The Tribes and Castes of the Central Provinces of India Volume I Part 1

The Tribes and Castes of the Central Provinces of India.

Volume I.

by R.V. Russell.

PREFACE

This book is the result of the arrangement made by the Government of India, on the suggestion of the late Sir Herbert Risley, for the preparation of an ethnological account dealing with the inhabitants of each of the princ.i.p.al Provinces of India. The work for the Central Provinces was entrusted to the author, and its preparation, undertaken in addition to ordinary official duties, has been spread over a number of years. The prescribed plan was that a separate account should be written of each of the princ.i.p.al tribes and castes, according to the method adopted in Sir Herbert Risley's _Tribes and Castes of Bengal_. This was considered to be desirable as the book is intended primarily as a work of reference for the officers of Government, who may desire to know something of the customs of the people among whom their work lies. It has the disadvantage of involving a large amount of repet.i.tion of the same or very similar statements about different castes, and the result is likely therefore to be somewhat distasteful to the ordinary reader. On the other hand, there is no doubt that this method of treatment, if conscientiously followed out, will produce more exhaustive results than a general account. Similar works for some other Provinces have already appeared, as Mr. W. Crooke's _Castes and Tribes of the North-Western Provinces and Oudh_, Mr. Edgar Thurston's _Castes and Tribes of Southern India_, and Mr. Ananta Krishna Iyer's volumes on Cochin, while a Glossary for the Punjab by Mr. H.A. Rose has been partly published. The articles on Religions and Sects were not in the original scheme of the work, but have been subsequently added as being necessary to render it a complete ethnological account of the population. In several instances the adherents of the religion or sect are found only in very small numbers in the Province, and the articles have been compiled from standard works.

In the preparation of the book much use has necessarily been made of the standard ethnological accounts of other parts of India, especially Colonel Tod's _Annals and Antiquities of Rajasthan_, Mr. J.D. Forbes'

_Rasmala or Annals of Gujarat_, Colonel Dalton's _Ethnology of Bengal_, Dr. Buchanan's _Eastern India_, Sir Denzil Ibbetson's _Punjab Census Report_ for 1881, Sir John Malcolm's _Memoir of Central India_, Sir Edward Gait's _Bengal and India Census Reports_ and article on Caste in Dr. Hastings' _Encyclopaedia of Religion and Ethics_, Colonel (Sir William) Sleeman's _Report on the Badhaks_ and _Ramaseeana or Vocabulary of the Thugs,_ Mr. Kennedy's _Criminal Cla.s.ses of the Bombay Presidency_, Major Gunthorpe's _Criminal Tribes of Bombay, Berar and the Central Provinces_, the books of Mr. Crooke and Sir H. Risley already mentioned, and the ma.s.s of valuable ethnological material contained in the _Bombay Gazetteer _ (Sir J. Campbell), especially the admirable volumes on _Hindus of Gujarat_ by Mr. Bhimbhai Kirparam, and _Parsis and Muhammadans of Gujarat_ by Khan Bahadur Fazlullah Lutfullah Faridi, and Mr. Kha.r.s.edji Nasarvanji Seervai, J.P., and Khan Bahadur Bamanji Behramji Patel. Other Indian ethnological works from which I have made quotations are Dr. Wilson's _Indian Caste_ (_Times_ Press and Messrs. Blackwood). Bishop Westcott's _Kabir and the Kabirpanth_ (Baptist Mission Press, Cawnpore), Mr. Rajendra Lal Mitra's _Indo-Aryans_ (Newman & Co., Calcutta), _The Jainas_ by Dr. J.G. Buhler and Mr. J. Burgess, Dr. J.N. Bhattacharya's _Hindu Castes and Sects_ (Thacker, Spink & Co., Calcutta), Professor Oman's _Mystics, Ascetics and Saints of India, Cults, Customs and Superst.i.tions of India_, and _Brahmans, Theists and Muslims of India_ (T. Fisher Unwin), Mr. V.A. Smith's _Early History of India_ (Clarendon Press), the Rev. T.P. Hughes' _Dictionary of Islam_ (W.H. Allen & Co., and Heffer & Sons, Cambridge), Mr. L.D. Barnett's _Antiquities of India_, M. Andre Chevrillon's _Romantic India_, Mr. V. Ball's _Jungle Life in India_, Mr. W. Crooke's _Popular Religion and Folklore of Northern India_, and _Things Indian_, Captain Forsyth's _Highlands of Central India_ (Messrs. Chapman & Hall), Messrs. Yule and Burnell's _Hobson-Jobson_ (Mr. Crooke's edition), Professor Hopkins' _Religions of India_, the Rev. E.M. Gordon's _Indian Folk-Tales_ (Elliot & Stock), Messrs. Sewell and Diks.h.i.+t's _Indian Calendar_, Mr. Brennand's _Hindu Astronomy_, and the late Rev. Father P. Dehon's monograph on the Oraons in the _Memoirs of the Asiatic Society of Bengal_.

