THE SPERMATOGENESIS OF AGALENA N.KVIA. 149 



numerous rod-like chromosomes which often overlap one another 

 and knowing also the possibility that all of the chromosomes 

 may not necessarily lie in the plane of a given section and may 

 therefore be counted twice, it seems to me unwise to place much 

 reliance upon the finding of an apparently unmated chromosome 

 in the telophase of the last spermatogonia. Again, the split and 

 unsplit odd chromosome of the rest stage and of the growth 

 period may be two univalent heterochromosomes or accessory 

 chromosomes before and after conjugation, as in Agalena. In 

 the primary spermatocytic division of Epeira the split, odd chro- 

 mosome which is carried to but one pole is probably the two 

 univalent accessory chromosomes and while Berry surmised that 

 the halves of this single element are separated in the secondary 

 spermatocytic division, it is possible that each half of the odd 

 chromosome, or, as I believe, each univalent element of the two 

 accessory chromosomes, splits lengthwise and is equally dis- 

 tributed to the opposite poles. While Berry finds a single chro- 

 matin mass in half of the spermatids, I find in Agalena two 

 accessory chromosomes in half of the spermatids. These, how- 

 ever, eventually fuse into a single mass, so there is no real 

 discrepancy here. Our main point of issue is with the origin of 

 the accessory chromosomes. If it be granted that they originate 

 from two spermatogonial chromosomes rather than from one, 

 then it is possible to interpret all of Berry's figures in such a way 

 that the odd chromosome of Epeira and the two accessory 

 chromosomes of Agalena will be seen to be nearly identical 

 in behavior and fate. The forms of the ordinary chromosomes 

 of Epeira in the prophase and metaphase of the first maturation 

 division are somewhat obscure in the figures but are described 

 as V's, rings, rods and crosses. Now one who is familiar with 

 the double V-shaped chromosomes in other forms can readily see 

 how they might appear as represented, especially if they are over- 

 stained or closely packed together. Furthermore, my own sec- 

 tions of Epeira show plainly that the definitive form of the chro- 

 mosomes in the equatorial plate of the first maturation division is 

 that of a double V. 



It may not be out of place to mention here the two chief errors 

 in my own previous work ('05). In the first place I failed to 



