¹ See, for example, Litt, Edgar, Ethnic Politics in America: Beyond Pluralism (Glenview, Ill.: Scott, Foresman, 1970)Google Scholar, and Glazer, Nathan and Moynihan, Daniel Patrick, Beyond the Melting Pot: The Negroes, Puerto Ricans, Jews, Italians and Irish of New York City, 2nd. ed. (Cambridge, Mass.: MIT Press, 1970)Google Scholar; also the volumes of readings edited by Brett W. Hawkins and Robert A. Lorinskas, The Ethnic Factor in American Politics and by Bailey, Harry A. Jr., and Katz, Ellis, Ethnic Group Politics (both Columbus, Ohio: Charles E. Merrill, 1970 and 1969 respectively)Google Scholar. Historically and comparatively in multiethnic societies, there have been a number of other situations beside those noted here; see Philip Mason's encyclopedic study Patterns of Dominance and also his Race Relations (both London: Oxford University Press, 1970)Google ScholarPubMed.

² See Kwavnick, D., “The Roots of French-Canadian Discontent,” Canadian Journal of Economics and Political Science, 31 (November, 1965), 509–523 CrossRefGoogle Scholar. On the strategies of various French Canadian political leaders, see Michael Oliver, “The Strategy of French Canadian Participation in Canadian Politics,” a paper presented at the International Political Science Association Conference, Ann Arbor, Michigan, September 1960.

³ Kashmir has a population that is more than two-thirds Muslim, but its internal politics have been of a rather special character that does not typify the rest of the country. Otherwise Assam (as constituted in 1961) has the highest proportion of Muslims (23.3 per cent) of any of the states. As to districts, only Murshidabad in West Bengal and Malappuram in Kerala have a majority of Muslims. In a sense there did remain for a short time after Partition in 1947 an area of Muslim dominance, for the “communally reserved” constituencies set up under the Government of India Act of 1935 provided legislative representatives for the Muslim community. India continued to be governed under this act until the new Constitution came into effect in 1950.

⁴ On the decision to eliminate reserved seats for Muslims, see Austin, Granville, The Indian Constitution: Cornerstone of a Nation (Oxford: Clarendon Press, 1966), pp. 144–156 Google Scholar. Reserved seats were retained for the Scheduled Castes (Harijans, or Untouchables) and Scheduled Tribes.

⁵ There are exceptions, unfortunately thus far few. The major effort to date has been the work of the Centre for the Study of Developing Societies in New Delhi and that of the opinion pollster E. P. W. da Costa. Some of the findings of the Centre's 1967 election survey have been published in Asian Survey's symposium issue on “Elections and Party Politics in India,” 10 (November, 1970)Google Scholar. Ramashray Roy of the Centre analyzes a survey done for the 1969 election in his The Uncertain Verdict (Bombay: Orient Longmans, 1972)Google ScholarPubMed. As an example of da Costa's work, see “The Indian General Elections 1967, the Structure of Indian Voting Intentions: January 1967, A Gallup Report with Analysis” (New Delhi: Indian Institute of Public Opinion, 1967)Google Scholar.

⁶ See for instance Dogan, Mattei and Rokkan, Stein, eds., Quantitative Ecological Analysis in the Social Sciences (Cambridge: M.I.T. Press, 1969)Google Scholar, and Merritt, Richard L. and Rokkan, Stein, eds., Comparing Nations: The Use of Quantitative Data in Cross-National Research (New Haven: Yale University Press, 1966)Google Scholar.

⁷ Chandidas, R. examines the problem in considerable detail in “Ecological Correlates of Politics: Problems of Data Equivalence,” a monograph presented to the Data Confrontation Seminar held at the Inter-University Consortium for Political Research, University of Michigan, 1969 (mimeo.)Google Scholar.

⁸ This is the procedure used by Morris-Jones, W. H. and Gupta, B. Das in “India's Political Areas: Interim Report on an Ecological Investigation,” Asian Survey, 9 (June, 1969), 399–424 CrossRefGoogle Scholar; also by Zagoria, Donald S., “The Ecology of Peasant Communism in India,” American Political Science Review, 65 (March, 1971), 144–160 CrossRefGoogle Scholar; and Elkins, David J., “Social Mobilization, Social Structure, and Politics: Evidence and Qualifications,” a paper presented at the annual meeting of the American Political Science Association, Chicago, 1971 Google Scholar. An earlier exercise along these lines is reported on in Government of India, Census of India 1961, Vol. I, India, Part I-A(i) and (ii), Levels of Regional Development in India, by Mitra, A. (Delhi: Manager of Publications, 1965 and 1966)Google Scholar. Much of the Chandidas project was devoted to fitting seats for members of the national Parliament to district areas, a much more difficult task, because these parliamentary boundaries do overlap district lines; see his “Ecological Correlates.”

