In this project, we examined how a common insecticidal seed treatment (thiamethoxam) influences an important group of non-insect pests (slugs) and their predators in corn and soybeans under no-till management. In particular, we explored whether thiamethoxam can travel up the food chain from seedlings to slugs to their ground beetle predators. In laboratory experiments, slugs were highly tolerant of thiamethoxam, readily feeding on treated corn or soybean seedlings and showing no signs of distress. However, slugs that were fed on neonicotinoid-infused seedlings were poisonous to the ground beetle, Chlaenius tricolor. These results suggested that neonicotinoid insecticides could worsen slug problems by relaxing top-down biocontrol, a hypothesis that we tested in field experiments in 2012. In soybeans, thiamethoxam seed treatment increased slug activity-density, undermined stand establishment, and increased the variability of yield. Seed treatment had mixed effects on predator activity-density and predation on sentinel caterpillars. In corn, seed treatments similarly failed to influence yield, but we did not observe differences in slug or predator activity. However, early season flooding and heavy weed pressure at that field site complicated data collection and interpretation. Our results suggest that neonicotinoid seed treatments may worsen slug problems, especially in soybeans, and that farmers could benefit from using these insecticides in an IPM framework only where they are needed to control target pests. For the outreach component of this project, we shared our results and associated video with over 2,000 agricultural professionals and scientists through conferences, extension meetings, and field days.
No-till farming is an important conservation practice in the Northeast, reducing soil erosion, improving soil quality, and saving fuel. At its roots, the no-till philosophy allows crop fields to better mimic natural ecosystems by decreasing disturbance and allowing plant residues to degrade naturally. By enhancing habitat for natural enemies, no-till also holds potential to improve intrinsic regulation of invertebrate and weed pests (Stinner 1990); however, this potential is rarely achieved because of widespread use of preventative insecticide treatments that disrupt natural enemy communities (e.g. Brust et al. 1985).
The most widely used insecticides in current field crop production are neonicotinoid seed treatments. Since their introduction in the mid-2000’s they have become incredibly prevalent (Jeschke et al. 2011); in fact, local growers have difficulty obtaining untreated seed unless they plan well in advance and make a special request (J. Tooker, pers. obs.). These insecticidal seed coatings are absorbed by germinating seedlings, transported throughout plant tissues, and persist for at least 6-7 weeks. They are applied to almost all corn seeds and an increasing proportion of soybean seeds, contrary to the principle of integrated pest management (IPM) that insecticide treatments should only occur when pest populations exceed economic thresholds. Importantly, early season insect pests in corn and soybeans are notoriously variable in space and time, often failing to materialize. As noted by one extension entomologist, “…widespread ‘prevention’ of an insect problem that does not exist and may not develop is not cost effective and may have unintended consequences” (Steffey 2008).
Unintended consequences of these seed treatments could occur both on and off the farm. For instance, concerns about neonicotinoid toxicity to pollinators have led to their suspension in several European countries (US EPA 2011). Treated seeds can be toxic to omnivorous natural enemies if they feed upon the seeds (Mullin et al. 2005), and can poison aquatic life should they enter waterways (Syngenta 2011). One area of potential concern that has not been explored is the relationship between seed treatments and slugs, major pests of no-till crops in the Northeastern U.S. (Hammond and Byers 2002). Current management options for slugs are very limited and farmers require better options; a survey conducted by our lab revealed that 80% of no-till farmers consider slugs their most challenging pest problem (Douglas and Tooker 2012).
Slugs are not insects (they are mollusks) and a previous study suggests that neonicotinoids are not lethal to slugs at usual doses (Simms et al. 2006). As a result, we hypothesized that slugs could transmit neonicotinoids from crops to natural enemies, such as ground beetles (Barker 2004). Transfer of insecticides to natural enemies via slugs could represent an unrecognized source of toxicity for an important group of predators. In this project, we explored whether neonicotinoids are indeed transferred up food chains, and the consequences of neonicotinoid seed treatments for slug populations under field conditions. Given that many slug predators are generalists that consume multiple prey types, we predicted that their disruption could undermine biocontrol of insect pests and weeds as well as slugs. Finally, we sought to educate farmers about the importance of natural enemies and potential unintended consequences of seed treatments to motivate more interest in an IPM approach.
