¹ We shall have occasion to mention certain non-American party systems, but shall do so only in terms of what is generally known about them, and thus does not require supporting evidence or documentation. The empirical data with which we shall deal are all American.

² Key, V. O. Jr., Politics, Parties and Pressure Groups, 3rd ed. (New York, 1952), pp. 314–22Google Scholar.

³ Cf. Merriam, Charles E. and Gosnell, Harold F., The American Party System, 3rd ed. (New York, 1940), p. 8Google Scholar; and McKean, Dayton D., Party and Pressure Politics (Boston, 1949), pp. 17–18Google Scholar. Merriam and Gosnell, to be sure, say later that such a totalitarian party system is not really a “party system,” in the usual sense of that term, at all.

⁴ I.e., party systems operating within a total governmental structure that permits organized party opposition. This meaning of “democratic” party systems clearly is the one assumed by Key and other writers, and is one with which the present writers have no quarrel.

⁵ Cf. Key, V. O. Jr., Southern Politics in State and Nation (New York, 1949)Google Scholar, Ch. 11.

⁶ “Toward a More Responsible Two-party System,” this Review, Vol. 44 (Sept., 1950, Supplement), p. 18Google Scholar. Cf. Merriam and Gosnell: “One of the outstanding characteristics of the American party system is that it is a two party system …” Op. cit., p. 2; or Bone: “… a … conspicuous feature … of the American party system is its biparty nature.” Bone, Hugh, American Politics and the Party System (New York, 1949), p. 356Google Scholar. Professor Schattschneider speaks somewhat more cautiously of the American two-party system: “… American politics is dominated and distinguished by the two-party system …. The two-party system is the most conspicuous feature of American political organization.” Schattschneider, E. E., Party Government (New York, 1942), pp. 66–67Google Scholar, emphasis added.

⁷ Cf. Schattschneider, pp. 162–65; and Key, Politics, Parties and Pressure Groups, Ch. 11.

⁸ One of the clearest examples of the dependence of present expectations upon past election results in classifying party systems was given by Professor Lipson in a recent issue of this Review. He stated that the following conditions must be present if a party system is to be called “two-party”: “(1) Not more than two parties at any given time have a genuine chance to gain power. (2) One of these is able to win the requisite majority and stay in office without help from a third party. (3) Over a number of decades two parties alternate in power.” “The Two-party System in British Politics,” this Review, Vol. 47, pp. 337–58, at p. 338 (June, 1953). Clearly, the expectations involved in Lipson's (1) depend upon the observations of past voting behavior which support his (2) and (3). See also the definitions, explicit and implicit, of “the” American two-party system in the works cited in footnotes 3, 4, and 7, above.

⁹ See, for example, the process by which the present writers arrived at the dividing-line between the “two-party” states and the other states in Table I, below. The specific dividing-line employed in that table would very likely be inappropriate for a classification of, say, the provincial party systems of Canada. The process for determining the dividing-line—as opposed to the specific location of the line itself—should be equally applicable to Canadian and American party systems.

¹⁰ See, for example, the graph in Key, , Politics, Parties and Pressure Groups, p. 225Google Scholar.

¹¹ 1892—Weaver (Populist), 11 per cent; 1912—Roosevelt (Progressive), 27.5 per cent; and 1924—LaFollette (Progressive), 16.6 per cent.

¹² 1892—Weaver (Populist), 22; 1912—Roosevelt (Progressive), 88; 1924—LaFollette (Progressive), 13; 1948—Thurmond (States' Rights), 39.

¹³ Cf. the graphs in Bean, Louis, How to Predict Elections (New York, 1948), pp. 15, 19Google Scholar.

¹⁴ The raw data for Table I and the subsequent tables were drawn from the World Almanac (published by the New York World-Telegram and Sun), various issues from 1915 to 1953. The period chosen for studying election results is from 1914 to 1952, and was selected for two reasons: It eliminates most of the effects of the 1912 split in the Republican party—an abnormal and temporary situation; and it provides a period of four decades, which, in the opinion of the authors, is a long enough period to indicate trends. The three offices were chosen partly because of the ready availability of statistics about them, and partly because in many states they are the only offices for which all the voters of the state vote.

¹⁵ In those states in which third-party or independent candidates (both listed in the tables as “3-P”) won at least one election (Alabama, California, Florida, Louisiana, Minnesota, Mississippi, Nebraska, Oregon, South Carolina, and Wisconsin), their victories were added to the second-parties' victories to show “total-opposition” percentages.

¹⁶ Cf. Key, Southern Politics in State and Nation, Chs. 4, 10; and Heard, , A Two-Party South? (Chapel Hill, N. C., 1952), p. 98Google Scholar.