Extrapolating the rates of population increase for the 1951–61 decade as given in the table on plate XI.A.3, one derives a set of hypothetical "doubling times" for the major administrative units of South Asia, ranging from a low of 9 years to a high of 136 years or, if we eliminate from consid- eration the then union territories of India, from 23 years for Assam to 77 years for Jammu and Kashmir. For South Asia as a whole the doubling time was approximately 35 years. Throughout the 1960s growth rates tended to rise; but during the 1970s in most of South Asia the acceleration of population growth appears to have been arrested. As of 1977 according to the Population Reference Bureau, the doubling time is about 30 years for the entire region, 33 years for India itself, only 24 years for Pakistan (West Pakistan in 1961), 26 years for Ban- gladesh (East Pakistan in 1961), 32 years for Afghanistan, 30 years for Nepal and 35 years for Sri Lanka. At this writing, despite some recent improvements, the short-run demographic prospects for South Asia do not yet seem bright.

Sources

Censuses

India (1961); Pakistan (1961).

General References

Population Reference Bureau (1975); United Nations, Demo- graphic yearbook (1965).

Government Document

Ceylon, Department of Census and Statistics, Statistical ab- stract . . . 1965.

Other Works

S. N. Agarwala (1967); A. J. Coale and E. M. Hoover (1958); K. Davis (1951); India (Republic), Office of the Registrar General, Vital statistics . . . (1961, listed under Government Documents); N. K. Sarkar (1957); G. T. Trewartha and G. S. Gosal (1957).

XI.A.4 and 5. Urbanization, 1872, 1901, 1931, and 1961

Although South Asia is one of the world's primary hearths of urbanism, cities having arisen within the Indus civilization in the third millenium B.C., and although cities like Varanasi and Delhi have probably been continuously inhabited since they were established—as Kāśī and Indraprastha, respectively —in about the 8th century B.C., the level of urbanization in the region today is low by world standards. Overall, towns and cities accounted for approximately 16.8% of the total popula- tion in 1961 and 18.5% a decade later. So large is South Asia's population base, however, that those percentages represented totals of 95 million urban dwellers in 1961 and 133 million in 1971. India alone had an urban population of 109 million in 1971, ranking fourth among the world's nations (behind China, the United States, and the Soviet Union in descending order).

Although four of the maps presented on plates XI.A.4 and 5 relate only to towns and cities of 50,000 or more, they con- vey a fairly good impression of the distribution of urban pop- ulation as a whole. The actual proportion of the total censused urban population represented on those maps is about 34% for 1901, 36% for 1931, and 57% for 1961; data are inadequate for deriving a percentage for 1872. Because of considerations of scale, however, it was not feasible to map all urban places, and the 50,000 threshold was in our judgment the lowest that would permit a uniform presentation of the data for the four selected census years. As the maps make clear, most of the urban growth within modern South Asia has been concen- trated in the last several decades. Although figures for the year 1872, when India's first census was taken, are far from com- plete, it is probable that fewer than 25 million persons were living in towns and cities throughout the region in that year. By 1901, the figure had risen to about 32.5 million (11.9%, estimates having been made for noncensused areas). The des- ultory pattern of change in the six decades before 1931 is borne out by the generally low ranges of growth depicted for towns and cities of 50,000 and over for the periods 1872–1901 and 1901–31, as shown on the maps for 1901 and 1931. In- deed, not a few towns and cities, mainly within the Gangetic Plain, suffered absolute declines in population during one or both of those periods. Surprisingly, the number of such towns was significantly greater during 1901–31 than in the preced- ing generation. On the whole there appears to have been, dur- ing the period under consideration, notable stagnation in the urban life of the chief inland centers developed under the Mu- ghals and the several regional powers who succeeded them, with a concomitant shift in urban activity toward the three presidency towns of Bombay, Calcutta, and Madras. The rele- vant data for specific cities are supplied in the table on plate XI.A.4, giving the population and rank of the leading cities of South Asia at each of the four dates for which maps have been made.

