Let us therefore leave these doubtful analogies to one side. To find out if men were primitively inclined to the confusions imputed to them, we should not study animals or children of to-day, but the primitive beliefs themselves. If the spirits and gods of nature were really formed in the image of the human soul, they should bear traces of their origin and bring to mind the essential traits of their model. The most important characteristic of the soul is that it is conceived as the internal principle which animates the organism: it is that which moves it and makes it live, to such an extent that when it withdraws itself, life ceases or is suspended. It has its natural residence in the body, at least while this exists. But it is not thus with the spirits assigned to the different things in nature. The god of the sun is not necessarily in the sun, nor is the spirit of a certain rock in the rock which is its principal place of habitation. A spirit undoubtedly has close relations with the body to which it is attached, but one employs a very inexact expression when he says that it is its soul. As Codrington says,[130] "there does not appear to be anywhere in Melanesia a belief in a spirit which animates any natural object, a tree, waterfall, storm or rock, so as to be to it what the soul is believed to be to the body of man. Europeans, it is true, speak of the spirits of the sea or of the storm or of the forest; but the native idea which they represent is that ghosts haunt the sea and the forest, having power to raise storms and strike a traveller with disease." While the soul is essentially within the body, the spirit passes the major portion of its time outside the object which serves as its base. This is one difference which does not seem to show that the second idea was derived from the first.