XI.A.1 and 2. Population Density, 1872, 1901, 1931, and 1961

The series of eight maps making up atlas plates XI.A.1 and 2 portrays the changing density of population within South Asia at two different levels of analysis. For the areas of India and Pakistan, the larger maps show data by minor adminis- trative divisions (districts and, before independence, princely states as well); the smaller maps relate to larger territorial units (provinces, major princely states, and state agencies to 1931; states, union territories, provinces, and frontier agencies in 1961). For Ceylon, Nepal, and Afghanistan, data, where available, are aggregated at differing levels as noted on the maps themselves. A uniform legend applies to all the maps in the series to facilitate comparison between time periods or, for a given date, between different levels of aggregation. The appended tables indicate densities, by country, for each cen- sus year from 1872 to 1961, interpolating where necessary for Ceylon to make the data comparable in date to those for India; for the pre-independence period data are also aggre- gated within India for territories under direct British rule and for the princely states.

The principal primary sources of the data presented are the censuses of India for each of the four reference years; the Pakistani and Nepali censuses of 1961; and Ceylonese cen- suses of 1871, 1901, 1931, 1953, and 1963. Although the 1872 Census of India did not cover the princely states, estimates for the populations of most such states were put forward for that year or another close to it in a statistical abstract for the pe- riod 1865–66 to 1874–75.

For Ceylon the 1961 data we have presented are interpola- tions between the data of the 1953 and 1963 censuses. Because of the unavailability of certain of the Ceylonese censuses re- ferred to above we have occasionally had to use secondary references in which the desired data are reported; data for 1901 and 1931 were derived from the 1911 and 1946 censuses, in which statistics for the previous census periods were re- peated, and 1871 data were derived from the Ceylon Blue Book for the year 1873. This will explain certain apparent in- consistencies in the citation of sources below. In Nepal a cen- sus of sorts was conducted in 1931; but its reliability is judged to be so low that we have not tried to use it. At late as 1961, Afghanistan had not yet had a national census; for that coun- try the density figures mapped are based on government esti- mates of total populations by provinces for the year 1966.

The range of population density encountered within South Asia is exceedingly great. Most of Baluchistan and a signifi- cant part of Afghanistan, for example, have fewer than fifteen persons per square mile, well below the threshold figure of fifty that marks the upper limit of the lowest class interval of our maps. This is less than one-hundreth the densities encoun- tered in 1961 in more than a dozen districts in East Pakistan, West Bengal, Kerala, and elsewhere. At the extremes the range runs from about one person per square mile in Chagai district in Baluchistan in 1931 to 73,642 per square mile in the en- tirely urban district of Calcutta in 1961.

To comprehend the underlying causes for the regional var- iations, one should study the maps of plates XI.A.1 and 2 in conjunction with those relating to such physical factors as ter- rain (I.B.1) and climate (I.C.1), and those depicting man's modification of those factors through the development of ir- rigation works, transportation facilities, and industry, all of which augment the productive capacity of the areas where they are situated (cf. maps in subsections XI.B–E). At all periods one perceives a marked relative concentration of South Asia's population on the alluvial and coastal plains, the Indo- Gangetic Plain alone accounting for roughly half the South Asian total throughout the period under review. Further, within any given type of terrain one may observe a general tendency for population density to be positively correlated with rainfall. Exceptions to this tendency are, of course, attributable to irri- gation in relatively dry regions such as the Punjab and Sind, while at the opposite extreme a superabundance of rainfall, as in parts of Assam and adjoining regions, may account for large areas of still uncleared forest with relatively low popula- tion density.

Although the broad regional pattern of population distribu- tion has not altered profoundly over the past century in rela- tive terms, there are several types of localities in which growth has been particularly rapid: dry areas into which irrigation has been newly introduced; densely forested areas that have re- cently undergone large-scale clearing; and, on a more local- ized basis, areas in which industrial and/or mining activities, plantations, and, for certain capital cities, the growth of the bureaucracy have given rise to substantial immigration streams. In general there has been a tendency for an infilling of the areas of relatively low density. If one examines the maps of plate XI.A.8, relating to the population as of 1971, one may discern not only a continuation, but an acceleration of that trend over the most recent intracensal period. The extent and implications of the regionally varying rates of population growth will be considered below in our analysis of plate XI.A.3, which portrays differential growth rates more clearly than the map series showing density in absolute terms.

