It will be seen that the average nasal index of the people investigated ranges from 69. 1 in the tall, light-skinned, and long, narrow-nosed Lambadis, who speak an Aryan language, to 95.1 in the short, dark-skinned, and short, broadnosed Paniyans; and that the indices recorded range between a minimum of 59.2 in a Lambadi and a maximum of 108.6 in a Paniyan. Mr Thurston has, however, measured a Paniyan woman who possessed a nose 31 mm. in height and 37 mm. in breadth, the nasal index being 119.4. The Sheik Muhammadans of Madras claim to be descendants of emigrants from the north, and to be distinct from the converted Dravidians. Their claim is no doubt justified; but wellmarked signs of admixture of Dravidian blood are conspicuous in some members of their communities, whose dark skin and high nasal index betray their non-Aryan descent. This miscegenation is clearly brought out by Thurston in the figure given on page 89, in which we have a series of triangles representing in two-thirds natural size the minima, average, and maxima nasal indices of individuals belonging to the poorer classes of Brahmans of Madras City, of Tamil Pariahs, and of Paniyans. There is obviously far less connection between the Brahman minimum and the Paniyan maximum than between the Brahman and Pariah maxima and the Paniyan average. The frequent occurrence of high nasal indices, resulting from short, broad noses, in Brahmans has

already been accounted for in the quotations I have made from Sir A. Lyall and Mr. H. A. Stuart.

One is accustomed to regard the problem of Indian ethnography as of only moderate complexity, as is seen in the following abstract of a paper by Mr. Risley.'

There are three main types in the population of India at the present day :

1. A leptorhine, pro-opic,' dolichocephalic type, of tall stature, light build, long and narrow face, comparatively fair complexion, and high facial angle.

This type is most marked in the Panjab.

Their exogamous groups are eponymous, names of Vedic saints or heroes.

2. A platyrhine, mesopic,' or nearly platyopic, dolichocephalic type, of low stature, thickset, very dark complexion, relatively broad face, usually low facial angle. This type is most distinct in Chota Nagpore and the Central Provinces.

Its sections are totemistic, like those of North American Indians; that is, they are names of animals, plants, or artificial objects, to all of which some form of taboo applies.

3. A mesorhine, platyopic,' brachycephalic type of a low or medium stature, sturdy build, yellowish complexion, broad face, and low facial angle.

This type is found along the northern and eastern frontiers of Bengal.

Their exogamous groups are very curious, being mostly nicknames of the supposed founder of the sept, such as "the fat man who broke the stool," and others less fit for publication.

1 H. H. Risley, "The Study of Ethnology in India," Journ. Anth. Inst., xx., 1891, p. 235.

The terms pro-opic, mesopic, and platyopic have reference to the height or prominence of the bridge of the nose; for further details see the chapter on

measurements.

Diagrams of the Variations in the Height and Breadth of the Noses of the Poorer Classes of Brahmans of Madras City, of Tamil Pariahs, and of Paniyans, two-thirds Natural Size; after Thurston.

1. Leptorhine, Pro-opic Dolichocephals.

If it be accepted that Karl Penka has proved the typical Aryan to be dolichocephalic, there would seem to be some grounds for believing that in the dolichocephalic leptorhine type of the Panjab and north-western frontier at the present day, we may recognise the descendants of the invading Aryans of three thousand years ago, changed, no doubt, in hair, eyes, and complexion, but retaining the more enduring characteristics of their race in the shape of their head, their stature, and the finely-cut proportions of their nose. Survivals of fair, or rather reddish hair, grey eyes, and reddishblond complexion are, moreover, still to be found, as Penka has pointed out, and as Risley himself has seen, among the Kaffirs from beyond the Panjab frontier.

Anyway, the striking preponderance of dolichocephaly in the Panjab and the North-Western Provinces, and its gradual increase as we travel up the Ganges Valley towards the traditional Aryan tract, tend both to strengthen Penka's hypothesis and to enhance the credibility of early Indian legends. These facts go also to show that Penka is mistaken in supposing that the Indian branch of the Aryans became brachycephalic on their way to India. Had this been so, the dolichocephaly which now distinguishes them could only have been derived from crosses with the black race, and the Aryans could hardly have become dolichocephalic in this way without also becoming platyrhine.

2. Platyrhine, Mesopic Dolichocephals.

The measurements show the current distinction between the Dravidians and Kolarians, on which stress has been laid by Dalton and others, to be a purely linguistic character, not corresponding to any appreciable difference of physical type.

The hypothesis of the north-eastern origin of the so-called Kolarians urged by Colonel Dalton, and recently advanced by Mr. J. F. Hewitt, must also be abandoned as inconsistent with the dolichocephalic skull of the typical representatives of the group.

Whatever the Kolhs may be, they certainly are not a Mongoloid race.

3. Mesorhine, Platyopic Brachycephals.

All of the groups which come within this category are demonstrably of more or less pronounced Mongolian descent; and we may conveniently call them Mongoloid.

The type is essentially a frontier type, and its influence can in no case be traced far into the interior of India.

The Kochh or Rajbansi, a large tribe of Bengal, who now pose as an outlying branch of the Rajputs, are, indeed, commonly supposed to have some strain of Mongolian blood among them, but Risley doubts if this opinion is well founded. A slight degree of platyopy is, it is true, met with among them, but this may equally well be accounted for on the supposition of their affinity to the platyrhine type.

The nasal index of the dolichocephalic tribes that are of non-Aryan descent requires a further analysis, and it appears to me that Mr. Thurston's researches suggest that the problem is more complex than is generally admitted.

On looking at the table on page 86, we are struck with the fact that three tribes, the Badagas, Todas, and Kotas, have the least variation of any in the range of their nasal indices. They have lived an isolated existence on the plateau of the Nilgiri Hills until the settlement of the English in recent times, and we may with safety regard them as a fairly primitive non-Aryan people. The owners of the greatest variation (exceeding a range of 30 units) constitute a group of Tamil classes made up of Brahmans, Pattar