MINIMIZING BEE BYCATCH IN JAPANESE BEETLE TRAPS

Native and introduced bees were attracted to and captured in commercially available Japanese beetle, Popillia japonica Newman, traps baited with floral lure components (geraniol, eugenol, phenethyl propionate [3:7:3]). Studies in Rhode Island found that Bombus impatiens Cresson was significantly more attracted to geraniol alone and as a component in a floral lure blend than to either eugenol or phenethyl propionate alone. Xylocopa virginica L. was more selective in being primarily attracted to traps baited with higher amounts of geraniol in 2016. In 2016, 32 baited and eight unbaited traps captured 856 B. impatiens and 124 X. virginica in 11 days of trapping in Rhode Island. In 2017, a total of 3,749 bees were captured in 124 traps in Rhode Island over 56 days and 90 in 32 traps in Ohio over 42 days. Removing geraniol from the floral lure reduced the bee capture but did not significantly reduce the Japanese beetle capture in 2017 and 2018. In 2018, a total of 708 bees were captured in 100 traps in Rhode Island over 56 days, 401 bees were captured in 68 traps in Tennessee over 60 days, and 34 bees were captured in 32 traps in Ohio over 58 days. Removing geraniol from Trécé dual lures significantly reduced bee captures in Rhode Island in 2018. Green, black, brown, and red traps captured significantly fewer bees than clear or standard yellow vane and green cage traps in 2018 in Rhode Island and Tennessee. The results suggest that using all green traps with a lure composed of eugenol and phenethyl propionate and the Japanese beetle male sex pheromone can effectively capture Japanese beetles while minimizing the bycatch of bees.


vi LIST OF TABLES AND FIGURES
This manuscript "Minimizing Bee Bycatch in Japanese Beetle Traps" is prepared for submission to the journal Environmental Entomology.

INTRODUCTION
Traps used to monitor Japanese beetles, Popillia japonica Newman (Coleoptera: Scarabaeidae), are known to also capture nontarget organisms, including bees (Hamilton et al. 1970). The Japanese beetle is an introduced pest that feeds on over 300 species of plants (Vittum et al. 1999  trapping for adult Japanese beetles has been conducted for many years. Over 12,000 traps are deployed and monitored during the Japanese beetle flight season (CDFA 2016 (CDFA 2014, ODA 2016. Traps are used in delimitation surveys by state regulatory 3 agencies after accidental introductions, and in some cases, to certify nurseries as Japanese beetle free. (National Plant Board, 2016). Landscape managers and homeowners also use traps to monitor and/or control Japanese beetles.
Several companies produce lures and traps to monitor and control Japanese beetles. It has been known for some time that Japanese beetle traps and lures will attract and kill bees (Hamilton et al. 1970(Hamilton et al. , 1971Ladd et al. 1974 Our objectives were to: 1) evaluate the attractiveness of the Japanese beetle floral lure blend (geraniol, eugenol, and phenethyl propionate) and each of these components at different amounts to determine which were most attractive to bees, and 2) evaluate lure components and trap colors and designs to determine which would reduce bee captures yet still capture Japanese beetles.  (Ladd et al. 1981) were tested: 1) 0.7 ml geraniol, 2) 1.6 ml eugenol, 3) 0.7 ml phenethyl propionate, 4) 3 ml in a ratio (0.7 geraniol:1.6 eugenol:0.7 phenethyl propionate), and 5) control. All lure components were released from 5 ml pads in 15 ml polycons. The experiment was conducted from 11 -19 August 2016 in an unmowed old field site with more floral resources than the mowed site. Traps (same type as at the mowed site) were checked daily at 0900, 1100, 1300, 1500 h to 6 compensate for higher bee captures and to release them unharmed after a shorter period of capture than at the mowed site. Each time the traps were checked, the bees were identified and counted. Captured bees were released 100 m from the study site.

MATERIALS AND METHODS
Trap collection jars were removed each day after the last collection to prevent traps from collecting bees in the evening.