Barker, GM, ed. (2004) Natural Enemies of Terrestrial Molluscs. CABI Publishing, UK.
Brust, GE, BR Stinner, and DA McCartney (1985) Tillage and soil insecticide effects on predator-black cutworm (Lepidoptera: Noctuidae) interactions in corn agroecosystems. J. Econ. Ent. 78(6): 1389-1392.
Douglas, M. and J. Tooker. 2012. Slug (Mollusca: Agriolimacidae, Arionidae) ecology and management in no-till field crops, with an emphasis on the mid-Atlantic region. Journal of Integrated Pest Management 3(1): C1-C9.
Jeschke, P, R Nauen, M Schindler, and A Elbert (2011) Overview of the status and global strategy for neonicotinoids. J. Ag. Food Chem. 59, 2897-2908.
Hammond, R and R Byers (2002) Agriolimacidae and Arionidae as pests in conservation-tillage soybean and maize cropping in North America. In, ed. Barker, GM, Molluscs as Crop Pests. CABI Publishing, UK.
Simms, LC, A Ester, and MJ Wilson (2006) Control of slug damage to oilseed rape and wheat with imidacloprid seed dressings in laboratory and field experiments. Crop Protection 25: 549-555.
Steffey, KL (2008) Managing insects in high-production soybeans: Forethought or afterthought? p. 15-18 in, Proceedings of the 2008 Illinois Crop Protection Technology Conference, University of Illiois Extension.
Stinner, BR (1990) Arthropods and other invertebrates in conservation-tillage agriculture. Ann. Rev. Ent. 35: 299-318.
Syngenta (2011) Pesticide label for Cruiser Maxx seed treatment.
US EPA (2011) “Colony collapse disorder: European bans on neonicotinoid pesticides.” U.S. Environmental Protection Agency. Accessed May 24, 2011. At: http://www.epa.gov/opp00001/about/intheworks/ccd-european-ban.html
Objective 1: Determine whether insecticidal seed treatments can be transferred from crop plants to natural enemies via slugs.
We completed laboratory experiments to address this objective in 2011 for soybeans and in 2013 for corn. These experiments used the beetle Chlaenius tricolor rather than Pterostichus melanarius (as originally proposed) because preliminary experiments revealed C. tricolor to be a more reliable slug predator.
Objective 2: Determine how insecticidal seed treatments influence pest communities and predation of multiple pest guilds in corn and soybeans under field conditions.
We conducted soybean and corn field experiments addressing this objective in 2012. We initially intended to measure predation in the field on three types of pests: 1) slugs, 2) caterpillars, and 3) weed seeds. However, our proposed method of enclosing slugs in a salt moat turned out not to be feasible in a field setting. We did measure predation on caterpillars and weed seeds.
Objective 3: Educate farmers about neonicotinoid seed treatments and alternative practices through extension materials, videos, and presentations.
We have worked toward this objective continuously since the project began (for more detail, see Results section below). One part of this objective that we did not complete was the proposed fact sheets on early-season IPM in no-till corn and soybeans. We determined that our originally-proposed fact sheets would overlap too strongly with existing resources; we do intend to create a more focused extension resource on seed treatments but this is still in progress.
a. Seeds, slugs, and beetles
We conducted two laboratory experiments, one in soybean and one in corn, to investigate whether neonicotinoid seed treatments can travel up the food chain from seedlings to slugs to ground beetles. For soybean, we used a single soybean variety (A1016495, FS HiSOY® RR2, Growmark, Inc.) that was treated in one of four ways to represent the range of commercially available soybean seed treatments: 1) untreated control; 2) fungicide-alone (mefenoxam and fludioxonil); 3) fungicide plus low rate insecticide (thiamethoxam at 0.0756 mg ai/seed); and 4) fungicide plus high rate insecticide (thiamethoxam at 0.152 mg ai/seed). For corn, we used a single variety (TA451-00, T.A. Seeds) that was treated with a similar range of treatments: 1) untreated control; 2) fungicide-alone (mefenoxam, fludioxonil, and azoxystrobin); 3) fungicide plus low rate insecticide (thiamethoxam at 0.25 mg ai/seed); 4) fungicide plus high rate insecticide (thiamethoxam at 1.25 mg ai/seed). We collected gray garden slugs (Deroceras reticulatum) and the ground beetle Chlaenius tricolor in the vicinity of State College, PA. We kept slugs and beetles in the lab and fed them with organic cabbage and softened cat food, respectively, until using them in experiments.
