1 UNIVERSITY OF ILLINOIS LIBRARY AT URBANACHAMPAIGN GEOLOGY The person charging this material is re- sponsible for its return to the library from which it was withdrawn on or before the Latest Date stamped below. Theft, mutilation, and underlining of book, ar. reasons for disciplinary action and may result In dismissal from the University. To renew call Telephone Center, 333-840O UNIVERS.TY OF IIUNOIS UBRARY AT URBANA-CHAMPAIGN i* M apr 2&mr «> * JUL 0*3 1991 L161— O-1096 c FIELDIANA: GEOLOGY A Continuation of the GEOLOGICAL SERIES of FIELD MUSEUM OF NATURAL HISTORY VOLUME 20 /.S NATURAL < HISTORY ^1 FIELD MUSEUM OF NATURAL HISTORY CHICAGO, U.S.A. urniAPV I IDB1I ^T^ XT ' v I/.3L0 TABLE OF CONTENTS PAGE 1. Revised Classification of Pteraspididae with Description of New Forms from Wyoming. By Robert H. Denison 1 2. Catalogue of Type and Referred Specimens of Fossil Echinozoa (Echi- noidea and Cyclocystoidea) in Field Museum of Natural History. By- Julia Golden and Matthew H. Nitecki 43 3. Catalogue of Type and Referred Specimens of Echinozoa (Edrioaste- roidea) in Field Museum of Natural History. By Julia Golden and Matthew H. Nitecki 53 4. Catalogue of Type and Referred Specimens of Fossil Stelleroidea in Field Museum of Natural History. By Julia Golden and Matthew H. Nitecki 65 5. Redescription of Ischadites clrodi (S. A. Miller, 1892) a Lower De- vonian Receptaculitid. By Matthew H. Nitecki 77 ('). A Strange Arthropod from the Mazon Creek of Illinois and the Trans Permo-Triassic Merostomoidea (Trilobitoidea.) By Frederick R. Schram. 85 ^FIELDIANA Geology Published by Field Museum of Natural History Volume 20, No. 1 September 18, 1970 Revised Classification of Pteraspididae with Description of New Forms from Wyoming Robert H. Denison Curator of Fossil Fishes Field Museum of Natural History In 1967 in connection with the description of a new pteraspid from Utah, I reviewed and revised the classification of this family. Now, in studying extensive collections in Field Museum from the Bighorn Mountains of Wyoming, I have felt the necessity to review this classification again, and particularly to re-evaluate the charac- ters which have been used to distinguish genera and species. The classification that I have arrived at is essentially the same as that of my 1967 paper, and I will repeat it with minor changes, additions, and further justifications. My own studies have been largely re- stricted to North American material, and so I have attempted no detailed review of European forms. Most of the additions concern the genus Protaspis and its relatives, which are the common pteras- pids of the western United States. Important new information about the structure of these forms has been provided by the material from the Bighorn Mountains of Wyoming, and by a restudy of old mate- rial from Utah and from Beartooth Butte, Wyoming. I have in- cluded all of the genera in one family, Pteraspididae, and feel no need to elevate it to the rank of suborder or order, as did Tarlo (1962, p. 265), Obruchev (1964, p. 63), and Stensio (1964, p. 361), although this is necessary if Doryaspis is placed in a family of its own. As stated below, I feel that the purposes of classification are satisfied equally well by isolating Doryaspis from other pteraspids in a sep- arate subfamily, which requires only the use of another subfamily, Pteraspidinae, to include typical pteraspids. Library of Congress Catalog Card Number: 71-136029 Publication 1106 1 2 FIELDIANA: GEOLOGY, VOLUME 20 I have given particular attention in this study to the structure of the branchial openings and the rostral plate. In the classifications of Stensio (1958, 1964) and Tarlo (1961), the rostral structure has been used as the primary character for distinguishing genera. Among the pteraspids from the western United States there are also impor- tant differences in the rostrum, yet, in spite of this, I feel that the majority of the species are closely related, as indicated by their broad, highly vaulted dorsal shields, posterior branchial openings, and usually scale-like dorsal spines. This suggests that many rostral characters have been independently acquired in the American species as a result of an evolution paralleling that found in Europe. For this reason I have given the rostral characters less weight, and have, for example, included in the subgenus Protaspis species with rostra suf- ficiently different to distinguish the genera of Stensio and Tarlo. The posterior branchial openings, which particularly characterize Pro- taspis, occur in certain European species. I have not studied these forms, so cannot come to any definite conclusions as to whether they are closely related to American species or the result of parallel evolu- tion. For this reason I have provisionally kept them subgenerically distinct. Specimens whose catalogue numbers have the prefix PF are in the collection of Field Museum of Natural History. The drawings have been made by Museum staff artist, Dr. Tibor Perenyi, many from my sketches. Family Pteraspididae Dorsal shield composed of rostral and pineal plates, a dorsal disc with a dorsal spine attached to its posterior part, as well" as paired orbital, branchial, and cornual plates, the latter occasionally reduced, absent, or two pairs. Ventral shield formed by a large ventral disc, variably developed paired lateral, oral, and sometimes postoral plates, and in Doryaspis a pseud orostrum. Branchial openings more or less posteriorly placed at or near the lateral margins of the dorsal shield, typically at the posterior ends of the branchial plates, and commonly bounded posteriorly by cornual plates. Two or three pairs of sensory canals radiate from the center of ossification of the dorsal disc. Plates and scales consist of a tangentially laminated basal layer attaining only moderate thickness, a middle layer characterized by relatively large chambers, and a superficial layer consisting of narrow dentine ridges, or in some forms of tubercles. DENISON: PTERASPIDIDAE Subfamily Pteraspidinae Rostral plate extending anteriorly over the mouth; no pseudo- rostrum attached to the ventral shield. il! brp— J+Ul W\ 1 1& /// / / \\ bro- £&l_dd— . >%-i' — co Fig. 1. Region of branchial opening of Protopteraspis primaeva (Kiaer), based largely on PF 1078 (X 7). A, ventral view, right side; B, dorsal view, left side. bro, branchial opening; brp, branchial plate; co, cornual plate; dd, dorsal disc. Protopteraspis Leriche, 1924 Type species. — Protopteraspis gosseleti Leriche. Protopteraspis Leriche, 1924, Bull. Soc. Beige Geol., Pal., Hydr., 33, p. 149 (footnote), (subgenus of Pteraspis). Simopteraspis White, 1950, Bull. Brit. Mus. (Nat. Hist.), Geol., 1 , p. 76 (sub- genus of Pteraspis). Zascinaspis Stensio, 1958, Traits de Zool., 13, fasc. 1, p. 255; Tarlo, 1961, Acta Palaeont. Polonica, 6, pp. 378-379. Diagnosis. — Small to medium-sized pteraspids. The rostrum short and bluntly rounded, with its ventral preoral surface short and lacking a preoral field.1 Branchial openings well in advance of the postero-lateral corners of the dorsal shield, and bounded by branchial and cornual plates (fig. 1). 1 1 have adopted the following terminology for the ventral surface of the ros- trum: Ventral preoral surface, the entire ventral surface of the rostrum anterior to the mouth, and including in some genera the preoral field, a clearly defined area elevated above the rest of the rostral surface and lacking ornamentation of dentine ridges; and the ascending lamella, the posterior part of the ventral preoral surface which rises dorsally anterior to the mouth. 