Landscape Context Affects Aphid Parasitism by Lysiphlebus testaceipes (Hymenoptera: Aphidiinae) in Wheat Fields

Landscape Context Affects Aphid Parasitism by Lysiphlebus testaceipes (Hymenoptera: Aphidiinae)... Abstract Winter wheat is Oklahoma’s most widely grown crop, and is planted during September and October, grows from fall through spring, and is harvested in June. Winter wheat fields are typically interspersed in a mosaic of habitats in other uses, and we hypothesized that the spatial and temporal composition and configuration of landscape elements, which contribute to agroecosystem diversity also influence biological control of common aphid pests. The parasitoid Lysiphlebus testaceipes (Cresson; Hymenoptera: Aphidiinae) is highly effective at reducing aphid populations in wheat in Oklahoma, and though a great deal is known about the biology and ecology of L. testaceipes, there are gaps in knowledge that limit predicting when and where it will be effective at controlling aphid infestations in wheat. Our objective was to determine the influence of landscape structure on parasitism of cereal aphids by L. testaceipes in wheat fields early in the growing season when aphid and parasitoid colonization occurs and later in the growing season when aphid and parasitoid populations are established in wheat fields. Seventy fields were studied during the three growing seasons. Significant correlations between parasitism by L. testaceipes and landscape variables existed for patch density, fractal dimension, Shannon’s patch diversity index, percent wheat, percent summer crops, and percent wooded land. Correlations between parasitism and landscape variables were generally greatest at a 3.2 km radius surrounding the wheat field. Correlations between parasitism and landscape variables that would be expected to increase with increasing landscape diversity were usually positive. Subsequent regression models for L. testaceipes parasitism in wheat fields in autumn and spring showed that landscape variables influenced parasitism and indicated that parasitism increased with increasing landscape diversity. Overall, results indicate that L. testaceipes utilizes multiple habitats throughout the year depending on their availability and acceptability, and frequently disperses among habitats. Colonization of wheat fields by L. testaceipes in autumn appears to be enhanced by proximity to fields of summer crops and semi-natural habitats other than grasslands. cereal aphid, biological control, Hymenoptera, parasitoid Winter wheat is Oklahoma’s most widely grown crop, with more than 2 million ha planted annually (Epplin et al. 1998, USDA NASS 2017). Winter wheat in Oklahoma is planted during September and October, grows from fall through spring, and is harvested in June. Several aphid species infest wheat fields in this region, the most common and important being greenbug, Schizaphis graminum (Rondani); bird cherry-oat aphid, Rhopalosiphum padi (L.); and English grain aphid, Sitobion avenae (F.) (Heteroptera: Aphididae). Apart from the use of insecticidal seed treatments, which are used preventively in Oklahoma by an increasing number of producers, insecticide applications are infrequent, and primarily used to control insects such as fall armyworm in autumn, and cereal aphids, primarily greenbug, in autumn or spring (Royer et al. 2015). During the wheat growing season, winter wheat fields are typically interspersed in a mosaic of lands in other uses. Grasslands (pasture and rangeland) with varying levels of management are the most abundant land use type. Grasslands range from semi-natural lands that have never been cultivated and have high plant species diversity to highly managed lands planted to a single grass species. Fallow fields will mostly be planted to summer crops (soybean, corn, sorghum, and cotton) in spring, and fields may also be planted to other winter crops, such as canola and barley. Riparian areas and other semi-natural lands are also present in the landscape. Based on previous studies that investigated pest suppression in agricultural landscapes, the spatial and temporal configuration of landscape elements, which contribute to agroecosystem diversity are likely to influence populations of natural enemies, and possibly, biological control of insect pests by determining the availability of resources for beneficial insects (Rusch et al. 2016). In Oklahoma the parasitoid Lysiphlebus testaceipes (Cresson; Hymenoptera: Aphidiinae) is highly effective at reducing aphid populations in winter wheat (Webster and Phillips 1912, Giles et al. 2003, Jones et al. 2007, Royer et al. 2015). The parasitoid’s effectiveness is thought to be the result of features of its biology and ecology. It attacks multiple aphid species during cold but non-lethal fall and winter months where daytime temperatures often exceed the species’ activity threshold (Jones et al. 2007), it has high attack and reproductive rates (Jones et al. 2003, Giles et al. 2003), it sterilizes the aphids it attacks (Hight et al. 1972, Eikenbary and Rogers 1974), and it dislodges aphids from the plant as it forages, making them subject to mortality from predation and the environment (Losey and Denno 1998). L. testaceipes has been observed to keep greenbug populations below the economic injury level in wheat (Eikenbary and Rogers 1974, Giles et al. 2003), and its impact has been successfully incorporated in pest sampling programs and management guidelines for greenbug (Giles et al. 2003; Royer et al. 2004, 2015). Even though a great deal is known about the biology and ecology of the species, there are gaps in knowledge that limit understanding of the L. testaceipes/aphid system and predicting when and where L. testaceipes will be effective at controlling infestations in wheat. The structure of the landscape surrounding a particular agricultural field has been shown to influence populations of predatory insects in wheat fields (Elliott et al. 1998), and parasitoid abundance and parasitism levels of herbivorous insects in other agricultural ecosystems (Schmidt et al. 2004, Thies et al. 2005). However, only one study has specifically addressed effects of landscape context on L. testaceipes. Brewer et al. (2008) demonstrated that adding sunflower to a strip cropping system including wheat in southeastern Wyoming increased parasitism levels of Russian wheat aphid by L. testaceipes. Our objective was to determine the relative influence of landscape structure on parasitism of cereal aphids by L. testaceipes in wheat fields early in the growing season when aphid and parasitoid colonization occurs, and later in the growing season when within field aphid and parasitoid population processes may predominate over colonization. Our hypothesis was that the landscape context within which a wheat field is embedded influences the level of parasitism in a wheat field through its presumed effect on resource availability to L. testaceipes in other habitats, and this influence would be most pronounced early in the growing season (autumn) when colonization of wheat fields by aphids and L. testaceipes from other habitats primarily determines abundance and parasitism. Landscape context would be predicted to have less influence later in the growing season (spring) when aphid and parasitoid population dynamics within wheat fields predominates over colonization in determining abundance and parasitism. Materials and Methods Field Study We utilized aphid-infested sentinel barley plants in pots to quantify aphid parasitism within wheat fields in north central Oklahoma over a 3-yr period (Fig. 1). Potted wheat plants grown in a greenhouse and infested with parasitoid free bird cherry-oat aphids were set out in commercial wheat fields during autumn and spring. The technique has been successfully used to compare relative parasitism rates within and between wheat fields (Brewer et al. 2008). Each year, approximately 24 wheat fields were selected to achieve broad coverage of the area with randomization partially restricted by the availability of cooperating farmers and that study fields were separated by a minimum of 5 km. Fig. 1. View largeDownload slide Approximate boundaries of the geographic area where field studies were conducted during the 2008–2009, 2009–2010, and 2010–2011 wheat growing seasons. Fig. 1. View largeDownload slide Approximate boundaries of the geographic area where field studies were conducted during the 2008–2009, 2009–2010, and 2010–2011 wheat growing seasons. To determine if landscape effects on parasitism existed during autumn and spring, the field study described below was repeated in autumn of three consecutive years, 2008–2010, and in spring of two consecutive years, 2009 and 2010. Studies were initiated in October, soon after wheat plants emerged from the soil, and during mid-March when aphid and parasitoid activity is common (Giles et al. 2003). Approximately 10 barley seeds (variety Eight Twelve) were planted in a 2:1 mixture of peat moss and fritted clay in 15.2 cm diameter plastic pots, and a 14 cm diameter by 35 cm high circular clear plastic cage was placed on each pot. Each cage was vented in the side and top with fine muslin cloth and pressed about 2.5 cm into the planting mixture. Fifteen pots containing caged barley plants (hereafter referred to as sentinel plants) were placed in each of 12 cubicle shaped cages (90 × 90 × 40 cm) covered with fine mesh screen on the four sides, a plywood bottom, and a clear plastic top. The double caging was done to ensure that aphids on sentinel plants remained parasitoid free until they were uncaged in the field. Two-week old sentinel plants in cages were infested with ca. 50 parasitoid free bird cherry-oat aphids. Approximately 10 d later, when aphid counts in cages averaged ca. 750 per pot, sentinel plants were transported to wheat fields where they were stationed as described below. One of the 15 sentinel plants from each cage (total of 12 plants) was retained in the greenhouse as a check to ensure that parasitoids had not entered cages and parasitized aphids while they were being grown in the greenhouse. In each wheat field, seven uncaged sentinel plants were placed into the soil and arranged 25 m apart in a T-shaped pattern. Sentinel plants were placed 5 m from a field edge, and every 25 m along a transect perpendicular to the field edge until five sentinel plants had been placed. Sentinel plants were also placed 25 m to the right and left of the fourth plant at 90-degree angles. Sentinel plants were left in fields for 3 d and then caged, returned to the greenhouse and maintained for 7 d at 21°C (±5°C), 16:8 (L:D) h, and ambient (uncontrolled) humidity to allow any parasitoids to undergo development to the pupal stage. After 7 d the barley plants from each pot were cut, separately placed in an emergence canister, and held for an additional 7 d to allow parasitoids to emerge as adults. Adult parasitoids in each emergence canister were counted and identified to species. In wheat fields in Oklahoma, L. testaceipes is the dominant cereal aphid parasitoid, usually accounting for over 95% of cereal aphid parasitism. During our study L. testaceipes accounted for over 98% of total parasitism in all fields in autumn and spring. Thus, the number of adult L. testaceipes per sentinel plant provided a useful measure of parasitism in a field. Check plants that remained under greenhouse conditions were processed identically to the experimental sentinel plants. No parasitoids were recovered from check plants during the 3-yr study indicating that it was very unlikely that contamination of aphid infested sentinel plants by parasitoids occurred in the greenhouse. We recorded several attributes for each wheat field. We recorded whether insecticidal seed treatment was used on the seed planted in a field, whether the field was in a no-till or conventional till system, whether the field was in a crop rotational system or in continuous wheat, the wheat plant growth stage, and aphid abundance in the field at the time sentinel plants were deployed. Wheat plant growth stage for each field was estimated using a 0–10 scale (Zadoks et al. 1974) on the day that sentinel plants were deployed. Aphid abundance in each field was estimated by sampling with a Backpack Model 24 D-Vac (Rincon-Vitova Insectaries, Inc., Ventura, CA) fitted with a standard 33 cm diameter collecting unit, fine mesh organdy collecting bag, and fiberglass collar. Sampling by D-Vac has been shown to provide useful estimates of aphid abundance in cereals (Hand 1986). A sample from each field consisting of three subsamples was taken within the area where the sentinel plants were deployed by walking three equally spaced linear transects. The first transect was situated parallel to and about 5 m to the left or right of the transect along which five of the seven sentinel plants had been stationed. The other two transects were parallel to and approximately 25 m to the right and left of the first transect. Approximately every 5 m the D-vac collecting unit was placed straight down over growing wheat plants to just above the soil surface until 20 such placements had been made. Each time the collecting unit was placed down it was held in position slightly above the soil surface for 5 s. After 20 stops, the sampling bag was removed from the D-vac and all arthropods in it were transferred to a labeled plastic bag. Bags were brought to the laboratory and placed in a freezer. Aphids were counted at a later date. Frozen aphids are difficult to identify to species so we did not distinguish among cereal aphid species. The mean number of aphids for the three transects sampled per field, each of which consisted of 20 D-vac placement subsamples (n = 3) provided our estimate of aphid abundance for each wheat field. Landscape Data Landscape context for each field was quantified for each of three circular areas centered on the focal wheat field with radii of 0.8, 1.6, and 3.2 km extracted from the USDA NASS Cropland Data Layer from the appropriate year. The Cropland Data Layer was acquired for the year in which winter wheat was present in the layer for the study fields for that year, and was used to quantify the amount and distribution of all land use types. The Cropland Data Layer differentiates crop types with accuracy rates typically above 85% (https://www.nass.usda.gov/Research_and_Science/Cropland/sarsfaqs2.php date accessed 5 March 2017). The crop-specific data is available at https://nassgeodata.gmu.edu/CropScape/ (date accessed 5 March 2017). Distances probably encompassed the range of potential dispersal ability of L. testaceipes (Thies et al. 2005). In addition to NASS data, we collected fine spatial scale ground survey data of the grass species in the wheat field boundaries. All grass species within one 15-m transect at an arbitrary location within each of two accessible field edges (e.g., adjacent to a road) were recorded for each field. One of the edges sampled was the one adjacent to where sentinel plants were deployed in the field. In addition, percent cover by Johnson grass, Sorghum halepense (L.), was estimated visually for an area of approximately 200 m2 in the boundary adjacent to the field. The cropland data layer was re-classified to retain eight land use categories: wheat, summer crops, winter crops other than wheat, fallow, grassland (pasture and rangeland), wooded, built areas and roads, and water. Aggregating land uses into fewer categories than represented in the original NASS data was desirable for calculating meaningful landscape metrics because many categories would have been represented by very small areas and metrics would be subject to high variability. We quantified landscape structure relative to each field in each year using the following landscape metrics: the proportion of the total area in each land use type, patch density, perimeter to area fractal dimension, Shannon’s patch diversity, and contagion (McGarigal and Marks 1995, McGarigal 2014). Landscape metrics quantify various characteristics of landscape structure that can be ecologically significant (O’Neill et al. 1988). Patch density, Shannon’s patch diversity, contagion, and perimeter to area fractal dimension are four among dozens of metrics. These four metrics have straight-forward interpretations, and the latter three were found by Ritters et al. (1995) in a study of several landscapes to be good quantitative descriptors of landscape structure that