b. Slug-plant interactions
For the soybean experiment, we created microcosms, each comprising a 16-oz clear plastic container with 2.5cm of potting soil, planted with four seeds/container (n = 16/seed treatment). The next day, into half of the microcosms in each treatment we introduced four juvenile slugs (mean mass: 0.045 ± 0.003 g); the remaining microcosms served as slug-free controls. Later we repeated this experiment with slightly larger slugs. In this case the slug treatment comprised a single, medium-sized juvenile slug (mean mass: 0.22 ± 0.09 g; n = 34/seed treatment with slugs and n = 24/seed treatment without slugs).
For the corn experiment, each microcosm comprised a 32-oz clear plastic container with 5cm of potting soil, into which we planted one seed/container (n = 60/seed treatment). Three days later, into 32 of the microcosms in each treatment we introduced two juvenile slugs (mean mass: 0.038 ± 0.012g).
In all experiments, once microcosms were assembled we placed them in a growth chamber (21oC, 14:10 L:D) for seven days, recording daily the status of slugs (alive/dead) and seedlings (undamaged, damaged, killed). In the corn experiment, because each microcosm had only one seedling, to gather more information we recorded damage on a 0 to 4 scale (0=no damage; 1= up to 25% leaf area removed; 2=25-50% leaf area removed; 3=50-75% leaf area removed; 4=75-100% leaf area removed). After one week, we recovered slugs and seedlings. Slugs were weighed wet, but we placed them on a moist paper towel before weighing to minimize differences due to hydration status. We separated seedlings into living and dead components, with a seedling scored as dead if slugs had eaten completely through the stem. We rinsed these components in water to remove soil and dried them in a drying oven at 65oC, then weighed them.
c. Slug-beetle interactions
To determine whether slugs can transmit seed-applied insecticides from soybean seedlings to ground beetle predators, we conducted another experiment using slugs generated in our slug-plant experiments. After we weighed slugs at the end of seven days of feeding, we transferred them to new 16-oz plastic containers with ~1cm of moist potting soil (one slug per container). The sides of the containers were coated with Fluon to keep slugs in the arena where beetles could attack them (Symondson 1993).
We starved beetles for one week prior to the experiment (well within the range of normal starvation for carabids, Bilde & Toft 1998), and then randomly assigned beetles to containers and introduced them in the early evening (n = 17 to 19/seed treatment for soybeans and n = 14/seed treatment for corn). We observed interactions in containers for the first few hours after beetles were introduced, recording the status of slugs and beetles every 15 to 30 minutes. We then moved containers to a growth chamber (21oC, 14:10 L:D). For soybean, we observed beetles for the next seven days. Because we found that beetle symptoms were quick to manifest, in the corn experiment we observed beetles for only one day and then terminated the experiment. We recorded the status of slugs (alive/dead), and recorded beetle flip-time as a measure of beetle coordination (Lundgren and Wiedenmann 2002). For each trial, we flipped a beetle on its back using forceps and recorded to the closest half second the time necessary for the beetle to right itself, ending a trial after 30 seconds if the beetle failed to flip over. To reduce variability, we flipped each beetle four times and averaged those four values as the flip time for that day (Lundgren and Wiedenmann 2002). In addition, we recorded beetle mortality and took qualitative notes on beetle behavior.
a. Study site and crop management
To explore the relevance under field conditions of seed-applied insecticides for interactions between plants, slugs, and predators, we conducted two field experiments arranged in randomized block designs at Penn State’s Russell E. Larson Agricultural Research Station at Rock Springs (Pennsylvania Furnace, PA). We established both experiments in fields under long-term no-till management that had not received insecticide applications for at least one year.
