An essential component of vector-borne disease monitoring programs is mosquito surveillance. Surveillance efforts employ various collection traps depending on mosquito species and targeted life-history stage, i.e., eggs, larvae, host-seeking, resting, or gravid adults. Surveillance activities often use commercial traps, sometimes modified to accept specific mosquito species attractants. The advent of widely available and affordable 3D printing technology allows the construction of novel trap designs and components. The study goal was to develop and assess a cost-effective, multipurpose, 6-volt mosquito trap integrating features of both host-seeking and gravid mosquito traps to collect undamaged live specimens: a multifunctional mosquito trap (MMT). We tested the MMT in comparison to commercial traps, targeting gravid Aedes albopictus, host-seeking Ae. albopictus, and total number of host-seeking mosquitos regardless of species. Field evaluations found the MMT performed as well as or better than comparable commercial traps. This project demonstrates an easy to construct, inexpensive, and versatile mosquito trap, potentially useful for surveying multiple mosquito species and other hematophagous insects by varying attractants into the MMT.
Active mosquito surveillance is an essential component of vector-borne disease control programs. As a pillar of the Global Vector Control Response 2017–2030, vector surveillance and monitoring help reduce vector-borne disease (WHO 2017). A key component of successful monitoring programs is surveillance traps near at-risk human populations (Roiz et al. 2018). The type of trap deployed is often driven by the need to collect as many mosquitoes as possible while explicitly targeting a known vector species at a given life-history stage: egg, larva, host-seeking adult, resting adult, or ovipositing adult (Silver 2008). Consequently, mosquito surveillance traps use various attractants, including light, odor, visual stimuli, thermal, movement, or sound (Sudia and Chamberlain 1962, Kline 2002, Silver 2008, Hoel et al. 2011). For adult trapping, traps can passively collect specimens by trapping individuals who are attracted to enter the trap or actively collect specimens through powered fan suction and usually target either host-seeking or gravid (ovipositing) female mosquitoes (Silver 2008).
Suction surveillance traps feature an electric suction fan designed to pull mosquitoes into a collection chamber. The most widely used suction trap is the Centers for Disease Control and Prevention light trap (CDC-LT), initially developed in the 1960s (Sudia and Chamberlain 1962, Silver 2008). The CDC-LT features an incandescent light source as the main attractant; however, CO2 attractant in the form of dry ice or a CO2 compressed gas cylinder is often used to significantly augment this trap (Silver 2008). The CDC-LT is a good trap for catching a diversity of host-seeking mosquitoes (McDermott and Mullens 2018). On the other hand, the BG-Sentinel (BGS) (Biogents AG, Regensburg, Germany) is a suction trap that has become the “gold standard” for species-specific traps (Kroeckel et al. 2006). Originally designed to target Aedes aegypti (L.) through visual and chemical cues, it has also proven effective in collecting Aedes albopictus (Skuse) (Ritchie et al. 2006, Meeraus et al. 2008, Lacroix et al. 2009). The CDC-LT and BGS mainly catch host-seeking adult mosquitoes.
While the CDC-LT and BGS collect specimens from downdraft suction action, other traps, such as the CDC gravid trap, use an updraft fan intake chamber positioned above a water tray to capture gravid female Culex (Reiter 1987). The updraft suction fan design pulls mosquitoes up into a collection chamber and improves upon downdraft suction fan designs because the latter pull mosquitoes past the fan blade, potentially degrading mosquito catch quality for identification (Wilton and Fay 1972, Wilton 1975). The BGS requires higher suction, using a more extensive and heavier 12-volt battery to pull mosquitoes into the collection bag position and above the fan, potentially damaging collected specimens. Many mosquito species respond to suction fan air current by flying upward, potentially reducing mosquito catch in downdraft suction designs (Wilton 1975, Silver 2008). An updraft suction fan does not reduce total mosquito catch even when fan power is low, whereas downdraft suction fans lose catch if similarly underpowered (Wilton 1975).
