pi. 20, fig. 2, diameter 65 mm.; L. Eucymaiocerns requienianus (d'Orbigny) 1840, pi. 10, fig. 2, diameter 280 mm. {=N. pUcatus Fitton, 1835); M, Cymatoceras necTcerianus Pictet, 1859, pi. 16, fig. lb, diameter 93 mm. ; N, Deltocymatoceras rugatus (Fritsch and Schlonbach) 1872, pi. 15, fig. 2, height 80 mm.; 0, Epicymatoceras vaelsensis Binekhorst, 1861, pi. 5c, fig. 2b, diameter 105 mm.; F, Cymatoceras albensis (d'Orbigny) from Pictet, 1859, pi. 17, fig. lb, diameter 95 mm. KUMMEL: POST-TRIASSIC NAUTILOID GKNF.RA 419 Figure 23 420 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (Department de I'Aube), where the species was found. "But the drawing- given by d'Orbigny is faulty; first its dimensions do not agree with the description; then the drawing being reduced to one-third the natural size, the specimen which it represents should have been at least 360 nun. in diameter; besides, the ornaments are wanting in the drawing on a great part of the last whorl. . . . "Consider then the figure by d'Orbigny as only approximate. D'Orbigny has restored a great many of his plates, for which, in my opinion, he is very much to blame. . . . "In the d'Orbigny collection no specimen is specially marked as the type. But d'Archiac (Hist, du progres de la geologic, vol. iv, p. 295) quotes N. pseudoelegans, and he has given to the Museum a specimen marked iype. This specimen is small (greatest diameter 150 mm.), and agrees neither with the figure nor with the dimensions given in d'Orbigny 's original description. "Perhaps the true type was contained in the collection of the geologist Clement Mullet, who showed d'Orbigny over the local- ity where N. pseudoelegans abounds?" In accordance with the above statement Foord and Crick (1890a, p. 551) in the leg- end of their figure 3, state that the drawing is from a specimen in the d'Orbigny Collection in the Museum of Natural History, Paris. However, Foord (1891, p. 255, fig. 59 a, b) labelled the same figure as drawn from d'Orbigny 's type specimen. The type specimen is a much inflated conch, very involute and with a small umbilicus. The whorl section is wider than high with a broadly rounded venter. Whorl sides are also broadly rounded. The test bears prominent ribs that are sig- moidal on the whorl sides and form a sinus on the venter. The suture is only slightly sinuous with shallow ventral and lateral lobes. The siphuncle lies below the center of the septa. There are approximately 64 species of Cymatoceras recorded to date. All of these are characterized by the ribbing and the slightly sinuous septa. The shape of the whorl section is quite variable in this genus (Fig. 23). The type species, C. pseudoele- gans, has a depressed, globular conch ; one of the mo.st com- pressed conchs is that of C. neocomiensis (d'Orbigny) (Fig. 23 J). Between these two extremes there are all gradations of KUMMEL : POST-TRIASSIC NAUTILOID GENERA 421 conch patterns with well rounded venters. Wtereas most of the species are very involute, as in the type, there is a wide range of variation to rather evolute species as C. patens (Kner, 1850) and C. tskaltsithelensis (Rouchadze, 1931). Some of the species develop sharply rounded to subangular shoulders. Cymatoceras tourtiae (Sehliiter) has a subquadrate whorl section with sub- angular ventral and umbilical shoulders ; the venter is low and arched (Fig. 236'). The juvenile volutions, however, have broadly arched venter with well rounded ventral shoulders (Sehliiter, 1876, pi. -±6, fig. 4). Cymatoceras deslongchampsianum (d'Or- bigny) has an arched venter with rounded ventral shoulders but subangular umbilical shoulders with a sloping, slightly arched umbilical wall (Fig. 237i). Nautilus vaelsensis JMnckhorst (1861, p. 15, and Sehliiter 1876, p. 57) is one of the most aberrant cymatoeeratids known to me and is made the type of a new genus (p. 43 9). The conch is very compressed and evolute. The whorl sides are slightly convex with broadly rounded umbilical shoulders and angular ventral shoulders; the venter is flattened. The test bears the typical cymatoceratid ribbing. The characters of the ventral area, that is, the flattened venter and the angular ventral shoulders, are also found in Heminautilus. This latter genus, however, is very involute and the whorl sides are generalh" slightl}^ inflated near the umbilical shoulders. Cymatoceras tourtiae also has the flattened venter and angular shoulders, but the conch is inflated, involute and the whorl section subquadrate. The position of the siphuncle in the type species (C. pseudo- elegans) is below the center — that is, closer to the dorsum than the venter (d'Orbigny, 1840, pis. 8, 9; Foord, 1891, p. 256). It appears that this condition prevails in a majority of the species of Cymatoceras (e.g. C. neocomiensis, C. radiatus, C. gahhi, C. colomhiana, C. elegans, C. kayeanus, C. kossmati, and C. ceno- manensis). The siphuncle lies above the center, that is closer to the venter than the dorsum in C. atlas, C. tenuico status, and C. hunstantonensis. A specimen of C. loricatus (Sehliiter) figured by Griepenkerl (1889) has a near marginal (ventral) siphuncle. The suture of C. pseudoelegans is only slightly sinuous with shallow ventral and lateral lobes. Paracymatoceras differs from Cymatoceras essentially only in the greater sinuosity of its suture (Fig. 24). As noted by Miller and Harris (1945, p. 10) these two genera are more or less gradational in respect to the 422 BULLETIX : MUSEUM OF COMPARATIVE ZOOLOGY suture. Those specimens with more sinuous sutures, and thus transitional to Paracymatoceras, include C. colombiana (Fig. Fig. 24. Diagrammatic representation of sutures of A, Cymatoceras col- onibiana Durham 1946, fig. 3G ; B, Paracymatocerufi texanum (Shumard) 1860, from Miller and Harris 1945, fig. ?.A; C, Cymatoceras pseudonegama (Spengler) 1910, pi. 12, fig. 7d; D, C. semilohatiis (Spcngler) 1910, pi. 11, fig. 4b; E, C. hilJi (Shattuok) from Miller and Harris 1945, fig. 2A ; F, C. kayeanum (Blanford), from Spengler, 1910, pi. 12, fig. 72; G, C. tskaltsithelensis (Eouchadze) 1931, fig. 6; E, C. kos.'), C. kossmati (Fig. 24 H), and C. virgatum (Fig. 24 /). The type species does not have an annular lobe, which ap- l^ears to be the case in the majority of species of Cymafoceras. However annular lobes in the center of the dorsal lobe are present in ('. I'adiafiis, C. loehlichi, C. colomhiana, and C. tour- tiae. Most species of Cymafoceras have broad ventral saddles but some as C virgatum (Spengier) and C. tstxaltsithelensis (Rou- cliadze) have very pronounced ventral lobes. Cymatoceras is the most common, and has a larger number of species than any of the other genera of the Cymatoceratinae. All members of this subfamily have in common the characteristic ribbing. The other genera are differentiated on the basis of spe- cialization of the ribbing, shape of conch, or suture. Frocyjiuifo- ceras has a depressed, rapidly expanding conch. M^ith vhorl sides and venter flattened. Eucymatoceras has the inflated rounded conch as in Cymatoceras pseudoelegans but the ril)s form Y- shaped patterns on the venter and on the whorl sides. Paracyuw- toceras is like Cymatoceras excei)t for the greater sinuosity of the suture which has a well developed lateral lobe and a saddle at the umbilical suture. Cymatonautilus has a depressed evolute conch with concave lateral and ventral areas. Syrionautilus has acute, widely spaced, and nearly straight ribs. Anglonautilus has a depressed whorl section with large transverse undulations on the venter. Hemi)wi(tihis has a much compressed conch, with a flattened venter, and a xevy sinuous suture. Epicymatoceras has a compressed, evolute conch with subquadrate whorl section. Deltocyynatoceras has a subtriangular whorl section. As stated above in the discussion of the subfamily Cymato- ceratinae, Xeocymatoceras Kol)ayashi (19o4a) is thought to fall within the range of variation of Cymatoceras and is placed in .synonymy of that genus. Cymatoceras is characteristic of the Cretaceous. Steuer (1921) has described Cymatoceras perstriatus (Steuer) from yevy late Jurassic and earliest Cretaceous strata of Argentina. Cymato- ceras tsukushicnse (Kobayashi, 1954a i conies from Oligocene strata of Japan. These two are the only non-Cretaceous species of Cymatoceras known to the writer. Cymatoceras is world- wide in distribution (Figs. 21, 22). 