an English town, or such and such localities in a county. Books like those of Mr. Charles Booth on "Poverty" will show by maps what neighbourhoods in certain large towns are occupied by what economic stratum of the people. But this is far from being always possible. In London there are poor neighbourhoods and good neighbourhoods, but they often conjoin astonishingly even within the same street. It may even be possible in the United States of America, where there has never been a feudal system, to divide the people into ranks; but in England (1) late changes have done away with many of the social barriers, chiefly educational, which formerly divided some Englishmen from others; (2) racially most English folk are such a blend that blood castes are hardly possible. It has occasionally been maintained that the titled classes and their kindred are mainly Norman, and the rest of the country Saxon and Celtic, but some careful genealogical study would usually soon dispose of this theory.

So that anthropologists ought to be exceedingly grateful to Manu for so rigorously ordering the separation of race from race. The result is that India is an ethnological museum such as is probably nowhere else to be found, and in most parts even one small village will show different castes, i.e., usually different races in little, existing side by side, perhaps each in its street or couple of streets, and offering remarkable opportunities of study of the circumstances which plant certain races in certain parts, of the conditions under which one race will tolerate another as neighbour, of the kind of habitations which will denote their occurrence in the village, of the geographical limits of one caste and the beginnings of another.

There is no part of the Presidency which is so homogeneous but that careful note will shed a light on the conditions of racial occurrence. In Sind it will repay careful enquiry to observe what places are preferred by the Hindu minority. These are mostly the larger towns and villages; but isolated merchants will occasionally face a lonely life entirely surrounded by Maho.

medans. The Amils, or clerical branch of the Sind Banias, are almost entirely confined in their origin to five or six headquarters Shikarpur, Hyderabad, Tatta, and a few other places which were once court capitals. The Sind Mahomedans, besides their numerous minor denominations, are broadly divided into Baluchis and Sindhis. How far do these live together in one village, and how far alone, and what kind of habitat do they severally prefer? A preponderance of the Baluchis will be found in Northern Sind, specially west of the Indus, through which they seem to have migrated south, and the more original Sindhis will, I think, be found to prefer the neighbourhood of the great river, while the Baluchis are better inured to the higher lands. But that is merely a rough delimitation. Coming to Northern Gujarat, the chief feature of village life is the almost universal cohabitation in one village of Kanbi and Koli, the former as an oligarchic owner of the best land, the latter as his field-labourer. The occurrence of a pure Koli village is not uncommon, but a community absolutely and unmixedly Pátidár if found would be an interesting anthropological fact. The Bania will of course be found numerous wherever the position of a village, say at a railway junction or a crossing of trade routes, facilitates commerce; but the occurrence of less numerous castes, like the Brahmans, of various sub-castes, the Dheds, or the Bhangis is a matter of some interest. We expect the Brahmans to muster strong round important religious centres, and so they do; but their occurrence in some villages and not in others is not a matter of chance, and the underlying laws would be interesting to discover. The Dhed as a village servant is very common in Gujarat, and his congener the Mhár or Mádar holds a similar place further south. Why are some villagers favoured by his presence and not others? Again, that invaluable treasure, the Bhangi, or scavenger, though more numerous in Gujarat than elsewhere in the Presidency, may or may not be found in any given Gujarat village. A map of his localization would be of use not only to be anthropologist, but to the sanitary profession.

To come further south, in Broach we find the Kanbi replaced by a rural Borah who is apparently a converted Kanbi, and the Koli largely by the Bhil-(further south again by other semiaboriginals, Dublas, Chhodras, and others), -but the economic relationship is much the same. Further south still, in Surat, the Hindu Kanbi partly regains his place, but does not flourish, and we find large areas of almost pure wild tribe. One peculiar type of Brahman, the Anavla, makes his appearance, and is mostly confined to certain definite villages and tracts. In Thana (following the coast, in Khandesh further inland) begins the Maratha, who claims Rajput blood, otherwise we might think him merely a Marathi-speaking Kanbi (if indeed Kanbi and Rajput are more than occupationally different). Official ethnology sees in the Maratha a Scythian strain which is not admitted in the Gujarat Kanbi. He does mostly all his own field labour too, his Mhár assistants seldom reaching the dimensions of the subordinate Koli population of Gujarat.

