The spermatogenesis of Peripatus (Peripatopsis) balfouri. 303 



these or some of them subdivide still further (compare especially 

 Figs. 73, 77 — 81, 83, 84, 88, in these figures each granule has been 

 drawn with great care). Evidences of such secondary segmentation 

 may be seen in all stages of the synapsis period. At first it occurred 

 to me that this might be a phenomenon of particular morphological 

 importance, — a transverse splitting of each granule comparable in 

 value to the longitudinal splittiog which takes place immediately after. 

 But a careful study of these appearances has convinced me that this 

 segmentation, though it is in a certain sense a transverse division of 

 granules (i. e. a division at right angles to the long axis of the 

 chromosome), is in reality only a delayed continuation of the process 

 of decondensation of the chromatin into its component granules. This 

 is thus a process the reverse of the one by which the chromosome in 

 the preceding prophase had built itself up out of separate chromatin 

 granules; and it is very probable, though I cannot prove it, that the 

 chromosome in the synapsis breaks up into as many granules, as had 

 in the prophase come together to form the chromosome. Thus the 

 apparent subdivision of these granules is probably only a separation 

 of granules, which had before been closely apposed but not actually 

 united together. 



These granules resulting from this gradual decondensation of the 

 chromatin of the chromosomes vary considerably in volume, even in 

 the same nucleus, as the Figs. 76 — 81 show clearly. Each granule 

 has roughened contours which would show that it is composed of a 

 mass of smaller microsomes ; probably the number of such microsomes 

 would vary with the volume of the granule. The word "chromatin 

 microsome" has been used rather loosely by different writers, but I 

 would prefer to limit it to the smallest visible particles, and for 

 larger aggregations of such particles to use the term "granule". In 

 no periods of the synapsis and telophase stages do these granules 

 break up into their ultimate microsomes, and even in the following 

 rest stage no such subdivision seems to be fully carried out. 



As to the number of chromatin granules which make up a chromo- 

 some of the early synapsis. Cases were taken where a whole bi- 

 valent chromosome could be seen plainly on lateral view, and its 

 granules drawn carefully with the camera lucida, then counted. In 

 four bivalent chromosomes so studied the number of granules found 

 in each component univalent chromosome were: 8, 11; 7, 12; 8, 12; 

 8, 6. This count shows that the two univalent chromosomes which 

 form a bivalent one have not the same number of granules, which is 



