The quaternary paleontology and paleoecology of crystal ball cave, millard county, utah: with emphasis on the mammals and the description of a new species of fossil skunk

Discussion--Several species of large horses have been recognized from the late Pleistocene of western North America. The Rancho La Brea asphalt deposits have yielded a single species of large horse (Savage 1951) usually referred to E. occidentalis (Merriam 1913, Stock 1963, Willoughby 1974), although the validity of this name has been questioned (Miller 1971). Based on comparative material and measurements made by Willoughby (1974), the large Crystal Ball Cave horse is distinct from the Rancho La Brea horse in having more transversely broad phalanges (see figures 6, 7, and 8) and carpals with relatively larger articulation surfaces. The Crystal Ball Cave specimens are distinctly larger than E. niobrarensis based on measurements given me by A. H. Harris (1983 personal communication) and in Harris and Porter (1980). A. H. Harris (1983 personal communication) also provided me with measurements of E. pacificus (although the validity of this species has been questioned by Savage 1951) from Fossil Lake, Oregon, and phalanges of this species match well in size with the large Crystal Ball Cave horse but are not as transversely broad.

Gazin (1936) listed measurements of the type specimen of E. scotti, and of all specimens and data seen, only it has phalanges that are as transversely broad as the Crystal Ball Cave specimens. The second phalanx (LACM 123683) is slightly larger than the E. scotti type but has identical proportions (see figure 7), and the third phalanx (BYUVP 7595), although smaller because it is of a subadult, has the same proportions as the anterior third phalanges of the E. scotti type (see figure 8). Dalquest (1964) stated that E. scotti was very heavily built, and this would suggest that the foot and toe bones are broad compared with other species of Equus. The large carpals from Crystal Ball Cave mentioned above, especially the cuneiform and magnum, are broad and have much larger articulation surfaces than the Rancho La Brea horse. Based on this limited information in the literature, the largest carpals listed above compare most favorably with E. scotti also.

E. scotti was originally named and described from Texas by Gidley (1900), and most specimens have been found in that state (Dalquest 1964, Gidley 1903, Johnston 1937). Hopkins et al. (1969) recovered a left metatarsal from the Late Pleistocene American Falls Lake Beds of southeastern Idaho that they referred to E. scotti. It is therefore not unlikely that E. scotti lived in Utah. A large horse was represented at Smith Creek Cave by a single vestigial metapodial (Miller 1979), but no attempt was made to identify it to species.

BYUVP 7588 is not as laterally broad as LACM 123683 but is too large to belong with the smaller species. The epiphysis is not fully fused, showing that it represents a subadult. It is the only bone from Crystal Ball Cave that matches well with the Rancho La Brea horse, although it is slightly smaller. But since it may differ by only individual, foot, or age variation from the better represented E. cf. scotti, it is tentatively referred to that species.

Equus ? conversidens

Material--Right M3/ (LACM 123677), thoracic vertebra (BYUVP 7687), 3 right pisiforms (BYUVP 7536-7538), left pisiform (BYUVP 7539), 2 right cuneiforms (BYUVP 7540, 7541), 4 right lunars (BYUVP 7543, 7545-7547), partial left lunar (BYUVP 7548), 4 right scaphoids (BYUVP 7551-7558), 4 partial left scaphoids (BYUVP 7555-7558), 2 right trapezium-trapezoids (BYUVP 7559, 7560), 2 right magnums (BYUVP 7562, 7563), partial right magnum (BYUVP 7564), left magnum (LACM 123678), 2 partial right unciforms (BYUVP 7565, 7566), proximal tibia epiphysis (BYUVP 7570), distal epiphysis of right tibia (BYUVP 7571), partial distal epiphysis of left tibia (BYUVP 7572), right calcaneum (LACM 123679), left calcaneum (BYUVP 7573), right astragalus (BYUVP 7575), right juvenile astragalus (BYUVP 7574), left astragalus (LACM 123680), right navicular (BYUVP 7576), left navicular (BYUVP 7577), left cuboid (BYUVP 7579), right meso-ento (BYUVP 7578), proximal portion of left metatarsal (BYUVP 7567), 2 distal metapodial epiphyses (BYUVP 7568, 7569), 6 first phalanges (BYUVP 7580, 7581, 7583, LACM 123684, 123685), 3 partial first phalanges (BYUVP 7582, 7584, 7585), 5 second phalanges (BYUVP 7589, 7593, 7594, LACM 123684, 123685), 4 partial second phalanges (BYUVP 7587, 7590-7592), 5 third phalanges (BYUVP 7597, 7600, 7601, 7605, 7606), 2 partial third phalanges (BYUVP 7602, 7603), juvenile third phalanx (BYUVP 7610), 11 proximal sesamoids (BYUVP 7611-7621). Phalanx measurements are listed in tables 4, 5, and 6.

