mo HARVARD UNIVERSITY Library of the Museum of Comparative Zoology BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, IN CAMBRIDGE VOL. 112 CAMBRIDGE, MASS., U. S. A. 1954 - 1955 The Cosmos Press, Inc. Cambridge, Mass., U. S. A \ CONTENTS PAGE No.* 1. — The Ant Genus Strumigenys Fred. Smith in the Ethiopian and Malagasy Regions. By Wil- liam L. Brown, Jr. August, 1954 1 No. 2. — Deep Water Elasmobranchs and Chimaeroids from the Northwestern Atlantic Slope. By Henry B. Bigelow and William C. Sehroeder. Sep- tember, 1954 35 No. 3. — Status op Invertebrate Paleontology, 1953. By Bernhard Kummel, Editor. October, 1954 . 89 No. 4. — Revision op the Chrysomelid Subfamily Aiila- coscelinae. By P. Monros. November, 1954 . . 319 No. 5. — The Comparative Biology of Reproduction in the Wood-Boring Isopod Crustacean Limnoria. By Robert J. Menzies. December, 1954 . . 361 No. 6. — The Genus Eustala (Araneae, Argiopidae) in Central America. By Arthur M. Chickering. March, 1955 ..." 389 Bulletin of the Museum of Comparative Zoology AT HAEVARD COLLEGE Vol. 112, No. 1 THE ANT GENUS STRUMIGENYS FRED. SMITH IN THE ETHIOPIAN AND MALAGASY REGIONS By William L. Brown, Je. Museum of Comparative Zoology, Harvard University CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM August, 1954 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863- The current volume is Vol. 112. Breviora (octavo) 1952 - No. 35 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. 33 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 — Vol. 1, no. 17 is current. Proceedings of the New England Zoological Club (octavo) 1899 1948 — Published in connection with the Museum. Publication terminated with Vol. 24. These publications 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. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 1 THE ANT GENUS STRUMIGENYS FRED. SMITH IN THE ETHIOPIAN AND MALAGASY REGIONS By William L. Brown, Jr. .Museum of Comparative Zoology, Harvard University CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM August, 1954 No. 1. — The Ant Genus Strumigenys Fred. Smith in the Ethiopian and Malagasy Regions By William L. Brown, Jr. This is a part of my revision of the dacetine ant genus Strumigenys Fred. Smith, planned to include the entire Strumi- genys world fauna. For information concerning the characters and relationships of Strumigenys, the reader should consult my recent general references on tribe Dacetini (Brown, 1948, 1953). The 1953 reference also contains a detailed discussion of the standard measurements most useful in dacetine studies and the indices derived from these measurements. To recapitulate briefly : TL or "total length" is the sum of the exposed lengths of the head with mandibles, alitrunk, petiole, postpetiole and gaster. HL is the maximum measurable length of the head proper, seen in perfect dorsal full-face view, including all of occipital lobes and clypeus. ML is the exposed length of the closed mandibles from dorsal view (same position from which HL is measured). WL, or Weber's length of alitrunk, is the diagonal distance from base of cervix to metapleural angles, as seen from the side. CI is the cephalic index, or HL/maximum width of headXlOO. MI, mandibulo-cephalic index, is HL/MLXlOO. Error of measurement for the head and mandibles should not normally exceed ±0.01 mm.; errors of indices as calculated from raw measurement units should not exceed ±1. The cooperation of numerous individuals in the entire dace- tine project is cited in detail in my 1953 reference, but I should like to acknowledge here the most valuable loans of material and other aid rendered by the following : Dr. George Arnold, of Bulawayo ; Prof. Francis Bernard, of the Universite d'Alger; Dr. Ch. Ferriere, of the Museum d'Histoire Naturelle, Geneva ; Prof. Guido Grandi, of the University of Bologna ; Prof. Ed. Handschin, of the Naturhistorisches Museum of Basel ; Dr. Harlow B. Mills, of the Illinois Natural History Survey; Dr. E. S. Ross, of the California Academy of Sciences; Dr. George Salt, of Cambridge University ; and Dr. Neal A. Weber, of Swarthmore College, Pennsylvania. Aside from a handful of obscure species in Microdaceton Sant- schi, Smithistruma Brown, Miccostruma Brown, Cocliomyrmex Wheeler, and perhaps one or two other small genera at present 4 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY known from North Africa, the Ethiopian-Malagasy clacetines fall into two distinct, relatively common, and presumably dom- inant genera : Strumigenys and Serrastruma Brown. Serra- struma underwent drastic preliminary revision in a recent paper (Brown, 1952), and it now appears that the number of species may have to be reduced still further by synonymy, since only four to six of those names appear to represent distinct entities. These few Serrastnima species are all very much alike, and all are exceptionally variable in a tribe which is otherwise outstand- ing in the constancy of species characteristics. Serrastnima appears to be a relatively recently evolved group of Ethiopian origin; its ancestors are probably to be looked for in Smithi- struma species like those of the alberti group. Serrastrama is easily the commonest, and apparently the dominating African dacetine genus, and it seems likely (on the assumption that it competes for the same food as other clacetines, namely collembola and a few other small cryptobiotic arthropods) that its presence is the chief cause of the scarcity of other dacetine groups below the Sahara. The other genus fairly well developed in Africa is Strumi- genys, which survives as fourteen known, valid species in the Ethiopian Region, plus one in Madagascar. Two of the Ethiopian species have become established as tramps outside Africa- Madagascar (rogeri and scotti), and these will be discussed be- low. All of the African Strumigenys species clearly belong to one ancestral stock, of which the most generalized known species is S. grandidieri of Madagascar, although the close interrelation- ship of these specie.s is masked by extensive morphological radia- tion and might not therefore be guessed at without one 's having the complete series of forms in intergrading array. Undoubtedly, other species from these regions remain undescribed and uncol- lected, but in spite of the incompleteness of the record, it seems clear that the Afro-Madagascan Strumigenys fauna is a very limited one compared to the two other distinct faunas of the genus. The New World and Indo-Australian faunas share, roughly equally, at least 100 described and undescribed species that I have been able to verify to date, leaving out those that are clearly synonyms. The reason for the paucity of the African Strumige?iys fauna brown : ant genus Strumigenys FRED. SMITH 5 is not entirely clear, especially in the absence of detailed ecologi- cal data, but it seems probable that competition between Strum- igenys and Serrastruma is at least partly responsible. In the other tropical dacetine faunas, Strumigenys is usually clearly the dominant genus, though Smith istruma is abundant in some of these regions and competes for much the same food (chiefly or entirely collembola of entomobryoid and isotomoid affinities). Strumigenys and Smithistruma differ, however, in the basic form of the mandibles, and to a certain extent also in details of predatory behavior (Wilson, 1954; Brown, 1954; unpublished data of both authors), so that competition between the two groups is probably only partial. On the basis of evidence I have given elsewhere (especially in the 1953 reference), it seems very likely that the long-mandibulate life-form (Strumigenys) is ancestral to the short-mandibulate, or smithistrumiform type. The long-mandibulate type is better fitted for foraging in the open, while the short-mandibulate forms are more suited to cryptic hunting; the former tends to concentrate more toward the tropics, while the latter is predominant, in the Northern Hemisphere, at least, in the warm temperate belts. In Africa, however, the chief short-mandibulate genus is Ser- rastruma, which exists successfully through most of the con- tinent from South Africa to the Sahara, avoiding only extreme montane and desert habitats. The Serrastruma mandibles, a modified and somewhat elongate version of those of Smithistruma, may be a very efficient prey-catching instrument, but there is probably more than this behind the success of Serrastruma. Probably ecological tolerances and fertility are involved strongly here, as suggested by the extreme morphological variability of the species and the rather larger-than-average nest populations, of which accurate estimates are just beginning to come in. As mentioned already, the Afro-Madagascan Strumigenys so far described all belong to a single group (the rogeri group, here so named) representing a common stock. This group, excepting the two probably historically-migrant tramp species, is restricted to Africa and Madagascar, and its relationships with groups of other faunas are not particularly close. With the single excep- tion of S. rufobrunea, a widespread and very variable species, the African Strumigenys show the narrow ranges of intraspecific b BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY variation usually expected of members of the genus, and ecologi- cal specialization, so far as known, follows morphological varia- tion. Strumigenys in Africa, as elsewhere, seems to tend to fill the available niches by speciating to specialized types, which then stabilize and become relatively constant and therefore are easily separable one from another. Serrastruma, on the other hand, fills many or all of the same niches by producing only a few, but very plastic species. It is interesting to note that the three truly widespread tramp dacetines [Strumigenys rogeri Emery, Trichoscapa membrani- fera Emery and Quadristruma emmae (Emery)] are all either certainly or very likely of African origin. In fact, Quadristruma emmae, though not yet recorded from Africa proper, is now seen, through the discovery of the intergradient Strumigenys tetraphcmes (q.v. infra), to be a direct offshoot of the S. rogeri group and hardly separable from that group except by the detail of the loss of the two smallest antennal segments. In a similar way, Trichoscapa membranifera seems close to certain old African stocks of short-mandibulate genera (Smithistruma, Cod- iomyrmex loveridgci Brown, Miccostruma), and this species is actually known from various humanly-disturbed areas in Egypt and elsewhere in North Africa. S. rogeri is widespread in the West Indies, the Pacific, and in the hothouses of the British Isles, and it has long been thought to be a native Neotropical ant, but I can now show that it is African as its morphological relatives are ; the West African S. sidfurea proves to be its synonym. The findings in the Dacetini are in agreement, as concerns tramping ability, with the patterns of colonization now emerging for the ants in general. It would seem, despite certain outstand- ing exceptions to the contrary, and taking into account the chaotic state of the taxonomy of ants, that the worldwide "tramp habitats" in the tropics and subtropics are occupied mainly by species and genera of African origin. The degree to which this is true will only be surely revealed after much difficult basic taxonomic labor, involving the synonymy of many species and varieties described originally on a purely geographical basis. The origins will be made known chiefly by analyzing group relation- ships, and then searching for the particular species concerned in its presumed native area, for it is frequently true that in its native brown : ant genus Strumigenys FRED. SMITH 7 range, an ant elsewhere common as a tramp will be diffusely distributed, or even rare. At present writing, this appears to be the case with all the dacetines mentioned above. The stages of tramping through the agency of human com- merce appear to be fairly obvious. The critical step involves the chance transplantation of a continental propagule to a favorable offshore island with a limited native fauna. If such an island is not already "saturated" with competing tramps, the chances greatly favor the establishment of a very dense population there within a relatively short time. With the pressures of normal continental extraspecific competition, predation and parasitism removed or greatly lessened at the new insular habitat, a dense population is virtually inevitable. From an insular colony, small in area by geographic restriction, but dense in structure, and exposed to intensive contacts with commercial transportation, the probability that new propagules will be transported to new colonial sites is enormously increased over what it originally was in the continental distribution. It seems likely that this is the usual pattern of dispersal of potent tramp species among the smaller insects and certain other invertebrates (the notorious and well-studied giant snail, Achat ina fulica, for instance; Be- quaert, 1950). The study of this problem in the ants is a fascin- ating one, but very difficult due to the present very great pro- portion of unrecognized synonyms among the tramp species. From the fragmentary information we have at present, the tramp ants of the tropics and subtropics seem, as has already been mentioned, to be predominantly African in origin. Perhaps, as seems very likely in the better-known but restricted case of the dacetines, this apparent predominance of African forms reflects a relatively potent evolutionary situation centering in the Dark Continent. In the case of the dacetines, Africa cer- tainly would appear to have been the chief center of major- group evolution, at least during the more recent geologic past, as well as the present. The largest and most generalized member of the rogeri group is S. grandidieri Forel, from Madagascar. This species shows many similarities in general habitus (as well as in details like the median longitudinal cephalic sulcus, the antennal scrobe posteriorly limited near the compound eye, the preocular notch, 8 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the size, and general head shape) to generalized members of the Indo- Australian fauna (chyzeri group) and to American species like 8. mandibular).* Fr. Smith and S. precava Brown. The closest relationships of 8. grcmdidieri are, however, clearly with the African members of the genus. The generalized rogcri group characters include well developed preocular notches combined with rather large eyes, fairly long mandibles with the usual pair of apical teeth forming a fork, and two additional preapical teeth, and a rather strikingly depressed posterior part of the mesonotum. Important tendencies of specialization within the group in- clude, in different lines, reduction of eye size and loss of the preocular notch, reduction of body size, shortening of mandibles and antennae, and reduction even to loss of the more distal of the two preapical teeth. The reduction of the distal preapical tooth is unique among dacetines in that it takes place asym- metrically, with the tooth on the left mandible diminishing more strongly than that on the right. The tooth on the left may even disappear entirely, while that on the right persists in a more or less reduced state in all but one species (irrorata) . All stages in the reduction of the distal preapical teeth are found among rogcri group species, and each stage furnishes a useful species- constant character. The eye-notch character and the dentitional asymmetry have received scant attention from most authors, especially Santschi, and in consequence these features are frequently not even men- tioned in past descriptions. The published figures of African Strumigenys, as well as the mensural citations, are also largely untrustworthy. In the matter to follow, I have not tried to correct by specific mention all of the numerous published errors of detail. Instead, there are emphasized below the essential characters, and especially the measurements, of all species studied at first hand. Special attention has been given to the construction of the key. Type material of all species treated has been directly examined unless otherwise specifically men- tioned. The species included here that were described prior to 1922 have been catalogued (under Strumigenys s. str.) in the follow- ing references: Emery, 1922, Genera Insectorum, Fasc. 174, pp. brown : ant genus Strumigenys fred. smith 9 320, 322; Wheeler, 1922, Bull. Amer. Mus. Nat. Hist., 45: 917- 918, 1034. In these lists are included Strumigenys reticulata Stitz and S. ludovici Forel, both of which have since been transferred to Serrastruma (Brown, 1952). S. reticulata is a synonym of Ser- rastruma lujae (Forel), while ludovici is perhaps a prior name for Serrastruma alluaucli (Santschi). Key to the Species of Strumigenys of the Ethiopian and Malagasy Regions, Based on the Worker Caste Notes on the use of the key. Since most of the species are still known only from scanty material, this key should be taken only as a preliminary guide. Larger series may well reveal that the allowances I have made for potential infraspecific variation, while generous, may in some cases be transgressed. It should also be emphasized that the mere fact that a given specimen does not readily run out in the couplets below is no assurance that the would-be identifier has discovered an undescribed species. It is regrettable that past authors have described in error certain "new" forms solely on the basis that they would not run out in the existing keys. On the other hand, it does seem likely that additional species of Strumigenys beyond those here treated occur in Africa. The identifier using this key must be prepared to measure the dimensions of the head and exposed mandibles to a satisfactory degree of precision. The tolerances of error for these measure- ments are about ±0.01 mm. A good-quality manipulator set in a mechanical stage, under magnification of not less than 60 diameters, is recommended for these measurements. It is also advisable to consult a full discussion of dacetine measuring techniques (Brown, 1953). It goes almost without saying that specimens must be clean and undamaged. Dirt or adhesive lodged in the preocular notches or mandibles can cause serious error, especially at couplets 1, 3 and 10. Furthermore, at couplets 3 and 10 in cases where any doubt occurs, the mandibles should be opened in order to see the dentition properly in at least a portion of any nest series. At couplet 1, the eye-notches should not be confused with the 10 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY antennal scrobes ; the latter are broad, usually elongate sulci running for some distance along the sides of the head in such a position as to readily receive the retracted antennae, while the preocular notches run more or less vertically, perpendicular to the long axis of the head, immediately in front of the compound eyes. Specimens should not be mounted flat on the surface of cards; card mounts obscure the details of periocular structure and mandibular dentition, and hence have been the greatest single source of confusion visiting this genus to date. Point mounts, utilizing fine pennant-shaped pieces of card, are by far preferable. At couplets 3 and 10, the possibility must be considered that rare atavistic specimens may preserve the minute vestige of a left preapical tooth close to the base of the dorsal apical tooth where this (distal) preapical tooth is normally absent in a given species. I have seen one specimen possibly belonging to this category (Bernard ms. new species, couplet 4 and p. 16 ). Such specimens can usually be detected if a part of a normal series, but uniques will always be difficult, and must be checked by means of dimensions and proportions and other characters given in the descriptions. 1. Ventrolateral border of the head receding sharply at the anterior margin of the eye to form a distinct preocular notch or groove; eye oriented more or less anteriorly and usually more or less detached and narrowly rounded in front 2 Ventrolateral border of head not or at most extremely feebly impressed at the anterior margin of the eye; eye oriented entirely laterally, the facets forming a flat or gently convex disc, or the eye reduced to a very few minute facets 7 2. Larger, length of head proper over 1.0 mm.; each mandible with two short, oblique, truncate preapical teeth (Madagascar) . grandidieri Forel Smaller, length of head proper under 1.0 mm.; preapical teeth acute. . .3 3. Left mandible with a single preapical tooth, the distal preapical tooth normally completely lacking (see note above on use of key) 4 Left mandible with two preapical teeth, the distal tooth smaller than the proximal, but still quite distinct and acute 5 4. Larger, length of head proper > 0.70 mm.; CI > 72, MI < 54; upper angles of propodeal lamellae low and more or less rounded (Kenya) .... londianensis (Patrizi) brown : ant genus Strumigenys feed, smith 11 Smaller, length of head proper < 0.60 mm. ; CI < 72, MI > 54 ; upper angles of propodeal lamellae normally dentiform and acute (Congo N. Angola) new species, Bernard ms. 5. Larger, length of head proper > 0.65 mm. ; MI < 42 ; compound eyes exceptionally large and prominent (Transvaal) pretoriae Arnold Smaller, length of head proper < 0.65 mm.; MI 42 or more; compound eyes smaller to moderate in size and prominence 6 6. MI 43-49; mandibular shafts distinctly arcuate (W. Africa to Natal and Angola) rufobrunea Santschi MI 51 or more; mandibular shafts nearly straight (W. Africa; wide- spread in tropics of both hemispheres, especially in the Pacific and W. Indies ; also British greenhouses) rog.eri Emery 7. Normally exposed portions of the antennal scapes very broad, less than three times as long as their greatest width; CI about 90 or slightly more (Uganda) t.etraphanes new species Normally exposed portions of antennal scapes more than three times as long as wide ; CI well under 90, and usually less than 85 8 8. Greatest diameter of compound eye distinctly greater than greatest width of antennal scape 9 Compound eye very small, its greatest diameter less than, or at least not greater than, greatest width of antennal scape 10 9. Head narrower, CI < 73; mandibles longer, MI > 45 (Seychelles Is.; Sao Tome I.) sootti Forel Head broader, CI > 73 ; mandibles shorter, MI < 40 (Natal) marleyi Arnold 10. Left mandible with only a single preapical tooth, the distal preapical tooth normally completely lacking 11 Both mandibles each with two preapical teeth, though in some cases, the distal preapical tooth on one or both sides may be reduced to a minute denticle 1- 11. Right mandible with two preapical teeth, the distal tooth small and normally covered by the dorsal apical tooth of the left mandible at full closure; CI between 70 and 80 (Uganda to Angola) . .dextra new species Eight mandible with only one preapical tooth, the distal preapical teeth of both mandibles lacking; CI > 80 (Zululand) irrorata Santschi 12. Combined length of head and mandibles when closed > 0.85 mm.; CI < 75 ; MI 50 or slightly more (Natal) liavilandi Forel Combined length of head and closed mandibles < 0.85 mm. ; CI > 75 ; MI usually under 50 13 12 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 13. Combined length of head and mandibles 0.70 mm. or less; CI ca. 81; MI ca. 37-38; most hairs on dorsum of head nearly or quite obicular, appearing like round, shining, convex scales or studs (British E. Africa) stygia Santsclii Combined length of head and closed mandibles more than 0.70 mm., usually 0.75 mm. or slightly more; CI 77-80± ; MI usually > 38; hairs on dorsum of head more or less broadened apically, but not orbicular. 14 14. MI > 43; each humerus with a flagellate hair; HL 0.53 mm. or less. (E. and S. Africa) tragaordlii Santschi MI < 43, usually 41 or less; humeri without flagellate hairs; HL usually greater than 0.53 mm. (E. and S. Africa) arnoldi Forel SYSTEMATIC TREATMENT BY SPECIES Stbumigenys grandidieri Forel Strumigenys Grandidieri Forel, 1892, Ann. Soc. Ent. Belg., 36: 517, worker. Type locality: Andrangoloaka Forest, Madagascar. Syntypes: Mus. Hist. Nat., Geneva ; Mus. Comp. Zool. Harvard University. Worker. Descriptive notes are based on two very similar syn- types in the Museum of Comparative Zoology, one of which was measured: TL 5.02, HL 1.33, ML 0.69, WL 1.30 mm. ; CI 76, MI 52. Because of its size, this species could hardly be confused with any other African-Malagasy Strumigenys. Head massive, deeply and broadly excised behind, with a distinct, narrow median dorsal sulcus running from clypeus to posterior excision. Antennal scrobes ending, or at least becoming extremely indistinct, at about the level of the eyes. Eyes large, but not so large relatively as in pretoriae, convex, prospicient; preocular notch broad and deep, involving the dorsolateral cephalic border. Mandibles broad, robust, slightly depressed, inserted close together and slightly diverging at full closure. Apical fork of two stout acute subequal teeth, without intercalary tooth or denticle. Each mandible with two short, truncate preapical teeth, directed obliquely anteriorly, subequal in size and scarcely longer than broad, very close to the apex and to each other. Alitrunk slender, pronotum convex, its anterior margin nar- rowly rounded and sharply marginate, without humeral angles. Mesonotum reasonably distinct, although the promesonotal suture is obsolescent ; posterior mesonotum forming a long slope down BROWN : ANT GENUS StrumigemjS FRED. SMITH 13 to the strong metanotal groove, from which the nearly plane, platform-like propodeal dorsum rises slightly but rather abruptly; dorsum and declivity of propodeum meeting at ap- proximately a right angle. Propodeal teeth long, spiniform, strongly elevated and divergent ; lateral borders of declivity without infradental lamellae, but with three or four fine vertical rugules on each side. Petiolar node shorter than its slender peduncle, about as broad as long seen from above and narrowly rounded above as seen in lateral view profile ; petiolar appendages reduced to in- significant vestiges. Postpetiole transversely elliptical, strongly convex, half again as broad as the petiolar node and less than half as broad as the gaster, with only strongly reduced ventral appendages. Gaster with 9-13 widely spaced, distinct basal costulae extending about 1/5 the length of the basal segment. Gaster otherwise and most of mandibles smooth and shining. Remainder of body, including pleura and both nodes, densely punctulate, opaque. Pine superimposed regulation on dorsum of head and alitrunk, most distinct on anterior pronotal margin. Ground pilosity sparse, short, very narrowly spatulate and subappressed on head ; available specimens may be partially rubbed. Row of 4 (or 6) longer, suberect, apically very feebly spatulate hairs bordering occipital excision; another pair on vertex. Scape hairs fine, curved apicad; clypeal border hairs narrowly spatulate, curved mesad. A pair of long, erect, very weakly clavate hairs on postpetiole, and a few on posterior half of gaster. Color light ferruginous (possibly somewhat faded) ; mandibles and head lighter and more yellowish; vertex trans- versely, nodes and gaster slightly darker than the rest. Female and male unknown to me. As already mentioned in my introductory remarks, 8. grand- idieri appears to be the most generalized and most primitive member of the rogeri group. Furthermore, it has a "primitive look" backed by size, head structure, form of alitrunk and propodeal armament having much in common with presumed primitive forms of other faunal groups. The species remains known only from the Madagascan type collection, and is the only species of the genus so far recorded from Madagascar proper. 14 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Strumigenys londianensis (Patrizi) Prosoopomyrmex londianensis Patrizi, 1946, Boll. 1st. Ent. Univ. Bologna, 15: 295, figs. 1, 2, worker. Type locality: Londiani, Kenya (by tau- tonymy). Additional orig. loc. : Mau Forest, Kenya. Holotype and para- types: 1st. Ent. Univ. Bologna; paratypes in Brit. Mus. (Nat. Hist.). Strumigenys (Proscopomyrmex) londianensis, Arnold, 1948, Occas. Pap. Nat. Mus. S. Bhodesia, 2 (14) : 227. Proscopomyrmex? londianensis, Brown, 1948, Trans. Amer. Ent. Soc., 74: 128. Strumigenys londianensis, Brown, 1949, Mushi, 20: 15. Worker. Notes based on a single paratype from Londiani, by courtesy of Prof. Grandi. TL 4.2, HL 0.88, ML 0.45, WL 0.88 ram. ; CI 75, MI 51. Patrizi gave a short description and large, detailed figures in dorsal and lateral views. In addition to marked contradictions involving dimensions and proportions between the description and figures, and between these and the paratype before me, I note the following : 1. In the paratype, the mandibles are shorter and heavier than as figured, not so strongly arcuate, and the teeth shorter and thicker and set more nearly at a right angle to the shafts. On the right mandible, a short distal preapical tooth is present, but is small and partly covered by the dorsal apical tooth of the left mandible at full closure. The left mandible lacks a distal preapical tooth, but the proximal preapical tooth is well de- veloped on both mandibles. 2. In the paratype, the compound eyes are rounded anteriorly as seen from above, and are not pointed ; from lateral view, the eyes appear roughly circular. In front of the eyes is a well- marked vertical groove, extending even into the dorsolateral cephalic borders above and well across the postbuccal surface below. Greatest diameter of eye (ca. 0.07 mm.) greater than maximum width of scape (ca. 0.05 mm.). 3. In the paratype, the posterior descending mesonotal slope is gently and evenly concave, not interrupted by a suture-like impression as shown in Patrizi 's figure 2. 4. In the same figure, the propodeal lamella is much too abruptly terminated ventrally, as is clear even without reference to a specimen. In the paratype, the dorsal angle of the lamella brown : ant genus Strumigenys fred. smith 15 is lower and more blunt than as shown in the figure, and the excision between upper and lower angles is concavely rounded, not subangular. 5. In the paratype, numerous short, spaced basal costulae meet the anterior border of the first gastric segment. Also, the petiole of the paratype bears feeble vestiges of posterolateral spongiform appendages left out in the figure, and there are small, inconspicuous, reclinate-spatulate hairs on the dorsum of the head of the paratype. The postpetiolar disc is smooth and shining, but dirty. This species is closely related to 8. rufobrunea and Strumi- genys new species of Bernard ms., but is larger than either of these and differs in numerous details, especially the very different propodeal lamellae. It is a forest species still known only from the two original Kenyan collections. Strumigenys pretoriae Arnold Strumigenys (Proscopomyrmex) pretoriae Arnold, 1949, Oecas. Pap. Nat. Mus. S. Bhodesia, 2 (15): 267, fig. 8, worker. Type locality: Pretoria, Transvaal. Holotype : Nat. Mus. S. Rhodesia, Bulawayo ; paratype in Agric. Res. Inst., Univ. Pretoria. Worker. Notes based on the single paratype worker, sent through the courtesy of Prof. J. C. Faure for my study. TL 2.8, HL 0.73, ML 0.29, WL 0.70 mm. ; CI 73, MI 39. Mandibles stout, gently arcuate, with dentition as in Figure la. Eyes exception- ally large and convex, narrowly rounded anteriorly and sep- arated here from the head by a deep, narrow preocular notch as shown in Figure lb. Posterior mesonotum depressed below level of propodeal dorsum; propodeal lamella with a short but acute, elevated tooth above, convex below (Figure lc). Petiole with a long, narrow peduncle having a spongiform border be- neath ; node broader than long, with moderate posterior spongi- form fringe. Postpetiolar node transverse, smooth and shining at least in the middle. Basigastric costulae distinct, fanning from bilateral sources and extending almost to the midlength of the basal segment. Head and promesonotum with moderately abundant small, reclinate spoon-shaped hairs ; gastric dorsum with sparse erect remiform hairs in transverse rows. Color light ferruginous. 16 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The very large eyes mark this form at once among the Old World Strwmigenys, and rival those of the Amazonian S. tococae Wheeler. E. K. Hartwig collected the types while "beating for thrips, " so it is possible that, like S. tococae, the eye develop- ment may be correlated with strongly epigaeic foraging habits, or perhaps even a snbarboreal mode of life. Such a correlation holds well for a number of neotropical dacetine ants. Figure 1. Strumigewys pretoriae Arnold, paratype worker; a. apex of mandible; b, eye and periocular region, dorsal view; c, outline of propodeum, lateral view; all to same scale. Strumigenys new species, Bernard ms. Worker. Notes based on two workers collected at different localities in the Belgian Congo by N. A. Weber (see below). This distinctive little species is widespread in the Congo drain- age system and is apparently common in the southern part of this region, as shown by several small series loaned by Prof. Bernard from the Machado Angolan collections. Prof. Bernard has in- dicated a desire to describe this species from his more abundant material in his forthcoming work on the Angola ant fauna. TL 1.9-2.0, HL 0.52, ML 0.29-0.31, WL 0.44-0.45 mm.; CI 69-70, MI 57-60. This very slender species has long, narrow, gently arcuate mandibles and small compound eyes with deep, narrow preocular notches, marking its close relationship to rogeri, rufobrimea and lo?idianensis. The apical teeth are long and slender, longer than in rogeri and rufobrunea, and the proximal preapical teeth of both mandibles are also quite well developed. In the case of the distal preapical tooth, however, only that of the right mandible is normally present, and even this is difficult to see at full closure because it is then covered by the dorsal apical tooth of the left mandible. In one of the brown : ant genus Strumigenys FRED. SMITH 17 specimens I have seen, the left mandible bears an extremely minute projection at the proximal side of the base of the dorsal apical tooth, and this projection may be an artifact, a structural defect, or the vestige of a distal preapical tooth. I do not think it wise to mount this specimen for examination by transmitted light until further material is available, as this is one of only two taken by Dr. Weber and deposited in American museums. In any case, the projection is so very small as to be insignificant, and it was not present in any of the remaining samples I have examined. Localities for material examined : BELGIAN CONGO : 10 miles E. Stanleyville, 1 worker (Weber, No. 2225). Beni to Irumu, Ituri Forest, 1 worker (Weber, No. 2129B). ANGOLA: Collections by A. de Barros Machado, all from vegetable debris of the soil in gallery forest of various river tributaries of the Congo system; Nos. 54-5, 1130-29, 1195-24, forest of Luachimo R., near Dundo; No. 408-1, forest of R. Sanga, branch of R. Luachimo, near Dundo ; No. 403-2, forest of R. Tchimana, branch of R, Tchikapa ; No. 1430-20, left bank of R, Kasai, NE corner of Angola. This form is most like S. londianensis, from which it differs very markedly in size, form of propodeal lamellae, and other features. Strumigenys rufobrunea Santschi Strumigenys rufobrunea Santschi, 1914, Boll Lab. Zool. Portici, 8: 373, worker, female. Type locality: Conakry, French Guinea, by present selection. Additional orig. loc. : Olokomeji, Nigeria. Lectotype, by present selection, the female from Conakry in Naturhist. Mus., Basel ; other original syntypes deposited with lectotype. Strumigenys (Proscopomyrmex) faurei Arnold, 1948, Occas. Pap. Nat. Mus. S. Ehodesia, 2 (14) : 226, figs. 12, 12a, worker, female. Type locality (orig. designation): Sordwana, Zululand. Other orig. Iocs.: Richards Bay and St. Lucia Lake, Zululand. Holotype and paratypes: Nat. Mus. S. Rhodesia, Bulawayo. Paratypes: Agric. Res. Inst., Univ. Pretoria; Mus. Comp. Zool. Harvard Univ., etc. NEW SYNONYMY. Strumigenys petiolata Bernard, 1953 (1952), Mem. Inst. Fr. Afr. Noire, 19 (1): 254, fig. 14 H, I, J, worker. Type locality: Mt. Nimba, French Guinea, 700 M, in termitary in forest. Holotype: Mus. Hist. Nat. Paris. NEW SYNONYMY. 18 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY 8. rufobrunea is a common and highly variable species ranging widely in tropical Africa. When Santschi first described the species, he overlooked the important eye notches, thus mislead- ing Arnold, who described the same form more accurately under the name faurei. The recent description and figures of 8. petiolata Bernard are in several respects incomplete and self- contradictory, and also fail to show eye notches. I believe this omission is due to the deplorably poor method of card-mounting practiced on ant specimens in many European museums; many of the choice samples sent me from the Old World have been rendered worthless by semi-immersion in adhesive that is virtu- ally impossible to remove in any solvent without damage to pilosity and other important details. One is frequently aston- ished at the gross inaccuracy of what are supposed to be scientific descriptions, but this astonishment is both explained and magni- fied anew when the slovenly preparations upon which these de- scriptions are so coolly based become available for re-examina- tion. It must also be admitted that some American preparations on the more desirable point mounts are so poorly done as to be equally worthless. In my opinion, no descriptive myrmecological work, however pretentious, can be trusted if the material upon which it is based is improperly prepared. To return to S. petiolata : Bernard considered this form closely related to 8. reticulata Stitz (actually a synonym of Serrastruma lujae), to S. hindenburgi Forel (an Argentinian species as dis- similar from petiolata as its geographical remoteness suggests), and to 8. grandidieri. Nothing is said about the really closely related forms like londianensis, rogeri, and the senior synonyms rufobrunea and faurei. Though I have not seen the petiolata type, and in spite of the confused circumstances surrounding its description, I believe that it is only a fairly large specimen of rufobrunea well within the variation outlined below. I have been able to compare directly the type of S. rufobrunea, work- ers and females of the type series of S. faurei, and a large amount of other material from diverse localities, and the notes below summarize the specific characters of the two female castes. Special emphasis is placed on variation within the species. Worker. Measurements are based on 25 workers from at least 11 separate colonies from 8 localities listed below, excepting brown : ant genus Strumigenys fred. smith 19 the Angola samples. HL 0.50-0.62, ML 0.22-0.30 mm.; CI 75-81, MI 43-49. Examples from single collections : Khor Aba, Anglo- Egyptian Sudan (Weber, 1470), HL 0.62, ML 0.30 mm.; CI 81, MI 48, one worker. Gross Batanga, Cameroon (Schwab), HL 0.60, ML 0.30 mm. ; CI 80, MI 49, one worker. 8. faurei type series, St. Lucia Lake and Richards Bay, Zululand, HL 0.52-0.56, ML 0.25-0.26 mm. ; CI 75-79, MI 45-48, 12 workers. Ituri Forest, Belgian Congo (Weber, 2124, 2129A), HL 0.50-0.51, ML 0.22- 0.24 mm.; CI 76, MI 43-47, 4 workers. Haut Mbomu, French Equatorial Africa (Weber, 2187, 2192), HL 0.50, ML 0.24 mm.; CI 76-77, MI 48, two workers. Same locality (Weber, 2188), HL 0.60, ML 0.28 mm.; CI 76, MI 47, one worker. Fort Portal, Uganda (Weber, 2095), HL 0.55, ML 0.25 mm.; CI 76, MI 45, one worker. Although none was measured from the six series collected by Machado in the Congo and Angola (see below), the workers here show a similar range of variation so far as can be determined by simple inspection. The larger workers often have broader heads and deeper, narrower, more distinct eye notches, but exceptions occur both ways. Larger workers also frequently have the pronotum evenly punctulate, while smaller ones usually have feeble longitudinal rugulation predominating on the pronotum ; all intergrades occur. The postpetiolar disc varies from smooth and shining in most series to finely longitudinally striate in the faurei types and certain Angolan samples; here again, intergrades are found. The proximal and distal preapical teeth are present on both mandibles, the distal being considerably smaller than the proxi- mal. The shafts of the mandibles are gently but very distinctly arcuate, clearly more strongly so than in rogeri. Color varies from light to deep ferruginous, and certain Angolan samples are nearly black. Female. Lectotype, HL 0.52, ML 0.23 mm. ; CI 73, MI 44. A single dealate specimen from Kawanda Experiment Station, near Kampala, Uganda (soil sample under elephant grass, Pennisetum purpureum (G. Salt), HL 0.60, ML 0.28 mm.; CI 78, MI 47. Two dealates from faurei type series, allonidal, HL 0.57-0.58, ML 0.27-0.28 mm.; CI 80-81, MI 46-47. Total ranges for the above 4 females; HL 0.52-0.60, ML 0.23-0.28 mm.; CI 78-81, MI 44-47. Variation in these and a few Angolan females parallels that of the workers. 20 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Male unknown. In addition to the localities cited above, six series have arrived after the main work on this species was completed, sent by Prof. Bernard from the collections of Machado, all taken in vegetable detritus in gallery forest in the southern Congo drainage area : BELGIAN CONGO, between Tchikapa and Luluabourg, 100 km. east of Tchikapa. Gallery forests of the rivers Luachimo, Tchimana and Sanga, mostly near Dundo, ANGOLA ; Machado Nos. 180-19, 403-2, 408-1, 1195-24, 1210-12, 1248-29. From the available ecological data, it would seem that 8. rufobrunea can occupy a great range of habitats, but it appears most often to be found in rainforest or gallery forest, or tree- shaded parts of the savannah. It is the commonest member of the genus in Africa, if collection frequency is any guide. This species could be confused only with 8. rogeri, but rogeri has longer, straighter mandibles and differs in other ways also. Strumigenys rogeri Emery Pyiamica giindlachi Roger, 1862, Berlin, ent. Zeitschr., 6: 253, fig. 18a, female, nee worker. Type locality: Cuba. Types in Zool. Mus. Univ., Berlin. Strumigenys rogeri Emery, 1890, Bull. Soc. Ent. Ital., 22: 68, pi. 7, fig. 6, worker. Type locality : St. Thomas, West Indies. Holotype : Mus. Civ. Stor. Nat., Genoa. Forel, 1893, Trans. Ent. Soc. Loud., p. 378, worker, female, St. Vincent, W.I., biology. Wheeler, 1908, Bull. Amer. Mus. Nat. Hist., 24: 145, pi. 12, fig. 13, worker, Puerto Rico. 1913, Bull. Mus. Comp. Zool., 54: 496. Wheeler and Mann, 1914, Bull. Amer. Mus. Nat. Hist,, 33: 40, Haiti. Menozzi and Russo, 1930, Boll. Lab. Zool. Portici, 24: 163, Moca, Dominican Republic. Santschi, 1931, Rev. Ent., Rio de Janeiro, 1: 275, Pinar del Rio, Cuba. M. R. Smith, 1936, Jour. Agr. Univ. Puerto Rico, 20: 856, fig. 12, Puerto Rico, ecology. Donisthorpe, 1915, British Ants, p. 341 (1927, 2nd ed., p. 393), synonymy and refs., summary records from greenhouses in British Isles. Brown, 1948, Trans. Amer. Ent. Soc, 74: 113, Fiji, Hawaii. Other locality records from West Indies occur in the literature, but are repetitious and will not be cited here. Strumigenys incisa Godfrey, 1907, Notes R. Bot. Garden Edinburgh, 17: 102, fig., worker. Type locality: Hothouse, R. Bot. Garden, Edinburgh. Types: locality unknown. Synonymy teste Donisthorpe, 1915, loo. cit. Strumigenys sulfurea Santschi, 1915, Ann. Soc, Ent. France, 84: 261, worker. Type locality: Samkita, Gabon. Syntypes: Naturh. Mus., Basel. NEW SYNONYMY. brown : ant genus Strumigenys fred. smith 21 Worker. One syntype had an HL of 0.57 mm. ; CT 72, MI 53. Ten workers from various West Indian and Hawaiian localities : HL 0.58-0.62 mm., ML 0.30-0.34 mm. ; CI 70-74, MI 55. Except for the very slightly smaller size, the syntype agrees very well with abundant material available to me from the West Indies and various Pacific islands. The extra-African material, as would be expected of a tramp species stemming from a single female, or at least from a restricted-locality stock, is unusually uniform. There can be little doubt of the origin of S. rogeri in Africa. It has no close relatives in the New World endemic fauna, and it is clearly very close to 8. rufobrunea, S. londian- ensis and other African species, previous accounts to the con- trary notwithstanding. 8. rogeri is so well known, and so frequently described and figured in the literature, that no new description is required here. To emphasize a few points : the eyes are small, but the preocular notches are distinct, deep and narrow. Mandibles robust, very nearly straight, lying very close together when closed, the preapical teeth two on each mandible, stout and moderate in length, close to apex, the distal tooth much the shorter of the two. Body and head slender; ground pilosity arranged as in related species, inconspicuous. A pair of slender, erect remiform hairs on vertex, a transverse row of four on occiput, one on each side of the mesonotum, and several on the nodes and gastric dorsum ; humeral angles each with a contorted flagellate hair. Mesonotum sharply depressed posteriorly, the low portion forming a continuous convexity with the propodeal dorsum. Propodeal lamellae distinct, with short, acute teeth above. Postpetiole more or less smooth and shining. Gastric dorsum with 5 or 6 coarse but not long costulae on each side at base. Color uniform light or yellowish ferruginous. The female of S. rogeri differs from its worker in the usual ways, and the compound eyes are unusually large. It can be dis- tinguished from the 8. rufobrunea female by means of its slightly narrower head and longer, straighter mandibles, which are a little more than half as long as the head proper. In addition to specimens from most of the localities cited in the synonymy, I have seen material from Micronesia (H. S. Dybas), Jamaica (H. B. Mills), Trinidad (N. A. Weber). 22 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY I am fortunate also in having important notes on the biology of 8. rogeri made by E. 0. Wilson during his stay in Cuba in the summer of 1953 ; these have been turned over to me in their entirety, and I have abstracted them for use here so that the habits of at least one member of the group should be illuminated along with the routine revisionary data. It should be borne in mind that the following notes were made in a habitat other than the native one for the species, and that the majority of feeding notes of this kind necessarily come from observations made in artificial circumstances. Wilson took his observation colony at San Vicente, Pinar del Rio, Cuba, from a small nest under a rotten limb lying on well shaded ground. The galleries extended into the wood itself. Transferred to a small plaster observation nest, the workers readily captured numbers of entomobryoid collembolans prof- fered ; campodeids up to four times the length of the ants were also always accepted and, like the entomobryoids, were fed to the larvae. Also observed to be chewed by the larvae after cap- ture were a small psocopteran, a small ichneumonid wasp, and a small, injured embiopteran that had previously been ignored by a colony of Smithistrwna nigrescens Wheeler. A symphylan and a pseudoscorpion, one each, were accepted and eaten by the larvae, but only after lengthy contact with the ants. Other specimens of these last two groups seem to have been refused by the larvae after capture on some occasions. A small polydes- mid millipede was also captured, but soon rejected by larvae and workers. Consistently avoided or ignored when offered in the intimate confines of the observation nest were mites, nasute and other termites, small isopods, poduroid collembolans, adult staphylinid and sylvanid beetles of small size, a small campodei- form beetle larva, and dead mosquitoes, though the beetles mentioned disappeared from the nest and may possibly have been eaten. Drosophila adults were caught by the adults, but later discarded. Entomobryoid collembolans seemed to be the usual and pre- ferred prey fed to the larvae, although campodeids were never refused. In feeding habits, therefore, 8. rogeri follows the generic habit of collembolan predation but, like some other wide- spread dacetine species, it will also accept a variety of other brown : ant genus Strumigenys fred. smith 23 small arthropods, particularly campodeids, when available. In hunting, or when disturbed, the workers and females open the mandibles to slightly more than 180°. In view of the fact that collembolan preclation is now known to furnish the basic food supply for widely differing Strumigenys species in all three of the major faunas of the world, it seems reasonable to assume that it is general for the genus and also that it is a primitive habit for the genus in the phylogenetic sense. The cases in which the exclusive diet of entomobryoids, isotomoids and symphypleonans is exceeded, show an erratic choice of additional prey, and it does not seem possible at the moment to predict for a given species just what additional kinds of arthropods may be taken over and above the collembolan groups named. This circumstance suggests that widened prey preferences may be secondarily acquired. Relatively common and successful species seem to accept a wider variety of non- collembolan prey than do the ants which are rarer and ecolog- ically more restricted. The rejection of poduroids as prey seems, however, to be nearly or quite universal among Strumigenys and the other dacetine genera so far studied. The generality of collembolan predation can probably be extended to all the African-Malagasy species of Strumigenys. Strumigenys scotti Forel Strumigenys scotti Forel, 1912, Trans. Linn. Soc. Lond., Zool., (2) 15: 159, worker. Type locality: Mare aus Cochons, ca. 1000 ft., Silhouette, Sey- chelles Is. Syntypes: Mus. Hist. Nat., Geneva. Worker. One syntype examined through the courtesy of Dr. Ch. Ferriere: TL 2.6, HL 0.67, ML 0.31, WL 0.66 mm.; CI 66, MI 47 ; scape L 0.40 mm. Twelve workers selected from two large nest series from Makambrera, ea. 1300 M., Sao Tome Island (B. Malkin) : HL 0.59-0.63, ML 0.30-0.32 mm.; CI 68-71, MI 49-51. The differences in size and proportions between the two samples available are not excessive when one considers that the first is limited to a single example ; furthermore, the localities, one on an island in the Gulf of Guinea and the other in the Indian Ocean, are probably secondary ones populated within recent times from the African mainland. As in the case of 24 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY S. rogo-i, such insular populations are at once more limited in their variability and denser in their structure than the hypo- thetical parent continental stock, the latter remaining uncollected to the present day. Though I was not able to compare directly the syntype of 8. scotti with the Sao Tome sample because the latter arrived after the type had been returned to Switzerland, my extensive notes on the type reveal no significant differences of the sort that usually distinguish species in this group. In this case offshore colonization may well have proceeded from two dif- ferent segments of the mainland population. 8. scotti is a medium-sized species without preocular notches. It is similar to S. havilandi in general size and appearance, but differs in a number of details, the most readily apparent being the much larger eyes in scotti, especially as compared to the more slender antennal scapes (see key). The mandibles are incurved at their bases and have well developed distal and proximal pre- apical teeth, though the distals are shorter than the proximals. Posterior mesonotum depressed, continuous with plane or feebly convex propodeal dorsum except for a feebly marked metanotal groove. Propodeal teeth sharp, strongly elevated, with narrow, concave lamellae beneath. Petiole with a rather long peduncle having a narrow ventral spongiform strip ; posterior appendages of node nearly obsolete. Postpetiole transverse-elliptical, convex above, smooth and shin- ing, its ventral spongiform appendages moderate in bulk. Basi- gastric costulae short, indistinct, grouped bilaterally. Ground pilosity of head composed of inconspicuous narrow spatulate reclinate hairs, generally distributed except for the occipital 2/5, where they are much reduced and partly replaced by 6 slender, erect, blunt or subclavate specialized hairs, of which there are also a pair on the mesonotum, a smaller pair on the posterior mesonotum, and groups becoming more numerous on the nodes and gaster. Humeri each with a long, finely flagelliform hair. A female from one of the Makambrera colonies : HL 0.61, ML 0.31, forewing L 2.3 mm. ; CI 72, MI 50. Eyes very large. Strumigenys marleyi Arnold Strumigenys havilandi race marleyi Arnold, 1914, Proc. Ehodcsia Sci. Assoc, 13: 31, pi., fig. 10, worker. Type locality. Durban, Natal. Syntypes: brown : ant genus Strumigenys fred. smith 25 Nat. Mus. S. Rhodesia, Bulawayo. 