Harper . — C ell- Division in Sporangia and Asci. 51 1 
aster-rays. I am inclined to think this may have significance 
as indicating a resemblance in constitution between rays and 
membrane. It might be taken to indicate a connexion of the 
rays with the cell-surface such as is assumed, by Heidenhain 
and others, to exist between the astral rays and cell-boundary 
in the animal cell. It might be further assumed then that the 
rays give an anchorage for the spindle poles as claimed by 
Van Beneden and others, and that the daughter nuclei are 
drawn up to the walls in the position indicated above by the 
contraction of these rays during the early anaphases. However, 
such a connexion of the rays to the membrane is plainly 
lacking in the first and second divisions and in the divisions in 
the ascus of Erysiphe , as I have pointed out in my earlier 
paper. The spore-plasma at this stage again shows rather 
sharp boundaries against the foamy plasma above and below 
it, a condition which is obscured during the second and third 
divisions by the appearance of a considerable number of 
vacuoles in it. These vacuoles perhaps represent the nuclear 
sap set free by the disappearance of the nuclear membranes 
during division. Ultimately, however, this sap is either ab- 
sorbed or extruded into the foamy epiplasm, and at the period 
just before the beginning of the delimitation of the spores, 
the spore-plasma is quite dense and homogeneous. There is no 
indication at this stage of a tendency to aggregate in denser 
layers about the nucleus. Such aggregation as was described 
in my former paper 1 represents in reality a later stage after 
the plasma- membrane is established, although I was not able 
to demonstrate the membrane at this stage from the prepara- 
tions I then had. In Lachnea it is perfectly clear that the 
spore-boundaries are in no way outlined at the stage shown 
in Fig. 43. There is nothing to indicate that the nuclei are 
centres of any particular limited areas of cytoplasm about 
them. The further reconstruction of the daughter nuclei 
proceeds in the fashion described for Erysiphe. In Fig. 43 
we have a stage in which the chromatin has been distributed 
through the nucleus and a nucleolus has appeared. The 
1 Ber. d. D. Bot. Gesellsch., Bd. xiii, p. 67, 1895. 
