EXPERIMENTS WITH COCONUT PALMS 23 
The important fact clearly brought out by the data was the wide 
variation in production between individual palms. This held true 
even for palms of the same age, receiving the same cultural and fer- 
tilizer treatment, and grown under apparently very uniform soil 
conditions. Such differences could be ascribed only to inherent 
differences in the palms themselves. 
Ten palms produced three times as many nuts as were produced 
by 10 others in the youngest group of palms. Twenty-five palms, 
somewhat more than one-quarter of the group, produced more than 
twice as many nuts as were produced by 25 other palms. In the 
intermediate-age group, 25 palms, one quarter of the group, pro- 
duced nearly three times as many nuts as were produced by another 
quarter. In the older group, the 50 most productive palms, constitut- 
ing one-sixth of the whole, produced 4.7 times as many nuts as were 
produced by the 50 least productive palms. 
In the youngest group, although no highly productive palms 
were found in the untreated plat, the distribution of palms of low 
and high yield through the other plats was such as to indicate that 
the wide variation in individual productivity was due mainly to 
some factor far more potent than the fertilizer treatment. In the 
older groups the distribution among the plats of the high and 
low yielding palms showed a tendency to produce much or little 
unrelated to the kind of fertilizer applied. In the intermediate-age 
group, although the low-production palms showed a decided in- 
crease in yield under fertilization, in contrast with the absence of 
any similar response to fertilization on the part of the high-produc- 
tion palms, the former continued to maintain their relative positions 
as low yielders in comparison with the other palms. Fertilization 
as employed in the experiments failed to transform low-production 
palms into palms of high production. 
The tendency to produce much or little becomes evident early in 
the life of the palm. The high yielders are likely to mature early 
and so come into production ahead of the low yielders. A long 
record is unnecessary for a comparative classification of palms as to 
productivity. Production recorded through several years should 
suffice for a general classification, and extremes of high and low pro- 
duction should be indicated in an even briefer period. 
The pronounced, inherent differences in productivity between 
palms indicate the line along which lies the greatest promise of im- 
provement in coconut production. No longer should any chance 
nut whatever serve for planting. Xor should the selection of seed 
nuts from the best palms be sufficient. The flowers from which 
such nuts developed may easily have been fertilized by pollen from 
the poorest drone tree in the plantation. Xature has provided 
against self-fertilization in the coconut. Man should see to it that 
for propagation purposes only the best palms, considered in respect 
to both quality and quantity production, serve as pollen parents and 
mother plants, so that each nut planted may be of pedigreed stock 
and carry a double inheritance of desirable characteristics. The 
results would repay the effort manyfold. 
U. S. GOVERNMENT PRINTING OFFICE: 1931 
