Simpler is better: fewer non-target insects trapped with a four-component chemical lure vs. a chemically more complex food-type bait for Drosophila suzukii

Introduction

Fermented food type lures such as wine, vinegar, beer, and fermenting fruit/sucrose/yeast are broadly attractive to a diversity of insects, including moths (Yamazaki, 1998; Sussenbach & Fiedler, 1999; Laaksonen et al., 2006; Pettersson & Franzén, 2008), vespid wasps (Dvorak & Landolt, 2006; Noll & Gomes, 2009; De Souza et al., 2011; Landolt et al., 2014), and natural enemies of crop pests (Aluja, 1999; Thomas, 2003). Whereas these baits are useful for survey or biodiversity studies, they can be detrimental to beneficial insects and difficult to use for the purpose of monitoring a single pest species; numerous non-target insects are captured and have to be sorted. This problem has long been recognized in trapping systems using other attractants as well (e.g., Weber & Ferro, 1991; Meagher & Mitchell, 1999; Yee et al., 2005; Uchida et al., 2007; Leblanc et al., 2010).

The spotted wing drosophila (SWD), Drosophila suzukii (Matsumura) (Diptera: Drosophilidae), was recently introduced and spread throughout much of North America and Europe, where it has become a significant pest of many fruit crops. Traps with baits are widely used to determine SWD presence, and work is underway to incorporate trap protocols into integrated pest management (IPM) programs (Dreves, 2011). These efforts rely heavily on fermented food baits as attractants (Landolt et al., 2012; Lee et al., 2012). Several problems have been noted with the fermented bait traps for SWD, including the difficulty of sorting through large numbers of non-target insects to find and tally SWD (e.g., Lee et al., 2012; Cha et al., 2013). This process adds time and cost to any trapping program.

Using a simpler and more defined chemical attractant rather than a chemically complex food material may have the practical benefit of a reduced number of non-target insects captured in SWD traps. In locating an attractive odor source, insects often respond to a specific blend of a small subset of the numerous volatile compounds released from the source (Bruce et al., 2005; Cha et al., 2008). We hypothesize that different insects may be sensitive to different subsets of volatile compounds from such materials, and that a blend isolated and optimized for one species may not be optimally attractive to other insects attracted to the original source. Thus, a chemically complex odor such as a fermented food-type bait is likely to produce volatile cues that are attractive to numerous insect species, as opposed to a simple chemical blend optimized for a single species. The attraction of SWD to fermented food baits appears to be due to a relatively small fraction of the volatile chemicals emitted by the bait. Cha et al. (2012, 2014) isolated and identified a four-component chemical lure from the headspaces of Merlot wine and rice vinegar, and showed that it is comparable to the original wine/vinegar bait in attractiveness to SWD (Cha et al., 2013, 2014). In these studies, 16 headspace compounds were identified as eliciting responses from SWD antennae, out of a much larger number of compounds emitted from the wine and the vinegar. Only four of the 16 compounds were found to be necessary for SWD attraction. As a result, we predict that this four-component SWD lure may be more selective than the original bait assuming non-target insects that are attracted to wine and vinegar are responding optimally to some other combination of volatile chemicals. Specifically we hypothesize that the SWD chemical lure developed from wine and vinegar odors is less attractive to non-target insects than the wine/vinegar bait.

We report here the results of seven experiments conducted in the states Washington, Oregon, and New York (USA) that tested this hypothesis. In the first four experiments conducted in Washington, we targeted non-drosophilid non-target insect pests that respond to fermented food baits and at times overwhelm SWD monitoring traps; they are known to us in the area and at discrete times of the season. Specifically, we evaluated responses of muscoid filth flies, noctuid cutworm and armyworm moths, and vespid yellowjacket wasps, common non-target insects, to the four-component SWD chemical lure and a wine/vinegar bait. In the following three experiments, we compared responses of SWD and non-target drosophilids to the chemical lure and a wine/vinegar bait in New York, Oregon, and Washington.

Materials and methods

Results

Experiment 1. Filth flies at a dairy

Large numbers of false stable flies (total 5 140 flies) were trapped. Numbers of these flies were 5.2-fold higher in traps baited with wine/vinegar compared to the SWD chemical lure (F_1,9 = 72.17, P<0.0001; Figure 1A). Although there were fewer little house flies captured (total 133 flies), the pattern was similar, with 66% more flies in traps baited with wine/vinegar than in traps baited with the SWD lure (F_1,9 = 10.72, P = 0.0096).