Ethnological works on the people of the Central Provinces are not numerous; among those from which a.s.sistance has been obtained are Sir C. Grant's _Central Provinces Gazetteer_ of 1871, Rev. Stephen Hislop's _Notes on the Aboriginal Tribes of the Central Provinces_, Colonel Bloomfield's _Notes on the Baigas_, Sir Charles Elliott's _Hoshangabad Settlement Report_, Sir Reginald Craddock's _Nagpur Settlement Report_, Colonel Ward's _Mandla Settlement Report_, Colonel Lucie Smith's _Chanda Settlement Report_, Mr. G.W. Gayer's _Lectures on Criminal Tribes_, Mr. C.W. Montgomerie's _Chhindwara Settlement Report_, Mr. C.E. Low's _Balaghat District Gazetteer_, Mr. E.J. Kitts' _Berar Census Report_ of 1881, and the _Central Provinces Census Reports_ of Mr. T. Drysdale, Sir Benjamin Robertson and Mr. J.T. Marten.

The author is indebted to Sir J.G. Frazer for his kind permission to make quotations from _The Golden Bough_ and _Totemism and Exogamy_ (Macmillan), in which the best examples of almost all branches of primitive custom are to be found; to Dr. Edward Westermarck for similar permission in respect of _The History of Human Marriage_, and _The Origin and Development of the Moral Ideas_ (Macmillan); to Messrs. A. & C. Black in respect of the late Professor Robertson Smith's _Religion of the Semites_; to Messrs. Heinemann for those from M. Salomon Reinach's _Orpheus_; and to Messrs. Hachette et Cie and Messrs. Parker of Oxford for those from _La Cite Antique_ of M. Fustel de Coulanges. Much a.s.sistance has also been obtained from Sir E. B. Tylor's _Early History of Mankind_ and _Primitive Culture_, Lord Avebury's _The Origin of Civilisation_, Mr. E. Sidney Hartland's _Primitive Paternity_, and M. Salomon Reinach's _Cultes, Mythes et Religions_. The labours of these eminent authors have made it possible for the student to obtain a practical knowledge of the ethnology of the world by the perusal of a small number of books; and if any of the ideas put forward in these volumes should ultimately be so fortunate as to obtain acceptance, it is to the above books that I am princ.i.p.ally indebted for having been able to formulate them. Other works from which help has been obtained are M. Emile Senart's _Les Castes dans I'Inde_, Professor W. E. Hearn's _The Aryan Household_, and Dr. A.H. Keane's _The World's Peoples_. Sir George Grierson's great work, _The Linguistic Survey of India_, has now given an accurate cla.s.sification of the non-Aryan tribes according to their languages and has further thrown a considerable degree of light on the vexed question of their origin. I have received from Mr. W. Crooke of the Indian Civil Service (retired) much kind help and advice during the final stages of the preparation of this work. As will be seen from the articles, resort has constantly been made to his _Tribes and Castes_ for filling up gaps in the local information.

