Despite the adjustments made to the 1931 census data on tribal population, no comparable attempt was made to stan- dardize for this atlas the 1961 data for "scheduled tribes." The reason for this inconsistency lies in the legal significance at- tached to the designation "scheduled tribes," according to the Indian constitution of 1950. Since seats in both the central and state parliaments are reserved for the tribal peoples in strict accordance with their numbers, and since a number of other social and political benefits are similarly extended with due regard to the size of the tribal population, it is important that the precise numbers of tribals be mapped as given by the cen- sus itself. Whether or not one agrees with a particular census classification for a given group, one must recognize the impor- tant sociolegal reality the census figures delineate. This reality is presented in the patterns of maps (e) and (f) of plate X.A.7.
Comparing maps (a) and (e), providing data on tribals at the district level (maps (b) and (f) are much more difficult to compare because of the pronounced differences in the extent of the major administrative divisions thereon), one notes a broad similarity of patterns within the area of post-indepen- dence India they both cover. The most heavily tribal areas, in- cluding many districts in which tribals constitute an absolute majority of the population, were and still are in the far north- eastern parts of the country, in the Chota Nagpur and other portions of northeastern peninsular India, in a tribal belt stretching from west to east across central India, in the south- ern Aravallis, and in the northern reaches of the Western Ghats. These areas generally constitute regions of rugged ter- rain, large parts of which remain under forest (cf. plates I.B.1 and I.C.2). They may be considered areas of refuge, relatively unattractive to the characteristically plains-dwelling, plough- cultivating peasantry in the mainstream of Indian culture. The future maintenance of a mainly or largely tribal population in such areas presumably is contingent on the degree to which population pressures elsewhere force Indian farmers to push into new, relatively inaccessible localities and, conversely, on the degree to which the Indian and state governments protect the established tribal populations against such incursions.
Where the 1961 map does show a relative decline in the tribal population from the percentages indicated in 1931, any of a number of possible explanations are possible: (a) that the difference may stem in part from differences between the censuses in classification (most notably in the cases of Uttar Pradesh, Jammu and Kashmir, and several union territories where no lists of scheduled tribes were promulgated, thus mak- ing the scheduled tribal population "nil" by administrative fiat); (b) that significant detribalization has occurred and groups once considered tribal have shifted out of the tribal category; (c) that, owing to rates of natural increase lower for the tribal population than for the population at large (a func- tion of higher rates of mortality), that group suffered a rela- tive decline within the larger society; (d) that immigration of nontribal settlers into largely tribal regions, coupled perhaps with emigration of tribals to areas of plantations, mines, cities, and other economic attractions outside the largely tribal areas, also led to a relative decline of tribals in the latter; and, finally, (e) some combination of two or more of the foregoing.
That there has, on the whole, been a relative decline is indi- cated by the fact that whereas within the 1961 area of India (deducting figures for the areas of Burma and Pakistan, which were included in the 1931 maps) tribals totaled roughly 7.9% of the 1931 population, by 1961 their proportion had fallen to 6.8%. In absolute terms, of course, the numbers of tribals rose from about 22 million in 1931 (21.8 million according to the census versus 22.1 million as adjusted for this work) to 29.9 million. But to specify those areas in which a significant real decline, as opposed to an arbitrary statistical decline, has taken place is a task we have not ventured to embark upon, given the complexity of the problem.
It remains only to note those relatively few instances in which the tribal proportion of the total population apparently increased from 1931 to 1961. Where the increase was dra- matic, as in certain Himalayan districts of Punjab and in Himachal Pradesh, one may reasonably assume an arbitrary extension of the list of scheduled tribes to ethnic groups not classified as tribal in 1931. Otherwise, the problem of explain- ing specific increases is likely to be comparable in difficulty to that of explaining specific areas of decline.
