has given so much consideration to the inscriptions and their interpretation. His objections afford an opportunity for going still deeper into the problems and they help us in arriving at, or at any rate near, the final solution. I have re-examined the whole question and the letters on the statues with reference to the criticisms of Dr. Barnett. I shall essay to answer his objections and shall also mention new facts bearing on the question which I have come across in consequence of or during the controversy.

I must also thank here Mahamahopadhyaya Haraprasad Shastri who gave several hours to the study of the letters on the stone. It is only an immediate scrutiny of the inscriptions which can give a sure basis for conclusions. No paperimpression can bring to clear light the difficult lettering in all its detail. Here the testimony of a savant of Mr. Shastri's position is invaluable as scholars abroad cannot physically see tae monuments.

We have now the further advantage of the testimony of Mr. Green, the stone expert. To bring about finality in the controversy as to the actual forms of the disputed letters and the interrelation between the letters and the drapery lines, I thought of obtaining the independent judgment of a techinical expert in stone who would examine the letters as mere incisions and trace their forms according to his technical knowledge of the chisel. For this purpose I sought the help of Messrs. Martin & Co. of Calcutta, the well-known firm of architects and engineers, who very kindly asked Mr. F. Green, their expert in charge of the construction of the Victoria Memorial, to examine the lines and letters on the statues and give his professional opinion. The result of his examination is published along with his drawings of the letters in controversy. (See Note below) I also give a tracing of the letters, in full size, kindly prepared by Mr. Bishun Swarup, Superintending Engineer, Patna, from paper casts of the inscriptions.

Dr. Barnett questions my assertion that the finishing touch to the statues was given after the letters had been engraved.

Mr. Green's examination proves conclusively that the letters on the statue of Aja had been drawn and cut before the drapery lines. This is abundantly clear even in the relief impression published in the Journal s e, for instance, the broken character of the lines between the mātrās of ni and the next letter. Where the drapery lines coincided with the lettering the letterportion has been redeepened and consequently bears a lower level than the drapery line and other portions of the letters. In the other inscription, on that of Vața Nandi, the lines and the lettering were chalked out together. This is proved by the fact that the drapery line No. 4 (counted from the bottom) coming against the curved body of the first letter, stops absolutely just against it.1 In this record also both deep and shallow levels in the same letters occur, the deep portions being the result of the method of distinguishing letter-forms from drapery lines. The bend of the line over the head of letters v and t and its higher altitude over the vertical bar of S indicate the contemporary existence of the letters and their influence on the drawing and cutting of the drapery lines just above the letters. The result is that both inscriptions are contemporary with the construction of the statues. The letters are not placed "upon "the lines (as Dr. Barnett says); they hang from the line. Placing of letters on lines is the system of European scripts, not Indian. "The accurate way in which the letters are placed upon the cross-lines" leads, if to any, to the conclusion that those lines were accurately placed below the lettering.

The decision of this point has a tremendous bearing on the question of the age of the script. No responsible scholar can allege that these statues are post-Mauryan. The dated technique of Asokan and Sungan monuments and the wide difference between the two will compel any art critic to place our statues before the S'ungan times. The same will be the con

Mahamahopadhyaya Haraprasad Shastri thinks that the line is recommenced after a break on the other side of the curve which gives to the arm of the letter the appearance of an arrow-head. He has therefore omitted the lower part of the arrow-bead in his drawing (compare his drawing with Mr. Bishun Swarup s sketch). The break of the drapery line is absolute in any case.

clusion of an unbiased observer who has observed both the Parkham and the Patna images. The Parkham is the only monument which comes in line with the art and the material of the Patna images and the Parkham bears an inscription in a known script which has been declared to be Mauryan (Vogel, Catalogue of Mathura Museum, p. 83. "It has an inscription in Maurya Brahmi "). It is thus impossible to assign a post-Mauryan age to the statues as statues. Hence Dr. Spooner, and I understand also Sir J. Marshall, are of decided opinion that the statues are old (Mauryan). Epigraphists thus have to face the fact that the statues are at any rate not later than the Mauryan period. At the same time, their preconceived ideas lead them to think that the script of the inscriptions is late. They as well as archeologists are consequently forced to postulate that the statues are old but the inscriptions later and subsequent. But the real situation is that if the statues are not post-Mauryan the inscriptions cannot be post-Mauryan, both being simultaneous and contemporary in execution. The statues bearing the "Mauryan polish" and " Mauryan art" bear inscriptions which are not of the Mauryan script. Then what is their script?

