¹ Of course, it must be added that any votes won by a splinter party might be considered as a success since the party itself, by definition, is a new party, and starts from zero. For the purposes of this Note, however, as will be seen, the electoral consequences will be considered as being profitable only when the combined vote of the parent party and the new splinter party (or parties) exceeds that of the formerly united party.

² See Mackie, T. T. and Rose, Richard, The International Almanac of Electoral History, 2nd edn (London: Macmillan, 1982)CrossRefGoogle Scholar, which is also the source for the party names used in this Note. Since there are some problems with the Belgian data in the Mackie-Rose handbook, I have resorted to the figures reported in Dewachter, Wilfried, ‘Changes in a Particratie: the Belgian Party System from 1944 to 1986’, in Daalder, Hans, ed., Parly Systems in Denmark, Austria, Switzerland, the Netherlands and Belgium (London: Frances Pinter, 1987), pp. 285–363.Google Scholar Of course, as far as party splits are concerned, the use of these data – which often fail to catalogue very minor parties or parties that fail to win legislative representation – is likely to bias my sample towards the more successful splinter parties.

³ It should also be noted that I have normally included only those fissions and fusions which are explicitly identified as such in the Mackie-Rose handbook, and thus some additional cases may have been neglected. One arguably quite important case which has also been excluded is the split in the right-wing Swiss National Action in 1971 which gave rise to the Republican Movement. Since the former had just begun to mobilize support when the split occurred, any comparison between its support in 1967 (when it won 0.5 per cent of the vote) and the support for the rump party (2.0 per cent) and the Republican Movement (3.5 per cent) in 1971 is likely to be misleading.

⁴ Dates here refer to the year of the first general election contested by the new party.

⁵ I exclude Norway and the United Kingdom where fusions did occur, but which are not included in these data, simply because of the inability to disentangle the consequences of fusion from those of fission.

⁶ See Sartori, Giovanni, Parties and Party Systems: A Framework for Analysis (Cambridge: Cambridge University Press, 1976).Google Scholar

⁷ See, for example, Klingemann, Hans-Dieter, ‘Election Programmes in West Germany, 1949–1980: Explorations in the Nature of Political Controversy’, in Budge, Ian, Robertson, David and Hearl, Derek, eds, Ideology, Strategy and Party Change: Spatial Analyses of Post-War Election Programmes in 19 Democracies (Cambridge: Cambridge University Press, 1987), pp. 294–323.CrossRefGoogle Scholar

⁸ While these mean figures necessarily exclude the case of the SDP/Alliance in Britain, it can be noted that the emergence of the SDP in Britain and its alliance with the Liberals in 1983 was electorally rewarding to a certain degree. In 1979, prior to the split/alliance, the combined vote of Labour and the Liberals was 50.7 per cent; in 1983, following the split/alliance, the combined vote of Labour, the Liberals and the SDP was 53.0 per cent, an increase of 2.3 per cent.

⁹ The left is defined as incorporating the historic ‘class left’, that is the traditional socialist and communist parties and their offshoots, as well as the ‘new left’ parties, while the right is defined as incorporating all other parties. Where, as is occasionally the case, a party of the left merges or forms an alliance with a party of the right, this is regarded as a left merger rather than a right merger.

¹⁰ The index of disproportionality is calculated on the basis of figures reported in Arend Lijphart, ‘The Political Consequences of Electoral Laws: a Critique, Re-Analysis, and Update of Rae's Classic Study’, paper presented to fourteenth World Congress of the International Political Science Association, Washingdon, DC, 1988. In this index, which is derived by summing the absolute value of all vote-seat share differences and dividing by two, a higher figure implies more disproportionality.

¹¹ Lijphart, , ‘The Political Consequences of Electoral Laws’, Appendix 2.Google Scholar

¹² Note that in two instances (Iceland in 1987 and Ireland in 1987) there is a simultaneous split from two parties.

¹³ As before, the post-fission vote represents the sum of the share won by the splinter party and the rump party.