The Development and Evaluation of Insect Traps for the Asian Citrus Psyllid, Diaphorina citri (Hemiptera: Psyllidae), Vector of Citrus Huanglongbing

Simple Summary Citrus cultivation is affected in many parts of the world because of a devastating disease, huanglongbing (HLB) or citrus greening. In Florida, nearly all commercial citrus is compromised due to HLB, and the disease has spread to other citrus-growing regions of the United States, California and Texas. In California, testing Asian citrus psyllids (ACPs) for the HLB pathogen has been an essential part of integrated pest management. ACP and HLB surveys are essential for disease management in areas where HLB is not widespread. We developed improved ACP traps that can be deployed in the field along with the standard yellow sticky traps. The reusable traps were designed with Rhinoceros computer software and a 3D printer. These traps can be deployed for several months and provide a dynamic sampling mechanism for an improved disease survey strategy. In the present study, ACPs from the 3D-printed traps are collected in a preservative and appear suitable for HLB testing. The evaluation of traps in Florida and California under laboratory, greenhouse, and field conditions indicates that the 3D-printed traps can capture ACPs with about the same efficiency as the sticky traps. They are easy to handle and provide an important field tool for HLB management. Abstract Citrus huanglongbing (HLB) is a severe problem for citrus cultivation. The disease management programs benefit from improved field tools suitable for surveying the ACP vector (Asian citrus psyllid, Diaphorina citri Kuwayama (Hemiptera: Psyllidae)) and the associated pathogen. In the present study, we utilize three-dimensional (3D) printers and design tools to develop traps that can capture and preserve ACPs. Three novel, 3D-printed traps were designed and evaluated: stem trap, and cylinder traps 1 and 2. The traps and yellow sticky cards were deployed weekly for 8 months in 2 non-commercial citrus groves in Florida; in California, the traps were evaluated for 12 months in field cages and 4 citrus groves. The stem traps captured lower numbers of ACPs at all experimental sites compared to the cylinder traps. Capture rates in the cylinder traps were comparable to the sticky trap, making the device a viable tool for monitoring field ACPs. The two main advantages of using the reusable 3D traps over standard methods of ACP and HLB surveys include dynamic sampling that can be conducted year-round and the capture of ACPs that can be preserved and tested. Improved trapping may facilitate quick management decisions and mitigate HLB.


Introduction
Huanglongbing (HLB) or citrus greening is considered the most serious disease of citrus worldwide, because of the extensive damage to citrus industries. The challenges are groves. The performance of the three trap models was compared with each other and with the standard yellow sticky traps. The 3D-printed traps provide several advantages, including dynamic ACP surveys and the capture of well-preserved ACPs. The information generated may be useful for better management of this economically significant disease.

ACP Traps
We designed insect traps, printed them using a 3D printer, and evaluated them for ACP capture in laboratory and field conditions. The two basic designs described in this work are "stem" traps and "cylinder" traps. All traps included a main body with a yellowcolored exterior to attract ACPs visually. The traps have entry holes either on top of the main body or at different levels on the unit's exterior surface to facilitate the entry of ACPs inside the trap. The interior surface of the trap is slippery and connected at the bottom to a capture vial filled with a preservative. All traps were fitted with a clear plastic dome, about 100 mm in diameter, to allow sunlight to pass through. The light inside the traps attracts ACPs into the dome and then into the trap's body. The interior surfaces were made slippery by applying a milky white fluoropolymer resin suspension of polytetrafluoroethylene ("Fluon", marketed as "Insect-a-Slip"; BioQuip, Inc., Rancho Dominguez, CA, USA). It is essential to have sunlight penetrate the trap through the dome since ACP is attracted to light. We reapplied "Fluon ® " as needed to maintain the slippery texture on the interior surface of the trap. A plastic cover was fitted at the bottom of the cylindrical traps to exclude spiders and other arthropods near the base of the trap. After each experiment, or at least once in eight weeks, the traps were washed with detergent and water to clean the body of the trap, and remove dirt and grime from the clear dome.

