944 Davis.—A Comparison of the Redtiction Divisions of 
exhibited by the protoplasts, which in favourable material may be scarcely 
evident. 
As in the previous studies, iron-alum haematoxylin was employed as 
the stain throughout the investigation. 
Oenothera Lamarckiana. 
Oenothera Lamarckiana has been made the subject of a lengthy study 
by Geerts (’ 09 ), which considers the development and fertilization of the 
embryo sac as well as the formation, of the pollen. Gates’s (’ 08 ) account of 
the reduction phenomena during pollen formation in O. rubrinervis con¬ 
siders a type so close to Lamarckiana as to make possible a close comparison 
with the latter form. There are important differences between the writer’s 
conclusions and those of the investigators cited above in various details of 
the accounts of presynapsis, synapsis, and the form and behaviour of the 
chromosomes during the reduction mitoses. These differences become at 
once apparent in comparing our figures, and only certain of the most marked 
points of disagreement will be referred to in the text. 
Presynapsis. The daughter nuclei following the last mitosis in the 
archesporium of the anther (PI. LXXI, Fig. i) exhibit clearly the fourteen 
sporophytic (somatic) chromosomes characteristic of the type. The chro¬ 
mosomes become smaller as the nucleus, increasing in size, passes into the 
resting condition, the development of the nuclear reticulum apparently 
drawing material from the chromosomes and so reducing their bulk that 
they can finally only be recognized in later stages as deeply staining bodies 
(Fig. 2). These are the chromatic bodies described by the author for biennis 
and grandiflora, and there can be no doubt that they are derived from the 
chromosomes of the preceding mitosis ; they apparently correspond to the 
prochromosomes of Overton, Rosenberg, and others. 
In the earlier presynaptic stages the chromatic bodies may frequently 
be counted as fourteen, but as the reticulum becomes more pronounced they 
cannot be distinguished from other structures, such as small nucleoli and 
granules which appear in the nucleus. The chromatic bodies of Lamarckiana , 
like those of biennis and grandiflora , are distributed irregularly throughout 
the nuclear cavity, and there is no evidence of groupings in pairs which 
might be interpreted as indicating the presence in the nucleus of two 
parallel spiremes such as have been described by Overton (’ 09 ), Lundegard 
(’ 09 ), and Rosenberg (’ 09 #). On the contrary, their position on the linin 
strands suggests rather an end-to-end relationship in keeping with the 
arrangement of the chromosomes in the spireme preceding the heterotypic 
mitosis. 
The development of the nuclear reticulum makes it impossible to follow 
the chromatic bodies further, since their substance merges with the strands 
of the network and their form becomes lost. This is a gradual process 
