Hooker, J. J. 2007. Bipedal browsing adaptations of the unusual Late Eocene–earliest Oligocene tylopod Anoplotherium (Artiodactyla, Mammalia). Zoological Journal of the Linnean Society 151 (3): 609-659.
Well, "camel" isn't technically correct. The animal in question, Anoplotherium, is a member of the Tylopoda, a clade of hoofed mammals whose only surviving members are the Camelidae (camels and llamas), but which was once more diverse.
The title of this paper certainly caught my eye when table of contents for the new issue of Zoological Journal of the Linnean Society arrived in my e-mail just now, and the article did not disappoint. Anoplotherium, which I was previously only vaguely aware of as a name to occassionally appear on lists of extinct artiodactyls, turns out to have been a rather interesting creature indeed. For a start, it was apparently one of the earliest fossil mammals to be described, losing out only to Palaeotherium (a distant relative of horses), Mammut (the mastodon), Megalonyx and Megatherium (both giant sloths). Since Cuvier named it in 1804, however, it garnered relatively little attention, largely because of the poor preservation and/or fragmentary nature of the material available. To add insult to injury, the most complete remains were embedded in a gypsum matrix that added to the difficulty of working with them.
The current paper was inspired by the discovery of a far more complete fossil skeleton in the Hamstead Member of the Bouldnor Formation on the Isle of Wight in England (though I wouldn't know how far that is from the old stomping grounds of Darren Naish). The story of the skeleton's discovery is worth the read in itself - pieces of the single individual were recovered by successive amateur palaeontologists over a period of 35 years, as the sea-side cliff it was embedded in eroded away and further pieces of it were exposed or, in some cases, fell out of the cliff.
Anoplotherium differs significantly from living artiodactyls (even-toed hoofed mammals) in a number of ways, some of which you can make out in the reconstruction above, taken from the paper. For a start, it has a longer and thicker tail than any living artiodactyl. Cuvier (mistakenly, as it turned out) compared the tail of Anoplotherium to an otter, leading him to later suggest that Anoplotherium might have been semi-aquatic. This idea was later expanded by Gervais, who suggested that it may have had webbed feet! Later authors, however, agreed that this reconstruction was functionally unlikely. The true significance of the tail, however, will be explained in due course.
The feet are particularly unusual - despite having the typical artiodactyl foot arrangement (with the midline of the foot running between the third and fourth digits, rather than down the third digit as in most mammals including humans), Anoplotherium had only three toes (digit V having been lost). Digit II is the smaller toe you can see above, and stuck out somewhat from the other toes. I should point out here that the new specimen belongs to the species Anoplotherium latipes - another species, A. commune, differed in having a much smaller digit II. However, this appears to be the only significant difference between the two species, and Hooker makes the suggestion that the two 'species' may be different sexes of a single species (the correct name for which would then be A. commune). He suggests that the form with the larger digit II might have been the male, and that the larger digit might have been used in inter-male conflict, allowing the front legs of opposing males to hook together when grappling, much like the tines of deer antlers. Unfortunately, Hooker isn't able to conclusively prove this interesting idea, and I think that a fair degree of skepticism is required for now.
And if all this wasn't enough, there's the matter of the spine and hips. Among other features, Anoplotherium has very widely flared ilia (the plate-like side parts of the pelvis) and a vertebral column seemingly adapted for bearing weight towards the posterior end. It appears that Anoplotherium was quite able to carry itself as a biped! Most likely, it would have only maintained a bipedal stance (as shown in the reconstruction to the left, again from the paper) when feeding or, possibly, fighting. Normal movement would have still been conducted quadrupedally. The unique tail would have provided balancing support, giving Anoplotherium a tripod stance. Admittedly, Anoplotherium differs in this respect from other bipedal browsers, many of which, such as gorillas and chalicotheres, have either a very small tail or none at all. As Hooker explains, this probably represents phylogenetic constraints - most of these other browsers derived from clades in which the tail had already been reduced to the point of insignificance, requiring the development of alternative methods of balance. Other bipedal browsers also developed grasping forelimbs that may have allowed for an extra degree of support - an option that would not really have been available to the hoofed ancestors of Anoplotherium.
If the idea of a bipedal artiodactyl seems unbelievable to you, Anoplotherium is not the only one. The living gerenuk (Litocranius walleri) stands bipedally when feeding (as shown in the photo above from Images of Anthropology), and is even able to take a few tottering steps in this position. The gerenuk has even less features showing its bipedality than Anoplotherium does - Hooker's comment is that "It remains difficult to understand... how Litocranius can function as a bipedal browser with such minimal adaptations... relatively small body size may be a facilitating factor."