Experiment 5. Drosophilids at a cherry orchard in Washington

Totals of 1 594 female and 375 male SWD and 18 733 non-target drosophilids were captured (9.5% SWD and 90.5% non-target drosophilids). Numbers of male and female SWD in traps baited with wine/vinegar were not different compared to numbers in traps baited with the SWD chemical lure (females: F_1,9 = 0.91, P = 0.37; males: F_1,9 = 1.89, P = 0.20; Table 1, Figure 2A). In contrast, numbers of total non-target drosophilids captured with wine/vinegar were 7.6-fold and significantly greater than numbers in traps baited with the SWD chemical lure (F_1,9 = 50.79, P<0.0001; Figure 2A). Among the six non-target drosophilid species captured, D. melanogaster species group, D. obscura species group, and D. immigrans accounted for 57.6, 41.2, and 1.1%, respectively, totaling 99.9% of all the non-target drosophilids captured. Numbers of the D. melananogaster and D. obscura species groups and D. immigrans in traps baited with wine/vinegar were 15.7-, 4.3-, and 3-fold, respectively, greater than numbers in traps baited with the SWD chemical lure (D. melanogaster: F_1,9 = 71.83, P<0.0001; D. obscura: F_1,9 = 29.14, P = 0.0004; D. immigrans: F_1,9 = 8.45, P = 0.017). Numbers of D. busckii, D. hydei, and Chymomyza spp. trapped were low, with no differences between the two trap treatments (for all three comparisons P>0.08; Table 1).

Discussion

Our results demonstrate that significantly fewer non-target insects are attracted to the four-component SWD chemical lure (acetic acid, ethanol, acetoin, and methionol) compared to its original source material of wine/vinegar (Cha et al., 2012, 2014). For non-Drosophilidae, this reduction was consistent and often striking. As predicted, traps baited with the SWD chemical lure consistently captured lower numbers of pestiferous filth flies, cutworm and armyworm moths, and yellowjackets compared to the traps baited with wine/vinegar, and in a variety of agricultural landscapes. All of these non-target insect species are seasonally but routinely abundant in temperate agricultural landscapes, with the potential to respond and at times overwhelm SWD monitoring traps.

Generally, non-target drosophilid flies also responded less to the chemical lure than to the trap bait. Three species or species groups (D. melanogaster species group, D. obscura species group, and D. immigrans) that made up the bulk of the non-target drosophilid trap catch were trapped in much greater numbers with the food bait than with the chemical lure; the exception being the D. obscura species group and D. immigrans in Oregon. Additionally, D. putrida, D. simulans, D. tripunctata, and Chymomyza spp. were reduced in traps with the chemical lure. Overall, non-target Drosophilidae were reduced 84.7, 70.3, and 37.2% in the chemical-lure traps over food-bait traps in Washington, New York, and Oregon, respectively. In no cases were the numbers of non-target drosophilid species captured with the SWD chemical lure greater than the numbers captured with wine/vinegar bait. Taken together, our results support the hypothesis that a feeding attractant that is isolated, identified, and optimized for a single pest species is chemically simpler and more specific to the target insect compared to the attractive starting material. We also conclude from this work that the chemically defined SWD lure should be more useful than a food bait as a monitoring tool for SWD because of reduced time and expertise required to maintain traps and sort trap catches.

Of the drosophilids captured in Washington, Oregon, and New York, three species or species groups (D. immigrans, D. melanogaster species group, and D. obscura species group) were dominant and present at all three sites. The responses of these three species or species groups in terms of greater attraction to wine/vinegar than the SWD lure were similar in Washington and New York. However, in Oregon, this was true only for D. melanogaster species group and no differences between treatments were observed for D. immigrans and D. obscura species group. Also, more Chymomyza spp. flies responded to wine/vinegar than to the SWD lure in New York, but not in Washington. These discrepancies could be due to conditions specific to different locations, such as differences in abundance and physiological status of flies, crop phenology and species, presence and absence of competing resources and resource signals, or differences in odor perception in locally adapted fly populations. At the Washington site, 90.5% of all drosophilids captured were non-targets, whereas at the Oregon and New York sites 6.2 and 15.3% of all drosophilids captured were non-targets, respectively. At the Washington cherry site, no fresh fruits were available in the surrounding area of the trapping site, whereas at the Oregon blackberry and New York raspberry sites, ripe fresh fruits were available at the time of trapping. More work is needed to determine specific causes for the observed discrepancies between locations. A similar explanation may also apply to why the SWD lure was less attractive than the wine/vinegar bait to SWD in New York, whereas the SWD chemical lure was similar to wine/vinegar in attractiveness to SWD in Oregon and Washington in the current study, and in four USA states (Connecticut, Mississippi, Oregon, and Washington) and a European location (Dossenheim, Germany) in previous studies (Cha et al., 2013, 2014).

