Duchesneodus uintensis, (Peterson, 1931)

Mihlbachler, Matthew C., 2008, Species Taxonomy, Phylogeny, and Biogeography of the Brontotheriidae (Mammalia: Perissodactyla), Bulletin of the American Museum of Natural History 311 (1), pp. 1-475 : 414-419

D. uintensis ( Peterson, 1931) Megacerops Leidy, 1870b

M. coloradensis Leidy, 1870b (sensuMihlbachleretal.,2004b)

M. kuwagatarhinus Mader and Alexander, 1995

BRONTOTHERE ORIGINS AND CLADISTIC BIOGEOGRAPHY

The fossil mammal records of North America, Europe, and Asia reveal an intricate history of faunal interchanges, resulting in a complex web of phylogenetic relationships among the faunas of these continents. Biogeographic hypotheses are implicit in any phylogeny; however, most prior attempts to reconstruct patterns of Cenozoic mammalian intercontinental dispersals, attempts to relate them to tectonic and global climate trends, and to explain such dispersal patterns using biogeographic theories, such as MacArthur-Wilson island biogeography ( MacArthur and Wilson, 1967), are not tied to explicit hypotheses of phylogeny. Rather, immigrants, and thus dispersal events, are usually inferred on a local basis by origination events,i.e., first appearances ( Stucky, 1990, 1992; Webb, 1969, 1977, 1989; Webb and Opdyke, 1995; Woodburne and Swisher, 1995). However, at least two processes result in localized originations of taxa. These include immigration as well as endemic speciation. Differentiating immigrants from those taxa arising from endemic speciation is not possible without considering phylogeny. Few have bothered to explain how immigrants were differentiated from localized speciation, and those who have are not explicit. For instance, Webb and Opdyke (1995) explained how they identified North American immigrant genera; ‘‘Immigrants (first appearances) were recognized on the basis of two criteria: (1) the absence of that genus or closely related genera in prior North American records and (2) the presence of related forms (sister group) earlier in another continental fauna.’’ ( Webb and Opdyke, 1995: 188). Plainly, their criteria for identifying immigrants are based on inferences of phylogeny, but without the rigor of explicit reference to a cladogram. Stucky (1992) has recognized that results of such studies need to be reevaluated in a cladistic context.

Optimizing geographic location (e.g., continent) a posteriori on a cladogram identifies the minimum number of dispersal events implicit in any phylogenetic hypothesis. Beard (1998) examined the geographic origins of major mammalian higher taxa (e.g., orders) by mapping geographic location (continent) on the cladograms of numerous placental mammal clades. Beard (1998) proposed the ‘‘East of Eden’’ model that views ‘‘North America as a biogeographical cul-de-sac that received repeated innoculations [sic] of more…taxa from Asia.’’ ( Beard, 1998: 25). Thus, Asia is postulated as the primary source for North American mammalian diversity. However, Beard’s focus on the geographic origins of higher taxa is not entirely satisfactory as an analysis of North American–Asia dispersal trends, because the emphasis on nodes representing the origins of higher taxa, as opposed to all nodes on a cladogram, results in arbitrary emphasis on certain nodes, while ignoring other nodes where immigration events might also be indicated. Ultimately, all of the nodes on a cladogram represent speciation events that, as historical events, are not fundamentally different from other speciation events regardless of whether they are arbitrarily associated with higher taxonomic names.

A nonarbitrary approach for examining the frequency, timing, and polarity of intercontinental dispersal events would be to reconstruct geographic histories on speciescomprehensive cladograms in which all known members (or as many as can possibly be sampled) of a clade are included. This approach utilizes all of the available evidence and is the only way to guarantee the discovery of all the dispersal events implied by the known members of a clade. Here, I compare the history of dispersals implied by brontothere phylogeny to previous ideas about brontothere dispersal, and to more generalized ideas about the history of North America–Asia biotic interchanges.

The trees resulting from both analyses of ordered and unordered data imply numerous intercontinental dispersals. The 3967 trees produced by the ordered character data imply 12 dispersals within the Brontotheriidae . An example of one of these trees is seen in fig. 201, with continental location optimized onto it as a character, showing 12 dispersals within the Brontotheriidae . (An additional dispersal of uncertain polarity also occurs at the base of the tree relating to Lambdotherium and Danjiangia .) The 603,705 MPTs resulting from the analysis of unordered data imply 9–12 dispersal events within the Brontotheriidae . Figure 202 is one tree from the unordered character search showing the minimum number of nine dispersals allowed by that analysis. To be clear, the cladograms are generated purely from morphological characters with the geographic character mapped on a posteriori. Brachydiastematherium transylvanicum , the sole brontothere taxon from eastern Europe included in the analysis, is coded as Asian because the locality from which it comes was biogeographically part of Asia and not Europe.

