But the medieval philosophy had asserted accidental relations between distinct sorts of _natures_, the form of living dog and the form of dead matter, for example. No one could know _a priori_ what effect an accidental relation would produce, and all accidental relations between different pairs of natures were different: at the most there was analogy between them. Every different nature had to be separately observed, and when you had observed them all, you could still simply write an inventory of them, you could not hope to rationalize your body of knowledge. Let us narrow the field and consider what this doctrine allows us to know about the wood of a certain kind of tree. We shall begin by observing the impressions it makes on our several senses, and we shall attribute to it a substantial form such as naturally to give rise to these impressions, without, perhaps, being so rash as to claim a knowledge of what this substantial form is. Still we do not know what its capacities of physical action and passion may be. We shall find them out by observing it in relation to different 'natures'. It turns out to be combustible by fire, resistant to water, tractable to the carpenter's tools, intractable to his digestive organs, harmless to ostriches, nourishing to wood-beetles. Each of these capacities of the wood is distinct; we cannot relate them intelligibly to one another, nor deduce them from the assumed fundamental 'woodiness'. We can now see why 'substantial forms' were the _betes noires_ of the seventeenth-century philosophers. It was because they turned nature into an unmanageable jungle, in which trees, bushes, and parasites of a thousand kinds wildly interlaced. There was nothing for it, if science was to proceed, but to clear the ground and replant with spruce in rows: to postulate a single uniform nature, of which there should be a single science. Now neither probatology nor cynology could hope to be [18] universal--the world is not all sheep nor all dog: it would have to be hylology; for the world is, in its spatial aspect, all material. Let us say, then, that there is one uniform material nature of things, and that everything else consists in the arrangements of the basic material nature; as the show of towers and mountains in the sunset results simply from an arrangement of vapours. And let us suppose that the interactions of the parts of matter are all like those which we can observe in dead manipulable bodies--in mechanism, in fact. Such was the postulate of the new philosophers, and it yielded them results.