For corn, we planted variety TA451-00 (T.A. Seeds) either with commercially applied fungicide (mefenoxam, fludioxonil, and azoxystrobin) and insecticide (thiamethoxam, 1.25 mg ai/seed; n = 6 plots) or with only the fungicide (fungicide-only control; n = 6 plots). We planted corn on May 4th at a rate of ~30,000 seeds/acre with 76-cm row spacing. Plots abutted one another, but we conducted all of our sampling in a defined area in the center of each plot, 24 rows by 18 m, leaving a buffer of at least 6 m to adjacent plots or edges.
For soybeans, we used variety HS31A03 (GROWMARK, Inc.), which was inoculated with Bradyrhizobium japonicum (Optimize ®, Novozymes Bio Ag Inc., Brookfield, WI). We planted this variety either with commercially applied fungicide (mefenoxam and fludioxonil) and insecticide (thiamethoxam, 0.152 mg ai/seed; n = 6 plots) or without a seed coating (untreated control; n = 6 plots). We originally intended to use a fungicide-only control but were unable to find the appropriate seed; fortunately our lab experiments suggested that slugs and natural enemies are not influenced by soybean fungicides, so that any differences we see are likely due to the insecticide. We planted the soybeans on May 18th at a rate of ~180,000 seeds/acre with 76-cm row spacing, and collected all samples in a defined area in the center of each plot, 20 rows by 22 m, leaving a buffer of 6 m as above.
b. Stand establishment, early season herbivory, and yield
To assess the influence of the seed treatment on crop growth and productivity, we measured plant populations and herbivore damage during three early growth stages (soybean: VC, V1, V3; corn: V1, V3, V5). On each sampling date, we counted the number of plants in 3 m of row at six randomly chosen locations per plot. At the same locations, we examined the first 15 seedlings for evidence of herbivory, recording the approximate percentage of leaf area removed using a semi-quantitative scale (0: 0%; 0.4: <10%; 1: 10-25%; 2: 25-50%; 3: 50-75%; 4: >75%). For soybeans at VC only, we recorded the number of seedlings visibly killed, as occurs when slugs chew through the plant stem near the soil surface as seedlings emerge. To measure yield, we harvested soybeans on November 9th, taking two samples per plot, each sample comprising four rows, 30.5-m long. We harvested corn on November 28th, also with two samples per plot, each sample comprising two rows, 30.5 m long.
c. Invertebrate activity-density
In both experiments, to assess the influence of the treatments on slug activity-density, we used square-foot pieces of roofing material (Owens Corning Rolled Roofing, color: Shasta White) as artificial slug shelters. While not a measurement of absolute density, the number of slugs under shelter traps provides a relative measure of slug activity (Byers, Barratt, and Calvin 1989) and documents patterns of slug activity over the season. We installed six shelters per plot in random locations shortly after planting and checked these shelters weekly until harvest. We identified slugs to species in the field (Chichester and Getz 1973, McDonnell et al. 2009).
To measure activity-density of ground-dwelling predators and provide an additional measure of slug activity-density, we installed four pitfall traps per plot. Within the inner sampling area of each plot, we installed traps in a rectangular arrangement with at least 6 m between traps. Each trap comprised a 16-oz plastic deli container (11.5-cm diameter, 8-cm tall, Reynolds Del Pak ®) that we sunk into the ground so that the edge was level with the soil surface. We removed the lip from an identical container and placed it inside the first so that it could be easily removed to empty the trap without disturbing the surrounding soil. A white plastic plate (18-cm diameter) supported by nails (8.5-cm long) served as a trap cover and the killing agent was 50% propylene glycol. We opened traps for 72 h, with the first sample occurring roughly three weeks after planting and samples continuing once/month for the rest of the growing season. Upon collection, we passed samples through a 1-mm mesh strainer and then rinsed and stored them in 80% ethanol. We identified slugs to genus, and natural enemies to family or order.
