a new temnospondyl amphibian from the late ... - GeoScienceWorld

0 downloads 0 Views 185KB Size Report
Feb 22, 2000 - nomen dubium. Characters of R. cosgriffi include its small size combined with relatively small laterally-facing orbits, relatively high skull, lack of ...
J. Paleont., 74(4), 2000, pp. 670–683 Copyright 䉷 2000, The Paleontological Society 0022-3360/00/0074-0670$03.00

A NEW TEMNOSPONDYL AMPHIBIAN FROM THE LATE TRIASSIC OF TEXAS JOHN R. BOLT

AND

SANKAR CHATTERJEE

Department of Geology, Field Museum of Natural History, Roosevelt Road at Lake Shore Drive, Chicago, Illinois 60605, ⬍[email protected]⬎, and Museum of Texas Tech University, Lubbock, Texas 79409, ⬍[email protected]⬎ ABSTRACT.—A skull representing a new genus and species of Late Triassic temnospondyl, Rileymillerus cosgriffi, is described from the Cooper Canyon Formation, upper Dockum Group of Garza County, Texas. Rileymillerus resembles Latiscopus disjunctus in size and proportions, but the very poorly preserved unique type specimen of L. disjunctus indicates that the taxon should be considered a nomen dubium. Characters of R. cosgriffi include its small size combined with relatively small laterally-facing orbits, relatively high skull, lack of lateral line canals, lateral exposure of the palatine on the skull surface, and lack of otic notch/quadrate angle. No postcranial material can be definitely associated, although we describe a partial vertebral column that might pertain to R. cosgriffi. Relationships of R. cosgriffi are uncertain. The possibilities of a close relationship to Almsauridae, Tupilakosauridae or (especially) Brachyopoidea are explicitly examined, but for the present we consider R. cosgriffi as Temnospondyl incertae sedis. Characters described in the text have been converted to the tripartite (part, feature, state) standardized format developed for the PRESERVE web site, and are presented as a 125-character data matrix in the Appendix.

INTRODUCTION

of western Texas and adjacent New Mexico has long been an important source of data regarding Late Triassic geology and paleontology. Vertebrate fossils of the Dockum represent numerous supraspecific groups of fish and tetrapods (Murry, 1986), although historically most interest has been focused on amniote groups. Anamniote tetrapods, all temnospondyl amphibians, occur at many Dockum sites. Almost all belong to one of three species of Metoposauridae (Hunt, 1993). The sole known exception until recently was the small skull (overall length about 35 mm) described by Wilson (1948) as Latiscopus disjunctus. This specimen comes from ‘‘Otis Chalk Quarry 1’’ of Gregory (1945), about five kilometers north of the town of Otis Chalk, Howard County, Texas. The associated fauna at this site includes mostly the archosauromorph Trilophosaurus (Gregory, 1945). Latiscopus, though clearly a temnospondyl and equally clearly not a metoposaur, has proven difficult to associate with any other amphibian family, surely because of the poor condition of the unique type skull. However, given the productivity of the Dockum and the continuing vertebrate paleontological collecting activity in it, eventual discovery of additional Latiscopus material has always seemed likely. That expectation has now been more or less realized: we here report on another, very Latiscopus-like, skull from the Dockum. As discussed below, ‘‘Latiscopus-like’’ is the best comparison we can make with the type, because of the poor condition of the latter. The new skull comes from Chatterjee’s (1983) Post Quarry, a site about 15 km southeast of Post, Garza County, Texas, and roughly 100 km north of Otis Chalk Quarry 1. The Post Quarry has produced a very different and exceptionally interesting terrestrial vertebrate fauna (e.g., Chatterjee 1985, 1986a, 1986b) which includes the putative early bird Protoavis texensis (Chatterjee, 1991, 1997, 1998).

T

HE DOCKUM GROUP

MATERIALS AND METHODS

TTU P 9168 is a single skull, preserved in a matrix of soft red mudstone with veins of calcite. Preservation of the bone and its surface is generally good, with no evidence of water wear although the bone is soft and easily abraded. The polished areas now visible are the inescapable result of handling during study. The skull is incomplete and somewhat crushed. About one-quarter of the right lower jaw was found, displaced into the intermandibular area and closely associated with the right palatine. The two were removed as a unit, then separated and prepared

individually. The left lower jaw was left in place to strengthen the specimen, which otherwise is supported only by poorly consolidated matrix. Preparation was mostly by needle and pin vise. The matrix in some areas was softened by application of 10 percent acetic acid in an inert paste; the loosened matrix was then removed mechanically. During preparation, the skull was supported in Carbowax (polyethylene glycol), which was later removed by washing with water. Abbreviations for institutions: TTU, Texas Tech University; TMM, Texas Memorial Museum, University of Texas. SYSTEMATIC PALEONTOLOGY

Order TEMNOSPONDYLI Zittel, 1888 Genus RILEYMILLERUS new genus Type species.—Rileymillerus cosgriffi new species. Diagnosis.—Small temnospondyl (estimated midline length from posterior extremity of postparietals on skull roof to tip of premaxillae ca. 35 mm), orbits laterally directed and approximately 20 percent of estimated skull length (orbit diameter ca. 7 mm). Exoccipital condyles distinctly double, stalked, slightly dorsal and posterior to quadrate condyles in lateral view. No squamosal embayment or tabular horn; in occipital view, line of junction between occiput and skull roof/cheek is a smooth curve from quadrate to midline. Quadratojugal and squamosal have an extensive occipital exposure, with no squamosal-quadratojugal trough. The medial edge of the occipital flange of the squamosal has no contact with the lamina ascendens of the pterygoid. The lamina ascendens arises from the dorsal surface of the pterygoid at the level of the basipterygoid joint. It is oriented in an approximately coronal plane; its lateral margin lies anterior to the occipital flange of the squamosal. The lacrimal is absent. The palatine is exposed on the skull surface as a dermally sculptured lateral exposure of the palatine at the anteroventral corner of the orbit. The anterior margin of the palatal ramus of the pterygoid projects anteriorly into the interpterygoid vacuity, at an inflection whose anterior edges meet at an obtuse angle of about 210 degrees. Etymology.—The generic name refers to Riley C. Miller, who generously permitted the junior author to collect at the Post Quarry. The species is named for the late John Cosgriff, a lifelong student of Triassic temnospondyls. Occurrence.—Post Quarry, Lat. 33⬚31⬘17⬙N; Long. 101⬚18⬘54⬙W, 14.5 kilometers southeast of Post, R. C. Miller

670

BOLT AND CHATTERJEE—NEW TRIASSIC TEMNOSPONDYL FROM TEXAS

671

FIGURE 1—TTU P 9168, holotype skull of Rileymillerus cosgriffi. 1, Dorsal view, anterior to right; 2, ventral (palatal) view; 3, left lateral view; 4, right lateral view; 5, occipital view. Scale bars ⫽ 1 cm.

Ranch, Garza County, Texas. Cooper Canyon Formation, upper Dockum Group, early Norian, Late Triassic. RILEYMILLERUS COSGRIFFI new species Figure 1 Diagnosis.—As for genus, by monotypy. Discussion.—In size and overall shape, Rileymillerus most closely resembles Latiscopus disjunctus Wilson, 1948. L. disjunctus is a small temnospondyl from the Dockum Group of Howard County, Texas. It is known from a single specimen that Wilson made the basis for the monotypic family Latiscopidae. Latiscopus is slightly smaller than Rileymillerus (postorbital skull table length as measured in the midline is ca. 14 mm for Latiscopus, 18 mm for Rileymillerus). But overall, the size and shape of Latiscopus, including small laterally facing orbits and absence of otic notches, are very similar to those of Rileymillerus. This combination of features is otherwise unknown among North American Triassic temnospondyls, and is rare in temnospondyls in general. In view of this similarity, and their geographic and stratigraphic proximity, the types of the two genera were compared in detail.

TMM 31025–182, the holotype of Latiscopus disjunctus, consists of a skull with lower jaws in place, missing the tip of the snout. None of the original surface of the bone is present, and over most of the skull and lower jaws the bone is missing entirely. In occipital view, all of the dermal bones bordering the occiput are broken, so that none of the original bone margins are present. All of these bones are now exposed as cross sections through the posteriormost part of the skull table and cheeks, so it cannot be determined whether any of them originally had occipital flanges. The quadrates and the posterior ends of the lower jaws are broken and exposed in section. The exoccipital condyles are missing, and the ascending columns of the epipterygoids are preserved in vertical section. From Wilson (1948), we suppose that most of this damage was due to overenthusiastic preparation. However it occurred, the result is a specimen that is virtually impossible to compare in detail with any other. The sole point in which detailed comparison can be made with Rileymillerus is in the nature of sutures, a few of which are partially preserved on the skull roof and were figured by Wilson. They are all highly interdigitated, unlike those of Rileymillerus.