Rai Bahadur Hira Lal was my a.s.sistant for several years in the taking of the census of 1901 and the preparation of the Central Provinces District Gazetteers; he has always given the most loyal and unselfish aid, has personally collected a large part of the original information contained in the book, and spent much time in collating the results. The a.s.sociation of his name in the authors.h.i.+p is no more than his due, though except where this has been specifically mentioned, he is not responsible for the theories and deductions from the facts obtained. Mr. Pyare Lal Misra, barrister, Chhindwara, was my ethnographic clerk for some years, and he and Muns.h.i.+ Kanhya Lal, late of the Educational Department, and Mr. Aduram Chandhri, Tahsildar, gave much a.s.sistance in the inquiries on different castes. Among others who have helped in the work, Rai Bahadur Panda Baijnath, Diwan of the Patna and Bastar States, should be mentioned first, and Babu Kali Prasanna Mukerji, pleader, Saugor, Mr. Gopal Datta Jos.h.i.+, District Judge, Saugor, Mr. Jeorakhan Lal, Deputy-Inspector of Schools, and Mr. Gokul Prasad, Tahsildar, may be selected from the large number whose names are given in the footnotes to the articles. Among European officers whose a.s.sistance should be acknowledged are Messrs. C.E. Low, C.W. Montgomerie, A.B. Napier, A.E. Nelson, A.K. Smith, R.H. Crosthwaite and H.F. Hallifax, of the Civil Service; Lt.-Col. W.D. Sutherland, I.M.S., Surgeon-Major Mitch.e.l.l of Bastar, and Mr. D. Chisholm.

Some photographs have been kindly contributed by Mrs. Ashbrooke Crump, Mrs. Mangabai Kelkar, Mr. G.L. Corbett, C.S., Mr. R.L. Johnston, A.D.S.P., Mr. J.H. Searle, C.S., Mr. Strachey, Mr. H.E. Bartlett, Professor L. Scherman of Munich, and the Diwan of Raigarh State. Bishop Westcott kindly gave the photograph of Kabir, which appears in his own book.

Finally I have to express my grat.i.tude to the Chief Commissioner, Sir Benjamin Robertson, for the liberal allotment made by the Administration for the publication of the work; and to the publishers, Messrs. Macmillan & Co., and the printers, Messrs. R. & R. Clark, for their courtesy and a.s.sistance during its progress through the press.

September 1915.

PART I.

INTRODUCTORY ESSAY ON CASTE

1. The Central Provinces.

The territory controlled by the Chief Commissioner of the Central Provinces and Berar has an area of 131,000 square miles and a population of 16,000,000 persons. Situated in the centre of the Indian Peninsula, between lat.i.tudes 1747' and 2427' north, and longitudes 76 and 84 east, it occupies about 7.3 per cent of the total area of British India. It adjoins the Central India States and the United Provinces to the north, Bombay to the west, Hyderabad State and the Madras Presidency to the south, and the Province of Bihar and Orissa to the east. The Province was const.i.tuted as a separate administrative unit in 1861 from territories taken from the Peshwa in 1818 and the Maratha State of Nagpur, which had lapsed from failure of heirs in 1853. Berar, which for a considerable previous period had been held on a lease or a.s.signment from the Nizam of Hyderabad, was incorporated for administrative purposes with the Central Provinces in 1903. In 1905 the bulk of the District of Sambalpur, with five Feudatory States inhabited by an Uriya-speaking population, were transferred to Bengal and afterwards to the new Province of Bihar and Orissa, while five Feudatory States of Chota Nagpur were received from Bengal. The former territory had been for some years included in the scope of the Ethnographic Survey, and is shown coloured in the annexed map of linguistic and racial divisions.

