Now, suppose that instead of these hypothetic gemmules we endow the units with memory in ever so slight a degree, how simple the explanation of all these facts becomes! What an unit has learned to do under given conditions it can do again under like circumstances. Memory _does_ pass from one unit to another, or we could not remember anything as men that happened in childhood, for we are not physically composed of the same materials. It is not at all necessary that an unit should remember it remembers any more than we in reading are conscious of the efforts we underwent in learning our letters. Few of us can remember learning to walk, and none of us recollect learning to talk. Yet surely the fact that we do read, and walk, and talk, proves that we have not forgotten how. Bearing in mind, then, the fundamental laws that the offspring is one in continuity with its parents, and that memory arises chiefly from repetition in a definite order (for we cannot readily reverse the process--we cannot sing the National Anthem backwards), it is easy to see how the oft-performed actions of an individual become its unconscious habits, and these by inheritance become the instincts and unconscious actions of the species. Experience and memory are thus the key-note to the origin of species. Granting that all living matter possesses memory, we must admit that all actions are at first conscious in a certain degree, and in the "sense of need" we have the great stimulation to action. In Natural Selection, as expounded by Mr. Darwin, there is no principle by which small variations can be accumulated. Take any form, and let it vary in all directions. We may represent the original form by a spot, and the variations by a ring of dots. Each one of these dots may vary in all directions, and so other rings of dots must be made, and so on, the result not being development along a certain line, but an infinity of interlacing curves. The tree of life is not like this. It branches ever outwards and onwards. The eyes of the Argus pheasant and peacock have been formed by the accumulation, through long generations, of more and more perfect forms; the mechanism of the eye and hand has arisen by the gradual accumulation of more and more perfect forms, and these processes have been continued along definite lines. If we grant memory we eliminate this hap-hazard natural selection. We see how a being that has once begun to perform a certain action will soon perform it automatically, and when its habits are confirmed its descendants will more readily work in this direction than any other, and so specialisation may arise.