264 VEGETABLE PHYSIOLOGY 
changed in appearance and become the first thickening 
layer of cellulose. The occurrence of the rows of granules 
frequently leads to the striated appearance which can be 
noticed on the walls of many fibres, particularly those of 
the bast of the fibro-vascular bundles. 
In all cases therefore the formation of cellulose can be 
traced to the self-decomposition of the protoplasm, though 
whether the granules are actually cellulose or an inter- 
mediate substance is still uncertain. 
A very similar phenomenon is observable in the forma- 
tion of starch grains. In this case, as we have seen, we may 
either have to deal with the general protoplasm of the cell, 
or, as is usual in reservoirs and in ordinary leaf parenchyma, 
with a definite plastid, either a chloroplast or a leucoplast. 
These structures, however, may be regarded as specially 
differentiated protoplasmic bodies. We have already 
discussed their behaviour and the formation of the starch 
grain by them. Building themselves up at the expense of 
sugar and probably of various nitrogenous compounds, 
either brought to them or remaining in their substance, 
they break down again to a certain extent, splitting 
off a quantity of starch, which is deposited in the interior 
of the plastid, sometimes at one point, sometimes at 
several. As the process goes on, successive lamine or 
shells of starch are continually deposited round the original 
grain or granule till the structure of the fully formed starch 
grain is reached. In this case the process is somewhat 
clearer than the corresponding one in that of cellulose, as 
there is little doubt that each shell is composed of starch 
at the moment of its deposition. 
The formation of starch is in these cases a secretion by 
the plastid, just as that of cellulose is a secretion by the 
protoplasm of the cell. The formation of the small starch 
grains by the general protoplasm of cells in which no 
plastid is present is of a similar character, though it is not 
so long continued and the formation of successive lamine 
does not take place. 