A total of 3,749 bees were captured in seven experiments conducted in
Rhode Island in 2017. The most common bees captured were Bombus spp. and X.
virginica with 2,440 and 701 captured respectively (Table 1). Bee captures at the Rhode Island study sites in the first four wk period (29 June -24 July 2017) were much lower (645) than bee captures (3,104) in the second four wk period (27 July -24 August 2017). Ninety bees were captured at the Ohio study site (Table 1).

A total of 708 bees were captured across three experiments conducted in
Rhode Island in 2018. The most common bees captured were Bombus spp., other Apidae, and Halictidae with 482, 75, and 75 captured, respectively (Table 2). In Tennessee, a total of 401 bees were captured across two experiments in 2018 and the most common bees captured were Bombus spp. and Halictidae with 299 and 47 captured respectively (Table 2). Thirty-four bees were captured in the Ohio experiment and the most common were Bombus spp. and other Apidae with 14 and eight captures respectively ( Table 2).

Attraction of Bees to Individual Floral Lure Components and the Three
Component Blend at Mowed and Unmowed Sites, 2016. Bombus impatiens was more attracted to geraniol, both alone and as a component in the floral lure blend, than to either eugenol or phenethyl propionate alone at the mowed site (F = 19.74; df = 4,15; P < 0.01) ( Table 3). Xylocopa virginica was also attracted to geraniol, but only at the higher concentrations present in the single-compound treatment; it was not as attracted to the three-component blend (F = 10.46; df = 4,15; P < 0.01) ( Table 3).
The attraction of B. impatiens to geraniol alone was not significantly different from the three component blend or PEP at the mowed site (Table 4). There was also no significant difference in the attraction of X. virginica to geraniol or PEP at the unmowed site (Table 4).

Attraction of Bees and Japanese Beetles to Commercially Prepared Lure
Components, 2017. As in 2016, geraniol alone and as a component of the floral lure blend continued to capture more bees than any other treatments in Rhode Island (F = 4.15; df = 7,24; P < 0.01) ( Table 5). In Ohio, bee captures were not numerous enough to show significant treatment differences (F = 1.41; df = 7,24; P = 0.25) ( Table 6).
Although significant treatment differences were found in Japanese beetle trap captures among the lure components in Rhode Island (F =15.06; df = 7,24; P < 0.01) and Ohio (F = 91.98; df = 7,24; P < 0.01), no significant differences were detected in captures between the eugenol and PEP combination and the three lure component blend (Tables   5 and 6).

Components and the Three Component Blend in Varying Amounts in Mowed
and Unmowed Sites, 2017. There were no significant differences in bee captures at the mowed site where any amount of geraniol was present in a treatment regardless of quantity; however, bee captures in eugenol and PEP alone were also not significantly different from some geraniol containing treatments (F = 5.69; df = 7,24; P < 0.01) (Table 7). At the unmowed site, all geraniol containing treatments captured significantly more bees than any other treatments (F = 24.15; df = 7,24; P < 0.01) 13 (  (Table 9). However, in Tennessee and Ohio, removal of geraniol did not result in lower bee captures (Tables 10 and 11 respectively). All green topped traps captured significantly fewer bees than yellow top and green cage traps regardless of the presence of geraniol in Tennessee (F = 4.68; df = 3,28; P < 0.01) (Table 10) and Ohio (F = 17.35; df = 3,28; P < 0.01) (Table 11). In all three study locations, there was no significant differences in Japanese beetle capture between traps with or without geraniol regardless of the trap color; RI:(F = 0.66; df = 3,28; P = 0.58); TN:(F = 2.81; df = 3,28; P = 0.06); OH:(F = 2.77; df = 3,28; P = 0.06) (Tables 9-11).

Attraction of Bees and Japanese Beetles to Trécé Japanese Beetle
Pheromone Lures, 2018. Japanese beetle traps baited with Trécé Japanese beetle pheromone lures were not significantly more attractive to bees than were unbaited (F = 9.49; df = 3,28; P < 0.01) (Table 12). Yellow topped traps captured significantly 14 more bees than the all green traps regardless of the presence of the Japanese beetle pheromone lure (F = 9.49; df = 3,28; P < 0.01) ( Table 12). The presence of the pheromone lure significantly increased the Japanese beetle captures in yellow topped traps and all green traps with the yellow topped traps capturing more than the all green traps (F = 27.25; df = 3,28; P < 0.01) ( Table 12).  Table 13).

