Ecological Archives A025-035-A1

Heidi Liere, Tania N. Kim, Benjamin P. Werling, Timothy D. Meehan, Douglas A. Landis, and Claudio Gratton. 2015. Trophic cascades in agricultural landscapes: indirect effects of landscape composition on crop yield. Ecological Applications 25:652661.

Appendix A. Detailed sampling protocols 2011 and 2012.

2011 Sampling protocol

In August 2011, we established natural enemy exclusion experiments in an effort to measure the effect of natural enemies on aphid abundance while controlling for variability in local conditions across the landscape gradient. To achieve this goal and to evaluate the effects of landscape composition on trophic cascades on soybean plants, we established one small garden (15 m × 8 m) at the edges of soybean fields (n = 10 gardens in Wisconsin; n = 10 gardens in Michigan).  Prior to seeding each garden the soil was tilled and fertilized (12N-12P-12K fertilizer at 3.4 kg / 120 m²). Because we were initially interested in looking at landscape effects on various crop species (soybean, corn, and collard), these three species were seeded along ten longitudinal rows: two rows of soybean, three of corn, three of collards and two additional rows of soybean. We followed commercial recommendations for row and plant spacing for the three crops (corn: 20 cm between rows and 30 cm between plants; collards: 76 cm between rows and 30 cm between plants; soybean: 76 cm between rows and 7.6 cm between plants). We only present the soybean results in this study (corn and collard results will be presented in other manuscripts).

The cage exclusion experiments began the first week of August when we selected three pairs of soybean plants (2 m apart) per garden that were naturally infested with > 20 soybean aphids (SBA). We started the experiment during early pod formation to avoid confounding effects of excluding both natural enemies and pollinators. For all the experimental paired plants, we left the SBA untouched but manually removed all natural enemies, NE (predators and parasitoids). We covered one plant per pair with a full cage (tomato cage, 0.6m × 1m, completely covered with a fine "no-see-um" mesh sleeve, Kaplan Simon, Braintree, MA); the other plant in the pair was left open. SBA abundances were counted on the experimental plants at 10 and 20 days after infestation. To estimate NE abundance at the field-level, we placed three large (15 × 30 cm) two-sided sticky cards per garden mounted in 1 m poles and replaced them every two weeks (July –September). We then estimated the average natural enemy abundance captured in a two-week period, per sticky card per experimental field. In early October, we harvested soybean pods after they naturally dried and recorded bean mass per plant and measured the percent of water content using a grain moisture tester. We then standardized the mass of our samples by correcting for percent humidity using 13% as the standard (13–15% is the recommended humidity content for harvesting soybeans Hurburgh, 2008)).

Corrected mass (g) = sample mass (g) * (sample % humidity / 13 % humidity)

2012 Sampling protocol

In 2012, we implemented the cage exclusion experiments on soybean within grower managed farming operations enabling us to double our sample size (n = 20 sites in Wisconsin, 20 sites in Michigan). Varieties of soybean planted by growers varied across our sites as did the management practices (e.g., fertilization, time of planting), but to our knowledge, none were treated with insecticides for the duration of our experiment. We used large PVC-framed cages that covered approximately 10 soybean plants (1.3 m × 0.6 m × 1.3 m in height). The frame of the "closed cages" treatment was completely covered with no-see-um netting to exclude even the smallest insects. To control for any potential cage effects, we partially covered our control, "open cages" treatment with the same type of netting (the netting covered the roof of the cage and only half its sides, leaving a 50 cm opening on every side). No-see-um netting has not been shown to have observable effects on soybean yield when fully uncovered plants were compared with partially covered plants (Costamagna et al. 2008).

In mid-August, we placed two sets of paired cages in each field; each pair consisted of a fully enclosed cage (i.e., closed cage treatment) with soybeans infested with aphids without natural enemies and a partially enclosed cage (i.e., open cage treatment) with soybeans infested with aphids exposed to natural enemy control. The cages in each pair were 2 m apart and 30 m away from the other pair in the field, and at least 50 m from the field edge. To ensure that all experimental plants started with similar densities of SBA, we first manually removed all existing aphids and natural enemies and then pinned an SBA infested soybean leaf (containing ~ 20 SBA from laboratory-reared or field-cage reared colonies) to the central soybean plant. Four small yellow sticky cards (5 cm × 10 cm, Great Lakes IPM) were placed within each cage to capture winged aphids in order to reduce 'artificial' increase of aphid population inside full cages. SBA on five, marked experimental plants inside the cage were counted twice in September. To measure natural enemy abundances, we placed one large (15 × 30 cm) two-sided sticky card mounted on a 1 m pole next to each pair of cages for two weeks (sticky cards were placed out in field twice, July and September). We then estimated the average natural enemy abundance captured in a two-week period, per sticky card per experimental field. When the bean pods were dry and ready for harvest (end of September-beginning of October), we collected the seeds from all experimental plants and estimated yield on a per-plant basis using the same procedure as in 2011.

Literature cited

Costamagna, A. C., D. A. Landis, and M. J. Brewer 2008. The role of natural enemy guilds in Aphis glycines suppression. Biological Control 45:368–379.

Hurburgh, C. 2008. Soybean drying and storage. University Extension, Iowa State University, Iowa, USA.

[Back to A025-035]