Habitat-based drivers of abundance and richness in an intensively farmed agricultural landscape

by

Aleksandra Dolezal

A Thesis presented to The University of Guelph

In partial fulfilment of requirements for the degree of Master of Science in Integrative Biology

Guelph, Ontario, Canada

© Aleksandra Dolezal, July 2019

ABSTRACT

HABITAT-BASED DRIVERS OF ARTHROPOD ABUNDANCE AND RICHNESS IN AN

INTENSIVELY FARMED AGRICULTURAL LANDSCAPE

Aleksandra Dolezal Advisor:

University of Guelph, 2019 Dr. A. S. MacDougall

Habitat is critical for sustaining arthropod populations, but its influence in agricultural landscapes where are declining is unclear. In a heavily farmed area of Ontario, I examined habitat effects on arthropods at two scales: landscape (habitat isolation) and local

(farm cover type -crop, prairie, forest). Additional analyses tested differences in arthropod communities among cover types, how prairie addition affects leaf defoliation, whether crop type influences arthropods (soy, corn, orchard, mixed organic), and whether arthropod richness differs between farm and non-farm sites. I found arthropod richness and abundance to be determined locally, via the presence of restored high-quality prairie habitat. Prairie addition did not affect leaf defoliation but influenced the spatial distribution of in crops. Crop type influenced arthropod groups, especially herbivores, which preferred mixed organic cover. Lastly, farm and non-farm sites revealed similar family-level richness, suggesting that conventional farms can support arthropod if habitat is not limiting. iii

ACKNOWLEDGEMENTS

This thesis would not have been possible without the researchers who came together to initiate an extraordinary project that was the basis for my work. It’s their guiding idea and the efforts of all the participating scientists and non-scientists which make this project an on-going success. Being a part of this collaborative effort has been a great privilege.

I want to express my gratitude to my supervisor Dr. Andrew MacDougall, who gave me the opportunity to do my MSc under his guidance. I am especially thankful that he encouraged me in skill-building activities such as workshops and conferences which added great experiences to my time as a graduate student. I am also grateful to my committee members Kevin McCann and Alex Smith for giving me input throughout my research journey. I would also like to acknowledge Steve Marshall whose passion for entomological research has motivated me to start my own journey as an entomologist. I look forward to following his work.

I wholeheartedly thank the MacDougall lab team for creating a positive and supportive working atmosphere which cannot be taken for granted. A big thanks to Dr. Ellen Esch for all her consulting days in the office for stats help. Thank you to all my field technicians who helped during the summer of 2017. Thank you Brianna Maher for being so hardworking, Sam Kloke for teaching me plant in the field, Brock Roth for your dedication and work ethic, and

Bernal Arce for your expertise in all things arthropods. I learned from all of you as I hope you took a piece of my work with you. Thank you Margeret Visser for putting in the time to edit my work. I cannot forget Kevin, the love of my life, who held me together when I fell apart. iv

I also want to thank the awesome participants in the ALUS program who gave us the permission to do research on their farms. I hope my study contributes to the conservation of these lands. Bryan Gilvesy was a huge contributor to the passion and work these participants had, as well as Casey Whitelock and Alyssa Cousineau. Thank you to Sage Handler of the Nigel

Raine lab for identifying my bee species to genus level, you don’t know how many hours this saved me, Sage.

I am thankful for the funding provided by the OMAFRA Highly Qualified Personnel

Scholarship that made this research possible. Tall Grass Ontario, and Food from Thought also provided partial support for this work through travel grants. Lastly, I would like to thank my family for their love and support for all my various endeavors over the years, and not freaking out when bags of arthropods appeared in the freezer. Planting wildflowers, caring for pets, running barefoot outside…. the childhood you gave me has made me who I am today.

“Nature holds a key to our aesthetic, intellectual, cognitive and even spiritual satisfaction”

-E.O. Wilson v

TABLE OF CONTENTS

Abstract…………………………………………………………………………………………....ii Acknowledgements ...... iii Table of Contents ...... v List of Tables ...... vi List of Supplementary Tables ...... vii List of Figures ...... viii List of Supplementary Figures ...... x List of Appendices ...... xii 1. Introduction ...... 1 2. Methods...... 4 2.1 Site selection and description ...... 4 2.2 Arthropod and habitat sampling ...... 9 2.3 Main analysis ...... 12 2.3.1 Landscape and local factors of habitat on 13 agricultural sites ...... 12 2.3.2 Within-prairie habitat factors ...... 14 2.4 Complementary analyses ...... 16 2.5 Statistical analysis ...... 17 2.5.1 Landscape and local factors of habitat on 13 agricultural sites ...... 18 2.5.2 Within-prairie habitat factors ...... 19 2.5.3 CA1: Arthropod community composition among cover types ...... 20 2.5.4 CA2: Assessment of arthropod-derived ecosystem service mediated by habitat ...... 20 2.5.5 CA3: Crop type differences for arthropod abundance and richness (family-level) ..... 21 2.5.6 CA4: Regional family pool comparison ...... 21 3. Results ...... 22 3.1 Landscape and local factors of habitat on 13 agricultural sites ...... 23 3.2 Within-prairie habitat factors ...... 25 3.3 CA1: Arthropod community composition among cover types ...... 27 3.4 CA2: Assessment of arthropod-derived ecosystem service mediated by habitat ...... 27 3.5 CA3: Crop type differences for arthropod abundance and richness ...... 28 3.6 CA4: Regional family pool comparison ...... 28 4. Discussion ...... 29 5. Conclusion ...... 39 6. References ...... 40 List of Tables ...... 66 List of Figures ...... 72 List of Appendices ...... 84 vi

LIST OF TABLES

Table 1: Results of linear models between percentage of landcover type of buffer radii (500, 1000, 1500, 2000 m) and total arthropod abundance and richness for 13 farms in Southern Ontario, Canada...... 66

Table 2: Mantel test r-values (Pearson’s product-moment correlation) for the correlation between similarity matrices for functional group abundance and richness (family-level) in conventional farms and similarity matrices for spatial connectivity of farms...... 67

Table 3: Summary of F- and p- values from analysis of arthropod abundance and family-level richness across four focal functional groups in response to cover type. Significant effects (p ≤ 0.05) in bold...... 68

Table 4: Results of Generalized Linear Mixed Models with negative binomial function of field observations evaluating the effect of plant-related variables on arthropod abundance...... 69

Table 5: Results of Generalized Linear Mixed Models with negative binomial function of field observations evaluating the effect of plant-related variables on arthropod family-level richness...... 70

Table 6: Results from ANOVA summary statistics analyzing factors prairie size and prairie age for arthropod abundance and richness (family-level) response...... 71

vii

LIST OF SUPPLEMENTARY TABLES

Table S1: List of study sites and cover type variables. Farms located in Southern Ontario, field research taken from May – July 2017. Annual mean temperature and annual mean precipitation values are taken from WorldClim 1.4: current conditions data at 30 arc-seconds (~1 km)...... 85

