The results of this field trial in Mali show that under local conditions a single application of ATSB solution by plant-spraying markedly reduced the relative abundance of An. gambiae s.l. populations and their longevity. Within a week after spraying, densities of adult females and males at the treatment site were reduced by around 90% and remained low throughout the remainder of the monitoring period. Clearly, the ATSB treatment was highly effective in killing the "older" more dangerous females as shown in table 1. Reducing the proportion of "older" females is a key factor in reducing malaria transmission . The pronounced impact of the ATSB is comparable to that demonstrated in ATSB field trials in Israel [12–16] and establishes that this method for mosquito control is also highly effective for targeting and killing major malaria vectors in semi-arid areas of Africa.
By using a dye marker in the ASB solution applied at the control site, as in previous studies in Israel [12–16], we demonstrated that a high proportion of the local An. gambiae s.l. populations were making contact with and feeding on the solution sprayed on local plants. The observed marking rates for females (56.4%) and males (62.2%), however, represent only minimal rates of contact as the dye marker persists for only about two days due to digestion processes while the mosquitoes can sugar-feed throughout their lifespan. The finding of marked mosquitoes on the last day of collection highlights that the sprayed ASB solution was still present at the very end of the trial. The low percentage (< 5%) of mosquitoes caught with colored ATSB from the treatment site indicates that a high percentage died before again flying where they could be caught, with the possibility that some may have exhibited behavioral changes after feeding on the bait that would have altered their probability of capture .
The results demonstrate how ATSB is effective when applied to various types of vegetation located in the vicinity of local mosquito populations, including that which exists around natural larval habitats and is likely used by both newly-emerged and older mosquitoes as outdoor resting sites. This approach is similar to the most recent studies in Israel [16, 32] but differs from initial studies of ATSB plant-spraying in Israel where ATSB was selectively sprayed on flowering plants known to be highly attractive to mosquitoes as sugar sources [12, 13]. The two approaches are both highly effective and potentially complimentary but the method used in this study and a recent study in Israel  is technically simpler as it does not require a priori knowledge the most attractive plants. It only requires some basic skills in identifying larval habitats and general types of vegetation that may be used by mosquitoes as outdoor resting sites .
The preparation of ATSB solution is technically quite simple. Four of the key ingredients are readily available at the local community level: water, unrefined brown sugar, beer, and ripe/overripe fruit. While initial studies in Israel used nectarines [14, 15] and plums in Florida , guava and honey melons were used instead based on local availability at the time of our studies and on our comparative tests of the attraction of An. gambiae s.l. to various local fruits and seed pods in Mali (unpublished). Even fruits that are close to rotting and are therefore unsuitable for trade and human consumption can be used, and leftover products can be used to feed domestic animals and fowl. As the chemical identity of the attractive ingredients in the fruits has not been determined, at this point it is not possible to substitute a synthesized chemical attractant. Two of the ingredients must be purchased, the BaitStab™ for preservation and stabilization, and the oral toxin, but both are very inexpensive. At the study area in Mali, the boric acid was purchased at the local market.
Rather than using Spinosad ("Tracer™"; Dow Agrosciences, Calgary, Canada) as the oral toxin as in the proof-of-concept studies in Israel [14, 15], we instead used boric acid, which is highly lethal to mosquitoes [24, 31]. Preliminary laboratory testing in Bamako confirmed high toxicity to An. gambiae. The advantage of using boric acid is that it is very inexpensive, readily available, is stable (in contrast to Spinosad which decays by UV), and has a mammalian toxicity level about as low as table salt . The boric acid proved highly effective in our initial field trial. This is not surprising because boric acid and a number of different insecticides have been used for many years as oral toxin for the control of ants, cockroaches, fruit flies, and house flies. Studies by Allan  have recently shown that, when delivered orally, a wide variety of different insecticides are effective against mosquitoes, with apparently no repellency effects. The study concluded that baits with oral toxins for mosquitoes using a phagostimulant, such as sucrose, are effective in causing mortality . Longer-term, operational strategies using ATSB solutions with mixtures of 2 or more different insecticides may help minimize the emergence of resistance in local populations of mosquitoes, which is of course already a concern for the insecticides associated with LLIN and IRS use for malaria vector control in Africa [36, 37].
This first field trial of ATSB methods in Mali begins to explore some of the ultimate impacts of the ATSB approach for malaria vector control in Africa. In addition to the ATSB plant-spraying methods tested here, it also might be possible to deliver the same ATSB solution using very simple bait stations that have proven successful in Israel [14, 15]. Ultimately, we expect that strategies will emerge for co-use of both plant-spraying and bait stations to achieve maximal killing of local vector populations. As the malaria vectors in Africa, An. gambiae, An. funestus and to a lesser degree An. arabiensis, show a pronounced tendency to rest inside houses where they feed on humans , it may also be possible to use ATSB methods directly outside or inside houses. Though there were indications of a differential impact on An. arabiensis in this trial (i.e., none remained after ATSB treatment), the numbers identified by PCR were too small to detail with certainty that the ATSB treatment had a more pronounced impact on this malaria vector which is well-known to be more exophilic than An. gambiae.
