The coffee leaf-miner Leucoptera meyricki Ghesq. is the most serious insect pest of Coffea arabica in Kenya. Until recently, sprays containing parathion, methyl-parathion or diazinon have been recommended for its control.Field trials were carried out there during 1962 with these three insecticides and two others, fenthion and fenitrothion, that had shown promise in laboratory trials described in an earlier paper. The greater part of the investigation was carried out on the larval stage. Single-tree plots were used, arranged in randomised blocks, and sprays were in most cases applied by knapsack machines. Spraying in all trials was timed to coincide with the presence of large numbers of half-grown larvae. Suitable dates could be forecast because the generations of L. meyricki follow a denned cycle with a single stage dominant at any one time. Mortalities of larvae were assessed, four days after application of the insecticide, by dissecting freshly mined leaves selected at random. Where necessary, a spray of coarse droplets of a poison bait was applied before a trial in order to reduce populations of parasites and mortality due to them to a low level.At a concentration of 0·02 per cent., a spray of parathion gave a. significantly higher kill than one of methyl-parathion. Diazinon, fenthion and fenitrothion were available as emulsifiable concentrates containing 60, 55 and 50 per cent, active ingredient (w/v), respectively. These were tested at equivalent dilutions. At a dilution of 1¼ pints concentrate in 100 gal. water (affording concentrations of 0·094, 0·086 and 0·078 per cent, active ingredient, respectively), which was the volume per acre required to give full cover, fenitrothion gave a significantly higher kill than fenthion or diazinon, there being no significant difference between mortalities achieved by the latter two insecticides. The addition of a supplementary wetting agent to sprays of fenitrothion increased kill significantly, as did a reduction in spray volume with fenitrothion and parathion. The possible reasons for this are discussed.The field trials described generally confirm the results obtained in earlier laboratory tests and support the view that any insecticide which shows promise in the laboratory will also give good results in the field.A small-scale trial on potted seedling coffee plants was carried out to test by the dipping technique the ovicidal properties of concentrations of parathion, diazinon, fenthion and fenitrothion equal to one-half or two-thirds those necessary to give satisfactory field control of mining larvae. Mortality of eggs varied with their age, but in all tests with the first three insecticides it exceeded 90 per cent. In a comparison between fenthion and fenitrothion, the former was significantly more toxic than the latter, which killed only 38·1 and 82·1 per cent., respectively, of eggs ½ and 5½ days old.The correct timing of sprays against Leucoptera is very important and should follow about seven days after the peak of moth numbers.The insecticides used in these experiments are considered in the light of the characteristics, which are listed, that are desirable in those to be recommended for the control of Leucoptera on coffee. Those recommended, with rates (Ib. per acre), are parathion (0·4), fenitrothion (0·9) and fenthion (1·0). Diazinon and methyl-parathion are no longer recommended.