50 CALIFORNIA ACADEMY OF SCIENCES. 



they may originate. The evidence on which I base my 

 belief that they arise by this last method is threefold : 

 First, the cells of the test are wholly unlike the cells of 

 the endothelial lining of the stolonic vessels, even at the 

 tips of the vessels where the endothelium is thickened 

 because of the growth that is there taking place. Both 

 the endothelial cells and the test cells are shown in figs. 

 30 and 31, ed. v. The former invariabl}' contain large, 

 clear nuclei, usually spherical in regions where growth is 

 taking place, but sometimes slightl}^ flattened in the plane 

 of the membrane to which they belong. These nuclei 

 are from 4 ,". to 6 ;'■. in diameter. Each contains one 

 large distinct nucleolus. The cell protoplasm, which is 

 in considerable quantity, stains more distinctly than does 

 the nuclear matter. There is no cell membrane, and the 

 cells are irregularly stellate in form, this form being de- 

 termined in part, no doubt, by mutual contact, though in 

 preserved specimens they do not appear closely crowded ; 

 in fact (fig. 31, pi. iii) there are often seen irregular 

 spaces between them. This is probably due to slight 

 shrinkage. 



The nuclei of the test cells are, on an average, about 

 one-half the size of those just described, i.e., 2^ //. in 

 diameter. (Figs. 29, 30 and 31, in. c, 111. r.' and m. c") 

 A nucleolus can usually, though by no means always, be 

 detected; but nearly the whole body of the nucleus stains 

 deeply, so that the nucleolus is never seen in the midst of 

 a large clear space, as in the case of the nuclei of the 

 endothelium. In by far the greater number of instances 

 the cell-body is not seen at all, excepting by the greatest 

 care and with the aid of high powers and favorable light. 

 When recognizable, it is sometimes disposed in a uniform 

 layer around the nucleus, giving the cell, as a whole, an 

 approximatel}' spherical form ; but, more commonly, one 



