EXPERIMENTS WITH COCONUT PALMS 
15 
by omitted harvests, whereas one was omitted within the third and 
two within the fifth period. Figure 6 shows the average produc- 
tion per palm for the plats as a unit for each period, the heavy shad- 
ing indicating the average when all the palms were included and the 
light shading when only producing palms were included. The pro- 
duction was fairly uniform for the first three periods, and the 
average harvest showed only a small proportion of nonproducing 
palms, whereas this proportion greatly increased in the last two 
periods. 
Since fertilizer applications were begun in June, 1912, and termi- 
nated in June, 1916, benefits resulting from the application would be 
expected to appear during the first or the second period, that is, 
between July, 1913, and June, 1917. The production of the plats as 
a unit during the second period, however, was lessthan that during 
the first period, and gained slightly during the third period, one to 
three vears after the final application. (Fig. 6.) Production thus 
s 
8 
7 
g 
5 
l 
n . 
July 
August September October November December Jonuory February March 
April 
Moy 
Figure 7. — Seasonal production curve for coconuts, as shown by harvest records at 
Boquillas plantation, 1913 to 1924 
failed to indicate any progressive benefit from the fertilizers as a 
whole. 
The yield of each plat for each period is given in Table 3 and is 
shown graphically in Figure 8. The diagrams of plat yields for the 
first and second periods closely resemble each other in sequence. 
If it be assumed that the increased production shown by the three 
plats which received complete fertilizer was due to the fertilizer, then 
the conclusion would naturally follow that the three plats each of 
which received two rather than three fertilizer elements were injured 
by the fertilizers received, since their production fell short of that 
of the check by as much as that of the others exceeded it. Since it 
is clearly incredible that of three fertilizing elements each combina- 
tion of two should prove to be injurious, whereas the three together 
should prove to be beneficial, the differences are without particular 
significance. 
A more detailed examination of the production throws light on 
the situation. If each plat is divided transversely into three sec- 
tions A, B, and C, containing, respectively, palms Nos. 1 to 8, 9 
to 16, and 17 to 24 in double rows, the records of productions for the 
first and the second period considered separately or together will 
