32 BULLETIN 56, HAWAII EXPERIMENT STATION 
of consideration. The leaves had the lowest osmotic pressure of any 
of the different parts of the plant. Reference to Table 7 shows that 
this was also true of the total sugars. 
DISCUSSION OF RESULTS 
The results obtained in this investigation lead to a number of 
practical conclusions. That the stem of the canna plant functions 
strongly in influencing the size of the offspring was confirmed by 
these results, which point, moreover, to the importance of the 
mature stem in storing starch within its own rootstock. The starch 
content of the rootstock is lowered when normal dying back of the 
leaves of the mature stem is unduly hastened by unfavorable weather. 
The fact that the immature stem fails to contribute much plant food 
to the rootstock and the mature stem continues to function at its 
maximum until it is well along in the mature stage emphasizes the 
importance of prolonging the life of the latter group. 
A comparison of stem growth at the station and in Waimea shows 
that at the station the stem maintains its maximum rate of sugar 
production for a short time only, whereas in Waimea it persists for 
a much longer period. Although the loose, loamy soil of Waimea is 
probably better adapted to a root crop than are the heavy soils of 
the station, the climate, which prolongs the life of the mature stem, 
is especially favorable to the crop. The use of such artificial aids as 
windbreaks to protect the stem from damage would seem to be war- 
ranted. 
The significance of the stem in maintaining the new growth of the 
hill is also brought out. Attempts at the station to renew growth 
on fields where the stems had been severely checked by drought re- 
sulted in a small crop of extremely stunted character, whereas at 
Waimea, where the stem growth had persisted through a long, dor- 
mant period, the new growth produced fairly good-sized rootstocks. 
Usually, the starch content of a rootstock of the immature stage 
is low and steadily increases until the rootstock is well along in the 
mature stage. This fact would seem to warrant carefully surveying 
the field before harvesting, which is ordinarily done at a definite age, 
say, 15 months. If a large part of the growth is in the immature 
stage, the crop should be allowed to grow for an additional period 
so as to increase the starch content of the young rootstocks. On the 
other hand, when a dormant period prevents new rootstocks from 
forming at, say, 14 months, nearly all the rootstocks will be in the 
mature stage, and the crop may as well be harvested at once because 
the new rootstocks will not attain any appreciable starch content for 
several additional months. Inability of the crop to make renewed 
growth with the production of rootstocks of fairly good size after the 
top growth has been destroyed would seem to be sufficient reason for 
harvesting, regardless of the age of the crop. Since, under normal 
conditions, a field of canna is irregularly but continuously growing, 
it can never be definitely stated when the entire hill is mature, and 
a criterion other than age should be used to determine the best time 
to harvest. 
Not only should the classification of the growth of a field be deter- 
mined before harvesting but the approximate starch content of the 