The main portion of the Province may be divided, from north-west to south-east, into three tracts of upland, alternating with two of plain country. In the north-west the Districts of Sangor and Damoh lie on the Vindhyan or Malwa plateau, the southern face of which rises almost sheer from the valley of the Nerbudda. The general elevation of this plateau varies from 1500 to 2000 feet. The highest part is that immediately overhanging the Nerbudda, and the general slope is to the north, the rivers of this area being tributaries of the Jumna and Ganges. The surface of the country is undulating and broken by frequent low hills covered with a growth of poor and stunted forest. The second division consists of the long and narrow valley of the Nerbudda, walled in by the Vindhyan and Satpura hills to the north and south, and extending for a length of about 200 miles from Jubbulpore to Handia, with an average width of twenty miles. The valley is situated to the south of the river, and is formed of deep alluvial deposits of extreme richness, excellently suited to the growth of wheat. South of the valley the Satpura range or third division stretches across the Province, from Amarkantak in the east (the sacred source of the Nerbudda) to Asirgarh in the Nimar District in the west, where its two parallel ridges bound the narrow valley of the Tapti river. The greater part consists of an elevated plateau, in some parts merely a rugged ma.s.s of hills hurled together by volcanic action, in others a succession of bare stony ridges and narrow fertile valleys, in which the soil has been deposited by drainage. The general elevation of the plateau is 2000 feet, but several of the peaks rise to 3500, and a few to more than 4000 feet. The Satpuras form the most important watershed of the Province, and in addition to the Nerbudda and Tapti, the Wardha and Wainganga rivers rise in these hills. To the east a belt of hill country continues from the Satpuras to the wild and rugged highlands of the Chota Nagpur plateau, on which are situated the five States recently annexed to the Province. Extending along the southern and eastern faces of the Satpura range lies the fourth geographical division, to the west the plain of Berar and Nagpur, watered by the Purna, Wardha and Wainganga rivers, and further east the Chhattisgarh plain, which forms the upper basin of the Mahanadi. The Berar and Nagpur plain contains towards the west the shallow black soil in which autumn crops, like cotton and the large millet juari, which do not require excessive moisture, can be successfully cultivated. This area is the great cotton-growing tract of the Province, and at present the most wealthy. The valleys of the Wainganga and Mahanadi further east receive a heavier rainfall and are mainly cropped with rice. Many small irrigation tanks for rice have been built by the people themselves, and large tank and ca.n.a.l works are now being undertaken by Government to protect the tract from the uncertainty of the rainfall. South of the plain lies another expanse of hill and plateau comprised in the zarmindari estates of Chanda and the Chhattisgarh Division and the Bastar and Kanker Feudatory States. This vast area, covering about 24,000 square miles, the greater part of which consists of dense forests traversed by precipitous mountains and ravines, which formerly rendered it impervious to Hindu invasion or immigration, producing only on isolated stretches of culturable land the poorer raincrops, and spa.r.s.ely peopled by primitive Gonds and other forest tribes, was probably, until a comparatively short time ago, the wildest and least-known part of the whole Indian peninsula. It is now being rapidly opened up by railways and good roads.

2. Const.i.tution of the population.

Up to a few centuries ago the Central Provinces remained outside the sphere of Hindu and Muhammadan conquest. To the people of northern India it was known as Gondwana, an unexplored country of inaccessible mountains and impenetrable forests, inhabited by the savage tribes of Gonds from whom it took its name. Hindu kingdoms were, it is true, established over a large part of its territory in the first centuries of our era, but these were not accompanied by the settlement and opening out of the country, and were subsequently subverted by the Dravidian Gonds, who perhaps invaded the country in large numbers from the south between the ninth and twelfth centuries. Hindu immigration and colonisation from the surrounding provinces occurred at a later period, largely under the encouragement and auspices of Gond kings. The consequence is that the existing population is very diverse, and is made up of elements belonging to many parts of India. The people of the northern Districts came from Bundelkhand and the Gangetic plain, and here are found the princ.i.p.al castes of the United Provinces and the Punjab. The western end of the Nerbudda valley and Betul were colonised from Malwa and Central India. Berar and the Nagpur plain fell to the Marathas, and one of the most important Maratha States, the Bhonsla kingdom, had its capital at Nagpur. Cultivators from western India came and settled on the land, and the existing population are of the same castes as the Maratha country or Bombay. But prior to the Maratha conquest Berar and the Nimar District of the Central Provinces had been included in the Mughal empire, and traces of Mughal rule remain in a substantial Muhammadan element in the population. To the south the Chanda District runs down to the G.o.davari river, and the southern tracts of Chanda and Bastar State are largely occupied by Telugu immigrants from Madras. To the east of the Nagpur plain the large landlocked area of Chhattisgarh in the upper basin of the Mahanadi was colonised at an early period by Hindus from the east of the United Provinces and Oudh, probably coming through Jubbulpore. A dynasty of the Haihaivansi Rajput clan established itself at Ratanpur, and owing to the inaccessible nature of the country, protected as it is on all sides by a natural rampart of hill and forest, was able to pursue a tranquil existence untroubled by the wars and political vicissitudes of northern India. The population of Chhattisgarh thus const.i.tutes to some extent a distinct social organism, which retained until quite recently many remnants of primitive custom. The middle basin of the Mahanadi to the east of Chhattisgarh, comprising the Sambalpur District and adjoining States, was peopled by Uriyas from Orissa, and though this area has now been restored to its parent province, notices of its princ.i.p.al castes have been included in these volumes. Finally, the population contains a large element of the primitive or non-Aryan tribes, rich in variety, who have retired before the pressure of Hindu cultivators to its extensive hills and forests. The people of the Central Provinces may therefore not unjustly be considered as a microcosm of a great part of India, and conclusions drawn from a consideration of their caste rules and status may claim with considerable probability of success to be applicable to those of the Hindus generally. For the same reason the standard ethnological works of other Provinces necessarily rank as the best authorities on the castes of the Central Provinces, and this fact may explain and excuse the copious resort which has been made to them in these volumes.

