60 Acton . — On the formation of sitgars 
Little further change takes place in the protoplasm of the 
cells after this till the opening of the flower. Figs. 5 and 6 
may be regarded as showing the structure of one of the cells 
at the period when the secretion is being most actively 
poured out. 
The nectariferous secretion first makes its appearance in 
the floral tube some time before the opening of the flower, 
but how long before seems to vary greatly with the rate of 
growth even in flowers of the same plant. After the flower 
is fully open further changes rapidly take place in the gland 
cells, the vacuoles increase largely in size, all the granules 
disappear, and oil-drops begin to be formed in the now 
diminished protoplasm (Figs. 5 and 6 ). Tannin was first 
observed in the vacuoles at this stage (by the iron and 
chromic acid reactions), and the cell-sap is more distinctly 
acid than in earlier stages. No starch-grains or solid matter 
giving a carbohydrate reaction with iodine, or chlor-zinc- 
iodine, could be detected in any of the stages, nor do any of 
the dextrins which colour with iodine (erythrodextrins) seem 
to be present. 
The gland-cells appear to contain sugar even after the 
withering of the perianth, but after tannin begins to be copiously 
formed, it is very difficult to draw certain inferences from the 
reaction with Fehling’s solution, as many varieties of tannin 
readily reduce the fluid. The subsequent changes in the cells 
were not followed out in detail, as they have no connection 
with the formation of sugar, but it may be stated here that, 
owing to the rapid increase in size of the septa after withering 
of the perianth, the epithem and epithelium cells split away 
from one another, and the cell-walls quickly lose their distinct 
outline, and undergo a mucilaginous degeneration. 
In the mature capsule the positions formerly occupied in the 
septa by glands are indicated by small cavities with a more or 
less ragged outline containing fragments of cellulose and 
mucilage attached to the lining cell-walls. 
Observations on the gland-cells in Noihoscordum bulgari - 
cum, LindL, Ornithogalum nutans , Lk., and Allium sp. ? 
