No abstract

Considerable technological advances have been made towards the generation of genetically modified mosquitoes for vector control. In contrast, less progress has been made towards field evaluations of transformed mosquitoes which are critical for evaluating the success of, and hazards associated with, genetic modification. Oceanic islands have been highlighted as potentially the best locations for such trials. However, population genetic studies are necessary to verify isolation. Here, we used a panel of genetic markers to assess for evidence of genetic isolation of two oceanic island populations of the African malaria vector, Anopheles gambiae s.s. We found no evidence of isolation between the Bijagós archipelago and mainland Guinea-Bissau, despite separation by distances beyond the known dispersal capabilities of this taxon. Conversely, the Comoros Islands appear to be genetically isolated from the East African mainland, and thus represent a location worthy of further investigation for field trials. Based on assessments of gene flow within and between the Comoros islands, the island of Grande Comore was found to be genetically isolated from adjacent islands and also exhibited local population structure, indicating that it may be the most suitable site for trials with existing genetic modification technologies.

The vertical distribution of spider mites, 'Fe/rtrr~~c~lrir.s spp.. on cotton is largely deterniincd by the phenological stage of crop growth. Early i n the season, plants have few leaves a n d thc mites are located close to the mainstem terminal: during the rapid phase of vegetative Frowth. the mites arc located increasingly Farther from the terniinal: as vegetative growth decreases latc i n the se;ison. the mites are again found closer to the terniinal. The small size a n d potcntially large numbcrs of mites found o n cotton nialie conventional monitoring proccdurcs economically impractical. A presence-absence monitoring procedure is presented which relates the proportion of mite-infested niainstcni node leiives at the most infested node to the nuniber of niites per leaf which can in tiirii be converted to an area estimate. An optimized sampling prograni was dcvclopcd by incorporating the spider mite distribution inforination into a presence-absence sequential sampling plan. Sonic monitoring shortcut5 that result in incorrect treatment decisions arc discussed.

Diagnostic test, level III.