1917, Ann. S. Afr. Mus., 14: 378, worker. Strumigenys marleyi Arnold, 1926, Ann. S. Afr. Mus., 23: 286, worker. Worker. Two syntypes, TL 2.4-2.5, IIL 0.62, ML 0.22, WL 0.60 mm. ; CI 76-77, MI 35-36. Related to 8. scotti, but mandibles much shorter and more broadened at the base, recalling-, in less extreme version, certain species of the Labidogenys complex of the Indo-Australian Region (S. biroi Emery, 8. emdeni Forel) ; the trend is surely a convergent one. Distal preapical teeth much smaller than the proximals, the right slightly larger than the left. Apical fork with ventral tooth slightly shorter than dorsal and bearing on its ventral basal surface a small acute adventitious tooth and a minute intermediate denticle. Arnold's figure is highly diagrammatic, and shows a small median clypeal protuberance corresponding to the approximate position of the projecting labral lobes. Eyes weakly convex, with 6-7 facets in the greatest diameter, the diameter being distinctly greater than the maximum width of the scape. Head and promesonotum covered with small inverted-spoon- shaped hairs, generally distributed dorsally, but no specialized erect hairs. Sparse, short, posteriorly inclined remiform hairs arranged symmetrically on nodes and gastric dorsum. S. marleyi is known only from the type collection, made by Arnold in a nest of Pheidole punctulata Mayr. This Strumigenys is similar to S. scotti in overall size and form, and in the size of the eyes ; on the other side, it seems related by tendencies in the development of mandibles and pilosity to the smaller form, S. arnoldi. Strumigenys havtlandi Forel Strumigenys havilandi Forel, 1905, Mitt, naturh. Mus., Hamburg, 22: 13, nota, worker. Type locality: "Natal, 5300 feet." Syntypes: Mus. Hist. Nat., Geneva ; Natal Mus., Durban ; Mus. C'omp. Zool. Harvard Univ. Santschi, 1913, Bull. Soc. Ent. France, p. 1257, worker, in key. Arnold, 1917, Ann. S. Afr. Mus., 14: 377, worker Worker. Two syntypes, TL 2.5-2.8, HL 0.62-0.65, ML 0.33-0.34, WL 0.67-0.70 mm. ; CI 71-73, MI 52-53. Mandibles very feebly arcuate, their shafts noticeably tapered apicad ; both preapical 26 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY teeth of each mandible very slender, long- and acute, only a little shorter than dorsal apical tooth. Compound eye minute, scarcely half as wide as the maximum width of the robust antennal scape. Alitrunk in profile very shallowly concave in the posterior mesonotal region, its outline almost straight. Hairs shorter, broader at apices, more numerous and more generally distributed on posterior cephalic dorsum than in 8. scotti. In addition to the usual sparse erect spatulate hairs, the gastric dorsum bears sparsely distributed short, linear-spatulate appressed hairs. I have not seen the sexual forms of this species, or records other than the original one. Strumigenys arnoloi Forel Strumigenys Arnoldi Forel, 1U13, Ann. Soc. Ent. Belg., 57: 114, worker. • Type Locality: Bulawayo, S. Rhodesia. Syntypes: Nat. Mus. S. Rho- desia, Bulawayo; Mus. Hist. Nat., Geneva'? Strvmigenys avnoldi Arnold, 1917, Ann. S. At'r. Mus., 14: 376, pi. 8, figs. 117, 117 ad, worker. I have not seen type material, but have examined a specimen from Pretoria (J. C. Faure), determined by Dr. Arnold, the original collector, and thus presumed authentic. Also available are 3 workers and a dealate female, preserved rather poorly, from Eldoret, Kenya (S. Patrizi). Tbese two samples are closely similar despite the geographical separation. Worker: TL 2.1-2.2, HL 0.54-0.55, ML 0.21-0.22, WL 0.54-0.55 mm.; CI 77-78, MI 38-41. Female: HL 0.60, ML 0.24 mm.; CI 76, MI 39. S. arnoldi is similar to #. tragaordhi, but differs in having a longer, slightly narrower head proper, while its mandibles are both relatively and absolutely shorter. The spoon-shaped pilosity of the anterior cephalic dorsum is broader and more conspicu- ous, and extends more abundantly to the promesonotum. In these samples of arnoldi, the only specialized erect hairs of the alitrunk are one clavate pair astride the mesonotum ; no special- ized humeral hairs. Dr. Arnold took the types of this species under a stone in a nest of Bothroponera krugeri Forel. Strumigenys tragaordhi Santschi Strvmigenys Tragaordhi Santschi, li)13, Bull. Soc. Ent. France, p. 2.17, worker, original diagnosis in key. brown : ant genus Strumigenys fred. smith 27 Strumigenys traegaordhi Santsehi, 1914, Medd. GSteborgs Mus. Zool. Afd., 3: 28, fig. 4, worker. Type locality: Pietermaritzburg, Natal, by present selection. Additional orig. loc: Sweetwaters, Natal. Syntypes : Naturh. Mus., Basel. Worker. Two lectotopic syntypes, courtesy of Prof. Ed. Handschin, HL 0.51-0.52, ML 0.23-0.25 mm. ; CI 79-80, MI 45-48. This is a rather " average-looking" small Strumigenys. Mandi- hles weakly arcuate, somewhat broader than as shown in Sant- sehi 's figure and gently tapered from base toward apex, enclosing a large oval space at full closure. Distal preapical teeth of both mandibles very small, that of the right slightly larger. Ground pilosity of head composed of narrow inverted-spoon-shaped hairs distributed over the entire dorsal surface, but becoming small and inconspicuous on the extreme occiput. Vertex with a pair of slender, curved, erect, remiform hairs, and a transverse row of four of these on the posterior occiput. Exposed scape L 0.29 mm. ; funiculus L 0.44 in the larger of the two syntypes. Alitrunk slender, resembling that of stygia, but promesonotum slightly more convex, propodeal dorsum straight in profile. A flagellate hair on each humeral angle ; mesonotum straddled by two pairs of erect remiform hairs, and the usual sparse remi- form pilosity on nodes and gaster. Postpetiolar disc smooth and shining. Color light ferruginous, head very slightly darker. Various authors have recorded this species from widely sep- arated East African localities, but the determinations remain unconfirmed. Strumigenys dextra new species Holotype worker. TL 1.6, HL 0.41, ML 0.17, WL 0.40 mm.; CI 76, MI 42. In general size and appearance resembling S. amolcli and 8. tragaordhi, but a little smaller even than the latter. Eyes very small, almost but not quite as wide as the maximum width of the antennal scape, feebly convex and laterospicient, without a preocular notch or groove. Mandibles gently arcuate, gradually and weakly tapered from base toward apex; dorsal apical tooth decidedly longer than ventral apical and proximal preapical ; all of these teeth well developed, slender and acute. Distal preapical tooth present on the right mandible only, small, concealed at full closure by the dorsal apical tooth 28 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY of the engaged left mandible ; no trace of distal preapical tooth on left mandible. Scape (L 0.63 mm.) gently curved at base, feebly and gradually incrassate at midlength; funiculus (L 0.90 mm.; apical segment L. 0.61 mm.) with greatly reduced, indistinct second and third segments. Alitrunk in profile gently convex above, with a feeble im- pression in the region of the posterior mesonotum ; metanotal groove virtually obsolete. Propodeum with small but acutely triangular teeth, continued below as feebly concave infradental lamellae. Petiole distinctly and robustly pedunculate ; node with gently sloping, bicarinulate anterior face, rounded above, its spongiform appendages reduced to cariniform vestiges. Post- petiole convex, smooth and shining above, with fairly well de- veloped ventral appendages. Gastric dorsum smooth, with a few coarse costulae extending over about the basal quarter of the first segment. Dorsum of head, mesonotum, propodeum and petiole coarsely retieulopunctate, opaque. Pronotum rather coarsely longitudinally rugulose, with broad interspaces weakly shining; posterior sides of alitrunk smooth, shining; appendages, including mandibles, finely and superficially sculptured, more or less opaque. Ground pilosity of head reduced and inconspicuous, except for a triple row of anteriorly curved spoon-shaped hairs along each dorsolateral cephalic margin from frontal lobe to posterior limit of scrobe. A similar single row, curved apicad, on the anterior border of each scape, and a few hairs forming a clypeal fringe. Ground pilosity reduced, scanty and inconspicuous. Specialized erect hairs remiform to claviform : one pair on ver- tex ; a curved row of 4 along occipital border ; one pair straddling mesonotum; a pair on each node; about six transverse rows of 4 each on gastric dorsum, smaller toward gastric apex. Legs and gula with fine, short, flattened reclinate pilosity. Color yellowish ferruginous throughout. Holotype taken with 6 paratype workers in a soil sample under elephant grass (Pennisetum purpureum) at Kawanda Ex- periment Station, 5 miles north of Kampala, Uganda (G. Salt, No. SS 30). In the same and adjacent samples were taken other presumably hypogaeic ant species {P oner a coeca Santschi, Sole- nopsist sp., and Strumigenys tetraphanes new species). BROWN: ant genus Si nt migenys fred. smith 29 Two additional workers, not paratypes, were seen from Bnsnia, at the Kenya-Uganda boundary (N. A. Weber, No. 2080) and a single worker from Haut Mbomn, Ubangi Shari, French Equa- torial Africa (Weber, No. 2177). Combined measurements for the Kawanda and Bnsnia series, 9 workers, TL 1.5-1.6, IIL 0.39- 0.43, ML 0.17-0.18, WL 0,38-0.43 mm.; CI 71-76, MI 40-43. Internidal variation only very slight. Busnia series with slightly broader hairs on the lateral cephalic borders than in the type series. In addition, I have belatedly examined four small series from among the material collected by A. de Barros Machado in Angola, all from the vegetable detritus of the soil of gallery forests of various tributaries of the Congo system : Camissombo, 87 km. south of Dundo, 850 M. (rain forest), No. 1419-6. Lua- chimo Forest, near Dundo, No. 1248-29. R. Sanga, Dundo, No. 408-1. Left bank R. Kasai, NE corner of Angola, No. 1430-20. This little species, related to arnoldi and the other small forms without eye notches, can be distinguished readily by means of the asymmetrical mandibular dentition, the lateralized develop- ment of the cephalic ground pilosity, and the proportions and relatively coarse sculpture. It is widely distributed and ap- parently rather common in central Africa. Strumigenys stygia Santschi Strumigenys stygia Santschi, 1013, Bull. Soe. Ent. France, p. 257, worker, original diagnosis in key. 1914, Voy. Alluaud et Jeannel Afr. Or., Hym., 2: 113, fig. 20, worker. Type locality: Cave A at Shimoni, coastal British E. Africa. Syntypes : Naturh. Mus., Basel. Worker. Two syntypes, courtesy of Prof. Handschin, HL 0.49- 0.50, ML 0.18-0.19 mm.; CI 81, MI 37-38. See key. Mandibles short, stout, feebly arcuate ; dorsal apical and proximal preapical teeth long and slender ; distal preapical teeth small, that on right larger than the one on the left. Head broad, and set with broadly suborbicular to orbicular, shining, pale whitish, stud-like hairs over its entire dorsal surface. A row of 4 stubby curved suberect hairs along the occipital margin. Promesonotum broad, de- pressed, with an indistinct median carinula ; farther posteriad. mesonotum narrowed and dorsally weakly impressed, forming with propodeal dorsum a gentle convexity, at the summit of 30 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY which is the feebly marked metanotal groove. Each side of mesonotum with a stout clavate hair. Postpetiole superficially sculptured, subopaque to opaque. Anteunal scape (exposed L 0.26 mm.) distinctly broadened at about midlength, its greatest width about 0.05 to 0.06 mm.; sharply narrowed just before apex ; funiculus L 0.36 mm. Color rather uniform yellowish-ferruginous. This species is known to me only from the type series. Strumigenys tetraphanes new species Holotype worker. TL 2.15, HL 0.54, ML 0.19, WL 0.50 mm. ; CI 91, MI 36. Head broad, in general form like that of Quadristruma eury- cera (Emery) (1897, Term. Fiizetek, 20: pi. 14, fig. 17) of New Guinea ; dorsum depressed, only feebly convex. Posterior ex- cision rather deep ; occipital lobes rounded, produced laterally at greatest breadth of head bluntly and subangularly ; anterior to this, the lateral borders converging strongly; preocular laminae feebly converging, almost parallel. Eyes small, only very feebly convex and without preocular notch, situated on ventral scrobe borders at about cephalic midlength. Clypeus much broader than long, obtusely angulate behind, free border broadly rounded, but feebly emarginate and faintly impressed in the center. Mandibles short, robust, resembling those of Strumigenys mocsdryi Emery (loc. cit., fig. 15) in size, form and position at full closure, but not in dentition. Dorsal apical tooth about 0.13 mm. long, very slender, sharp, feebly recurved; ventral apical tooth straight, slender, more than half as long as the dorsal tooth and feebly diverging from it ; no intercalary teeth or denticles. The large spiniform preapical tooth, situated a little distad of midlength of mandible, is about % the length of the dorsal apical tooth and similar in shape and feeble re- curvature. About midway between apical and preapical teeth is a minute but acute denticle (=distal preapical tooth), between 0.01 and 0.02 mm. long, that of the right mandible slightly larger than that of the left. Antennal scape 0.27 mm. long, broadly expanded anteriad, brown : ant genus Strumigenys fred. smith 31 in shape like a naval cocked hat or the space enclosed by a low normal curve and its baseline. The scapal expansion is not so extreme as in Quadristruma curycera, but is considerably more so than in Q. emmae (Emery) or any species of Strumigenys so far described. Maximum breadth of scape about 0.11 mm. ; expanded portion feebly convex dorsally. Funiculus 0.34 mm. long, of which the apical segment occupies slightly more than 3/5 ; basal segment longer than broad, II, III and IV very short, broader than long; IV as long as or slightly longer than II— {-III. Alitrunk distinctive in form. Pronotum and anterior mesono- tum together forming an almost perfect circle as seen from above, this surface rather strongly depressed dorsally and with blunt, overhanging lateral pronotal margins; humeral angles not de- veloped. The surface of this promesonotal disc is divided by the feeble, sulciform, semi-circular promesonotal suture ; pronotum with an anteromedian and bilateral, and the mesonotum with an anterior, dorsal convex area. Behind the promesonotal disc, posterior mesonotum immediately and strongly depressed and narrowed, and as seen from above continued posteriad by pro- podeum to form with it an oblong, parallel-sided section slightly shorter than the promesonotal disc and less than half as broad ; this section as seen from the side forming one continuous, convex dorsal outline, and without lateral margins as seen from above. Metanotal groove visible on dorsum as a darkened line. Propodeal declivity steep ; teeth short, half as long as the distance between the centers of their bases and only moderately acute, continued below by feebly concave infradental lamellae which are almost as broad as the height of the teeth. Petiolar peduncle laterally compressed and longer than its node. Node small, anteroventrally compressed, broader than long; as seen from the side high and narrowly rounded at the dorsal apex. Petiolar appendages reduced to fine vestiges, mid- ventral strip represented only as a low, non-spongiform carina. Postpetiole forming a transverse ellipse, about twice as broad as long and nearly twice as broad as the petiolar node; strongly convex dorsally ; appendages fairly veil developed, but largely restricted to the venter. Gaster slightly narrower than head, somewhat depressed ; anterior spongiform margin medially emarginate. 32 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Gastric costulae coarse, radiating from bilateral origins, about 7 or 8 on each side, those nearest the middle oblique, enclosing a free narrow median triangular area ; longest costulae extending nearly 1/3 the length of the long basal tergite ; gastric dorsum otherwise appearing very finely and superficially reticulate (perhaps due in part to a film of secretion), but still very strongly shining. Sides of posterior alitrunk shining, but feebly roughened as on the gastric surface. Mandibular teeth shining. Remainder of body densely punctulate, opaque ; postpetiole with very feeble superimposed longitudinal rugulosity. Dorsum of head from about midlength to posterior clypeal border thickly set with conspicuous, heavy, suborbicular, sub- appressed, inverted-spoon-shaped hairs ; a double row of the same extends posteriad along each dorsolateral cephalic border as far as the blunt lateral occipital angles, and a single row of the same, large and very conspicuous (6-7 hairs) lines the anterior border of each scape. Similar, but very much smaller and less conspicuous hairs on the clypeus and posterior half of the cephalic dorsum; contrast between the pilosity of anterior and posterior parts of head quite striking, as in certain other species of the S. rogeri group (e.g., scotti). Hairs on anterior clypeal border similar to the foregoing, intermediate in size, subap- pressed, 4 on each side of the middle. Alitrunk with a sparse and altogether insignificant complement of minute appressed hairs, the surface appearing nude except for a prominent pair of short, erect, strongly clavate mesonotal hairs. A pair of the same is directed posteriorly from the postpetiole, and there are about twelve (some possibly broken off) on the gastric dorsum, diminishing in size apicad. Legs and scapes with small ap- pressed spatulate hairs. Underside of head with fine subap- pressed pubescence ; inner borders of mandibles with a few long fine hairs. « Color ferruginous yellow, dorsal surfaces a trifle darker. Holotype a unique worker taken in a soil sample from under elephant grass at Kawanda Experiment Station, 5 miles north of Kampala, Uganda, on Feb. 15, 1949 (G. Salt) with Strumigenys dextra new species (q.v.) and other ants. Holotype deposited in Museum of Comparative Zoology, Harvard University. This aberrant member of the rogeri group is of more than brown : ant genus Strumig eny 's fred. smith 33 usual interest because it shows affinities to the tropicopolitan tramp species Quadristruma emmae (Emery), and in fact nicely links this little form to the rogeri group, and thus to genus Strumig enys. The fact that Q. emmae has only four segments in the antennae still separates it from Strumigenys, but the ex- treme reduction of the second and third funicular segments in S. tetraphanes makes this difference largely an academic one. A reappraisal of the genera of subtribe Strumigeniti may well reveal that Quadristruma Brown (1949) should be merged with Strumigenys, but whatever the eventual decision, it seems plain now that emmae is African, not Papuasian, in origin, and that the rogeri group is directly ancestral. The other Quadristruma species, Q. eurycera (Emery), may not belong to the same stock as emmae, and its resemblances to emmae and tetraphanes may well be convergent ones. The final generic review of the Dacetini will see this relationship altered in some manner. S. tetraphanes is readily separated from the other members of the rogeri group by means of its broad head and short mandi- bles, and above all by means of its very broad antennal scapes. Strumigenys irrorata Santschi Strumigenys irrorata Santschi, 1913, Bull. Soc. Ent. France, p. 257, worker, original diagnosis in key. 1914, Medd. Goteborgs Mus. Zool. Af'd., 3: . 29, fig. 5, worker. Type locality: Lake Sibayi, Zululand. Holotype: Naturh. Mus., Basel. I have not been able to view the unique worker type directly, but Prof. Handschin has very kindly sent me pencil sketches in answer to my inquiries. These confirm that both mandibles lack the distal preapical tooth, although both mandibles are shown as having long proximal preapical teeth. The compound eye is portrayed as small, its greatest diameter less than the greatest scape width ; the drawing shows no obvious trace of a preocular notch. As measured from Santschi 's figure (loc. Git.), which may or may not be accurate in showing proportions, the CI would be about 84 and the MI about 41. Santschi gave the length as 1.5 mm. total, which is probably too low. 34 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY REFERENCES Bequaert, J. C. 1950. Studies in the Achatininae, a group of African land snails. Bull. Mus. Comp. Zool., 105: 1-216, 81 pis. (cf. pp. 50-94). Brown, W. L., Jr. 1948. A preliminary revision of the higher Daeetini (Hymenoptera: Formicidae) . Trans. Amer. Ent. Soc, 74: 101-129, 2 figs. 1952. Revision of the ant genus Serrastruma. Bull. Mus. Comp. Zool., 107: 65-86. 1953. Revisionary studies in the ant tribe Daeetini. Amer. Midi. Naturalist, 50: 1-137, 10 text-figs., 3 pis. (cf. pp. 3-4, 7-15, 16). 1954 (1953). A preliminary report on dacetine ant studies in Aus- tralia. Ann. Ent. Soc. Amer., 46: 465-471. Wilson, E. O. 1954 (1953). The ecology of some North American dacetine ants. Ann. Ent. Soc, Amer., 46: 479-495. Publications Issued by or in ( "onnection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 -- The current volume is Vol. 112. Breviora (octavo) 1952 — No. 35 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. 33 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 - Vol. 1, no. 17 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 -- Published in connection with the Museum. Publication terminated with Vol. 24. These publications 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 ( 'omparative Zoology, ( 'ambridge 38, Massachusetts. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 2 DEEP WATER ELASMOBRANCHS AND CHIMAEROIDS FROM THE NORTHWESTERN ATLANTIC SLOPE by Henry B. Bigelow and William C. Schroeder CAMBRIDGE, MASS., U.8.A. PRINTED FOR THE MUSEUM September, 1954 No. 2 — Deep Water Elasmobranehs and Chimaeroids From the Northwestern Atlantic Slope1 By Henry B. Bigelow and William C. Schroeder CONTENTS PAGE Introduction 38 Elasmobranehs Sharks 40 Family Scyliorhinidae, Cat Sharks 40 Apristurus profundorum 40 Family Squalidae, Spiny Dogfishes 40 Centroscyllhim fabricii, Black Dogfish .... 40 Etmopterus princeps 46 Centroscymnus coelolepis, Portuguese Shark 47 Batoids Family Rajidae, Skates 52 Raja bathyphila 52 Raja fyllae 54 Raja jenseni 60 Raja laevis, Barn Door Skate 62 Raja mollis 63 Raja radiata, Thorny Skate 64 Raja senta 65 Raja spinicauda 66 Chimaeroids Family Chimaeridae 70 Hydrolagus affinis, Deep Water Chimaera . 70 Family Rhinochimaeridae 71 Rhinochimaera and Harriotta 71 Rhinochimaera atlantica 72 Harriotta raleighana 81 References 84 1 Contribution No. 692 from the Woods Hole Oceanographic Institution. 38 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY INTRODUCTION The slope of the bottom is so steep from the edge of the continent at the 150-200 fathom level down to the 800-900 fathom level off the northeastern United States and off Nova Scotia that the intervening zone is not wider than about 25 miles, anywhere between the offing of Chesapeake Bay and the Laurentian Chan- nel, with a minimum breadth of only 7-10 miles. Nevertheless this zone, narrow though it be, is of great interest ecolog- ically, because it is the site of the transition from the bottom- dwelling animals of the shallow waters of the continental shelf, to those of the Atlantic abyss. The explorations by the "Blake," and especially those by the "Albatross" in the 1880 's, yielded a wide variety of bottom-liv- ing fishes from the lower part of the slope at depths of 800 fath- oms and more; they also sampled the bathypelagic communities of the overlying waters, many additional members of which have been brought to light subsequently. The fishes, too, of the north- eastern American shelf have been studied intensively at many hands for many years, not only from the descriptive-taxonomic standpoint, but also from the standpoint of the commercial fish- eries. But the "Albatross" and "Fish HaAvk" devoted less attention to the middle part of the slope. (They made only 63 trawl hauls at 250-500 fathoms, contrasted with 117 hauls deeper than 700 fathoms). And no subsequent attempt has been made to con- tinue the scientific investigation of the bottom-living fishes in deep water in the northwestern Atlantic since 1887 when the "Albatross" was transferred to the Pacific, whether because of the expense that would have been involved, or because of a shift of interest to other subjects. And very little commercial fishing, either with hook and line or with the trawl, has ever been at- tempted deeper than about 250 fathoms along the slope with which we are concerned, partly because of the difficulty and expense of operating commercial gear in greater depths, and partly because it has not seemed likely that saleable food fishes would be found on bottom there in worthwhile quantities. Consequently, while it seemed safe to forecast, from the fish faunae of shallow water and of deeper, what species of bottom fishes might be expected along the mid zone of the slope, the BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 39 belt in question was a mare ineognitum, to all intents and purposes, as regards the quantities of any species to be found there. In the hope of filling this gap in our knowledge, trawling campaigns were carried out by the Woods Hole Oceanographic Institution along the slope between the offings of southern New England and of La Have Bank, Nova Scotia, on the "Caryn" in June and September 1949; between the offings of Montauk, Long Island, New York, and of La Have Bank from the dragger "Cap'n Bill II" in June and July 1952; and between the offings of Virginia and of La Have Bank on "Cap'n Bill II" in June and July 1953. Otter trawls 35 feet wide were used in 1949, 35 and 50 ft. trawls in 1952, and 60 ft. trawls in 1953. One hundred and forty-three successful hauls were made in the three summers combined, dragging on bottom for 30-60 minutes, between 200 fathoms and 730 fathoms. And a total catch of about 25,000 bottom fishes in all the hauls combined seems good evidence that the nets worked with at least moderate efficiency. The present report deals with the catches of elasmobranchs and of chimaeroids. The following species of sharks and skates had been reported, through 1953, from depths greater than 200 fathoms on the American Atlantic slope north of Chesapeake Bay: Apristurus profundorum Goode and Bean 1895; Centroscyllium fabricii (Reinhardt) 1825; Etmopterus princeps Collett 1904; Centro- scymnus coelolepis Bocage and Capello 1864; Raja bathyphila Holt and Byrne 1908 ; Raja jenseni Bigelow and Schroeder 1950 ; R. mollis Bigelow and Schroeder, 1950 ; R. laevis Mitchill 1817 ; R. senta Garman 1885 ; and R. spinicauda Jensen 1914. On the other hand, one species — Scymnodon melas — described by Big- elow, Schroeder and Springer (1953, p. 233) as new must now be relegated to synonymy (p. 51). The trawling operations of "Cap'n Bill II" add R. fyllae Liit- ken 1887 to this list. And it is likely that R. hyperborea Collett 1878 will be found off Labrador or off Newfoundland sooner or later. But the number of trawl-hauls that have been made along the 300-550 fathom zone between the Hudson Canyon and the offing of La Have Bank, Nova Scotia, is now so large that the list of elasmobranchs to be found on the bottom there is not likely to be increased much in the future. 40 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY SHARKS Family SCYLIORHINIDAE Cat Sharks Apristurus propundorum (Goode and Bean) 1895 This deep-water shark had been known from two specimens only, both of them from the offing of Delaware Bay, until the summer of 1952 when "Cap'n Bill II" trawled it at 7 stations (8 specimens) scattered along the slope from the offing of New York (Lat. 39°46'N, Long. 71°35'W) to the offing of southern Nova Scotia (Lat. 42°39'N, Long. 63°54'W). And with 8 more, taken in 1953 at 6 stations, the records for it are numerous enough now and distributed widely enough to show that it is to be found generally, though sparsely, along the slope from the offing of Nova Scotia to that of Virginia (most southerly station, Lat. 37°39'N, Long. 74°06,W; most northerly, Lat. 42°40'N, Long. 63°52'W) . The depths of capture range between 360-420 fathoms and 640-720 fathoms for the two years combined. The tempera- ture on bottom, at the stations where it was taken, was 3.7° to 4.5 °C in 1953, and presumably was about the same in 1952, though no bottom readings were taken that year at the particular localities in question. Points of taxonomic interest are discussed in our earlier paper (Bigelow, Schroecler and Springer, 1953, p. 214). Family SQUALIDAE Spiny Dogfishes Centroscyllium fabricii (Reinhardt) 1825 Black Dogfish Earlier captures of the black dogfish on the slopes of the Nova Scotian Banks, and of Georges Bank, had not been numerous enough to suggest the existence of more than a sparse population there, though a widespread one as had been known for many years. [n fact, only two specimens had come into our hands, in the Museum of Comparative Zoology, at the time when our earlier discussion of its occurrence in the western Atlantic was written (Bigelow and Schroeder, 1948a, p. 486). But catches of 69 black dogfishes by "Caryn" in 1949, and of 715 and 371 of them by "Cap'n BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 41 Bill II" in 1952 and 1953 respectively, show that our previous estimate of its numbers off Nova Scotia and off Georges Bank must be revised upward. This seems, indeed, to be the most plentiful shark there at depths greater than those frequented by the common spiny dogfish (Squalus acanthias). No doubt the reason for the quantitative discrepancy between the older records for the species in Nova Scotia slope waters, and the catches of it made there in 1949, 1952 and 1953 is that all of the former were based on fish taken with hook and line by the halibut fleet which did not fish much below 200 fathoms, i.e. not deep enough to sample more than the uppermost fringe of the population of this deep-water shark. The number of specimens at hand is now large enough to throw some light on the quantitative occurrence of the species, both geographic and bathymetric. The average catches, per successful haul,1 within the depth range where fabricii was taken at all, were about 24 specimens to the eastward of longitude 66 °W for the years 1949, 1952 and 1953 combined (49 hauls) ; between 4 and 5 specimens from longitude 66° to longitude 69°59'W (26 hauls) ; 2 specimens from longitude 70° to longitude 71°59' (10 hauls) ; and about 1 specimen westward and southward thence to the offing of Virginia (12 hauls). Similarly, the largest catches made in any one haul were 92 and 95 specimens east of longitude 66° ; 29 between longitudes 66° and 69°59'W; 8 between longitudes 70° and 71°59,W; and 3 farther to the west and south. The regularity, too, with which fabricii was taken shows a correspond- ing gradient from northeast to southwest, for while 97 per cent of the hauls made in 1952 and 1953, within its preferred depth zone, yielded it to the eastward of longitude 66°, only 65 per cent did so along the slope off Georges Bank (longitudes 66° to 69°59') ; 60 per cent in the offing of southern New England (longitudes 70° to 71°59') ; and 50 per cent farther westward and southward. Evidently, then, the center of abundance for fabricii, in American waters, lies at least as far to the eastward and north- ward as the offing of western Nova Scotia, perhaps farther still in that direction. Even here, however, the black dogfish — like other fishes — vary widely in numbers from place to place within 1 We take no account of such of the hauls as obviously were unsuccessful for one reason or another. 42 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY O m o C\J o 3 CD c3 o CD ft O CO CD 1-Q CD Sh CD o3 CD CD ^5 CD co CD © > o is . o o in o C\J o to °CVi °CM BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 43 sliort distances, independent of the precise depth, as is illustrated by catches ranging from 3 to 92 per haul within a distance of about 5 miles along the 340 to 520 fathom zone between longitudes 63°58' and 64°06' in 1952; from 15 to 95 per haul within about 10 miles in that same general region in 310 to 625 fathoms be- tween 63°50' an 64°00' in 1953 ; and from 5 to 40 within a few miles to the southwestward (longitudes 64°10' to 64°17' ; 280 to 475 fathoms) that same summer. But a plot of the catches made per haul (Fig. 1) shows that these were not haphazard, for in each year there was a definite center of abundance, concentrating at about longitudes 63° 50' to 64° 10', extending for something like 30 miles along the slope between 310 and 545 fathoms, with the bottom less productive both to the northeastward and to the southwestward. And a similar concentration was encountered some 35-40 miles to the southwestward (longitude 65° to 65°10/) in 1952 at about that same depth (310-490 fathoms). This precise locality was not revisited in 1953. The discovery of these rather definite centers of population is of general interest as evidence that the relative productivity of the bottom, in fishes, may vary as widely from place to place, far down the slope, as every fisherman knows that it does on the great fishing banks on the continental shelf. It may be worth adding that an average catch of 21 specimens per haul, with the trawls used, would work out at about 4 speci- mens per acre, the richest catches (92 and 95 per haul) at about 18 per acre — if the trawls caught all the specimens lying in their path, which they certainly did not; how much to add for failure in this respect would be sheer guesswork. If the catches can be taken at face value, fabricii is only about Yq-Yq as numerous along the slope of Georges Bank and off southern New England as it is off Nova Scotia, and still less numerous farther to the west and south. Nevertheless, captures of five specimens off Delaware Bay in 3 hauls (Lat. 38°41'N-38°47'N), of one off the coast of Maryland (Lat. 38°05'N), and of one off the coast of Virginia (Lat. ' 37°39'N) show that fabricii ranges farther to the southward, along the American Atlantic slope, than had been known previ- ously. Off Nova Scotia, the shoalest capture of fabricii was in a haul at 44 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 185-220 fathoms (3 specimens) ; the largest catches (92 and 95 specimens) were made between 310 and 360 fathoms; and the species seems to have been distributed rather uniformly down to about 600 fathoms, to judge from average catches of about 28 per haul at 301-400 fathoms, of 22 at 401-500 fathoms, and of 33 at 501-600 fathoms. But catches of only 15 in a haul at 610- 625 fathoms, and 6 in one made at 660-705 fathoms1 suggest that fabricii is less plentiful deeper than 550 fathoms than shoaler. But average numbers of specimens taken per haul at different depths may be deceptive, if taken by themselves, for they conceal the fact that the numbers caught, in individual hauls, varied about as widely within representative depth zones as they did geographically (Fig. 1), no doubt for the same reason. Maximum and minimum numbers caught per haul, east of longitude 66°W, in 1949, 1952 and 1953 combined, at depths greater than 300 fathoms, are as follows : Depth in Number of Fabricii caught fathoms hauls per haul 301-350 12 3-95 351-400 6 4-71 401 - 450 7 0-59 451 - 500 6 1-40 501 - 550 4 3-62 551-600 1 5 601 - 650 1 15 651 - 700 1 6 Farther to the west and south fabricii is not only less numer- ous, but the upper boundary to its vertical range lies deeper down the slope, progressively. Thus the shoalest hauls in which it was taken were at 355-400 fathoms (1 specimen) and 360-420 fathoms (2 specimens) along the slope of Georges Bank, at 390-460 fath- oms (3 specimens in 3 hauls) in the offing of southern Massa- chusetts, and 485-520 fathoms, southward from the offing of New York. i These were the only hauls to the eastward of longitude C6C deep as 600 fathoms, either in 1952 or in 1953. that worked as BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 45 The deepest catches were made at 660-705 fathoms off Nova Scotia (6 specimens in one haul) ; at 600-670 fathoms off Georges Bank (5 specimens in 1 haul) ; at 710-730 fathoms off southern Massachusetts (5 specimens in 1 haul) ; at 630-675 fathoms (3 specimens in 1 haul) and at 640-720 fathoms (2 specimens in 1 haul) farther to the westward and southward. As these were the deepest successful hauls, with one exception, that were made in the respective sectors of the slope, it is a question for the future how much deeper the range of fabricii may extend. In 1953, when the temperature was taken on bottom with maximum-minimum thermometers at almost all the stations where trawl hauls were made, all the captures of fabricii occurred where bottom temperatures ranged between about 4.5°C, and about 3.3°C, the great majority between 4.5° and 3.9°. And even if the range of fabricii reaches down as deep, say, as 1200 fathoms — of which there is no evidence — its lower thermal boundary along the American slope would not be colder than about 2.5°C. At the other extreme, specimens that work their way up the Nova Scotian slope to 200 fathoms or so, such as were brought in by the halibut fleet years ago, are likely to meet water at least as warm as 6-7 °C. Thus, the range of temperature within which it occurs in greatest numbers in American waters is only about 2°C, the extreme range there perhaps 4-5 °C for the entire population. Consequent on the steepness of the slope, the maximum breadth of the zone where fabricii occurs the most regularly is only some- thing like 25-30 miles off the Grand Banks and off Nova Scotia, even if it ranges down to 1000 fathoms there ; 10-12 miles along Georges Bank and off southern New England; and perhaps 15 miles in the offing of Maryland and of Virginia. The catches of fabricii made east of longitude 66 °W averaged considerably larger in 1952 than in 1953 as tabulated below. But with the catches in individual hauls ranging from 1 to up- wards of 90 in each of the two years, the discrepancy may not be wider than can be credited to what we may call "fisherman's luck," for want of a better name. 46 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Catches east of longitude 66 °W, at depths greater than 300 fathoms Percentage of hauls Average Successful Hauls with with Total per Maximum hauls fabricii fabricii catches haul catch 1952 19 19 100% 610 32 92 1953 15 14 93% 333 22 95 Among 25 specimens that were opened in 1953, 10 were empty; the others contained remains of fishes, decapod crustaceans, (shrimps), schizopod or euphausiid, squids, and octopus.1 Etmopterus princeps Collett 1904 We have already reported (with description) the capture of 47 specimens of this deep-water shark along the slope between the offings of southern New England and western Nova Scotia (Bigelow, Schroeder and Springer, 1953, p. 47). Earlier records for princeps had been from the Faroes-Hebrides region (Collett, 1904, p. 3; 1905, p. 28) and from the offing of the Straits of Gibraltar (Koefoed, 1927, p. 21). The trawlings of 1953 yielded 7 more, all of them from the offing of western Nova Scotia. The largest numbers taken were 9 (once), 8 (once) and 7 (once) while 1 was the number taken the most often (fourteen times). The hauls that took princeps, in the three seasons, were at depths ranging from 310-320 fathoms down to 490-540 fathoms, with the largest catches (7, 8 and 9) at 480-520 fathoms, 420- 480 fathoms, and 390-440 fathoms respectively, which suggests that princeps is about equally plentiful, downward, across the depth zone along which it occurs off our coasts. And it is to be expected considerably deeper there as well, for it has been taken at 1134 fathoms off the Straits of Gibraltar. The 18 hauls that yielded it in 1949 and 1952 were scattered all along from the offing of La Have Bank (longitude 63°38/W) to the offing of Cape Cod (longitude 70°05'W), with the largest catches (3, 7, 8 and 9 individuals) off Nova Scotia. Similarly, 5 of the 7 specimens taken in 1953 were from the general offing of La Have Bank, only 2 off the eastern part of Georges Bank, although many successful hauls were made there and to the i Stomach contents identified by Dr. Benjamin Laevitt. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 47 westward along the 300-550 fathom zone that summer. It is a question for the future whether this difference is evidence of a corresponding fluctuation in the western boundary of princeps, in numbers large enough for the trawl to pick it up, or whether we are dealing with a matter of pure chance. In either case, there is nothing to suggest that this shark ranges westward beyond the offing of Cape Cod, a limitation we cannot explain as due to temperature, for readings taken by "Cap'n Bill II" in 1953 on bottom along the 300-550 fathom zone were very nearly the same to the westward of longitude 70° W (3.5°-4.8°) as to the eastward (3.3°-4.5°). Centroscymnus coelolepis Bocage and Capello 1864 Portuguese shark The trawlings of 1953 yielded three specimens of this species : two juvenile males, 565 mm. and 690 mm. long, and an adult female of 1035 mm. which gave premature birth on deck to 15 embryos 266-282 mm. long, 5 of them males and 10 females. The number of embryos is in line with Vaillant's (1888, p. 66) record of 13 to 15 for this species. Moreau (1891, p. 9), it is true, has reported only 5 for a Mediterranean female, identified as coelolepis, but her small size (599 mm.) suggests that she be- longed to some other species in reality. It has been a tacit assumption, until recently, among students of sharks, that the denticles which develop on any given part of the body at different stages of growth are similar in shape ; hence, that the shape of the denticles is a more or less reliable specific character — or even a generic one. But Radcliffe (1916, p. 267) has pointed out that the denticles to be seen on specimens of different ages differ somewhat in shape in some species of the genus Carcharhinus. Tortonese, too, (1952, p. 386, fig. 1; p. 387) has recently found that the denticles on the sides of a juvenile male, 270 mm. long, seemingly referable to Centroscymnus coelolepis by other characters, were strongly tridentate, whereas those on the adult coelolepis have evenly rounded edges, as has been known for many years. And the "Cap'n Bill II" series of that species enables us to corroborate his very interesting finding. Thus those on the body of our adult female 1035 mm. long are 48 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY evenly ovoid in outline, with concave crowns, and regularly over- lapping (Fig. 2D) ; they have been pictured and described simi- larly for this species by Bocage and Capello (1866, pi. 3, fig. 3), by Vaillant (1888, pp. 64-65), by Garman (1913, pi. 14, fig. 8), and by us (1948a, p. 495, fig. 94A). But the denticles on the embryos to which she gave birth1 are tridentate, weakly so on the top of the head, where they are moderately widely spaced, but much more strongly tridentate on the sides of the body where they are more widely spaced as well (Fig. 2A). Fig. 2. Centroscymnus coelolepis. Dermal denticles from side of trunk below first dorsal fin. A, Embryo male, 282 mm. long, from adult female of 1035 mm., about 13 x. B, Juvenile male, 565 mm. long, about 8 x. C, Juvenile male, 690 mm. long, about 6 x. B, Adult female 1035 mm. long (mother of embryo shown in A), about 5 x. i Embryos well formed, but with only the tips of the teeth exposed, while the denticles had not yet erupted through the skin. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROms 49 It is evident, too, from partly grown specimens of different sizes that the margins of the denticles that are developed succes- sively, during growth, are less and less dentate. The smaller of the denticles, for example, on the flanks and belly of a juvenile male 565 mm. long are strongly tridentate still, but the larger (i.e. younger) ones only Aveakly so, and with one overlapping the next more or less widely in most cases (Pig. 2.6). At this stage the denticles on the sides below the first dorsal fin range from about 1.2 mm. to about 2 mm. in length. At a slightly more advanced stage, as represented by a male of 690 mm., the smallest (i.e. the oldest) of the denticles on the trunk are tridentate (Fig. 20), but many of the larger (i.e. younger) ones now have entire margins, though with pointed tips still, while denticles of intermediate sizes show various inter- mediate stages and the denticles on the body now overlap so regularly that they hide the skin. Up to this stage the denticles vary considerably in size from one to the next, with those on the sides below the first dorsal fin ranging from about 1 mm. to about 2.4 mm. in length. But those on the adult female of 1035 mm. (Fig. 2/)) are all about equal in size over any given region of the body. And they no longer include any tridentate forms except near the tips of the pectoral fins. This seems an appropriate place to remind the reader that on adult coelolepis (females, at least; adult males have not been seen) the denticles are largest (about 3-3.5 mm. long) on the mid sector of the trunk, belly as well as sides and back, and are progressively smaller not only forward, but rearward as well, with those on the caudal peduncle on the average only about 70 per cent as long as those on the sides below the first dorsal fin. The discovery that the denticles are of one shape on young specimens of this particular species of shark, but of a strikingly different shape on adults, opens vistas that are interesting from the standpoint of elasmobranch morphology in general, as well as from that of taxonomy. To begin with, it poses the question, how long is the normal life span of a dermal denticle in relation to the life of the individual ? So far as we can learn, no special attention has been paid to this point, nor is it safe to reach any conclusion a priori 50 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY here, for while the teeth of sharks, which correspond in all essentials with the placoid scales or dermal denticles, are replaced periodically, the dorsal fin spines — equally homologous with the denticles — are retained throughout life in such sharks as have them. Among bony fishes, the general rule is that a scale, once formed, persists throughout the life of the fish, growing in size meantime, at a rate that keeps pace (more or less) with the rate of growth of the fish. In fact, the determination of the ages of fishes by the markings on their scales, a familiar procedure in fisheries biology nowadays, is based on belief in the validity of this rule. But Hertwig's (1874, p. 358) observation that new dermal denticles are formed between pre-existing ones in some sharks,1 not only during embryonic development but during later life as well, shows that the situation is not so simple for elasmo- branchs. Counts of about twice as many denticles within a given per- centage of the total length of a specimen, below the first dorsal fin, of C. coelolepis 565 mm., 690 mm., and 1035 mm. long as on an embryo of 277 mm. (for details, see p. 49) show that the final number is developed, in this particular species, by the time a given specimen is about half grown — how much earlier is not known. It also seems evident, from the shapes and relative sizes of the denticles, and from their irregular arrangement, that several generations of them are represented on small examples as illustrated by our specimens of 565 and 690 mm. (Pig. 2, B, C). But the uniformity in size and shape of the denticles on adults, and the regularity with which they are arranged, suggest that the ovoid ones that first develop on specimens a little more than half grown represent the final generation, which persists through- out the later life of the individual concerned. We can go no farther than this from information available up until this time. We think it probable, too, that a similar succession in the shapes of the denticles takes place between young specimens and adults, in the case of Centroscymnus owstoni Garman 1906, from Japan, a near relative of the North Atlantic C. coelolepis. When adult, its denticles resemble those of adult coelolepis very closely in shape. But they are tridentate over the trunk as a whole on two i It is not clear, from the context, whether Ilertwig made these particular observations on Mustelus, on Acanthias, on Dalatias or on Heptranchias. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 51 Japanese squalids, 260 mm. and 287 mm. long-, in the Museum of Comparative Zoology, which agree very closely in all other respects (including very minute denticles, very oblique lower teeth, and no median tooth in the lower jaw) with two adult males of owstoni, 785 mm. and 805 mm. long, in the Museum. Incidentally, the two small specimens just mentioned came from Japan labelled Scymnodon squamulosus. But they differ sharply from squamulosus as pictured by its describer (Giinther, 1878, pi. 2, fig. B) both in their very oblique lower teeth, and in pectorals reaching back as far as the level of the origin of the first dorsal fin, for the lower teeth of squamulosus are shown as nearly erect, and the tips of its pectorals as falling short of the level of the origin of the first dorsal by a distance about as great as the length of the snout in front of the mouth. On the other hand, the successional history of the denticles of Centroscymnus coelolepis strengthens the specific validity of G. cryptacanthus Regan 1906, from Madeira, the denticles on the type specimen of which (about 700 mm. long) were tridentate with 3 parallel keels on the trunk anterior to the first dorsal, but "the others smooth, with rounded edges and with a rounded depression on the free surface of each" (Regan, 1906, p. 437). Present realization that it is not safe to define species of sharks by the shape of their denticles — unless the ages of the specimens in question be taken into account — forces us, next, to conclude that the juvenile squalids, 330-462 mm. long on which Bigelow, Schroeder and Springer (1953, p. 233) based the new species Scymnodon melas actually represent nothing more than an early stage in the growth of Centroscymnus coelolepis. Thus their denticles correspond very closely to the smallest and most tri- dentate of those of our 565 mm. coelolepis (p. 49, Fig. 2B), though they contrast so strongly with the rounded denticles of the adult that we would never have guessed that they belonged to the same genus, much less to the same species, had it not been proven that the denticles of the young of coelolepis are similarly tridentate, to be replaced by ovoid ones later. The somewhat wider spacing of the denticles on the supposed melas than on partly grown coelolepis points merely to an earlier stage in the growth of the individuals. Neither can we find any better excuse for retaining the species, whether in proportional dimensions, in the fins, or 52 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY in the shapes and number of teeth. And while we had thought its black color distinctive, as contrasted with the chocolate hue of the adult coelolepis, embryos of the latter, of proven parentage, are a very deep blue, while half-grown males are black, with only a faint tinge of chocolate to forecast the color of the adult. The previous records for this species for the western Atlantic have been scattered along the slope, from the Grand Banks of Newfoundland to the offing of Nantucket. It is therefore inter- esting — as it was unexpected — that the three caught by ' ' Cap 'n Bill II" were all taken between the offings of New York and Delaware Bay, as follows: Length in mm. 1035 Sex 5 Lat. 38°41'N Long. 73°01'W Depth in fathoms 570-610 690 $ 39°09'N 72°21'W 485-520 555 S 38°43'N 72°56'W 630-675 Our failure to take coelolepis off Georges Bank, or off Nova Scotia, where it was reported so often by the halibut fishermen 3rears ago (Bigelow and Schroeder, 1948a, p. 498), and where about 90 successful hauls were made in 1949, 1952 and 1953 at depths greater than 200 fathoms, including about 40 hauls deeper than 400 fathoms, is something we cannot explain. Earlier captures of coelolepis in American waters were from depths of 150-250 fathoms — all of them made with hook and line. But we have already noted the likelihood that it would be found much farther down the slope in the western side of the ocean, as it is in the eastern, if fished for there with suitable gear. And the "Cap'n Bill II ' trawlings bear out this expectation, by extending its known range in American waters down to 630 fathoms at least. BATOIDS Family RAJIDAE Skates Raja bathyphila Holt and Byrne 1908 The capture by "Cap'n Bill II" of a female R. bathyphila, 360 mm. long, on the seaward slope of Georges Bank (Lat. 40°04'N, Long. 68°34'W), at 370-450 fathoms in 1952, and of a BIGELOW AND SCHROEDER : ELASMOBRANCHS AND OHIMAEROIDS 53 juvenile male of 370 mm. at a neighboring- station (Lat. 40°10'N, Long. 68°16'W) at 490 fathoms, added to earlier locality records for it at one station off Chesapeake Bay, at 3 stations off southern New England, and at one station off Brawns Bank, in depths of 885 to 1188 fathoms,1 show that this deep water skate is generally distributed along the mid zone of the continental slope, between the offings of Chesapeake Bay and of southern Nova Scotia. Knowledge as to its status farther to the northeastward along the American slope is to be desired to tell us how wide the gap may be between the geographic ranges occupied by it in the western side of the Atlantic, and in the eastern, where it is known only from the Irish slope (type specimen), so far as we are aware. The depths of capture recorded for it so far (370-1188 fathoms on the American slope, 673 fathoms on the Irish) suggest that its center of population lies deeper than 400-500 fathoms, with the upper limit little shoaler, if at all, than 350 fathoms. The most distinctive character of bathyphila is that its entire lower surface is darker than the upper surface. Indeed, it is unique in this respect, among the skates that are known from the North Atlantic outside of the Gulf of Mexico. Within the Gulf it is paralleled in this respect by the newly described R. fuliginea Bigelow and Schroeder 1954. But the latter differs from bathyphila so sharply in a more convex anterior outline as well as in the prickliness of the lower surface of its tail and of the upper sides of its pelvic fins that there is little likelihood that the one species could be taken for the other. For a comparison of bathyphila with the several skates of the Pacific and Indian Oceans that are similarly dark colored below, see Bigelow and Schroeder 1953, p. 161. A character, almost equally diagnostic for bathyphila among skates of its geographic province, but one that has not been stressed previously, is the narrowness of its mouth. In the four specimens that we have measured2 the breadth of the mouth ranges from 4.9 to 5.9 per cent as great as the total length. The only other skates known from the northwestern Atlantic, outside of the Gulf of Mexico, that approach it closely in this respect are Raja garmani Whitley 1939 (about 5.7-5.9 per cent) ; R. fyllae i For details, see Bigelow and Schroeder, 1953, p. 159. - The two "Ciip'n Bill II" specimens, and two others, the proportional dimen- sions of which are given in Bigelow and Schroeder 1953, p. 161. 54 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY Liitken 1887 (5.8 per cent in eastern Atlantic specimen, 6.1-7.0 per cent in four of the ' ' Cap 'n Bill II ' ' specimens, 397-497 mm. long) ; and Breviraja plutonia (Garman) 1881 (5.1, 5.6 per cent, in two specimens taken off Jacksonville, Florida). And there is no likelihood of confusing baihijphila with any of these, quite apart from its coloration. Thus the anterior contour of the disc of bathy- phila, half-grown and larger, differs widely from that of fyllae, and the arrangement of thorns and prickles is noticeably different in the two species, as we have pointed out elsewhere (1953, p. 159). The plain coloration of the upper surface of bathyphila contrasts as strikingly with the pattern of dark rosettes on garmani as does the dark lower surface of bathyphila with the pale lower surface of garmani. And no one, we fancy, would be likely to mistake a bathyphila, wedge shaped in front, with tail only moderately long, and a hard snout, for a plutonia, with its much more convex anterior outline, much longer tail and soft snout. The male bathyphila taken in 1953 has six large thorns along the mid line of the back from the scapular region to the level of the axils of the pectorals, with the single median row reaching back as far as the axils of the pelvics ; the female has seven large medians between scapular region and level of axils of pelvics, rearward from which there are 2-3 less regular rows. Thus these specimens, 370 and 360 mm. long, respectively, represent a stage in development between the 463 mm. female pictured by us in Part 2 of Fishes of the Western North Atlantic and the somewhat smaller (323 mm.) male (1953, Figs. 30, 31A). They agree so closely with our earlier account of this species in all other respects that no further discussion of them seems needful from the taxonomic standpoint. Raja fyllae Liitken 1887 This skate was included in our general survey of the raj ids of the western North Atlantic (Bigelow and Schroeder, 1953, p. 194) because of its presence in west Greenland waters, where it has been known for many years, as it has in boreal and subarctic latitudes in the eastern side of the Atlantic. The cruises of "Cap'n Bill II" have now extended the known range of fyllae to the Nova Scotian slope, and even to the southwestern slope of BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CfllMAEROIDS 55 Georges Bank, in depths of 290-600 fathoms. It must, indeed, be very generally distributed (though in small numbers) along this depth zone between the offing of Halifax and the seaward slope of Browns Bank (longitudes 63°47'W to 65°06'W) for it was taken in 8 of the 19 successful hauls that were made there in 1952 (16 specimens) as well as in 8 of the 16 successful hauls Fig. 3. Raja fi/llae. Left, male, 505 mm. long, off La Have Bank, M.C.Z. No. 37905; and right, female, 430 mm. long, same general locality, M.C.Z. No. 37898. made there in 1953 (13 specimens). And we may expect it to prove equally widespread around the Newfoundland Banks when the 300-600 fathom depth zone has been explored there more fully. But it seems to be much less numerous farther to the westward — unless at a considerably greater depth — for only 56 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY three hauls (3 specimens) yielded it on the Georges Bank slope (longitudes 67°02'W to 68°54'W), out of the 24 successful hauls that were made in 1952 and 1953, combined, at depths greater than 300 fathoms, between longitudes 66°W and 70°W. And it was not taken at all in the many hauls made farther to the west- ward and southward. The Nova Scotian-Georges slope specimens (Fig. 3) agree so closely with the specimen from southwest of Ireland, figured by us (Bigelow and Schroeder, 1953, Fig. 39), both in shape of disc, in proportional dimensions including length of tail, and in dermal armature, that no cleavage is apparent between the populations inhabiting North American and north European waters. But the additional material allows us to expand our earlier descrip- tion of the species in the following particulars. A — Length of tail. In 7 of the larger specimens, 407-505 mm. long, including 2 mature males, the tail, measured from the center of the cloaca, occupies 58-61 per cent of the total length, and the ratio between length of tail and length of body is between 1.35 and 1.56. In 5 juveniles of both sexes, 198-272 mm., this ratio is 1.69-1.80. In B. erinacea and in B. ocellata (only skates with which fyllae might be confused in American waters) the ratio, length of tail to length of body, is about 1.00-1.35. B — Shape of disc. The anterior contour of an adult male 505 mm. long, in the "Cap'n Bill II" series (Fig. 3) parallels closely that of a somewhat larger male (555 mm. long) from West Greenland that was pictured first by Liitken (1898, PI. 2), and subsequently by Clark (1926, PI. 22, fig. a) in being deeply concave on each side, abreast of the spiracles, a point worth mentioning since these are the largest males that have yet been seen, so far as we are aware. Females from off Nova Scotia, 397 and 430 mm. long, agree very closely in the shape of their discs with a female of 452 mm. taken southwest of Ireland that we have pictured elsewhere (Bigelow and Schroeder, 1953, p. 195, Fig. 39). Students of skates have learned, long since, to expect con- siderable variation in dermal armature from specimen to speci- men in nearly every species, both for the larger thorns and for the smaller prickles as well, though each species shows a basic arrangement that is characteristic of it. Our predecessors have BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 57 observed, already, that fyllae is no exception to this ride ; and the Nova Scotian specimens afford an additional illustration. Thus the areas bare of prickles that tend to develop on the upper surface of its disc, as this skate grows, vary considerably in their extent among- the females, independently of the sizes of the latter. In one, 430 mm. long, there is a bare area behind each spiracle, reaching rearward to the shoulder region, and there are no prickles on the upper surface of the pelvics. But the regions rearward from the spiracles are prickly on another female of 397 mm., and also the upper surface of the posterior lobe of each pelvic fin, much as they are on a female of about this same size taken off Ireland (Bigelow and Schroeder, 1953, Fig. 39), and in all of the smaller specimens of both sexes taken off Nova Scotia. The males, as they mature, tend to lose any prickles they may have had earlier on the pelvics, and also to lose part of the larger thorns from the mid-dorsal belt of the disc between the levels of the pectoral girdle and of the axils of the pectorals. It was known, previously, that the upper surface of the tail, which is set with prickles (besides the thorns) on small specimens, loses most of these prickles with growth. It now seems that this alteration is a more regular one than previous observations had suggested, for while the tail is uniformly and densely prickly above on the Nova Scotian juveniles, to 270 mm. long, it is wholly bare of prickles along a definite median band throughout its length back to the first dorsal fin on the larger Nova Scotian specimens of 397-505 mm. The shape, too, of the tail alters in fyllae with growth, from arched above in the young to flat above in adults, along the median band that has become bare of prickles. A feature which seems to have escaped scientific attention is that the mucous pores on the nuchal region of fyllae are arranged in a pattern that is conspicuous on small specimens as well as on large. The only other skate of the western North Atlantic in which the pores in this region are noticeable is R. laevis, but they are black pigmented in the latter, whereas they are pale in fyllae. Thirty to thirty-eight series of teeth have been recorded for fyllae. In the larger of the Nova Scotian specimens, 397-505 mm. long, the number of series varies more widely in the upper jaw (30-38) than in the lower jaw (30-32). It is not known whether 58 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY this same disparity obtains for the European population of this species. The northeastern Atlantic and west Greenland representatives of fyllae so far seen have been ashy gray to chocolate brown above, the adults uniform, the young marked more or less dis- tinctly with darker spots. One of the larger of the Nova Scotian females 430 mm. long, and 2 others half to three-quarters grown, are similarly of a uniform ashy gray tint above. But 5 other females are conspicuously marked, above, with an oblong whitish blotch, longer than wide, between the eyes, and with a similar but vaguely outlined pale blotch on the inner posterior part of each pectoral fin, a pattern of which we find no previous report for fyllae. One large male and 1 half grown lack these blotches but 1 large and 2 half to three-quarters grown males are so marked for R. fyllae. The lower surface is uniformly grayish white on the largest Nova Scotian male (505 mm.) both on tail and on disc. But the lower surface of other large specimens is more or less smoky around the outer posterior edges of the pectorals, in the region of the cloaca, and on the anterior parts of the pelvics; their tails are variously dark-blotched or mottled below ; the region of the gill openings is smoky on one of them ; and there is a smoky prepelvic blotch on three of them. Thus the dark markings on the lower surface vary as widely from specimen to specimen in extent and in arrangement among the Nova Scotian population as among the Greenland and eastern Atlantic populations. The fact that the claspers are small, still, on one "Cap'n Bill II" male of 442 mm., but seem ready to function on another of 505 mm., on which the alar thorns have also developed in 2-3 rows, suggests that males of the Nova Scotian population mature, sexually, at a length of perhaps 475-500 mm. The only other western North Atlantic skates that fyllae re- sembles closely in shape of disc, and in the general arrangement of its dermal armature, are R. erinacea Mitchill 1825, and R. ocel- lata Mitchill 1815. We have pointed out already (1953, p. 196) that it differs from both of these in a longer tail. But the measure- ments of the fyllae that we now have at hand show that the distance from the axils of the pelvics to the first dorsal fin as employed in our key (1953, p. 150) is not as reliable a criterion BIGELOW AND SCHROEDER : ELASMOBRANCHS AND OHIMAEROIDS 59 as it seemed. Thus, while this distance is greater than from axils of pelvics to fronts of orbits in most of our fyllae, it is only about as great as to the mid levels of the orbits in a few, and is some- what shorter than to the rear edge of the orbits in one adult male,1 as it is in erinacca and ocellata also. And while the total length of the tail (measured from center of cloaca) is the greater relative to the length of the body in fyllae, there is virtually no discontinuity in this respect, as appears from the following table. Ratio, total length of tail (from center of cloaca) to length of body number of total lengths max. min. average specimens mm. fyllae 1.80 1.35 1.57 12 198—505 erinacea 1.35 1.06 1.22 9 209—505 ocellata 1.10 0.97 1.05 4 414—809 The distance, however, from the origin of the first dorsal fin to the center of cloaca is longer than from the cloaca to the tip of the snout in all the specimens of fyllae that we have measured, but shorter than from cloaca to snout both in erinacea and in ocellata, as follows: Ratio, distance from first dorsal to cloaca relative to distance, cloaca to tip of snout number of total lengths max. min. average specimens mm. fyllae 1.36 1.03 1.19 10 202—505 erinacea 0.97 0.81 0.88 5 209—505 ocellata 0.79 0.71 0.74 5 423—641 The interorbital space is narrower on fyllae, averaging 6.73 per cent (6.12-7.30) of the width of disc on 13 specimens ex- amined, than it is on erinacea with an average of 8.01 per cent (7.32-8.50) on 11 specimens, or ocellata with an average of 9.12 per cent (8.82-9.32) on 4 specimens. For specimens that cannot be identified, positively, by length of tail alone, the teeth should usually prove diagnostic, for there are only 30 to 38 series in the upper jaw in fyllae contrasted with i Abnormal ? 