which are entirely foreign to the original sūtras, ' and confirms decisively the conclusion that the text was intended for Rājas.' May it not also be held to indicate that writers belonging to Brahman caranas were beginning to recognise the necessity of teaching to human beings artha as well as dharma, the way of building up and maintaining a State, as well as the way of performing sacrifices and maintaining order amongst the classes?

I observe that Vishnu, too, contains a tolerably lengthy chapter on artha, namely No. III.; and Apastamba briefly describes the duty of a king in II. 10, 25-26. The 'Gentoo Code,' as we have seen, has a long chapter on artha prefixed to the body of the work; founded apparently upon a work of Pacheshputtee Misr on the duties of a king.

The curious passage with which Apastamba is brought to a conclusion, shows clearly that a knowledge of the Vedas, however extensive, will not suffice for all purposes, and therefore must be supplemented with knowledge to be derived from other sources.

II. 11, 29, 11, tells us: The knowledge which Çūdras and women possess is the completion (of all study).' By this (according to Bühler) is meant 'dancing, music, and other branches of the Artha çãstra.' But, surely, dancing, music, &c., are of the sixty-four sciences of Kāma.

Apastamba goes on to say, 'It is difficult to learn the sacred law from the Vedas, but by following the indications it is easily accomplished,' and then gives the indications,' as thus: 'He shall regulate his

course of action according to the conduct which is unanimously recognised in all countries by men of the three twice-born castes, who have been properly obedient, who are aged, of subdued senses, neither given to avarice, nor hypocrites. Acting thus, he will gain both worlds.' This appears to be a quotation of a proverbial saying. The actual ending of the work is this: Some declare that the remaining duties must be learnt from women and men of all castes.' Have we here (as I suppose) a recognition of the necessity of learning artha and kāma from any who teach them, e.g. from the Hetæra?

I have shown at p. 31, above, where Çudras and women are to get information as to their duty, namely, from the epics and similar compositions. But, whatever else they may find here, certainly they will not find law. An interesting passage in the Mahābhārata (Vana Parva, 312) shows that the author of it, like the author of Apastamba, Nārada, and others, greatly distrusted the Holy Law, and preferred that usage which Manu declares to be 'highest dharma.' In it Yudhistira, after solving with preternatural sagacity a string of enigmas propounded by Yama, tells him. what is the path,' as thus: Argument leads to no certain conclusion: the çrutis are different from one another; there is not even one Rishi whose opinion can be accepted as infallible: the truth about religion and duty is hid in caves; therefore, that alone is the path along which the great have trod.'

If not even one Rishi exists whose opinion can be accepted as infallible, may I not be pardoned for

declining to believe in guesses at the meaning of Vijñāneçvara's speculations upon the meaning of the Yajnavalkya' recollection' of a Rishi's teachings?

This passage may usefully be compared with the above-mentioned passage from Apastamba, as also with the passages remarked on above, at pp. 32, 47, and 82.

CHAPTER VIII.

THE JOINT FAMILY.

PERHAPS the commonest phrase in the reports of the Madras High Court, of cases involving questions of Hindu law, is the joint family' or 'family.' Thus we find it stated in Norton's Leading Cases, at p. 173, I., that Joint undivided family is the ordinary status of the Hindoo. Sometimes this has been termed joint-tenancy, sometimes coparcenary, sometimes co parcenary with a benefit of survivorship.' And at II. 461, of the same: The ordinary status of a Hindu family is that of coparcenary; insomuch so, that this is always presumed until the contrary is shown.' In order, therefore, to understand the principles upon which the Madras High Court administers its law to Indian litigants, in affairs of inheritance, succession, and the like, it is essentially necessary to comprehend the views that the Madras High Court from time to time takes of the composition and nature of the Indian 'family.' But to do this is by no means an easy task. Not only are the views of the Court constantly changing; even the views of individual members of it appear to undergo frequent modification and amendment, and it is not too much to say that at Madras the whole subject of the structure of Indian society is

wrapt in as much of uncertainty and obscurity at the present moment as it was in the days of the elder Strange and Ellis. This statement will be justified by the third part of this work.

If we would attempt scientifically to reconstruct the Hindu law for Madras, or rather to construct for the first time a code of Indian usages for Madras, the first pre-requisite of success would be a thorough examination of the families of various forms at present existing in Madras, including, e.g., the old-fashioned Brahman family of secluded villages, the polyandrous family of the Western Coast, the ordinary agricultural family of the interior, and the modern trading family of the coast. In this chapter I purpose indicating, quite roughly and briefly, the character of the examination which I would suggest in this behalf.

In the first place it is proper to observe that the ambiguous word 'family' is (or may be) extremely misleading. It may be taken (according to its context) to mean the whole collection of slaves or servants in one house; or, all the individuals forming one household, under one head; the descendants of a common ancestor; a race of men, and many other things. It may even mean one small baby: or the whole population of this world. The word (legally) is not a term of art; indeed it is not known to the English law, though 'familia' was used in technical senses in Latin legal writings. And, as I have pointed out in my Prospectus (p. 187, n.), the concept appears to be foreign to the Sanskrit language. At all events, I have never succeeded in learning a San