424 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY DISTEIBUTION OF SPECIES OF THE GENUS CYMATOCERAS Species Stratigraphic Geographic Distribution Distribution C.alhensis (d'Orbigny) 1850 C. altavensis (Pomel) 1889 C. andranofotsyense Colliguon, 1951 C.anguliferous (Schliiter) 1876 C. articulatus (Pulteney) 1813 C. atlas (Whiteaves) 1876 C? averilli (Andersou) 1938 r. bayfieldi (Foord and Crick) 1890 C. hifurcatum (Ooster) 1858 C. broitzemenfiis (Miiller and Wollenunui) 1906 CJcampbelli (Meek) 1861 C. carlottensis (Whiteaves) 1900 C. cenomamensis (Schliiter) 1876 C. colombiana Durham, 1946 C. compressns (Tavani) 1942 C. crehricostatus (Blanford) 1861 C. deslongchampsianum (d'Orbigny) 1840 C. eichwaldi (Karakasch) 1907 C. e.legans (J. Sowerby) 1816 C.elegantoides (d'Orbigny) 1840 Cretaceous England, France Cretaceous Algeria (Neoeomian) Cretaceous Madagascar (Albian) Cretaceous Germany (Cenomauian) Cretaceous England Cretaceous England, (Upper) France Cretaceous California (Lower) (U.S.A.) Cretaceous England (Senonian) Cretaceous Switzerland Cretaceous Germany (Senonian) Cretaceous Vancouver Island, B.C. Cretaceous Maple (Upper) Island, B.C. Cretaceous Europe (Cenomanian) Cretaceous Colombia (Aptian) Oi-etaceous Somaliland (Albian) Cretaceous India (Albian) Cretaceous Europe (Neocomian- Cenomanian) Cretaceous Crimea (Lower) Cretaceous England (Cenomanian) Cretaceous England, (Cenomanian) France KUMMEL : POST-TRIASSIC NAUTELOLD GENERA 425 Species Stratigraphic Geographic Distribution Distribution C. farmosus (Blanford) 1861 G.gabhi (Anderson) 1902 C. guilielmitelli (Ooster) 1858 Chilli (Shattuck) 1903 C. hunstantonensis (Foord and Crick) 1890 C. huxleyanus (Blanford) 1861 C. imhricatus (Crick) 1907 C. interstriaius (Strombeek) 1863 C. kayeanus (Blanford) 1861 C.kossmati (Spengler) 1910 C. loeblichi (Miller and Harris) 1945 C. loricatus (Sehliiter) 1876 C. madagascarensis (Yabe and Shimizu) 1924 C. manuanensi^ (Crick) 1907 C. mikado (Krenkel) 1910 C. nebrascense (Meek and Hayden) 1862 C. nedkerianus (Pietet) 1847 C. negama (Blanford") 1861 C . neocomiensis (d'Orbigny) 1840 C . neohispanicum (Burckhardt) 1925 Cretaceous India, (Senonian) Madagascar Cretaceous California (Lower) (U.S.A.) Cretaceous Switzerland Cretaceous Texas (Lower) (U.S.A.) Cretaceous England (Cenomanian) Cretaceous South India, (Turonian) Madagascar Cretaceous Natal, Africa Cretaceous Germany (Upper) Cretaceous India, (Albian- Madagascar Cenomanian- Turonian) Cretaceous India (Albian) Cretaceous Texas (Lower) (U.S.A.) Cretaceous Germany (Senonian) Cretaceous Madagascar (Upper) Cretaceous Natal, (Albian) Africa Cretaceous Tanganyika (Xeocomian) Cretaceous Montana (Upper) (U.S.A.) Cretaceous Europe (Aptian, Albian) Cretaceous India (Albian) Cretaceous Europe, (Lower) Crimea Cretaceous Mexico (Aptian) 426 BULLETIN : MUSELTM OF COMPARATIVE ZOOLOGY Species Stratigiaphic Geographic Distiibutiou Distribution C. occlusus Crick, 1907 C. patens (Kner) 1850 C. perstriatus (Steiier) 1921 C. pseudoathu (Yabe and Sliiniizir) lOill C. psetidoelegans (d'Orbignyj 1840 C. psmidonegama Spengler, 1910 C. pseiuloneokomienae Shiniizii. 19ol C.radiatus (J. Sowerby) 18'J2 C. sal'cdavtim Colliguon, 1949 C. saussvreanus (Pictet) 1847 C. scecqureniiis (Tavani) 19412 C. semilohaius Spengler, 1910 C. snniiindatus (Foord) 1891 C.sliarpei (Schliiter) 1876 C' sinuato-pUcatiis (Geiuitz) 1843 C.i^traticostatu^ (Crick) 1907 C.sucientie (Whiteaves) 1879 C. tenvicostatiis (Schliiter) 1870 C. tourtiac (Schliiter; 1876 C. Ukalt.siihrJensis (Rouchadze) 1931 C. tsulcushiensc (Kobayashi; 1954 C. iindulnfifnrmis Spath, 1927 Cretaceous Zululand < 'retaceous Poland, ( Senonian ) Germany Jurassic ( Upper ) Argentina Cretaceous (Lower) Cretaceous Japan (Upper) Cretaceous Europe Cretaceous India (Albian) Cretaceous Japan (Aptian) Cretaceous Europe ( 'retaceous Madagascar (Albian) Cretaceous R^ritzerland (Albian) Cretaceous Sonialiland (Albian) Cretaceous India (Albian) Cretaceous England (Cenouianian) Cretaceous Germany (Cenonianian) Cretaceous Germany (Upper) Cretaceous Natal, Africa Cretaceous Sucia Island (Upper) (iroup, B. C. Cretaceous (Germany (Cenomaniairi Cretaceous (iermauy (Cenonianian) Cretaceous Eussia (Aptian) Oligoceiie Japan Cretaceous Tanganyika (Xeocomian) KUMMEL : POST-TRIASSIC NAUTILOID GENERA 427 Species Stratigiaphic Distribution Geographic Distribution C.ventroplicatus (Foord) 1891 C. virgatus (Spengler) 1910 C. woodsi (v. Hoepen) 1921 Cretaceous England (Cenomanian) Cretaceous India (Albian") Cretaceous South Africa (Upper) 1927 Genus PaRACYMATOCEKAS Spatli, Paniciimatoceras Spath, 1927a, pp. 21, 25. Paracymatoceras Miller and Harris, 1945, p. 9. Paracymatoccras Kummel, 19o3a, p. 4. Tirpe species. Nautilus aspcr (Oppel) Zittel, 1868, pi. 3, lig. 1 (by original designation). Illustrated on Plate 19, figures 1, 2 and Text Figure 23(t, of this report. Most of the other genera of the Cymatoceratinae are dis- tinguished from Cymotoceras on differentiations of the conch form or sculpture. Paracymatoceras differs from Cyrnatoceras essentially only in the greater elaboration of the suture. It is also generally agreed that in respect to the .suture Cymatoceras and Paracymatoceras intergrade (Fig. 24). The conch form, degree of involution, and shell sculpture agree perfectly with Cymato- ceras. The type species — P. as per — which is of Upper Jurassic age, has a rounded ventral groove on the most adoral part of the living chamber (Pig. 236'). This feature is not present in any of the five Cretaceous species of Paracymatoceras known to date. Aside from the ventral groove in the type species, the conch form in the other species assignable to Paracymatoceras are completely similar to or within the i-ange of variation of Cymatoceras. Six species can be assigned to this genus i-anging in age from Upper Jurassic through the Cretaceous. Four of the species are Lower Cretaceous in age. The genus does not appear to be a com- mon element in nautiloid faunas but it has been recorded from India, Europe, Texas, and Mexico (Figs. 21. 22). Some juvenile forms from the Lower Cretaceous PawpaAv formation of Texas with perfectly smooth conchs have been recorded by Kummel (1953a). These specimens are of interest in emphasizing the difficulties in identifying weakly-ribbed Paracymatoceras and differentiating these from Hercoglossa. 428 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY DISTRIBUTION OF SPECIES OF THE GENUS PAEACYMATOCERAS Species Stratigraphic Geographic Distribution Distribution Jurassic Europe (Upper) Cretaceous Crimea (Lower) Cretaceous Durango, (Aptian) Mexico Cretaceous South India, (Albian, Madagascar Maestrichtian) Cretaceous Texas ( Lower ) (U.S.A.) Cretaceous South (Senonian) India; East Indies? P.asper (Oppel) 1865 P . karpinskyi (Karakasch) 1907 P.? milleri Humphrey, 1949 P. rota (Blanford) 1861 P. texanum (Shumard) 1860 P.trichinopolitensis (Blanford) 1861 Genus PeOCYMATOCERAS Spath, 1927 Procymatoceras Spath, 1927a, pp. 21, 25. Procymatoceras Miller and Harris, 1945, p. 11. Type species. Nautilus stibtruncatus Morris and Lycett, 1850, pi. 1, fig. 2 (by original designation). The genotype described and figured by Morris and Lycett does not seem to be available. Several topotype specimens are in the collections of the British Museum (Natural History) one of which is illustrated on Plate 17 and the cross section of the same specimen in Text Figure 25. From a study of topotype specimens and the original description of the tj'pe species the following diagnosis of Procymatoceras can be made. The conch is robust, rapidly expanding, very involute to occluded. Whorl section generally wider than high. On the living chamber the venter and whorl sides are flattened. The ventral and umbilical shoulders are well rounded. The whorl sides converge towards the venter. The earlier volutions are more rounded in outline. The surface of the conch bears sinuous ribs that form a broad sinus on the venter. The ribs appear to be fasciculate growth lines and may not be homologous with those of typical Cymatoceras. The ribbing is most distinct on the living chamber. The sutures are only slightly sinuous with shallow ventral and dorsal lobes. KUMMEL: POST-TRIASSIC NAUTILOID GENERA 429 When Spath (1927a) first established Procymatoceras he in- cluded only the type species and P. haheri (Morris and Lycett) both from the Middle Jurassic of Great Britain. The topotype specimen of P. subtruncatus illustrated on Plate 17, and on Text Fig. 25. Procymatoceras subtruncatus (Morris and Lycett). Diagram- matic cross section of topotype represented by figures 1, 2, Plate 17, X 0.50. Figure 25, clearly shows the robust involute character of the conch and the subtrapezoidal outline of the whorl section. Neither the suture nor the siphuncle are visible. Nautilus in- tumescens Waagen has been referred with question to Procyma- toceras. I agree with Spath (1927a, p. 33) that it is doubtful whether Noetling's (1896) specimens from Baluchistan assigned to N. intumescens Waagen are correctly identified. The only 4:JU UULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY other species that appears to belong to this genus is Paraceno- ceras cosiatum Scott (1943). Genus CymATONAUTILUS Spath, 1927 Cymatonautdiis Hpath, 1927a, p. l21. Cymatonautilus Miller and Harris, 1945, p. 9. Type species. Nautilus julii (Baiigier MS) d 'Orbigny, 