Nonsuction traps, commonly known as passive traps, are generally designed to be visually stimulating and lure mosquitoes into collection chambers with odor attractants, then collect mosquitoes with sticky cards or kill them with a pesticide. Their main advantage is the ability to collect mosquitoes without the need for power: they can be deployed over long periods, i.e., days to weeks, are less expensive, and have more commonly been used to target egg-laying, gravid mosquitoes where large numbers of traps should be deployed (Barrera et al. 2014, Johnson et al. 2015). The CDC autocidal gravid oviposition trap (AGO) and the gravid Aedes trap (GAT) (Biogents AG, Regensburg, Germany) attract gravid container-breeding Aedes mosquitoes into collection chambers with water-filled bases, killing the mosquitoes with sticky cards (Mackay et al. 2013, Barrera et al. 2014, Ritchie et al. 2014). In addition to surveillance, AGOs and GATs are useful for mosquito control, demonstrating reductions in female Ae. aegypti and Ae. albopictus. However, further large-scale trials are recommended to assess suppression of mosquito-borne pathogens (Johnson et al. 2017). As such, passive traps can target host-seeking mosquitoes for pathogen surveillance. The Sentinel Mosquito Arbovirus Capture Kit (SMACK), using a CO2 attractant, identified Ross River virus and Barmah Forest virus in samples collected from North Queensland, Australia, and collected hundreds of mosquitoes (Johnson et al. 2015). A more recent design of a host-seeking passive trap is the Silva trap (Silva et al. 2019). This trap features an ultraviolet light source with a collection chamber; however, no suction fan, which, when tested, performed as well as a CDC-LT (Silva et al. 2019).
A lack of access to surveillance traps in remote locations inspires light traps constructed out of local materials, often without standardization (Silver 2008). One method to assist in building standardized traps is to use three-dimensional (3D) printed components. In recent years, 3D-printing technology has dramatically increased in availability and affordability for many different applications. In the case of mosquito traps, this technology allows the development of prototype components, allowing researchers to rapidly modify traps to suit their needs (Hoshi et al. 2019). Printing a 3D trap similar to the CDC-light trap reduces the approximate per trap cost from $106 to $12.97 (Hoshi et al. 2019). Thermoplastics used in 3D printing are relatively inexpensive, reducing per trap costs for surveillance programs after the initial purchase of 3D printers (Hoshi et al. 2019). Different plastics are available for 3D printing, depending on the application. Polylactic acid (PLA) is a popular inexpensive plastic made from renewable sources (i.e., corn starch, sugar cane); however, it is biodegradable and not suitable for environmental exposure (Anderson 2017, Wickramasinghe et al. 2020). This plastic is a standard for building prototype parts due to its low cost; however, outdoor insect traps require environmentally durable 3D printable plastics. For this purpose, acrylonitrile butadiene styrene (ABS) and glycol-modified polyethylene terephthalate (PETG) are readily available, environmentally durable thermoplastics suitable for outdoor use (Anderson 2017, Wickramasinghe et al. 2020). As a structural material, PETG prints better, is stronger, more durable, and has greater color availability. However, ABS is less expensive and lighter weight. An immense benefit to 3D printed traps is the rapid replacement of parts damaged during trap operation. Additionally, the flexibility of 3D printing allows rapid modifications of trap design (Hoshi et al. 2019). Overall, there is a lack of published literature about 3D-printed mosquito surveillance traps. This deficit represents an opportunity for researchers to develop and widely distribute novel trap designs that local surveillance agencies can economically produce.
Furthermore, no trap design targets both gravid and host-seeking mosquitoes. The goal of this study was to design, build, and assess an inexpensive, easy to construct, resilient, 6-volt multipurpose trap, able to collect gravid or host-seeking mosquitoes during 24-h sample periods: a “multifunctional mosquito trap” or MMT. Here we present an adaptable trap design built using 3D printing technology and conventional construction techniques to target gravid or host-seeking mosquitoes.
