¹ I wish to acknowledge helpful criticisms of several Ohio colleagues: Ben A. Arneson, Wilfred E. Binkley, E. Allen Helms, and Howard White. While their comments were most beneficial to me, they should not be taxed with any responsibility for the content of this article.

² The county offices included in the computations underlying the graph were those listed as “county offices” in Ohio Election Statistics; legislators and judges are not so listed. Over the period covered by the graph there was change from time to time in the offices filled in presidential years.

³ Just prior to the 1940 election the ballot form was changed from a straight party-column ballot to a double ballot with one ballot containing the presidential candidates and another the remainder of the candidates. The move was a stratagem to deprive state and local candidates of the benefit of the Roosevelt popularity. From Figure 1 it would seem unlikely that the change in ballot form had a great immediate effect.

⁴ These calculations are on the assumption that all local offices were contested. In 1944 almost 30 per cent of Ohio county offices were filled without contest in either the primary or the general election. The proportion filled without contest, however, declined as the presidential voting became closer.

⁵ Gray, Wood, The Hidden Civil War, The Story of the Copperheads (New York, 1942), p. 137Google Scholar.

⁶ Porter, G. H., Ohio Politics During the Civil War Period (New York, 1911), pp. 106, 157Google Scholar.

⁷ Pennell, W. W. and Vance, J. R., “Holmes County Rebellion—Fort Fizzle,” Ohio Archeological and Historical Quarterly, Vol. 40, pp. 23–51 (01, 1931)Google Scholar.

⁸ Adams, Auglaize, Clermont, Darke, Fairfield, Holmes, and Perry.

⁹ Athens, Coshocton, Greene, Guernsey, Highland, Jackson, Lake, Ross, Vinton.

¹⁰ A similar analysis of those counties Republican in 1944 by 50–54.9 per cent and under 50 per cent urban showed that in 1944 the 1861 Democratic counties filled 71.1 per cent of the county posts with Democrats; the 1861 Unionist counties gave 22.2 per cent of their county offices to the Democrats. Of the twelve counties under 50 per cent urban and 50–54.9 per cent Republican in 1940, the 1861 Democratic counties filled 80.6 per cent of their county offices with Democrats; the 1861 Unionist counties gave only 25.0 per cent of their local offices to the Democrats.

¹¹ Lest “Germanism” and the peculiar behavior of counties of German origin in the presidential voting of the 1940's be accepted as accounting for the success of Democratic local candidates in these counties in resisting the presidential tide, it is in order to examine those counties whose local Republican candidates withstood the great upsurge in Democratic presidential strength in 1936. Fourteen counties under 50 per cent urban in 1940 gave Roosevelt in 1936 a plurality by between 50–54.9 per cent of the two-party vote. The nine 1867 Unionist counties of this group filled 72.8 per cent of their county offices with Republicans; the five 1867 Democratic counties filled 46.7 per cent of their county offices with Republicans.

¹² Local histories constitute a source that remains to be exploited by students of politics. Although the production of local chronologies does not seem to attract first rate historical talent, local histories, perhaps fortuitously, record data that are often most revealing when viewed through an appropriate analytical framework.

¹³ In fact, it also seems that ghostly descendants remain of partisan groups that once were far more corporeal. In 1944 Ohio had 28 rural counties that were predominantly Republican in the presidential voting in that year, that is, less than 40 per cent Democratic. Of the 28 counties 21 had been Unionist in 1867. In the seven 1867 Democratic counties, the Democrats in 1944 had nominees for 68.2 per cent of the county posts; in the twenty-one 1867 Unionist counties the Democrats had nominees for only 35.4 per cent of the county posts. In the seven 1867 Democratic counties some sort of Democratic group or tradition must have lingered on to produce competition for county office in spite of the fact that the counties had long since moved over to the Republican side of the fence.

¹⁴ The relationships shown in Figure 1 probably reflect the peculiar circumstances of Ohio, yet somewhat similar relationships appear in the Iowa data. That may be seen by placing on a chart the following series, in which the Democratic percentage of county offices is listed first after the date and then the percentage of counties with Democratic presidential pluralities: 1920, 12.0, 0.0; 1924, 15.4, 0.0; 1928, 12.9, 6.1; 1932, 36.5, 93.9; 1936, 34.8, 81.8; 1940, 21.9, 35.4; 1944, 16.7, 33.3; 1948, 20.3, 53.6. The figures were built up from the listing of county officials in the Iowa bluebook. The 1928 and 1932 figures may not be exactly comparable with those of other years as the bluebook did not list county supervisors for those years.