国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

whom he measured, but from indications afforded by hair- and eye-colour it is obvious that the Galcha are very
closely allied to the Wakhi, and stand in closer relation to the Pamir peoples than to the Kirghiz group. Their
percentages are as follows, and can be compared with Tables 13 and 14. Hair: black, 9·4 per cent.; dark brown,
50·2 per cent.; red, 1·9 per cent. (one individual); medium and fair, 37·7 per cent. Eyes: dark, 11 per cent.;
medium, 60·3 per cent.; light, 20·7 per cent. The ΣΔ shows that the Galcha are definitely opposed to the
Chinese, the Ladakhi, and the Indo-Afghans. The fact that the ΣΔ for the Kafir is 4·97 indicates that
the element which differentiates the latter people from the Pamir folk proper is evidently not Iranian.

Pathans.—The affinities of the Pathans in terms of ΣΔ are shown in Tables 16, 18, and 19. As regards
resemblances, the connexion of the three Indo-Afghan peoples is at once obvious; moreover, the fact that the
ΣΔ for the Kafir falls under 3·50, taken with the further fact that Table 18 shows the Pathan to be far
removed from the Kirghiz group and from several important tribes of the Desert group, leads us inevitably
to the conclusion that the element which differentiates the Kafir from the Pamir and Desert peoples is
Indo-Afghan.

Biloch.—The same three tables show the affinities of the Biloch, whose higher degree of brachycephaly
brings them into relation with the Sarikoli and Mastuji. It has been shown that the Mastuji are closely related
to the Kafir, and the Sarikoli to the Mastuji. We may conclude, therefore, that the Biloch contain an appreciable
Iranian element, and, possibly, that the Mastuji and Sarikoli may possess a slight, but very slight, Indo-Afghan
strain. The latter supposition is problematical, but receives a little support from the fact that these two tribes
have on the whole darker hair than the Wakhi, who seem to stand nearest to the comparatively pure Iranian as
exemplified in the Galcha.

Dard.—The most interesting point with regard to this people is that they appear closely related to the
Chitrali, while their comparatively low ΣΔ for the Kafir, 3·10, is invalidated by a Δ of 1·40 for the nasal length.
Now, it will be remembered that the ΣΔ for the Kafir and Chitrali was only 3·92, but that Δ for the nasal
length was 1·00. We find, therefore, that the Dards differ from the Kafir in the very same respect as the Chitrali,
though to a greater degree. The averages for the absolutes are: Dards, 53; Chitrali, 51; Pathan, 50; and
Kafir, 46. It is evident that some strain of Indo-Afghan blood enters into the composition of the Chitrali,
and it is difficult to account for the shortness of nose among the Kafirs, a characteristic which is shared, though
to a less degree, by the Sarikoli and Mastuji. This shortness of nose cannot be due to Galcha influence, since
the average of that people is 52. Measurements of more of the surrounding tribes are necessary before this point
can be elucidated. The Dards show a considerable dissimilarity to the Kirghiz and Desert groups.

Ladakhi.—It is clear that we are badly in need of the facial measurements to define accurately the relation of
the Ladakhi to the tribes under discussion. As far as the measurements obtainable go, this people show
affinities with certain of the Pamir group, of the Desert group, and the Chinese and Loplik. At the same time
they display a definite divergence from the Kirghiz group. They are probably a very mixed people, and contain
Indo-Afghan, Tibetan, and possibly Pamir elements.

Tibetans.—The most noticeable point about the Tibetans is that their ΣΔ in relation to all tribes is very low.
It is strange that the highest should be that expressing their relation with Niya, one of the villages of the
southern desert, which appears to be an aberrant member of the Desert group, and which might be supposed
to owe its peculiarities to Tibetan influence. On the other hand, there seems to be some kinship between the
closely inter-related members of the Desert group and the Tibetans, as was suggested, with respect to Khotan, in
my previous paper, to which allusion has been made above.

Final Conclusions.

To sum up, the measurements show that the majority of the peoples surrounding the Taklamakan desert
have a very large common element. Further, that this element is seen in its purest form in the Wakhi. The
fact that the Wakhi display so close a relationship with the Galcha proves that the basis of the Taklamakan
population is Iranian. At the north-western edge of the desert an intrusive element, which can be sharply
differentiated from the Iranian, makes its appearance—the Turki element. Besides this there seems to be some
common bond between the peoples of the desert and of Tibet. This probably means that the Iranian element
has penetrated to Northern Tibet, though it is not unlikely that Tibetan (modified Mongolian) influence has been
exercised, to a slight degree, upon the Desert peoples. In any case the relationship with Tibet requires