WHAT KINDS OF RATIONALE ARE NECESSARY OR APPROPRIATE FOR explanation in comparative politics? A rationale is essential for explanation, underlying any attempt to construct a theory, since it is the only means by which one can indicate why the event to be explained should follow from the precedent events held to explain it - something which a simple succession of events in itself, no matter how regular, can supply. In this article, I shall outline an answer to this question, and then indicate a framework of investigation inside which the basic requirements of political explanation can be met, and at the same time made relevant to strategies for research. There are, I suggest, only three rationales from which explanatory statements need to be derived, and these can be combined with one another in ways which are easily enough put into practice, but which also have important implications for the kinds of explanation which it is possible to achieve. I hope to come out, therefore, with a general framework into which any valid explanation can be put, and to suggest that by putting our explanations into such a framework, we will be better able to compare them with one another, and thus to ‘cumulate’ our ability to understand political phenomena drawn from different areas. I do not pretend to overcome the well-known problems in the way of a ‘science of politics’, and will indeed argue that within the pattern of explanation which I suggest such problems are insoluble.