were relatively independent of one another. For the central Oklahoma landscapes in our study the three metrics were correlated among themselves, and were also correlated with patch density. Patch density measures the number of patches per km2 and indicates average patch size for a landscape. The perimeter to area fractal dimension is dimensionless and increases with increasing patch boundary curvilinearity. Contagion, measures the amount of clumping of patch types within a landscape as a percentage of the maximum. Maximum contagion for a given landscape would be achieved when each landscape element type occurred as a single contiguous patch. High contagion indicates highly aggregated and poorly interspersed patches. Shannon’s patch diversity index is based on information theory (Shannon and Weaver 1949) and measures landscape composition, not shape or configuration. Large values of Shannon’s diversity index indicate a greater number of landscape element types present (patch richness) in the landscape, greater evenness in area of the patch types present, or both. The number of patch types present did not vary much among landscapes in our study because most patch types were present. The minimal variation observed for patch richness indicates that variation in Shannon’s patch diversity index in our landscapes primarily reflected variation in the evenness in percent of area of various patch types. Large values for fractal dimension, patch density, and Shannon’s patch diversity, and small values for contagion generally indicate high landscape diversity (see Turner et al. 2001 for more information). Measured attributes of vegetation in field edges and landscape metrics calculated using Fragstats Version 4 were used to quantify landscape structure. Fragstats derived variables were calculated at each of the three hierarchically increasing spatial extents (0.8, 1.6, and 3.2 km radii) for each field each year. GIS operations were accomplished using ERDAS Imagine version 2014 including exporting data for calculating landscape metrics with Fragstats. Data Analysis Correlation was used to evaluate pairwise relationships among variables. When one (or both) variables were categorical, such as tillage type, which was coded numerically as zero for no-till and one for conventional tillage, spearman rank correlation coefficients were calculated. Pearson correlation coefficients were calculated when both variables were continuous. The magnitude of correlation coefficients was used to determine the spatial extent to account for the greatest amount of variation in parasitism. SAS PROC CORR (SAS Institute 2004) was used to calculate correlation coefficients. Many of the landscape variables were correlated so principal components analysis was used to derive a set of linearly independent regressors for use in regression modeling. Principal components were rotated using varimax rotation (Dillon and Goldstein 1984). The number of rotated standardized principal components retained for use as independent variables in regressions was determined by the scree method. The scree method involves plotting the eigenvalue associated with each principal component in successive order and determining the point beyond which the smaller eigenvalues form an approximately straight line. The components retained are those associated with eigenvalues that fall above the straight line formed by the smaller eigenvalues (Dillon and Goldstein 1984). The components that were retained were used as regressors in models relating parasitism to landscape context. Only linear (first order terms) were included in models. Standardized components were interpreted based on magnitudes of component loadings on the original variables. Stepwise multiple regression was used to construct models using the components as regressors. Component loadings were entered as independent variables in stepwise multiple regression models with the mean number of L. testaceipes per sentinel plant for each field as the dependent variable. F-tests were used to determine the significance of regression models with α for inclusion of a regressor in a model set at 0.15. The α = 0.15 level for inclusion was chosen so that moderately influential regressors were not overlooked during model selection. Significance of the overall regression model was maintained at α ≤ 0.05. Regression modeling was accomplished using PROC REG (SAS Institute 2004). Results General Patterns Seventy-one fields were studied in autumn during the three growing seasons, however data for one field was not used because all but one sentinel plant were destroyed by wildlife. For autumn, the proportion of the seven sentinel plants from each field that had one or more L. testaceipes per sentinel plant ranged from 0 to 1 among fields with a mean of 0.28, and the mean number of L. testaceipes per sentinel plant ranged 0 to 100.4 among fields with a mean of 11.6 (Table 1). During spring both the average number of L. testaceipes per sentinel plant and the proportion of sentinel plants with L. testaceipes were greater by 10-fold and threefold, respectively, than in autumn (Table 1). Aphid abundance in fields was also higher (greater than sevenfold) in spring than in autumn (Table 1) as was expected since sentinel plants were placed in wheat fields in autumn as soon as possible after emergence of wheat plants in the field from the soil. It was not possible to identify 24 fields each year with identical planting dates, so there was variation in wheat plant growth stage among fields. Variation in planting date probably accounted for a major portion of the variation in aphid abundance among fields. Table 1. Summary statistics (mean, SE, minimum, and maximum) for parasitism by L. testaceipes, aphid abundance, and other variables measured for n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Mean  SE  Minimum  Maximum  Autumn   Parasitism    Number of L. testaceipes / sentinel plant  11.6  2.62  0.0  100.2    Prop. sentinel plants with L. testaceipes  0.28  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  27.7  15.06  0.0  1040.0  Spring   Parasitism    Number of L. testaceipes / sentinel plant  179.9  22.64  0.0  576.7    Prop. sentinel plants with L. testaceipes  0.88  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  195.6  47.48  12.3  1680.3   Within-field & Field Edge    Tillage (0 = conventional, 1 = no till)  0.32  0.06  0.0  1.0    Crop rotation (0 = no, 1 = yes)  0.26  0.05  0.0  1.0    Grazed (0 = no, 1 = yes)  0.67  0.06  0.0  1.0    Grass species Richness  7.7  0.27  4.0  13.0    % Johnson grass coverage  34.1  2.66  0.0  75.0  Variable  Mean  SE  Minimum  Maximum  Autumn   Parasitism    Number of L. testaceipes / sentinel plant  11.6  2.62  0.0  100.2    Prop. sentinel plants with L. testaceipes  0.28  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  27.7  15.06  0.0  1040.0  Spring   Parasitism    Number of L. testaceipes / sentinel plant  179.9  22.64  0.0  576.7    Prop. sentinel plants with L. testaceipes  0.88  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  195.6  47.48  12.3  1680.3   Within-field & Field Edge    Tillage (0 = conventional, 1 = no till)  0.32  0.06  0.0  1.0    Crop rotation (0 = no, 1 = yes)  0.26  0.05  0.0  1.0    Grazed (0 = no, 1 = yes)  0.67  0.06  0.0  1.0    Grass species Richness  7.7  0.27  4.0  13.0    % Johnson grass coverage  34.1  2.66  0.0  75.0  View Large Table 1. Summary statistics (mean, SE, minimum, and maximum) for parasitism by L. testaceipes, aphid abundance, and other variables measured for n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Mean  SE  Minimum  Maximum  Autumn   Parasitism    Number of L. testaceipes / sentinel plant  11.6  2.62  0.0  100.2    Prop. sentinel plants with L. testaceipes  0.28  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  27.7  15.06  0.0  1040.0  Spring   Parasitism    Number of L. testaceipes / sentinel plant  179.9  22.64  0.0  576.7    Prop. sentinel plants with L. testaceipes  0.88  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  195.6  47.48  12.3  1680.3   Within-field & Field Edge    Tillage (0 = conventional, 1 = no till)  0.32  0.06  0.0  1.0    Crop rotation (0 = no, 1 = yes)  0.26  0.05  0.0  1.0    Grazed (0 = no, 1 = yes)  0.67  0.06  0.0  1.0    Grass species Richness  7.7  0.27  4.0  13.0    % Johnson grass coverage  34.1  2.66  0.0  75.0  Variable  Mean  SE  Minimum  Maximum  Autumn   Parasitism    Number of L. testaceipes / sentinel plant  11.6  2.62  0.0  100.2    Prop. sentinel plants with L. testaceipes  0.28  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  27.7  15.06  0.0  1040.0  Spring   Parasitism    Number of L. testaceipes / sentinel plant  179.9  22.64  0.0  576.7    Prop. sentinel plants with L. testaceipes  0.88  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  195.6  47.48  12.3  1680.3   Within-field & Field Edge    Tillage (0 = conventional, 1 = no till)  0.32  0.06  0.0  1.0    Crop rotation (0 = no, 1 = yes)  0.26  0.05  0.0  1.0    Grazed (0 = no, 1 = yes)  0.67  0.06  0.0  1.0    Grass species Richness  7.7  0.27  4.0  13.0    % Johnson grass coverage  34.1  2.66  0.0  75.0  View Large Approximately 32% of fields studied were no till and 68% were conventional till (Table 1). Sixty-six percent of fields were grazed by cattle during winter months (a common practice in Oklahoma; Epplin et al. 1998). Grazing is associated with use of conventional tillage, because grazing cattle on no-till fields causes high levels of soil compaction. Therefore, the percent of fields in no-till systems was inversely related to grazing. The number of grass species in field edges ranged widely among fields, as did the percentage cover by Johnson grass in field edges. Landscape metrics varied substantially among the 70 fields for land areas in the three radii measured (Table 2). For example, the perimeter to area fractal dimension ranged from 1.20 to 1.45 for a radius of 0.8 km and from 1.31 to 1.45 for areas with a radius of 3.2 km. Percent of total land area planted to wheat ranged from 2.7 to 86.6 for the 0.8 km radius and from 10.1 to 68.5% for the 3.2 km radius. Although landscape metrics varied considerably for each landscape extent, the range of each metric was similar across the 0.8 to 3.2 km spatial extents. Table 2. Summary statistics (mean, SE, minimum, and maximum) for landscape metrics measured at three radii centered on each of n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Mean  SE  Minimum  Maximum  Radius 0.8 km   Patch density  26.3  1.56  7.7  57.3   Shannon’s patch diversity  1.31  0.03  0.61  1.88   Fractal dimension  1.33  0.01  1.20  1.45   Contagion  53.0  1.22  32.5  77.5   % Wheat  40.8  2.36  2.7  86.6   % Summer crops  24.7  2.63  0.11  82.6   % Winter crops (other than wheat)  0.2  0.07  0.0  4.1   % Fallow  1.9  0.24  0.0  10.1   % Grassland  23.0  1.69  1.7  52.4   % wooded  2.4  0.42  0.0  16.2   % manmade (built areas and roads)  6.5  0.53  2.0  24.3   % water  0.4  0.10  0.0  5.0  Radius 1.6 km   Patch density  24.9  1.28  7.8  48.3   Shannon’s patch diversity  1.46  0.03  0.87  1.87   Fractal dimension  1.37  0.004  1.30  1.46   Contagion  51.5  0.89  34.5  69.4   % Wheat  35.9  1.77  11.6  78.1   % Summer crops  24.5  2.08  0.99  68.4   % Winter crops (other than wheat)  0.3  0.05  0.0  2.9   % Fallow  2.6  0.20  0.3  8.4   % Grassland  27.0  1.89  6.5  67.9   % Wooded  2.8  0.33  0.0  9.9   % Manmade (built areas and roads)  6.2  0.31  2.7  13.4   % Water  0.7  0.12  0.0  5.9  Radius 3.2 km   Patch density  24.8  1.26  6.0  52.4   Shannon’s patch diversity  1.51  0.23  1.07  1.90   Fractal dimension  1.38  0.003  1.31  1.45   Contagion  50.6  0.78  35.4  66.2   % Wheat  34.7  1.51  10.1  68.5   % Summer crops  22.4  1.71  2.34  59.7   % Winter crops (other than wheat)  0.2  0.04  0.01  2.0   % Fallow  2.9  0.17  0.6  7.1   % Grassland  29.6  1.74  9.2  66.8   % Wooded  3.1  0.26  0.1  10.5   % Manmade (built areas and roads)  6.1  0.23  3.4  11.7   % Water  0.8  0.10  0.02  3.8  Variable  Mean  SE  Minimum  Maximum  Radius 0.8 km   Patch density  26.3  1.56  7.7  57.3   Shannon’s patch diversity  1.31  0.03  0.61  1.88   Fractal dimension  1.33  0.01  1.20  1.45   Contagion  53.0  1.22  32.5  77.5   % Wheat  40.8  2.36  2.7  86.6   % Summer crops  24.7  2.63  0.11  82.6   % Winter crops (other than wheat)  0.2  0.07  0.0  4.1   % Fallow  1.9  0.24  0.0  10.1   % Grassland  23.0  1.69  1.7  52.4   % wooded  2.4  0.42  0.0  16.2   % manmade (built areas and roads)  6.5  0.53  2.0  24.3   % water  0.4  0.10  0.0  5.0  Radius 1.6 km   Patch density  24.9  1.28  7.8  48.3   Shannon’s patch diversity  1.46  0.03  0.87  1.87   Fractal dimension  1.37  0.004  1.30  1.46   Contagion  51.5  0.89  34.5  69.4   % Wheat  35.9  1.77  11.6  78.1   % Summer crops  24.5  2.08  0.99  68.4   % Winter crops (other than wheat)  0.3  0.05  0.0  2.9   % Fallow  2.6  0.20  0.3  8.4   % Grassland  27.0  1.89  6.5  67.9   % Wooded  2.8  0.33  0.0  9.9   % Manmade (built areas and roads)  6.2  0.31  2.7  13.4   % Water  0.7  0.12  0.0  5.9  Radius 3.2 km   Patch density  24.8  1.26  6.0  52.4   Shannon’s patch diversity  1.51  0.23  1.07  1.90   Fractal dimension  1.38  0.003  1.31  1.45   Contagion  50.6  0.78  35.4  66.2   % Wheat  34.7  1.51  10.1  68.5   % Summer crops  22.4  1.71  2.34  59.7   % Winter crops (other than wheat)  0.2  0.04  0.01  2.0   % Fallow  2.9  0.17  0.6  7.1   % Grassland  29.6  1.74  9.2  66.8   % Wooded  3.1  0.26  0.1  10.5   % Manmade (built areas and roads)  6.1  0.23  3.4  11.7   % Water  0.8  0.10  0.02  3.8  View Large Table 2. Summary statistics (mean, SE, minimum, and maximum) for landscape metrics measured at three radii centered on each of n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Mean  SE  Minimum  Maximum  Radius 0.8 km   Patch density  26.3  1.56  7.7  57.3   Shannon’s patch diversity  1.31  0.03  0.61  1.88   Fractal dimension  1.33  0.01  1.20  1.45   Contagion  53.0  1.22  32.5  77.5   % Wheat  40.8  2.36  2.7  86.6   % Summer crops  24.7  2.63  0.11  82.6   % Winter crops (other than wheat)  0.2  0.07  0.0  4.1   % Fallow  1.9  0.24  0.0  10.1   % Grassland  23.0  1.69  1.7  52.4   % wooded  2.4  0.42  0.0  16.2   % manmade (built areas and roads)  6.5  0.53  2.0  24.3   % water  0.4  0.10  0.0  5.0  Radius 1.6 km   Patch density  24.9  1.28  7.8  48.3   Shannon’s patch diversity  1.46  0.03  0.87  1.87   Fractal dimension  1.37  0.004  1.30  1.46   Contagion  51.5  0.89  34.5  69.4   % Wheat  35.9  1.77  11.6  78.1   % Summer crops  24.5  2.08  0.99  68.4   % Winter crops (other than wheat)  0.3  0.05  0.0  2.9   % Fallow  2.6  0.20  0.3  8.4   % Grassland  27.0  1.89  6.5  67.9   % Wooded  2.8  0.33  0.0  9.9   % Manmade (built areas and roads)  6.2  0.31  2.7  13.4   % Water  0.7  0.12  0.0  5.9  Radius 3.2 km   Patch density  24.8  1.26  6.0  52.4   Shannon’s patch diversity  1.51  0.23  1.07  1.90   Fractal dimension  1.38  0.003  1.31  1.45   Contagion  50.6  0.78  35.4  66.2   % Wheat  34.7  1.51  10.1  68.5   % Summer crops  22.4  1.71  2.34  59.7   % Winter crops (other than wheat)  0.2  0.04  0.01  2.0   % Fallow  2.9  0.17  0.6  7.1   % Grassland  29.6  1.74  9.2  66.8   % Wooded  3.1  0.26  0.1  10.5   % Manmade (built areas and roads)  6.1  0.23  3.4  11.7   % Water  0.8  0.10  0.02  3.8  Variable  Mean  SE  Minimum  Maximum  Radius 0.8 km   Patch density  26.3  1.56  7.7  57.3   Shannon’s patch diversity  1.31  0.03  0.61  1.88   Fractal dimension  1.33  0.01  1.20  1.45   Contagion  53.0  1.22  32.5  77.5   % Wheat  40.8  2.36  2.7  86.6   % Summer crops  24.7  2.63  0.11  82.6   % Winter crops (other than wheat)  0.2  0.07  0.0  4.1   % Fallow  1.9  0.24  0.0  10.1   % Grassland  23.0  1.69  1.7  52.4   % wooded  2.4  0.42  0.0  16.2   % manmade (built areas and roads)  6.5  0.53  2.0  24.3   % water  0.4  0.10  0.0  5.0  Radius 1.6 km   Patch density  24.9  1.28  7.8  48.3   Shannon’s patch diversity  1.46  0.03  0.87  1.87   Fractal dimension  1.37  0.004  1.30  1.46   Contagion  51.5  0.89  34.5  69.4   % Wheat  35.9  1.77  11.6  78.1   % Summer crops  24.5  2.08  0.99  68.4   % Winter crops (other than wheat)  0.3  0.05  0.0  2.9   % Fallow  2.6  0.20  0.3  8.4   % Grassland  27.0  1.89  6.5  67.9   % Wooded  2.8  0.33  0.0  9.9   % Manmade (built areas and roads)  6.2  0.31  2.7  13.4   % Water  0.7  0.12  0.0  5.9  Radius 3.2 km   Patch density  24.8  1.26  6.0  52.4   Shannon’s patch diversity  1.51  0.23  1.07  1.90   