^138:1 I have adopted the reading vikkhayayanti, because it is the most difficult, and therefore explains most easily the various corruptions, or it may be emendations, that have crept into the text. Sankara explains vikkhadayanti by vidravayanti, and this shows that he too must have read vikkhayayanti, for he could not have explained vikkhadayanti, which means they uncover or they deprive of their clothing, by vidravayanti, they drive away. It is true that vikkhayayanti may be explained in two ways; it may be the causative of kha, to cut, but this meaning is not very appropriate here, and quite inadmissible in another passage where vikkhayayanti occurs, whereas, if derived from vikh (oixomai) in a causative sense, Sankara could hardly have chosen a better explanation than vidravayanti, they make run away. The root vikh, vikkhayayati is recognised in Panini III, 1, 28, and in the Dhatupatha 28, 129, but it has hitherto been met with in this passage only, and in Brihadaranyaka, Up. IV, 3, 20. Here also the author speaks of a man who imagines that people kill him or do him violence, or that an elephant chases him or that he falls into a pit. Here we have hastiva vikkhayayati, and Sankara, at least as printed by Dr. Roer, explains this by vikkhapayati, vikkhadayati, vidravayati; dhavatity arthah. Much better is Dvivedaganga's commentary, as published by Dr. Weber, Satap. Brahm. p. 1145, Kadakid enam hasti vikkhayayativa vidravayativa; vikha gatau, gupudhupavikhipanipanibhya aya iti (Pan. III, 1, 28) svartha ayapratyayah. In the Dictionary of Boehtlingk and Roth the derivation from kha, to cut, is preferred; see Nachtrage, s. v. kha.