Turning from the designation of "tribal" by ethnic affiliation to that of "tribal" by religion, as recognized in both cases by the censuses of 1931 and 1961, we see in the case of religion that the pattern of change appears far more striking than what we have just described. A comparison of maps (c) and (d) of plate X.A.7, relating to 1931, with maps (g) and (h), for 1961, is here in order, One notes that, whereas in 1931 adher- ents to what were then considered to be tribal religions were encountered over a rather large part of India and constituted
Even allowing for a significant degree of conversion to Christianity among certain tribes and possibly some additional conversion to Buddhism, a real shift away from tribal forms of religion of the magnitude indicated is to be doubted. Rather, what the census data seem to suggest is a reclassification as Hindus of large numbers of tribals who would have been clas- sified in 1931 as belonging to religions that were more "tribal" than "Hindu." This is not to deny that there is in fact a grad- ual acculturative process of Hinduization at work among the greater part of India's tribal population and that real religious shifts toward Hinduism took place in the period from 1931 to 1961. But in the absence of a clear dividing line between what is a tribal religion and what is one or another form of folk Hinduism, there is little point in trying to specify the precise magnitude of this shift. It is reasonable, however, to suppose that tribals whose religion was reported as something other than Hindu in 1961 made a clear specification to that effect.
The analysis of the maps relating to "depressed castes" in 1931 and "scheduled castes" in 1961 on plate X.A.8 is beset with the same sorts of difficulties as were noted above for the maps showing the distribution of tribal peoples. Here too the comprehensiveness of the terms varies between census periods and, in both 1931 and 1961, from one administrative division to another; and here too handling the 1931 data posed some- what greater problems than handling the data of 1961, which we accepted as given, because "scheduled castes," like "sched- uled tribes" in the latter period, were a sociolegal reality the precise distribution of which had important practical con- sequences.
For the 1931 maps, (a) and (b), the general sources are the census statistics on depressed castes provided in the keys to the social maps of the "Report" volumes of the several In- dian provinces, major princely states, and states agencies. But several minor adjustments in or additions to those data have been made: (1) In order to avoid double counting, certain groups (Agarias, Bayars, Khangars, Kols, etc.) shown as de- pressed castes in the United Provinces, but taken by us to be tribal in accordance with their designation in other areas, have been deducted from the depressed caste totals. (2) In Burma, for which depressed caste data are not provided, the percent- age of the total population in depressed castes was arbitrarily assumed to be one-fifth that of the Hindu population. (3) In the Andaman Islands, for which no census total was given for depressed castes, our figures were derived by totaling the data for individual groups for which statistics were available.
It should be noted that a group considered depressed in one province/state or even one part of a province/state need not be so considered in another. (Hence, for example, the wide variation in the percentages shown for depressed castes in the socially and culturally similar states of Travancore and Co- chin.) In this respect the maps of depressed castes in 1931 are less standardized than those relating to tribal population, for which given groups were considered as tribal irrespective of where they were located (e.g., Santals on the plains of Bihar or Kols in the United Provinces). Further, we must note that, whereas in 1931 depressed castes referred to a portion of the Hindu population only, the usage for scheduled caste in 1961 extended also to Sikhs. This accounts for the relatively low 1931 percentages in the Punjab. It also explains why the per- centages on map (a) of plate X.A.8 may be lower than those for the most numerous single depressed caste, as shown on plate X.C.3, map (d), for which Hindus and Sikhs with a common caste name have been counted together.