The Kolis of the south are either mountaineers or fishermen,1 and their peristent name opens up vide vistas of enquiry as to the identity of all the folk who call themselves Koli, from the borders of Sind (and beyond, if the Kori weaver is a Koli) down to Ratnagiri. What is the origin of the marked Agri, or salt-working race, round Bombay? Are they an evolution of the Koli, or of the Kanbi, or of the aboriginal, or of none of these? Either their occupation or their blood has made of them a distinct type.

The south also raises the Chit pávan puzzle. The origin of this race is still unsolved, but it is not difficult to detect their habitations, where they live in any number, by their healthy love of little gardens round their houses, their neat clean houses themselves, and their nearer approach to sound sanitary arrangements than most other castes show.

1 Isolated ferrymen above Ghauts.

In this connection it is interesting to learn to recognize by the look of the dwelling the caste by which a village or street is nhabited. The Broach village Bhil (unlike the Mahi-Kantha Bhil, who objects to a neighbour, and lives in a cottage) unex. pectedly favours a long well-built line of continuous slopingroofed houses. (I might add that the odour of the locality wil also betray the resident.) The Parsee likes an upper story when he can get it, and much whitewash. In the Carnatic the Lingayat house is said to be open right through its depth, while the Mahomedan dwelling with a similar exterior is carefully partlywalled across. By a certain amount of experience I fancy myself now able to detect the exact degree of untidiness which marks the Talvár (or Bedar) and the Kurubar (or Dhangar) street of houses in a Dhárwár village. A few days ago I was at the Hubli Criminal Tribes Settlement, a perfect storehouse of odds and ends of wild life, and was interested to find that I had correctly diagnosed the difference of caste in tribes who had side by side erected their huts of mere sticks and coverlets, and of sticks and clay (assigned them for the purpose) respectively, as if they had been birds making their nests, by unfailing instinct, out of certain hereditary materials. (Compare, in higher life, the persistence of the Dutch house at the Cape, or of "Gothic" or of Portuguese church architecture in India.)

My point, then, is that the local distribution of castes or races in India has only tentatively been worked out or mapped, and that any enquirer who would follow it out in detail either in a town or a district, or even in a few villagea, would produce some interesting results. Bombay City may present rather too big a field, containing as it does a more variegated patchwork of races than any area in India, but even there it will be remarkable how distinct castes, not excluding the British, have as a rule each their own quarter. Substituting trades for castes this phenomenon is, of course, far from unknown in Europe, and was formerly there even far more the rule than now, as mere London street names will indicate (Bread Street, Milk

Street, Ironmonger Lane, Booksellers' Row, and so forth). Even Karachi is almost too haphazard a jumble of races to be profitably caste-mapped; but I should like to take a plaa of, say, Ahmedabad, and colour it first Hindu and Mahomedan, and then caste by caste. The Jains, the Boras, the Vaishnavs, the Brahmans of many kinds, the weavers of several kinds, the Kumbhars, the Marathas, the Kolis, the Patidars of two kinds, and countless other races, would be found to fill each in So with Surat, or Sholapur, or its own place in the mosaic.

Poona, or any large up-country town, and so, in rather less infinite variety, would almost the smallest village yield interesting results.

And the same, in broader feature, would be the advantage of The race-maps prevalence-maps for extended rural areas. which, e.g., the Imperial Gazetteer provides, only loosely touch the subject. The census reports give the statistics, but can hardly be expected to show all the "oecology" in the graphic way which would fully illustrate it.

An example of the result which these enquiries might give is, I believe, that it would be found that certain castes do not readily live together in one village, and that other castes are very much complementary to each other (like some Beloch tribes and their "Hamsayas").

Even the British, from far Europe, taken in the lump, follow certain laws for the locality of their incidence in India. They flock, of course, to the cool places-Malabar Hill, Bandra, Colaba in Bombay,-(in Byculla, which was once a suburb, they have been hemmed in, and they will soon, perhaps, leave it, as they have left so many old centres, as the indigenous population surrounded them)-the hill-stations up-country, including a favourite climate like Bangalore, the great ports and industrial centres like Bombay and Calcutta, including the planting districts, the great railway colonies, and, of course, the cantonments. Where Mahomedans, earlier invaders, are in a minority, it is