BRONTOTHERE ORIGINS

Geographic location is optimized on all the cladograms in such a way that indicates a North American origin for Brontotheriidae . The primitive North American brontothere genera, Eotitanops and Palaeosyops , have previously been interpreted as immigrants from Asia ( Stucky, 1992; Woodburne and Swisher, 1995; Beard, 1998); however, there is little evidence that these early brontotheres or their ancestors resided in Asia. A few fossils from Asia have been tentatively identified as primitive brontotheres similar to Eotitanops and Palaeosyops . Cladistically, Eotitanops is the most primitive brontothere. Presently there are no indisputable Eotitanops fossils known from Asia. ‘‘ Eotitanops ?’’ dayi Dehm and Oettingen-Spielberg (1958) is not clearly a member of Eotitanops ; the known specimens are far too incomplete to determine their taxonomic identity or evaluate their phylogenetic position. A small Eotitanops -like brontothere from Japan has been preliminarily reported by Miyata and Tomida (2003); this material may ultimately have a strong bearing on brontothere origins. Palaeosyops -like fossils have been reported from Asia, but they are based on very fragmentary specimens ( Gabounia, 1977; Xu et al., 1979). An upper premolar (P4) identified as Palaeosyops sp. by Xu et al. (1979) is an indeterminate specimen that could belong to one of several relatively small brontotheres; however, the thin enamel of that specimen suggests something more derived than Palaeosyops . On the other hand, the upper molar specimen identified as Palaeosyops sp. by Gabounia (1977) from the Zaysan Basin of Kazakstan most certainly indicates a Palaeosyops -like brontothere in Asia. Based on its wide ectoloph but otherwise plesiomorphic condition, it would fit onto the cladograms in the same position as Palaeosyops . Based on this specimen, if Palaeosyops were recoded as ‘‘polymorphic’’ for geographic location (i.e., having occurred on both continents), the geographic character would be optimized the same way on the cladograms, still suggesting a North American origin for Brontotheriidae , and subsequent dispersal of Palaeosyops into Asia.

To determine the origin of the Brontotheriidae , it is necessary to look outside the family to its sister taxa. If the North American genus, Lambdotherium is the sister taxon of the Brontotheriidae , as it is historically presumed to be, then it is likely that the Brontotheriidae originated in North America. In contrast, Beard (1998) suggested that ‘‘the ancestry of the Bridgerian brontotheres [ Eotitanops and Palaeosyops ] must be sought outside of North America, where only the autapomorphous Lambdotherium is known from earlier rock’’ ( Beard, 1998: 27). Regarding the autapomorphous condition of Lambdotherium , autapomorphies are phylogenetically uninformative and, thus, cannot influence cladistically derived hypotheses of biogeography. Currently the best hypothesis is a North American origin for brontotheres.

Recent discoveries of Lambdotherium -like mammals in Asia demand reconsideration of a possible Asian origin for brontotheres. Danjiangia pingi Wang (1995) , from late early Eocene Yuhuangding Formation of Liguanqiao Basin of China was considered the primitive sister of Lambdotherium by Beard (1998) and a more plausible stem taxon for the Brontotheriidae (this is in contrast to the original assessment of Wang [1995], who suggested that it is an early Chalicothere). However, these statements by Beard (1998) are contradictory. If Danjiangia is the closest relative of brontotheres, and Lambdotherium and Danjiangia are sisters, as Beard suggests, then Lambdotherium must also be the sister of brontotheres. Meng et al. (1998) described dental fragments of an unnamed Lambdotherium -like mammal from the late Paleocene Bayan Ulan fauna of China that add additional uncertainly to the geographic origin of the Brontotheriidae . The present cladistic analysis places Lambdotherium as the sister of the Brontotheriidae , with Danjiangia one step removed. A more comprehensive analysis, utilizing more outgroup taxa is needed to determine the phylogenetic position of Brontotheriidae within Perissodactyla and, simultaneously, to develop the best hypothesis for the geography of its origin.

CLADISTICALLY IMPLIED INTERCONTINENTAL DISPERSALS

The recovered phylogenies imply numerous intercontinental dispersal events. Moreover, the frequency of implied dispersals is greater than had been previously surmised for this family. All trees resulting from the analysis of ordered data imply 10 dispersals occurring on internal branches within the Brontotheriidae , while all trees resulting from the analysis of unordered data imply between 7–10 dispersals occurring on internal branches. Two additional ‘‘autapomorphic’’ dispersals occur on distal branches; Metatelmatherium ultimum and Eubrontotherium clarnoensis are found in both North America and Asia. Therefore, the total number of dispersals implicit in all of the cladistic results ranges from 9–12. Problematic taxa not included in the phylogenetic analysis, particularly cf. Palaeosyops ( Palaeosyops sp. of Gabounia, 1977), Eotitanops? dayi , Pakotitanops latidentatus , and Mulkrajanops moghliensis , may indicate additional dispersal events.