4 FIELDIANA: GEOLOGY, VOLUME 20 Discussion. — Although in 1967 I followed Leriche and White in ranking this as a subgenus, I now feel warranted in elevating it (as well as subgenera Pteraspis and Althaspis) to generic rank. White (1961, pp. 257-258) prefers to use Simopteraspis, and to retain Pro- topteraspis for P. gosseleti alone because the preoral region of the latter is unknown. However, P. gosseleti and the type species of Simopteraspis, Pteraspis leathensis, are so similar in other respects that at one time White (1950, p. 82) believed that they might turn out to be conspecific, so I feel justified in placing them in the same genus. The species referred to Simopteraspis by White in 1950 are small, primitive forms with blunt rostra and typical branchial openings (fig. 1). In addition, each has a small pineal plate isolated from the orbitals, lacks postoral plates, and has the posterior extensions of the supraorbital canals (pineal canals of Stensio) forming a V-shaped loop on the dorsal disc. I refer also to Protopteraspis the species placed in Zascinaspis by Stensio in 1958 and Tarlo in 1961 (Brachi- pteraspis heintzi Brotzen, B. grossi Brotzen, and Pteraspis carmani Denison), which differ mainly in their larger size and in characters correlated with size, such as the relatively larger ventral preoral surfaces, and the wider pineal plates in contact with the orbitals. The presence of postoral plates in P. carmani may be related to size also. The extension of the supraorbital canals onto the dorsal disc is not known in the species referred to Zascinaspis, and may be char- acteristic only of small primitive species; it is retained in Althaspis whitei (Denison, 1955, fig. 110A), and in a modified form in one speci- men of Pteraspis rostrata (White, 1935, fig. 68). Pteraspis Kner, 1847 Type species. — Cephalaspis rostratus Agassiz. Pteraspis Kner, 1847, Haidinger's Naturw. Abh., 1 , p. 165. Scaphaspis Lankester, 1865, Rept. Brit. Assoc. Adv. Sci., Trans. Sect., 1864, p. 58 (in part). Podolaspis Zych, 1931, Fauna Ryb Dewonu i Downtonu Podola, p. 89. Brachipteraspis Brotzen, 1936, Ark. Zool., 28A, no. 22, pp. 31-35. Mylopteraspis Stensio, 1958, Traits de Zool., 13, fasc. 1, p. 255. Parapteraspis Stensio, 1958, Traits de Zool., 13, fasc. 1, p. 265. Plesiopteraspis Stensio, 1958, Traite de Zool., 13, fasc. 1, p. 265. Diagnosis. — Moderate-sized pteraspids. The rostrum somewhat elongate and with a preoral field developed on at least the posterior part of its ventral preoral surface. Branchial openings well in ad- DENISON: PTERASPIDIDAE brp • ro — JL Fig. 2. Left branchial opening of Pteraspis 1 priscillae (Denison), dorsal view, PF 867 ( X 5). bro, branchial opening; brp, branchial plate; dd, dorsal disc. vance of the postero-lateral corners of the dorsal shield, and bounded by branchial and well-developed cornual plates. Discussion. — As used here, the genus Pteraspis is restricted to the species referred to the subgenus Pteraspis in my 1967 paper. Its most important distinguishing feature is the presence on the ventral surface of the rostrum of a preoral field lacking dentine ridges and probably representing the locus of sensory organs. Its branchial openings are basically similar to those of Protopteraspis (fig. 1). White (1961, p. 270) has questioned the reference of Protaspis priscillae Denison (1953, pp. 348-349, figs. 82-83) to Protaspis. Further preparation and study of the type and only specimen has shown that he was correct, for there appear to be slender branchial openings (fig. 2, bro) well in advance of the posterior corners of the shield with the usual relations of those of Pteraspis, except that dis- tinct cornual plates cannot be identified. Little of the ventral surface 6 FIELDIANA: GEOLOGY, VOLUME 20 of the rostrum can be seen, but there is a median unornamented area that is presumably a preoral field, so this species can be referred to Pteraspis or possibly to Rhinopteraspis. It shows resemblances to another elongate, slender species, Pteraspis dewalquei Fraipont, as figured by White (1960, fig. 3), though the ventral surface of the rostrum of the latter is unknown, and its generic reference is also uncertain. Althaspis Zych, 1931 Type species. — Pteraspis elongata Zych (non Alth) = Althaspis samsonowiczi Tarlo. Podolaspis (Althaspis) Zych, 1931, Fauna Ryb Dewonu i Downtonu Podola, p. 89. Brotzenaspis Stensio, 1958, Traite de Zool., 13, fasc. 1, pp. 265, 272. Pseudopteraspis Stensio, 1958, Traite de Zool., 13, fasc. 1, p. 254. Pteraspis (Cymripteraspis) White, 1960, Bull. Inst. Roy. Sci. Nat. Belg., 36, no. 6, p. 8 (footnote). Loricopteraspis Tarlo, 1961, Acta Palaeont. Polonica, 6, p. 382. Diagnosis. — Mostly moderate-sized or rather large pteraspids. The rostrum more or less elongate, but lacking a preoral field, and Fig. 3. Rostral plate of Althaspis whitei (Denison), ventral view, PF 1195 (X 6). dla, ventral surface of dorsal lamina, lacking basal layer and exposing cancelli anteriorly; ifc, pores of infraorbital canal; vlr, ventro-lateral face; vps, ven- tral preoral surface. DENISON: PTERASPIDIDAE 7 having its ventral preoral surface covered with dentine ridges. Bran- chial openings well in advance of the postero-lateral corners of the dorsal shield, and bounded by branchial and cornual plates. Discussion. — An irregular dentine ridge arrangement on the ven- tral side of the rostrum was used by Tarlo (1961, p. 382) to distin- guish Loricopteraspis, but this is hardly evidence that this area was formed of a number of separate synchronomorial units as he assumed. Pteraspis whitei (Denison, 1955, p. 450-454) may show an irregular development of the ventral rostral ridges (fig. 3), though this is a small species with the rostrum only slightly elongate; it may be con- sidered a primitive species of Althaspis. White (in Allen, Halstead, and Turner, 1968, pp. 151-152) still maintains the generic or subgeneric distinction of Cymripteraspis, though this differs from typical Althaspis only in the presence of a small plate with concentric ridges on the preoral border. Halstead (in Allen, Halstead, and Turner, 1968, p. 145) has attempted to ex- plain this as an oral plate appressed to the preoral border, an expla- nation denied by White. In any case, it appears at present to be a rather insecure basis for the erection of a distinct genus or subgenus. Brotzenaspis pteraspidoides Stensio (1958, fig. 152B) is a small form with a rostrum of only moderate length. Stensio's figure shows the ventral surface of the rostrum to be covered with more or less transverse ridges, which indicates a relationship to Althaspis. Tarlo (1961, p. 378) referred it to Zascinaspis (Protopteraspis of this paper), but that seems to be incorrect. Rhinopteraspis Jaekel, 1919 Type species. — Palaeoteuthis dunensis Roemer, 1855. (For synonymy, see Tarlo, 1961, pp. 368-370.) Diagnosis. — Moderate-sized to large pteraspids. The rostrum much elongated, with a preoral field formed by a separate plate. Branchial openings in advance of the postero-lateral corners of the dorsal shield, but the cornual plates reduced or absent. Discussion. — The International Commission on Zoological No- menclature (1957, p. 43) placed Rhinopteraspis Jaekel on the official list, and rejected the earlier names, Archaeoteuthis Roemer and Palae- oteuthis Roemer, as invalid. The type species was designated as Palaeoteuthis dunensis Roemer, and it was also placed on the official list. Since this decision, Tarlo (1961, p. 370) has indicated that Rhinopteraspis dunensis is a synonym of the earlier name, Stegano- 8 FIELDIANA: GEOLOGY, VOLUME 20 dictyum cornubicum M'Coy, 1851, as already