1850, p. 328 (by original designation). The type species of this unusual Juj-assie cymatoceratid was originally briefly diagnosed but not illustrated by d'Orbigny (1850, p. 328) who gave Baugier the credit as author of this species. However, Baugier 's name apparently was only a manu- script name, and was never published by him. The first com- plete description and illustration of Nautilus jidii was published by Gueranger (1865, p. 189, pi. 3). Gueranger's figures arc reproduced here on Plate 18, figui'cs 1-4. This genus can be diagnosed as follows : Conch widely umbili- cate, robust, whorls subquadratic, being slightly wider than high. The whorl sides are flattened and with a broad lateral groove. The venter is likewise flattened and with a median groove. The conch bears sinuous ribs that ol:)liquely cross the suture and form a deep ventral sinus. The suture forms a shal- low ventral lobe and a broad concave lateral lobe. The siphuncle is subcentral in position, being closer to the dorsum than the venter. The type specimen came from strata of Callovian age in France. Petitclerc (1926) has described and illustrated a specimen of the same species from strata of Bathonian age in France. Nautilus mojsisoric.n Neumayr (1870) from Upper Jurassic strata of Austria most probably also belongs in Cyma- tonautiJus. The most distinctive- features of this genus are the wide umbilicus and the grooved venter and whorl sides. Most species of the Cymatoceratinae are involute forms. Genus AXGLONAUTILUS Spath, 1927 Anglonautilus Spath, 1927a, pp. 21, 25. Anglonautilus Miller and Harris, 1945, p. 8. Type species. Nautilus undulatus J. Sowerby, 1813, pi. 40, upper figure (by original designation). KUMMEL : POST-TRIASSIC NAUTILOID GENERA 431 Sowerby's type specimen does not appear to be in the British Museum, but the specimen referred to by Foord (1891, p. 245) labelled ^'Nautilus uiidulatus, M. C. 182" in Sowerby's hand- writing is figured here, Plate 20, figures 1, 2, and Text Figure 26. This specimen is also numbered B.M.N.H. 66336. Dr. L. F. Spath kindly informs me that this specimen is unlocalized, but Fig. 26. Anglonautilus undulatus (Sowerby). Diagrammatic cross sec- tion of adoral part of living chamber of topotype represented by figures 1, 2, Plate 20, X 1. undoubtedly is a topotype from Nutfield, Surrey. This specimen is incomplete consisting only of two camerae of phragmocone and one-third whorl of living chamber. However, it shows clearly the most diagnostic feature of the genus, that is the large, fold- like undulations which are most prominent on the venter and decrease rapidlj^ on the flanks. The suture forms a very shallow ventral lobe and a l)road shallow lateral lobe. The siphuncle is subcentral in position being nearer the dorsum than the venter. Tn addition to the type species which is known from Aptian to Cenomanian strata in England and various localities in Europe {]ide Foord, 1891, p. 244), Nautilus suhalbensis SinzoA\ (1913) and N. hegudensis Kilian and Reboul (1915) should be included in Anglonautilus. The former species is from Albian strata in the Crimea and the latter from Hauterivian strata in southern France. Genus EuCYMATOCERAS Spath, 1927 Eucymatoceras Spath, 1927a, pp. 21, 22, 25. 432 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY E uoymatoceras Miller and Harris, 1945, p. 9. Type species. Nautilus plicatus Fitton, 1835, p. 129 (by original designa- tion). The type specimen was not available for study but several conspecific forms are in the British Museum. One of these speci- mens from the Lower Greensand, Atherfield, Isle of Wight (B.M.N.H. C335) is illustrated on Plate 21, and the cross section Fig. 27. Eucymatoceras plicatus (Fitton). Diagrammatic cross section of plesiotype represented by figures 1, 2, Plate 21, X 1. in Text Figure 27. The only previous illustrations of the type specimen are the original woodcut in Fitton (1835, p. 129) and figures in d'Orbigny (1840, pi. 10; Nautilus requienianus d'Orbigny 1840 =N. plicatus Fitton 1835) and in Uhlig (1883. pi. 3). KUMMEL : POST-TRIASSIC NAUTILOID GENERA 433 The genus can be diagnosed as follows: Conch subglobular, involute, umbilicus closed, whorl section rounded, broader than high. Suture only slightly sinuous. The siphuncle is small and subcentral. Surface of conch with prominent ribs that form a deep angular V-shaped ventral sinus and on the whorl sides a similar but asymmetrical salient. The most diagnostic feature is the V-shaped pattern of the ribs by which it can easily be distinguished from any other genus of the Cymatoceratinae. At present only three species of Eucymafoceras have been recorded, namely N. plicatus Fitton, the type species, and N. sieveni Karakasch (1907, p. 30, pi. 2, fig. 13; pi. 8, fig. 12) from Lower Cretaceous formations of the Crimea; and N. stschiirouskii Milaschevitch, (1877, p. 125, pi. 1, figs. 11, 11a) from the Lower Cretaceous of Russia. All of these species have the depressed, subglobular, rounded conchs. From the great variability in conch form in Cymatoceras it is doubtful if the conch form in the known species of Eucymatoceras can be considered a diagnostic character. Genus SyrIOXAUTILUS Spath, 1927 Syrionautilus Spath, 1927a, pp. 21, 25. SyrionautUus Miller and Harris, 1945, p. 11. SyrionautUus Aynimelech, 1946, pp. 523-528; 1947a, p. 690. Type species. Nautilus lihanoticus Foord and Crick, 1890, p. 404, fig. 6 (by original designation). All the specimens of Nautilus libanoticus in the British Mu- seum (Natural Histor}^) upon which Spath established his genus SyrionautUus are more or less crushed and distorted. One of the syntypes with an attached mandible is figured here on Plate 7, figure 3 (previously figured by Foord, 1891, p. 371, fig. 82.). The general form of the conch is like that of Cymato- ceras but the character of the ribbing is quite distinct. In Syrio- nautilus the ribs are acute and separated by interspaces rather exceeding their own width. The type of ribbing is very similar to that of Proclydonautilus spirolohus of the Upper Triassic. The form and spacing of the ribs in SyrionautUus are the dis- tinguishing characters that serve to separate it from allied genera of the Cymatoceratinae. The position of the siphuncle and nature of the suture are not known. The type and only 434 BULLETIN : ilLWEUM OF COMPARATIVE ZOOLOGY species so far referable to this genus comes from Senonian formation of Syria. Recenth' Avnimelecli (1946, 1947a) has published two short notes on additional specimens of Syrionaiitilus lihanoficus from Palestine. As with the types, this new material is also incom- plete and generally crushed. However, the best and nearly com- plete specimen was iigured by Avnimelech (1947a, pi. 15). Even though this figured specimen is also compressed it clearly shows Syrio7iaufilus to have an involute, compressed conch and arched venter. In other words, it has a typical cymatoceratid conch but differs in the character of the ribs. Genus He^^LIXAUTILVS Spath, 1927 Heminautilus Spath, 1927a, pp. 22, 25. Vorticoceras Seott, 1940, pp. 1074. 1075. Heminautilus Scott 1943, p. 89. Heminautilus Durham, 1946, ii. 432. Plati/tiautihis Yabe and Ozaki, 1953, pp. 55-61. Tijpe species. Nautilus saxbii Morris, 1848, p. 106 (woodcut) ; PI. 10, figs. 1, 2 of this report (by original designation). When Spath established this genus he designated Nautilus saxhii Morris (1848, p. 106, woodcut) as genotype and thought this species was the same as X. JaUierimuts d'Orbigny (1841). The latter species was first very In-iefl}' diagnosed but not illus- trated by d'Orbigny in 1941 in the Revue Cuvierienne (p. 318). and was listed the same year in his Paleontologie Francaise. Terrains Cretaces (p. 620). Nautilus laUierianus was again cited by d'Orbigny in his Prodrome (1850, t. 2, p. 112), and N. saxhii was listed as a synonym. In none of these publications is there an illustration and only the first (1841, p. 318) has a very brief ddagnosis. A. laUierianus was next described and illustrated from Switzerland by Pietet and Campiche (1859). Douville (1916) gave a very complete description with several figures of specimens of A. laUierianus from Aptian strata east of Suez. Douville likewise placed A. saxhii in synonymy of A. laUierianus. Foord (1891) recognized both of these species. The type specimen of Nautilus saxhii is illustrated on Plate 10, figures 1, 2, and the cross section in Text Figure 28. The only previous illustrations of this specimen are the woodcuts in Morris (1848. p. 106) and Foord (1891, p. 311, fig. 68). Since KUMMEL: POST-TRIASSIC NAUTILOID GENERA 435 d'Orbigny's N. lallierianus has not been illustrated, and tho description is very brief, it is impossible to tell whether it is eonspecifie with N. saxbii. Nautilus lallierianus, Donville (1916) is sufficiently distinct from N. saxhii Morris in the form of the septa and ornamentation