44 SOCIAL SCIENTIST following four marriage rites are lawful for the Brahmins: 1) Brahma 2) Prajapatya, 3) Arsa, and 4) Daiva, the Asura and Raksasa rites for the K§atriyas, and the Gandharva and Paisaca rites respectively for the Vaisyas and Sudras.5 The main division as far as the marriage rites are concerned, is between the Brahmins to whom belong the former four rites, and the rest of the castes to whom belong the latter four rites. The Nibandhakaras make a significant observation to the effect that a bride who is married according to any one of the former four rites, adopts the gotra of her husband, while a bride who is married according to any one of the latter four rites, remains in the gotra other father: The Smrti passages that condemn cross-cousin marriages are explained in a peculiar manner by the Sm. C. and the Par. M. When a woman is married in one of the four forms, Brahma & c. she passes into the gotra of her husband, becomes a sapinda to the husband's family and so she is severed from her father's family (as to gotra and sapmda relationship); but when a woman is married in the asura, gandharva and other forms, she does not pass over into the gotra of her husband, but remains in the gotra of the father and her sapinda relationship with her father and mother continues. Therefore the son of such a woman if he marries the daughter of his mother's brother, would be marrying a girl who is a sagotra and sapinda of his mother. The Sm. C. and the Par. M. and other works say that the smrti texts forbidding marriages with maternal uncle's daughter refer to a person whose mother was married in the gandharva, asura and the other two forms, but not to a person whose mother was married in the brahma and the three other approved forms.. .. c This explanation of the prohibition or permission of cross-cousin marriage inadvertently admits that the Brahmins who followed the gotra system were patriarchal, while the rest of the castes were matrilineal. The option allowed to the Ksatriyas and Vaisyas not only proves that gotra system was totally foreign to them, but clearly suggests that they followed a matrilineal clan system; for it is ridiculous to suppose that the lower three castes were without any clan system. This is also proved by Panini's rule IV. I. 147, which is explained by V S Agrawala as follows: .... According to Panini, one's designation after the gotra name of one's mother (gotra-stri) implied censure (IV. I. 147), because it was supposed that the mother's name would be adopted only in the event of the father's name being unknown (kasika, Pitur-asaimvijnane matra vyapadeSo, patyasya kutsa). . . . 7 But the explanation that to be known by the gotra of one's mother was derogatory is at variance with the metronymic by which Panini himself was known: . . . . Patanjali quotes a karika describing Panini as Daksiputra (Dak^iputrasya Panineh, 1-75) after the name of his mother who was of the Dak§a gotra. Dakseya also would be Panini's metronymic.8 V S Agrawala thinks that the popular opinion tnrned in favour of