Every time M.P.s in the House of Commons divide to vote lists are made by the tellers of all those who voted Aye and all those who voted No. Thirteen years worth of these division lists covering the period 1861-1926 are now available in machine readable form. The aim of this thesis has been to apply a range of exploratory methods to the voting data to obtain new information about political behaviour. We start with a review of statistical techniques that previous studies have used to analyse legislative votes. These range from simple summary statistics to sophisticated models of political relationships. The data that has been used in this study is then described and a short outline of the historical background is given. Initial data analysis includes the measurement of participation rates and the use of rebellion scores to measure party cohesion. Insight into the development of the party system is gained by investigations of both patterns of telling and the practice of pairing. A major part of the thesis describes the application of Shepard-Kruskal ordinal scaling to the data. This technique enabled us to examine the behaviour of individual M.P.s as well as parties and to see how this behaviour changed through time. To aid interpretation of the scaling configurations we made two extensions to the basic method. These were designed to quantify how well points are placed in the configurations. A third extension allows the inclusion of a weighting scheme. Finally we describe how we have used correspondence analysis and biplots. The latter enabled M.P.s voting patterns to be compared over different categories of divisions.