Table S2: Total plant species and associated functional groupings found within prairie cover.. 86

Table S3: Definitions of arthropod functional groups...... 94

Table S4: Functional group categorization and literature search...... 95

Table S5: Proportion of agricultural and semi-natural landcover at four buffer radii (500 m, 1000 m, 1500 m, and 2000 m) of 13 farms in Southern Ontario, Canada. Landcover data was extracted from Agri-Canada Annual Crop Inventory...... 111

Table S6: Lab methods for plant tissue quality analysis as provided by SGS Labs, Guelph, Ontario...... 112

Table S7: Summary of arthropod functional group abundance data. All data is presented as the pooled number of individuals caught per site for both yellow pan traps and sweep nets samples...... 113

Table S8: Summary of arthropod functional group richness (family-level) data. All data is presented as the pooled number of individuals caught per site for both yellow pan traps and sweep nets samples...... 114

Table S9: Results of arthropod family identity in agricultural and non-farm (Naturalist) sites and the shared or unique families between them...... 115

viii

LIST OF FIGURES

Figure 1: Total arthropod abundance and richness (family-level) sampled across (A) 22 total farms and (B) 13 farms as they relate to 9 key functional groups. The focal functional groups are given in colour (herbivores = purple, predators = yellow, = blue, pollinators = green) while all other functional groups are in grey. Bars are annotated to show the fraction of the total arthropod community comprised by each functional group when the fraction was > 0.1...... 72

Figure 2: richness (family-level) as related to percent agricultural land use at a buffer radius of 500 m...... 73

Figure 3: (A) Abundance (B) and richness (family-level) of arthropods across three different cover types (crop = green, prairie = yellow, forest = brown). Error bars show ± 1 SE, and lower- case letters indicate significance at p<0.050 as determined by post-hoc tests...... 74

Figure 4: (A) Abundance (B) and richness (family-level) of arthropods belonging to the 4 focal functional groups across three different cover types (crop = green, prairie = yellow, forest = brown). Error bars show ± 1 SE, and lower-case letters indicate significance at p<0.050 as determined by post-hoc tests...... 75

Figure 5: Ordination of arthropod community composition using non-metric multidimensional scaling (NMDS) showing variability between cover types on farm sites (n=13). Farm sites near each other have a community composition that is more similar than farms that are separated by greater distance. Ellipses represent 95% CI around the centroid...... 76

Figure 6: Average soybean leaf defoliation for farms which had prairie adjacent to the crop field (n=7) and those without adjacent prairie (n=3). Error bars show ± 1 SE...... 77

Figure 7: Relationship between abundance and richness of arthropod functional groups in farms with prairie present (n=7) and farms with prairie absent (n=3). (A) Mean abundance of arthropod functional groups in prairie absent farms. (B) Mean abundance of arthropod functional groups in prairie present farms. (C) Mean richness (family-level) of arthropod functional groups in prairie ix

absent farms. (D) Mean richness (family-level) of arthropod functional groups in prairie present farms...... 78

Figure 8: Relationship between abundance of arthropod functional groups in crop fields and distance from edge from farms with (A) farms with prairie present adjacent to crop (n=7) and (B) farms with no prairie present (n=3)...... 79

Figure 9: Relationship between leaf defoliation and distance from edge on farms with prairie present adjacent to crop (n=7) and farms with no prairie present (n=3)...... 80

Figure 10: (A) Mean abundance of arthropod functional groups for different crop types. (B) Mean richness (family-level) of arthropod functional groups for different crop types...... 82

Figure 11: Data from non-farm iNaturalist observations in Southern Ontario compared to 13 agricultural field sites. Each bar in the graph represents the summed count of total arthropod families in each dataset and segments in the graph represent the functional group composition of that total...... 82

Figure 12: Number of arthropod families unique or shared from data of non-farm iNaturalist observations in Southern Ontario compared to 13 agricultural field sites. Each bar in the graph represents the summed count of total arthropod families in each dataset and segments in the graph represent the functional group composition of that total...... 83

x

LIST OF SUPPLEMENTARY FIGURES

Figure S1: Hierarchical framework of habitat-based drivers of arthropod abundance is shown. Habitat occurs at multiple scales: landscape distance-based and local farm cover based. A specialized plot-level analysis was conducted to explain my local-scale prairie cover which consisted of plant-related factors of plant quantity, plant tissue quality, and plant composition. 84

Figure S2: Location of the 22 study sites within the Southern Ontario, Canada region. Farms strictly used in my main analysis are coloured in green. Farms used in complementary analyses include farms coloured in black...... 88

Figure S3: Monthly temperature and precipitation data from historical records (in red) and within my sampling year (in blue) during sampling months May-July 2017...... 89

Figure S4: Quadrant sampling design for arthropod and plant measurements. All sampling occurred from May-July 2017 on farms located in Southern Ontario, Canada...... 90

Figure S5: Conceptual figure of sampling design used. For the majority of cases, a 5x 5 grid would be overlaid onto the satellite image of each farm cover (crop, prairie, wood). In cases where covers were unsymmetrical or where a 5 x 5 grid could not be easily overlaid, instead 25 grid squares that spanned the whole cover were overlaid. In all cases, five sampling locations were chosen using a random number generator...... 91

Figure S6: Sample-based rarefaction curves of arthropod richness were drawn three different cover types (prairie, wood, crop). Solid lines represent the estimated number of species for each cover and the shaded area between the lines indicates the estimated 95% confidence interval. . 92

Figure S7: (A) Mean abundance (± SE) of arthropod community per cover type and (B) Mean richness (family-level) (± SE) of arthropod community per cover type. Bars with the same letter are not significantly different (ANOVA, followed by comparisons between cover type using Tukey’s HSD test, α = 0.05)...... 93 xi

Figure S8: Correlation matrix of landcover data carried out to assess the relationship among predictor variables. Squares in blue indicate a strong negative correlation, squares in red indicate a strong positive correlation...... 118

Figure S9: Correlation matrix of plant resource data carried out to assess the relationship among predictor variables. Squares in blue indicate a strong negative correlation, squares in red indicate a strong positive correlation...... 119

Figure S10: Locations of insect and iNaturalist observations logged from 05 May - 17 July 2017 spanning a 10,000 km² region of Southern Ontario, Canada...... 120

Figure S11: Summary of arthropod (A) abundance and (B) richness (family-level) per farm from a total of 22 farms in Southern Ontario, Canada...... 121

xii

LIST OF APPENDICES

Appendix 1: Conceptual figures ...... 84 Appendix 2: Site selection and description ...... 85 Appendix 3: Sampling methods ...... 90 Appendix 4: Functional group categorization and definitions ...... 94 Appendix 5: Supplementary methods and results ...... 111

1. Introduction

Arthropods are critical components of agricultural landscapes providing valuable ecosystem services relating to nutrient recycling, pollination, biological control, and acting as a food source for higher trophic levels (Losey & Vaughan, 2006; Zhang et al., 2007). Declines of arthropod populations are reported globally (Dirzo et al., 2014; Hallmann et al., 2017; Sánchez-

Bayo & Wyckhuys, 2019). This includes the losses of beneficial groups such as predatory and pollinating arthropods whose reductions may intensify insect pest attack on crops and pollination deficits (Altieri, 1999; Deguines et al., 2014). These trends are closely related to the intensification of farm practices in many parts of the world (Benton et al., 2003; Albrecht et al.,

2007; Geiger et al., 2010; Phalan et al., 2011) suggesting a widening dichotomy between a critical need for arthropod services on farms and farm practices that may reduce their presence and benefits.