Drosophilid composition at the Washington cherry site and Oregon blackberry site was similar with five species in common, with the exception of Chymomyza spp. that were captured only at the Washington site. Two species captured in Washington and Oregon, D. buskii and D. hydei, were not captured in the New York raspberry site. However, a greater diversity of drosophilids was captured at the New York site. Seven species (D. putrida, D. robusta, D. simulans, D. tripunctata, L. maculosa, Scaptomyza spp., and Z. indianus) were not captured in Washington and Oregon, but they were captured in New York. It is not currently known whether the differences in drosophilid communities we observed are due to geographic location (i.e., eastern vs. western USA), fruit species present, or fruit phenology.

Zaprionus indianus, the African fig fly (AFF), is a generalist drosophilid that is native to the Afro-tropical region, was introduced into Brazil by 1999 (Vilela, 1999), and spread rapidly in South America (Leão & Tidon, 2004). AFF was found in Florida, USA, in 2005 (Steck, 2005) and since then has been found in much of eastern North America (van der Linde et al., 2006; Biddinger et al., 2012; Van Timmeren & Isaacs, 2014). This study is the first to report the presence of AFF in New York. As AFF has expanded its range, an increasing variety of fruits is used as larval hosts (van der Linde et al., 2006) and for this reason there is increasing interest in detection and monitoring of this species (Pasini et al., 2013; Epsky et al., 2014), although there is some question as to whether they can act as primary pests in undamaged fruit (Steck, 2005).

Most of the non-drosophilid non-target insects captured in this study were attracted to both wine/vinegar bait and the SWD lure. However, greater numbers, and at times much greater numbers, were captured using the wine/vinegar bait than the SWD lure. This might suggest that these non-target insects were cuing on other wine and vinegar volatiles in addition to all or part of the four volatiles used in the SWD lure. Both F. canicularis and M. stabulans are attracted to fermented molasses or sucrose solutions and ethanol was identified as the sole attractant emanated from a fermented sucrose solution (Hwang et al., 1978). With ethanol as one component of the SWD lure, however, our results suggest that there are other wine and vinegar chemicals that might be involved in the chemical attraction of these flies to fermented sweet baits. Previous research showed that M. configurata, a pest of mustards and canola, and X. c-nigrum, a pest of many vegetable crops, are attracted to the combinations of acetic acid and 3-methyl-1-butanol or acetic acid and 3-methyl-1-pentanol (Landolt, 2010). In addition, V. germanica and V. pensylvanica, stinging hazards to humans, are attracted to combinations of acetic acid and isobutanol, as well as to heptyl butyrate (Davis et al., 1969; Landolt, 1998). Strong trap response of these pestiferous species to wine/vinegar bait suggests that attractive and selective lures for these pests may be possible based on chemicals from the wine and vinegar. This supports the idea that as much as a chemically well-defined attractant is important for IPM, use of chemically complex baits such as fermentation baits may be important for surveillance and attractant development purposes (Brockerhoff et al., 2013).

We recognize that the data for non-target non-drosophilids were collected in agricultural settings where SWD might not be a great concern, such as in dairy, mint, sagebrush, and grass fields. There were two rationales for taking this approach. First, to test the hypothesis on the effect of chemical complexity on non-target attraction, we sought to ensure adequate numbers of these non-target insects at the time of trapping. In other words, a valid test of this hypothesis might not be possible when the trapping potential of an attractive trap is limited by a low insect density. Second, we sought to test the hypothesis at the species level. For example, considering non-target insects at above the species level (e.g., genus, family, order, etc.) could complicate the interpretation, especially when two types of lures attracted similar number of a group of non-target insects (e.g., flies) but with different species composition due to differences in the species-specific response to different lures (Cha et al., 2013). Thus, to test our hypothesis, we selected species of insects that can be identified with confidence, respond to fermented food baits, are abundant as pests, and are known to occur seasonally in the area. It is also true that social wasps, noctuid moths, and muscoid flies are highly mobile, widespread, and generally abundant; thus, they are all likely to be contaminants in SWD traps in many locations.

Interestingly, although not counted at the species level, the results of the non-target non-drosophilids captured at the New York raspberry site still support our conclusion. Specifically, traps baited with the wine/vinegar bait captured significantly more non-target dipteran, hymenopteran, lepidopteran, and coleopteran insects, including some wasp parasitoids, than traps baited with the SWD lure. Taken together, these results suggest that the traps baited with the SWD chemical lure, composed of acetic acid, ethanol, acetoin, and methionol, will likely capture significantly fewer non-target insects than traps baited with more complex fermentation baits. Ultimately, using a combination of more selective chemical lure and more selective trap design (e.g., small trap openings physically reduce the number of larger non-target insects captured; Cha et al., 2013) can improve the cost-to-benefit ratio of monitoring by reducing labor required to sort the trap catches while also decreasing the ecological cost associated with trapping beneficial non-target insects.