In both experiments, to gain an additional measure of ground-dwelling predator activity and function in the experiment, we deployed waxworm caterpillars (Galleria mellonella;) as sentinel prey (Lundgren et al. 2006). In proximity to the first three pitfall sample dates, we deployed ten caterpillars in each plot, separated by at least 3 m. We pinned caterpillars (0.21 ± 0.06 g) through their final abdominal segment to a small piece of modeling clay that was buried in the field so that the caterpillar rested on the soil surface. To exclude vertebrates, we enclosed each caterpillar in a cylindrical, hardware cloth cage (9.5-cm tall, 11.5-cm diameter, mesh size: 1.3 cm) topped with a plastic lid. On each sample date, we assessed both diurnal and nocturnal predator activity: one sample started at 8:30, and the other at 20:30. After the first 12 h, we replaced any caterpillars that had been attacked or were missing or compromised. We recorded caterpillars as whole and alive, partially eaten, or missing at each time point. Caterpillars that were dead but showed no sign of predation were excluded from the analysis.
We also measured predation on weed seeds using weed seed cards (3 cards/plot, separated by at least 6 m) at three times during the season. Each seed card comprised a 7.5 X 8.5 cm rectangle of 60 grit sandpaper with 50 giant foxtail seeds attached using spray adhesive. In the field, seed cards were enclosed in wire cages to exclude vertebrates (as above). We left seed cards in the field for 48 rain-free hours, then collected them in plastic bags and brought them back to the lab to count the remaining seeds under the microscope.
e. Statistical analysis
To analyze our results, we used the statistical software R (version 3.0.2) and SAS (version 9.2). For most of our analyses we used linear mixed-effects models in R (package lme) with seed treatment as a fixed effect and trial or block a random effect. In some cases time was also included as a repeated measure. Where necessary, we transformed response variables to meet the assumptions of the analysis; for instance, we square-root transformed pitfall counts to normalize residuals. Additional details of the analysis are noted in the Results section below.
To share our results with farmers and agricultural support professionals, in 2012 and 2013 we reached out through field days, winter extension meetings, and conferences. Our audiences ranged from no-till farmers with slug problems, to farmers looking to gain pesticide recertification credits, to NRCS and extension agents, to fellow land-grant researchers. To better illustrate the effects of seed treatments on natural enemies, in 2011 we created a short video clip showing the abnormal behavior of ground beetles that were poisoned after eating neonicotinoid-infused slugs. We used this video to enrich our outreach presentations. At field days, we brought both live and pinned examples of natural enemies so that farmers could see them first-hand. In the fall of 2013 we also discussed our findings with an employee in the Environmental Protection Agency’s Office of Pesticide Programs.
Bilde, T. and S. Toft. 1998. Quantifying food limitation of arthropod predators in the field. Oecologia 115: 54-58.
Byers, R. A., B. I. P. Barratt, and D. Calvin. 1989. Comparison between defined-area traps and refuge traps for sampling slugs in conservation tillage crop environments. In: I. F. Henderson, ed. Slugs and Snails in World Agriculture, BCPC Monograph 41, 187-192.
Chichester, L. F., and L. L. Getz. 1973. The terrestrial slugs of northeastern North America. Sterkiana 51: 11-42.
Lundgren, J. G. and R. N. Wiedenmann. 2002. Coleopteran-specific Cry3Bb toxin from transgenic corn pollen does not affect the fitness of a nontarget species, Coleomegilla maculata DeGeer (Coleoptera: Coccinellidae). Environmental Entomology 31(6): 1213-1218.
Lundgren, J. G., J. T. Shaw, E. R. Zaborski, and C. E. Eastman. 2006. The influence of organic transition systems on beneficial ground-dwelling arthropods and predation of insects and weed seeds. Renewable Agriculture and Food Systems 21(4): 227-237.
McDonnell, R. J., T. D. Paine, and M. J. Gormally. 2009. Slugs: A Guide to the Invasive and Native Fauna of California. University of California, Division of Agriculture and Natural Resources, Publication 8336.
