Telmatherium, Marsh 1872

Mader, Bryn J., 2008, A species level revision of Bridgerian and Uintan brontotheres (Mammalia, Perissodactyla) exclusive of Palaeosyops, Zootaxa 1837 (1), pp. 1-85 : 52-62

Genus TELMATHERIUM Marsh 1872

= Leurocephalus Osborn et al. 1878

= " Telmatotherium " Marsh 1880

= Manteoceras Hatcher 1895

Age. Bridgerian.

Subage. Twinbuttean.

Type species. T. validus Marsh 1872 .

Included species. Genus is monospecific.

Diagnosis. Medium to moderately large-sized (length P2 to M3 approximately 160–190 mm) brontotheriine brontothere with a small but distinctive horn-like prominence developed on the frontonasal boundary over facial concavity and a deep pit-like fossa in the middle of the sagittal crest near the back of the skull.

Discussion. In 1872 O.C. Marsh named a new brontothere taxon, Telmatherium validus , based upon "the greater portion of a skull with teeth, and portions of several other skeletons." Marsh did not identify these materials by catalog number but Osborn (1929, p. 160) identified the skull as YPM 11120. Osborn (1929) cited this skull as the holotype of T. validus as did Mader (1989), who believed the type cited by Osborn to be accurate. Marsh, however, had based the species upon a type series, even though he treated the skull as his principal specimen and gave locality information for it alone. Thus, YPM 11120 would more properly be regarded as a lectotype for the species, although neither Osborn nor Mader (1989) formally designated it as such. In 1998, however, Mader, cognizant that a type series existed, listed YPM 11120 as the “ lectotype ” of T. validus , effectively establishing it as such, although he did not explicitly state that it was his intention to fix the type. If subsequent researchers should conclude that the type of T. validus was not properly fixed by Mader (1998), then, as the first reviser to directly address this problem, I now formally select YPM 11120 as the lectotype for the species T. validus Marsh. The taxonomic purpose for this designation (as per the requirements of the International Code of Zoological Nomenclature, Article 74.7.3, Ride et al. 1999) is to avoid confusion and nomenclatural instability should it be determined that the specimens of the original type series belong to different genera. This is a distinct possibility given that three contemporaneous brontothere genera existed at the time (Twinbuttean) and that early authors, such as Marsh, often had an imprecise conception of the taxa they described.

Five years after Marsh's description of Telmatherium validus Henry Fairfield Osborn and William Berryman Scott made an amateur collecting trip to the Bridger (Green River) Basin to celebrate their graduation from Princeton ( Howard 1975, pp. 258–259). This first "Princeton Expedition" ( Osborn 1929; p. 338) resulted in the collection of a second specimen of Telmatherium (YPM-PU 10027, the right side of a skull and lower jaw). Osborn et al. (1878) identified this specimen as a new genus, Leurocephalus (type species L. cultridens ), but subsequent authors (e.g., Earle 1892; Osborn 1929; Mader 1989) have recognized that it is a specimen of Telmatherium .

The first complete skulls of Telmatherium were discovered in 1893 by an American Museum expedition to the Washakie Basin led by J. L. Wortman. In a letter from the field dated June 1, 1893, Wortman wrote Osborn that he believed that the skulls represented a new genus for which he suggested three possible names Manteoceras, Manteotherium , or Manteocephalus, (= "prophecy" + "horn", "animal", or "head").

Osborn (1895) disagreed with Wortman's assessment of these skulls, however, and instead identified them as specimens of Palaeosyops vallidens Cope (a taxon based on a lower jaw, AMNH 5098), which Osborn referred to the genus " Telmatotherium ". Osborn's interpretation of " Telmatotherium " was very broad, and he included under this generic name Telmatherium and all of the valid dolichorhinine genera. Hatcher (1895) recognized that " Telmatotherium " as defined by Osborn consisted of several distinct taxa and proposed that the type jaw of Cope's Palaeosyops vallidens be given the new generic name Manteoceras ("prophet horn"), following the original suggestion of Wortman.