claimed by Lankester in 1868. Since cornubicum was the type species of Steganodictyum, strict adherence to the law of priority would require the use of M'Coy's name for the genus and species. However, Steganodictyum would appear to qualify as a nomen oblitum, as defined in Article 23b of the International Code of Zoological Nomenclature, though this is not true of the specific name. If Lankester and Tarlo are correct, we may then refer to the genus as Rhinopteraspis, and its best known species as R. cornubica (M'Coy), but the type species remains Palaeo- teuthis dunensis Roemer, here accepted as a synonym of R. cornubica. As I suggested in 1967 (p. 34), Belgicaspis Zych, whose type spe- cies is Pteraspis crouchi Lankester, may belong to Rhinopteraspis or be closely related. Its preoral field is formed by a separate plate, as indicated by White (1960, p. 8 footnote), its rostrum is long and slender, though variable in shape (White, 1961, pi. 36, figs. 4-11), and its cornual plates are reduced or absent. As noted above (p. 6), Protaspis priscillae Denison may belong to Rhinopteraspis. Protaspis Bryant, 1933 Type species. — Pteraspis bucheri Bryant, 1932. Diagnosis. — Moderate-sized to large pteraspids with the dorsal shield rather broad or highly vaulted. The rostrum very short to moderate in length, with the development of its ventral preoral sur- face dependent on the rostral length. The branchial plates long, the cornual plates reduced or absent, and the branchial openings at the postero-lateral corners of the shield. The dorsal spine usually small, recumbent, and scale-like. Discussion. — Glossoidaspis Branson and Mehl, 1931 predates Pro- taspis Bryant, but was based only on a dorsal disc which shows no generically diagnostic characters, so it must remain incertae sedis. As used here, the genus Protaspis is restricted to those species in which the branchial openings are at the postero-lateral corners of the shield. This requires the removal of certain species formerly re- ferred to Protaspis in which the branchial openings have a more anterior position. I have not given the ventral preoral surface much weight in classifying Protaspis because its development is clearly related to the length of the rostrum. Thus, in species with a short, broad rostrum (figs. 14, 19) the ventral preoral surface is small, as in Protopteraspis, while in species with longer rostra (fig. 9) the ventral preoral surface may be considerable and a preoral field may be devel- oped, as in Pteraspis. The genus Protaspis has been subdivided into DENISON: PTERASPIDIDAE 9 a number of subgenera, based largely on differences in ornamentation, on details of the structure of the external gill openings, and, in one case, on the aberrant development of the dorsal spine and orbital plates. Subgenus Protaspis Bryant, 1933 Type species. — Pteraspis bucheri Bryant, 1932. Diagnosis. — Dorsal shield of moderate breadth. Ornamentation consisting of ridges of dentine arranged approximately parallel to lines of growth. Branchial openings facing posteriorly or postero- dorsally, and bounded externally by the branchial plates and some- times by the dorsal disc, with cornual plates absent or poorly de- veloped. Protaspis (Protaspis) bucheri (Bryant), 1932 Pteraspis bucheri Bryant, 1932, Proc. Amer. Phil. Soc, 71 , pp. 232-238, text- fig. 1; pi. 1; pi. 2, fig. 1; pi. 3, figs. 1-2. Protaspis bucheri Bryant, 1933, Proc. Amer. Phil. Soc, 72, pp. 294-296, text- fig. 1A; pi. 1; pi. 2, figs. 1-2; pi. 5, fig. 1; 1934, Proc. Amer. Phil. Soc, 73, p. 152, pi. 22; pi. 26, fig. 2; 1935, Proc. Amer. Phil. Soc, 75, p. 122, pi. 13, fig. 1; White, 1935, Phil. Trans. Roy. Soc. London (B), 225, p. 438; Deni- son, 1953, Fieldiana: Geol., 11, p. 334. Protaspis brevirostris Bryant, 1933, Proc. Amer. Phil. Soc, 72, pp. 297-298, pi. 8, figs. 1-2; 1935, Proc. Amer. Phil. Soc, 75, p. 123, pi. 15; Denison, 1953, Fieldiana: Geol., 11, p. 334. Occurrence. — Lower Devonian, Beartooth Butte formation, Bear- tooth Butte, Park County, Wyoming. Diagnosis. — Rostrum short (RL/TL= 0.11-0.19) • and broadly rounded anteriorly. Ornamentation coarse with 4.5 to 6.5 ridges per millimeter. Discussion. — Certain of Bryant's figures (1932, pi. 1; 1933, pi. 1; 1935, pi. 15), as well as one specimen in Field Museum (PF 2168), show that the branchial plates extend to the postero-lateral corners of the dorsal shield, where they are simply terminated and usually truncate. The branchial openings must have been at their poste- rior ends. 1 The measurements used in determining rostral ratios are: RL, length of rostral plate measured in the midline; RW, maximum width of rostral plate; and TL, total length of dorsal shield, measured in midline, but excluding the dorsal spine. 10 FIELDIANA: GEOLOGY, VOLUME 20 Fig. 4. Dorsal shield of type specimen of Protaspis (Protaspis) mcgrewi, n. sp., PF4337 (X 3/4). Protaspis (Protaspis) dorfi (Bryant), 1932 Pleraspis dorfi Bryant, 1932, Proc. Amer. Phil. Soc, 71, pp. 238-240, text-fig. 3; pi. 2, fig. 2; pi. 4. Protaspis dorfi Bryant, 1933, Proc. Amer. Phil. Soc, 72, pp. 296-297, text-fig. IB; pi. 3; pi. 5, fig. 2; pis. 6-7; 1934, Proc. Amer. Phil. Soc, 73, pp. 152- 153, pi. 23; pi. 24, fig. 1; Denison, 1953, Fieldiana: Geol., 11 , pp. 324, 334, fig. 74. Occurrence. — Lower Devonian, Beartooth Butte formation, Bear- tooth Butte, Park County, Wyoming. Diagnosis. — Rostrum short (RL/TL= 0.14-0.18) and narrowly rounded anteriorly. Ornamentation very fine with 8 to 12 ridges per millimeter. DENISON: PTERASPIDIDAE prp 11 Fig. 5. Rostral plate of Protaspis (Protaspis) mcgrewi, n. sp., ventral view restored from PF 4763 (about X 2). al, ascending lamella; dla, ventral surface of dorsal lamina; prp, median preoral wall; vlr, ventro-lateral face; vps, ventral pre- oral surface. Discussion. — Bryant's figure of the type specimen (1932, pi. 4) indicates that the branchial plates and openings were developed as in Protaspis bucheri. Protaspis (Protaspis) mcgrewi,1 new species Type. — Field Museum, PF 4337, a dorsal shield, complete except for its right postero-lateral corner (figs. 4, 6A, 7A). Referred specimens. — Dorsal shields, PF 4758-9, 4763 (figs. 5, 6B), 4914, 4917, 4356, and part of 4911; incomplete juvenile dorsal shield, PF 4491; ventral disc, PF 4361; rostral plates, PF 5265, 5267, 5269, 5278, 5284-5, 5288, 5295. Occurrence. — Lower Devonian, Beartooth Butte formation, Cot- tonwood Canyon, east of Lovell, NE H, sec. 4, T56N, R93W, Big- horn County, Wyoming. Diagnosis. — Dorsal shield broader or more highly vaulted than in P. bucheri and P. dorfi. Rostrum of moderate length and broadly rounded anteriorly, longer relatively than in P. bucheri or P. dorfi 1 After Dr. Paul O. McGrew who first called my attention to the rich occur- rence of fossil fishes in Cottonwood Canyon. 