Anthropology.net

Earlier today, I shared the news that two of the Leakeys published their Ilert Homo fossil findings in the latest Nature. The paper is out now, under the title, “Implications of new early Homo fossils from Ileret, east of Lake Turkana, Kenya.” Here is the abstract, I’m bolding what I think are the most important parts,

“Sites in eastern Africa have shed light on the emergence and early evolution of the genus Homo. The best known early hominin species, H. habilis and H. erectus, have often been interpreted as time-successive segments of a single anagenetic evolutionary lineage. The case for this was strengthened by the discovery of small early Pleistocene hominin crania from Dmanisi in Georgia that apparently provide evidence of morphological continuity between the two taxa. Here we describe two new cranial fossils from the Koobi Fora Formation, east of Lake Turkana in Kenya, that have bearing on the relationship between species of early Homo. A partial maxilla assigned to H. habilis reliably demonstrates that this species survived until later than previously recognized, making an anagenetic relationship with H. erectus unlikely. The discovery of a particularly small calvaria of H. erectus indicates that this taxon overlapped in size with H. habilis, and may have shown marked sexual dimorphism. The new fossils confirm the distinctiveness of H. habilis and H. erectus, independently of overall cranial size, and suggest that these two early taxa were living broadly sympatrically in the same lake basin for almost half a million years.“

I’ve read the paper and for a two page report on the fossils, they spend a remarkable majority of it making the case that these two fossils indicate the co-occurrence of H. habilis and H. erectus. Instead of describing and documenting the fossils they focus almost entirely on the claim that these two hominins occupied the same evolutionary time frame and ecology.

The two fossils used to make this case are KNM-ER 42700, a hominin calvaria and KNM-ER 42703, a right hominin maxilla fragment that were both found by their team. The photo I put up above are the respective fossils.

The second paragraph is where things become fudged up. The authors begin to justify how they classified each fossil instilling me with some doubt,

“Although [KNM-ER 2700] is closer in overall size to H. habilis we assign [it] to H. erectus.”

It is often smart to cautiously represent one’s work in journals like Nature and to use careful ways to describe what you say, but there’s something about reading that sentence that made me raise an eyebrow. It’s not like that paragraph got any better,

“A multivariate analysis of the calvarial dimensions confirms the affinities of KNM-ER 42700 with H. erectus. Some characteristics often considered diagnostic of this species (for example, a thick cranial vault and supraorbital torus, and strong occipital angulation) are lacking in KNM-ER 42700…”

I don’t get it. KNM-ER 42700 is not the same size as H. erectus nor does it have the characteristic morphology of H. erectus. They even say it themselves. So why the hell is it an erectus? Am I missing something here?

KNM-ER 42703’s classification as H. habilis is flawed too. To classify this maxilla, they get a crazy amount of detail out of these heavily cracked out teeth. I don’t understand how anyone could get measurements out of teeth so damaged. Don’t believe me, click on the photo of the teeth to your right. See how worn down some of the teeth are, like the crowns of the teeth? They even acknowledge that the teeth,

“exhibit heavy occlusal and interpromixmal wear as well as postmortem weathering. Little to no cusp morphology is present.”

Despite this, Spoor et al., decide there is enough measurements to do a principal components analysis (PCA) to determine if KNM-ER 42703 falls as erectine or habiline. A quick side note, PCA is a statistical test used to reduce multidimensional data sets to lower dimensions for analysis.

What gets me, and what should get anyone with an undergraduate education in statistics is, why they did a PCA with only two variables?! If I had to report on the measurements of these molars, I woulda done a bivariate plot if I were them. A bivariate plot graphs the relationship between two variables that have been measured on a single sample of subjects, a much more fitting test than PCA. The fact they used a PCA tells me either they don’t know what sort of tests to run or they specifically used PCA to render their analysis obscure.

I won’t get into the whole sexual dimorphism stuff they touch on. I’m really confused on how their fossils fall into the anti-angenesis camp. When I take their word choice, the quality of the teeth they decided to measure, and the statistical tests they ran, I can’t help but think that they are warping the data and observations to make it work for them. And by no means am I saying that angenesis was the way humans evolved. They could be right… but these two fossils don’t prove it to me.