PLACE OF HYPOTHESIS IN EXPEEIMENTAL SCIENCE. 555

hypothesis which introduces a larger measure of unity into our conception of the objective conditions of the said pheno- mena. Similarly, physics originated in the attempt to define the essential nature of material substance. Thales saw all things in water, Anaximenes in air, Heraclitus in fire, Anaxagoras supposed matter to be constituted of minute invisible particles, but apparently failed to conceive of these particles as essentially homogeneous. An immense stride was taken in this direction when Democritus attempted to explain the secondary qualities of matter as resulting from differences in the arrangement of the ultimate particles. In modern times all the secondary qualities except flavour and odour have been resolved into modes of molecular motion. Had not the thinkers to whom we owe the molecular theory believed in a " real essence " underlying material phenomena, and in the possibility of defining that " real essence " had they been, in a word, imbued with the principles of modern empirical philosophy, and content to abide in the outward shows of things, noting and registering such rela- tions as were uniform physics had never been. The causes of events in the sense explained may or may not be conceived as events or sequences of events, but the relation between cause and effect is always one of coexis- tence ; nor does science rest content with a cause which is merely an event or a sequence of events. The first hypothesis concerning the cause of an event usually consists in treating it as the index of some other but insensible event suggested by analogy. Logicians of the empirical school regard analogy as distinct from induction, and treat it with scant respect ; but, as Lotze has pointed out (Logic, 274), hypothesis ought to rest upon analogy. Young is said by a flash of genius to have seen, u in the darkness which resulted from the fusion of two beams of light in Grimaldi's experiment, the analogue of the stillness produced by the neutralisation of one wave by another. This was the germ of the undulatory theory of light. In the case of the molecular theory of sound the analogy was more obvious. There the experimentum crucis was the ringing of a bell in a receiver nearly empty of air, the sound being scarcely audible. It was then clear that in so far as the vibrations of the resonant surface failed of affecting the ear, it was for want of propagation by the air, and the analogy of the ripple propagated in a gradually widening circle round an object dropped into water came to hand at once. If the latter was explicable as an agitation of particles, each of which after communicating its motion to its neighbour re-