of the conch to keep these two species separate. The plaster casts of the specimen of N. lallieri- anus figured by Pictet and Campiche (1859, pi. 19, figs. 6a, 6c) are in the Museum of Comparative Zoology. These show no surface ornamentation whatsoever. The tricarinate character of the venter is not as well preserved as the figures by Pictet and Campiche would lead one to believe. The genus TIeniinaiitilus can be diagnosed as follows: Conch involute, very compressed, whorls much higher than wide. Whorl sides flattened and strongly converge towards a narrow, flattened or slightly concave venter. The ventral shoulders are angular. The venter on the earlier volutions is rounded. The umbilical shoulders are broadly rounded. Greatest width of the conch is just ventral of the umbilical shoulders. The suture is rather sinuous with a ventral lobe, a subangular saddle on the ventral shoulder, broad deep lateral lobe and a narrowly rounded saddle on the UDibilical shoulder. The si})huncle is subcentral in posi- tion lying closer to the dorsum than the venter. The conch bears sinuous ribs that curve strongly aborally over the ventral half of the whorl side. The ribbing may be very weak and in fact is absent on the type of Nautilus saxhii (which is probably a juvenile). Spath originally placed Heminautilus in his Paracenoceratidae with question and made note of the possibility that it more properly belonged in his Cymatoceratidae. The Paracenocera- tinae are characterized by differentiation of the periphery, gen- erally associated with increase in siiniosity of the suture lim-. and a ventral lobe (S])ath, 1927a, p. 25). Plowever, the sutures of Heminautilus are much more sinuous (Fig. 20). The Para- cenoceratinae have smooth shells, devoid of ribbing. Tithoiio- ceras, and Aulaconautilus have longitudinal ribs or carinae on the ventral area but no sculpture on the flanks. Sinuous cyma- toceratid-like ribbing is characteristic of Heminautilus. The ribbing is well preserved on the specimens of H. lallierianus fig- ured by Douville (1916, pi. 17, figures 2-6) and on //. rangei Hoppe (1922, pi. 4. figures 2, 8). Among the specimens illus- 436 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY trated by Douville there is some variability in the strength and patterns of the ribs. On the earlier volutions the ribs are low, broad, and rounded and decrease toAvards the venter. The mature specimen of Douville 's (1916) figure 5, plate 17, has a smooth outer volution, whereas the larger specimen of figure 6a, b, shows prominent ribs that are very broad and separated by narrow concave interspaces. In Heminautilus etheringtoni Dur- ham (1946) the ribs are most prominent on the living chamber Fig. 28. Hemiimutilus saxbii (Morris). Diagrammatic cross section of holotype represented by figures 1, 2, Plate 10, X 1. which is about one-third of a volution in length. Topotype speci- mens of Heminautilus etheringtoni kindly loaned to me by J. Wyatt Durham show rather distinct ribbing on early volutions, with a single large topotype smooth except for prominent growtli KUMMEL: POST-TRIASSIC NAUTILOID GENERA 437 lines. A crushed and weathered specimen, not referred to by Durham (1946), from near the top of the basal Velez Sandstone measures approximately 145 mm. in diameter and has very prominent "cymatoceratid" ribs. The ventral area likewise appears to have a median longitudinal rib. With the material available it cannot be determined whether or not this specimen is conspecific with //. ethcringtoni. They do not come from the same horizon within the Aptian. No data are available on the degree and extent of variation in the sculpture of this or most species of nautiloids. It is very rare that really representative populations can be collected. Hemiriaufilns tijosiensis (Yabe and Ozaki) 1953, shows prom- inent growth lines on tiie venter that form a deep V-shaped I)attern. On the lateral whorl sides low radial folds with broader intervals are present on the type (and only) specimen of this species. These authors proposed a new genus - — Platynautilus — for their new species, but it is clearly a representative of Heminautilus and thus Platynautilus is placed in synonymy. Heminautilus stantoni (Scott) has broad, low, rounded and widely spaced ribs reaching their greatest prominence on the middle of the flanks. Heminautilus saxhii, the type specimen which is figured here on Plate 10, figures 1, 2, is perfectly smooth but is probably not a mature specimen. The suture in most species of the Cymatoceratinae are only very slightly sinuous. Paracymatoceras has a well-developed, deep, lateral lobe followed by a saddle on the umbilical shoulder. The suture across the venter generally develops a ventral saddle, not a lobe. Cyinatonautilus julii has a deep ventral lobe, and a broad lateral lobe with the beginnings of a saddle on the umbili- cal region. "Nautilus" vaelsensis (Binckhorst, 1861) is a very compressed form with a subrectangular whorl section and a wide, shallow umbilicus. The truncate venter and compressed conch are very reminiscent of Heminautilus except in the dif- ference of involution. The suture of ''Nautilus" vaelsensis has a very shallow ventral lobe and a broad, shallow, lateral lobe. The six known species of Heminautilus are from Lower Cre- taceous formations in Arkansas, England, France, Switzerland, Egypt, Palestine, Colombia, and Japan. 438 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Stratigraphic Geographic Distribution Distribution Cretaceous Colombia (Aptian) Cretaceous Europe. (Xeocomian) EgJTt, Ethiopia Cretaceous Palestine (Albian) Cretaceous England, (Aptian) France Cretaceous Arkansas (Aptian) (U.S.A.) Cretaceous Japan (Aptian?) DISTEIBUTION OF SPECIES OF THE GENUS HEISnNAUTILUS Species S. etheringtoni Durham, 1946 H.lallierianus (d'Orbigny) 1840 E.rangei (Hoppe) 1922 H. saxbii (Morris) 1848 E.stantoni (Scott) 1940 H. tyosiensis (Yabe and Ozaki) 195.3 Genus DeLTOCYMATOCERAS n. gen. Type species. Nautilus leiotropis Schliiter, 187G, p. 175, pi. 48, figs. 1, 2 (Plate 22, figures 1, 2 of this paper). The adaptive radiation within the Cymatoceratinae is ex- pressed in patterns that are homeomorphous with evolutionarj- trends in other phyletic lines within the Nautilidae. Modification of rib patterns, conch shape, and suture are the more obvious characters which express evolution within the subfamily. The present genus is established for those cymatoceratids with a more or less triangular outline of the whorl section. The type species- — Nautilus leiotropis, and N. rugatus (Fritsch and Schlonbach) are the only species known to date that can be placed in this genus. The triangular outline of the whorl sectioji serves readily to differentiate this genus from all others Avithin the Cymatoceratinae. The type species has a large, inflated, involute conch with broadly arched and strongly convergent whorl sides. The venter is marked by a slight rounded, smooth, keel-like ridge. There are no distinct ventral shoulders, the ventral region grading evenly onto the flanks. The flanks have prominent cymatocera- tid ribs that bifurcate near the middle of the whorl sides. The ribs do not cross the keel-like ridge on the venter. The suture forms a narrow, somewhat pointed, ventral saddle, a broad, KUMMEL : POST-TRIASSIC NAUTILOID GENERA 439 leather shallow lateral lobe, and a prominent saddle on the um- bilical shoulder. The position of the siphuncle is not known. The distinctive subtriangular whorl section with the smooth keel-like ridge on the venter set Deltocymatoceras leioiropis apart from all other eymatoeeratids. In conch form and suture pattern Ddlocymatoceras is homeomorphous with AnguUihcs of the Hercoglossinae, being distinguished bv its ril)bing. Aiigu- Jithes is most probably a development of Cimomia, a form with a more rounded whorl shape. Deltocymatoceras is best visualized as a sagittate derivative of Cjfmatoceras which show's much vari- ation in conch shape. The type species of Deltocymatoceras comes from Upper Cre- taceous (Emseher-Mergel) of German}-. The original is in the Museum of the University of Bonn {fide Schliiter, 1876, p. 175). Deltocymatoceras rugeitiis comes from Tapper Cretaceous strata of Bohemia. Genus EpirYMATOCER.AS n. gen. Ti/pe species. Xaiitilus vaelsensis Bmckhor.st, 1861, p. 15, pi. 5c, figs. 2a, b, c This new genus is established for a species of Upper Creta- ceous nautiloid that has a most unusual conch shape for post- Triassic nautiloids. The species has been described and figured by Binckhor.st (1861, p. 15, pi. 5c, figs. 2a, b, c) and Schliiter (1876. p. 177, pi. 51, figs. 3-4). Schliiter 's figure is reproduced here on Plate 23, figs. 1, 2. Xautilns raclscvsis has an evolute and greatly compressed conch. The umbilical diameter is approximately 25 per cent of the diameter of the conch. The whorl section is subquadrate, being