Discussion--In addition to the fossils of large horses from Crystal Ball Cave (referred to E. cf. scotti) are numerous bones of small horses. Some of these compare well with E. conversidens, the species to which most small Pleistocene North American horse fossils have been assigned, while others do not. Considerable time has been spent evaluating the size and morphologic variation among these bones and comparing the results with descriptions and measurements in the literature. But both complexities within this collection and disagreements regarding valid species in the literature have prevented positive species identification of these small horse bones.

E. conversidens (Owen 1869) has been considered by some to be the only species of small Pleistocene horse in North America (Harris and Porter 1980, Miller 1971), and most other named species of small Pleistocene horses have at some time been synonymized with this species (Dalquest and Hughes 1965, Hibbard 1955, Hibbard and Taylor 1960). However most workers presently recognize at least two species. Owen (1869) named E. tau at the same time he named E. conversidens. Poor illustrations of the type specimens have caused some workers to consider E. conversidens and E. tau synonymous (Hibbard 1955). But Dalquest (1979) and Mooser and Dalquest (1975), after researching the early descriptions (the type specimen of E. tau is lost), considered these two species distinct. The teeth that Mooser and Dalquest (1975) assigned to E. tau are smaller than those of E. conversidens, and the metapodials are longer and more slender. Skinner (1942) assigned a first phalanx from Papago Springs Cave, Arizona to E. tau because it was much narrower than those of E. conversidens from the same assemblage. But based on his measurements this phalanx is narrower transversely than anteroposteriorly, making it doubtful of being horse at all.

Hay (1915) named E. francisci, which was synonymized with E. conversidens by Hibbard and Taylor (1960). But Lundelius and Stevens (1970) reprepared the metatarsal of the type specimen and found it to be distinctly longer and narrower than that of E. conversidens. Lundelius and Stevens (1970) therefore considered E. francisci a valid species, and they synonymized E. quinni (based on the similar long metatarsal) and Onager zoyatalis (based on dental similarities) to it. Dalquest (1979) considered E. francisci, as well as E. littoralis, E. achates, and E. quinni, to be synonymous with E. tau, and he considered E. conversidens and E. tau the only two valid species of small Pleistocene North American horses.

Based on an illustration in Lundelius and Stephens (1970), the M3/ of E. francisci is distinctly wider transversely than that of E. conversidens although they are of similar anteroposterior length. LACM 123677, although quite worn, has the same width and length as the E. francisci type and has an enamel pattern most similar to it also. Dalquest (1979) synonymized E. francisci with E. tau, but the M3/ of the lectotype of E. tau illustrated by Mooser and Dalquest (1975) is not transversely broad like the E. francisci type and Crystal Ball Cave M3/. Unfortunately the only phalanx measurements given in the literature are for E. conversidens, except the questionable first phalanx assigned to E. tau by Skinner (1942).

The only phalanges from Crystal Ball Cave that compare well with measurements of E. conversidens phalanges in the literature are three of the five second phalanges (see figure 7). The other two second phalanges (BYUVP 7593, 7594) are distinctly smaller than any assigned to E. conversidens yet have complete epiphyseal fusion. All nine first phalanges are from individuals intermediate in size between those represented by the two sets of second phalanges, and all are small compared with the first phalanges assigned to E. conversidens in the literature (see figure 6). Six of the seven third phalanges articulate well with the three larger second phalanges yet are smaller than the third phalanges assigned to E. conversidens in the literature (see figure 8). The other third phalanx (BYUVP 7600) is larger than any assigned to E. conversidens and too large to articulate with any of the second phalanges under discussion.

It is important to consider sexual dimorphism, individual variation, and variation among different feet of the same individual to see how much variation is expected within a species. Willoughby (1974), in a table of bone measurements from 25 species and races of Equus, listed mean dimensions for both sexes with respect to two characters: metacarpal mid-width and metacarpal mid-width divided by length. Metacarpals of males had a mid-width of 3.1% to 7.3% greater than females and a mid-width divided by length of 2.3% to 6.9% greater than females. Species with more sexual dimorphism in metacarpal width tended to also have more dimorphism in width relative to length, so male metacarpals tend to be more robust and just slightly longer than female metacarpals. These measurements show that sexual dimorphism is not great in Equus and certainly not sufficient to have caused the variability seen among the small Crystal Ball Cave equids.