In a surprising turn of events Osborn (1929) concluded that the lectotype lower jaw of Palaeosyops vallidens Cope did not represent the same taxon as Wortman's skulls after all but rather represented the same genus as the holotype skull of " Telmatotherium " cornutum , the type species of Dolichorhinus . Both Manteoceras and Dolichorhinus were named in the same paper ( Hatcher 1895) and, if Osborn was correct, Dolichorhinus could be recognized as a junior synonym of Manteoceras .

Osborn (1929) attempted to avoid this situation by contending that Hatcher had " Telmatotherium " vallidens, sensu Osborn (1895) , in mind when he named Manteoceras and not Palaeosyops vallidens Cope. Osborn supported this argument by pointing out that Hatcher's figures of " Telmatotherium " vallidens were based on Osborn's figures of Wortman's skulls and that Hatcher's diagnosis of Manteoceras (which was based entirely upon cranial characters) was inapplicable to the lectotype lower jaw of Palaeosyops vallidens .

Osborn accepted Manteoceras as the valid generic name for Wortman's skulls (because this is what Hatcher had intended), but could not recognize Manteoceras vallidens as the type species. Osborn was forced, therefore, to look in the subsequent literature for an appropriate type species. In 1899 Matthew had published a faunal list in which the name Palaeosyops manteoceras appeared. This was a new trivial name meant to apply to Wortman's skulls, which Matthew took from an unpublished Osborn manuscript. Matthew, however, did not designate a type or provide a diagnosis. As Osborn (1929) pointed out, Palaeosyops manteoceras Matthew was, therefore, a nomen nudum.

In his Bibliography and Catalogue of the Fossil Vertebrata of North America, Hay (1902) listed as the type of the genus Manteoceras Hatcher , Palaeosyops manteoceras Osborn and gave as figure references Osborn's and Hatcher's illustrations of Wortman's skulls. Hay did not provide a diagnosis for the species. According to Osborn (1929) the giving of the figure references was essentially the same as designating a type because they specifically identified the specimens to which the name was to be applied. The figures in question illustrated two skulls, AMNH 1569 and AMNH 1570, of which Osborn (1929) selected AMNH 1569 as a lectotype.

Thus, by highly circuitous reasoning, Osborn (1929) concluded that the type species of Manteoceras was M. manteoceras Hay , because M. manteoceras was a synonym of " Telmatotherium " vallidens, sensu Osborn. As I have previously pointed out ( Mader 1989), if Osborn was correct in establishing that the lectotype jaw of Palaeosyops vallidens represents the same genus as Hatcher's Dolichorhinus , then the conclusion that Manteoceras and Dolichorhinus are synonyms is inescapable. Hatcher clearly stated that the type of Palaeosyops vallidens was to be given the new generic name Manteoceras and, regardless of what Hatcher's concept of that genus might have been, the name stays with the type originally specified. The type species of Manteoceras , therefore, is Palaeosyops vallidens Cope not " Telmatotherium " vallidens, sensu Osborn.

Because the recognition of Manteoceras and Dolichorhinus as synonyms could create confusion, Mader (1989) recommended that any author attempting to synonymize the genera formally choose Dolichorhinus as the name having priority (see recommendation 24A in the International Code of Zoological Nomenclature, Ride et al. 1999). Manteoceras would thus be a junior synonym of Dolichorhinus rather than a junior synonym of Telmatherium (as it is treated here). I continue to recommend this assignment of nomenclatural priority, but note that it is now less critical because, even if Manteoceras and Dolichorhinus are synonymous, both terms would be junior synonyms of Sphenocoelus (see Discussion section for the genus Sphenocoelus ).

Osborn (1929) recognized both Telmatherium and Manteoceras as valid genera. In practice he referred fragmentary specimens of Telmatherium to the genus Telmatherium and more complete specimens to the genus Manteoceras . Although Osborn cited a number of differences between Telmatherium and Manteoceras , these differences are either trivial or based upon comparisons with the unrelated brontothere genus Metatelmatherium (see below).

Telmatherium disappeared from the fossil record at the end of the Bridgerian, but Osborn (1908; 1929) regarded Uintan aged specimens now referred to the genus Metatelmatherium as a later form of Telmatherium . Interestingly, for both of the genera that Osborn split Telmatherium into ( Telmatherium and Manteoceras ), he recognized Uintan descendents that are all specimens of Metatelmatherium ultimum . Thus Osborn recognized Telmatherium ultimum (now Metatelmatherium ultimum ) as the supposed Uintan successor of T. validus and T. cultridens , and Manteoceras uintensis (= Metatelmatherium ultimum ) as the presumed successor of M. manteoceras and M. washakiensis (see section on Metatelmatherium ).