Although arthropod declines in farm landscapes are well-described, the causal factors regulating these changes are not fully clear. One central issue is the role of habitat loss relative to the obvious impacts of insecticides (Robinson & Sutherland, 2002; De Clerck-Floate &

Cárcamo, 2011). Habitat loss, in many agricultural landscapes, derives from decades of change towards fewer farms with larger fields supporting crop monocultures, fewer indigenous floral and nesting resources, and reduced availability of habitat refuges against factors including spraying and drought (Tillman et al., 2011). Habitat is a rigid limiting factor for many arthropod populations, providing stable food resources, nesting sites, and overwintering refuge (Benton et al., 2003; Duelli & Obrist, 2013; Schellhorn, Gagic & Bommarco, 2015). In many farm landscapes lacking habitat there will be a loss of any arthropods unable to persist in crop 1

monocultures. Many farm regions support some degree of habitat complexity such as grassland borders, forest patches, and field edges with wildflowers. An increasing issue of interest is using often small amounts of high-quality habitat in marginal edge areas (Paterson et al., 2019). How do these refuge areas affect arthropod groups? The impacts appear to be highly variable by habitat type (closed forest vs open edge or restored areas), by arthropod functional group including whether such areas are a greater benefit to herbivores or predators, and the type of crops planted (Tscharntke et al., 2012; Ma et al., 2019). This complexity in terms of responding to limited habitat can apply to non-targeted species (i.e., not subject to pest control), the most well studied and rapidly declining group (Ramsden et al., 2017). It can also apply to insect pests, including possibly providing unintended beneficial alternative host plants for parts of their life cycles and refuges from spraying. Additionally, there are intricate relationships between pests and natural enemies mediated by habitat – sometimes in remnant patches the former does better, sometimes the natural enemies control the pests (Tscharntke et al., 2007). In total, these complexities indicate that we still lack a fundamental understanding of the habitat-based dynamics that regulate arthropod groups – beneficials and pests - in agricultural landscapes.

Many of the uncertainties with habitat remnants and arthropods center on scale- dependent mechanisms of habitat size, distance among required habitat, plant cover type, and plant resources, which are rarely tested together and may act differently on some arthropod groups over others (Harvey & MacDougall, 2014, 2015; Grainger & Gilbert, 2017). Habitat impacts in agroecosystems can unfold at landscape and local scales (Soberon, 2007), each of which I test. At the landscape scale, distance among habitat types may shape arthropod communities strictly by influences on dispersal (closer sites have more similar arthropod communities than sites farther away), but especially if preferred habitat such as semi-natural 2

areas (prairie strips, forest, hedgerows, field edges) are relatively rare compared to crop monocultures (Tscharntke et al., 2012). At the local scale, the presence of different habitat types- specifically crop fields, forest, and permanent cover of herbaceous species found in field edges or restored marginal lands- can be important determinants for sustaining populations of arthropods (Hanski & Ovaskainen, 2000). Species richness of arthropods is expected to increase in habitat with greater resources and patch size due to the increased likelihood of supporting more life history strategies (Lawton, 1983; McCoy & Bell, 1991). In total, understanding how habitat - be it landscape, local or both - affects arthropods in farm landscapes is an urgent focus, given how and beneficial functional groups are both important for the future of food.

My research tests the influence of scale-dependent habitat factors on the abundance and richness (family-level) of arthropod communities (see Appendix 1: Fig. S1) with a focus on four agriculturally important functional groups (herbivores, predators, parasitoids, pollinators) on 22 farms across a 10,000 km2 agricultural landscape in Southern Ontario, Canada. My sampling was structured around the examination of distance-based spatial constraints at the landscape-scale and the influence of cover type at the local-scale within farms. My local-scale work on cover type included restored high-diversity prairie on peripheral marginal lands (e.g., riparian strips, field edges), as part of a farmer-led national program growing services on farms (ALUS Canada,

2019). To that end, I conducted additional analyses at the plot-level in prairies – the most influential local factor that I detected (i.e., the presence of prairie for arthropods) – investigating effects of plant species composition and related issues of tissue quantity and quality given that I found variation in these factors largely absent in my sampling for crop fields and forest.

3

I also conducted several complementary analyses relating to drivers of arthropod richness

(family-level) and abundance on farm landscapes. First, I tested if arthropod community composition differed among cover types (crop, prairie, forest), specifically if the two semi- natural habitats (prairie and forest) widely differed. Second, I tested the potential functional implications of restored prairie habitat, given that they could support both pest and predator populations with unclear implications for crop damage (Tscharntke et al., 2007). That is to say, whether prairie habitat potentially harbors high herbivore populations that lead to higher leaf defoliation, or the alternative where populations of beneficial arthropod predators curtail leaf defoliation. Third, I tested the influence of crop type on arthropod abundance and richness

(family-level), and whether herbivore groups would show stronger responses due to host- specificity or beneficial groups due to nectar and resources by different crops. Finally, given the widespread expectation that arthropods are declining with agricultural intensification

(Sánchez-Bayo & Wyckhuys, 2019) I pooled my farm data for family-level richness and contrasted it with a regional database from non-farm areas in the surrounding region sampled during the same time period. I was interested if total differences in richness (family-level) were apparent and which families were absent from farms (e.g., ground nesting , butterflies).

