Indoor-resting blood-fed Anopheles gambicie s.l. were collected from selected localities in Ethiopia. Anopheles arabiensis was the only sibling species identified. Paracentric chromosomal inversion polymorphisms were observed on the 2R and 3R arms. A significant excess of heterozygotes was obtained in samples from Siile carrying the 3Ra inversion. Similarly, significant variation was seen in the frequencies of the 3Ra inversion between the study localities. The observed intraspecific inversion polymorphisms relative to indoor resting behaviour are discussed.

Chromosomal inversion polymorphism was investigated in samples of Anopheles gambiae sensu lato collected from different localities of Tanzania. An. merus was found monomorphic, while two widespread inversion polymorphisms were recorded in both An. gambiae sensu stricto (2Rb/+ and 2La/+) and An. arabiensis (2Rb/+ and 3Ra/+). Two additional polymorphic inversions (2 Rbr and 2Rq), generally associated, were observed in An. arabiensis from the Rift Valley. Observed karyotype frequencies were in agreement with those expected according to the Hardy-Weinberg equilibrium suggesting conditions of panmixia in both An. gambiae s.s. and An. arabiensis in the studied areas.

A study of the mosquito fauna of the Mwea Rice Irrigation Scheme, Kenya, carried out during 1984 and 1985 identified 13 species, with Anopheles gambiae s. l. comprising 65.71% of the total collection. X-chromosome identification of 652 semi-gravid females resulted in 649 (99.5 %) positive identifications of An. arabiensis, implicating it as the noxious member of the An. gambiae complex present. Inversion polymorphism of chromosomes 2Rb and 3Ra karyotypes was studied in one village population. Seasonal population changes were found to be heavily dependent on rice field surface water. Artificial irrigation for rice farming greatly extended the breeding period per year by linking up the two wet seasons. Adult females fed more frequently on bovids than humans. The mean percentage gonotrophic condition of all females caught in the villagesof Karima, Thiba, Mucii-wa-Urata and Rurumi by pyrethrum spray-sheet collections, miniature light-traps and exit window traps was 69 % bloodfed, 18 % unfed and 13 % gravid. These proportions were statistically shown to be within predictable limits over time and space. The sporozoite infection rate in adult females was 0.55% by manual dissections and 1.2% by the ELISA technique.

Between May 1984 and December 1985, the population of immature stages of Anopheles arabiensis was sampled by the dip method at sites close to two villages in Mwea Irrigation Scheme, Kenya. Larval presence corresponded to rainy seasons and the flooding phase of the rice cycle. No larvae were found during the postharvest dry season. The instar composition, age distribution and survivorship curves for the different stages were determined. Pupal productivity for Mwea rice fields was estimated to be 1% of the total immature population.

Electrophoretic variation was determined at eight enzyme loci in Anopheles arabiensis from four villages in Mwea Irrigation Scheme, Kenya. Seven loci had polymorphic alíeles of which at least two were common. Pooled allele frequencies at the Pgm, β-Had and Odh loci were in Hardy-Weinberg equilibrium, while those at the Ao, Idh, Adh, Est-1 and 6-Pgd loci showed significant deviations. The frequency of heterozygotes at the Adh, Est-1 and 6-Pgd loci was less than expected, while at the Ao locus, there was an excess. A 2 × 4 contingency χ2-test for each of the loci not in Hardy-Weinberg equilibrium indicated an association between the respective gene frequencies and the villages. The suggested explanation for these observations is that each village constitutes a distinct population, and that the pooled data introduced the Wahlund effect. Allele frequencies at individual loci per village population are most likely at Hardy-Weinberg equilibrium. Larger samples per village need to be examined to facilitate a goodness of fit χ2-test between observed and expected frequencies.