Suppose now (these assumptions being granted) that we have a baby before us who sees a candle-flame for the first time, and, by virtue of a reflex tendency common in babies of a certain age, extends his hand to grasp it, so that his fingers get burned. So far we have two reflex currents in play: first, from the eye to the extension movement, along the line 1--1--1--1 of Fig. 3; and second, from the finger to the movement of drawing back the hand, along the line 2--2--2--2. If this were the baby's whole nervous system, and if the reflexes were once for all organic, we should have no alteration in his behavior, no matter how often the experience recurred. The retinal image of the flame would always make the arm shoot forward, the burning of the finger would always send it back. But we know that 'the burnt child dreads the fire,' and that one experience usually protects the fingers forever. The point is to see how the hemispheres may bring this result to pass. [Illustration: FIG. 4.--The dotted lines stand for afferent paths, the broken lines for paths between the centres; the entire lines for efferent paths.] We must complicate our diagram (see Fig. 4). Let the current 1--1, from the eye, discharge upward as well as downward when it reaches the lower centre for vision, and arouse the perceptional process _s^1_ in the hemispheres; let the feeling of the arm's extension also send up a current which leaves a trace of itself, _m^1_; let the burnt finger leave an analogous trace, _s^2_; and let the movement of retraction leave _m^2_. These four processes will now, by virtue of assumption 2), be associated together by the path _s^1--m^1--s^2--m^2_, running from the first to the last, so that if anything touches off _s^1_, ideas of the extension, of the burnt finger, and of the retraction will pass in rapid succession through the mind. The effect on the child's conduct when the candle-flame is next presented is easy to imagine. Of course the sight of it arouses the grasping reflex; but it arouses simultaneously the idea thereof, together with that of the consequent pain, and of the final retraction of the hand; and if these cerebral processes prevail in strength over the immediate sensation in the centres below, the last idea will be the cue by which the final action is discharged. The grasping will be arrested in mid-career, the hand drawn back, and the child's fingers saved.