proposition in two recently published papers that the concentration of sugar in nectars of adapted to pollination tends to be unusually high compared with that for most insect-pollinated is contradicted, with evidence. Increased cross-pollination when hummingbirds are forced to visit many flowers, difficulties in rapidly ingesting viscous nectar, and need for water as a dietary article are suggested factors in keeping down nectar-sugar concentrations. VERY RECENTLY, in this journal, Percival (1974) has provided us with an invaluable compilation of data on the floral ecology of an area of coastal scrub in southeast Jamaica. In her paper, notable contributions are made of information concerning flower morphology, phenology, sugar rewards to visitors, seed set, and other features of the interaction between the plants and the birds and insects that visit their flowers. conclusions that she draws from this impressive mass of data are generally acceptable with one striking exception. It seems important to draw attention to this contested conclusion because it has significance for the presently developing quantitative approach to pollination biology that is being made by way of energetics (cf. Heinrich and Raven 1972). Percival (1974: 126) writes The more concentrated nectars attracted the more specialized pollinators, namely the hummingbirds and butterflies. Leaving aside the question whether hummingbirds and butterflies are more specialized pollinators than, for example, bees, her statement appears to be based on her Table 5 (p. 115) which shows that none of the nectars from hummingbird flowers for which she obtained sugar-concentration data contained less than 11% sugar (sucrose-equivalent, weight by total weight). However, neither did the nectars from any other kind of flower except those inhabited only by ants and thrips, and, unlike most flower types, the included no representatives with over 30% nectar-sugar. This circumstance strongly suggests that Percival's quoted deduction is wrong, at least as far as the are concerned. suggestion is confirmed by my calculation of means and ranges for each flowerpollination type from her data (table 1), a manner of expressing her data that, inexplicably, she does not use in her paper. It is clear that no statistically significant differences are to be seen between the sugar concentrations of the different flower pollination types in coastal southeast Jamaica, and, if anything, the occupy a belowaverage position. TABLE 1. Nectar-sugar concentrations of various flower pollination types in southeast Jamaica, calculated from data in Percival (1974). Pollination type Number of spp. Mean (%) Range (%) Hummingbird 9 21 13-30 Butterfly 13 22 13-44 Solitary bee 6 25 16-50 Wasp 8 24 13-49 Fly 5 20 16-23 Beetle Ant 25 20 5-50 Thrip 11 22 5-40 Honey bee 14 21 19-27