664 A MANUAL OF PHYSIOLOGY 



posterior root and their collaterals end by arborising around the 

 dendrites of the cells of the anterior horn. The excitation is, there- 

 fore, able to pass from the telodendrions of the posterior root fibres 

 through the dendrites of the anterior horn cells towards their cell 

 bodies, but not in the opposite direction, and in general the direction 

 of conduction is from the dendrites towards the cell-body. 



Varieties of Neurons. Nearly all the nerve-cells of the cerebro- 

 spinal axis agree with the cells of the anterior horn in the possession 

 of an axon and one or more dendrites, although sometimes the den- 

 drites are scanty in number and insignificant in size. In the cerebral 

 cortex the typical cells are of pyramidal shape. From the base 

 comes off the axon, and from the angles dendritic processes, a par- 

 ticularly massive dendrite proceeding from the apex of the pyramid 

 towards the surface of the brain. 



Sometimes an axon, instead of ending in an arborization which 

 comes into relation with the dendrites of another nerve-cell, or, as is 

 more frequently the case, with the dendrites of more than one cell, 



FIG. 230, 



Cells from the Gasserian ganglion of a developing guinea pig. The originally bipolar 

 cells are seen changing into cells apparently unipolar. The same process occurs in the 

 cells of the spinal ganglia. (Van Gehuchten.) 



breaks up into a sort of basket-work of fibrils surrounding the cell- 

 body. The cells of Purkinje, for instance, in the cerebellum, are sur- 

 rounded by such pericellular baskets (Fig. 232). The cells of the spinal 

 ganglia have two axons, which in the embryo arise one from each 

 end of the bipolar cell, but in the adult are connected to the cell by 

 a single process (Fig. 230). The great majority of them have no 

 dendrites, unless, as some have supposed, the peripheral process 

 really represents a dendrite. Another kind of cell which seems un- 

 doubtedly to be of nervous nature is the 'granule-cell.' Granule- 

 cells are much smaller than the nerve-cells we have been describing. 

 Their processes are much less easily followed, but all appear to give 

 off an axon and several dendrites. They contain a relatively large 

 nucleus (5 to 8 /A in diameter), with only a mere fringe of cell- 

 substance. The nucleus, unlike that of a large nerve-cell, stains 

 deeply with haematoxylin. Some parts of the grey matter are crowded 

 with these granule-cells, e.g., the nuclear layer of the cerebellum and 

 the substantia gelatinosa, or substance of Rolando, which caps the 

 posterior horn in the cord. In other parts they are more thinly 



