Abstract: The malaria mosquito Anopheles gambiae Giles s.s. has a diet consisting of nectar meals taken from plants and blood from warm-blooded animals, particularly humans. Foraging theory predicts that diets include only those items that maximize energetic intake (a proxy for fitness), yet previous studies indicate that for mosquitoes specializing on human blood, such as the yellow fever mosquito, Aedes aegypti, and An. gambiae, sugar has a negative effect on their fitness. The objective of this dissertation was to find how the presence of sugar- and blood-hosts in the environment affect the foraging decisions made by An. gambiae mosquitoes, their reproductive success, and their potential to transmit malaria. The experiments reported on in this dissertation were conducted to study the mosquito in a more ("semi"--)natural environment than more commonly used laboratory cage environments, to more aptly reflect the energetic expenditures that come with meal-, mate-, and oviposition-site seeking behaviour in nature. Chapter 1 gives a description of the mesocosm, and develops the rationale for using such set-ups. The importance of male mosquitoes to population processes has long been overlooked, but their mating capability is strongly dependent upon access to sugar sources. Chapter 2 investigates the insemination rates of females per the presence or absence of environmental sugar; a matrix population model suggested that An. gambiae populations will not be viable in environments devoid of sugar sources, due to the hampering of male mating ability. Rather than maximizing energy intake and thereby directly increasing fitness, female sugar-feeding may reflect a behavioural constraint, e.g., finding and locating a swarm of males and selecting a mate may require energy best provided by a sugar meal, or sugar may be required to enable (blood- )host seeking. Thus, in Chapter 3 the sugar-feeding behaviour of females is placed in a behavioural context. Results indicated that females may take either a blood or a sugar meal as their first meal before mating and subsequently seek a blood meal. In Chapter 4 the opportunistic nature of the first meal of this species is investigated further, by altering nectar-bearing plant abundance and blood host accessibility, and assessing the meal choices of large- and small-bodied 1-d-old mosquitoes. With unrestricted access to a human throughout the night, blood was the preferred meal type. When access to a human was limited by the use of a bed net, a proportion of mosquitoes switched to sugar, and smaller (i.e., containing less energetic reserves) females were more likely to do so. When cohorts of An. gambiae were presented with various "sugar-poor" plants or "sugar-rich" plants, they experienced different, age-dependent, levels of mortality, and the plant community affected their human-biting rates. This resulted in mosquitoes in environments where sugar was less readily available having a higher vectorial capacity - a measure of their ability to propagate malaria (Chapter 5). Measures of fitness slightly favoured mosquitoes in sugar-rich rooms, supporting the notion of opportunistic sugar-feeding by this highly efficient vector of Plasmodium.