518 University of California Publications in Zoology t v L - n 



portions throughout the greater part of its length, but leaving 

 one or both of the ends intact (pi. 26, fig. 12), a process which 

 is the reverse of the one previously described. Cells in which 

 the split is developing can readily be distinguished from those 

 in the amphitene stage, by the absence of polarization, by their 

 larger size, and by their position in the testes near the region -in 

 which the maturation divisions take place. They are separated 

 from the amphitene cells by the extensive and clearly defined 

 region of the polarized loops. 



This and the subsequent stages in amphibians have been 

 described so frequently since first pointed out by Fleming (1887) 

 that they need not be discussed in detail here. According to 

 Hermann (1889), Meves (1897), Eisen (1902), Janssens (1901, 

 1905), Kingsbury (1902), and other careful investigators, the 

 longitudinal halves of the loop are separated in the first matura- 

 tion division. This seems to be true in the case of Aneidcs. The 

 split loops shorten, thicken, and at the same time become twisted 

 to form what Janssens designates the streptotene stage (pi. 26, 

 fig. 13). They finally form heterotypical chromosomes, the halves 

 of which, corresponding to the halves of the split loops, pass to 

 opposite poles during the following anaphase. 



A number of spindles representing the first maturation div- 

 ision were examined and some drawn with the aid of the camera 

 lucida (pi. 26, fig. 14). The number of chromosomes was four- 

 teen, though fifteen were counted in two cases. In the latter, 

 however, from the position of the chromosomes it seems probable 

 that a chromosome had just divided and it is possible that four- 

 teen was the original number in all cases examined. Both sections 

 and entire cells were employed in making these counts, the latter 

 being found particularly valuable. While they are frequently 

 somewhat dense and often pressed out of shape, it is hardly 

 possible that any of the chromosomes could be lost without rup- 

 turing the cell membrane. In that case the loss would be detected. 



DISCUSSION AND CONCLUSIONS 



As previously mentioned, the usual number of chromosomes 

 found in the equatorial plate of the spermatogonial spindle is 