Since 1931, towns and cities have mushroomed over most parts of South Asia. Not only has the number of urban places with 50,000 or more inhabitants increased dramatically, but the rates of growth in the period 1931–61 are in general far greater than those of previous periods. The acceleration in the rate of urban growth was most pronounced among the larger cities. Whereas in 1931 there were no more than a half dozen cities with populations over 500,000, by 1961 the number had risen to twenty-nine. In the same period the number of cities with over a million population rose from two to nine. For a presentation of the pattern of urban growth in the period 1961–71 see plate XI.A.8.

Two maps on plate XI.A.5 indicate for the major adminis- trative subdivisions of South Asia the degree of urbanization as of 1931 and 1961, the measure of urbanization being the proportion of urban population to total population. Not count- ing areas like Nepal, for which 1931 data are lacking, or such small highly urbanized areas as the former province, now un- ion territory, of Delhi, the range in urbanization shown on these two maps runs from 3.4% (for Assam) to 22.1% (in the Western India States Agency) in 1931 and from 4.7% and 5.2% (for Himachal Pradesh and East Pakistan, respec- tively) to 28.2% (for Maharashtra) in 1961. Rates of urban- ization over 25% were still exceptional as late as 1961, while on the contrary rates below 10% were common.

In interpreting and comparing the rates of urbanization pre- sented in the maps and tables of plates XI.A.4 and 5 one ca- veat should be kept in mind. The definition of what constitutes an urban place varies from one country to another as well as from one census to another. In Ceylon, for example, urban places were all those legally classified as "municipal areas" or "urban council areas" irrespective of their size. In India and Pakistan, on the other hand, while legal status or the existence of urban amenities provided bases for designating settlements as urban, size has always been an important additional, if not the principal, determinant for urban classification, the thresh- old population of 5,000 being almost universally applied. But, whereas size was a sufficient basis for classifying a place as urban in India before 1961, two additional criteria were then added: having a population density of at least 1,000 per square mile, and having at least three-fourths of the working popula- tion engaged in nonagricultural pursuits.¹ While the uniform- ity in determining levels of urbanization obtained from the application of these rules was probably greater for 1961 than for preceding census years, for certain states comparability of 1961 data with those for previous years is decidedly vitiated. The requirement that an urban place have at least three-fourths of its working force in nonagricultural pursuits in particular resulted in the "declassification" of large numbers of settle- ments from urban to rural status. Thus, despite the creation of 497 new towns in the period 1951–61, there was a net decrease in the total number of towns from 3,060 to 2,700 in that dec- ade. The state most affected was Uttar Pradesh, where the number of towns fell from 486 to 267; but Maharashtra, Andhra Pradesh, Mysore, Gujarat, and Rajasthan also suffered rather significant losses. The total population of the declassi- fied towns came to nearly 4.4 million persons, over a fourth of whom were accounted for by Uttar Pradesh alone. Had the 4.4 million persons living in declassified settlements been added to India's 1961 urban population it would have risen to 83.3 million, and the rate of urbanization would have gone from 18.0% to 19.0%. While the overall impact of the definitional change is not great, its differential regional effect is significant for particular states.

Sources

Censuses

Ceylon (1921), (1946), (1953), (1963); India (1872), (1901), (1911), (1921), (1931), (1941), (1951), (1961); Nepal (1961); Pakistan (1951), (1961).

General References and Other Works

Same as for plates XI.A.1 and 2.

XI.A.6 and 7. Pattern of Internal and External Migration in the Pre-Independence and Post-Independence Periods

A question on "place of birth" has been included in every Indian census since 1881. Although data are therefore avail- able to enable us to plot the pattern of migration for a number 1 The full definitions of "urban" for each Indian census from 1891 to 1961 and the manner of their application are presented in the 1961 Census of India, vol. I, Part II-A(i), pp. 51–52. Similar statements appear in other censuses consulted in preparing plates XI.A.4 and 5. of different periods, considerations of space and economy of effort have led us to focus on two benchmark years, 1931 and 1961, in treating that subject on plates XI.A.6 and 7. Addi- tionally, 1951 data on the pattern of residence of displaced persons forced to leave or choosing to leave India or Pakistan at the time of partition or after are presented on plates X.A.2 and 4. Apart from our exposition of certain patterns of inter- nal migration within South Asia, we have also portrayed on the two world maps on plates XI.A.6 and 7 a variety of data relevant to the direction and volume of migration from South Asia to other parts of the world and between countries within South Asia itself.

