>. pas ’ ‘i ben 54 PASS fe me STO Bah : fet tm Th key et SA rere ‘ ° - a a - oF, >: anew >. agen : Rested » Dt te ik op 2 heh . he CN a! * ee es : Ao ha! - abel see aus hae RW y . \. ; eh a eet, " a rate © Tene Ps need ting ae a hte ah tn ‘ : Pe te eee ts “a wi woacnl 5 oe ees pee . © ~~" + heath oul poiard Me Non ee Oe a on, — ener: oa *- - 2 : - Se San we Ce a em ‘ mt vie ok a. = Met Pi OA atl Cth yt - - ~ e > Pty 7 YS fy ea i a Pye : : eet taht = “me r 2 G re Fy Ay ‘s = 8 = n~ 8 S A) ‘& i) 3 ~ i) ~_— AY ol ro) cS) a Los, —_— 3 fe a) =| ) > HiG.- LS. DENISON: HOLLAND QUARRY SHALE FISHES 591 their arrangement is different from that known in other species of the genus. The largest and most characteristic, here referred to sim- ply as the lateral plate, lies not far from its correct position on the right side of the type (fig. 124, B). It has a pointed anterior end, relatively long medial and lateral edges, and a long, slender projection Fic. 188. Anterior part of ventral disc of Pteraspis carmani, PF 1922 (x 2), showing growth anomaly. on its posterior end (fig. 139, A). Its probable relations to neighbor- ing plates are shown in figure 140 (la). The posterior projection lies between two small postero-lateral plates. The medial side of the anterior point presumably contacts the antero-lateral plate. The position of the lateral side of the anterior point is difficult to deter- mine; it may abut against the rostral plate as shown, or it may lie farther posterior against the orbital and branchial plates. The ven- tral lateral sensory line traverses the lateral plate and presumably meets the infraorbital line in its anterior part, though the course of the latter is difficult to determine in available specimens. The medial postero-lateral plate is preserved in the type (figs. 124, B; 140, mpl). It occupies the notch between the posterior proc- ess of the lateral plate and the ventral disc, and carries the ventral lateral line from one to the other. A number of isolated small plates (fig. 139, C) showing lateral line pores have been provisionally iden- tified as this plate. No lateral postero-lateral plate is seen in the type, but the presence of a notch laterad to the posterior process of the lateral plate indicates that there must have been one. A small elon- gate plate, PF 2030 (fig. 139, B), may be this plate, and in PF 2025 a similar one is associated with a probable medial postero-lateral plate. 592 FIELDIANA: GEOLOGY, VOLUME 11 In the type a small pyritic nodule has obscured the right antero- lateral plate, but what is probably the left one is present, though in- completely exposed and displaced somewhat forward. A number of isolated antero-laterals (fig. 1389, E) show this to be a subrhombic Fic. 139. Oro-pharyngeal plates of Pteraspis carmani (X 6). A, left lateral plate, PF 2037; B, (?)lateral postero-lateral plate, PF 2030; C, medial postero-lat- eral plate, PF 2031; D, oral plate, PF 2025; E, left antero-lateral plate, PF 2043. plate with the infraorbital lateral line traversing one corner. Its cor- rect position (fig. 140, al) is difficult to determine precisely, largely because the continuation of the infraorbital canal on the oral-lateral or rostral plates cannot be traced. Postoral plates.—In Pteraspis vogti (Kiaer, 1928, fig. 2) there are no postoral plates, and the oral plates articulate directly with the anterior edge of the ventral disc. In Pteraspis rostrata, White (1935, DENISON: HOLLAND QUARRY SHALE FISHES 593 figs. 41-47) found one pair of postoral plates lying between the ven- tral disc and oral plates. Stensié (1958, p. 261, fig. 140, B) has indi- cated that there is a complicated development of these plates in a Podolian pteraspid. In the type of Pteraspis carmani (fig. 124, B) there are numerous small plates lying between the ventral disc and oral plates, suggesting a complicated development of the postoral Fic. 140. Partial restoration of the oro-pharyngeal region of Pteraspis car- mani, ventral view (X 3/2). al, antero-lateral plate; br, branchial plate; Ja, lateral plate; lpl, lateral postero- lateral plate; mpl, medial postero-lateral plate; ol, oral-lateral plate; or, oral plate; orb, orbital plate; po, postoral plates; ro, rostral plate; srl, subrostral lamina; vd, ventral disc. plates in this species also. Unfortunately they are disarranged and it is impossible to reconstruct this region with any confidence. Oral-lateral plates.