Stem Traps
The stem trap (Figure 1) was designed based on our observations that ACP walks along stem-like structures. The traps had 2 rows of 16 equidistant circular holes (11 mm) intersected by 4 mm stem-like projections, which extended from the bottom exterior surface of the trap ( Figure 1A). The projections passed through the center of each entry hole and were connected to the interior surface of the trap. ACPs land on the exterior surface of traps, walk up these stems leading them into the body of the trap. A tripod hanger screwed into the top of the trap body was fitted with a plastic bolt at the center. A rain skirt was printed using a 3D printer and hot-glued onto the dome to prevent rainwater from entering the trap's interior through the entry holes. A 10 mm hole was drilled on top of the clear plastic dome, and a hanging nut was hot-glued on top. The dome was screwed on the top, and a funnel was attached to the bottom of the trap body. The preservation vial caps (Thermo Fisher Scientific, Inc., 2019, Cat. No. 14-956-7E, Pittsburg, PA, USA) were drilled and hot-glued onto the bottom of the trap funnel. Just before the trap deployment, capture vials containing preservation fluid were fitted into the caps. All the components were printed using a 3D printer in the stem traps, except the clear plastic dome at the top and the capture tubes attached to the funnel. A previous version of the stem traps described here was used for the trapping of the tomato psyllid, Bactericera cockerelli [22,24]; the stem trap described in the current manuscript has many novel features and modifications compared to the one described earlier in the tomato psyllid study.

Cylinder Traps
Cylinder traps were assembled using commercially available parts in combination with custom-made funnels printed using a 3D printer ( Figure 2). The external surface of the traps consisted of clear, cylindrical, plastic postal tubes (three-inch diameter, obtained from Uline Inc., Catalog #S-14121, Pleasant Prairie, WI, USA). These were cut into eightinch-long pieces and sanded (3M Center, 100 grit: 332U, 320 grit: 738U, St. Paul, MN, USA) to facilitate better paint adhesion. Sanded tubes were treated with yellow spray paint (Rust-Oleum, 7747, sunburst yellow, Vernon Hills, IL, USA), fluorescent paint containing barium sulfate (Krylon, 3104, fluorescent lemon-yellow, Cleveland, OH, USA), and a liquid rubberized coating (Clear Flex Seal, Swift Inc., 2019, Weston, FL, USA). This created a treated surface on which ACP could easily move (monitored in a laboratory setting). The design features of the 3D printed funnel include (1) entry holes, (2) overhang, and (3) funnel. Twenty-four equidistant entry holes (4 mm × 4 mm) allowed adult ACPs to enter the trap funnel. Due to the size of the orifice, larger insects that could not squeeze through the holes were eliminated from the capture tubes. The rim of the trap extruded downward 20 mm, creating a 4 mm thick overhang that provided shade for the entry holes and reduced the amount of light that entered the trap from below. The 3D-printed funnel was designed to pressure fit into the painted cylinder, and a clear plastic dome was hot-glued on top. A small hole was drilled into the top of the transparent dome, and a small ring of metal wire was hot-glued on the top to facilitate the deployment of the trap on a citrus tree.

Cylinder Traps
Cylinder traps were assembled using commercially available parts in combination with custom-made funnels printed using a 3D printer ( Figure 2). The external surface of the traps consisted of clear, cylindrical, plastic postal tubes (three-inch diameter, obtained from Uline Inc., Catalog #S-14121, Pleasant Prairie, WI, USA). These were cut into eightinch-long pieces and sanded (3M Center, 100 grit: 332U, 320 grit: 738U, St. Paul, MN, USA) to facilitate better paint adhesion. Sanded tubes were treated with yellow spray paint (Rust-Oleum, 7747, sunburst yellow, Vernon Hills, IL, USA), fluorescent paint containing barium sulfate (Krylon, 3104, fluorescent lemon-yellow, Cleveland, OH, USA), and a liquid rubberized coating (Clear Flex Seal, Swift Inc., 2019, Weston, FL, USA). This created a treated surface on which ACP could easily move (monitored in a laboratory setting). The design features of the 3D printed funnel include (1) entry holes, (2) overhang, and (3) funnel. Twenty-four equidistant entry holes (4 mm × 4 mm) allowed adult ACPs to enter the trap funnel. Due to the size of the orifice, larger insects that could not squeeze through the holes were eliminated from the capture tubes. The rim of the trap extruded downward 20 mm, creating a 4 mm thick overhang that provided shade for the entry holes and reduced the amount of light that entered the trap from below. The 3D-printed funnel was designed to pressure fit into the painted cylinder, and a clear plastic dome was hot-glued on top. A small hole was drilled into the top of the transparent dome, and a small ring of metal wire was hot-glued on the top to facilitate the deployment of the trap on a citrus tree.
Plastic self-standing transport tubes with screw caps were utilized to collect and preserve captured insects (Thermo Fisher Scientific, Cat. No. 22-010-1228, Pittsburg, PA, USA). A clear rubber cap (Uline, S-15014, Pleasant Prairie, WI, USA) was pressed to the underside of each cylinder trap to keep the collection vial area clean and free from other arthropods. The vial cap was drilled and hot-glued onto the bottom of the trap funnel.

