SEX-RATIOS IN GAMBUSIA HOLBROOKI. 187 



reticulum of inconspicuous fibrils; the chromatin flakes are 

 small and seen only with difficulty. 



The early Spermatogonial chromosomes in multiplication-divi- 

 sions in the cyst (Figs. 3-5) are simple rod-shaped elements, 

 with apparently terminal spindle-fiber attachment. In no case 

 was a J-, U-, or V-shaped chromosome observed in a polar view 

 of the plates. The chromosomes in the spermatogonial meta- 

 phase plates are usually so numerous and so densely massed that 

 a dependable count is almost precluded. The number, however, 

 is not less than thirty nor more than thirty-six; in a single 

 widely-spread plate 36 chromosomes were clearly seen. That 

 this is the diploid number is suggested by the fact that in the 

 aiakinesis stage preceding the first spermatocyte division, counts 

 of the tetrads give 15 and 18 as the number of quadrivalents. 

 The diploid number of chromosomes in the male germ-cells would 

 then be 36 if no heterochromosome were present, or 35 if one 

 heterochromosome were present. On the latter point it will not 

 be possible to speak definitely, since the II. -Spermatocyte plates 

 are always too crowded and dense to show the exact number 

 of chromosomes. It is apparent, however, that approximately 

 1 8 univalent chromosomes are present. Lagging chromosomes 

 are frequently seen in anaphase stages viewed from the side, 

 but not constantly enough to give a clue as to the number, size 

 and form of possible heterochromosomes. Occasionally preco- 

 cious chromosomes appear in early maturation-stages. If one 

 were to trust the uncertain help of analog}-, however, it might 

 be claimed that the male Gambusia has one heterochromosome. 

 (Cf. Schmidt, '20.) With this assumption, it would be quite 

 within the bounds of probability that the spermatogonial chromo- 

 some number is really 35, the plate which showed 36 merely 

 possessing a lagging chromosome belonging to the other daughter 

 cell. 



By repeated mitotic divisions the cysts become filled with the 

 final spermatogonia (Figs. S-io). The precise number of fis- 

 sions of the spermatogonium forming a cyst is not known : com- 

 putations indicate that at least 10, and often 12 fissions occur 

 before the cysts are ready for spermatogenesis. In a given cyst 

 the cells are all in the same stage of maturation and thus, as in 



