The effect of attack by the Coreid bug Pseudotheraptus wayi Brown on the yield of coconut palms was studied in Coast Province, Tanganyika, in 1957–60. Alternate plots in a line of six one-acre plots, each of 50 palms, were sprayed weekly with a mixture containing 0·1 per cent. DDT (to control P. wayi) and 0·1 per cent. malathion (to prevent any increase in scale insects). To facilitate collection of the immature nuts (vidaka) shed by the palms, the plots were kept cleared of surface vegetation. Since this is not a normal practice in local coconut plantations, its effect on yield was checked by leaving the vegetation untouched on six similar, unsprayed, plots immediately adjacent to the first six. All vidaka falling on the latter were collected and recorded daily, and the numbers showing damage by P. wayi were noted. The ripe coconuts on all twelve plots were harvested by climbers, and then counted, at eight-week intervals. Since the nuts take about 12 months to ripen, yields in the first year were in effect those of untreated plots.

The yields from unsprayed plots showed that removing the vegetation was beneficial; the three cleared plots produced 490 and 1,159 more nuts per plot in the second and third years, respectively, than their adjacent, uncleared, partners, whereas in the first year they produced fewer. The mean annual yields from these three cleared but unsprayed plots were successively 817, 2,134 and 3,129 nuts per plot. The corresponding yields from the three cleared plots that were sprayed were 1,102, 3,310 and 2,992. Thus, protection against P. wayi appeared to increase the yield in the second year, but to have no effect in the third, although vidaka damaged by P. wayi were fewer on sprayed plots throughout.

A series of estimates of ‘ total production ’, defined as the sum of the vidaka shed in any period and the ripe nuts subsequently harvested, was obtained by adding the total vidaka collected in successive eight-week periods to the ripe nuts gathered in the harvest taken 11½ months after the middle of each period. These showed that the initially greater yield from sprayed plots came about because a greater proportion of the ‘ total production ’ was converted into ripe nuts, that is, fewer vidaka were shed, doubtless as a result of the diminished attack by P. wayi. The subsequent drop in yield to about the level of the unsprayed plots was due not to any substantial change in this proportion but mainly to a drop in ‘ total production ’.

It is inferred that the number of ripe nuts produced by the palm is subject to an upper limit, which is determined by agronomic factors and greatly exceeded by the number of immature nuts formed. The excess nuts may be shed at an early stage, either for physiological reasons or as a result of biotic factors such as insect attack. Thus there exists a compensating mechanism whereby yields may remain constant under different levels of insect attack and yet show immediate responses to changes in soil or climatic conditions. This could account for the fact that, in the area studied, clearing of plots produced an apparently permanent increase in yield, and insect control only a doubtful and temporary one. It is concluded that, under present conditions, increases in coconut yields are more likely to result from improved methods of cultivation than from pest control with insecticides.