Cylinder Traps
Cylinder traps were assembled using commercially available parts in combination with custom-made funnels printed using a 3D printer ( Figure 2). The external surface of the traps consisted of clear, cylindrical, plastic postal tubes (three-inch diameter, obtained from Uline Inc., Catalog #S-14121, Pleasant Prairie, WI, USA). These were cut into eightinch-long pieces and sanded (3M Center, 100 grit: 332U, 320 grit: 738U, St. Paul, MN, USA) to facilitate better paint adhesion. Sanded tubes were treated with yellow spray paint (Rust-Oleum, 7747, sunburst yellow, Vernon Hills, IL, USA), fluorescent paint containing barium sulfate (Krylon, 3104, fluorescent lemon-yellow, Cleveland, OH, USA), and a liquid rubberized coating (Clear Flex Seal, Swift Inc., 2019, Weston, FL, USA). This created a treated surface on which ACP could easily move (monitored in a laboratory setting). The design features of the 3D printed funnel include (1) entry holes, (2) overhang, and (3) funnel. Twenty-four equidistant entry holes (4 mm × 4 mm) allowed adult ACPs to enter the trap funnel. Due to the size of the orifice, larger insects that could not squeeze through the holes were eliminated from the capture tubes. The rim of the trap extruded downward 20 mm, creating a 4 mm thick overhang that provided shade for the entry holes and reduced the amount of light that entered the trap from below. The 3D-printed funnel was designed to pressure fit into the painted cylinder, and a clear plastic dome was hot-glued on top. A small hole was drilled into the top of the transparent dome, and a small ring of metal wire was hot-glued on the top to facilitate the deployment of the trap on a citrus tree.
Plastic self-standing transport tubes with screw caps were utilized to collect and preserve captured insects (Thermo Fisher Scientific, Cat. No  Adult ACPs are attracted to the yellow exterior, land on the trap's surface, and move upward towards the entry holes that appear lit from the outer surface of the trap. Once ACPs enter the trap, the viewing angle changes, and because of the added shading mechanism mentioned before, if the ACPs were to look back towards the entry holes, they would appear darkened and hence unappealing. Two versions of the cylinder trap were evaluated in this study. The original cylinder trap 1 was modified into cylinder trap 2 by incorporating three significant changes: (1) entry holes were angled inward at 35-45 • , (2) the bottom interior surface of the external brim structure was curved up and inward to create a sill, and (3) the thickness of the 3D-printed funnel was reduced from 3 mm to 1.5 mm. Reduced filament was required to print the funnel in cylinder trap 2 ( Figure 3). The sill designed at the bottom of the brim helps land ACPs that may find their way out from the trap. Landing allows ACPs to reorient and climb back up into the trap. In the cylinder traps, only the funnel part with the entry holes was printed using a 3D printer; the rest of the parts of the trap were purchased from commercial sources. Adult ACPs are attracted to the yellow exterior, land on the trap's surface, and move upward towards the entry holes that appear lit from the outer surface of the trap. Once ACPs enter the trap, the viewing angle changes, and because of the added shading mechanism mentioned before, if the ACPs were to look back towards the entry holes, they would appear darkened and hence unappealing.
Two versions of the cylinder trap were evaluated in this study. The original cylinder trap 1 was modified into cylinder trap 2 by incorporating three significant changes: (1) entry holes were angled inward at 35-45°, (2) the bottom interior surface of the external brim structure was curved up and inward to create a sill, and (3) the thickness of the 3Dprinted funnel was reduced from 3 mm to 1.5 mm. Reduced filament was required to print the funnel in cylinder trap 2 ( Figure 3). The sill designed at the bottom of the brim helps land ACPs that may find their way out from the trap. Landing allows ACPs to reorient and climb back up into the trap. In the cylinder traps, only the funnel part with the entry holes was printed using a 3D printer; the rest of the parts of the trap were purchased from commercial sources.

3D Design and Printing
Traps were designed using Rhinoceros, a computer-aided design software to produce a stereolithography (STL) file (Robert McNeel & Associates 2019, Seattle, WA, USA). This was converted to a G-code using Cura LulzBot Edition slicing software (Aleph Objects, Inc., 2019, Loveland, CO, USA) and printed using LulzBot Taz 5 and 6 fused depositional desktop 3D printers (Aleph Objects, Inc., 2019, Loveland, CO, USA). Polylactic Acid (PLA) plastic filament (Polymaker LLC, 2018, PolyLite™, Bluffton, SC, USA) in a "True Yellow" shade was selected for printing as it most closely matched the yellow-colored traps successfully used in other ACP trapping studies [27,28].

