Mitochondrial DNA diversity was studied at four loci in six natural populations of the tsetse fly Glossina pallidipes from Zimbabwe, Mozambique, Kenya, and Ethiopia. Single-locus diversity varied from 0.39 at 12S to 0.65 at COII. A total of 32 haplotypes was found with a mean of 6.4 +/- 2.9 per locus. To study breeding structure, diversity at two loci, COII and 16S2, was evaluated in 18 populations sampled from an area of approximately 1,611,000 km2 and in three laboratory cultures. Twenty-six haplotypes were detected at the two loci and mean haplotype diversity over all natural populations was 0.63. A high degree of population subdivision was detected within and among the Ethiopian and Kenya populations. The Zimbabwe and Zambia populations showed much less variation and differentiation than the northern populations. A population in Mozambique showed high levels of haplotype variation and affinities closest to populations in eastern Kenya, some 1700 km to the north. Analysis of variance of haplotype frequencies showed that 51.5% of the total lay within populations, 13% among populations within five nested groups, and 35.5% among the five groups. Wright's FST was 0.485, Nei's GST was 0.33, and Weir and Cunningham's theta = 0.45. Ecological data show that G. pallidipes is highly vagile. The large amount of genetic differentiation may be explained by genetic drift that occurred in scattered, relict populations during the rinderpest panzootic of the late 19th and early 20th centuries.

Variation at mitochondrial and microsatellite loci was used to study the breeding and dispersal structure of Glossina morsitans centralis, in six natural populations from Botswana, the Caprivi Strip (Namibia), Zambia, and in a laboratory culture derived from Singida, Tanzania. Only seven mitochondrial haplotypes were found. Mean diversity averaged over the six natural populations was 0.216 +/- 0.085. The fixation index FST = 0.866 indicated a high degree of genetic differentiation among populations. Fifty-three alleles were detected among six microsatellite loci and six natural populations. Mean microsatellite diversity was 0.702 +/- 0.091. Depending on the estimating model used, fixation indices varied from 0.15 to 0.225 confirming that G. m. centralis populations are strongly subdivided. For all FST estimates, positive correlations were detected between pair-wise genetic distance measures and geographical distances. The difference in fixation indices estimated from mitochondrial or nuclear loci was explained by the greater sensitivity of mitochondrial genomes to genetic drift. Population differentiation can be explained by genetic drift and the subsequent recovery of extant populations from small, discontinuous populations. These data confirm genetically the collapse and retreat of G. m. centralis populations caused by the rinderpest epizootic of the late 19th and early 20th centuries.

To study the population structure of Glossina morsitans morsitans Westwood (Diptera: Glossinidae), polymerase chain reaction (PCR) and single-strand conformational polymorphism (SSCP) methods were used to estimate mitochondrial DNA diversity at four loci in six natural populations from Zambia, Zimbabwe and Mozambique, and in two laboratory cultures. The Zambian and Zimbabwean samples were from a single fly belt. Four alleles were recorded at 12S and 16S1, and five alleles at 16S2 and COI Nucleotide sequencing confirmed their singularities. Chi-square contingency tests showed that allele frequencies differed significantly among populations. Mean allele diversities in populations averaged over loci varied from 0.14 to 0.61. Little loss in haplotype diversity was detected in the laboratory cultures thereby indicating little inbreeding. Wright's fixation index F ST in the natural populations was 0.088 ± 0.016, the correlation of haplotypes within populations relative to correlations in the total. A function of its inverse allows an estimate of the mean equivalent number of females exchanged per population per generation, 5.2. No correlation was detected between pairwise genetic distance measures and geographical distances. Drift explains the high degree of differentiation.

Breeding structure of Glossina morsitans submorsitans Newstead was evaluated by using genetic markers in mitochondrial DNA where diversity was scored at two loci in five natural populations from The Gambia and two populations in Ethiopia (form ugandensis Vanderplank), countries separated by c. 5450 km. Twenty six haplotype combinations were found, of which 17 were shared among two or more populations. Nine haplotypes were found in The Gambia and 23 haplotypes in Ethiopia. There were 12 unique haplotypes. Only six haplotypes were shared between the two countries. Populations in The Gambia (h e = 0.26 ± 0.04) showed less than a third of the diversity of populations in Ethiopia (h e = 0.84 ± 0.03). This suggests recovery from an earlier reduction in population. In a nested analysis of molecular variance of haplotype frequencies, 65% of the variance was due to differences within populations, 34% to differences between populations grouped by country, and only 1% was due to differences among populations within countries. In terms of gene flow, the fixation index F ST = 0.35, which leads to an estimate by Wright's island model of less than one reproducing migrant per generation exchanged between the eastern and western submorsitans populations. Nei's genetic similarity measure showed a deep division between Gambian and Ethiopian populations.