The Spermatogenesis in Peutatoma up to the Formation of the Spermatid. 39 



form. The nucleolus is in direct contact with strands of chromatin, 

 and its surface is usually a point of apposition for these elements. 

 This apposition would show that the chromatin derives substance from 

 the nucleolus; for the mode of origin of the latter in the synapsis 

 shows that it is not a derivative of the chromatin. The nucleolus is 

 larger in the large than in the small generation of cells, which accords 

 with my assumption of its nutritive value. 



The cytoplasm is less dense than in preceding stages, and usually 

 vacuolated at the periphery. Such differences in the cytoplasmic 

 structure at different stages of the spermatogenesis probably cor- 

 respond to different states of metabolic activity. One or several yolk 

 globules frequently occur in the cytoplasm, these are mature yolk 

 globules and stain deeply, so that it would appear that the formation 

 of the yolk is completed before the rest stage is attained; in ovo- 

 genesis the formation of yolk usually continues through the latter 

 period also, so that it may be concluded that since in the spermato- 

 genesis only little yolk is produced, enough is formed in the synapsis 

 and telophase to serve for the subsequent nourishment of the cell. 



The idiozome {Id. Z, Figs. 95—100) now usually forms a cap 

 surrounding the nucleus on the greater part of its surface; the first 

 steps of this process were seen in the telophase. It usually does not 

 completely envelope the nucleus. Its substance has attained its 

 maximum amount, and is coarsely-granular in appearance, much 

 darker on preparations than the cytoplasm. It is more abundant 

 (absolutely rather than relatively) in cells of the larger generation. 

 According to the plane in which a cell is sectioned, this mass may 

 appear to form a cap around the nucleus, or a single mass at one 

 pole, or two masses at opposite ends of the nucleus; it usually but 

 not always lies close to the nuclear membrane. The granules com- 

 posing it vary in size and form, and it has no limiting membrane nor 

 separate zones, nor do its component granules show any radial ar- 

 rangement. With all the stains employed it may be readily distin- 

 guished from the cytoplasm. Since minute yolk globules are of frequent 

 occurrence in the idiozome, it is very difficult to make sure of the 

 presence of centrosomes in it. But in two cases at least (Figs. 95, 

 96) were found in each cell two small granules of exactly equal size, 

 lying in a clear space in the idiozome; and in one of these (Fig. 95) 

 a doubly-contoured line connected the granules. So these granules 

 very probably represent centrosomes, and the clear zone surrounding 

 them might correspond to the inner zone of the Amphibian idiozome. 