12 FIELDIANA: GEOLOGY, VOLUME 20 Fig. 6. Region of branchial opening of Protaspis (Protaspis) mcgrewi, n. sp. (about XI). A, dorsal view of PF 4337; B, ventral view of PF 4763. bro, bran- chial opening; brp, branchial plate; dd, dorsal disc. (RL/TL= 0.22-0.23). Dorsal spine relatively long and projecting. Ornamentation coarse with about 5.5 ridges per millimeter in the midline of the dorsal disc. Length of the dorsal disc (excluding the dorsal spine) as much as 137 millimeters. Description. — The dentine ridges (fig. 7A) typically have a smooth, gently convex crest, but locally may be flat-crested or tuberculate. There is a tendency for them to be grouped in bands separated by what are probably growth lines, and in some bands the ridges may be irregularly arranged, or rarely transverse or diagonal within the bands. In the central area of the dorsal disc the ridges, where un- worn, are finer, more convex, tuberculate, and with crenulate mar- gins. The rostrum has a well-developed ventral preoral surface (fig. 5, vps), covered medially with short to elongate, crested dentine ridges with prominent lateral projections, reminiscent of those of Traquairaspis; this area is bounded laterally by borders with rela- tively broad dentine ridges, or in the figured specimen, with bands of transverse or diagonal ridges similar to those of subgenus Cosmaspis. Ascending lamellae mark the posterior edge of the ventral preoral surface, while medially there is a preoral wall (fig. 5, prp) connecting the ventral preoral surface with the dorsal lamina of the rostral plate; this structure will be described more fully in Protaspis (Protaspis) brevispina. The pineal plate usually contacts the orbitals, and dis- tinct suborbital plates are probably not present below the orbitals Fig. 7. Ornament of dorsal shield (X 6). A, P. (Protaspis) mcgrewi, n. sp., PF4337; B, P. (Protaspis) brevispina, n. sp., PF4343; C, P. (Cyrtaspidichthys) ovata (Bryant), PF 1523; D, Lampraspis tuberculata, n. sp., PF 4913; E, P. (Cos- maspis) transversa, n. sp., PF 4334; F, P. (Cosmaspis) sp., PF 951. 13 14 FIELDIANA: GEOLOGY, VOLUME 20 Fig. 8. Dorsal shield of type specimen of Protaspis (Protaspis) brevispina, n. sp., PF4343 (X 3/4). and between the rostral and branchial plates. In all specimens the postero-lateral corners of the dorsal disc form the medial margins of the branchial openings, which are directed postero-medio-dorsally (fig. 6, bro). Protaspis mcgrewi resembles P. bucheri in its ornament and rostral shape, but differs in its broader (or more vaulted) dorsal shield, in the greater length of the rostrum, and possibly in the greater length of the dorsal spine. Its closest relative is Protaspis (Cosmaspis) transversa (p. 21), which it resembles in many features, most espe- cially in the occasional occurrence of bands of transverse or diagonal ridges. However, this is rare in mcgrewi, but characteristic of trans- versa, which also differs distinctly in the structure of its branchial openings. DENISON: PTERASPIDIDAE 15 Protaspis (Protaspis) brevispina, new species Type. — Field Museum, PF 4343, a dorsal shield crushed on the left side (figs. 7B, 8, 11A). Referred specimens.— Dorsal shields, PF 4340, 4344, 4346-8, 4352, 4762, 4764, 4766, 4768, 4770, 4910, 4912, 4918, 4929-30, 4933-5, 4940; associated dorsal and ventral discs, PF 4345, 4365, 4923; dorsal discs, PF 4372, 4388, 4937, 4939, 4986-7, 4989, 4991, 4999, 5062; ventral discs, PF4360, 4368, 4779, 4785, 4947; rostral plates, PF 5274 (fig. 10A), 5279-80, 5296 (fig. 10B-C), 5298, 5299 (fig. 9); branchial plates, PF 4376-7, 4380-1, 5314, 5322 (fig. 11B); incom- plete articulated juvenile individual, PF 4959. Occurrence. — Lower Devonian, Beartooth Butte formation, Cot- tonwood Canyon, east of Lovell, NE yA, sec. 4, T56N, R93W, Big- horn County, Wyoming. Diagnosis. — Shield moderately slender. Rostrum of moderate length (RL/TL= 0.20-0.24), narrowly rounded anteriorly, and with a well-developed preoral field covered with a fine reticular ornament. vir na Fig. 9. Rostral plate of Protaspis (Protaspis) brevispina, n. sp., ventral view, restored largely from PF 5299 ( X 2}4). al, ascending lamella; dla, ventral surface of dorsal lamina; na, possible narial notch; pf, preoral field; prp, median preoral wall; spr, outline of paired rostral space, based on dissection of PF 5298; vlr, ven- trolateral face. 16 FIELDIANA: GEOLOGY, VOLUME 20 Dorsal spine short, projecting little or not at all. Ornament rather fine with 6 to 8 ridges per millimeter near the midline of the dorsal disc. Dorsal shield estimated to have attained a length of 175 mil- limeters. Description. — This species differs from P. mcgrewi in its narrower proportions, longer and more narrowly rounded rostrum, short dorsal spine, and usually by its finer dentine ridges. From P. dorfi it may be distinguished by its coarser ridges and relatively longer rostrum. The more or less complete dorsal shields referred to this species have a length of 107-142 millimeters. The largest specimen referred here, PF 4365, is estimated to have had a dorsal shield 175 milli- meters long on the basis of the length of its ventral disc, which is 140 millimeters. The dentine ridges (fig. 7B) are usually smooth with flat or gently convex crowns and crenulate margins, but in some specimens the crests may be slightly tuberculate. Near the center of the dorsal disc the ridges may be similar, but sometimes their crests are cut by cross grooves which give the ridges the appearance of a row of overlapping pointed scales. The rostrum is somewhat variable in shape and proportions, but ventrally has a long preoral surface (fig. 9, pf) with fine reticular, probably non-dentinous ornament, comparable to the preoral field of Pteraspis. The posterior edge of the preoral field has a slight, rounded median process, bounded on either side by an ascending lamella (fig. 9, al). In many specimens there is a rostral space (fig. IOC, spr), widely open posteriorly, between the dorsal and ven- tral laminae of the rostrum. This is the usual condition in pteraspids with somewhat elongate rostra, and has been figured by Stensio (1964, figs. 46B-C, 91A-C). In other specimens of Protaspis brevi- spina, especially PF 5299, there is a nearly vertical preoral wall (fig. 9, prp) in the midline at the posterior edge of the ventral lamina of the rostrum, and connecting it with the dorsal lamina. A longi- tudinal section of another specimen (PF 5274) shows that in the midline where the preoral wall is developed (fig. 10 A, prp), the rostral space is absent and the rostrum is filled with spongy bone. On either side of the preoral wall the rostral space is present (fig. 10B, spr) ; its extent has been determined by dissection of another specimen (PF 5298) and indicated on Figure 9 (spr). This is important in the interpretation of pteraspid structure since it indicates the presence of a pair of pockets extending anteriorly from the oral cavity, and occupied in life by paired structures. These structures were prob- ably nasal sacs, and they could have opened posteriorly into the DENISON: PTERASPIDIDAE 17 die dla Fig. 10. Vertical longitudinal sections through rostral plates of Protaspis (Protaspis) brevispina, n. sp. ( X 3). A, sagittal section of PF 5274; B, parasagittal section of same rostrum; C, approximately sagittal section of PF 5296. dla, dorsal lamina; pf, preoral field; prp, median preoral wall; spr, rostral space. buccal cavity, or perhaps postero-laterally to the exterior at the cor- ners of the mouth, where notches are sometimes seen in the rostral plate (figs. 9, 19, na) . A preoral wall has also been found in Protaspis (Protaspis) megrewi and Protaspis (Cosmaspis) transversa. Its ab- sence in other species and in some specimens of Protaspis brevispina suggests that it formed only at maturity, perhaps as a part of the rostral plate itself, or perhaps as a distinct plate. The pineal plate of Protaspis brevispina contacts the orbitals, and usually there are separate suborbital plates separating the branchial and rostral plates. The branchial openings (fig. 11, bro) are similar to those of P. megrewi with their medial margins formed dorsally by the postero-lateral corners of the dorsal disc. Below this, one or more small platelets, possibly cornuals (fig. 11,? co), line the medial side of the branchial duct in some specimens. Subgenus Europrotaspis White, 1961 Type species. — Protaspis (Europrotaspis) crenulata White, 1961. Europrotaspis White, 1961, Bull. Brit. Mus. (Nat. Hist.), Geol., 5, pp. 268- 270 (new subgenus). 