Before we terminate this brief discussion of population den- sity, a note of caution is in order. Density per se conveys little meaning. The significance of a specific density figure hinges on its relationship to carrying capacity—that is, to the ability of the territorial unit to which the figure applies to support life at a reasonable standard, given the area's level of technological development. Thus, areas of low absolute density may in fact be dangerously overpopulated if their carrying capacity is also very low, while in theory—though not in practice, in the South Asian context—areas of very high density may not be over- populated at all. A consideration of the meaning of Rajasthan's density, 153 persons per square mile in 1961 (the lowest of any state in India), may prove instructive. To the uncritical mind, considering India's average 1961 density of 439 persons per square mile, Rajasthan might appear relatively underpop- ulated; on the other hand, its density was nearly sixty times that of the physically similar and comparably productive state of Nevada in the United States (2.6/sq. mi. in 1960). By global standards, virtually all of South Asia may be judged overpopulated. That is, whatever the physical environment may be, it can support the numbers of people it does only be- cause most of them live at levels of consumption that are among the world's lowest. There is, however, nothing immu- table about this situation. Real population pressure, we have noted, is a function not merely of numbers of people and the productive capacity of the land, but also of the technology applied to production. The demographic future of South Asia, then, will depend not only on its future population growth, but also on its ability to cope with that growth by increasing its productive capacity.

Sources

Censuses

Ceylon (1911), (1921), (1946), (1953), (1963); India (all censuses from 1872 to 1961); Nepal (1952-54), (1961); Pakistan (1951), (1961).

General References

Ceylon blue book for the year 1873; Goa, Daman, and Diu . . . (1965); Kabul Times annual (1967).

Government Document

Great Britain, Parliament, House of Commons, East India, Sta- tistical Papers and Abstracts, Statistical abstract . . . (1865–66/ 1874–75)

XI.A.3. Population Growth, 1901–61

In both India and Pakistan the 1961 census publications in- cluded data showing the growth of population for the areas of the states, union territories, or provinces then composing the nation over each decade since 1901. Thus, the census had to recast data for the administrative units, often down to the lo- cal level, for each of the enumeration years from 1901 to 1951 into the territorial frame existing in 1961. Thanks to that la- borious undertaking in the two countries, we are able to pre- sent in plate XI.A.3 a map showing absolute and relative pop- ulation growth, by major administrative subdivisions, for a sixty-year period over most of South Asia. Data for Ceylon were also available for the entire period. Within India our map omits data for union territories because of considerations of scale.

For each territorial unit portrayed on the map, two stan- dardized graphs are presented. As the legend indicates, the graph to the left depicts the total population in millions at each of seven censuses. It also specifies the density per square mile in 1921 and in 1961. The right-hand graph utilizes a logarithmic scale to show percentage growth rate for each of six intercensal periods. The curve of population growth is scaled by index numbers taking the population of the year 1921 as the base (100) for comparison with all other census years. A curve of a given slope indicates a particular inter- censal growth rate no matter where it may fall on any of the right-hand graphs. The growth range (or the fact of a popu- lation decrease) for each intercensal decade is given by the varying colors under the curve. Figures in the lower right cor- ner of the graph show percentage growth for the periods 1901–21 (upper figure) and 1901–61 (lower figure).

South Asian demographers point to the year 1921 as the approximate starting point of the region's population explo- sion. As is shown on the synoptic graphs for South Asia as a whole, the area's population grew relatively slowly in the pe- riod before 1921 and with increasing rapidity thereafter. The apparent slight deceleration in the rate of growth for the dec- ade 1941–51 may (despite the great Bengal famine of 1943) be spurious, since it is conceded that the 1941 census returns for India, and particularly for Bengal, were significantly in- flated because, within both the Hindu and the Muslim com- munities, certain census enumerators and higher officials used the census to advance their communal political goals. It is noteworthy that from 1921 onward not a single state of India or province of Pakistan shows a population loss in any inter- censal decade, though the gains for Indian Punjab and East Pakistan in the period 1941–51 are minimal as a consequence of their net population losses at the time of partition. By con- trast, nine of the eighteen areas depicted showed an absolute loss in either 1901–11 or 1911–21, Uttar Pradesh (then the United Provinces) registering losses in both periods. Most of these losses occurred during the decade 1911–21, which was marked by an unusually high number of natural catastrophes and some 12 to 14 million deaths in the influenza epidemic of 1918. Over South Asia as a whole growth during that period was negligible, while within the present area of India itself there was actually a decline of nearly one million persons (0.31% of the 1911 population).