Howe (1970), in a study of Equus (Plesippus) simplicidens, showed that individual variation in bone size can be greater than previously thought. Because the large number of specimens at Nebraska's Broadwater Quarry fell into a single size curve with no gaps, he concluded that they all represent a single species, and he synonymized a number of species which had previously been named based on limited material at other sites. Table 5 of Howe (1970) shows that the largest metacarpal and metatarsal lengths and widths average 32% larger than the smallest corresponding measurements, and none are more than 36% larger. Even with a sample size of 97 to 190, the metapodials measured by Howe (1970) show less variation than do the few second and third phalanges from Crystal Ball Cave.

Isolated front and rear phalanges are usually indistinguishable and therefore have an additional degree of variation. Front and rear phalanx measurements were taken from recent E. caballus and E. burchelli specimens, and the larger measurements for each species averaged 4.2% larger than the smallest corresponding measurements with a maximum of 9.4% larger. But even this much variability, in addition to sexual and individual variation, does not adequately account for the great size range among the small Crystal Ball Cave equids.

Six measurements of the 5 second phalanges from Crystal Ball Cave (excluding those referred to E. scotti) show that the largest measurements are 24% to 43% larger than the smallest corresponding measurements with an average of 31.5% larger. Eleven measurements of the 9 third phalanges from Crystal Ball Cave show that the largest measurements are 7% to 122% larger than the smallest corresponding measurements with an average of 50.7% larger. Considering the second and third phalanges separately, each have enough variation to make it marginal whether they could all be assigned to the same species considering sexual, individual, and foot variation. The variation seems even more extreme when one considers that the smallest second phalanges (BYUVP 7593, 7594) are from much smaller individuals than the smallest third phalanx, and the largest third phalanx (BYUVP 7600) is from a larger individual than the largest second phalanx. This is far more variation than can be accounted for by the sexual, individual, and foot variation for a single species as discussed above, and it suggests that multiple species of horse smaller than E. cf. scotti are represented at Crystal Ball Cave.

Finding a dividing line between two species in this material is nearly impossible, however. Most of the material could be assigned to a species of horse 15% smaller than E. conversidens, but the two smallest second phalanges (BYUVP 7593, 7594) and the largest third phalanx (BYUVP 7600) seem too far from the mean to belong to this supposed species. Until more phalanx measurements are available for small Pleistocene horses other than E. conversidens, it is difficult to determine how many species are represented by the smaller Equus fossils from Crystal Ball Cave and whether most of the material represents an unusually small variety of E. conversidens, a species distinct from E. conversidens such as E. tau and/or E. francisci, or both.

Order Artiodactyla
Family Camelidae
Camelops cf. hesternus

Material--Right scaphoid (LACM 123686), left scaphoid (LACM 123687), left lunar (BYUVP 7624), left magnum (BYUVP 7625), right unciform (BYUVP 7626), distal fragment of metapodial (BYUVP 7629), 2 first phalanges (BYUVP 7627, LACM 123689), proximal portion of first phalanx without epiphysis (LACM 123691), partial proximal epiphysis of first phalanx (BYUVP 7638), 3 second phalanges (LACM 123692, BYUVP 7630, 7632), 3 proximal portions of second phalanges (BYUVP 7633, 7634, 7637), 3 partial proximal portions of second phalanges (BYUVP 7628, 7635, 7636), 3 third phalanges (BYUVP 7639, 7641, 7642). Six sesamoids (BYUVP 7644-7649) are probably of Camelops but may represent Bison. Phalanx measurements are listed in tables 7, 8, and 9.

Discussion--Webb (1965, 1974) recognized only four valid genera of late Pleistocene North American camels: Titanotylopus, Camelops, Hemiauchenia (=Tanupolama), and Paleolama (in order of decreasing size). Titanotylopus is somewhat common and Camelops is very common in late Pleistocene assemblages of western North America, but neither has been found in the east (Webb 1974). Hemiauchenia is found in late Pleistocene deposits throughout the Americas (Webb 1974) and is commonly associated with Camelops (Miller 1979). Paleolama has only been found in Florida, Texas, and southern California in Pleistocene deposits of North America (Miller 1976). Miller (1982) identified Camelops and Hemiauchenia from Crystal Ball Cave.