2. Methods

2.1 Site selection and description

In this study, twenty-two farm study sites were selected from a 10,000 km² intensely farmed agricultural region of Southern Ontario, Canada in Hillsburgh (43°47'30.7N

80°08'35.0W), Norfolk (42°52'17.4N 80°18'26.8W), and Elgin Counties (42°46'59.6N

4

81°10'54.9W) (see Appendix 2: Fig. S2). Approximately 92% of this region supports crop monocultures, especially soybean (41.7%), corn (28.7%), and wheat (12.5%) (Government of

Canada, 2017). The remainder of the landscape, not including settlement, is largely forest and non-crop cover such as oldfield plant communities along roadsides and crop field margins. The region has a temperate climate strongly influenced by the proximity to Lake Erie, with 7.75 ℃ mean annual temperature, 946 mm mean annual precipitation (Fick & Hijmans 2017; Appendix

2: Table S1), and frost-free days averaging between 160 to 170 from April to October

(OMAFRA Publication 811, 2017; Government of Ontario, 2019). Mean monthly temperature and precipitation data during my sampling year were consistent with historical climate data

(years 1970-2000) (Appendix 2: Fig. S3). Soils have developed on glaciolacustrine deposits with surface textures silty clay or silty clay loam, and some areas of silty or sandy loam (Presant,

1984). For each of the twenty-two farms, I conducted standardized arthropod sampling by cover type depending on which covers were present on the farm (e.g., some farms lacked restored prairie). The farms ranged in size from 2.37– 115.07 acres, with crop, prairie and forest cover of various sizes (if present) per farm (Appendix 2: Table S1).

All farms were part of a national farmers program (ALUS, 2019) seeking to “grow” ecosystem services in agricultural landscapes including providing good quality habitat to beneficial . When present, prairie grassland covered 2% to 21% of total farm area. Four farms were part of ALUS but had not yet planted prairie, providing me the opportunity to test farms with and without this restored cover type. All sites with prairie were planted with a similar mix of grasses and forbs, although species dominance, diversity, and composition of the established plant community sometimes varied among farms (Appendix 1; Table S2). Overall, forb species varied the most among farms while several C4 grasses tended to be present at all 5

sites (e.g., big bluestem (Andropogon gerardii), Indian grass (Sorghastrum nutans), and switchgrass (Panicum virgatum)). Given how prairie cover was different on each farm, I conducted a specialized series of analyses using finer-scale plot-level data to investigate how plant-based resources relating to plant quantity, tissue quality, and plant composition influenced on-farm arthropod abundance and richness (family-level).

From the twenty-two farms, my primary focus (“Main Analysis” [MA]) involved thirteen farms that represented the dominant landcover profile of the region – conventional crop monocultures with soy, corn, or wheat. Additionally, these farms all had forests and small marginal cropland areas (e.g., poorly drained soils, steep slopes) that had been converted to permanent prairie grasslands. Using these thirteen MA farms, I tested my main hypotheses regarding whether landscape versus local factors of habitat determined the abundance and richness (family-level) of arthropods. Landscape-scale spatial factors were assessed as distance among farms, and distance among similar nearby semi-natural habitat (described below). If distance drives large spatial turnover in arthropod communities over my 10,000 km2 study area, whether it be among farms or among semi-natural habitat, then I would predict spatial separation among these factors should determine arthropod abundance and richness (family-level) regardless of cover type variation. For example, my most isolated farms would likely have the lowest numbers of species, especially if semi-natural habitat is lacking. Local-scale habitat factors were assessed based on aspects of habitat type of crop, prairie, and woodland (described below). If local habitat drives arthropod abundance and richness (family-level) over my approximate 10,000 km2 study area, then I would predict ‘all insects to be everywhere’ (because spatial factors are not limiting) with differences shaped instead by cover type on farm. In short, this latter prediction turned out to be the case. 6

Given that local farm cover type largely ‘explained everything’, I conducted a series of complementary analyses to flesh out some of the underlying details explaining this outcome.

First, I was interested in testing arthropod community composition overlap between three on- farm cover types (crop, prairie, forest) (“Complementary Analysis 1” [CA1]). I was especially interested in the forests which were present on all farms. Forests have received significant attention on agricultural landscapes, including a recent study by Hallmann et al., (2017) testing long-term changes in flying insect biomass in wooded areas of heavily farmed regions of

Germany. On one hand, forests may serve as important refuges for arthropods in heavily cropped regions by providing sanctuary from spraying as well as food and nesting resources (e.g., floral resources in early spring via “spring-ephemeral wildflowers (Dailey & Scott, 2006)). On the other hand, there may be strong compositional differences between forest and open-field arthropods based on microclimate: arthropods are ectothermic and thus highly sensitive to temperature and humidity (Heinrich, 1996) such that large percentages of forest arthropods in farm landscapes may not commonly venture forth (Duflot et al., 2015). To test for differences between arthropod community composition by cover type I used the Bray-Curtis dissimilarity index (Bray & Curtis, 1957) from my thirteen MA farms. These farms were used because they contained all three cover types and with equal sampling effort.

Second, my results of MA and CA1 pointed to one clear finding – the presence of semi- natural restored prairie habitat largely explained the variation I observed in arthropods across my study region. Further, this habitat housed several groups of beneficial arthropods including predatory insects known to attack crop pests. Thus, I tested whether the occurrence of prairie had detectable differences in arthropod abundance, richness (family-level), and arthropod-derived ecosystem services (“Complementary Analysis 2” [CA2]). In terms of services, I tested whether 7

the occurrence of prairie influenced crop leaf damage and arthropod distribution in edges or interior of crop fields directly adjacent to the prairie. For this analysis, I compared data from three soybean “complementary farms” that did not have the occurrence of prairie to seven “main analysis” farms that also grew soybean but had prairie cover (more detailed sampling methods provided below).

Third, I compared the influence of other forms of farm cover-type on the arthropod richness (family-level), abundance, and functional group differences (“Complementary Analysis

3” [CA3]). To do this, five of my “complementary farms” grew crops other than soy, corn, wheat

(three farms: apple orchards, mixed organic, sorghum) or lacked crops altogether (two farms:

100% restored prairie). The lack of sampling power did not allow me to fully test the difference in beneficial arthropods on these farms versus my core thirteen. However, I conducted non- parametric assessments of arthropod richness (family-level), abundance, and functional group differences between these complementary farms and my main analysis farms.

Finally, I compared broad difference in the arthropod families I sampled from my thirteen

MA farms, versus the total family pool for the region observed in non-farm areas

(“Complementary Analysis 4” [CA4]). I was especially interested whether farm richness (family- level) was highly depauperate versus the regional pool and which major functional groups might be missing. An increasing number of studies are suggesting large insect declines globally, especially in farmed landscapes (Benton, Bryant & Cole, 2002). Given that I work in one of the more intensively farmed regions of central North America, with crop cover >90% of my study area, an obvious prediction would be that the richness (family-level) of arthropods on my farms would be greatly reduced compared to more habitat-rich areas regionally. Relatedly, it would

8

also seem likely that the arthropods found on farms would be a non-random collection of functional groups with life history traits that may favor persistence in areas dominated by crop monocultures. For example, ground nesting bees (: ) are often assumed to be rare or absent on farms because they lack nest sites and nest construction materials

(Williams & Kremen, 2007). To test these possibilities, I assembled a database representing the

“regional family pool” for Southern Ontario using data extracted from iNaturalist, a community science program (https://www.inaturalist.org/ - details described below).

