Of course I do not desire to be understood to deny the existence or currency of such Draviḍa terms of art as Pangu (share) and 'Piri' (divide). But, from my experience of the use of these and connected expressions, I gather that they denote and connote the joint holding and subsequent partition not of the lands of a Household, but of the lands of village. Thus, at the present day, many of the villages in the Chingleput district of Madras are divided (as regards the arable lands) into a number of equal pangus (shares or allotments), which once may have been held and enjoyed by as many proprietors and their families; whilst now one proprietor owns two or more pangus, another perhaps ten or more, and a

1 A troublesome composite word, of constant occurrence in Tamil deeds, is Ullittar. I have never been able to satisfy my mind as to what it really denotes and connotes. The first part of it means 'within,' and the second 'those who placed' (or 'are placed'). Wilson's Glossary says the word sometimes means the direct descendants of a common ancestor; and one is naturally tempted to think it may indicate a body of agnates living together in the hand of a Father of a Family. But it would be rash to do so. At present, I should prefer to connect the word with the land and the village community. Ul-kudi' seems to be one holding land' within the village.' 'Ul-manei' is an abode 'within the village.' Confer Ullavan,'' Ulpatti,' &c. Possibly the phrase may mean all connected with a man by claims to a particular share in a village, actually held by him as dominus, or something like this. Ordinarily, according to Wilson, it means partners in a business; coparceners: sometimes it is used for heirs generally.

Another unsatisfactory word, used habitually (I believe) by Dravidas. everywhere, is 'Vārasudār,' which comes, according to Wilson, from the Arabic Wāris, and is equivalent to one who has a claim to a share in an ancestral estate. Strange that men supposed to govern themselves unconsciously by the Sanskrit rules of the Mitakṣarā, should have recourse to Arabic for a word equivalent to co-sharers or joint successors. Can it be that we are all mistaken-that the Dravidas never heard of the Mitakşara and its theoretic developments, and, having no convenient general term of their own, borrowed Varasudār from their Muhammadan conquerors?

third only a half or a small fraction of a single pangu. These allotments, with various appurtenant rights, as of pasturage over the common land, have been freely alienable and partible under British domination; and so it has come about that on the one hand many of them have been bought by prospering families and added on to their existing holdings, on the other hand many of them have been split up by partition.

In Wilson's Glossary the following terms of art connected with the sharing of a village may be found, namely:

Pangu A share in a coparcenary village.

=

Pangali One who holds such share; a co

parcener.

=

Pangupirinthavargal Those who have divided such share amongst them.

Pangumālei = A list or roll of such shares, showing the amount of land cultivated by each member of the community, the changes of property, the original divisions, the quality of the lands, and whether cultivated by the proprietors or by migratory culti

vators.

Panguvikrayam = Sale of such share.

Panguvali = A village held in common by a certain number of coparceners, amongst whom the lands are distributed at various times, according to the votes of the majority of the sharers, and are held in severalty for a given time under such distribution.

The more general word pangu is represented in

some parts of South India by the word karei. Thus, I have shown in my Madura Country that from a report of the Collector, of January 10, 1815, it appeared that, at that time, the privileged landholders of the greater part of the Madura and Dindigul districts, who paid their land-tax in the form of a share of the crop, were known as holders of' kareis' or shares of villages. When they did not themselves cultivate the land, they received ten per centum of

1 In the Papers on Mirasi Right (Madras, 1862), will be found a considerable amount of information about the Karei system. According to Ellis, the term pasung-karei' used to denote that particular joint tenure of the cultivated lands, which was anciently universal throughout the Tamil country, and still prevails in many villages in every part of it, but especially in that known to the natives by the name Tondei Mandalam. Under this system, the meerassy right to any particular spot of cultivated land in the village is not vested in any individual.' But there is a periodic redistribution of lands among the shareholders. The other most prevalent system was the arudi-karei, under which each holder enjoyed a right over his own particular fields.

It is to be regretted that Ellis was prevented from doing for Madras what Mr. Seebohm has done for England in his admirable English Village Community. Many of the distinctive marks of the open or common field system once prevalent in England' will be at once recognised by the observer as existent in South India. For example, we have here the open fields divided up into little narrow strips; the Kāni or Tamil acre, measured off with a pole of varying length, but not differing greatly from the English pole of 16 feet; the turf balks; the scattered and intermixed holdings; the periodic redistribution of holdings, superseded generally by fixed holdings; co-operation in ploughing; the right to graze cattle over the whole of the arable land, when not under crop; the common lands; the system of boundaries; the services; the different classes, corresponding roughly to the landlord, the tenants in villenage, the cottiers, and the prædial slaves; the township situated in the midst of the fields; the rights to cut fuel, take fish, &c. It would be highly interesting to learn by inquiry that the Dravidas, who, according to Manu, are degraded Kṣatriyas, had worked out for themselves a thousand years ago a system of agricultural life very similar to the system once prevalent in England. In any case scientific inquiry into the nature of the Dravida system could not fail greatly to facilitate the study of Indian usage.

the crop raised by outsiders who did cultivate it. Their right was not lost by neglect to cultivate for one year. If a karei-holder wished to part with his karei-a thing almost unknown-he must offer it first to his relations, next to the other karei-holders, lastly to strangers. And his right, if sold, probably would be worth on the average twenty years' purchase. From information elsewhere obtained I was enabled to add that in a karei village the kareis, or allotments of arable land, were theoretically equal in extent and value; but in order to avoid all cause for dissatisfaction, they were originally made only for a term of years, at the end of which a new allotment took place, and the proprietors all exchanged holdings with one another. The allotment did not extend to the pasturage, which remained always common.

Looking to what we know of the history of the Madura and Dindigul districts, it is impossible to doubt that many (if not all, or most) of these karei villages must have been established by clans that came down from the North one after another, in consequence of the pressure of over-population, war, or other disturbing cause; most of them, probably, under the guidance of a Poligar or other military chief. And if each karei was originally allotted to a single family, we have here a certain resemblance to the state of things described by Doctor Hearn in the Aryan Household, and it becomes possible that the ordinary family of these villages of the present day may in many essentials resemble the Aryan 'Joint Undivided Family.'

But, however closely any existing agricultural family in Madura or Dindigul may be found to resemble this particular form of family, it must not be forgotten that the farmers of Madura and Dindigul are not Aryan by descent, but Dravida. So that their progenitors have not borne any part in, or been in any way connected with, the particular states of society contemplated and provided for by the authors of Manu, Narada, and other smṛtis. And, any development they may have effected in their internal social organism cannot (so far as appears) have been affected in any, the slightest degree by Sanskrit writings. The real character of their Family is quite unknown, and remains to be ascertained by observation.

To the east of Madura, and on the Ramnad coast, occurs a family of a very different character, that of the Maravans, who formerly were the soldiers and dependents of the Sethupati, or Chief who guarded the Isthmus of Rameçvara. From the Marava-jati-varna of Taylor it appears (see my Madura Country) that this tribe is still divided into seven clans, of which the Sembu-nāṭṭu is the principal: and its usages are peculiar, and specially noteworthy. Properly speaking, every Maravan should be a warrior, and hold lands on a strictly military tenure, on condition of his being ready at a moment's notice to follow his lord, wherever led, equipped for battle. Not so very long ago an ordinary foot-soldier, carrying a sword and a spear, was granted for his support a piece of land capable of yielding him, per annum,