2 I 2 
Agnes Robertson. 
feature of the proembryo nuclear divisions of Gymnosperms. In- 
Fig. 23 two nuclei remain in the pollen-tube, while in Fig. 22 
there are two nuclei in the upper part of the archegonium 
which have the appearance of having come from the pollen-tube. 
It seems not unlikely that in both cases these nuclei have resulted 
from the division of the second male nucleus. The next stage is 
the separation of the four daughter nuclei, and the initiation of 
wall formation between them, (Fig. 24). I have not found 
anything corresponding to the extensive free cell-formation which 
Arnoldi 1 has described as occurring in Cephalotaxus , and Jager 2 in 
Taxus. In the four-nucleate stage which I have drawn we see the 
second male nucleus, which is undivided, remains in the pollen- 
tube with a vegetative nucleus close to it. I have seen exactly the 
same state of things in another pollen-tube which was in contact 
with a multicellular proembryo. In the next stage of which I have 
a preparation the proembryo consists of seven cells, six of which 
form one tier, open above, while a single one is placed at the tip 
(Fig. 25). Further divisions result in the cutting off of rosette 
cells from the upper tier, and the formation of a cluster of cells at 
the tip instead of a single one. At first all the tiers are short as 
indicated in Fig 26 which is a proembryo of September 19th. In 
Fig. 29 the seeds of this date are shewn as they appear to the 
naked eye. Fig. 27 represents the condition exactly a month later, 
when the suspensors have begun to elongate in both directions, 
apparently demolishing the rosette cells in their upward growth. 
Fig. 28. is a diagrammatic longitudinal section of a seed at this- 
date. 
I can detect no starch in the unfertilised egg, the egg in which 
the nuclei are fusing, or the bi-nucleate proembryo, but it appears 
in considerable quantity at the four-nucleate stage. The sudden 
appearance of the starch is difficult to account for. Plastids must 
either have been present in the egg or have been brought in in the 
dense cytoplasm of the sperm cell. At the stage drawn in Fig 25 
the tip cell and the lower parts of the cells of the upper tier are 
packed with starch, while the grains in the upper part of the egg 
cell are few and scattered. Fig 26 shews a stage in which the 
entire proembryo, with the exception of the rosette cells is densely 
filled with starch grains; in the rosette cells comparatively few are 
to be found. When the suspensor tier has elongated, as in 
1 W. Arnoldi. Loc. cit. 
2 L. Jager. Loc. cit. 