—These plates, first named by Kiaer (1928, p. 123), and renamed ‘“‘preorogonial’’ by Stensié (1958, p. 258), lie between the antero-lateral and rostral plates, and form the sides of the mouth region. In Pteraspis vogti (Kiaer, 1928, fig. 2) there is a single pair lying medial to the infraorbital lines, which appear to pass directly from the lateral plates to the rostral plates. In Pter- aspis rostrata, White (1935, p. 409) found two pairs of oral-lateral plates, a very small anterior pair and a large posterior pair, the latter traversed by the infraorbital canal. On the left side of the type of Pteraspis carmani (fig. 124, B) there are two plates of which the pos- terior is identified as the antero-lateral, and the anterior as the oral- lateral, both displaced somewhat antero-laterally. The oral-lateral is a small, subtriangular plate that probably lay laterally against the 594 FIELDIANA: GEOLOGY, VOLUME 11 ventral lamina of the rostral plate, as shown in figure 140 (ol). Whether or not it was crossed by the infraorbital sensory canal can- not be determined with certainty. Oral plates—At least eleven of the oral plates are preserved in the type (fig. 124, B), and a number of isolated ones have been found in the collection. They agree in general with those described in Pteraspis vogtt and P. rostrata. They are long and narrow, slightly curved, broader posteriorly, and tapering anteriorly. On their ven- tral or outer surface they are ornamented with more or less trans- verse dentine ridges that curve forward at the sides into longitudinal ridges. On the inner or dorsal side the oral plates have a strongly developed ridge that may extend the full length of the plate or may be restricted to its anterior end. The longitudinal dentine ridges on one side of the plate extend anteriorly onto this ridge (fig. 139, D), where they fan out to form what was called the ‘‘tooth plate’ by Kiaer (1928, p. 123). Scales.—The most complete account of pteraspid scales has been given by White (1935, pp. 412-418), who has based his description on the articulated specimens from the Wayne Herbert Quarry in Herefordshire, England. Since there are no specimens with articu- lated scales in the Carman collection, I have relied on White’s de- scription for location and orientation of the numerous isolated scales. All ridge scales were originally symmetrical and strongly arched, but in the Holland Quarry shale they have often been flattened and distorted. A number of different types may be distinguished, but all intergradations between them occur. One type includes large ridge scales (PF 2061-62, fig. 141, A) with an exposed surface as much as 16.5 mm. long. These are of moderate width and taper to a point posteriorly; they have a deeply rounded notch in the middle of an otherwise convex anterior edge, and a narrow overlap area along this edge. These are possibly ventral ridge scales, since in Protaspis dorfi the largest scales occur here (Denison, 1958, fig. 74). Another type (PF 2059-60) differs only in its smaller size (length 8-9 mm.) and relatively broader proportions. These scales grade into two types. The first of these (PF 2055-56, fig. 141, B) is small, about 6 mm. long, relatively very broad, and bluntly pointed pos- teriorly; the anterior edge is always convex. The second type in- cludes more slender ridge scales (PF 2063-64, fig. 141, C) that are sharply pointed posteriorly and may have the anterior edge either notched or convex. This type, in turn, grades into the long, very slender, fulcral scales of the caudal region (PF 2065-66, fig. 141, D). Fic. 141. Scales of Pteraspis carmani (X 5). A-—D, ridge scales; E—-H, flank scales; I, aberrant ridge scale; J, inner side of ridge scale. A, PF 2061; B, PF 2055; C./PP 20635) D> PE. 2065-76, eb b-.2047) EP BY 20513-G, PR 2048; se P2053 Tt PF; 2069;' J, BE 2057. 595 596 FIELDIANA: GEOLOGY, VOLUME 11 Flank scales have typically a rounded anterior border, a pointed posterior end, and a nearly symmetrical, subrhombic shape (PF 2047— 49, fig. 141, E). Rarely both the anterior and posterior ends are pointed, in which case the scales are approximately diamond-shaped (PF 2051-52, fig. 141, F). Not uncommonly the anterior border is notched, and in this case there is a slight to marked asymmetry of development on either side of the notch (PF 2053-54, fig. 141, H). Some of these asymmetrical scales correspond to the so-called double scales described by White (1935, p. 414; 1950, p. 79) in Pteraspis rostrata and P. leathensis. The anterior overlapped area of flank scales may be slightly or strongly developed. The ornamentation of both ridge and flank scales consists basi- cally of a number of narrow rows of short, longitudinal ridges. Each row is parallel to the anterior edge of the scale and is thus either con- vex anteriorly or, when the scale has an anterior notch, has this convexity modified by a median concavity. Within each row the short dentine ridges are arranged mainly antero-posteriorly though in some scales the ridges tend to flare towards the sides. The ridges have round to sharp crests and are usually faintly denticulate. In the posterior and ontogenetically original row the ridges are two or more times as long as in the more anterior rows. Variations in this pattern are not uncommon and include such things as occasional ridges that extend across two rows, rows that do not extend across the whole width of the scale, and denticulate areas where the ridges are subdivided and the row pattern is not apparent. Transverse ridges parallel to the anterior edge, such as are characteristic of Pteraspis leathensis and P. rostrata toombsi, have been noted only in a few scales of P. carmani (PF 2048, fig. 141, G), and in these they are usually incomplete mesially. Along the posterior edges where they overlap the scales behind, the dentine ridges extend onto the inner surface, and may cover a relatively large area at the pos- terior tip (fig. 141, J). One aberrant scale, PF 2069 (fig. 141, I), is worthy of mention. It is relatively short (length=7.9 mm.) and broad (width=5.0 mm.), considerably arched, and apparently symmetrical. Its most striking feature is that most of the surface is covered with dentine ridges radi- ating from a point near the center of the scale. A few transverse ridges lie in the mid-line at one end of the radiating ridges, while the opposite end of the scale is covered by six narrow rows of short, antero-posteriorly directed ridges. Lateral line canals.