Field Trials in Florida
Field experiments in Florida were conducted in two non-commercial citrus groves: Ocklawaha (29.037281° N, 29.037013° W) and Merritt Island (28.464417° N, 80.712989° W).

3D Design and Printing
Traps were designed using Rhinoceros, a computer-aided design software to produce a stereolithography (STL) file (Robert McNeel & Associates 2019, Seattle, WA, USA). This was converted to a G-code using Cura LulzBot Edition slicing software (Aleph Objects, Inc., 2019, Loveland, CO, USA) and printed using LulzBot Taz 5 and 6 fused depositional desktop 3D printers (Aleph Objects, Inc., 2019, Loveland, CO, USA). Polylactic Acid (PLA) plastic filament (Polymaker LLC, 2018, PolyLite™, Bluffton, SC, USA) in a "True Yellow" shade was selected for printing as it most closely matched the yellow-colored traps successfully used in other ACP trapping studies [27,28].

Field Trials in Florida
Field experiments in Florida were conducted in two non-commercial citrus groves: Ocklawaha (29.037281 • N, 29.037013 • W) and Merritt Island (28.464417 • N, 80.712989 • W). We used a randomized complete block design and designated eight blocks for trap assessments at each of the two sites. The four treatments consisted of three 3D-printed prototypes (stem trap, cylinder traps 1 and 2) and yellow sticky cards, used as a standard (13 cm × 18 cm; Alpha Scents Inc., West Linn, OR, USA). Within each block, 5 m treatment plots were designated. Traps were rotated through each of the four plots during subsequent replications. Traps were hung at breast height and deployed on trees having new flush growth, observable ACP activity, and optimal sunlight during daytime hours.
In Florida, the first set of field experiments were conducted using the stem trap, cylinder trap 1, and yellow sticky cards, deployed at 5-9 day intervals between 1 May 2018 and 26 June 2018. Capture vials were monitored after each experiment, and the number of ACP trapped was documented. Clear plastic food wrapping was folded and cut around sticky cards to contain the gummy substance during transport. The number of ACP trapped during each deployment was counted from the sticky cards and preservation vials for Merritt Island (n = 48) and Ocklawaha (n = 72). In the 2nd experiment, 4 types of traps, including the cylinder trap 2, were deployed weekly for 5-9 days between 17 July 2018 to 5 December 2018 at Merritt Island (n = 64) and Ocklawaha (n = 128).
Data analysis was performed using R statistical computing software (R Core Team, 2018, stats v3.5.2). The null hypothesis was that all traps collect the same numbers of ACP. The data generated follow a Poisson distribution. The Poisson means (ACPs captured per trap, per evaluation period) were compared within each site using the Poisson test [29]: . The field trees inside the cages were pruned periodically to keep the plant size small. Clean ACPs were maintained inside the field cages for research projects. The field cages were equipped with a double-door entry system to prevent ACPs from escaping. This provided an ideal setting for conducting pilot studies using newly designed traps. Each field cage was estimated to contain about 20,000 to 100,000 ACPs (negative for CLas). In each trial, we evaluated three types of traps (stem trap, cylinder traps 1 and 2) in the field cages for four weeks. A total of six trials was carried out from April 2018 to May 2019. For each trial, we utilized 2 field cages and deployed 48 traps. Sixteen traps of each type were used and positioned at the height of about three-to-six feet, deployed around the inner perimeter of the cages. Each tree had four traps hanging from the branches inside the field cages. Since the comparison was between the three trap models previously selected in Florida, we did not utilize yellow sticky traps in these experiments. At the end of each trial, the traps were removed from the trees, and the number of ACPs captured was recorded. The mean number of ACPs captured in each trap for the duration of the trial were compared using a one-way ANOVA and Tukey's HSD (honestly significant difference test; p ≤ 0.05) [30].

Trap Evaluations in Citrus Groves in Temecula, CA
We conducted field evaluations of cylinder trap 1 in four citrus orchards in the Temecula area of southern California (Riverside County). After the initial visual observation of ACP activity on the citrus trees, four sites were chosen. Location 1 had grapefruit trees and was situated next to an organic grove. Location 2 had lemon trees; many abandoned citrus trees surrounded the grove. Locations 3 and 4 had tangerines (cultivar "Tango") and sweet oranges (cultivar "Valencia"), respectively, and well-managed groves surrounded the selected experimental blocks. Six trials were conducted in locations 1 and 2 between March 2019 and February 2020, and three trials were conducted in locations 3 and 4 between September 2019 and February 2020. Each trial was carried out using 16 traps (cylinder trap 1) deployed on trees in border rows and 8 yellow sticky traps used as controls. We did not conduct trap evaluation trials for about six weeks when pesticides were applied in any of the test sites. The ACP capture vials were collected at the end of each trial and brought to the laboratory for documentation. Yellow sticky traps were covered with plastic wrap before transporting and brought to the lab for counting the trapped ACPs. The mean ACP capture from each trap type at each location was compared using the Mann-Whitney rank sum test [31]. Cylinder traps were washed with detergent and water before the next deployment.