⁹ See, for instance, the lists of predictor variables employed in Morris-Jones and Das Gupta, “India's Political Areas,” and Zagoria. The large number of variables available at district level also permits the use of statistical techniques like principal components analysis to ameliorate the problems of multicollinearity among the predictors. See Gupta, Biplab Das, “Socio-Economic Classification of Districts: A Statistical Approach,” Economic and Political Weekly, 6 (August 14, 1971), 1763–1774 Google Scholar; and also Morris-Jones, W. H. and Gupta, B. K. Das, “More Dealings with Districts,” a paper presented at the annual meeting of the Association of Asian Studies, New York City, March 1972 Google Scholar.

¹⁰ Many of these shortcomings are pointed out by Morris-Jones and Das Gupta in “India's Political Areas.”

¹¹ Since W. S. Robinson's classic essay in 1950, the literature on the “ecological fallacy” has become fairly extensive. Hayward R. Alker, Jr. surveys this literature in his “A Typology of Ecological Fallacies,” in Dogan, and Rokkan, , pp. 69–86 Google Scholar. For Robinson's original piece, see ”Ecological Correlations and the Behavior of Individuals,” American Sociological Review, 15 (August, 1950), 351–357 CrossRefGoogle Scholar.

¹² Elkins; Heginbotham, Stanley J., “The 1971 Revolution in Indian Voting Behavior,” Asian Survey, 11 (December, 1971), 1133–1152 Google Scholar; Weiner, Myron, “Political Development in the Indian States,” in State Politics in India, ed. Weiner, Myron, (Princeton: Princeton University Press, 1968), pp. 3–58 CrossRefGoogle Scholar.

¹³ For instance, from the yearbook put out by the Publications Division of the Ministry of Information and Broadcasting, Government of India India: A Reference Annual.

¹⁴ Benjamin, Roger W. and Blue, Richard N., with Coleman, Stephen, “Modernization and Political Change: A Comparative Aggregate Data Analysis of Indian Political Behavior,” Midwest Journal of Political Science, 15 (May, 1971) 219–261 CrossRefGoogle Scholar.

¹⁵ Ibid. This method is also limited to Parliamentary seats, as the state Legislative Assembly seats are far too small (usually 1/6 of a Parliamentary seat) to be compared individually with districts.

¹⁶ Brass, Paul R., “Ethnic Cleavages and the Punjab Party System,” a paper presented at the annual meeting of the Association of Asian Studies, New York City, March 1972 Google Scholar. Yet another method is to select a sample of Legislative Assembly seats that do fit the census boundaries exactly; the difficulty here is that most of the seats do not fit and so must be thrown out of the analysis. See Blair, Harry W., “Elections and Aggregate Data Analysis at the Block Level in Bihar, 1967–1969,” also a paper presented at the March 1972 meeting just notedGoogle Scholar.

¹⁷ Government of India, Census of India 1971, Series I—India, Paper 2 of 1972: Religion, by Sekhar, A. Chandra (Delhi: Manager of Publications, 1972), pp. 16, 95 Google Scholar. Bihar's Muslim population grew by 13.9 per cent during 1931–41, decreased by 7.3 per cent during 1941–51, then increased by 32.3 per cent during 1951–61. Apparently what had occurred during the 1950s was a significant reverse migration from East Pakistan back into Bihar after the large outward flow during and after the Partition of 1947. See Government of India, Census of India 1961, Vol. IV, Bihar, Part I-A(i), General Report on the Census, by Prasad, S. D. (Patna: Bihar Secretariat Press, 1968), pp. 484–485 Google Scholar. The Hindu-Muslim ratio of about 7-to-1 in 1961 and 1971 fits roughly into what Mason would call a “competitive” type; cf. Patterns of Dominance, pp. 60–65.

¹⁸ Purnea did in fact border East Pakistan after the Partition, but a slice of about 700 square miles was transferred to West Bengal in 1956, as a part of the States Reorganization Act. See Franda, Marcus F., West Bengal and the Federalizing Process in India (Princeton: Princeton University Press, 1968), pp. 8–61 CrossRefGoogle Scholar.

¹⁹ Data from which the map was derived were obtained from the district census handbooks that formed part of the 1961 census, e.g., Superintendent of Census Operations, Bihar, Census 1961, Bihar District Census Handbook 1: Patna, by Prasad, S. D. (Patna: Secretariat Press, 1966), pp. 90–93 Google Scholar. Data for the 1971 census by block had not become available by the time of writing this article.

