________________
162
of the archetype and which represent later additions. Only those which occur in the Third South-Indian recension, on the one hand, and in the Jaina-recension and/or the Vulgata, on the other, can positively be traced back to the archetype. One way of editing the text is by first giving only the Gathās which are clearly authentic, with the remaining Gathās following in appendices arranged according to the position in the stemma of the MS or MSS in which the Gathās occur. The basic text would then probably contain not more than approximately 500 Gathās. The alternative, which I have opted for here, is to take one complete text which is close to that of the archetype as the basis, in this case a combination of the Jaina-recension and the Vulgata. This text will include Gathās of which the authenticity is doubtful. Some of these may possibly go back to the archetype. In the part edited here altogether four Gathās are missing in the Third South-Indian recension. It is almost certain that one of these, viz. *4, originally formed part of the text of the Third South-Indian recension and consequently of that of the archetype (see the note given there). It is equally certain that Gathā *40 is a later addition to the text in the North-Indian branch. This follows from the occurrence of a relatively modern word, bh amāgana, which is otherwise found only in Apabhraía. This argument will turn up again in connection with 77 (absent in Ma, Ti, Tp and T (and W)) which contains the Apabhransa verb nada- 'to nag', and with 92 (absent in Ma, Ti and Tp (and W); but found in T) which contains the word phaladiā (see Bha phalagayā), a formation, like bhamāçana, typical of Apabhraía. The status of the two remaining Gathās, *18 (absent in all the South Indian recensions) and * 37 (absent in Ma, Ti and Tp (and W); but found in T), remains unclear.
As already noted earlier, for the text of each individual Gathā often numerous variants are available. The elimination of variants, involved in reconstructing the text of the archetype, is generally settled with reference to the stemma. Thus, the text almost automatically contains that which Ma, Ti and Tp have in common with Bh and R and/or with the Vulgata MSS. Not infrequently two different, but stemmatically equivalent, variants are available. In such cases the basic argument in the elimination is that of the lectio difficilior.
Occasionally, however, matters may be more complicated. Certain pro
