Analysis of the results obtained from suction-trap catches of the coffee leaf-miners, Leucoptera meyricki Ghesq. and L. caffeina Wshbn. on an unsprayed block of coffee at the Coffee Eesearch Station, Lyamungu, on the slopes of Kilimanjaro, are given.

By thermal summation methods it was possible to delimit successive generations of moths with some degree of accuracy and the generation catches were used for regression analysis with various meteorological factors. With L. meyricki, a negative correlation was shown between the rate of increase from one generation to the next and with both maximum and mean daily temperature, solar radiation and duration of sunshine taken as means over the period of each generation. In each case, the highest degree of correlation was found between the meteorological data in one generation and the rate of increase in the next.

When the generation catches were plotted on a logarithmic scale it was found that linear regression lines could be fitted to the periods of both rise and fall of the annual outbreak. With L. meyricki, the rate of increase of the rise was negatively correlated with, and the rate of decrease of the fall positively correlated with, the size of the generation present at the beginning of the outbreak. It is therefore possible, knowing the size of the minimum generation, to predict the size of the coming outbreak. No such relationship could be proved with L. caffeina, probably because the catches of the key generations were very small, but when the catches of the two species were pooled the same relationship held. It is probable that parasites are the main factor governing the annual fluctuations.