Electric nets have been used successfully for the development of control tools for tsetse flies for nearly 40 years [23, 29, 57], yet have been used little for mosquito research [27, 28]. Results presented here show that e-nets can be used to study the oviposition behaviour of malaria vectors. Importantly, it was found that reducing the voltage to prevent sparking doubled the catch, which confirms earlier findings by Torr and colleagues . It is uncertain whether it is the visual, acoustic or chemical cues associated with the sparking that reduces the catch. When a single e-net was used next to an artificial pond, similar numbers of mosquitoes were collected on both sides of the net indicating that the mosquitoes approached the target from both directions. In order to quantify the total number approaching an attractive source, such as a water body, a complete square of e-nets surrounding the water was found useful. Field tests need to evaluate the performance of the e-nets for studying gravid mosquitoes under open field conditions, especially during rainy seasons the normal periods of maximum malaria transmission. Previous work on host-seeking An. arabiensis has shown that e-nets covered with a small roof work well even when it rains .
Sticky boards proved to be a simple method for collecting mosquitoes that were stunned after colliding with the net and fell to the ground since they effectively retained specimens and protected them from predation by ants. However, it was found that a transparent film was also attractive to gravid mosquitoes, even when used as sole collection device without any e-nets and without a water source nearby. Adding an artificial pond behind the transparent film sticky board increased the number of females trapped on the board confirming that water vapour is a strong attractant for oviposition site seeking mosquitoes [40, 58, 59].
In search of an alternative collection material under e-nets, the black fibreglass gauze coated with insect glue proved as attractive to gravid mosquitoes as transparent film. Both surfaces were conspicuously shiny for the human eye compared to the yellow film that appeared matt and might act as a visual cue for gravid females. Previously, black flies of all physiological stages have been successfully trapped with glue coated aluminium plates [60–62] and in a recent study, Harris and colleagues  utilized this principle to collect gravid mosquitoes from water surfaces using glue-coated transparencies. Many insects, including mosquitoes, respond to reflectance of water surfaces to locate water bodies to lay their eggs, often using horizontally polarized light reflected from the water surface as orientation cues [64–69]. Surfaces with high polarized light reflectance might be promising as trapping devices alone or in combination with a gravid trap for monitoring African malaria vectors. Nevertheless, it can not be excluded that a chemical cue associated with the insect glue attracted the gravid females and there is need to further investigate the properties of the glue-coated surfaces used in this study.
The low number of mosquitoes on the yellow sticky film and the high number of eggs laid in the adjacent pond suggest that this material does not have the same visual properties for a mosquito as the transparent film and black glue boards and does not attract mosquitoes. Oviposition site-seeking females fly straight to the pond to lay their eggs, and then fly off again, without landing close to the aquatic habitat before or after egg-laying. This might be due to the light colour  and the lack of reflectance. It is unlikely that it has to do with the actual colour of the board since mosquitoes have dichromatic vision, which results in good contrast sensitivity but poor colour resolution . It is known that mosquitoes respond to contrasts [64, 70] and gravid females are attracted by dark surfaces rather than light coloured ones .
The number of mosquitoes collected with transparent sticky boards was approximately twice the number collected with yellow sticky boards. It is likely that transparent films overestimated the number of mosquitoes that approached the pond when they were used in combination with e-nets but sticky boards made of the yellow film can serve as effective collection device. On the other hand, the attractiveness of the boards mounted with transparent sticky film might be exploited further in future for the development of new trapping devices for gravid malaria vectors.
For the development of new interventions (e.g. auto-dissemination of larvicides [72–74]) and monitoring tools (e.g. ovi-traps and gravid traps [18–20, 38, 47–53]) targeting gravid malaria vectors it is important to know if and where gravid females land during oviposition. Notably, very few studies have investigated this and all these studies used relatively small cages (less than 1m3) except one which was implemented under field conditions . Gravid females were most commonly observed laying their eggs directly, either laying eggs when on the water surface or on the lip of the oviposition cup [59, 64, 75]. Occasionally, oviposition from flight has been described when the oviposition cup was placed over a black surface . Here, for the first time, experiments in large semi-field systems are described that investigate if and where An. gambiae s.s. lands to lay her eggs. The results indicate that gravid females primarily land directly on the water surface to lay eggs. Since no eggs were found in ponds with both detergent and sticky sides, which prevents directly egg-laying on the water surface, there is no evidence for eggs being dropped in flight onto the water from these experiments. The relatively large number of eggs found associated with females caught on the spray glue applied on the water surface was probably due to stress induced oviposition on the surface .
Similar numbers of eggs were laid in ponds treated with different sticky materials at their edges, though the number of adults caught on the edges differed, the number of adults caught there was small. This suggests that even mosquitoes caught at the edge might have landed there to rest before or after laying eggs, rather than to lay whilst seated on the edge of the pond. In the case of the pond with spray glue at the edge attraction of female mosquitoes cannot be excluded since the numbers were significantly higher than for the other two treatments and the glue made the pond edge appear very shiny.
Some caution must be exercised when interpreting the data since the artificial ponds used in this study had a sharp vertical edge which was not utilized by gravid females to sit on and lay eggs. This might have been different if ponds with a slope would have been used. Previous cage experiments have shown that An. gambiae s.s. and Anopheles arabiensis laid a large proportion of eggs on water saturated slopes rather than the free-standing water when given a choice [58, 75, 76]. Nevertheless, even then it was observed that these eggs were laid whilst on the water surface rather than during flight .
The finding of this study that An. gambiae s.s. lays its eggs directly on the water surface supports the observations made on An. arabiensis by Harris and colleagues  in the field using transparencies floating on the water on the edge of natural habitats. The finding that gravid An. gambiae s.s. lay their eggs directly on the water surface is encouraging for two reasons. Firstly, it lends support to the principle that gravid females could be used to transfer larvicides from a resting site to a breeding site [72–74, 77]. Secondly, it may lead to the development of a gravid trap where mosquitoes are attracted to a water source and trapped there .
Sticky materials and the detergent used in this study were shown to be useful methods for collecting mosquitoes when landing to lay eggs. Of all the tools tested the detergent and the spray glue directly applied to the water surface was most effective at collecting gravid females under semi-field conditions. Transparencies and sticky screens did not work as well which might be due to obstruction of water vapour coming off the pond by the transparency or due to visual obstruction of the water surface area. The latter two might have been useful tools for testing the attraction of female vectors towards a water source that was treated with putative oviposition semiochemicals or natural infusions  but due to their reduced trapping efficiency e-nets might be the best alternative for analysing such odour-oriented behaviour. Detergents and spray glue, though powerful in arresting approaching females, might interfere with the presented chemical or infusion. Therefore, further research would be required to present these in combination for attracting and trapping gravid female mosquitoes. The use of these tools under natural conditions also needs to be further evaluated.