stated but it is clear that accused's brain is liable to derangement. Though this, by itself, is no factor to determine the guilt or innocence of the accused on the date of the occurrence, it is, in my opinion, a circumstance which has always been regarded as bearing on the question of sentence."

"Bentham, in his Rationale of Punishment, page 20, says: Punishment is elevated to the first rank of benefits when it is regarded not as an act of wrath or vengeance against a guilty or unfortunate individual who has given way to mischievous. inclinations, but as an indispensable sacrifice to the common safety." To inflict the sentence of death on the accused, who is a person of weak intellect, who at least once recently showed symptoms of insanity and who committed the murder under an unfounded but deep-rooted belief that, by doing so, he would save the lives of his wife and daughter, would, in my opinion, be an act of wrath or vengeance and would not serve "general prevention any more than if the lesser punishments were

inflicted on him."

66

On these grounds I think the proper sentence to be inflicted on the accused is that of transportation for life."

The following is Mr. Justice P. R. Das's judgment :

"For the reasons which have been given by Sultan Ahmad J., I am of opinion that the proper sentence that ought to be passed on the appellant is one of transportation for life, and I order accordingly."

From the foregoing judgments, we are able to glean the undermentioned facts bearing upon the belief in witchcraft which is prevalent among the aboriginal tribes living on the Chotā Nagpore plateau :

(1) That witches, can, by means of their spells, cause sickness and spread epidemics of cholera.

(2) That the recovery of the sick patient can be brought about by killing the offending witch.

(3) That witches can transform themselves into dogs and sheep. (This belief in the power of the witches of transforming themselves into various beasts and birds is widespread. The Khons of Orissa and the Oraons of Chota Nagpore believe that the witches can assume the shapes of tigers and panthers." This belief is also prevalent even in enlightened England, Scotland and Ireland. In the county of Yorkshire in England, it is believed that the witches can change themselves into cats. In the counties of Devonshire and Yorkshire in England, and in Wales and Scotland, the belief prevails that witches can metamorphose themselves into hares. In the county of Cumberland in England, it is believed that they can assume the shape of red deer. Then again, in Scotland they are credited with the possession of power whereby they can transform themselves into ravens

and in Ireland they are believed to have the power of metamorphosing themselves into cattle.) 1

THE RESULTS OF THE ETHNOGRAPHICAL SURVEY OF BOMBAY.

BY R. E. ENTHOVEN, ESQ., C. I. E., I. C. S.

(Read on 23rd November 1920.)

In May 1901 the Government of India issued orders for the commencement of the Ethnographical SurHistory of the Pombay Ethnogra- vey of eight provinces, and proposed that phical Survey. the enquiries into the origin, social configuration, customs and occupations of the numerous castes, and tribes should be spread over a peroid of four or five years. These enquiries were to follow closely the lines of certain questions approved by Messrs. Nesfield, Ibbetson, and Risley at a Conference held in 1885. They were to be carried out by the Superintendents in addition to their ordinary official duties.

1 Vide my article "Indian Folk-Beliefs about the Tiger; Part 1" in The Journal of the Anthropological Society of Bombay, Vol. VIII, pages

114-115.

It was not possible to allow more than a few thousand rupees annually to each province for the work of the Survey; and before the work was half done, i. e., in 1909 even this small financial provision was withdrawn. The Survey has since been practically dependent on the voluntary labours of the Superintendent in charge, with such assistance as he was able to secure without the expenditure of funds. This voluntary assistance has been forthcoming from three sources. Certain scholars such as Mr. D. R. Bhandarkar and Mr. B. A. Gupte have placed their knowledge at the disposal of the Superintendent and assisted the work of the Survey in the capacity of Honorary Assistants. A number of gentlemen have furnished valuable materials in the capacity of Honorary Correspondents, among whom I would like to mention the learned Honorary Secretary of the Anthropological Society, Shams-ul-ulema Dr. J. J. Modi, as well as Rao Bahadur R. C. Artal, Rao Bahadur S. T. Bhandare, Rao Bahadur P. B. Joshi and Mr. J. A. Saldanha. Finally, certain local Committees organised by the Superintendent during the course of Census operations in 1901 for the purpose of investigating and reporting on caste questions have supplied materials of considerable value and interest. With this assistance it has been possible to complete the Survey of the tribes and castes in the Presidency, excluding Sind, by the beginning of the present year. The work has involved the preparation of nearly 300 articles. These have been published in draft form, and circulated for criticism before being finally embodied in the Survey record.

