Two conventional larvicide applicators, a Stihl^® SR 420 backpack sprayer (Andreas Stihl Ag & Co. KG, Waiblingen, Germany) and a truck-mounted Curtis Dyna-Fog Ag-Mister LV- 8™ (Curtis Dyna-Fog, Westfield, IN, USA) low-volume (LV) sprayer were evaluated for their efficacy to deliver spray droplets in open and cryptic cups. Tests were carried out in an open parking lot at Rutgers University. Cryptic cups (i.e., ovicups inserted into pipes) were placed in parallel, perpendicular and oblique positions relative to the equipment’s spray direction. Cups were placed 1 m distance apart in three replicates. Kromekote papers (Mohawk Fine Papers, Inc., Cohoes, NY, USA) measuring 5×7.5 cm were placed in the bottom of each cup. A 2% solution of Red 40 granular dye (Glanbia Nutritionals, Carlsbad, CA, USA) was sprayed using the backpack and truck-mounted sprayers. Experiments were conducted on the day when the prevalent wind speed was minimal and was found to be between 2 to 5 km/h. Backpack sprayer was calibrated to flow rate of 0.12 L/min and application was performed from the distance of 3 m from the cup placement. The low-volume sprayer was adjusted to flow rate of 8.3 L/min and application was performed at 10 m from the cup placement. Both sprayers moved at the speed of 8 km/h and perpendicular to the wind direction. Papers were collected after 30 min of application and analyzed by DropVision AG system (Leading Edge Associates, Inc., Waynesville, NC) to determine droplet size and density. Each experiment was repeated three times.

Data were first tested for normal distribution using Shapiro-Wilk test. Data are presented as mean ± SE unless otherwise stated. Mann-Whitney Rank sum test was performed to compare eggs received by open and cryptic cups in tunnel experiment. Fisher Exact test was used to compare oviposition receiving and non-receiving, cryptic and open cups. Mann-Whitney Rank sum tests were performed for pairwise analysis to examine statistically significant differences for droplet density and median droplet size between open and cryptic cups. Efficacy of larvicide and pyriproxyfen autodissemination was expressed as percent larval and pupal mortality, respectively. In all bioassay experiments, control mortality was adjusted with open and cryptic cups mortality using Abbott’s formula [16]. Mann-Whitney tests were used to determine the significant difference in larval mortality between open and cryptic cups during larvicide application, and in pupal mortality during autodissemination experiments. Percent pyriproxyfen contamination was calculated using the following formula: %pyriproxyfencontamination=NumberofcontainersshowingpyriproxyfenactivityTotalnumberofcontainers×100

The backpack application of B. thuringiensis larvicide produced 75.8±7.0% Ae. albopictus mortality in open ovicups, whereas cryptic cups showed virtually no measurable mortality at 0.3±0.3 (Mann-Whitney, T = 1867, p≤0.001). The truck-mounted treatment was similarly ineffective against cryptic cups (0.1±0.1% mortality) in comparison to open cups (19.4±6.1%) (Mann-Whitney, T = 1499, p = 0.001). Cups placed in the control plot did not show larval mortality in these trials.

Asian tiger mosquitoes can use almost any small, water-holding container for larval development; however, this species showed a strong oviposition preference for cryptic over open cups in our field cage experiment suggesting a higher contribution of cryptic habitats in population buildup. The major difference between open and cryptic cups is their accessibility, open cups have a highly apparent and accessible top entrance, whereas cryptic cups are hidden within a pipe and only accessible from side entrances. Greater Ae. albopictus larval prevalence in corrugated extension tubes than open (i.e., exposed) containers in urban residential areas has been previously reported [17]. Higher larval prevalence was found in containers such as tires, water bottles, tin cans, gutter hoses, corrugated pipes, and water drainage pipes etc. [6,18–22] which displayed some degree of a cryptic nature.

If our hypothesis is correct that cryptic habitats hold much or most of the Ae. albopictus population, effective control strategies to curtail container mosquitoes in these habitats would be vital. Yet the droplet data demonstrated for the first time that cryptic cups are nearly impenetrable to the conventional backpack and truck mounted applications routinely used for larvicide applications [9,23]. Irrespective of their position relative to the spray application path, cryptic cups were difficult to reach by aqueous droplets. Similarly, cryptic cups were nearly invulnerable to our Bti applications as reflected in almost negligible larval mortality compared to open cups. The side entrance of cryptic cups makes them almost inaccessible to insecticidal droplets suggesting that the efficacy of conventional sprayers against container mosquitoes may be limited and their efficacy is overestimated when evaluated against open cups only as the majority of the population preferred cryptic habitats.