The specimens listed above fall within the range of variation of Camelops hesternus measurements from Rancho La Brea, southern California (Webb 1965) and Selby and Dutton, eastern Colorado (Graham 1981). T. E. Downs (1984 personal communication) provided me with 8 first phalanx measurements of Titanolopus sp., 21 of Camelops hesternus, and 21 of Hemiauchenia sp. from southern California deposits. Those of Titanolopus range from 105 to 138 mm in length with an average of 121 mm, those of Camelops hesternus range from 105 to 125 mm in length with an average of 116 mm, and those of Hemiauchenia range from 91 to 110 mm in length with an average of 94 mm. The two complete first phalanges from Crystal Ball Cave, both of which are of adults based on epiphyseal fusion and bone density, measure 114 and 117 mm in length (see table 7). Although there is some overlap in first phalanx length between these genera, the Crystal Ball Cave specimens clearly match best with Camelops.

Savage (1951) recognized four valid species of Camelops: C. hesternus and C. huerfanensis, which are larger, and C. sulcatus and C. minidokae, which are smaller; and Webb (1965), in his detailed description of Camelops, supported this system. Based on limb bone measurements given by Savage (1951), C. minidokae was about 14% smaller than C. hesternus. C. huerfanensis can only be distinguished from C. hesternus and C. sulcatus can only be distinguished from C. minidokae based on dental characters (Graham 1981, Savage 1951). Both C. minidokae and C. sulcatus are too small to match the Crystal Ball Cave specimens, and both are known only from pre-Wisconsinan deposits (Kurten and Anderson 1980).

C. hesternus and C. huerfanensis are very similar and may be conspecific (Hopkins 1955, Savage 1951). Both are known from the late Pleistocene, and both are known from Idaho (Gazin 1935, Hopkins 1955, Hopkins et al. 1969) and Colorado (Cragin 1892, Graham 1981). C. hesternus is the only species of Camelops reported from Utah. A Camelops hesternus skull was recovered from a lava tube 87 miles (140 km) east-southeast of Crystal Ball Cave (Romer 1928, 1929) and dated at 11,075 =225 Y.B.P. (Nelson and Madsen 1979). Camelops cf. hesternus was reported from the Silver Creek fauna in north-central Utah (Miller 1976). Camelops sp. was reported from Smith Creek Cave (Harrington 1934, Stock 1936, Miller 1979), but the only material mentioned is a right navicular (Miller 1979), and no attempt was made to identify it to species.

Since the Crystal Ball Cave specimens match measurements of C. hesternus by T. E. Downs (1984 personal communication), Graham (1981), and Webb (1965), and since C. hesternus is the only species reported from the state of Utah, the Crystal Ball Cave specimens are referred to this species. But since the only diagnostic character to distinguish C. hesternus from C. huerfanensis is a dental feature not applicable to the Crystal Ball Cave specimens (Hopkins 1955, Savage 1951), C. huerfanensis cannot be positively eliminated on the basis of these foot elements.

Hemiauchenia cf. macrocephala

Material--Distal right portion of metapodial (LACM 123688), first phalanx (LACM 123690), partial proximal portion of first phalanx (BYUVP 7640), second phalanx (BYUVP 7631). Phalanx measurements are listed in tables 7, 8, and 9.

Discussion--Two genera of small camels are recognized from the Pleistocene of North America: Hemiauchenia and Paleolama (Webb 1974). Based on illustrations of Hemiauchenia (=Tanupolama) macrocephala (=stevensi) by Stock (1928) and Paleolama mirifica by Webb (1974), the metapodials of H. macrocephala are 63% longer but 3% transversely narrower at the distal end than those of P. mirifica. The Crystal Ball Cave metapodial fragment is 12% transversely narrower than the H. macrocephala specimens illustrated by Stock (1928) and measurements from the Vallecito Creek and Ringold sites of southern California provided by T. E. Downs (1984 personal communication). The first phalanges from Crystal Ball Cave fall well within the range of Hemiauchenia specimens reported by T. E. Downs (1984 personal communication), McGuire (1980), and Schultz (1937). Nothing was available to compare the second phalanx with, but it is from the same size of camel as the other elements. The Crystal Ball Cave specimens clearly match the more narrow-legged Hemiauchenia rather than the more broad-legged Paleolama.

Webb (1974) synonymized the North American genus Tanupolama with the South American genus Hemiauchenia and recognized six valid species. Of these, only H. macrocephala is found in the late Pleistocene of North America. H. macrocephala represents the synonymy of a number of previously named North American species (Webb 1974), and it is the best-known Pleistocene llama (Kurten and Anderson 1980). Since only this species matches the age and locality of the Crystal Ball Cave assemblage, and since the Crystal Ball Cave specimens match specimens from other sites assigned to this species, the four Crystal Ball Cave specimens are referred to H. macrocephala. Characters separating this species from others of Hemiauchenia are almost entirely dental (Webb 1974), however, and are therefore not applicable to the Crystal Ball Cave material.