—As far as it has been determined, the arrange- ment of the lateral line canals of the shield is similar to that of Pter- DENISON: HOLLAND QUARRY SHALE FISHES 597 aspis rostrata as described by White (1935, pp. 421-426). The course of the infraorbital canal on the ventro-lateral rim of the rostral plate has not been found, and only an occasional pore of the ventral trans- verse commissures has been seen. I have been unable to identify any pores of the postoral canal on the anterior part of the ventral disc, though some of the irregularities of the dentine ridge pattern here may indicate the presence of such pores. Form of the dentine ridges——There is a correlation in Pteraspis carmant between the width of the dentine ridges and the shape of their crests. The extremely broad ridges that usually occur on the antero-median part of the ventral disc have nearly flat crests with- out any tuberculation; their edges are finely scalloped by the basilar foramina that open into the very narrow grooves between the ridges. The broad ridges of other plates, and on the antero-lateral parts of the ventral disc, have gently rounded crests that are usually faintly tuberculate; their margins are also crenate. Very fine ridges such as occur on the postero-lateral parts of the dorsal and ventral discs, on the branchial process of the orbital plate, and on both laminae of the branchial plate, are sharply crested and separated by widely open grooves; they are commonly crenate. Occasionally and locally the ridges are divided into discrete denticles or partially divided so that they present a beaded appearance. In scales the ridges are sub- divided into short lengths. In the posterior part of each scale the ridges are sharply crested, while in the anterior part they are gen- erally broader and less sharply or rarely round crested. The inden- tations on the sides usually do not extend to the top of the crests. The dorsal spines with a scale-like ridge pattern have the ridges shaped like those of scales. EUARTHRODIRA PHL YCTAENASPIDAE Aethaspis ohioensis, new species Type.-—CNHM =PF 1661, a nearly complete cranial roof with associated intero-lateral, spinal, and incomplete anterior ventro- lateral and antero-ventral plates (fig. 142). Referred specimens.—Incomplete impression of a cranial roof, PF 1662; incomplete cranial roof, inner side, PF 1850; left para- nuchal, inner side, PF 1852; median dorsals, PF 1663-64; posterior dorsals, PF 1855-57; anterior laterals, PF 1666, 1846; anterior ventro- 598 FIELDIANA: GEOLOGY, VOLUME 11 laterals, PF 1848-49; anterior ventro-laterals and spinals, PF 1667— 1668; spinals, PF 1665, 1853; posterior ventro-lateral, PF 1847; undetermined plates, PF 1851, 1854. Horizon and locality—As given on page 555. Diagnosis.—A small species whose cranial roof has a median length, excluding the rostral and pineal, of 25 mm. in the type, and whose median dorsal plate is 16 to 23 mm. in median length. The rostral and pineal plates are not fused to the rest of the cranial roof. The postorbital plates contact the nuchal plate, separating the cen- trals and preorbitals. The central sensory canal is restricted to the postorbital plate and does not extend onto the central plate. The nuchal plate has little development of an antero-median projection and of lateral notches for the paranuchals. The anterior lateral plate has a relatively long ventral edge and the inner wing is little developed. The spinal plate is relatively long compared to the anterior ventro- lateral; it projects behind the posterior edge of the anterior ven- tro-lateral and is provided with several relatively strong spinelets on its medial edge. The surface of the dermal bones is ornamented with small, widely spaced tubercles, rather uniform in size, and lack- ing any pronounced linear orientation. Description and discussion.