2.2 Arthropod and habitat sampling

Prior to field sampling, I randomly selected five sampling locations within each farm cover type (crop, prairie, forest) to establish a 5 x 5 m quadrant that arthropods and plants would be sampled (5 quadrants x [20 crop cover + 18 prairie cover + 22 wood cover] = 300 quadrants).

To do this, I overlaid a 5 x 5 grid onto a satellite image of each farm cover using the Agricultural

Information Atlas (AgMaps, 2017), then used a random number generator to select five sampling locations in the grid. In situations where a 5 x 5 grid could not be easily overlaid on top of a cover, such as strip prairies or unsymmetrical covers, I overlaid 25 grid squares that spanned the whole cover and five sampling locations were chosen using a random number generator. The position of all sampling locations was pre-recorded using the UTM coordinate system and later used to find locations in the field using a Trimble GeoXT 2008 GPS (horizontal resolution of 2 meters set to zone 17 north and WGS 1984). See Appendix 3 for full details of sampling methods.

9

Arthropod field sampling occurred between 15 May to 4 July 2017. This sampling window was selected to maximize the number of farms in my study, while occurring prior to the period of insecticide application regionally for corn, soy, and wheat. The recommended application period date is 21 pre-harvest interval days for corn, 28 pre-harvest interval days for soybean, and 30 pre-harvest interval days for wheat (OMAFRA Publication 812, 2018). Based on these recommended dates, insecticide application would occur in late July or early August on farm sites, thus after my sampling period. Given the extent of my sampling period and the phenological shifts that may occur in arthropod communities over the summer growing season

(Welch & Harwood 2014), there was a risk of compositional turnover that might bias my analyses. To test for this, differences in arthropod community composition between sampling dates was assessed using the Bray-Curtis dissimilarity index (Bray & Curtis, 1957) on the arthropod community. I conducted a permutational multivariate analysis of variance

(PERMANOVA) to test the null hypothesis of no differences in the arthropod community among different sampling dates using the adonis function in vegan package of R (Oksanen et al., 2016).

Overall, I found no significant influence of sampling date on my arthropod community dissimilarity (F= 0.78, R² = 0.014, p = 0.774), therefore indicating that my farms were statistically comparable despite being sampled at different times.

Arthropods were sampled within each 5 x 5 m quadrant using two yellow pan traps placed at different heights (H1 at soil level and H2 at canopy level) and filled with soapy water to reduce surface tension (Leong & Thorp, 1999). Sampling occurred on sunny days between

10:00 a.m. and 4:30 p.m., at temperatures above 15 ⁰C, and with wind speeds less than 20 km/h.

Yellow pan trap samples were placed in whirl-pak bags, stored at 4 ⁰C until arthropods were

10

transferred to 90% ethanol for future identification. Yellow pan traps were used because they attract a wide range of taxa which include flying and ground dwelling arthropods (Kirk, 1984;

Marshall et al., 1994).

Yellow pan trap samples were complemented with sweep-netting to capture the arthropod community at a broader cover-level habitat scale. Sweep net samples consisted of a W-transect standardized by beat frequency, net size, and occurred at the same time in each cover for five minutes by three field technicians. This sampling method was chosen because it is the recommended strategy for growers to scout their fields for arthropods (OMAFRA Publication

811, 2017). Each sweep-net sample was placed into individual top-closure polyethylene bags, labeled according to its respective date, site, cover, and stored at -20° C for future identification.

In total, 60 sweep net samples were conducted because not all farms contained all cover types (1 sweep net x (20 crop cover + 18 prairie cover + 22 wood cover)).

In the laboratory, arthropods were identified to family-level at minimum, and lowest taxonomic level when possible, using a combination of taxonomic keys, photographic guides and natural history references. Bee specimens were identified to genus at minimum, and lower taxonomic level when possible, by the Nigel Raine laboratory. Although identification to relatively coarse levels of family or genus captures less diversity versus identification to the species-level, the large number of arthropods collected (total 27,080) necessitated this approach.

Family-level taxonomic resolution has been found to be sufficient for detecting significant patterns of community composition and habitat preference in temperate systems, including agroecosystems (Cardoso et al., 2004; Timms et al., 2013). After identification I classified arthropods into 9 functional groups (Appendix 4; Table S3) by life stage to account for

11

arthropods with ontogenetic shifts in feeding styles (e.g., Syrphidae larvae are predators, but adults are pollinators). Functional group classification was based on the most common feeding mode within the family and determined by an extensive literature search (Appendix 4; Table S4).

The primary goal was to examine key drivers of abundance and richness (family-level) of four groups of arthropods considered agriculturally important: (1) herbivores because they are the primary consumers, linked closely to the plant community and can be significant pests, (2) predators, because they are important pest control agents, (3) parasitoids, because they are also valuable pest control agents often used for biological control and (4) pollinators due to the important pollination service they provide. Five other “secondary” functional groups were categorized but grouped into a “total” arthropod functional group for subsequent analyses. These functional groups were: kleptoparasite, aquatic, decomposer, anthophilous, and uncommon.

Individuals categorized into group “uncommon” were arthropods of unresolved feeding guilds or when less than three individuals were collected total at a farm site, regardless of cover.

2.3 Main analysis

2.3.1 Landscape and local factors of habitat on 13 agricultural sites

My landscape-scale hypothesis was that distance among farms and, more specifically, distance among nearby semi-natural habitat, would strongly determine arthropod richness

(family-level) and abundance on a farm. For this analysis, arthropod abundance and richness data were compared among the farms (n=13).

More distant farms could have different arthropod communities due to spatial turnover.

Alternatively, there may be little influence of distance if “everything is everywhere” which might 12

be possible given that the dispersal dynamics of most arthropods are unknown except in a few cases (e.g., pollinators). To test this, I calculated the between-farm distance (km) of each farm to every other farm site using the ArcGIS measure tool. This resulted in 13 distance values per farm which were averaged into one estimate and used as a measure of distance. A Mantel test was used to examine if there were relationships between distance in geographic space and arthropod family or abundance space (McCune & Mefford, 1999). The Mantel test is widely used in ecological studies to examine whether populations are similar across sites (Koenig & Knops,

1998).

The distance of semi-natural areas could also be affecting abundance and richness

(family-level) with the assumption that this habitat (i) likely supports greater numbers of arthropods due to higher availability of floral and nesting resources and the possibility that these areas provide refuge against crop spraying, and (ii) that more nearby semi-natural habitats means greater likelihood that these arthropods disperse into farms. To test this, I calculated the landscape-level landcover surrounding the centroid of each of my 13 farms at four distances

(500, 1000, 1500, and 2000 m radii). The radii were chosen to capture likely dispersal distances for less mobile (e.g., parasitoids) and highly mobile (e.g. flying predators) functional groups

(Schellhorn, Bianchi & Hsu, 2014). Landcover data were obtained from Agri-Canada annual crop inventory (AAFC, 2017) and re-classified into aggregate percent agricultural landcover and percent semi-natural landcover (see Appendix 5: Table S5). Specifically, agricultural landcover contained the following specific land use types: corn, wheat, soybean, rye, ginseng, tobacco, fruit and berry crops, orchard, vegetable, beans, oats, barley, fallow, vineyard, potato, peas. Semi- natural land contained the following specific land use types: coniferous, broadleaved and mix

13

wood, grassland, shrub land, wetland. This analysis was conducted using ArcGIS version 10.3.1

(ESRI, 2014).