672

JOURNAL OF PALEONTOLOGY, V. 74, NO. 4, 2000

Wilson (1948) makes very modest claims regarding the quality of the specimen, but his diagnosis and description nonetheless give a misleading impression of what can actually be seen. The diagnosis is: ‘‘Skull small, highly arched, triangular, not ornamented; parasphenoid broadly joined to the pterygoids, cultriform process broad, occipital condyles double(?), teeth conical, large pulp cavity; [tabular] cornua absent, otic notch very small, orbits large, lateral and dorsally directed’’ (Wilson, 1948, p. 359). Our comments on the diagnosis are: 1) None of the original surface is present, so presence or absence of ornamentation is undeterminable. 2) A continuous sheet of bone extends from one pterygoid to another across the basis cranii, so it is safe to say that the pterygoid and parasphenoid were in contact and the basipterygoid joint was not moveable. But the original surfaces of the pterygoid and parasphenoid are not preserved, the two bones are not distinguishable from one another, and no suture between them is visible. 3) Damage to the occipital condyle(s) makes it impossible to determine whether there was one condyle or two. 4) The shape of the teeth and presence of a large pulp cavity would not distinguish this species from any number of others, and in any case the pulp cavity in the few teeth where it can be seen is not particularly large. 5) Presence or absence of cornua (⫽tabular horns) cannot be determined due to damage. 6) There is no indication of the presence of an otic notch or squamosal embayment. 7) The orbits are not large; on the contrary, considering the size of the skull, they are remarkably small. For Latiscopus, maximum anteroposterior orbit diameter is ca. 5.5 mm; for Rileymillerus, it is estimated (the orbit is slightly deformed by crushing) as ca. 7 mm. As a percentage of postorbital skull table length, the orbit diameter in Latiscopus is about 3.6 percent, that of Rileymillerus about 3.9 percent. For all practical purposes, they are identical. Wilson’s description also contains a number of questionable statements about Latiscopus, which we would amend as follows. There is no evidence that the parasphenoid covered the entire ventral surface of the exoccipitals because there is only matrix, rather than bone, between the anteriormost preserved portion of the exoccipitals (the base of the condyle(s)) and the posterior part of the parasphenoid as preserved. Nothing of the cultriform process is preserved except the (rapidly tapering) posteriormost part of its base; the proportions of the cultriform process are therefore unknowable. Besides this remnant of the cultriform process and the adjacent, badly damaged left palatal ramus of the pterygoid, none of the bones bordering the interpterygoid vacuities is visible, and the palate has been excavated far enough to render it unlikely that any others are preserved; the size and shape of the interpterygoid vacuities are thus conjectural. The remnant of the palatal ramus of the pterygoid is in an approximately horizontal plane, while the quadrate rami follow the same course as in Rileymillerus. The relationships of the lamina ascendens are poorly shown, on the left side only, and it is impossible to say whether it reached the skull roof dorsally, although it might have reached the cheek. It is oriented like the lamina ascendens of Rileymillerus. The base of a large labyrinthine tooth, presumably a vomerine fang, is partly exposed on the right side near the tip of the snout. A number of workers have commented on the relationships of Latiscopus, generally without having seen the specimen. With the type in hand, it is clear that, due to its extremely poor preservation, Wilson’s original assignment—to ‘‘Stereospondyli’’— is about the best that can be done. The morphology of the few

sutures preserved in Latiscopus suggests at least a specific difference from Rileymillerus, in view of their similar size. But given the condition of the Latiscopus type, no more detailed comparison with any other specimen is likely ever to be possible. In view of the inadequacy of the type, we consider Latiscopus disjunctus to be a nomen dubium. DESCRIPTION

Skull reconstruction.—We here deal with reconstruction of general skull features. Details are given under the appropriate headings below. The overall condition of the skull as preserved is shown in Figure 1. In dorsal view, the skull is seen to be somewhat skewed, with its entire left side including the occiput pushed anteriorly relative to the right and crushed inward toward the right as a unit. Crushing has broken the left mandible, and displaced it toward the midline. Because of crushing, the original dorsal outline of the lateral skull margin is uncertain. The straight-sided reconstruction in Figure 2 is a compromise outline. Proportions of the postorbital region, in all views, were taken mainly from the relatively undistorted right side. Proportions of the antorbital region were perforce taken from the left side, but corrected for skewing and bone displacement as far as possible by reference to the right postorbital region. Dermal skull roof.—Ornament of the usual temnospondyl pitand-ridge type is present on most of the dermal roofing bones (Figs. 1, 2.1). The ridges are somewhat antero-posteriorly elongate on the frontal, parietal and dentary, and radiate outward and upward from a common center near the lower edge of the angular; but there are no ‘‘zones of intensive growth’’ such as observed in some other temnospondyls (Bystrow, 1935). This is consistent with both the small size of the skull and its ‘‘normal’’ proportions with the orbits near the center of the skull. There are no traces of lateral line canals either on the skull roofing bones or on the mandible. The right frontal has rotated sharply downward approximately around its suture with the parietal, and has rotated outward about 45⬚ around its long axis. Much of the left side of the skull roof has been displaced to the right and overrides part of the right side. Displacement is greatest anteriorly, and diminishes posteriorly to disappear at the approximate level of the supratemporalpostparietal suture. As a result, the lateral part of the left parietal is progressively overridden from back to front, and the right frontal is almost invisible in dorsal view. Sutures are difficult to trace, due to a network of cracks that extends through every one of the roofing bones. Most of the determinable sutures are visible only because the bones on either side have partly separated. Besides making it possible to trace the suture, this separation displays sutural morphology in three dimensions. The exposed sutures are often bevelled, with one of the adjacent bones sending a flange under the other, but they are not complex: in external view suture lines are fairly straight to sinuous. The midline suture is easily traceable for most of its length. The small parietal foramen is well preserved; its greatest diameter represents no more than 2.5 percent of the estimated quadrate-premaxilla distance measured parallel to the midline. The anterior end of the right parietal forms a sinuous outline and is apparently unbroken, so we locate the suture here; it is not in contact with the displaced and tilted right frontal. Around the right orbit, only the jugal and postorbital preserve their orbital surfaces, and the jugal-postorbital suture is the only one determinable. Most of the circumorbital sutures shown in Figure 2.1 and 2.3 were taken from the left side, where the orbital rim is little disturbed and the bones are intact around its circumference except for a small ventral gap, clearly due to breakage,

BOLT AND CHATTERJEE—NEW TRIASSIC TEMNOSPONDYL FROM TEXAS

673

FIGURE 2—Rileymillerus cosgriffi skull reconstructions. 1, Dorsal view; 2, ventral view; 3, left lateral view. Scale bars ⫽ 1 cm. Abbreviations for this and subsequent figures: An ⫽ angular, Ar ⫽ articular, De ⫽ dentary, Ec ⫽ ectopterygoid, Ex ⫽ exoccipital, fct ⫽ foramen for chorda tympani, Fr ⫽ frontal, fs ⫽ unnamed small foramen, Ju ⫽ jugal, LAS ⫽ lamina ascendens, LEP ⫽ lateral exposure of palatine, Mx ⫽ maxilla, Na ⫽ nasal, Pa ⫽ parietal, Pal ⫽ palatine, Pof ⫽ postfrontal, Ppa ⫽ postparietal, Pra ⫽ prearticular, Prf ⫽ prefrontal, Ps ⫽ parasphenoid, Pt ⫽ pterygoid, Ptq ⫽ pterygoid quadrate ramus, Qj ⫽ quadratojugal, Qu ⫽ quadrate, Sa ⫽ surangular, Sq ⫽ squamosal, Ta ⫽ tabular, Vo ⫽ vomer, * ⫽ pterygoid quadrate ramus dorsal division.

674

JOURNAL OF PALEONTOLOGY, V. 74, NO. 4, 2000

between the prefrontal and jugal. The prefrontal-postfrontal suture is reasonably clear on the dorsal surface and can be traced into the orbital rim. From this point down to the prefrontalmaxilla suture the anterior orbital wall is interrupted only by a crack. The postfrontal-postorbital suture is not definitely determinable within the orbit, although a small part of that suture is visible distal to the orbit. The jugal-postorbital suture is clearly visible both externally and within the posterior orbital wall. The orbital face of the right jugal shows a deep recess for the ventral extremity of the postorbital. The position of the jugal-quadratojugal suture is conjectural. Beneath the orbit the jugal evidently ended as a more or less sharp-pointed process, the anterior tip of which is broken off. The orbital rim is completed in the reconstructions by restoring the short anterior extension of the jugal and a ventral extension of the prefrontal, separated by a small lateral exposure of the palatine. There is no evidence for an exposure of the maxilla within the orbit. The posterior end of the left maxilla is missing. Due to partial or complete absence of the maxilla along the jugal-maxilla suture, the jugal part of the sutural surface is visible in ventral view on both sides of the skull as a shallow groove. The groove ends just behind the orbit, marking the posterior terminus of the maxilla; immediately posterior to this point the jugal forms a smoothly rounded edge, which thins posteriorly. The strongly striated ventral surface of the alar process is visible in Figure 1.2, immediately adjacent to the posterior (left) end of the scale bar. It is not visible in any of the reconstructions, because it is covered by the pterygoid or ectopterygoid. Anterior to the orbit few sutures can be traced with confidence. The anterior ends and the lateral borders of both frontals are clear. A large flange projecting from the anterolateral corner of each frontal underlay part of the nasal. The sutures bordering the prefrontal are reasonably clear. A short segment of the nasalprefrontal suture is tentatively identified on the left, indicated by a short flange from the nasal which originally underlay the prefrontal. The nasal is clearly incomplete, and we can find no trace of the external narial opening. There is no indication of a lacrimal, and we conclude that it is absent. The left maxilla is separated by a conspicuous gap from the bones dorsal to it. We interpret this gap as a separation along a suture, so the maxilla has no preorbital dorsal process. Lying within the gap just anterior to the orbit and between the prefrontal and maxilla is a tiny area of sculptured bone about 0.5 mm high and just over 2.0 mm long. It is closely associated with the lateral edge of the palatine, a small part of which is visible in the anterior floor of the orbit. On this evidence as well as the lateral face of the separated right palatine, this small piece of sculptured bone is restored as a lateral exposure of the palatine (LEP of Bolt, 1974), which forms a small segment of the anterior orbital rim (Fig. 2.3). The ventral margin of the prefrontal immediately anterior to the orbit and dorsal to the LEP bears a shallow groove. The anterior extremity of the maxilla is likely missing because the preserved tip appears to be neither sutural nor part of the external narial border. Occiput.—Although the left suspensorium is missing, the occiput is otherwise well preserved, with relatively minor distortion due to crushing. Thus the reconstruction in Figure 3 represents the original condition of the skull quite closely. There is no squamosal embayment or tabular horn. As preserved, the proportions of the paroccipital process of the exoccipital and the size of the posttemporal fossa differ markedly between the left and right sides. On the left, the paroccipital process curves smoothly upward to its junction with the tabular; on the right, the process is a linear buttress between the base of the vertical column and the tabular, and is shorter than that on the left. As we interpret the situation, the curvature