Symondson, W.O.C. 1993. Chemical confinement of slugs: an alternative to electric fences. Journal of Molluscan Studies 59: 259-261.
In the soybean experiment with medium-sized slugs, seed treatments did not protect seedlings from slug damage (Fig. 1A; Treatment F3,116 = 0.24, P = 0.87), nor did they influence slug survival (>85% in all treatments; Fisher’s Exact Test, P = 0.46) or feeding behavior as reflected by mass gain (Fig. 1B; Treatment F3,116 = 0.42, P = 0.74). The experiment with smaller slugs yielded similar results (data not shown). However, slugs that fed on neonicotinoid-treated seedlings were toxic to the ground beetle Chlaenius tricolor. Over 60% of C. tricolor that ate a single slug from low or high thiamethoxam treatments showed symptoms of impairment, such as partial paralysis, flipping slower than control beetles, twitching, or uncoordinated movement (Fig. 1C; Fisher’s Exact Test, P < 0.01).
Results in corn were broadly similar. Seed treatments did not protect seedlings from slug damage, and in fact there was a trend for greater damage in the high rate insecticide treatment toward the end of the experiment (Fig. 2A; Treatment*Day P = 0.08). Slug survival was similar across treatments (>90%; Fisher’s Exact Test, P = 0.19), and while slugs did not gain very much mass, seed treatment did not appear to influence their mass gain (Fig. 2B; Treatment F3,107 = 0.94, P = 0.42). Neonicotinoid-infused slugs were again poisonous to over 60% of C. tricolor individuals that ate a single slug(Fig. 2C; Fisher’s Exact Test, P < 0.01).
In 2012, a warm and wet spring followed an unusually warm winter. As a result, slugs were very abundant and active at both of our sites as crops were emerging, creating good conditions for testing our hypotheses. The only downside was that the wet spring caused severe flooding at our low-lying corn site, complicating field operations, data collection, and interpretation.
In soybeans, seed treatment marginally reduced herbivore damage to leaves at VC (Fig. 3A; F1,5 = 5.86, P = 0.06), but also tended to increase the percentage of seedlings killed outright (Fig. 3B; F1,5 = 4.30, P = 0.09). This second type of damage is commonly caused by slugs as they chew through the stem as seedlings emerge. Interestingly, plant populations at VC were ~20% lower in plots planted with thiamethoxam-treated seed compared to controls (Fig. 3C; F1,5 = 27.5, P < 0.01), a difference that persisted through V3. Consistent with the idea that slugs caused the poor establishment, pitfall trap captures of slugs in treated plots were consistently higher than in controls (Fig. 4; Treatment F1,10 = 5.25, P = 0.04), and early-season pitfall captures of slugs were negatively correlated with soybean establishment (Pearson’s r = -0.73, P < 0.01). Activity-density of slugs measured via shelter traps did not differ between treatments (P = 0.25), but this data was quite variable. Soybean yield did not differ significantly between treatments (Fig. 3D; Kruskal Wallis ANOVA, P = 0.75). Notably, yield was more variable where seed treatment was present (Levene’s Test, P < 0.01).
The major natural enemies captured in pitfall traps were ground beetles (Carabidae), ants (Formicidae), rove beetles (Staphylinidae), wolf spiders (Lycosidae), and parasitic wasps (Hymenoptera). There was a trend for ground beetle adults to be more abundant in untreated plots on the first sample date, but this was not significant when the full season’s data was taken into account (Fig.5; Treatment*Date F3,30 = 0.53, P = 0.67). Similarly, ground beetle larvae were numerically more abundant in control plots but not significantly so (Fig. 5; Treatment F1,10 = 2.03, P = 0.18). Thiamethoxam seed treatment reduced early season pitfall catches of rove beetles by 37% (Fig. 5; Treatment*Date F3,30 = 3.78, P = 0.02), and tended to reduce catches of parasitic wasps (Fig. 5; Treatment F1,10 = 4.09, P = 0.07). No significant differences were seen for ants (Fig. 5; Treatment F1,10 = 0.30, P = 0.60). Activity-density of wolf spiders was actually higher in treated plots (Fig. 5; Treatment F1,10 = 5.85, P = 0.04).