Evaluation of Trap Color and Design (Translucent Funnel) to Reduce Bee
Captures at an Unmowed Site, 2017. There were no significant differences in bee captures in the yellow top and green cage Japanese beetle traps without lures and the yellow bee vane traps (Table 14). Traps with green tops again captured significantly fewer bees than yellow or clear top traps (F = 16.69; df = 6,21; P < 0.01) (Table 14).
Japanese beetles were significantly more attracted to traps baited with the Trécé dual lure than traps without the Trécé dual lure regardless of the trap color or design (F = 34.53; df = 6,21; P < 0.01) ( Table 14).

Evaluation of Trap Color and Design (Glass Funnel) to Reduce Bee
Captures at a Unmowed Site, 2017. The yellow bee vane traps with no lures captured as many bees as the standard Japanese beetle trap with or without the dual lure (F = 10.48; df = 6,21; P < 0.01) (Table 15). Traps with green tops again captured significantly fewer bees than yellow or clear top traps with or without lures (Table 15).
Japanese beetles were significantly more attracted to traps baited with the Trécé dual lure than traps without the Trécé dual lure regardless of the trap color or design (F = 35.98; df = 6,21; P < 0.01) ( Table 15).

Attraction of Bees and Japanese Beetles to Different Color Trécé Japanese
Beetle Traps, 2018. The clear Japanese beetle traps continued to be the most attractive to bees in Rhode Island, capturing significantly more bees than the yellow top and green cage traps (F = 30.18; df = 5,30; P < 0.01) ( Table 17). The yellow topped traps also captured significantly more bees than green, red, black and brown traps in Rhode Island (Table 17). Similar results were recorded in Tennessee where clear traps and yellow topped traps were equally attractive to bees and captured significantly more bees than the green, red, black, and brown traps (F = 9.94; df = 5,30; P < 0.01) (  (Tables 17 and 18).

DISCUSSION
This study found that bycatch of pollinating bees can be minimized by using the floral lure combination of eugenol and phenethyl propionate in the standard ratio  (Tatsuka et al. 1990, Schlotzhauer et al. 1996. Geraniol (along with nerol, citral, nerolic acid, geranic acid, and farnesol) is also a component of Nasanov pheromone, which is used by honey bees (Apis mellifera L.) to find their nest entrance, in swarming and to mark rewarding flowers (Winston 1987, Trhlin andRajchard 2011). Geraniol has been found in the mandibular glands of Andrena helvola (L.) and Andrena labiata (Fabricius) (Tengo and Bergstrom 1976). Because geraniol is a component of honey bee pheromones, it was expected that traps baited with geraniol would capture honey bees; however, Free (1962) found that five drops of geraniol were repellent to honey bee recruits. Free (1962) did not describe the volume of the five geraniol drops, but five drops from a dropper in our lab yielded 0.16 ml. The commercially available Japanese beetle floral lures have 0.76 ml of geraniol, which may explain the apparent repellency of floral lures to honey bees.
Eugenol and phenethyl propionate (7:3) was the standard lure adopted by State and Federal agencies for surveying Japanese beetles for many years (Ladd and McGovern 1980 B. terrestris were found foraging in habitats between 1,500 and 1,750 m from the nest. The explanation for these differences has to do with colony size, that is, larger colonies need to range over a larger area to find sufficient food (Goulson 2010).
Bumble bee species that live in small colonies may be particularly vulnerable where large numbers of Japanese beetle traps are deployed. California annually deploys over 12,000 Japanese beetle traps around airports to intercept Japanese beetles coming in from the eastern U.S (CDFA 2016 We have determined that geraniol is the main component in the Japanese beetle floral lure that is most attractive to native bees and that removing it will not significantly reduce Japanese beetle capture. We also have determined that entirely green, brown, black, or red traps are the least attractive to native bees. Companies and regulators can now offer a 'bee friendly' trap for monitoring and control purposes.