84 BOTANICAL GAZETTE [FEBRUARY 
intermediate in position between the two parents, though no fractioni- 
zation of any characters was observed. 
In killing material various fixing fluids were used, the most 
satisfactory being 1 per cent. chromacetic, a chromacetic solution made 
up as follows: 1 per cent. chromic acid, 70°°; glacial acetic acid, 0.5°°; 
water, 30°°; and the following chromosmacetic formula: 1 per cent. 
chromic acid, 70°; glacial acetic acid, 0.5°°; 2 per cent. osmic acid, 
oP", Water, 30°. 
Sections were cut varying in thickness from 3 to top. The 
stains producing the best results were found to be Haidenhain’s 
iron-alum-haematoxylin with orange G as a plasma stain, and Flem- 
ming’s triple stain. The former is most satisfactory for the study of 
karyokinetic figures, particularly the chromosomes, but safranin and 
gentian violet give equally good results for synapsis stages. 
Conditions of degeneration of pollen and tapetum in anther 
development 
Unless otherwise stated, this account will refer to conditions 
observed in O. Jata. The morphology of the Lamarckiana hybrid is 
very similar, so that the description would apply equally well to the 
latter, except where the pollen abortion in O. data causes abnormalities. 
Fig. 1 shows a transverse section of a single loculus in the “resting” 
pollen mother cell stage. The epidermal cells when normally devel- 
oped have thick, highly cutinized walls at this early stage. Occa- 
sionally, when failure of anther development sets in at an earlier stage 
this cutinization does not take place, the cells remaining thin-walled. 
There are several “middle layers,”? generally about four. These 
are narrow radially and often show the effects of pressure from the 
large cells of the tapetum. The latter forms a well-organized layer 
of cells enclosing the sporogenous tissue. The granular appearance 
of the tapetal cells, owing to deeply-staining inclusions which fill the 
cytoplasm, easily distinguishes them from the middle layers. There 
are usually one to three longitudinal rows of pollen mother cells in a 
loculus, but the rows are not regular and a cross-section may show 
as many as five mother cells. In the great majority of cases the 
2 There is some variation in the use of this term. It is here applied to all the 
layers between the epidermis and the tapetum, because they are undifferentiated and 
all have the same developmental history. 