60 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 38 to 64 in erinacea, and 72 to 110 in ocellata. And the arrange- ment of the thorns on the tail usually is distinctive for large specimens, and in most cases for half grown, for ocellata and erinacea of these sizes usually have a narrow naked band along the mid line which has not been the case in any fyllae we have seen. But we must admit that specimens might come to hand for which depth of capture would be the only reliable criterion for identification that we have been able to discover. In any case there is little danger of confusing fyllae either with erinacea or with ocellata in the field, for it has never been taken in North American waters shoaler than 290 fathoms, and neither of the others deeper than 85-87 fathoms (two erinacea trawled by "Cap'n Bill II" off the southern edge of Georges Bank in July 1953). Raja jenseni Bigelow and Schroeder 1950 The captures of a female jenseni, 625 mm. long and in good condition, by "Cap'n Bill II" on the slope of Browns Bank, Lat. 42°19'N, Long. 64°59'W, in a trawl haul at 390-440 fathoms, and of a second female, of 695 mm., off Delaware Bay (Lat. 38°47'N, Long. 72°54'W) at 585-595 fathoms, call for mention, for the only specimens of this deep water skate that had been reported previously were a male, of 223 mm. from the slope of Georges Bank, 1255 fathoms; a female (the type) of 541 mm. off southern New England, 1043 fathoms; and a female of 850 mm. (now fragmentary) from off Halifax, Nova Scotia, brought in by a fisherman from 200 fathoms;1 all are now in the U. S. Na- tional Museum, and all were taken many years ago. The "Cap'n Bill II" specimens agree so closely with the type specimen (Bigelow and Schroeder, 1950, PI. 1; 1953, Fig. 45) that their specific identity is evident at a glance. There are 29 large thorns in the mid dorsal row on the 625 mm. female, 8 of them anterior to the level of the axils of the pectorals. On the 695 mm. specimen, the total number in the median row is 26, 6 of which are anterior to the level of the axils of the pectorals, so spaced as to suggest that 3 or 4 others had been lost. These counts contrast with a total of 24 or 25 on the type (7 anterior to i For the nomenclatural history of jenseni, description, and illustrations, see Bigelow and Schroeder, 1950, p. 385, PI. 1 ; 1953, p. 213. Figs. 45, 40. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND OHIMAEROIDS 61 the axils of the pectorals) and with a total of 31 on the smaller male (8 anterior to the axils of the pectorals). Evidently there is a small variation in number from specimen to specimen, inde- pendent of the sizes of the latter. Enough specimens have been seen to show that it is characteristic of jenseni for the mid-dorsal thorns to be widely spaced anterior to the pelvic girdle, but increasingly closely spaced thence rearward onto the tail. On the 625 mm. female the thorns on the disc are about 20 mm. apart; those abreast of the axils of the pelvics about 10 mm. apart. The next 8 thorns along the tail are in contact at their bases, but the thorns thence rearward are separated, one from the next, by short interspaces. Also, the thorns are successively smaller, rearward along the posterior third of the tail. As on the type, there is no thorn in the interspace between the two dorsal fins on either of the "Cap'n Bill II" specimens, but only a few prickles. The presence, on each side, of 3 or 4 thorns on the scapular region on the small male of 223 mm., and of 3 on the type, 541 mm. long, but of 2, only, on both of the "Cap'n Bill II" specimens of 625 mm. and 695 mm. indicates that one or two of the thorns in this group are lost with growth. There may be either two post- ocular thorns or three on different specimens,1 but every specimen that has been seen has had one preocular thorn, only, on each side. The tip of the snout and the rostral ridge are about as rough on the "Cap'n Bill II" specimens as on the type, and the small prickles are about as densely distributed and cover about the same areas. A point worth emphasis is that while the sides of the tail of jenseni are densely prickly there are no prickles on the lower surface of its tail below the level of the lateral folds. The lower surface of the disc is naked everywhere. Fifty-eight and sixty-six series of upper teeth have been reported previously, for the two specimens for which they have been counted. The "Cap'n Bill II" female 625 mm. long has 60, but the 695 mm. specimen has only 51. Evidently the number is widely variable on this species, irrespective of size, much as in R. erinacea, and in R. ocellata. The teeth resemble those of the type specimen closely (Bigelow and Schroeder, 1953, Fig. 46D) both in shape and in arrangement. i Three on the 850 mm. female ; two on each of the other specimens that have been examined. 62 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY We have forecast already that the dark markings on the lower surface of the disc, in jenseni, would prove widely variable (Bige- low and Shroeder, 1953, p. 216), and the "Cap'n Bill II" speci- mens corroborate this expectation. The ground-tint in both cases is yellowish white. On the 625 mm. specimen this is clouded with ashy gray around the posterior marginal zone of the pectorals and on the outer posterior parts of the pelvics; in the region of the cloaca and forward on either side of the abdomen; also over an irregular area on each side close behind the mouth. On the 695 mm. female the lower surfaces of the pelvics, the abdominal region in general, the lower surfaces of the pectorals, an irregular area on either side inward and forward from the gill region, and another smaller, inward from the nostril, are dark sooty gray. This pattern simulates, rather closely, the dark markings on some specimens of R. hyperborea (Bigelow and Schroeder, 1953, p. 209, Fig. 44), from which jenseni differs quite sharply by having a more simple dermal armature and fewer teeth (Bigelow and Schroeder, 1953, p. 213). The lower surface of the tail has been plain dark ashy gray on all the specimens of jenseni that have been seen thus far. The localities whence jenseni has been reported (see above) are scattered sufficiently to show that it is generally distributed along the upper part of the slope between the offings of middle Nova Scotia and of Delaware Bay. But we have yet to learn how much farther its geographic range may extend in either direction. And the paucity of captures, with the great depths at which the "Albatross" specimens were taken, suggests that while a sparse population exists as shoal as 200-600 fathoms, the center of abundance for this species lies deeper down the slope than has been sampled yet by adequate methods of fishing. Raja laevis Mitchill 1817 Barn Door Skate This skate has been recorded as deep as 235 fathoms, but it occurs most regularly, and in greatest numbers on the continental shelf in depths shoaler than 70-80 fathoms. Hence captures of it by "Cap'n Bill II" at 265-305 fathoms off Nova Scotia (2 specimens in 2 hauls, Lat. 42°23'— 42°28'N; Long. 64°31'— BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 63 64°52'W), and at 300-410 fathoms off Nantucket (1 specimen in 1 haul, Lat. 39°54'N, Long. G9°56'W) are of interest, as extend- ing the known depth-range of the species downward, somewhat. Other records of it, from the traAvlings of 1952 and 1953 were from 220-255 fathoms on the seaward slope of Georges Bank (1 specimen, Lat, 40°11'N, Long. 68°20'W), and from 175-225 fathoms off southern New England (1 specimen, Lat. 40°04'N, Long. 72°12'W, 190-225 fathoms, and 2 specimens in 1 haul, Lat. 39°56'N, Long. 71°22'W, 175-200 fathoms). Raja mollis Bigelow and Schroecler 1950 The type — and only known — specimen of this species, trawled on the slope off southern Nova Scotia, Lat. 41°53'N, Long. 65°35'W, at a depth of 858 fathoms by the "Albatross" in 1883, had rested, unnoticed, in the IT. S. National Museum until 1950. We can now add three more records for mollis from that same general region, namely, a juvenile male, in the neighborhood of 195 mm. long (tip of tail lost) taken by the "Caryn" in June 1949, at Lat. 41°25'N, Long. 65°54'W, in a trawl fishing between 415 and 490 fathoms ; a second male, of 267 mm. taken by " Cap'n Bill II," July 26, 1952, at Lat, 42°40'N, Long. 63°51'W, from 465-480 fathoms; and a third, 298 mm. long, July 12, 1953, from Lat, 42°40'N, Long. 63°52'W, at 415-420 fathoms. The fact that "Caryn" and "Cap'n Bill II" took only three specimens of mollis in the 39 hauls they made at depths deeper than 300 fathoms in the Nova Scotian sector during the three summers combined, coupled with the depth (850 fathoms) at which the type specimen was trawled, makes it likely that the center of population for this species is along the lower part of the continental slope. But it is anyone's guess how far its geo- graphic range may extend, either to the northeastward, or to the southwestward. Mollis falls in the moderately short-tailed subdivision of the genus, the distance from the level of the axils of its pelvics to the first dorsal fin being much shorter than from the axils of the pelvics to the tip of the snout, and the total length of its tail from the cloaca being only about 1.1 times as long as the body from center of cloaca to tip of snout. Its most noticeable external characteristics are that it has no large thorns anywhere, posterior 64 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY to the scapular region, but that the upper surface of its disc and tail are densely prickly, except close along the outer posterior edges of the pectorals, and that the entire breadth of the lower surface of its tail is also densely prickly, except for the extreme tip.1 The only slight differences worthy of mention between the larger of the two "Cap'n Bill II" specimens and the type is that the former has two minute thorns close in front of each eye (only one in front of each eye on the type) ; and that while there is a very small postocular thorn (as on the type) on one side of its head, none is to be seen on the other side among the close-set prickles that roughen the skin there. There are 54 series of teeth in the upper jaw of the "Caryn" specimen about 195 mm. long, 60 in both the ''Cap'n Bill II" specimens of 267 and 298 mm., and 60 on the type specimen. On the type specimen, the rostral projection from the cranium, reaching nearly to the tip of the snout (easily felt) is soft throughout its length. It is soft on the ' ' Cap 'n Bill II ' ' specimens also and this state is so unusual, among the members of the gen- era Raja and Breviraja, that we considered it as perhaps the outstanding feature of the species mollis, in our earlier discussion (1953, p. 237). But the rostral cartilage of the very small specimen (195 mm. long) collected by "Caryn" in 1949 is rather stiff to the touch, raising the question as to whether its degree of firmness is a matter of the stage in growth, or whether it perhaps is affected by preservation. Eaja radiata Donovan 1807 Thorny Skate Previous knowledge of the range of the thorny skate, and the numbers in which it has been taken on the Newfoundland Banks, in the inner parts of the Gulf of Maine, and on Georges Bank (Bigelow and Schroeder, 1953, p. 262) had led us to expect a plentiful representation of this species among the catches that were made off Nova Scotia, and along the seaward edge and slope of Georges Bank by the "Cap'n Bill II" in 1952 and 1953. But the results proved otherwise for it was taken in 10 hauls, only, in these two sectors combined, out of a total of 51 hauls that fished i For detailed comparison with other species of Raja, see our earlier account (1953, p. 237). BIGELOW AND SCHROEDER : ELASMOBRANCHS AND OHIMAEROIDS 65 successfully there between 120 fathoms and 400 fathoms and none was caught in 34 deeper hauls. The total number of specimens, too, was only 12, the maximum catch in one haul only 2. This, contrasted with the large catches that have been made on the Newfoundland Banks, on Georges, and in the western side of the Gulf of Maine (Bigelow and Schroeder, 1953, p. 262) is evi- dence that while it has been taken as deep as 430-490 fathoms off New York by "Cap'n Bill II" (Lat. 39°26'N, Long. 72°12'W, July 1953) and at 459 fathoms near Spitsbergen, its center of abundance lies shoaler than the zone along which the great ma- jority of the "Cap'n Bill II" hauls were made. But the sizes of the specimens taken at different depths make it likely that the thorny skate reproduces itself, indifferently, down to the greatest depth to which it occurs, for those taken shoaler than 300 fathoms ranged from 196 mm. to about 860 mm. in length, those taken deeper, from 145 mm. to about 865 mm. And while one from 240-270 fathoms contained an e^ about ready for deposition, an- other from 400-460 fathoms contained a mass of immature eggs. This skate was taken at 8 stations out of 44 along the 120-400 fathom depth zone off southern New England in the two years combined; but, again, the largest catch was only 2 specimens, the total catch 11 and only 1 was caught in 26 stations deeper than 400 fathoms. Five specimens, also, Avere taken in 3 hauls fishing between 253 and 490 fathoms off Long Island, New York (longitudes 72°12'-72°19') in June-July 1953 (it had been re- ported nearer land there, previously). But it seems that the so- called "Hudson Trough" off New York about marks the bound- ary to its regular occurrence in that direction at any depth, for "Cap'n Bill II" did not take it in any of the hauls that she made southward thence to the offing of Virginia in the summer of 1953, though we have seen a specimen from 74 fathoms off Charleston, South Carolina (Bigelow and Schroeder, 1953, p. 263). Raja senta Garman 1885 This skate, ranging from the Newfoundland Banks and the inner part of the Gulf of St. Lawrence to the offing of South Carolina, was already known to occur widespread, though in small numbers, from the offing of western Nova Scotia to the offing of 66 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY New Jersey; chiefly between 50 fathoms and 250 fathoms, but recorded as deep as 478 fathoms off South Carolina. Therefore, it was no surprise to take 1-2 specimens in each of 6 hauls at 125-340 fathoms along the slope of Georges Bank (Long. 66°51'W and 69°43'W), 1 or 2 in each of 7 hauls at 150-225 fathoms off southern New England (Long. 70°-72°01'W), 1 off Maryland (Lat. 38°38'N, Long. 73°10'W), at 190-200 fathoms, and 1 off Virginia (Lat. 37°38'N, Long. 74°14'W) in 1952 and 1953. The series ranges from 120 mm. in length to about 577 mm. ; the larg- est is a male with well developed claspers. Raja spinicauda Jensen 1914 Definite locality records for this cold-water skate had been limited, previously, to Barents Sea, to the waters east of Iceland, to southwestern Greenland and the Greenland side of Davis Strait, to the continental slope off eastern Newfoundland, and to Hermitage Bay on the southern Newfoundland coast. But the fact that an egg case, apparently of this species to judge from its external sculpture and from the embryo contained within it, had been brought in from Banquereau Bank, and that a similar case (empty) had been trawled on the southwestern slope of Georges Bank,1 had made it likely that the range of spinicauda extended southward and westward along the upper part of the continental slope as far as the general offing of Cape Cod. And this likelihood has now been corroborated by the capture of a juvenile male 392 mm. long by "Cap'n Bill II" on the slope of Georges Bank (Lat. 42°17'N, Long. 65°06'W) at 320-360 fathoms, besides two egg cases (one with identifiable embryo) off southern Nova Scotia (Lat. 42°44'N, Long. 63°17'W, 410-420 fathoms) in 1952, and of a third egg case, with well advanced embryo, on the slope of Georges (Lat. 40°43'N, Long. 66°42'W, 405-430 fathoms) in 1953. Spinicauda had been recorded previously from 77-88 fathoms (140-160 meters) off Iceland, 120-404 fathoms off west Green- land, and 120-140 fathoms in Newfoundland waters. The depths recorded for the partly grown male taken by "Cap'n Bill II" in 1952 (320-360 fathoms), as well as for the egg cases taken i See Jensen (1948, pp. 50-52, and 53, Fig. 3) for records far Barents Sea, Iceland and Greenland; Bigelow and Schroeder (1953, p. 272, 276) for those for American waters. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 67 by "Caryn" in 1949 (260-350 fathoms) and by the "Cap'n Bill II" in 1952 and 1953 (405-430 fathoms) suggest that spinicauda occupies a somewhat deeper zone in the southernmost part of its range than in the northern part. The bottom temperature was 3.3° in 1953, at the only locality where spinicauda was taken (egg case with embryo) that summer, and 3.9°-4.4° along the general depth zone where either free-living spinicauda, or its egg case, was taken in 1949 or in 1952, years when the temperature was not recorded. These values suggest that the upper limit to its thermal range may be slightly higher off Georges Bank and off Nova Scotia than off the east coast of Newfoundland, where the bottom water, at the depth where it was taken (91-127 fathoms) was between -1.5° and +0.65°, or in west Greenland and Icelandic waters where it has been found in temperatures of 1.7° to 3.8°. Spinicauda is given so distinctive an aspect by its wedge shaped anterior contour with very long snout, and by a generally prickly upper surface, but with larger thorns confined to a single row of 21-26 (three examined by us had 22, 23 and 24 thorns, re- spectively) along the mid line of the tail with one between the first and second dorsal fins, that it could hardly be mistaken for any other skate known from the western North Atlantic. The only noticeable respect in which our juvenile male differs from the adult is in a somewhat longer tail, a growth feature that also is known for several other skates. More in detail, the distance from the center of the cloaca to the first dorsal fin, relative to the distance from center of cloaca to tip of snout, is about 1.2 times as great on the "Cap'n Bill II" juvenile male, as on the mature male, 1,236 mm. long, from Newfoundland, the proportional dimensions of which are given in Part 2 of The Fishes of the Western North Atlantic" (Bigelow and Schroeder, 1953, p. 272). And this divergence between small specimens and large is about what might have been expected ; the correspond- ing ratio, for example, in length of tail between young and adult (measured similarly) is about 1.2-1.3 for R. erinacea; about 1.2 for R. fyllae; about 1.1 for R. laevis; 1.1-1.3 for R. radiata; and 1.1-1.3 for R. senta. In the adult spinicauda the rostral projection from the front of the cranium (reaching very nearly to the tip of the snout) 68 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY is slender and rodlike along the anterior % of its length, and its cartilaginous nature is evident, if the skin above it be slit, and its sheath of fibrous tissue be spread apart. It is of the same shape in our juvenile male. But in this case the histologic nature of its anterior part remains to be learned, for it is only close to the cranium that it is visible on an X-ray photograph. And we have not felt free to dissect the single specimen, or to treat it with a stain selective for cartilage. Ishiyama (1952, p. 2) has already pointed out that X-rays must be used with caution as tests for cartilaginous skeletal elements in rajids. We have found, for example, that while the outlines of the cranium, of the jaw cartilages, of the vertebral column, and of the basal cartilages of the paired fins show clearly in our X-rays of various skates and rays, those of the pectoral fins may, or may not. Thus, the pectoral radialia are sharply outlined on our X-rays of Gymnura, but the only visible evidence of them on X-rays we have studied of various species of Raja, Breviraja, Cruriraja, Psammobatis and Sympterygia is along the narrow axial strands of calcification. Consequently, X-ray photographs can be relied on as tests for the length of the rostral cartilage only if they show its anterior termination sharply.1 The egg cases trawled by "Cap'n Bill II" in 1952 and 1953 — identity established by the embryos contained in them — resemble those referred provisionally to this species by Jensen (1914, p. 33 ; 1948, pp. 55, 56) and by us (1953, p. 272) very closely. They are unique so far as known, among the egg cases of Atlantic skates, in the sculpture of their surfaces with a large number of low, longitudinal ridges, each close-set with a single series of several hundred stiff rod-like structures with complexly dissected tips (Fig. 4). The three egg cases we have examined measure about 90 by 130 mm. ; about 92 by 135 mm. ; and about 86 by 136 mm., not counting the terminal horns. The embryo that one of them contained, 185 mm. long to its first dorsal fin, and 242 mm. to the tip of its tail, still bears the large yolk sac. But it is so nearly ready for birth that its identity as spinicauda is obvious. Embry- onic characters — still persisting — apart from the yolk sac are that only the tips of the single line of caudal thorns (22 in i See, for example, Bigelow and Schroeder, 1948, Figs. 1, 2 ; 1953, Fig. 62. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 69 number) show through the skin; that the length of the tail, from center of cloaca to first dorsal fin, relative to the distance from center of cloaca to snout, is about 1.2 times as great as in the fe^^tf^l^^-tyt^fr^yr^:,^^ Fig. 4. Raja spinioauda. Portion of egg ease trawled on the Nova Scotian slope in 410-420 fathoms, Lat. 42°44'N, Long. 63°17'W, M.C.Z. No. 37691, to show sculpture; above, about 7 x, below, about 11 x. juvenile male of 390 mm. ; that the tip of the tail still carries the embryonic filament which with the caudal fin fold extends 35 mm. from the rear base of the second dorsal; and that the thorn is not yet visible between the two dorsal fins. 70 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY CHIMAEROIDS Family CHIMAERIDAE Hydrolagus affinis (Brito-Capello) 1867 Deep Water Chimaera This chimaera was described originally from deep water off Portugal. The first report of it, on our side of the Atlantic, was by Gill in 1878 (as Chimaera plumb ea n. sp.), from one brought in from 350 fathoms off La Have Bank. So many of them were brought in during the next few years, by schooners long-lining for halibut, that Goode and Bean (1895, p. 31) characterized it eighteen years later as "very common in the deep water on the outer edge of the banks north of Georges Banks."1 And while one specimen, only, is known to have been taken on the American slope during the past quarter century, we have risked the pre- diction that "it would be found on the offshore slopes in un- diminished numbers if it were sought at the proper depth" (Bige- low and Schroeder, 1953, p. 544). This, then, was one of the fishes we expected the trawlings of "Caryn" and of "Cap'n Bill II" to yield. But they did not catch a single specimen. The failure to take this particular species can hardly be blamed on insufficient coverage of the zone explored, for 38 successful hauls reaching deeper than 200 fathoms were made along the slope of Georges Bank, 26 of which fished deeper than 300 fathoms, 19 deeper than 400 fathoms and 3 deeper than 500 fathoms; 49 hauls reaching deeper than 200 fathoms along the Nova Scotian slope, 39 of them deeper than 300 fathoms ; 21 deeper than 400 fathoms and 7 hauls deeper than 500 fathoms, in the summers of 1949, 1952, and 1953 combined. The fact that all reports of affinis, for the western side of the Atlantic, have been based on specimens caught with hook and line might suggest that our method of fishing was at fault. But the otter trawl is so efficient an apparatus that we think it more likely that affinis actually is far less numerous along the Georges Bank and Nova Scotian slopes today than it was, there, 50-70 years ago. i For reports of it, in the Western Atlantic previous to 1953, see Bigelow and Schroeder, 1953, p. 544. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND OHIMAEROIDS 71 Family RHINOCHIMAERIDAE Rhinochimaera and Harriotta The most striking feature of these bizarre chimaeroids is their very long pointed snout, supported chiefly by the correspondingly long upper rostral cartilage. This cartilage follows the same course in Harriotta as is pictured for it in Rhinochimaera by Garman (1904, PL 1, fig. 2) and by Dean (1904, PI. 1, fig. 4). We may add to our earlier account (1953, p. 549) that it is sep- arated from the skin along the inner part of the snout by a mass of white, semigelatinous, pulpy tissue. The cartilage is in close contact with the overlying skin toward the tip of the snout for a short distance in Rhinochimaera of both sexes, also in females and young males of Harriotta, and for a longer distance in mature males of the latter, the tips of the snouts of which are hard. The cartilage as we have pointed out (1953, p. 549) is more flexible in the vertical plane around the crest of its curvature than elsewliere, so that it is easily bent down there mechanically, even on preserved specimens. But we have yet to learn whether the fish can direct the outer part of the snout upward or down- ward voluntarily. Three genera of rhinochimaerids are known: Neoharriotta Bigelow and Schroeder 1950, type species N. pinnata (Schnaken- beck) 1929 1 (West Africa south of the equator), with separate anal fin ; Harriotta Goode and Bean 1895 (North Atlantic, Japan, and off lower California) and Rhinochimaera Garman 1901 (North Atlantic and Japan), without separate anal. Characters that we have accepted previously (1953, p. 549) as alternative between Harriotta and Rhinochimaera are the nature of the dental plates (smooth in Rhinochimaera but with grinding ridges and knobs in Harriotta), and whether the upper margin of the caudal fin is smooth (Harriotta) or is armed with a longitudinal row of denticulate structures (Rhinochimaera) . Examination of the representatives of each that were taken during the recent trawling trips of "Caryn" and of "Cap'n Bill II" shows that the first of these criteria can be relied upon, except for newly hatched specimens, the dental plates of which are smooth, or very nearly so in Harriotta as well as in Rhinochimaera. In fact we iGiven erroneously as 1931 in Bigelow and Schroeder 1953, p. 550. 72 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY have found no reliable feature to separate newly hatched speci- mens of the North Atlantic representatives of Harriotta and of Rhinochimaera generically, though it is easy to do so specifically. It proves, also, that the presence of denticulations along the upper side of the caudal fin is reliable as a distinctive criterion for adult males of Rhinochimaera, as contrasted with Harriotta, but not for young males, or for females, as is pointed out below (p. 80). And it must remain an open question, until growth series of Rhinochimaera can be studied, as to how early these denticula- tions develop. All we can say, in this regard, is that while the upper margin of the caudal is so thick and fleshy in adult Rhino- chimaera of both sexes that the horny rays are entirely concealed, it is so thin on a female R. atlantica 457 mm. long that these rays are clearly visible. The development, also, of a double series of hard knobs along the terminal %-% of the snout of maturing males in Harriotta raleighana and in its close relative H. chaeti- rhamphus (Tanaka) 1909, of Japan,1 but not in Rhinochimaera seems more properly a generic character than a specific. Another character, not proposed previously as alternative, between Har- riotta and Rhinochimaera, but which seems to be so, is the shape of the outer part of the long upper rostral cartilage which is sub- triangular (base uppermost) in cross section in Harriotta but is nearly as thick along its lower side as along its upper side in Rhinochimaera. Rhinochimaera atlantica Holt and Byrne 1909 This chimaeroid, previously known from the type specimen only, from the Irish Atlantic slope, and from a few empty egg cases presumably referable to it, was represented by 8 specimens in the "Cap'n Bill II" collections. These range from a newly hatched male, 151 mm. in total length, to large adults of both sexes 1060-1315 mm. long. According to Holt and Byrne (1910, p. 19, PI. 3) atlantica differs from pacifica (Mitsukuri) 1895, of Japan, in a relatively shorter second dorsal fin, which they picture as shorter than the distance from the gill openings to the origins of the pelvic fins, but which they characterize as longer i The illustration of the only other member of the genus, H. ctirtis-jamesi Townsend and Nichols 192."). from off lower California (Townsend and Nichols, 11)25, Fig. 2) suggests that the specimen was a female; and in any case it was too small (G inches) to show this character, even if actually a male. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 73 than that distance in pacifiea, as it is shown on Dean's (1904, PI. 1, fig. 1) illustration. But the second dorsal of pacifiea is shown as only about as long as from gill opening to pelvic origin by Mitsukuri (3 895, PI. 1), and as shorter than that distance by Garman (1904, PI. 1, fig. 1). The ratios, tabulated below, between length of base of second dorsal and distance from gill opening to pelvics, for the "Cap'n Bill II" series of atlantica, and for two specimens of pacifiea in the Museum of Comparative Zoology (one of these was the basis of Garman 's illustration) are further evidence that the Atlantic population cannot be separated from the Japanese on this basis. And we may note, in passing, that the point of origin cannot be located with precision either for the second dorsal, or for the pelvics because of the shapes of these fins. Species atlantica Length, to rear base of 2nd dorsal mm. 727 Sex Ratio, base 2nd dorsal to distance gill opening to pelvics 1.0 : 0.98 < i 755 $ 1.0 : 1.09 1 1 770 9 1.0 : 1.0 1 1 825 9 1.0 : 1.08 i i 856 9 1.0 : 1.0 1 1 .880 9 1.0 : 1.0 pacifiea < 1 575 590 1.0 1.0 : 1.0 : 1.0 According to Holt and Byrne (1910, p. 19, Fig. 3, and Footnote) a second difference between atlantica and pacifiea is that the "posterior limbs" of the forked canal that runs rearward along the lower surface of the snout, end blind in atlantica, but join the median anterior loop of the angular canal1 in pacifiea. Ex- amination, however, of the two specimens of pacifiea in the Museum of Comparative Zoology shows that they actually end blind in that form, as they do in atlantica, though the blind termination may lie ver}' close to the median loop of the angular canal, on one side of the head or the other, on some specimens, both of pacifiea and of atlantica, perhaps on both sides in some cases. Garman 's (1904, PI. 2, fig. 2) illustration of the lower i Nomenclature according to Garman (1SS8) and Bigelow and Schroeder (1953, p. 530, Fig. 119A). 74 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY • ■"••*W ■-.■># © o 3 § 53 S - -4— » | § g .5 ^ 3 °.sf S «s a S £ a ^ ^ 03 o a - ^ ^ -a '5s 3 fcUD o o CO S * S 03 S t o CO 03 GO • 2 s S 03 o 1j «H © O 03 03 g . 3 « «5 •** ^3 _•> -3 £ . 03 t« M of J ^ g 03 a O 03 » 3 ° H " -2 I- -* ^ ^ *# o s ci « 03 00 03 p CO -2 o g 'So g 03 a ® 6 - » - ^ 1 °" g v ^ » 3 09 03 O ^ ° 03~ £ 03 CO te a « O cs co © '■S ^ £ P o to « S is 03 e« 03 03 « a 03 i-s CJO S s a £ BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 75 surface of the snout of pacifica is not clear in this respect. Neither have we found anything- to differentiate the Atlantic form (Fig. 5) from the Japanese, either in proportional dimen- sions, in the shapes, sizes or relative positions of the fins, in the nature of the dorsal spine (described for atlantica on p. 76), or in the shapes of the dental plates, which are as free from any trace of grinding prominences in the one as in the other. But our specimens do suggest that the caudal filament may average significantly longer on Japanese specimens than on Atlantic. Thus it is almost as long as the second dorsal fin in one of our pacifica and is pictured as about 70 per cent as long as that fin by Dean (1904, PI. 1, fig. 4), whereas it is only about 10 per cent as long as the base of the second dorsal on one of our atlantica specimens on which it seems to be intact, while it is pictured as very short, indeed, by Holt and Byrne (1910, PI. 3) for the type specimen. It seems, too, that the denticulations (single or paired) with which the upper margin of the caudal fin is armed, on large specimens, are not only more numerous on the Japanese form than on the Atlantic, but that they are regularly present on the females of the former, as well as on the males, but only on some of the females of the latter. Thus there are 47 and 52 series, respectively, on the two adult males of pacifica that we have examined, with ' ' over fifty ' ' reported for it by Dean (1904, p. 6), while his illustrations of an adult female (Dean, 1904, PI. 1, figs. 1, 4) show them as rather prominent and as distributed continuously all along the caudal. But they number only 25 and 30, respectively, on our two adult Atlantic males ; only one of our four large Atlantic females shows traces of them all along the caudal; the caudals of two of them are perfectly smooth ; and they are much smaller on females where they are to be detected at all than they are on adult males. On the other hand, they are larger, relatively, on the large Atlantic males (Fig. 5) than on the Japanese. In estimating the taxonomic significance of these differences, we face the puzzling fact that the caudal denticulations of one of the Japanese males resemble those of the Atlantic males in their low, rounded form, with the interspaces smooth edged between them. But they are sharp pointed (whether single or in pairs) on the other Japanese male of about the same size, with the interven- 76 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ing spaces interrupted, in most eases, by 1-3 much smaller, pointed denticles, as pictured by Garman (1904, PI. 4, fig. 2) many years ago for this same specimen, and by us more recently (1953, Fig. 122). It must remain an open question, until half-grown males can be examined, whether this difference between two specimens, both of which seem to be mature, represents two extremes of individual variation, or whether one of these conditions (and if so, which of them) precedes the other, as a given male becomes active, sexually. To sum up, the North Atlantic population of Rhinoehimaera seems to differ certainly from the Japanese only in a secondary sexual character that is subject to considerable variation from specimen to specimen, and perhaps in the proportional length of a structure (caudal filament) so susceptible to damage that estimates of its length are largely a matter of guesswork for a large proportion of the specimens that have come to hand. Were we facing the question de novo we would hesitate to separate the Atlantic form from the Pacific, specifically, on such slender bases. However, since the name atlantica is in use already, we are content to let it stand for the time being. Reduction to the rank of subspecies may be its ultimate fate. The general morphology of Rhinoehimaera pacifica, internal as well as external, was made well known by Garman 's (1904) and Dean's (1904) studies, accompanied by beautiful illustra- tions. The following details for atlantica, additional to Holt and Byrne's (1910) account, are based on the "Cap'n Bill II" specimens. The rear edges of the dorsal fin-spine are perfectly smooth on all the large specimens, as noted by Holt and Byrne (1910, p. 18) . But they are rough with minute serrations along the outer third of their length on a female 457 mm. long ; evidently they are lost during later growth. Holt and Byrne's illustration (1910, PI. 3) suggests, further, that the spine is attached to the fin right out to its tip (they do not comment on this), not free toward the tip as it is in Harriotta (p. 83). And this seems to be the normal state, being true of one of our large females, and of the largest male as well. But the union must be a feeble one, for the membrane is free from the fin along its outer %-% on all our BIGELOW AND SCHROEDER : ELASMOBRANCHS AND OHIMAEROIDS 77 Fig. 6. Diagrams showing pattern of mucous canals on head. A, Ehino- chimaera atlantioa, same specimen as in Figure 5, left side. B, Same speci- men, right side. C, Dorsal view of male, 1034, mm. long to upper termination of caudal fin, M.C.Z. No. 38243. D, Earriotta raleighana, female, 840 mm. long to upper termination of caudal fin, M.C.Z. No. 38247, left side. E, Same specimen, right side. F, Same specimen, dorsal view. 78 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY other specimens, including a female only 450 mm. long, without any clear sign that the separation had been a forcible one. Our specimens corroborate Holt and Byrne's (1910, p. 21) suggestion that the pattern of mucous canals on the sides of the head is a variable character, as it is in Harriotta (p. 83). It seems usual for the jugular canal and the oral canal to branch off separately from the orbital canal below the eye, with the two separated by a longer or shorter interspace. But it is only on one side of the head that this is the case on one of the large females, and on one of the large males ; on the other side of the head of each of these individuals the jugular canal and the oral canal branch from the orbital as a single trunk which bifurcates some little distance outward from its point of departure from the orbital (Fig. 6A, B). It proves, too, that the canal pattern on the crown is variable, also. Normally, the cranial canals of the two sides are connected across the back of the head by a cross- canal known as the aural.1 But the aural is interrupted, midway, both on one of our specimens of pacifica (see Garman, 1904, PL 2, fig. 1) and on our 457 mm. female of atlantica, with its two parts overlapping. And one of the large males of atlantica shows a still more aberrant state, with the left-hand cranial canal re- curving forward toward the mid line of the head to end blind, with the aural canal interrupted (Pig. 6C). The large pores, also, on the sides of the head, vary in number, not only from specimen to specimen, but between the two sides of the head on some specimens. A count of 13 below the cranial canal on the left-hand side of the head of our largest male atlantica, in the region between eye and frontal tenaculum, but of 10 on the right-hand side, may serve as an illustration. The presence of a few sharp denticles on the mid line of the back between second dorsal fin and caudal, also on the nape, on an atlantica so small (151 mm. long) as evidently to have been hatched recently, deserves mention, also. Rhinochimaera atlantica resembles its relative Harriotta ra- leighana so closely in general appearance that the one might be mistaken for the other on cursory examination, especially since the two are likely to be taken together in trawl hauls at appropriate depths on the slope, as happened on "Cap'n Bill II" i Nomenclature according to Garman (1888, Pis. 1, 2, 4) and Bigelow and Schroeder (1953, p. 530, Fig. 119). BIGBLOW AND SCHROEDER : ELASMOBRANCHS AND OHIMAEROIDS 79 on four occasions off southwestern Nova Scotia. But it is easy to tell partly grown specimens apart, on closer inspection. Thus the glossy smoothness of the dental plates of Rhinochimaera con- Fig. 7. Tracings of pectoral fins, adjusted to equal lengths along outer margin, to show difference in shape ; solid line, Harriotta ralcighana, female about 908 mm. long to upper termination of caudal fin, M.C.Z. No. 37726; broken line, Rhinodhimaera atlantica, female about 1282 mm. to upper termination of caudal fin, M.C.Z. No. 37735. trasts strongly with the grinding ridges and on the dental plates of Harriotta, soon after eyes of Rhinochimaera are noticeably sma Harriotta at all stages in growth as appears comparative table : Length, to termination of upper caudal1 knobs that develop hatching ; also, the Her than those of from the following Species Rhinochimaera atlantica Harriotta ralcighana mm. 1034 1130 1184 1282 1290 1295 283 447 758 840 885 908 Sex $ Ratio, horizontal diameter of eye to distance eye to base of dorsal spine 1.0 : 2.80 s 1.0 : 2.66 9 1.0 : 3.17 9 1.0 : 2.43 9 1.0 : 3.03 9 1.0 : 3.07 9 1.0 : 1.78 S 1.0 : 1.65 2 1.0 : 1.63 9 1.0 : 2.00 9 1.0 : 1.85 9 1.0 : 2.13 1 Approximate measurement 80 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The pectoral fins are noticeably narrower toward the base in Rhinochimaera than in Harriott a but with more broadly rounded tip (Fig. 7). The rear edges of the dorsal fin-spine, which are serrate in small specimens of both fish, and are serrate to maturity in Harriotta, are perfectly smooth on large examples of Rhino- chimaera, male as well as female. And the lateral mucous canal, which runs nearly straight rearward from its point of departure from the occipital canal in Rhinochimaera, or slopes slightly downward (Fig. 5A), bows upward at first, then turns downward — rearward in Harriotta. More conspicuous differences between specimens of the two fish that are nearing sexual maturity are that the upper margin of the caudal fin thickens in Rhinochimaera, while developing a row of prominent denticulate structures in the males (also in some ffmales), but with the tip of the snout continuing smooth and soft, whereas in Harriotta the upper caudal continues thin and smooth-edged in both sexes, but the tip of the snout stiffens in the males (not in females), bends upward, and develops a double series of rounded knobs. It is probable that R. atlantica is more widespread and more plentiful in the eastern side of the Atlantic than the paucity of the previous records (one adult specimen and a few egg cases from the Irish Atlantic slope) might suggest, for this is certainly the case in the west, where "Cap'n Bill II" took it at 6 stations off southwestern Nova Scotia, at 1 station on the slope of Georges Bank, and at 1 station in the offing of New York. A table follows of localities, depths, and number, size and sex of the specimens : Length to termination of upper caudal1 Total length Sex Lat. N. Long. W. Depth in fathoms Date 1512 1.11 $ 42' '22' 64 c ) — ' .i.i 290-340 7/15/53 436 4;") 7 9 42' '40' 63' '51' 465-480 7/26/52 1034 1060 S 38' '52' 72c '51' 415-440 6/27/53 1130- 1130 $ 42c '40' 63 c '54' 520-545 7/12/53 1184 1202 9 40' '07' 68' '30' 420-480 7/13/52 1282 1315 9 42' '14' 65' '10' 490-530 7/28/52 1290 1315 9 40' '11' 68' '16' 480-490 7/14/52 1295 1315 9 42' '38' 64c '10' 460-475 7/13/53 i Approximate measurement 2 Filament lost BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS SI It seems that Rhinochimacra is considerably less plentiful than Harriotta off Nova Scotia and off New England judging from the fact that four times as many specimens of Harriotta (32) as of Rhinochimaera (8) were taken by "Cap'n Bill II." Harriotta raleighana Goode and Bean 1895 Nine specimens of this long-nosed chimaeroid had been re- ported previously from the western side of the Atlantic, from the trawlings made by the "Albatross" in the 1880 's (Goode and Bean, 1895, p. 33) , by the Prince of Monaco in 1913 (Koule and Angel, 1933, p. 75) and by "Caryn" of the Woods Hole Oceano- graphic Institution in 1949 (Bigelow and Schroeder, 1953, pp. 551, 552) at localities scattered along the continental slope from the offing of Chesapeake Bay to the offing of Halifax, Nova Scotia. Three specimens, also, were taken in the eastern side of the Atlantic by the "Michael Sars" in 1910 (Koefoed, 1927, p. 29), one of them near the Canaries, the other two west of Scot- land. We can now report the capture, by "Cap'n Bill II," of 26 specimens of both sexes in 1952, and of 6 more in 1953, ranging from partly grown to adults. Specimens of Harriotta taken by "Caryn" in 1949, and by "Cap'n Bill II" in 1952 and 1953, arranged in latitudinal sequence, south to north Length in mm. Lat. N. 38°43' Long. W. 72°56' Depth fath. 630-675 Date June 30, 1953 No. 1 Sex 9 to termina- tion of upper caudal1 256 to tip of fila- ment 329 39°49' 70°05' 710-730 July 28, 1953 1 $ 375 456 40°07' 68°30' 420-480 July 13, 1952 1 9 850 900 40°10' 68° 16' 490 July 14, 1952 3 $ 702-748 742-837 41°25' 65° 54' 415-490 June 19, 1949 2 $ 9 126-730 155-770 42°14' 65°10' 490-530 July 28, 1952 5 $ 9 157-447 283-550 42°16' 65°08' 370-420 July 28, 1952 2 9 885-925 9462 42°22' 64°55' 290-340 July 15, 1953 1 9 840 948 42°38' 64°04' 440-460 June 17, 1949 1 $ 735 773 42°38' 64°10' 460-475 July 13, 1953 1 $ 758 836 42°39' 63° 58' 520 July 26, 1952 5 S 9 283-756 373-870 42°39' 64°00' 610-625 July 12, 1953 1 9 865 1025 42°40' 63°51' 465-480 • July 26, 1952 3 S 9 188-716 249-792 42°40' 63°54' 520-545 July 12, 1953 1 $ ? 775 42°40' 64°00' 440-450 June 17, 1949 1 $ 705 818 42°41' 63°49' 465-480 July 26, 1952 4 S 250-721 295-815 42°41' 64°02' 385-400 July 27, 1952 3 S 9 741-908 832-935 i Approximate measurement - Tail lost on 925 mm. specimen 82 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The fact that 30 of the total of 36 specimens were taken in the offing of southwestern Nova Scotia, and that 6 of the 12 success- ful hauls that were made there in 1952, at the appropriate depths, yielded Harriotta, shows that this chimaeroid is more plentiful along this part of the slope than earlier records for it might have suggested. The proportion of hauls yielding it in this region was somewhat lower in 1953 (4 out of a total of 10) , though the trawlings of that summer seem to have heen equally successful in general, to judge from the catches of fishes of other kinds. It appears to be less plentiful to the westward of longitude about 66°W than to the eastward, for "Cap'n Bill II" took it in only 4 hauls (6 specimens) along the slope of Georges Bank, and to the westward, out of a total of 41 hauls that she made there at 400 to 730 fathoms, in the summers of 1952 and 1953 combined. On the Nova Scotian slope the upper limit to the regular occurrence of Harriotta appears to lie at about 400 fathoms although a few specimens were taken as shoal at about 340 fath- oms. "Cap'n Bill II" took ralcighana down to 710-730 fathoms, the Prince of Monaco at 728 fathoms (1332 meters, Boule and Angel, 1933, p. 75), the "Albatross" trawled it at 1081 fathoms off Marthas Vineyard, at 991 fathoms off New York, at 707 fathoms off New Jersey and at 781 fathoms off Chesapeake Bay. Depth records for it in the eastern Atlantic are 1014 fathoms (1853 meters) west of Scotland, and 1423 fathoms (2603 meters) near the Canaries (Koefoed, 1927, p. 29). The external aspect of H. ralcighana has been made well known by the successive studies of Goode and Bean (1895, p. 32), Koefoed (1927, p. 29), Roule and Angel (1933, p. 75), and Bigelow and Schroeder (1953, p. 551). Examination of the "Cap'n Bill II" series adds the following details. A — Dermal denticles and dorsal fin-spine. No trace of the juvenile dermal denticles is to be seen (or felt) on specimens more than about 485 mm. long to the rearmost visible rays of the upper side of the caudal fin. A female of 300 mm. (measured similarly), still has one pair of supra-oculars, one pair on the mid- line of the back in the space between the first and second dorsal fins, and 4 pairs between the second dorsal and the caudal. The increase in the size of the dorsal fin-spine with growth is not accompanied by a corresponding increase in the size of the BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 83 serrations along the rear margin of the spine, hence the latter does not feel any rougher, to the touch, on large specimens than on those of medium size. And the serrations may be partly obliterated by maturity, on some individuals, perhaps wholly so. The "Cap'n Bill II" series also verify earlier accounts of the spine as free from the margin of the fin along at least its outer half. This contrasts with the condition in Rhinochimaera atlantica, where the spine is smooth edged from a very early stage in growth, and where it is attached to the fin-margin, right out to its tip, or nearly so (p. 76). B — Caudal filament. When intact, the caudal filament may be as much as 33-44 per cent as long as from snout to last visible caudal fin-ray among the smaller specimens (191-263 mm., total length) . On three of our larger males (680-750 mm., total length) the filament is 111-116 mm. long, and 160 mm. on our largest female (1025 mm., total length). But it is so thread-like toward its tip that there is always a possibility that part of it may have been lost, even on specimens on which it seems to be intact. C — Pattern of mucous canals on head. Cumulative evidence is conclusive, that the pattern of mucous canals on the head varies too widely to have much significance in taxonomy, not only be- tween different individuals, but even between the two sides of the head of a single individual in many cases. On three of the larger males, for example, the jugular canal and the oral diverge jointly from the orbital canal on one side of the head, but separately on the other side with an interspace between them (Fig. 6D, E). When they arise jointly they may separate at once or they may run for a longer distance or a shorter as a joint trunk before they diverge, one from the other. On one specimen of each sex the jugular canal fails to connect with the orbital on either side of the head, while on one male it ends blind on the right-hand side but connects with the orbital on the left-hand side. D — Snout, and sexual tenacula. In H. raleighana the snouts of the males are similar to those of the females, up to a length of 450 mm. or so (to last visible caudal rays) ; and no trace is to be seen of the frontal tenaculum, or of the prepelvic tenacula, although the prepelvic pockets are already formed on newly hatched specimens. But the tip of the snout has begun to curve 84 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY upward, its terminal knobs have formed, the sexual tenacula have developed in the males, and the claspers seem to be ready to function by the time a length of about 650 mm. (to last caudal rays) has been reached. The largest male yet recorded is about 836 mm. in total length (758 mm. to last visible caudal ray). A female 990 mm. in total length including a caudal filament of about 125 mm. as scaled from Roule and Angel's (1933, PI. 4, figs. 34-34a) illustration, and another 1025 mm. in total length, including caudal filament of 160 mm. (Cap'n Bill II" specimen) are the largest yet seen of that sex. REFERENCES Bigelow, Henry B., and William C. Schroeder 1948. New genera and species of batoid fishes. Journ. Marine Research, vol. 7, no. 3, pp. 543-566. 1948a. Fishes of the western North Atlantic. Part 1. Mem. Sears Foundation for Marine Research, no. 1, part 1, pp. 59-576. 1910. New and little known cartilaginous fishes from the Atlantic. Bull. Mus. Comp. Zool., vol. 103, pp. 385-408, 7 pis. 1953. Fishes of the western North Atlantic. Part 2. Mem. Sears Foundation for Marine Research, no. 1, part 2, x-f- 562 pp., plus index. 1954. A new family, a new genus, and two new species of batoid fishes from the Gulf of Mexico. Breviora, Mus. Comp. Zool., no. 24, 16 pp., 4 figs. Bigelow, Henry B., William C. Schroeder, and Stewart Springer 1953. New and little known sharks from the Atlantic and from the Gulf of Mexico. Bull. Mus. Comp. Zool., vol. 109, no. 3, pp. 213-276, 10 figs. Booage, Jose Vinoente Barisoza du, and Felix de Brito Capello 1864. Sur quelques especes inedites de Squalidae . . . Proc. Zool. Soc. London, pp. 260-263. 1866. Apontamentos para a ichthyologia de Portugal. Peixes Plagio- stomos. Pt. 1. Esqualos. Lisboa. 40 pp., 3 pis. Brito Capello, Felix de 1867. Descripc,ao de dois peixes novos provenientes dos mares de Portugal. Jorn. Sci. Math. Phys. Nat, Acad. Lisboa, vol. 1, no. 4, pp. 314-317, pi. 3, fig. 1, pi. 5. BIGELOW AND SCHROEDER : ELASMOBRANCHS AND CHIMAEROIDS 85 Clark, Eobert S. 1926. Eays and skates . . . Sci. Invest. Fisher. Board Scotland, 1926, no. 1, 66 pp. Atlas, 36 pis. COLLETT, R. 1878. Fiske fra Nordhaus-expeditionens sidste togt, Sommeren 1878. Vidensk. Selsk. Christiania, Forhand. (1878) 1S79, no. 14, 106 pp. 1904. Diagnoses of four hitherto undescribed fishes from the depths south of the Faroe Islands. Videns. Selskabs. Forhand. Christiania, 1904, no. 9, 7 pp. 1905. Fiske . . . "Michael Sars" togter i Nordhavet 1900-1902. Rept. Norwegian Fisher, and Marine Invest., vol. 2, part 2, no. 3, 147 pp., 2 pis. Dean, Bashford 1904. Notes on the long snouted chimaeroid of Japan. Journ. Coll. Sci. Imp. Univ. Tokyo, vol. 19, art. 4, 20 pp., 2 pis. Donovan, Edward 1807. The natural history of British fishes ... 5 vols., 120 pis., London. Garman, Samuel 1881. Report on the selachians. Bull. Mus. Comp. Zool., vol. 8, pp. 516-523. 1885. Notes and descriptions taken from selachians in the U. S. Na- tional Museum. Proc. U. S. Nat. Mus., vol. 8, pp. 39-44. 1888. On the lateral canal system of the Selachia and Holocephala. Bull. Mus. Comp. Zool., vol. 17, pp. 57-119, 53 pis. 1901. Genera and families of the chimaeroids. Proc. New England Zool. Club, vol. 2, pp. 75-77. 1904. The chimaeroids . . . Bull. Mus. Comp. Zool., vol. 41, pp. 245-272, 15 pis. 1906. New Plagiostomia. Bull. Mus. Comp. Zool., vol. 46, pp. 201-208. 1913. The Plagiostomia (sharks, skates, and rays). Mem. Mus. Comp. Zool., vol. 36, xiii, 515 pp., and atlas, 75 pis. Gill, Theodore 1878. A new species of chimaera found in American waters. Bull. Philos. Soc, Washington, vol. 2, (1874-1878), 1875-1880, p. 182. Goode, G. B., and T. H. Bean 1895. Oceanic Ichthyology . . . Smithsonian Contrib. Knowl., vol. 30, xxxv+26*+553 pp.; vol. 31, Atlas, 123 pis. Gunther, Albert 1878. Report on the deep sea fishes. "Challenger" Rept. Zool., vol. 22, part 57, lxv+268 pp., 66 pis. 86 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Hertwig, Oscar 1874. Ueber Bau unci Entwickelung der Placoidschuppen und der Zahne der Selaehier. Jena Zeitschr. Naturw., vol. 8, pp. 331-404, pis. 12, 13. Holt, E. W. L., and L. W. Byrne 1908. Second report on the fishes of the Irish Atlantic slope. Fisheries Ireland Sci. Invest. (1906), no. 5, 63 pp., 5 pis., 2 figs. 1909. Preliminary note on some fishes from the Irish Atlantic slope. Ann. Mag. Nat. Hist., ser. 8, vol. 3, pp. 279-280. 1910. Third report on the fishes of the Irish Atlantic slope. Fisheries Ireland Sci. Invest. (1908), no. 4, 26 pp., 4 pis. ISHIYAMA, EEIZO 1952. Studies on the skates and rays . . . found in Japan and adjacent regions. Journ. Shimonseki Coll. Fisheries, vol. 2, no. 2, 34 pp., 4 pis. Jensen, A. S. 1914. The selachians of Greenland. Mindeskr. Jap. Steenstrup, vol. 2, no. 30, 40 pp., 1 pi. 1948. Contributions to the ichthyofauna of Greenland. Spolia Zool. Mus. Hauniensis, Skrifter . . . Universitetets Zoologiske Museum Kobenhavn, vol. 9, 182 pp., 4 pis. KOEFOED, ElNAR 1927. Fishes from the sea bottom. Bept. Sci. Bes. "Michael Sars." N. Atlant. Exped. 1910, vol. 4, part 1, 147 pp., 6 pis. LUtken, Chr. 1887. Korte Bidrag til Nordisk ichthyographi. 4. Vidensk. Medd. naturh. Foren. Copenhagen, pp. 1-4, pi. 1. 1898. The ichthyological results of the expeditions of the "Ingolf". Danish "Ingolf " Exped., vol. 2, part 1, 39 pp., 4 pis. Mitchill, S. L. 1815. The fishes of New York, described and arranged. Trans. Lit. Philos. Soc. New York, vol. 1, pp. 355-492, pis. 1-6. 1817. The fishes of New York . . . Supplement . . . Amer. Monthly Mag. Grit. Bev., vol. 2, pp. 241-248, 321-328. 1825. The hedgehog-ray. Amer. Jour. Sci., vol. 9, pp. 290-293, pi. 6. Mitsukuri, K. 1895. On a new genus of the chimaeroid group Harriotta. Zool. Mag., Tokyo, vol. 7, no. 80, pp. 97-98, pi. 16. BIGELOW AND SCHROEDER: ELASMOBRANCHS AND CHIMAEROLDS 87 MOREAU, EMILE 1891. Histoire naturelle des poissons de la France. Supplement, 144 pp. Paris. Radclitfe, Lewis 1916. The sharks and rays of Beaufort, North Carolina. Bull. U. S. Bur. Fisher., vol. 34, pp. 239-284, pis. 38-49. Regan, C. T. 1906. Descriptions of some new sharks in the British Museum col- lection. Ann. Mag. Nat. Hist., ser. 7, vol. 18, pp. 435-440. Reinhardt, Johannes 1825. Ichthyologiske Bidrag . . . Overs. Dansk. Vidensk. Selsk. Forh. Kabenhavn (1824-1825), pp. 2-3. Roule, Louis, and Fernand Angel 1933. Poissons provenant des campagnes du Prince Albert I de Monaco. Result, camp. Sci. . . . Albert Ier, Prince Souverain de Monaco . . . Fasc. 86, pp. 1-78, 4 pis. SOHNAKENBECK, W. 1929. Uber einige Meeresfische aua Siidwestafrika. Mitt. Zool. Staats- inst. Zool. Mus. Hamburg, vol. 44, pp. 23-45. Tanaka, Shigeto 1909. Descriptions of one new genus and ten new species of Japanese fishes. Journ. Coll. Sci. Tokyo, vol. 27, art. 8, 27 pp., 1 pi. Tortonese, Enrico 1952. Studi sui Plagiostomi . . . Arch. Zool. Ital., vol. 37, pp. 383- 398. Townsend, C. H., and J. T. Nichols 1925. Deep sea fishes of the ' ' Albatross ' ' Lower California expeditions. Bull. Amer. Mus. Nat. Hist., vol. 52, pp. 1-20, pis. 1-4. VArLLANT, L. 1888. Expeditions scientifiques du " Travailleur " et du "Talisman" pendant les annees 1880-1883, Poissons, part 1, 406 pp., 28 pis. Paris. Whitley, Gilbert P. 1939. Taxonomic notes on sharks and rays. Australian Zool., vol. 9, part 3, pp. 227-262. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 Bernhard Kummel, Editor CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM October, 1954 Publications Issued by or in Connection with THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE Bulletin (octavo) 1863 - - The current volume is Vol. 112. Breviora (octavo) 1952 — No. 35 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. 33 is current. Occasional Papers of the Department of Mollusks (octavo) 1945 - Vol. 1, no. 17 is current. Proceedings of the New England Zoological Club (octavo) 1899- 1948 -- Published in connection with the Museum. Publication terminated with Vol. 24. These publications 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. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 Bernhard Kummel, Editor CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM October, 1954 No. 3. Status of Invertebrate Paleontology, 1953 Bernhard Kummel, Editor Documentation and interpretation of data on fossil inverte- brates have actively engaged many scientists for more than a hundred years. The historical development and methods of documentation can be divided into four phases, (a) Description of the faunas of a region; Sowerby's Mineral Conchology and d'Orbigny's Paleontologie Francaise are classic examples of this type, (b) Monographs of zoological groups by stratigraphic horizons, (c) Extensive monographs of animal lineages through all or part of their history, (d) Condensed works or phylogenies covering whole classes or phyla. Needless to say, contributions falling in categories (a) and (b) are primarily documentary, hence must occupy much of the attention of our profession. All interpretation of fossil invertebrates must ultimately stem from this type of work. Contributions in categories (c) and (d) are fewer in number and more difficult to attain. It is in these broad syntheses that we seek ultimate meaning in paleozoology, evolu- tion, and geologic history. Great advances in recent years in genetics and neontology have broadened understanding of population structures, processes of isolation, and intraspecific variation. Whereas the paleontolo- gist can contribute little towards the understanding of the mech- anism of evolution, he can and does contribute the dimension of time and thus supplies substantial evidence for the interpretation of evolutionary patterns. Current taxonomic arrangements of the invertebrates have been built up by many individual special- ists working generally with small and geologically restricted groups. The projected Treatise on Invertebrate Paleontology under the editorship of Raymond C. Moore represents the largest organized effort in the history of our science to summarize knowledge of fossil invertebrate animals. The Treatise will contain data on the morphology, systematics, and evolution of all of the inverte- brate phyla known to paleontology, and it will provide a sub- stantial foundation for further advances. The purpose of the present symposium is to assess in broad perspective the status of invertebrate paleontology. It brings together specialists who will present thumb-nail summaries on the 92 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY status and problems of taxonomy, evolutionary patterns, and geologic history of many of the invertebrate phyla, especially for the non-specialist. The total sum of these contributions is a partial review of current thought and interpretation of fossil invertebrates. These papers were presented orally at the annual meeting of The Paleontological Society in Toronto, Canada, November 10, 11, 1953. Symposium Committee J. Wyatt Durham N. D. Newell B. Kummel, Chairman CONTENTS PAGE I. Toward a more ample invertebrate paleontology. N. D. Newell 93 II. Protozoa. H. E. Thalmann 99 III. Coelenterata. J. W. Wells 109 IV. Echinodermata Pelmatozoa. E. C. Moore 12" Eleutherozoa. J. W. Durham 15 V. Mollusca Pelecypoda. N. D. Newell 161 Gastropoda. J. B. Knight, E. L. Batten and E. L. Yochelson 173 Cephalopoda. B. Kummel 181 VI. Arthropoda Trilobita. H. B. Whittington 193 VII. Graptolithina. O. M. B. Bulman 201 VIII. On development, evolution and terminology of ammonoid suture line. O. H. Schindewolf 217 IX. Macroevolution and the problem of missing links. A. Petrunke- vitch 239 X. Evolution of late Paleozoic invertebrates in response to major oscillations of shallow seas. E. C. Moore 259 XI. Systematic, paleoecologic and evolutionary aspects of skeletal building materials. H. A. Lowenstam 287 Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 I. Toward a More Ample Invertebrate Paleontology By Norman D. Newell The American Museum of Natural History and Columbia University, New York It is essential in any assessment of invertebrate paleontology to keep in mind that it originated, and to a large extent has developed, as a stratigraphic tool. In North America, particu- larly, we have tended to be concerned with the uses of inverte- brate fossils in the solution of geologic problems rather than with the meaning, in the broadest sense, of the fossils. The emphasis has been that of an applied rather than a pure science. Let it be granted that the discovery of the useful attributes of fossils forms one of the grand chapters in the early history of geology, leading directly to the chronological classification of the geological record and all that this implies in the story of life. Although this is, indeed, a tremendous contribution to human knowledge, there is much more to be learned from fossils. Fossil invertebrates, because of their ubiquity, convenient size, and relative ease of recognition, have always been favored over other fossils by field geologists, and it is not surprising that the study of their distribution has become an essential part of stratigraphy. Without the data from these fossils our knowledge of the details of geologic succession, correlation, and past geog- raphies would indeed be very sketchy. The debt which strati- graphic geology owes to paleontology, especially invertebrate paleontology, is great, and the debt which paleontology owes to stratigraphy is equally great. However, it is too often implied that chronology is the chief or the only contribution to human knowledge that our science can make. Nothing can be farther from the truth. The scope of our subject is vast, covering the enumeration, history, distribution and ecology of a score of animal phyla and a hundred or so orders, many of which have not yet really been subjected to intensive inquiry. The literature is diffuse and the numbers of genera are legion. Since the total number of investiga- 94 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY tors is woefully inadequate as compared with the size of the task, there has inevitably been a tendency toward extreme spe- cialization along taxonomic, stratigraphic, and even provincial lines. Under the circumstances, our work has not, in general, been characterized by broad syntheses. There are, however, many signs that we have begun the ascent from an intellectual plateau in our subject since the second World War. Increasing effort to sharpen the utility of fossil invertebrates as stratigraphic guides is leading competent investi- gators to recognize some of the blind alleys where the labors are not rewarded by commensurate returns. Many are beginning to attack problems of broader significance. In the development of commercial micropaleontology by the petroleum industry, we very probably are witnessing the final stage in maturing of theories of paleontologic correlation. The continuous growth of knowledge in this subject has not resulted in revolutionary new develop- ments since the general recognition of the importance of biologic facies. Increasing numbers of stratigraphers are impressed by the fact that stratigraphic relationships of separate outcrops within a sedimentary basin are sometimes more readily deter- mined by simple tracing and matching of strata than by com- parison of suites of fossils. Many of these investigators are giving more attention to fossils as indicators of environment, a field rich in unexploited possibilities. In many cases the evolutionary changes within a fauna during deposition of a stratigraphic sequence are so slight that for all practical considerations the fauna ranges without appreciable change from bottom to top of a succession. The application of statistical methods in the analysis of such a faunal sequence promises to permit further refinement of paleontological zonation. However, in a majority of cases successive modifications induced by shifting environment may be more conspicuous than changes resulting from evolution. Although statistical methods will certainly aid in the recognition of subspecies and in study of obscure trends, the evaluation of these trends, of course, must remain largely subjective. Mathe- matical treatment of fossils does not offer a general solution of our problems. Occasionally the field geologist, on submitting his collection of fossils to the paleontologist for study, is disappointed because STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 95 the material fails to elicit the desired information as to geologic age. It may be that the collection is poorly preserved or frag- mentary or, more commonly, the most distinctive guide fossils are lacking in the material. This may seem to reflect on the competence of the paleontologist or to cast doubt on the whole principle of paleontologic zonation. In very many cases the geologist in search of refinement of zonation of a stratigraphic sequence makes demands that cannot be satisfied by the materials at hand. So-called facies fossils, on the other hand, may lead to valuable inferences with respect to past environments. We suspect that often the wrong answers are sought from the ma- terials at hand. If the signs of change in emphasis are correctly read, I venture to prophesy that developments in invertebrate paleontology dur- ing the next few years will be characterized by increasing empha- sis on the more biological aspects of invertebrate paleontology. Even now, it is hardly appropriate to describe paleontology as the "hand-maiden of geology". Its role in earth sciences is at least as significant as those of petrology or structural geology, PALEOZOOLOGY Zoogeography Geological Emphasis Biological Emphasil Fig. 1. Biological and physical aspects of paleozoology. 96 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY neither of which is primarily esteemed for useful applications to other disciplines. Paleontology is constituted of diverse matters (Fig. 1) many of which have little to do with geologic problems, per se. Evolu- tion, ecology, systematics, comparative anatomy, and zoogeog- raphy of fossil invertebrates offer a wide range of fields of investigation outside the usual scope of stratigraphic geology. Geology can and does benefit from the more biological branches of paleontology; and, of course, the stratigraphic aspects of paleontology are absolutely fundamental for the establishment of chronological sequence and correlation in the history of life ; but balance must be sought in the application of our energies if we are to achieve the most complete understanding of the his- tory of life on earth. The biological aspects of invertebrate paleontology, which have not received as much attention as the physical aspects, now offer the greatest opportunities for investi- gation. By general agreement, paleontology is a full member of the family of earth sciences. It must be stressed, however, that it really involves much more of biological than physical sciences. Tangible evidence of a strong trend in our profession toward a general broadening of the field of interest is manifest in two great group projects of incalculable importance to invertebrate paleontology. They involve active collaboration of more than 150 investigators. They are : The Treatise on Marine Ecology and Paleoecology under the leadership of Harry S. Ladd, and The Treatise of Invertebrate Paleontology under the leadership of Raymond C. Moore. These two modern syntheses will open new vistas to students of invertebrate paleontology. They will provide inspiration and rich sources of information in the quest for a broader understanding of fossil invertebrates. We are in the midst of a new flowering of interest in organic evolution in which genetics, systematics, and paleontology join hands, and the vitality of this movement is making a deep im- pression on our own science. In these notable efforts we see evidence of a gradual widening of appreciation and interest in many of the fundamental prob- lems of invertebrate paleontology. These are exciting develop- ments in which attention is increasingly focused on the fact that fossils are the remains of organisms rather than being simply STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 97 stratigraphic tools. The new emphasis inevitably will lead to a more ample science. The symposium of The Paleontological So- ciety, planned and organized by Dr. Bernhard Kummel, clearly indicates interest in the fundamental problems of our science. If we must limit our attention to the strictly "practical" points of view, many fundamental problems of paleontology may seem utterly esoteric. Many of us, however, give at least lip service to the thesis that pure science is also intensely practical. ' ' The whole of man 's experience has demonstrated that the prac- tical results required for tomorrow depend essentially on the 'impractical' free curiosity of today." The broad problems of paleontology are important to us, if I may paraphrase Warren Weaver (1953), because they have "depth and sweep, because they are esthetically attractive, because they are instances of man's mind seeking to meet the challenge of the universe." reference Weaver, Warren 1953. Fundamental questions in science. Sci. Amer., vol. 189, pp. 47-51. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 II. Protozoa By Hans E. Thalmann Stanford University, Stanford, California In the field of fossil Protozoa great advances have been achieved during the past hundred years. At that time it was relatively easy for a paleontologist to overlook, singly, the different branches of the modern science of micropaleontology. Today even a spe- cialist will have to devote his time either to a single class of microfossils or, as is more often the case, to a selected group within this class. The doctrine of mechanism, Darwin's ideas on natural selection and organic evolution, improvement of optical instruments from the simple microscope to the modern electron- microscope, refinement of methods and techniques and last, but not least, the paleontologist's incurable curiosity to penetrate deeper into the secrets of Nature, have contributed to the accu- mulation of knowledge of the most primitive forms of life buried in sediments since pre-Cambrian time. Truly enough, we are not yet standing on top of the pyramid of knowledge, but each new contribution adds another stone that broadens the foundation on which we build up understanding^ and intelligently the sci- ence of fossil Protozoa, by unraveling their geologic history, their trend of evolution and phylogenetic relationships, and by solving the problems of paleoecology and paleogeography. In the following pages an attempt is made to summarize the status quo of our knowledge of the fossil Protozoa. The limited space available for such a thumb-nail summary does not permit one to give full credit to the legion of authors to whom every paleontologist is indebted for his contributions, small or large. Phytomonadina. So far only one representative, Chlamydomo- nas, is known from the Upper Eocene, plus Phacotus which is known from the Upper Miocene and forms limestones in the Pleistocene, thanks to its calcareous tegument. Xanthomonadina. Although forming siliceous cysts, this group 100 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY is not yet definitely known as fossil. With the help of modern optical instruments, however, there is every possibility that they might turn up when searched for in bituminous deposits. Euglenida. Phacus and Lepocinclis are known from bitumi- nous sediments of Eocene age, and Traehelomonas has been found in the same habitat of Pliocene age. Dinoflagellata. Great progress has been made during the last two decades thanks to the intensive research work by G. Deflandre and 0. Wetzel. Nearly half a hundred genera, three-quarters of which are exclusively fossil, are presently known, most of them organically conserved thanks to the fossilization of their cellulose thecae. They are now traced back to the Jurassic, but their phylogenetic relationships are still only vaguely known. The question of whether the siliceous Dinoflagellata might be an- cestral to the Radiolaria, and the calcareous Dinoflagellata to the Foraminifera, is not yet settled. Ebriidea. These marine planktonic flagellates with siliceous endoskeleton have only recently become known as fossils, appear- ing suddenly in the lower Tertiary. They are already represented by about 20 genera, and more representatives are likely to be found in pre-Tertiary beds with the help of modern optical in- struments and more refined techniques. Silicoflagellata. Relatively little progress has been made in silicoflagellate research. It is an open field for investigation, and the question, whether animal or plant, has not been settled. These marine flagellates with siliceous tests of only a few microns are known since Middle Cretaceous. Curiously enough, fossil Silico- flagellata are at present better known than the Recent ones, and definitely show evolutionary tendencies from geologically older to younger forms. Recent work by G. D. Hanna and Y. T. Mandra, although concerned only with a few Californian sedi- ments already indicates that the Silicoflagellata might have future use as stratigraphic tools and ecologic indicators. C occolithophoridea. Although known as fossils since Ehren- berg (1836), the true nature of these extremely minute, rock- forming, planktonic, calcareous flagellates ("coccospheres" or "coccoliths") was only recognized about 50 years ago by Loh- mann (1902). They still are obscure as to their origin, appearing since Liassic time, but modern research work with the help of STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 101 the polarization- and electron-microscope will certainly enhance our knowledge of this class of Protozoa, which today are grouped into the orders of Heliolithae (3 families with 25 genera), and Ortholithae (3 families with 6 genera). Recent studies indicate their possible stratigraphic value. Some "Problematica" : Cera- tolithus Kamptner, Nannoconus Kamptner, Lithostromatium Deflandre and Peritrachelina Deflandre, all about 12-20 microns, are still not taxonomically pigeon-holed (Flagellata incertae sedis). Chrysomonadina. Formerly assigned to the C occolithophori- dea, they are now separated as a distinct group. Minute siliceous cysts belonging to Archaesphaeridium Deflandre and Micram- pulla G. D. Hanna are knoAvn from Cretaceous beds, others from Eocene lignites, Miocene fresh-water diatomites and Pleistocene peats. The taxonomic position of Palaeostomocystis Deflandre, a planktonic flagellate in Jurassic and Cretaceous beds is not yet settled. Chitinozoa. Discovered in insoluble residues of Ordovician and Silurian limestones by Eisenack, the Chitinozoa are also of un- known systematic position. Their youngest representatives so far are reported from the Middle Devonian but more intensive studies are necessary before their true nature and taxonomic place are fully understood. Thecamoebina. Fossil Thecamoebina have been reported to occur in Middle Eocene deposits, but recent studies by Bolli and Saunders (1954) indicate that most, if not all, so-called fossil forms might in fact be Recent representatives, often erroneously attributed to the Foraminifera. Similarity between tests of The- camoebina and certain genera of the Saccamminidae (Foramini- fera) seems to be responsible for this error. The presence of thecamoebian forms as "fossils" apparently is due to contami- nation from fresh-water rivers, lakes and swamps. "Fossil" representatives of the "foraminiferal" genera Leptodermella and Milletella should now be allocated to the Recent thecamoebian C entropy xis, Lagunculina and TJmulina to Difflugia, and some species of Proteonina to either Pontigulasia or Difflugia. Bolli and Saunders' studies rule out the Thecamoebina as fossils. Tintinnoidea (Calpionellidae) . Great progress has been made in the recognition of the true systematic position of these lori- 102 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY cated oligotrichous Infusoria, which as planktonic microfos- sils were populating the seas during Upper Jurassic and Lower Cretaceous. G. Colom (1948) has excellently summarised our present knowledge of these tiny marine, pelagic and free-swim- ming, therefore widely distributed, organisms of the old Tethys region. They suddenly appear at the base of the Tithonian and persist as rock-formers associated with Nannoconus, coccoliths, and radiolarians, into the Barremian where they abruptly disap- pear. Their presence in sublithographic limestones, where the calcified loricae appear brightly white in a dense grey ground- mass, indicates a warm tropical environment. Few, if any, fossil tintinnoids have been reported from the "Western Hemisphere but a search for them in beds of Jurasso-Cretaceous age will undoubtedly reveal their presence and might help to correlate dense sublithographic limestones when megafossils or Foramini- fera are scarce or absent. Radiolaria. The taxonomy of the Radiolaria, based on the composition of their siliceous skeleton, has recently been sum- marized by A. S. Campbell (1952). These wonderful microscopic animals of the seas have for a hundred years been the object of study for their beauty and multiformity, although still prac- tically nothing is known regarding the formation of the test in these sarcodine Rhizopoda. For the geologist and paleontologist there are only two of the four groups of Radiolaria regarded as fossils of importance, the peripylean Spumellina and the monopy- lean Nassellina, occurring as microfossils in all types of siliceous rocks from Cambrian to Recent all the world over. Their ex- clusively marine planktonic mode of life, although passively transported by currents, is responsible for their world-wide distri- bution, from littoral to deep-sea sediments. They are excellent indicators of paleoecological conditions if and when properly evaluated. Their stratigraphic use is at present confined to local correlations, but further research indicates that many radiolarian genera and species might become good guide-fossils and help out, as in the case of calpionellids and coccolithophorids, when mega- fossils or other microfossils are absent or of little value. Foraminifera. Without any doubt the greatest advances made in the study of fossil Protozoa have been made in the order of the Foraminifera during the last hundred years or so. This prog- STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 103 ress is evident when d'Orbigny's classification of 1852 is com- pared with the modern taxonomic treatment of the testaceous rhizopods in the modern handbooks of Galloway (1933), Glaess- ner (1945), Cushman (1948), and Sigal (1952). Sigal's and Glaessner's classification seems at present to be the most comprehensive and logical one, grouping the order of Foraminifera conveniently into three suborders : the Uniloculini- dea (gelatinous, chitinous or slightly agglutinated tests, not yet found fossil), the Biloculinidea (globular proloculus followed by a tubular chamber), and the Pluriloculinidea (uni- or plurilocu- lar proloculus followed immediately by a series of simple or complicated chambers). These three suborders include seven superfamilies with altogether 61 families and many subfamilies comprising approximately 1200 valid genera and subgenera — thus contrasting strikingly with the seven "orders", few families and less than 100 genera of d'Orbigny in 1852. The tendency in modern foraminiferal taxonomy is overwhelmingly in the direc- tion of a natural classification. For such an end the Foraminifera are excellently suited, thanks to their abundance in all sediments of the geological column which allows morphogenetic and phylo- genetic studies of their evolutionary pattern and geologic distri- bution. It has to be admitted, however, that the ultimate goal in every branch of paleontology has not yet been reached in the Foraminifera : a true natural classification. But at least one important principle is now settled, namely, that the primitive forms were non-septate, and thus precede geologically as well as phylogenetically the cohorts of septate, and especially septate- spirally-coiled, more highly developed forms. Much research work will have to be done with regard to the more complex families in order to understand their phylogenetic relationships. This is evident in Hofker's new attempt at a natural classification (1951), that is by using micro-anatomical studies of the tiny tests in order to establish the phylogenetic concatenation of the genera and families of Foraminifera, fossil and Recent. In establishing true natural relationships it is of utmost importance, according to Hofker, to study and follow in time sequence, the nature and behavior of the dentate buccal apparatus or "toothplate" and its connection either with the protoforamen or with the deuteroforamen in the chambers of 104 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY the foraminiferal test, and the pattern and distribution of the pores in the test-walls. Glaessner (1954) has summarized the new aspect of foraminiferal morphology and taxonomy initiated and so strongly advocated by Hofker. Studies by Arnold (1954) on the evolution of the Foraminifera, stressing their variation and isomorphism, will also shed new light on taxonomy of the group, as will Wood's (1948) interesting investigations of the structure of the wall of the foraminiferal test. A new field of study is now opened with the use of the electron-microscope where magnifications up to 4000 times reveal highly interesting morpho- logical characters (Jahn, 1953). The principle of phylo-morphogenetic analysis, initiated by the late Tan Sin Hok on Cycloclypeus and other larger Foramini- fera, is gaining more and more ground leading to many hitherto unsuspected genetic relationships in different groups of smaller and larger Foraminifera, and will become increasingly more im- portant for stratigraphic correlation purposes. The abundance of material for such studies and for biometrical analysis, and the relatively easy technical preparation of it, is hardly surpassed by any other group of fossils. The geologic history of the Foraminifera dates from the Cam- brian. Presence of calcareous forms in Cambrian time has now definitely been disproved, and it is not before the beginning of Middle Devonian that the first calcareous genera appear. The Lower Paleozoic is the time of the arenaceous genera and species (Ordovician, Silurian and Devonian). "With the sudden evolution and development of the calcareo-arenaceous endothyroids and fusulinids at the end of the Mississippian and throughout the Pennsylvanian, the development of calcareous perforate tests in Foraminifera increases rapidly, especially when the Triassic- Jurassic boundary is reached ; here more than half of the families of Foraminifera have their origin, and persist, with few excep- tions (Orbitolinidae, Spirocyclinidae, Stomiosphaeridae, Gumbe- linidae, Meandropsinidae, Hantkeninidae, Orbitoididae, Discocy- clinidae, Victoriellidae and Miogypsinidae) to Recent time. The spectacular explosive and sudden evolution and develop- ment of the rock-forming Fusulinidae during Pennsylvanian and Permian, and their complete extinction at the end of the Permian is a unique feature of the younger Paleozoic, repeated again, to a STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 105 lesser degree, by the Alveolinidae during Upper Cretaceous to Oligocene times, the Orbitolinidae during Cretaceous, the Orbi- toididae from Upper Cretaceous to Miocene, the Discocyclinidae during Eocene, and the Nummulitidae during the Paleogene. Wide lacunae of knowledge of the Foraminifera (excepting Fusulinidae) still exist generally for the whole of the Paleozoic, Triassic, Jurassic and Lower Cretaceous periods. Many of the pluriloculinid families have their roots in either Triassic or Juras- sic beds and unless we can learn more about the assemblages and faunal composition of these periods the phylogeny of the Fora- minifera will remain incomplete. The large families of Lagenidae and Rotaliidae play an important role during the older Mesozoic, aside from some prominent arenaceous families, and many ances- tral genera will still have to be detected in years to come. Future foraminiferal studies will have to emphasize the importance of the Lower Paleozoic and Lower Mesozoic Foraminifera, and it is a healthy sign that recent monographs are devoted, in increasing numbers, to these so-far neglected faunas. Since a great share of the total crude oil production in the world is found in Upper Cretaceous and Tertiary formations, much more attention has been paid to the foraminiferal assem- blages of these beds in practically every country where oil wells are drilled or exploration for hydrocarbons is under way. Much more is, therefore, known about smaller and larger Foraminifera (except for the Paleozoic Fusulinidae) of the last sixty or seventy millions of years in our geologic history than of the preceding five hundred millions. During the Cretaceous the important planktonic Foraminifera, especially Globigerinidae, appear and these widely distributed "Ammonites of the Tertiary" have since proved to be of enormous stratigraphic value (Globigerina, Globorotalia, Hantkenina, Orbulina, Pseudotcxtularia, etc.) for interregional and intercontinental correlation of contemporaneous beds and formations or, as is the case especially for Pliocene and Pleistocene sediments, for interpreting paleoclimatic and paleo- ecologic conditions. The value of orbitoids and nummulites for world-wide correlations has been known for a long time and the number of benthonic forms, as Globotruncana, Bolivinoides, Flabellina, certain Rotalia species, etc., as time-markers is ever increasing. The Globotruncanae with their short-lived and explo- 106 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY sive development are at present one of the most useful tools for world-wide stratigraphic correlation of Upper Cretaceous sedi- ments from Cenomanian to the end of Maestrichtian. In applied micropaleontology, nowadays, the paleoecologic significance of the Foraminifera is of great importance in recon- structing the bionomics of former habitats and, most valuable of all, in elucidating lateral and vertical facies changes in the sedi- ments. Studies on Recent faunal assemblages in different habi- tats, nekrocoenosis as well as biocoenosis, demonstrate the depend- ance of most of the Foraminifera on varied biological factors ranging from depth and temperature, light penetration, food supply, salinity, calcium carbonate content, nutritious salts, etc. to factors regulating the rate and mode of sedimentation, nature of bottom sediments and influence of litho-f acies of the embedding sediments. In order to establish biochronological zones and paleoecological conditions, each sedimentary basin has to be studied as a unit, covering for each formation present, all facies and habitats from reef and near-shore to deep-sea environment. Only after adjoin- ing basins have been treated similarly as a unit, will it become permissible to generalize or to make inter-basinal or inter-regional correlations and paleogeographic deductions about the individual formations. Up to now quite a few basins have been studied in such a way, and future progress in everything concerned with Foraminifera will be dependent on the amount of team work undertaken, in cooperation with specialists in other fields of paleontology, and thus shall the geologic history of the earth become fully understood and the documents which Nature pro- vides us in the form of large and small fossils shall be intelligently deciphered. Likewise, concerted efforts of every micropaleontol- ogist to reach and achieve a natural classification of the Foram- inifera and other classes of Protozoa, and a purification of their nomenclature will be a conditio sine qua non for the next hundred years. As Albert Gaudry said in 1890 : . . . "nous ne devons desesperer de rien ; peu a peu nous decouvrirons dans les couches terrestres les ancetres des creatures qui nous entourent, si dedicates qu'elles soient." STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 107 SELECTED BIBLIOGEAPHY Arnold, Z. M. 1954. Variation and isomorphism in Allogromia laticollaris: a clue to foraminiferal evolution. Cushman Found. Foram. Research, Contr., vol. 5, pt. 2, pp. 78-87. Bolli, H. M., and J. B. Saunders 1954. Discussion of some Thecamoebina described erroneously as Fo- raminifera. Ibid., vol. 5, pt. 2, pp. 45-52. Campbell, A. S. 1952. An introduction to the study of the Radiolaria. The Micro- paleontologist, vol. 6, no. 2, pp. 29-44. Colom, G. 1948. Fossil tintinnids: loricated Infusoria of the order of the Oligo- trichia. Jour. Paleontology, vol. 22, pp. 233-263. Cushman, J. A. 1948. Foraminifera, their classification and economic use. 4th edit., Cambridge (Harvard Univ. Press). 605 pp. Deplandre, G. 1952a. in: Grasse, P.-P., Traite de Zoologie, tome I, premier fascicule. Paris (Masson). 1952b. in: Piveteau, J., Traite de Paleontologie, tome premier. Paris (Masson). Ehrenberg, C. G. 1838. Die Infusorienthierchen als vollkommene Organismen. Leipzig. 548 pp. Galloway, J. J. 1933. A manual of Foraminifera. Bloomington (Principia Press). 483 pp. Gaudry, A. 1890. Fossiles secondaires, vol. 2, Paris, p. 37. Glaessner, M. F. 1945. Principles of micropaleontology. Melbourne (Univ. Press). 1954. New aspects of foraminiferal morphology and taxonomy. Cush- man Found. Foram. Research, Contr., vol. 5, pt. 1, pp. 21-25. Hofker, J. 1951. The Foraminifera of the Siboga-Expedition. Part III. Leiden (Brill). 108 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY JAHN, B. 1953. Elektronenmikroskopische Untersuchungen an Foraminiferen- sehalen. Zeitschr. f. wiss. Mikroskopie u. f. mikrosk. Technik, Bd. 61, Heft 5, pp. 294-297. LOHMANN, H. 1902. Die Coccolithophoridae, etc. Arch. f. Protistenkunde, vol. 1, pp. 89-165. SlGAL, J. 1952. Ordre des Foraminifera, in: Piveteau, J., Traite de Paleontolo- gie, vol. 1, pp. 133-178; 192-301. Wood, A. 1948. The structure of the wall of the test in the Foraminifera; its value in classification. Quart. Jour. Geol. Soc. London, vol. 104, pt. 2, pp. 229-255. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 III. Coelenterata By John W. Wells Cornell University, Ithaca, N. Y. These remarks are admittedly somewhat subjective, and some points are to be taken as stimulants to cast light in still dark places. I have divided my time amongst three aspects of coelen- terate paleontology : classification, evolution, and paleoecology. The classification (Fig. 1) as a whole is in no better shape than that of most other phyla. Even the name of the phylum is not wholly free from question. The coelenterated animals are obvi- ously divisible into two distinct groups: (1) the Cnidaria (which most people mean when they say Coelenterata), and (2) the Ctenophora. Opinion favors recognition of two phyla (Hyman, 1940), and my comments are wholly confined to the Phylum Cnidaria, divided by tradition and on morphological grounds into three geologically ancient classes: (1) Hydrozoa, (2) Scy- phozoa, and (3) Anthozoa. It is, of course, in the ranks below the classes that we find divergent ideas of classification, and the deeper we go the more we have a feeling of working our way through a dark cellar full of cobwebs, a feeling not peculiar to the Cnidaria. In the Hydrozoa only the order (or class?) Stromatoporoidea has much of a chronological record, but rather than being thereby on firm ground, this extinct group of reef dwellers has given the most trouble at the highest systematic level. They have been claimed for the Foraminifera (Hickson, 1934; Parks, 1935), and sponges (Heinrich, 1914; Twitchell, 1929) as well as for the Cnidaria. The sum of morphological characters favors strongly a hydrozoan connection (Lecompte, 1951), but evidence of this position is largely negative, except that the fundamentally tra- becular structure of the coenosteum is like unto only that of cnidarians. The pelagic siphonophores are best left to the neontologists. Their only fossil representatives are vellelids and porpitids from 110 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the Paleozoic (Caster, 1942), significant only as being indicative of the antiquity of this group. The Scyphozoa, like the siphonophores, are notorious amongst paleontologists for their lack of hard parts. They appear here and there in the geologic column as stray and aggravatingly ^ ^ HYDROIDA TRACHYLINA < o ? _ M o a MILLEPORINA STYLASTERINA SIPKONOPHORA >- i - STR0MAT0P0R0IDEA < o M LUCERNARIINA CARYBDEINA _ __ O a C0R0NATA ? — o SEMEOSTOMEAE — < en RHIZOSTOMEAE » ? — — CERIANTIPATHARIA ALCYONARIA a. < TELESTACEA ALCYONACEA GORGONACEA o 'Z. o COENOTHECALIA PENNATULACEA ? < o M STOLONIFERA O t- z < TABULATA ZOANTHACTINIARIA ZOANTHINARIA RUGOSA HETEROCORALLIA - PTYCHODACTIARIA — CORALLIMORPHARIA — SCLERACTINIA ACTINIARIA ? 7 > Fig. 1. Classification and chronology of the Cnidaria. STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 111 vague impressions from the Precambrian on, but the fossil mate- rial signifies little in the systematics of the jellyfish. Perhaps I may mention here the conulariids, another perplexing group, possessed of a thin, phosphatic skeleton, whose tetrameral sym- metry has suggested to some workers a scyphozoan relationship. Indeed, it is hard to think that such a form as the Ordovician Conchopeltis, the only conulariid for which traces of the tentacles have been found, was not a jellyfish. But then, Conchopeltis may not have been a conulariid. In the Anthozoa we find a really significant chronological rec- ord, and it is on the tabulate, rugose ("tetracorals"), and scler- actinian ("hexacorals") corals that the most work has been done. These are all clearly anthozoan cnidarians. The scleractinians are clearly from the same stock as the corallimorph anemones rather than directly from actiniarian anemones, as indicated lately by studies of nematocysts (Carlgren, 1940, 1943, 1945). But the relations of the extinct tabulates and rugosans to skele- tonless stocks are still speculative ; both are cryptogenic and prob- ably polyphyletic groups appearing first in the Ordovician with- out any known phyletic past. In the corals, as in other organic groups, the systematic value of morphological characters is the chief classificatory problem. Modern classifications are, or ought to be, phylogenetic, and to discover the bases for such classifications, characters must be found which show a certain degree of constancy and which also reflect the fundamental organic plan of the animal. In the rugose and scleractinian corals these characters are to be found in the septa, which reflect the mesenteries, which in turn are the primary structures of the anthozoan polyp. The mode of insertion, devel- opment, and microstructure of the septa are thus of basic signifi- cance. This was first recognized some time since for the scleractinian corals (Pratz, 1882; Vaughan and Wells, 1943) and is now finally being found to lie at the roots of rugosan system- atics (Wang, 1950; Lecompte, 1952). The structure of the septa, when more thoroughly understood than at present, will enable us to thread our way through dense overgrowth of homeomorphs, "morphic equivalents," "radicals," morphogenetic trends, and confusion due to ecologic plasticity. The tabulates have not yet revealed such constant but truly basic structures, a situation 112 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 113 reflected in their chaotic systematica, wherein strongly differing forms are artificially lumped together by homeomorphy rather than phylogeny, on the feeble basis of negative characters (Le- compte, 1952). Thorough analysis of their microstructures, a very difficult task in all Paleozoic corals, should help to clear up the taxonomy of the tabulate group or groups, and may or may not sustain the often-made suggestion of a relationship with the alcyonarians (Hickson, 1924, int. at.). The general evolutionary pattern of the major cnidarian groups is only hinted at in the structural series of living forms, and has found little or no actual confirmation from the fossil forms. It should be remembered that a calcareous skeleton, some- times calcific, sometimes aragonitic, is an evolutionary develop- ment which has appeared at different times in different cnidarian lines, each time being a secondary acquisition consequent on sedentary habits. This scattering appearance of a major struc- tural type seems peculiar to the Cnidaria, in which it has occurred on widely-differing plans, in about half of the major groups. As shown by Figure 2, it appears, solely from the living struc- tural and ontogenetic series, that the hydrozoans are the most primitive cnidarians and hence presumably appeared first, that the scyphozoans came next, followed by the structurally advanced anthozoans. The fossil record does indicate, at least, that the three classes were probably differentiated by the beginning of the Cambrian. Supposed jellyfish occur in Precambrian rocks, and reputed skeletonless zoanthactiniarians are reported from the Cambrian (Clark, 1913; Dollfus, 1875). It is otherwise in the groups with a respectable paleontological record. The evolutionary pattern of the stromatoporoids and tabulates is extremely obscure, partly because they have not really been thoroughly studied, and partly because they seem to have been remarkably self-satisfied organisms. The Milleporina and Stylas- terina have as yet too few known fossil forms to allow any gener- alizations on their history, except that they are a relatively recent development. Of all the cnidarian groups, only two have an abundant fossil record nearly from their first inception — the Rugosa and the Seleractinia. Only the latter is represented by any living forms, 114 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 1 I 1N303« xaviiugi Sn030V13B0 oissvanr oissviai STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 115 which is a fortunate situation, for it sheds some light on the gen- eral evolution of the other anthozoan corals. The principle which serves in the classification of the Scleractinia — that the septal structure is constant within a group and that it is progressively modified in each group, often in more than one direction with the passage of time — is also the guide to the broad evolutionary history of the group (Fig. 3). The concept of morphogenetic trends (Lang, 1917, 1923, 1938; Vaughan and Wells, 1943) also sheds light on the evolution of lesser categories, especially at the generic level. Here, while septal structures remain relatively constant, the form of the corallum changes systematically, often in several directions and on two levels (first-order and second- fSSB\ ME ANDROID INTRATENTACULAR BUDDING PLOCOIO \ fl^l p222ij EXTRATENTACULAR BUDDING PlOCOID \ CERIOIO TRENDS IN COLONY-FORMATION Fig. 4. First-order morphogenetic trends in scleraetinian corals, condi- tioned by modes of colony-formation. 116 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY order trends), from solitary to colonial, from relatively simple to fairly complex colonial structures, the latter being the termina- tion of trends from which little more than mere survival may be expected. Figure 4 shows some of the potential first-order trends in colony formation from a solitary founder polyp. But, if we add to these such seemingly genetic, second-order trends as the encrusting, massive, ramose, and foliate habits, which may oper- ate on nearly all of the first-order forms, something like 45 distinct colonial forms are possible, and the discouraging specter of homeomorphy looms ever and ever larger. For instance, Montlivaltia, a common Mesozoic genus of solitary forms, now stands revealed as a monostomatous progenitor from which, by various first- and second-order trends, many homeomorphic colonial types arose at different times from different species, and many good species of Montlivaltia are known from the late Trias- sic well into the Cretaceous. At present the colonial montlival- tians are generically divided on the basis of modes of colony- formation, but it is obvious that these "genera" are no more than form-genera, and the tracing of polyphyletic developments in this subfamily alone will require very large collections, much time, and painstaking study. One of the most debated questions in the evolution of the anthozoan corals is the relation of the scleractinians to the rugo- sans. It boils down to whether the scleractinians were filiated from the rugosans (Schindewolf, 1942), or whether they were independently derived from skeletonless anemone stocks : the scleractinians from the stock which led to the living corallimorphs and actiniarians (Carlgren, 1918; Stephenson, 1921), and the rugosans from earlier stocks leading to the zoanthids (Duerden, 1902, Hyman, 1940). Chronologically and to some extent eco- logically, the scleractinians succeeded the rugose corals and it is tempting to see in them revivified rugosans. It would seem that the ancestral stock should be found in the late Paleozoic corals, but it is generally admitted that the rugose corals of the Permian are the specialized end-forms of a long-lived line of solitary forms (Schindewolf, 1942). To derive the Scleractinia from these, we have to assume (1) a possible change in skeletal substance from calcite to aragonite, (2) a complete change in the microstructure of the septa, and (3) a change in the mode of mesenterial and STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 117 H O < UJ _j o V) < w o o Z3 O 3 to H O 5 -u -*j a OS a o o fan 118 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY septal insertion. Further, this necessitates acceptance of a mono- phyletic derivation by typostrophic proterogenesis from some early ontogenetic stage of the tylolytic late Permian corals. All of this is assuredly a large order, even for typostrophism. In addition, the Scleractinia, like the Rugosa, are apparently not monophyletic (Fig. 5). At the earliest appearance of both orders, we perceive that each already contains widely divergent groups on the subordinal level, such as the columnariids in the Rugosa, and the astrocoeniids in the Scleractinia. Survival of aspects of the rugosan mode of septal insertion in seleractinians (Schinde- wolf, 1942; whence Cloud, 1948, and Moore et at., 1952), indica- tive of transitional characters, seems to be more apparent than real, even more curious modifications in scleractinian septal insertion being overlooked. The question is not likely to be settled for some time. In recent years the ceaseless search for petroleum has resulted in the discovery of more and more ancient reef structures, espe- cially in the Paleozoic rocks. This has focused attention on the natural history of bioherms and biostromes as organic associa- tions. Some indication of the extent of work on ancient and modern reefs is given by a recent bibliography (Pugh, 1950) which, although incomplete, still lists some 1200 titles. While relatively few of these contribute much to real ecological under- standing of reefs, from them emerges the realization that Paleo- zoic rugose corals were not important as reef constructors, and that the real reefs of the Paleozoic were largely the work of algal and algoid organisms (mostly of uncertain affinities), aided to some extent by tabulate and stromatoporoid corals. Further, few of these structures had the wave-resistant qualities of modern oceanic reefs, possibly because few, if any, of the known Paleo- zoic reefs were oceanic. This understanding results from closer paleoecological and sediment logical studies of Paleozoic reefs and from the extensive ecological work done on Recent corals during the past 25 years (Yonge, 1940). Earlier analyses (Vaughan, 1911, int. al.) of ancient reef associations drew conclusions that were somewhat too broad. This does not mean that ecological principles devel- oped from the study of living corals are not applicable to Paleo- zoic forms. Indeed, such studies show rather clearly why some STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 119 Paleozoic corals were significant on reefs and why others were not. Simple comparison of the growth-forms of rngose and scler- actinian corals suggests that the rugose corals were unfitted for the turbulent environment of reef structures in the cumatic zone. They lacked adhesiveness, and the rootlets developed in some forms were only poor makeshifts, whereas the tabulates and stromatoporoids, found in genuine reef communities, gained stability from their encrusting or strongly adherent habit, just as all important Recent hermatypic corals are stoutly adherent basally to their substrates. Further comparison shows definite superiority in colony-formation and compactness in modern hermatypic corals, as well as in the extinct stromatoporoids and SALINITY (PPM) TEMPERATURE PC.) ILLUMINATION (LUXES) (CLEAREST WATERS) oo DEPTH (METERS) Fig. 6. Some ecological restrictions of the hermatypic corals. tabulates. A few rugose corals were colonial, but their colonies never amounted to much. This requires some explanation beyond the limited range of modes of colony-formation available to rugo- sans by extratentacular budding, and their failure to develop edge-zone. Much of the answer is found in the ecological differ- ences between living reef (hermatypic) and non-reef (aherma- typie) corals. Modern reef corals owe most of their success on reefs to an acquired symbiosis (Yonge, 1931), probably developed during the Mesozoic, with unicellular algae (zooxanthellae), a 120 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY e^ s g o g 3W tO G3 o p-H | 3 •2 'S S3 ,a a 03 03 gS ^s 73 1 2 T3 O 03 i— I o 13 43 a B3 03 ft P ft o S 5~ O 2a & s s o S ft a o 03 -t^ 43 BQ •»• e a, a o g o S *> - u .2 •£ no 43 be a 03 03 rf IS 15 03 r-H o 3 S O 03 93 03 ft £ ft o S 03 § 8 o 1-1 S J- ft S O 03 +J "3 O •<^> o « _ ft- CQ a -M -t-= '-"1 Beceni ermaty Large colonie 03 O 43 be 43 o s o 43 ledium high 03 03 03 t-l ft s 'is ft o 03 be 43 ftl 43 tn 43 e © ^> a 5 -2 (13 e3 £ 03 03 erate arm 4= o 43 Solit few co 03 ft 03 03 © US O 03 03 ri c3 temp to w O 03 S 03 03 s*> 'a 13 03 03 +J 03 c3 'd o 43 o a s -1 as 3 03 £ el 03 & o Z, a ^S *-H ^ s 4= O aS • i-H a! o 03 temp to c ft oS rO Si «. • r-H O *H 4^ O a ® ft? 1 CG g 03 03 cS "3 43 03 a ^ «*H BQ 03 03 •H c3 f-> 03 c3 03 03 3 a o a o 4= 33 03 03 ^- a o3 O be •£ ater erat -t-a 43 be 03 w o a CO 8 O 6 3 1 CO 03 A ft a P S) S3 i- o be o o x be STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 121 curious relationship which has conferred on the corals advantages of increased metabolism or faster growth, with the concomitant potential for immensely larger colonies or populations within a given area — this to forms already structurally better fitted to the reef environment, with, however, restriction to the better- illuminated sites in warmer sea climates (Fig. 6). Recent non- reef corals lack zooxanthellae, their colonies are small, they are widespread even in temperate to cold sea climates, and are restricted mainly by sediment controls. The rugosans are ana- logous to the ahermatypic corals to a considerable degree. Comparisons of some eeologic factors are suggested by the chart (Fig. 7), ].iartly speculative and indicative of areas for further application of ecological principles. It may be suggested that the concept of warm, shallow, clear, well-illuminated tropical seas as the environment of Paleozoic corals in general, by closer analogy with the ecology of Recent corals, applies largely to the stromatoporoids and tabulates, and that the major requirements of the rugose corals were satisfied by temperate to warm (8°- 18°C.) waters and not -too-rapid accumulation of sediments. But there is still much to be learned about the "ecological structure" of corals, living and fossil. REFERENCES Carlgren, 0. 193 8. Die Mesenterienanordnung der Halcuriiden. Lunds Univ., Arsskr., N.F., Avd. 2, vol. 14, 37 pp., 25 figs., 1 pi. 1940. A contribution to the knowledge of the structure and distribution of the cnidae in the Anthozoa. Idem, vol. 36, 62 pp., 16 figs. 1943. East-Asiatic Corallimorpharia and Actiniaria. Kungl. Svenska Vetenskapsakad., Handl., 3rd ser. vol. 20, pp. 3-43, 32 figs. pis. 1,2. 1945. Further contributions to the knowledge of the cnidom in the Anthozoa, especially in the Actiniaria. Lunds Univ., Arsskr., N.F., Avd. 2, vol. 41, 24 pp. 1949. A survey of the Ptychodactiaria, Corallimorpharia, and Acti- niaria. Kungl. Svenska Vetenskapsakad., Handl., 4th ser., vol. 1, 121 pp., 4 pis. 122 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Caster, K. E. 1942. Two siphonophores from the Paleozoic. Paleontographica Ameri- cana, vol. 3, pp. 57-90, 7 figs., pis. 4, 5. Clark, A. H. 1913. Cambrian holothurians. Amer. Nat., vol. 47, pp. 488-507. Cloud, P. E. 1948. Some problems and patterns of evolution exemplified by fossil invertebrates. Evolution, vol. 2, pp. 324-350, 4 figs. Dollfus, G. 1875. Note sur des empreintes attribuables a une actinie (?Palaeactis vetula) dans les schistes cambriens des Moitiers-d 'Allonne. Mem. Soc. nat. Sci. natur.-math. Cherbourg., 2e ser., vol. 9, pp. 224-232, pi. 3. DUBRDEN, J. E. 1902. Eelationships of the Rugosa (Tetracoralla) to the living Zoan- theae. Ann. Mag. Nat. Hist., 7th ser., vol. 9, pp. 381-398, 12 figs. Heinrich, M. 1914. TJeber den Bau und das System der Stromatoporoidea. Centralbl. Min., pp. 732-736. HlCKSON, S. J. 1924. An introduction to the study of recent corals. Univ. Manchester Pub., Biol. Ser. No. 4, vii + 257 pp., 110 figs. 1934. On Gypsina plana, and on the systematic position of the stroma- toporoids. Quart. Jour. Micr. Sci., vol. 76, pp. 433-480, pis. 26, 27, 13 figs. Htman, L. H. 1940. The invertebrates: Protozoa through Ctenophora: New York, xii + 726 pp., 221 figs. Lang, W. D. 1917. Homoeomorphy in fossil corals. Proe. Geol. Assoc, vol. 28, pp. 85-94. 1923. Trends in British Carboniferous corals. Proc Geol. Assoc, vol. 34, pp. 120-136, figs. 15, 16. 1938. Some further considerations on trends in corals. Proc. Geol. Assoc, vol. 49, pp. 148-159, figs. 25-28, pi. 7. Leoompte, M. 1951. Les stromatoporoides du Devonien Moyen et Superieur du bassin de Dinant. Premiere Partie. Mem. Inst. Roy. Sci. Nat. Belgique, vol. 116, 215 pp., 35 pis. 1952. Madreporaires paleozoiques. In: Trait e de Paleontologie, ed. by J. Piveteau, vol. 1, pp. 419-538, 229 figs. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 123 Moore, R. C, C. G. Lalicker, and A. G. Fischer 1952. Invertebrate fossils. New York, xiii + 766 pp. Parks, W. A. 1935. Systematic position of the Stromatoporoidea. Jour. Paleont., vol. 9, pp. 18-29, pla. 6, 7. Pratz, E. 1882. Ueber die verwandtschaftlicher Beziehungen einiger Korallen- gattungen mit hauptsachlicher Berucksiehtikung ihrer Septal- Struktur. Palaeontographica, vol. 29, pp. 81-122, 1 pi. Pugh, W. E. 1950. Bibliography of organic reefs, bioherms, and biostromes. Seis- mograph Service Corp., xxxi -f- 139 pp. SOHINDEWOLF, 0. H. 1942. Zur Kenntniss der Polycoelien und Plerophyllen ; eine Studie ueber den Bau der "Tetrakorallen" und ihre Beziehungen zu den Madreporarien. Abh. Reichsamt Bodenforsch., N.F., heft 204, 324 pp., 155 figs., 36 pis. Stephenson-, T. A. 1921. On the classification of Actiniaria. Quart. Jour. Micr. Sci., vol. 65, pp. 493-576, 20 figs. TWITCHELL, G. B. 1929. The structure and relationships of the true stromatoporoids. Amer. Midi. Nat., vol. 11, pp. 270-306, 2 figs., 9 pis. Vatjghan, T. W. 1911. Physical conditions under which Palaeozoic coral reefs were formed. Bull. Geol. Soc, Amer., vol. 22, pp. 238-252. Vaughan, T. W., and J. W. Wells 1943. Revision of the suborders, families, and genera of the Sclerac- tinia. Geol. Soc. Amer., Spec. Paper No. 44, xv 4- 363 pp., 39 figs., 51 pis. Wang, H. C. 1950. A revision of the Zoantharia Rugosa in the light of their minute skeletal structures. Phil. Trans. Roy. Soc. London, vol. 234B, pp. 175-246, 4 figs., pis. 4-9. Yonge, C. M. 1931. The significance of the relationship between corals and zooxan- thellae. Nature, Aug. 22, pp. 1-7. 1940. The biology of reef -building corals. Brit. Mus. (N.H.), Great Barrier Reef Exped., 1928-29, Sci. Repts., vol. 1, no. 13, pp. 353- 391, 6 pis. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 IV. Echinodermata: Pelmatozoa By Raymond C. Moore University of Kansas, Lawrence, Kansas INTRODUCTION The purpose of this paper is to assess the present status of knowledge concerning fossil pelmatozoans, taking into account various important contributions which have been published dur- ing the past half century. In 1900 Bather organized his own and others' studies on all main groups of echinoderms in a well- illustrated volume of Lankester's "Treatise on Zoology"; this furnishes a very convenient starting point. Bather's lucid de- scriptions of morphological characters and concise outline of classification as then conceived have served during subsequent decades as an indispensable reference and, although a revision which might incorporate work done since 1900 never was pre- pared, the book has been reprinted again and again. Appropri- ately titled sections of the present paper contain brief statements of what seem to be the significant features in Bather 's discussion and furnish records of various contributions which have appeared in later years. In 1900, as generally now, echinoderms were divided into the two main groups called Pelmatozoa (more or less permanently fixed during life) and Eleutherozoa (free-moving) ; it is unim- portant that Bather designated these assemblages as "grades," rather than subphyla or superclasses. The Pelmatozoa were dis- tributed in four classes, respectively named Edrioasteroidea, Cystidea, Blastoidea, and Crinoidea. Jaekel's (1899) proposal to remove from among cystoids the peculiarly flattened forms which he named Carpoidea was not considered by Bather until after 1900 but the group was accepted by Schuchert (1904) and others as an independent class and now generally is so treated (although the name Heterostelea Jaekel, 1899, is used by some). Other classes called Eocrinoidea (Jaekel, 1913) and Paracri- 126 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY noidea (Regnell, 1945) have been introduced also. The relegation of blastoids to rank as a subclass of the cystoids, as recently pro- posed by Regnell (1945), is a backward step which is not accepted. CARPOIDEA The carpoids, which include ovoid to unevenly angular primi- tive echinoderms having a dorso-ventrally much flattened theca composed of few to many plates and generally a very peculiar Fig. 1. Analysis of thecal structure of regular eoerinoids and rhombiferan cystoids. In upward order, the alternately disposed circlets commonly are designated as basals, infralaterals (outlined by heavy black lines), laterals, radials (outlined by heavy black dots), and orals. Numbers for identification of individual plates are indicated according to the system of Forbes (1848), based on Ehombifera. The exterior of plates other than the orals typically (or primitively) bears a pattern of ridges which is confluent across mid- points of suture lines with the markings on adjacent plates, and among rhombiferan cystoids the pore rhombs are similarly arranged. STATUS OF INVERTEBRATE PALEONTOLOGY, 105.