Due to the role of mosquitoes in disease transmission and their impact on human well-being through their biting behaviour, both commercial and scientific interest exists for efficient trapping devices. During the last century, a number of different mosquito traps and collection methods were developed (reviewed by  and ). Recently, variations of the CDC light trap , the OBET [4, 5] and Mbita trap [6–8], electric nets  and different traps featuring counterflow geometry [10–14], have been used to evaluate the attractiveness of various complex host odours, individual volatile organic compounds, or mixtures thereof.
While full body odour was often successfully used to attract mosquitoes [15–17], synthetic baits were developed to improve ease of use and to allow standardizing the attractant. As mosquito host-seeking behaviour is governed by semiochemicals, baits can contain a number of chemical attractants, e.g. 1-octen-3-ol, lactic acid, ammonia, [13, 18, 19] and means to increase humidity and temperature.
For Anopheles gambiae sensu stricto (henceforth termed An. gambiae), there is currently no combination of trapping device and bait available that can successfully compete with the human landing catch (HLC) as the standard method for population surveillance in the field . Due to the possible exposure of field workers to infectious mosquito bites, cost, and tediousness of the HLC, this method poses both ethical and logistical problems (reviewed by ).
Recently, it was shown that rapid testing of candidate odour baits is possible in semi-field systems [21, 22]. The partly controlled environment helps to yield statistically powerful results quickly in advance of full field evaluation, and it increases the potential to characterize mosquito responses to traps.
While there is no consensus on the exact role of CO2 in the behaviour of An. gambiae sensu lato, this compound is frequently used in trapping systems [16, 17, 23]. Anopheles gambiae responds strongly to combinations of human odour and CO2  or human foot odour and CO2 [21, 24]. This robust synergistic effect makes CO2 an important constituent of odour mixtures, although the practical value is limited due to technical and logistical problems under rural conditions .
Plant-derived essential oils can be used as mosquito repellents [25, 26]. The repellency of plants themselves was surveyed during ethnobotanical studies in western Kenya by Seyoum et al , where traditional usage included direct burning of plant material and placement of repellent plants within houses. Initial experiments were conducted under semi-field conditions and later confirmed in field studies, where both potted plants and direct burning of Corymbia citriodora, Ocimum kilimandscharicum and Ocimum suave had a repellent effect. Using the latter method, the effect was comparable to commercial mosquito coils.
In this study, the trapping efficiency of two counterflow trap designs, the MM-X and BGS, was evaluated under semi-field conditions in Ifakara, Tanzania. Experimental baits included human foot odour and combinations of human foot odour with either carbon dioxide or lemongrass leaves.