There is a danger of seeing too much in a nerve-cell when examining it under the highest powers of the microscope, and of endeavouring to picture in too much detail the arrangements which regulate the flow of impulses. Its markings are so complicated as to suggest to the mind of the observer that it is itself a microcosm—a nervous system in miniature. Neuro-fibrillæ appear to offer many alternative paths within the cell. It is unlikely that such a way of looking at the unit of structure is the right one. A certain motor cell of the spinal cord is connected by its axon with thirty or forty separate muscle-fibres; but there is no reason for thinking that the fibres ever contract save as a single group. The axon consists of parallel fibrillæ, but these do not appear to be needed as separate conductors; an impulse travels down the fascicle. It does not appear to be necessary in the case of a motor cell, and presumably the statement holds good for the large cells of the cerebellum and cerebrum to picture any arrangement for the simultaneous conduction in its axon of several impulses, or for the conduction of one impulse along one of its fibrillæ and a different one along another. What is necessary is that this particular efferent path Z should be accessible from every other part of the nervous system—from A to Y. If, merely for the sake of filling the space which would otherwise be blank in the mental picture, we imagine a pericellular net connected by thorns with the body and dendrites of the nerve-cell Z, then the net is the meeting-ground of all the routes through which Z is called into action. A nerve-wave from any of the neurones A to Y, breaking upon this net, passes along the thorns into the protoplasm of Z. In size a granule of the cerebellum presents a marked contrast to a motor cell of the spinal cord; yet it is formed on essentially the same plan. From its minute round body (about 8 µ in diameter) four or five slender dendritic processes are drawn out. Each dendrite ends in a little bunch of twigs, resembling fingers curved over the palm. Its single slender axon runs towards the surface of the cortex. As the granules lie at a considerable depth, this course is, for those which distribute to the most superficial layers, a long one. They pass from the granular to the molecular layer between the big cells of Purkinje. When the axon has reached a certain level in the molecular layer, it divides into two threads which run for a great distance, right and left.