Miller (1982) reported Hemiauchenia from Crystal Ball Cave based on the same material reported here. Miller (1979) reported ? Hemiauchenia sp. from Smith Creek Cave based on a left cuboid, the proximal portion of a scapula, and a juvenile metapodial. Hemiauchenia is better represented than Camelops at Smith Creek Cave by a ratio of 3 to 1, but Camelops is better represented than Hemiauchenia at Crystal Ball Cave by a ratio of 7 to 1. This difference seems even more dramatic in light of the selection for smaller bones at Crystal Ball Cave but not at Smith Creek Cave. Although this difference could be explained by slight age differences in these faunas, human intervention, or chance preservation, I feel it is more likely due to habitat differences between these two genera of camels.

Kurten and Anderson (1980) stated that ". . . Hemiauchenia had a long stride and was highly cursorial. It was a plains-dweller and probably fed primarily on grass." About Camelops they stated: "Although primarily a grazer, Camelops, with its long neck and legs, was probably an occasional browser." Although these two camels are thought to have been plains-dwelling grazers, it is interesting to speculate about their habitat differences. Webb (1974) presented strong evidence that Hemiauchenia gave rise to the mountainous living South American llamas. Camelops, on the other hand, probably resembled the living dromedary camel (Kurten and Anderson, 1980) which prefers flat plains habitats. The fact that Camelops is by far the better represented camel at Crystal Ball Cave, located in a small outlier surrounded by a flat valley, while Hemiauchenia is better represented at nearby Smith Creak Cave, located in a canyon at the base of a high mountain, suggests that Hemiauchenia preferred higher elevations and/or more rugged terrain than Camelops.

Family Cervidae

cf. Cervus elaphus

Material--First phalanx (BYUVP 7811).

Discussion--Several cervid phalanges from Crystal Ball Cave are intermediate in size between Cervus and Odocoileus. BYUVP 7811 (60.2 mm long) is the largest of these and is much closer in size to Cervus. In comparison with the others it is distinctly larger and more robust, yet high bone porosity suggests that it is of a sub-adult. Navahoceros fricki is another late Pleistocene cervid found as close to Utah as Arizona and Wyoming, and its size is intermediate between Odocoileus and Cervus (Kurten and Anderson 1980). No character has been described to differentiate phalanges of Navahoceros and Cervus, and no comparative material of Navahoceros was available to the author. C. elaphus was recovered from Smith Creek Cave (Miller 1979) and has been reported living in the Snake Range in Recent times (Hall 1981), so the phalanx is referred to this species.

Odocoileus hemionus

Material--Partial right dentary with P/3,/4, M/1 (BYUVP 7651) and anterior left dentary with P/3,/4, M/1,/2 (BYUVP 7650, probably from the same individual), partial right dentary with P/3 (BYUVP 7652), left patella (BYUVP 7934). Of 21 first and 51 second phalanges of non-camelid artiodactyls, most compare best in size and proportions with Odocoileus.

Discussion--BYUVP 7650 and 7651 are of a juvenile and compare best in size and degree of hypsodonty with juvenile individuals of O. hemionus. The P/4's in these dentaries have 3 lobes rather than 2, a condition seen in juveniles of Odocoileus but not Antilocapra. The P/3 of BYUVP 7652 is identical to adult O. hemionus and distinctly larger and less hypsodont than A. americana. The first and second phalanges from Crystal Ball Cave that compare best with Odocoileus have a slightly larger mean size than those of Recent O. hemionus living in Utah. This demonstrates that the Crystal Ball Cave specimens are of O. hemionus rather than the smaller O. virginiana (Hall 1981), and it suggests that deer decreased in size at the end of the Pleistocene much like Ovis canadensis did (Harris and Mundel 1974).

Based on numbers of phalanges, Odocoileus is the best represented artiodactyl in the Crystal Ball Cave assemblage; but Antilocapra americana is now the dominant artiodactyl of the local fauna. Odocoileus sp. was reported at Smith Creek Cave by Goodrich (1965), but no material was found by Miller (1979). Mule deer now live in Smith Creek Canyon (Miller 1979) and sometimes come down to Gandy at night to feed in cultivated fields (J. C. Bates 1984 personal communication). The replacement of Odocoileus by Antilocapra, suggested by comparison of the Crystal Ball Cave assemblage with the living community, shows that plant communities preferred by deer apparently moved upward in altitude from Snake Valley to higher elevations in the Snake Range at the close of the Pleistocene.