3. The word 'Caste.'

The word 'Caste,' Dr. Wilson states, [1] is not of Indian origin, but is derived from the Portuguese _casta_, signifying race, mould or quality. The Indian word for caste is _jat_ or _jati_, which has the original meaning of birth or production of a child, and hence denotes good birth or lineage, respectability and rank. _Jatha_ means well-born. Thus _jat_ now signifies a caste, as every Hindu is born into a caste, and his caste determines his social position through life.

4. The meaning of the term 'Caste.'

The two main ideas denoted by a caste are a community or persons following a common occupation, and a community whose members marry only among themselves. A third distinctive feature is that the members of a caste do not as a rule eat with outsiders with the exception of other Hindu castes of a much higher social position than their own. None of these will, however, serve as a definition of a caste. In a number of castes the majority of members have abandoned their traditional occupation and taken to others. Less than a fifth of the Brahmans of the Central Provinces are performing any priestly or religious functions, and the remaining four-fifths are landholders or engaged in Government service as magistrates, clerks of public offices, constables and orderlies, or in railway service in different grades, or in the professions as barristers and pleaders, doctors, engineers and so on. The Rajputs and Marathas were originally soldiers, but only an infinitely small proportion belong to the Indian Army, and the remainder are ruling chiefs, landholders, cultivators, labourers or in the various grades of Government service and the police. Of the Telis or oil-pressers only 9 per cent are engaged in their traditional occupation, and the remainder are landholders, cultivators and shopkeepers. Of the Ahirs or graziers only 20 per cent tend and breed cattle. Only 12 per cent of the Chamars are supported by the tanning industry, and so on. The Bahnas or cotton-cleaners have entirely lost their occupation, as cotton is now cleaned in factories; they are cartmen or cultivators, but retain their caste name and organisation. Since the introduction of machine-made cloth has reduced the profits of hand-loom weaving, large numbers of the weaving castes have been reduced to manual labour as a means of subsistence. The abandonment of the traditional occupation has become a most marked feature of Hindu society as a result of the equal opportunity and freedom in the choice of occupations afforded by the British Government, coupled with the rapid progress of industry and the spread of education. So far it has had no very markedly disintegrating effect on the caste system, and the status of a caste is still mainly fixed by its traditional occupation; but signs are not wanting of a coming change. Again, several castes have the same traditional occupation; about forty of the castes of the Central Provinces are cla.s.sified as agriculturists, eleven as weavers, seven as fishermen, and so on. Distinctions of occupation therefore are not a sufficient basis for a cla.s.sification of castes. Nor can a caste be simply defined as a body of persons who marry only among themselves, or, as it is termed, an endogamous group; for almost every important caste is divided into a number of subcastes which do not marry and frequently do not eat with each other. But it is a distinctive and peculiar feature of caste as a social inst.i.tution that it splits up the people into a mult.i.tude of these divisions and bars their intermarriage; and the real unit of the system and the basis of the fabric of Indian society is this endogamous group or subcaste.