Consistent with the mixed or neutral response of natural enemies to seed treatment, predation on sentinel caterpillars was similar across treatments, though there was a tendency for lower predation in treated plots on the first sample date (Fig. 6; Treatment*Date F2,20 = 1.63, P = 0.22). Ants were the group most commonly observed attacking caterpillars, though we also observed other enemies attacking caterpillars, such as ground beetles, wolf spiders, and harvestmen (Opiliones). Predation on weed seeds did not differ by treatment on any of the three sample dates (Treatment P = 0.28; Treatment*Time P = 0.19).
Taken together, our soybean field results present a bit of a puzzle. In line with our hypothesis, slugs were more abundant in treated plots compared to controls, and soybean establishment was actually worse where the seed treatment was used. However, we did not see a clear signal in the natural enemy data to suggest that natural enemy abundance was reduced by the seed treatment. Given our laboratory results, it could be that slug enemies are being exposed to sublethal doses of the insecticide that impair their ability to hunt slugs without outright killing them. This is a possibility we are continuing to explore through additional experiments.
Corn plants were heavily damaged by pests (slugs and cutworms) early in the season. Leaf damage to seedlings at V3 was similar between treatments (Fig. 7A; P = 0.59), and even though the high rate of thiamethoxam is labeled for cutworm control, the percentage of seedlings severed by cutworms at V3 was similar between treatments (Fig. 7B; P = 0.74). Plant populations were generally on the low side (~27,000 plants/acre at V3), but seed treatments did not significantly improve or reduce stand establishment (Fig. 7C; P = 0.41), and corn yield did not differ by treatment (Fig. 7D; P = 0.54). One of the outcomes of the abundant rain in the early part of the season was luxurious weed growth in some of the plots; as a result weed biomass was very high and negatively correlated with yield (Pearson’s r = -0.70, P = 0.01).
The first planned pitfall sample in corn could not be completed because many plots were under several inches of water. We collected a pitfall sample as soon as plots had dried out sufficiently, which was in early June, ~5 weeks after planting. Then, pitfall sampling resumed monthly for the rest of the season. Seed treatment did not appear to influence slug activity-density in pitfall traps over the season (Fig. 8; Treatment F1,10 = 1.10, P = 0.32), not did slug activity-density as measured under shelter traps differ by treatment (Treatment P = 0.83). The dominant natural enemy taxa were similar to the soybean experiment. Catches of ground beetle adults and larvae did not differ significantly between treatments, though there was again a trend for fewer larvae in treated plots (Fig. 9; Adult Treatment F1,10 = 0.52, P = 0.48; Larvae Treatment F 1,10 = 4.26, P = 0.07). Ants, rove beetles, wolf spiders, and parasitic wasps did not differ significantly by treatment (Fig. 9; Ants Treatment F1,10 = 0.006, P = 0.94; Rove beetle Treatment F1,10 = 0.86, P = 0.38; Wolf spider Treatment F1,10 = 0.02, P = 0.89; Parasitic wasp Treatment F1,10 = 0.46, P = 0.51).
Similarly, predation on sentinel caterpillars did not differ by treatment on any of the three sample dates (Fig. 10; Treatment F1,25 = 0.36, P = 0.56), nor did disappearance of weed seeds differ by treatment (Treatment P = 0.49, Treatment*Time P = 0.61).
Our results for corn are somewhat difficult to interpret given complications of the experiment (very wet spring, with high weed pressure in some plots). Because of flooding we were not able to sample natural enemies during the very early part of the season when we predicted they would be most affected. Given what we were able to measure, we did not see strong differences in plants, slugs, or enemies across treatments. Overall our corn experiment did not show a strong cost or benefit of thiamethoxam seed treatment under our experimental conditions.
