in Li Hum Mart agon : IL. Spermatogenesis . 195 
already (I, p. 451). It is clear that such observations are of 
importance, for if well established they will settle the question 
as to the natural or artificial character of the contraction we 
are dealing with. 
It has already been said that the nuclei of the pollen- 
mother-cells remain in the contracted condition for several 
days. It is therefore easy to find material of the right age. 
Anthers are chosen from buds which stand midway between 
the highest and lowest buds of such an inflorescence as that 
described on page 190. The only difficulty in their prepara- 
tion lies in cutting sufficiently thin sections from fresh 
material. This can generally be done in elder-pith with a dry 
razor. The sections were mounted in sap expressed from the 
perianth-leaves of the bud, and were covered with a glass slip. 
They were examined with a homogeneous immersion objective 
of 2 mm. focal length, N.A. 1.30, from Zeiss’ apochromatic 
series ; eye-pieces 4 and 6 were used. The Powell and 
Lealand condenser was lowered until the angle of the incident 
light — from an incandescent gas-burner— -was sufficiently 
narrow to give a clear outline to the unstained nucleus 1 . 
Notes were taken of the observations made in this way on 
three separate occasions, and outline drawings of the pollen- 
mother-cell and its nucleus were kept. Within the rather 
obscure wall of the p>ollen-mother-cell was seen a clear space, 
well outlined against the granular protoplasm filling the rest 
of the cell. One side of this space was occupied by a rather 
opaque spherical ball. 
It now became necessary to identify the features of this 
outline. On two mornings the section was roughly fixed and 
stained by Strasburger’s well-known method of allowing 
methyl-green dissolved in one per cent, acetic acid to run in 
under the cover-slip. This very imperfect fixing method 
produced deformation in the shape of the clear space, but the 
1 This is the arrangement of optical conditions which I have found to give the 
best results. When the aperture of the incident light was decreased by the use 
of an iris diaphragm, the definition of the image was less perfect than that obtained 
by lowering the condenser. 
P 