°, 127 tail-like stem appendage that narrows to a point, are chiefly Cambrian and Ordovician fossils but some specimens assigned to this group occur in rocks as young as Devonian. Typical genera include Trochocystites, Mitrocystites, Anomalocystites, Placocyst- ites, Cothurnocystis, Lagynocystis, Ceratocystis, and Dendrocyst- ites. All of these forms known to Bather in 1000 were assigned to two families of an order of eystoids called Amphoridea, a heterogeneous assemblage which no longer is recognized. The Carpoidea contain such strange, peculiarly modified echinoderms of several sorts that interpretation of them as representatives of a lineage standing widely separated from other groups is inescap- able. Important additions to knowledge of them have been made by Bather (1913), Chauvel (1939-41) and Hecker (1940), but in general they remain poorly known and not at all adequately understood. Gislen (1927) has published on some carpoids but also has undescribed specimens. Fig. 2. Plate arrangement of regular eoerinoids. A, Macrocystella, Trema- docian (?U.Cam.), England; position of anus indicated by "A"; arrows denote brachioles attached to radials. B, Lichenoides, Middle Cambrian, Czechoslovakia, showing proximal parts of brachioles on both radials and laterals and epispires along sutures; basals are very small, irregular, and their exact number unknown; position of anus not determined. (After Ubaghs.) 128 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY EOCRINOIDEA The Class Bocrinoidea (Jaekel, 1918) was distinguished on the basis of noncystidean features of the theca associated with the occurrence of biserial brachioles such as characterize the cystoids. A typical example is Cryptocrinus, which has an ovoid theca com- posed of three circlets of plates below the small tegmen where five pairs of brachiole facets are found. Because neither pore rhombs nor diplopores penetrate the plates, this echinoderm was placed by Bather in an order called Aporita and classed with the cyst- oids. Detailed studies of Cryptocrinus by Yakovlev (1917-27) led him to conclude that it was derived from rhombiferan cystoids but assignable to a primitive group of crinoids. Regnell (1945) refers Macrocystella and other genera of the Macrocystellidae to the Eocrinoidea, whereas Bather thought that, despite lack of pore rhombs, they belonged in the order Rhombifera of the cystoids. The plates are imperforate but their outer surface bears ridges and grooves disposed approximately normal to suture lines between the plates and so spaced that the surface markings are confluent from one plate to another (Figs. 1, 2A) . Another genus, Lichenoides, which was included in the Macrocystellidae by Bather but now separated in a family of its own, is a remark- able eocrinoid that recently has been studied carefully and described by Ubaghs (1953). Unlike other known early pelmato- zoans, Lichenoides is demonstrated to have possessed two circlets of brachiole-bearing plates near the summit of the theca (Fig. 2B). The restoration of this fossil by Jaekel (1918), copied in various textbooks, is erroneous in showing only a single circlet of such plates with five brachioles all together, and there are other important inaccuracies. Margins of the thecal plates carry sur- fieial canals that lead to pores located along the suture lines ; they constitute so-called epispires (Hudson, 1915) which superficially resemble cystoidean pore rhombs but differ in the lack of internal canals or sacs. It is noteworthy that primitive crinoids such as Palaeocrinus and Carabocrinus among inadunates and Cleiocrinus among camerates have seemingly identical structures. The eocrin- oids have a known range from Lower Cambrian to Middle Ordovician. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 129 PARACRINOTDEA Regnell (1945) introduced the Paracrinoidea as a class of pelmatozoans which is denned by lack of polymeric symmetry of their generally numerous thecal plates, absence of distinction between dorsal and ventral parts of the theca, and the occurrence of uniserial armlike appendages. Typical genera are Comarocyst- ites, considered by Bather as a representative of the cystoids (Rhombifera), and Canadocystis; a number of other forms are doubtfully included in the group. Clearly, these fossils, all of Middle Ordovician age in so far as known, differ markedly from crinoids and they seem to be distinct from true cystoids. The pore-rhomb structures that characterize all thecal plates of Comarocystites are highly developed but unusual. EDRIOASTEROIDEA The edrioasteroids are distinctive, somewhat aberrant pelmato- zoans, characterized by the many-plated flexible nature of their upper surface which bears the mouth, anus, and generally curved ambulacra. They first appear in Cambrian rocks and persist into the Carboniferous. Most of them are discoid, but a few, such as Pyrgocystis, have a stalklike cylindrical form. One genus, Astro- cystites, rather strikingly simulates a blastoid in appearance. First described by Billings in 1854, a dozen genera had become known by 1900 and since that time some 15 additional ones have been defined. They are currently divided among seven families. Chief publications on the edrioasteroids since 1900 are a series of papers by Bather (1915), Bassler (1935-36), and Cuenot (1948). CYSTOIDEA The early Paleozoic pelmatozoans called cystoids were under- stood in 1900 to embrace a wide variety of supposedly most primitive thecate echinoderms. A majority of them are charac- terized by very large numbers of irregularly arranged plates and associated lack of well-defined symmetry. Also, the plates com- monly are perforated by rounded pores or indented by slits that open on the interior of the theca or communicate with thin-walled saclike canals. Slender armlike appendages, which are almost 130 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY exclusively biserial in structure and unbranched, may arise from the tegminal area, or subvective systems leading to the mouth may extend across the surface of thecal plates. Bather divided this assemblage into orders called Amphoridea, Aporita, Diploporita, and Rhombifera and he arranged among them genera which now are removed to the classes called Carpoidea, Eocrinoidea, and Paracrinoidea. Many forms that remain classified as cystoids exhibit constancy of plate arrangement but not well-developed radial symmetry. A few are both regular and symmetrical and among these some offer problems in classificatory assignment. Now known genera of cystoids are approximately 90 in num- ber, of which some 40 belong to the order Diploporita, character- ized by abundance of relatively small plates pierced by almost universally paired pores, and the remainder are assigned to the order Rhombifera. The latter mostly have larger and far fewer plates, which in many genera exhibit a regularity of arrangement that permits application of a scheme of individual plate designa- tions, and determination of homologous thecal elements. Instead of pores, the rhombiferan cystoids possess slitlike openings or internal parallel canaliculate structures that cross sutures be- tween the plates approximately at right angles ; in simplest form, the groups of slits or canals occupy rhomb-shaped areas located half on one plate and half on an adjoining plate but the external appearance of specialized pore rhombs (as in pectinirhombs) is considerably modified. An evident trend in evolution is toward reduction in the total number of pore rhombs and their localiza- tion on particular parts of the theca. Advance in knowledge of kinds of cystoids, during the last half century, is indicated by tabulation of genera recorded in 1900 (25 diploporites and 32 rhombiferans) as compared with present-day numbers (approxi- mately 40 and 50, respectively). The most important published studies since 1900 are papers by Bather (1913) on Middle and Upper Ordovician cystoids of the Girvan district, southern Scot- land; Jaekel (1918) on general structure and taxonomy; Chauvel (1939-41) on Ordovician forms from France; Hecker (1940) on fossils from northwestern Russia; and Regnell (1945-51) on Ordovician and Silurian cystoids from Sweden and Belgium. Diploporita. Some of the diploporite cystoids display note- worthy resemblance to other groups of echinoderms, for example, STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 131 the five spirally curved subvective pathways on the summit of Gomphocystis, which is strikingly like the pattern on the upper surface of various edrioasteroids ; also Proteroblastus and Astero- blastus which rather closely simulate Mastoids. Bather actually classified Asteroblastus as a primitive sort of Mastoid, in spite of the presence of diplopores and occurrence of an indefinite number of irregularly arranged plates. The recently published Traite de Paleontologie (tome 3, 1953) conveniently but ambiguously places this genus and the family Asteroblastidae both in the chap- ter on cystoids and in that on Mastoids. Actually, the diploporite group of cystoids seems to comprise an assemblage that disap- peared without issue. Their known range is from lower Middle Ordovician to Devonian. Rhombifera. The rhombiferan cystoids merit special notice because many of {hem have features closely similar to regular eocrinoids, suggesting genetic relationship to various groups of crinoids, on one hand, and to Mastoids, on the other. Although homologies of several sorts have been noted by several workers between pore-rhomb structures and the hydrospires of Mastoids, neither Bather nor anyone else seems seriously to have explored the possibilities that beginnings of evolutionary differentiation leading to typical eublastoids and to such diverse crinoids as dicyclic inadunates like Palaeocrinus and Porocrinus, and all kinds of dicyclic and monocyclic camerates, may exist within this type of cystoids or the eocrinoids. One purpose of this paper is to point out what seems to be previously overlooked ways in which several puzzling characters of plate arrangements which are diagnostic features of these noncystoid pelmatozoans may be explained. Accordingly, the organization of several representa- tive rhombiferan genera will be illustrated and discussed briefly. Rhombiferans of the family Echinosphaeritidae are judged to be the most archaic representatives of the order both because they are among the oldest known and because they have the most numerous thecal plates, virtually all of which are occupied com- pletely by an unspecialized type of pore rhombs. The exterior of unusually well-preserved specimens may be nearly smooth, as in Echinosphaerites, or very strongly marked ridges running trans- verse to plate sutures may form a distinctive pattern over all of the theca, as in Caryocystites, Orocystites, and Heliocrinitet. 132 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Forms like some of these genera may very well correspond to ancestors of the Cheirocrinidae and other families having more stable structures, which also are more specialized, but because c aTal^h OXAj&^&fo E 4\__/' ' ' " 2 " 7 ~^ p Fig. 3. Plate arrangement of rhombiferan cystoids (laterals and radials differentiated as in Figure 1 and position of anus marked by "A"). The diagrams illustrate stability in placement of pore rhombs on plates 1+5. 12 + 18, and 14 + 15 and they demonstrate tendency of various plates to change in shape and position. A, EcMnoenorinites, M.Ord.; B, Callocystites, M.Sil. ; C, Schizocystis, MJSil.; D, Lepadocystis, U.Ord.; E, Cheiroorinits, M.Ord. ; F, Glyptocystites, MLOrd. (All modified from Bather.) STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 133 homologous elements of these latter are better recognizable, atten- tion will be confined to them. The most regular, simply constructed sort of cheirocrinid has a.theca composed of four 5-plate circlets disposed in alternating position beneath a tegmen formed of five small plates correspond- ing to orals. The arrangement of these elements and the system of numbers for designation of them introduced by Forbes (1848) are illustrated in Figures 3 and 4. Pore rhombs of unspeeialized sort occur on a majority of the plates in the most primitive genera, as for example in Glyptocystites (Fig. 3F), but they are progressively reduced in number and increasingly modified in type as seen in genera which must be interpreted as advanced. 4 1 OTO B Fig. 4. Plate arrangement of rhombiferan cystoids (laterals and radials differentiated as in Figure 1 and position of anus marked by "A"). -4, Pleurocystites, M.Ord.; B, Lovenicystis, U.Sil.; C, Proctocystis, M.Ord. (Modified from Bather, A, and Regnell, B, C.) 134 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The most persistent pore rhombs are those located on plates 1 -f- 5, 12 + 18, and 14 -f- 15. In several genera of this group, such as Ecliinoencrinites, Glyptocystites, Cheirocrinns, Pleuro- cystites, and Proctocystis, thecal plates are marked externally by parallel ridges and grooves that meet suture lines transversely in such a manner as to be confluent with similar markings on neighboring plates. Their pattern exactly corresponds to that of simple pore rhombs in forms like Echinosphaerites and they closely resemble the plates of the eocrinoid Macrocystella, also. It is altogether reasonable to interpret them as traces or deriva- tives of once-functional pore rhombs. They may dwindle into marginal corrugations and along suture lines be represented by regularly spaced pits that hold bundles of ligamentous fibers. They may disappear entirely. These observations of plate fea- tures in rhombiferan cystoids, and likewise in eocrinoids, prob- ably have significance in understanding the more or less identical characters of many crinoid plates. Another important generalization derivable from comparative study of these cystoids relates to placement of the food-gathering subvective structures. In several genera they lie recumbent on the thecal plates radiating from the tegminal region, becoming rather closely similar to blastoids in such forms as Lepocrinites, Proteroblastas, and Cystoblastus. The brachioles also may be free appendages rising from the summit part of the theca. In spite of stability expressed by the almost invariable enlarge- ment of plate 3 in the lowermost circlet and constancy in place- ment of pore rhombs, various genera show noteworthy downward and upward shifts in position of some thecal plates along with change in the plate outlines. In this way, the complement of plates in a given circlet may be enlarged or reduced and the latter sort of change may be effected also by complete disappear- ance of thecal plates. Thus, modifications in the architectural pattern of the whole theca are introduced. It is interesting to learn that the position of the anal opening is almost invariably above or between plates 7 and 8 of the next-to-lowermost (infra- lateral) circlet. Before leaving the cystoids, it is desirable to notice forms which are grouped in the family Caryocrinitidae, for these have mostly very regularly arranged plates in three circlets below the tegmen STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 135 and they exhibit distinct trimerous (or hexamerous) symmetry. Hemicosmites and Caryocrinites are examples (Fig. 5). Pore- rhomb canals, which are well developed on the inner side of plates, do not appear externally, although their presence is * 13 19 *- / ' \ "*--» Fig. 5. Plate arrangement of hexamerally symmetrical rhombiferan cys- toids showing inferred correlation with elements of other regular rhom- biferans (as illustrated in Figs. 3 and 4) ; arrows mark location of brachiole facets. A, Hemicosmites, M.Ord. ; B, Caryocrimtes, M.Ord.-M.Sil. (Modified from Bather.) 136 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY marked by rows of tubercles running from the center to angles of the plates; these tubercles comprise the thin covering of slightly produced pores at borders of the rhombs. Bather applied nomenclature of dicyclic crinoids to these cystoids, designating the lowermost circlet as infrabasals, the intermediate one as basals, and that next to tbe tegmen as radials and interradials (although neither distribution of brachioles, or arms, nor disposi- tion of plates in this third circlet furnish basis for distinction of radials from interradials). The Avhole pattern of thecal structure is explainable readily by downward migration of one plate from the circlet of "laterals" (probably no. 13) in the 4-circlet regular Rhombifera and upward migration of the others. A feature that suggests affinity with camerate crinoids (as noted by Bather) is the subtegminal placement of the proximal part of food passage- ways leading to the mouth. Caryocrinites possesses typical pin- nulate biserial arms of camerate type (Springer, 1926, pi. 33, fig. 37). BLASTOIDEA The Class Blastoidea is generally distinguished from cystoids by the very regular pentameral symmetry of the budlike theca composed of plates which ordinarily are reckoned to be reduced in number to 13 (three basals, five radials, and five interradially disposed deltoids). Actually five lancet plates occur also, placed along the midline of each radial, but they are concealed by numer- ous small side plates of the recumbent ambulacra. Abundant threadlike brachioles are joined to edges of the ambulacra. Bather divided the group into "grades" called Protoblastoidea and Eublastoidea, the former including Blastoidocrinus, Asteroblas- tus, and Asterocystis, whereas the latter comprises all others. Although Bassler (1938), and Bassler and Moodey (1943) have followed this classification, adding Mesocystis to the Protoblast- oidea, most students assign all of these named genera, except Blastoidocrinus, to the cystoids and generally thejT do not recog- nize a division (order) called Protoblastoidea. Probably the most noteworthy contributions to the knowledge of blastoids during the last half century are the studies of Hudson (1907), who defined the order Parablastoidea ; Jaekel (1918), Avho recognized the importance of hydrospire openings in classifying the Eublast- Fig. 6. A, plate arrangement of a regular rhombiferan cystoid, Cystoblas- tus, M.Ord. B, hypothetical evolutionary derivative of a ^re-Cystoblastus type and C, a typical eublastoid in which correlation with elements of pre- sumed rhombiferan ancestral stock is indicated by differentiation of circlets as in Figure 1. Arrows denote position of ambulacra and "A" location of the anus. The Mastoid radials are identified as rhombiferan infralaterals, lancet plates as rhombiferan radials, and deltoids as rhombiferan laterals. 138 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY oidea by introducing suborders called Fissiculata and Spiracul- ata; Wanner (1924-49), who described many remarkable Mas- toids from Permian rocks of Timor; Cline (1936-44), who spe- cially studied some Mississippian genera ; and Bergounioux (1953), who surveyed morphology and taxonomy of the class. On the whole, however, advance in this period is considerably smaller than one might expect. An absorbing problem on which no real progress has been made for a very long time relates to the origin of the Mastoids. That the group has strong cystidean affinities is very evident — so much so, indeed, that Regnell (1945) ill-advisedly proposed to rank the Blastoidea as a subclass of the cystoids. Yet no one thinks that evolution leading to the eublastoids can be traced from such blastoid-like cystoids as Proteroblastus, Asteroblastus, or Cystoblastus. What then are the roots of Mastoids? Observa- tion of the structure and evolutionary trends shown by the regu- lar rhombiferan cystoids leads me to suggest a previously unexplored line of attack. Remembering that thecal elements of the 4-eirclet cheirocrinids show tendencies to shift positions and shapes, as clearly illustrated by Cystoblastus (Fig. 6A), is it conceivable that the ultimately very stable features of eublastoid structure may be derived from this source? A postulate in this direction is offered diagrammatically by Figures 6B and 6C, in which a hypothetical intermediate cystoid-blastoid, descended from a somewhat C y sto blast 'us -like ancestor, is succeeded by a form having all of the structural characters typical of the true blastoids. If this concept accords with the actual course of evolu- tionary change, it becomes evident that such elements as the lancet plates are fundamental parts of blastoid anatomy, being derived from the uppermost circlet (so-called radials, nos. 15-19) of the regular rhombiferans, whereas the deltoids are upwardly moved laterals (nos. 10-14) in interradial position. Fusion of two pairs of lowermost-circlet plates of the presumed rhombiferan ancestor yields the characteristic basal circlet pattern of the eublastoids, with the small (azygous) basal in the antero-right position. The hydrospires, as perhaps most simply shown in Codaster, are strictly homologous to the pore rhombs in the Rhombifera, transecting the sutures between radials and deltoids approximately at right angles, and study of them in thin sections STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 139 under polarized light demonstrates crystallographic continuity of different parts of each hydrospire with thecal plates which they adjoin. Thus, derivation of the blastoids from rhombiferan cys- toids of seemingly very unlike thecal organization becomes a rational hypothesis. CRINOIDEA By far the largest and most varied division of the Pelmatozoa consists of the crinoids. Without doubt, also, this class contains the most highly evolved members of the subphylum, as well as the only representatives found in post-Paleozoic deposits. More than 5,000 species of fossil crinoids have been described, which is sev- eral times the number of all other pelmatozoans combined. Naturally, therefore, the Crinoidea have preeminent importance in paleontological study of fixed echinoderms. Crinoids are distinguished by almost universal well-developed symmetry of the theca (calyx), and presence almost without exception of relatively strong, prominent arms which may be many-branched and very commonly provided also with innumer- able branchlets (pinnules). A more or less elongate stem com- posed of generally discoid, centrally pierced plates (columnals) serves as a means of attachment, although many crinoids (espe- cially modern comatulids) are stemless in adult life. Strati- graphic range of the class is from Tremadocian ( ?Upper Cambrian) to Recent. In 1900, all main kinds of crinoids were fairly well known but the classification accepted by Bather and generally adopted by other workers of the time is radically different from that prevail- ing now. Of course, much new information has been gained from study of innumerable genera which then were unknown and the interpretation of various morphological features has changed. Crinoid genera known in 1900 numbered about 250, whereas additional genera described since 1900 are about 500. Bather divided all crinoids into subclasses called Monocyclica (having only a single circlet of plates below the radials) and Dicyclica (having two circlets below the radials). The monocyclic group contained orders named Inadunata, Camerata, and Adunata, whereas the dicyclic group contained orders named Inadunata, Camerata, and Flexibilia, thus duplicating some of the ordinal 140 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY designations. Not only by this classification but by Bather's discussions the conclusion is expressed that monocyclic inadun- ates (like Pisocrinus) and monocyclic camerates (like Actino- crinites) are much more closely related to one another than respectively to dicyclic inadunates (like Cyathocrinites) and dicyclic camerates (like Bhodocrinites) . This was denied by Springer (1913) and has come almost universally to be rejected. Wachsmuth and Springer (1885) divided the Inadunata into assemblages called Larviformia (dominantly monocyclic) and Fistulata (dominantly dicyclic) which, although widely accepted in publications as late as 1943 (Bassler and Moodey), have been displaced by differently defined orders named Disparida, Hybo- crinida, and Cladida (Moore and Laudon, 1943; Jaekel, 1918; Regnell, 1948; Ubaghs, 1953). The Camerata are divided into orders called Diplobathrida and Monobathrida (Moore and Lau- don, 1943 ; Ubaghs, 1953). Post-Paleozoic crinoids are grouped in the Subclass Articulata. It is impractical to review here the voluminous literature on crinoids which has appeared during the past half century and accordingly mere mention of selected contributions that seem to have most significance must suffice. These include in foremost place: (1) many long and short papers by Frank Springer (1900-26), especially his monographs on the Flexibilia and American Silurian crinoids which not only furnish organization of a vast amount of accumulated studies but importantly extend the frontiers of knowledge; (2) A. H. Clark's (1908-41) addi- tions to understanding of the morphology and classification of articulate crinoids, especially in monographs on living forms; (3) Kirk's (1911-50) discussion of eleutherozoic crinoids and numerous short papers on various genera; (4) James Wright's (1913-53) comprehensive studies of British Carboniferous cri- noids, including description of many new forms; (5) Wanner 's (1916-49) invaluable enrichment of information pertaining to Permian crinoids and penetrating interpretation of their rela- tionships; (6) Jaekel's (1918) taxonomic survey and discussion of the phylogeny of all pelmatozoans ; (7) Goldring's (1923) large memoir on Devonian crinoids of New York; (8) Gislen's (1924-38) work on fossil and living articulates; (9) Sieverts- Doreck's (1927-52) numerous papers; (10) W. E. Schmidt's STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 141 (1930-42) studies of Devonian and Carboniferous crinoids of Germany; (11) work by Moore (1940-52), Moore and Laudon (1943-44), and Moore and Plummer (1937-40) describing many late Paleozoic crinoids and revising classification ; (12) Teichert's (1949) report on Australian Calceolispongia; (13) Termier and Termier's (1949) effort to correlate structural characters of cri- noid groups; and (14) Ubagh's (1943-53) thorough investigation of some European genera and able general account of crinoids representing all types. Collectively, these and the work of many not mentioned may be accounted at least to have matched the progress made in the investigation of crinoids from the time of J. S. Miller's (1821) "Natural History of the Crinoidea, or Lily-shaped Animals" down to 1900. Origin of some crinoid stocks. The oldest known crinoid seems to be a rather poorly described inadunate (called Dendrocrinus cambriensis Hicks) from Tremadocian rocks of England, vari- ously assigned to uppermost Cambrian or lowermost Ordovician. Very little later and almost coincidentally, many representatives of the Inadunata, Camerata, and some Flexibilia made their appearance. The origin and interrelationships of these crinoid groups have been studied again and again without arriving at any firm conclusions but it is agreed that in each group various stable features are distinguishable from the very beginning. For ex- ample, among flexible crinoids this applies to the peculiarities of ray structure and constant arrangement of infrabasal plates (two large and one small, the latter invariably in right posterior position except in rare aberrant individuals). In the monocyclic camerate group called Tanaocrinina, the insertion of an anal plate (tergal) in the circlet of radials and the hexagonal outline of the basal circlet, which very predominantly consists of three equal plates, are constant distinguishing features, whereas radials in contact with one another all around and a pentagonal basal circlet are similarly constant attributes of remaining monocyclic camerates (Glyptocrinina). The phylogenetic significance of these things, not to mention many others, remains almost wholly conjectural. Bather gave detailed attention to homologies observed between various seemingly little related crinoids and to correspondence in structures of some crinoids and cystoids, but he rejected the 142 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Fig. 7. Plate arrangement of lower part of cup in dicyclic camerate crinoids showing inferred correlation with thecal elements of eocrinoids and rhombiferan cystoids as differentiated in Figures 1-4. A, Archaeocrinidae and Ehodocrinitidae, M.Ord.-Miss. ; B, Ptychocrinidae, Dimerocrinitidae, and Lampterocrinidae, M.Ord.-L.Dev. If derivation here suggested is correct, the brachials of each ray are developed in series with plates corresponding to laterals (as in Lichenoides) ; also, the plan of plates shown in A is the more advanced, which is contrary to opinion expressed by Moore and LaudoD (1943). STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 18* ^ U9 143 Fig. 8. Plate arrangement of lower part of cup in monocyclic camerate crinoids showing inferred correlation with thecal elements of eocrinoids and rhombiferan cystoids as differentiated in Figures 1-5. A, Glyptoerinina, U.Ord.-Perm. ; B, Tanaocrinidae, U.Ord.; C, Actinocrinitidae, Miss.-Perm. ; D, Plan of basal circlet in all families of Tanaocrinina except Tanaocrinidae, Dichocrinidae, and Acrocrinidae, showing orientation of pentastellate lumen and (by dotted lines) position of sutures between plates in a 5-plate circlet. If derivation here suggested is correct, the brachials of each ray are de- veloped in series with plates corresponding to infralaterals of eocrinoids and rhombiferan cystoids; also, the ancestral stock that gave rise to the Tanao- crinina (Figs. B-D) must have developed the hexameral organization of caryocrinitids, whereas the Glyptoerinina descended from a normal penta- meral type of eocrinoid. 144 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY postulate that rhorabiferan cystoids (or pelmatozoans of this t3^pe, including forms now classed as eocrinoids) could be ancestral to any crinoid group. I undertake here to demonstrate by diagrams (Figs. 7, 8) that the respective dorsal cup patterns of all types of camerate crinoids are directly derivable from eocrinoids or conceivably from regular rhombiferan cystoids of sorts that be- long either to the Cheirocrinidae or Caryocrinitidae ; this calls merely for longitudinal shifting of thecal plates in a manner clearly shown within the cystoid assemblage. It is interesting to observe that, whatever the ultimate stable arrangement of cup plates in the crinoids may be, the orientation of the pentagonal or pentastellate lumen at the center of the lowermost circlet (con- tinuous with the stem) remains constant, although where this opening is circular, orientation of the chambered organ cannot be detected. A radical innovation which is introduced by this approach to explaining camerate cup architecture is the deduc- tion that infrabasal plates of dicyclic cups are identical to basal plates of monocyclic forms and not elements of originally dif- ferent circlets. It follows that monocyclic camerates are not descendants of dicyclic camerates produced by diminution and ultimate disappearance of the lowermost circlet (so-called infra- basals). It is pertinent to call attention to the persistence of stereom folds or ridge patterns on cup plates of very many of the camerates, both dicyclic and monocyclic, reflecting the arrange- ment of pore rhombs belonging to their presumed eocrinoid or cystoid ancestors. Much more lengthy discussion than can be given in this paper is needed in order to analyze the many implications of homologies which are newly suggested here. Such discussion may be under- taken in later writing. For the present, I assert only that the postulates now offered cannot plausibly be extended to embrace any of the Flexibilia nor possibly more than a very few kinds of inadunates (perhaps Hybocrinida, Carabocrinidae, Palaeo crinidae, and some others). No trace of affinities with any cys- toids, eocrinoids, paracrinoids, or carpoids is recognized in the many kinds of flexible crinoids, disparid inadunates, or most of the cladids. If all camerates are descended from eocrinoid- or cystoid-like ancestors, whereas most other crinoids are not, the Class Crinoidea is polyphyletic. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 145 REFERENCES Bassler, R. S. 1935. Classification of the Edrioasteroidea. Smithson. Misc. Coll., vol. 93, no. 8, pp. 1-11, pi. 11. 1936. New species of American Edrioasteroidea. Ibid., vol. 95, no. 6, pp. 1-33, pis. 1-7. 1938. Pelmatozoa Palaeozoica. Fossilium Catalogus, pt. 83, W. Junk, 'S Gravenhage. Bassler, R. S., and M. W. Moodey 1943. Bibliographic and faunal index of Paleozoic pelmatozoan echino- derms. Spec. Paper Geol. Soc. America, no. 45, pp. 1-734. Bather, F. A. 1900. Echinoderma. in Lankester, E. R., Treatise on zoology, A. & C. Black, London, pt. 3, pp. 1-344, illus. 1908. Jiingeres Palaozoicum von Timor, genus Schisoblastus. Neues Jahrb. Min. Pal., Beil.-Bd., vol. 25, pp. 303-319, pi. 10. 1912. Notes on Hydreionocrinus. Trans. Edinburgh Geol. Soc, vol. 10, pp. 61-76, pi. 8, figs. 1-14. 1913. The Trenton crinoid, Ottawacrinus. Canada Geol. Survey, Bull. Victoria Mus., vol. 1, pp. 1-16, pi. 1. 1913. Caradocian Cystidea from Girvan. Trans. Roy. Soc. Edinburgh, vol. 49, pt, 2, pp. 359-529, pis. 1-6, figs. 1-80. 1915. Studies in Edrioasteroidea, I-LX. Collected reprints from Geol. Mag., Wimbledon, Eng. 1917. The base in the camerate monocyclic crinoids. Geol. Mag., dee. 6, vol. 4, pp. 206-212, figs. 1-9. Bergounioux, F. M. 1953. Classe des blastoides. Traite de Paleontologie, ed. J. Piveteau, Masson, Paris, vol. 3, pp. 629-650, figs. 1-36. Chauvel, J. 1939- Recherches sur les cystoi'des et les carpoides armoricains. Soc. 1941. geol. et miner. Bretagne, vol. 5. Clark, A. H. 1908- Numerous short papers on crinoid morphology and taxonomy. 1914. 1915- Monograph of the existing crinoids. Bull. U. S. Nat. Mus., vol. 1941. 82, pts. l-4a. Cline, L. M. 1936- Blastoids of the Osage group, Mississippian. Jour. Paleont., 1937. vol. 10, pp. 260-281, pis. 44, 45 ; vol. 11, pp. 634-649, pis. 87, 88. 1944. Class Blastoidea, in Shimer, H. W. and R. R. Shrock, Index fossils of North America. Wiley, New York, pp. 133-137, pis. 50, 51. 146 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY CUENOT, L. 1948. Anatomie, ethologie, et systematique des echinodermes ; Traits de Zoologie, ed. P. Grasse, Masson, Paris, vol. 11, pp. 3-275, figs. 1-312. 1953. Classe des heterosteles, classe des cystides. Traite de Paleontolo- gie, ed. J. Piveteau, Masson, Paris, vol. 3, pp. 599-628, figs. 1-12, 1 31. Forbes, E. 1848. On the Cystidae of the Silurian rocks of the British Islands. Mem. Geol. Survey Great Britain, vol. 2, pp. 483-534, pis. 11-23. GlSLEN, T. 1924. Echinoderm studies. Zool. Bidr. Uppsala, vol. 9, pp. 1-316, figs. 1-351. 1927. A new Spanish carpoid. Ark. Zool., vol. 19, pt. 2, pp. 1-3. 1934. A reconstruction problem. Lunds Univ. Arsskr., n. f., avd. 2, vol. 45, no. 11. 1938. A revision of the Recent Bathyerinidae. Ibid., vol. 34, no. 10. GOLDRING, W. 1923. Devonian crinoids of New York. Mem. New York State Mus., vol. 16, pp. 1-670, pis. 1-60. Heoker, R. 1940. Carpoidea, Eocrinoidea, und Ophiocistia des Ordovmums des Leningrader Gebietes und Estlands. Acad. Sci. U.R.S.S., Trav. Inst. Paleont., vol. 9, pt. 4, pp. 5-82, pis. 1-10. Hudson, G. H. 1907. On some Pelmatozoa from the Chazy limestone of New York. Bull. N. Y. State Mus., vol. 107, pp. 97-152, pis. 1-10. 1911. Studies of some early Siluric Pelmatozoa. Ibid., vol. 149, pp. 195-272. 1915. Some fundamental types of hydrospires with notes on Porocrinus smithi Grant. Ibid., vol. 177, pp. 163-165. Jaekel, O. 1899. Stamingesehichte der Pelmatozoen, 1, Thecoidea und Cystoidea, Berlin, pp. i-x, 1-442, pis. 1-18. 1900. Ueber Carpoideen, eine neue Klasse von Pelmatozoen. Zeitschr. deutseh. geol. Gesell., vol. 52, pp. 661-677. 1918. Phylogenie und System der Pelmatozoen. Palaont. Zeitschr., vol. 3, pp. 1-128, figs. 1-114. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 147 Kirk, E. 1911. The structure and relationships of certain eleutherozoic Pelma- tozoa. Proc. U. S. Nat. Mus., vol. 41, pp. 1-137, pis. 1-11. 1929- Numerous short papers on crinoids in various journals. 1950. Moore, R. C. 1940. Relationships of the family Allageerinidae. Bull. Denison Univ. Sci. Lab., vol. 35, pp. 55-137, pis. 2, 3, figs. 1-14. 1950. Evolution of the Crinoidea in relation to major paleogeographic changes in earth history. Internat. Geol. Congr., Rept. 18th Sess., London, pt. 12, pp. 27-52, figs. 1-18. Moore, R C, and L. R. Laudon 1943. Evolution and classification of Paleozoic crinoids. Spec. Paper Geol. Soc. America, no. 46, pp. 1-153, pis. 1-14, figs. 1-18. 1944. Class Crinoidea. in Shimer, H. W., and R. R. Shrock, North American Index fossils, Wiley, New York, pp. 137-209, pis. 52-79. Moore, R. C, and F. B. Plummer 1937. Upper Carboniferous crinoids from the Morrow subseries of Arkansas, Oklahoma, and Texas. Bull. Denison Univ. Sci. Lab., vol. 32, pp. 209-313, pis. 12-16, figs. 1-37. 1940. Crinoids from the Upper Carboniferous and Permian strata in Texas: Bull. Univ. Texas, no. 3945, pp. 1-468, pis. 1-21, figs. 1-78. PrVETEATJ, J. 1953. Classe des edrioasteriodes. Traite de Paleontologie, ed. J. Pive- teau. Masson, Paris, vol. 3, pp. 651-657, figs. 1-10. Regnell, G. 1945. Non crinoid Pelmatozoa from the Paleozoic of Sweden. Lunds Geol. Miner. Inst., Medd. 108, pp. 1-255, pis. 1-15, figs. 1 30. 1948. Swedish Hybocrinida. Arkiv f. zoologi, vol. 40A, no. 9. 1951. Caradocian-Ashgillian cystoid fauna of Belgium. Mem. Inst. Roy. Sci. Nat. Belgique, vol. 120, pp. 1-47, illus. Schmidt, W. E. 1930. Die Echinodermen des deutsches Unterkarbons. Jahrb. Preuss. geol. Landesanst., pp. 1-92, pis. 1-3, figs. 1-20. 1934- Die Crinoiden des rheinischen Devons, Teil 1. Ibid., Abh., n. f., 1942. Heft 163, pp. 1-149, pis. 1-34; Teil 2. Abh. Reichanst, f. Boden- forschung, n. f., H. 182, pp. 1-253, pis. 1-26. SOHUCHERT, C. 1904. Siluric and Devonic Cystidea. Smithson. Misc. Coll., vol. 27, pp. 201-272, pis. 34-44. 1913. Cystoidea. Maryland Geol. Survey, Lower Devonian, pp. 227-248, ols. 32-36. 148 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY SlEVERTS (DOREOK.), H. 1927. Ueber die Crinoidengattung Marsupites. Abh. Preuss. geol. Landesanst., n. f., heft 108, pp. 1-73, pis. 1-5. 1934- Numerous papers on crinoids, blastoids, etc., in various journals. 1952. Springer, Frank 1901. Uintacrinus, its structure and relations. Mem. Mus. Comp. Zool., Harvard Coll., vol. 25, no. 1. 1911. On a Trenton echinoderm fauna at Kirkfield, Ontario. Mem. Geol. Survey Canada, no. 15P, pp. 1-68, pis. 1-5. 1913. Crinoidea. in Zittel-Eastman, Textbook of Palaeontology, Mac- Millan, London, 2d ed., vol. 1, pp. 173-243, figs. 267-346. 1920. Crinoidea Flexibilia. Smithson. Inst. Pub. 2501, pp. 1-486, pis. A-C, 1-75, figs. 1-51. 1921. Dolatocrinus and its allies. Bull. U. S. Nat. Mus., vol. 115, pp. 1-78, pis. 1-16. 1923. On the fossil crinoid family Catillocrinidae. Smithson. Misc. Coll., vol. 76, no. 3, pp. 1-41, pis. 1-5. 1926. Unusual fossil crinoids. Proc. U. S. Nat. Mus., vol. 67, art. 9, pp. 1-137, pis. 1-26. 1926. American Silurian crinoids. Smithson. Inst. Pub. 2871, pp. 1-239, pis. 1-33 ; numerous other short papers. Teiohert, Curt 1949. Permian crinoid Calceolispongia. Mem. Geol. Soc. America, vol. 34, pp. 1-132, pis. 1-26, figs. 1-24. Termier, H., and G. Termier 1948. Les echinodermes du Paleozoique inferieur. Revue Sci, Paris, no. 3298. 1949. Hierarchie et correlations des caracteres chez les crinoides fos- siles. Bull. Service Carte Geol. de 1 'Algerie, ser. 1 (Paleont.), no. 10, pp. 1-69, pis. 1-8. Ubaghs, Georges 1943. Note sur la morphologie, le biologie, et la systematique du genre Mespiloorinus. Bull. Mus. Nat. Hist. Belg., vol. 19; no. 15. 1945- Contribution a la connaissance des crinoides de l'Eodevonien de 1947. la Belgique. Ibid., vol. 21, no. 16; vol. 23, no. 4. 1950. Le genre Spyridiocrinus. Ann. Paleont., vol. 36. 1953. Classe des crinoides. Traite de Paleontologie, ed. J. Piveteau, Masson, Paris, vol. 3, pp. 658-773, figs. 1-166. 1953. Notes sur Lichenoides priscus Barrande, eocrinoide du Cambrien moyen de la Tchecoslovaquie. Bull. Inst. Roy. Sci. Nat. Belgique, vol. 29, no. 34, pp. 1-24, figs. 1-12. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 149 Waohsmuth, Charles, and Frank Springer 1885. Revision of the Palaeoerinoidea. Proc. Philadelphia Acad. Sci. Wanner, Johannes 1916. Die permischen Echinodermen von Timor, Teil 1. Palaont. von Timor, Lief. 6, Teil 11, pp. 1-329, pis. 96-114, figs. 1-88. 1924. Die permischen Krinoiden von Timor. Mijn. nederl. Oost-Indie, Jahrb., Verhandel. 1921, Gedeelte 3, pp. 1-34-8, pis. 1-22, figs. 1-61. 1924. Die permischen Echinodermen von Timor, Teil 2. Palaont. von Timor, Lief. 14, Abh. 23, pp. 1-81, pis. 1-8, figs. 1-31. 1930. Neue Beitrage zur Kenntnis der permischen Echinodermen von Timor, IV. Flexibilia. Dienst Mijnb. nederl. -Indie, Wetensch. Mededeel. 14, pp. 1-52, pis. 1-4. 1937. Ibid., VIII-XIII. Palaeontographica Suppl.-Bd. 4, Abt. 4, Lief. 2, pp. 57-212, pis. 5-14, figs. 1-82. 1949. Ibid., XVI. Palaeontographica Suppl. Bd. 4, pp. 1-56, pis. 1-3. Wright, James 1913- Numerous papers on British crinoids, incl. 1939, Scottish Car- 1953. boniferous Crinoidea. Trans. Roy. Soc. Edinburgh, vol. 60, pt. 1, pp. 1-78, pis. 1-12, figs. 1-86. 1949- A monograph of the British Carboniferous Crinoidea. Mon. 1953. Palaeontograph. Soc, London. Yakovlev, N. N. 1917- Some new data on Cryptocrinus and the connection between the 1918. Crinoidea and Cystoidea. Ann. Soc. Pal. Russia, vol. 2. 1927. Sur l'homologie dans la structure de la face ventrale du calice de Cystoidea et de Crinoidea. C. R. (Dokladi) Acad. Sci. U.R.S.S.; 1930. Le genre Petschoraorinus et le passage des crinoides dicycliques aux crinoides monocycliques. Ibid. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 195-1 TV. Echinodermata: Eleutherozoa1 By J. Wyatt Durham Museum of Paleontology, University of California Berkeley 4, California. Inasmuch as all classes of eleutherozoan Echinodermata are well represented in the Recent faunas, a realistic appraisal of our knowledge of fossil Eleutherozoa requires a comparison with the living fauna. According to A. H. Clark (1948) there are about 1500 species of Asteroidea, the same number of Ophiuroi- dea, and 771 species of Echinoidea in the Recent fauna (no estimate for Holothuroidea). Mortensen (1928-1951) in his comprehensive work lists 867 species and 88 varieties of living Echinoidea. Pratt (1935) lists 1100 species of Asteroidea, 1600 species of Ophiuroidea, 600 species of Echinoidea, and 650 species of Holothuroidea as the approximate numbers of these organisms that occur in the Recent faunas. Because of his greater familiarity with the Echinodermata it seems probable that Clark's estimates are more nearly correct than those of Pratt. Clark also indicates that most of the species are inhabitants of shallow water areas, from low tide down to depths of about 1500 feet. Thus the occurrences of most living Eleutherozoa would compare favorably with the depth facies of much of the fossil record. A general survey of the literature (not exhaustive) indi- cates that as of 1952 there were described from the fossil record at least 300 species of Asteroidea (first record in the Ordovician), at least 179 species of Ophiuroidea (first record in the Ordovi- cian, including about 74 species referable to the Auluroidea auct.), at least 3 species of Somasteroidea (restricted to Ordo- vician), and about 7200 species (about 4500 Trregularia) of Echinoidea (first record in the Ordovician). D. L. Frizzel (per- sonal communication, June 19, 1953) states that he recognizes i A contribution from the Museum of Paleontology of the University of Cali- fornia, Berkeley 4, California. 152 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY /750i /500- 1000- 750- ^ 500 250 0 Fig. 1 <3 Living Eleutherozoarv E chiriodenmt a Worldwide. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 153 7200 7/00 CU O CL. E r> 2 <3 § CD 0 ^ a 1 o ^ , 159 limestone. All other records are based on sclerites. All authenti- cated records of fossil holothurians are from Devonian or younger rocks. The supposed pre-Devonian records cannot be referred to the group. Mortensen (1928-1951; esp. vol. 5, pt. 2, pp. 565-573) has sum- marized data on fossil echinoids as well as that derived from the living echinoids and presented a number of important con- clusions on their phylogeny. The first of these is that Bothrioci- daris, considered by Jackson and others to be the ancestral stock of the Echinoidea is not a true echinoid, but more probably a cystoid. The lack of an open ambulacral furrow in the Middle Ordovician Bothriocidaris and its occurrence in the Middle Ordo- vician Myriastiches and Upper Ordovician Aulechinus (which Mortensen considers as ancestral to later echinoids), as well as other morphological reasons, militate against Bothriocidaris being considered the ancestor of the echinoids or even being considered a member of the Class Echinoidea. Mortensen further points out that the Paleozoic arehaeocidarids and lepidocentrids have living descendants and that there is no sharp distinction between the Paleozoic and later echinoids. He also notes, although he does not recognize it in his classification, that the Irregular echinoids are derived from several, perhaps 3 or 4, sources among the Regular echinoids. Mortensen also indicates that the Order Camarodonta (and perhaps others) among the Regular Echinoids is polyphyletic and thus an un- natural group, although he continues to use it. These facts, as well as others not noted here, point to the need for a revision of the currently used major classification of the Echinoidea. Of considerable significance is the establishment of the pres- ence of growth lines (Durham, 1951; Zoeke, 1952) in the plates of several different kinds of echinoids (and presumably, there- fore, in all). The growth lines show the changes in shape the plates have undergone during ontogeny, and inasmuch as they are complete and not interrupted in all cases observed, they ap- pear to cast considerable doubt on the occurrence of resorbtion as a factor in growth and change of shape in the various plates of the test (however this is not evidence against the complete re- sorbtion of plates around the peristome). 160 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY LITERATURE CITED Clark, A. H. 1948. Animals Alive. 472 pp., ill., New York (D. Van Nostrand Co.) Clark, H. L. 1946. The echinoderm fauna of Australia, its composition and its origin. Carnegie Inst. Wash., Publ. 566, pp. 1-567. 1948. A report on the Echini of the warmer Eastern Pacific, based on the collections of the Velero III. Univ. S. Calif. Publ., Allan Hancock Pacific Exped., vol. 8, no. 5, pp. 225-352, pis. 35-71. Durham, J. W. 1950. 1940 E. W. Scripps Cruise to the Gulf of California. Part II. Megascopic, paleontology and marine stratigraphy. Mem. Geol. Soc. Amer., vol. 43, pt. II, pp. 1-216, pis. 1-48. 1951. Mode of growth in some echinoids. Bull. Geol. Soc. Amer., vol. 62, p. 1434 (abstract). Durham, J. W., and W. A. Roberts 1948. Cretaceous asteroids from California. Jour. Pajeont., vol. 22, pp. 432-439, pis. 65-66. Eaton, J. E., U. S. Grant, IV, and H. B. Allen 1941. Miocene of Caliente Range and environs, California. Bull. Amer. Assoc. Petrol. Geol., vol. 25, pp. 193-262, pis. 1-9. Fisher, W. K. 1911- Asteroidea of the North Pacific and adjacent waters. Bull. U. S. 1930. Nat. Mus., vol. 76, pt. 1 (1911), pp. 1-419, pis. 1-122; pt. 2 (1928), pp. 1-245, pis. 1-81; pt. 3 (1930), pp. 1-356, pis. 1-93. Grant, U. S., IV, and L. G. Hertlein 1938. The west American Cenozoie Echinoidea. Univ. Calif., Los Angeles, Publ. Math. Phys. Sci., vol. 2, pp. 1-226, pis. 1-30. MORTENSEN, Th. 1928- A monograph of the Echinoidea. 5 vols, in 17, Copenhagen (C. A. 1951 Reitzel). Pratt, H. S. 1935. A manual of the common invertebrate animals. Pp. 1-854, figs. 1-974, Philadelphia (P. Blakiston's Son and Co., Inc.). ScmJCHERT, C. 1915. Revision of Paleozoic Stelleroidea with special reference to North American Asteroidea. Bull. U. S. Nat. Mus., vol. 88, pp. 1-311, pis. 1-38. Spencer, W. K. 1951. Early Paleozoic starfish. Philos. Trans. Roy. Soc. London, Ser. B, vol. 235, pp. 87-130, pis. 2-8. Zoeke, M. E. 1952. Sur la croissance du squelette des Clypeaster fossiles. Compt. Rend. Acad. Sci. Paris, vol. 234, pp. 1999-2002. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Statics of Invertebrate Paleontology, 1953 V. Mollusca: Pelecypoda By Norman D. Newell The American Museum of Natural History and Columbia University, New York The pelecypods form a truly diverse group of invertebrates of great geological antiquity for which more than 7500 generic and subgeneric names have been proposed, some 2000 to 3000 of which may prove to be useful after elimination of homonyms, synonyms and unrecognizable categories. Classification of the fossil forms involves special difficulties because knowledge of internal shell characters is required, espe- cially details of the hinge and muscle insertions which generally are not clearly visible. The majority of pelecypod shells are, and apparently have always been, formed mainly of the unstable mineral aragonite, hence shell characters of examples from forma- tions older than the Tertiary commonly are obscured or destroyed as a result of recrystallization, solution and other diagenetic changes. Therefore, an understanding of morphological details in the older pelecypods is dependent on exceptionally well-pre- served specimens which are even more essential for the pelecypods than for other molluscs. For this, and for other reasons, existing classifications of pelecypods have not adequately taken into account the early geological history of the group, and many of the current views with respect to pelecypod taxonomy are only crude inferences based mainly or solely on comparative morphol- ogy of living forms and therefore are not applicable to the greater part of the history of the class. Many diverse classifications of the bivalves have been proposed by both paleontologists and neontologists and each has stressed one or another line of evidence. None really has taken full ad- vantage of existing knowledge.1 Classifications in which higher categories are based primarily on soft anatomy have been some- i Haas (1938) and Iredale (1939) have reviewed the voluminous literature on the most significant attempts to develop a satisfactory classification of the pelecypods. 162 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY what favored by zoologists as compared with those based on skeletal characters. It has frequently been assumed that muscles and organs are somehow more fundamental than skeletal parts, and hence are the best indicators of relationship. On the con- trarj-, all lines of evidence must be considered, anatomy, morphol- ogy, ontogeny, and paleontology, and of these the evolutionary records of the stratigraphic succession provide, after all, the most valuable clues in the search for phylogeny. Although this will seem obvious nowadays to many systematists, there are still a few who are not acquainted with the richness of the fossil record and who are not impressed by the vastness of geologic time. After the pioneer efforts of Linnaeus and Lamarck, the group- ing of the bivalves followed expediency for many decades. Then, with remarkable insight, the great Viennese paleontologist, Neumayr, developed a comprehensive classification in 1883 based primarily on features of the pelecypod hinge (Neumayr, 1883, 1891). His conclusions soon gained wide recognition and adop- tion, partly because of Neumayr 's great personal prestige, but even more because of the practical value of his results which have to some extent influenced all subsequent work on the group. Six years after the appearance of Neumayr 's first work on pelecypod systematics, Dall (1889) introduced an outline of his classification which was to prove influential, particularly in America. Fundamentally, his work, which was not fully devel- oped until 1895 (Dall, 1895), was not radically different from that of the Austrian paleontologist, but it incorporated many improvements. Dall followed Neumayr 's lead in emphasizing the importance of the pelecypod hinge in classification of the group. He also recognized the value of shell form and structure, adductor asymmetry, and development of siphons as taxonomi- cally very significant. Unfortunately, he was somewhat intolerant of the pioneer contribution of Neumayr and he produced some confusion by the introduction of several new terms for approxi- mately the same groups recognized in the earlier work. While Neumayr and Dall were working on pelecypod systematics, Pel- seneer (1889), a French zoologist, proposed a wholly new classi- fication in which the higher categories were based on details of gill structure, but unlike Neumayr and Dall he gave practically STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 163 no consideration to the succession of fossil forms and he relegated shell characters to a subordinate place among bases of classifica- tion. Consequently, his conclusions were at considerable variance with previous arrangements and they tended to bring together many dissimilar kinds of shells. This classification, with minor alterations, is still favored by a few neontologists, but a majority now use a somewhat uncritical and heterogeneous synthesis by Thiele (1934) of the works of Neumayr, Pelseneer, and Dall. In 1895 Dall, considering the work of Neumayr and Pelseneer, arrived at a taxonomic arrangement which, with but little modi- fication, is embodied in the final edition of the Eastman-Zittel Textbook of Paleontology published in 1913 and has been gen- erally followed by American paleontologists. However, this clas- sification, since it appeared half a century ago in essentially its final form, does not agree in many respects with the conclusions of many investigators about phylogeny. The last really notable attempt to build a classification of pelecypods (founded on somewhat new lines) was that of Douville (1912). Douville was one of the leading paleontologists of the time and he made use of much more ample paleontological data than were available in the earlier efforts. Although very imper- fect and incomplete, his classification represents a real advance and has been widely used by British and French paleontologists. Neumayr, Pelseneer, and Dall did not have access to sufficient data on the succession of pelecypod faunas to enable them to arrange their higher categories in a temporal sequence. Further- more, none of these students displayed any considerable under- standing of the principles of animal evolution. Their conclusions were based mainly on comparative morphology with only general recourse to the faunal sequence. In no sense were they based on phylogenetic analyses. In his work on the pelecypods, Douville took advantage of the principle of adaptive radiation which already had become a basic aspect of vertebrate phylogeny. He also took into account valuable ontogenetic data worked out by the French paleontolo- gist, Bernard (1895-1897), but generally not appreciated by other investigators. It is unfortunate that the American and German students of pelecypods have tended to overlook Douville 's con- tribution. 164 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The fundamental general studies over pelecypod taxonomy- were mainly between 1889 and 1912, and later general classifica- tions are based in one way or another on the earlier efforts ; hence they are not really new. Since they do not make use of steadily accumulating paleontologic evidence, they do not repre- sent very much progress toward an understanding of the general Areas Scallops Oysters Actinodonfs Fig. 1. Probable phylogeny of the Arcacean pelecypods and their aniao- myanian relativea since the Ordovician period, illustrating marked parallel- ism in the development of hinge characters. This is the ' ' sedentary ' ' branch of Douville. Representation is generalized; particular genera are not in- tended. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 165 phylogeny of the Pelecypoda. It now seems clear that the ma- jority of existing families and superfamilies of pelecypods origi- nated before the Jurassic period. The faunas of older times, although still poorly known, are much better understood today than they were in Douville's time, and it is probable that the origins of many of the higher categories will be worked out when available data on the older pelecypod faunas are evaluated. Parallel evolution is common in the pelecypods where separate but related lines pass through closely similar morphological stages. Convergence between more distantly related groups is also a complicating factor which may result in an artificial association of basically unlike groups. Attention to the strati- graphic sequence of forms enables us in many cases to discrim- inate between convergent and parallel trends (Fig. 1). For example, some of the Recent Arcaceans possess taxodont denti- tion and because of this have been placed by most students close to the nuculoids. However, as pointed out by Douville, the Mesozoic and Paleozoic ancestors of the Arcaceans are not at all like the nuculoids. They possess a very different dentition of the type (cyrtodont-parallelodont) which characterizes many primitive Pteriaceans, Mytilaceans, and allied pelecypods. The taxodont dentition of the Recent Arcaceans represents a very late convergence toward the basically unlike nuculoids. The Arcaceans and their ancestors possess a distinctive com- plex ligament type which I have termed duplivincular (Newell, 1937, 1942). This is found in living pelecypods only among Arcacea, and in the ancestry of Pteriaceans, scallops, and the marine mussels. In all but the Arcacea the duplivincular liga- ment was replaced before the Jurassic by other kinds of liga- ments. Ligament evolution in these groups has followed parallel trends. In the majority of living Pteriidae the ligament is similar to that of living oysters, but as shown by the sequence of fossils, it was developed separately in the two lines from different ances- tors that possessed the duplivincular ligament. This is a trend which has independently characterized several lineages. Attention to ligament characteristics, almost completely ig- nored by pelecypod systematists in the past, promises to provide additional clues to phylogeny. For example, in both the pterioids and the scallops the ancestral duplivincular ligament is replaced 1G6 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY by a so-called external ligament possessing one or more resilia. In the former this change occurred during the Permian, in the latter during the De\ronian. These grades of ligament structure, arrived at independently through selection and parallel adapta- tion,1 certainly are useful in recognition of higher categories, but they may not be of equal value in different groups. The same applies to many other characters of pelecypods, among which may be enumerated the structure of gills, respiratory siphons, size and distribution of adductor muscles, possession of a byssus, shell microstructure, and dentition. The experience of the majority of students since Neumayr has tended to confirm his conclusions with respect to the relative conservatism of pelecypod dentition, and the emphasis given hinge teeth and sockets in classification of pelecypods probably has been justified, even though other characteristics must also be given weight. Independent development in separate stocks of closely similar dentitions, as in the Nuculaceans and Arcaceans cited above, probably is not common and in any case can be evaluated by reference to the temporal sequence of forms. The teeth and sockets of the pelecypod hinge lie below the hinge axis. They do not serve as fulcral points, hence they are not closely analogous to similar convergent structures in brachio- pods and ostracodes. Their main function in pelecypods is to guide the valves as they close so that they will always fit snugly at the margins. Obviously a good marginal fit between the valves is of high selection value in those pelecypods that are not buried and protected by the substratum. R. Anthony (1905) has shown that the position of the hinge axis within the ligament is not permanently fixed. By means of a smoked paper and a stylus cemented to one valve of a living pelecypod (Fig. 2), he was enabled to record the successive, often erratic, wandering of the hinge axis during opening and closing of the valves. In the first two examples of Figure 2, dental articu- lation is loose and the ligament flexible, hence the valves deviate laterally in the sagittal plane. These forms normally are bur- rowers so that a truly efficient hinge is not needed. In the other i E. R. Truenian, who is systematically investigating pelecypod ligaments, has shown experimentally that there is a direct correlation between mechanical efficiency of the ligament and the mode of life (Truenian, 1953). Hence it seems practically certain that variations in ligament structure are adaptive and that the observed evolutionary trends in ligaments are also adaptive. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 167 R. Anthony, 1905. II ScUifO-ttt6&u*>*> Iff IV. Mniil*tlf*AT£LLAC£A ir If APLYSIACEA 1 PERACLIDACEA SACCOGLOSSA \BASO ORDER T STYLOMM h Fig. 1. Proposed phylogeny of the Gastropoda. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 175 was concentrated in the attempt to discover missing links and to construct hypothetical ancestral types, all with little attention to or understanding of fossils. Paleontologists, on the other hand, busy describing genera and species, were seldom concerned with supra-familial categories and were content to have their fossils placed in divisions erected for living species. Paleozoic gastropods suffered the most from this haphazard treatment because they differ most from the living forms. Mesozoic species were generally classified more rationally. Those of Cenozoic age, which differ little from those living today, were for the most part correctly classified but placed without roots into the past. The great difference in Paleozoic gastropods as compared with living ones can be seen by observing the distribution of the major categories in time as shown in Figure 1. Note the concentration of Archaeogastropoda in the Paleozoic era and the Mesogastropoda and Neogastropoda together with the Opisthobranchia and Pul- monata in the late Mesozoic and Cenozoic eras. It has only been within the past decade that sufficient informa- tion about the anatomy, embryology, and physiology of critical living species has been available. This information combined with previous observations — and above all with a deeper understand- ing of the more ancient fossils — has made possible the recogni- tion of the probable ancestral groups of the major ordinal cate- gories and some understanding of the evolution among them. Let us turn to an example of what can be done by correlating the anatomy of the soft parts with conchology. Among the Mono- placophora, which range from early Cambrian to Devonian, is a group of cap-shaped shells, the Tryblidiidae. For many years they were considered to be primitive patellids, which they do resemble superficially in shape. In 1938, Wenz made the sugges- tion that those ancient forms were actually very primitive bilat- erally symmetrical animals. He based this idea on the presence of symmetrically paired dorsal muscle scars as contrasted with the horseshoe-shaped muscle scar of the patellid shell. If Wenz was correct, as we believe he was, it follows that these primitive forms did not undergo torsion of the soft parts as do the patellids but were quite similar in internal organization to the chitons (Polyplacophora) . 176 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY This was a revolutionary idea in gastropod systematics. It was arrived at by discounting superficial shell resemblances and con- centrating on what might be learned from the record of the soft parts that was preserved in the shell. In this case it was merely the scars left by the pedal muscles, but it was enough to give a clue to the probable internal anatomy of monoplacophorans. The cap-shaped shell of Patella and patellid like forms has been achieved by many unrelated rock-clinging gastropods ; it is most often an ecological adaptation and does not necessarily point to relationships. Unfortunately, except in cap-shaped shells, muscle scars are extremely difficult to observe, even in Recent gastropods. Other features in the shell often correlate with specific features of the anatomy. One of the most important shell features that can be correlated with anatomical characters is the emargination or channel that marks the position of the anal tube. Since the ctenidia (primary gills) and other pallial organs have fairly definite relationships to the position of the anus, this may give information as to whether there was the primitive pair of gills or only a single gill. An anterior canal or notch accompanied by certain related features commonly indicates an inhalent siphon. A heterostrophic nucleus has been shown to characterize certain opisthobranchs inhabiting the present-day plankton. These ex- amples indicate the type of inferential data that is important in the estimation of the anatomy of extinct gastropods. If we then reconstruct the soft anatomy of our early gastropods inferentially, we can discuss their probable evolution in terms of three principal adjustments : flexibility in the Isopleura and in- creased motility and enhanced sanitation in the Anisopleura. To return to the monoplacophoran, we have inferred that it had the anatomy of a bilaterally symmetrical organism, much like that of the chiton. The major evolutionary advance within the Isopleura was the replacement of the single monoplacophoran shell by a series of eight transverse plates. This gave the animal greater flexibility and permitted it to cling closer to the irregular surfaces of the rocks. So successful was this adjustment to its narrow environment that the chitons, appearing for the first time in the late Cambrian, survive today with very little change and virtually no adaptive radiation. It may be appropriate to say at this point that most zoologists STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 177 who are unfamiliar with the fossil record and who may not be aware of the existence of the Monoplacophora regard the chitons as a separate class reflecting characters of the ancestors of the gastropods, if not all mollusks. As paleontologists we prefer to be guided by the fossil record and suggest the Monoplacophora for the ancestral role. Whether the Isopleura be regarded as a subclass or a separate class from the Anisopleura is relatively unimportant; if the current separation is maintained, we feel that the Monoplacophora should be included with the chitons (see Figure 1). Stomach- Muscle^ scars Mouth -A/entricle -Auricle Ctenidium Anus A. Primitive Isopleuran (left side view) Stomach Ventricle —Auricle B. Primitive Anisopleuran (right side view) Reconstructions from Knight (1952). Fig. 2. Schematic reconstructions of two gastropods. The introduction of the phenomenon of torsion gave rise to the Anisopleura with greatly increased motility and a newly important problem of sanitation. Figure 2A represents the left side of a primitive monoplacophoran restored with an organiza- tion similar to that of a chiton, shown with the shell transparent. Note that the anus lies in a posterior mantle cavity between a pair of ctenidia. As in the chiton, the posterior position of the anus obviates the need of any elaborate provisions for sanitation. Note 178 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY also that the broad, low shell with its equally broad, low foot and its rather complex muscle attachment, permit little movement of the shell and probably allowed only very sluggish progression. Figure 2B represents a restoration of a gastropod-like mollusk of the early Cambrian that is thought to be a primitive bellero- phont, an advanced, relatively tall monoplacophoran, so to speak, that has undergone torsion. This is a view of the right side and is also drawn as being transparent. As is well known, torsion takes place in an early stage of the trochophore larva. This tor- sion occurs in such a way that the primitively posterior mantle cavity, with the anus lying between a pair of ctenidia, develops in a forward position over the head, just as if the shell with the visceral hump and contained organs had been twisted 180° in a counterclockwise direction relative to the head and foot. Note that the broad, low shell in Figure 2A has become an ele- vated one with the curvature that must have initiated coiling. The elevated shell may have provided for a long, slender neck such as that possessed by most living Anisopleura, which — com- bined with the reduction of the shell muscles to a single pair — would have permitted free movement of the shell relative to the foot. The anterior mantle cavity now permits the withdrawal of the body headfirst into the shell. Surely the traditionally slow- moving snail has become an active race horse as compared with the probably almost stationary monoplacophoran. But the problem of keeping the ctenidia free from fecal matter is now more difficult. Let us now consider briefly the series of solutions that gave those gastropods which possessed them com- petitive advantage. The first solution was the development of an anal emargination (a sinus or slit) for the egress of fouled ciliary currents without contaminating the ctenidia that lay on each side. This solution was used by the bellerophonts and pleurotomarians. Next — after the introduction of asymmetrical coiling — came the sup- pression of the right-hand ctenidium and associated organs. As a result the ciliary currents entered the mantle cavity from the left anterior side and departed by the right posterior immediately after passing over the anus, which had migrated to that position after the loss of the right ctenidium. Such a development oc- curred several times within the archaeogastropods and led to the STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 179 origin of other orders. Still further adaptations, also leading in part to the origin of new orders, were the gradual return of the anus to a posterior position in the adult of many opisthobranchs and — the last step — loss of the shell with almost full bilateral symmetry in the adult. A final adjustment not discussed here is the adaptation for life on the land with conversion of the mantle cavity into a lung and the loss of ctenidia in the pulmonates. In summary, this classification shows that while the living Iso- pleura have undergone very little change since late Cambrian time, the Anisopleura have radiated outward to become one of the most successful animal groups. Living snails have become extremely numerous and varied and have invaded all habitats and all environments from high mountains to the depths of the oceans. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 V. Mollusca: Cephalopoda By Bernhard Kummel Museum of Comparative Zoology, Cambridge, Mass. Few invertebrate classes possess (or are plagued with) the volume of literature and diversity of interpretation, or are so susceptible as media of philosophical speculation as the fossil cephalopods. A complete review of the literature and progress in the study of fossil cephalopods is beyond the scope of this paper. Therefore I will center attention on the geologic distribution of fossil cephalopods and current ideas on classification and evolu- tion. The extent and degree of cooperation towards a uniformity of nomenclature — both morphological and taxonomic — that has prevailed in the past few years among the European and Ameri- can authors responsible for the Cephalopoda for the Treatise has been truly inspiring. The results of this labor will place the study of these animals on a new threshold. It is a pleasure to acknowledge the data on Jurassic ammonoids furnished by W. J. Arkell ; the data on Cretaceous ammonoids furnished by C. W. Wright ; the data on Paleozoic ammonoids furnished by A. K. Miller; the data on the Dibranchiata furnished by L. Bairstow; and data on nautiloid genera by Curt Teichert. Of the three major groups of Cephalopoda, I will confine most of my remarks to the Ammonoidea. Recent tabulations indicate that there are approximately 3000 valid genera of fossil cephalopods with more than an additional thousand generic names which go into synonomy. Within the Nautiloidea there are now recognized approxi- mately 700 genera. Their evolutionary pattern is one of intense radiation in the Ordovician during the initial phases of their evolution, followed by a rapid and steady decline until the present day with our one remaining relic genus. The main pattern of distribution of nautiloid genera as illustrated on Figure 1 has been known for a long time. Morphologically the nautiloids are a highly diversified group but many of the orders that made their 182 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY appearance in the Ordovician became extinct by the Silurian and Devonian. During the latter half of the Paleozoic the evolutionary rate of the nautiloids was greatly diminished. A second surge occurred in the Upper Triassic which has a diversified nautiloid 275 vvwvww 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 NAUTILOID GENERA i i 1- 111 k::v :v::::^: [ u.-e 0RD. SIL DEV. MISS PENN PERM JURA. CRET. Pa E. QM.PLR Cenozoic Fig. 1. Bar graph showing number of nautiloid genera in each geologic period. (Data from C. Teichert, personal communication.) STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 183 fauna, but nothing compared to that of the early Paleozoic. The evolutionary pattern of the nautiloids in the Triassic is merely a culmination of trends begun back in the Carboniferous. The Per- mian was not a period of crisis for the nautiloids as it was for the ammonoids, but in the late Triassic (Rhaetian) there was almost complete extinction of the nautiloids. No "Rhaetian nautiloids are known, but a single specimen from the Carnian of Now Zealand offers strong evidence that an offshoot of the Syringonautilidae survived into the Lias (Spath, 1927, p. 23 ; Kummel. 1953) ; from this family all post-Triassic nautiloids are derived. Tho Jurassic to Recent history of the nautiloids duplicates the general pattern shown by the Paleozoic forms; that is, there is an initial experi- mental phase of extensive radiation from which arose several more stable stocks, none of which, however, was long lived. In the Cenozoic only three new genera appear, of which only Nautilus survives. Up until recently the classification of nautiloids has been com- pletely dominated by the scheme proposed by Hyatt in the 1900 edition of the Zittel-Eastman Textbook of Paleontology. Today there is no doubt in anyone 's mind that Hyatt 's scheme is unten- able. The rejection of Hyatt's classification is the rejection of a systematic arrangement based on a single character. The tenta- tive scheme recently proposed by Flower and Kummel (1950) summarizing existing thinking can unquestionably be improved but it has had the very desirable effect of crystallizing attention on the more critical problems. There is wide agreement among nautiloid specialists that the number of orders proposed by Flower and Kummel should be reduced, but no agreement as yet as to how this should be done. Of the 13 orders recognized by these authors the Ellesmeroceratida, Endoceratida, Michelino- ceratida, Ascoceratida, Discosorida, and Oncoceratida are gen- erally accepted as well defined groups and appear to need only minor revisions. The recognition of an Upper Triassic Bisiphytes (Family Nautilidae) which shows clear affinities to the Syringo- nautilidae changes the derivation of all post-Triassic nautiloids to the Triassic Syringonautilidae rather than the Paranautilidae as previously proposed (Kummel, 1953). This change necessi- tates the suppression of the Centroceratida and the joining of these families into the Nautilida. The Rutoceratida can most 184 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY probably also be included in the Nautilida but clear evidence of this relationship is not available (R. H. Flower, personal com- munication). Flower (1952) would derive the Rutoceratida from the Oncoceratida, The Liroceratidae, Ephippioceratidae, Para- nautilidae, Clydonautilidae, and Gonionautilidae, all formerly placed in the Nautilida, may now need to be placed in a new order. At first it would seem that they would be logically included in the Barrandeoceratida ; however, connecting forms from the Devon- ian have not been recognized as yet. The status of the Solenoch- ilida, Tarphyceratida, Barrandeoceratida, and Bassleroceratida is a source of disagreement among specialists. The crux of the difficulties and confusion in nautiloid systematics and evolution lies in the great host of early Paleozoic families. Taxonomic sta- bility of these forms will be approached as our knowledge and understanding of their complex morphology and modes of adap- tation is increased. Agreement on the systematics of early Paleo- zoic nautiloids is near but is not quite yet a reality. Because of this no phylogenetic chart is presented. It is significant to note that in the past few years very few new genera have been established in spite of the fact that several large new faunas have been described. It would be pleasant to think that possibly we are near the limit of the total number of genera in this order. This observation is encouraging in that maybe now more effort and thought can be concentrated on critical paleo- biological problems. Ever since William Smith's great discovery of the usefulness of fossils for correlation and geochronology, the ammonites have played a leading role as a stratigraphic tool. The pioneer work of d'Orbigny, Oppel, Quenstedt, Mojsisovics, Waagen, Neumayr and many others firmly established the ammonites as the leading actor in the drama of Mesozoic stratigraphy. Likewise the nature of the ammonoid shell and its manner of growth has lent itself beautifully to the pliable minds of evolutionary theorizers. Few people have had more influence on the study of fossil cephalopods than Alpheus Hyatt whose whole philosophy revolved around the recapitulation doctrine — a principle not accepted today by most cephalopod students. However, few text books of zoology or geology fail to illustrate some of Hyatt's cephalopods demon- strating recapitulation in action. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 185 The recent taxonomic house cleaning of ammonoid genera enables the construction of the bar graph (Figure 2) on the geologic distribution of ammonoid genera. There is a total of approximately 1800 genera of ammonoids; of these 172 are from 140 130 AMMONOID GENERA 199 Fig. 2. Bar graph showing number of ammonoid genera in each geologic period. (Data on Paleozoic ammonoids from A. K. Miller, that on Jurassic ammonoids from W. J. Arkell, and that on Cretaceous ammonoids from C. W. Wright, personal communications.) 186 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY the Paleozoic, 370 from the Triassic, 735 from the Jurassic and 493 from the Cretaceous. There are an additional 1000 generic names that fall into synonomy. The largest number of these involve Jurassic ammonites. The overall distribution of genera confirm commonly held views. In spite of specific instances of "monographic highs" and lack of record, it is the opinion of the Treatise ammonoid authors that the data reasonably reflect the history of this group of ani- mals. In fact, C. W. Wright (personal communication), who compiled the list of Cretaceous ammonites, found his data most illuminating as a commentary on the validity of the Cretaceous stages. The most unusual aspects of the evolutionary history of the ammonoids are the three periods of crisis ; during the first two such crises the group nearly became extinct and from but few surviving stocks even larger radiations developed until the final extinction of the group in the late Cretaceous. The two earlier periods of crisis are in the late Permian and in the late Triassic. Several important aspects in the history of the study of am- monoids have greatly influenced our interpretation of ammonoid phylogeny and taxonomy. It is indeed remarkable that not until the 1860 's did the splitting of the genus Ammonites really begin. The recognition of generic categories among the ammonoids de- veloped rapidly after that date. Since this early period there have been proposed many taxonomic schemes for the Ammonoidea. Most of these, however, were based on a single character; the suture, sculpture, siphuncle, and length of body chamber have all been used in various manners with little success. Unit systems of classification invariably ended up by bringing together hetero- chronous homeomorphs. Present interpretation of the relationships of Permian and Triassic ammonoids illustrates this point well. Figure 3 is a bar graph of total genera and new genera (dark stipple) of Permian and Triassic ammonoids. The average evolutionary tempo of Permian ammonoids is only slightly higher than that of the Carboniferous. The most striking feature is the presence of only seven genera in the Upper Permian. It should be kept in mind that the number of known fossiliferous Upper Permian deposits is indeed few. The ammonoid faunas of such classic areas as the STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 187 Salt Range, Djulfa, Timor, the Alps, Madagascar, and East Greenland are not abundant, The radiation of the ammonoids in the Lower Triassic is well demonstrated with its 128 genera. This radiation was not a sudden affair. Lowermost Triassic de- posits, that is, those of Otoceratan age, are as scarce as marine Upper Permian strata. The smaller bar graph of Figure 3 shows the total genera and new genera for the six ages of the Lower Triassic. Only one genus carries on into the Triassic from the Upper Permian. It is not until the middle Scythian that a really large diversified fauna is encountered. Figure 4 is a diagrammatic interpretation of the phylogenetic relationships of Permian and Triassic ammonoids. Of the two main groups of Paleozoic ammonoids, the goniatitids are repre- sented by seven families in the Permian and the prolecanitids by five families. It has long been thought that many Middle and Upper Triassic ammonoids had their ancestry among these goniatitid families. However, the extensive researches of L. F. Spath over the past 30 years have demonstrated in a convincing manner the improbability of any of these goniatitid families be- ing ancestral to any Triassic stocks. There are numerous cases of homeomorphy between many of these goniatitid genera and Mid- dle and Upper Triassic genera. Permian ammonoids of the prolecanitid stock are included in two superfamilies. One of these — the Pronoritaceae — continues into the Triassic with a very modest development but did not give rise to any new stock. The ancestry of Triassic ammonoids is to be found in the Xenodiscaceae. The most common and abundant lowermost Triassic ammonoids are the ophiceraticls which show a wide range of variability in conch form. From this extremely variable ophiceratid stock arose directly or indirectly more than a dozen families of Lower Trias- sic ammonoids. This radiation is a continuation of that started at the base of the Triassic with the ophiceratids, and represents the height of the experimental phase in the eruptive comeback of the ammonoids after their near extinction in the late Permian. Most of the genera and families of this eruptive phase are short- lived but from a few of them arose the eight long-lived super- families that include most of the Middle and Upper Triassic ammonoids. Thus with the Anisian begins the stabilized phase in the evolutionary history of Triassic ammonoids. 188 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The great decline and near extinction of the ammonoids in the Rhaetian has been known for a long time. From a peak of 140 genera in the Carnian there are left only a few species of 5 genera in the Rhaetian. There is a sharp decline in evolutionary development already in the Norian. O (/) CO < IT q: Q- _L c o o CO Prohungaritan !.*::;::#*:':::'::C;::;: m i Columbitan 1 ■i .1 Owenitan mm i Flemingitan , 1 Gyronitan i 20 Otoceratan —iiio 30 40 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Fig. 3. Bar graph showing total number of ammonoid genera and new genera (dark stipple) in each stage of the Permian and Triassic. A single stock carries through into the Lias to give rise to the great host of Jurassic and Cretaceous ammonoids. The evolu- tionary patterns of the ammonoids in these two periods is quite unlike that in the Paleozoic or Triassic. Two persisting, slowly evolving, root-stocks - - the Lytoceratina and the Phylloceratina STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 189 — are the evolutionary reservoirs from which the great host of Jurassic and Cretaceous ammonoids are derived. The Lytocera- tina and Phylloceratina are characteristic of the ancient Tethys ; they are the leiostracous or smooth shelled forms. The trachyos- tracous or ornamented genera are most abundant in the epicon- Fig. 4. Diagram showing inferred phylogeny and geologic distribution of Permian and Triassic ammonoids. tinental seas adjoining the Tethys. Some 30 years ago Salfeld proposed his theory of Iterative Evolution in which he inter- preted the innumerable throngs of the Ammonitina as arising through repeated radiations of offshoots from the conservative 190 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY stocks. Deciphering the countless threads connecting the Am- monitina and these conservative stocks is still evasive. More Pig. 5. Diagram showing inferred phylogeny and geologic distribution of Mesozoic ammonoids. (Adapted from Arkell 1950, Wright 1952, and Kummel 1952). STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 191 progress has been made with Cretaceous ammonites than with those of the Jurassic. When and if the various Jurassic and Cretaceous superfamilies can be more definitely tied to one or the other of the conservative root-stocks there will be need for modification of the suborder Ammonitina. The Dibranchiata cephalopods have not received as wide atten- tion as the nautiloids and ammonoids. Mr. Leslie Bairstow kindly furnished tentative and approximate data on the time 141 100 90 80 70 60 50 40 30 20 10 DIBRANCHIATE GENERA LM.U PERM. L M.I U. RIAS. L. M. U. LIRA. U. CRET. PdeoEoOM. PJIR CENOZOIC Fig. 6. Bar graph showing number of dibranch genera in each geologic period. (Data from L. Bairstow, personal communication.) distribution of genera of Dibranchiata. The most striking feature of the bar diagram of Figure 6 is the great number of Recent genera ; of the approximately 250 genera of Dibranchiata, 138 are known only from the Recent. The great diversity of Recent fauna, and the fact that many of the genera are very different from the nearest fossil forms known, imply the previous existence of many genera that are not actually known as fossils (L. Bair- stow, personal communication). The incompleteness of the fossil 192 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY record of the dibranchs does not make them a suitable group for the study of evolutionary rates. The present status of our knowledge of the fossil Cephalopoda presents an encouraging framework to direct and guide future research. The great success to which the cephalopods, especially ammonoids, have been used in stratigraphy has resulted in the major emphasis being placed on them as "tools" of stratigraphy. It is becoming more and more clear that a greater understanding of the animal and its ecology and adaptive range will add greatly to the value of the Cephalopoda as "tools" of stratigraphy and as "tools" for study of evolutionary processes. REFERENCES Arkell, W. J. 1950. A classification of the Jurassic ammonites. Jour. Paleontology, vol. 24, pp. 354-364. Flower, R. H. 1952. The ontogeny of Centroceras, with remarks on the phylogeny of the Centroceratidae. Jour. Paleontology, vol. 26, pp. 519-528. Flower, R. H., and B. Kummel 1950. A classification of the nautiloids. Jour. Paleontology, vol. 24, pp. 604-616. Kummel, B. 1952. A classification of the Triassic ammonoids. Jour. Paleontology, vol. 26, pp. 847-853. 1953. The ancestry of the family Nautilidae. Breviora, No. 21, pp. 1-8. Spath, L. F. 1927. Revision of the Jurassic cephalopod fauna of Kachh (Cutch). Mem. India Geol. Survey (Palaeontologia Indica), n. ser., vol. 9, mem. 2, pp. 1-84, pis. 1-7. Wright, C. W. 1952. A classification of the Cretaceous ammonites. Jour. Paleontology, vol. 26, pp. 213-222. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 VI. Arthropoda: Trilobita By Harry B. Whittington Museum of Comparative Zoology, Cambridge, Masa. The following notes express my own views and some that W. R. Evitt and I have developed jointly, and are not necessarily those of others working on Trilobita for the Treatise of Invertebrate Paleontology. I acknowledge with thanks the helpful discussions I have had with many friends, especially Drs. A. R. Palmer, C. Poulsen, F. Rasetti, and C. J. Stubblefield. Many articles have been written on the classification of trilo- bites, but there is still no satisfactory arrangement. Major diffi- culties are : (1) About half of the 600 or so Cambrian genera are incompletely known or based on poor material, and many have never been placed in any family. In particular the Upper Cambrian forms grade into each other and are hard to separate. Specialist or not, one welters in a morass of names and concepts. (2) Post- Cambrian trilobites seem to fall readily into families and higher groups, but there is a great gap between them and their late Cambrian predecessors. This gap seems to be real, and not merely because many students have worked on either Cambrian or later trilobites, but rarely both. (3) There is a lack of knowledge of the entire, and especially the ventral exo- skeleton, and of the ontogeny of representatives of many families. Attempts were made to base major divisions in trilobite classi- fication on the number of thoracic segments or the size and segmentation of the pygidium. The latter has been used recently by Hupe (1953). Beecher's (1897) classification, using the cephalic sutures, depended on an interpretation that accepted Haeckel's theory of recapitulation. I know of no example in trilobites of ontogeny recapitulating phylogeny, i.e., I know of no ontogenetic stage that resembles a geologically older adult. The use of part of a single morphological character as a basis for an ordinal classification seems unwise (cf. Stubblefield 1936, p. 432, etc.) and in this case has been shown to be unworkable. 194 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY It is to be hoped that this classification will no longer be used or taught in the classroom. In 1936 Stubblefield suggested that many proparian trilobites might be permanently neotenous forms. Stormer (1942) took up this suggestion vigorously, and considered that three orders of trilobites (modifications of those of Beecher) were derived from the supposedly primitive olenellids by arrested development (a partial neoteny) of the anterior cephalic segments. In addition to rejecting Beecher 's orders, I consider that certain of Stormer 's premises are possibly erroneous, and his classification is no more usable than Beecher 's. Hupe (1950) has claimed that partial neoteny is general in trilobites. His evidence is drawn from the study of representatives of two groups, proparia and opis- thoparia, each of which is here regarded as being comprised of genetically unrelated genera. Broad morphological trends of change in the cephalon during the Palaeozoic are revealed, but scarcely seem to me to admit of the conclusions arrived at. We do not yet know a single line of evolution with data on the on- togenies of successive genera, and are in no position to assess the importance of neoteny as a process in trilobite evolution. I sus- pect that adaptation to particular ecological niches (depth, bot- tom conditions, food supply, etc.) is a far more important factor in trilobite evolution, but to discern its operation is difficult. If we can avoid being mesmerized by the importance that has so far been attached to cephalic sutures in classification, a dif- ferent approach is possible, i.e., to consider all the exoskeletal characters (including ventral, and especially axial characters, as Stubblefield [1936] urged), ontogeny (as Stormer [1942] did), and the distribution of genera in space and time. Henningsmoen (1951) has recently made suggestions along these lines and they appear promising. The Cambrian superfamilies shown in Figure 1 are those used recently by Hupe (1953). The superfamilies outside the stippled area are agreed upon by most students. The ptycho- parioid and allied superfamilies, included within the more heavily stippled area, seem to constitute a major natural group of trilo- bites, but the subdivisions of this group, adopted for convenience here, are not generally agreed upon (cf. Rasetti, 1951, pp. 198- 202). The post-Cambrian superfamilies of Figure 1 are the STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 195 Fig. 1. Eange in time of superfamilies of trilobites. Each black area varies in width in proportion to the number of genera recorded at any one time. Total number of genera in each superfamily is given. Data on Cambrian trilobites largely from Hupe (1953). Tremadoc is included in the Ordovician, and Lower, Middle, and Upper divisions of this period and of the Cambrian are indicated. Proportions of length of periods after Newell (1952). Correlations of strata and ranges of genera necessarily generalized. The Ptychoparioidae and allied superfamilies of the Cambrian are included in the more heavily stippled area, and post-Cambrian superfamilies possibly derived from them in the lighter, diagonally stippled area. 196 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY groups of families indicated by the shading in Figure 2. In making these superfamilies, besides using axial and ventral char- acters of adult exoskeletons, W. R. Evitt and I here suggest, as a result of our unpublished studies of protaspids of Ordovician trilobites, that one might group together those families in which early ontogenetic stages have the same general characters, al- though the adults are distinctly different. This leads, for example, to grouping together as the Cheiruroidae the calymenids, homalo- notids, phacopids, encrinurids, cheirurids, and pliomerids. Per- haps more striking is the Asaphoidae, here regarded as including the asaphids and remopleuridids, and possibly nileids, telephinids and cyclopygids. The lichids and odontopleurids — distinctive groups, the origin and relationships of each of which is uncertain — are shown in Figure 1, though they are here regarded as of family, and not superfamily, rank. The trilobites were marine animals and apparently dominated Cambrian seas. Olenelloids appear first, and are confined to the Lower Cambrian. Other groups occur in the later Lower Cam- brian, and some extend only into the Middle Cambrian. These are replaced by the more varied and numerous Upper Cambrian forms, derivatives of the Middle Cambrian Ptychoparioids. "Spindle", "burst", and "truncated burst" patterns, shown by Cooper and Williams (1952) to characterize brachiopod evo- lution, are equally evident in the trilobites. A lengthy pre- Cambrian history evidently precedes the known history, and the seemingly abrupt appearance of trilobites is apparently owing to their becoming able to form a mineralized exoskeleton. This Cambrian picture is very generalized, owing to difficulties of correlation, and the extinctions may not be as sudden as they appear to be. There is a marked gap after the end of Cambrian time, only four Cambrian superfamilies ranging into the Ordovician in greatly reduced numbers. Following this great extinction, a number of new superfamilies arise suddenly in the Lower Ordo- vician, but no new ones arise after this time. Only one Ordovician superfamily, here called the Asaphoidae, is known to arise in the Upper Cambrian. The picture in the Ordovician is one of a last evolutionary burst before new classes, e.g., of Mollusca and Vertebrata, arise to crowd out the trilobites, and the post-Ordo- STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 197 CAMBtoRDOVICIAN DEVON. 1 GARB. PBM HYSTRICURIDAEl PLETHOPELTIMci CERATO- PYGIDAE' KOMASPIDAE, LEIOST- EGIIDAE ENDYMIONIIDAE SCUTELLIDAE LICKIOAE ODONTOPLEURIDi PLIOMERIDAE — SHUMARDIIDA Fig. 2. Range in time of families of post-Cambrian trilobites. Con- structed in the same manner as Figure 1. Tremadoe, Lower, Middle and Upper subdivisions of the Ordovician are indicated separately. The shading links those families that have been combined into a superfamily in Figure 1. Leiostegiidae is used in the sense of Whittington, 1953, and Lecanopygidac Lochman, 1953, is not shown. 198 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY vician history is one of lingering survival. Figure 2, composed in the same manner as Figure 1, indicates the range in time of post-Cambrian families. The gap at the base of the Ordovician is more clearly shown, and only in four cases do we, know possible Cambrian ancestors of dominantly Ordo- vician families. Striking is the asaphid "burst", probably occur- ring simultaneously in unrelated groups in Baltoscandia, America, and Asia. "Bursts" of this type may have taken place in other trilobite families. The great "burst" of new proparian forms begins in early Canadian (i.e. Tremadoc) time and ex- tends into the Champlainian, and we have as yet no notion of the ancestral stocks. Equally puzzling are the Lichidae and Odontopleuridae, arising suddenly, "ready made", apparently unrelated to each other or to other families. The possibility that groups hitherto soft-shelled were acquiring the power to mineral- ize the exoskeleton cannot be overlooked. The general decline of trilobites from the mid-Ordovician onwards is evident, though inadequate knowledge of Silurian trilobites may falsify the pic- ture. The diversification of certain groups in the Devonian is real, but perhaps over-emphasized as a result of the intensive systematic work of Drs. R. and E. Richter. Undoubtedly, other "bursts" are in part "monographic", and new ones will inevi- tably be added. It seems reasonable to suggest that a natural classification should attempt to make orders from groups of related super- families. If the superfamilies as shown in Figure 1 are accepted tentatively, the chief stumbling block in the way of the erection of orders is the lack of known connections between the super- families. At least four orders might be suggested : one to include eodiscoids and agnostoids, which are perhaps more like each other than either resembles any other trilobite group ; a second to include olenelloids and redlichioids ; a third for the corynexo- choids ; and the fourth including the group of ptychoparioids and allied superfamilies included in the heavily stippled area of Figure 1. A lighter, diagonally stippled area includes the post- Cambrian superfamilies which may be derived from ptycho- parioids and their allies, and which consequently may be regarded as belonging to the same order. A separate order may be neces- sary for the cheiruroids, the post-Cambrian proparian trilobites, STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 199 the origins of which are at present obscure. Omitted from these groups are lichids, odontopleurids, and various isolated genera and small families of uncertain affinities. It does not seem ap- propriate to formally propose such orders at the present time, my intention being merely to outline what seems to be a promis- ing line of approach. This approach emphasizes our lack of knowledge in critical areas, and it is this lack which, in my opinion, precludes any satisfactory classification of trilobites at this time. REFERENCES Beechbr, C. E. 1897. Outline of a natural classification of the trilobites. Am. Jour. Sci., ser. 4, vol. 3, pp. 89-106, 181-207, pi. 3. Cooper, G. A., and A. Williams 1952. Significance of the stratigraphic distribution of brachiopods. Jour. Paleont., vol. 26, pp. 326-337. Henningsmoen, G. 1951. Remarks on the classification of trilobites. Norsk, geol. tidskr., vol. 29, pp. 174-217. Hupe, P. 1950. Etude statistique de 1 'evolution du cephalon chez les trilobites Proparia et Opistoparia. Bull. Soc. Geol. France, ser. 5, vol. 20, pp. 9-24. 1953. Classe des Trilobites, in Traits de Paleontologie, vol. 3, pp. 44-246, ed. J. Piveteau. Paris. LOCHMAN, C. 1953. Analysis and discussion of nine Cambrian trilobite families. Jour. Paleont., vol. 27, pp. 889-896. Newell, N. D. 1952. Periodicity in invertebrate evolution. Jour. Paleont., vol. 26, pp. 371-385. Rasetti, F. 1951. Middle Cambrian stratigraphy and faunas of the Canadian Rocky Mountains. Smith. Misc. Coll., vol. 116, no. 5, pp. 1-277, pis. 1-34. 200 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY ST0RMER, L. 1942. Studies on trilobite morphology. Part II. Norsk geol. tidskr., vol. 21, pp. 49-164, pis. 1, 2. Stubblbfibld, C. J. 1936. Cephalic sutures and their bearing on current classifications of trilobites. Biol. Rev., vol. 11, pp. 407-440. Whittington, H. B. 1953. North American Bathyuridae and Leiostegiidae. Jour. Paleont., vol. 27, pp. 647-678, pis. 65-69. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 VII. frraptolithina1 By 0. M. B. Bulman Sedgwick Museum, Cambridge, England Graptolites are extinct colonial organisms, and graptolite phylogeny is expressed in the complicated relations of a succes- sion of colonies. Not only has every individual theca a complete ontogeny (accurately recorded in its growth-lines) but each adult represents a stage in the development of the astogenic unit. Despite the stratigraphical importance of the group, precise knowledge of graptolite phylogeny is disappointingly slight. The general succession of graptolite faunas is, however, well-estab- lished, and it reveals a fairly steady process of simplification in the colony and reduction in number of constituent individuals, with phases of elaboration in the skeleton of individual zooids. Only recently have we begun to get sufficiently exact details concerning structure and mode of growth to afford a basis for investigation of the complex problems presented. These notes are confined to the order Graptoloidea, or true graptolites, but to understand their structure some preliminary reference to the ancestral order Dendroidea is necessary. Here the three types of thecae recognized by Wiman have been shown by Kozlowski to be related to an internal stolon system and probably to have been secreted by only two kinds of individual (plausibly interpreted as male and female). Thus the term "stolotheca" ( Wiman 's "budding-individual") really desig- nates no more than the proximal, immature portion of the auto- theca, since continuity of growth-lines demonstrates that each stolotheca is part of a unit which includes the succeeding auto- theca (Fig. 1 A, B) . Consideration of the growth-lines and the stolon system also indicates that there was not (as there is in Rhabdopleura, Fig. 1 E) a terminal "leading bud" behind which successive individuals are proliferated; but that each i I am grateful to Professor L. Sterrner of the University of Oslo for permis- sion to publish this material which is based on a lecture given in Oslo in December 1951. 202 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY theca (a6) gymn (si) pect(st) 11"// / D Pig. 1. Comparison of organization and method of budding in a dendroid (A,B), a graptoloid (C,D) and Btobdopleura (E). al, a2 etc., autothecae; b2, b3 etc., bithecae; si, s2 etc., stolothecae; st, stolon; gymn., gymnocaulus; pect., pectocaulus. In A, B and D, one complete unit (including stolotheca and protheca) is shaded. Fig. A represents the growing end of a branch, and a3, b3 and sS are all incompletely developed. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 203 autotheea in turn has represented the terminal bud of its branch, and at a certain stage in its development there are produced from back on its stolon two buds, male bitheca and female auto- theea (Fig. 1 A, B). The budding processes in Bhab do pleura, on the one hand, and the graptolites on the other, exactly parallel the monopodial budding with terminal growing points, and the sympodial budding, described (L. Hyman) in calyptoblastean hydroids. In the Graptoloidea, branch organisation is considerably sim- plified by the loss of bithecae (the autothecal zooid, on Koz- lowski's interpretation, becoming hermaphrodite), but the stolon Fig. 2. Mode of branching in a dendroid (D. flabelliforme) . A, branching division with production of two stolothecae (s4 and s'4) in place of stolo- theca and bitheca. B, approximate zones of branching in the rhabdosome of D. flabelliforme. xl. system may be considered to persist to a limited extent in an unchitinised condition. The occurrence of a small "unconform- ity" between the growth-lines of adjacent thecae (Fig. 10) even in early stages of colonial development, and the dual construction of each interthecal septum, indicates that each theca buds from the preceding theca as a slightly modified inheritance of the dendroid method of budding (Fig. ID). Bifurcation in a dendroid branch results from the formation of two stolothecae in place of the normal stolotheca and bitheca 204 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (Fig. 2A). Branching is often very regular; more so in Dit- ty onema than in most dendroids and in D. flabelliforme than in other species of that genus (Fig. 2J3). In the Graptoloidea, branching is generally strikingly regular, but since only one type of individual is present in the colony, it clearly results from the formation of two buds in place of one under some periodic stimulus. In relation to the Dendroidea, there is a break here in the evolutionary sequence for which we still lack the material to investigate. The differences among branched dichograptid colonies are largely due to the spacing and timing of these double buds. There seems to be a general tendency towards reduction in the number of branches in a colony, though few lineages are well established and the recent work of Australian paleontologists suggests that the phyletic relations of the Dichograptidae are much more involved than has generally been recognized. Cer- tainly the two-stiped Didymograptus is the successful type, judged by its increasing preponderance in the upward succes- sion; but branch reduction may have taken place either proxi- mally or distally — i.e., Didymograptus may be either neotenic or gerontomorphic, if one may use such terms of colonial proc- esses. Reduction in number of branches implies a reduction in total number of individuals in a colony, which over the whole series may be extreme. A large rhabdosome of Dictyonema flabelliforme may have comprised the astonishing total of not less than 35,000 individuals (autothecal and bithecal) ; a large dichograptid might have 3000 individuals (all of one kind) ; most leptograptids, dicellograptids and diplograptids have from 100 to 200 individ- uals (usually nearer the lower limit) ; while the Silurian Mono- graptus averages about 50 and the late M. leintwardinensis consists of only 10 to 17 individuals in a colony. With the establishment of the simplified autothecal and pauci- ramous rhabdosome of the typical graptoloid, several structural features make their appearance — some rhabdosomal and some thecal. Rhabdosomal changes, apart from stipe-reduction, are most conspicuously concerned with orientation relative to the sicula (which itself remains strikingly constant) and the nema suspen- STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 205 Oncogr Cryptogr. 22 2' Fig. 3. Progressive changes in the proximal end of the rhabdosome in dichograptids, Ieptograptids and dicellograptids, showing progressive delay in formation of the ' ' double bud ' \ 206 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY sion; the rhabdosome passes from pendent through horizontal to reclined and eventually scandent. This change in direction of growth is an intermittently expressed tendency which affects all graptolites ever since D. flabelliforme reversed the normal orientation of a Dictyonema rhabdosome to begin the story of graptoloid evolution. It proceeds side by side with a sequence of changes in the development of the proximal end of the rhab- Fig. 4. Monograptus argenteus (Nicholson) x2 showing change in thecal character from hooked (proximal end) to straight (distal end) ; growth- lines on enlargements xlO of selected thecae are partly conjectural, and are inserted to illustrate the form and development of the thecae. dosome which may be expressed as a progressive delay in the formation of the two buds from which any bilateral rhabdosome develops (didymograptid, leptograptid, dicellograptid or diplo- graptid). A temporarily stable condition is reached when the first four thecae of the rhabdosome alternate in origin and there are three crossing eanals (as is very usual in dicellograptids and STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 207 diplograptids), the two buds being carried on th21 (Fig. 3). Further evolution to aseptate and uniserial rhabdosomes is men- tioned later. Thecal changes include many types of elaboration and what may be termed "thecal differentiation." The simplest expression of the latter is a distal increase in size, the thecae remaining constant in shape. Even the Dichograptidae show such a distal increase, which contrasts with the uniformity in size of dendroid thecae throughout a rhabdosome. In its extreme expression, it is coupled with thecal elaboration, and successive thecae may undergo progressive change in size and form, as in the many bi-form monograptids (Fig. 4). The available evidence suggests that new types spread gradually along the rhabdosome with time. The linear propagation of buds in one of the simplified Graptoloidea makes such changes more regular than comparable changes in, for example, a compound coral ; and it has been com- pared with the introduction of new characters in a metamerically segmented animal. Less commonly does a new character seem to operate throughout the life of a colony at its first appearance ; and occasionally the distal thecae of a rhabdosome may show new features which did not originate proximally, though such features seem rarely conspicuous. In seeking to attach significance to these thecal changes, it may be added that so far as we are aware, they reflect no dif- ference in habitat ; all graptolites are believed to be epiplanktonic, and the remains of many different types occur side-by-side in the rocks. As regards their transmission, it may be noted that whatever modifications the first and subsequent thecae may show, the sicula (the first individual of the colony) remains throughout the Order almost as constant in form as it is in orientation. Every subsequent zooid is produced by asexual budding, the colony being subject to a common genetic control. In stipe-reduction, change in direction of growth (from pend- ent to scandent) and the various thecal changes (especially among monograptids), the graptolites furnish numerous instances of "programme evolution", for parallel changes undoubtedly occur independently in many different lines of descent, though few detailed phylogenies have as yet been established by which this assertion can be illustrated. 208 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY There seem to have been three major events in the geological history of the Graptoloidea : I. Derivation of the Graptoloidea from the Dendroidea in late Tremadocian times. II. Origin of the biserial scandent rhabdosome early in the Ordovician. III. Origin of the uniserial scandent rhabdosome slightly but appreciably above the base of the Silurian. I. Derivation of the Graptoloidea. This involves loss of dis- sepiments and bithecae, and was preceded by a change in mode of life, the benthonic rooted habit giving place to a nema-attach- ment to floating weeds. Such a change had in fact occurred in Dictyonema flaoelli forme and that varied collection of its im- mediate descendants the Anisograptidae. Of these, the best known are Clonograptus (Fig. 5 A, B) and Bryograptus (Fig. 5 C, D), but Anisograptus (Fig. 5 F), Adelograptus (Fig. 5 E) and Staurograptus (Fig. 5 H) are also important. They vary in attitude of branches from pendent to reclined, and in number of primary stipes from two to four; but all are united in the pos- session of a nema attachment along with typical dendroid branch structure, and they differ from Dictyonema flabelliforme chiefly in loss of dissepiments. The change from anisograptids to true dichograptids requires the loss of bithecae and is usually accompanied by some reduction in number of branches. Stipe-reduction may come first — Aniso- graptus itself in one of its varieties has only four branches and leads to the triradiate, unbranched Triograptus (Fig. 5 0) while apparently retaining the dendroid branch structure. Or bithecae may be lost while the number of branches shows little change and may even increase — for there is much to suggest that the Ordovician Clonograptus and Bryograptus species have a simpli- fied graptoloid branch while retaining the rhabdosome form of their Tremadoc ancestors. The transition from dendroids to graptoloids is clearly a gradual one (in contrast to the other two episodes) and the Graptoloidea may well be polyphyletic, though all seem ultimately traceable (through various anisograptids) to Dictyonema flaoelli forme. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 209 By the middle of the Lower Ordovician the dominant grapto- loid type had become the two-stiped, horizontal or reclined didymograptid. Then progressive delay in the origin of the two buds from which the two separate stipes develop (Fig. 3) leads to the proximal end characteristic of most leptograptids and all dieellograptids of the Middle Ordovician, wherein the first four thecae of the rhabdosome alternate and there are three crossing- canals. Further, these leptograptids and dieellograptids show the substitution of a new type of theca — the sigmoid leptograp- Fig. 5. Representative members of the family Anisograptidae. A, B, Clonograptus (flexilis and tenellus) ; C, D, Bryograptus (Tcjerulfi and patens) ; E, Adelograptus (hunnebergensis) ; F, Anisograptus (matanensis) ; G, Triograptus (canadensis); H, Staurograptus (dichotomus) . x% approx. tid type — for the simple straight dichograptid type. In Dicello- graptus and Dicranograptus this may become further modified in a distinctive manner, the distal end becoming isolated and introverted (Fig. 6 B, C). With the development of the leptograptids, branched colonies 210 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY reappear (Pleurograptus, Nemagraptus and Amphigraptus) . No satisfactory explanation of this has yet been given. Since the details of branch formation are still unknown, parallel descent from multiramous dichograptid ancestors cannot be altogether excluded, nor even some kind of " pseudobranching " analogous to the monograptid cladia; but an atavistic interpretation is the one most usually accepted. D E Fig. 6. Selected types of graptoloid thecae. A, leptograptid ; B, dicello- graptid (simple) ; C, dicellograptid and dicranograptid (more complex) ; D, E, extreme types of monograptid (triangulate and lobate). Growth-lines, mainly conjectural, inserted for greater clarity. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 211 II. Origin of the biserial scandent rhabdosome. The Dicello- graptus — Dicranograptus series had reached a condition as re- gards their proximal end (in the alternating origin of the first four thecae) which is practically identical with that of the diplo- graptids. Moreover, Dicranograptus (with its Y-shaped rhab- dosome) has long been suggested as an intermediate between Dicellograptus and the biserial graptolites. Nevertheless, the diplograptids are the earlier group, and if Dicranograptus has any significance in this story it can only be as a contributory fac- tor in a polyphyletic ancestry. The earliest diplograptid known is Glyptograptus dentatus from the Extensus Zone in Britain ; by Upper Arenig times it is found in Scandinavia, and by the beginning of the Llanvirn (Bifidus Zone) it had reached Australia and America. Against a background of Lower Ordovician graptolites, its appearance is quite startling ; and its structure is so novel that it must surely represent an unusually large but successful mutation. It pos- sesses certain distinctive features in its proximal end which are gradually lost in its descendants (but curiously enough reappear in the only Dicranograptus of which we have as yet any detailed knowledge). The significance of these features (the --'-shaped thl2 with its initially upward growth, and the long downward- growing initial part of th21) is unknown, but somewhat similar features are recognizable in the nearly contemporaneous Lasio- graptus hystrix and to a less extent in early Climacograptus scharenbergi. In fact, the bulk of the true diplograptids may well have a monophyletic origin in G. dentatus (using the term diplograptid to include Diplograptus sensu lato, Climacograptus, and some as yet undetermined portion of the lasiograptids). So long as the rhabdosome consisted of two separate uniserial stipes, it was not mechanically possible for the formation of the two buds representing the bases of the two stipes to be much delayed — in fact, the presence of three crossing-canals and the alternating origin of the first four thecae is about as far as this trend can proceed. But there is no such necessary limitation in the scandent biserial rhabdosome, where the septum separating the two stipes may arise at any point along the rhabdosome (Fig. 1 C,D). Its earliest appearance is between th22 and th22 (Fig. IB) and the septum is then said to be "complete"; but it is 212 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Cryptograptus ?Glossograptus Generalized Diplogr- (complete septum") Glyptogr. dentatus Oncograptus PCardiograptus ? Skiagraptus Fig. 7. Progressive changes in the proximal end of biserial scandent graptolites. A-D, diplograptids illustrating further delay in production of the "double bud"; E, F, proximal end of other scandent biserial graptolites. Precise relations of the thecae in retiolitids remain unknown. STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 213 progressively deferred as more and more of the proximal thecae alternate in origin, and may never occur at all if the thecae alternate throughout the rhabdosome. It may be noted that the thecae in diplograptids, though vari- ously modified, never show the extremes that we encounter among the Dicellograptus — Dicranograptus series; in particular, no diplograptid so far as I am aware shows any trace of apertural isolation. Thecae remain for the most part fairly generalised and the angularly sigmoidal thecae of Ample rograpt us and Climacograptus represent the most extreme modifications. It is from the more conservative stock (which persists into the Silurian system) that the great wealth of monograptids must be derived. In addition to the diplograptids proper, there are numerous other scandent biserial forms, such as Cardiograptus, Skiagrap- tus, Cryptograptus, Glossograptus, Trigonograptus, Gymnograp- tus and the retiolitids, which are all very imperfectly understood as yet. Many of these appear to have been side lines (some even from the dichograptids) incidental to the main evolutionary story ; the retiolitids, however, are probably derived from diplo- graptids (perhaps more than once), though the chitinous skeleton is so extensively reduced, often to a few structural rods and girders, that it is practically impossible to interpret their proxi- mal end in terms of zooidal succession and development. III. Origin of the uniserial scandent rhabdosome . Failure to produce the pair of buds, or rather, the production of a single bud in place of the pair, cannot by itself convert a diplograptid rhabdosome into Monograptus, for even in the simplest and most generalised diplograptid the first four thecae of the rhabdosome are alternating in origin (Fig. 8A). This is beautifully illus- trated by a form discovered by Dr. Strachan but not yet de- scribed, which comprises a uniserial rhabdosome yet possesses thl2 in its normal diplograptid relation. The change can only effectively come after these proximal thecae have substituted a linear serial arrangement for the alternating proximal arrange- ment of their ancestor, either by reduction or loss of thl2 and any later "second-order" thecae (Fig. 8C), or perhaps more probably by a "merging" of the thecae of the two series (Fig. 8D) as postulated by Wiman. Such merging may be combined with the pronounced tendency to an upward direction of growth 214 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY which affects the biserial Petalograptus and Cephalograptus. Starting from a diplograptid with the most generalised proxi- mal end (a complete septum) it would only be necessary to lose one theca, or to get thl2 into series between thl1 and th2x, for subsequent failure of the "double bud" to result immediately in a monograptid (Fig. 8C, D). From such an ancestor, there- 2 ,'\ 3"< Generalized Diplogr. type Cephalogr. type reduction and loss of thl2 "merging' of ih.\f Fig. 8. Diagram illustrating the problem of monograptid origin. fore, the origin of Monograptus is essentially a question of the disposal of thl2; rearrangement (or loss) of numerous proximal thecae in a form with a delayed septum is more difficult to under- stand. Yet the well-defined species groups of Dimorphograptus suggest that, if this was their origin, the change must have STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 215 occurred independently in several lines of descent. It may there- fore be thought that Dimorphograptus (with a uniserial portion often of considerable length) is less likely to be an intermediate stage than an atavistic form (as Wiman suggested) or a short- lived side line. But it must be admitted that Dimorphograptus is essentially pre-monograptid in its range, and its significance in relation to Diplograptus and Monograptus remains uncertain. With the origin of the monograptids only a little more than halfway through the evolutionary history of the Graptoloidea, the major phases of structural evolution of the graptolite rhab- dosome are passed, and for the next thirty million years the uniserial scandent rhabdosome remains the dominant graptolite form. Thecal elaboration begins anew1 and may lead to extremes exceeding anything known from the Ordovician. The sequence of thecal elaboration and the structure of ' ' pseudobranched " rhabdosomes of genera like Cyrtograptus present many unsolved problems ; but the basic form of the colony had in its very sim- plicity attained a complete stability persisting to the extinction of the Graptoloidea. i In these monograptid elaborations there is an exaggerated development of the dorsal margin of the theca leading to extroverted thecal forms, in contrast to the dominant development of the ventral lip which produces the introverted Ordovi- cian type of elaboration. Compare Fig. 6 B,C, with 6 D,E. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 VIII. On Development, Evolution, and Terminology of Ammonoid Suture Line By 0. H. Schindewolp University of Tubingen, Germany INTRODUCTION In comparative anatomy it is a matter of course to use a termi- nology that is based on the homology of organs. In osteology of vertebrates, for instance, elements of the same origin and de- velopment, but often of quite unlike shape receive the same designations. Elements of different origin, however, i.e., merely analogous organs, are sharply distinguished in terminology. These principles are not yet sufficiently established in the terminology of ammonoid suture line. Here the denominations mostly have been based on the mature sutures, without regard to, or at least without fully utilizing their ontogenetic de- velopment. There are, however, several different modes in the development of suture lines, so that it is impossible to base the designations of the lobes merely on their position and shape in the adult suture. This of course would be much simpler, but it would veil the existing differences and thus the natural relations of the forms exhibiting these different suture lines. This method, therefore, is not feasible. The first authors realizing these failings and initiating a ra- tional terminology of suture line based' on its individual develop- ment were F. Noetling (1905, 1906) and R. Wedekind (1913 a,b ; 1916; 1918). The attempt of Noetling, though ingenious in itself, was a failure, since it involved two wrong suppositions : (1) The foundation, on which the ontogenetic development and phylogenetic evolution of suture line rise, is the primary suture (i.e., the first genuine suture line following the prosuture), but not the prosuture, a larval organ with widely different features. (2) A genetic terminology of suture line cannot be based, as Noetling proposed, on highly differentiated ammonites; it has to 218 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY start from the basic structures of the earliest ammonoids, the undifferentiated goniatites. Wedekind has wisely avoided these two mistakes and has estab- lished a terminology unsurpassed until now in consistency and simplicity. I have further elaborated the principles introduced by him, and in this paper I shall discuss in a summary form some general questions concerning the suture line and certain related topics. Agoniatites Mimagoniat. Oyroceratites Lobobactrites Bactrites Fig. 1. Suture lines (partly adolescent and adult) and shell shapes of the earliest representatives of goniatites. (Adapted from Schindewolf 1933.) ACKNOWLEDGMENTS The writer is deeply indebted to the Geological Society of America for a very generous invitation to attend its 1953 Annual Meeting at Toronto. This gave him the welcome opportunity to deliver this lecture in the symposium of the Paleontological STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 219 Society. Dr. B. Kummel very kindly undertook the laborious task of revising the manuscript and of bringing it into the present form. The author wishes to render him sincerest thanks for his trouble and for his stimulating interest. THE EARLIEST GONIATITES According to my conviction, the ammonoids descend from straight orthocone nautiloids via Bactrites, Lobobactrites, Gyro- ceratites, and Mimagoniatites (Schindewolf 1932, 1933, 1935, 1939a). The Michelinoceratida are an old root stock giving rise to many other groups of nautiloids, to belemnoids and, as it seems to me, also to ammonoids. It may be debated whether Bactrites and Lob o~b act rites are to be included in nautiloids or in am- monoids. I prefer to place them in the ammonoids on account of the extreme ventral position of their thin siphuncle and its connection with a ventral lobe, which are important diagnostic features of ammonoids. Bactrites and Lobobactrites (Fig. 1) thus would be the earliest and most primitive representatives of goniatites, still retaining the straight orthocone shell of Michelinoceratidae. The suture line of Bactrites is characterized by the possession of a small ventral lobe; otherwise, it is simply straight. In Lobobactrites the shell is laterally compressed and on each of the flattened sides a broadly rounded lateral lobe is added, while the dorsum is occupied by a dorsal saddle. Gyroceratites (Fig. 1), in my opinion a descendant of Lobo- bactrites, has a loosely coiled shell with egg-shaped protoconch and a large umbilical perforation. The dorsal side of the whorls is still without an impressed concave zone. The suture line is the same as in Lobobactrites. Mimagoniatites (Fig. 1) continues this trend of evolution. In the beginning of its shell it resembles Gyroceratites, but soon the whorls begin to cover each other and to develop a concave zone at the dorsal side. The primary suture and the next sutures in the region of the umbilical perforation still retain the dorsal saddle, which in later developmental stages, as soon as the concave zone appears, is replaced by a dorsal or internal lobe. By the inception of this lobe the typical trilobate goniatitic suture line is completed. The earliest goniatites with tightly coiled shells are Agoniatites 220 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY J U, U,U+U2 L £ M f JUL A. AtA}E f STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 221 (Fig-. 1) and Werneroceras (Pig. 3). In them the protoconch is spirally coiled, the umbilical perforation is lacking, and the whorls from the very beginning possess a concave dorsal zone. As a further progress of evolution the primary suture already is of the type exhibited by the later and adult sutures of Mima- goniatites: it is characterized by a dorsal lobe. It seems to me that this gradual perfection of suture line and of involution is a strong argument in favour of a derivation of the coiled ammonoids from uncoiled orthocone nautiloids. This theory was first advocated by A. Hyatt (1884) and then sup- ported and further elaborated by myself. Having corrected Hy- att's misconception of the protoconch in nautiloids and ammo- noids I was able to prove that the small egg-shaped protoconch of michelinoceratids, bactritids, and all ammonoids is the exact homologue of the coarse conical initial chamber of coiled nauti- loids. This statement is incompatible with the view held by L. F. Spath (1933, 1936), that the coiled or typical goniatites evolved Fig. 2. Suture lines of Anarcestes (Anarcestes) lateseptatus (Beyr.) from the lower Middle Devonian (a, b), Subanarcestes macro cephalus Schdwf. from the lower Middle Devonian (c), and Raymondiceras simplex (Raym.) from the Upper Devonian (d, e). (Adapted from Schindewolf 1933 and 1934.) Fig. 3. Development of suture line of Werneroceras ruppachense (Kays.) from the lower Middle Devonian. (Adapted from Schindewolf 1933.) Fig. 4. Development of suture line of Foordites platypleura (Freeh) from the lower Middle Devonian. (Adapted from Schindewolf 1933.) Fig. 5. Phylogenetic evolution of suture line in Manticoceratidse (and its predecessor Archoceras) from the lower Upper Devonian, a, Archoceras — b, Ponticeras — c, Manticoceras — d, Koenenites — e, Timanites — /, Pharciceras. (After Wedekind 1913b and Schindewolf 1951.) Fig. 6. Development of suture line of Cheiloceras sp. from the Upper Devonian. (Adapted from Schindewolf 1929.) Fig. 7. Development of suture line of Eeticuloceras reticulatum (Phill.) from the Lower Pennsylvanian. (Adapted from Schindewolf 1929.) Fig. 8. a-e, Development of suture line of Sporadoceras miinsteri (v.B.) from the Upper Devonian. (Partly after Perna 1914.) /, Adult suture of Discoclymenia cucullata (v.B.) from the Upper Devonian. (After Schinde- wolf 1951.) 222 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY from coiled Silurian and Ordovician nautiloids of the general type of Barrandeoceras, Tarphyceras or Paleonautilus. The cup- shaped initial chamber, the rapidly expanding first whorls, the mode of involution, the position of the siphuncle and other char- acters of Barrandeoceras and related coiled nautiloids are so profoundly different from those in coiled ammonoids, that such a derivation is impossible. On the other hand, there is in every respect a gradual transition from michelinoceratids through bactritids to the coiled ammonoids. The progressive reduction of the umbilical perforation and the transformation of the proto- conch to be observed from Lower Devonian to Middle Devonian species of Anarcestes and from them to the Middle Devonian Werneroceras definitely postulate such forerunners as Mimago- niatites and Gyroceratites. The only difficulty so far was that this evolutionary trend could not exactly be proved by the geological distribution of the respective genera. But since H. K. Erben (1953) recently recorded the occurrence of Lobobactrites, Gyroceratites, and Mimagoniatites in the Lower Devonian of the Hartz Mountains, these difficulties are eliminated, so that there are no more ob- stacles to regarding Lobobactrites etc. as being the ancestors of the typically coiled goniatites. As a matter of fact the tightly coiled genera Agoniatites and Werneroceras are geologically younger, and Anarcestes, at least, is not known from beds older than those containing Lobobactrites, Gyroceratites, and Mimagon- iatites. The interpretation by Spath, who considered these genera as uncoiled derivatives of Anarcestes, is thus no longer tenable. But these controversies do not affect the question of our suture line. We repeat that all the earlier normally coiled goniatites are characterized by a trilobate suture consisting of the ventral or external lobe (E), the lateral lobe (L), the dorsal or internal lobe (I), and the two saddles E/L and I/L separating them. This elementary suture is the basis from which all the more dif- ferentiated sutures of the later ammonoids can be derived. I have termed the three lobes composing this suture the basic lobes or protolobes, since they and they alone are common to all the host of goniatites, clymenoids, ceratites, and ammonites. In the later representatives of ammonoids these primary lobes are sup- plemented by a varying number of additional lobes which may STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 223 be called secondary or metalobes. The ventral lobe B and the dorsal lobe I are fixed in the plane of symmetry of the conch ; their position is invariable. The lobe E is connected with the ventrally arranged siphuncle ; the lobe I is combined with the impressed zone of the whorls. Variable, however, is the position of the lateral lobe, which has no such limitations. In a few forms (Gyroceratites, Mimago- niatites, Agoniatites) it occupies, as a broadly rounded lobe, the middle of the flanks from the very beginning, i.e., from the primary suture. It then shows a position commonly taken as characteristic for the lateral lobe in a stricter morphologic sense. In most cases, however, the lateral lobe is situated in the primary suture on or at the umbilical seam. Usually small and shallow, it is separated from the external lobe by a broadly vaulted saddle on the flanks. Then, in later ontogenetic stages this lobe mostly is broadened and shifted to the middle of the flanks (e.g., Werner oceras, Fig. 3). In some few genera (Anarcestes, Fig. 2a, b; Subanarcestes, Fig. 2c; Raymondiceras, Fig. 2d, e; Arckoceras, Fig. 5a), how- ever, it may retain the primary position next to the umbilicus up to the mature suture. Its position thus is not, strictly speaking, lateral. But the gradual shifting of the lateral lobe in related goniatites proves that this lobe at the umbilicus is homologous with the lateral lobe of the adult stages in other genera. It may likewise be termed lateral in the somewhat broader sense that it is the protolobe lying laterally from the plane of bilateral symmetry, indicated by the ventral and dorsal lobe. In any case the characteristic lateral position of this lobe on the flanks is nearly always a secondary one. THE DIFFERENTIATED GONIATITES The further elaboration of the suture line consists of the in- sertion of new lobes. The main process is that of subdividing the saddles. In the apex of a saddle a new lobe appears, shallow in the beginning and then gradually deepening. In the simple trilobate suture of early goniatites and in the likewise trilobate primary suture of the more advanced goniatites two saddles are present: the internal saddle (I/L) and the external saddle (L/E). Each of these saddles may be subdivided and thus give rise to new lobes. 224 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY JUL JUL A E J U, U3 U2 L E J U, Uz L £ 14 STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 225 The lobes which are derived from the internal saddle are termed umbilical lobes (abbreviated U), since they are formed in the region of the umbilicus; the lobes originating by the sub- division of the external saddle are called adventitious lobes (A). These lobes arise in a very definite order. The umbilical lobes are formed alternately, U2 follows Ui at the ventral side, U3 comes in dorsally to U2, and so on. The adventitious lobes are inserted in a simple ventrad sequence. In these differentiated goniatites again the fate of the lateral lobe is different. In the same manner as in most of the early goniatites, it is situated in the primary suture immediately next to the umbilicus and is crossed by the umbilical seam. It retains this position throughout its ontogeny in the case that one or more adventitious lobes are formed. Then the middle of the flanks is occupied by the adventitious lobes. This case is by far the most common among late Paleozoic goniatites. Some examples are shown in Figures 6-10. On the other hand, if no formation of adventitious lobes takes place, usually a larger number of umbilical lobes arise which displace the lateral lobe from its original position. The second or third umbilical lobe, respectively, is then situated on the um- bilical seam and the lateral lobe is shifted to the middle of the flanks or still more outwards. This type of suture line is realized in only a few groups of goniatites, as for instance the Mantico- Fig. 9. Development of suture line of Anthracoceras paucilobus (Phill.) from the Lower Pennsylvanian. (After Schindewolf 1951.) Fig. 10. Development of suture line of Imitoceras sp. from the Lower Mississippian. (After Schindewolf 1951.) Fig. 11. Development of suture line of Merocanites applanatus (Freeh) (and asiaticus [Karp.]) from the Mississippian. (Adapted from Karpinsky 1896 and Schindewolf 1929.) Fig. 12. Development of suture line of KonincTcites sp. from the Lower Triassic of the Salt Range (Pakistan). Geol. Dept. Tubingen Ce 1043/3. Fig. 13. Development of suture line of Strenoceras (Strenoceras) sub- furcatum (Ziet.) from the Middle Jurassic. (Adapted from Schindewolf 1953.) Fig. 14. Development of suture line of Spiroceras bifurcati (Qu.) from the Middle Jurassic. (After Schindewolf 1951.) 226 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY ceratidae (Fig. 5), the Prolobitidae, and the Prolecanitidae (Fig. 11). The adult sutures of such goniatites, though they originated in two profoundly different ways, may sometimes be very similar in shape and formal composition (Fig. 15). But owing to their different origin the homology of lobes is quite different. What is the lateral lobe in the one suture is an adventitious lobe in the other, and the lateral lobe of the latter suture corresponds to an umbilical lobe of the first one. It is therefore necessary to regard the actual development of suture line in order to deter- mine the true homologies and to gain a proper designation of the U-Type A -Type u, u2 Fig. 15. Diagrammatic confrontation of the U- and A-type of suture development. lobe elements. This is the more necessary since the two trends of lobe formation are significant for definite groups of goniatites and thus obviously are important for tracing the phylogenetic rela- tions. For brevity, we will call the type with adventitious lobes the A-type (Fig. 15). Its main characteristics are: One or more adventitious lobes are formed, but only one umbilical lobe is present ; the lateral lobe retains the initial umbilical position it held in the primary suture. The other type may be termed the U-type (Fig. 15) : A larger number of umbilical lobes, but no adventitious lobes are formed; the lateral lobe is removed outwards from its original position. The U-type is limited to a STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 227 small stem of goniatites, while by far the majority of late Paleo- zoic forms, the Tornoceratidae, Cheiloceratidae, the Goniatitidae with all their derivatives, follow the A-type. THE MESOZOIC AMMONOIDS Differing from the Paleozoic goniatites (Fig. 12), the primary suture in Triassic ceratites is quadrilobate, i.e., it contains one element more than that of the former. It is to be interpreted in such a manner, that by accelerated development a first umbilical lobe (Uj) (which in Devonian and Carboniferous goniatites, if at all, was inserted in a much later stage of development) has been added , to the three protolobes. The lateral lobe is, from the beginning, situated on the middle of the flanks and afterwards is shifted still more outwards. The primary suture of such a Triassic ceratite (Fig. 12a) corresponds thus in every detail to a later developmental stage of e.g. Merocanites (Fig. lie). The further elaboration of the suture line proceeds exactly in the same way by the addition of two other umbilical lobes (U2 and U3). In later stages a crinkling of the lobes comes in. The umbilical lobes Ui and U3, situated around the umbilical seam, sometimes are strongly enlarged and unevenly differentiated by serration. They may thus simulate a more or less large series of "auxiliary" lobes, which are nothing else than parts of the two umbilical lobes, produced by digitation. The suture line of the Triassic ceratites follows the U-type, and it can only be derived from goniatitic sutures representing the same type of lobe development, as is the case, for instance, in the Prolecanitidae. Now it is the general conviction, which can be supported by my own studies of Triassic and Permian ammonoids, that the Lower Triassic ceratites are descendants of the Permian Xenodiscidae, which on their part go back to the Carboniferous Prolecanitidae. So our observations and interpre- tations of suture line lead to the same conclusions as those drawn from other morphological features. They can be used to reinforce and to define more exactly our phylogenetic conceptions. So far as I can see, all the Triassic ceratites represent the U-type of lobe formation; I have never observed a case of the A-type. It seems to me, therefore, that the A-type prevalent in the Paleo- zoic was extinguished at the Permian-Triassic boundary and left 228 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY no descendants in the Mesozoic (except, at most, some few surviv- ing forms in the Lower Triassic). In the Triassic and, as we will see, also in the Jurassic and Cretaceous the U-type alone is dominant. The evolutionary cycle of Jurassic and Cretaceous ammonites originates in one of the lineages of Triassic ceratites. So it is to be expected that the same type of lobe formation will be con- tinued by them, and that indeed is the case. The primary suture fi ^ -< <*J £ Ammon. "> «>» <«J ^ ^> *<» ^ I Cera t. \ £ s f-\ £ Pro/. Gon. d <> U-Type A -Type CS «5 i. k4£o/7.>J — " C* Fig. 16. Diagrammatic representation of the phylogenetie relations of ammonoids. of these forms (Figs. 13, 14) is quinquelobate ; by a further acceleration of development, two umbilical lobes are represented in it besides the three protolobes. Their primary suture is thus equivalent to a more advanced stage in the development of the Triassic forerunners. The lateral lobe is situated on the flanks; no adventitious lobes are formed between it and the external lobe. Summarizing these observations on the suture line we gain a STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 229 very simple picture of the broad phylogenetic relationships among ammonoids represented in Figure 16. Of course, this is only a rough scheme, omitting all details, and needing further refinement. PRIMARY SUTURE AND PROSUTURE The evolution of the first genuine or primary suture and of its further ontogenetic differentiation shows a definite trend as summed up on the right side of Figure 17. In Devonian am- monoids, the fully coiled goniatites and the clymenoids, the primary suture is trilobate ; it consists of the three protolobes I, Manticoceras Sudeticeras Schlotheimia Oxynoticeras Fig. 17. Comparison of prosutures (Pro) and primary sutures (Pri) of various ammonoid genera. (Adapted from Schindewolf 1929 and 1942.) L, and E only. This simple type of suture is retained until the adult stage in early Devonian ammonoids ; the lobes only may become deeper. In more advanced Devonian forms one or more metalobes are added during development, The insertion of a first umbilical lobe by subdividing the internal saddle usually takes place between the 20th and the 25th suture. In Carboniferous goniatites the initial suture is of the same type as that of Devonian forms, but by acceleration of develop- ment the umbilical lobe already appears in the next suture or the one immediately following. This quadrilobate stage becomes permanent as the primary suture in Triassic ceratites, and a second umbilical lobe is added in about the 12th to the 16th 230 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY suture. Finally, by acceleration, this quinquelobate stage be- comes the primary suture of Jurassic and Cretaceous ammonoids. It may be further differentiated by the insertion of a third um- bilical lobe in the 12th to the 15th suture line. This is a very impressive example of the acceleration of development, the skip- ping of old developmental stages and the insertion of new ones into increasingly earlier stages of ontogeny. The primary suture, however, is not the absolutely first suture line ; in the development of every ammonoid it is preceded by another suture belonging to the septum closing the protoconch. It is totally different from the primary suture and has been termed the prosuture (Schindewolf 1928, 1929). In its ventral and dorsal parts it is characterized by having saddles instead of the external and internal lobe so highly significant for the normal ammonitic suture. Often the differences between the prosuture and the primary suture are so great that the one in nearly all cases has lobes, whereas the other shows saddles. The primary suture then may look like a reversed prosuture. Furthermore, the number of elements between the ventral and the dorsal side of the two sutures is quite independent. It may be the same (Fig. 17, Manticoceras, Oxynoticeras), or it may be larger (Fig. 17, Schlotheimia) or even smaller in the primary suture as compared with that in the prosuture (Fig. 17, Sudcticeras) . Finally, the phylogenetic evolution of the prosuture from Devonian to Jurassic and Cretaceous ammonoids (Fig. 18, left side) is entirely different from that of the primary suture. It is not necessary to go into the details here, since I have given full descriptions in my earlier papers (Schindewolf 1929, 1942). It may suffice to state that the special character of the prosuture is maintained throughout the whole history of ammonoids, and that the sequence of the newly inserted lobes is quite different from that of the later sutures. Because of these manifold differences in shape and evolution, it is impossible to deduce the primary suture from the prosuture, whereas the former is closely connected with the following suture lines and forms the basis for their further differentiation. There is a definite break between the two sutures; the morphological features of the prosuture do not find a continuation in the indi- vidual development. It is therefore considered as an ephemeral STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 231 Prosufuren Sufuren Ju ra Trias Karbon Devon J U,U3U2 L E Fig. 18. Phylogenetie evolution of the prosuture and primary suture in ammonoids. (After Schindewolf 1942.) larval organ and we have to assume that the animal must have undergone a kind of metamorphosis between the secretion of the first and the second individual suture line. Only the latter, Fig. 19. Median dorsoventral sections through the initial shells of Gonia- tites (Paragoniatites) newsomi Smith from the Mississippian (a), and Polyptychites sp. from the Lower Cretaceous (&). (After Schindewolf 1939a.) 232 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY our primary suture, is a genuine suture line in the usual sense. A similar difference, like that between the prosuture and the primary suture, is also expressed in the shape of the two respec- tive septa. This may be illustrated by the two median dorso- ventral sections in Figures 19a and b. In Goniatites newsomi (Fig. 19a) the proseptum delimiting the protoconch differs from the second septum by (1) its greater thickness, (2) its concave instead of convex curvature, and (3) above all by the shape of the ventral part of the septum, which is represented by a minute dent of the shell wall only. The second septum, however, in its ventral part is well developed and bears a hammer-like siphonal funnel of exactly the same nature as in the following septa. There is thus a distinct break between the first and the second septum, whereas the latter is closely connected morphologically with the following ones. The only difference is the adoral spur at the dorsal side of the siphonal funnel in contrast to the retro- fleeted funnels of the next following septa. It may be due to the lack of room ; a backward extension of the siphonal funnel is prevented by the immediate neighbourhood of the proseptum. Polyptychites (Fig. 19b) from the Lower Cretaceous exhibits a similar profound difference between the first septum and the second septum. Further median sections of goniatites, ceratites, and ammonites, all showing the same phenomena, have been figured by Schindewolf (1939a and 1942). I cannot agree, therefore, with J. C. A. Bohmers (1936), who assumes two prosepta which are said to be different from the re- maining septa. It may be that in some cases the second septum shows a somewhat intermediate character between the first and the third. The main break, however, seems to occur from the first to the second septum, thus coinciding with the differences between the prosuture and the primary suture. PARTITION OF LOBES Besides the two main trends in the formation of suture line, by subdivision of saddles, there are still several other modes of minor importance. It is not our intention to describe all the different processes forming median saddles, the so-called " statur- al'' lobes, which could be translated as seam-lobes, or the strange anomalies to be observed in the insertion of umbilical lobes in a STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 233 u f Fig. 20. Development of suture line of Marathonites dieneri Smith from the Lower Permian. (Adapted from Schindewolf 1939b.) 234 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY certain group of Jurassic ammonoids. But one more process should be mentioned and that is the partition of lobes. It is much rarer than the subdivision of saddles and is subordinate to it insofar as it occurs on the platform of the one or the other of the main types in the elaboration of suture line. Nevertheless, it is very- characteristic for several distinct groups of ammonoids, as for instance the Agathiceratidae and Popanoceratidae among the Paleozoic forms, the genera Metaplacenticeras and Coilopoceras in ammonites. In these forms one or several lobes are affected by the sub- division. From the base of these lobes two saddles rise and split up the original lobe into three components, which with growing saddles appear as independent lobes. But of course they are not autonomous lobes in the sense of the independent lobes newlj inserted by subdivision of saddles. They are only parts of pre- existing lobes and are to be marked as such. This kind of development has been studied in detail by myself in the genus Marathonites (Schindewolf 1931, 1939b). Since it and the popanoceratids are on the whole derivatives of the main goniatite group with the A-type of suture line, I assume now that Marathonites also will follow this type. The earliest ob- served stage of development (the primary suture unfortunately is not accessible) then is to be interpreted as noted in Figure 20a. Concerning the number and position of lobes it corresponds to the suture lines of Imitoceras (Fig. 10), Reticuloceras (Fig. 7), Anthracoceras (Fig. 9), and other Carboniferous genera. All the lobes are still undivided ; only the median saddle in the ex- ternal lobe makes its first appearance. In the stage of Figure 20b the adventitious lobe A has become trifurcated and shows a development as in the permanent suture of Proshumardites. In the following ontogenetic steps (Fig. 20c-/) the saddles in the lobe are growing higher and the dorsal element is divided once more into two parts by a saddle rising from its base. So in the suture of Figure 20g, the adventitious lobe is represented by a large complex consisting of four ap- parently independent lobes. In a similar manner the umbilical lobe and a little later also the lateral lobe is split up into a series of components. In the umbilical lobe it is, however, the ventral of the three elements that is bifurcated once more in the same STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 235 way as the dorsal one of the adventitious and the lateral lobes. The adult suture of Marathonites and of the other popanocera- tids thus contains only five lobes, though it is composed of a large number of elements. It is self-evident that in terminology of lobes this type of development must be distinguished from the mode of subdividing the saddles. THE SUTURE LINE OP SPIROCERAS In conclusion, I wish to demonstrate a unique, strange case which I recently analyzed and which furthermore may prove the necessity of carefully studying the development of suture line. It concerns the criocone genus Spiroceras of the Middle Jurassic. The mature suture line (Fig. 14a") is strongly reduced. It seems to contain only three lobes. Besides the ventral and the dorsal lobes there is only one definite lobe which, without knowledge of its development, would be identified with the lateral lobe. Each of the two saddles between these lobes bears in its apex a somewhat deeper indentation. They have been interpreted either as mere crinkles or as reduced lobes, as an adventitious or an umbilical lobe, respectively. The development of suture line, however, proves that each of these interpretations is absolutely erroneous. The primary suture of Spiroceras (Fig. 14a) agrees completely with that of the other Jurassic and Cretaceous ammonites (Fig. 13a). It is com- posed of the elements I, Ui, U2, L, and E. They show the same position and relative size as usual. In the course of development, however, the porportions are profoundly changed. The initially very shallow lobe U2 quickly becomes deeper and larger, whereas the neighbouring lobes are gradually reduced in size. The large lobe of the adult suture, supposed to be the lateral, is thus in reality the accelerated lobe U2, while the two insignificant crinkles are to be homologized with the lateral and the first umbilical lobe ! This is a very surprising result, which by no means could have been deduced from the adult suture line. It corroborates our thesis that a scrupulous study of ontogeny is necessary for a consistent interpretation of suture line. Only in this way can the true homologies of lobes be recognized and distinguished from mere analogies. If such investigations could be done in all crucial 236 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY cases, the suture line would prove still more valuable for reveal- ing the natural relationship of ammonoids than is the case today. REFERENCES BOHMERS, J. C. A. 1936. Bau und Struktur von Schale und Sipho bei permischen Am- monoidea. Diss. Univ. Amsterdam, 125 pp., 2 pis. Erbbn, H. K. 1953. Goniatitacea (Ceph.) aus dem Unterdevon und dem unteren Mitteldevon. Neues Jahrb. Geol. Palaontol., Abh., vol. 98, pp. 175-225, pis. 17-19. Hyatt, A. 1883- Genera of fossil cephalopoda. Proc. Boston Soc. Natur. Hist., 1884. vol. 22, pp. 253-338. Karpinsky, A. 1896. Sur 1 'existence du genre Prolecanit.es en Asie et sur son de- veloppement. Bull. Acad. imp. Sci. St.-Petersbourg, ser. 5, vol. 4, pp. 179-194. NOETLING, F. 1905. Untersuchungen iiber den Bau der Lobenlinie von Pseudosage- ceras multilobatum NOETLING. Palseontographica, vol. 51, pp. 155-260, pis. 19-27. 1906. Die Entwickelung von Indoceras baluchistanfinse NOETLING. Ein Beitrag zur Ontogenie der Ammoniten. Geol. u. palaeontol. Abh., n.ser., vol. 8, pp. 1-96, pis. 1-7. Peena, E. 1914. Die Ammoneen des oberen Neodevon vom Ostabhang des Siid- urals. Mem. Com. geol. St.-Petersbourg, n. ser., vol. 99, 114 pp., 4 pis. SCHINDKWOLF, O. H. 1928. Zur Terminologie der Lobenlinie. Palaeontol. Zeitschr., vol. 9, pp. 181-186. 1929. Vergleichende Studien zur Phylogenie, Morphogenie und Termi- nologie der Ammoneenlobenlinie. Abh. preuss. geol. Landesanst., n.ser., vol. 115, 102 pp., 1 pi. 1931. Uber den Ammoniten-Sipho. Sitz.-Ber. preuss. geol. Landesanst., vol. 6, pp. 197-209, pi. 5. 1932. Zur Stammesgesehichte der Ammoneen. Palaeontol. Zeitschr., vol. 14, pp. 164-181. 1933. Vergleichende Morphologie und Phylogenie der Anfangskam- mern tetrabranchiater Cephalopoden. Eine Studie iiber Herkunft, Stammesentwicklung und System der niederen Ammoneen. Abh. preuss. geol. Landesanst. n.ser., vol. 148, 115 pp., 4 pis. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 237 1934. Tiber eine oberdevonische Ammoneen -Fauna aus den Rocky Mountains. Neues Jahrb. Mineral., etc., Beil.-Bd., ser. B, vol. 72, pp. 331-350. 1935. Zur Stammesgesehiehte der Cephalopoden. Jahrb. preuss. geol. Landesanst, vol. 55 (1934), pp. 258-283, pis. 19-22. 1939a. Tiber den Bau karbonischer Goniatiten. Palaeontol. Zeitschr., vol. 21, pp. 42-67, pi. 4. 1939b. Zur Kenntnis von Pericleites RENZ und verwandter palaozoia- cher Ammoneen. Jahrb. preuss. geol. Landesanst., vol. 59 (1938), pp. 423-455. 1942. Evolution im Lichte der Palaontologie. Bilder aus der Stam- mesentwicklung der Cephalopoden. Jenaische Zeitschr. Med. u. Naturw., vol. 75, pp. 324-386. 1951. Zur Morphogenie und Terminologie der Ammoneen Lobenlinie. Palaontol. Zeitschr., vol. 25, pp. 11-34, pi. 1. 1953. Tiber Strenoqeras und andere Dogger-Ammoniten. Neues Jahrb. Geol. Palaontol., Monatsh., 1953, pp. 119-130. 1954. Tiber die Lobenlinie der Ammonoideen. Neues Jahrb. Geol. Palaontol., Monatsh., 1954 (in press). Spath, L. F. 1933. The evolution of the Cephalopoda. Biol. Reviews, vol. 8, pp. 418-462. 1936. The phylogeny of the Cephalopoda. Palaeontol. Zeitschr., vol. 18, pp. 156-181, pi. 9. Whdhktktd, R. 1913a. Beitrage zur Kenntnis des Oberdevon am Nordrande des Rhein- ischen Gebirges. 2. Zur Kenntnis der Prolobitiden. Neues Jahrbl Mineral., etc., 1913, vol. 1, pp. 78-95, pi. 8. 1913 b. Die Goniatitenkalke des unteren Oberdevon von Martenberg bed Adorf. Sitz-Ber. Ges. naturf. Freunde Berlin, 1913, pp. 23-77, pis. 4-7. 1916. Tiber Lobus, Suturallobus und Inzision. Centralbl. Mineral., etc., 1916, pp. 185-195. 1918. Die Genera der Palaeoammonoidea (Goniatiten). Palaeonto- graphica, vol. 62, pp. 85-184, pis. 14-22. Bulletin of the Museum oi Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 IX. M acroevolution and the Problem of Missing Links By Alexander Petrunkevitch Yale University, New Haven, Conn. The problem of missing links had its inception in Darwin's theory of evolution. As soon as his Origin of Species by Means of Natural Selection was published, it was generally assumed that if this theory is correct then there must be an unbroken chain of connecting links between living animals and their fossil ancestors. Because of the direct bearing of the theory on the question of the origin of man, the search for missing links was taken up by anthropologists, paleontologists and zoologists. It has continued unabated for almost a century and presumably will continue to occupy our attention for many years to come. However, with the universal acquiescence in the idea that all species of plants and animals, man included, are the product of evolution and not of creation, the problem of missing links assumed an aspect different from its original one which concerned the proof of evolution as against special creation. At present evolution is an accepted fact. What remains in dispute is only the method of evolution. On the answer to this question there is considerable disagreement. Opinions differ not only about the ancestry of various categories of animals, but also about the probability of the eventual discovery of now missing connecting links, as well as about the possibility or impossibility of explain- ing the origin of larger categories by a process of speciation as attempted by Darwin. The reason for such disagreement lies not in insufficiency of evidence, but in the undeniable fact that, with the exception of observations made in breeding experiments, all evidence relating to ancestry is of an indirect nature. As such, it is subject to doubt, error and different interpretation. The truth of this statement becomes immediately apparent when one considers the question of the relationship between the coelenterates and the flat-worms. Basing their opinion on an analysis of extensive data gradually accumulated during more 240 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY than a century of careful research, the majority of zoologists derive the flat-worms from the coelenterates, especially from the ctenophores or comb- jellyfish which they resemble in many respects. The greater complexity of organization of flat-worms is generally explained by progressive evolution and by the formation of new characters better adapted to crawling. On the other hand, the Yugoslavian zoologist Hadzi (1953), using the same data, derives the coelenterates from the flat-worms. He explains the origin and the greater simplicity of structure of the former as a result of regressive evolution. According to his theory, rhabdocoele Turbellaria gave up their mode of life as crawling animals, became sessile and gradually lost all structures which were now useless to them. At the same time they acquired a new plan of structure, becoming radially symmetrical out of originally bilateral animals. Developing his theory to its logical end, Hadzi reverses the current idea of the evolution of coelen- terates also within that group itself. He considers the Anthozoa to be the oldest group which developed first and then produced the Scyphozoa and finally the Hydrozoa. According to his in- terpretation, the latter represent the acme of regressive evolution. The ctenophores Hadzi derives from another class of flatworms, namely the Polyclada. Hadzi 's ideas cannot be dismissed lightly. The fact that they are diametrically opposed to the ideas promulgated by other zoologists is not caused by insufficiency of knowledge or by faulty logic. In the absence of direct evidence which could be furnished only by breeding, both his and his opponents' ideas are based entirely on similarities of structure. Such similarities can be arranged in a series of either increasing or diminishing complex- ity of organization. The controversy could be settled, but even then only partially, by paleontological evidence. If it were pos- sible to show that one of these groups of animals appeared at an earlier geological period, then the later group could not pos- sibly be the ancestral one. But this would be insufficient as proof of the correctness of the opposite contention, namely that the older group produced the younger one. There is another possi- bility, besides that of direct descent, that can account for the existence of similar structures in groups otherwise differing in many ways. It is the simple assumption that both groups are STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 241 descendants of the same ancestral stock, but that their relation- ship to each other is that of "cousins" and not of parents and children. "We are all familiar with such cases when animals origi- nally considered to be direct ancestors were later demoted to the status of more or less distant relatives representing branches of a common ancestral stem. This applies not only to larger and smaller categories of animals, but also to closely related genera as exemplified by the history of the derivation of man from anthropoid apes. Leaving out of consideration the direction of evolution, i.e., the question as to whether in a given case we have to do with progressive or regressive changes, the strength of an argument in favor of direct ancestral relationship grows with the increase of similarities in number and in degree. If speciation is assumed to be the only method of evolution, operating since the appear- ance of animal life on earth, then it should not be difficult to find connecting links between species possessing the greatest number of similar structures and the greatest degree of simi- larity. The theory demands that new species be produced by old species and in that way produce new genera ; or better, that new species become eventually so different from their parental ancestors that the erection of a new genus for their reception becomes necessary; that some of the new genera produced by speciation require the erection of new families and so on until through the same process of continued speciation new orders, classes and phyla are produced. Let us assume for the sake of simplicity that a new species can be produced through the trans- formation of a population owing to the gradual spread of a single new mutation — the extinction of the parent species and the survival of the new species being the result of natural selec- tion. With the exception of special cases, the number of deaths corresponds to the size of a population regardless of its com- position, decreasing or increasing with the population. On the other hand, the number of fossils is always much smaller than the number of individuals which die, because the majority of the bodies disintegrate before they can become fossilized. Conse- quently fossils represent only a random sample of a living population. If that population consists of a much greater num- ber of individuals of the parental type with the characters of 242 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the original species than of individuals with modified characters, as one would expect to be the case at the beginning of speciation, then all fossils of that generation may be entirely of the original, parental type, because the ratio between living and fossilized individuals is so great that the chance of mutants becoming fossilized is greatly reduced and approximates zero. But when in the course of time and owing to the survival of the fittest, the mutants have reached the point at which their number in the living populaion is equal to that of the parental type, i.e., when the population consists of 50 per cent of the old type and 50 per cent of the mutant type, then each random sample of the population will also contain the same proportion of both types. Environmental conditions and chances of fertilization being the same for both types, both should be represented by an equal num- ber of fossils. From now on, with an increase of mutants over old types, fossils should also show an increasing number of mutants. The number of fossils of the old type will now grow smaller and smaller until their complete disappearance at an earlier date than the actual extinction of the living species. Thus for the total span of time required for the production of a new species the number of fossils of the parental type should be the same as that of the mutant type. In other words, connecting links between any two species standing in a direct relationship of ancestor and descendant should be as common as the two species themselves. If, as Darwin's theory assumes, all higher categories of classification have been produced by speciation of the kind with which we are now familiar (owing to the extensive work of geneticists), then the ratio of connecting links between all higher categories should also be the same as between species all the way back to the first appearance of fossils in the strata of the earth, forming an unbroken chain between them. There should be no missing links at all except in cases of animals whose fossilization is either impossible on account of the softness of their body, causing its rapid decay, or is indeed so rare for the same reason, that if such fossils were ever found it would be impossible to decide whether they represent the typical or the mutant type. The above analysis is correct only on the assumption that speciation is a continuous process. That mutation is a constant STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 243 process is generally accepted today. Thus Dobzhansky (1951) says on page 73, "The mutation process constantly and unremit- tingly generates new hereditary variants — gene mutations and chromosomal changes." But if it could be shown that in each line of evolution there are long periods of rest between periods of speciation, intervals during which newly produced species con- tinue their existence as such and no new mutations or heritable changes of any kind are produced, then, of course, the ratio between typical and mutant fossils would be considerably greater. It would depend on and reflect the duration of each pause. But the greater ratio would apply only to the total number of fossils during the combined periods of pause and speciation. The ratio for the latter alone would still remain the same, only the horizon containing the fossils of both types would be re- duced to a thinner layer corresponding to the time between the beginning of speciation and the extinction of the parental species. In cases of common species there should be no difficulty in finding connecting links. The existence of such pauses was assumed by de Vries, the creator of the mutation theory (1901). In his address before the society of German naturalists and physicians he considered the average duration of each interval between the periods of mu- tation to be no less than 4000 years. Accepting the then current idea that the age of life on earth amounted to 24,000,000 years, de Vries arrived at the conclusion that 6000 mutations sufficed to complete the evolution of now living phanerogams from their unicellular ancestors to the present state of their organization, (de Vries, 1901, p. 63). Although the total time allotted to the existence of life on earth has been greatly increased since de Vries 's time, owing to new methods of measurement devised by students of radioactivity, its span seems still to be much too small to account for the evolution of all animals by the slow process of speciation. Species are not produced suddenly, by a single muta- tion. Even with the selection intensity equal to 0.5, it requires 30 generations to bring a mutation frequency to 100 per cent (Lerner, 1950). The time required for the production of a new species varies considerably. But with the exception of special cases, speciation is much slower. Haldane, in his controversy with Lunn (1935, p. 164), admits that "selection is the only agency 244 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY which can lead to important changes within times measured in millions of generations." Simpson (1944, p. 69), in discussing the classical case of the evolution of the horse, writes : ' ' In larger populations dominated by selection, a million years may not suffice to introduce a new species and relatively seldom suffices for evolution of generic rank. Note the previous estimate that the actively and progressively evolving tertiary horses took on an average 8,000,000 years for evolution from one genus to the next." If we put the average time for the production of a vertebrate species at only 500,000 years, and regard the process of evolution as being continuous, it would have to be completed from fish to man within only about 625 steps from one species to another. If we accept intervals between speciation equal to the time required for speciation, we would have to accept that only 312 species-transformations separate man from fish. But the slow tempo of speciation is not the only reason for my belief that the evolution of the higher categories of animals could not have been accomplished by speciation. If it depended entirely on speciation, then formation of new species either by production of new characters or by modification or loss of old species characters would be insufficient for achieving the evolu- tion of higher categories. To do this, every animal of whatever phylum, class, order or family would have to be first completely dedifferentiated to the state of organization of the ancestor of the corresponding category. Production of new characters required for the establishment of a new phylum by speciation is a step toward complication of organization, not toward its simplifica- tion. Nobody could maintain that our ancestors were more complexly built than we are. To be sure, we are familiar with numerous cases of regressive evolution leading to secondary simplification of structure, due to various causes, such as para- sitism, loss of locomotion with adaptation to sessile existence, etc., but such dedifferentiation is not identical with that which is prerequisite for the evolution of a new phylum or class. A creature like the parasitic cirripede Sacculina or the parasitic copepod Lernaea (the adult females of which have no resemblance whatsoever to other, free living species of Crustacea because in the course of development their organization becomes reduced to nothing but reproductive glands and nutritive tubes), would still STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 245 have to return to the state not of their own nauplius which can- not reproduce, but to that of their ancestor which still could not only reproduce and develop into an adult of its own species, but which was so primitive that it still possessed the potential ability of developing into a new order of Entomostraca. If that ancestor of Entomostraca were to engage in the evolution of a phylum other than that of Arthropoda, to which it belongs, then it would have to dedifferentiate further yet to the state of the ancestor which still possessed the potential ability to start the evolution of any other phylum, including even the vertebrates. The necessity of dedifferentiation of this kind before a species can resume creative speciation in a new line may be called the law of precession. It involves the formation of a chain of regres- sive links in addition to that of progressive links. Of course, nobody imagines that a Recent primate, carnivore or monotreme could ever start evolving a new phylum, because all Recent mam- mals are so complex that one could not expect fundamental changes in their anatomical structure, but attempts are still made to derive vertebrates from various groups of invertebrates, arthropods from annelids, and practically all classes of arth- ropods from trilobites. If phylogeny through speciation were possible, then the law of precession would have to be taken into account, the time required for the evolution of the animal king- dom would have to be extended to include the time required for dedifferentiation. There should be a chain of transitional links similar to that of progressive evolution, but reversed in the order of sequence and representing the steps of regressive evolution, preceding and prerequisite for each start on a new line of pro- gressive evolution. It is well known that not only species, but genera and even families present transitional characters, while orders, classes and phyla are sharply delimited. There are no transitions between scorpions and spiders, between beetles and butterflies, between fairy-shrimps and daphnids; there is none between Crustacea, Arachnida and Hexapoda ; between Arthropoda and Echinoder- mata and so on. Some striking resemblances exist, but they are only external resemblances. Some cockroaches look like beetles, but are in every other way typical cockroaches. A Carboniferous arachnid of the order Ricinulei resembles some beetles to such 246 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY an extent that its discoverer Buckland actually mistook