—Aethaspis ohioensis is a very small arthrodire, particularly when compared to the other two species of the genus, A. major and A. utahensis, from the Water Canyon for- mation of Utah (Denison, 1958). I believe, however, that most of the known specimens are not juvenile. The known cranial roofs have most of their plates fused, the only exceptions being the rostral, pineal, postnasals, and postmarginals; these are not attached in the Ohio species though they are firmly fused in known specimens of A. major and A. utahensis. There is some reason for believing that the unfused condition of the most anterior plates is primitive, and that fusion took place in larger individuals and in more advanced species and genera. The ornamentation of Aethaspis ohioensis is not particularly dis- tinctive. It has been shown elsewhere (Denison, 1958, p. 470) that the character of arthrodire ornamentation may change considerably in old individuals as larger tubercles are developed near the plate margins and overgrow the smaller primary tubercles near the center of the plates. Secondary as well as primary tubercles occur in A. ohio- ensis; this is shown best on the nuchal plate of PF 1662 (fig. 144), where minute primary tubercles are scattered between moderate- sized secondary tubercles near the center of the plate. Fic. 142. Aethaspis ohioensis, type, PF 1661 (xX 5/2). A, cranial roof; B, an- terior ventro-lateral, intero-lateral, and spinal plates. 599 600 FIELDIANA: GEOLOGY, VOLUME 11 The cranial roof (figs. 142-144) shows clearly that this species belongs to Aethaspis. An elongate nuchal that meets the preorbitals and separates the centrals is not known in any other euarthrodire. The contact of the preorbitals with the nuchal is a character of A. = 2. ee, Senses WZ SS \ —E———— | | PAN NS) ZA\\ 7a ZB ‘ ——— eee Fic. 148. Restoration of the cranial roof of Aethaspis ohioensis (X 3). CE, central; MG, marginal; NU, nuchal; PAN, paranuchal; PRO, preorbital; PTO, postorbital; RC, endoskeletal rhino-capsular ossification; cc, central canal; ife, infraorbital canal; lc, main lateral line; mp, middle pit line; poc, preopercular canal; pp, posterior pit line; soc, supraorbital canal. ohioensis that distinguishes it from A. major and A. utahensis. Like- wise, the restriction of the central sensory canals to the postorbitals, and the absence of any extensions onto the central plates is distine- tive of the Ohio species. The preorbitals are not as greatly elongate posteriorly as in the other species of the genus. The pattern of the cranial roofing bones has been determined from PF 1662 (fig. 144), where the sutures are discernible. Neither the sutures nor the bone radiation can be seen in the type. The dermal bones of the snout (rostral, pineal, and postnasals) are not known in A. ohioensis, but the type (fig. 142, A) shows at its anterior end the imperfectly preserved endoskeletal, rhino-cap- DENISON: HOLLAND QUARRY SHALE FISHES 601 sular ossification, as well as the subnasal shelf of the posterior endo- cranial ossification. Though the preservation is so poor that it does not permit a reconstruction, the extent and shape of the anterior part of the cranium are indicated (fig. 143). The snout is relatively short Fic. 144. Photograph of a cast of an incomplete natural mold of a cranial roof of Aethaspis ohioensis, PF 1662 (X 3). and wide, being intermediate between the broad-snouted Kujdanowi- aspis and the narrow-snouted Aethaspis major and A. utahensis. The inner side of the cranial roof is shown incompletely in PF 1850, and it agrees in most respects with that of Kujdanowzaspis (Stensi6, 1945, fig. 8, B). The endocranium is absent and may not have been ossified in this specimen, but the median depressed area that it occu- pied is clearly marked by the surrounding ridges on the inner side of the dermal bones. On the right side a prominent ridge on the medial part of the paranuchal plate presumably enclosed the endo- lymphatic duct. It occupies the position of the depression on the dorsal face of the endocranium of Kujdanowiaspis in which this duct passed into the paranuchal plate (Stensid, 1945, fig. 1). 602 FIELDIANA: GEOLOGY, VOLUME 11 Fic. 145. Plates of the trunk shield of Aethaspis ohioensis (X 3). A, median dorsal, PF 1664; B, spinal, PF 1853; C, posterior dorsal, PF 1855; D, posterior dorsal, PF 1856; E, anterior ventro-lateral, PF 1849; F, posterior ventro-lateral, PF 1847. The median dorsal plate (figs. 145, A; 146, C) is similar to that of other Aethaspis and is of the short, broad type that typifies the Actin- olepinae. Its highest point is near the center, and from here a low external ridge extends about half the distance to the posterior edge. The small, anterior, external, unornamented areas (fig. 146, C, 2) that occur in other Aethaspis are indicated in PF 1664. The inner surface has no median keel, and even the small median ridge that occurs in A. major is lacking. In the central part of the inner surface i WZ = Fic. 146. Restorations of plates of the trunk shield of Aethaspis ohioensis. A, right anterior lateral, based on PF 1666 (x 5); B, right posterior ventro-lateral, based on PF 1847 (x 5); C, median dorsal, based on PF 1663-64 (x 5/2); D, right anterior ventro-lateral (AVL), antero-ventral (AV), intero-lateral (JL), and spinal (SP), based on PF 1661, 1667-68 (xX 5/2). end, limits of endoskeleton of trunk shield; s. ADL, s.AVL, s.PDL, s.PVL, and s.SP, overlap areas for, respectively, the anterior dorso-lateral, anterior ventro- lateral, posterior dorso-lateral, posterior ventro-lateral, and spinal plates; x, ex- ternal unornamented area. 603 604 FIELDIANA: GEOLOGY, VOLUME 11 there is a deep depression which is abruptly terminated posteriorly but which opens into the general concavity of the antero-medial part of the plate. Three small, bilaterally symmetrical plates, PF 1855-57, are