In the example tree in fig. 201, nine of these events imply dispersal from North America to Asia, while only three imply dispersal in the opposite direction. Other MPTs resulting from the same analysis show conflicting dispersal patterns within the Brontotheriita . These discrepancies relate to the wildcard behavior of Notiotitanops . Ultimately, the polarities of any of the three possible dispersal events occurring within the Brontotheriita are uncertain. Despite these few discrepancies, the analysis of ordered data shows a dispersal pattern that indicates predominantly dispersal to Asia from North America. In contrast to the East of Eden model of Beard (1998), no tree shows a predominance of dispersal from Asia to North America.

The minimum age of a dispersal event is constrained by the age of the taxon (either a terminal taxon or a clade) that results from the dispersal event. In other words, the dispersal could have occurred any time before the appearance of the immigrant and its descendants but not after. Most brontothere dispersal events occurred no later than the middle Eocene. The following interpretation is based on the tree in fig. 201. The appearance of Bunobrontops in Asia results from a dispersal from North America by the middle Eocene. Another dispersal from North American into Asia occurring by the Irdinmanhan land mammal age involves the appearance of Desmatotitan and Acrotitan in Asia. The dispersal events involving the appearances of Microtitan , Qufutitan , Metatelmatherium , and Epimanteoceras also occurred before or during the Irdinmanhan land mammal age. The existence of Metatelmatherium ultimum in the Irdinmanhan of Asia and in the Late Uintan of North America strongly pinpoints the Irdinmanhan and late Uintan as the time of dispersal.

Horned brontotheres ( Brontotheriina ) appear to have originated in Asia, with Protitan as the most basal member. Subsequent dispersal of members of the Brontotheriina back into North America involved Uintan taxa Protitanotherium emarginatum , Pollyosbornia altidens , and Diplacodon elatus . Fig. 201 depicts three subsequent back-dispersals to Asia involving members of the Brontotheriita , Parabrontops , Dianotitan , and Eubrontotherium . The timing of the dispersal of Eubrontotherium clarnoensis , either from Asia to North America or vice versa is difficult to evaluate. In Asia it is known from late Eocene Ulangochuian deposits, but its age in North America is controversial, with opinions ranging from Uintan to Chadronian. Likewise, the age of Dianotitan is poorly constrained, but possibly suggests a middle Eocene dispersal before or during the Irdinmanhan or Sharamurunian land mammal ages. Finally, the late Eocene appearance of Parabrontops in Asia indicates dispersal before or during the Ulangochuian; this is the latest possible dispersal event implied by this tree. As noted above, the dispersal pattern and polarity within the Brontotheriita varies in other trees depending on the position of the wildcard, Notiotitanops , and the timing of these events are poorly constrained in comparison to those dispersal events that lie in more resolved parts of the consensus trees resulting from the analysis of ordered character data.

The number, polarity, and nodal positions of the dispersal events implied by the analysis of unordered data are more difficult to evaluate due to the excessive number of MPTs and large number of erratic wildcard taxa. The example shown in fig. 202 implies five dispersal events from North America to Asia, with an additional four occurring in the opposite direction. Other trees resulting from this analysis show differing dispersal patterns and the predominant direction of dispersal is indeterminate from the analysis of unordered character data.

CONCLUSION

Three general conclusions regarding brontothere biogeography are apparent in the phylogenetic results. (1) Contrary to the earlier statements of an Asian origin, a North American origin seems likely, but this is contingent upon Lambdotherium being the sister-taxon of the Brontotheriidae . A more comprehensive analysis with a more extensive outgroup needs to be performed. Brontotheres have yet to be sufficiently incorporated into analyses of perissodactyl phylogeny.

