Gregory . — Spore- Formation in Leptosporangiate Ferns . 455 
average equal numbers of gametes of each kind, that is to say, a symmetrical 
result, without suspecting that this fact must correspond with some sym- 
metrical figure of distribution of the gametes in the cell-divisions by which 
they are produced V 
Rosenberg’s earlier observations upon the hybrid Drosera longifolia x 
D. rotundifolia , whose parents differ from one another in the number of 
chromosomes, gave somewhat indefinite and variable results 2 . The chromo- 
somes derived from the two parents do not apparently present any 
morphological differentiation by which their lineage can be recognized ; 
but, even in the absence of this, Rosenberg’s later results 3 agree entirely 
with the expectation based upon the theory of the individuality of the 
chromosomes and the union in pairs of paternal and maternal chromosomes 
in synapsis. 
The reduced number of chromosomes in D. longifolia is ten, in D. rotun- 
difolia twenty. In the somatic mitoses of the hybrid thirty (i. e. 10 + 20) 
chromosomes occur. In the pollen-mother-cells and in the embryo-sac 
mother-cell there always occur twenty chromosomes, of which ten are large, 
ellipsoidal, with a central constriction, and ten are smaller, without 
a constriction 4 . At metaphase the ten large chromosomes occupy the 
equator of the spindle, the small ones are distributed somewhat irregularly 
upon both sides of the equatorial plate. The large chromosomes divide 
and the daughter-chromosomes move towards opposite poles of the spindle, 
where they are enclosed by a nuclear membrane ; with them may be 
included any of the small chromosomes which lie near the poles, but some 
are usually left behind in the cytoplasm. 
Rosenberg suggests the following explanation. D. rotundifolia is 
represented in the hybrid by only ten chromosomes, D. longifolia by 
twenty. Each of the chromosomes derived from the parent D. rotundifolia 
fuses with one derived from the other parent, D. longifolia , giving rise 
to the ten double chromosomes. The remaining ten chromosomes of 
D. longifolia find no corresponding chromosomes of the other parent, and 
remain as the ten small chromosomes, which do not divide at the first 
division. Accepting the views of Strasburger and Guignard upon the 
division of the chromosomes by two longitudinal fissions, Rosenberg 
supposes that the fusion in synapsis has been lateral instead of end to end, 
as is usually the case. It perhaps seems more probable that further work 
in the light of Farmer and Moore’s interpretation of the changes undergone 
in the prophase of the heterotype division will reveal similar phenomena in 
the Drosera hybrid. 
On the hypothesis that the segregation of characters occurs at the 
reduction-division, we shall expect that the mitoses in a Mendelian 
1 Bateson (’02), p. 30 . 
3 Rosenberg (’04). 
1 Rosenberg (’03). 
4 Ibid. 
