EXPERIMENTS WITH COCONUT PALMS 7 
1920, in the high-production group 17 of 25 palms came into produc- 
tion, whereas in the low-production group only 5 of 25 palms did so. 
In 1921 all palms in the former group fruited, whereas in the latter 
group seven palms did not fruit until the following year. 
As has been stated, the original plats were crossed at right angles 
by new plats beginning in May, 1926, to test the effect of salt on 
yield of nuts. The palms which bore the numerals 3, 4, 5, or 6 in 
unit's place constituted the salted plat, and those the numerals of 
which ended in 0, 1, 8, or 9 constituted the check. The 36 palms in 
the former averaged 262.8 nuts per palm in four years, whereas 
the 35 palms in the latter averaged 260.8 nuts in the same period, 
which was a difference of only 0.8 per cent. This very uniform 
average production indicated the plats as units to be fairly com- 
parable. In 1927, the earliest possible date at which any effect of the 
treatment on production could be expected to appear, the salted 
palms produced an average of 63 nuts per palm and the check palms 
/S20 
I sit 
I9ZZ 
192.5 
!9Z<t- 
/S25 
/326 
/S27 
Figure 3. — Average nut production per palm, San Jose" plantation, 1920-1927, of 
the 25 most productive and of the 25 least productive palms, the former in solid, 
the latter in shaded hlocks 
61.9. In comparison with the yields of the preceding year these were 
reductions of 5.5 and 11 nuts per palm, respectively. In the half 
year 1928, during which the yields were recorded, the salted palms 
averaged 37.3 nuts per palm, whereas the check palms averaged 32.9 
nuts, which was a difference of 4.4 nuts, or 13.4 per cent in favor of 
the salted palms. Unfortunately, the hurricane terminated the 
experiment in September, 1928. In the brief duration of the experi- 
ment the recorded production indicated a possible benefit from the 
application of salt, but the period was too brief to warrant drawing 
conclusions. 