In 1908, Osborn described a new species of Manteoceras from the highest level of Washakie A, which he named M. washakiensis (type AMNH 13165, a skull). Osborn (1908; 1929) distinguished M. washakiensis from M. manteoceras by its larger skull size, larger cheek teeth, more complete cingula on the cheek teeth, morphology of the canine (obtuse, recurved, and with a heavy posterior cingulum), better developed "deuterocone" (= protocone) on P2 ( Osborn 1929, p. 371, stated that the internal lobes of P2 and P3 are broadening, with a shelf for the development of the deuterocone), shorter face, moderate zygomata, and relatively inconspicuous horns. Under the heading "Progressive Characters" Osborn (1929, p. 371) also stated that the molar series of M. washakiensis is relatively longer than that of M. manteoceras , in which the premolar length is 76% that of the molar length (averaged for six individuals), while in M. washakiensis the premolars are only 71% of the molar length. Osborn regarded this as a progressive character because there is an increase in the relative length of the molar series over the course of brontothere evolution.

In my opinion, all of the characters (both cranial and dental) used by Osborn to distinguish Manteoceras washakiensis from M. manteoceras are quite minor and are insufficient to establish a new taxon. I also find no basis for Osborn's statement that the type skull of M. washakiensis is larger than skulls that he referred to M. manteoceras . The type of M. washakiensis lacks the premaxillary region and thus Osborn (1929) was only able to provide an estimated skull length (= 490 mm) for it. The amount of bone missing from the type skull and the degree to which it is crushed make this a very rough estimate, and as a result, any taxonomic conclusions based upon skull length are necessarily suspect. In addition, in a table (1929, p. 364) Osborn listed skull lengths and estimates of skull lengths for specimens that he referred to M. manteoceras that equal or exceed Osborn's estimated value for the skull length in the type of M. washakiensis . Thus AMNH 12204, identified as a female M. manteoceras by Osborn, has an estimated skull length of 490 mm (equal to that of M. washakiensis ) and AMNH 1545, identified as a male M. manteoceras , has a skull length of 523 mm (which exceeds that of M. washakiensis ).

I also find no evidence that the molar series is longer in the type of Manteoceras washakiensis than in skulls that Osborn referred to M. manteoceras . As previously noted, Osborn stated that the length of the premolar series in the type skull of M. washakiensis is 71% that of the molar length. This is very similar to the value obtained from my own data which results in the premolar length being 70% of the molar length. Osborn's statement that in M. manteoceras the premolar length was 76% that of the molar length, however, was based on an average value for six individuals. If these six individuals are among the specimens that Osborn listed in the table noted above (1929, p. 364), then these specimens would be AMNH 12683, AMNH 1511, AMNH 2353, AMNH 1569, AMNH 1570, and AMNH 1545; the only specimens for which the length of both the premolar series and molar series are reported. Although Osborn was correct in reporting that the average value for premolar length relative to molar length is 76% for these specimens, in one specimen (AMNH 1545) the premolar length is 70% of the molar length, which is virtually the same as in the type of M. washakiensis .

a Based, whenever possible, on an average of left and right measurements.

b Insufficient data.

c Excluding Diastema Length and Basilar Length of Skull.

a Based, whenever possible, on an average of left and right measurements.

b There is no variance in one of the groups being compared.

a Based, whenever possible, on an average of left and right measurements.

b Insufficient data for t -test.

c Separate t -test.

a Based, whenever possible, on an average of left and right measurements.

b Insufficient data.

c Excluding Diastema Length.

a Based, whenever possible, on an average of left and right measurements.

b Insufficient data.

c Excluding Diastema Length and Basilar Length Skull.

