Several trapping techniques have been deployed in sampling malaria vectors population in the world 1 2 3 . Human landing catch has been considered as the gold standard method in mosquito sampling for surveillances and control programmes to estimate the infectivity rates, species abundance and mosquitoes dynamics 4 . With the increase of ethical issues in using humans as subject in collection of disease vectors, alternative simple methods, such as odour-baited traps, need to be developed and evaluated for their effectiveness 5 . Host odours provide olfactory cues by which haematophagous insects locate host for their blood meal 6 . Mosquitoes are the most important disease vectors in sub-Saharan Africa 7 . Odour-baited traps have shown to be effective to different disease vectors in tropical areas, including tsetse flies 8 9 , mosquitoes 10 and ticks 11 . Several studies have shown the attractiveness of malaria vectors to different hosts 12 13 . In the Anopheles gambiae complex, studies have shown different odour preference among the sibling species: Anopheles gambiae s.s. is attracted by semiochemicals from humans sweat 14 , Anopheles arabiensis notably varies from being attracted by humans or bovine odour, depending on the geographical location 15 . In several arid parts of sub-Saharan Africa, An. arabiensis, is a major malaria vector 16 . Anopheles arabiensis is regarded as more exophilic than other sibling species of the An. gambiae complex 15 . Several trapping tools, baited with odour, have shown great efficiency when evaluated for mosquitoes sampling in different areas, such as the mosquito magnet^® 17 and the MMX-trap 10 . In rural areas, there is a need to develop a trapping system, which uses the host-seeking strategies of the vectors and can be incorporated in simple tools/devices for vectors sampling 18 19 . The exploitation of both ecology and behavioural aspects of vectors are important in reducing the vector-human contact by developing targets or odour-based sampling tool 8 19 .

The purpose of this paper was to report the findings of evaluating resting boxes baited with cow odour against human landing catch in irrigation schemes of lower Moshi, northern Tanzania.

A total number of 9,558 of Anopheles gambiae s.l were collected. Of these 42.5% (n = 4060) by BBU, 24.9% (n = 2386) by HLC, 17.5% (n = 1668) were collected by BBCO, and 15.1% (n = 1444) by UB. All An. gambiae s.l collected in this area were regarded as An. arabiensis 13 . HLC did better in mosquitoes sampling than UB and BBCO (Figure 2).

In abdominal conditions, the percentage of unfed collected were 48.3%, 97.0%, 41.7% and 4.8%, fed were 0.0%, 3.0%, 16.2% and 4.6% semi-gravid were 22.7%, 0.0%, 0.0% and 92.0%, gravid were 29.0%, 0.0%, 42.1% and 0.6% for BBU, HLC, BBCO and UB respectively.

In evaluation of these resting boxes, a total of 128 (32 days) experiments were done. In analysis two days (eight experiments) were not considered for analysis due to delay in mosquito collection in the morning (i.e. mosquitoes were collected at 07:30 am instead of 06:00 am). General linear model univariate analysis method was used, position of the houses had no effect on mosquito density catch (DF = 3, F = 0.753, P = 0.387); the sampling techniques had significant impact on the caught mosquito densities as shown in Figure 2. The effect of days was not significant in the model (DF = 29, F = 5.095, P = 0.08). Mosquito density fluctuation between sampling tools found to be significant (X² = 50.806, DF = 29, P = 0.007) by the Kruskal-Wallis test.

The results of this study have demonstrated the usefulness of using simple tools for sampling disease vectors in surveillance and control programmes in disease endemic areas. These findings have been supported by previous studies of using different tools as observed in other studies 20 21 22 . BBU performed better than traditional method of human landing catch, the mostly known sampling method believed to perform well in malaria vector sampling in all positions 4 . The use of BBU in surveillance studies will give better results in areas with a higher population of An. arabiensis than HLC, hence protecting human from been bitten by infected mosquitoes during sampling exercise. These results have given an insight in proposing the need of re-evaluating the mosquito sampling method according to ecological factors or species abundance if already known. These boxes were able to collect mosquitoes outdoor without interfering with norms and values of the community that made the method to be more useful in the community. These boxes are freely obtained and affordable for surveillance studies in rural areas where modern trapping systems are limited. The complex natural animal odours used are available and not costly.

The sampling of mosquitoes was not affected by other factor except trapping technique (Figure 2). This shows that in area with An. arabiensis, animal odours can be used to replace human landing catch technique, which was not powerful enough (Figure 3). Other methods of mosquito sampling, such as the pyrethrum spray catch (PSC) which samples indoor resting mosquitoes, needs well trained personnel, expensive instruments and entering peoples house fortnightly 4 . Light traps (LT) that sample host-seeking mosquitoes need electrical power to operate the traps that might not be available. These PSC and LT are performed indoors while An. arabiensis mosquitoes are more exophilic, exophagic and zoophilic 15 ; hence PSC and LT might not perform well in areas with a high proportion of An. arabiensis. This study shows that cattle urine odour baited resting boxes had better results than traditional method (HLC) in collection of mosquitoes per day for 30 days (Figure 3).

This comparative study of HLC and odour-baited resting boxes enabled direct comparison of collection methods for An. arabiensis alternative to HLC. The main advantage of the odour-baited resting boxes is that it samples host-seeking mosquitoes and protects human from the risk of being bitten by infected mosquitoes 23 . The preference of cattle odour by An. arabiensis has been documented elsewhere 24 . It has also been recorded that the odour-baited traps promotes species-specific responses 13 , which was further documented in this study (Figures 2 and 3).

This study showed similarity in attraction between BBCO and BBU in the proportions of unfed mosquitoes, which means that odour from any parts of the body can play the same role in surveillance studies for collection of An. arabiensis (Figure 2). In these sampling techniques, the high proportion of unfed mosquitoes was found in HLC, fed in BBCO, semigravid in UB and gravid in BBCO. This trend still convincing that, these techniques created specific abdominal status attraction 13 . The attraction of BBU to high mosquito density in different abdominal conditions should be observed critically to show the components responsible for such attraction, looking for example at what happens when urine ages. This method may be incorporated in surveillance studies for identifying the infectivity rates of malaria vectors and planning control measures for reducing malaria transmission in disease endemic areas 25 .

The use of complex natural odour baited resting boxes should be considered for further evaluation This will be more appropriate in planning for the intervention or control measures against An. arabiensis. The ecological characteristics should be used to deploy the appropriate odour for vector sampling.

The authors declare that they have no competing interests.

EJK conceived and designed the study, data collection and analysis; and writing up of the first draft of this manuscript. BJM and CPM involved in data collection. SM and EEK provided support on the overall study and commented on drafts of this manuscript. AMM co-designed and was responsible for data collection and analysis. All authors read and approved the final version of this manuscript.