(2) Several successive intercontinental brontothere dispersals occurred in the middle Eocene, roughly corresponding to the Uintan and Irdinmanhan land mammal ages. The dispersal patterns implied by the phylogenetic results differ markedly from previous ideas about brontothere dispersal. Granger and Gregory (1943) viewed the Asian brontotheres largely as a separate radiation from North American horned brontotheres, both of them having arisen from a North American Telmatherium ancestry. In general, paleontologists have oversimplified brontothere paleobiogeography. For instance, Stucky (1992), Webb and Opdyke (1995), and Woodburne and Swisher (1995) identified the arrival of Duchesneodus in North America as the sole immigration event for brontotheres. The number of brontothere dispersals is greater than previously thought; however, the timing of the majority of brontothere-dispersal events, in the Irdinmanhan and late Uintan land mammal ages (late middle Eocene), is consistent with Woodburne’s and Swisher’s (1995) finding that this time period, with the Wasatchian (early Eocene), represent the two most significant periods of Asian–North American faunal interchanges of the Tertiary. of dispersal events are thought to have gone from Asia into North America. The East of Eden model follows classic island biogeography theory whereby immigrant taxa are more likely to hail from a larger land mass (Asia) and disperse to the smaller land mass (North America) ( Beard, 1998). Dispersal in the opposite direction, from North America into Asia, is more strongly supported by the analysis of ordered character data, thus calling into question the validity of island biogeography theory in explaining continent-scale diversity dynamics. The predominant dispersal polarity is inconclusive according to the analysis of unordered data; nevertheless, the results offer no positive support for the East of Eden model. It is possible that brontotheres went the opposite direction with respect to the dispersal-polarity tendencies of other clades, or that the middle and late Eocene interchanges were predominantly reversed in comparison to other times such as the early Eocene, when dispersal trends appear to have been predominantly consistent with the East of Eden model ( Beard, 1998). Presently there are no clear answers for the inconsistency of brontothere dispersal trends with the East of Eden model. Dispersal trends in other mammalian clades need to be rigorously examined in a cladistic context before intercontinental overland dispersal patterns and their effect on biodiversity can be precisely understood.

ACKNOWLEDGEMENTS

I remain in awe of how the former curators and staff of the American Museum and other natural history institutions built such wonderful and scientifically important collections out of nothing more than scraps of old bones that mostly lay scattered across the unpopulated badlands of the world. For the past several years I have had the pleasure of discovering many wonders in these collections. I wish to thank Malcolm McKenna, Jin Meng, and Mark Norell for welcoming me into the venerable Columbia/American Museum vertebrate paleontology program to study brontotheres, my favorite extinct creatures since the third grade. The staff, students, and volunteers at the Division of Paleontology at the American Museum were all enormously helpful during my years there: (in no particular order) Jeanne Kelly, Robert Evander, Denny Dively, Chris Norris, Chris Collins, Carl Mehling, Jim Clausen, Michael Ellison, Rick Edwards, Tom Rothwell, Robert Asher, Julia Clarke, Sunny Hwang, Diego Pol, Jonathan Geisler, Yaoming Hu, Xu Xing, Benjamin Burger, Bolortsetseg Minjin, Susan Bell, Ivy Rutzky, Alejandra Lora, Lorraine Meeker. I also wish to thank Robert Emry, Robert Purdy (NMNH), Lyndon Murray, Daniel Brinkman, Marilyn Fox (YPM), Chris Beard, Alan Tabrum, Mary Dawson (CMNH), John Flynn, William Simpson, William Turnbull (FMNH), Jaelyn Eberle (UCM), Logan Ivy, Richard Stucky (DMNH), Patricia Holroyd, Robert Feranec (UCMP), Lyndon Murray, Dennis Ruez (TMM), Xiaoming Wang, Sam Macleod (LACM), Alexandre Agadjanian, (PIN), Li Qian, Wang Yuanqing, Deng Tao (IVPP), Dou Wenxiu (VM), and Takenori Sasaki (UMUT) for access and help with specimens in their respective institutions and/or hospitality. Hiromichi Hirano (Waseda University) helped me track down Protitanotherium koreanicum specimens. Collaborative research with Nikos Solounias, Spencer Lucas, and Robert Emry was of tremendous value in the development and undertaking of this project. Others who provided valuable information and advice are Paul Olsen, Florent Rivals, Bryn Mader, Luke Holbrook, Mark Siddall, Don Prothero, Jeremy Hooker, Demberelyin Dashzeveg, and Joel Cracraft. Mariko Mihlbachler aided in the production of the figures. Spencer Lucas provided the photo of Epimanteoceras praecursor . Finally, this work could not have been completed without funding from Columbia University, the American Museum of Natural History, the Evolving Earth Foundation, the Geological Society of America, the Paleontological Society, and the New York College of Osteopathic Medicine.

Figures

Fig. 202. One of 603,705 trees produced from the analysis of unordered character data, with a geographic character mapped onto it a posteriori, showing nine intercontinental dispersal events within the Brontotheriidae, the minimum allowed by this analysis.

Fig. 201. One of 3967 trees produced from the analysis of ordered character data with a geographic character mapped onto it a posteriori showing 12 possible dispersal events within the Brontotheriidae occurring between North American and Asia.