His followers of the Yellow school called themselves Gelugpa (or Gelukpa), ‘adherents of virtue’ (or, Gal-danpa from their monastery); their principal characteristic being that they adhered to the purer discipline. The chief point of interest in connexion with Tsong Khapa is the bearing of his reformation on the development of the Avatāra theory already mentioned (see pp. 190, 265). It is said that Tsong Khapa himself, like Gautama Buddha, had two chief pupils, and that he appointed these two to succeed him with equal authority as Heads of the orthodox sect. He is also credited with having been the first to promulgate the doctrine that no election of successors to his two pupils would at any time be needed, as each of them on dying would be constantly re-born in a supernatural manner. There is, however, no historical foundation for such a statement. Indeed, according to the opinion of some, the two Grand Lāmas were merely the lineal successors of the two eminent Lāmas, Atīsha and his pupil Brom Ton (see p. 273). After all, it seems most likely that the whole Avatāra theory was an invention of some shrewd Head Lāma, who, perceiving that the strict enforcement of celibacy would prevent any hereditary succession, like that possible in monasteries of the Red school, and foreseeing that it would be necessary to prevent the suicidal divisions to which the intrigues of an election to the Headship of monasteries—especially of Grand Lāma monasteries—would be likely to give rise, bethought himself of a compromise between hereditary succession and election. After more than one trial, the system was found to work so well that it was eventually adopted with little modification by all Northern Buddhists, and even by those of the Red sect. The date of its invention is as uncertain as the name of the inventor. All that can be said is, that it cannot be traced back to an earlier period than the fifteenth century. And here we must again guard against the confusion of thought likely to arise from the usual practice of translating Avatāra by ‘incarnation.’ We have seen that the doctrine of transmigration (gati) through various embodiments, as applicable to all beings, is a fundamental dogma both of Brāhmanism and of Buddhism, though in Buddhism transmigration properly means a mere continuous transformation and reconstruction of the elements (Skandhas) of being (p. 109). The idea is very dimly, if at all, adumbrated in the Mantra portion of the Veda. It is more clearly traceable in one of the Brāhmaṇas, and distinctly enunciated in the Ćhāndogya Upanishad (V. x. 7) thus:—‘He whose conduct has been good, quickly attains to some good embodiment as a Brāhman, Kshatriya, or Vaiṡya. He whose conduct has been bad, assumes an inferior embodiment, as a dog, a hog, or a Ćaṇḍāla.’ In Manu the theory is fully developed.