Introduction] CRITICAL PRINCIPLES FOLLOWED [xix A has no omissions) contain 289 ślokas, or one śloka less than 5x58 (= 290). These two Adhyāyas contain no vargas with fewer than four ślokas, and only four as short as that. The omissions in the text of A do not begin to appear till the middle of the third Adhyāya (iii. 62). We find that four ślokas are here omitted (85ed-89ab) at the end of the seventeenth and the beginning of the eighteenth varga, and two more in the nineteenth (95d-97); yet the end of the nineteenth varga is indicated by the figure 19 after the same śloka¹ in A as in B. In other words, though the three vargas in A contain only nine ślokas, or an average of three for each, the old number 19 remains at the end of the third varga in A as well as B. The same thing occurs over and over again in the A MSS. The conclusion that A is an abridgement of the original text is cor- roborated by another feature in the numbering of the A MSS. It is only the A group that, in addition to numbering the vargas, also numbers the ślokas continuously from the beginning of the book to the end. This seems almost certainly due to the fact that, by the omission of certain passages, the number of the varga (if occurring within such passages) also disappeared, and that the figures indicating the end of the vargas thus became inter- mittent in the shorter recension. Consequently a copyist would, I think, naturally come to adopt the additional method of numbering the slokas of the whole work consecutively. It is obvious that this method of enumera- tion affords many opportunities of error. As a matter of fact h shows frequent mistakes in this part of its numbering. Hence the figures of the final śloka, 1049, in h, as well as 1281 in m², do not correspond to the actual total of ślokas contained in those two codices. 7. Critical principles followed in constituting the Text. In view of the existence of these two recensions, what method, it may be asked, have I adopted in constituting the text? Having arrived at the conclusion that the longer recension represents on the whole the original extent of the Brhaddevata, I have admitted the additional matter of B to the text, with the exception of a few occasional ślokas or half-ślokas, which are clearly spurious, and which have therefore been relegated to the critical notes. Every one of the additional lines of B has, however, been indicated, both in text and translation, by a B fixed to it. Similarly, the comparatively few ślokas which are peculiar to pre- A are indicated by that letter. Hence the reader can see at a glance what lines are peculiar to either recension. I may here add that the text of the 1 That is, in hd m³ m³ as well as m¹. Mitra does not indicate the vargas in his edition; but it is to be assumed that his A MSS. agres with h in this respect as in others.