of the left paroccipital process had been exaggerated by crushing, which also resulted in enlargement of the left posttemporal fossa. On each side, because of lateral crushing, the base of the vertical column of the exoccipital has broken away from its original attachment point and now lies medial and slightly anterior to it. The base of the right exoccipital has been displaced further medially than that of the left. The right quadrate region has likely been crushed slightly medially. This is consistent with the observed elevation of the right tabular region above the adjacent surface of the skull roof, presumably due to its ventral connection with the paroccipital process and the latter’s resistance to lateral crushing. The left tabular region, lacking the paroccipital brace, is bent strongly downward compared with the right, and the tabular is not elevated above the surface. We have moved the quadrate region outward in the reconstruction, to accommodate restoration of the exoccipital columns to their original positions. Reconstruction of the paroccipital processes and the posttemporal fenestrae is based mostly on the right side. Taking the skull roof as a reference, the left exoccipital condyle is lower than the right one, and slightly anterior to it. The basicranial unit is tilted downward on the left side, which is consistent with the stronger downward curvature of the left tabular region. There is evidence that the left condyle has been displaced forward through skewing of the entire basicranial region: the absence of the left suspensorium creates a ‘‘window’’ through which much of the left pterygoid is visible. The palatal ramus is broken just anterior to the ascending process, and foreshortened by more than a millimeter where the anterior part of the ramus overrides the posterior. Taken together, these observations suggest that the left exoccipital condyle has moved down and forward. In the reconstructions, therefore, the left condyle and pterygoid are moved back and up to match the right side. The right and left exoccipitals are separated by a small unossified gap (basioccipital fenestra of Bystrow and Efremov, 1940) that extends forward several millimeters, as far as the foramen magnum has been cleaned. This space represents the position of the basioccipital, which was evidently unossified in this area because there is no indication in occipital view of a basioccipital-exoccipital suture. The basioccipital fenestra is partly roofed by a submedullary process from each exoccipital. Upward crushing of the parasphenoid and basisphenoid (see below) likely moved the exoccipitals inward, but there is no indication of any large displacement. The reconstructions therefore show the condyles only slightly farther apart than as preserved. The ascending column of the left exoccipital is so cracked and crazed as to be distinguishable with difficulty from an occipital flange of the postparietal. The right ascending column is well preserved except for its dorsal portion. The occipital surface of the column is flat and nearly featureless, (Fig. 3), with no evidence of a distinct facet that might have received a flange from the postparietal. The apparent presence of such a flange on the left side is misleading, both because it is too broad for the flat part of the ascending column and because its antero-posterior thickness as seen through the left posttemporal fenestra is more appropriate to the ascending column than to the postparietal. The ascending column does not bear a distinguishable inwardly directed lamellar process, although damage to the dorsal ends of both exoccipitals might have destroyed it. Between the top of the right ascending column and the border of the postparietal there is a 1 mm interval of matrix plus bone fragments. Although these fragments are shown as part of the ascending column in our reconstruction, this is conjectural as is the position of the exoccipital-postparietal suture. The proximal part of the right paroccipital process is covered by bone fragments. Their source is unclear, although all those visible in occipital view are thin

BOLT AND CHATTERJEE—NEW TRIASSIC TEMNOSPONDYL FROM TEXAS

FIGURE 3—Rileymillerus cosgriffi skull reconstruction, posterior (occipital) view. Scale bar ⫽ 1 cm.

enough to suggest that they are not part of the paroccipital process itself. The paroccipital process is composed solely of exoccipital; there is no sign of an ossified opisthotic. Lying alongside the body of the exoccipital and just ventral to the origin of the paroccipital process is an unidentified short rodlike piece of bone. Its expanded ‘‘head’’ is in the plane of the paroccipital process but covered by the ventral-most of the displaced fragments. The same area of the left exoccipital is plainly visible and shows no trace of such a structure. It may or may not be part of a stapes; no stapedial foramen is visible. The position of the exoccipital-pterygoid suture is obscure because of extensive cracking and some shifting of bones in the basicranial region. The suture between the paroccipital process and the tabular is indicated in Figure 3; on both sides there is a suture (or break?) in this region, but there is no plausible suture anywhere else on either process. Granted this distal location for the suture, the left paroccipital process meets a short parotic process from the tabular. As preserved, this process appears to be absent on the right tabular, but we presume this is attributable to crushing. The area of the right tabular in contact with the paroccipital process has been displaced outward along the line of the process, and the parotic process may as a consequence have been displaced, shortened, or both. The presumed postparietal-tabular suture is exposed in occipital view on the right side, although it cannot be traced onto the skull table (Fig. 3). The occipital face of the right suspensorium is beautifully preserved, and sutures are clear. The occipital face of the squamosal and quadratojugal is unsculptured and flat to slightly convex, but never concave. There is thus no squamosal-quadratojugal trough in the sense of Warren and Black (1985, p. 304), as ‘‘a vertical concavity of the supraquadrate area . . . composed of squamosal, or squamosal and quadratojugal, or both . . . combined with the quadrate.’’ Both squamosal and quadratojugal make a smooth transition from cheek to occiput. There is no crista falciformis of the squamosal, defined by Warren and Hutchinson (1988, p. 858) as ‘‘a flattened flange of bone on the otic-occipital margin of the squamosal, which projects toward the tabular horn’’ (or simply posteriorly, in cases such as Rileymillerus where there is no tabular horn). The free medial edge of the occipital flange of the squamosal appears to be essentially undamaged and undistorted. The suture between the quadrate and the quadrate ramus of the pterygoid is partly visible in occipital view. At the dorsal end of this suture is a small fossa, broadly open above and apparently developed in the quadrate. There is no sign of a paraquadrate foramen in either occipital or lateral view. The occipital flange of both the squamosal and the quadratojugal is pierced by a small foramen, but in either case this is too small to be a paraquadrate foramen.

675

The quadrate ramus of the pterygoid is comprised of ventral and dorsal divisions. The ventral division may be thought of as the quadrate ramus proper. It arises from the pterygoid at the level of the quondam basipterygoid joint. It runs posteriorly and ventrally as a massive buttress whose heavily striated lateral face, visible on the left side of the skull, is sutured to the medial side of the quadrate. It does not project posteriorly beyond the quadrate as a free flange. The dorsal division of the quadrate ramus can be thought of as the dorsal free margin of the ventral division. The two divisions, oriented at nearly 90⬚ to one another, meet in a straight, sharp edge which is especially pronounced posteriorly. Their relationship, unfortunately preserved only on the right side of the skull, appears to be undisturbed. As preserved, the dorsolateral and posterior borders of the dorsal division are free, and this portion of the quadrate ramus does not make contact with any other bone. The space enclosed between the quadrate, the dorsal (and a small part of the ventral) division of the quadrate ramus, and the medial edge of the squamosal may in life have been filled by a cartilaginous continuation of the quadrate. The occipital face of the quadrate shows no sign of the ‘‘hyoid tubercle’’ (Shishkin et al., 1996) found there in some other Triassic temnospondyls. The lamina ascendens pterygoidei has a distinctive relationship to the quadrate ramus. In most Triassic temnospondyls, the lamina ascendens represents a dorsal continuation of the quadrate ramus. If a transition in orientation of the lamina ascendens occurs, from the roughly anteroposterior orientation of the quadrate ramus, to a transverse orientation of the medial portion of the lamina, it is generally a gradual one; see diagrams in Warren and Black (1985). This is not the case in Rileymillerus, where the lamina ascendens arises abruptly from the dorsal surface of the pterygoid at the anterior end of the quadrate ramus. There is no smooth transition between quadrate ramus and lamina ascendens. The lamina ascendens is thin and almost perfectly vertical. In dorsal view, it runs from anteromedial to posterolateral, its posterior surface making about a 70⬚ angle with the sagittal plane, and its medial portion curving slightly posteriorly. This may be called the ‘‘definitive’’ lamina ascendens. The medial edge of the lamina is difficult to see. However, the posterior surface of the right lamina can be seen well enough to suggest that a thickened ‘‘dorsal column’’ (Warren and Black, 1985) is absent. From its base, the lamina fans out dorsally in both medial and lateral directions. The dorsomedial extremity of the lamina is about 2.5 mm from the skull midline, and just posterior to the narrow base of the cultriform process. The lamina’s dorsolateral extremity is missing on the left side. As preserved, both lamina ascendens fail to reach the underside of the dorsal skull roof; this is likely their original condition, considering the good preservation of the right side of the skull. The lateral edge of the right ascending ramus is in contact with dermal bones of the cheek–the squamosal and possibly the tabular. The squamosal appears to send a flange down along the posterior face of the ascending ramus; the approximate borders of the flange(s) from the squamosal and tabular(?) are shown in Figure 3. This squamosal flange is deep (anterior) to and distinct from the free medial margin of the squamosal’s occipital flange. Unsurprisingly in view of the manner of the lamina ascendens’ origin, no lamina obliqua, which in some Triassic temnospondyls marks the boundary between quadrate ramus proper and the lamina ascendens (Warren and Black, 1985), can be identified in Rileymillerus. The presence or absence of a palatoquadrate fissure, and details of its extent, are considered by a number of authors to be important characters in Triassic temnospondyls. Following Warren and Black (1985) and Damiani and Warren (1996), the palatoquadrate fissure may be defined as a gap between the occipital process of the squamosal or the combined squamosal and