MATERIALS AND METHODS
Multifunctional mosquito trap design
The multifunctional mosquito trap (MMT) uses an updraft suction collection chamber inspired by the Frommer updraft gravid trap (Model 1719, John W. Hock Company, Gainesville, FL) combined with a 5-gallon bucket base similar to the CDC AGO. A novel intake manifold integrates both trap elements (Fig. 1). This intake manifold consists of a 3D printed intake component attached to a commercially available Viagrow 6-inch mesh pot bucket lid insert (Hydro Generation Inc., Atlanta, GA).
The 3D components were initially designed in Tinkercad software (Autodesk Inc, San Rafael, CA) then imported into Cura software (Fargo Additive Manufacturing Equipment 3D, Fargo, ND) for printing on a Lulzbot Mini 2.0 3D printer (Fargo Additive Manufacturing Equipment 3D). The 3D printing of trap components used either black or white PETG 3D filament (Esun, Shenzhen, China) depending on the part.
Field trapping evaluations
We assessed the multifunctional mosquito trap for collecting gravid, host-seeking female Ae. albopictus, and total host-seeking female mosquitoes in 3 separate field evaluations against single standard function commercial traps. A field evaluation for collecting gravid Ae. albopictus mosquitoes compared the MMT to the Frommer updraft gravid trap (Frommer-GT) and GAT (Fig. 2). Another field evaluation compared host-seeking Ae. albopictus collections of the MMT and BG-Sentinel 2 trap (BG2) using the BG-lure (Biogents, Regensburg, Germany) chemical attractant alone (Fig. 2). The final field evaluation compared total female mosquito collections between the MMT and CDC-LT using dry ice as an attractant (Fig. 2).
We selected 3 field sites for gravid Ae. albopictus collection on North Carolina State University (NC State) property in Wake County, NC. Each of the 3 field sites received 3 traps (1 MMT, 1 Frommer-GT, and 1 GAT) along the tree line with 30-m separation between traps. Traps were rotated at each site every 72 h over 9 days to homogenize variation due to trap placement within the sites. Previous sampling using host-seeking BG-2 traps with BG-lure found Ae. albopictus at all sites. Two sites, Site 1 (8758338°35′18.8406″W, 4262783°19′23.9109″N) and Site 2 (8759453°0′56.8645″W, 4262616°6′22.7069″N), are located at the NC State Lake Wheeler Road Field Laboratory, a 1,500-acre property, and another sample site, Site 3 (8760444°8′39.4755″W, 4271500°26′39.6741″N), was near Method Road Buildings (Fig. 3). A 4 g/liter oak leaf infusion (OLI) prepared 5 days before initiation of the field evaluation provided an attractant for gravid Ae. albopictus mosquitoes (Obenauer et al. 2009, Reiskind and Janairo 2018). We made OLI by mixing in a water cooler 320 g (16 g/liter) of locally collected dried willow oak leaves (Quercus phellos L.) bundled in a fine mesh cloth, 4.8 g (0.3 g/liter) of brewer's yeast (MP Biomedicals, Cat no. 903312, Fisher Scientific, Waltham, MA), and 4.8 g (0.3 g/liter) of egg albumin (Cat no. A388-500, FisherChemical, Waltham, MA) into 20 liters of tap water, then incubating the 16 g/liter oak leaf infusion for 5 days. Finally, we diluted the OLI 1:4 in tap water, resulting in a 4 g/liter oak leaf infusion on the day of use. The Frommer-GT and GAT received 3 liters of 4 g/liter OLI during field evaluations; however, the MMT received 6 liters of OLI due to the more extensive reservoir. We recognize that the larger OLI volume may increase the MMT's attractiveness, but we were interested in trap effectiveness as they were intended to be deployed. We inspected gravid traps daily, replacing batteries and collection chambers for the Frommer-GT and MMT during each visit. For the GATs, collection and replacement of sticky cards occurred on 72-h intervals, rotating all traps at each site during this inspection. To homogenize OLI attractiveness and account for trap liquid volume differences due to OLI evaporation, during each rotation, we combined and reallocated OLI from each trap at the site with 6 liters of excess OLI maintained at room temperature in our laboratory. We identified mosquitoes from each trap and dissected females to determine gravid status after each collection period. The number of collected gravid female Ae. albopictus was the measured response during each 72-h sample period due to the difficulty and expense of replacing sticky cards for the GATs.
