Pollen Development in Lactuca. 387 
such as 15 and 36, are particularly convincing, and the original preparations 
still more so, as showing that the lateral pairing of threads is a progressive 
process beginning at one end of the threads at one side of the nucleus, and 
proceeding towards the other end, the double threads so formed becoming 
at the same time polarized in their arrangement. When this process is 
partly completed, the filaments may be seen to be lying side by side in the 
paired portion, but diverging widely in the unpaired portion of their length. 
This would scarcely be the case if the process were simply one of the 
approximation of the two halves of a single chromosome. 
Agar’s (1911) figures of the lepto-zygotene and the zygo-pachytene 
stages in the spermatogenesis of Lepidosiren are equally clear and convincing, 
for they show the lateral pairing of filaments taking place in precisely the 
same way. It is probable that the polarization of threads in the bouquet 
stage is associated with the presence of a centrosome. In higher plants, on 
the contrary, in the absence of a centrosome, there is no such polarization, 1 
. and it is possible that this may be associated with the telosynaptic method 
of pairing. 
As the primary purpose of this paper was not a study of synapsis, 
reference will be made to only a few of the recent papers on this subject. 
In plants, the most important recent paper is that by Digby (1919) on 
Osmunda. In a very careful investigation of the meiotic and premeiotic 
divisions in this plant, she emphasizes the fact that the somatic chromo- 
somes in the premeiotic mitoses undergo a split in telophase as a first 
stage in the process of alveolization, and that in the reverse series of 
prophase events each chromosome appears to be formed by the re- 
approximation and fusion of these two split halves or threads. The double 
thread which every one has observed in the postsynaptic stages of the 
heterotypic prophase in Osmunda is the crux of the matter for the 
telosynaptic or parasynaptic interpretation. She interprets it as the same 
doubling which appears in premeiotic prophases, i. e. as the approximation 
of the two halves of a univalent spireme, and not the pairing of two univalent 
filaments. Strong and critical evidence is brought in support of this view, 
but it unfortunately, probably owing to the nature of the material, lacks 
one crucial fact which v/ould be absolutely determinative : the number of 
loops or strands before and after this stage has not been determined, and is 
probably not possible of determination owing to their arrangement. 
The Gordian knot of interpretation might be cut, as Hogben (1920) 
has done, by assuming constant differences between plants and animals in 
their meiotic processes. But if one compares Miss Digby’s figures (1919, 
PL IX, Figs. 47, 48) with Gregoire’s (1907) earlier study of Osmunda 
(Figs. 24, 25) on the one hand, and with Janssens’s (1905, PI. Ill, Fig. 15) on 
1 The slight indication of polarization of the threads at one transient stage in Osmunda may 
represent an intermediate condition. 