My local-scale hypothesis was that on-farm habitat type (crop, prairie, forest) would be influencing arthropod abundance and richness (family-level). By studying the habitat use of arthropods in agricultural landscapes, insights can be gained into how arthropod populations can be regulated. To investigate local habitat use, I compared arthropod abundance and richness

(family-level) per cover type within farm (n = 39; 13 farms x 3 cover types).

Due to size differences among cover types, there was a risk that I over-sampled smaller areas while under-sampling larger ones. I thus evaluated arthropod sampling effectiveness by constructing sample-based rarefaction curves using the specaccum function in the vegan package of R (Oksanen et al., 2016). Accumulation curves (Appendix 3; Fig. S6) tended towards saturation and sampling was thus deemed sufficient for comparisons among cover types. Corn, soy, and wheat were pooled to create the crop cover type as they represent the most common rotated conventional cropping system in Southern Ontario. Further, arthropod abundance (F =

0.23, p = 0.798) and richness (family-level) (F = 0.55, p = 0.592) did not differ between crop type (Appendix 3: Fig. S7).

2.3.2 Within-prairie habitat factors

As mentioned, my analyses showed that the presence of local prairie habitat largely explained most of my variation in arthropod occurrences. As such, I examined finer-scale measures within the prairie for plant quantity, quality, and composition to investigate effects on arthropod abundance and richness (family-level).

14

To do this, plant variables were sampled in July in 1m2 plots inside or adjacent to the 5 x

5 m quadrant where arthropods were sampled (see Appendix 3: Fig. S4). For this analysis, arthropod abundance and richness (family-level) data were the value per plot from farms with prairie cover (n = 90; 18 farms x 5 plots per prairie). The following variables were measured: photosynthetically active radiation (PAR intercept – the ratio of light above the canopy versus ground level with 1.0 being full light and 0 being full light interception), litter (% cover), bare ground (% cover), and plant species composition (% cover). I measured photosynthetically active solar radiation (PAR) by placing an LI-190SA quantum sensor coupled to an LI-250 Light Meter

(LI- COR) on the ground surface in the center of each plot and at the canopy surface. I took the intercept between the two measurements representing the proportion of available light reaching the ground. I allowed the readings to stabilize for 10 seconds before recording a measurement.

Litter, bare ground, and plant species composition were measured as the percent cover in each plot. Plant species composition was used to calculate plant richness (species-level) and percent cover of plant functional groups: C3 grasses, C4 grasses, and forbs (all in % cover). To the southwest corner adjacent to the 5 x 5 m quadrant, in a 1 x 0.2 m plot, aboveground plant biomass was cut using garden shears. Following this harvest, samples were dried in a 65°C oven for 24 hours, weighed, then homogenized in a Wiley Mill using a 1 mm grinding plate and sent for plant nutritional quality analyses by SGS-AgriFood labs in Guelph, Ontario (Appendix 5:

Table S6). Plant tissue quality variables included: carbon to nitrogen ratio (C:N), organic carbon, phosphorous, actin-depolymerizing factors (ADF) proteins, and nitrogen. These variables were chosen due to their influence on plant palatability to arthropod herbivores (see Appendix 4:

Table S6 for details). Similarly, light, bare ground cover, and litter can be important factors

15

affecting development, distribution and behavior of many insects (Michener, 2000; Schowalter,

2006).

2.4 Complementary analyses

Soybean farms were revisited in July to conduct leaf defoliation assessment, comparing differences between farms that had prairie cover (n=7) to those without prairie cover (n=3). On average, farms with prairie cover had the same sized crop fields as farms without prairie cover (t

= 2.23, p = 0.066). For this analysis, I was interested whether the occurrence of prairie was associated with detectable differences in crop leaf damage ad arthropod abundance. I was also interested whether distance from the edge of prairie altered arthropod distribution (i.e., more herbivores closer to the prairie) and if this was true, if this affected leaf defoliation in the same way (e.g., more towards the prairie). Several papers note that soybean yield is reduced when defoliation occurs during bloom to pod formation (McWilliams et al., 1999; Pedersen, 2007;

Owen, 2013; Ohnesorg & Hunt, 2015) and therefore sampling in July was essential when soybeans are in their reproductive stages (Pedersen et al., 2014). Measurement of crop leaf damage occurred in 5 randomly placed 1 x 1 m plots in each crop field, varying distances from edge to interior of field within dispersal capabilities of functional groups (randomly placed 1- 90 m from field edge). In each plot 20 leaves were randomly chosen from different plant heights

(top, middle, bottom) to assesses crop leaf damage by estimating leaf defoliation at 5 % intervals

(Ohnesorg & Hunt, 2015). Values of leaf defoliation from each plot were averaged resulting in a mean leaf defoliation (%) value for each crop plot distance.

16

Finally, I assembled a database representing the “regional family pool” for Southern

Ontario using data extracted from iNaturalist, a community science program

(https://www.inaturalist.org/). A search query was created filtering family-level observations by

(i) place: Southern Ontario region, (10,000 km² region from non-farm locations such as conservation areas, hiking trails, parks, forests; see Appendix 5, Fig. S10) (ii) taxon: and insects, and (iii) date range: 2017-05-15 to 2017-07-04 (overlapping collection period for farm sites). In total, 10,742 family-level observations were obtained from iNaturalist that fit this search query and combined into a “regional family pool” list consisting of 194 families. Each arthropod family was categorized into a functional group as discussed in the above arthropod sampling methods. Observations from iNaturalist were compared to my 13 agricultural field sites, with a total of 12,558 observations pooled into an “agricultural family pool” list consisting of 183 families. The data from this analysis were only used as a comparison for my on-farm arthropod sampling because community science data has the potential to over-emphasize charismatic and easy-to-identify functional groups such as butterflies and under-emphasize small-bodied arthropods such as parasitic which are considered hyper-diverse in temperate and tropical regions (Eagalle & Smith, 2017).

2.5 Statistical analysis

All analyses were conducted in Rstudio version 3.5.1 (Rstudio Team, 2015; R

Development Core Team, 2018). For all analyses, assumptions of normality and equal variance were verified using residual plots, Shapiro-Wilk test of normality (Shapiro & Wilk, 1965), and

Bartlett’s test of equal variance (Snedecor & Cochran, 1989). Non-normal data were evaluated

17

using non-parametric tests. Significance of all factors was evaluated with Type II tests using the

Anova function in the car package (Fox & Weisberg, 2011).