18 FIELDIANA: GEOLOGY, VOLUME 20 I. i .' brp-- FlG. 11. Region of external branchial opening of Protaspis (Protaspis) brevi- spina, n. sp. (about X 2). A, dorsal view, PF4343; B, ventral view, PF 5322. bro, branchial opening; brp, branchial plate; ? co, possible cornual plates; dd, dor- sal disc. Diagnosis. — Similar to subgenus Protaspis except that each bran- chial opening is bounded medially by a small narrow cornual plate at the postero-lateral corners of the dorsal disc. Discussion. — When White (1961, p. 268) erected the subgenus Europrotaspis, the structure of the branchial openings of the typical Protaspis was unknown. It now appears that the branchial openings are very similar in the two subgenera. In typical Protaspis, where cornual plates are known, they are small elements lying in the medial walls of the posterior ends of the branchial ducts, but they do not appear on the dorsal surface of the shield. In Europrotaspis (fig. 12, co), especially in P. (Europrotaspis) arnelli from Podolia (Brotzen, 1936, fig. 3), they differ only in extending to the dorsal surface of the shield, where they lie against the postero-lateral corners of the dorsal disc and form the medial edges of the branchial openings. White used other characters to distinguish Europrotaspis. One was the large, flat preoral field of P. (Europrotaspis) crenulata, which he contrasted with the very narrow ventral preoral surface of "Pro- taspis cf. bucheri" (of Denison, 1953, fig. 78, here referred on p. 29 to Oreaspis dunklei n. gen., n. sp.). However, a well-developed pre- oral field may occur in those species of Protaspis which have a ros- trum of moderate length, as P. brevispina (fig. 9) . White also mentions the very large ridge scales of Europrotaspis, but all of his scales are DENISON: PTERASPIDIDAE 19 \, \ Fig. 12. Region of external branchial opening of Protaspis (Europrotaspis) crenulala White, after White, 1961. bro, branchial opening; brp, branchial plate; co, cornual plate; dd, dorsal disc. isolated and their relative size uncertain; moreover, the ridge scales of Protaspis dorfi and P. (Cosmaspis) transversa are known to be rela- tively large. The asymmetrical ridge scales of P. (Europrotaspis) cren- ulata described by White can hardly be accepted as sufficient evidence for the presence of lateral keels, for they are isolated, and lateral keels are unknown in other pteraspids. I conclude that Europrotaspis, as far as it is known, differs only in minor ways from the subgenus Protaspis. Subgenus Gigantaspis N. Heintz, 1962 Type species. — Gigantaspis isachseni N. Heintz, 1962. Giganthaspis A. Heintz, 1937, Skr. Svalbard og Ishavet, nr. 72, p. 18 (nomen nudum); F0yn and Heintz, 1943, Skr. Svalbard og Ishavet, nr. 85, p. 14 (nomen nudum); Dineley, 1960, Geol. Mag., 97, pp. 18, 20, 22-23, 30 (no- men nudum). Gigantaspis N. Heintz, 1962, Norsk Polarinst., Arbok 1960, pp. 23-24. Diagnosis. — Similar to the subgenus Protaspis, from which it is distinguished by the very slight vaulting of the dorsal disc, the up- right dorsal spine, the relatively longer rostral plate, and by the slit- like branchial openings facing upward between the posterior ends of the branchial plates and the dorsal disc. Discussion. — Gigantaspis, the characteristic pteraspid of the Kapp Kjeldsen division of the Wood Bay series of Spitsbergen, was recog- nized as a close relative of Protaspis by N. Heintz. She gave it generic rank, but I find that it fits well in the genus Protaspis as used 20 FIELDIANA: GEOLOGY, VOLUME 20 here, and differs from the subgenus Protaspis only in relatively minor characters that certainly warrant no more than subgeneric distinc- tion. The branchial openings are fundamentally the same in both subgenera; in both there are no visible cornual plates (except possibly for small platelets in Protaspis brevispina) , but in Gigantaspis the openings face more dorsally and are more slit-like. The dentine ridges of subgenus Protaspis may be "fine, smooth and entire," as described for Gigantaspis. The rostal plate is relatively longer than in any known species of subgenus Protaspis; based on the restoration of P. (Gigantaspis) bocki (N. Heintz, 1982, fig. 1A), RL/TL=0.30, compared to 0.11 to 0.24 in species of subgenus Protaspis. An up- right dorsal spine is unknown in subgenus Protaspis, though it occurs in Eucyclaspis and doubtfully in P. (Europrotaspis) arnelli. Subgenus Cyrtaspidichthys Whitley, 1940 Type species. — Cyrtaspis ovatus Bryant, 1932. Cyrtaspis Bryant, 1932 (not of Fischer, 1853), Proc. Amer. Phil. Soc, 71, p. 241; 1933, Proc. Amer. Phil. Soc, 72, p. 304; White, 1935, Phil. Trans. Roy. Soc. London (B), 225, p 438. Cyrtaspidichthys Whitley, May 1940, Australian Nat., 10, p. 243; Denison, 1953, Fieldiana: Geol., 11, p. 318. Eucyrtaspis White and Moy Thomas, June 1940, Ann. Mag. Nat. Hist. (11), 5, p. 507. Diagnosis. — Dorsal shield broad or highly vaulted. Ornamenta- tion consisting of oval dentine tubercles arranged in rows approxi- mately parallel to the lines of growth (fig. 7C). Branchial openings directed postero-medially at the posterior ends of the branchial plates. Protaspis (Cyrtaspidichthys) ovata (Bryant), 1932 Cyrtaspis ovatus Bryant 1932, Proc. Amer. Phil. Soc, 71 , pp. 242-243, text-fig. 4, pi. 6, fig. 1; 1933, Proc Amer. Phil. Soc, 72, pp. 304-305, text-fig. 5; pi. 5, fig. 3; pi. 14, figs. 1-2; 1935, Proc. Amer. Phil. Soc, 75, pp. 123-124, pi. 11, fig. 2, pi. 16. Cyrtaspis papillatus Bryant, 1933, Proc. Amer. Phil. Soc, 72, p. 306, pi. 15, figs. 1-2; 1934, Proc. Amer. Phil. Soc, 73, pp. 153-154, pi. 25. Cyrtaspis falcatus Bryant, 1933, Proc. Amer. Phil. Soc, 72, p. 307, pi. 17, fig. 1. Protaspis (Cyrtaspidichthys) ovatus Denison, 1953, Fieldiana: Geol., 11 , p. 336. Occurrence. — Lower Devonian, Beartooth Butte formation, Bear- tooth Butte, Park County, Wyoming. Diagnosis. — Shield very broad (or highly vaulted). Rostrum of moderate length (RL/TL=0.25 to 0.28) and rather sharply rounded anteriorly. DENISON: PTERASPIDIDAE 21 Discussion. — Specimens figured by Bryant (1933, pi. 14, fig. 2; 1935, pi. 16). as well as those in Field Museum (PF 1523, 1601. 2170). show that the branchial openings were fundamentally as in the sub- genus Protaspis. The branchial plates are very broad and their posterior ends extend slightly beyond the postero-lateral corners of the dorsal disc, with the result that the branchial openings must have opened medially as well as posteriorly. Protaspis (Cyrtaspidichthys) sculpta (Bryant), 1933 Cyrlaspis sculptus Bryant, 1933, Proc. Amer. Phil. Soc, 72, pp. 306-307, pi. 16, figs. 1-2. Protaspis (Cyrtaspidichthys) sculptus Denison, 1953, Fieldiana: Geol., 11, p. 336. Occurrence. — Lower Devonian, Beartooth Butte formation, Bear- tooth Butte, Park County, Wyoming. Diagnosis. — Shield less broad (or less vaulted) than in P. (C.) ovata. Rostrum relatively shorter (RL/TL=0.20) and bluntly rounded. Subgenus Cosmaspis,1 new subgenus Type species. — Protaspis (Cosmaspis) transversa, new species. Diagnosis. — Dorsal shield broad or highly vaulted. Ornamenta- tion mostly arranged in bands more or less parallel to lines of growth, while within the bands the individual dentine ridges are arranged diagonally, or sometimes transversely or irregularly with respect to the bands. Branchial openings behind the dorsal disc at the poste- rior tips of the branchial plates. Protaspis (Cosmaspis) transversa, new species Type. — Field Museum, PF 4334, a nearly complete dorsal shield (figs. 7E, 13, 16). Referred specimens.