To speak only of physics in the first place, this science presents different characteristics from those of astronomy. The speculative perfection of a science is measured by two correlative although distinct considerations, by the more or less complete co-ordination of the laws, and by the more or less accurate prevision of facts. Now, under one aspect or the other, even supposing that physics should make very important progress, it will always remain very much behind astronomy. Indeed, the celestial science presents an almost perfect unity; physics, on the contrary, is composed of several branches which are almost isolated from one another, and each one taken by itself cannot even reduce all its laws to a more general law. And, as to the second point, while a very small number of direct observations allows of rational and exact prevision of the whole of the celestial phenomena, physics only renders possible predictions which are generally founded upon experience at once immediate and within easy reach. Undoubtedly some parts of physics allow of the use of mathematical analysis (we shall see presently under what conditions). Nevertheless, the part played by experience is infinitely greater in physics than in astronomy. So it is in the former science that we first meet with the inductive method, which is afterwards used and developed in the other positive sciences. Although deduction continues to fulfil an important part, it already ceases to predominate here, because, says Comte, in it the institution of true principles begins to become more troublesome than the development of accurate consequences.[134] The inductive method implies these essential processes; 1∘ observation properly so called, that is to say the direct examination of the phenomenon such as it appears naturally: 2∘ experimenting, which is usually defined as the examination of the phenomenon more or less modified by artificial circumstances instituted by us in order to study it better; 3∘ comparison, that is to say the gradual consideration of a succession of analogous cases, in which the phenomenon becomes more and more simple. Of these three processes astronomy only makes use of the first. Physics cannot use the third which is reserved for biology; but it avails itself of the first and institutes the second. This is a fresh confirmation of the law established by Comte: to the complexity and increasing difficulty of the sciences, corresponds an increasing development of the processes of the positive method applicable to them.