Beyond this initial field trial, the full impacts of ATSB need to be determined by field assessments on a larger scale and of longer duration at the village and/or district levels with designs that measure impact not only on vector densities and vector longevity, but also measure malaria parasite transmission (e.g., EIRs), and malaria burden in human populations (e.g., incidence and prevalence of malaria cases). It is also important to determine the additive effects of ATSB when used in combination with existing vector control methods including LLINs and IRS, as it is not likely that ATSB methods alone would be sufficient to meet programmatic goals for malaria vector control. It is also important to determine the full impact on all mosquito species, not just the malaria vectors. In Mali, for example, Culex quinquefasciatus is locally abundant and serves as a major nuisance-biting mosquito and a vector of filariasis in some areas .
The range of environments in Africa where ATSB methods can be used effectively remains to be determined. They will likely work best in arid and semi-arid areas where natural flowering plants are limited, as the effectiveness of the ATSB methods depends on their ability to outcompete natural plant sources of sugar available to mosquito populations . Though this first field trial was conducted in a semi-arid area of Mali, the actual study sites containing multiple ponds that were highly productive larval habitats surrounded by both natural vegetation and rice paddies were, in general, fairly typical of An. gambiae s.l. habitats over a range of environments found in malaria endemic areas of Africa. The ATSB methods may also work well in urban setting and in environmentally altered environments that lack biologically diverse groups of indigenous flowering plants that would naturally sustain mosquito populations. They may also be effective for use in large-scale irrigation areas where rice, for example, is cultivated (rice plants apparently do not provide a source of sucrose for mosquitoes).
There are three further considerations worth noting. First, to optimize performance of ATSB plant-spraying and bait station methods, there is a need to determine the coverage of plant spraying needed and also the density of bait stations needed to achieve effective control. In this first field trial in Mali, spraying just a series of 1 m2 spots of vegetation every 3 m around breeding sites was apparently sufficient. Second, a logistical consideration is that heavy rains will wash off ATSB sprayed on plants and so re-applications during the rainy seasons may be needed. This is one reason why bait stations are equipped with covers [14, 15]. During some periods of the year it may be feasible to use both plant-spraying and bait stations but this will depend on local circumstances. Third, ATSB approaches have only minimal risks to humans. Ongoing studies in Israel are determining potential impacts of ATSB on non-target insects. Strategies for spraying ATSB on non-flowering plants may be better than spraying the most attractive flowering plants, in terms of minimizing damage to non-target insects. Honey bees or any of the many species of pollinating bees may be affected and so in Israel suitable metal grids for bait stations have been developed that allow mosquitoes to pass but keep honey bees out (G. Müller and Y. Schlein, unpublished data). Overall, spraying non-flowering vegetation seems to be environmentally safe, except that non-biting midges (Diptera: Chironomidae) feed in similar proportions to the mosquitoes. The ATSB methods may pose only limited environmental risks for the following reasons: 1) whole classes of pollinating insects orient using optical targets rather than scents, 2) beneficial predatory insects will not be harmed because they do not feed on sugar, and 3) pollinating insects are typically absent from some of the prime target areas for ATSB treatment, such as rice fields and areas around mosquito larval habitats where there is minimal flowering vegetation.
In conclusion, this first field trial of ATSB methods in Mali provides a strong indication that such strategies will be very effective for malaria vector control in Africa. If tested further and found to be effective across a range of malaria endemic environments in Africa, it is likely that ATSB approaches could soon be added as a major component of IVM-based malaria vector control programmes. ATSB methods differ from and potentially complement LLIN and IRS methods, which focus on indoor-feeding and resting mosquitoes, because they have so far proven effective in outdoor habitats for killing all physiological states of females and at the same time also kill male mosquitoes. By targeting sugar-feeding mosquitoes in outdoor environments it is likely that their use will on the micro-scale overlap significantly in space with other key life history strategies of malaria vectors including mating and oviposition, both of which are temporally associated with sugar-feeding. Thus, in terms of malaria vector control in Africa, the ATSB methods when used operationally will likely reduce both total numbers of recently emerged female anophelines before they enter houses to feed on humans and the proportion of females exiting houses to oviposit and then returning to houses to re-feed on humans. This strategy to broaden the currently narrow segments of vector populations targeted for control (i.e., those females that feed and rest indoors) is certainly consistent with the broad-based IVM concepts being promoted and implemented throughout Africa.