The main reason for Osborn's recognition of Manteoceras washakiensis as a species distinct from M. manteoceras appears to be the former's higher stratigraphic occurrence. Osborn stated that the "decidedly progressive characters (of M. washakiensis ) beyond those of M. manteoceras ... perfectly accord with its somewhat higher geologic level." In my opinion, the "decidedly progressive characters" noted by Osborn are illusory and the supposed higher stratigraphic level is not by itself, sufficient grounds for the recognition of a taxon.

The genus Telmatherium ( Fig. 19 ) is known from contemporaneous deposits in the Green River and Washakie Basins. Most specimens are from the Green River Basin, but since the sample size from that location is relatively low (no more than thirteen individuals for any given variable) the sample was supplemented with five individuals from the Washakie Basin for the purpose of statistical analysis.

The coefficient of variation ( Table 12) for most variables in the combined Green River and Washakie sample of Telmatherium lie within the range of 4 to 10, and the average value for all variables (excluding diastema and skull length) is 6.5. Diastema length and skull length are eliminated from consideration because of the high variability in diastema length in Pre-Chadronian brontotheres in general (see Summary Statistics tables in the present paper and in Mader 1991) and because the sample for skull length consisted of only a single individual. Thus, the variation present in all of the specimens of Telmatherium from the Green River and Washakie Basins is no more than is ordinarily encountered in a single extant mammalian species.

Cluster analysis of all variables, however, shows remarkably clear delineation into two size groups ( Fig. 20 ). T -tests ( Tables 13 and 14) confirm that, for most variables, the means of these size groups are significantly different. Approximately three quarters of the variables for which a value of T could be calculated had a probability of less than.05 and over one half had a probability of less than.01. Although the t -test for the length of the right third upper premolar fails to show a significant difference, the t -test for the length of the left P3 and t -tests for the widths of both P3's do show a significant difference. Thus, the t -test result for the length of the right P3 is probably a chance occurrence. The only variables that truly appear to have no significant difference between the two size groups are the length of the diastema and the lengths and widths of both second upper premolars. It should be noted that there is a 77% chance that at least one significant result in this study is due to error (see Methods). Given the great quantity of significant results, however, it seems that the statistical reality of the two size groups is firmly established.

The summary statistics for each of the two size groups of Telmatherium are presented in Tables 15 and 16. Almost one-third of the variables in both size groups (excluding the basilar length of the skull in the large-size group) have values of V that are below 4 (values rounded to the nearest whole number). In addition, although the average value of V for the smaller size group is within the expected range for a single species, the average value for the larger size group is rather low (4.1). These results suggest that in both size groups there is less variation than is typically encountered in extant mammalian species (although most of the samples are probably large enough to show most of the variation present).

Thus, although two size groups are clearly present, the variation within each is probably too low for them to be regarded as separate species. Despite the fact that cluster analysis strongly differentiates between these size groups, an inspection of the size ranges reported in Tables 15 and 16 will show that in absolute terms, the size difference between the groups is quite minimal. I conclude, therefore, that the two clusters in the dendrogram in Figure 20 probably represent two size groups within a single species (perhaps males and females).

It would have been desirable to analyze canine size and horn size in Telmatherium to determine whether these variables were bimodal within each of the size groups (as might be expected if they were two different species). Unfortunately, the canines were very poorly preserved in most of the specimens of Telmatherium measured and horn size is difficult to measure precisely because there is no definite horn base (my impression, however, is that horn size does not vary appreciably in any specimens of Telmatherium ).

If bimodality could have been established for either canine or horn size I would be willing to regard the two size groups as distinct species because there would then be evidence that males and females were present in each size group. If the two size groups of Telmatherium were to be interpreted as being two species, then the type species ( T. validus ) would be represented by the larger size group while the name T. cultridens would be applied to the smaller-size group. Manteoceras washakiensis would remain a junior synonym of T. validus but M. manteoceras would become a junior synonym of T. cultridens .

Figures

FIGURE 19. Skull of Telmatherium validus in A, anterior; B, dorsal; C, ventral; and D, lateral views. After Osborn, 1929.

FIGURE 20. Cluster dendrogram for specimens of Telmatherium from the Green River and Washakie Basins resulting from a cluster analysis of all variables listed in Table 12. a, type of Telmatherium washakiensis; b, type of T. validus; c, type of T. cultridens; d, Osborn’s lectotype of T. manteoceras.