iden- tified as posterior dorsal plates and presumably lay in the mid-line be- hind the median dorsal plate (Denison, 1958, pp. 482, 518). PF 1855 (fig. 145, C) resembles a plate from Utah that was believed to be a posterior dorsal of a juvenile Aethaspis (op. cit., p. 482, fig. 98, C). It is remarkable for having a posterior median slot immediately be- hind a high median ridge. This is of the type referred to as posterior dorsal, type 2 (op. cit., p. 482). PF 1856-57 (fig. 145, D) are rela- tively narrower posterior dorsals that lack the posterior slot. As in PF 1855, there is a median ridge that is high posteriorly and dis- appears at the gently arched anterior end. This type does not agree well in shape with the posterior dorsal, type 1, of Aethaspis major (op. cit., fig. 93, A), and may have occupied a more posterior position. The anterior lateral plate (fig. 146, A) differs considerably from that of Aethaspis major. The ventral or spinal edge is relatively longer, the center of ossification is more posterior, and the external surface is relatively flat, with the inner wing only slightly differen- tiated. In all these features it is more primitive than A. major and approaches the condition of Kujdanowiaspis. The anterior ventro-lateral plates (figs. 142, B; 145, E; 146, D) are similar to those of A. major and A. utahensis. One specimen, PF 1848, is much smaller and relatively much narrower than the others. If this belongs to a juvenile A. ohioensis it would indicate that there was considerable proportional change during growth. The intero-lateral and antero-ventral plates are present in the type (fig. 142, B), but the preservation is such that their shape and struc- ture cannot be determined. The spinal plate (figs. 142, B; 145, B; 146, D) is relatively longer than in A. major and A. utahensis. As has been mentioned previously (Denison, 1958, p. 487), a long spinal appears to occur in small Aethaspis and a relatively short one in larger forms, and A. ohioensis agrees with this correlation. The rela- tive proportions of spinal and anterior ventro-lateral plates in Aeth- aspis are as follows: Length Length _Length SP AVL SP Length AVL Aethaspis major, PF 939........... 69 mm. 55 mm. 0.80 Aethaspis utahensis, PF 321........ 36 32 0.89 ACNASDIS SPs ER O4 Uae oe ee, 30 39 nie Aethaspis ohioensis, PF 1667....... iby 21 fe DENISON: HOLLAND QUARRY SHALE FISHES 605 The posterior ventro-lateral plate of Aethaspis ohioensis (figs. 145, F; 146, B) has a relatively long ventral lamina. In A. major and A. utahensis this lamina was restored as being quite short (Den- ison, 1958, figs. 99, A, 114, C), but since the known plates are incom- plete posteriorly, this may be incorrect. The lateral lamina in A. ohioensis is short antero-posteriorly as in other Aethaspis, but it is relatively higher than in the other species. Aethaspis ohioensis is believed to be more primitive than A. major and A. utahensis in the following characters: the small size; the broad, unreduced snout with its dermal bones not fused to the cranial roof; the moderate elongation of the preorbital plates; the long spinal edge and little-developed inner wing of the anterior lateral plate; and the relatively long spinal plate. In most of these features it approaches Kujdanowiaspis, which is perhaps the most primitive of known Arc- tolepida. Euarthrodire indet. A minute dermal jaw element, PF 2086 (fig. 147), resembles one (PF 526) from the Water Canyon formation of Utah that was con- sidered (Denison, 1958, p. 496, fig. 102, D—E) to be a supragnathal of an undetermined arctolepid. Its size is about what would be Fic. 147. Dermal jaw element of undetermined euarthrodire, PF 2086 (x 10). expected for a gnathal of Aethaspis ohioensis. Its reference to the latter could be defended on the grounds that it is the only arthrodire known from the Holland Quarry shale. However, because very little is known of the jaws of Lower Devonian arthrodires, and because PF 2086 shows possible specializations, I consider it more prudent to leave this specimen unnamed. This small gnathal has lost one tip, and where it is broken it is seen to be deeply excavated on the unexposed side. It is believed that this side attached to the palato-quadrate or Meckel’s cartilage. The opposite surface (fig. 147) is nearly flat except for four faint 606 FIELDIANA: GEOLOGY, VOLUME 11 longitudinal ridges. One of the sides adjacent to the ridged surface is slightly convex and is covered with low, blunt, and apparently worn denticles. The ridged and denticulate surfaces are of a dark color and together are probably the biting surface. The side oppo- site the denticulate surface is slightly concave and smooth. The unbroken end is depressed below the ridged biting surface and forms a sort of neck. : This gnathal is quite different from the shearing type, as exempli- fied by Coccosteus and Dinichthys, and is comparable only with crush- ing dentitions such as occur in ptyctodonts. A crushing dentition is presumably a specialization in arthrodires, and it is surprising to find it in such an early form. With the exception of the Water Canyon formation specimen, no