676

JOURNAL OF PALEONTOLOGY, V. 74, NO. 4, 2000

quadratojugal and the posterolateral margin of the lamina ascendens of the pterygoid. Damiani and Warren (1996, p. 295) suggested that presence of a palatoquadrate fissure reflects ‘‘failure of the quadrate to ossify fully so that a gap is present between the quadrate and ascending ramus of the pterygoid and the adjacent occipital exposures of the squamosal and quadratojugal.’’ This definition of the palatoquadrate fissure appears to reflect the general usage of the term in the literature. In Rileymillerus, the transversely oriented, definitive lamina ascendens is not in a position to form the medial border of the palatoquadrate fissure. We therefore consider that the term ‘‘palatoquadrate fissure’’ is not applicable to Rileymillerus. Palate.—Dermal bones of the palate that are present in the prepared specimen include the left vomer, palatine and ectopterygoid, substantial portions of both pterygoids including their basipterygoid regions, and the parasphenoid including the cultriform process. The right palatine was removed from the intermandibular area and prepared separately. The left vomer and palatine have apparently been shifted very little from their original locations, except that their medial edges have been crushed ventrally along with the anterior end of the right frontal. This now lies directly above the posteromedial extension of the vomer, from which it is separated by a piece of bone apparently broken off from the cultriform process of the parasphenoid. The rest of the cultriform process slopes gently downward as it runs posteriorly from the broken piece. The angle of slope may have been accentuated by crushing that displaced the anterior end of the cultriform process dorsally. In other respects the cultriform process appears to be close to its original orientation. The left lower jaw has been crushed upward and inward so that it covers the lateral part of the palate, especially the ectopterygoid and posterior portion of the palatine. Much of the palatal ramus of the left pterygoid posterior to the jugal processus alaris can be seen. Absence of the left cheek and suspensorium made it possible to prepare the dorsal surface of this part of the pterygoid. Its medial edge and the medial portion of the ventral surface were prepared in the narrow space remaining between the cultriform process and the inner margin of the left mandible. The ventral surface of the right palatal ramus can be seen where it extended between the basipterygoid region and the processus alaris of the jugal. This part of the palatal ramus had broken off just anterior to its origin from the basipterygoid region, and is now lying free in the matrix near its original location. Its anterior end is adjacent to the alar process of the jugal, on which it is now supported by a short length of copper wire cemented in place. These posterior parts of the palatal rami, the basipterygoid regions, and the quadrate rami as described above, are the only identifiable areas of the pterygoids. The more anterior part of the right pterygoid is missing. Most of the palatal ramus of the left pterygoid may be preserved, but it is hidden by the left lower jaw. With the exception of the pterygoid-quadrate suture, none of the sutures of the pterygoid with other bones can be determined with confidence. The posterior end of the exoccipital-pterygoid suture is visible on the left side. It is not shown in the reconstruction because given the condition of the parasphenoidal basal plate, it is uncertain whether this suture was originally visible in ventral view, or in other words, whether the parasphenoid ‘‘underplated’’ the exoccipitals. The posterior margin of the pterygoid is a continuous smooth curve from the quadrate ramus to the basipterygoid region; in other words, there is no sign of the pterygoid incisure found in tupilakosaurids (Shishkin, 1973; Warren, 1999). Only the ‘‘free’’ portion of the palatal ramus is visible, plus the small area that underlies the processus alaris (the free portion may be defined as the part of the palatal ramus extending between its base in the basipterygoid region and the

posterior border of the processus alaris). The pterygoid likely did not extend much anterior to this level. The medial margin of the separated right palatine is not sutural, and similarly there is no indication of a sutural surface on the left vomer and visible portion of the left palatine, where they border the interpterygoid vacuity. The palatal ramus of the pterygoid appears to have originally been in approximately the same plane as the vomer and palatine. The left palatal ramus is in that orientation now. The palatal ramus was flat, with no arching or vaulting. The anteromedial edge of the palatal ramus just posterior to the jugal processus alaris has a distinctive shape. The posterior and anterior portions of the edge meet here at an obtuse angle of nearly 210⬚. The distinctive shape of the pterygoid here is reflected in the outline of the interpterygoid vacuity. The left vomer and palatine are in contact with the maxilla in what appears to be their original anteroposterior location, although partly obscured by the anterior end of the lower jaw which lies directly below the internal naris. The anterior margin of the vomer appears to be intact laterally, but broken medially. The medial margin of the vomer is thin, and clearly damaged in places; we cannot be certain how much of it represents the original margin, although it must be very close to it. The medial suture with the palatine is clear. A high flange of the vomer runs upward along the inner side of the maxilla and extends to the ventral part of the nasal. The anteroposterior extent of this flange is unknown, as the posterior one-half of the dorsal surface of the vomer is matrix-covered. The vomer forms the entire anterior border of the internal naris, and the vomer and palatine together form its medial border. Their relationships along this border are clear from both the skull and the separated right palatine. Lateral to the internal naris, we have not been able to determine the details of the palatinevomer contact (if any) and their relationships to the maxilla. In the skull, this region is partly covered; in the prepared portion sutures are unclear, and the anterolateral portion of the separated right palatine is broken. In view of this uncertainty, we have not indicated any sutures along the lateral border of the internal naris (Fig. 2). The right palatine preserves the small, sculptured LEP (Fig. 2.3). Almost all of the area normally occupied by the ectopterygoid is hidden by the left lower jaw. The maximum anterior extent of the ectopterygoid, pterygoid, or both is indicated by a sutural area that spans the ventral surface of the right palatine just posterior to the last tooth. The maximum posterior extent of the ectopterygoid is marked by the processus alaris of the jugal because no marginal palatal bone extends posterior to that point on the (articulated) left side. No ectopterygoid-pterygoid suture is visible, possibly because it is covered by matrix. We nonetheless believe that an ectopterygoid is present because the alternative is to suppose a most unusual shape as well as dentition for the pterygoid. Anterior to the processus alaris, a short segment of the medial margin of the ectopterygoid(?) is visible. The parasphenoidal cultriform process is complete, and all but its anterior tip is exposed in ventral view. As shown in Figure 2.2, the ventral surface of the cultriform process is slightly and smoothly convex downward, with no keel. Posterior to the cultriform process, the parasphenoidal basal plate has been crushed upward into the basisphenoid region, resulting in reduction of the parasphenoid in this area to small plates, some of which have been displaced or lost. Consequently the sutures of parasphenoid with the pterygoid and exoccipital cannot be determined, and the overall shape of the basal plate is indeterminate. The ventral surface of the parasphenoid shows no sign of a transverse ridge or ‘‘pockets’’ of Watson (1962) for muscle insertion, although preservation here is so poor that we cannot be certain of their absence.

BOLT AND CHATTERJEE—NEW TRIASSIC TEMNOSPONDYL FROM TEXAS Lower jaw.—The separated right lower jaw fragment consists of approximately the central two-thirds of the dentary, plus an obviously displaced and incomplete bone on the medial side of the dentary. This bone bears two tooth bases on its dorsal border, and is thus likely a coronoid. The left lower jaw is nearly complete. It has been crushed inward and partly broken at approximately midlength, and there has been some consequent displacement of bones. The occlusal surface and the anterior one-half of the medial surface of the jaw are mostly obscured by matrix, which was left in place to support the anterior end of the jaw and protect the fragile vomerine fang. At least part of the symphysial region appears to be missing, as does a part of the postglenoid area. On the lateral face of the mandible, relations of most bones are reasonably clear. The dentary extends well posterior to the tooth row (Fig. 4.1). The relationships of angular, surangular and dentary to one another are partly known from preserved sutures, and partly inferred from sculpturing on the angular. There are short segments of preserved suture along the mostly eroded angular-surangular contact. At the level of the last dentary teeth, the dorsal margin of the surangular is poorly visible due to damage and to matrix coverage. For the same reasons, we cannot determine the extent and shape of the coronoid contribution to the lateral margin of the adductor fossa. From the relationship of the coronoid teeth seen in medial view, we suspect that there is no prominent ‘‘coronoid process’’ in Rileymillerus. The posterior extremity of the surangular as preserved carries two very large, deep pits that are connected beneath a thin bridge of bone. On the ventral face of the mandible, the groove for the splenials is well marked on the underside of the dentary, although parts of the splenials are missing or covered and the suture between pre- and postsplenial cannot be determined. On the medial aspect of the lower jaw, we were unable to locate any Meckelian fenestrae. This may simply be due to matrix coverage, inadequate preservation, or both. Only the posterior one-quarter of the medial side is sufficiently visible and well preserved to permit description. The chorda tympani foramen lies entirely within the articular (Fig. 4.3). It is clearly visible just ventral to the articular-prearticular suture, which is easily traceable for most of its extent but becomes obscure posteriorly. A vertical parting-line running ventrally from the anteroventral border of the foramen to the angular-prearticular suture is an obvious crack, even though it occupies the position of the articular-prearticular suture in many temnospondyls (cf. Jupp and Warren, 1986). Directly ventral to the chorda tympani foramen, a fenestra is developed within the angular-prearticular suture. The posterior border of this fenestra is missing due to breakage; anteriorly, the border has rounded edges. The fenestra is floored by bone of the prearticular. In dorsal view, little can be seen but the articular region and the lateral (surangular) wall of the adductor fossa. This view is of interest primarily because the articular seems, at first sight, to be bipartite (Fig. 4.2). However, the posterior half of this ‘‘bipartite’’ articular is a normal-appearing articular, and is definitely in place with a visible suture between it and the prearticular. The anterior half on the contrary does not demonstrate a normal-appearing suture along any of its contact surfaces. Its currently upward-facing surface, however, appears to be a joint surface. The most plausible explanation for these observations is that the bone lying anterior to the (true) articular is the left quadrate. The quadrate has thus dropped into the adductor fossa and its upper portion has rotated forward 180 degrees, so that the originally ventrally directed articular surface now faces dorsally. The triangular area of broken bone between articular and quadrate is likely a broken part of the latter. Perhaps partly because of damage associated with this unusual displacement of