The causes of the post-1921 population explosion in South Asia are manifold. In part they are natural, the period since 1921 having been one in which serious drought and flooding were much less common than in the previous half-century. (This is convincingly documented in graphic and cartographic from in the Report on the 1951 Census of India, and in Tre- wartha and Gosal [1957].) But in part the causes are those now common to most of the lesser developed portions of the world: a pronounced drop in the average rates of mortality with little decline in the average rates of birth. While data on vital rates in South Asia are open to question (Sri Lanka/ Ceylon excepted) because of the incomplete registration of both births and deaths, the general trends depicted in the lower left corner of plate XI.A.3 make clear the fundamental cause and the general magnitude of the growing rates of natural in- crease (birth rate minus death rate) of the region.

As was noted in the discussion of plates XI.A.1 and 2, there are significant variations in the rates of population increase within South Asia from one area to another. For the period 1901–61 growth rates range from 54% (Uttar Pradesh) to 158% (Assam), the figure for South Asia as a whole being 80%. The corresponding figures for the period 1921–61 are 47% (Jammu and Kashmir), 130% (Assam), and 72%. The fastest growth occurred, for the most part, in areas of rela- tively low density that have been subject in recent times either to a marked increase in irrigated acreage or extensive forest clearing for new agricultural settlement. Such areas are found mainly in the arid and semiarid portions in the northwest of the subcontinent, in the very humid northeast, and in various other wet, hilly regions of peninsular India. The growth rates recorded for West Pakistan, Rajasthan, and Assam illustrate this pattern. Kerala, though high in population density through- out the period under review, also grew rapidly because of new land clearing. Other high growth rates are attributable, at least in part, to industrial expansion and the immigration to which such expansion gave rise. West Bengal is a case in point; but growth in that state has also largely been a function of population movements out of East Pakistan at the time of and after the partition of India. Particularly low growth rates over the period 1921–61 are associated with areas whose popula- tion was already rather dense as of 1921 and in which subse- quent industrial expansion has not been great. Such areas are regions of chronic net outmigration, which, whatever their rates of natural increase may be, significantly dampens the overall rate of population increase. East Pakistan, Madras, and Uttar Pradesh are generally representative of this pattern, as are the plains portions of Bihar and Orissa, for which sepa- rate data are not presented.

Discounting aid from abroad, the future ability of South Asia to feed its growing population will depend in part on the expansion of its natural resource base and in part on its up- grading productivity per man and per acre. The possibilities for expansion of the resource base are distinctly limited be- cause relatively high proportions of the total acreage are al- ready under cultivation over much of South Asia. But the pos- sibilities for intensifying production through irrigation, greater use of fertilizer, double-cropping, and use of improved seeds are substantial. Such packages of agricultural inputs have re- sulted in truly remarkable changes in agricultural output in those areas, most notably East and West Punjab, where they have been widely adopted. It is in these areas that the Green Revolution may be said to have occurred; and there appears to be no compelling reason why that revolution should not ultimately diffuse throughout the greater part of South Asia.

But in the final analysis a solution to South Asia's popula- tion problem must, in our judgment, also involve a sharp re- duction in the rate of natural increase through bringing birth rates more nearly into line with the already fairly low death rates. There are, however, certain elements inherent in the demographic and social patterns that militate against this hap- pening in the near future. These factors are largely apparent in the graphs at the bottom of plate XI.A.3, which portray the sex ratio, age structure, and civil condition of India in 1931 and 1961 as well as of the United Kingdom in 1961 and the United States in 1960. Among the features the graphs reveal are the near universality of marriage in India, particularly among females; the very young average age of marriage (note