158. The facts expressed in the “law of isolated conduction” are important, and are difficult of explanation; but it is obvious that they cannot be referred to the presence of the medullary sheath. Nor indeed will any insight into the propagation of stimulation through the central axis be intelligible until we have reformed our anatomical theories, and taken the Neuroglia into account. The theory which connects every fibre directly with a cell, and every cell with another by anastomosis--even were it demonstrated--would not explain the law of isolated conduction. Butzke cogently remarks[183] that such a disposition of the elements should render all neural paths invariable; whereas the fact is that they are very variable. We learn to perform actions, and then we unlearn them; the paths are traversed now in one direction, now in another. Fluctuation is the characteristic of central combinations. And for this fluctuating combination of elements a corresponding diversity is required in the possible channels. This seems to be furnished by the network of the Neuroglia. See the representation copied from Butzke’s plate, and note how the cell-process blends with the meshes of the Neuroglia. Is it fanciful to regard this network of fibrils as having somewhat the relation of capillaries to blood-vessels? Did we not experimentally know that the capillaries are terminal blood-vessels, we should not suspect it from mere examination of the structure. 159. Having insisted that our knowledge is insufficient for any explanation of the “law of isolated conduction,” I can only suggest a path of research which may lead to some result. What we know is that some stimulations are propagated from one end of the cerebro-spinal axis to the other in definitely _restricted_ paths, while others are _irradiated_ along many paths. In the succeeding chapter this will be more fully considered; what we have here to note is that the manifold irradiations of a stimulation have an anatomical substratum in the manifold sub-divisions of the network of fibrils and the amorphous substance in which they penetrate. [Illustration: Fig. 26.--_Nerve-cells with processes terminating in neuroglia._] 160. In conclusion, I would say, let no one place a too great confidence in the reigning doctrines respecting the elementary structure of the nervous system, but accept every statement as a “working hypothesis” which has its value in so far as it links together verified facts, or suggests new research, but is wholly without value in so far as it is made a basis of deductions not otherwise verified. Hypotheses are indispensable to research, but they must be accompanied by vigilant scepticism. Imagination is only an enemy to Science when Scepticism is asleep. CHAPTER VIII. THE LAWS OF NERVOUS ACTIVITY.