nearly twice as high as wide. The ventral shoulders are angular, the venter narrow and flattened. The lateral flanks are only slightly inflated. The umbilical shoulders are broadly arched. Tlu^ suture inscribes a very shallow ventral lobe and a shallow lateral lobe. The position of the siphuncle is not known. The conch bears fine sinuous ribs that curve backwards toward the ventral shoulder and form a slight sinus on the venter. Epicymatoceras is distinguished from all other genera of the Cymatoceratinae by its compressed, evolute conch and flattened venter. It is the ribbing which attaches this genus to the Cyma- 44U BFLLETIX : MFSEUif OF COMPARATRT: ZOOLOGY toceratinae. Among the Cj^matoceratinae the only group some- what similar is Heminautilus which is likewise compressed in conch form and with a flattened venter. However, in Heminau- tilus the whorl section is much more inflated near the umbilical shoulders and the flanks more convergent. Likewise this genus is very involute and with a more sinuous suture. Most of the remaining genera and species of the Cymatoceratinae have more involute and inflated conchs. The presence of cymatoceratid ribbing readily differentiates Epicymatoceras from other subfamilies of the Nautilidae. There is some similarity in conch shape of Epicymatoceras with that of Pseuclocenoceras of the Nautilinae. However, in that genus the smooth conch has a well rounded ventral area ,steep umbilical shoulders, and is more involute. The type and only species of this genus has been identified from Upper Cretaceous (Maestrichtiau) strata at Vaels near Achen (Germany) and at Limbourg (Belgium) just south of Achen. Subfamily HERCOGLOSSINAE Spath, 1927 Spath (1927a) originally erected this family to include the following genera: Pseiidaganides Spath of the Jurassic and lowermost Cretaceous; Hercoglossa Conrad, Upper Cretaceous to Eocene; Hercoglossoceras Spath, Liassic ; Pseudonautilus Meek, Upper Jurassic; Paraturia Spath {=Aturoidea Vredenburgj, Cretaceous to Eocene ; and Deltoidonautilus Spath, Eocene. Later he (Spath 1927b) included Cimomia Conrad, and Stenzel (1940) added his genus Woodringia. Restudy of this group alters the above interpretation in many important details. The Jurassic hercoglossids, namel}' Pseudaganides and Pseudonau- tilus are interpreted as a distinct phjdetic line from the Cre- taceous-Eocene trend involving Hercoglossa itself (see page 389). Hercoglossoceras is considered to be a synonj-m of Pseudaganides. It is merely an early, highly' compressed ''agauitiei" and is part of the radiation from the cenoceratid complex which went in for highly modified sutures (Pseudagauitinae). Adaptation to- wards modified (highly sinuous) sutures is a trend that has occurred many times in the evolutionary history of the nautil- oids. Within the Upper Paleozoic few nautiloids are known with KUMMEL : POST-TRIASSIC NAUTILOID GENERA 441 highly sinuous sutures. The most extreme development is that found in Permoceras hitauniensis (Haniel) from strata of Leonard age in Timor. Grypoceras {Plummeroceras) plummeri KuDimel (1953c) of the Grypoceratidae from beds of Leonard age in Texas has a deep ventral lobe and lateral lobe. When Miller and Collinson (1953) established the genus Permoceras thev also erected the familv Permoceratidae for their monotypic genus. It appears more likely that Permoceras is an aberrant development from the Domatoceras-Grypoceras evolutionary line and should be included in the Grypoceratidae. In the Triassic, the first large-scale adaptation towards highly modified sutures appears in the Clydonautilidae, Gonionautili- dae, and Siberionautilidae all of the Upper Triassic. The sutural patterns seen in these groups are closely similar to those of the Pseudaganides-Pseudonautilus trend of Jurassic age and to the Hcrcoglossa-Aturoidea trend of Cretaceous-Eocene age. The details have been discussed previously (p. 389) and the evidence seems quite strong that in the Mesozoic there were three separate radiations built largely around modification of the suture, the first in the Triassic, the second in the Jurassic, and the third in the Cretaceous. A similar adaptive trend also took place in the Cymatoceratinae. Since the basic sutural type must surely be an adaptive character it is no more than to be expected that repetition of similar forms will be produced. For the aboA'e reasons Pseudaganides and Pseudonautilus are considered a separate phyletic development and placed in the subfamily Pseudaganidinae. Deltoidonautilus Spath is considered to be a synonym of Aii- gidithes Montfort (Kuramel 1953a) ). Spath placed AnguUthes in the Nautilidae but it is here considered part of the Hercoglossi- nae. Woodringia Stenzel (1940) is placed as a synonym of Herco- glossa since the character upon which it was distinguished — a slight ventral lobe in the ventral saddle — is of no more than specific importance in other genera of this subfamily. Thus, as now interpreted, the Hercoglossinae include the fol- lowing genera: Hercoglossa Conrad, Cimomia Conrad, Angu- Uthes Montfort, and Aturoidea Vredenburg. These genera are all characterized by involute, smooth, conchs with modified sutures. As has been shown by Miller (1947, 1949) on scA^eral occasions this group of genera appears to form a linear series BULLETIN: ML'SEUM OF COMPARATIVE ZOOLOGY KUMMEL : PO^^T-TRIASSK XAUTILOID GENERA 443 444 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of increasing sinuosity in the suture ranging from Cimomia to Hercoglossa to Aturoidea. Angulitlies is probably a special de- velopment out of Cimomia maintaining an approximate cimomid tj'pe of suture but modifying the whorl section to a subtriangular shape rather than rounded. Cimomia is readily derived from Eutreplioceras with which it is more or less gradatioual. Since the genera of this subfamily (with the exception of Angulitlies) are separated on the basis of increasing sinuosity of the suture, it is understandable that there should l)e gradational forms. These gradational forms are found especially between Entrepho- ceras and Cimomia and between Cimomia and Hercoglossa. It has been generally assumed that each of these units had a single time of origin and thereafter were distinct genetic stocks (Miller. 1949). An alternative interpretation is, however, jDOssible; that is, that there were several times in the Cretaceous and Tertiary when radiations towards sinuosity of the suture oc- curred and what is normally included in the genus Cimomia is a group of transitional species and not part of a single linear trend. The kind and nature of the data available, however, is not sufficient to test this suggestion. The eutrephoceratid-evolv- ing stock is the parent stock of several radiations (generic groups) of rather diverse morphology. The fact that the genetic potential of the group shows variability in conch shape and suture, is no reason why any number of these variants in suture could not become in themselves geneticalh' distinct, to become part of the species complex included in Cimomia. All four of these genera occur in the Cretaceous and range into the Eocene or Oligocene. Cimomia probably includes some Upper Jurassic species. In numbers of species, Hercoglossa has 21, Cimomia 44, Angulithes 32, and Aturoidea 11. The relatively large number of species is largely due to ''monographic" highs, especially in Tertiary faunas. Geographically, species of this subfamily are vevy wide- spread. The distribution of the Cretaceous species of the sub- family is illustrated on Figure 29, those of the Tertiary on Figure 30. The maps show very well the general overlap in range of all of the genera. None of the forms is sufficiently com- mon at any of the localities indicated so that collections could be made for detailed population studies. KUMMEL : POST-TRIASSIC NAUTILOID GENERA 445 Genus Ci:\rOMIA Conrad, 1866 Cim<»nia Conrad, 1866a, p. 102. Cymomia Conrad, 1866b, p. 19. Cymmomea Conrad, 1868, j). 