with the body and its parts alone, as is done below. It would not be possible to produce between any two languages the same amount of detailed comparisons as can be made between Dravidian and Uralian without giving reason to believe that those languages were themselves related. Because however Schrader's article contained only a fraction of the available material, it was possible for Lewy to use this method with some plausibility, and his arguments seem to have had considerable effect, since no further notice was taken of Schrader's article from the Finno-ugrian side.

A work published shortly afterwards by W. von Hevesy purporting to prove that the Muṇḍā or Kolarian languages of India are related to Finno-ugrian, is deserving only of being passed over in silence.¹³ Unfortunately it seems to have impressed Schrader, a fact which tends to throw doubt on his own competence to deal with the matter. In consequence we find him returning to the subject with an article entitled On the "Uralian" element in the Drāviḍa and Muṇḍā Languages,¹⁴ in which he attempts to reconcile his own doctrines with those of Hevesy. They cannot, of course, be reconciled if it assumed that Dravidian and Uralian are related languages in the proper sense of the term, i.e. descended from the same original language, since the Dravidian and Kolarian languages are obviously unrelated. Schrader had, however, in the first place been undecided whether to assume genetic relationship or prehistoric mutual influence, and now influenced by Hevesy, he is inclined to believe that the latter theory can be applied to both groups of languages. This leads him to some implausible theorizing about two different streams of "Uralians" having entered India and influenced the one the Dravidian and the other the Kolarian languages before presumably their own languages died out. This and the alternative theories he offers¹⁵ do not carry much conviction, nor is there any reason why the common material of Dravidian and Uralian should not be explained by the theory of genetic relationship, while the evidence for a relationship between Uralian and Kolarian is easily seen on examination to be illusory.

In returning to the subject in this paper Schrader adds somewhat to his material by giving a list of words denoting parts of the body which he regards as common to Dravidian and Uralian. This list contains seventy items, out of which between a quarter and a third constitute acceptable equations. In view of the large percentage of error in this list, it is unlikely to create much impression

------------------------------------------------------------------------------------------------------------------

¹³ Finnisch-ugrisrhes aus Indien, Wien, 1932. See the review by R. L. Turner, JRAS., 1934, pp. 799 ff.

¹⁴BSOS., VIII., 751-762.

¹⁵ Op. cit., p. 762.