By-Catch
Non-target insects or by-catch for the 3D-printed stem traps and cylinder trap 1 were tallied for the California field trials, but not for the field trials conducted in Florida. Major arbitrary categories tallied captured insects; most of the Hemiptera were identified to the genus or species level. We did not document the by-catch for the yellow sticky cards.

Field Trials in Merritt Island and Ocklawaha, Florida
Data from the first set of trials conducted between 1 May 2018 and 26 June 2018 suggest that there was no difference between the mean number of ACPs trapped by the sticky cards (10.1) and the mean number of ACPs captured by the cylinder trap 1 (9.6) in Ocklawaha (p = 0.300) (Figure 4). The same results were not found in Merritt Island, where the mean number of ACPs trapped by the sticky cards (7.5) was greater than the cylinder trap 1 (6.2) (p = 0.016). However, because the 95% confidence intervals calculated for the ratio of these two means are close to one, the means are similar. Conversely, the 95% confidence intervals calculated for the ratio of the mean number of ACPs trapped by the sticky trap (10.1, 7.5) and the stem trap (1.6, 0.9) are not close to one in both Ocklawaha and Merritt Island, and thus the means are very different. Additionally, the mean number of ACPs trapped by cylinder trap 1 (9.6, 6.2) was greater than the mean for the stem trap (1.6, 0.9) in both locations (p < 0.001).   to reject the null hypothesis and conclude that the ratio of the two means associated with the sticky card (5.5, 16.1), and cylinder trap 1 (4.2, 11.3) is greater than one for the traps deployed in Ocklawaha (p < 0.001) and Merritt Island (p < 0.001). Therefore, the mean number of ACPs trapped by the sticky cards was greater than cylinder trap 1. However, because the 95% confidence intervals calculated for the ratio of these two means are close to one, the means are similar. As in previous data, the 95% confidence intervals calculated for the ratios of the mean number of ACPs trapped by the sticky card (5.5, 16.1) and the stem trap (0.8, 2.4) are not close to one. Hence, the means are very different.

Trap Evaluations in Field Cages in Pomona, California
ACPs captured in the Pomona field cage trial were counted, and the data were lyzed by comparison using ANOVA and Tukey HSD. The cylinder traps performed b than stem traps in all the trials conducted. Relative values indicate that cylinder tr Data suggest insufficient evidence to reject the null hypothesis, thus the ratio of the two means associated with cylinder trap 1 (11.3) and cylinder trap 2 (10.7) is equal to one for the traps deployed in Merritt Island (p = 0.274). Therefore, the mean number of ACPs trapped by cylinder trap 1 equals cylinder trap 2 in Merritt Island. The same cannot be said for Ocklawaha, where the ratio of the two means associated with cylinder trap 1 (4.2) and cylinder trap 2 (3.4) is not equal to one (p = 0.004). However, because the 95% confidence intervals calculated for the ratio of these two means are close to one, the means are similar. Finally, the means for the number of ACPs trapped by cylinder trap 1 (4.2, 11.3) and cylinder trap 2 (3.4, 11.3) were greater than the mean for stem trap (0.8, 2.4) ( Figure 5). The Poisson means were compared within each site using the Poisson test.

Trap Evaluations in Field Cages in Pomona, California
ACPs captured in the Pomona field cage trial were counted, and the data were analyzed by comparison using ANOVA and Tukey HSD. The cylinder traps performed better than stem traps in all the trials conducted. Relative values indicate that cylinder traps 1 and 2 performed equally well and were superior to the stem traps (Supplementary Table S2, Figure 6).  Figure 6).

Evaluation of Cylinder Trap 1 in Citrus Groves in Temecula, CA
For the Temecula field trial evaluations, we used cylinder trap 1 along with yellow sticky traps. The stem traps were determined to be inferior to cylinder traps in previous evaluations, and hence not included in the field trial. Since cylinder traps 1 and 2 had very similar capture rates, only cylinder trap 1 was used. The results indicate that the performance of cylinder trap 1 is equivalent to the yellow sticky trap in all four locations ( Figure  7). In locations one and two, the border row adjoining the organic grove or abandoned citrus had more ACPs than traps placed at other locations. However, when the means were calculated, the performance of cylinder trap 1 was at least equivalent to the yellow sticky trap.