To assess the MMT for collecting host-seeking Ae. albopictus, we conducted a 16-day field evaluation at the NC State Lake Wheeler Field Laboratory Site 1 (Fig. 3). This assessment placed an MMT and BG2 along the tree line with a 60-m separation between traps. We provisioned each trap with a BG-lure attractant and twice conducted 8-day 24-h sample periods with trap rotations every 4 days. We inspected each trap daily, collecting 24-h catch samples and replacing batteries. In this case, we used the number of female Ae. albopictus collected in each 24-h sample period as the response variable.
For the 3rd field evaluation, we used a dry ice attractant to assess the MMT's effectiveness at collecting host-seeking mosquitoes regardless of species. We compared the MMT to the CDC-LT without a light bulb, a standard set up for collecting large numbers of host-seeking mosquitoes with minimal bycatch of nontarget insects (Silver 2008). Again, trap assessment took place at Site 1, located at the NC State Lake Wheeler Field Laboratory (Fig. 3). We placed each trap with 60-m separations in new locations within Site 1. We inspected traps each morning, provisioning each trap with approximately 3.5 lbs (∼1.59 kg) of dry ice, collecting mosquito catch, and replacing batteries. We tied the CDC-LT to overhanging tree branches for sampling, with the intake positioned approximately 1.5 m above the ground. In contrast, the MMT, when positioned in the environment, has an intake at 0.38 m high, which does not require hanging to function. We collected mosquitoes daily 6 times over 24-h sample periods, and we switched trap positions after 3 sample days; therefore, each trap had equal surveillance time at both locations within Site 1. In this field evaluation, we measured the response in total female mosquitoes collected.
Mosquitoes were frozen upon collection, later identified under a dissecting microscope (Model SZ61, Olympus, Shinjuku, Japan) using The Mosquitoes of the Mid-Atlantic Region: An Identification Guide (Harrison et al. 2016). Dissection of mosquitoes assessed gravid status for all female mosquitoes collected during gravid trapping tests.
Data analysis was conducted in R (R Core Team 2017), and figures produced using the package ggplot2 (Wickham 2009). A split-plot analysis of variance (ANOVA) model analysis using the linear mixed effect models package lmertest (Kuznetsova et al. 2017) compared gravid Ae. albopictus collections. Trap comparison for BG-lure (BG2 vs. MMT) or dry ice (CDC-LT vs. MMT) as attractants compared total host-seeking female mosquito collections using pairwise t-tests (t.test in R) for R Core Team (2017).
There was no difference among traps for gravid Ae. albopictus (F = 0.711, df = 2,16; P = 0.506). This field evaluation collected an average of 4.48 ± 0.96 gravid Ae. albopictus mosquitoes per 72-h sample period per trap (Fig. 4). When removing the largest outlier for the Frommer-GT data set, 25 gravid female Ae. albopictus, the cumulative trap average count reduced to 3.69 ± 0.58 gravid Ae. albopictus. When put in perspective, this assessment collected 3 to 4 gravid Ae. albopictus per 72-h sample period per trap, regardless of type. There was a significant difference between sample sites (F = 5.615; df = 2,16; P = 0.014), as expected; therefore, the sample site is a blocking factor in statistical analysis. There was no difference among locations (A, B, or C) within each sample site (F = 0.585; df = 2,16; P = 0.569) or interactions between location and trap (F = 0.965, df = 4,16; P = 0.453).
Host-seeking female Ae. albopictus field collections did not differ between BG2 and MMT traps using BG-lure (t = 1.377; df = 0.189; P = 1.377) (Fig. 5). There were, on average, 28 ± 4.44 female Ae. albopictus mosquitoes per 24-h sample period when grouping collected mosquitoes, regardless of the trap.