2.5.1 Landscape and local factors of habitat on 13 agricultural sites

To test my hypothesis that distance among semi-natural habitat affects arthropod abundance and richness (family-level), I performed linear models with the lm function with the package “stats” (R Core Team, 2019). Fixed factors included land use (% agricultural and % semi-natural) at each buffer radii. Given the high degree of multicollinearity between agricultural and semi-natural landcover across the 4 buffer zones (Appendix 5: Fig. S8), each buffer zone was analyzed separately.

To test my hypothesis that geographic distance among farms affects arthropod abundance and richness (family-level), I performed Mantel tests using the mantel function in the package

“ape” (Paradis & Schliep, 2018) to assess spatial autocorrelation of measured geographic location of farms and arthropod community data (n=13; summed response value per farm)

(Peres-Neto & Legendre, 2010). This test evaluates the similarity between two matrices: one measuring ecological distance (e.g. differences in arthropod family richness or abundance) and one as geographical distance (Legendre & Legendre, 2012). I used Monte Carlo permutations with 999 randomizations to test for significance (Oksanen, 2014).

To test my hypothesis that cover type affects arthropod abundance and richness (family- level), I performed a linear mixed model using the lmer function in the “lme4” package (Bates et al., 2015). Cover type was used as a fixed effect and farm site was treated as a random effect to account for site specific differences. Here, arthropod abundance and richness (family-level) was

18

aggregated by cover type within farm (n = 39; 13 farms x 3 covers). When there was a significant effect of cover type, Tukey’s post-hoc tests were run using the emmeans call to evaluate the response variable against the cover types while accounting for the random effect of farm identity. I evaluated total arthropod responses as well as responses belonging to the four focal functional groups in this manner.

2.5.2 Within-prairie habitat factors

To test my hypothesis that plant-based resources affect arthropod abundance and richness

(family-level), I used generalized linear mixed-effect models with the glmer.nb function, as my data were over-dispersed, in the ‘MASS’ package in R (Venables & Ripley, 2002). Fixed effects included the following plant-based resource variables relating to plant quantity, quality, and composition: total plant biomass, PAR intercept, litter % cover, bare ground % cover, C:N, phosphorous, ADF, organic carbon, nitrogen, C3 grasses, C4 grasses, forbs, and plant richness.

For this analysis arthropods collected from pan traps were used where plant data was also collected (n = 90, 18 farms x 5 plots in prairie cover). Since plants indirectly provide prey resources for predator groups, an additional predictor variable of herbivore abundance or herbivore richness was used in predator or parasitoid models (whether testing abundance or richness data, respectively).

In all models farm site was used as a random effect. I ran separate models for arthropod abundance and richness (family-level) for the total arthropod community (all functional groups pooled) and my four focal functional groups. Predictor variables used in my models were not found to have high collinearity (r >0.06) and therefore no variables were removed (Appendix 5:

Fig. S9). 19

2.5.3 CA1: Arthropod community composition among cover types

To study differences in arthropod community composition (including taxa of all functional groups) among cover type (crop, prairie, forest), I calculated the Bray-Curtis dissimilarity index (Bray & Curtis, 1957) for each farm cover type. The Bray-Curtis dissimilarity index is a version of the Sørensen index (Magurran, 2004) modified to include abundance. I visualized the dissimilarity in reduced dimensionality using non-metric multi-dimensional scaling (NMDS) using the R package “vegan” (Oksanen et al., 2013). A Monte Carlo randomization test was performed on the final stress value to calculate goodness of. I tested the significance of cover type with a permutation test using the adonis function in “vegan” (Oksanen et al., 2018).

2.5.4 CA2: Assessment of arthropod-derived ecosystem service mediated by habitat

To test my hypothesis that the occurrence of prairie on a farm will affect response variables leaf defoliation and arthropod abundance, I compared responses between farms with prairie (n=7) to farms without prairie (n=3). For these analyses, means were calculated from crop cover plots from each farm (prairie present n= 35 (5 plots x 7 farms), prairie absent n= 15 (5 plots x 3 farms). I used a Welch Two-Sample T-test that assumes unequal variances to account for our unequal sampling design. I used the function t.test in the package “stats” with the argument var.equal=FALSE to assume unequal variance (R Core Team, 2019). Analyses were conducted separately for each response using fixed effect “presence” (with prairie absent or prairie present as a categorical variable).

20

To test my hypothesis that the occurrence of prairie on a farm will affect arthropod dispersal (of my four focal functional groups) and leaf defoliation from edges to interior of crop fields, I analyzed the same crop plot data in the previous analysis with distance as a fixed effect.

First, I analyzed whether the abundance of my focal functional groups (herbivores, predators, parasitoids, pollinators) differed with plot distance (1-90 m from a crop field edge) using generalized linear mixed models with the glmer.nb function using farm as a random effect. This statistical analysis was appropriate to deal with my unbalanced sample sites and my over dispersed data. Second, I analyzed whether leaf defoliation differed with plot distance using the same statistical analysis.

2.5.5 CA3: Crop type differences for arthropod abundance and richness (family-level)

I used a Kruskal-Wallis test to investigate if on-farm crop type affects arthropod abundance and richness (family-level). A Kruskal-Wallis test is a non-parametric alternative to a one-way ANOVA test and used in the situation where there are more than two groups being compared. Here, my response variable was the summed arthropod abundance or richness

(family-level) in each crop type per farm. Separate analyses were run for the total arthropod community, then for herbivore, predator, parasitoid, and pollinator functional groups. This analysis was performed using the kruskal.test function in the package “stats” implemented in R

(R Core Team, 2019).

2.5.6 CA4: Regional family pool comparison

Since the number of families per site (non-farm iNaturalist vs farm field sites) was pooled

I could not use statistical analyses to compare the two sites. Rather, I broadly compared the two

21

sites using visual bar plots and compared differences between total number of families and between functional groups in each site. Further, I described family differences between the two sites in descriptive tables and whether families were found only in farms sites, only in non-farm iNaturalist sites, or shared between sites.

3. Results

I tested how scale-dependent drivers of habitat affect arthropod abundance and richness

(family-level) in agricultural landscapes of Southern Ontario (see Appendix 1: Fig. S1). My results show that arthropods responded to cover type at the farm level. The main habitat driver was prairie grassland cover, which produced two times greater arthropod abundance than crop fields or forests. Within prairie strips, plant composition was the primary factor at the plot-level.

