Below is a K = 3 craniometric admixture plot.
The K = 3 analysis produces an archaic component (blue), a Mongoloid modern component (green), and a non-Mongoloid modern component (red).
The archaic component is on average modal for Australians, Melanesians (Tolai), Tasmanians, some Africans (Teita, Zulu), and some Polynesians (Easter Island).
Once again, Broken Hill I has less of the archaic component than Upper Cave, Djebel Irhoud 1, Steinheim, Shanidar 1, La Chapelle, La Ferrassie I, and Predmost III. In this analysis it even has less of the archaic component than Markina Gora, Djebel Qafzeh 6, and Skhul V.
The Dogon of Africa have a lot of the archaic component, but on average the non-Mongoloid modern component is modal for them. The same is true of Bushmen.
Buriats are the most Mongoloid population in Howells’ main data set.
The Mongoloid component is on average modal for Eskimos, but they have a lot of the archaic component.
The Chinese specimens in Howells’ data set are from South China, and they show some of the Mongoloid component, but not much.
In some cases presence of the Mongoloid component is actually due to Mongoloid admixture, but in other cases it’s spurious. One of the main distinguishing characteristics of Mongoloid cranial morphology is brachycephaly, and as a result brachycephalic non-Mongoloid individuals will often show up as having some of the Mongoloid component. This error is especially frequent among the brachycephalic Alpine Caucasoid specimens from Berg, Austria.