For effective management of the Asian tiger mosquito, an alternate vector management strategy is needed which can specifically target cryptic habitats. Autodissemination is emerging as a potential new strategy with a promise to deliver insecticides to both exposed and hidden habitats. Autodissemination exploits the behavior of the insect as a vehicle to deliver toxicants to the pest habitats. Consider the goat moth, Coccus coccus, which also lives in cryptic, hidden habitats. Pre-infecting adults with entomopathogenic nematodes and then releasing them to locate and deliver the parasites to these hard-to-reach target sites resulted in 86% larval mortality, whereas conventional methods produced only 4% larval mortality [24]. Similarly, the ability of autodissemination to transfer lethal concentrations of insecticides has been successfully demonstrated in the laboratory and the field for Ae. albopictus and other mosquito vectors [12,25–29]. Previous studies, however, have used only open cups to assess autodissemination efficacy and did not examine efficacy in cryptic habitats. We have tested and demonstrated that autodissemination can deliver toxic concentrations of insect growth regulator to both open and cryptic cups.

We have developed multiple designs for autodissemination stations that exploit the oviposition behavior of gravid Ae. albopictus to deliver pyriproxyfen to larval habitats [12,14,30]. Attracted to the station by an oak infusion, the mosquito enters the station but is unable to reach the water and oviposit. In the current design, when they exit they are forced to walk across an oil and powder fabric bands impregnated with pyriproxyfen. In addition to its insecticidal role, the oil formulation serves additional roles as a sticker to enhance attachment of the powder formulation, and to facilitate quick release into the larval habitat during oviposition.

In two years of field studies, we have shown here that autodissemination stations are useful against both open and cryptic cups. We observed higher pupal mortality and pyriproxyfen contamination in cryptic than open cups, however, this difference was not significant in 2013. Although pyriproxyfen reaches both cryptic and open cups in 2013, overall field efficacy was at best modest that year. Our 2013 plots were located in a suburban area with unmistakably low mosquito populations; because autodissemination relies entirely on wild adult mosquitoes to transfer the insecticide, the outcome from low populations cause low mortality. Greater autodissemination efficacy was obtained in the 2014 field study because stations were modified to perform for a longer period, the number of stations per block was doubled, and the trial was conducted in an inner-city area where adult populations were chronically high. Pyriproxyfen autodissemination in cryptic cups was consistently higher than open cups during the entire study period. Mortality in 2014 was comparable to previous autodissemination field experiments [30], though the pupal mortality in cryptic cups was relatively low to reach the desired level of effective control recommended by vector control authorities. Efficacy may have been underestimated because field samples were filtered to remove organic debris which likely reduced pyriproxyfen given this chemical’s propensity for adsorption to different substrates [31].

Pyriproxyfen dissemination was further confirmed by residue analysis of the sentinel cups, while previous autodissemination field studies have tended to rely on larval bioassay to show pyriproxyfen presence and activity but lacked direct evidence of transfer to sentinel cups [25,27,28]. Although, open and cryptic cups have shown the presence of pyriproxyfen, we found a surprisingly large difference in pyriproxyfen concentration between the 2013 and 2014 samples. One possible explanation for such a huge difference may be variation in the samples. During 2013, we analyzed week 8 samples for residue analysis. Whereas in 2014, week 4 samples were analyzed instead due to low pupal mortality in week 8 and week 12 samples. There is a possibility that pyriproxyfen may have accumulated in field sentinel cups over time in 2013, which eventually resulted in higher concentrations. However, this requires further investigation to establish the relationship of various factors i.e. age of water sample, organic components, temperature, and mosquito density that can affect pyriproxyfen residual concentrations.

Autodissemination system is a low maintenance and easy to operate which can cover an entire mosquito season once deployed. Pyriproxyfen is a highly effective juvenile hormone analog, besides its pupicidal action, pyriproxyfen sterilizes adult females and decreases spermatogenesis in males [32], shows ovicidal activity [13], and can terminate egg diapause prematurely [33]. The impact of pyriproxyfen on non-targets is minimal as dissemination is tightly targeted to small, scattered, container habitats. Additionally, the amount carried by females is minuscule which further reduces insecticide load in the environment [14].