5. The subcaste.

The subcastes, however, connote no real difference of status or occupation. They are little known except within the caste itself, and they consist of groups within the caste which marry among themselves, and attend the communal feasts held on the occasions of marriages, funerals and meetings of the caste _panchayat_ or committee for the judgment of offences against the caste rules and their expiation by a penalty feast; to these feasts all male adults of the community, within a certain area, are invited. In the Central Provinces the 250 groups which have been cla.s.sified as castes contain perhaps 2000 subcastes. Except in some cases other Hindus do not know a man's subcaste, though they always know his caste; among the ignorant lower castes men may often be found who do not know whether their caste contains any subcastes or whether they themselves belong to one. That is, they will eat and marry with all the members of their caste within a circle of villages, but know nothing about the caste outside those villages, or even whether it exists elsewhere. One subdivision of a caste may look down upon another on the ground of some difference of occupation, of origin, or of abstaining from or partaking of some article of food, but these distinctions are usually confined to their internal relations and seldom recognised by outsiders. For social purposes the caste consisting of a number of these endogamous groups generally occupies the same position, determined roughly according to the respectability of its traditional occupation or extraction.

6. Confusion of nomenclature.

No adequate definition of caste can thus be obtained from community of occupation or intermarriage; nor would it be accurate to say that every one must know his own caste and that all the different names returned at the census may be taken as distinct. In the Central Provinces about 900 caste-names were returned at the census of 1901, and these were reduced in cla.s.sification to about 250 proper castes.

In some cases synonyms are commonly used. The caste of _pan_ or betel-vine growers and sellers is known indifferently as Barai, Pansari or Tamboli. The great caste of Ahirs or herdsmen has several synonyms--as Gaoli in the Northern Districts, Rawat or Gahra in Chhattisgarh, Gaur among the Uriyas, and Golkar among Telugus. Lohars are also called Khati and Kammari; Masons are called Larhia, Raj and Beldar. The more distinctly occupational castes usually have different names in different parts of the country, as Dhobi, Warthi, Baretha, Chakla and Parit for washermen; Basor, Burud, Kandra and Dhulia for bamboo-workers, and so on. Such names may show that the subdivisions to which they are applied have immigrated from different parts of India, but the distinction is generally not now maintained, and many persons will return one or other of them indifferently. No object is gained, therefore, by distinguis.h.i.+ng them in cla.s.sification, as they correspond to no differences of status or occupation, and at most denote groups which do not intermarry, and which may therefore more properly be considered as subcastes.

t.i.tles or names of offices are also not infrequently given as caste names. Members of the lowest or impure castes employed in the office of Kotwar or village watchmen prefer to call themselves by this name, as they thus obtain a certain rise in status, or at least they think so. In some localities the Kotwars or village watchmen have begun to marry among themselves and try to form a separate caste. Chamars (tanners) or Mahars (weavers) employed as grooms will call themselves Sais and consider themselves superior to the rest of their caste. The Thethwar Rawats or Ahirs will not clean household cooking-vessels, and therefore look down on the rest of the caste and prefer to call themselves by this designation, as 'Theth' means 'exact' or 'pure,'