A

45

Before entering into the question of paleography I would take up "the serious linguistic difficulties "proposed by Dr. Barnett. He objects to the change of j into ch (Acho=Ajo) in an inscription coming from Patna as the change is a Paisa chi characteristic and the Paisachi country is the North-Western Frontier (Grierson, Z.D.M.G., 46, 49; J.R.A.S., 1912, 711). I had admitted

1 Professor Foucher writes:-"As to the Patna statues, I have no hesitation in telling you that I would, from the point of view of their artistic treatment, ascribe them to the second century B.C., and from the analogy of the Parkham and Bharhut images, identify them as yakshas, that comes to say that I share the common opinion prevailing on the subject". At the same time the learned Professor admits that the judgment is open to revision :-" Of course, I admit that this judgment, as every human one, is open to revision." I agree with the Professor that before we revolutionize all our ideas about the history of Indian Art positive proof should come forward. This proof can be afforded, in my humble opinion, by a solution of the inscriptions to the complete satisfaction of epigraphists.

this characteristic but demurred to its being exclusive to Paisachi. Two examples given by me in support of the change in non-Paisachi areas are regarded by the learned Doctor as " disputable". One of them is vrach=vraj (Aśoka) and the other, pachana (Pali literature). Taking vraj, to show that no room for dispute really exists, I quote a sūtra of the Prakrit Grammar Prakrita-Manjari where it is expressly recognized "cho vrajanṛityoh" (VII. 44, page 117, ed. Nirnaya Sāgara). For the second example, I beg to refer to Müller's Grammar of the Pali Langnage, p. 38 (1884), where to exemplify the process of "hardening of a soft consonant", i.e., the Paisachi process, Müller gives "pacheti " (pra-j), and pachana. Here in the inscription the change is similarly in aj. Another point of objection is that the change is too sporadic to justify the change in the name of the king side by side with unchanged consonants. But phonetic laws make no distinction between names of kings and gods and names of commoners and animals. If there is a law of sporadic change it may choose any victim. Take for instance the Paisachi changes in the very name of the mighty Kubera on the Bharhut railing as Kupira (L. B. I. 794) while side by side other letters enjoy immunity from this vulgar profanity. Similarly see Bhagapata for Bhagavata at Amaravati (L.B.I.1271). Here Bhaga is free like Bhage in our inscription. Take again the commoner Fitura for Vidhura and the animal Mugapakiya (=pankhiya) at Bharhut; also see Makhadeva Maghādeva in the Jātaka (J.R.A.S. 1912, p. 406). These amongst numerous are all cases of Paisachisms outside the Paiśāchi jurisdiction, and sporadic, side by side, with unchanged consonants. Dr. Barnett's objection to the change of v into p in sapa is similarly answered. See examples given by Müller at p. 32 (lāpa=lāva, paiāpati=prajāvati, palā pa=palāva, chhāpa= sava, Sapadāna Savadāna, Supāṇa=suvana, dog, dhopana= dhovana). See also Erapato for Airavata at Bharhut (L.B.I,752).

=

The change of j into ch which later Prakrit Grammarians regard as characteristic of the North-Western dialect, is known to the official Pali ".J.B.O.R.S., V, p. 102.

All these are sporadic changes side by side with unhardened consonants. Inscriptions prove that nowhere was the strict or ideal Prakrit of the grammarian spoken. Prakrita or vulgar tongue had its own way when Prakrit languages were alive. There are forms and changes in the very names of kings and rulers in official records which followed the convenience of the tongue of populace or idiosyncracies of individual scribes and engravers and not the philological laws of grammarians. How gramma. rians' phonetics was at a discount may be seen from Bṛihāsvātimitra and Bahasatimitra, one appearing in a near relative's record and the other on coins, the original of the two being one and the same Brihaspatimitra; Gudaphara, Galaphara, Gudapharna, all on the coins of the same Gondophares, Mālavāņa Jaya and Malavahna Jaya (C.M. 172) on coins of the same time and people are examples of popular phonetics of the tongue as against iron phonetics of books. Ramains prove that it was the former which ruled in every-day life and not Vararuchi and Pischel. In these circumstances it is enough for my position to show from literature and stone that changes of j into ch and v into p did take place in Magadha and that the so-called Paisachisms were not confined to the Pais'acha country.

The remaining linguistic question raised by Dr. Barnett is about the nominative endings in Bhage, Acho and chhonidhise. This situation he considers "manifestly impossible". He seems to think that every case-ending should have followed either what the grammarian calls the ardha-Mugadhi e-ending or the Māgadhi o-ending. To see that the situation far from being impossible, is very common, take instances in Aśoka's inscriptions: "Satiyaputo Kelalaputo Tamba pamni Amtiyoye" at Kalsi, "rajuko pradesike" at Shahbazgarhi (against rājuke cha pradesike at Girnar), at Shahbazgarhi in section X" Devana priye" and then "Devanam priyo," in section XI dhrama sam slave dhramasamvibhago (against dhammasamstavo vā dhammasamvibhago of Girnar), Devānam priye at Girnar in the beginning of section XII while a few words after Devanampiyo and throughout 'piyo