558 j. M. EIGG :

may seem superfluous. I need therefore only remind the reader that the essence of the law of gravitation consists in its connecting facts so disparate as the fall of a stone and the persistence of the planets in their orbits as effects of one and the same cause ; that the molecular theory of heat identifies the cause of the sensation, no matter what the sensible ante- cedent may have been, with an agitation of the insensible particles of matter ; that Lavoisier's theory of combustion traces processes so apparently diverse as calcination, com- bustion and acidification to the operation of the same force, viz., oxygen. To sum up the matter : physical science presupposes the determination of phenomena by objective conditions or causes ; observation establishes empirical rules to the effect that certain phenomena have hitherto, so far as experience has extended, coexisted with or ensued upon certain other phenomena ; experiment purifies these rules from all unes- sential elements by the process of elimination analysed by Mill ; the result is the formulation of conditional uniformi- ties of coexistence, antecedence and sequence ; then begins the process of induction in the strict sense, which consists in framing an hypothesis based on analogy concerning the nature of the causes which determine the uniformity in ques- tion ; the hypothesis, when framed, is tested by attempting to deduce the phenomena from it with the help, if necessary, of experiments to test the adaptability of the theory to spe- cial cases. This last process is sometimes called verification, and no doubt in a certain relative sense it is so. Verifica- tion, in the sense of strict proof, it certainly is not, since even if the hypothesis stands the test its absolute truth is not thereby established. The principal use of the process commonly known as veri- fication is not to verify, but to disprove. It is a criterion rather of error than of truth. If we were to adopt the view that mere deduction of the phenomena from the hypothesis, together with the fact that the theory has been so articulated as to meet all special cases as yet given by nature or in- vented by experiment and to serve as an instrument of pre- diction, amounted to absolute proof, we should land ourselves in the absurd position that what is proved to-day may be disproved to-morrow, since hypotheses which explain certain facts perfectly well are sometimes superseded by others which explain them not a whit better, but happen to have a wider range of applicability ; it is very difficult, in any given in- stance, to say that the resources of analogy have been exhausted. That some limit to the powers of the human