Because of the marked changes in the political map of India between 1931 and 1961, the shifts in population attendant on the 1947 partition, and the inclusion or noninclusion of certain Sikh groups in the depressed or scheduled caste populations, meaningful comparisons of maps (b) and (e) at the level of major administrative divisions cannot easily be made. Com- parisons at the district level, as portrayed on maps (a) and (c), are somewhat easier, but there too the value of the exer- cise will often be vitiated by the varying inclusiveness of the categories of depressed and scheduled castes. To get a truly meaningful idea of how scheduled caste groups are faring demographically relative to the population as a whole, one would have to compare figures for various dates at the level of the individual caste. Although this is possible, given the fact that the Indian census continues to report data on individual scheduled castes (and on scheduled tribes, but not other caste groups), we have not undertaken such an inquiry. For what- ever it may be worth, however, we note that, excluding the area of Burma, the total depressed/scheduled caste population
To account for the decline just noted, several explanations may be advanced. First, because of the economically inferior position of depressed castes in general, their rates of mortality may remain significantly higher than those of the population as a whole, and their rates of natural increase may, assuming no marked differences in birthrate, be commensurately lower. Second, as specific groups, either throughout India or in par- ticular regions, advance their corporate social status by adopt- ing more "Sanskritized" behavior (i.e., behavior emulating that of the ritually higher castes), attaining higher levels of educa- tion, wealth, or political power, and thereby overcoming sub- jection to various forms of discrimination, they will be recog- nized by the government as having moved out of the depressed category and no longer be "scheduled." Third, many persons in depressed castes have converted to religions other than Hin- duism in an effort to free themselves from the stigma and dis- abilities their low-caste status imparts. In former times Islam and Sikhism were common vehicles for such attempts at up- ward social mobility, and both religions still so operate in vary- ing degrees (despite the paradoxical retention of official sched- uled caste recognition for various castes among Sikhs, a conflict between theory and practice that is also applicable to other religious groups in the social setting of South Asia). Recently, however, Christianity and Buddhism have had relatively greater appeal, the former having its greatest impact in the Telugu- speaking areas now comprising the state of Andhra Pradesh and in parts of Punjab, and the latter having profoundly re- duced the scheduled caste population of Maharashtra and ad- joining districts of Madhya Pradesh, where well over 2.5 mil- lion conversions took place in the period 1956–61. Map (d), which shows the proportion of the population in the area of Buddhist conversions who were in scheduled castes in 1951, will help the reader appreciate the regional impact of that movement. Some rough calculations lead us to suggest that conversions of low-caste individuals to Buddhism and Chris- tianity, with allowances for natural increase among converting families, between the years 1931 and 1961 have reduced the number of scheduled castes in the latter year by close to 4 million persons. Were these persons to be added to the afore- mentioned 1961 scheduled caste total of 70 million, the abso- lute growth in that segment of the population would have been between 45% and 50%, rather than the aforementioned figure of 40%. Even the larger figure, however, is well below the general rise of 68% for the total Indo-Pakistani population, which suggests the true importance of factors other than con- version in the trend toward lower proportions of depressed castes within society as a whole.
For maps (f) and (g) we have totaled the percentages of the 1961 population enumerated in scheduled castes and tribes to obtain a composite view of the fraction of the total pop- ulation who are officially recognized as being disadvantaged groups within the larger society. At the level of the state or union territory this fraction may constitute more than nine- tenths of the total, as in Nagaland or Tripura, which are both overwhelmingly tribal. For the whole of India and of Pakistan the percentages are 21.5% and 5.8% respectively. Since the areas in which the tribal population is particularly prominent tend to complement those more extensive areas in which sched- uled castes are relatively prominent, there are relatively few areas of India in which either group or the two groups com- bined do not comprise at least 15% of the total population. For the most part such areas are those in which minority reli- gions are particularly prominent. We have commented on most such areas earlier in this text and in that for plates X.A.3–6. The greater part of Gujarat, however, is exceptional. That area, appears to be one in which movements for upward mo- bility among depressed caste groups have been singularly successful.
India, Census (1931), (1951), (1961); Pakistan, Census (1961); J. M. Mahar, ed. (1972), though not utilized for our compilations, is a relevant and highly useful work.
X.A.9. Holy Places of South Asia
For a part of the world in which there is scarcely a village without its local shrine or sacred precincts and in which the places of pilgrimage are legion, any attempt to map the more important holy places will of necessity be highly subjective and open to criticism. Nevertheless, we have here sought to plot the locations of those holy places deemed most sacred to each of the major religions found in South Asia. And to attempt to differentiate two different levels of importance among certain