164 THE CELL 



at the centre, and diminishes as the surface is approached. The 

 hilum is richest in water, whilst the superficial layer, bordering 

 on the protoplasm, is most dense in composition. To this cause 

 we can trace the fissures which occur in the hilum of the starch 

 grain as it dries, and which extend outward from it towards the 

 periphery (Nageli V. 17). 



As has been already mentioned, the starch grains of plants do 

 not, as a rule, arise directly in the protoplasm, but in certain 

 special differentiation products of it, the starch -forming corpuscles 

 (amyloplasts, and chlorophyll bodies). According to the investiga- 

 tions of Schimpfer (V. 27), the special variety of stratification 

 which occurs in the grain depends upon whether it is situated in 

 the interior or upon the surface of one of these corpuscles. In the 

 first case, the starch lamella? arrange themselves evenly around 

 the hilum since they receive equal accretions on every side from 

 the starch -forming corpuscle. In the second case, that portion of 

 the grain, which adjoins the free surface of the amyloplast, is 

 under less favourable conditions for growth, for the surface of the 

 grain, which is directed towards the centre of the starch-forming 

 corpuscle, acquires the most substance, and in consequence the 

 layers are thicker at this point, and grow gradually thinner as 

 they approach the opposite side. 



Hence the hilum, about which the layers are arranged, becomes 

 pushed further and further beyond the surface of the amyloplast, 

 assuming a more and more eccentric position in the stratification. 



That the starch grains grow by the deposition of new layers 

 upon the surface, that is by apposition, may be deduced from a 

 statement of Schimpfer's. He observed, that around the corroded 

 centres of starch grains whose surfaces had been dissolved away 

 new layers had been deposited. 



Strasburger is of opinion that starch grains may be occasionally 

 produced in the protoplasm itself, without the intervention of 

 special starch-forming corpuscles. He found them in the cells of 

 the medullary rays of Gcmiferre, during their early stages of 

 development, as minute granules, embedded in the strands of the 

 plasmic network. As they grew larger they were to be plainly 

 seen situated in the plasmic cavities. These cavities have highly 

 refracting walls, upon which microsomes are situated. 



One of the most remarkable of the internal plasmic products 

 is the nematocyst (Fig. 70), which functions in Gaelenterata as 

 a weapon of attack, in the cnidoblasts, which are distributed 