677

FIGURE 4—Rileymillerus cosgriffi, reconstructions of posterior one-third of left lower jaw. 1, Medial view, anterior to left; 2, dorsal view, anterior to left; 3, lateral view, anterior to right; 4, posterior view, dorsal at top. Scale bars ⫽ 1 cm (1, 2) or 0.5 cm (3, 4).

the quadrate, the location of the articular-surangular suture is uncertain. A possible course for the suture is indicated by an indistinct groove that can be traced only as far as a small foramen (poorly visible in Fig. 4.2; labelled ‘‘fs’’ in Fig. 4.4). The articular itself is clearly in its original position relative to the

678

JOURNAL OF PALEONTOLOGY, V. 74, NO. 4, 2000

surrounding dermal bones. The glenoid fossa is thus at approximately the same dorsoventral level as the dentary (Fig. 4.1). Preparation of the incompletely preserved posterior tip of the mandible shows a recess in posterior view, surrounded mostly or entirely by dermal bones. The articular may or may not form part of the dorsal border of the recess, because sutures cannot be completely traced. It is likely that the recess was occupied by cartilage of the articular because, according to Jupp and Warren (1986), the posterior end of the articular is never ossified in Triassic temnospondyls. We cannot tell how much of the posterior ends of the prearticular, angular, and surangular around the recess may be broken, and how much is a natural termination. The small area of the recess, combined with the rapid tapering of the posterior end of the mandible, suggests that little of the postglenoid area is missing. Jupp and Warren (1986) distinguished two types of postglenoid area (PGA) in Triassic temnospondyls. Their two types of PGA may differ in up to nine characters. However, they warned that no one character is diagnostic for each PGA type, and that ‘‘a character possessed by one type of PGA may occasionally be present in the other’’ (p. 103). Such a shorthand typological system can be useful for characterizing species to which it clearly applies, but misleading when used for morphologies in which, perhaps, few of the features that define the ‘‘type’’ are present. The latter appears to be the case for Rileymillerus, which we are unable to place into either one of Jupp and Warren’s PGA types. Dentition.—The maxilla as preserved has a total of about 24 teeth plus tooth spaces, excluding the row of teeth in the ectopterygoid region that might or might not pertain to the ectopterygoid. The dentition is not well exposed on the left dentary, but there is room for about 35 teeth plus tooth positions there. The incomplete right dentary has 29 teeth plus tooth positions, most of which are occupied by teeth. The total number of marginal teeth plus tooth spaces on the dentary is thus likely to be in the range of 35 to 40. All of the marginal teeth are tall, slender, and sharp pointed, with bases circular in cross section and the crowns slightly recurved and compressed labiolingually. No keels or serrations are visible. There are no caniniform marginal teeth. However, several dentary teeth immediately anterior to the orbit are twice as tall as the maxillary teeth just posterior to them. There is thus a possibility that the dentary teeth are consistently larger than the opposing maxillary teeth, although the marginal dentition is not well enough preserved to make it certain that this is the case. Teeth that are broken near the base show labyrinthine folding. Five small presumed coronoid teeth are visible on the medial surface of the left lower jaw just anterior to the adductor fossa. Because most of the coronoid region is inaccessible, its dentition cannot be completely described nor can these teeth be assigned to a specific member(s) of the coronoid series. When compared with Figure 2.2, Figure 4.1 and Figure 4.2 suggest that the dentary tooth row extended somewhat farther posteriorly than the maxillary tooth row. This was not necessarily the case. Aside from the fact that Figure 2 and Figure 4 are reconstructions done independently from difficult material, the two small posteriormost teeth in Figures 4.1 and 4.2 were reconstructed as functioning teeth based on empty tooth positions. These positions may not have carried functioning teeth at the time the animal died. No palatal denticles were found. The vomer and palatine bear fang teeth, plus smaller teeth that are comparable in size to the adjacent marginal teeth. A single vomerine fang tooth is in place and intact at the anterior border of the internal naris. It is about twice as long as the largest marginal tooth, with a proportionately large basal diameter. This was part of an alternately replacing pair, as shown by the pit for the other fang immediately anterior to it. The pit is clearly associated with a labyrinthine

tooth, and part of the base of the intact fang shows weak labyrinthine folding. The vomerine fang is strongly recurved and has a faint anterior keel only, and no trace of serrations on the keel. In addition to the fang-pair, the vomerine dentition includes two series of marginal-sized teeth. One series extends posteriorly in an irregular row that begins medial to the anterior fang position and continues near the intervomerine suture. The other borders the medial margin of the internal naris, between the posterior fang position and the suture with the palatine (Fig. 2.2). Most of the left palatine dentition is obscured by the lower jaw. The separated right palatine preserves its entire dentition. There are two adjacent fang tooth positions on this palatine; the anteriormost, which is unoccupied, lies immediately posterior to the border of the internal naris. The posterior position is occupied by a fang tooth which is slightly smaller than the vomerine fang, less strongly recurved, and shows no anterior or posterior keel. Its base shows labyrinthine infolding, and is flattened where it adjoined the maxilla. The palatine dentition also includes a few teeth anterior and posterior to the fang pair (Fig. 2.2). The anterior teeth continue posteriorly the vomerine series along the medial margin of the internal naris. The teeth posterior to the fang pair are continuous with the reconstructed ectopterygoid tooth row. Figure 2.2 shows our interpretation of the ectopterygoid dentition. Members of the ectopterygoid tooth row are comparable in size to maxillary teeth. What can be seen of the ectopterygoid tooth row is visible where the posterior end of the maxilla is missing, beginning at a level just posterior to the anterior orbital wall. These teeth are pleurodont, attached to the medial surface of a pronounced ridge of dermal bone (not shown in Fig. 2.2), and the row ends at the same posterior level where the maxilla ended. The first impression thus is that this is part of the maxillary tooth row, driven inward by crushing. This is contradicted, however, by the fact that the teeth are attached to the ectopterygoid, and that the lateral surface of the ridge to which they are attached is not ornamented. Other interpretations are possible; the ‘‘ectopterygoid’’ could be the posterior end of the maxilla, in whole or part, with an extensive palatal flange that now occupies the usual position of the ectopterygoid. This is more plausible for the anterior portion of the ‘‘ectopterygoid’’ tooth row than for the posterior portion. Based especially on this posterior portion, we have reconstructed a row of ectopterygoid teeth. There is no sign of an ectopterygoid fang tooth, which may be present but covered. Vertebrae.—No postcranial material can be definitely assigned to Rileymillerus. However, some possible Rileymillerus vertebrae (TTU P 9170) were found at the site, in proximity to the skull but not definitely associated. These vertebrae are part of a mass of small bones that includes two partial and badly damaged lower jaws, ribs, and a number of unidentified bone fragments. The vertebrae are in three short articulated segments, which are separated by intervals of matrix and other bones. The articulated segments are not in perfect alignment with one another, but could be part of the same column. Fifteen centra are visible. Centra are notochordal, approximately 2.5 mm in diameter, and about one-half that in length. They thus give a strong impression of diplospondyly. Unfortunately, it is impossible to definitely determine the number of centra per vertebra. All of the neural arches are missing or inaccessible, and the few that are partly visible do not give enough information to settle the question. The nature of the neurocentral junction is obscure in all examples. The ribs, arranged en echelon along the vertebrae, are bicipital and curved, with a shaft roughly semicircular in cross section. Each rib extended over at least 10 centra; however, successive rib heads are not well enough preserved to enable us to determine the number of centra spanned.