In conducting a trap field evaluation using dry ice, we found no difference (t = −2.183; df = 5; P = 0.081) in total female mosquito collections between the CDC-LT and MMT (Fig. 5). However, after removing a single sample period outlier from each trap type, the MMT collected an average of 109.8 ± 13.71 mosquitoes per 24-h sample period, which is significantly greater than the CDC-LT (t = −7.703; df = 4; P = 0.002), which collected an average of 25.4 ± 13.82 female mosquitoes per 24-h sample period. A majority (92.42%) of collected mosquitoes were Ae. albopictus regardless of trap when using a dry ice attractant, likely due to the use of only 1 survey location, which previously demonstrated a high Ae. albopictus density.
The multifunctional mosquito trap performed as well as or better than comparable single function traps. For most of the tests, there was no significant difference between traps, and most of the point estimates (mean mosquitoes captured) were similar, suggesting increasing the sample size would not likely result in statistical significance. The one exception was comparing the MMT to the CDC-LT, in which the MMT outperformed the CDC-LT in collecting host-seeking female mosquitoes when removing 1 outlier from each trap. However, over 90% of the mosquitoes caught in that field experiment were Ae. albopictus. A further comparison between the CDC-LT and MMT under other common pest mosquito monitoring settings, including flood water and saltmarsh habitats, could help clarify any differences in trap effectiveness. Likewise, it would also be important to compare the MMT to the BGS and gravid traps in a more tropical, Ae. aegypti–dominated urban landscape.
Mosquito surveillance traps represent a significant expense for mosquito control and surveillance activities. The cost of producing the multifunctional mosquito trap is $31.60 as of 2019, not including batteries and chargers' cost or attractants. A cost savings compared with the direct purchase of commercial traps, BG2 $213 and CDC-LT $106, similar to that noted by Hoshi et al. (2019) for a 3D print light trap. Surveillance programs already using CDC-LT would only need to produce the MMT, as this trap uses the same gel-cell rechargeable 6-volt battery. Additional cost savings advantage can also be realized in the repair of printed components as they degrade from use, versus purchasing brand new CDC-LT or BGS. There is a caveat to cost savings, since Hoshi et al. (2019) and this study only include material costs and do not account for the total cost of producing commercial traps (i.e., materials, manufacturing, rent, salaried employees, patent protection, shipping, taxes, and of course, profit).
The multifunctional mosquito trap displays several advantages as a multipurpose surveillance trap. Much like the BG2, it breaks down into components for transportation and storage, with all components fitting into its 5-gallon bucket with the lid closed. It is easily placed in the environment and enhanced with multiple attractants. Attractants in the MMT can be used singly or combined due to multiple reservoirs incorporated into the design. While only used to hold attractive oviposition media in this study, the large reservoir base can hold a wide range of bulky attractant material such as rodent litter (Le Goff et al. 2017) or bottles of CO2-producing yeast-sugar mixtures (Jerry et al. 2017). In targeting gravid mosquitoes, the MMT's larger base can hold liquids for a much more extended period, similar to an AGO. As compared with the smaller GAT or Frommer-GT, the large reservoir volume allows for longer-term operation of the trap. The smaller reservoir accommodates self-contained attractants such as BG-lure as demonstrated in this study, octenol cartridges, or similar small volume odor attractants. As an adaptable trap, the MMT's 3D printed intake manifold can be modified to accept new attractants that do not physically fit into the current trap design.
The updraft suction design used in the MMT is similar to previously used updraft designs (Wilton and Fay 1972, Kimsey and Chaniontis 1984, Silver 2008). Updraft suction designs collect mosquitoes as they fly upward in response to the air current (Wilton and Fay 1972). Additionally, the robust ABS plastic collection chamber protects specimens both in the field and during transportation, similar to the Silva trap (Silva et al. 2019). The mesh nets used in downdraft traps potentially damage specimens leading to difficulties in species identification. The high quality of specimens collected in the MMT, many of which were still alive after a 24-h sampling period, can be used for disease detection, morphological, behavioral, and other live insect studies.
