The aim of this study was to assess the costs of resistance associated with the homozygous state of the G119S mutation based on life-history traits of An. gambiae in an insecticide-free laboratory environment.
No significant differences in larval mortality were found for the two strains (Table 2a, b). However during pupation the mortality of the resistant strain was greater than that of the sensitive strain (Table 2c). When data from each experiment and food treatment were combined, ~ 31% of all AcerKis pupae died during pupation (265/868), compared with 20% for the Kisumu strain (162/802). This represents a substantial cost for resistant individuals as it directly decreases the proportion of individuals able to contribute to the next generation. Furthermore, the difference between the two strains was seen across food treatments and included the higher food conditions where pupal mortality was lowest (Figure 1). This indicates this cost is likely to be expressed even in environments favourable for mosquito development.
Previous laboratory studies have found increased levels of pre-adult mortality for C. quinquefasciatus mosquitoes bearing the G119S mutation [19–21]. However in these studies larval and pupal mortality were not reported separately. A re-inspection of the data in  found no significant difference in mortality for pupae of the insensitive acetylcholinesterase strain and the sensitive strain (P. Agnew, pers. comm.), whereas substantial pupal mortality was observed in  for the insensitive acetylcholinesterase strains (M. Weill, pers. comm.).
AChE has been identified as having a role during insect development . For example, its activity has been reported as maximal in the pre-pupal stage of Drosophila melanogaster  and at adult emergence for Apis mellifera queens . Hence, the pupal mortality observed for the AcerKis strain may be due to its modified AChE causing disruption to developmental processes during metamorphosis. Further data are required to verify if this is a general effect extending to other mosquitoes and/or in which conditions it is most likely to be expressed.
Comparing the frequency of the G119S mutation in fourth instar larvae and adults emerging from the same sites would provide a means to test whether pupal mortality is a cost experienced in natural mosquito populations. Furthermore, such data should be refined to examine the relative frequencies of homozygous and heterozygous individuals in the respective populations. This would provide useful information as to the relative strength of selection acting on mosquitoes with one or two copies of the allele. Based on what is already known from Culex mosquitoes , it can be anticipated that there will be less pupal mortality for heterozygous individuals and their mortality will vary relatively more from site to site than for homozygous individuals as the dominance of the mutation is environmentally variable .
The fitness of resistant and sensitive mosquitoes not only depends on their respective chances of reaching adulthood but also on the reproductive success they achieve as adults. Two important life-history traits influencing reproductive success are age and size at maturity. In these experiments AcerKis individuals tended to reach pupation earlier than those of the Kisumu strain (Table 3a), but they also emerged as smaller adults (Table 3b). The possibility that AcerKis adults were smaller because they pupated earlier can be discounted as the difference in the weights of the strains remained when individuals from each strain were matched for ages at pupation (mean [AcerKis - Kisumu] = -5.438 μg, t = 5.895, d.f. = 125, p < 0.001; comparison of mean dry weights of each strain when matched for age at pupation, experiment, food treatment, replicate food group within experiment, and sex).
Earlier ages at pupation indicate potential fitness benefits for the resistant strain. For example, shorter developmental times could lead to reduced generation times and reduced risk of larval mortality due to predation or their aquatic environment drying out. The reproductive success of resistant adult males could also benefit from earlier emergence if it increases their encounter rate with previously unmated females. In contrast, the smaller adult size of resistant individuals indicates potential fitness costs for resistant mosquitoes. For example, smaller females are likely to be less fecund and to experience greater risks of mortality while blood-feeding due to the need to feed more frequently. The influence of adult size on the reproductive success of Anopheles males is less clear, varying considerably among studies [37–39].
Previous studies involving Culex mosquitoes have produced mixed results concerning age and size at maturity. Later ages at pupation and smaller adult size were found for resistant individuals in a field study involving C. pipiens . An individual's resistance status did not influence either trait in a laboratory study involving Culex quinquefasciatus , but in different experimental conditions resistant individuals were found to be smaller . Age and size at maturity were not reported, but the mating success of resistant male C. quinquefasciatus was less than that of sensitive males when both compete for females .
The above studies do not show a clear pattern for the effects of the G119S mutation on the age and size at maturity of mosquitoes. They also suggest any effects it has on these traits are likely to be obscured by variation in environmental conditions experienced during larval growth.
The G119S mutation in its homozygous state was not found to influence the time that adult mosquitoes could survive by metabolising reserves accumulated during larval development (Table 3c, Figure 4). This suggests metabolic costs are not particularly associated with this form of insecticide resistance, whereas such costs are important in other forms of insecticide resistance, e.g., esterase overproduction .