The Cicadellid, Empoasca lybica de Berg, is an important pest of cotton in the Sudan Gezira, where over 300,000 acres of cotton are grown annually under irrigation. Cotton is sown in mid-August, and the plants are uprooted and burnt the following May. The life-cycle of E. lybica from egg to gravid adult takes 16–24 days, and the adults live for up to 40 days. There is no diapause. During the 100 days from late August to early December when breeding on cotton is of economic importance, a single male and female could give rise to some 50,000 progeny.

During May to July, when crops are confined to irrigated gardens and river banks, E. lybica is widely distributed in such places and can be found also on tree hosts, which are numerous especially in the southern Gezira and along river banks. There is circumstantial evidence of displacement over long distances, and the great majority of catches of E. lybica in sticky traps were made before the increase in population on cotton that occurs from September onwards.

Of the 53 species of host-plants that have been recorded, only Solanum dubium, Rhynchosia memnonia, Hibiscus spp. and Abutilon spp. are of importance in the ecology of E. lybica. The first two especially are common weeds in fallows, which comprise more than half the land under rotation. Populations of E. lybica in Gezira fallows at the time of cotton germination tended to be greatest where pre-sowing rains (i.e., those falling from 1st July to 15th August) were highest. Correspondingly, initial infestation of cotton was highest in seasons and places receiving the most pre-sowing rains, although density of infestation in any place was affected by sowing date and proximity to irrigated fields and gardens which supported weed host-plants.

In order to develop a system of sampling for infestations of E. lybica in the cotton crop, the distribution of nymphs on cotton plants was examined. It was found that nymphs were most numerous in the leafiest zones of the plant and a random choice of leaves seemed an appropriate means of sampling for infestation. The distribution of nymphs within and between cotton fields was also investigated and a standard sampling procedure adopted.

Peak infestations on cotton could not be predicted from the level of initial colonisation, or from surveys a month later. Peak infestations were usually inversely related to the level of initial colonisation, especially when comparisons were made between seasons, as at the Gezira Research Farm. That is to say, high levels of initial infestation, which occurred in seasons of good pre-sowing rains, tended to be followed by low rates of increase, and in years of poor pre-sowing rains, initial infestations tended to be low and rates of increase high.

The relationship of these findings to those of Cowland & Hanna (1950) and Hanna (1950) are discussed; the hypothesis that pre-sowing mud-splash is a major factor controlling numbers of E. lybica in the Sudan Gezira is discounted, although it is accepted that this factor temporarily reduces populations.

The rate of increase of infestations of E. lybica was found to be positively correlated with the concentration of nitrogen recorded 2–4 weeks previously in the cotton leaf. This concentration affected not only the rate of increase of the initial colonisers, but also the rate of recovery of populations during November and December after spray-applications of DDT. The nitrogen concentration in the leaf was increased by nitrogenous fertiliser, with a corresponding increase in infestations of E. lybica. It was also found to be negatively correlated with pre-sowing rains, which, if low, prevent the nitrate in the top 12 in. of Gezira soil being washed to lower levels, but the data presented provide no evidence that the relationship is causal.

It is concluded that localities and seasons of poor pre-sowing rains favour a high rate of increase of small populations of E. lybica because of high nitrogen concentration in cotton leaves during September and October. This tendency is augmented by application of nitrogenous fertiliser. A regression equation relating the peak infestations of E. lybica with pre-sowing rainfall and with nitrogenous fertiliser is given and the infestations computed from this are shown not to differ significantly from those recorded in the Gezira as a whole, and in the four main divisions of it separately, during the eight years 1949–1956.