— Dorsal shields, PF 4335-6, 4338-9, 4349, part of 4352 (fig. 15), 4355, 4755, 4757, 4760, 4915, 4924, 4926, 4931- 2, 4944, 5497; dorsal discs (juvenile), PF4375, 4394, 4403, 4408, 4412, 4416, 4772-4, 4974, 4977, 4981, 4984, 5020, 5050-1, 5059, 5077; ventral discs, PF 4362-3, 4780, 4936; ventral discs (juvenile), PF 4450-1, 4453-4, 4486, 4994, 5016, 5122, 5148-9, 5175, 5177-8, 5180, 5184; dorsal and ventral discs, PF4369; rostral plates, PF 4942, 5262, 5268, 5276-7, 5281 (fig. 14), 5291-4; orbital plates, PF 5211, 1 From kocthos, ornament; and aains, shield. 22 FIELDIANA: GEOLOGY, VOLUME 20 Fig. 13. Dorsal shield of type specimen of Protaspis (Cosmaspis) transversa, n. subgen., n. sp., PF 4334 (X 3/4). 5215, 5221, 5233-4, 5237, 5243, 5252, 5259, 5260; branchial plates, PF4382, 4384, 5300, 5307, 5311, 5315-6, 5319, 5324; dorsal spine, PF 5349; more or less articulated specimens, PF 4389 (fig. 17), 4390-1; juvenile articulated or associated specimens, PF 4490, 4492, 4494, 4496, 4953, 4969. Occurrence. — Lower Devonian, Beartooth Butte formation, Cot- tonwood Canyon, east of Lovell, NE }4, sec. 4, T56N, R93W, Big- horn County, Wyoming. Diagnosis. — As for genus. Dorsal shield attaining a length of 140 millimeters. Description. — The ornament of bands of cross ridges is particu- larly characteristic of the subgenus Cosmaspis. In P. (C.) transversa DENISON: PTERASPIDIDAE 23 Fig. 14. Rostral plate of Protaspis (Cosmaspis) transversa, n. sp., ventral view, PF 5281 (about X 2). dla, ventral surface of dorsal lamina; ma, matrix; prp, median preoral wall; vps, ventral preoral surface. (fig. 7E) the individual cross ridges are more commonly arranged diagonally than transversely within the bands, and are typically gently convex and smooth-topped with crenulate margins. The bands are approximately 1.5 to 2.5 millimeters wide, while the indi- vidual ridges within the bands are 4.5 to 5.5 per millimeter. In the central or juvenile area of the dorsal disc the ornament is different and resembles that of typical pteraspids, for the ridges are not ar- ranged in bands, but are concentric and parallel to lines of growth. They are finer, 6 to 8 per millimeter, and where unworn are sharp- crested or tuberculate. Outside of this zone is an intermediate zone where the ridges may occur diagonally in bands, where they are of intermediate coarseness, and where they grade from tuberculate into smooth crests. The rostrum, though variable in proportions, may be described as of moderate length (RL/TL=0.20 to 0.27), broad (RW/RL= 1.75 to 2.38), and broadly rounded anteriorly. Ventrally, it has a narrow preoral surface (fig. 14, vps) covered with irregularly-shaped, serrate tubercles, arranged in patches posteriorly, irregularly ante- iorly, and in rows laterally. Dorsally, the posterior border of the rostral plate is usually (though not in the type) deeply indented at each side to receive the prominent anterior process of each orbital plate. Beneath each orbital plate there may be a separate suborbital 24 FIELDIANA: GEOLOGY, VOLUME 20 A B Fig. 15. Region of branchial opening of Protaspis (Cosmaspis) transversa, n. sp., PF 4352 (X 1M)> A, dorsal view; B, ventral view, bro, branchial opening; brp, branchial plate; dd, dorsal disc. plate separating the rostral and branchial plates (PF 4334, 4338, 4775) ; in others (PF 4391, 4924) this space is occupied by a ventral process of the orbital plate. The pineal and orbital plates may or may not contact each other. The dorsal spine is long, projecting, and recumbent. Each branchial plate extends a short distance behind the postero- lateral corner of the dorsal disc. In this region, the medial side of its dorsal lamina (fig. 15A) bends down to meet the inner margin of its ventral lamina (fig. 15B), and so surrounds the branchial duct, which opens at the posterior end of the branchial plate. The bran- chial duct extends forward from here for some distance in a deep groove between the dorsal and ventral laminae of the branchial plate. For a short distance anterior to the point where the branchial plate joins the dorsal disc, the branchial duct is bounded medially by one or more plates of thick, spongy bone, extending from the lateral mar- gin of the dorsal disc to the medial margin of the ventral lamina of the branchial plate (fig. 16,? co) . This plate or plates has much of the relations of the cornual plate of Protaspis arnelli, as figured by Brot- zen (1936, fig. 3:1), and may be called by that name provisionally. One large articulated specimen, PF 4389, is crushed but shows much of the structure of the oral region and tail. At least 18 oral plates (fig. 17, op) are exposed but considerations of symmetry indi- brd br,Pd brpv Fig. 16. Oblique section through left branchial duct of Protaspis (Cosmaspis) transversa, n. sp., PF 4334. b, breaks; brd, branchial duct; brpd, dorsal lamina of branchial plate; brpv, ventral lamina of branchial plate; ? co, possible cornual plate; dd, dorsal disc. Fig. 17. Attempted restoration of the oral region of Protaspis (Cosmaspis) transversa, n. sp., ventral view, based on PF 4389 (about X 1). brp, branchial plate; Ip, lateral plates; ol, oral-lateral plate; op, oral plates; orp, orbital plate; pop, postoral plates; vd, ventral disc; vps, ventral preoral surface. 25 26 FIELDIANA: GEOLOGY, VOLUME 20 Fig. 18. Region of branchial opening of Protaspis (Eucyclaspis) erroli Deni- son, somewhat restored from PF 3814 (about X 2 }/£). A, ventral view; B, dorsal view; C, posterior view, bro, branchial opening; brp, branchial plate; ? co, possible cornual plate; dd, dorsal disc. cate that there were more originally. Between them and the ventral disc are two rows of postoral plates (fig. 17, pop) . The posterior row bounding the ventral disc consists of a small median plate with one or two elongate plates bounding it on either side. The anterior row consists of a large, subtriangular median element with a small plate at either side. At each antero-lateral corner of the ventral disc are two lateral plates (fig. 17, Ip). The posterior and dorsal part of the tail is not preserved. The median ventral edge is formed by a series of ridge scales, of which the first is very large, the second even larger, while those behind gradu- ally decrease in size posteriorly. The flank scales are also relatively large anteriorly and of the usual shape found in pteraspids. As indicated above (p. 14), Protaspis (Protaspis) mcgrewi, though distinguished subgenerically from Protaspis (Cosmaspis) transversa, was surely closely related. P. mcgrewi rarely and locally shows bands of diagonal or transverse ridges, and agrees well in many other characters. Protaspis (Cosmaspis) sp. Protaspis sp., representing an undescribed subgenus, Denison, 1953, Fieldiana: Geol., 11, p. 350, fig. 84A. DENISON: PTERASPIDIDAE 27 Occurrence. — Lower Devonian, Card member of Water Canyon formation, Cache County, Utah. Discussion. — The few fragments from the Water Canyon forma- tion described by me in 1953 can now be referred to the subgenus vlr Fig. 19. Rostral plate of Protaspis (Eucyclaspis) erroli Denison, ventral view, restored from PF 3814 (X 2l/Q. al, ascending lamella; dla, ventral surface of dorsal lamina; na, possible narial notch; sla, sutural area for lateral plate; vlr, ven- trolateral face; vps, ventral preoral surface. Cosmaspis. They differ from P. (C.) transversa in that the cross ridges within the bands of ornamentation are more regular and trans- verse and have flatter crests (fig. 7F) . Also, the rostrum (PF 952) appears to be relatively longer. Subgenus Eucyclaspis,1 new subgenus Type species. — Protaspis (Protaspis) erroli Denison. Diagnosis. — A Protaspis in which the branchial ducts open poste- riorly at the postero-lateral corners of the rather broad shield (fig. 18), and are bounded by the branchial plates ventrally and laterally, by the dorsal disc dorsally, and by probable cornual plates medially. The dorsal spine is slender and erect, and each orbital plate has a process projecting postero-laterally from its lateral margin behind the orbit. The ornament consists of very fine ridges more or less concentric to lines of growth. Protaspis (Eucyclaspis) erroli Denison, 1967 Protaspis (Protaspis) erroli Denison, 1967, Jour. Linn. Soc. (Zool.), 47, pp. 34- 37, text-figs. 1-2, pi. 1. Occurrence. — Lower Devonian, near base of Grassy Flat member of Water Canyon formation, Logan Canyon, Cache County, Utah. 1 From tvKvicXos, well rounded; aaxis, shield. 