In 1013 the late Dr. Theodor Bloch, Ph.D., of the Calcutta Museum, made an attempt to decipher the inscriptions on these two statues. But they baffled his attempt. The word Namdi could be deciphered by him with portions of the other words. He, however, did not publish the result of his researches as he was not sure of his interpretation. It was at that time that the paleography of the records was carefully examined. I did this work under Dr. Bloch's supervision but the result of my investiga tions, too, were not published at that time at his request.

In 1913 Dr. D. B. Spooner, then Superintendent of the Eastern Circle, consulted me about the date of these two sculptures. He was of opinion that they were specimens of Mauryan Art and thought so because of the high polish on them. When I pointed out to him, the peculiarities of what I then considered a later script used in the short records on the monuments, he told me that most probably the inscriptions were later in date than the sculptures. I did not agree with him at that time but it seems to me now that probably Dr. Spooner was correct in assigning a later date to the inscriptions than the sculptures on which the records are incised, if the inscriptions turn out to be post-Mauryan.

As to the reading of the i scriptions I agree entirely with Mr. Jayaswal in his reading of the inscription on the statue of Varta-Nandin. There is only one defect in it. In this record the second syllable of the first word is ba and not pa. The meaning is not affected in the least, as the word in both cases remains to be the same (sarva). I examined the original very carefully once again in Calcutta and I find that the top bar of the square ba is partly distinct and in part faintly traceable on the stone. In 1903 Dr. Bloch and I read wrongly the first word as yakha (Skt. Yakṣa), because we failed to discern the vertical upper limb of the sa which is faintly discernible on the stone.

I am afraid I cannot agree entirely with Mr. Jayaswal in his reading of the inscription on the statue of Aja for the following reasons :

(1) The syllables read by him as bha in bhage and dhi and se in chhonidhise are not sure results. The first syllable of bhage may be bha because it has some resemblance to the Mauryan and later Mauryan bha but the absence of the right upper vertical which is characteristic of this consonant is missing. Even in the Bhattiprolu records the upper vertical is present though it is on the left instead of right. On the other hand there is a short covered hook attached to the left upper corner which cannot be explained.

(2) The syllable read by Mr. Jayaswal as dhi in the last word of this record appears to me to be vi. Va lost the upper vertical line on the top in the first century B.c. This vertical is to be found in the later Mauryan inscription from Mathura1 but it disappears in the inscription of Sōḍāsa.2

(3) The syllable read by Mr. Jayaswal as se in the same word appears to be Ko of the first century B.C. or A.D. The Keil stroke, which is the precursor of the serif, is visible on the original. The form of the syllable as well as the form of other letters in the same inscription indicate that it should not be read as se. The oldest known form of the palatal sibilant is to be found in the record, probably pre-Mauryan, incised in one of the caves at Ramgarh in Sirguja State. This form very nearly resembles the later form of Punic shin. Had the last syllable of the word read by Mr. Jayaswal as Chhonidhise been really se then the resemblance between the Jogimara form of the palatal sa and this one would have been easily noticeable.

3

The forms of the majority of letters in both of the inscriptions show that the records should be regarded as later in date. I shall take them in order and show their resemblance to forms in other inscriptions of undisputed date. The

1

Epi. Ind., Vol II, p. 198, No. 1.

2 lbid., p. 199, No. 2.

3 Arch. Annual, 1903-4, p. 128. pl. XLIII b.

second letter of the first word on the statue of Aja is read by Mr. Jayaswal as g1. It is correct but the form of the letter is late. The Mauryan ga has an acute angle at its top whereas the top of this letter is round. With this compare ga in Phaguyassa in an inscrip:ion which must have been incised at the same time as the record of the year 72 of the reign of Soḍāsa.

(5) The vowel in Aco very closely resembles in form the same vowel in the Sarnath inscription of the first century B.C. or A.D. Cf. A in Asvaghosasya. 2

(6) In the same word the form of ca in the second syllable is certainly much later. The Mauryan form is quite different. The form in the Patna inscription resembles that in a Mathura inscription of the year 52 of the Kuṣaṇa era, cf. cha in váchakasya in 1. 2.

(7) The form of cha in Choni is also later. The Asokan form consists of a circle bisected by a vertical straight line which projects above the upper periphery of the circle whereas here we have a vertical straight with two ellipsoid curves attached to its lower extremity, one on each side. The Asokan form persisted for a long time, cf. vachiputrasa. 1 The form in the Patna inscription resembles the Kuṣaṇa form, cf. Chatra in the Kuşana Buddhist inscription from Sarnath. "

Examined palæographically the inscription on the statue of Varta-Nandin also poiut to the same conclusion. The following facts indicate that this inscription also should not be regarded earlier than the first century B.C. :—

(a) The triangle instead of a round figure or circle as the base of kha in khate.

(2) An isosceles triangle as ra in choniriko.

(c) The curvature in the base line of na in choni; cf. also

the form of na in Namdi.

1

Epi. Ind, Vol. II, p. 200, No. V.

2 Ibid., Vol. VIII, p. 171, inscription e and ƒ.

3 Ibid., Vol. II., p. 203, No. XVIII.

Ibid., p. 199, No. I.

Ibid., Vol. VIII, p. 176ff.

(d) Two right angles instead of a semi-circle in the back of da.

A careful scrutiny of the original inscribed surface enables me to assert that the records were incised on the statues after the finishing of the sculptor's work.

We do not know any other examples of pre-Mauryan art and consequently we cannot make comparison. It appears to me, however, that the statues were finished and exhibited in a gallery of the sort described in the Pratima naṭakam. Long after, when people had begun to forget who the Śaiṣunākas were, somebody connected with the Art gallery had the names chiselled on the monuments in an inconspicuous place.

NOTE ON THE A OVE.

1. The letter which I read as bh is not explained by Mr. Banerji and cannot be explained on the theory of a late script. The letter has to remain unidentified, as in Mr. Banerji's note, on the late theory, and consequently the whole word (bhage) unread.

2. The peculiar composition of ga is not noticed. It is composed of two parts. Then it is not correct to say that the Aśoka ga is always angular; see Siddapur in Bühler's Chart.

3. Kha has a quadrilateral base, not triangular, which again is impossible to be explained on the late script theory.

4. If viko is read instead of my dhise (or dhiso), the result is a senseless word; chhoni-viko gives no meaning. Then, two different forms of va are to be read in the inscriptions, one with the top-bar and the other without it, which is inadmissible. On the late script theory the letter which I read as dhi cannot be read at all. I agree that the Ramgarh inscription (Jogimara) is preAśokan, not pre-Mauryan. But to call it "the oldest known" is to beg the very question. The Kalsi and Girnar sa's are nearer the Saiśunāka letter. It is radically different from the later ka's, being a three-stroke letter.

The Kushan and Western letters preserve the tradition of older forms and very probably a different style than that adopted officially under Aśoka.

5. Mr. Banerji leans to the conservative view, but he fails to read all the letters on his hypothesis and to give any sensible meaning to his new readings wherever he differs from me. Until and unless all the letters can be identified and explained on the hypothesis of a late script, I am not prepared to accept that hypothesis.

My arguments on the evolution of the letters based on the stroke-effort have not been considered by Mr. Banerji. I think that it is axiomatic that a three-stroke letter must be older in origin than a one-stroke or two-stroke representative thereof.

I am, however, very glad to see that Mr. Banerji agrees with me in the reading of the proper names and in the general result.