²⁰ The first four general elections included both the state Legislative Assembly and the national Parliament. Because of the chronic instability that characterized state politics after the 1967 elections, the Assembly was dissolved in 1968 and new elections were held in 1969. Elections for the Parliament, due by the five-year constitutional requirement to be held at least by 1972, were called a year early by Mrs. Gandhi and thus were held in 1971. The Assembly held its sixth election in March 1972.

²¹ The United States House of Representatives elected in 1972, for example, had only 11 black members out of the total 435, while the population of the country is about 11 per cent Negro. For an analysis of Muslim representation in another state of India (Andhra Pradesh), see Khan, Rasheeduddin, “Muslim Leadership and Electoral Politics in Hyderabad: A Pattern of Minority Articulation,” Economic and Political Weekly 6 (April 10, 1971), 783–794 and (April 17, 1971), 833–840, esp. 837Google Scholar.

²² A study based on the first three elections concluded, partly on the basis of data similar to those given here for the 1952–1962 period (i.e., showing a rising proportion of the vote for Muslim candidates), that the Muslim community was being successfully integrated into the Indian political system. See Krishna, Gopal, “Electoral Participation and Political Integration,” Economic and Political Weekly, 2 (Annual Number, February 1967), 179–190 Google Scholar. See also, in rebuttal, , Ahmed, Imtiaz, “Indian Muslims and Electoral Politics,” Economic and Political Weekly, 2 (March 11, 1967), 521–523 Google Scholar; and more recently Ahmed's, “Secularization,” Seminar, 144 (August 1971), 22–26 Google Scholar; and in addition, Kothari, Rajni Politics in India (Boston: Little, Brown, 1970), pp. 245–247 Google Scholar.

²³ From 1532 to 1983 for MLA's during 1952–1972 (the maximum was 2153 in 1969), and from 197 to 421 for MP's during 1952–1971.

²⁴ On nuisance candidates in Bihar, see Blair, Harry W., “Caste, Politics and Democracy in Bihar State, India: The Elections of 1967” (Ph.D. thesis, Duke University, 1969), pp. 146–148 Google Scholar.

²⁵ To cut down on the number of petty candidates, the government requires that the nomination deposit be forfeited if a candidate does not get at least one sixth (16⅔%) of the vote. This discouragement seems ineffective, however, for 1301 MLA entrants lost their deposits in 1967, 1366 in 1969, and 1254 in 1972.

²⁶ Blair, Harry W., “Ethnicity and Democratic Politics in India: Caste as a Differential Mobilizer in Bihar,” Comparative Politics, 5 (October, 1972), 107–127 CrossRefGoogle Scholar. This pattern of inequality has continued through the 1972 election as well; see Singh, N. K., “Bihar: Many Faces of Caste Politics,” Economic and Political Weekly, 7 (April 8, 1972), 748–749 Google Scholar.

²⁷ The 1961 population data are being projected backward in time, a practice which necessarily entails the possibility of inaccuracies, just as does the forward projection to 1972 later on in this paper. In effect, it is being assumed that the Muslim population of Bihar is expanding at roughly the same rate as the population as a whole. This was distinctly not the case in the 1940s and early 1950s (cf. note 17, above) but is, one hopes, not an unreasonable assumption for the period since then. A caveat should also be added concerning constituency boundary changes over time. Because the geographical delimitation of seats was changed for the 1957 elections and then again for the 1967 elections, statistically precise comparisons cannot be made across all six elections for all seats, though a great many seats have remained almost intact geographically ever since 1952.

²⁸ It might be pointed out that while the Congress was electing so many Muslim MLA's in the 1950s, Muslim representation on the vitally important Bihar Pradesh Congress Executive Committee was significantly diminishing. In the 1930s and 1940s, Muslim membership on this committee averaged about 20 per cent, but in the 1950s it decreased to the level of 5–6 per cent. See Roy, Ramashray, “Dynamics of One-Party Dominance in an Indian State,” Asian Survey, 8 (July, 1968), 553–575 CrossRefGoogle Scholar; and his “Caste and Political Recruitment in Bihar,” Caste in Indian Politics, ed. Kothari, Rajni (New Delhi: Orient Longmans, 1970), pp. 228–258 at 242–244 Google Scholar.

²⁹ The ordinal categories used were 0–5 per cent, 5–10 per cent, deciles up to 40 per cent over 40 per cent. A test for accuracy was made by comparing the estimates arrived at in this fashion with direct census data on Muslim population for 69 MLA seats scattered around the state that actually do match up with the census units. The correlation (r) between the Muslim population estimated from the isoplethic map and the actual Muslim population for the 69 seats was .939. Using r², we could say that the degree of “congruence” was 89 per cent.