Our results contribute to a complex debate about the use and consequences of neonicotinoid seed treatments. Audiences from farmers to regulatory authorities and the public are increasingly concerned about how these widely used products are influencing non-target organisms such as pollinators and aquatic life; our results highlight a novel potential for neonicotinoids to move in the environment through non-target herbivores to their natural enemies. Our results and outreach activities have made at least a small impact in encouraging farmers to consider using IPM for early-season insect pests (see Farmer Adoption). In addition, the Environmental Protection Agency is currently doing an expedited review of neonicotinoids, and we have opened a line of dialogue with an agency representative so that our findings may contribute to the assessment of seed treatment costs and benefits. It is our hope that as we continue to share our results through extension work and scientific publications, our work will contribute to agricultural sustainability.
Education & Outreach Activities and Participation Summary
The primary way we conducted outreach for this project was through J. Tooker’s extension responsibilities. In the course of over 30 extension events we reached over 2,000 agricultural professionals, mainly in Pennsylvania but also in Maryland and Virginia (Table 1). Slugs are a serious concern for the region’s growers and so Tooker integrated into his extension programming the results from this project, illustrated with graphs and charts along with our short video of ground beetles under the influence of neonicotinoid-infused slugs. Douglas, the graduate student grantee for this project, assisted by supplying materials and helping to present at three field days in 2012. We especially enjoyed outreach events that allowed for informal discussion with attendees; in these settings we had productive dialogue that led farmers and crop consultants to reconsider their reflexive use of seed treatments and led us to new directions in our work (see Farmer Adoption and Areas Needing Additional Study).
We also shared our results with scientists through presentations at scientific conferences (Table 2). We were fortunate to present to diverse audiences including entomologists, agronomists, malacologists (mollusk-researchers), and biocontrol researchers.
In addition, Douglas assisted another NESARE graduate student grantee (Ian Grettenberger) in obtaining video footage of diverse natural enemies attacking crop pests. These videos can be viewed on a dedicated youtube channel: http://www.youtube.com/channel/UCP1bMuIKoBhQiGgXUP33b9Q?feature=watch. In the future, our lab will incorporate these videos into our extension programming and Grettenberger will incorporate them into the web version of a new fact sheet he is writing.
We are currently working on a manuscript to publish the results from part of this project in a scientific journal.
Per acre costs for seed treatments are not readily available but estimates are around $10-20/acre (M. Douglas, pers. obs.). While this is a relatively small cost in the scheme of crop production, it is still a cost, and many farmers are operating with slim profit margins. Because insecticidal seed treatments failed to increase yield in either of our experiments, our results suggest that these seed treatments do not consistently improve economic returns. Further, although farmers and others tend to perceive seed treatments as “insurance” that decreases the risk of crop production, we found that soybean yield was actually more variable with seed treatments than without. In line with IPM, farmers could decrease their production costs by using insecticidal seed treatment only where target insect pests are expected based on scouting and/or field history, and forgoing insecticidal seed treatment in areas where early-season insect pest pressure is low. Finally, our results suggest that in soybeans, seed treatments may even have a hidden cost in the form of increased slug pressure.
We have had the opportunity to assess the farmer response to our findings during outreach activities. Most farmers and consultants were interested in our results, though they varied in their openness to changing management practices. One challenge that farmers brought to our attention is that seed treatments are the “default” option in corn and increasingly in soybeans, so obtaining untreated seed is sometimes difficult. While we knew this was the case going into the project, we were surprised by the extent of the situation (for example we had trouble finding enough untreated seed for our own research). In soybeans, there is still an opportunity for farmers to choose untreated seed and at least a few farmers have gone this route in response to our outreach activities. For instance, Gerard Troisi, a crop consultant who assists a number of no-till farmers in central Pennsylvania, convinced several of his clients to forgo insecticidal seed treatments on soybeans in 2013.
Areas needing additional study
Our work suggests several questions for future research, including:
1) What concentrations of neonicotinoids exist in soil, plants, slugs, and predators in the laboratory and the field?
2) Do neonicotinoid seed treatments alter the structure of invertebrate food webs under field conditions?
3) What are the long-term effects of neonicotinoid seed treatments in crop rotations where they are used every year?
4) How might generalist and specialist natural enemies be differently affected by trophic transfer of insecticides?
5) Are there aspects of farming systems that can buffer potential negative effects of insecticides?