To facilitate comparison between the patterns of the pre- independence and post-independence periods, as exemplified by the data for 1931 and 1961, we have kept the scales, leg- ends, and manner of presentation of the data for each period as consistent as possible. On both the upper maps, data relate to major administrative units: provinces of British India, ma- jor princely states, and states agencies in 1931; Indian states and union territories and Pakistani provinces in 1961. Be- cause of changes of administrative boundaries within India and Pakistan between 1931 and 1961, the regional groupings of those units are not identical for the two years, though broadly similar patterns have been maintained.

For each territorial unit portrayed on the maps of internal migration two columns or pairs of columns are presented, those on the left relating to immigration and those on the right relating to emigration. The difference in the heights of the columns, or in their total areas (where two pairs of columns are shown), indicates net immigration into or emigration from the territorial unit to which they relate. On the 1931 map, for example, one sees in Assam an extreme surplus of immigrants over emigrants, while for Bihar & Orissa the opposite situation obtains. Looking at the regional breakdown shown within the bar representing immigration into Assam, one notes that the two principal source areas are the provinces of Bengal and of Bihar & Orissa; similarly, scrutiny of the bar representing emi- gration from Bihar & Orissa reveals that the chief destinations of the emigrants from that area are Bengal and Assam. The breakdown within the bars for Bengal, in turn, depicts both the substantial emigration to Assam and the much larger mi- gratory stream from Bihar & Orissa. Throughout both maps, each bar segment depicting emigration from one area has its counterpart in a bar segment showing immigration into another.

A careful study of the maps of internal migration within South Asia will yield many valuable insights into other phe- nomena. Excluding from consideration the migratory streams between India and Pakistan related to the partition of the sub- continent, the size and composition of which were determined by the social composition of the population on both sides of the border before 1947, one finds that areas of pronounced net immigration are virtually always regions of relative op- portunity. Such areas are characterized by significant growth in one or more sectors of the economy or, at the very least, have undergone such growth during the lifetime of a large proportion of the migrants. The development of plantations in Assam and of industry in Bombay/Maharashtra and Bengal/ West Bengal, largely with the aid of an immigrant labor force, are cases in point. Conversely, areas of pronounced net emi- gration, again discounting the effects of partition, are usually, though not invariably, areas of relative stagnation. An obvious exception to the latter generalization is Punjab, which in 1961, despite its relatively favourable economic situation, showed a substantial surplus of emigrants to other states over immigrants from other states. The explanation lies in the venturesome spirit for which Punjabis, especially Sikhs, are renowned. The emigrants from that area are to be found throughout India in a wide variety of occupations: military, commercial, indus- trial, and agricultural. (Much of the land reclamation in the swampy Tarai regions of Uttar Pradesh, for example, is cur- rently being undertaken by Punjabi immigrants.)

The interpretation of migratory patterns, however, is not always simple, since several states comprise both limited areas of economic growth and more widespread areas of economic retardation. Bihar, for example, contains a major zone of rapid development stretching from the Jharia coalfields southward to the iron and steel center of Jamshedpur; it is largely to this region that its not insignificant stream of immigrants from other states (as well as an additional stream of intrastate mi- grants, not shown on our map) is attracted. But the plains re- gion of northern Bihar, containing most of the state's popu- lation, is one from which substantial emigration has been in progress for decades.

Among the maps in this atlas that might well be studied in conjunction with those portraying internal migration are those depicting overall population growth and urban growth (i.e., all the remaining map plates of subsection XI.A). Addition- ally, the atlas plates showing the distribution of plantation crops (XI.B.2, map (b)), industrial growth (all of subsection XI.C), and a synoptic view of the economy as a whole (XI.E.1) are worthy of note.