268 LILLIE. [VOL. I. 



In the description' of the centrosomes I shall follow the first 

 method entirely. See Figs. 3-5, radial sections through the 

 entire central spindle, and "jAE, horizontal sections through 

 the outer centrosome and spheres drawn on a much larger 

 scale. 



The prophase of the second maturation spindle is inaugurated 

 by the origin of a new set of radiations around the sphere (Fig. 

 3). Within the sphere are two centrosomes united by delicate 

 threads, the beginnings of the central spindle. Each centro- 

 some is composed of several granules. The radiations are 

 generally attached to the sphere, but they can sometimes be 

 traced in part to the centrosomes. 



As the spindle elongates, the sphere becomes elliptical, as 

 though stretched by the central spindle, the ends of which abut 

 against it (Fig. 4). The centrosomes have become still more 

 subdivided, and have increased in bulk. The fibers of the 

 central spindle have also increased greatly in number and dis- 

 tinctness. The rays surrounding the sphere are beginning to 

 disappear. They are, plainly, inserted in the elliptical sphere. 

 This is the stage which MacFarland has figured in such detail 

 for Diaulula. He calls the whole sphere the centrosome, be- 

 cause the rays are inserted in it. The history of this body, 

 which was not sufficiently investigated by MacFarland, makes 

 it plain that it is not the centrosome, but the inner splierc. 

 The insertion of rays is therefore in itself no criterion of a 

 centrosome. 



In a slightly later stage the rays have been entirely resolved 

 into vesicular cytoplasm ; and the sphere has been stretched 

 out into the peripheral fibers of the central spindle. The cen- 

 trosomes are yet more subdivided (Fig. ?A), and the mantle 

 fibers are just beginning to form. 



Now follow some very important centrosome phenomena ; 

 when the spindle has elongated a little more the radiations of 

 the asters begin to develop, those of the inner aster before 

 those of the outer in the radial position of the spindle. In a 

 radial section through such a stage, one is inclined at first to 

 think that the inner centrosome has disappeared. More care- 

 ful observation shows, however (Fig. 5), that the center of 