and Thethwar is one who has not degenerated from the ancestral calling. Salewars are a subcaste of Koshtis (weavers), who work only in silk and hence consider themselves as superior to the other Koshtis and a separate caste. The Rathor subcaste of Telis in Mandla have abandoned the hereditary occupation of oil-pressing and become landed proprietors. They now wish to drop their own caste and to be known only as Rathor, the name of one of the leading Rajput clans, in the hope that in time it will be forgotten that they ever were Telis, and they will be admitted into the community of Rajputs. It occurred to them that the census would be a good opportunity of advancing a step towards the desired end, and accordingly they telegraphed to the Commissioner of Jubbulpore before the enumeration, and pet.i.tioned the Chief Commissioner after it had been taken, to the effect that they might be recorded and cla.s.sified only as Rathor and not as Teli; this method of obtaining recognition of their claims being, as remarked by Sir Bampfylde Fuller, a great deal cheaper than being weighed against gold. On the other hand, a common occupation may sometimes amalgamate castes originally distinct into one. The sweeper's calling is well-defined and under the generific term of Mehtar are included members of two or three distinct castes, as Dom, Bhangi and Chuhra; the word Mehtar means a prince or headman, and it is believed that its application to the sweeper by the other servants is ironical. It has now, however, been generally adopted as a caste name. Similarly, Darzi, a tailor, was held by Sir D. Ibbetson to be simply the name of a profession and not that of a caste; but it is certainly a true caste in the Central Provinces, though probably of comparatively late origin. A change of occupation may transfer a whole body of persons from one caste to another. A large section of the Banjara caste of carriers, who have taken to cultivation, have become included in the Kunbi caste in Berar and are known as Wanjari Kunbi. Another subcaste of the Kunbis called Manwa is derived from the Mana tribe. Telis or oilmen, who have taken to vending liquor, now form a subcaste of the Kalar caste called Teli-Kalar; those who have become shopkeepers are called Teli-Bania and may in time become an inferior section of the Bania caste. Other similar subcastes are the Ahir-Sunars or herdsmen-goldsmiths, the Kayasth-Darzis or tailors, the Kori-Chamars or weaver-tanners, the Gondi Lohars and Barhais, being Gonds who have become carpenters and blacksmiths and been admitted to these castes; the Mahar Mhalis or barbers, and so on.

7. Tests of what a caste is.

It would appear, then, that no precise definition of a caste can well be formulated to meet all difficulties. In cla.s.sification, each doubtful case must be taken by itself, and it must be determined, on the information available, whether any body of persons, consisting of one or more endogamous groups, and distinguished by one or more separate names, can be recognised as holding, either on account of its traditional occupation or descent, such a distinctive position in the social system, that it should be cla.s.sified as a caste. But not even the condition of endogamy can be accepted as of universal application; for Vidurs, who are considered to be descended from Brahman fathers and women of other castes, will, though marrying among themselves, still receive the offspring of such mixed alliances into the community; in the case of Gosains and Bairagis, who, from being religious orders, have become castes, admission is obtained by initiation as well as by birth, and the same is the case with several other orders; some of the lower castes will freely admit outsiders; and in parts of Chhattisgarh social ties are of the laxest description, and the intermarriage of Gonds, Chamars and other low castes are by no means infrequent. But notwithstanding these instances, the principle of the restriction of marriage to members of the caste is so nearly universal as to be capable of being adopted as a definition.

8. The four traditional castes.

The well-known traditional theory of caste is that the Aryans were divided from the beginning of time into four castes: Brahmans or priests, Kshatriyas or warriors, Vaishyas or merchants and cultivators, and Sudras or menials and labourers, all of whom had a divine origin, being born from the body of Brahma--the Brahmans from his mouth, the Kshatriyas from his arms, the Vaishyas from his thighs, and the Sudras from his feet. Intermarriage between the four castes was not at first entirely prohibited, and a man of any of the three higher ones, provided that for his first wife he took a woman of his own caste, could subsequently marry others of the divisions beneath his own. In this manner the other castes originated. Thus the Kaivarttas or Kewats were the offspring of a Kshatriya father and Vaishya mother, and so on. Mixed marriages in the opposite direction, of a woman of a higher caste with a man of a lower one, were reprobated as strongly as possible, and the offspring of these were relegated to the lowest position in society; thus the Chandals, or descendants of a Sudra father and Brahman mother, were of all men the most base. It has been recognised that this genealogy, though in substance the formation of a number of new castes through mixed descent may have been correct, is, as regards the details, an attempt made by a priestly law-giver to account, on the lines of orthodox tradition, for a state of society which had ceased to correspond to them.

9. Occupational theory of caste.