BOLT AND CHATTERJEE—NEW TRIASSIC TEMNOSPONDYL FROM TEXAS On balance, we are inclined to accept these centra as belonging to diplospondylous vertebrae. However, diplospondyly is usually associated with aquatic anguilliform locomotion, which is inconsistent with the fact that the Post Quarry assemblage is entirely terrestrial and includes no fish remains. Despite this, we consider the centra most likely to pertain to Rileymillerus because none of the other known members of the Post Quarry fauna has diplospondylous centra. If these centra do represent Rileymillerus, they must belong to a smaller individual than TTU P 9168. The maximum distance across occipital condyles in the Rileymillerus skull is about 5.5 mm, which is more than twice the diameter of the centra. DISCUSSION

Relationships of Rileymillerus.—We would prefer to assess the phylogenetic relationships of Rileymillerus by incorporating our data into one or more of the published data matrices for temnospondyls and analyzing the resulting matrix with a standard phylogenetics package such as PAUP. This is not possible at present, for several reasons. Many of the phylogenetic studies in the literature use families as terminal taxa. The evidence for considering these families as monophyletic groups is thus unavailable from the published matrices, and there is no possibility of testing these family-level groupings by adding data from other taxa. Many of the characters themselves are poorly stated, and in particular they are often subjective. Such adjectives as ‘‘large,’’ ‘‘broad,’’ ‘‘well-developed,’’ may be meaningful for the author(s) of a study, but state assignments based on such characters are not reliably reproducible by other observers. Many character statements are unduly complex, brigading what should be several statements into a single character. These could in principle be broken down into several more simply stated characters. In practice, we found this impossible because we were confined to using literature descriptions to determine state assignments, and the literature was too often inadequate to permit reliable assignments. Finally, many of the published character tables dealing with Triassic temnospondyls include few or no Paleozoic taxa. Since there is no reason to assume a priori that the Triassic temnospondyl taxa form a monophyletic group, much of the published data is of little help in determining the position among temnospondyls as a whole, of a taxon which happens to be Triassic. We could of course have performed the sort of analysis we originally intended, using whatever characters were available in the literature. This seems a pointless exercise because, for the reasons just given, we feel that the result would be spurious precision. We have chosen instead to present a more realistic, though less ‘‘precise,’’ discussion of the relationships of Rileymillerus. In the course of this study several colleagues, including one of the reviewers, have suggested a possible relationship between Rileymillerus and the small temnospondyl Almasaurus habbazi from the Late Triassic of Morocco (Dutuit, 1972, 1976). The suggestion is tempting because Almasaurus is of comparable age and has similar proportions including relatively small orbits, and has also been difficult to place taxonomically. But Dutuit’s (1976) detailed description of Almasaurus does not reveal any characters that suggest particularly close relationship. Some characters of Almasaurus, such as the more ‘‘normal’’ relationships of the lamina ascendens of the pterygoid, absence of a lateral exposure of the palatine, presence of a lacrimal, and a large otic notch, suggest that Almasaurus and Rileymillerus are not closely related. Characters and descriptions in the literature indicate that the relationships of Rileymillerus may be with the Brachyopoidea (chigutisaurids plus brachyopids; see Damiani and Warren,

679

1996), and/or some of the families most closely related to it. Of the characters used by Damiani and Warren (1996) to investigate relationships of and within the Brachyopoidea, several that are clearly stated and of restricted distribution among temnospondyls appear to provide the strongest indications of relationships (our numbers): 1) Absence of lacrimal. This is restricted to brachyopoids, tupilakosaurids and rhytidosteids. 2) ‘‘Brachyopoid’’ lamina ascendens of the pterygoid. According to Damiani and Warren (1996, p. 294), this ‘‘refers to the form of the ascending ramus in brachyopids which differs from, and is hence derived with respect to, other temnospondyls in this analysis in that it arises from the dorsal surface of the pterygoid well anterior to its posterior margin.’’ This is the condition in Rileymillerus. It occurs in brachyopoids, tupilakosaurids, rhytidosteids, dvinosaurids, and saurerpetontids, according to Damiani and Warren (1996). The significance of this character is uncertain, because the lamina ascendens of Rileymillerus differs from that of other temnospondyls, including brachyopoids, in that it is entirely anterior to the occipital part of the squamosal. 3) Lateral exposure of the palatine (LEP). Distribution of this is not completely known, but Damiani and Warren considered that it is found in at least some brachyopids, and tupilakosaurids (see also Warren, 1999). We agree with Damiani and Warren that the LEP is entirely a part of the palatine, rather than a ‘‘palatolacrimal’’ as suggested by Shishkin (1973). 4) Lamina ascendens of pterygoid lacks a dorsal column. Considered by Damiani and Warren to be a brachyopid synapomorphy. We do not use the ‘‘vaulted’’ palate as a character in this discussion, even though by some definitions it is present in Rileymillerus. The problem is that it has been inconsistently defined and, presumably, inconsistently used. Damiani and Warren (1996, p. 294) stated it thus: ‘‘pterygoid downturned forming deep U-shaped palate . . . refers to the sharp downturning of the lateral margins of the pterygoid to form a deeply vaulted palate.’’ They found this character in brachyopids and a number of related families. Warren and Black (1985, p. 305) defined and discussed a similar character as follows: ‘‘Brachyopoid palate . . . 9–Quadrate ramus of pterygoid downturned . . . . The characteristic shape of the brachyopid palate is well known (Watson, 1956; Welles and Estes, 1969) and was shown to be present also in the chigutisaurs by Warren (1981). It is formed by a downturning of the quadrate ramus of the pterygoid.’’ The latter sentence exemplifies the problem. If a ‘‘vaulted’’ palate is characterized only by a downturning of the quadrate rami, Rileymillerus has a vaulted palate. This type of vaulted palate will likely be found in any temnospondyl in which the quadrate is ventral to the exoccipitals. Alternatively, a ‘‘vaulted’’ palate could (and we believe should) be considered as one in which the lateral free margin of the palatal ramus of the pterygoid, i.e., the portion adjacent to the subtemporal fossa, is downturned. This type of vaulted palate is clearly absent in Rileymillerus. We suggest that ambiguity in this character can be avoided by breaking it into two characters as in Table 1, namely exoccipital condyles dorsal to quadrate condyles, and lateral free margin of palatal ramus of pterygoid downturned. Although we cannot prove that the vertebrae described above are diplospondylous or that they pertain to Rileymillerus, the vertebrae do at least suggest the possibility of a close relationship between Rileymillerus and tupilakosaurids. The case for such a relationship does not appear strong. The pterygoid incisure, a tupilakosaurid autapomorphy (Warren, 1999) is absent in Rileymillerus. The large vomerine fangs of Rileymillerus may or

680

JOURNAL OF PALEONTOLOGY, V. 74, NO. 4, 2000

may not conform to Warren’s character ‘‘tusks of the palate hypertrophied’’ (p. 147), but the palatine fangs seem unlikely to do so. None of the other diagnostic tupilakosaurid characters listed by Warren provides strong support for a relationship with Rileymillerus, and some (‘‘dorsal column of the ascending ramus of the pterygoid,’’ p. 147) contradict it. The absence of lateral line sulci is unexpected in a brachyopoid. From the literature, sulci appear to be present in all brachyopoids where the skull surface is well preserved. We do not consider this a bar to relationship, because it might reflect nothing more than terrestriality. Two characters appear to strongly contradict a proposed relationship between Rileymillerus and brachyopoids and the families most closely related to them: 1) Squamosal-quadratojugal trough. Damiani and Warren (1996) found this presumed derived character in brachyopids and a number of families related to them, but it is unequivocally absent in Rileymillerus. 2) Elongation of postglenoid area of lower jaw. Damiani and Warren considered this a synapomorphy of Brachyopidea. Although ‘‘elongation’’ is a subjective character, we believe that Rileymillerus likely lacks it under any reasonable definition. However, since the postglenoid area of the Rileymillerus mandible is obviously incomplete, the state of this character cannot be definitely determined. We conclude that evidence for a relationship between Rileymillerus and brachyopoids, although suggestive, is not compelling. We therefore assign Rileymillerus to ‘‘Temnospondyli incertae sedis,’’ a surprising result in view of the known diversity of Triassic temnospondyls and the large number of both specimens and studies that are now available. Regardless of what the relationships of Rileymillerus may prove to be, there is only one way to rigorously test any hypothesis of relationships and to develop strongly supported alternative hypotheses. That is, to develop and analyze a data matrix with adequately formulated and scored characters for a reasonably large number of other temnospondyls. As discussed above, such a data matrix does not yet exist; it is well beyond the scope of this paper to comb the literature for the required data and probably next to impossible anyway, given the inadequacy of many specimens and descriptions. As a contribution to what is necessarily a joint effort on the part of many colleagues to develop such a database, we have translated much of our descriptive matter above into character statements and presented them in tabular form in Table 1 of the Appendix. Character data matrix for Rileymillerus.—The character data matrix uses the three-element format developed for the PRESERVE web site and recently presented in Lombard and Bolt (1999). Under this system, each character is stated in terms of an anatomical part, a feature of that part, and a set of possible states for the part plus feature combination. The three elements are distinguished by typographical convention: PART feature state, e.g., PREFRONTAL bone present. Parts are individual skeletal elements or groups of such elements (SQUAMOSAL; MANDIBLE). A single feature, applying to only one attribute of a part, is used in each character. Features and states are so constructed as to be taxon neutral, without implied hypotheses of process or function, and unambiguous, so that their meaning is as clear as possible to other researchers. When the state set comprises absent/present, ‘‘absent’’ is always represented by 0 (zero), ‘‘present’’ by 1. These numeric scores do not equate to ‘‘primitive’’ or ‘‘derived.’’ Features are assigned to an individual part when that is sufficient to describe the feature. Thus the maxilla facial process is assigned to MAXILLA because it is confined to that bone. If a given feature involves more than two

parts, a collective regional part name such as SKULL or MANDIBLE is used, e.g., SKULL lateral line canals absent. Column 1 of Table 1 organizes the characters by anatomical regions. In Column 2, each character is stated as a part, a feature, and the state scored for Rileymillerus. Column 3 is the numerical score, usually a 1 or 0. Column 4 presents the state set for each character and the numeric value assigned to each state, usually ‘‘absent (0), present (1).’’ Each character associated with a given part is hierarchically ranked by indentation, according to the inclusiveness of the character. Thus the character ‘‘PARASPHENOID bone present’’ is more inclusive than the character ‘‘PARASPHENOID cultriform process present,’’ which conveys additional information about the parasphenoid bone. Ontogenetic stage and environment.—Although small, Rileymillerus is undoubtedly postmetamorphic. This is indicated by its overall high level of ossification, including the exoccipital and quadrate (cf. Boy, 1974). The development of labyrinthine folding in the teeth is also consistent with this. Unusually for a Late Triassic temnospondyl, Rileymillerus is apparently terrestrial. Evidence for this comes from the terrestrial nature of the associated fauna, the absence of lateral line sulci, and the high degree of ossification despite its small size. ACKNOWLEDGMENTS