Fractal dimension  1.38  0.003  1.31  1.45   Contagion  50.6  0.78  35.4  66.2   % Wheat  34.7  1.51  10.1  68.5   % Summer crops  22.4  1.71  2.34  59.7   % Winter crops (other than wheat)  0.2  0.04  0.01  2.0   % Fallow  2.9  0.17  0.6  7.1   % Grassland  29.6  1.74  9.2  66.8   % Wooded  3.1  0.26  0.1  10.5   % Manmade (built areas and roads)  6.1  0.23  3.4  11.7   % Water  0.8  0.10  0.02  3.8  View Large Many landscape metrics were correlated. Correlations among metrics for 3.2 km radius land areas (Table 3) were similar to correlations for the two smaller radii (not shown). Of particular note, the correlation between Shannon’s patch diversity index and contagion was −0.90, which indicates that evenness in the proportion of each patch type was strongly negatively related to the extent of aggregation of particular patch types in the landscape. The presence of correlation among the majority of landscape metrics indicates that their use as independent variables in regression modeling would result in multicollinearity. Table 3. Correlation among variables describing landscape context within 3.2 km radius areas centered on each of n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Shannon diversity  Fractal dimension  Contagion  % Wheat  % Summer crops  % Grass  % Wooded  Patch density  0.60*  0.26*  −0.32*  −0.25*  0.50*  −0.36*  0.28*  Shannon diversity    −0.38*  −0.90*  −0.45*  0.77*  −0.52*  0.45*  Fractal dimension      −0.50*  −0.09  0.10  −0.16  0.49*  Contagion        0.46*  −0.56*  0.33*  −0.59*  % Wheat          −0.36*  −0.35*  −0.48*  % Summer crops            −0.72*  0.10  % Grass              0.13  Variable  Shannon diversity  Fractal dimension  Contagion  % Wheat  % Summer crops  % Grass  % Wooded  Patch density  0.60*  0.26*  −0.32*  −0.25*  0.50*  −0.36*  0.28*  Shannon diversity    −0.38*  −0.90*  −0.45*  0.77*  −0.52*  0.45*  Fractal dimension      −0.50*  −0.09  0.10  −0.16  0.49*  Contagion        0.46*  −0.56*  0.33*  −0.59*  % Wheat          −0.36*  −0.35*  −0.48*  % Summer crops            −0.72*  0.10  % Grass              0.13  View Large Table 3. Correlation among variables describing landscape context within 3.2 km radius areas centered on each of n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Shannon diversity  Fractal dimension  Contagion  % Wheat  % Summer crops  % Grass  % Wooded  Patch density  0.60*  0.26*  −0.32*  −0.25*  0.50*  −0.36*  0.28*  Shannon diversity    −0.38*  −0.90*  −0.45*  0.77*  −0.52*  0.45*  Fractal dimension      −0.50*  −0.09  0.10  −0.16  0.49*  Contagion        0.46*  −0.56*  0.33*  −0.59*  % Wheat          −0.36*  −0.35*  −0.48*  % Summer crops            −0.72*  0.10  % Grass              0.13  Variable  Shannon diversity  Fractal dimension  Contagion  % Wheat  % Summer crops  % Grass  % Wooded  Patch density  0.60*  0.26*  −0.32*  −0.25*  0.50*  −0.36*  0.28*  Shannon diversity    −0.38*  −0.90*  −0.45*  0.77*  −0.52*  0.45*  Fractal dimension      −0.50*  −0.09  0.10  −0.16  0.49*  Contagion        0.46*  −0.56*  0.33*  −0.59*  % Wheat          −0.36*  −0.35*  −0.48*  % Summer crops            −0.72*  0.10  % Grass              0.13  View Large Aphid Abundance in Wheat Fields Cereal aphid abundance was estimated for each wheat field at the times sentinel plants were deployed. Aphid abundance in autumn was correlated with some landscape and within field variables. Autumn aphid abundance was correlated with whether the field was used for cattle grazing (r = 0.38; n = 70; P = 0.001), but not with any other of the within-field variables measured. Since cattle were not placed on fields until after our autumn study was complete, this effect may have been due to planting date, which is earlier for dual purpose wheat fields than for fields used only for grain production (Epplin et al. 1998). Aphid abundance in autumn was positively correlated with % grassland in the landscape (r = 0.28; n = 70; P = 0.02), but not with any other landscape metric. For spring, aphid abundance was negatively correlated with both tillage type (conventional vs. no-till) (r = −0.47; n = 46; P = 0.001) and crop rotation (r = −0.58; n = 46; P < 0.001). Since no-till is commonly practiced in conjunction with crop rotation in central Oklahoma, the two variables were highly correlated (r = 0.84; n = 70; P < 0.001). Aphid abundance in wheat fields in spring was positively correlated with fractal dimension (r = 0.29; n = 46; P = 0.05) and negatively correlated with % summer crops (r = −0.30; n = 46; P = 0.04); the correlation with % grassland for spring was positive (r = 0.24) but not significant. Parasitism of Bird Cherry–Oat Aphids by L. testaceipes Both the average number of L. testaceipes per sentinel plant and the proportion of sentinel plants with L. testaceipes were correlated with landscape variables for autumn and spring. Correlations for the number of L. testaceipes per sentinel plant were generally greater in magnitude and more often significant than correlations for the proportion of sentinel plants with L. testaceipes. This was not surprising since only seven sentinel plants were deployed per field which limits interpretation of this measure. Therefore, we limit further analysis of the relationship of parasitism with landscape variables to the number of L. testaceipes per sentinel plant. Neither aphid abundance nor any within field or field edge variable was correlated with the number of L. testaceipes per sentinel plant in autumn (not shown). In spring, there was a significant correlation between aphid abundance in wheat fields and number of L. testaceipes per sentinel plant (r = 0.35; n = 46; P = 0.02). None of the other within field or field edge variables was significantly correlated with the number of L. testaceipes per sentinel plant in spring. For landscape metrics, significant correlations occurred for the number of L. testaceipes per sentinel plant with patch density, fractal dimension, Shannon’s patch diversity, percent wheat, percent summer crops, and percent wooded land (Table 4). Correlations with percent wheat were negative for all spatial extents in autumn, but were not significant in spring. Correlations with patch density were positive in autumn but negative in spring (Fig. 2). Correlations with percent summer crops and fractal dimension were positive for all spatial extents and often significant in both autumn and spring (Fig. 2). Correlations for contagion were negative for all spatial extents in both seasons, and were significant in spring. Correlations with Shannon’s patch diversity index were positive for all spatial extents in spring, but significant only at 1.6 km. The correlation with percent wooded land was positive and significant at 3.2 km, but not at finer spatial extents. Table 4. Correlation of landscape metrics with the number of L. testaceipes per sentinel plant and the percent of sentinel plants with L. testaceipes present for autumn (n = 70) and spring (n = 46) Metric  Autumn  Spring  No. L. testaceipes  % with L. testaceipes  No. L. testaceipes  % with L. testaceipes  Radius 0.8 km   Patch density  0.26*  0.23  −0.43*  −0.05   Shannon’s patch diversity  0.13  0.07  0.01  0.22   Contagion  −0.10  −0.09  −0.29*  −0.26   Fractal dimension  0.10  0.06  0.11  0.11   % Wheat  −0.34*  −0.25*  0.01  −0.15   % Summer crops  0.39*  0.18  0.17  0.12   % Grassland  −0.13  0.01  0.24  0.04   % Wooded  −0.04  −0.04  0.09  0.15  Radius 1.6 km   Patch density  0.32*  0.30*  −0.50*  −0.12   Shannon’s patch diversity  0.21  0.17  0.33*  0.10   Contagion  −0.12  −0.15  −0.29*  −0.43*   Fractal dimension  0.18  0.18  0.39*  0.25   % Wheat  −0.31*  −0.35*  0.01  −0.17   % Summer crops  0.29*  0.16  0.27  0.10   % Grassland  −0.08  0.08  0.20  0.00   % Wooded  0.07  0.08  0.22  0.22  Radius 3.2 km   Patch density  0.32*  0.29*  −0.61*  −0.21   Shannon’s patch diversity  0.20  0.15  0.26  0.16   Contagion  −0.14  −0.18  −0.30*  −0.28   Fractal dimension  0.26*  0.31*  0.29*  0.20   % Wheat  −0.34*  −0.40*  −0.01  −0.20   % Summer crops  0.31*  0.13  0.27  0.10   % Grassland  −0.07  0.12  0.25  0.05   % Wooded  0.24*  0.19  0.07  0.18  Metric  Autumn  Spring  No. L. testaceipes  % with L. testaceipes  No. L. testaceipes  % with L. testaceipes  Radius 0.8 km   Patch density  0.26*  0.23  −0.43*  −0.05   Shannon’s patch diversity  0.13  0.07  0.01  0.22   Contagion  −0.10  −0.09  −0.29*  −0.26   Fractal dimension  0.10  0.06  0.11  0.11   % Wheat  −0.34*  −0.25*  0.01  −0.15   % Summer crops  0.39*  0.18  0.17  0.12   % Grassland  −0.13  0.01  0.24  0.04   % Wooded  −0.04  −0.04  0.09  0.15  Radius 1.6 km   Patch density  0.32*  0.30*  −0.50*  −0.12   Shannon’s patch diversity  0.21  0.17  0.33*  0.10   Contagion  −0.12  −0.15  −0.29*  −0.43*   Fractal dimension  0.18  0.18  0.39*  0.25   % Wheat  −0.31*  −0.35*  0.01  −0.17   % Summer crops  0.29*  0.16  0.27  0.10   % Grassland  −0.08  0.08  0.20  0.00   % Wooded  0.07  0.08  0.22  0.22  Radius 3.2 km   Patch density  0.32*  0.29*  −0.61*  −0.21   Shannon’s patch diversity  0.20  0.15  0.26  0.16   Contagion  −0.14  −0.18  −0.30*  −0.28   Fractal dimension  0.26*  0.31*  0.29*  0.20   % Wheat  −0.34*  −0.40*  −0.01  −0.20   % Summer crops  0.31*  0.13  0.27  0.10   % Grassland  −0.07  0.12  0.25  0.05   % Wooded  0.24*  0.19  0.07  0.18  *Statistically significant correlations (P < 0.05). View Large Table 4. Correlation of landscape metrics with the number of L. testaceipes per sentinel plant and the percent of sentinel plants with L. testaceipes present for autumn (n = 70) and spring (n = 46) Metric  Autumn  Spring  No. L. testaceipes  % with L. testaceipes  No. L. testaceipes  % with L. testaceipes  Radius 0.8 km   Patch density  0.26*  0.23  −0.43*  −0.05   Shannon’s patch diversity  0.13  0.07  0.01  0.22   Contagion  −0.10  −0.09  −0.29*  −0.26   Fractal dimension  0.10  0.06  0.11  0.11   % Wheat  −0.34*  −0.25*  0.01  −0.15   % Summer crops  0.39*  0.18  0.17  0.12   % Grassland  −0.13  0.01  0.24  0.04   % Wooded  −0.04  −0.04  0.09  0.15  Radius 1.6 km   Patch density  0.32*  0.30*  −0.50*  −0.12   Shannon’s patch diversity  0.21  0.17  0.33*  0.10   Contagion  −0.12  −0.15  −0.29*  −0.43*   Fractal dimension  0.18  0.18  0.39*  0.25   % Wheat  −0.31*  −0.35*  0.01  −0.17   % Summer crops  0.29*  0.16  0.27  0.10   % Grassland  −0.08  0.08  0.20  0.00   % Wooded  0.07  0.08  0.22  0.22  Radius 3.2 km   Patch density  0.32*  0.29*  −0.61*  −0.21   Shannon’s patch diversity  0.20  0.15  0.26  0.16   Contagion  −0.14  −0.18  −0.30*  −0.28   Fractal dimension  0.26*  0.31*  0.29*  0.20   % Wheat  −0.34*  −0.40*  −0.01  −0.20   % Summer crops  0.31*  0.13  0.27  0.10   % Grassland  −0.07  0.12  0.25  0.05   % Wooded  0.24*  0.19  0.07  0.18  Metric  Autumn  Spring  No. L. testaceipes  % with L. testaceipes  No. L. testaceipes  % with L. testaceipes  Radius 0.8 km   Patch density  0.26*  0.23  −0.43*  −0.05   Shannon’s patch diversity  0.13  0.07  0.01  0.22   Contagion  −0.10  −0.09  −0.29*  −0.26   Fractal dimension  0.10  0.06  0.11  0.11   % Wheat  −0.34*  −0.25*  0.01  −0.15   % Summer crops  0.39*  0.18  0.17  0.12   % Grassland  −0.13  0.01  0.24  0.04   % Wooded  −0.04  −0.04  0.09  0.15  Radius 1.6 km   Patch density  0.32*  0.30*  −0.50*  −0.12   Shannon’s patch diversity  0.21  0.17  0.33*  0.10   Contagion  −0.12  −0.15  −0.29*  −0.43*   Fractal dimension  0.18  0.18  0.39*  0.25   % Wheat  −0.31*  −0.35*  0.01  −0.17   % Summer crops  0.29*  0.16  0.27  0.10   % Grassland  −0.08  0.08  0.20  0.00   % Wooded  0.07  0.08  0.22  0.22  Radius 3.2 km   Patch density  0.32*  0.29*  −0.61*  −0.21   Shannon’s patch diversity  0.20  0.15  0.26  0.16   Contagion  −0.14  −0.18  −0.30*  −0.28   Fractal dimension  0.26*  0.31*  0.29*  0.20   % Wheat  −0.34*  −0.40*  −0.01  −0.20   % Summer crops  0.31*  0.13  0.27  0.10   % Grassland  −0.07  0.12  0.25  0.05   % Wooded  0.24*  0.19  0.07  0.18  *Statistically significant correlations (P < 0.05). View Large Fig. 2. View largeDownload slide Number of L. testaceipes per seven sentinel plants versus patch density and fractal dimension for autumn (top) and spring (bottom). Fig. 2. View largeDownload slide Number of L. testaceipes per seven sentinel plants versus patch density and fractal dimension for autumn (top) and spring (bottom). Correlations between the number of L. testaceipes per sentinel plant and landscape variables for autumn were generally greatest at the 3.2 km radius and correlations in spring were similar at 1.6 and 3.2 km (Table 4), therefore, we used landscape data measured at the 3.2 km scale in regression modeling for autumn and spring. Aphid abundance in autumn, and other variables measured in wheat fields and field edges, were uncorrelated with the number of L. testaceipes per sentinel plant and therefore were not used as predictors in regression models. Aphid abundance in wheat fields in spring was correlated with parasitism and was included as a predictor in stepwise regressions. Based on the scree method, we included five of eight factors (Supplementary Appendix 1) as predictors of the number of L. testaceipes per sentinel plant in stepwise regression models for autumn and spring. The best fitting first-order regression model for the number of L. testaceipes in autumn (F = 6.72; df = 4, 65; P = 0.0001) was  Lt= 11.56 + 4.90·F1– 7.76·F3+ 4.93·F4 +5.66·F5(R2= 0.30) (1) where Lt is the number of L. testaceipes per sentinel plant, and F1, F3, F4, and F5 are factors derived from principal components analysis (Supplementary Appendix 1). For spring, the best fitting regression model (F = 13.34; df = 42, 3; P = 0.0001) was  Lt= 179.95 – 27.43·F1+ 40.39·F4– 95.52·F5(R2= 0.49) (2) where Lt is the number of L. testaceipes per sentinel plant and F1, F4, and F5 are factors. Regression models for L. testaceipes parasitism in wheat fields in autumn and spring were interpreted based on standardized factor loadings on the original landscape variables. For the autumn regression model (Eq. 1) four factors, F1, F3, F4, and F5 were entered into the model using the selection criterion of P < 0.15 for inclusion of a variable. The regression coefficient for F1 was positive, and the largest loadings on F1 were for % summer crops (0.75) and % grassland (−0.71), indicating that the presence of high acreage of summer crops increased parasitism in wheat in autumn, whereas the presence of large amounts of grassland was detrimental to parasitism. Factor F3 had a negative regression coefficient and was dominated by a large loading on % wheat (1.23), indicating a negative effect of wheat mono-cropping on parasitism in autumn. The positive regression coefficient for F4 combined with the large factor loading on patch density (1.18) indicates a positive influence of small patch (field) size in the landscape on parasitism. Finally, the positive regression coefficient for F5 combined with the dominant positive loading on fractal dimension (1.19) indicates a positive effect of curvilinear patch boundaries, characteristic of semi-natural and natural lands, on parasitism. In spring, three factors were entered into the regression, F1, F4, and F5 (Eq. 2). In contrast to autumn, F1 had a negative regression coefficient indicating a reversal of the relationship between percent of landscape planted to summer crops and the percent grassland from that in autumn. F4 had a positive regression coefficient indicating a positive response of parasitism to increasing fractal dimension. The coefficient of F5 was negative indicating a negative response of parasitism to increasing patch density, which is opposite of the response observed in autumn. Discussion Variables representing landscape composition as well as spatial configuration were correlated with parasitism levels by L. testaceipes in wheat fields. Correlations were observed both in autumn when colonization of wheat fields by cereal aphids and L. testaceipes first occurs, and in spring when within field processes might be expected to predominate over colonization in determining L. testaceipes population density and the resulting cereal aphid parasitism. It is notable that L. testaceipes responds to landscape composition (e.g., % wheat and Shannon’s patch diversity index) as has been observed for braconid parasitoids of aphids in several studies (e.g., Roschewitz et al. 2005, Thies et al. 2005, Plecas et al. 2014, Zhao et al. 2015), but also to landscape configuration (e.g., fractal dimension and contagion), which has been less frequently measured (but