¹ For a convenient and comprehensive summary of definitions, see Fairlie, John A., “The Nature of Political Representation” this Review, Vol. 34 (April-June, 1940), pp. 236–48; 456–66Google Scholar.

² An effort at conceptual clarification is made by De Grazia, Alfred, Public and Republic —Political Representation in America (New York, 1951)Google Scholar.

³ Finer, Herman, The Theory and Practice of Modern Government (New York, rev. ed., 1949), p. 219Google Scholar.

⁴ In his “Speech to the Electors of Bristol” (1774), Works, Vol. II, p. 12Google Scholar.

⁵ Cf. Beer, Samuel H., “The Representation of Interests in British Government,” this Review, Vol. 51 (Sept. 1957), p. 613Google Scholar, who points out how little general legislation was proposed or enacted in those days.

⁶ For the conception of “bounded rationality” as well as the notion that roles constitute some of the premises of decision-making behavior, we are indebted to Simon's, Herbert A. writings, notably Models of Man (New York, 1957)Google Scholar. Our own formulations of the concept of role are developed in Wahlke, John C. and Eulau, Heinz, Legislative Behavior: A Reader in Theory and Research (Glencoe, 1959)Google Scholar.

⁷ For a perspicacious discussion of ambiguities in representation, see Gosnell, Harold F., Democracy—The Threshold of Freedom (New York, 1948), pp. 124–42Google Scholar.

⁸ Most theories of functional or proportional representation are motivated or supported by tacit and untested assumptions about the relationship of legislators' behavior to the process by which they are selected. This is merely a special case of the general democratic assumption that political responsibility is the mechanism par excellence for bringing legislators' actions in line with the expectations of the represented.

⁹ See, for instance, Turner, Julius, Party and Constituency: Pressures on Congress (Baltimore, 1951)Google Scholar; or MacRae, Duncan Jr., Dimensions of Congressional Voting (Berkeley, 1958)Google Scholar.

¹⁰ See, for instance, Matthews, Donald R., The Social Background of Political Decision-Makers (Garden City, 1954)Google Scholar; or Hyneman, Charles S., “Who Makes Our Laws?” Political Science Quarterly, Vol. 55 (December, 1940), pp. 556–81CrossRefGoogle Scholar.

¹¹ See Eulau, Heinz, “The Ecological Basis of Party Systems: The Case of Ohio,” Midwest Journal of Political Science, Vol. 1 (August, 1957), pp. 125–35CrossRefGoogle Scholar.

¹² The samples for the four legislatures are 91 per cent in Tennessee, 94 per cent in California and Ohio, and 100 per cent in New Jersey. The four states composing the total sample represent different regions of the country, different ratios of metropolitan and non-metropolitan population, and different degrees of party competition. The interviews, using fixed schedules, uniform in all four states and including both open-ended, focussed-type questions as well as olosed, or fixed-answer type questions, averaged about two hours.

¹³ The reduction in the number of respondents from the total samples is, of course, due to the open-endedness of the question. Hence not all respondents could be used in the construction of the role types as they emerged from representatives' own definitions, and in the analysis.

¹⁴ In constructing stylistic and areal-focal role orientation types, the responses to the question were coded in terms of (a) characterization of job; (b) objectives of job; and (c) criteria of decision. Each total answer was broken up into individual statements and coded in terms of manifest content rather than latent meanings, though meaning was taken into consideration in locating manifest statements. Role orientation types were constructed by combining relevant manifest statements which seemed to make for a major orientational dimension. In general, data concerning criteria of decision yielded the stylistic orientation, and data concerning the objectives of the job yielded the areal orientation.

¹⁵ Constitutional Government and Democracy (Boston, rev. ed., 1950), p. 297Google Scholar.

¹⁶ In the years before the second World War, public opinion polls several times sampled expectations in this regard. Relevant poll questions were: (1) Do you believe that a Congressman should vote on any question as the majority of his constituents desire or vote according to his own judgment? (2) Should members of Congress vote according to their own best judgment or according to the way the people in their district feel? (3) In cases when a Congressman's opinion is different from that of the majority of the people in his district, do you think he should usually vote according to his own best judgment, or according to the way the majority of his district feels? In three of four polls, 61, 63 and 66 per cent, respectively, of the respondents said the Congressman should vote the way people feel. In the fourth poll, only 37 per cent gave this answer. See Cantril, Hadley, ed., Public Opinion, 1935–1946 (Princeton, 1951), p. 133Google Scholar.

¹⁷ As the Trustee orientation includes responses stressing traditional moral values, it might be assumed that these virtues—such as following one's conscience or what one feels to be “right”—are more valued in rural Tennessee than in the three more urbanized states. But inspection of the frequency with which this attitude appears in Tennessee as against the other states does not reveal significantly different distributions of relevant responses: California—18%; New Jersey—8%; Ohio—28%; and Tennessee—23%.

¹⁸ Of the 46 Tennessee respondents who mentioned an areal orientation, only four came from competitive and five from semi-competitive districts.

¹⁹ Competition in district was severally defined in the four states on the basis of past election returns. Space limitations prevent us from specifying the criteria here. They may be obtained from the authors.

²⁰ χ ² = 9.238 for the entire array, where d.f. =4, p ≥ .05. If the middle categories are omitted and only competitive and one-party districts are compared with respect to state and district orientation alone, χ ² = 7.12; d.f. = 1; p<.01.

²¹ However, this finding may be spurious. It might be less a function of the political character of the district than of its ecological character. Competitive districts are, more often than not, located in metropolitan areas, while one-party districts are more frequent in non-metropolitan areas. It seemed advisable, therefore, to control the districts' political character by their ecological character. For this purpose, the districts were divided on the basis of the 1950 Census specifications. The hypothesis concerning the relationship between political character of district and areal orientation was clearly maintained in both metropolitan and non-metropolitan districts. However, while the pattern proved similar in both ecological categories, a greater proportion of district-and-state-oriented representatives appeared in the non-metropolitan than in the metropolitan areas, suggesting a pull towards greater dichotomization of areal orientations in the metropolitan environment. In view of the intimate connection in industrialized states between metropolitan and state-wide problems, this result is not surprising. It seems that the state is more salient as a focus of attention for representatives from metropolitan districts (no matter what their political character) than from non-metropolitan districts.

²² He might, of course, receive instructions from a state-wide clientele such as a pressure group or political party, but these constitute other dimensions of his attention foci.