it for a beetle and gave it the generic name Curculioides, Curculio being a well known genus of beetles. But Curculioides was soon proven to have all the characteristic structures of an arachnid, which could not possibly be confused with the characteristic structures of beetles. Nor can resemblances between so-called homologous structures be regarded as evidence of close relationship and used as examples of connecting links. This lamentable overestimation of their value is still broadly in vogue. Invertebrate zoologists and paleontologists cling tenaciously to the old concept of homol- ogy as defined by Owen. In the case of arthropods they regard as sufficient evidence of homology the presence of even widely dissimilar appendages on the same segment of the body, counted in an early embryonic stage from its anterior end backward. Vertebrate anatomists, beginning with Owen himself, did not use this method because of the fluctuation in the number of cervical vertebrae in birds, but never doubted the homology of such structures as the foreleg on a mammal and the wing of a bird. The segmentation of the arthropod head is still a subject of controversy. While we may safely homologize the segmenta- tion of an insect leg with that of an arachnid leg, although they may not belong to the same body segment, we may not do this so easily when comparing the arachnid chelicerae with the crusta- cean second antennae even though they are regarded as belonging to the same cephalic segment. The fact is that under the influence of experimental zoology the concepts of homology and analogy have undergone considerable change and the very distinction between them is now not simple to define. This is further compli- cated by the phenomena of so-called homomorphism (Novikoff, 1953), the existence of similar organs in unrelated classes of ani- mals, and by cases of accidental and experimental development of eyes and of appendages on segments of the body to which they do not normally belong. All this and similar facts make the use of "homologous" organs as an argument in favor of direct ancestry or even of close relationship, not nearly as convincing as it was thought to be some time ago. They are suggestive of some relationship, but by no means a proof of it. We have to admit and to face the fact that such connecting links representing gradual transition from one phylum to another, from one class STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 247 to another and from one order to another do not exist in in- vertebrates. Vertebrate anatomists seek to prove that such tran- sitions exist between the classes of vertebrates. They derive the higher classes of vertebrates from fishes on the basis of struc- tural evidence. But even if such transitions among vertebrates can be demonstrated, there may be some other explanation for their existence than that offered by evolution through speciation — an explanation which would bring them in line with the in- vertebrates. It seems clear that if we are to solve the problem of evolution within the available geological time, we must either admit the possibility of a much greater speed of speciation or restrict the scope of speciation and accept the existence of another mode of evolution in the case of the higher categories. The speed of selec- tion is undoubtedly correlated with the speed of reproduction. Other factors are known which also have a bearing on the speed of speciation. But the paleontological evidence is rather in favor of slow speciation, while the evolution of higher categories often showrs a remarkably higher speed. Moreover, evolution by specia- tion offers no other answer for the lack of connecting links be- tween higher categories, except the lame one of the extreme rarity of such fossils. On the other hand, an answer involving the as- sumption of another mode of evolution was given several times, although its bearing on the problem of missing links seems to have been overlooked. Goldschmidt (1933, 1940) called this mode macroevolution ; Lunn (1935) spoke of it as major evolution; Simpson (1944) referred to it as megaevolution. In view of the objections raised by geneticists, zoologists and paleontologists to the existence of another mode of evolution differing from that of speciation, we must first consider wherein phylogeny differs from ontogeny besides the already mentioned lack of connecting links, and then show why macroevolution gives an answer to the above questions. When one considers the development of any animal, one is forced to regard the regularity of the process as something much more striking than its casual and slight irregularities, as if the development followed a predetermined design. The sequence of the stages of development is fixed, some following one upon the other, some beginning at the same time and either running 248 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY synchronously to completion or if asynchronously, yet at a pre- determined rate and time required for completion. If ever a deviation of the development from its prescribed course takes place, it is either corrected by the intervention of regulatory forces or else leads to the formation of monstrosities or even causes the death of the embryo. In closely related species out- ward differences in development appear only toward maturity, even though microscopic study and breeding reveal differences in the egg. The less two animals are related to each other, the earlier some differences in their embryonic development appear. In species belonging to different families, differences may appear shortly after the appearance of the first rudiments of organs ; in different orders — even shortly after gastrulation. It is now known that all such changes are the result of corresponding, invisible changes in the genes either through mutation, recom- bination or loss. This knowledge is relatively recent and did not exist in the past century. When Ernst Haeckel (1866) proposed his Biogenetic Law according to which ontogeny is an abbreviated recapitulation of phylogeny, he conceived it on the evidence of the regularity of development and of striking similarities between early embryos of distantly related species, as had been pointed out even earlier by von Baer. Haeckel (1877) further elaborated his theory by postulating the existence of a hypothetical ancestor of all metazoa. He imagined that it was a free-living, marine animal having the simple structure of a gastrula and proposed for it the name Gastraea. Haeckel 's theory attracted a number of enthusiastic supporters, but was later attacked from all sides and almost discarded as a fanciful speculation. Some of Haeckel's examples were indeed fanciful and the objections to them are valid and remain in force. Other objections require careful re- examination. It was pointed out, for example, that the de- velopment of Recent animals takes place along narrow lines which are similar in only closely related groups, but differ widely in the case of different phyla; that free-living animals morpho- logically similar to the later embryonic stages do not exist; that the gastrula stage can be easily explained as a result of special mechanical and physiological requirements of growth and nour- ishment ; that free-living larvae with the structure of a gastrula, such as the pluteus, bipinnaria, auricularia, tornaria, trocho- STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 249 shaera, veliger, etc., are as highly specialized as the corresponding adults and therefore cannot be used as evidence of relationship or, as Yves Delage (1898, p. 331) puts it in discussing the origin of Prochordata, that they represent at best a very distant rela- tionship of the Prochordata to the Echinodermata, as if the latter were separated from the ancestors of the former consider- ably before these acquired their distinctive characteristics. After serious examination of the value of these objections on the basis of evidence adduced by others and that assembled by myself in the course of long study of arthropods and especially of arachnids, I have come to the conclusion that most of the objections are wrong because of a literal interpretation of Haeckel's Biogenetic Law which, when properly interpreted in the light of modern knowledge, is perfectly sound and universal in its application. The objection that no living or fossil animal has the structure of the hypothetical gastraea endowed with the ability of self-perpetuation is quite correct as far as it goes, but it is also true that if it ever existed in the past, a fossilized gastraea, on account of its small size and perishable constitution would be not only much rarer than even a fossilized medusa, but also would be much more difficult to find, unless preserved in something like the Old Red Sandstone of Rhynie Chert and handled in a similar way, i.e., in thin sections through the rock, and studied with the aid of a microscope under high power. It is quite true that free-living larvae having the structure of a gastrula are all specifically distinct and that the invagination of the blastoderm, leading to the formation of a gastrula can be explained as a response to physico-chemical stimuli, but such an explanation does not take into account the genetic composi- tion of the hereditary apparatus of the nuclei. The fact that pressure on the elastic wall of a thin rubber balloon may produce an invagination similar to that produced by gastrulation does not prove that pressure is the only requirement in both cases. All attempts to explain even such a simple phenomenon as loco- motion on the basis of simple physical laws have broken down completely unless the voluntary element is taken into account. The free-living larvae of Echinodermata, Annelida, Mollusca, etc., are indeed specifically distinct, showing modifications of both the adaptive and the passive type, but their possession of 250 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY specific characters does not change the fact that they all have the same basic structure of a gastrula. The objection that there never have been any free-living animals comparable to the later embryonic stages is not only true in this respect, but far from being a refutation is, on the contrary, a direct confirmation of the Biogenetic Law, because it is quite true that beyond the gastrula stage the development of any animal species repeats the development of its ancestors and is thus a recapitulation of its own line of descent. A species of Drosophila does not repeat in its ontogenetic development the evolution of a starfish or an octopus. It repeats only its own evolution. It follows the de- velopment of an arthropod through the stage of a gastrula to the point at which that arthropod ancestor became an insect, then the development of that insect ancestor to the point when it became a fly and finally the development of the fly ancestor through the stage of the Drosophila ancestor to the stage of the particular species under observation. Anything else is impos- sible and unthinkable, because it would contradict every principle of genetics. A deviation from this rule is possible only when some change in the genetic composition has taken place or some subsequent injury has interfered with normal development. The objection of geneticists to a mode of evolution different from that of speciation (based on the assumption that changes of the magnitude of macroevolutionary ones would be so rare and so disadvantageous that they would be eliminated by selec- tion or would be lethal from the start), is as unsupported by evidence as the assertion that if they were viable we would meet with such changes either in nature or in the laboratory. Both objections are based on observations of living animals, apply only to speciation, and are gratuitous when applied to macro- evolution which took place in past geological eras. Highly dis- advantageous non-heritable monstrosities of various kinds do not necessarily cause premature death, nor prevent monsters from giving birth to normal children. The celebrated Siamese twins, Cheng and Eng, married two sisters in America, had 22 children by them and died in 1874 at the ripe age of 62 years, one a few hours after the other, because they could not be sepa- rated from each other surgically. Human monsters with two heads, two chests and four arms, but a single pelvis and a single STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 251 pair of legs have been observed several times. One of the most celebrated cases of this type came to the notice of King James IV of Scotland, who took the twin monsters under his protection and gave them an education. They lived 22 years (Martin, 1880). Monsters among invertebrates are not uncommon. One finds them in medusae, in worms, in starfishes, in molluscs, in arthropods. Many of them become sexually mature and repro- duce. Under my own observation a female Cynthia moth with rudimentary wings copulated with a normal male and laid a normal batch of fertilized eggs. I have caught, in Saxony, a male beetle Prionus coriarius copulating with a normal female. The male had deformed elytra and lacked one leg. In Panama I watched a male spider Nephila clavipes copulating with a normal female. He lacked three of his legs and was of small size, while in the same web were several normal males of larger size. If one wished to list all monsters which lived to mature age one could write a whole book about them. But "hopeful monsters" of macroevolution, a humorous term proposed by Goldschmidt to emphasize their ability to survive, are not even real monsters. They differ from non-heritable monsters in that the changes which produce them involve only fundamental structural charac- ters, which are not necessarily a handicap and may have been even of advantage. These changes also differ from those produced in speciation by their magnitude and suddenness. Moreover they are restrictive in eliminating other changes of a similar nature in their own line of evolution and are therefore directive as pointed out by various writers. And in addition, evidence of their existence and of their independence from speciation is abundant, as may be seen in arachnids, at least in so far as the formation of suborders and families is concerned. I have dis- cussed the evolution of Arachnida in former contributions to our knowledge of their geological past, and of the principles which should govern the natural classification of that class. Here I want only to point out by the example of scorpions the difference between macroevolutionary and microevolutionary changes, and the fact that both are plainly preserved in fossil material. The macroevolutionary change on which I based the classification of Carboniferous scorpions involves the loss of thoracic sternites and the shifting into the place vacated by 252 BULLETIN : MUSEUM OF COMPAEATTVE ZOOLOGY 1 2 3 rl «"" <* j k- 0 (' 0 1(- *)/ OPISTHACANTHUS (RECENT) i sobuth us (carboniferous) PALAEOBUTHUS (PENNSYLVAN IAN) Fig. 1. The ventral surface of Opisthacanthus elatus (Gervais), a Recent scorpion of the family Scorpionidae. The arrangement of its coxae is the same as in the fossil family Eoscorpionidae. Fig. 2. The ventral surface of Palaeoluthus distinctus Petrunkevitch, a Pennsylvanian scorpion of the family Isobuthidae. The arrangement of its coxa? is the same as in Isobuthus, but the posterior edge of its abdominal sternites is straight as in Opisthacanthus. Fig. 3. The ventral surface of Isobuthus Tcralupensis (Th. and L.), a Carboniferous scorpion of the family Isobuthidae. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 253 EOSCORPIUS 5 ISOBUTHUS (PEN N SYLVAN I an) (CARBON I FE ROUS) HORMURUS 7 MICROLABIS (RECENT) (carboniferous) Fig. 4. The sterno-coxal region of Eoscorpius typicus Petrunkevitch, a Pennsylvanian scorpion of the family Eoscorpionidae. Notice that the arrangement of its coxae is the same as in the Recent Hormurus, while the structure of the combs resembles that of Isobutlius. Fig. 5. Isobutlius rakovnicensis Fritsch, a Carboniferous scorpion of the family Isobuthiclae. Notice that its disposition of coxae is the same as that of Hicrolabis, but its combs resemble those of Eoscorpius. Fig. 6. The sterno coxal region of Hormurus australasiae Fabricius, a Recent scorpion of the family Scorpionidae. Notice that the disposition of its coxae is the same as in Eoscorpius, while its combs resemble those of Micro labis. Fig. 7. Microlabis sternbergii (Corda), a Carboniferous scorpion of the family Isobuthidae. Notice that the disposition of its coxae is the same as that of Isobutlius, while its combs resemble those of Hormurus. Abbrevia- tions: C\, C2, Cz, d, coxa of leg 1, 2, 3 and 4; GO, genital opercula; -ST, sternum. 254 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY them of the second and third abdominal embryonic sternites represented in the adult by the genital opercula and the median piece of the combs. This change (Figs. 1-7) is possible in several different combinations, all of which are represented in the fossil material, but only two of which are shown in the figures and only one of which survived and produced all Kecent scorpions (Figs. 1, 4, 6). This is not the extreme modification. Yet all extinct modifications, even that of the original type in which only the first sternite was lost, survived long enough to produce other changes by speciation, namely the change in the shape of the posterior edge of the abdominal sternites (Figs. 1-3), and the change in the size and shape of the combs and in the number of their teeth (Figs. 4-7). The straight posterior edge of ab- dominal sternites is the original type found in all Recent scor- pions without exception and in almost all fossil scorpions. It is also found in the genus Palaeobuthus of the Family Isobuthidae to which the other two genera, Isobuthus and Microldbis, also belong. Isobuthus has bilobed sternites. In Microlabis the first and second sternites have a small median indentation, while the third sternite is clearly, though not deeply, bilobed. It seems to be an "ornamental" change rather than anything else. The combs are sense organs of unknown function, but being sense organs, they probably represent "adaptive" speciation. Eoscor- pius is a genus of the fossil family Eoscorpionidae from which all Recent families may be derived. Hormurus is a genus of the Recent family Scorpionidae. Figures 4 to 7 show similar, parallel changes in two different families, changes consisting in a reduc- tion of the number of teeth, and therefore independent of the macroevolutionary changes which gave rise to different families. In this connection I would like to point out that the chief difference between macroevolutionary and microevolutionary changes is not the magnitude of the change, but the nature of the trend which produced it and the fact that no dedifferentia- tion is required in the case of macroevolutionary changes. The magnitude in itself is a very relative concept. All evolutionary changes are of the saltatory, discontinuous type, even those which appear to be gradual. Moreover, a change in the genetic composition of chromosomes may produce an outwardly invisible change of considerable importance, such as color blindness for STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 255 example, but it is a saltatory change, nevertheless. When a series of small changes confronts us, we speak of them as gradual or transitional. When a change is considerable enough to make the difference glaring, we look for the missing connecting links, not realizing that the change in itself is the connecting link be- tween the parents with the original structure of their body and the offspring with the modified structure. This connection can be discovered only through a study of fundamental trends. Macroevolution leaves no evidence of its operation other than accomplished changes, nor any regressive links, for it is not subject to the law of precession. But it offers an indirect evi- dence of its existence by leaving unaffected all specific characters in the branches of the same line, characters subject to independ- ent speciation. Oparin (1953) has shown how self -perpetuating, living sub- stance of the nature of colloidal protein may have originated on earth. He has also explained how the original substance may have broken up into individual "droplets" and how these drop- lets may have formed clusters by coacervation. He also pointed out how such clusters may have become cellularized and thus given rise to primitive animals. Being a biochemist, Oparin left the problem of further evolution to zoologists and botanists. Whether or not his outline of the origin of life is correct in every detail, his approach to its solution is sound and we may accept his hypothesis as far as the formation of clusters. Finding fur- ther support in Haeckel's Biogenetic Law and in the principle of macroevolution, we may now attempt to get some insight into the phylogenetic evolution of animals. To begin with, instead of deriving the Metazoa from this or that class of Protozoa, as is usually done in phylogenetic trees, we may assume that in- dividual "droplets" developed a somewhat different genetic apparatus, remained single and produced the kingdom of Pro- tista, while clusters cellularized and some of them became the ancestors of the kingdom of plants, others changed into free-living moreae with a structure resembling that of a morula and became the ancestors of the kingdom of Metazoa. The next change must have resulted in the transformation of the morea first into a blastea and then into a gastraea, retaining the ability of asexual self-perpetuation, presumably by transverse division comparable 256 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY to stabilization in the scyphostoma larva of the medusae. Such a gastraea must have had the shape of an ellipsoid, a ciliated ectoderm, a nutritive endoderm and a mouth at one end of its longitudinal axis. As it still possessed the potentiality of eventu- ally developing into any of the future phyla, this stage may be properly called protogastraea. Its plan of symmetry must have been radial and its asexual reproduction made it still "immortal" in Weismann 's sense of this word as he applied it to the Protozoa. Now some molecular change in the genetic apparatus must have brought about a differentiation of the cells into reproductive ones and somatic ones, restricting the "immortality", or as I prefer to call it "the ability of self-perpetuation" to the repro- ductive cells, while the soma became the protective carrier of the reproductive cells, and the perishable residual of the organ- ism. Before this change, competition between individual proto- gastraeae was strictly intra-specific because there did not exist as yet any other species of Metazoa. But now, owing to the separation of the soma from the reproductive cells, a new type of competition arose. Under the influence of this competition the protogastraeae were changed into Metagastraeae with poten- tialities now restricted to the production of individual phyla. This macroevolutionary change which produced as many species of metagastraea as the number of future phyla, did not need to be synchronous in all cases. It may have occurred repeatedly as long as protogastraeae survived. But the complexity of the structure of animals found in the Cambrian makes it certain that the separation into phyla had been completed before that era. The difference between the various species of metagastraea at that time may have been only intracellular, in the structure of the genetic apparatus ; the outwardly visible specific differences were probably developed by speciation at a later period. It would be futile to speculate further as to how the present modes of reproduction involving complicated life cycles were evolved, because we have not enough evidence to do so. Such evidence could be assembled only by extensive study of comparative em- bryology and anatomy. For the present we may abide by the assumption of an independent origin of all phyla from specifically different metagastraeae, but all produced by the same type of protogastraeae. But even resigning myself to this limitation, I STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 257 believe that the thoughts which I have presented here in brief outline give us a much simpler picture of evolution, and are in closer agreement with the fossil evidence and with the available span of time that has passed since the first appearance of animal life than does the theory of evolution by speciation. REFERENCES Beer, G. R. db 1951. Embryos and ancestors. Revised edition. Oxford University Press. Delage, Y., and E. Herouard 1898. Traite de zoologie concrete. Vol. VIII, Les Procordes. Paris. DOBZHANSKY, T. 1951. Genetics and the origin of species. Columbia University Press. New York. Third edition. Garstang, W. 1922. The theory of recapitulation. Jour. Linn. Soc. London, Zoology, vol. 35. GOLDSCHMIDT, R. B. 1940. The material basis of evolution. Yale University Press. 1952. Evolution, as viewed by one geneticist. American Scientist, vol. 40. Haeckel, E. 1866. Generelle Morphologie der Organismen. Berlin. 1877. Biologische Studien: II Heft: Studien zur Gastrsea-Theorie. Jena. Hadzi, J. 1953. An attempt to reconstruct the system of animal classification. Systematic Zoology, vol. 2, no. 4. Lerner, M. M. 1950. Population genetics and animal improvement. Cambridge Univer- sity Press. Lunn, A., and J. B. S. Haldane 1935. Science and the supernatural. New York. Martin, E. 1880. Histoire des Monstres depuis l'antiquite' jusqu'a nos jours. Paris. 258 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY MAYR, E. 1942. Systematica and the origin of species. Columbia University Press. Mayr, E., E. G. Linslby and E. L. Usinger 1953. Methods and principles of systematic zoology. New York. Novtkoff, M. M. 1953. Regularity of form in organisms. Systematic zoology, vol. 2, no. 2. Oparin, A. I. 1953. Origin of life. Translated by S. Morgulis. New York. Petrunkevitch, A. 1952. Macroevolution and the fossil record of Arachnida. American Scientist, vol. 40. ScHINDEWOLF, O. H. 1950. Grundf ragen der Palaontologie. Stuttgart. Simpson, G. G. 1944. Tempo and mode in evolution. Columbia University Press. Stern, C. 1953. The geneticist 's analysis of material and the means of evolution. Scientific Monthly, vol. LXXVII. Vries, H. de 1901. Die Mutationen und die Mutationsperioden bei der Entstehung der Arten. Leipzig. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 112, No. 3 October, 1954 Status of Invertebrate Paleontology, 1953 X. Evolution of Late Paleozoic Invertebrates in Response to Major Oscillations of Shallow Seas. By Raymond C. Moore University of Kansas, Lawrence, Kansas INTRODUCTION The Problem Among the many factors which influence evolution of inverte- brate organisms, particularly those of benthonic habitat in a marine environment, are composite effects of broad oscillatory movements of shallow seas. When such seas transgress widely over portions of the conti- nental shelf, areas available for occupancy by marine organisms may be expanded vastly. During the enlargement of shallow seas and for a time after maximum expansion has been attained, competition for food and for a place of attachment is made easier. Concurrently, the opening of new territory for habitation by bottom-dwelling organisms may foster and then tend to stabi- lize adaptive changes. This should operate in roughly the same manner as alteration of human societies which has marked the spread of white men throughout the North American continent. Regression of a shallow sea introduces conditions opposite to those of transgression. Inevitably it shrinks space (Lebensraum) available to the shallow-water marine organisms. It causes crowding, and increases competition to survive. If retreat of the shallow sea is measured in terms of many hundreds or thou- sands of square miles, not only must populations be drastically reduced but the effects on those that continue to live may be profound. Weaker, less well adapted marine invertebrates are sure to be weeded out and only animals which can hold their own, perhaps expanding at the expense of those which disappear, belong to the remnant shallow-water fauna. It seems reasonable to construe times of marine regression as more significant in terms of accelerated evolution than times of marine transgression. 260 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Inadequacy of Evidence from Living Faunas For the most part a zoologist working on living marine organ- isms can only theorize about effects on life of the enlargement or reduction in area occupied by a given shallow sea. He cannot make surveys of bottom-dwelling populations in some chosen area in a manner serving to demonstrate change introduced by expansion or contraction of the selected seaway. Quantitatively significant advance or retreat of seas is much too slow to permit observation of its effects on bottom dwellers in any one region. Collection of reliable data would require millennia. Accordingly, the neo-zoologist can only compare assemblages of organisms in broad continental-shelf areas with those occurring in narrow shallow-sea belts, trying to interpret the origin of such differences as he finds. Although some of the differences may reflect evolu- tion at work, he cannot safely identify them, let alone evaluate them. Paleontological Investigation A paleontologist is able to study the record of life in three dimensions, for he can both observe areal distribution of organic assemblages represented by remains preserved in rock strata, and he can study the nature of closely adjacent different as- semblages higher or lower in the geologic column. The nature of this advantage is very well known. In most sections of conformable marine strata, fossils collected from two or more contiguous layers are properly inferred to represent samples of successive populations in an area continu- ously occupied by the sea. Whether the collections are derived from a former biocoenose or give evidence of a thanatocoenose, for present purposes makes little difference. The small fraction of the paleontological record reveals at most the nature of slowly changing conditions that reflect lapse of geologic time (excluding effects of possible physical changes in environment). Study of fossil collections from such conformable strata does not supply information on effects of advancing or retreating seas. If the faunal assemblage in one layer is known to represent part of an extremely wide-spread shallow-sea deposit, whereas a not-far-distant higher or lower assemblage represents a similar STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 261 fraction of organisms living in a much restricted seaway, this would begin to meet the conditions of the problem proposed for investigation. Unfortunately, it is almost impossible to find geo- logical conditions which can be so interpreted reliably. Evidence from Cyclic Deposits Another approach is the comparison of successive widely dis- tributed marine faunas, each of which is independent, repre- senting a shallow-sea transgression wholly distinct from others. Such independence may be demonstrated by the presence of a disconformity, indicating nondeposition and possibly erosion of previously formed deposits, which intervenes between the com- pared fossiliferous marine strata. It may be proved also by widespread occurrence of nonmarine deposits belonging to a time of considerable marine regression, where these subaerially formed beds are found between two sets of shallow-sea deposits. Late Paleozoic sedimentary rocks of the northern midcontinent region in the United States offer ideal examples of such cyclic sedimentation. They include many units consisting of very widely distributed marine strata that record extensive inundation of the continental platform, and they include equally widespread nonmarine deposits that occur between the marine units. The beds which were laid down on land incontrovertibly prove ab- sence of the sea at the time and place where they were formed. If fossil invertebrates found in one of these independent wide- spread marine deposits are lineal descendants of similar organ- isms found in a next-lower broadly distributed marine band, then the differences between them, if any, should measure the effects of sea retreat and re-advance between the times of exten- sive marine sedimentation. The examination of successive assemblages of marine fossils belonging to cyclic deposits in Kansas is comparable to study of a series of individual pictures ("frames") in a motion picture film. Each such picture furnishes a single glimpse of an ever- moving subject separated from the slightly different next picture, by blank film. The blanks (nonmarine record) of the strati- graphic succession are much longer than those of ordinary film, however. The cyclic aspects of Pennsylvanian strata and to a lesser 261! BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Phases SCALE IN FEET 300 25d— 200 — 150— 100— 50— jFloranee Is. BLUE SPRINGS SH. KINNEY. LS. WYMORE SH SPEISER SH EASLYCR.SH STEARNS SH Morrill Is — Floreno sh, •^Vr'Wcottonwooa: Is. ESKRIDGE SH. 1 | ' - Nevo !». Fig. 1. Composite geologic section of part of the Lower Permian succession in Kansas with accompanying graph showing nonmarine units and inferred depth of inundation in marine units (after ML K. Elias). STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 263 Fig. 2. Paleogeographic sketch maps showing inferred approximate loca- tion of seaways during (A) one of the several times of relative continental emergence (when Eskridge deposits were made in Kansas) and (B) the immediately following rather considerable shallow-sea transgression (when fusulinid-bearing Beattie strata were laid down in Kansas). 264 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY extent of Lower Permian deposits in the northern midcontinent region have been described in some detail (Condra and Upp, 1931; Jewett, 1933; Moore, 1936-1953; Moore and Thompson, 1949; Moore et al., 1951). Also the nature of faunal associations belonging to different subdivisions of the marine part of cyclic deposits has been delineated and interpreted (Elias, 1937 ; Moore, 1936, 1949, 1950, 1953). Therefore, discussion here is superfluous, but a part of the Lower Permian succession may be shown graphi- cally (Fig. 1). This is chosen because fusulinids derived from these rocks are employed in a later part of this paper as material for investigation of evolutionary changes. A method of illustrating the geographic magnitude of shallow- sea oscillations in Pennsylvanian and Permian time in the central United States is by means of paleogeographic maps. Figure 2 presents such maps for two adjoining segments of the Lower Permian succession in the midcontinent region. Taken together, they represent only a very small fraction of Permian time, for Figure 2A represents geographic conditions perhaps only 100,000 years before those depicted in Figure 2B. In early Permian time, the areas successively flooded and laid bare in the central United States were much smaller than in the preceding Pennsylvanian period. Although outlines of the seas shown in Figure 2 are conjectural, they emphasize the point that the marine oscillations affected many thousands of square miles. They should have had an influence on evolution of invertebrate animals living in the shallow seas. EVIDENCE FURNISHED BY SELECTED INVERTEBRATE GROUPS General Statement For the purpose of studying effects of broad shallow-sea oscilla- tions on the evolution of bottom-dwelling invertebrates, any group of them found preserved in a number of successive distinct cycles is appropriate for investigation. Some assemblages are likely to be better than others, however. The chief guides in choosing materials for study are : (1) availability of a reasonably large number of specimens belonging to the group in each of the compared marine deposits, (2) adequacy of fossil collections from precisely known stratigraphic positions, (3) thoroughness STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 265 of investigations leading to discrimination of significant morpho- logical and taxonomic interrelations, and (4) distinctness of differences that may be judged ascribable to evolutionary change. A fairly large sample of an invertebrate group in a specified shallow-sea deposit is required in order to appraise characters that possess value in defining evolutionary trends. Such a sample may be composite, coming from many localities, but it must be representative. Also, each of the two or more compared inde- pendent marine units needs to supply materials of this sort, for otherwise comparison leads to dubious conclusions or it is impos- sible. Genera, subgenera, and possibly species that exhibit at least moderately extended stratigraphic range are the only taxo- nomic units which can be considered suitable for this study. Adequacy of fossil collections is a self-evident requirement. Evidence that exists in the field is not usable until it is brought to the laboratory and subjected to careful analysis. In early geological studies, knowledge of detailed stratigraphic relations was lacking generally and its value was unappreciated. There- fore, most fossil collections from late Paleozoic strata in the northern midcontinent region which were made before 1920 have greatly impaired usefulness unless locality records permit accurate identification of the exact source of the fossils. Comprehensive research of so-called monographic sort is needed before almost any roup of Pennsylvanian or Permian in- vertebrates can be studied usefully with the object of recognizing evolutionary distinctions and trends. Seemingly, all kinds of benthonic organisms living in shallow seas of the North American continental platform had attained remarkable stability in late Paleozoic time and although faunas are varied and some of them large, their components mostly persist. As result, a Lower Per- mian fauna may resemble a Lower Pennsylvanian one so closely in a majority of its elements as to be distinguishable only by specialists. Few invertebrate groups have yet been examined thoroughly. It is recognized universally that invertebrate groups vary enormously, both within themselves and between time-defined segments within almost any single group, in the clearness and rapidity of evolutionary change which is discernible in their geologic history. Linguloid brachiopods furnish examples of un- 266 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Form Ratio 1.2 I.I 0.9 M. copha (Atokan) M. lepta (Desmoinesian) M. wyomingensis—~- (Missourian) M. miopetina — - 10 (M.Virgilian) M. pliopetina (L. Permian) M. copei — »— (L. Permian) M.arbala (Missourian) — M glossoidea (Virgilian) M petina (U. Virgilian) I i i_ 2.4 2.0 1.6 1.2 Ratio Lfi:L<* 0.8 0.4 Fig. 3. Evolutionary trends of myalinid shells represented by species from Pennsylvanian and Lower Permian strata of the midcontinent region. The form ratio (shell length to shell height) of species is plotted against the ratio of angle beta to angle alpha (inserted diagram A). Taking account of stratigraphic occurrence, two main trends seem to be derived from Myalina lepta of Desmoinesian age (data from Newell 1942). STATUS OF INVERTEBRATE PALEONTOLOGY. 195:$ 267 L$- . copha --, Atokan 130" 115 100° /"V Desmoinesian '.* —jLepta^J f.% ;-•., \ \/ \ / wyo'mLn'g'ensis -Desm.-Virg.~ / \y \ '. yglossbidea- L.Virgilian V. \ / N\ M. Virgilian « • J> ., \ mtopetlna arbala-M\ssb\ir'\Qn Fig. 4. Eange in values (shown by quadrilaterals) and mean values (black dots) of the alpha and beta angles of species of Myalina from Penn- sylvanian and Lower Permian rocks of the midcontinent region (data from Newell, 1942). 268 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY detectable alteration in the course of what must be ten-million- year intervals, whereas the complexly organized crinoids display profound and generally rapid evolutionary change. A satisfac- tory rating of late Paleozoic invertebrate groups according to Fig. 5. Inferred phylogenetic relations of some Pennsylvanian and Lower Permian myalinid species from the midcontinent region, based largely on data illustrated in Figures 3 and 4. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 269 distinctness of numerous evolutionary modifications is not now possible, although ammonoids, fusulinids, and crinoids surely would rank high. The fusulinids are a supremely useful group because of their almost incredible abundance, wide geographic distribution, occurrence in nearly every cycle of Pennsylvanian and Permian deposits, and considerable range of morphological differentiation. In the study here reported, attention was directed to chonetid brachiopods, fistuliporoid bryozoans, myalinid and pectinoid clams, and Lower Permian fusulinids. This choice was based on the availability of monographic studies completed or in prog- ress. Only the myalinids and fusulinids, however, are discussed in this paper. Myalinid Pelecypods Pelecypods of the group called myalinids are widely distrib- uted in late Paleozoic near-shore marine deposits of the northern midcontinent region. They are more or less common in nearly every cycle. A comprehensive study of these shells, based on large collections from precisely controlled stratigraphic positions and from very numerous localities, has been made by Newell (1942). His work has led to recognition of ten generic or sub- generic groups in the family Myalinidae and to differentiation of some 30 species of Myalina. Among characters found by Newell to be most useful for classi- fication and recognition of evolutionary trends are shape of the shell, including especially the angle between the umbonal ridge and the hinge line (termed alpha angle) and the similar angle between hinge line and growth lines at the posterior mar- gin (termed beta angle). In addition to a tendency of shells to become larger and thicker in the course of evolution, unidirec- tional trends in the nature of these angles is observed. Figures 3 to 5 furnish graphic indication of characters of some species belonging to the subgenus Myalina {Myalina). Evolutionary modifications can be detected and reasonable inferences as to genetic relationships among species can be drawn. The assigned stratigraphic range of most species, however, is too great to allow recognition of differences between specimens found in marine parts of the successive Pennsylvanian and Permian cyclic deposits. 270 BULEETIN : MUSEUM OF COMPARATTVB ZOOLOGY Fig. 6. Generalized section of Lower Permian deposits in Kansas showing longitudinal sections of representative fusulinids from various shallow marine strata (modified from Thompson, 1954). STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 271 Fig. 7. Generalized section of Lower Permian rocks of north-central Texas showing longitudinal sections of representative fusulinids from shallow marine strata. Heavy vertical bars represent nonmarine red deposits (modi- fied from Thompson, 1954). 272 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Fusulinids Study of Lower Permian fusulinids as markers of evolution in response to extensive marine oscillations was undertaken be- cause of the completion recently of comprehensive research by M. L. Thompson (1954) on these fossils. Primarily based on col- lections from Kansas (40 localities) and Texas (47 localities) representing every discovered fusulinid-bearing zone in the Wolf- campian part of the rock column, Thompson's investigation also included very numerous specimens from Oklahoma, New Mexico, Nevada, Utah, Arizona, and California. One new fusulinid hori- zon (Wreford limestone) has been discovered in Kansas since the time of Thompson's work (D. E. Hattin, personal communi- cation). The monograph by Thompson describes and illustrates 56 species (42 new) which are distributed among 11 genera. The stratigraphic occurrence of all forms is recorded precisely, and accordingly a good source of information is offered for investiga- tion of the sort here discussed. Cyclic deposits containing fossiliferous marine strata separated by generally unfossiliferous nonmarine beds characterize the Lower Permian succession in Kansas and Nebraska, parts of Oklahoma, and north-central Texas. They may occur in other states also but my attention in the present study has been re- stricted to the part of Thompson's paper concerned with Kansas- to-Texas fusulinids. These include species which are identified both in the north and south and those recorded as yet only from the Kansas region or from north-central Texas. Figures 6 and 7 show the Lower Permian succession and stratigraphic position of fusulinid-bearing rocks in Kansas and Texas, respectively. In order to obtain a common denominator for the Kansas and Texas sections, identifying as exactly as possible the equivalent marine rock units in each, two paleontological main tie points are employed. The lower one consists of fusulinids near the base of the northern and southern successions; they include Triticites confertus Thompson and Dunbar inella eoextenta Thompson, found in both, associated with slightly different but closely related fusulinids {T. pointensis Thompson, in Kansas, and D. en tenia Thompson, in Texas) and other species (Meekopora prosseri Ul- rich, M. mollis Moore and Dudley, and distinctive abundant crinoid ossicles including the so-far-as-known very short-ranged STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 273 Triceracrinus Bramlette). The higher tie point is located at the horizon of the Goldbusk limestone in Texas and Florence lime- stone in Kansas, each of which contains an assemblage of fusu- EXPLANATION 9"! Coleman EH opecies common to Kansas $ and Texas I I I 5pecies only known in Kansas or Texas ^M Marine deposits Ejj Nonmarine deposits 100 ft. Oft. £ Florence s- -*■ *tSW- corrtptexa, I Goldbusk % % S3-^er^lUei D8-wetkererusls I <5'2b- ernxLcixt-ta. D7c-koscnntcLrun, p ft Neva I? G/enrock -P.- Foraker e|e Americus -=- .i. Five Point -=- v— v/c - tcowrw n (5 U»D7b- obesa *■ -D7cl - tumida> \52-turtcL [ 1'T6-roc/csns6s 67- casnpa. - = Stbckwether T7- creekensis = \ Sl-bngissi/noulea | Camp Creek | Saddle Creek -{Waldnp3 -g Waldrip 2 -I Waldrip I D3-ex.ten.ta KT3- direct us D€-glenertsis I \nr l l T ¥ T5- ventrlcosus KDS-!uqhesensis\ T ^-nxeeki D4-a/7t£riccma \ D2- eoextenta. —. . , • T2-con,fertus 77- potfiterbSLs XD/- fi.ven.sis Fig. 8. Stratigraphic distribution of selected fusulinid species in the Kansas and Texas Lower Permian sections. The Florence and Goldbusk fusulinid-bearing strata are judged to be closely contemporaneous for, in addition to Schwagerina complexa, they contain Pseudoschwagerina texana and Pseudofusulina? moranensis, all three species being found in no other rocks of north central Texas or the Kansas region. Other species found near the base of the Permian deposits in Kansas and Texas indicate age equivalence of these rocks. Thus, an integrated scale in terms of fusulinids can be constructed. (T1-T7, species of Triticites; D1-D8 species of Dunbar- inella; S1-S4 and S'l-S'4, species of two inferred lineages of Schwagerina). 274 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY linids found nowhere else, although one of the species occurs also in Lower Permian rocks of western Texas and New Mexico. The Goldbusk and Florence species are distinctive forms named Pseudoschwagerina texana Dunbar and Skinner (also in Wolf- camp rocks of the Glass Mountains and Hueco limestone of the Sierra Diablo and Hueco Mountains), Schwagerina complexa Thompson and Pseudofusulina? moranensis Thompson. The strata containing this fusulinid assemblage must be very nearly equivalent in age, if not precisely so. A comparison of the Kansas and Texas sections showing placement of fusulinid species studied in preparing this paper is given in Figure 8 and a tabulation of marine units which tentatively are considered to have the same age follows. Correlation of Marine Lower Permian Stratigraphic Units in Kansas and North-central Texas (Units containing identical species of fusulinids marked by **; other fusulinid-bearing units marked by *) Kansas Texas Nolans limestone "Coleman Junction limestone Winfield limestone Sedwick limestone **Florence limestone **Goldbusk limestone *Wreford limestone *Camp Colorado limestone •Beattie limestone *Stockwether limestone **Neva limestone **Camp Creek shale (and lime- stone) **Glenrock limestone ** Saddle Creek limestone **Foraker limestone (Hughes **Waldrip shale (No. 3 limestone) Creek) **Foraker limestone (Americus) **Waldrip shale (No. 2 limestone) **Fivepoint limestone **Waldrip shale (No. 1 limestone) Comparison of fusulinid species belonging to any given generic assemblage with the object of distinguishing evolutionary trends naturally must take account first of relative age of the chosen fossils. Older forms may have given rise to younger ones but not the reverse. Morphological features must be examined in order to discover resemblances and differences. Average size of STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 275 individuals is not likely to be significant, although this may be a specific character along with others. Nearly all descriptions of fusulinid species are accompanied by more or less extensive tables of measurements which furnish record of observations concerning not only the whole shell but the proloculus and each successive volution. These data constitute a statistical delimita- tion of various morphological elements but it is difficult indeed to use them directly in making comparisons and in learning what may be most significant. Graphic analysis of such information rarely is given by authors and therefore a good deal of time was spent by me in trying to plot useful graphs. Some showed average measurements or computations and others showed values reported for individual specimens, thus indicating range of varia- tions. Ratios of two selected characters recorded by measure- ments can be plotted against a third set of measurements or against other ratios. Some work along this line seemed to be worth while but most of it was discarded. Triticites A somewhat arbitrarily chosen but seemingly useful starting point for the comparative study of fusulinid species consists in plotting average measurements for the diameter and wall thick- ness of the proloculus. Among Lower Permian species from Kansas and Texas described by Thompson, the proloculus diam- eter is found to range from less than 100 to nearly 250 microns and the wall thickness from 9 to 25 microns (Fig. 9). If the stratigraphically lowest species is postulated to be ancestral (or alternatively, closely similar to ancestral) to younger species found higher in the section of Kansas, Texas, or both, a line or lines drawn from the plotted point for the oldest species to points representing other species indicates the nature of evolutionary trends. The premise that evolutionary trends are consistent, rather than erratically reversible, underlies the conclusion that not all observed species belong to a single chain between the presumed progenitor and its youngest known descendant. Rather, the directions of evolutionary alteration may diverge from a parent stock so that graphic representation of inferred relationships has a branched pattern, as in Figure 9. It is obvious that an inherent assumption in the procedure 276 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY just described is that all of the forms considered are actually interrelated, all being derived from a source consisting of the oldest species of the group. The validity of this assumption is open to challenge, since one or more of the studied species may be immigrants to the Kansas or Texas areas, being derivatives of stock that resided in some "foreign" part of the Permian shallow seas such as New Mexico, Arizona, or Utah. Also, if Triticites rockensis Thompson from the Glenrock limestone of Kansas) is a lineal descendant of T. point ensis Thompson or T. confertus Thompson, found in stratigraphically lower marine TS T7 verttricosLLS creetcertsLs I coriferbus rockensis 77 poirttertsis 100 Triticites Proloculus 150 Diameter in yx. 200 250 Fig. 9. Proloculus of species of Triticites plotted according to diameter and wall thickness. Taking account of stratigraphic occurrence, three dif- ferent lineages are suggested. The letters T1-T7 inclusive correspond to those shown in Figure 8. deposits of Kansas, and if the line of descent to T. rockensis does not include T. ventricosus (Meek and Hayden) or T. meeki (Mol- ler), which occur in Kansas rocks older than the Glenrock lime- stone, a question is raised concerning the identity and place of occurrence of immediate forerunners of T. rockensis. Figure 10 illustrates morphological characters (form ratio, height of chambers, thickness of spirotheea, cumulative number of septa, and tunnel angle) of the Lower Permian species of Triticites from Kansas and Texas as measured at the fifth volu- tion, lines being drawn between plotted points for the individual STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 277 Fig. 10. Morphological features of the fifth whorl of Lower Permian species of Triticites arranged according to lineages suggested by Figure 9 (A, form ratio; B, height of chambers; C, thickness of spirotheca; D, number of septa, cumulative to end of fifth whorl; E, tunnel angle: 278 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY species in a manner corresponding; to that in Figure 9. Graphs which were constructed similarly for all volutions do not differ enough from the one presented here to warrant publication of them. Average values for form ratio (Fig. 10A) are almost the same in Triticites pointensis and T. rockensis but show ups and downs along the lines leading- to other species. This suggests that form 140 Diameter in jj. Fig. 11. Proloculus of Lower Permian species of Dunbarinella showing diameter and wall thickness. Taking account of stratigraphie occurrence, three lineages of these species are suggested. The letters D1-D8 inclusive correspond to those in Figure 8. ratio is not a very reliable indicator of evolutionary change because it moves in reverse directions or else the compared species are not actually a genetically related series. The same inference is appropriate in trying to interpret the graph of tunnel angles (Fig. 10E). On the other hand, comparison of values for cham- ber height, thickness of spirotheca, and number of septa (Fig. 10B-D) plausibly conform to a pattern of evolution having consistent trends. STATUS OF INVERTEBRATE PALEONTOLOGY, 1953 279 Dunbarinella Studies of the Kansas and Texas species of Dunbarinella de- scribed by Thompson (1954) were made in the same way as for Triticites. Graphic representation of proloculus characters (Fig. 11) suggests existence of four lines of development, but attention needs to be called to the record of D. eoextenta Thompson asso- Fig. 12. Form ratios of Lower Permian species of Dunbarinella showing variation during growth. ciated with D. fivensis Thompson in the Pivepoint limestone of Kansas and the presence of both D. eoextenta and D. extenta Thompson in rocks of the same or nearly identical age in north- central Texas. Thus, these three species which have notable dif- ferences in character of the proloculus are contemporaneous at least in part ; because they have other features that indicate close 280 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY relationship, it is probable that they are slightly divergent de- rivatives of a late Pennsylvanian ancestor. In wall thickness and diameter of the proloculus, D. extenta rivals the stratigraphi- cally higher D. glenensis Thompson and D. obesa (Beede). In order to illustrate ontogenetic changes, as well as the range of variation among the studied species of Dunbarinella, the values of computed form ratios for each volution have been plotted (Fig. 12). They show that conclusions based on comparison of a single growth stage may not be reliable. Since similar graphs of other characters mostly demonstrate acceleration or retardation in the rate of change from volution to volution, they offer little indica- tion that successive growth stages furnish record of the course of phylogeny. Diagrams prepared in the manner of Figure 12 seem to be useful for summarizing taxonomic aspects of morpho- logical features but they fail to indicate directly the nature of evolutionary trends which otherwise may be discernible in the group of species. Comparison of morphological characters of species of Dun- barinella, based on measurements of the sixth volution, which is nearest to the adult stage as far as data for all species are available, is shown in Figure 13. The line connecting D. fivensis with D. americana Thompson and D. hughesensis Thompson and that joining D. fivensis with D. extenta and D. glenensis are both consistent with evolution in fairly uniform though divergent di- rections. On the other hand, indicated trends from D. fivensis through D. eoextenta to D. tumida (Skinner) and 0. wether -ensis Thompson, and less clearly to D. obesa (according to the pattern suggested by Figure 11) seem to be anomalous. D. tumida and D. obesa, which occur with D. koschmanni (Skinner) (Figs. 8, 12, 13) in the Neva limestone of Kansas and northern Oklahoma, are conceivably derivatives of the older D. eoextenta but not from the also older D. americana and D. hughesensis group or the D. extenta and D. glenensis group. That some of these species are interlopers of unknown origin, rather than descendants of in- digenous midcontinent species, is very possible. The wide range in values of measurements representing most morphological ele- ments is a noteworthy feature of these fusulinids, leading one to think that recognition of definite evolutionary trends should be relatively easy. Such is not true. STATUS OP INVERTEBRATE PALEONTOLOGY, 1953 281 t3 si SSI i?| *■« Q..3 Q^ Q| 1 Fig. 13. Morphological features of Lower Permian species of Dun- barinella with differentiation of inferred lineages as shown in Figure 11. A, indicates diameter of the proloculus (in microns). B to F represent characters of the sixth whorl (B, form ratio; C, height of chambers, in microns; D, thickness of spirotheca, in microns; E, number of septa to end of sixth whorl; F, tunnel angle, in degrees). (Data from Thompson, 1954.) 282 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Schwagerina Another Lower Permian genus which is represented by nu- merous species (10 in the midcontinent area and 9 others from trans-Pecos Texas and country farther west) is Schwagerina. Characters of the proloculus (Pig. 14) suggest division of the midcontinent forms into two sharply distinguished groups. These respectively comprise 8. campa Thompson, 8. jewetti Thompson, S. vervillei Thompson, S. emaciata (Beede), and S. complexa Thompson in one assemblage and S. longissimoidea (Beede), S. Wall Thickness 30 Coleman Is. (Tex) 34 cotemanl Gold busk Is. (Tex.) Florence Is. (Kan.) S'4 compLexa Morrill Js. (Kan.) ^ vervUlec. ~j£uu*ata. Cottonwood Is. (Kan.) S'2b Red Erie's. **&***£ (Kan.) I Cottonwood Is. (Kan.) .100. I .120- S3 campensis Camp Creek sh. (Tex) 62 turki Foraker fm. (OklaJ