Like many suction traps, a 6-volt battery powers the MMT for a 24-h mosquito sampling. The use of 6-volt batteries, instead of 12-volt batteries, is an advantage. Six-volt batteries are lighter, charge faster, and are not restricted from travel on passenger airplanes, all challenges with 12-volt batteries.
Trap durability is another limitation during deployment, as suction traps are often damaged. While there was no testing for MMT environmental resilience, it does use a 25% rectilinear infill support structure for trap components in the PETG 3D printing process. The infill support structure creates durable components while reducing each print's thermoplastic volume (Fernandez-Vicente et al. 2016).
It takes time in training and experience to become proficient at printing the 3D components. New users of 3D printing technology should expect the failure of their initial prints, and we recommend dedicating a couple of weeks to learning the technology. Thankfully, there are many easy-to-follow online resources available to new users, which can rapidly assist with the learning process. The Cura and Blender software used in this study are free to download; however, there is an associated learning curve in becoming proficient in these software platforms. In addition to a learning curve for the technology, there is a production limitation regarding the length of time 3D components take to print. For example, a single MMT intake takes 13 h to print, which means multiple printers are needed for rapid production. Thermal plastic 3D printers useful in mosquito trap construction can vary in cost, $250–$1000, depending on automation of functions and size of building plate. Thermoplastics are relatively inexpensive, with a 1 kg 2.85-mm spool of PETG costing $30–$40. For perspective, 1 kg of PETG is enough plastic to make all of the required printed components for 3 MMTs. As 3D printing becomes more widely available, faster, less expensive, and simpler to implement, many of these limitations will likely be alleviated.
The current assessed MMT design does not include a light source for attracting mosquitoes during operation at nighttime, limiting specific settings. The CDC-LT is often run without lights to avoid nontarget insects bycatch (O'Brien and Reiskind 2013, Hoekman et al. 2016). Future refinements to the MMT include adding a light source; however, if an immediate light source solution is critical, then merely adding a light-emitting diode (LED) flashlight to the MMT would provide a temporary solution. Other variables that could affect trap performance include the height of capture port as well as the volume versus surface area of a gravid attractant, in part because of the limitations associated with using a manufactured 5-gallon tub as our base.
This study demonstrates that the MMT is an adaptable multipurpose suction trap useful in mosquito surveillance. As a multipurpose trap, the MMT potentially reduces trap costs for surveillance programs due to the low production cost and ability to target multiple species as gravid or host-seeking mosquitoes. Furthermore, the open-source architecture of computer software for trap design allows for a community of interested people who can make easily adopted improvements. Since completing this study, we developed an improved version of the MMT, removing the unnecessary central support column to reduce weight and cost; however, this trap is still untested in a field environment. Future improvements include incorporating a 6-volt LED light strip featuring ultraviolet and color lights for specific mosquitoes, a smaller pore size insect screen, and a component to increase “lightness” or reflectance resembling liquid water on the surface of the trap. The addition of these features will potentially expand the range of hematophagous insects collected by this trap. Additionally, future studies can include surveillance for a broader range of mosquito species, biting midges (Ceratopogonidae), black flies (Simuliidae), sand flies (Phlebotominae), and horn flies (Haemotobia irritans) by adjusting attractants or trap layout.
The MMT demonstrated the ability to collect mosquitoes using multiple attractants, a useful function in remote locations with limited access to commercial surveillance traps, and no access to common attractants such as dry ice. This study provides a tested trap design readily adaptable to the many different insect surveillance applications' specific needs.
Material has been reviewed by the Walter Reed Army Institute of Research. There is no objection to its presentation and/or publication. The opinions or assertions contained herein are the private views of the authors, and are not to be construed as official, or as reflecting true views of the Department of the Army or the Department of Defense.
Department of Entomology and Plant Pathology, Box 7613, North Carolina State University, Raleigh, NC 27695.
U.S. Army Medical Research Directorate—Georgia (USAMRD-G), Walter Reed Army Institute of Research, 16 Kakheti Highway, Tbilisi 0190, Georgia.