see Plecas et al. 2014), and may be important in determining arthropod population processes in agroecosystems. Large acreage of wheat surrounding a focal wheat field was negatively correlated with cereal aphid parasitism in autumn, but was uncorrelated with parasitism in spring (Table 4). Acreage of summer crops was positively correlated with parasitism in autumn, but negatively correlated in spring (Table 4). The autumn and spring regression models (Eqs. 1 and 2) reflect these differing associations of parasitism with landscape variables for autumn and spring. Additionally, patch density had a positive effect on parasitism in autumn, but a negative effect in spring. Conversely, fractal dimension had a positive effect in both seasons (Eqs. 1 and 2). In the agricultural landscapes of north central Oklahoma, both cropland and semi-natural land habitats contribute to the landscape scale population dynamics of L. testaceipes. The autumn and spring regression models reflect the differing associations of parasitism with landscape variables for autumn and spring. Differences in effects of landscape context among seasons suggest that the response of L. testaceipes to landscape context is very rapid, within a single growing season. L. testaceipes dynamics within a wheat field, and correspondingly, capacity for cereal aphid suppression, is partially dependent on population dynamics of the parasitoid in the landscape surrounding a wheat field during the course of the growing season. In the heavily agricultural landscapes of north central Oklahoma, cropland and semi-natural lands both contribute to determining the landscape scale population dynamics of L. testaceipes. Considered in this context, the change in regression coefficient (positive to negative) of parasitism to presence of summer crops in autumn and spring is likely due the fact that summer crops serve as habitat for L. testaceipes in early autumn when wheat is first planted, but are fallow fields in spring. The negative effect of amount of grassland on parasitism may be a consequence of the extremely arid and high temperature summer conditions in Oklahoma, which result in mostly dormant grasslands that harbor extremely low numbers of aphid hosts and therefore of parasitoids (Anstead 2000). Variation in the effect of wheat acreage on parasitism can be viewed similarly, where wheat fields are essentially devoid of cereal aphid hosts in early autumn but serve as habitat for cereal aphids and L. testaceipes in spring. The varying effect of increasing patch density on parasitism from positive in fall to negative in spring is difficult to interpret. If small patch size (high patch density) increased aphid parasitism by L. testaceipes in wheat fields it’s effect would most likely result from increasing access to habitats containing hosts or other resources. The magnitude of this effect might vary seasonally but would not be expected to shift from positive to negative. The influence of patch density on parasitism is probably indirect and related to other landscape characteristics, such as those captured by Shannon’s diversity index, that influence resource availability to L. testaceipes, are correlated with patch density (Table 3), and are not explicity accounted for in regression models. This result highlights the difficulty in teasing out the effects of numerous interacting factors on L. testaceipes ecology using multivariate methods. Even with this limitation the study has highlighted several landscape characteristics that, through their influence on resource availability influence the success of L. testaceipes as a biological control agent in wheat agreoecosystems in Oklahoma. In Oklahoma, the correlation of aphid parasitism by L. testaceipes with landscape variables in wheat fields did not decrease with increasing spatial extent, contrary to observations for braconid parasitoids of cereal aphids in European wheat agroecosystems (Thies et al. 2005). L. testaceipes was not recorded in their study even though it is widely established in European fauna (Stary 1988, Zikic et al. 2015). Interestingly, our finest spatial extent 0.8 km radius (1.6 km diameter) was closest to the maximum spatial extent (2.0 km diameter) at which landscape complexity significantly explained variation in parasitism in wheat fields in Europe (Thies et al. 2005). Our results suggest that L. testaceipes responds to landscape variation over a broad spatial extent. L. testaceipes shows a very limited response to plant volatiles (Lo Pinto et al. 2004, Fauvergue et al. 2006), and is considered to be both a habitat and host generalist with potential for very rapid population growth (Mackauer and Stary 1967, Jones et al. 2003). These life history characteristics might be expected to be associated with high levels of dispersal. Although life histories of most braconid aphid parasitoid species are not well known, L. testaceipes is probably exceptional in relation to breadth of host and habitat range (Stary et al. 1988, Pike et al. 2000). Because of its life history characteristics L. testaceipes appears to be well adapted to the agroecosystems of the Southern Plains, where agricultural landscapes are coarse grained and habitats are poorly interspersed. L. testaceipes is a key natural enemy of cereal aphids in wheat (Jones et al. 2007, Giles et al. 2017), and effective biological control frequently occurs within wheat fields prior to the aphid infestation reaching the economic threshold (Giles et al. 2003). L. testaceipes does not appear to disperse among a set of essential habitats seasonally and therefore does not exhibit the characteristics of a cyclic colonizer (Wissinger 1997). Rather L. testaceipes appears to utilize multiple habitats throughout the year depending on their availability and acceptability, and frequently disperses among habitats (Jessie 2017). In this sense L. testaceipes is better described as an r-selected species, with high reproductive and dispersal rates (Ehler and Miller 1978, Price and Waldbauer 1994). The ability of L. testaceipes to control aphids in wheat fields depends on rapid colonization of fields (Bortoloto et al. 2015) and high attack and reproductive rates (Jones 2005). Colonization of wheat fields in autumn is enhanced by proximity to fields of summer crops and semi-natural habitats other than grasslands. Based on this study, the optimal landscape to promote biological control of cereal aphids in wheat by L. testaceipes is one where summer crops, wheat, and semi-natural habitats exist in significant amounts, and are well interspersed. Future emphasis should be given to determining the specific resources within these habitats that are utilized by L. testaceipes. Supplementary Data Supplementary data are available at Environmental Entomology online. Acknowledgments We thank Tim Johnson for technical assistance with the project and for coordinating data collection and processing activities. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture (USDA). USDA is an equal opportunity provider and employer. References Cited Anstead, J. A. 2000. Genetic and biotypic diversity of greenbug Schizaphis graminum (Rondani) populations on non-cultivated hosts. Doctoral Dissertation . Oklahoma State University, Stillwater, OK. Bortolotto, O. C., A. D. O. M. Júnior, and Hoshino, A. T. 2015. Aphidophagous parasitoids can forage wheat crops before aphid infestation, Paraná State, Brazil. J. Insect Sci . 15. doi: 10:1093/jisesa/iev027 Brewer, M. J., T. Noma, N. C. Elliott, A. N. Kravchenko, and A. L. Hild. 2008. A landscape view of cereal aphid parasitoid dynamics reveals sensitivity to farm- and region-scale vegetation structure. European J. Entomol . 105: 503– 511. Google Scholar CrossRef Search ADS   Dillon, W. R., and Goldstein, M. 1984. Multivariate analysis: methods and applications . John Wiley and Sons, New York, NY. Eikenbary, R. D., and C. E. Rogers. 1974. Importance of alternate hosts in establishment of introduced parasites, pp 119– 133. In Tall Timbers Conference on Ecological Animal Control by Habitat Management; Tallahassee, FL. Ehler, L. E., and J. C. Miller. 1978. Biological control in temporary agroecosystems. Entomophaga . 3: 207– 212. Google Scholar CrossRef Search ADS   Elliott, N. C., Kieckhefer, R. W., Lee, J. H., and French, B. W. 1998. Influence of within-field and landscape factors on aphid predator populations in wheat. Landscape Ecol . 14: 239– 252. Google Scholar CrossRef Search ADS   Epplin, F. M., R. R. True, and E. G. Krenzer, Jr. 1998. Practices used by Oklahoma wheat growers by region. Okla. Curr. Farm Econ . 71: 14– 24. Fauvergue, X., R. Boll, J. Rochat, E. Wajnberg, C. Bernstein, and L. Lapchin. 2006. Habitat assessment by parasitoids: consequences for population distribution. Behav. Ecol . 17: 522– 531. Google Scholar CrossRef Search ADS   Giles, K. L., D. B. Jones, T. A. Royer, N. C. Elliott, and S. D. Kindler. 2003. Development of a sampling plan in winter wheat that estimates cereal aphid parasitism levels and predicts population suppression. J. Econ. Entomol . 96: 975– 982. Google Scholar CrossRef Search ADS PubMed  Giles, K. L., B. P. McCornack, T. A. Royer, and N. C. Elliott. 2017. Incorporating biological control into IPM decision making. Curr. Opin. Insect Sci . 20: 84– 89. Google Scholar CrossRef Search ADS PubMed  Hand, S. C. 1986. The capture efficiency of the Dietrick vacuum insect net for aphids on grasses and cereals. Ann. Appl. Biol . 108: 1744– 7348. Google Scholar CrossRef Search ADS   Hight, S. C., R. D. Eikenbary, R. J. Miller, and K. J. Starks. 1972. The greenbug and Lysiphlebus testaceipes. Environ. Entomol . 1: 205– 209. Google Scholar CrossRef Search ADS   Jessie, C. N. 2017. Diversity, relative abundance, movement, and fitness of insect predators in winter canola agricultural landscapes in the United States Southern Plains. Doctoral Dissertation . Oklahoma State University, Stillwater, OK. Jones, D. B. 2005. Winter ecology of the parasitoid Lysiphlebus testaceipes Cresson. Doctoral Dissertation . Oklahoma State University, Stillwater, OK. Jones, D. B., K. L. Giles, R. C. Berberet, T. A. Royer, N. C. Elliott, and M. Payton. 2003. Functional responses of an introduced parasitoid and an indigenous parasitoid on greenbug at four temperatures. Environ. Entomol . 32: 425– 432. Google Scholar CrossRef Search ADS   Jones, D. B., K. L. Giles, N. C. Elliott, and M. E. Payton. 2007. Parasitism of Greenbug, Schizaphis graminum, by the parasitoid Lysiphlebus testaceipes at winter temperatures. Environ. Entomol . 36: 1– 8. Google Scholar CrossRef Search ADS PubMed  Lo Pinto, M., E. Wajnberg, S. Colazza, C. Curty, and X. Fauvergue. 2004. Olfactory response of two aphid parasitoids, Lysiphlebus testaceipes and Aphidius colemani, to aphid-infested plants from a distance. Entomol. Exp. Applic . 110: 159– 164. Google Scholar CrossRef Search ADS   Losey, J. E., and R. F. Denno. 1998. Positive predator-predator interactions: enhanced predation rates and synergistic suppression of aphid populations. Ecology . 79: 2143– 2152. Mackauer, M., and P. Stary. 1967. Index of entomophagous insects (hym., ichneumonidae): index of world aphidiidae . Le Francois, Paris, France. McGarigal, K. 2014. FRAGSTATS: spatial pattern analysis program for categorical maps. http://www.umass.edu/landeco/research/fragstats/fragstats.html McGarigal, K., and B. J. Marks. 1995. FRAGSTATS. Spatial analysis program for quantifying landscape structure. USDA Forest Service General Technical Report PNWGTR-351 . USDA Forest Service, Pacific Northwest Research Station, Portland, OR. O’Neill, R. V., J. R. Krummel, R. H. Gardner, G. Sugihara, B. Jackson, D. L. DeAngelis, B. T. Milne, M. G. Turner, B. Zygmunt, S. W. Christensen, et al.   1988. Indices of landscape pattern. Landscape Ecol . 1: 153– 162. Google Scholar CrossRef Search ADS   Pike, K. S., P. Starý, T. Miller, G. Graf, D. Allison, L. Boydston, and R. Miller. 2000. Aphid parasitoids (Hymenoptera: Braconidae: Aphidiinae) of Northwest USA. Proc. Entomol. Soc. Wash . 102: 688– 740. Plećaš, M., V. Gagić, M. Janković, O. Petrović-Obradović, N. G. Kavallieratos, Ž. Tomanović, and A. Ćetković. 2014. Landscape composition and configuration influence cereal aphid–parasitoid–hyperparasitoid interactions and biological control differentially across years. Agric. Ecosyst. Environ . 183: 1– 10. Google Scholar CrossRef Search ADS   Price, P. W., and G. P. Waldbauer. 1994. Ecological aspects of pest management, pp. 33– 68. In R. L. Metcalf and W. H. Luckman (eds.), Introduction to insect pest management . Wiley, New York, NY. Riitters, K. H., R. V. Oneill, C. T. Hunsaker, J. D. Wickham, D. H. Yankee, S. P. Timmins, K. B. Jones, and B. L. Jackson. 1995. A factor analysis of landscape pattern and structure metrics. Landscape Ecol . 10: 23– 39. Google Scholar CrossRef Search ADS   Roschewitz, I., M. Hücker, T. Tscharntke, and C. Thies. 2005. The influence of landscape context and farming practices on parasitism of cereal aphids. Agric., Ecosyst. Environ . 108: 218– 227. Google Scholar CrossRef Search ADS   Royer, T. A., K. L. Giles, and N. C. Elliott. 2004. The cereal aphid expert system and Glance’ n Go Sampling, questions and answers. CR-7191 . Oklahoma Cooperative Extension Service, Stillwater, OK. Royer, T. A., B. B. Pendleton, N. C. Elliott, and K. L. Giles. 2015. Greenbug (Hemiptera: Aphididae) biology, ecology, and management in wheat and sorghum. J. Integ. Pest Mngmt . 6: 19. Google Scholar CrossRef Search ADS   Rusch, A., R. Chaplin-Kramer, M. M. Gardiner, V. Hawro, J. Holland, D. Landis, C. Thies, T. Tscharntke, W. W. Weisser, C. Winqvist, et al.   2016. Agricultural landscape simplification reduces natural pest control: a quantitative synthesis. Agric. Ecosyst. Environ . 221: 198– 204. Google Scholar CrossRef Search ADS   SAS Institute. 2004. SAS/STAT user’s guide, version 9.1 . SAS Institute, Cary, NC. Schmidt, M. H., C. Thies, and T. Tscharntke. 2004. Landscape context of arthropod biological control, pp. 55– 63. In G. M. Gurr, S. D. Wratten, and M. A. Altieri (eds.), Ecological engineering for pest management: advances in habitat manipulation for arthropods . CSIRO Publishing, Collingswood, VIC, Australia. Google Scholar CrossRef Search ADS   Shannon, C., and W. Weaver. 1949. The mathematical theory of communication . Univ. Illinois Press, Urbana, IL. Starý, P., J. P. Lyon, and F. Leclant. 1988. Biocontrol of aphids by the introduced Lysiphlebus testaceipes (Cress.) (Hym., Aphidiidae) in Mediterranean France. J. Appl. Entomol . 105: 74– 78. Google Scholar CrossRef Search ADS   Thies, C., I. Roschewitz, and T. Tscharntke. 2005. The landscape context of cereal aphid-parasitoid interactions. Proc. Biol. Sci . 272: 203– 210. Google Scholar CrossRef Search ADS PubMed  Turner, M. G., R. H. Gardner, and R. V. O’Neill. 2001. Landscape ecology in theory and practice . Springer-Verlag, New York, NY. USDA, NASS (National Agricultural Statistics Survey). 2017. http://www.nass.usda.gov/. Webster, F. M., and W. J. Phillips. 1912. The spring grain aphis or “greenbug”, U.S. Bur. Entomol. Bull . 110: 1– 153. https://archive.org/details/springgrainaphis110webs Wissinger, S. A. 1997. Cyclic colonization in predictably ephemeral habitats: a template for biological control in annual crop systems. Biol. Control . 10: 4– 15. Google Scholar CrossRef Search ADS   Zadoks, J. C., T. T. Chang, and Konzak, C. F. 1974. A decimal code for the growth stages of cereals. Weed Res . 14: 415– 421. Google Scholar CrossRef Search ADS   Zhao, Z. H., H. S. Sandhu, F. Gao, and D. H. He. 2015. Shifts in natural enemy assemblages resulting from landscape simplification account for biocontrol loss in wheat fields. Ecol. Res . 30: 493– 498. Google Scholar CrossRef Search ADS   Žikić, V., S., Stanković, M. Milošević, O. Petrović-Obradović, A. Petrović, P. Starý, and Ž. Tomanović. 2015. First detection of Lysiphlebus testaceipes (Cresson) (Hymenoptera: Aphidiinae) in Serbia: an introduced species invading Europe? N.W. J. Zool . 11: 97– 101. Published by Oxford University Press on behalf of Entomological Society of America 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US. This work is written by (a) US Government employee(s) and is in the public domain in the US. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Environmental Entomology Oxford University Press

Landscape Context Affects Aphid Parasitism by Lysiphlebus testaceipes (Hymenoptera: Aphidiinae) in Wheat Fields

Loading next page...
