Aedes aegypti (L.) is the main vector of important arboviruses that cause human disease in tropical and subtropical urbanized areas of the world, such as dengue (DENV), chikungunya (CHIKV), Zika (ZIKV), and yellow fever (YFV). Incidence of dengue has been high in recent years (2010–2023), with 10,912 travel associated cases in 56 states or territories and 34,619 locally acquired cases in 16 states or territories of the United States (CDC 2024). The number of reported dengue cases in the American Region has been increasing at record numbers since the 1980’s, with over 10 million cases recorded in 2024 (PAHO 2024). These numbers illustrate the challenge of effectively controlling Ae. aegypti and dengue. A key to controlling this mosquito species is having adequate tools to track its presence and abundance. It is important to define mosquito density thresholds based on Ae. aegypti females that would prevent local arbovirus transmission and serve as well-defined targets for mosquito control programs (Barrera 2022, Ong 2021). Several passive traps that attract and capture gravid females of Ae. aegypti have been designed to monitor this species (Lee et al. 2013, Barrera et al. 2014, Ritchie et al. 2014). These traps use the same principles for attracting gravid Ae. aegypti females as traditional ovitraps such as dark colors, standing water, and decomposing plant materials. A common trapping element is a sticky glue board placed inside a capture chamber where females mosquitoes land and get stuck while trying to reach the water, which is made inaccessible by means of a screen.

We compared the sensitivity of three sticky autocidal gravid traps: CDC Autocidal Gravid Ovitrap (AGO), Biogents Gravid Aedes Trap (GAT) (Biogents AG, Regensburg, Germany), and the Singapore Gravitrap (SGT) (Lee et al. 2013, Barrera et al. 2014, Ritchie et al. 2014). These traps are all dark, but they vary in size and components such as volume of water, capture surface, and plant materials (Figure 1). They also vary in the recommended length of time each type of trap should be refreshed and how frequently mosquitoes should be removed, identified, and enumerated. We made efforts to prepare and use the traps following their original instructions.

The AGO trap is a 19-l black plastic container that holds 10 l of water and a 30 g bundle of hay grass, a 3.8-l black plastic cylindric capture chamber and trap entrance that sits on top of a plastic lid, a 3/4” black polypropylene netting at the entrance to the trap to block large debris, and a fine mesh at the chamber’s bottom (Barrera et al. 2014). The traps were sampled and rotated weekly following a Latin Square design. No AGO trap maintenance was conducted because they only need servicing every eight weeks, which was the duration of this study (Acevedo et al. 2016). The BG-GAT trap is a 10-l black plastic container with 3 l of an alfalfa infusion, a translucent capture chamber, a black nylon screen placed between the base and the translucent chamber, a black funnel on top of the translucent capture chamber that serves as the entrance to the trap, and a sticky card inside the capture chamber (Eiras et al. 2014, Ritchie et al. 2014). Sampling and rotation of the BG-GAT traps were done weekly, with complete infusion and sticky board replacement every 2 weeks. The SGT trap is a 1.2-l black plastic container that has an inner adhesive glue lining above 500 ml of 10% hay grass infusion. A wire netting prevents escape of adults and two draining holes at the level of the wire net keeps water level below the net. Sampling, rotation, and replacement of the sticky lining of the SGT traps were done weekly. However, the infusion was replaced twice a week because of water evaporation (Lee et al. 2013). The glue used in the three types of traps came from the same manufacturer (UVR 32, Atlantic Paste and Glue, Bayonne, NJ). The study was conducted in Villa del Rey neighborhood in Caguas city, Puerto Rico (18° 12’ 52” N; 66 ° 03’ 9” W; 91 m.a.s.l). Villa del Rey is a residential neighborhood with one- or two-story houses, piped water supply, residential garbage pickup, and sewerage. We deployed a total of 60 traps (20 traps from each type) in each of 20 pairs of adjacent houses with three traps for each pair of houses. Traps were placed at the front of the houses. Distance between traps was 20–25 m. The presence and abundance of Ae. aegypti females were monitored by transferring the specimens from the sticky liner to a paper towel, where they were identified and counted in the field. The integrity of the traps was monitored twice a week to replace any missing traps or trap components.