gnathals of this type have yet been described from the Lower Devonian. ACANTHODII Onchus cf. peracutus Bryant The acanthodian fin spines from the Holland Quarry shale (PF 2087-96) are long, straight, slender, and gradually tapering to a point (fig. 148). The surface of the exserted portion is smooth, lacking the ridges or ribs usually present in Onchus spines, except that faint ridges may be present proximally on the postero-lateral edges. The inserted base is longitudinally striated. The pulp or main canal is widely open at the base (fig. 149, C, mc) and continues to be open for about the proximal third of the total length of the spine. In the distal two-thirds, the pulp canal is partly, then com- pletely enclosed by processes of the walls of the spine. Near the middle of the spine these processes do not quite meet, with the result that the pulp canal retains an external connection through a narrow slit (fig. 149, B). The processes that enclose the pulp canal are set in front of the postero-lateral edges of the spine so that there is a channel on the posterior surface of the spine all the way to the tip (fig. 148, B). The histological structure of these acanthodian spines (fig. 149) shows a general correspondence to that of elasmobranchs (Peyer, 1946, pp. 88-98). An elasmobranch spine is a complex structure, really a “spine within a spine,”’ with the inner spine or ‘‘Stammteil’’ (op. cit., p. 97) arising ontogenetically quite separately from the outer spine or “‘Mantelteil’’ which surrounds it. In Onchus cf. peracutus the pulp canal or main canal, as the terms are commonly used, Fic. 148. Onchus ef. peracutus, fin spines (X 2). A, PF 2090, anterior face; B, PF 2087, posterior face. Fic. 149. Transverse sections of fin spines of Onchus cf. peracutus (X 30). A, PF 2096, distal part of spine; B, PF 2096, near middle of spine; C, PF 2095, proximal part of spine. dt, dentine tubules; mc, main canal; pe, pulp canal of inner spine; po, pulp canals of outer spine; spz, inner spine or ‘‘Stammieil’’; spo, outer spine or “‘ Man- telteil.”’ 607 608 FIELDIANA: GEOLOGY, VOLUME 11 is the pulp cavity of the inner spine, and from it the dentine tubules radiate outward to the boundary of the outer spine. The pulp cavity of the outer spine is variously developed. Devononchus was charac- terized by Gross (1940, p. 15) in part by the presence of an upper canal, which is the main pulp canal of the outer spine. On the other hand, Onchus has a diffuse outer pulp, consisting of many irregular but mainly longitudinal canals from which the dentine tubules extend outward. This is the situation in Onchus cf. peracutus, except that there may be a small upper canal centered in the diffuse pulp of the proximal half of the spine. Reference of the Holland Quarry shale specimens to Onchus per- acutus is uncertain because of the poor preservation of the type speci- men from Beartooth Butte, Wyoming (Bryant, 1934, p. 149, pl. 18, fig.3). According to Bryant, the sides are ornamented with a median series of several sharp, longitudinal ridges, but the only ridges visible on the type are two or three fine ones on the postero-lateral edges. The surface of most of the spine is smooth, as in the Ohio specimens. The type is obliquely crushed and incompletely preserved, but it can be interpreted as having a shape and structure similar to that of the Ohio specimens. A second Beartooth Butte species, Onchus penetrans (Bryant, 1982, p. 252, pl. 10, fig. 4), differs in its curvature and fine longitudinal ridges. The spine described as Machaeracanthus minor (Bryant, 1934, p. 148, pl. 18, fig. 1) may be Onchus penetrans. The type is a poorly preserved natural cast, and so what Bryant believed to be a longitudinal ridge on the side could be one of the postero- lateral edges of an obliquely crushed spine. AGE OF THE HOLLAND QUARRY SHALE The vertebrate fauna of the Holland Quarry shale is comparable to that of the Beartooth Butte formation of Wyoming (Bryant, 1932; 1934; 1935) and the Water Canyon formation of Utah (Denison, 1958; 1958). Allocryptaspis occurs in all three formations, but not elsewhere. This is the largest of known Cyathaspidae and its size is suggestive of a late or post-Dittonian age (Denison, 1953, p. 296). Protaspis is the pteraspid of the Wyoming and Utah formations, but the reference of the Holland Quarry species to Pteraspis is purely ar- bitrary. Pteraspis carmanz is closely related to Protaspis, approaches it in most respects, and could well be nearly and directly ancestral toit. This species suggests, therefore, that the Holland Quarry shale is slightly older than the Beartooth Butte and Water Canyon forma- tions. The importance of pteraspids in Lower Devonian correlations DENISON: HOLLAND QUARRY SHALE FISHES 609 has been emphasized by White (1956) and Schmidt (1959). The earliest members of the family, small species with short, blunt rostra, are characteristic of the lower Dittonian. They are succeeded in the middle Dittonian and equivalent upper Gedinnian by larger species, typified by Pteraspis rostrata and P. crouchi, which show various spe- cializations. In the upper Dittonian and equivalent