Severe must have been the selection of those that were capable of such progress, and correspondingly rapid the advance and differentiation of the species. (11) Using stones as weapons, and finding that broken stones do most damage, and breaking them for that purpose, the progressive hunter necessarily makes some sparks fly; and if these fall amongst dry leaves or grass, he may light a fire. “In making flint implements sparks would be produced; in polishing them it would not fail to be observed that they became hot; and in this way it is easy to see how the two methods of making fire may have originated.”[9] But if the production of fire by friction had been suggested by the polishing of flints, it could hardly have been discovered before the neolithic stage; whereas hearths are known of much earlier date. And it may have happened earlier whilst some one was polishing an arrow or a spear with another piece of wood: a supposition which dispenses with the long inference from a warm flint to a flaming stick. It is a curious fact that to this day in Australia fire is sometimes made by rubbing a spear-thrower upon a shield;[10] but I lay no stress upon this, as if such a practice must be traditionary from the earliest discovery of the method. Either in the chipping of flints or in the polishing of spears it is far easier, and a more probable way, to learn the art of making fire than by observing that dried boughs or bamboos driven together by the wind sometimes catch fire; because those processes include the very actions which the art employs: imitation of nature is not called for. It is true that the natives of Nukufetan in the South Seas explain the discovery of fire by their having seen smoke arise from two crossed branches of a tree shaken in the wind;[11] but this, probably, is merely the speculation of some Polynesian philosopher. Volcanoes, too, have been pointed out as a possible source of fire; and, in the myth, Demeter is said to have lit her torches at the crater of Ætna—an action fit for a goddess. But were such an origin of fire conceivable with savages, it would not show how they came to make it themselves. Fire at first must have excited terror. Until uses were known for fire no one could have ventured to fetch it from a volcano, nor to make it by imitating the friction of boughs in the wind. Fires were accidentally lit by man again and again, and much damage done, before he could learn (_a_) the connection of events, (_b_) the uses of fire, (_c_) purposely to produce it, (_d_) how to control it. The second and fourth of these lessons are much more difficult than the mere making of fire; they are essential, yet generally overlooked. It seems necessary to suppose a series of accidents at each step, in order to show the effects of fire in hardening wood, hollowing wood, cooking game, baking and (later) glazing clay, and so forth. Perhaps a prairie-fire disclosed the advantages of cooking game, and many a