EXPERIMENTS WITH COCONUT PALMS 
9 
lections are made at brief or regular intervals. Collections were at 
somewhat variable intervals. Since an average interval of two and 
three-fourths months elapsed between pickings, there were four col- 
lections in some calendar years and five in others. For an analysis of 
the effect of fertilization on production, the harvests were divided 
into three periods, the first of 8, the second of 12, and the third of 
8 harvests. This arrangement placed one of the eliminated harvests 
between the first and second periods and the other within the second 
period. The first period extended from January, 1922, to August, 
Nuts 
Average of 
IZ harvests 
in Znd period 
Feb./92<t-Oct./926 
Average of 
8 harvests 
in 3rd period 
Average of 
d harvests 
in / st period 
Jan.l9ZZ-Auo.l9Zl Feb./92<t-Oct./926 Feb.l927-5ept./92o 
Figure 4. — Production per palm per picking at Corsica plantation 
1923, and the last period from February, 1927, to September, 1928, 
the seasonal range of the two periods being much the same. 
Since the first fertilizer application was in March, 1922, 17 months 
prior to the end of the first period, and since a coconut requires a 
year in which to mature, very little, if any, effect of the fertilizer 
would be expected to show in an increased number of nuts produced 
during the first period. The production during this period was 
used accordingly as a check on subsequent production. 
Figure 4 shows the production for the three periods, in nuts per 
palm per picking for each plat and for the 25 palms which gave the 
62849°— 31 2 