lower Siegenian comes the first appearance of species with long rostra, exemplified by Pteraspis leachi. It is in this zone that Protaspis makes its first ap- pearance. Finally there is the zone of the very long snouted Pteraspis dunensis, beginning with the lower Breconian and middle Siegenian, and continuing into the Emsian. If Protaspis indicates an upper Dittonian or lower Siegenian age, the slightly more primitive con- dition of Pteraspis carmani suggests that the Holland Quarry shale was a bit older. The arthrodire of the Holland Quarry shale, Aethaspis, is known elsewhere only from the Water Canyon formation. The Ohio species, A. ohioensis, is more primitive than those from the Water Canyon formation, which suggests that it is somewhat older. But Aethaspis is a specialized arctolepid, at least in comparison to Kujdanowiaspis (Denison, 1958, p. 545), and one would expect it to be younger than the latter. Kujdanowiaspis occurs in upper Gedinnian and lower Siegenian equivalents in Podolia and Spitsbergen. Aethaspis thus suggests an age no older than lower Siegenian. The acanthodian of the Holland Quarry shale, Onchus cf. pera- cutus, indicates little more than a general correspondence with the Beartooth Butte and Water Canyon formations. In summary, the vertebrates of the Holland Quarry shale favor a correlation with the upper Dittonian of England and the lower Siegenian of continental Europe. Correlation with the North Amer- ican standard section can be accomplished only indirectly. According to Cooper et al. (1942), the Siegenian (or lower part of the Coblenzian) is to be correlated with the Deerpark stage. This correlation, based on the vertebrate fossils, agrees with the stratigraphical evidence that Dr. Carman has summarized (1960, p. 4). The Holland Quarry shale was formed in the interval between the deposition of the Raisin River dolomite and the Sylvania sandstone. This interval covered the Helderberg, Deerpark, and part of the Onesquethaw stages. It should be noted that by American usage the Onesquethaw stage is referred to the Middle Devonian, and the Holland Quarry shale would occupy some of the upper part of the Lower Devonian. In Europe presumed Onesquethaw equivalents are referred to the Lower De- 610 FIELDIANA: GEOLOGY, VOLUME 11 vonian (Emsian), and the Holland Quarry shale would fall into the middle part of the Lower Devonian. ECOLOGY OF THE HOLLAND QUARRY SHALE Since the Holland Quarry shale is known from a single small out- crop that has never been well exposed and is now completely covered, there is little geologic information available for the determination of the manner of its deposition. Carman (1960, p. 2) has shown that the Holland Quarry outcrop could be interpreted as a filling of (1) a roughly circular pit, or (2) a valley. The presence of aquatic verte- brates favors the second interpretation, since they must have had access to stream or sea. The individual groups of vertebrates found in the shale are of little help in determining the depositional environment. Allocryptaspis belongs to the subfamily Poraspinae, which occurs in marginal marine and fluviatile deposits, but not in typical marine sediments (Denison, 1956, pp. 416-417). Pteraspis belongs to a family that occurs in fresh- water, marginal marine, and marine formations. Some pteraspids may have been adapted to particular habitats within this range; others may have been euryhaline (op. cit., pp. 417-418). Aethaspis belongs to the Euarthrodira, a group that may have been originally restricted to fresh waters but at the time of the deposition of the Holland Quarry shale was also found in marine and marginal marine deposits (op. cit., pp. 426-427). Onchus belongs to the Acanthodii, which in Lower Devonian times lived in marine, marginal, and fresh- water habitats (op. cit., p. 425). Though the individual vertebrates have little ecological signifi- cance, the whole assemblage, including the eurypterids and plants, shows such striking resemblance to that of the Beartooth Butte for- mation of Wyoming that it is safe to assume some similarity in envi- ronment. The Beartooth Butte formation has the appearance of a channel fill and has been interpreted as a stream or estuarine deposit. But a similar vertebrate assemblage occurs in the Water Canyon for- mation of Utah, and this is a widespread deposit, with gastropods, brachiopods, and ostracods, some of which are definitely marine. I believe that it is a marginal marine sediment deposited in a large bay whose waters were salt or brackish (Denison, 1956, pp. 413-415). If the Water Canyon formation is a marginal marine deposit, the same is probably true of the Beartooth Butte and Holland Quarry formations. The Holland Quarry shale may have been deposited in a channel or estuary opening into the transgressing sea to the north. DENISON: HOLLAND QUARRY SHALE FISHES 611 Its waters may have been brackish or salt, and the absence of any definitely marine invertebrates may be due to a foul, muddy bottom.! SUMMARY The Holland Quarry shale, a Lower Devonian formation known from a single small exposure in Lucas County, northwestern Ohio, has yielded