Wood and Lieberman (2002) attempt to systematize the variation in fossil A. boisei. Their hypothesis is that some kinds of craniodental variables may be expected to be more variable than others, by virtue of their function. In particular, they consider it likely that variables that reflect high magnitudes of mechanical strain during life will be more variable than those that reflect low magnitudes of strain. The strain connection is based on the notion that phenotypic plasticity may be a significant contributor to the phenotypic variance of traits, and especially for areas of the skull or dentition that are affected by strain during life. This hypothesis is predicated on the notion that the effects of such forces during life are actually variable, rather than constantly presenting the same degree of strain in different individuals. There is no discussion of this issue, because the empirical data ultimately support the connection. The areas of the skull under masticatory strain are more variable than other regions.
It is not clear exactly what accounts for this observation. For example, one hypothesis would be that the masticatory regions show a higher degree of sexual dimorphism than other parts of the skull. Another might be that there is ontogenetic variation between older and younger adults in these characters, and both are represented in the sample. Neither of these hypotheses nor others can be tested with the data presented.
But the causes of the variability are not necessarily relevant to the paper's conclusions about taxonomic problems. Most notably, they conclude that features related to masticatory strain are not well suited to testing taxonomic hypotheses in fossil species. Instead, they recommend basing judgments about species on features that are typically less variable in living analogues.
I think this conclusion is generally right, and it would be interesting to see the consequences of following it through (for example, as applied to earlier hominids). For example, would there be any disagreement about the status of A. africanus if this criterion were followed?
On the other hand, who is to say that a fossil species must resemble the pattern of variability of (essentially) four living species of hominoids? Would we be confortable applying this criterion to Dryopithecus? Proconsul?
The paper appears to present the strain story as a stopgap between splitting and lumping. The fossil instance at stake is the addition of remains from Konso to A. boisei. Wood and Lieberman conclude that the Konso A. boisei specimens do not markedly extend the range of variation in fossil A. boisei. In this, they appear to be picking an argument with Gen Suwa and colleagues (1997), who argued that the Konso specimens were different in many respects from earlier A. boisei. That paper wanted to make the argument that most hominid species may have had as-yet-unrecognized variability; this one wants to argue that most of the as-yet-unrecognized variability probably isn't taxonomically interesting. A minor point, but one that illustrates the predispositions of the authors in each case.
Interestingly, the paper also tests the notion that dental features should be less variable than skeletal features because dental features exhibit higher heritability. They find that although dental measurements do tend to be slightly less variable in the observed data, there is no significance to the relation. In other words, traits that are more heritable are not less variable within species in these data.
Wood BA and Lieberman DE. 2002. Craniodental variation in Paranthropus boisei: a developmental and functional perspective. Am J Phys Anthropol 116:13-25.