PROGRESS in knowledge takes place by the discovery of facts, and by drawing inferences from the facts discovered. It is commonly supposed that the facts existed before we discovered them; and this hypothesis stands the test of practical utility. But it is not supposed, except by the most careless thinkers, that the inferences which we draw from the facts—our laws and principles—are anything more than conceptual summaries of the facts and sequences of facts within our ken. (We use the word fact, not because in current usage it means that about which there can be no doubt, but because it includes a wider range of phenomena than the word “thing.” The fact that 25 per cent, of the peas, produced by breeding from hybrids resulting from the union of a yellow and a green pea, are green can hardly be described as a “thing.” Our inferences may be right or they may be wrong, but they did not exist before we made them. Whether we can draw any sharp distinction between fact and inference; or, to put it in terms of space, whether we can draw a line of demarcation and say where fact ends and where inference begins, is a question which concerns the modern biologist perhaps more vitally than any other; yet it is one which very few have definitely formulated, much less attempted to answer. The cause of the almost universal failure to provide a satisfactory answer is a habit of the mind, encouraged by text-books of logic, which drives it to classify things, often dichotomously, into two mutually exclusive categories. Music affords an example. Some folk hold that the music of certain authors is good, whilst that of others is bad. Others, however, say that taste is a purely relative matter, and that no one has a right to say that the music of a given author is good, or bad, as the case may be.