1918] SINNOTT—FOOD RESERVE 171 
dant. There is evidently no impediment to thorough diffusion 
in such tissues, and conversion may take place far from any center 
of water conduction. 
Where the wall of the storage cell has become heavily lignified, 
however, even though it is provided with pits (which in such cases 
are usually very small), diffusion seems to be much impeded, and 
the reserve food remains unchanged throughout the winter. This 
is well shown by the terminal cell in the medullary ray of a typical 
starch tree, which cell, filled with starch and surrounded by its 
lignified wall, abuts directly upon a starchless, fat-filled cell of the 
cambial region. There are small pits in the wall between these 
two cells, but the wall nevertheless seems to serve as an effective 
barrier between them in preventing rapid diffusion. The same 
circumstance may often be noted where a ray cell touches a vessel. 
Here the wall between the two is provided with many large pits 
(fig. 1), so that communication must be easy. The ray cell in this 
case is usually without starch in the winter and generally contains 
some fat. Its neighbors at either end, however, are often full of 
starch but contain no fat. The tangential ray cell walls, although 
provided with small pits, seem here also to be permeable with 
difficulty. This leads us to believe that heavy lignification of the 
wall is a decided hindrance to the ease of diffusion between cells, 
and that the pits in such a case are for some reason, perhaps 
because of their very small size, unable to perform their normal 
functions. 
If the wall of the storage cell is less heavily lignified, or is thinner 
or more abundantly pitted, entrance of water is evidently easier, 
and we have noted that in these cases starch is more completely 
converted and fat is more common. In some instances fat may be 
limited to the cells directly adjacent to a vessel. In others it may 
extend to adjoining cells, starch occurring only in the more isolated 
regions. In these cases fat may be observed extending in from the 
cambial region along the rays for a considerable distance, thus 
indicating that diffusion takes place between the cells of the phloem 
region and the ray cells and affording a marked contrast to condi- 
tions at the ends of the rays in starch trees. In the true fat trees 
diffusion is evidently still more easy, for fat occurs throughout the 
