In what, then, does the scientific method consist? It would be difficult to give a precise definition; it has, however, two main characteristics, the choice of facts and the treatment of facts. It does not seem to be generally recognised that scientific men do choose their facts; there are many people who suppose that all facts are of equal interest to scientific men, and that information respecting the number of nightingales heard in Hertfordshire during a certain month, for instance, is a contribution to scientific knowledge. It should be obvious, however, that a mere random collection of facts is very unlikely to aid either practice or theory. The aim of science is not to form catalogues, but to form theories describing phenomena, and to this end some facts are pertinent and a very great number are not. All men, faced with a problem of any kind, choose such facts for examination as they consider relevant. Sherlock Holmes often bewildered Watson by pondering over facts that Watson considered irrelevant, but Watson’s surprise was a proof that even he had a standard of relevance. The history of any science shows that the facts first chosen were those most likely to be repeated. Such facts obviously lead to statements which have a greater or less degree of generality. That an unsupported stone falls to the ground is a fact of this kind. The facts chosen by the man of science are those that permit generalisation. For this reason they usually differ entirely from the facts of interest to historians. After selecting, in accordance with this principle, the facts which are to be examined, the next step consists in establishing relations between sets of these facts. The precise expression of these relations is called a law of nature, to use a somewhat old-fashioned terminology. If now all the relations between certain sets of facts can be expressed in one general statement, that general statement is called a scientific theory. The ultimate aim of the scientific method is to create scientific theories. The scientific theory, however, usually introduces an element which has not been or cannot be directly observed, and also, as we have seen, usually proves to have been too hasty a generalisation. Its function is to co-ordinate known phenomena and to predict hitherto unobserved phenomena. The extent to which it does this is the measure of its success as a scientific theory, and, since the primary object of the scientific theory is to express the harmonies which are found to exist in nature, we see at once that these theories must have an æsthetic value. The measure of the success of a scientific theory is, in fact, a measure of its æsthetic value, since it is a measure of the extent to which it has introduced harmony in what was before chaos.