172 BOTANICAL GAZETTE [AUGUST 
rays and parenchyma, even in regions remote from centers of 
conduction. . 
In the cases of Cornus, Viburnum, and others which we have 
noted, where the degree of lignification of the walls of the storage 
cells seems to be a factor which determines the character of the 
food reserve, it probably operates by rendering easy or difficult 
the diffusion of water into the cell. 
That the starch in starch fibers is never converted into fat is 
evidently due to the fact that there can be little or no communica- 
tion between them and the water-conducting elements. The: 
willows are illuminating in this connection. Here the phloem, 
cortex, and medullary rays contain abundant fat in the winter and 
very little starch, so that the willows have often been regarded as 
fat trees. In the last annual ring, however, there frequently occur 
large numbers of starch-filled fibers, so that some writers have 
included the willows among starch trees. These fibers are almost 
absolutely pitless. Diffusion of water among them must thus be a 
slow process, a fact which probably explains this persistence of 
starch here where it is lost elsewhere in the wood. 
This hypothesis, that the character of the food reserve in a cell 
is dependent primarily upon the ease with which water or substances 
carried by water can pass from cell to cell, thus makes more under- 
standable the various facts which have been observed as to the type 
and distribution of reserves in the stems of woody plants from 
' season to season, and is the contribution of anatomy to the problem 
under discussion. With this as a basis we may allow ourselves to 
speculate a little as to just what factors are operative in causing the 
seasonal changes in the food reserve. That change of temperature 
alone is quite insufficient to account for these is shown by the fact 
that they take place in plants kept over winter in the greenhouse. 
The writer has also observed their occurrence in trees growing in the 
frostless area of the Gulf states. The changes are doubtless due 
in the last analysis to the action of enzymes, presumably diastase 
and lipase. There are evidently two quite distinct series of pro- 
cesses, those concerned with changes in starch, and involving the 
action of diastase, and those concerned with changes in fats, involv- 
ing the action of lipase. 