732. Cymomia Conrad, 1868, p. 732. Cimomia Spath, 1927b, pp. 424-428. Nautilus ( J avanoceras ) Martin, 1932, pp. 1, 2. Cimomia Miller and Thompson, 1933, pp. 305-313. Cimomia Stenzel, 1940, pp. 751-753. Cimomia, Miller, 1947, pp. 39-49. Cimomia Miller, 1951, pp. 38-47. Cimoviia Haas and Miller, 1952^ pp. 329-338. Type species. Naidilus hurtini Galeotti, 1837, from Eocene of Belgium, by original designation. Figured here on Plate 24. Most of the species of this genus are Tertiary in age (Paleo- cene — Lower Oligocene) and the Tertiary species have in recent years been thorouglhly discussed by Miller and Thompson '(1933), Stenzel (1940)', Miller (1947), Miller (1951), and Haas and Miller (1952). The very able diagnosis of the genus by Miller (1947, pp. 39-40) is quoted here: "Conch subglobular to subdiseoidal in shape and nautiliconic in its mode of growth. AYhorls broadly rounded laterallj^ rounded veutrally, and im- pressed dorsally. Umbilicus small : umbilical shoulders low and broadly rounded. Surface of conch marked by growth lines but does not bear ribs or nodes. Septa numerous and each suture forms a broad very shallow broadly rounded or slightly retuse ventral saddle, a broad shallow broadly rounded lateral lobe, a narrower higher rounded lateral saddle (located near the umbili- cal shoulder), a broad rounded lobe on the umbilical wall, and a fairly prominent internal lateral saddle that extends to a broad rounded dorsal lobe. Siphuncle small, circular in cross section, and orthochoanitic in structure ; its position varies con- siderably in different species, but in no case is it marginal." Cimomia is in a sense a morphological transitional form, es- pecially in its suture, between Eutrephoceras and Hercoglossa. In Eutrephoceras the suture is essentially' straight, or only slightly sinuous ; in Hercoglossa the suture has well defined lobes and saddles. In both of these genera there are transitional forms to Cimomia, reflected especially b}- the suture. That is, in Cimomia there is great A-ariation in the intensitv of the lateral U6 lULLETix : .\i[:sei;m of comparative zoology Figure 3 1 KUMMEL : POST-TRIASSIC XAUTILOID GENERA 447 lobe and saddle (Fig. 'S2) . There is likewise much variation in the shape of the conch (Fig-. 31). I agree heartily with Miller (1947, p. 41) that C. kuglen and C. vaughani which have slight lobes in the ventral saddle should be retained in Ciniomin and not separated from Hercoglossa, as Woodringia was, solely on this character. In both these species the median lobe is not developed until full maturity . Most previous authors who have discussed this genus were mainly concerned with the species which occur in Tertiary for- mations. There are in addition to these, several species of Cre- taceous and possibly Upper Jurassic age that should be assigned to this genus. There is first of all Nautilus romeroi Ihering (1903, pp. 195-196, fig. 1) of the "Etage Rocaneen" (Upper Cre- taceous?) of the territory of Rio Negro in central Argentina, previously mentioned by ]Miller and Thompson (1935, p. 570) and Miller (1947, p. 41). Recently Miller (1951) has recorded poorly preserved specimens which most surely belong to Cinio- mia from Maestrichtian formation of Senegal. These specimens are too poorly preserved to accurately identify specifically, but there is no doubt as to their generic affinities. From Libya, in Fig. 31. Cross sections of the couch of A, Hercoglossa harrisi Miller and Thompson, from Stenzel 1940, fig. 116, diameter 100 mm. ±; B, H. ulrichi (White) from Stenzel 1940, tig. 116, diameter 170 mm.; C, E? foriesianus Blanford 1861, pi. 13, fig. 2, diameter 75 mm.; I), H. gardnerae Stenzel 1940, fig. 116, diameter 145 mm.; E, E. splendens Stenzel 1940, fig. 120, diameter 73 mm.; F, Hercoglossa orUciilata (Tuomey) from Miller 1947, lig. 12, height 174 mm.; G, E. mcglameryae Miller and Thompson, from Miller and Collinson 1951, fig. IB, height 100 mm.; E, Aturoidea vieirai Miller 1951, fig. 16B, height 130 mm.; /, A. parl-insoni (Edwards) from Miller 1951, fig. 15A, height 180 mm., the genotype of Aturoidea; J, A. paui'tfcr (Cope), from Miller 1947, fig. 18, diameter 228 mm.; E. Cimomin vestali Miller and Thompson, from Stenzel 1940, fig. 118, diameter 126 mm.; L, C. haltomi (Aldrich) from Miller 1947, fig. 8, height 104 mm.; M, C. bticcinaeformis Haas and Miller 1952, fig. 2, height 29 mm.; N, C. vaughani (Gardner) from Stenzel 1940, fig. 118, diameter 59 mm.; 0, C. turcicus (Krumbeck) 1905, pi. 7, fig. 6b, diameter 120 mm.; P, C. haughti (Olsson) from Miller 1947, fig. 9, diameter 60 mm.; Q, C. suhrecta Miller and Thomp- son, from Stenzel 1940, fig. 118, diameter 120 mm.; R, C. angustus (Blan- ford), from Stoliezka 1866, pi. 93, fig. 42, diameter 40 mm. 448 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY •D Figure 32 KUMMEL : POST-TRIASSIC NAUTILOID GENERA 449 North Africa, Wanner (1902, p. 143) has described Nautilus jordani, from formations of Danian age, which is clearly a species of Cimomia. Alberici (1940, p. 187) lists the range of this species as Campanian to Danian. The lateral saddle next to the umbilical shoulder in both C. romeroi and C. jordani is well developed and more distinct than in many lower Tertiary species such as C. septenicastrensis Haas and Miller from Lower Eocene of British Somaliland and C. sahariensis (Keller) from Eocene? of French Sudan. Another Cretaceous species which should be placed in Cimomia is N. angustus Blanford (1861, p. 27, pi. 14, figs. 1, 2) from the Ootatoor group (Cenomanian) of south India. Stoliczka (1866, p. 209) records an additional specimen of this species from the Ariyalur group (upper Senonian) of south India. The type specimen of N. angustus is a much com- pressed form with a narrowly rounded Tenter, and is the most compressed species of Cimomia of which a record is available. Stoliczka 's specimen (1866, pi. 93, fig. 4, 4a) is a more in- flated form and more typical of most species of Cimomia. Speng- ler (1910, p. 145) expressed doubt as to whether Stoliczka 's speci- men is conspecific with that illustrated by Blanford. Even if these two specimens are not conspecific, they both belong in Cimomia. In both specimens the suture has the characteristic pattern found in Cimomia. rig. 32. Diagrammatic representation of sutures of A, EutrepJioceras victorianum (Teichert) 1947, fig. 7; B, E. d.el-ayi (Morton) from Reeside, 1924, fig. Id.; C, Cimomia wyllei (Newton) from Haas and Miller, 1952, fig. 5 ; i», EutrepJioceras sloani Reeside from Miller, 1947, fig. 6 ; E, Cimo- mi^94. Die tithonischen Ablageruugen von Theodosia ; ein Bcitrag zur Palaontoiogie der Krim. Moseou Soc. Imp. Xat. Bull., n.s., vol. 7. pp. 206-301, pis. 9-14 (1893). RoEilER, F. A. 1836. Die X'ersteinerungen der norddeutschen Oolithen-Gebirges. 218 pp., 16 pis., (in Atlas). 488 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY RONCHETTI, C. E. 1947. Appendice alia revisione dei Gen. Nautilus, Indoceras, e Bacndites del Neoeretacico della Libia. 1st. Geol. Paleont. Geog. Fisica, Univ. Milano, ser. P, pub. 50, pp. 1-4, fig. 3. ROUCHADZE, J, 1931. Notice siir les Belemnites et les Nautilus de I'Aptien de Koutais. Bull. Mus. Georgie, vol. 6 (1929-1930), pp. 125-138, pis. 1-3. Sanchez Roig, Mario 1951. La Fauna Jurasica de Vinales. Ann. Ciencias Medicas, Fis. Nat., Havana, vol. 89, pt. II, pp. 47-94, pis. 1-28. SCHAFHAUTL, K. E. 1852. Der Teisenberg oder Kressenberg in Bayern. Neuea Jahrb. Mineral. Geogn. Geol. Petrefakten-Kunde. Jahrgang 1852, pp. 129-175, pis. 3, 4. 1863. Siid-Bayems Lethaea Geognostica, — Der Kuessenberg und die siidlieh von ihrn gelegenen Hochalpen, geognostisch betrachtet in iliren Petrefacten. Leipzig, 487 pp., 86 pis., 7 maps (in Atlas) SCHENCK, H. G. 1931. Cephalopods of the genus Aturia from western North America. Calif. Univ. Pub., Dept. Geol. Sci., Bull., vol. 19, pp. 435 491, pis. 66-78. SCHLAGINTWEIT, 0. 1912. Die Fauna des Vracon und Cenoman in Peru. Neues Jahrb., Beil.-Bd. 33, pp. 43-136, pis. 5-7. SCHLOTHEIM, E. F. 1820. Die Petrefactenkunde. Cephalopoda. Pp. 45-88. SCHLUTEB, C. 1876. Cephalopoden der oberen Kreide. Palaeontographiea, vol. 24, pt. II, pp. 123-263 (1-144), pis. 36-55. Scott, G. 1940. Cephalopods from the Cretaceous Trinity group of the south- central United States. Texas Univ. Bull. 3945, pp. 969-1106, pis. 55-68. 1943. Palaeontology of Harrar Province, Ethiopia Part 4, Jurassic Cephalopoda and a Cretaceous Nautilus. Am. Mus. Nat. Hist. Bull., vol. 82, art. 3, pp. 55-94, text figs. 1-23, pis. 10-25. Sergio, Venzo 1933. II cattiano di Monte Brione presso Riva del Garda e la sua nuova fauna. Studi Trentini Sci. Nat. Trento, vol. 14, fasc. 3, pp. 188-213, pis. 1-2. KUMMEL : POST-TRIASSIC NAUTILOID GEXERA 489 Sharpe, D. 1853. On the age of the fossiliferous sands and gravels of Farringdon aud its neighbourhood. Quart. Jour. Geol. Soc. London, vol. 10, pp. 176-198, pis. 5-6. Shattuck, G. B. 1903. The Mollusea of the Buda limestone. U. S. Geol. Survey Bull. 205, pp. 1-36, pis. 1-25. ShIMIZU, S^iBUEO 1926. On two species of Nautiloidea from the Tertiary of Japan. Sci. Kept. Tohoku Imp. Univ., ser. 2 (geol.), vol. 9, pp. 25-27, pi. 8. 1931. The marine Lower Cretaceous deposits of Japan, with special reference to the Ammonite-bearing zones. Sci. Kept. Tohoku Imp. Univ., ser. 2 (geol.), vol. 15, no. 1, pp. 1-40, 4 pis., 1 text fig. Shumard, B. F. 1860. Descriptions of new Cretaceous fossils from Texas. St. Louis Acad. Sci. Trans., vol. 1, pp. 590-610. SiNZOW, I. 1913. Beitrage zur Kenntnis der unteren Kreideablagerungen des Nord- Kaukasus. Trav. Mus. Geol. Imp. Sci. St. Petersburg, vol. 7, pp. 93-117, pis. 4-6. SOKOLOV, D. X. 1928. Fossiles Mesozoiques de la Bolschezemelskaja Tundra et de Kash pur. Trav. Mus. Geol. Leningrad, vol. 3, pp. 15-62, pis. 3, 4. SORRENTIXO, StEFAXO 1932. Cefalopodi Maestrichtiani della Tripolitania. Bol. Soc. Geol. Italiana, vol. 51, pp. 184-209, pis. 4-6. SowERBY, J., and J. de C. Sowerby 1812- Mineral Conchology. 