GERM CELLS OF GRYLLOTALPA 303 



and arrangement here is very much as in the first maturation 

 division (plate 2, A, anaphase second maturation division). 

 Again the chromosomes reach the end of the spindle before the 

 mitochondrial threads divide. Plate 2, B, is a late telophase of 

 the second division. Here the mitochondria have reached the 

 ends of the spindle and lie near the reconstructed nucleus. It 

 will be noticed that there is no indication of the remains of a 

 sphere at either end of the spindle. In fact, I believe in this case, 

 at any rate, that the spindle and spheres completely disappear. 

 Plate 2, C, is a spermatid shortly after the second division, and 

 there are no indications of any such remains. I shall discuss 

 this fact in connection with the formation of the acrosome. 

 The spermatid now transforms into a spermatozoon, and it 

 remains to follow the mitochondria in this transformation. 



Plate 2, D and E, show an early stage, but here the mito- 

 chondrial mass remains very much the same as at the end of the 

 second division. Two dark bodies, however, are present at this 

 time. One of them is concerned in the formation of the acro- 

 some, but they will be discussed in another place. The first 

 change in the mitochondria is shown in plate 2, F. The granular 

 appearance of the threads is entirely lost and they become 

 arranged into a fantastic design which looks almost too dia- 

 grammatic to be true. The body which forms the acrosome has 

 moved toward the opposite pole of the nucleus, or at least, the 

 relative position of this body and the mitochondria has changed. 

 In plate 2, G, the threads seem to form a series of concentric 

 circles, with another thread running through the middle. This 

 is the first indication of the separation into two halves. The 

 axis cylinder has made its appearance at this time and is entirely 

 separate from the mitochondria. Its point of origin is a small 

 granule in contact with the nuclear wall. Whether this granule 

 is a centrosome, I cannot say. The two halves of the mito- 

 chondrial mass now begin to elongate (plate 2, H). The inner 

 circles lose their identity more or less as transformation pro- 

 ceeds. The axis cylinder, which has grown in length, now lies 

 along the median line which separates the two halves of mito- 

 chondria (plate 2, I). The single granule which was the point 



