I read that big study of hominin teeth that I mentioned earlier this week and it has turned out to be a let down. My disappointment lies in their systematics.

Martinón-Torres, et al. carried out both a cladistic and phenetic analysis on a lot of homin teeth, which sounds great at first. It sounds like they really wanted to carry out a thorough analysis.

In general, phenetic systematics aims to determine the relationships of organisms by measuring the amount of similarity, factoring both plesiomorphies or ancestral traits and apomorphies or derived traits equally.

Since plesiomorphies sometimes complicate things, cladisitics does away with them and factors only traits that make a clade unique. These identifying characteristics of a clade are called synapomorphies or shared, derived characters.

Thankfully for hominin teeth, there is the Arizona State University Dental Anthropology System (ASUDAS) which is currently the best determinitation of which plesiomorphic traits were present before the last common ancestor of the species group and which synapomorphic traits were present in the last common ancestor by establishing an outgroup or two. In other words, it is the industry standard.

Figure 2 from Martinón-Torres, et al., 2007Figure 2, their cladogram, is the only analysis that shows the European-Asian relationship — their big headliner. So when I flip back to the supplemental materials to see what traits were used in the cladistic analysis and read that they didn’t use ASUDAS defined dental traits, instead they establish their own traits especially for this study by,

“carry[ing] out a rigorous preselection of dental traits searching for those suitable for the cladistic approach and excluding those that are highly homoplasic and variable within a population,”

I begin to wonder what’s going on here?

How can this study be of any effect when they don’t apply to standards others use for their cladogram from homin teeth? What I mean is if one makes up their own system of traits to classify, in this case 9 traits, that haven’t been screened by others before, then getting the unique, interesting results (the whole Eurasian hominin dispersal conclusion) that you were looking for isn’t all too surprising.

Other factors to consider when new unestablished classification systems are used is how does one know that these new traits are or are not genetically correlated? Teeth aren’t the perfect genetic safe-box, as they authors tell us. They are the best phenotype to genetic correlator of any skeletal material that we have to work with as paleontologists. But that doesn’t make them flawless.  Teeth can vary and be modified a lot due to life history like wear.

Razib shares this sentiment,

“I do think it is important to be cautious about translating morphological continuity as clear and present evidence of genetic continuity in any preponderant fashion.”

So I wonder, as W. Henry Gilbert communicated to me,

“What would that say about the [traits] applicability to cladistic analysis?”

John Hawks isn’t completely impressed with this paper either,

“After working through the data supplements for the paper, I think that the analysis is much weaker in statistical power than it could be. In their analysis, they disregard much of the variation within these ancient samples…

The phenogram inexplicably omits Middle and Lower Pleistocene Africans entirely, and considers only australopithecines and habilines as the African sample….

There is no test of gene flow here, just an assertion.”

Could it be possible that Martinón-Torres, et al. setup a unique set of traits to push their Ceprano and Atapuerca hominids closer to the base of all humanity? Their cladogram sure shows it. SH, the nomenclature they used to group Atapuerca, lies awfully closer to H. sapiens in their cladogram that previously accepted. And Manzi, one of the co-authors, has attempted doing just this in the past. I wouldn’t be surprised if that’s the case. So very coy.