Evaluation of Cylinder Trap 1 in Citrus Groves in Temecula, CA
For the Temecula field trial evaluations, we used cylinder trap 1 along with yellow sticky traps. The stem traps were determined to be inferior to cylinder traps in previous evaluations, and hence not included in the field trial. Since cylinder traps 1 and 2 had very similar capture rates, only cylinder trap 1 was used. The results indicate that the performance of cylinder trap 1 is equivalent to the yellow sticky trap in all four locations (Figure 7). In locations one and two, the border row adjoining the organic grove or abandoned citrus had more ACPs than traps placed at other locations. However, when the means were calculated, the performance of cylinder trap 1 was at least equivalent to the yellow sticky trap.

By-Catch
In addition to ACPs, the 3D-printed traps (both stem traps and cylinder traps) also captured other arthropods under field conditions. The type of by-catch varied depending on the location, season, weather pattern, pesticide application status, and presence of other plant types adjacent to the citrus grove. The majority of the by-catch consisted of parasitic Hymenoptera, small Diptera, ants, thrips, and small beetles ( Table 1). Many of the Hemiptera were identified at least to the genus level. Psyllids other than ACPs collected in the early prototypes of stem traps included Acizzia uncatoides (Ferris and Klyver); Acizzia spp.; Bactericera maculipennis (Crawford); Calophya schini Tuthill; Calophya spp.; Ctenarytaina spatulata Taylor; Glycaspis brimblecombei (Moore); and two unknown species (Supplementary Table S3 Tables S3 and S4). evaluations, and hence not included in the field trial. Since cylinder traps 1 and 2 had very similar capture rates, only cylinder trap 1 was used. The results indicate that the performance of cylinder trap 1 is equivalent to the yellow sticky trap in all four locations ( Figure  7). In locations one and two, the border row adjoining the organic grove or abandoned citrus had more ACPs than traps placed at other locations. However, when the means were calculated, the performance of cylinder trap 1 was at least equivalent to the yellow sticky trap.  Cylinder traps were deployed in the Temecula area (Riverside County, California) from 2019-2020 (n = 71). Diaphorina citri Kuwayama (ACPs) were identified and counted; other insects were assigned to arbitrary categories, mostly insect orders. Many Hemiptera were identified at a genus or species level (Supplemental Tables S3 and S4). Diaphorina citri numbers are included for comparison.

Cost of 3D-Printed Traps
In our prototypes, cylinder traps required about 35 g of 3D-printer plastic filament. The cost of producing a cylinder trap, including the price of clear domes, tubes, end caps, collection vials, wire, spray paints, other surface treatments, and the cost of labor, was about USD 16.20 per trap. The stem trap required more 3D-printed components than the cylinder traps (about 200 g). The stem trap could be produced at USD 22.80 per trap.