It may at once be admitted that the work could not possibly have been completed even in this greatly extended pericd, had it not been for the very full materials available in the pages of the Bombay Gazetteers compiled by the late Sir James Campbell. Much of the work of the Survey has indeed consisted of the re-arrangement of these materials, which were drawn up originally on a district basis instead of dealing with tribes and castes as a whole. When compared with more recent information the original district accounts have been found to be remarkably accurate in detail. With the assistance described, and

in such leisure as was available to the Superintendent, in spite of the pre-occupation of more important official duties, a considèrable amount of new information has been collected regarding the tribes and castes of the Bombay Presidency. The Survey records thus contain much new matter, though the main source of information has been found in the pages of the Bombay District Gazetteers, which contain accounts remarkable both for their fulness and accuracy.

At the time of issuing the first of the draft monographs, I indicated that the tribes and castes recorded in the Census Tables of 1901, and numbering

Scope of work.

over 500 would be dealt with in three classes :

Class 1, being those containing one hundred thousand members, were to be described as fully as possible;

Class 2, being those between one hundred thousand and five thousand, were to be dealt with in less detail;

Class 3, being those under five thousand were for the most part to be described merely by re-arranging

the materials on the lines of the Ethnographical questions.

This scheme has been adhered to. The only important departure from the general scheme outlined by me in 1903 at the time of publishing the first Monograph, dealing with the Ahir, is the omission of the tribes and castes of Sind. The Survey volumes at present include only tribes and castes found in the Presidency proper.

In the course of revising existing materials for the caste accounts a number of new caste divisions have been discovered, both in connection with the groups inside which marriage is essential, and the smaller divisions inside which marriage

Special features of the Survey.

is forbidden. These are referred to in this work respectively as endogamous and exogamous groups. They are of special interest as the interior structure of a caste is frequently valuable evidence of its origin. It has also been found that certain caste names are synonymous. Castes have in such cases been re-grouped. In connection with castes of converts from Hinduism to Islam useful information has been obtained from the marriage registers of the Kázis, showing the cases in which inter-marriage between such castes is allowed. Many Musalman castes adhere to the Hindu practice of endogamy. Among those Musalman castes which inter-marry are found armourers, butchers, farriers, elephant-drivers, and similar occupational groups, which are probably descendants of the camp followers of the Mughal armies. It is interesting to note that contact with the Muhammadan armies has left its influence in the greater freedom with which inter-marriage is allowed. The social prejudice which prevents members of many Muhammadan castes from marrying outside the caste is a relic of the previously existing Hindu custom, such castes being clearly traceable to a Hindu origin before conversion. They are found chiefly in areas where Musalman rule has stimulated conversion, either by example, as in Bijapur, or by force as in North Kanara, once under the dominion of Hyder and Tipu. A great deal of new information has been obtained regarding the interesting Lingayat community, which is found to be based partly on religious and partly on caste distinctions. The late Mr. J. F. Fleet, C. I. E., supplied some valuable notes for this article, which forms the basis of a new account of Lingayats contributed by the Superintendent to Dr. Hasting's Dictionary of Religion. The special interest of the community lies in the fact that it exhibits the process by which a religious movement, starting with the abolition of caste distinctions, develops slowly into a community of which the most recent converts adhere strictly to the noninter-marrying caste distinctions of their Hindu ancestors. The Survey has brought to light a great deal of new information regarding totemistic divisions in the Deccan and Southern