462 BOTANICAL GAZETTE [December 



is 1 20 n by 80 ju, while the male nucleus is 45 li by 1 5 ju. The mem- 

 branes of the nuclei are resorbed at the place of contact, and the 

 contents of the male nucleus pass into the interior of the female 

 nucleus, thereby leaving a protuberance containing very little 

 cytoplasm. 



In Abies the chromatin could not be detected definitely until 

 the formation of two groups in the micropylar end of the egg 

 nucleus takes place. While the groups are still distinct, the indi- 

 vidual chromosomes become separate (fig. 28), each group con- 

 taining the haploid number of chromosomes. As the two spindles 

 unite, the chromosomes become paired (figs. 21-33); at first the 

 individuals of a pair approximate side by side (figs. 31, 32); soon 

 they twist about one another and jointly loop into the form of a C 

 (figs. 30, 32, 33). The chromosomes are very large, in some cases 

 exceeding 20 ix in length. There is abundant evidence that this is 

 a pairing, not a longitudinal splitting of chromatin elements. First, 

 the number of pairs is haploid. This is the number which would 

 necessarily result from a pairing of the double number of chromo- 

 somes already present; a splitting would, of course, result in 2X 

 pairs. Also, the twisting of the chromosomes about one another 

 is identical with their behavior in what is generally regarded as a 

 pairing during the prophase of the first reduction division. More- 

 over, there follows a transverse segmentation. If the diploid num- 

 ber of chromosomes should undergo two divisions, one longitudinal 

 and one transverse, an 8a - number of chromosomes would necessarily 

 result. The facts cannot be explained by the supposition that 

 there is a longitudinal split ; they are readily explained by regarding 

 this paired appearance as a true pairing. 



The segmentation. — The bending into C or V-shaped forms is 

 followed by a segmentation of each component of the pairs; in 

 other words, a transverse fission at the angle of the bent chromo- 

 somes (figs. 34, 40). The resulting 4X daughter segments are 

 approximately 10 n in length, or one-half the length of the pairs 

 before and during the looping process. At first the segments 

 remain more or less twisted about one another (figs. 34, 37, 41, 42), 

 and for some time retain a paired relation (figs. 43, 45). They may 

 be in the form of X's, or V's, or parallel rods. At the time when 