 
/lp/ou_press/landscape-context-affects-aphid-parasitism-by-lysiphlebus-testaceipes-0n30K3yGGC
Publisher
Entomological Society of America
Copyright
Published by Oxford University Press on behalf of Entomological Society of America 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US.
ISSN
0046-225X
eISSN
1938-2936
D.O.I.
10.1093/ee/nvy035
Publisher site
See Article on Publisher Site

Abstract

Abstract Winter wheat is Oklahoma’s most widely grown crop, and is planted during September and October, grows from fall through spring, and is harvested in June. Winter wheat fields are typically interspersed in a mosaic of habitats in other uses, and we hypothesized that the spatial and temporal composition and configuration of landscape elements, which contribute to agroecosystem diversity also influence biological control of common aphid pests. The parasitoid Lysiphlebus testaceipes (Cresson; Hymenoptera: Aphidiinae) is highly effective at reducing aphid populations in wheat in Oklahoma, and though a great deal is known about the biology and ecology of L. testaceipes, there are gaps in knowledge that limit predicting when and where it will be effective at controlling aphid infestations in wheat. Our objective was to determine the influence of landscape structure on parasitism of cereal aphids by L. testaceipes in wheat fields early in the growing season when aphid and parasitoid colonization occurs and later in the growing season when aphid and parasitoid populations are established in wheat fields. Seventy fields were studied during the three growing seasons. Significant correlations between parasitism by L. testaceipes and landscape variables existed for patch density, fractal dimension, Shannon’s patch diversity index, percent wheat, percent summer crops, and percent wooded land. Correlations between parasitism and landscape variables were generally greatest at a 3.2 km radius surrounding the wheat field. Correlations between parasitism and landscape variables that would be expected to increase with increasing landscape diversity were usually positive. Subsequent regression models for L. testaceipes parasitism in wheat fields in autumn and spring showed that landscape variables influenced parasitism and indicated that parasitism increased with increasing landscape diversity. Overall, results indicate that L. testaceipes utilizes multiple habitats throughout the year depending on their availability and acceptability, and frequently disperses among habitats. Colonization of wheat fields by L. testaceipes in autumn appears to be enhanced by proximity to fields of summer crops and semi-natural habitats other than grasslands. cereal aphid, biological control, Hymenoptera, parasitoid Winter wheat is Oklahoma’s most widely grown crop, with more than 2 million ha planted annually (Epplin et al. 1998, USDA NASS 2017). Winter wheat in Oklahoma is planted during September and October, grows from fall through spring, and is harvested in June. Several aphid species infest wheat fields in this region, the most common and important being greenbug, Schizaphis graminum (Rondani); bird cherry-oat aphid, Rhopalosiphum padi (L.); and English grain aphid, Sitobion avenae (F.) (Heteroptera: Aphididae). Apart from the use of insecticidal seed treatments, which are used preventively in Oklahoma by an increasing number of producers, insecticide applications are infrequent, and primarily used to control insects such as fall armyworm in autumn, and cereal aphids, primarily greenbug, in autumn or spring (Royer et al. 2015). During the wheat growing season, winter wheat fields are typically interspersed in a mosaic of lands in other uses. Grasslands (pasture and rangeland) with varying levels of management are the most abundant land use type. Grasslands range from semi-natural lands that have never been cultivated and have high plant species diversity to highly managed lands planted to a single grass species. Fallow fields will mostly be planted to summer crops (soybean, corn, sorghum, and cotton) in spring, and fields may also be planted to other winter crops, such as canola and barley. Riparian areas and other semi-natural lands are also present in the landscape. Based on previous studies that investigated pest suppression in agricultural landscapes, the spatial and temporal configuration of landscape elements, which contribute to agroecosystem diversity are likely to influence populations of natural enemies, and possibly, biological control of insect pests by determining the availability of resources for beneficial insects (Rusch et al. 2016). In Oklahoma the parasitoid Lysiphlebus testaceipes (Cresson; Hymenoptera: Aphidiinae) is highly effective at reducing aphid populations in winter wheat (Webster and Phillips 1912, Giles et al. 2003, Jones et al. 2007, Royer et al. 2015). The parasitoid’s effectiveness is thought to be the result of features of its biology and ecology. It attacks multiple aphid species during cold but non-lethal fall and winter months where daytime temperatures often exceed the species’ activity threshold (Jones et al. 2007), it has high attack and reproductive rates (Jones et al. 2003, Giles et al. 2003), it sterilizes the aphids it attacks (Hight et al. 1972, Eikenbary and Rogers 1974), and it dislodges aphids from the plant as it forages, making them subject to mortality from predation and the environment (Losey and Denno 1998). L. testaceipes has been observed to keep greenbug populations below the economic injury level in wheat (Eikenbary and Rogers 1974, Giles et al. 2003), and its impact has been successfully incorporated in pest sampling programs and management guidelines for greenbug (Giles et al. 2003; Royer et al. 2004, 2015). Even though a great deal is known about the biology and ecology of the species, there are gaps in knowledge that limit understanding of the L. testaceipes/aphid system and predicting when and where L. testaceipes will be effective at controlling infestations in wheat. The structure of the landscape surrounding a particular agricultural field has been shown to influence populations of predatory insects in wheat fields (Elliott et al. 1998), and parasitoid abundance and parasitism levels of herbivorous insects in other agricultural ecosystems (Schmidt et al. 2004, Thies et al. 2005). However, only one study has specifically addressed effects of landscape context on L. testaceipes. Brewer et al. (2008) demonstrated that adding sunflower to a strip cropping system including wheat in southeastern Wyoming increased parasitism levels of Russian wheat aphid by L. testaceipes. Our objective was to determine the relative influence of landscape structure on parasitism of cereal aphids by L. testaceipes in wheat fields early in the growing season when aphid and parasitoid colonization occurs, and later in the growing season when within field aphid and parasitoid population processes may predominate over colonization. Our hypothesis was that the landscape context within which a wheat field is embedded influences the level of parasitism in a wheat field through its presumed effect on resource availability to L. testaceipes in other habitats, and this influence would be most pronounced early in the growing season (autumn) when colonization of wheat fields by aphids and L. testaceipes from other habitats primarily determines abundance and parasitism. Landscape context would be predicted to have less influence later in the growing season (spring) when aphid and parasitoid population dynamics within wheat fields predominates over colonization in determining abundance and parasitism. Materials and Methods Field Study We utilized aphid-infested sentinel barley plants in pots to quantify aphid parasitism within wheat fields in north central Oklahoma over a 3-yr period (Fig. 1). Potted wheat plants grown in a greenhouse and infested with parasitoid free bird cherry-oat aphids were set out in commercial wheat fields during autumn and spring. The technique has been successfully used to compare relative parasitism rates within and between wheat fields (Brewer et al. 2008). Each year, approximately 24 wheat fields were selected to achieve broad coverage of the area with randomization partially restricted by the availability of cooperating farmers and that study fields were separated by a minimum of 5 km. Fig. 1. View largeDownload slide Approximate boundaries of the geographic area where field studies were conducted during the 2008–2009, 2009–2010, and 2010–2011 wheat growing seasons. Fig. 1. View largeDownload slide Approximate boundaries of the geographic area where field studies were conducted during the 2008–2009, 2009–2010, and 2010–2011 wheat growing seasons. To determine if landscape effects on parasitism existed during autumn and spring, the field study described below was repeated in autumn of three consecutive years, 2008–2010, and in spring of two consecutive years, 2009 and 2010. Studies were initiated in October, soon after wheat plants emerged from the soil, and during mid-March when aphid and parasitoid activity is common (Giles et al. 2003). Approximately 10 barley seeds (variety Eight Twelve) were planted in a 2:1 mixture of peat moss and fritted clay in 15.2 cm diameter plastic pots, and a 14 cm diameter by 35 cm high circular clear plastic cage was placed on each pot. Each cage was vented in the side and top with fine muslin cloth and pressed about 2.5 cm into the planting mixture. Fifteen pots containing caged barley plants (hereafter referred to as sentinel plants) were placed in each of 12 cubicle shaped cages (90 × 90 × 40 cm) covered with fine mesh screen on the four sides, a plywood bottom, and a clear plastic top. The double caging was done to ensure that aphids on sentinel plants remained parasitoid free until they were uncaged in the field. Two-week old sentinel plants in cages were infested with ca. 50 parasitoid free bird cherry-oat aphids. Approximately 10 d later, when aphid counts in cages averaged ca. 750 per pot, sentinel plants were transported to wheat fields where they were stationed as described below. One of the 15 sentinel plants from each cage (total of 12 plants) was retained in the greenhouse as a check to ensure that parasitoids had not entered cages and parasitized aphids while they were being grown in the greenhouse. In each wheat field, seven uncaged sentinel plants were placed into the soil and arranged 25 m apart in a T-shaped pattern. Sentinel plants were placed 5 m from a field edge, and every 25 m along a transect perpendicular to the field edge until five sentinel plants had been placed. Sentinel plants were also placed 25 m to the right and left of the fourth plant at 90-degree angles. Sentinel plants were left in fields for 3 d and then caged, returned to the greenhouse and maintained for 7 d at 21°C (±5°C), 16:8 (L:D) h, and ambient (uncontrolled) humidity to allow any parasitoids to undergo development to the pupal stage. After 7 d the barley plants from each pot were cut, separately placed in an emergence canister, and held for an additional 7 d to allow parasitoids to emerge as adults. Adult parasitoids in each emergence canister were counted and identified to species. In wheat fields in Oklahoma, L. testaceipes is the dominant cereal aphid parasitoid, usually accounting for over 95% of cereal aphid parasitism. During our study L. testaceipes accounted for over 98% of total parasitism in all fields in autumn and spring. Thus, the number of adult L. testaceipes per sentinel plant provided a useful measure of parasitism in a field. Check plants that remained under greenhouse conditions were processed identically to the experimental sentinel plants. No parasitoids were recovered from check plants during the 3-yr study indicating that it was very unlikely that contamination of aphid infested sentinel plants by parasitoids occurred in the greenhouse. We recorded several attributes for each wheat field. We recorded whether insecticidal seed treatment was used on the seed planted in a field, whether the field was in a no-till or conventional till system, whether the field was in a crop rotational system or in continuous wheat, the wheat plant growth stage, and aphid abundance in the field at the time sentinel plants were deployed. Wheat plant growth stage for each field was estimated using a 0–10 scale (Zadoks et al. 1974) on the day that sentinel plants were deployed. Aphid abundance in each field was estimated by sampling with a Backpack Model 24 D-Vac (Rincon-Vitova Insectaries, Inc., Ventura, CA) fitted with a standard 33 cm diameter collecting unit, fine mesh organdy collecting bag, and fiberglass collar. Sampling by D-Vac has been shown to provide useful estimates of aphid abundance in cereals (Hand 1986). A sample from each field consisting of three subsamples was taken within the area where the sentinel plants were deployed by walking three equally spaced linear transects. The first transect was situated parallel to and about 5 m to the left or right of the transect along which five of the seven sentinel plants had been stationed. The other two transects were parallel to and approximately 25 m to the right and left of the first transect. Approximately every 5 m the D-vac collecting unit was placed straight down over growing wheat plants to just above the soil surface until 20 such placements had been made. Each time the collecting unit was placed down it was held in position slightly above the soil surface for 5 s. After 20 stops, the sampling bag was removed from the D-vac and all arthropods in it were transferred to a labeled plastic bag. Bags were brought to the laboratory and placed in a freezer. Aphids were counted at a later date. Frozen aphids are difficult to identify to species so we did not distinguish among cereal aphid species. The mean number of aphids for the three transects sampled per field, each of which consisted of 20 D-vac placement subsamples (n = 3) provided our estimate of aphid abundance for each wheat field. Landscape Data Landscape context for each field was quantified for each of three circular areas centered on the focal wheat field with radii of 0.8, 1.6, and 3.2 km extracted from the USDA NASS Cropland Data Layer from the appropriate year. The Cropland Data Layer was acquired for the year in which winter wheat was present in the layer for the study fields for that year, and was used to quantify the amount and distribution of all land use types. The Cropland Data Layer differentiates crop types with accuracy rates typically above 85% (https://www.nass.usda.gov/Research_and_Science/Cropland/sarsfaqs2.php date accessed 5 March 2017). The crop-specific data is available at https://nassgeodata.gmu.edu/CropScape/ (date accessed 5 March 2017). Distances probably encompassed the range of potential dispersal ability of L. testaceipes (Thies et al. 2005). In addition to NASS data, we collected fine spatial scale ground survey data of the grass species in the wheat field boundaries. All grass species within one 15-m transect at an arbitrary location within each of two accessible field edges (e.g., adjacent to a road) were recorded for each field. One of the edges sampled was the one adjacent to where sentinel plants were deployed in the field. In addition, percent cover by Johnson grass, Sorghum halepense (L.), was estimated visually for an area of approximately 200 m2 in the boundary adjacent to the field. The cropland data layer was re-classified to retain eight land use categories: wheat, summer crops, winter crops other than wheat, fallow, grassland (pasture and rangeland), wooded, built areas and roads, and water. Aggregating land uses into fewer categories than represented in the original NASS data was desirable for calculating meaningful landscape metrics because many categories would have been represented by very small areas and metrics would be subject to high variability. We quantified landscape structure relative to each field in each year using the following landscape metrics: the proportion of the total area in each land use type, patch density, perimeter to area fractal dimension, Shannon’s patch diversity, and contagion (McGarigal and Marks 1995, McGarigal 2014). Landscape metrics quantify various characteristics of landscape structure that can be ecologically significant (O’Neill et al. 1988). Patch density, Shannon’s patch diversity, contagion, and perimeter to area fractal dimension are four among dozens of metrics. These four metrics have straight-forward