28 FIELDIANA: GEOLOGY, VOLUME 20 Fig. 20. Region of branchial opening of Oreaspis dunklei, n. gen., n. sp., slightly restored from U. S. Nat. Mus. 20524 (about X 3}^). A, dorsal view; B, ventral view, bro, branchial opening; brp, branchial plate; ? co, possible cornual plate; dd, dorsal disc; pbr, postbranchial plate; ppb, posterior process of branchial plate. Diagnosis. — A small species with the median length of the dorsal shield attaining 61 millimeters. The rostral plate broad (RW/RL= 2.14), bluntly rounded anteriorly, and with a narrow ventral preoral surface (fig. 19, vps). Ornament of sharp-crested, crenulate ridges, about 11 to 13 per millimeter. Discussion. — This species differs from all the other species referred to the genus Protaspis in having projecting processes on the orbital plates, and from all except the species of Gigantaspis in its erect dorsal spine. These characters are considered sufficiently aberrant to warrant subgeneric separation. Oreaspis,1 new genus Type species. — Oreaspis dunklei, new species. Diagnosis. — Pteraspids with each branchial opening situated dor- sally between the posterior part of the dorsal disc and branchial plate, and usually bounded posteriorly by both (fig. 20, bro) . Typically a scale-like post-branchial plate (fig. 20, pbr) is attached to a posterior projection of each branchial plate. Dorsal shield moderately broad or vaulted. 1 From opos, mountain; a Fig. 26. Rostral plate of Lampraspis tuberculata, n. sp., ventral view, PF 5261 (X 3). dla, ventral surface of dorsal lamina, lacking the basal layer anteriorly; ma, matrix; vlr, ventro-lateral face; vps, ventral preoral surface. 36 FIELDIANA: GEOLOGY, VOLUME 20 ratios of their width/length are unreliable. However, the rostral plate is relatively short (RL/TL= 0.19-0.22), broad, and bluntly or even squarely rounded anteriorly. As is to be expected in a short rostrum, the ventral preoral surface (fig. 26, vps) is narrow, and has a slight median posterior projection. This surface is mostly covered with a fine reticular ornamentation, but anteriorly there are fine tubercles. The pineal plate usually contacts the orbitals, though not in PF 4754. The orbital plates have a ventral process that separates the rostral and branchial plates. The dorsal spine is short and re- cumbent, and projects behind the dorsal disc in most, though not in all specimens. Psephaspis 0rvig, 1961 Type species. — Psephaspis williamsi 0rvig. Diagnosis. — Moderate to large-sized pteraspids with the shield usually relatively broad, and with the ornament consisting typically of denticles with crenulate margins. The branchial openings are at the postero-lateral corners of the shield, at the posterior ends of the branchial plates. Discussion. — Psephaspis was originally referred to the Drepa- naspida by 0rvig (1961, pp. 526-533), but was shown to be a pteras- pid by me in 1968 (p. 270). Its known characters indicate a close relationship to Protaspis, from which it is most clearly distinguished by its psammosteid-like ornamentation. Additional differences may be discovered when it is more completely known. It is the latest pteraspid, and the only genus of the family known to survive into the Middle Devonian. Psephaspis williamsi 0rvig, 1961 Psephaspis williamsi 0rvig, 1961, Ark. Zool., ser. 2, 12, nr. 33, pp. 526-533, figs. 6-10; Denison, 1968, Fieldiana: Geol., 16, p. 270, figs. 2, 7A. Occurrence. — ? Lower Devonian, upper part of Water Canyon formation, Cache County, Utah, and Bear Lake County, Idaho. Diagnosis. — A species attaining a large size, and with the den- ticles lower and less convex than in P. idahoensis. Psephaspis idahoensis Denison, 1968 Psephaspis idahoensis Denison, 1968, Fieldiana: Geol. 16, pp. 270-279, figs. 1, 3, 7B, 8. Occurrence. — Middle Devonian, basal unit of Jefferson formation, Lemhi County, Idaho. DENISON: PTERASPIDIDAE 87 Diagnosis. — A species attaining a smaller size, and having higher, more convex denticles than P. williamsi. Subfamily Doryaspidinae Rostral plate truncate anteriorly, and the mouth opening antero- dorsally between it and an elongate pseudorostrum, which is attached to the ventral shield. Doryaspis White, 1935 Type species. — Scaphaspis nathorsti Lankester. Dyreaspis A. Heintz, 1934, Naturen, 58, p. 244 (nomen nudum); 1935, Ann. Rept. Smithsonian Inst., 1934, p. 235 (nomen nudum). Doryaspis White, 1935, Phil. Trans. Roy. Soc. London (B), 225, p. 444; F0yn and Heintz, 1943, Skr. Norges Svalbard-og Ishavs-Unders0k.( 85, pp. 14- 15; Obruchev, 1964, Osnovy Paleontologii, 11, p. 69. Lyktaspis N. Heintz, 1968, Proc. 4th Nobel Symp., Stockholm, 1967, p. 73. Diagnosis. — An aberrant pteraspid with the rostral plate short and truncate anteriorly, and underlain by a long pseudorostrum attached to the ventral shield, and with the mouth opening antero- dorsally between the rostrum and pseudorostrum. The cornual plates project far laterally. The ornament consists of concentric ridges subdivided by a system of radiating grooves. Doryaspis nathorsti (Lankester) Scaphaspis nathorstii Lankester, 1884, K. Svenska Vetenskapsakad. Handl.' (n.f.), 20, no. 9, p. 5, pi. 1, figs. 1-3. Pteraspis nathorsti Woodward, 1891, Ann. Mag. Nat. Hist., (6), 8, p. 2, pi. 2, fig. 1. Doryaspis nathorsti White, 1935, Phil. Trans. Roy. Soc. London, (B), 225, p. 444, text-fig. 80, pi. 25, fig. 102; F0yn and Heintz, 1943, Skr. Norges Svalbard-og Ishavs-Unders0k., 85, p. 14, fig. 5A; Obruchev, 1964, Osnovy Paleontologii, 11, p. 69, fig. 44. Lyktaspis nathorsti N. Heintz, 1968, Proc. 4th Nobel Symp., Stockholm, 1967, pp. 73-79, figs. 1-4. Occurrence. — Lower Devonian, Lykta Division, Wood Bay Series, Spitsbergen. Discussion. — According to Errol White (in litt., July 2, 1968), Doryaspis Dejean 1835 (Coleoptera) is a nomen nudum, as are its three species, which apparently have never been validated, so Dory- aspis was available when used by White, in 1935. This being the 38 FIELDIANA: GEOLOGY, VOLUME 20 case, the new name, Lyktaspis, proposed by N. Heintz, was not necessary. Doryaspis is an aberrant genus and occupies an isolated position with relation to other pteraspids. This has been recognized by Tarlo (1962, p. 265) who placed it in a separate family, Doryaspididae, and in a separate order, Doryaspidida; by Obruchev (1964, p. 69) who used the family Doryaspididae; and by N. Heintz (1968, p. 79) who used the family Lyktaspididae, and with reservations the sub- order Lyktaspidida. In my own work with Heterostraci I have not felt the need, as have some students of the group, to elevate families to orders and to create new families, but have found that retaining the original family rank and using subfamily divisions has been ade- quate to express the grouping and relationships of genera. Using this