loaded cannon at one end and a bundle of straw at the other, when if a spark be dropped anywhere on this train, the flame runs along in both directions, explodes the cannon, and sets alight the straw. 167. Indeed we have only to remember the semi-liquid nature of the axis cylinder to see at once that it must conduct a wave of motion as readily in one direction as in another. A liquid transmits waves in any direction according to the initial impulse. There is consequently no reason for asserting that because the usual direction is centripetal in a sensory nerve, and centrifugal in a motor nerve, each nerve is _incapable_ of transmitting excitations in both directions. And I think many phenomena are more intelligible on the assumption that neural transmission is in both directions. If the eye is fixed steadfastly on a particular color during some minutes, the retina becomes exhausted, and no longer responds to the stimulus of that color: here the stimulation is of course centripetal. But if instead of looking intently on the color, the mind (in complete absence of light) pictures it intently, _this cerebral image is equally capable of exhausting the retina_; and unless we believe that color is a cerebral, not a retinal phenomenon (which is my private opinion), we must accept this as proof of a centrifugal excitation of a sensory tract. Another illustration may be drawn from the muscular sense. There may be a few sensory fibres distributed to muscles; but even if the observations of Sachs[188] should be confirmed, I do not think that all muscle sensations can be assigned to these fibres, but that the so-called motor fibres must also co-operate. When a nerve acts upon a muscle, the muscle reacts on the nerve; and when a nerve acts on a centre, the centre reacts on the nerve. The agitation of the central tissue cannot leave the nerve which blends with it unaffected; the agitation of the muscular tissue must also by a reversal of the “current” affect its nerve. Laplace points out how the movement of the hand which holds a suspended chain is propagated along the chain to its terminus, and if when the chain is at rest we once more set that terminus in motion, the vibration will remount to the hand.[189] The contraction of a muscle will not only stimulate the sensory fibres distributed through it, but also, I conceive, stimulate the very motor fibres which caused the contraction, since these fibres blend with the muscle.[190]