Discussion
Citrus HLB caused significant financial damage to citrus industries worldwide. Since there is no known cure for HLB, chemical control of the ACP vector is the primary mode of disease control. The early detection of HLB-associated pathogens is key to mitigating the spread of HLB, especially where the disease is not endemic. In California, the disease is not widespread and hence the industry is still in the disease exclusion mode. Infected trees are removed constantly to decrease the inoculum and prevent further spread of the disease. Enormous resources are utilized for coordinated area-wide monitoring of the ACP vector, and treatment programs for effective disease mitigation follow these.
For a successful integrated pest control program, population monitoring of the insect vector is necessary, since data on pest abundance and spatial distribution are required for effective disease management. Because of the importance of ACP surveys, extensive research has been directed at developing robust monitoring techniques. The presence and abundance of all ACP life stages in citrus trees can be determined by visual inspections of flush shoots [32]. For assessment of adult populations, stem tap sampling is recommended [33]. Other methods employed include sticky traps [27,28,[34][35][36], vacuum sampling [37], and sweep nets [38]. Based on comparative studies of the sampling methods reported earlier, yellow or lime-green sticky cards are considered the most effective in monitoring ACP populations, especially at a low density [27,35,36,39].
Yellow sticky traps are used as a standard method for monitoring the insect population in a grove. Integrated pest management programs to suppress ACPs involve the constant monitoring of ACPs, generally accomplished by a combination of yellow sticky cards and tap sampling [40]. Data collected from sticky traps are used for planning pesticide spray regimens to control various insects. However, if an ACP is present in a grove, it would be beneficial to preferentially capture the ACP and test it for the presence of HLB-associated CLas. In many disease-survey programs, ACPs are captured manually using aspirators or tap sampling, stored in vials containing alcohol, transported to a laboratory, and used for DNA extractions and testing for the pathogen [25]. While this is an effective method to obtain suitable quality DNA that can be used for molecular testing, manual capture of ACPs is not an efficient process. In addition, the manual method of capture provides pest status in the grove only for the days when ACPs are captured. In an extensive study conducted from samples collected in Pakistan where the high temperatures can reach up to 48 • C for brief periods during the summer, ACPs are generally devoid of the pathogen, as a result of the high temperatures [41]. During extreme weather conditions, the pathogen resides in the host plant's root system; as normal temperatures resume, the pathogen recolonizes the canopy and can be acquired by the ACP feeding on the plant. For an efficient HLB survey, capturing the ACP vectors during different seasons and evaluating the disease situation is essential.
Inexpensive, effective, and widely available traps improve the efficiency of ACP surveys. This is increasingly important in areas that are not yet widely affected by the disease, but where disease introduction is heavily monitored. Our results demonstrate that 3D-printed cylinder traps can attract and capture ACPs, and, compared to the yellow sticky card, these traps are efficient in preserving captured ACPs (Figure 8). At some sites, cylinder traps had trapping rates comparable to the yellow sticky cards, making these traps valuable for ACP population monitoring and disease surveys. In pilot studies, we reported that propylene glycol with salt (NaCl) was suitable for ACP preservation and pathogen detection in stored ACPs for up to six weeks [26]. Using a previous model of the stem trap designed and printed by the Division of Plant Industry, similar studies were conducted with the tomato psyllid, Bactericera cockerelli; the psyllids were well preserved and suitable for pathogen testing [24].
The 3D-printed trap units can be deployed strategically and in great numbers to cover a large sampling area [42]. Sample vials can be collected from the traps easily and replaced periodically. This reduces the amount of time an inspector has to spend at any given location and maximizes the total area that can be sampled per inspection cycle.
Using 3D printers was essential to the success of this project since it is possible to develop trap prototypes quickly, make required modifications, and evaluate improvements. Initial trap designs were based on data collected by observing ACP behavior in the laboratory setting. Later models incorporated other details, such as the identification of the optimal surface for ACP movement, selection of materials that are non-toxic to the environ-ment, ease of handling, improvement of surface texture so that the ACPs attracted to the trap can move freely on the outer surface of the trap, and the development of a slippery surface to promote movement of captured insects into the collection vial. The development of traps suitable for efficient ACP capture in the field is an iterative process. We described three trap models in this communication. Several modifications were made to the original prototypes to design a trap that is easy to deploy; has apertures large enough for ACPs, but too small to allow larger non-target insects to enter the trap body; is appealing to the target insects; and is durable, using material that the ACP prefers. Preventing other insects from inhabiting the crevices of the trap was incorporated into the later models. The traps can be utilized for other insects and modified depending on the intended use, mainly the type of target insect. The exclusion of specific arthropods may be accomplished by custom modifications, such as the aperture size and other parameters, which may vary depending on the target insect. However, 3D printing performed in a laboratory setting is not be suitable for large-scale production. Once a final design is selected, printing jobs can be outsourced to commercial companies to reduce the price per unit. Since 3D printing is becoming more popular and cheaper, the cost per unit will be significantly lower in the future [43].  The 3D-printed trap units can be deployed strategically and in great numbers to cover a large sampling area [42]. Sample vials can be collected from the traps easily and replaced periodically. This reduces the amount of time an inspector has to spend at any given location and maximizes the total area that can be sampled per inspection cycle.
Using 3D printers was essential to the success of this project since it is possible to develop trap prototypes quickly, make required modifications, and evaluate improvements. Initial trap designs were based on data collected by observing ACP behavior in the laboratory setting. Later models incorporated other details, such as the identification of Panel (B) (inset) shows the capture tube from the trap with insects. Panel (C) shows 28 ACPs (top Petri dish) and by-catch (bottom Petri dish) captured in 1 trap during a 6-week period. Psyllids were separated in the lab. ACPs were clearly recognizable six weeks after capture since they were not degraded. Panel (D) shows a yellow sticky trap on an adjacent tree; a magnifying lens was required to identify and count the 20 degraded psyllids on the sticky trap.
Other researchers have utilized 3D-printed traps for various arthropods [22,[44][45][46][47]. However, as with any design and manufacturing process, 3D printing does have limitations. Printing times for a standard FDM 3D printer are relatively long, as a single object can take hours to print entirely. Objects can frequently fail to print for many reasons. Print quality and print resolutions can vary significantly among different printer models. Plastic filaments sometimes are printer specific and may be sold in limited colors. We utilized a thermoplastic high-strength polymer, polylactic acid. The filaments selected for this study had a relatively low melting temperature, around 185-190 • C [48].
The stem trap design did not capture ACPs at a rate near to that of the sticky card or cylinder traps. While the stem structures may entice and direct ACPs to walk into the trap, ACPs can just as easily turn around and walk out of the trap. The stem trap is also prone to water intrusion, which reduces the effectiveness of the preservative. The stem trap requires more 3D-printed components than the cylinder traps and requires 12 h of printing and 200 g of plastic filament per trap, increasing the production costs. Due to the complex geometries associated with the stem trap design, it would be unsuitable for mass manufacturing techniques, such as injection molding.
To utilize the cylinder trap effectively for ACP surveys, enough traps should be deployed. Secondly, the traps should be cost-effective. Cylinder trap production can be escalated to reduce trap costs, since effective ACP surveys require the use of a sufficient number of traps per acre. The required numbers may vary according to disease incidence, ACP populations, management practices, etc. In Southern California, where the ACPs have established in citrus groves, 5-16 yellow sticky traps are placed per square mile (5 if a few citrus trees are present and 16 if the density of citrus is high) [25]. Alternate technologies, such as mold injections, can be utilized for producing a large number of traps. All trap components could be mold injected, including the plastic tubes and clear domes. The color of the plastic tubes used in the traps can be matched to a specific spectral signature that is most attractive to ACPs, eliminating the need to hand paint each tube.
Sétamou et al. [20] developed mesh-covered sticky traps to reduce the by-catch of other organisms. Yellow sticky traps often capture many non-target species in addition to small lizards, leaf material, and other debris in the grove. In the 3D-printed traps utilized in the current study, non-target species were monitored and documented (Table 1, Supplementary  Tables S3 and S4). The capture of non-target insects was less of an issue with 3D traps than with sticky traps [22].
For oriental fruit flies (Bactrocera dorsalis) and melon flies (B. cucurbitae), methyl eugenol (4-allyl-1,2-dimethoxybenzene-carboxylate) and cue lure [4-(p-acetoxyphenyl)-2-butanone], respectively, are used as effective kairomone lures in traps [49]. Such effective attractants are not known for ACPs. Other insect attractants that may increase capture rates include semiochemicals [50], lighting [51][52][53][54] auditory devices [55][56][57][58][59]. Despite research to find odorants attractive to ACP adults [60,61], there are still no effective semiochemicals for luring ACP adults in field population studies. Acetic acid has been reported to increase trap catches of both adult males and females, presumably acting as an ACP pheromone [62]. However, other researchers are yet to confirm the field efficacy of acetic acid. The chemosensory organs on the antennae of ACPs are depauperate [63]. Thus, yellow or lime-green sticky cards that serve as visual attractants for ACPs remain the gold standard in the population studies [27,28,40]. Many insects exhibit positive phototaxis and respond to different wavelengths of light [51]. We observed a similar trend with 3D-printed traps placed at the different sides in the field cages in Pomona, CA. Sunlight positively influenced the capture rate of adult ACPs in the traps. Further research with ACPs may lead to the identification of the correct light conditions for increasing ACP capture in 3D-printed traps.

