A method of estimating the damage inflicted on the developing fruit of the coconut palm by Pseudotheraptus wayi Brown is needed in order to assess the efficacy of measures for the control of this Coreid. The insect lives in the crown of the palm, populations are small in relation to the damage they cause, the adults take wing readily and the nymphs are agile and secretive. Direct estimation of populations by hand-collection, trapping, the use of knockdown sprays and marking-recapture methods, using paints or radioactive materials, were all unsuccessful, and indirect methods of estimation were therefore studied.

The growth, flowering and fruiting habits of the coconut palm are described. There is a single growing point, at which the leaves are produced in succession. From 12 to 18 inflorescences are formed each year; each is borne in a leaf axil and carries male flowers distally and female flowers proximally; the latter open, and are fertilised, almost together, after all the former have fallen off, and some 3–4½ weeks after the inflorescence has emerged from its spadix. All subsequent stages in the development of the fruiting body are dated from the latter event.

The young nut at first elongates rapidly and in the first six months attains over 80 per cent. of its final length; it is ripe at about a year.

In the absence of damage by insects, from 20 to 80 per cent. of the fruiting bodies are shed as nutlets, 1–2 weeks after fertilisation, the proportion depending on the variety of the palm and being less in those that bear few flowers and large nuts, but very few are shed at any other stage of their development. When an adult or a late-instar nymph of P. wayi feeds upon a developing fruiting body a characteristic lesion develops at the site of the puncture, which is concealed by the calyx, and if the fruiting body is less than three months old at the time (in which case it is locally termed a kidaka, plural vidaka) it is normally shed 4–7 days later. As the nut develops, it becomes less likely to be shed following attack by P. wayi, although the lesion formed may be invaded by fungi or by the weevil, Diocalandra frumenti (F.). Fruiting bodies attacked by the latter insect cease growth but are not shed. Experiments and field observations show that where damage by P. wayi is prevalent, a proportion of the vidaka are shed in the pre-fertilisation stage, even though they show no evidence of attack by the insect. The extent and implications of this phenomenon are not yet understood.

Fallen vidaka rot quickly and those that are found will reflect the damage that was inflicted 7–10 days previously. By collecting a sample of freshly fallen vidaka, removing the calyces and determining the percentage of them that shows damage by P. wayi, a measure is obtained that is termed the Vidaka Damage Rate (V.D.R.). The method advocated is to collect at random and at weekly intervals a sample of at least 100 vidaka per ten acres, excluding any with an overall length exceeding three inches. This size is reached at about 2½ months; older fallen nutlets invariably show damage from P. wayi or other causes. No significant differences have been found between values of the V.D.R. based on such collections and those based on collections from below individual palms, or rows of palms, taken at random. No significant difference was found between vidaka in the pre-fertilisation stage and those in the stage at which normal nutfall occurs as regards the proportion showing damage by P. wayi, and the former are therefore not excluded.

Comparison of the numbers of nymphs of P. wayi found by two trained entomologists at weekly intervals on 12 palms in each of two areas, one of which was treated with an insecticidal spray at monthly intervals and the other left untreated, and the V.D.R. determined a week later in the same area but on a larger sample of palms, showed a curvilinear association between the two quantities; the association between the logarithms of the quantities was linear but the correlation was not quite significant. In more detailed investigations, two 100-acre areas were used, one of which was undergoing insecticidal treatment, and in each a large but variable number of palms (from 100 to over 900) was examined during each week over a period from mid-October 1956 to the end of March 1957 by a team of climbers who had been trained to search for P. wayi. There was less variability in the data from the control than from the treated area; the latter showed a curvilinear relationship between the observed population of nymphs of P. wayi and the V.D.R. from the whole area one week later, but the correlation between the logarithms of the two variables was not significant.

If counts of first- and second-instar nymphs, which cause very little damage, were excluded from the data, together with any week in which no individuals were recorded, and the values of the V.D.R. were computed from that part of the area in which the insects were counted, then, when both variables were converted to logarithms, the transformed data showed a straight-line relationship with a highly significant correlation coefficient of −0·975. The regression equation indicates that where the number of palms that it would be necessary to search in order to find one 3rd- to 5th-instar nymph is 10, 100 or 1,000, the corresponding value of the V.D.R. would be 84, 13 or 2·4, respectively. It is concluded that where the V.D.R. is below 20, the population density giving rise to such a rate of damage is so low that its accurate determination by direct counts is unlikely to be possible.

Data on V.D.R. and yield were collected from a number of experimental plots and control areas; values from the latter were too uniform to give rise to significant correlations, but when a quarterly estimate of the V.D.R. in an experimental area, obtained from the totals of damaged and of undamaged fallen vidaka collected during successive three-month periods, was compared with the figure for yield obtained from harvesting the same area during the corresponding quarter the following year, a linear relationship was found between yield and V.D.R., the coefficient of correlation being significant in each of the five areas considered, and ranging from −0·7 to −0·984, and the values of a and b in the regression equation Y = a + bX, where Y is the yield in nuts per palm per harvesting and- X is the quarterly value of the V.D.R., ranged from 13·65 to 19·8, and −0·141 to −0·287, respectively. It is concluded that the V.D.R. is a good measure of expected yield, provided that the rate does not rise unduly in the interval between assessment and harvest; in such circumstances, the increased population of P. wayi that is reflected in the increased V.D.R. may damage the nuts at a later stage of their development and so reduce the yield below that expected.