7 vols., pis. 1-337 (1812-1822) by J. 1846. Sowerby; pis. 338-648 (1822-1846) by J. de C. Sowerby. Spath, L. F. 1921. Cretaceous Cephalopoda from Zululand. Ann. S. African Mus., vol. 12, pt. 7, no. 16, pp. 217-321, pis. 19-26. 1927a. Eevision of the Jurassic cephalopod fauna of the Kachh (Cutch). Mem. Geol. Surv. India (Palaeout. ludica), n.s., vol. 9, no. 2, pp. 1-84, pis. 1-7. 1927b. On the classification of the Tertiary nautili. Ann. Mag. Nat. Hist., ser. 9, vol. 20, pp. 424-428. 1930a. The Eotriassic invertebrate fauna of East Greenland. Medd. oni Gronland, vol. 83, pp. 1-90, pis. 1-12. 1930b. On the Cephalopoda of the Uitenhage beds. Ann. S. African Mus., vol. 28, pp. 131-157, pis. 13-15. -tyU BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 1934. Catalogue of the fossil Cephalopoda in the British Museum (Nat ural History). Part 4, The Amnionoiflea of tlic Trias. London, l)p. 1-521, pis. 1-18. 1935a. Additions to the Eotriassic invertebrate fauna of East Green- land. Medd. om Granlaud, vol. 98, pii. 1-115, pis. \--2o. 1935b. The Mesozoic palaeontology of British Sonialiland. X, Jurassic and Cretaceous Cephalopoda. Geol. Palaeont. British Somali- land, pt. 2, pp. 205-228, pis. 24-25. 1953. The Upper Cretaceous cephalopod fauna of Graham Land. Falkland Is. Dependencies Survey, Sci. Eept. no. 3, pp. 1-60, pis. 1-13. Spexgler, Erich 1910. Uutersuehungen liber die siidindisehe Kieideformation. Die Nautiliden und Belemniteu des Trii-liinopolydi8trikts. Beitr. PalJiont. Osterreich Ungarns und Orients, vol. 23, pp. 125-157, pis. 11-14 (26-29). 1919. Die Gebirgsgruppe des Plassen und Hallstattcr Salzberges ini Salzkammergut. Jahrb. Geol. Eeich., Jalirgnnj; 1918, vol. 68, pp. 285-474, pis. 14-18 (1-5). Stchepixsky, V. 1943. L'Oligocene mariu dc Gaziantep (Turquie meridiouale). Maden Tetkik ve Arama, Ankara, sene 8, sayi 2/30, pp. 223-248, pis. 1-4. Steinmann, G. 1895. Die Cephalopoden der Quiriquinaschichten. Neues Jahrb. Beil.- Bd. 10, pp. 64-94, pis. 4-6. Stexzel, H. B. 1935. Xautiloids of the genus At aria from the Eocene of Texas and Alabama. Jour. Paleont., vol. 9, pp. 551-562, pis. 63, 64. 1940. Tertiary nautiloids from the Gulf Coastal Plain. Texas Univ. Publ. No. 3945, pp. 731-794, pis. 35-42, 1948. Paleoecology of Tertiary nautiloids. Rept. Comm. on a Treatise on Marine Eeol. and Paleoecol., Div. Geol. Geog., Natl. Res. Council, no. 8, pp. 96-97. 1948. Ecology of living nautiloids. Kept. Comm. on a Treatise of Marine Ecology and Paleoecol., Div. Geol. and Geog., Nat. Ees. Council, pp. 84-90. 1952. Living Nautilus. Treatise on Invertebrate Paleontology, Pt. H, Cephalopoda, Nnntiloidea, pp. HI — 1119, Preprint. KUMMEL: POST-TKIASSIf XAUTlLulD GENERA 491 Stephknsox, Iv. W. lOil. Thp larger invertebrate fossils of the Xavarro group of Texas (exc-lusivo of corals and crustaceans and exclusive of the fauna of the Escondido formation). Texas Univ., Pub. 4101, pp. ]-«41, pis. 1 !t.l. .Stetee, Alejandro 19i:i. Estratos Juracicos Argentinos. Aetas Acad. Nac. Cienc. Cor doba, vol. 7, pp. 3.3-128, pis. 1-24. Strombeck, a. v. 1863. Ueber die Kreide am Zeltberg bei Liineburg. Zeitschr. deutsch. geol. GeselL, vol. 15, pp. 97-187, pi. 4. Tagliarixi, F. p. 1901. Monografia sui Nautili tlel Dogger inferiore di Monte San Giuliano (Erice). Gior. Sci. Xat. Econ., Palermo, vol. 23, pt. I, pp. 186-203, pis. 1(5. Tavani, G. 1942. Paleoutolugia della .Somalia, AEolluscki del Cretaceo della Somalia. Palaeontographica Italica, vol. 32, suppl. 4, fasc. 2, pp. 7-47, pis. 1 5 (10-14). Teichert, C. 1940. Coutributious to nautiloid nomenclature. Jour. Paleont., vol. 14, pp. 590-597. 1943. Eocene nautiloids from Victoria. Proe. Roy. Soe. Victoria, vol. 55. pt. 2 (n. s.j, pp. 257-264, pi. 11. 1944. The genus Aturia in the Tertiary of Australia. Jour. Paleont., vol. 18, pp. 73-82, pis. 14-16. 1947a. Notes on Eocene nautiloids from Victoria. Mining Geol. Jour., vol. ;!, Mu. 1, pp. 40-42. 1947b. New nautiloids from the older Tertiary of Victoria. Mining Geol. Jour., vol. 3, no. 2, pp. 48-51. , and B. C. Cotton 1949. A new Aturia from the Tertiary of South Australia. Rec. South Australian Mus., vol. 9, pp. 255-256, pi. 21. , and B. F. Glenister 1952. Fossil nautiloid faunas from Australia. Jour. Paleont., vol. 26, pp. 730 752, pis. 104-108. Tf.rquem, O. 1855. Paleontologie de I'etage inferieur de la formation Liasique (le la province de Luxembourg, Grand-Duche (Hollande) et de llctlange, Departement de la Moselle. Mem. Soc. Geol. de France. 2nd ser.. vol. 5. no. 3, pp. 219-343, pis. 1-15. 492 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY TuoMET, Michael 1854. Descriptions of some new fossils from the Cretaceous rocks of the southern states. Acad. Nat. Sci. Philadelphia, Proc, vol. 7, pp. 167-172. Uhliq, Victok 1883. Die Cephalopodenfauna der Wernsdorfer Schichten. Denksch. Akad. Wiss. Wien, vol. 46, abt. 2, pp. 127-290, pis. 1-32. Vacek, M. 1886. ijber die fauna der Oolithe von Cap. S. Vigilio. Abh. K. K. Geol. Eeich. Wien, vol. 12, no. 3, pp. 57-212, pis. 1-20. Vadasz, M. E. 1911. Die Juraschichten des Siidlichen Bakony. Eesultate Wiss. Erforsch. des Balatonsees, vol. 1, Pal. Abh. 3, pp. 1-89, pis. 1, 2. Valenciennes, Achille 1841. Nouvelles recherches sur le Nautile flanibe. Paris Mus. Natl. Hist. Nat., Archiv., vol. 2, pp. 257-314, pis. 8-11. Venzo, S. 1937. La fauna Cattiana delle glauconie Bellunesi. Mem. Inst. Geol. E. Universita di Padova, vol. 13, pp. 1-207, pis. 1-12. VlALLI, ViTTERIO 1937. Ammoniti giurassiche del monte Peller. Trento, Mus. Storia Nat. Venezia Tridentina, Mem. an 7, vol. 4, pp. 99-148, pis. 1-3, 5 figs. Vincent, Emxle 1907. Description d'une espeee de cephalopode tetrabranchial nouvelle, du Landenien de Belgique {At^iria linicentensis, E. Vincent). Ann. Soc. Eoy. Zool. Malac. Belgique, vol. 42, pp. 321 323, fig. 1. 1913. Contribution a la paleontologie des Falaises de Landana (Bas- Congo) ; Mollusques. Mus. Congo Beige. Ann., sec. 3, vol. 1, pp. 1-46, pis. 1-6. VOQL, V. 1908. iJber Eozane Nautilidcn. Foldtani Kozlony, vol. 38, pp. 568-582, 636-649. 1910. Neuere Beitrage zur Kenntnis der alttertiaren Nautiliden. Centralbl. Min., Geol. Pal., Jahrg. 1910, pp. 707-710. VOKES, H. E. 1937. Nautiloid cephalopods from the Eocene of California. Jour. Paleont.. vol. 11, pp. 3-9, pis. 1, 2. KUMMEL: POST-TBIASSIC NAUTILOID GENERA 493 Veedenburg, E. W. 1925. Descriptions of Mollusca from the post-Eocene Tertiary forma tion of northwestern India; Cephalopoda, Opisthobranehiata, Siphonostomata. Mem. Geol. Surv. India n. s., vol. 50, pt. 1, pp. 1-350, pis. 1-13. , and G. de P. Cotter 1928. A supplement to the Mollusca of the Eanikot series. Mem. Geol. Surv. India (Palaeont. Indica), n. s.. vol. 10, no. 4, pp. 1-75, pis. 1-9, and corr. slip. Waagen, W. 1873. Jurassic fauna of Kutch. The Cephalopoda. Mem. Geol. Surv. India (Palaeont. Indica), ser. 9, vol. 1, pt. 1, pp. 1-22, pis. 1-4. Wanner, J. 1902. Die Fauna der obersten weissen Kreide der libysehen Wiiste. Paleontographica, vol. 30, Abt. 2, pp. 91-152, pis. 13-19. Waring, C. A. 1914. Eocene horizons of California. Jour. Geol., vol. 22, pp. 782-785. Weaver, C. E. 1931. Paleontology of the Jurassic and Ctetaceous of west-central Argentina. Mem. Washington Univ. (Seattle), vol. 1, pp. 1-594, pis. 1-62. Wetherell, N. T. 1836. Observations on some of the fossils of the London clay, and in particular those organic remains which have been recently discovered in the tunnel for the London and Birmingham Rail- road. London and Edinburgh Phil. Mag. Jour. Sci., ser. 3, vol. t 9, pp. 462-469. White, C. A. 1884. On the nautiloid genus Enclimutoceras Hyatt and a description of the type species. U. S. Geol. Surv. Bull. 4, pp. 16, 17, pis. 7-9. Whiteaves, J. F. 1876. On some invertebrates from the coal-bearing rocks of the Queen Charlotte Islands collected by Mr. James Richardson in 1872. Canada Geol. Survey, Mesozoic fossils, vol. 1, pt. 1, pp. 1-92, pis. 1-10. 1900. On some additional or imperfectly understood fossils from the Cretaceous rocks of the Queen Charlotte Islands, with a revised list of the species from these rocks. Canada Geol. Survey, Mesozoic fossils, vol. 1, pt. 4, pp. 263-307, pis. 33 39. Whitfield, R. P. 1892. Gastropoda and Cephalopoda of the Raritan clays and Greensand marls. U. S. Geol. Surv., Mon. 18, pp. 1-402, pis. 1-50 [also issued as N. J. Geol. Surv., Paleont., vol. 2]. 494 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Yabe, H., and H. Ozaki 19n3. A new type of Cretaceous nautiloids from Tyosi Peninsula, Kwanto Eegion, Natl. Sei. Mus. (Tokyo), Bull. 32, pp. 55-61. , and S. Shimizu 1924. A new species of Nautilus, X. (Cymatoceras) pseudo-atlas Yabe and Shimizu, from the upper Cretaceous of Amakusa. Japanese Jour. Geol. Geogr., vol. 3, pp. 41-43, pi. 5. YOKOYAMA, M. 1913. On two new fossil Cephalopoda from the Tertiary of Izunio. Jour. Geol. Soc. Tokyo, vol. 20, pp. 33-35 (1-3), pis. 7, 8. Young, G., and J. Bird 1828. A geological survey of the Yorkshire Coast. 