In choosing the subject for this short paper I am guided not solely by the fact that the archæological observations which first drew my attention to it were gathered in that field of my Central-Asian explorations with which I have fortunately been able to associate my old friend Professor Rapson as one of the earliest and most helpful of my collaborators. What invests certain curious finds among modest burial remains of the Lop desert with a special quasi-personal interest for me is the distant and puzzling relation they bear to a much discussed question of Vedic and Avestic research, that of the sacred Soma and Haoma.

It is a question which was often touched upon in his lectures by that great scholar and teacher, Professor Rudolf von Roth, during the years 1881–4, when I had the good fortune, figuratively, to sit at his feet as an eager devoted pupil. The question as to the identity of the original Soma plant and its home which he had discussed just at that time in two short papers of masterly clearness, was not to be solved then, and still remains undecided. But Roth's main contention still holds good that a solution for it could be hoped for only by the study of relevant physical facts, if possible, on the ground of early Aryan occupation.

Since the investigations of Oldenberg on the language of the Lalitavistara, it is no longer a useful question to ask whether the prose portions or the verses are the older. There is more than one layer of verse as well as of prose. Still less is it sufficient to describe it as “ a poem of unknown date and authorship, but probably composed in Nepāl, by some Buddhist poet who lived some time between six hundred and a thousand years after the birth of Buddha ”. As Oldenberg has shown, there is an older layer of verse in fairly good Sanskrit, which rests on passages in a dialect closely related to Pāli, and which was hence easily sanskritized. There are also the poems in so-called mixed Sanskrit, mixed just because they were once in a dialect that resisted all efforts to fit them with a proper Sanskrit dress, and still later are the verses which may have been originally composed in Sanskrit.

In the Kharoṣṭhī records from Chinese Turkestan, of which the edition commenced by the Abbé Boyer, Professor Rapson, and Monsieur Senart, has recently been completed with important dissertations and index by Professor Rapson and Mr. P. S. Noble (Kharoṣṭhī Inscriptions, i-iii, Oxford, 1920-9), the vocabulary is composed in the main of recognizably Indian terms or of personal or topographical designations belonging to the locality ; but we can discriminate a relatively small number of words having other traceable origin or obscure signification. To the last mentioned group we mayassign the words mukeṣi and lote (loteya, lode).

These two terms, although they do occur apart, are apt to be found in more or less close conjunction; and the general sphere of their meaning may be ascertained by considering one of the passages where they are associated. We may take document No. 474 (p. 171) of the edition, which, except as regards spacing, majuscules, and some added punctuation, is followed in all respects.

The normal suffix of the future in Sanskrit was –syá-. The root had full grade, e.g. kartsyáti, jeṣyāti. It follows that in the futures of dissyllabic (seṭ) roots the suffix was preceded by i (< IE. ә), e.g. jani-ṣyāti, bhavi-ṣyāti. Even monosyllabic (aniṭ) roots ending in –r (-ͬ) had an i between the root and the suffix, e.g. mar-i-ṣyáti. The Ŗgveda adheres to this division except for one apparent exception, asiṣyát-, fut. part, of ásyati. But, despite ástra- and astá-, ásyati perhaps contains a dissyllabic root **esē(i) (cf. ásira- ‘ ray ’). Of this ásy-ati (which, as a transitive verb, probably does not contain the suffix –ye- of the intransitives of the type táp-yati) and ásira show full grade in the first syllable; á-sát ‘ he has slain ’ and sāy-aka- ‘ javelin ’ show full grade in the second syllable. In that case asi-ṣyá should be analysed like jani-ṣyá-.

The remarkable group of five rock-cut temples at Māmallapuram or Māvalivaram, to the south of Madras, has often been described. Popular imagination has associated these wonderful shrines with the Pāṇḍvas; thus it has happened that the one which is smallest in size has become known as the rath of Draupadi. Evidently this temple was in reality dedicated to some form of the goddess Durga, whose effigy, standing on the severed head of the Buffalodemon, is found carved upon the outer wall, whereas her vehicle in the form of a well-conceived but unfinished lion-statue may be seen at no great distance.