In order to assess the potential effectiveness of DDT spray against mosquito larvae living under natural conditions and to determine the characteristics of the most effective form of spray, the variables considered were, area dosage of DDT, oil solutions compared with oil-in-water emulsions, concentration of DDT in the spray liquid, and drop size.Spraying was carried out over small (20 ft.×40 ft.) areas of canal harbouring larvae of Anopheles maculipennis var. messeae, and A. claviger. The canal was densely packed with vegetation. The central channel was overgrown with Potomogeton natans, so dense that in some regions there was more area of leaf than of open water surface, and in places there were large masses of floating algae (a coarse species of Spirogyra). At the edge of the canal near the bank there was a margin of reeds averaging about 4 ft. wide, and from 1 to 4 ft. in height.Spraying was carried out with a specially designed spinning disc atomiser giving drops of nearly uniform size and enabling drop-size and dosage delivered to be independently varied over a wide range. The range of DDT dosages investigated was from 4 g./acre to 80 g./acre, delivered in solutions or emulsions of approximately 0·5 per cent. and 5 per cent. concentration and drop sizes of 0·4 and 1 mm. diameter. Assessment of surface dosage in relation to delivered dosage, by spraying dyed solutions, showed that about 30 per cent. of the smaller drops were lost in the wind, but that there was no appreciable loss of the coarser spray.For 5 per cent. oil solutions of DDT drop size in the range 0·4 to 1 mm. diameter had no effect on the kill, except at low dosages (< 10 g. DDT/acre) when the greater wind loss of small drops produced variable results. Tests on the effect of drop size were not carried out with emulsions.From the point of view of larvicidal action, the experiments have not shown that is advantageous to distribute DDT in high dilution. 4·3 per cent. DDT oil solution sprayed at 10 g. DDT/acre actually gave a slightly higher mortality than 0·43 per cent, solution at the same dosage area of DDT, but the difference may not be significant owing to uncontrollable variables. Under the conditions of these trials, both oil solutions were superior to 0·43 per cent. AMSO emulsions. Experiments on the effect of concentration at constant area dosage were made only in the reed margin of the canal and data were not obtained for open water.In the centre of the canal, mortalities greater than 90 per cent, were obtained for both oil solutions and AMSO emulsions at delivered dosages greater than 10 g. DDT/acre, the oil solution being slightly superior. In the reed margin the kill dropped sharply with AMSO emulsions at dosages below 40 g. DDT/acre, the fall-off was much less marked with oil solution. From considerations of the average mortality in all areas, a delivered dosage of 40 g. DDT/acre (small drops) or estimated surface dosage of approximately 30 g. DDT/acre, appears the optimum for economical usage of DDT.When areas were sprayed the first two larval stages suffered a higher mortality compared with the later larval stages and pupae, and the proportion of pupae and later instars thus rose sharply.There was no evidence that residual lethal effects of oil films extended beyond three days. The initial kill in these trials was followed by a very low population level, due, it is thought, to a natural decline in the reproductive potential of the population.