plex description; and, further, though we have not, as yet, the materials for making anything but a provisional classification of these languages, yet one or two of what appear to me to be errors in Max Miiller's classification may be corrected. Sarpa, for instance, is a mere dialect of Tibetan, and Changlo is very closely related; and yet Tibetan, Sarpa, and Changlo are all three placed in separate subdivisions. Burmese again, which is not classed with Tibetan, is certainly more like Tibetan than it is to some of the languages, such as Dhimal and Kachari, with which it is classed. Further, I think a more detailed classification than Max Müller's may with advantage be adopted, by dividing these languages into subclasses based upon differences of verbal structure or other characteristics.
My object in this paper is to give a brief sketch of the different groups into which the non-Aryan languages of India may be formed, derived from more recent materials than those which were available to Max Müller. These materials are described by Mr. Oust in certain communications made by him to the Philological Society, and published in the two last Annual Reports of that Society. Some of the books referred to he procured from India, as they were not to be had in this country. Another object I have in view is, to call attention to the very little that, after all, is known of most of these Indian languages, in the hope that, if the necessity for further investigation is prominently noticed by such a Society as the Asiatic, scholars in India may be moved to do something to that end. It is not to our credit that so little on the whole should be known of the languages spoken by our nonAryan fellow-subjects in India, or by those non- Aryans on the immediate frontiers of British India with whom our officers are frequently brought in contact. The Dravidian group, as was to be expected from its importance, has received the most attention; yet, even here, much remains to be done. It is supposed that this group extends almost continuously from Cape Comorin to the Ganges. The most northerly member of the group is supposed to be the Rajmahali; but we have only a very meagre vocabulary of the language, and