We are grateful to R. C. Miller for allowing the junior author to collect at the Post Quarry and B. Creisler for suggesting the generic name. We thank M. Donnelly for illustrations, J. Weinstein for photography, and W. Simpson and S. McCarroll for specimen preparation. The Texas Memorial Museum loaned us the holotype of Latiscopus. The senior author received support from the Maurice Richardson Fund and the Carl Kropf Fund at The Field Museum; the junior author is grateful for support from National Geographic Society grant number 2620–83 and Texas Tech University. We thank R. Hook and A. Warren for their thoughtful and helpful reviews. REFERENCES

BOLT, J. R. 1974. Evolution and functional interpretation of some suture patterns in Paleozoic labyrinthodont amphibians and other lower tetrapods. Journal of Paleontology, 48:434–458. BOY, J. A. 1974. Die Larven der rhachitomen Amphibien (Amphibia: Temnospondyli; Karbon-Trias). Pala¨ ontologische Zeitschrift, 48:236– 268. BYSTROW, A. P. 1935. Morphologische Untersuchungen der Deckknochen des Scha¨ dels der Wirbeltiere. I. Mitteilung. Scha¨ del der Stegocephalen. Acta Zoologica, 16:65–141. , AND I. A. EFREMOV. 1940. Benthosuchus sushkini Efr.—a labyrinthodont from the Eotriassic of Sharzhenga River. Travaux de l’Institut Pale´ ontologique de l’Academie de Sciences de l’URSS, 10(1):1–152. CHATTERJEE, S. 1983. An ictidosaur fossil from North America. Science, 220:1151–1153. . 1985. Postosuchus, a new thecodontian reptile from the Triassic of Texas and the origin of tyrannosaurs. Philosophical Transactions of the Royal Society of London B 309:395–460. .1986a. The Late Triassic Dockum vertebrates: their stratigraphic and paleobiogeographic significance, p. 139–150. In K. Padian (ed.), The Beginning of the Age of Dinosaurs. Cambridge University Press, . 1986b. Malerisaurus langstoni, a new diapsid reptile from the Triassic of Texas. Journal of Vertebrate Paleontology, 6:297–312. . 1991. Cranial anatomy and relationships of a new Triassic bird from Texas. Philosophical Transactions of the Royal Society of London Ser. B, 332:277–346. . 1997. The Rise of Birds: 225 Million Years of Evolution. Johns Hopkins University Press, Baltimore. . 1998. The avian status of Protoavis. Archaeopteryx, 16:99–122. DAMIANI, R. J., AND A. A. WARREN. 1996. A new look at members of the Superfamily Brachyopoidea (Amphibia, Temnospondyli) from the

BOLT AND CHATTERJEE—NEW TRIASSIC TEMNOSPONDYL FROM TEXAS Early Triassic of Queensland and a preliminary analysis of brachyopoid relationships. Alcheringa, 20:277–300. DUTUIT, J.-M. 1972. Un nouveau genre de Ste´ gocephale du Trias Supe´ rieur Marocain: Almasaurus habbazi. Bulletin du Muse´ um National d’Histoire Naturelle, Science de la Terre, 3e se´ rie 11(72):73–77. . 1976. Introduction a` l’e´ tude pale´ ontologique du Trias continental Marocain. Description des premiers ste´ gocephales recuellis dans le couloir d’Argana (Atlas Occidental). Me´ moires du Muse´ um National d’Histoire Naturelle 36:1–253. GREGORY, J. T. 1945. Osteology and relationships of Trilophosaurus. University of Texas Publication No. 4401:273–359. HUNT, A. P. 1993. Revision of the Metoposauridae (Amphibia: Temnospondyli) and description of a new genus from western North America, p. 67–97. In Morales, M. (ed.), Aspects of Mesozoic Geology and Paleontology of the Colorado Plateau. Museum of Northern Arizona Bulletin 59. JUPP, R., AND A. A. WARREN. 1986. The mandibles of the Triassic temnospondyl amphibians. Alcheringa, 10:99–124. LOMBARD, R. E., AND J. R. BOLT 1999. A microsaur from the Mississippian of Illinois and a standard format for morphological characters. Journal of Paleontology, 73:908–923. MURRY, P. A. 1986. Vertebrate paleontology of the Dockum Group, western Texas and eastern New Mexico, p. 109–137. In K. Padian (ed.), The Beginning of the Age of Dinosaurs. Cambridge University Press. SA¨ VE-SO¨ DERBERGH, G. 1935. On the dermal bones of the head in labyrinthodont stegocephalians and primitive Reptilia with special reference to Eotriassic stegocephalians from East Greenland. Meddelelser om Grønland, 98:1–211. SHISHKIN, M. A. 1973. The morphology of the early Amphibia and

681

some problems of lower tetrapod evolution. Trudy Paleontologicheskogo Instituta, Akademiya Nauk SSSR, 137:1–260. , B. S. RUBIDGE, AND J. W. KITCHING. 1996. A new lydekkerinid (Amphibia, Temnospondyli) from the Lower Triassic of South Africa: implications for evolution of the early capitosaurid cranial pattern. Philosophical Transactions of the Royal Society of London (B), 351: 1635–1659. WARREN, A. A. 1981. A horned member of the labyrinthodont superfamily Brachyopoidea from the Early Triassic of Queensland. Alcheringa, 5:273–288. . 1999. Karoo tupilakosaurid: a relict from Gondwana. Transactions of the Royal Society of Edinburgh: Earth Sciences, 89:145–160. , AND T. BLACK. 1985. A new rhytidosteid (Amphibia, Labyrinthodontia) from the Early Triassic Arcadia Formation of Queensland, Australia, and the relationships of Triassic temnospondyls. Journal of Vertebrate Paleontology, 5:303–327. , AND M. N. HUTCHINSON. 1988. A new capitosaurid amphibian from the early Triassic of Queensland and the ontogeny of the capitosaur skull. Palaeontology, 31:857–876. WATSON, D. M. S. 1956. The brachyopid labyrinthodonts. Bulletin of the British Museum (Natural History), 2:315–392. . 1962. The evolution of the labyrinthodonts. Philosophical Transactions of the Royal Society of London (B), 245:219–265. WELLES, S. P., and R. ESTES. 1969. Hadrokkosaurus bradyi from the Upper Moenkopi Formation of Arizona. University of California Publications in Geological Sciences, 84:1–56. WILSON, J. A. 1948. A small amphibian from the Triassic of Howard County, Texas. Journal of Paleontology, 22:359–361. ACCEPTED 22 FEBRUARY 2000

APPENDIX I Characters observed in Rileymillerus cosgriffi, TTU P 9168. See Discussion for explanation. 1. Regions

2. Characters (part—feature—state)

3.

4. Possible states

MANDIBLE

1. MANDIBLE bones present 2. MANDIBLE lateral line indication absent 3. MANDIBLE suture complexity in dermal sheathing bones straight to sinuous

1 0 0

4. MANDIBLE level of glenoid fossa relative to dentary tooth row approximately same level

0

5. MANDIBLE recess in posterior tip, surrounded by dermal bones of mandible present 6. ANGULAR bone present 7. ANGULAR foramen in joint with prearticular (苷chorda tympani foramen) present 8. ARTICULAR bone present 9. ARTICULAR para-articular foramen (⫽presumed chorda tympani foramen) present 10. ARTICULAR para-articular foramen (⫽presumed chorda tympani foramen) location entirely in articular 11. CORONOID bones present 12. CORONOID teeth present 13. DENTARY bones present 14. DENTARY teeth present 15. DENTARY tooth bases cross-sectional shape circular

1

absent (0), present (1) absent (0), present (1) straight to sinuous (0), interdigitated (1) approximately same level (0), elevated above tooth row (1), depressed below tooth row (2) absent (0), present (1)

1 1

absent (0), present (1) absent (0), present (1)

1 1

absent (0), present (1) absent (0), present (1)

1

in suture with prearticular (0), entirely in articular (1) absent (0), present (1) absent (0), present (1) absent (0), present (1) absent (0), present (1) circular (0), labio-lingually compressed (1), proximo-distally compressed (2) ⬍30 (0), 30–40 (1), 41 or more (2) straight (0), recurved (1) absent (0), present (1) absent (0), present (1) absent (0), present (1)

16. 17. 18. 19. 20.

DENTARY DENTARY DENTARY DENTARY DENTARY

number of marginal-tooth spaces 30–40 teeth profile in labiolingual view recurved teeth labyrinthine folding present teeth anterior keels absent teeth posterior keels absent

1 1 1 1 0 1 1 1 0 0

682

JOURNAL OF PALEONTOLOGY, V. 74, NO. 4, 2000 APPENDIX I Continued.

1. Regions SKULL

2. Characters (part—feature—state)

3.