interpretations, and the latter three were found by Ritters et al. (1995) in a study of several landscapes to be good quantitative descriptors of landscape structure that were relatively independent of one another. For the central Oklahoma landscapes in our study the three metrics were correlated among themselves, and were also correlated with patch density. Patch density measures the number of patches per km2 and indicates average patch size for a landscape. The perimeter to area fractal dimension is dimensionless and increases with increasing patch boundary curvilinearity. Contagion, measures the amount of clumping of patch types within a landscape as a percentage of the maximum. Maximum contagion for a given landscape would be achieved when each landscape element type occurred as a single contiguous patch. High contagion indicates highly aggregated and poorly interspersed patches. Shannon’s patch diversity index is based on information theory (Shannon and Weaver 1949) and measures landscape composition, not shape or configuration. Large values of Shannon’s diversity index indicate a greater number of landscape element types present (patch richness) in the landscape, greater evenness in area of the patch types present, or both. The number of patch types present did not vary much among landscapes in our study because most patch types were present. The minimal variation observed for patch richness indicates that variation in Shannon’s patch diversity index in our landscapes primarily reflected variation in the evenness in percent of area of various patch types. Large values for fractal dimension, patch density, and Shannon’s patch diversity, and small values for contagion generally indicate high landscape diversity (see Turner et al. 2001 for more information). Measured attributes of vegetation in field edges and landscape metrics calculated using Fragstats Version 4 were used to quantify landscape structure. Fragstats derived variables were calculated at each of the three hierarchically increasing spatial extents (0.8, 1.6, and 3.2 km radii) for each field each year. GIS operations were accomplished using ERDAS Imagine version 2014 including exporting data for calculating landscape metrics with Fragstats. Data Analysis Correlation was used to evaluate pairwise relationships among variables. When one (or both) variables were categorical, such as tillage type, which was coded numerically as zero for no-till and one for conventional tillage, spearman rank correlation coefficients were calculated. Pearson correlation coefficients were calculated when both variables were continuous. The magnitude of correlation coefficients was used to determine the spatial extent to account for the greatest amount of variation in parasitism. SAS PROC CORR (SAS Institute 2004) was used to calculate correlation coefficients. Many of the landscape variables were correlated so principal components analysis was used to derive a set of linearly independent regressors for use in regression modeling. Principal components were rotated using varimax rotation (Dillon and Goldstein 1984). The number of rotated standardized principal components retained for use as independent variables in regressions was determined by the scree method. The scree method involves plotting the eigenvalue associated with each principal component in successive order and determining the point beyond which the smaller eigenvalues form an approximately straight line. The components retained are those associated with eigenvalues that fall above the straight line formed by the smaller eigenvalues (Dillon and Goldstein 1984). The components that were retained were used as regressors in models relating parasitism to landscape context. Only linear (first order terms) were included in models. Standardized components were interpreted based on magnitudes of component loadings on the original variables. Stepwise multiple regression was used to construct models using the components as regressors. Component loadings were entered as independent variables in stepwise multiple regression models with the mean number of L. testaceipes per sentinel plant for each field as the dependent variable. F-tests were used to determine the significance of regression models with α for inclusion of a regressor in a model set at 0.15. The α = 0.15 level for inclusion was chosen so that moderately influential regressors were not overlooked during model selection. Significance of the overall regression model was maintained at α ≤ 0.05. Regression modeling was accomplished using PROC REG (SAS Institute 2004). Results General Patterns Seventy-one fields were studied in autumn during the three growing seasons, however data for one field was not used because all but one sentinel plant were destroyed by wildlife. For autumn, the proportion of the seven sentinel plants from each field that had one or more L. testaceipes per sentinel plant ranged from 0 to 1 among fields with a mean of 0.28, and the mean number of L. testaceipes per sentinel plant ranged 0 to 100.4 among fields with a mean of 11.6 (Table 1). During spring both the average number of L. testaceipes per sentinel plant and the proportion of sentinel plants with L. testaceipes were greater by 10-fold and threefold, respectively, than in autumn (Table 1). Aphid abundance in fields was also higher (greater than sevenfold) in spring than in autumn (Table 1) as was expected since sentinel plants were placed in wheat fields in autumn as soon as possible after emergence of wheat plants in the field from the soil. It was not possible to identify 24 fields each year with identical planting dates, so there was variation in wheat plant growth stage among fields. Variation in planting date probably accounted for a major portion of the variation in aphid abundance among fields. Table 1. Summary statistics (mean, SE, minimum, and maximum) for parasitism by L. testaceipes, aphid abundance, and other variables measured for n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Mean  SE  Minimum  Maximum  Autumn   Parasitism    Number of L. testaceipes / sentinel plant  11.6  2.62  0.0  100.2    Prop. sentinel plants with L. testaceipes  0.28  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  27.7  15.06  0.0  1040.0  Spring   Parasitism    Number of L. testaceipes / sentinel plant  179.9  22.64  0.0  576.7    Prop. sentinel plants with L. testaceipes  0.88  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  195.6  47.48  12.3  1680.3   Within-field & Field Edge    Tillage (0 = conventional, 1 = no till)  0.32  0.06  0.0  1.0    Crop rotation (0 = no, 1 = yes)  0.26  0.05  0.0  1.0    Grazed (0 = no, 1 = yes)  0.67  0.06  0.0  1.0    Grass species Richness  7.7  0.27  4.0  13.0    % Johnson grass coverage  34.1  2.66  0.0  75.0  Variable  Mean  SE  Minimum  Maximum  Autumn   Parasitism    Number of L. testaceipes / sentinel plant  11.6  2.62  0.0  100.2    Prop. sentinel plants with L. testaceipes  0.28  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  27.7  15.06  0.0  1040.0  Spring   Parasitism    Number of L. testaceipes / sentinel plant  179.9  22.64  0.0  576.7    Prop. sentinel plants with L. testaceipes  0.88  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  195.6  47.48  12.3  1680.3   Within-field & Field Edge    Tillage (0 = conventional, 1 = no till)  0.32  0.06  0.0  1.0    Crop rotation (0 = no, 1 = yes)  0.26  0.05  0.0  1.0    Grazed (0 = no, 1 = yes)  0.67  0.06  0.0  1.0    Grass species Richness  7.7  0.27  4.0  13.0    % Johnson grass coverage  34.1  2.66  0.0  75.0  View Large Table 1. Summary statistics (mean, SE, minimum, and maximum) for parasitism by L. testaceipes, aphid abundance, and other variables measured for n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Mean  SE  Minimum  Maximum  Autumn   Parasitism    Number of L. testaceipes / sentinel plant  11.6  2.62  0.0  100.2    Prop. sentinel plants with L. testaceipes  0.28  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  27.7  15.06  0.0  1040.0  Spring   Parasitism    Number of L. testaceipes / sentinel plant  179.9  22.64  0.0  576.7    Prop. sentinel plants with L. testaceipes  0.88  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  195.6  47.48  12.3  1680.3   Within-field & Field Edge    Tillage (0 = conventional, 1 = no till)  0.32  0.06  0.0  1.0    Crop rotation (0 = no, 1 = yes)  0.26  0.05  0.0  1.0    Grazed (0 = no, 1 = yes)  0.67  0.06  0.0  1.0    Grass species Richness  7.7  0.27  4.0  13.0    % Johnson grass coverage  34.1  2.66  0.0  75.0  Variable  Mean  SE  Minimum  Maximum  Autumn   Parasitism    Number of L. testaceipes / sentinel plant  11.6  2.62  0.0  100.2    Prop. sentinel plants with L. testaceipes  0.28  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  27.7  15.06  0.0  1040.0  Spring   Parasitism    Number of L. testaceipes / sentinel plant  179.9  22.64  0.0  576.7    Prop. sentinel plants with L. testaceipes  0.88  0.03  0.0  1.0    Aphids (no. per 20 D-vac placements)  195.6  47.48  12.3  1680.3   Within-field & Field Edge    Tillage (0 = conventional, 1 = no till)  0.32  0.06  0.0  1.0    Crop rotation (0 = no, 1 = yes)  0.26  0.05  0.0  1.0    Grazed (0 = no, 1 = yes)  0.67  0.06  0.0  1.0    Grass species Richness  7.7  0.27  4.0  13.0    % Johnson grass coverage  34.1  2.66  0.0  75.0  View Large Approximately 32% of fields studied were no till and 68% were conventional till (Table 1). Sixty-six percent of fields were grazed by cattle during winter months (a common practice in Oklahoma; Epplin et al. 1998). Grazing is associated with use of conventional tillage, because grazing cattle on no-till fields causes high levels of soil compaction. Therefore, the percent of fields in no-till systems was inversely related to grazing. The number of grass species in field edges ranged widely among fields, as did the percentage cover by Johnson grass in field edges. Landscape metrics varied substantially among the 70 fields for land areas in the three radii measured (Table 2). For example, the perimeter to area fractal dimension ranged from 1.20 to 1.45 for a radius of 0.8 km and from 1.31 to 1.45 for areas with a radius of 3.2 km. Percent of total land area planted to wheat ranged from 2.7 to 86.6 for the 0.8 km radius and from 10.1 to 68.5% for the 3.2 km radius. Although landscape metrics varied considerably for each landscape extent, the range of each metric was similar across the 0.8 to 3.2 km spatial extents. Table 2. Summary statistics (mean, SE, minimum, and maximum) for landscape metrics measured at three radii centered on each of n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Mean  SE  Minimum  Maximum  Radius 0.8 km   Patch density  26.3  1.56  7.7  57.3   Shannon’s patch diversity  1.31  0.03  0.61  1.88   Fractal dimension  1.33  0.01  1.20  1.45   Contagion  53.0  1.22  32.5  77.5   % Wheat  40.8  2.36  2.7  86.6   % Summer crops  24.7  2.63  0.11  82.6   % Winter crops (other than wheat)  0.2  0.07  0.0  4.1   % Fallow  1.9  0.24  0.0  10.1   % Grassland  23.0  1.69  1.7  52.4   % wooded  2.4  0.42  0.0  16.2   % manmade (built areas and roads)  6.5  0.53  2.0  24.3   % water  0.4  0.10  0.0  5.0  Radius 1.6 km   Patch density  24.9  1.28  7.8  48.3   Shannon’s patch diversity  1.46  0.03  0.87  1.87   Fractal dimension  1.37  0.004  1.30  1.46   Contagion  51.5  0.89  34.5  69.4   % Wheat  35.9  1.77  11.6  78.1   % Summer crops  24.5  2.08  0.99  68.4   % Winter crops (other than wheat)  0.3  0.05  0.0  2.9   % Fallow  2.6  0.20  0.3  8.4   % Grassland  27.0  1.89  6.5  67.9   % Wooded  2.8  0.33  0.0  9.9   % Manmade (built areas and roads)  6.2  0.31  2.7  13.4   % Water  0.7  0.12  0.0  5.9  Radius 3.2 km   Patch density  24.8  1.26  6.0  52.4   Shannon’s patch diversity  1.51  0.23  1.07  1.90   Fractal dimension  1.38  0.003  1.31  1.45   Contagion  50.6  0.78  35.4  66.2   % Wheat  34.7  1.51  10.1  68.5   % Summer crops  22.4  1.71  2.34  59.7   % Winter crops (other than wheat)  0.2  0.04  0.01  2.0   % Fallow  2.9  0.17  0.6  7.1   % Grassland  29.6  1.74  9.2  66.8   % Wooded  3.1  0.26  0.1  10.5   % Manmade (built areas and roads)  6.1  0.23  3.4  11.7   % Water  0.8  0.10  0.02  3.8  Variable  Mean  SE  Minimum  Maximum  Radius 0.8 km   Patch density  26.3  1.56  7.7  57.3   Shannon’s patch diversity  1.31  0.03  0.61  1.88   Fractal dimension  1.33  0.01  1.20  1.45   Contagion  53.0  1.22  32.5  77.5   % Wheat  40.8  2.36  2.7  86.6   % Summer crops  24.7  2.63  0.11  82.6   % Winter crops (other than wheat)  0.2  0.07  0.0  4.1   % Fallow  1.9  0.24  0.0  10.1   % Grassland  23.0  1.69  1.7  52.4   % wooded  2.4  0.42  0.0  16.2   % manmade (built areas and roads)  6.5  0.53  2.0  24.3   % water  0.4  0.10  0.0  5.0  Radius 1.6 km   Patch density  24.9  1.28  7.8  48.3   Shannon’s patch diversity  1.46  0.03  0.87  1.87   Fractal dimension  1.37  0.004  1.30  1.46   Contagion  51.5  0.89  34.5  69.4   % Wheat  35.9  1.77  11.6  78.1   % Summer crops  24.5  2.08  0.99  68.4   % Winter crops (other than wheat)  0.3  0.05  0.0  2.9   % Fallow  2.6  0.20  0.3  8.4   % Grassland  27.0  1.89  6.5  67.9   % Wooded  2.8  0.33  0.0  9.9   % Manmade (built areas and roads)  6.2  0.31  2.7  13.4   % Water  0.7  0.12  0.0  5.9  Radius 3.2 km   Patch density  24.8  1.26  6.0  52.4   Shannon’s patch diversity  1.51  0.23  1.07  1.90   Fractal dimension  1.38  0.003  1.31  1.45   Contagion  50.6  0.78  35.4  66.2   % Wheat  34.7  1.51  10.1  68.5   % Summer crops  22.4  1.71  2.34  59.7   % Winter crops (other than wheat)  0.2  0.04  0.01  2.0   % Fallow  2.9  0.17  0.6  7.1   % Grassland  29.6  1.74  9.2  66.8   % Wooded  3.1  0.26  0.1  10.5   % Manmade (built areas and roads)  6.1  0.23  3.4  11.7   % Water  0.8  0.10  0.02  3.8  View Large Table 2. Summary statistics (mean, SE, minimum, and maximum) for landscape metrics measured at three radii centered on each of n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Mean  SE  Minimum  Maximum  Radius 0.8 km   Patch density  26.3  1.56  7.7  57.3   Shannon’s patch diversity  1.31  0.03  0.61  1.88   Fractal dimension  1.33  0.01  1.20  1.45   Contagion  53.0  1.22  32.5  77.5   % Wheat  40.8  2.36  2.7  86.6   % Summer crops  24.7  2.63  0.11  82.6   % Winter crops (other than wheat)  0.2  0.07  0.0  4.1   % Fallow  1.9  0.24  0.0  10.1   % Grassland  23.0  1.69  1.7  52.4   % wooded  2.4  0.42  0.0  16.2   % manmade (built areas and roads)  6.5  0.53  2.0  24.3   % water  0.4  0.10  0.0  5.0  Radius 1.6 km   Patch density  24.9  1.28  7.8  48.3   Shannon’s patch diversity  1.46  0.03  0.87  1.87   Fractal dimension  1.37  0.004  1.30  1.46   Contagion  51.5  0.89  34.5  69.4   % Wheat  35.9  1.77  11.6  78.1   % Summer crops  24.5  2.08  0.99  68.4   % Winter crops (other than wheat)  0.3  0.05  0.0  2.9   % Fallow  2.6  0.20  0.3  8.4   % Grassland  27.0  1.89  6.5  67.9   % Wooded  2.8  0.33  0.0  9.9   % Manmade (built areas and roads)  6.2  0.31  2.7  13.4   % Water  0.7  0.12  0.0  5.9  Radius 3.2 km   Patch density  24.8  1.26  6.0  52.4   Shannon’s patch diversity  1.51  0.23  1.07  1.90   Fractal dimension  1.38  0.003  1.31  1.45   Contagion  50.6  0.78  35.4  66.2   % Wheat  34.7  1.51  10.1  68.5   % Summer crops  22.4  1.71  2.34  59.7   % Winter crops (other than wheat)  0.2  0.04  0.01  2.0   % Fallow  2.9  0.17  0.6  7.1   % Grassland  29.6  1.74  9.2  66.8   % Wooded  3.1  0.26  0.1  10.5   % Manmade (built areas and roads)  6.1  0.23  3.4  11.7   % Water  0.8  0.10  0.02  3.8  Variable  Mean  SE  Minimum  Maximum  Radius 0.8 km   Patch density  26.3  1.56  7.7  57.3   Shannon’s patch diversity  1.31  0.03  0.61  1.88   Fractal dimension  1.33  0.01  1.20  1.45   Contagion  53.0  1.22  32.5  77.5   % Wheat  40.8  2.36  2.7  86.6   % Summer crops  24.7  2.63  0.11  82.6   % Winter crops (other than wheat)  0.2  0.07  0.0  4.1   % Fallow  1.9  0.24  0.0  10.1   % Grassland  23.0  1.69  1.7  52.4   % wooded  2.4  0.42  0.0  16.2   % manmade (built areas and roads)  6.5  0.53  2.0  24.3   % water  0.4  0.10  0.0  5.0  Radius 1.6 km   Patch density  24.9  1.28  7.8  48.3   Shannon’s patch diversity  1.46  0.03  0.87  1.87   Fractal dimension  1.37  0.004  1.30  1.46   Contagion  51.5  0.89  34.5  69.4   % Wheat  35.9  1.77  11.6  78.1   % Summer crops  24.5  2.08  0.99  68.4   % Winter crops (other than wheat)  0.3  0.05  0.0  2.9   % Fallow  2.6  0.20  0.3  8.4   % Grassland  27.0  1.89  6.5  67.9   % Wooded  2.8  0.33  0.0  9.9   % Manmade (built areas and roads)  6.2  0.31  2.7  13.4   % Water  0.7  0.12  0.0  5.9  Radius 3.2 km   Patch density  24.8  1.26  6.0  52.4   Shannon’s patch diversity  1.51  0.23  1.07  1.90   Fractal dimension  1.38  0.003  1.31  1.45   Contagion  50.6  0.78  35.4  66.2   % Wheat  34.7  1.51  10.1  68.5   % Summer crops  22.4  1.71  2.34  59.7   % Winter crops (other than wheat)  0.2  0.04  0.01  2.0   % Fallow  2.9  0.17  0.6  7.1   % Grassland  29.6  1.74  9.2  66.8   % Wooded  3.1  0.26  0.1  10.5   % Manmade (built areas and roads)  6.1  0.23  3.4  11.7   % Water  0.8  0.10  0.02  3.8  View Large Many landscape metrics were correlated. Correlations among metrics for 3.2 km radius land areas (Table 3) were similar to correlations for the two smaller radii (not shown). Of particular note, the correlation between Shannon’s patch diversity index and contagion was −0.90, which indicates that evenness in the proportion of each patch type was strongly negatively related to the extent of aggregation of particular patch types in the landscape. The presence of correlation among the majority of landscape metrics indicates that their use as independent variables in regression modeling would result in multicollinearity. Table 3. Correlation among variables describing landscape context within 3.2 km radius areas centered on each of n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Shannon diversity  Fractal dimension  Contagion  % Wheat  % Summer crops  % Grass  % Wooded  Patch density  0.60*  0.26*  −0.32*  −0.25*  0.50*  −0.36*  0.28*  Shannon diversity    −0.38*  −0.90*  −0.45*  0.77*  −0.52*  0.45*  Fractal dimension      −0.50*  −0.09  0.10  −0.16  0.49*  Contagion        0.46*  −0.56*  0.33*  −0.59*  % Wheat          −0.36*  −0.35*  −0.48*  % Summer crops            −0.72*  0.10  % Grass              0.13  Variable  Shannon diversity  Fractal dimension  Contagion  % Wheat  % Summer crops  % Grass  % Wooded  Patch density  0.60*  0.26*  −0.32*  −0.25*  0.50*  −0.36*  0.28*  Shannon diversity    −0.38*  −0.90*  −0.45*  0.77*  −0.52*  0.45*  Fractal dimension      −0.50*  −0.09  0.10  −0.16  0.49*  Contagion        0.46*  −0.56*  0.33*  −0.59*  % Wheat          −0.36*  −0.35*  −0.48*  % Summer crops            −0.72*  0.10  % Grass              0.13  View Large Table 3. Correlation among variables describing landscape context within 3.2 km radius areas centered on each of n = 70 wheat fields in north central Oklahoma, for 2008, 2009, and 2010 Variable  Shannon diversity  Fractal dimension  Contagion  % Wheat  % Summer crops  % Grass  % Wooded  Patch density  0.60*  0.26*  −0.32*  −0.25*  0.50*  −0.36*  0.28*  Shannon diversity    −0.38*  −0.90*  −0.45*  0.77*  −0.52*  0.45*  Fractal dimension      −0.50*  −0.09  0.10  −0.16  0.49*  Contagion        0.46*  −0.56*  0.33*  −0.59*  % Wheat          −0.36*  −0.35*  −0.48*  % Summer crops            −0.72*  0.10  % Grass              0.13  Variable  Shannon diversity  Fractal dimension  Contagion  % Wheat  % Summer crops  % Grass  % Wooded  Patch density  0.60*  0.26*  −0.32*  −0.25*  0.50*  −0.36*  0.28*  Shannon diversity    −0.38*  −0.90*  −0.45*  0.77*  −0.52*  0.45*  Fractal dimension      −0.50*  −0.09  0.10  −0.16  0.49*  Contagion        0.46*  −0.56*  0.33*  −0.59*  % Wheat          −0.36*  −0.35*  −0.48*  % Summer crops            −0.72*  0.10  % Grass              0.13  View Large Aphid Abundance in Wheat Fields Cereal aphid abundance was estimated for each wheat field at the times sentinel plants were deployed. Aphid abundance in autumn was correlated with some landscape and within field variables. Autumn aphid abundance was correlated with whether the field was used for cattle grazing (r = 0.38; n = 70; P = 0.001), but not with any other of the within-field variables measured. Since cattle were not placed on fields until after our autumn study was complete, this effect may have been due to planting date, which is earlier for dual purpose wheat fields than for fields used only for grain production (Epplin et al. 1998). Aphid abundance in autumn was positively correlated with % grassland in the landscape (r = 0.28; n = 70; P = 0.02), but not with any other landscape metric. For spring, aphid abundance was negatively correlated with both tillage type (conventional vs. no-till) (r = −0.47; n = 46; P = 0.001) and crop rotation (r = −0.58; n = 46; P < 0.001). Since no-till is commonly practiced in conjunction with crop rotation in central Oklahoma, the two variables were highly correlated (r = 0.84; n = 70; P < 0.001). Aphid abundance in wheat fields in spring was positively correlated with fractal dimension (r = 0.29; n = 46; P = 0.05) and negatively correlated with % summer crops (r = −0.30; n = 46; P = 0.04); the correlation with % grassland for spring was positive (r = 0.24) but not significant. Parasitism of Bird Cherry–Oat Aphids by L. testaceipes Both the average number of L. testaceipes per sentinel plant and the proportion of sentinel plants with L. testaceipes were correlated with landscape variables for autumn and spring. Correlations for the number of L. testaceipes per sentinel plant were generally greater in magnitude and more often significant than correlations for the proportion of sentinel plants with L. testaceipes. This was not surprising since only seven sentinel plants were deployed per field which limits interpretation of this measure. Therefore, we limit further analysis of the relationship of parasitism with landscape variables to the number of L. testaceipes per sentinel plant. Neither aphid abundance nor any within field or field edge variable was correlated with the number of L. testaceipes per sentinel plant in autumn (not shown). In spring, there was a significant correlation between aphid abundance in wheat fields and number of L. testaceipes per sentinel plant (r = 0.35; n = 46; P = 0.02). None of the other within field or field edge variables was significantly correlated with the number of L. testaceipes per sentinel plant in spring. For landscape metrics, significant correlations occurred for the number of L. testaceipes per sentinel plant with patch density, fractal dimension, Shannon’s patch diversity, percent wheat, percent summer crops, and percent wooded land (Table 4). Correlations with percent wheat were negative for all spatial extents in autumn, but were not significant in spring. Correlations with patch density were positive in autumn but negative in spring (Fig. 2). Correlations with percent summer crops and fractal dimension were positive for all spatial extents and often significant in both autumn and spring (Fig. 2). Correlations for contagion were negative for all spatial extents in both seasons, and were significant in spring. Correlations with Shannon’s patch diversity index were positive for all spatial extents in spring, but significant only at 1.6 km. The correlation with percent wooded land was positive and significant at 3.2 km, but not at finer spatial extents. Table 4. Correlation of landscape metrics with the number of L. testaceipes per sentinel plant and the percent of sentinel plants with L. testaceipes present for autumn (n = 70) and spring (n = 46) Metric  Autumn  Spring  No. L. testaceipes  % with L. testaceipes  No. L. testaceipes  % with L. testaceipes  Radius 0.8 km   Patch density  0.26*  0.23  −0.43*  −0.05   Shannon’s patch diversity  0.13  0.07  0.01  0.22   Contagion  −0.10  −0.09  −0.29*  −0.26   Fractal dimension  0.10  0.06  0.11  0.11   % Wheat  −0.34*  −0.25*  0.01  −0.15   % Summer crops  0.39*  0.18  0.17  0.12   % Grassland  −0.13  0.01  0.24  0.04   % Wooded  −0.04  −0.04  0.09  0.15  Radius 1.6 km   Patch density  0.32*  0.30*  −0.50*  −0.12   Shannon’s patch diversity  0.21  0.17  0.33*  0.10   Contagion  −0.12  −0.15  −0.29*  −0.43*   Fractal dimension  0.18  0.18  0.39*  0.25   % Wheat  −0.31*  −0.35*  0.01  −0.17   % Summer crops  0.29*  0.16  0.27  0.10   % Grassland  −0.08  0.08  0.20  0.00   % Wooded  0.07  0.08  0.22  0.22  Radius 3.2 km   Patch density  0.32*  0.29*  −0.61*  −0.21   Shannon’s patch diversity  0.20  0.15  0.26  0.16   Contagion  −0.14  −0.18  −0.30*  −0.28   Fractal dimension  0.26*  0.31*  0.29*  0.20   % Wheat  −0.34*  −0.40*  −0.01  −0.20   % Summer crops  0.31*  0.13  0.27  0.10   % Grassland  −0.07  0.12  0.25  0.05   % Wooded  0.24*  0.19  0.07  0.18  Metric  Autumn  Spring  No. L. testaceipes  % with L. testaceipes  No. L. testaceipes  % with L. testaceipes  Radius 0.8 km   Patch density  0.26*  0.23  −0.43*  −0.05   Shannon’s patch diversity  0.13  0.07  0.01  0.22   Contagion  −0.10  −0.09  −0.29*  −0.26   Fractal dimension  0.10  0.06  0.11  0.11   % Wheat  −0.34*  −0.25*  0.01  −0.15   % Summer crops  0.39*  0.18  0.17  0.12   % Grassland  −0.13  0.01  0.24  0.04   % Wooded  −0.04  −0.04  0.09  0.15  Radius 1.6 km   Patch density  0.32*  0.30*  −0.50*  −0.12   Shannon’s patch diversity  0.21  0.17  0.33*  0.10   Contagion  −0.12  −0.15  −0.29*  −0.43*   Fractal dimension  0.18  0.18  0.39*  0.25   % Wheat  −0.31*  −0.35*  0.01  −0.17   % Summer crops  0.29*  0.16  0.27  0.10   % Grassland  −0.08  0.08  0.20  0.00   % Wooded  0.07  0.08  0.22  0.22  Radius 3.2 km   Patch density  0.32*  0.29*  −0.61*  −0.21   Shannon’s patch diversity  0.20  0.15  0.26  0.16   Contagion  −0.14  −0.18  −0.30*  −0.28   Fractal dimension  0.26*  0.31*  0.29*  0.20   % Wheat  −0.34*  −0.40*  −0.01  −0.20   % Summer crops  0.31*  0.13  0.27  0.10   % Grassland  −0.07  0.12  0.25  0.05   % Wooded  0.24*  0.19  0.07  0.18  *Statistically significant correlations (P < 0.05). View Large Table 4. Correlation of landscape metrics with the number of L. testaceipes per sentinel plant and the percent of sentinel plants with L. testaceipes present for autumn (n = 70) and spring (n = 46) Metric  Autumn  Spring  No. L. testaceipes  % with L. testaceipes  No. L. testaceipes  % with L. testaceipes  Radius 0.8 km   Patch density  0.26*  0.23  −0.43*  −0.05   Shannon’s patch diversity  0.13  0.07  0.01  0.22   Contagion  −0.10  −0.09  −0.29*  −0.26   Fractal dimension  0.10  0.06  0.11  0.11   % Wheat  −0.34*  −0.25*  0.01  −0.15   % Summer crops  0.39*  0.18  0.17  0.12   % Grassland  −0.13  0.01  0.24  0.04   % Wooded  −0.04  −0.04  0.09  0.15  Radius 1.6 km   Patch density  0.32*  0.30*  −0.50*  −0.12   Shannon’s patch diversity  0.21  0.17  0.33*  0.10   Contagion  −0.12  −0.15  −0.29*  −0.43*   Fractal dimension  0.18  0.18  0.39*  0.25   % Wheat  −0.31*  −0.35*  0.01  −0.17   % Summer crops  0.29*  0.16  0.27  0.10   % Grassland  −0.08  0.08  0.20  0.00   % Wooded  0.07  0.08  0.22  0.22  Radius 3.2 km   Patch density  0.32*  0.29*  −0.61*  −0.21   Shannon’s patch diversity  0.20  0.15  0.26  0.16   Contagion  −0.14  −0.18  −0.30*  −0.28   Fractal dimension  0.26*  0.31*  0.29*  0.20   % Wheat  −0.34*  −0.40*  −0.01  −0.20   % Summer crops  0.31*  0.13  0.27  0.10   % Grassland  −0.07  0.12  0.25  0.05   % Wooded  0.24*  0.19  0.07  0.18  Metric  Autumn  Spring  No. L. testaceipes  % with L. testaceipes  No. L. testaceipes  % with L. testaceipes  Radius 0.8 km   Patch density  0.26*  0.23  −0.43*  −0.05   Shannon’s patch diversity  0.13  0.07  0.01  0.22   Contagion  −0.10  −0.09  −0.29*  −0.26   Fractal dimension  0.10  0.06  0.11  0.11   % Wheat  −0.34*  −0.25*  0.01  −0.15   % Summer crops  0.39*  0.18  0.17  0.12   % Grassland  −0.13  0.01  0.24  0.04   % Wooded  −0.04  −0.04  0.09  0.15  Radius 1.6 km   Patch density  0.32*  0.30*  −0.50*  −0.12   Shannon’s patch diversity  0.21  0.17  0.33*  0.10   Contagion  −0.12  −0.15  −0.29*  −0.43*   Fractal dimension  0.18  0.18  0.39*  0.25   % Wheat  −0.31*  −0.35*  0.01  −0.17   % Summer crops  0.29*  0.16  0.27  0.10   % Grassland  −0.08  0.08  0.20  0.00   % Wooded  0.07  0.08  0.22  0.22  Radius 3.2 km   Patch density  0.32*  0.29*  −0.61*  −0.21   Shannon’s patch diversity  0.20  0.15  0.26  0.16   Contagion  −0.14  −0.18  −0.30*  −0.28   Fractal dimension  0.26*  0.31*  0.29*  0.20   % Wheat  −0.34*  −0.40*  −0.01  −0.20   % Summer crops  0.31*  0.13  0.27  0.10   % Grassland  −0.07  0.12  0.25  0.05   % Wooded  0.24*  0.19  0.07  0.18  *Statistically significant correlations (P < 0.05). View Large Fig. 2. View largeDownload slide Number of L. testaceipes per seven sentinel plants versus patch density and fractal dimension for autumn (top) and spring (bottom). Fig. 2. View largeDownload slide Number of L. testaceipes per seven sentinel plants versus patch density and fractal dimension for autumn (top) and spring (bottom). Correlations between the number of L. testaceipes per sentinel plant and landscape variables for autumn were generally greatest at the 3.2 km radius and correlations in spring were similar at 1.6 and 3.2 km (Table 4), therefore, we used landscape data measured at the 3.2 km scale in regression modeling for autumn and spring. Aphid abundance in autumn, and other variables measured in wheat fields and field edges, were uncorrelated with the number of L. testaceipes per sentinel plant and therefore were not used as predictors in regression models. Aphid abundance in wheat fields in spring was correlated with parasitism and was included as a predictor in stepwise regressions. Based on the scree method, we included five of eight factors (Supplementary Appendix 1) as predictors of the number of L. testaceipes per sentinel plant in stepwise regression models for autumn and spring. The best fitting first-order regression model for the number of L. testaceipes in autumn (F = 6.72; df = 4, 65; P = 0.0001) was  Lt= 11.56 + 4.90·F1– 7.76·F3+ 4.93·F4 +5.66·F5(R2= 0.30) (1) where Lt is the number of L. testaceipes per sentinel plant, and F1, F3, F4, and F5 are factors derived from principal components analysis (Supplementary Appendix 1). For spring, the best fitting regression model (F = 13.34; df = 42, 3; P = 0.0001) was  Lt= 179.95 – 27.43·F1+ 40.39·F4– 95.52·F5(R2= 0.49) (2) where Lt is the number of L. testaceipes per sentinel plant and F1, F4, and F5 are factors. Regression models for L. testaceipes parasitism in wheat fields in autumn and spring were interpreted based on standardized factor loadings on the original landscape variables. For the autumn regression model (Eq. 1) four factors, F1, F3, F4, and F5 were entered into the model using the selection criterion of P < 0.15 for inclusion of a variable. The regression coefficient for F1 was positive, and the largest loadings on F1 were for % summer crops (0.75) and % grassland (−0.71), indicating that the presence of high acreage of summer crops increased parasitism in wheat in autumn, whereas the presence of large amounts of grassland was detrimental to parasitism. Factor F3 had a negative regression coefficient and was dominated by a large loading on % wheat (1.23), indicating a negative effect of wheat mono-cropping on parasitism in autumn. The positive regression coefficient for F4 combined with the large factor loading on patch density (1.18) indicates a positive influence of small patch (field) size in the landscape on parasitism. Finally, the positive regression coefficient for F5 combined with the dominant positive loading on fractal dimension (1.19) indicates a positive effect of curvilinear patch boundaries, characteristic of semi-natural and natural lands, on parasitism. In spring, three factors were entered into the regression, F1, F4, and F5 (Eq. 2). In contrast to autumn, F1 had a negative regression coefficient indicating a reversal of the relationship between percent of landscape planted to summer crops and the percent grassland from that in autumn. F4 had a positive regression coefficient indicating a positive response of parasitism to increasing fractal dimension. The coefficient of F5 was negative indicating a negative response of parasitism to increasing patch density, which is opposite of the response observed in autumn. Discussion Variables representing landscape composition as well as spatial configuration were correlated with parasitism levels by L. testaceipes in wheat fields. Correlations were observed both in autumn when colonization of wheat fields by cereal aphids and L. testaceipes first occurs, and in spring when within field processes might be expected to predominate over colonization in determining L. testaceipes population density and the resulting cereal aphid parasitism. It is notable that L. testaceipes responds to landscape composition (e.g., % wheat and Shannon’s patch diversity index) as has been observed for braconid parasitoids of aphids in several studies (e.g., Roschewitz et al. 2005, Thies et al. 2005, Plecas et al. 2014, Zhao et