Conclusions
We developed two basic models of 3D-printed traps that can be used for ACP capture. The cylinder traps were more efficient in ACP capture than the stem traps. The 3D-printed traps are reusable, and the capture of non-target species in the traps is low compared to the sticky traps. Cylinder traps are at least as efficient as sticky traps and can be combined with sticky traps in routine ACP surveys. Since the 3D-printed traps can be deployed on a tree for an extended period, the captured insects represent dynamic sampling over a period of time. One of the major advantages of using a 3D-printed trap is the ability to test the captured ACPs for the presence of the HLB pathogen, CLas. Unlike sticky traps that can only be used for ACP surveys, 3D traps can also be used for HLB surveys since the captured psyllids are suitable for DNA-based testing.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/insects13030295/s1, Table S1: Field trials in Florida: a comparison of the different types of traps. Table S2: A comparison of the traps in the Pomona field cage experiment. Table S3: Identified Hemiptera (excluding Diaphorina citri Kuwayama) collected in early prototype stem traps deployed in Riverside and Santa Paula, CA, 2016-2017. Table S4  Data Availability Statement: Open-source files used for printing the cylinder traps will be made available on the Division of Plant Industry website after publication is accepted (https://www.fdacs.gov/ Agriculture-Industry/Science/3D-Printing-Lab, accessed on 13 March 2022).