2nd ed. Whitby. 324 pp., 17 pis. ZlETEN, C. H., V. 1830. Die Versteinerungen Wiirttembergs, 12 pts., 102 pp., 72 pl.s. (Stuttgart). For dates of publication Fee Crick, 1899, p. 554. ZiTTEL, K. A. VON. 1868. Palaeontologische Studien iiber die Granzschichten der Jura und Kreide-Formation im Gebietc der Karpathen, Alpen und Apenninen, I. Abtheilung, Die Cephalopoden der Stramberger Schichten. Mus. K. Bnyer. Staates Palaeont. Mitt, (begonnen von Oppel, forgesetzt von Zittel), vol. 2, pp. 33-118, pis. 1-24 in an atlns. ZWIERZYCKI, J. 1914. Die Cephalopodenfauna der Tendagurusehichten Deutsch-Ost afrika. Arehiv. Biontologie, Gesell. Xaturforsch. Freunde Bev lin, vol. 3, pt. 4, Teil 3, pp. 7-96, pis. 1-10. • 1915. Fossielei! Afkomstig van eenige Vindplaatsen op Sumatra. Jaat Mi.inwezen Neder. Oost-Indie, vol. 42, pp. 101-129 (1913). PLATES PLATE 1. CENOCEKAS ORBIUNVI Figures Pago 1, 2 Cenocevan orhignyi (Prinz ) 361 Two views of type species of the genus Ccriuccra.s (=A'aw- tilus inter 'medius d'Orbigny — non Sowerby). From d'Or- bigny. X 1. PLATE 1 PLATE 2. CENOCERAS TBECHMANNI Figures Page 1-3 Cenoeeras trechmanni (Knmmel » 349 Three views of holotype from Upper Triassic, Carnian, Bed C, Otomita, Hokonin Hills, New Zealand. B.M.N.H. C21947. X 1. <»J ''x^''-''^' V\/ ^ CM ^fy PLATE 2 PLATE 3. CENOCERAS STRIATUS Figures Page 1, 2 Cenoceras striatus (Sowerby) 362 Two views of plesiotype, X 0.60, from Lower Lias, locality unknown. B.M.N.H. 43852 (Sowerby Collection). PLATE 3 PLATE 4. CENOCEEAS SIMILLIMUS Figures Page 1, 2 Cenoeeras simillimvf; (Foord and Crick) 362 Two views of holotype, X 1, from Lias, Oharmouth, Dor set, England. B.M.N. H. 39887 (Harrison Collection). PLATE 4 PI.ATH 5. OENOCERAS TNORNATFS Figures Pago 1. 2 Ccnoci rtiK innnuitus (d 'Orbigny) 3H2 Two views of i)lesiotyi)e, X 0.75, from Inferior Oolite, Burton Bradstock, Dorset, England. B.M.N. II. C2843 (Baber Collection). \ PT.ATF 5 PLATE 6. CENOCERAS BURTONENSIS Figures Pnge 1, 2 Cenoccras burtnncnsis (Foord and Crick) ;J71 Two views of liolotype, X 0.75, from Inferior Oolite, Burton-Bradstock. Dorset. Eng;land. B.M.N.H. 2841 (Ba- ber Collection ) . "^- PLATE 6 PLATE 7. CENOCERAS AND SYRTOXArTILUS Figures Page 1, 2 Cenoceras excavatus (Sowerby) 369 Two views of holotype, X 1, from Inferior Oolite, Dorset. B.M.N. H. 4.3855 (Sowerby Collection). 3 Syriotmutiliis Ubanoiicus (Foord and Crick) 433 Ventral view of syntype, X 1, with mandible on surface, from Upper Cretaceous, Sahil Ahr.a, I^ebanon, Syria. B.M.N.H. 83663 (Lewis Collection). Previously figured in Cat. Foss. Ceph., B.M., I't. '1, p. 371, fig. 8'2 ; mandible figured, Ann. Mag. Xat. Hist. [(i|. vol. 5 (1890;, p. 405, fig. 6c; and Cat. Foss. Ceph. B.M., Pt. 2, p. 305, fig. 67c. \ CM / PLAIE 7 PLATE 8. PAEACENOCERAS AND CENOOEEAS Figures Page 1 Paracenoccrds hfXdf/onum (Sowerby; 4Q2 View of polished section of holotype from Corallian Oolite, locality unknown. See Plate 13 for viewt? of external fea- tures. B.M.X.H. 43S56 (Sowerby Collection MC 529f2). X 1. 2 Cenociras cjiavatux (Sowerby ) ;{(>y Mew (if polished section of holotype from Inferior Oolite, Dorset. See Plate 7 for views of external features. B.M.N.H. 43855 (Sowerby Collection). X 1. PLATE 8 PLATE 9. OBINAUTILUW AND CENOCEKAW Figures Page 1, 2 ObinaulUus jnilclira Kt)li;t vashi 386 Two views of liolotype from Palaeogene Nichinan forma- tion at a cutting between Obi and Aburatsu in Province Hyuga, Japan. X 1. From Kobayashi. 3-6 Cenoctrax roi nudum (Hyatt ) 3(j0 Four views of liolotype of IHgoniorrras Hyatt. Locality and horizon unknown; Hyatt labels it ''obviously from the Oolite." X 1. M.C.Z. 5:217. •C Mi ':% ^^^ PLATE 9 PLATE 10. TIEMINAUTILUS AND PSEUDOCENOCERAS Figures Page 1 , 2 llemiiumUlu.s Mixbii (Morris ) 434 Two views of holotype, X 1, from Lower Greensand, Atherfield, Isle of Wight. B.M.N. H. 47019 (Siixli.y Collec- tion). 3, 4 raeudoeenoctras Idrylllierluinu.s (d 'Orbiguy) 384 Two views of plesiotype, X 1, from Craie Chloritee, Rouen, Er:tiu-e. B.M.N.H. C1016. /WM m M \ m ZJ.. I M . if ■ / / /' ^^■- \'\ V % PLATE 10 PLATE 11. CARINONAUTILUS ARIYALURENSTR Figures Page 1-3 Carinonautilus ariyalurensis Spengler 385 Three views of holotype. X 1, from Ariyalur group (Up per Cretaceous), Trichinopoly district, India. From Spengler. -^ ^-J PLATE 11 PLATE 12. PSETTDONAUTILUS GEINITZT Figures Page 1, 2 Psevdonautiltts gcinitzi (Oppel) 397 Two views of the holotype, X 1, from Stramberger beds (Upper Jurassic) of Stramberg, Moravia. FroTu Zittel. fi;r^3S^*f*^-^ ^5*^^.. '■~-*-*^Ta<^gss?r-- PLATE 12 PLATE l.-.. PAR.A(^EN0CEKA8 HEXAGONUM Figures Page 1, - ParacenoceraH hcxac/onum (Sowerby) 402 Two views of liolotype, X 1, from Corallian Oolite, locality uuknown. R.M.N.H. 43856 (Soworhy roUection. MV 52i)-f2). PLATE 13 PLATE 14. AI^LACONAUTILUS AND SOMAT.TNAUTTLUS Figures Page 1, 2 AvlaronautiluN sr.rcarwafii.s (Pictet) 4()fl Two views of holotypc, X 0.75, from beds with Terc- bratulo dipliyoideft of Berrias (Ardeche), France. From Pictet. n, 4 Snmalinaiitilvs anticpnis (Daf-que) 411 Two views of holotype, X 1, from ITpper .Turassie strata of SomalilaTid. From Daeqiie. ^mmm^~. \i!. PLATE 14 I'LATE 15. TITHONOCERAS ZITTELI Figures Pago 1 . L' Tithonooerati zitteli Retowski 410 Two views of holotype, X 1, tioni I'lijier JuiMssie strata of Crimea. From Retowski. .*- - PLATE 15 I'LATE Ui. CVMATOCERAS PSEUDOELEGANS Figures Page 1, 2 Cymatoceru.s puc/ndoeleyans (d 'Orbiguy j 418 Two views of cast of holotype ? (see text page -ilS) in Museum of N.,t,„.j,i History. Paris. X 0.5. B.M.N. 11. C526(i. '^^ •^ f PLATE 16 PLATE 17. PEOCYMATOCERAS SUBTRUNCATUS Figures I'agf 1,2 Procymatoceras nubtruncutus (Morris and Lycett) 428 Two views of topotype, X 0.5, from Great Oolite, Min- chinhampton, Gloucestershire, England. B.M.N. H. 21275. PLATE 17 PLATP: ]8. CYMATONAUTILUS .FULII Figures Pago 1-4 Cymatoriautilus juiii (,d'Url)igny j 430 Four views of type species from strata of (Jallovian age near Montreuil-Bellay and OhaufPoui-, Fraiuc. P^om Gueranger. ,,5^*H*r PLATE 18 PLATE 19. PABACYMATOCERAS ASPER Figures Page 1, 2 Paracymatoceras asper (Oppel) 427 Two views of holotype, X 1, from the Stramberger beds (Upper Jurassic) of Stramberg, Moravia. Proni Zittel. '^^V.^-;. :;'<'V-- -'^ -'^ it^fV ^- :■■ . - c. v./ . % .?A^- PLA'I'K 19 PLATE 20. ANGLONAUTILUS UNDHLATT'S Figures Page 1, 2 AnglonauUlus undulatns (Sowerby) 430 Two views of topotype, X 1, from Lower Greensand, Nui- field, Surrey (fide L. F. Spath, p;Msonal communication). B.M.N.H. 6(53315. (Sowerby Collection MC 182). -r.-^^^^K-^as^ \ >■'> , ■• CvJ _2S-:- PLATE 20 PLATE 21. EUCYMATOCERAS PLICATUS Figures Page 1, 2 Eucymatoceras plicatus (Fitton) 431 Two views of plesiotype, X 0.73, from Lower Greensand, Atherfield, Isle of Wight. B.M.N.H. C335. C«J N.;^ ■i\ f X PLATE 21 PLATE 22. ])ELTOCYMATOCERAS LEIOTROPIS Figures Page 1, 2 Ddtacjiiiiatorcids Ic'ioiropia (.Schliite'r') 437 Two views of type specimen, X 0.4. frotri Tppcr Vvi.' tMceous I'oiin.-iticii of Essen, Gei-ni;in.v. From Scliliiter, 1876. ^^ " ' .%#*,. ^".si'r^ FLATK 22 I'LATK L'.l. KHlCVMATOCEJtAW VAEJ^SENSIW Figures Page 1, 2 Epicyniatoctra.s vatLsc.nnis (Biru-khoisf ) 43§ Two views of type species, X ], from Upper Cretaceous (Maestriehliaii) formation at l>imlioing, Belgium. From Schliiter, 187(i. -^r*^5;n Rruxelles, Belgium. X 0.0. From Burtin. PLATE 24 PLATE 25. ANGULITHES AND "BISIPHYTES" Figures Page 1, 2 An(julUh( li triangulari.!< Moiitfoit 452 Figure 1, from Montfort 1808, p. (i ; Figure 2, from Mont fort, 1802. pi. 49. 3 ' ' Bisiplnites" reticulatux Montfort ;5(il View of holotype of the genus Bisiphytes from beds of uncertain age near Somberuon, France. X 1. PLATE 25 PLATE 26. ANGULITHES TRIANGULAEIS Figures Page 1, 2 Angulithes triangularis (Montfort) 452 Two views of plpsioty])e, X 0.(55, from Craie Chloritee (Cenomanianj Escragnolles, France. B.M.N.H. 37097 (Astier Collection). v v^^_ X A PLATE 26 PLATE 27. HERCOGLOSSA ORBIOULATA Figures Page 1, 2 HcrcogJossa orbiciilata (Tuouiey ) 457 Two views of neoholotype from the Clayton foniijitit)ii on Pine Barren Creek, about 5 miles nortli of Alleiiton, Ala- bama, X 0.5. From Miller. CVJ ■>t >'4 >: ^-<^: /^ / PLATK 27 PLATE 28. ATUROIDEA AND PSEUDAGANIDES Figures Page 1. 2 Atnroidca parkinsoni (Edwards) 462 Lateral view (X 0.5) and septal view (reduced) of holo- type, from London Clay at Harwich, England. From Edwards. 3, 4 Pseiulaganides TcutclienMs (Waagen') 390 Two views of holotype from beds with Perisphinetes obtusicosta of Wanda, Kachh (Cutch) Paki-stan. From Waagen. PLATE 28 Harvard MCZ Libra liiliniiii 3 2044 066 303 868 Date Due oef-3-n992