approach, Doryaspis can be placed in a subfamily, Doryaspidi- nae, to distinguish it from all other pteraspids, which have been re- ferred to the Pteraspidinae. Pteraspididae incertae sedis Two small pteraspids from the Lower Devonian, Wood Bay Series, of Spitsbergen have been described by Natascha Heintz (1960) as Pteraspis ? lyktensis and Pteraspis ? minor. They are known only from isolated dorsal and ventral discs, and are distin- guished particularly by the following: their coarse, non-crenulated dentine ridges; the absence of a dorsal spine; the fact that the median longitudinal sensory canals meet a loop of the supraorbital (or pineal) canals on the dorsal disc; and the inclusion of the lateral longitudinal sensory canals on dorsal discs as small as 14 and 18 millimeters. The inclusion of the lateral sensory canals is no proof that these discs be- longed to adults, for many pteraspids continue to grow long after this stage, as has been clearly shown by White (1958, p. 230, figs. 4-8). However, the inclusion at such a small size indicates that the adults were small, and the extension of the supraorbital canals onto the dorsal disc is known only in small pteraspids. Pteraspis ? lykten- sis has been referred to a new genus, Grumantaspis, by Obruchev (1964, p. 69), and Pteraspis ? minor has placed in another new genus, Ennosveaspis, by Stensib" (1964, p. 363). What little is known about them hardly warrants their separation into two distinct genera, in fact, at present they are inadequately characterized generically. They may represent persistently primitive pteraspids, as suggested by N. Heintz (1960, p. 11). DENISON: PTERASPIDIDAE 39 Pteraspis dixoni White (1938, pp. 100-110, figs. 11-25) was re- ferred to a new genus Penygaspis, by Stensio (1958, p. 292) because of a presumably different manner of growth, as inferred from its den- tine ridge pattern. The distinction of its ridge pattern is that the central areas of the dorsal and ventral discs have more or less longi- tudinal ridges, while only the peripheral parts have the usual con- centrically arranged ridges. Stensio was probably correct in assuming that the central areas were mineralized at one time (synchronomori- ally), but this is also true of pteraspids with concentric central ridges, as will be shown later. One difference is that the central areas are relatively large, with a length on the order of 25 millimeters. The longitudinal arrangement of the central ridges is known to occur occasionally in other pteraspids, such as Protaspis (Europrotaspis) crenulata White (1961, pp. 277-278, pi. 43, fig. 4), and may be con- sidered a relic of cyathaspid ancestry. Until this species is more completely known it must remain incertae sedis, but it is probably primitive. Protaspis (Protaspis) tenuistriatus Denison (1953, pp. 343-348> figs. 80-81) does not preserve the posterior part of the shield, so the position of the external branchial openings is uncertain. The slight development of the branchial groove in the most posterior preserved part of the left branchial plate suggests that the branchial openings were posterior, but there is no way to tell whether they are the type found in Protaspis or in Oreaspis. REFERENCES Allen, J. R. L., L. B. Halstead, and S. Turner 1968. Dittonian ostracoderm fauna from the Brownstones of Wilderness Quarry, Mitcheldean, Gloucestershire. Proc. Geol. Soc. London, no. 1649, pp. 140-151, figs. 1-3, with discussion, pp. 151-153. Branson, E. B. and M. G. Mehl 1931. Fishes of the Jefferson formation of Utah. Jour. Geol., 39, pp. 509-531, figs. 1-2, pis. 1-3. Brotzen, F. 1936. Beitrage zur Vertebratenfauna des westpodolischen Silurs und Devons. I. Protaspis arnelli n. sp. und Brachipteraspis n. gen. latissima Zych. Ark. Zool., 28A, nr. 22, pp. 1-52, figs. 1-17, pis. 1-10. Bryant, W. L. 1932. Lower Devonian fishes of Bear Tooth Butte, Wyoming. Proc. Amer. Phil. Soc, 71, pp. 225-254, figs. 1-6, pis. 1-10. 1933. The fish fauna of Beartooth Butte, Wyoming. Parti. The Heterostraci and Osteostraci. Proc. Amer. Phil. Soc, 72, pp. 285-314, figs. 1-8, pis. 1-21. 40 FIELDIANA: GEOLOGY, VOLUME 20 1935. Cryptaspis and other Lower Devonian fossil fishes from Beartooth Butte, Wyoming. Proc. Amer. Phil. Soc, 75, pp. 111-128, figs. 1-2, pis. 1-18. Denison, R. H. 1953. Early Devonian fishes from Utah. Part II. Heterostraci. Fieldiana: Geol., 11, pp. 291-355, figs. 61-85. 1955. Early Devonian vertebrates from the K,noydart formation of Nova Scotia. Fieldiana: Zool., 37, pp. 449-464, figs. 109-113. 1967. A new Protaspis from the Devonian of Utah, with notes on the classifi- cation of Pteraspididae. Jour. Linnean Soc. London (Zool.), 47, pp. 31-37, figs. 1-2, pi. 1. 1968. Middle Devonian fishes from the Lemhi range of Idaho. Fieldiana: Geol, 16, pp. 269-288, figs. 1-12. Heintz, N. 1960. The Downtonian and Devonian vertebrates of Spitsbergen. X. Two new species of the genus Pteraspis from the Wood Bay series in Spitsbergen. Skr. Norsk. Polarinst., no. 117, pp. 1-13, figs. 1-4, pis. 1-2. 1962. Gigantaspis—a new genus of fam. Pteraspidae from Spitsbergen. A pre- liminary note. Arbok Norsk. Polarinst., 1960, pp. 22-27, fig. 1, pi. 1. 1968. The pteraspid Lyktaspis n.g. from the Devonian of Vestspitsbergen. In 0rvig, T., ed., Current problems of lower vertebrate phylogeny. Proc. 4th Nobel Symp., Stockholm, 1967, pp. 73-80, figs. 1-4. International Commission on Zoological Nomenclature 1957. Opinion 438. Validation under the plenary powers of the generic name "Rhinopteraspis" Jaekel, 1919 (Class Ostracodermi). Internat. Comm. Zool. Nomencl., Opinions and Declarations, 15, pt. 3, pp. 41-50. Lankester, E. R. 1868. On the discovery of the remains of cephalaspidian fishes in Devonshire and Cornwall; and on the identity of Steganodictyum, M'Coy, with genera of those fishes. Quart. Jour. Geol. Soc. London, 24, pp. 546-547. Obruchev, D. M. 1964. Fundamentals of Paleontology. Manual for paleontologists and geolo- gists of USSR. Vol. XI. Agnatha, Pisces. (In Russian.) Moscow, 522 pages, 673 figs., 45 pis. 0RVIG, T. 1961. Notes on some early representatives of the Drepanaspida (Pteraspido- morphi, Heterostraci). Ark. Zool., ser. 2, 12, nr. 33, pp. 515-535, figs. 1-10. Stensio, E. 1958. Les cyclostomes fossiles ou ostracodermes. In Grasse P.-P., ed., Traite de Zool., 13, fasc. 1, pp. 173-425, figs. 107-219. 1964. Les cyclostomes fossiles ou ostracodermes. In Piveteau, J., ed., Traite de Paleont., 4, vol. 1, pp. 96-382, figs. 1-125. Tarlo, L. B. H. 1961. Rhinopteraspis cornubica (McCoy), with notes on the classification and evolution of the pteraspids. Acta Palaeont. Polonica, 6, pp. 367-397, figs. 1-13, pis. 1-2. 1962. The classification and evolution of the Heterostraci. Acta Palaeont. Polonica, 7, pp. 249-281, figs. 1-15. White, E. I. 1935. The ostracoderm Pteraspis Kner and the relationships of the agnathous vertebrates. Phil. Trans. Roy. Soc. London, (B), 225, pp. 381-457, figs. 1-97, pis. 25-27. DENISON: PTERASPIDIDAE 41 1938. New pteraspids from South Wales. Quart. Jour. Geol. Soc. London, 94, pp. 85-115, figs. 1-26. 1950. Pteraspis leathensis White, a Dittonian zone-fossil. Bull. Brit. Mus. (Nat. Hist.), Geol., 1, pp. 69-89, figs. 1-25, pi. 5. 1958. Original environment of the craniates. In Westoll, T. S. ed., Studies on fossil vertebrates. London, pp. 212-234, figs. 1-8. 1960. Notes on pteraspids from Artois and the Ardenne. Bull. Inst. Roy. Sci. Nat. Belgique, 36, no. 6, pp. 1-16, figs. 1-6, pis. 1-3. 1961. The Old Red Sandstone of Brown Clee Hill and the adjacent area. II. Palaeontology. Bull. Brit. Mus. (Nat. Hist.), Geol., 5, pp. 243-310, figs. 1-61, pis. 33-48.