³⁰ Because of difficulties in the delimitation for the 1952 election, this analysis was not performed for the first general elections. No attempt was made to perform the same technique for Parliamentary constituencies, which are each six times the size of a State Assembly seat, because it was thought that the problem of “ecological fallacy” was too great. A number of techniques have been developed by geographers for working out correspondences between maps using different areal units. See, for instance, Robinson, Arthur H., “Mapping the Correspondence of Isarithmic Maps,” Annals of the Association of American Geographers, 52 (December, 1962), 414–425 CrossRefGoogle Scholar. (The terms “isarithmic” and “isoplethic” are often used interchangeably.) There has, however, been little interest in using such techniques on electoral data. For an exception, see Lewis, Peirce F., “Impact of Negro Migration on the Electoral Geography of Flint, Michigan, 1932–1962: A Cartographic Analysis,” Annals of the Association of American Geographers, 55 (March, 1965), 1–25 CrossRefGoogle Scholar.

³¹ It cannot, of course, be inferred from the increase in correlation that a greater proportion of Muslims voted for Muslim candidates in the 1962 election than in 1957, though that is probably the case. Given the statistics of Table 2, it is possible (albeit unlikely) that Muslims as individuals turned away from Muslim candidates in 1962. All we can say definitely is that Muslim candidates received more votes in general in areas more heavily populated with Muslims.

³² Because of the constituency delimitation problem mentioned above, correlations were not calculated for the 1952 elections. Again, it must be emphasized that Table 3 does not measure the relationship between voting by Muslims and voting for particular parties. It only measures the relationship between Muslim population and voting for the individual parties.

³³ As reported to the author, who was in Bihar at the time of the 1967 elections.

³⁴ Quraishi, Zaheer Masood, “Electoral Strategy of a Minority Pressure Group: The Muslim Majlis-e-Mushawarat,” Asian Survey, 8 (December, 1968), 976–987 CrossRefGoogle Scholar. Whatever success the Majlis had was rather short-lived, for it had more or less collapsed by the time of the 1969 poll. See the same author's “Emergence and Eclipse of Muslim Majlis-e-Mushawarat,” Economic and Political Weekly, 6 (June 19,1971), 1229–1234 Google Scholar; also Wright, Theodore P. Jr., “Muslims as Candidates and Voters in 1967 General Elections,” in Fourth General Elections in India, ed. Varma, S. P. and Narain, Iqbal, vol. 2 (Bombay: Orient Longmans, 1970), pp. 207–224 Google Scholar.

³⁵ On the Shoshit Dal, see Blair, , “Caste, Politics and Democracy,” pp. 334–338 Google Scholar.

³⁶ Varma, Vishwanath Prasad, “Analysis of Results,” in A Study of Mid-Term Elections in Bihar (1969), ed. Varma, V. P. (Patna: Institute of Public Administration, Patna University, 1970), pp. 76–77 Google Scholar; also his “District-wise Classification of Results,” in the same volume, pp. 94–95 and 101.

³⁷ There are only 13 seats out of the total 318 in Bihar that could be called in any way genuinely urban. The Jana Sangh won 2 of the 13 in 1972.

³⁸ Baxter, Craig disputes the existence of such a relationship in The Jana Sangh: A Biography of an Indian Political Party (Philadelphia: University of Pennsylvania Press, 1969), pp. 234–235 CrossRefGoogle Scholar. Burger, Angela also takes issue with the thesis in her Uttar Pradesh study, Opposition in a Dominant Party System (Berkeley: University of California Press, 1969), pp. 199–200.Google Scholar But see also Morris-Jones, and Gupta, Das, “India's Political Areas,” 416 Google Scholar; Graham, B. D., “A Report on Some Trends in Indian Elections: The Case of Uttar Pradesh,” Journal of Commonwealth Political Studies, 5 (November, 1967) 179–199 at 191 CrossRefGoogle Scholar; and Ahmad, Imtiaz, “Religion in Politics: A Comment,” Economic and Political Weekly, 7 (January 8, 1972), 81–86 at 84–85Google Scholar.

³⁹ For other electoral systems, evidence for the salience of religious cleavage appears mixed; in some it continues to be of major importance, while in others it is barely noticeable in electoral behavior. See Arend Lijphart's study of 10 Western systems, “Class Voting and Religious Voting in the European Democracies,” Survey Research Centre, Occasional Paper No. 8 (Glasgow: University of Strathclyde, 1971)Google Scholar.