The knuckle-walking anteater

Caley Orr (Personal page, Arizona State University) has an advance paper in AJPA examining convergent features in the wrists of knuckle-walking hominoids and the terrestrial giant anteater (Myrmecophaga tridactyla). Did you know that some anteaters were knuckle walkers? I certainly didn't, until I read this!

The background for this paper is the recent finding of certain features in the wrists of early hominids, specifically the distal radius of Australopithecus afarensis and Australopithecus anamensis, that appears similar to those found in chimpanzees and gorillas (Richmond et al. 2001). It has been argued that these features are primitive retentions from a knuckle-walking ancestor, a hypothesis viewed as consistent with the idea that all of the African apes descend from a single agent knuckle-walking species. The current paper gives a good synopsis of this evolutionary problem, including the alternative hypothesis that knuckle-walking evolved in parallel in the chimpanzee and gorilla lineages, so that the hominids did not descend from an age in knuckle-walker.

So why study anteaters? The idea is that different phylogenetics lineage that share a behavior ought to share convergent features to support that behavior. African apes have long fingers for suspensory climbing, which are talked out of the way for quadrupedal movement, hence walking on the knuckles. Giant anteaters have long claws for digging into insect colonies, and these long claws are talked out of the way for a quadrupedal walking. Thus the purpose of this study was:

1. To determine if the locomotion and hand postures of the giant anteater are appropriately analogous to the knuckle-walking of African apes;
2. To identify features of the Myrmecophaga hand and wrist that converge functionally with Pan and Gorilla, and that distinguish these taxa from their non knuckle-walking out groups and terrestrially digitigrade primates; and
3. Through the above analyses, to help determine the traits most likely to be adaptive to knuckle-walking, thereby suggesting which features of the early hominin and modern human wrists might be reliably indicative of a knuckle-walking ancestry (Orr 2005:3).

Did it work? The study did find some features of giant anteaters that made their wrists similar to those of chimpanzees and gorillas. But does this provide support for the idea that early hominids were descended from knuckle-walkers? Orr saves a critical piece of logic for the discussion:

A convergence study can only provide positive or equivocal evidence in testing hypotheses of adaptation for purported knuckle-walking features of the African Hominidae (or any other such study). That is, because all taxonomic groups have unique evolutionary histories, a convergence test cannot truly falsify a hypothesis of an adaptation for a particular lineage. However, convergence study can provide positive support for the hypothesis that structure X is an adaptation to function F, if X distinguishes F- performing taxa from their respective non-F-performing outgroups (Orr 2005:18, emphasis in original).

OK, so we should be wary. Comparing australopithecines to anteaters cannot falsify the hypothesis that hominids had knuckle-walking ancestors. In other words, it is not testing any hypothesis of human origins, although it may provide evidence consistent with one or more of them. Here's the summary of anteater-hominid resemblances:

Morphological features that appear in the hominin lineage shared by Myrmecophaga, Pan, and Gorilla, to the exclusion of their respective outgroups and digitigrade primates, are supported as adaptations to knuckle-walking and provide strong inference of a knuckle-walking last common ancestor (LCA) of Gorilla, Pan, and hominins. Only one such feature (proximal expansion of the non articular surface of the dorsal capitate) appears in the hominin lineage. Human capitates show the African apes state of proximal expansion, and the A. afarensis capitate (AL 333-40) shares the morphology of Gorilla, Pan, and Homo (Orr 2005:19).

Orr discusses the distal radius articular ridge that features in the arguments of Richmond and colleagues (2001), but does not designate this trait as one that necessarily reflects knuckle-walking as opposed to other kinds of vertical hand posture during locomotion, as found in cercopithecid monkeys.

If you're interested in the origins of knuckle-walking, and the question of whether early hominids were knuckle-walkers, this article is for you. As for myself, I think the issue is more likely to be settled with more fossil evidence of Miocene apes and their locomotor styles, rather than the examination of other mammalian lineages. It remains a mystery to me why early hominids should retain features useful only for a knuckle-walking, when their knuckles clearly could not have reached the ground. It seems more likely to me that there is some other function for which these characters might be adaptive related to early hominid locomotion, such as climbing, or other activities. Phylogenetic inertia is never very convincing, especially for early hominids, says they altered almost every other interface between their body and the environment in the pursuit of more perfect bipedal locomotion.


Orr CM. 2005. Knuckle-walking anteater: a convergence test of adaptation for purported knuckle-walking features of African Hominidae. Am J Phys Anthropol (advance before print).

Richmond BG, Begun DR, Strait DS. 2001. Origin of human bipedalism: The knuckle-walking hypothesis revisited. Yrbk Phys Anthropol 44:70-105.