back to the period prior to the differentiation of the Northern and Southern scripts of Brahmi. In the north the vertical bar evolved to 'the right and in the south to the left, very probably in the way as shown in the diagram on the plate attached,

We have thus two letters in the Aja inscription about the reading of which there is no controversy (bh and ch) which are unquestionably pre-Mauryan and removed from As'oka by centuries. There is the third letters which if my reading is accepted, and my reading alone gives a meaning, will have also to be admitted as pre-Mauryan.

Dr. Barnett does not suggest that our chh is late, nor does he demur to my assertion about t, t, v and d, in the other inscription being old. In fact the here, even according to Bühler's theory, is the oldest form (see his v reproduced in the plate). It is oldest also on the basis of my theory, circular and oval forms requiring less effort, came later than the straight base V. The d-form of our inscription is seldom, if ever, found again after As'oka (Delhi). About the form of kh Dr. Barnett is mistaken when he says that it is like the type of Mathura (150-100 B.C.), for the Mathura letter is a triangle while ours is four-sided (see Mr. Green's diagram and other drawings), and the body is to the left. Such a kh was unknown to Indian epigraphy up to this time. Thus the position about the second inscription is that four out of the undisputed letters are unquestionably early and about one further (h) it is impossible to allege a late date for want of a second example. In the first inscription, similarly leaving out the disputed two letters, at least four letters out of the remaining six are such which Dr. Barnett would not call late on his own theory. As to the remaining letters I shall presently show that their forms do go to a period beyond As'oka. But even on the basis of admitted old letters the question arises: the statues in age being on the evidence of the polish and the Parkham image at least Mauryan and the script being un-Mauryan, what is the script in question? It may be equally earlier than As'oka or later than As'oka. Lut

as we have records from every part of the country in number for the period later than As'oka, and to none of the post-Mauryan scripts the statue script corresponds,-no script is found where the "tailed" ch and barless bh, the crossbar a and the foursided th and the four-angled d occur together or occur at all— the statue script must necessarily be pre-As'okan.

That before Asoka's time scripts other than As'okan were in existence is proved by positive evidence. That is the evidence of the coins of the Akhaemenians current in North-Western India, the Persian sigloi. They are the only documents about whose pre-As'okan date there is not the least controversy. The Persian empire was destroyed in 331 B.c. by Alexander and their rule in India had come to an end much earlier, very probably about 400 B.C., in the time of Darius II who lost the greater portion of the outlying Persian dominions. Ktesias writing about 416-398 B.C. in Persia speaks of the "Indian king" and Alexander found the Punjab independent. The Persian coins in India thus would be dated between 500 B.C. and 400 B.C., or at the latest about 350 B.C. Now what do they prove? They completely destroy the theory of Bühler in respect of his supposed pre-As'okan development of Brahmi letters. According to his theory the sigloi letters ought to have been post-As'okan by centuries. But the coins are admittedly at least a century older than As'oka. Bühler had to concede "But the shape of the characters on the Persian sigloi makes it probable that even its (that of Maurya alphabet) more advanced forms existed before the end of the Akhaemenian rule in India 331 B.C." (I.P. 33). Bühler had also to face the occurrence of numerous forms in Aśokan inscriptions what on his hypothesis that Brahmi originated from a Semitic script of known forms between 800 B.C. and 500 B.C.-he would treat as later by centuries (" Kushana, Mathura, Andhra, Abhira," I. P., 7). In view of these facts he said about the "apparently or really advanced" types and "modern looking signs from Aśoka's cdicts" that

The existence of so many local varieties, and of so very numerous cursive forms, proves in any case that writing had had a long history in Asoka's

time, and that the alphabet was then in a state of transition. The use of the cursive forms together with archaic ones may possibly be explained by the assumption that several, partly more archaic and partly more advanced alphabets were simultaneously used during the third century B.c., and that the writers, intending or ordered to use lapidary forms, through negligence mixed them with the more familiar cursive letters, as has also happened of not rarely in later inscriptions. It is possible to adduce in favour of this 11 view the abovementioned tradition of the Dṛṣṭivada, according to which a larger number of alphabets was in use about 300 B.C. The conjecture gum would become a certainty, if it could be shown that the word seto, the

white (elephant),' which has been added to Dhauli edict VI in order to T). explain the sculpture above the middle:column, was incised at the same time at ( as the preceding edicts. The two characters of seto show the types of the Kusana and Gupta inscriptions. Though it is difficult to understand that, 4). in later times anybody should have cared to add the explanation of the relief, keeping exactly the line of the edict, the possibility of the assumption 0 M that this was actually done, is not altogether excluded.

rath

Then again,

Delhi

The forms of the Brahmi and Draviḍi, used during the first 600 years, are known at present only from inscriptions on stones, copper-plates, coins, seals and rings, and there is only one instance of the use of ink from the S third or second century B.C. The view of the development of the characters lina during this period is, therefore, not complete. For, in accordance with the air results of all palæographic research, the epigraphic alphabets are mostly more archaic than those used in daily life, as the very natural desire to vel employ monumental forms prevents the adoption of modern letters, and as, in the case of coins, the imitation of older specimens not rarely makes the Sa alphabet retrograde. The occurrence of numerous cursive forms together with very archaic ones, both in the Asoka edicts and also in later inscriptions, clearly proves that Indian writing makes no exception to the =S general rule

In other words, the As'oka letters do not furnish a sure) criterion to determine the age of the letters actually in use in and before As'oka's time; in business and daily life they were later in form; they were later in form even during the Persian rule, a century (if not more) before As'oka's time; and there were other scripts current side by side with As'okan and protoAs'okan scripts. This admission is quite enough for my purposes. When I show that there are letters in our inscriptions of whichin one, if not in two, are the "tailed" ch and the

}

na

Si

S