10 BULLETIN 958, U. S. DEPARTMENT OE AGRICULTURE. 
These maximum rates were reached during the week ended August 
31 in 1916 and September 6 in 1917, or after a growth period, from 
the time of planting, of 79 and 85 days, respectively. After these 
dates the gain per week gradually decreased. Since the first week 
in September marked the terminatiom of irrigation in both seasons, 
it is probable that the falling off in the rate of increase in weight per 
hill at such a comparatively early stage of the period of tuber devel- 
opment was due, in part at least, to the diminution of the moisture 
supply. There was no evidence of maturity of the plants at this 
time, the Rural New Yorker being a late variety which in this locality 
remains green and vigorous until killed by frost. In 1917 light frosts 
on the nights of October 7 and 8 killed the leaves of the plants on the 
experimental plats and a severe frost on October 1 1 completely killed 
the vines. The vines were also killed after the next to the last dig- 
ging in 1916, so that the last increase recorded in each season repre- 
sents the extent of tuber growth that had taken place after the death 
of the vines. 
An inspection of Table II shows that there was a gradual spread- 
ing out of the frequency distributions toward the high values as the 
season advanced until the digging of October 4, after which no 
important changes occurred. One of the most striking features 
brought out in this table is the ununiform rate of growth in the tuber 
population. The last two samples of the season showed that between 
11 and 12 per cent, or nearly one-eighth of the entir'e number of 
tubers, had failed to grow beyond the 25-gram class, the average 
weight of those remaining in this class being 8.2 grams. On the 
other hand, the data taken at each digging period show a rapid 
increase in the growth of the tubers of the larger sizes up to the 
time of the last harvest in September. These inequalities in growth 
are brought out more clearly in figure 7. It has previously been 
shown that practically all the tubers which grow to more than one- 
half inch in size set early in the season, probably at very nearly the 
same time; therefore, we must conclude that those which are small 
at the end of the season are of this relative size because of their 
failure to grow at a normal rate and not because they are of much 
more recent origin than the larger sizes. 
In Table III, which gives the weight of the tubers in each of the 
25-gram classes, the increases in size from week to week are shown 
by the gradual extension of the distribution and' a shifting of the 
mode toward the higher classes up to October 4, after which there 
was little change. The failure of part of the small tubers to grow 
sufficiently to exceed in weight the limits of the 25-gram class, pre- 
viously referred to, is further emphasized by this table, which shows 
very little change in the total weight of tubers in this class after 
September 20. The changes in the distribution with respect to 
weight resulting from tuber growth are shown in figure 8. 