al. 2015), but also to landscape configuration (e.g., fractal dimension and contagion), which has been less frequently measured (but see Plecas et al. 2014), and may be important in determining arthropod population processes in agroecosystems. Large acreage of wheat surrounding a focal wheat field was negatively correlated with cereal aphid parasitism in autumn, but was uncorrelated with parasitism in spring (Table 4). Acreage of summer crops was positively correlated with parasitism in autumn, but negatively correlated in spring (Table 4). The autumn and spring regression models (Eqs. 1 and 2) reflect these differing associations of parasitism with landscape variables for autumn and spring. Additionally, patch density had a positive effect on parasitism in autumn, but a negative effect in spring. Conversely, fractal dimension had a positive effect in both seasons (Eqs. 1 and 2). In the agricultural landscapes of north central Oklahoma, both cropland and semi-natural land habitats contribute to the landscape scale population dynamics of L. testaceipes. The autumn and spring regression models reflect the differing associations of parasitism with landscape variables for autumn and spring. Differences in effects of landscape context among seasons suggest that the response of L. testaceipes to landscape context is very rapid, within a single growing season. L. testaceipes dynamics within a wheat field, and correspondingly, capacity for cereal aphid suppression, is partially dependent on population dynamics of the parasitoid in the landscape surrounding a wheat field during the course of the growing season. In the heavily agricultural landscapes of north central Oklahoma, cropland and semi-natural lands both contribute to determining the landscape scale population dynamics of L. testaceipes. Considered in this context, the change in regression coefficient (positive to negative) of parasitism to presence of summer crops in autumn and spring is likely due the fact that summer crops serve as habitat for L. testaceipes in early autumn when wheat is first planted, but are fallow fields in spring. The negative effect of amount of grassland on parasitism may be a consequence of the extremely arid and high temperature summer conditions in Oklahoma, which result in mostly dormant grasslands that harbor extremely low numbers of aphid hosts and therefore of parasitoids (Anstead 2000). Variation in the effect of wheat acreage on parasitism can be viewed similarly, where wheat fields are essentially devoid of cereal aphid hosts in early autumn but serve as habitat for cereal aphids and L. testaceipes in spring. The varying effect of increasing patch density on parasitism from positive in fall to negative in spring is difficult to interpret. If small patch size (high patch density) increased aphid parasitism by L. testaceipes in wheat fields it’s effect would most likely result from increasing access to habitats containing hosts or other resources. The magnitude of this effect might vary seasonally but would not be expected to shift from positive to negative. The influence of patch density on parasitism is probably indirect and related to other landscape characteristics, such as those captured by Shannon’s diversity index, that influence resource availability to L. testaceipes, are correlated with patch density (Table 3), and are not explicity accounted for in regression models. This result highlights the difficulty in teasing out the effects of numerous interacting factors on L. testaceipes ecology using multivariate methods. Even with this limitation the study has highlighted several landscape characteristics that, through their influence on resource availability influence the success of L. testaceipes as a biological control agent in wheat agreoecosystems in Oklahoma. In Oklahoma, the correlation of aphid parasitism by L. testaceipes with landscape variables in wheat fields did not decrease with increasing spatial extent, contrary to observations for braconid parasitoids of cereal aphids in European wheat agroecosystems (Thies et al. 2005). L. testaceipes was not recorded in their study even though it is widely established in European fauna (Stary 1988, Zikic et al. 2015). Interestingly, our finest spatial extent 0.8 km radius (1.6 km diameter) was closest to the maximum spatial extent (2.0 km diameter) at which landscape complexity significantly explained variation in parasitism in wheat fields in Europe (Thies et al. 2005). Our results suggest that L. testaceipes responds to landscape variation over a broad spatial extent. L. testaceipes shows a very limited response to plant volatiles (Lo Pinto et al. 2004, Fauvergue et al. 2006), and is considered to be both a habitat and host generalist with potential for very rapid population growth (Mackauer and Stary 1967, Jones et al. 2003). These life history characteristics might be expected to be associated with high levels of dispersal. Although life histories of most braconid aphid parasitoid species are not well known, L. testaceipes is probably exceptional in relation to breadth of host and habitat range (Stary et al. 1988, Pike et al. 2000). Because of its life history characteristics L. testaceipes appears to be well adapted to the agroecosystems of the Southern Plains, where agricultural landscapes are coarse grained and habitats are poorly interspersed. L. testaceipes is a key natural enemy of cereal aphids in wheat (Jones et al. 2007, Giles et al. 2017), and effective biological control frequently occurs within wheat fields prior to the aphid infestation reaching the economic threshold (Giles et al. 2003). L. testaceipes does not appear to disperse among a set of essential habitats seasonally and therefore does not exhibit the characteristics of a cyclic colonizer (Wissinger 1997). Rather L. testaceipes appears to utilize multiple habitats throughout the year depending on their availability and acceptability, and frequently disperses among habitats (Jessie 2017). In this sense L. testaceipes is better described as an r-selected species, with high reproductive and dispersal rates (Ehler and Miller 1978, Price and Waldbauer 1994). The ability of L. testaceipes to control aphids in wheat fields depends on rapid colonization of fields (Bortoloto et al. 2015) and high attack and reproductive rates (Jones 2005). Colonization of wheat fields in autumn is enhanced by proximity to fields of summer crops and semi-natural habitats other than grasslands. Based on this study, the optimal landscape to promote biological control of cereal aphids in wheat by L. testaceipes is one where summer crops, wheat, and semi-natural habitats exist in significant amounts, and are well interspersed. Future emphasis should be given to determining the specific resources within these habitats that are utilized by L. testaceipes. Supplementary Data Supplementary data are available at Environmental Entomology online. Acknowledgments We thank Tim Johnson for technical assistance with the project and for coordinating data collection and processing activities. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture (USDA). USDA is an equal opportunity provider and employer. References Cited Anstead, J. A. 2000. Genetic and biotypic diversity of greenbug Schizaphis graminum (Rondani) populations on non-cultivated hosts. Doctoral Dissertation . Oklahoma State University, Stillwater, OK. Bortolotto, O. C., A. D. O. M. Júnior, and Hoshino, A. T. 2015. Aphidophagous parasitoids can forage wheat crops before aphid infestation, Paraná State, Brazil. J. Insect Sci . 15. doi: 10:1093/jisesa/iev027 Brewer, M. J., T. Noma, N. C. Elliott, A. N. Kravchenko, and A. L. Hild. 2008. A landscape view of cereal aphid parasitoid dynamics reveals sensitivity to farm- and region-scale vegetation structure. European J. Entomol . 105: 503– 511. Google Scholar CrossRef Search ADS   Dillon, W. R., and Goldstein, M. 1984. Multivariate analysis: methods and applications . John Wiley and Sons, New York, NY. Eikenbary, R. D., and C. E. Rogers. 1974. Importance of alternate hosts in establishment of introduced parasites, pp 119– 133. In Tall Timbers Conference on Ecological Animal Control by Habitat Management; Tallahassee, FL. Ehler, L. E., and J. C. Miller. 1978. Biological control in temporary agroecosystems. Entomophaga . 3: 207– 212. Google Scholar CrossRef Search ADS   Elliott, N. C., Kieckhefer, R. W., Lee, J. H., and French, B. W. 1998. Influence of within-field and landscape factors on aphid predator populations in wheat. Landscape Ecol . 14: 239– 252. Google Scholar CrossRef Search ADS   Epplin, F. M., R. R. True, and E. G. Krenzer, Jr. 1998. Practices used by Oklahoma wheat growers by region. Okla. Curr. Farm Econ . 71: 14– 24. Fauvergue, X., R. Boll, J. Rochat, E. Wajnberg, C. Bernstein, and L. Lapchin. 2006. Habitat assessment by parasitoids: consequences for population distribution. Behav. Ecol . 17: 522– 531. Google Scholar CrossRef Search ADS   Giles, K. L., D. B. Jones, T. A. Royer, N. C. Elliott, and S. D. Kindler. 2003. Development of a sampling plan in winter wheat that estimates cereal aphid parasitism levels and predicts population suppression. J. Econ. Entomol . 96: 975– 982. Google Scholar CrossRef Search ADS PubMed  Giles, K. L., B. P. McCornack, T. A. Royer, and N. C. Elliott. 2017. Incorporating biological control into IPM decision making. Curr. Opin. Insect Sci . 20: 84– 89. Google Scholar CrossRef Search ADS PubMed  Hand, S. C. 1986. The capture efficiency of the Dietrick vacuum insect net for aphids on grasses and cereals. Ann. Appl. Biol . 108: 1744– 7348. Google Scholar CrossRef Search ADS   Hight, S. C., R. D. Eikenbary, R. J. Miller, and K. J. Starks. 1972. The greenbug and Lysiphlebus testaceipes. Environ. Entomol . 1: 205– 209. Google Scholar CrossRef Search ADS   Jessie, C. N. 2017. Diversity, relative abundance, movement, and fitness of insect predators in winter canola agricultural landscapes in the United States Southern Plains. Doctoral Dissertation . Oklahoma State University, Stillwater, OK. Jones, D. B. 2005. Winter ecology of the parasitoid Lysiphlebus testaceipes Cresson. Doctoral Dissertation . Oklahoma State University, Stillwater, OK. Jones, D. B., K. L. Giles, R. C. Berberet, T. A. Royer, N. C. Elliott, and M. Payton. 2003. Functional responses of an introduced parasitoid and an indigenous parasitoid on greenbug at four temperatures. Environ. Entomol . 32: 425– 432. Google Scholar CrossRef Search ADS   Jones, D. B., K. L. Giles, N. C. Elliott, and M. E. Payton. 2007. Parasitism of Greenbug, Schizaphis graminum, by the parasitoid Lysiphlebus testaceipes at winter temperatures. Environ. Entomol . 36: 1– 8. Google Scholar CrossRef Search ADS PubMed  Lo Pinto, M., E. Wajnberg, S. Colazza, C. Curty, and X. Fauvergue. 2004. Olfactory response of two aphid parasitoids, Lysiphlebus testaceipes and Aphidius colemani, to aphid-infested plants from a distance. Entomol. Exp. Applic . 110: 159– 164. Google Scholar CrossRef Search ADS   Losey, J. E., and R. F. Denno. 1998. Positive predator-predator interactions: enhanced predation rates and synergistic suppression of aphid populations. Ecology . 79: 2143– 2152. Mackauer, M., and P. Stary. 1967. Index of entomophagous insects (hym., ichneumonidae): index of world aphidiidae . Le Francois, Paris, France. McGarigal, K. 2014. FRAGSTATS: spatial pattern analysis program for categorical maps. http://www.umass.edu/landeco/research/fragstats/fragstats.html McGarigal, K., and B. J. Marks. 1995. FRAGSTATS. Spatial analysis program for quantifying landscape structure. USDA Forest Service General Technical Report PNWGTR-351 . USDA Forest Service, Pacific Northwest Research Station, Portland, OR. O’Neill, R. V., J. R. Krummel, R. H. Gardner, G. Sugihara, B. Jackson, D. L. DeAngelis, B. T. Milne, M. G. Turner, B. Zygmunt, S. W. Christensen, et al.   1988. Indices of landscape pattern. Landscape Ecol . 1: 153– 162. Google Scholar CrossRef Search ADS   Pike, K. S., P. Starý, T. Miller, G. Graf, D. Allison, L. Boydston, and R. Miller. 2000. Aphid parasitoids (Hymenoptera: Braconidae: Aphidiinae) of Northwest USA. Proc. Entomol. Soc. Wash . 102: 688– 740. Plećaš, M., V. Gagić, M. Janković, O. Petrović-Obradović, N. G. Kavallieratos, Ž. Tomanović, and A. Ćetković. 2014. Landscape composition and configuration influence cereal aphid–parasitoid–hyperparasitoid interactions and biological control differentially across years. Agric. Ecosyst. Environ . 183: 1– 10. Google Scholar CrossRef Search ADS   Price, P. W., and G. P. Waldbauer. 1994. Ecological aspects of pest management, pp. 33– 68. In R. L. Metcalf and W. H. Luckman (eds.), Introduction to insect pest management . Wiley, New York, NY. Riitters, K. H., R. V. Oneill, C. T. Hunsaker, J. D. Wickham, D. H. Yankee, S. P. Timmins, K. B. Jones, and B. L. Jackson. 1995. A factor analysis of landscape pattern and structure metrics. Landscape Ecol . 10: 23– 39. Google Scholar CrossRef Search ADS   Roschewitz, I., M. Hücker, T. Tscharntke, and C. Thies. 2005. The influence of landscape context and farming practices on parasitism of cereal aphids. Agric., Ecosyst. Environ . 108: 218– 227. Google Scholar CrossRef Search ADS   Royer, T. A., K. L. Giles, and N. C. Elliott. 2004. The cereal aphid expert system and Glance’ n Go Sampling, questions and answers. CR-7191 . Oklahoma Cooperative Extension Service, Stillwater, OK. Royer, T. A., B. B. Pendleton, N. C. Elliott, and K. L. Giles. 2015. Greenbug (Hemiptera: Aphididae) biology, ecology, and management in wheat and sorghum. J. Integ. Pest Mngmt . 6: 19. Google Scholar CrossRef Search ADS   Rusch, A., R. Chaplin-Kramer, M. M. Gardiner, V. Hawro, J. Holland, D. Landis, C. Thies, T. Tscharntke, W. W. Weisser, C. Winqvist, et al.   2016. Agricultural landscape simplification reduces natural pest control: a quantitative synthesis. Agric. Ecosyst. Environ . 221: 198– 204. Google Scholar CrossRef Search ADS   SAS Institute. 2004. SAS/STAT user’s guide, version 9.1 . SAS Institute, Cary, NC. Schmidt, M. H., C. Thies, and T. Tscharntke. 2004. Landscape context of arthropod biological control, pp. 55– 63. In G. M. Gurr, S. D. Wratten, and M. A. Altieri (eds.), Ecological engineering for pest management: advances in habitat manipulation for arthropods . CSIRO Publishing, Collingswood, VIC, Australia. Google Scholar CrossRef Search ADS   Shannon, C., and W. Weaver. 1949. The mathematical theory of communication . Univ. Illinois Press, Urbana, IL. Starý, P., J. P. Lyon, and F. Leclant. 1988. Biocontrol of aphids by the introduced Lysiphlebus testaceipes (Cress.) (Hym., Aphidiidae) in Mediterranean France. J. Appl. Entomol . 105: 74– 78. Google Scholar CrossRef Search ADS   Thies, C., I. Roschewitz, and T. Tscharntke. 2005. The landscape context of cereal aphid-parasitoid interactions. Proc. Biol. Sci . 272: 203– 210. Google Scholar CrossRef Search ADS PubMed  Turner, M. G., R. H. Gardner, and R. V. O’Neill. 2001. Landscape ecology in theory and practice . Springer-Verlag, New York, NY. USDA, NASS (National Agricultural Statistics Survey). 2017. http://www.nass.usda.gov/. Webster, F. M., and W. J. Phillips. 1912. The spring grain aphis or “greenbug”, U.S. Bur. Entomol. Bull . 110: 1– 153. https://archive.org/details/springgrainaphis110webs Wissinger, S. A. 1997. Cyclic colonization in predictably ephemeral habitats: a template for biological control in annual crop systems. Biol. Control . 10: 4– 15. Google Scholar CrossRef Search ADS   Zadoks, J. C., T. T. Chang, and Konzak, C. F. 1974. A decimal code for the growth stages of cereals. Weed Res . 14: 415– 421. Google Scholar CrossRef Search ADS   Zhao, Z. H., H. S. Sandhu, F. Gao, and D. H. He. 2015. Shifts in natural enemy assemblages resulting from landscape simplification account for biocontrol loss in wheat fields. Ecol. Res . 30: 493– 498. Google Scholar CrossRef Search ADS   Žikić, V., S., Stanković, M. Milošević, O. Petrović-Obradović, A. Petrović, P. Starý, and Ž. Tomanović. 2015. First detection of Lysiphlebus testaceipes (Cresson) (Hymenoptera: Aphidiinae) in Serbia: an introduced species invading Europe? N.W. J. Zool . 11: 97– 101. Published by Oxford University Press on behalf of Entomological Society of America 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US. This work is written by (a) US Government employee(s) and is in the public domain in the US.

Journal

Environmental EntomologyOxford University Press

Published: Apr 12, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off