पृष्ठम्:बृहद्देवता.djvu/७

Introduction] ACCOUNT OF THE MS. MATERIAL [xi nizable principle; for the reading chosen is not only constantly opposed to the predominant evidence of the MSS., but is frequently by no means the best a priori. Based mainly on one MS. (or two which are practically identical), it shows haphazard selections and additions from MSS. of an- other recension. I have estimated that the edition contains one error of some kind on the average for every śloka, making a total of about 1200. I may refer the reader to p. 85 as a specimen, where a passage of three ślokas (93-95), as a comparison with my text (iii. 95-97) will show, contains twelve mistakes, and is consequently quite unintelligible. Since Mitra gives no information whatever about the MS. material he used, his critical notes are hardly of any value as far as his own edition is concerned. They have, nevertheless, proved quite useful to me in the light of the MSS. which have been at my disposal. Judging by his notes, I infer that Mitra employed seven MSS. in the constitution of his text. These MSS. were sold after his death. Mahamahopädhyāya Haraprasada S'astrī very kindly took much trouble in tracing them for me to their present owner; but my strenuous efforts to obtain access to them by purchase or loan have proved absolutely futile. 3. General account of the MSS. and their relationship. For the purposes of my edition nine MSS. have been directly available. Eight of these fall into two well-defined groups, representing a shorter and a longer recension of the Brhaddevata. The former group I have designated A, the latter B (suggestive of the Berlin and Bodleian codices which belong to it). The ninth MS. belongs, as will be seen, in a sense to both groups. To these nine may be added the seven which were used by Mitra. It is quite certain from his critical notes, in the light of the material accessible to me, that four of his MSS. belong to the A, and three to the B group. Thus the evidence of altogether sixteen MSS. has been available, directly or indirectly, in the constitution of my text. Of this total, eight belong to A. This group has two subdivisions: the first (A¹) comprising the Haug MS. (h), two of Max Müller's (m², m³), Bhau Dāji's (d), and Rajendralala's No. 3 (r³: called by him ga); the second (A³) com- prising Rajendralala's Nos. I, 4, 6 (r¹r¹r6 = ka gha ca). The latter three are practically identical, for their readings are almost invariably the same, and often agree in opposition to the readings of all the other MSS. of both families. The readings of A are on the whole inferior to those of A¹, being often clearly alterations of the original text; yet it is in some cases apparent that A2 has preserved the correct reading as compared with all the other MSS. of both recensions. It is on A2 that Mitra