McAllister—Cytology and Embryology. 
645 
chromosomes do not become immediately associated in pairs 
after fusion, in these forms, they gradually come to be so paired. 
These facts suggest a possible explanation of the appearance 
of blend hybrids and of hybrids which segregate according to 
Mendelian ratios. The blends may occur in those cases in 
which the parental chromosomes become associated early in the 
diploid generation and thus exert a mutual influence upon one 
another. The hybrids which segregate in the second generation 
may be those whose chromosomes do not become associated until 
synapsis. The association of the parental chromosomes is here 
so intimate that the mutual influence which they exert upon one 
another is far greater than that which results from the less inti¬ 
mate association of the chromosomes in th© somatic nuclei. 
SUMMARY. 
1. Evidence found in the pollen mother cells of Smilacina 
racemosa indicates that no preparation for reduction takes place 
in the last division preceding synapsis but that the last presynap- 
tic division is an ordinary somatic cell division. 
2. The observed phenomena during synapsis suggest very 
strongly that a side by side pairing and fusion of leptonema 
spirems occurs, and that the fusion is complete at the time of the 
recovery from synapsis. 
3. The second contraction stage which follows a period of 
uniformly distributed spirem results necessarily from the grad¬ 
ual contraction of the spirem which is attached to the nuclear 
membrane in places. 
4. There is no approximation of the limbs of loops to form 
the double heterotypic chromosomes. 
5. The double heterotypic chromosomes are formed by the 
transverse segmentation of a longitudinally split spirem, the 
line of the split of which probably represents the line of ap¬ 
proximation of the two parental spirems at the time of synapsis. 
6. The flrst division of the embryo-sac mother cell of S. 
racemosa results in two unequal cells, the inner being much the 