4. Possible states

21. SKULL bones present 22. SKULL lateral line indication absent 23. SKULL squamosal embayment absent 24. SKULL cross-sectional shape of transition from skull roof to cheek, posterior to orbits continuous curve 25. SKULL ornament present 26. SKULL ornament zones of intensive growth absent 27. SKULL ornament type pit and ridge ‘temnospondyl-type’

1 0 0 0

absent (0), present (1) absent (0), present (1) absent (0), present (1) continuous curve (0), top and side meet at an angle (1) absent (0), present (1) absent (0), present (1) ‘anthracosaur-type’ (0), pit and ridge ‘temnospondyl-type’ (1), little to none (2) anterior one-quarter (0), second one-quarter (1), third one-quarter (2), posterior one-quarter (3) straight to sinuous (0), interdigitated (1) absent (0), present (1) absent (0), present (1) absent (0), present (1) absent (0), present (1) ectopterygoid teeth ⬍90% (0); 90% ⱖ ectopterygoid teeth ⱕ 110% (1), ectopterygoid teeth ⬎110% absent (0), present (1)

1 0 1

28. SKULL position of center of orbit relative to overall skull length (quadrate-snout) second one-quarter

1

29. SKULL suture complexity in skull roofing bones straight to sinuous

0

30. SKULL paraquadrate foramen absent 31. SKULL posttemporal fenestrae present 32. ECTOPTERYGOID bone present 33. ECTOPTERYGOID teeth present 34. ECTOPTERYGOID teeth average basal diameter relative to average basal diameter of adjacent marginal teeth 90% ⱖ ectopterygoid teeth ⱕ 110% 35. ECTOPTERYGOID teeth distributed in single row parallel to maxilla present 36. ECTOPTERYGOID tooth attachment type pleurodont lingual to attachment ridge 37. EXOCCIPITAL bone present 38. EXOCCIPITAL condyles double present 39. EXOCCIPITAL condyles stalked present 40. EXOCCIPITAL condyle anteroposterior position relative to quadrate condyle posterior 41. EXOCCIPITAL condyle dorsoventral position relative to quadrate condyle dorsal 42. EXOCCIPITAL basioccipital fenestra between exoccipitals at posterior extremity of braincase present 43. EXOCCIPITAL submedullary processes partly roof basioccipital fenestra present 44. EXOCCIPITAL lamellar process absent 45. EXOCCIPITAL paroccipital process present 46. JUGAL bone present 47. JUGAL border to orbit present 48. JUGAL processus alaris present 49. LACRIMAL bone absent 50. MAXILLA bone present 51. MAXILLA preorbital dorsal process absent 52. MAXILLA inner surface in contact with dorsolateral flange of vomer present 53. MAXILLA border to orbit absent 54. MAXILLA teeth present 55. MAXILLA tooth bases cross-sectional shape circular

0 1 1 1 1

56. 57. 58. 59. 60.

MAXILLA teeth anterior keels absent MAXILLA teeth posterior keels absent MAXILLA teeth labyrinthine folding present MAXILLA teeth profile in labiolingual view recurved MAXILLA fang peaks (‘peak’ ⫽ 1–3 adjacent teeth at least 1.5 the diameter of adjacent teeth) absent 61. NASAL bone present 62. NASAL inner surface in contact with dorsolateral flange of vomer present 63. OPISTHOTIC bone absent 64. PALATINE bone present 65. PALATINE lateral exposure on skull surface (LEP) present 66. PALATINE joint with pterygoid absent 67. PALATINE participation in medial border of internal naris present 68. PALATINE participation in lateral border of internal naris present 69. PALATINE fang tooth (at least 25% greater diameter and/or height than adjacent marginal teeth) present 70. PALATINE fang tooth location immediately posterior to internal naris present 71. PALATINE number of fang tooth attachment loci immediately posterior to internal naris two 72. PALATINE fang tooth labyrinthine folding present 73. PALATINE fang tooth anterior keel absent 74. PALATINE fang tooth posterior keel absent

1 1

1

acrodont (0), pleurodont lingual to attachment ridge (1) absent (0), present (1) absent (0), present (1) absent (0), present (1) anterior (0), posterior (1), approximately same level (2) dorsal (0), ventral (1), approximately same level (2) absent (0), present (1)

1

absent (0), present (1)

0 1 1 1 1 0 1 0 1

absent absent absent absent absent absent absent absent absent

0 1 0 0 0 1 1 0

absent (0), present (1) absent (0), present (1) circular (0), labio-lingually compressed (1), mesio-distally compressed (2) absent (0), present (1) absent (0), present (1) absent (0), present (1) straight (0), recurved (1) absent (0), present (1)

1 1 0 1 1 0 1 1 1

absent absent absent absent absent absent absent absent absent

1

absent (0), present (1)

1

one (0), two (1), ⬎ two (2)

1 0 0

absent (0), present (1) absent (0), present (1) absent (0), present (1)

1 1 1 1 0

(0), (0), (0), (0), (0), (0), (0), (0), (0),

(0), (0), (0), (0), (0), (0), (0), (0), (0),

present present present present present present present present present

present present present present present present present present present

(1) (1) (1) (1) (1) (1) (1) (1) (1)

(1) (1) (1) (1) (1) (1) (1) (1) (1)

BOLT AND CHATTERJEE—NEW TRIASSIC TEMNOSPONDYL FROM TEXAS

683

APPENDIX I Continued. 1. Regions

2. Characters (part—feature—state)

SKULL (cont.)

77.

80.

83. 85. 87. 89.

102. 104. 107. 110. 113.

75. PALATINE single row of approximately marginal-sized teeth along medial border of internal narial opening present 76. PALATINE single row of approximately marginal-sized teeth posterior to internal naris and running parallel to maxilla present PARASPHENOID bone present 78. PARASPHENOID cultriform process shape in cross section convex downward 79. PARASPHENOID basal plate with ‘‘pockets’’ or transverse ridge presumably for muscle insertion absent PARIETAL bone present 81. PARIETAL parietal foramen present 82. PARIETAL parietal foramen greatest diameter relative to overall skull length (quadrate-snout distance measured parallel to the midline) ⱕ2.5% POSTFRONTAL bone present 84. POSTFRONTAL border to orbit present POSTORBITAL bone present 86. POSTORBITAL border to orbit present PREFRONTAL bone present 88. PREFRONTAL border to orbit present PTERYGOID bone present 90. PTERYGOID lamina ascendens present 91. PTERYGOID lamina ascendens contact with occipital flange of squamosal absent 92. PTERYGOID lamina ascendens contact with flange of squamosal anterior to occipital flange present 93. PTERYGOID lamina ascendens origin from dorsal surface of pterygoid body 94. PTERYGOID lamina ascendens orientation in an approximately coronal plane

95. PTERYGOID lamina ascendens dorsal column absent 96. PTERYGOID free portion of palatal ramus inflected anteriorly to project into interpterygoid vacuity present 97. PTERYGOID quadrate ramus with dorsal division present 98. PTERYGOID quadrate ramus with free flange that projects posteriorly beyond the quadrate absent 99. PTERYGOID joint with vomer absent 100. PTERYGOID free portion of palatal ramus ‘‘vaulted,’’ by downturning of its lateral margin adjacent to subtemporal fenestra absent 101. PTERYGOID pterygoid incisure absent QUADRATE bone present 103. QUADRATE hyoid tubercle absent QUADRATOJUGAL bone present 105. QUADRATOJUGAL occipital flange present 106. QUADRATOJUGAL squamosal-quadratojugal trough absent SQUAMOSAL bone present 108. SQUAMOSAL occipital flange present 109. SQUAMOSAL crista falciformis absent TABULAR bone present 111. TABULAR horn absent 112. TABULAR parotic process for contact with paroccipital process present VOMER bone present 114. VOMER participation in anterior border of internal naris present 115. VOMER participation in medial border of internal naris present 116. VOMER anterolateral corner of dorsal surface produced into a dorsolateral flange present 117. VOMER fang tooth (at least 25% diameter and/or height than adjacent marginal teeth) present 118. VOMER fang tooth location immediately anterior to internal naris present 119. VOMER number of fang tooth attachment loci immediately anterior to internal naris two 120. VOMER fang tooth labyrinthine folding present 121. VOMER fang tooth anterior keel present 122. VOMER fang tooth anterior keel serrations absent 123. VOMER fang tooth posterior keel absent 124. VOMER single row of approximately marginal-sized teeth along medial border of internal narial opening present 125. VOMER single row of approximately marginal-sized teeth along intervomerine suture present

3.

4. Possible states

1

absent (0), present (1)

1

absent (0), present (1)

1 2

absent (0), present (1) flat (0), keeled (1), convex downward (2), convex upward (3) absent (0), present (1)

0 1 1 0

absent (0), present (1) absent (0), present (1) ⱕ2.5% (0), ⬎2.5% ⱕ 5% (1), ⬎5% (2)

1 1 1 1 1 1 1 1 0

absent absent absent absent absent absent absent absent absent

1

absent (0), present (1)

1

0 1

from quadrate ramus (0); from dorsal surface of pterygoid body (1) in an approximately coronal plane (0); makes transition from parasagittal posteriorly, to approximately coronal anteriorly (1); mostly parasagittal (2) absent (0), present (1) absent (0), present (1)

1 0

absent (0), present (1) absent (0), present (1)

0 0

absent (0), present (1) absent (0), present (1)

0 1 0 1 1 0 1 1 0 1 0 1 1 1 1 1

absent absent absent absent absent absent absent absent absent absent absent absent absent absent absent absent

1

absent (0), present (1)

1 1

absent (0), present (1) one (0), two (1), ⬎ two (2)

1 1 0 0 1

absent absent absent absent absent

1

absent (0), present (1)

0

(0), (0), (0), (0), (0), (0), (0), (0), (0),

(0), (0), (0), (0), (0), (0), (0), (0), (0), (0), (0), (0), (0), (0), (0), (0),

(0), (0), (0), (0), (0),

present present present present present present present present present

present present present present present present present present present present present present present present present present

present present present present present

(1) (1) (1) (1) (1) (1) (1) (1) (1)

(1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1)

(1) (1) (1) (1) (1)