A Generalized Linear Mixed Model analysis was conducted to compare average weekly captures of female Ae. aegypti attracted to the three types of sticky gravid traps. We used a negative binomial distribution function with log link to account for the count and over dispersed nature of the data. A first-order autoregressive covariance was used to account for repeated measures (weeks). The model included trap types as the main factors and trap ID as a random effects factor to account for individual trap variation. The results of the analysis showed significant differences in the average captures among each trap type during the eight weeks of the study (F_2,471= 80.2; P< 0.001; Figure 2). Model estimated means revealed substantial disparities in capture rates among trap types. Captures of female Ae. aegypti per trap/week were 11.1 (9.1 – 13.5; 95% CI) in AGO traps, 7.2 (5.9 – 8.7) in GAT traps, and 1.7 (1.4 – 2.1) in SGT traps. These results indicated that GAT traps and SGT traps captured 35.5% and 84.7% fewer females of Ae. aegypti than AGO traps.

The results showed that Ae. aegypti was present in all 20 sites during the eight weeks of observations. Monitoring mosquitoes in three trap types in 20 sites for eight weeks produced a total of 480 observations. The AGO, GAT, and SGT traps did not to catch specimens in one, nine, and 58 out of the 160 observations conducted on each trap type (trap-wks), respectively. Trap failures were one, six, and one for the AGO, GAT, and SGT traps, respectively. Trap failures included missing and disturbed traps. Despite the absence of females of Ae. aegypti at some sites and weeks in the traps, all three traps were able to detect the presence of this mosquito at each of the 20 sites during the eight weeks of observations.

Field studies using AGO traps in other locations in Puerto Rico found similar numbers of female Ae. aegypti per trap per week to the present study. A study conducted in Jacksonville, FL (Cilek et al. 2017) using a modified AGO trap captured fewer specimens of Ae. aegypti (1.05 ± 0.28). Discrepancies among studies are expected because of differences in local abundance of mosquitoes and modifications of the traps. For example, the Jacksonville study used AGO traps with three liters of an oak infusion instead of the ten litters and a pack of hay grass. GAT traps captured similar number of female Ae. aegypti in Cairns, Australia (5.7 ± 6.1) to those in the present study (Ritchie et al. 2014). Captures conducted with the SGT in Singapore were much lower (0.07–0.12) than in the present study, possibly reflecting the lower Ae. aegypti abundance due to an aggressive control program (Lee et al. 2013, Ong et al. 2020, Ho 2023). It is not surprising that AGO traps captured more specimens of female Ae. aegypti and detected the presence of this species on more trapping instances than the smaller gravid traps compared in this study (Harrington et al. 2008). AGO traps are bigger, holding a larger volume of water, and have a greater sticky capture surface. Capturing more specimens can be useful but other aspects of a trap are also important including its sensitivity to detect the presence of Ae. aegypti, cost of traps that could affect the deployment density, and logistics of maintenance and deployment of the surveillance network of traps. Missing the presence of this mosquito species may overestimate the impact of vector control or missed important detections of Ae. aegypti in areas where this mosquito may be expanding its distribution. While the results of this brief study showed that all three traps were adequate to monitor the presence of Ae. aegypti, increasing the number of smaller traps is required to improve their sensitivity. This study has the limitations of a relatively short duration of the field assays and that it was conducted in only one urban area. Longer comparative studies would be needed to establish how trap captures in each of the three trap types correlate with each other. One limitation of this study is that the plant materials used as attractants (hay grass, alfalfa) and their presentation (infusion, hay bundle) varied for each type of trap. We compared traps based on their original descriptions and field usage to provide comparative values for interpreting published reports.