plants, eurypterids, and vertebrates. The vertebrates are described in this paper. The Cyathaspidae are represented by Allo- cryptaspis laticostatus, new species. The most abundant form belongs to the Pteraspidae and has been named Pteraspis carmani, new spe- cies. Its generic reference is arbitrary, and it is closely related and may be ancestral to Protaspis. Some aspects of its growth and indi- vidual variation have been considered. The arthrodire, a relatively rare element in the fauna, is Aethaspis ohioensis, new species, a small and primitive member of the genus. Spines of acanthodians have been referred to Onchus cf. peracutus Bryant. The vertebrates indicate that the Holland Quarry shale should be correlated with the upper Dittonian or equivalent lower Siegenian of the European Lower Devonian. This is probably of the same age as the Deerpark stage of the North American standard section. The vertebrate assemblage is similar to those of the Beartooth Butte for- mation of Wyoming and the Water Canyon formation of Utah. I believe that the Holland Quarry shale was deposited in salt or brackish water, possibly in a channel or estuary opening into the transgressing sea to the north. 1As this was in press, Dr. E. Kjellesvig-Waering found a piece of Holland Quarry shale containing ostracods, and Dr. E. S. Richardson, Jr., found another with not only ostracods, but also gastropods, articulate and inarticulate brachio- pods, tubicolous annelids, and a crinoid. 612 FIELDIANA: GEOLOGY, VOLUME 11 REFERENCES BROTZEN, FRITZ 1936. Beitrage zur Vertebratenfauna des westpodolischen Silurs und Devons. I. Protaspis arnelli n. sp. und Brachipteraspis n. gen. latissima Zych. Ark. Zool., 28A, no. 22, pp. 1-52, figs. 1-17, pls. 1-10. BRYANT, WILLIAM L. 1932. Lower Devonian fishes of Bear Tooth Butte, Wyoming. Proc. Amer. Phil. Soc., 71, pp. 225-254, figs. 1-6, pls. 1-10. 1934. The fish fauna of Beartooth Butte, Wyoming. Parts II and III. Proc. Amer. Phil. Soc., 73, pp. 127-162, figs. 1-8, pls. 1-26. 1935. Cryptaspis and other Lower Devonian fossil fishes from Beartooth Butte, Wyoming. Proc. Amer. Phil. Soc., 75, pp. 111-128, figs. 1, 2, pls. 1-18. CARMAN, J. ERNEST 1960. The stratigraphy of the Devonian Holland Quarry Shale of Ohio. Field- jana: Geol., 14, pp. 1-5, fig. 1. CoopER, G. ARTHUR, ET AL. 1942. Correlation of Devonian sedimentary formations of North America. Bull. Geol. Soc. Amer., 53, pp. 1729-1794, fig. 1, chart 4. DENISON, ROBERT H. 1953. Early Devonian fishes from Utah. Part II. Heterostraci. Fieldiana: Geol., 11, pp. 291-355, figs. 61-85. 1956. A review of the habitat of the earliest vertebrates. Fieldiana: Geol., 11, pp. 359-457. 1958. Early Devonian fishes from Utah. Part III. Arthrodira. Fieldiana: Geol., 11, pp. 461-551, figs. 86-116. FAHLBUSCH, KLAUS 1957. Pteraspis dunensis Roemer. Eine Neubearbeitung der Pteraspidenfunde (Agnathen) von Overath (bez. Kéln). Palaeontographica, 108, Abt. A, pp. 1-56, figs. 1-27, pls. 1-7. Gross, WALTER 1940. Acanthodier und Placodermen aus Heterostius-Schichten Estlands und Lettlands. Ann. Soc. Reb. Nat. Invest. Univ. Tartuensis Const., 46, pp. 1-89, figs. 1-17, pls. 1-9. KIAER, JOHAN 1928. The structure of the mouth of the oldest known vertebrates, pteraspids and cephalaspids. Palaeobiologica, 1, pp. 117-134, figs. 1-7, pls. 12-13. 1932. The Downtonian and Devonian vertebrates of Spitsbergen. IV. Sub- order Cyathaspida. Skr. Svalbard Ishavet, 52, pp. 1-26, figs. 1-12, pls. 1-11. KIAER, JOHAN, and HEINTZ, ANATOL 1935. The Downtonian and Devonian vertebrates of Spitsbergen. V. Sub- order Cyathaspida. Part I. Tribe Poraspidei, Family Poraspidae Kiaer. Skr. Svalbard Ishavet, 40, pp. 1-138, figs. 1-57, pls. 1-40. KJELLESVIG-WAERING, ERIK N. In press. Eurypterids of the Devonian Holland Quarry Shale of Ohio. Field- iana: Geology. LERICHE, MAURICE 1926. Deuxiéme note complémentaire sur les “‘ Pteraspis’’ de Wihéries (P. du- nensis F. Roemer). Bull. Soc. Belg. Géol., 35, pp. 19-29, figs. 1-3, pls. 1-4. DENISON: HOLLAND QUARRY SHALE FISHES 613 PEYER, BERNHARD 1946. Die schweizerischen Funde von Asteracanthus (Strophodus). Schweiz. Palaont. Abh., 64, no. 5, pp. 1-101, figs. 1-33, pls. 1-11. SCHMIDT, WOLFGANG 1959. Grundlagen einer Pteraspiden-Stratigraphie im Unterdevon der Rhein- ischen Geosynklinale. Fortschr. Geol. Rheinland u. Westfalen, 5, pp. 1-82, figs. 1-13, pls. 1-4. STENSIO, ERIK A. 1945. On the heads of certain arthrodires. II. On the cranium and cervical joint of the Dolichothoraci (Acanthaspida). K. Svenska Vetenskapsakad. Handl., (3), 22, no. 1, pp. 1-70, figs. 1-14. 1958. Les cyclostomes fossiles ou ostracodermes. Jn GRASSE, P.-P., Traité de Zoologie, 13, Fasc. 1, pp. 178-425, figs. 107-219. WHITE, ERROL I. 1935. The ostracoderm Piteraspis Kner and the relationships of the agnathous ie ie la Phil. Trans. Roy. Soc. London, (B), 225, pp. 381-457, figs. 1-97, pls. 25-27. 1950. Pteraspis leathensis White, a Dittonian zone-fossil. Brit. Mus. (Nat. Hist.), Bull. (Geol. Ser.), 1, pp. 69-89, figs. 1-25, pl. 5. 1956. Preliminary note on the range of pteraspids in western Europe. Bull. Inst. Roy. Sci. Nat. Belgique, 32, no. 10, pp. 1-10, figs. 1-8. Gt