Crop fields were generally depauperate while forests tended to have widely different arthropod communities presumably due to habitat structure and microhabitat conditions that differed significantly from open prairie and crop areas. In terms of service provision, the addition of prairie adjacent to a crop field did not increase leaf defoliation, despite prairies having higher herbivore abundances. Crop type was found to be an important factor influencing arthropod groups, with herbivore groups preferring mixed organic crop type. Regional factors relating to spatial distance were not key and, surprisingly, my comparison with iNaturalist data from surrounding non-farm sites had similar arthropod richness (family-level) as agricultural sites although with some differences in arthropod functional groups.

Over the 44-day sampling period a total of 27,080 arthropods were caught from 22 farms with 223 families. Of these totals, 18,209 individuals belonged to one of the four focal functional

22

groups, consisting of 70 families of herbivores, 64 families of predators, 33 families of parasitoids, and 17 families of pollinators. These key functional groups accounted for 82 % of arthropod abundance and 82 % among the richness total numbers collected (Fig. 1A). Of my remaining five functional groups, 25 families consisted of anthophilous groups, 9 families of decomposer, 4 families of aquatic, 4 families of kleptoparasite, and 16 families of uncommon groups. The most abundant families caught (> 1000 individuals) were: Chloropidae (Order:

Diptera), Cicadellidae (Order: ), Anthomyiidae (Order: Diptera), and Chrysomelidae

(Order: Coleoptera), which were all categorized into the herbivore functional group. Summary of arthropod abundance and richness data for all farm sites can be found in Appendix 5 (Table S7,

Table S8, Fig. S110).

My landscape and local analysis focused on 13 farms, of which, 14,175 arthropods were collected from 183 families. Of these totals, 11,525 individuals belonged to one of the four focal functional groups, consisting of 52 families of herbivores, 52 families of predators, 24 families of parasitoids, and 10 families of pollinators. These key functional groups accounted for 81 % of arthropod abundance and 80 % richness among the total numbers collected (Fig. 1B). Of my remaining five functional groups, 23 families consisted of anthophilous groups, 9 families of decomposer, 4 families of aquatic, 3 families of kleptoparasite, and 6 families of uncommon groups.

3.1 Landscape and local factors of habitat on 13 agricultural sites

Percentage semi-natural landcover at the landscape-scale was not a major driver of arthropod abundance or richness (family-level) at any buffer radius (Table 1). Similarly, percent

23

agricultural land use at the landscape-scale did not show any significant effects, with one exception. At the 500 m buffer radius, parasitoid richness was highest with low agricultural landcover, and declined with higher agricultural landcover (F = 7.12, p = 0.022, Fig. 2).

Although farm distance ranged from less than 1 km to greater than 60 km apart, farms closer to each other were not more similar in total arthropod abundance or richness than farms farther away (abundance: r = 0.086, p = 0.152; richness: r = 0.072, p = 0.146). This result was true for the abundance and richness responses of each focal functional group (Table 2).

Basically, the same collections of arthropods tended to be represented at all farms.

Cover type had a significant influence on arthropod abundance and richness (abundance:

X2 = 19.56, p < 0.001; richness: X2 = 41.84, p < 0.001; Fig. 3). More specifically, prairie had greater arthropod abundance than either the crop or forest covers (p = 0.002 and p = 0.003, respectively) but there was no significant difference between overall arthropod abundance between the crop and forest cover type (p = 0.949). For arthropod richness, crop cover type had the lowest richness, while prairies and forests had greater richness than crops (p < 0.01 for both), and forest and prairie cover were marginally significant (p = 0.054).

Abundance and richness of the four focal functional groups also varied significantly between cover types (Table 3, Fig. 4). In all instances, there was greater abundance and richness in the prairies as compared to the crop fields, while differences between forest-type covers versus both prairies and crop fields depended on the functional group. For example, herbivore abundance was not significantly different between crop or forest covers (Fig. 4A) but herbivore richness was significantly different between all cover types (Fig. 4B).

24

3.2 Within-prairie habitat factors

My previous results indicate that within-farm prairie cover type is a large driver of arthropod abundance and richness, but I further investigated how plant resources at a finer scale

(plot-level) influenced the arthropod community. By measuring plant resources of cover quantity, quality, and plant composition, I found that plant composition was a more important driver of arthropod abundance and richness, however, responses varied by functional group

(abundance: Table 4, richness: Table 5).

The abundance of the total arthropod community was positively associated to increasing forb percent cover (Z = 2.22, p = 0.026) and greater solar radiation at ground level (PAR) (Z =

2.09, p = 0.036). Herbivore abundance was positively associated with percent cover of all plant functional groups (C3 grass, forb, and C4 grass). In fact, only the herbivore functional group responded positively to greater C4 grass percent cover (Z= 2.14, p = 0.033), which is the plant functional group most pronounced in prairie grassland habitat. Herbivore abundance also increased with increasing PAR (Z = 1.98, p = 0.048), which means that more herbivores were found where light was able to penetrate the canopy of prairie cover. Predator abundance was similarly affected by the same predictor variables as the herbivore functional group, with the exception they were not influenced by C4 grass percent cover and were positively influenced by increasing bare ground percent cover (Z = 2.36, p = 0.018). The strongest predictor variable for predators was the abundance of herbivores (Z = 3.95, p < 0.001). Similarly, parasitoid abundance, being natural enemies of herbivores, was also positively influenced with greater herbivore abundance (Z = 2.99, p = 0.003). Additionally, parasitoid abundance was positively influenced by litter percent cover (Z = 2.16, p = 0.031). Pollinator functional groups were

25

negatively associated with C3 grass percent cover (Z = -2.67, p = 0.008) and bare ground percent cover (Z = -2.15, p = 0.032). All other variables, including plant tissue quality and quantity

(plant biomass), did not significantly influence arthropod functional group abundance.

Just as with arthropod abundance, richness of the total arthropod community was positively associated with increasing forb percent cover (Z = 3.00, p = 0.003) and increasing solar radiation at ground level (PAR) (Z = 2.80, p = 0.005). Total arthropod richness additionally showed positive association to C3 grass percent cover (Z = 2.47, p = 0.014), and litter percent cover (Z = 2.20, p = 0.028), variables which did not significantly affect total arthropod abundance. Herbivore richness was only positively associated with one variable, which was an increasing amount of PAR (Z = 2.25, p = 0.011), or the amount of light penetrating to ground level. Predator richness was positively associated with all predictor variables in the abundance model, apart from PAR. Parasitoid richness was positively associated with both the increasing percent cover of C3 grasses (Z=2.10, p=0.036), bare ground (Z = 2.00, p = 0.045) and most strongly by herbivore richness (Z = 3.22, p = 0.001). Lastly, pollinator richness showed no significant positive or negative relationships to any predictor variable – similar numbers of families were trapped in all parts of the prairie.

Even though prairie grasslands sampled ranged from age 1 to 10 years and size from

0.485 to 23.37 acres, I found no strong effect of age or size on arthropod abundance and richness for any functional group (see Table 6). On the contrary, predator abundance and predator richness showed a marginally significant response to prairie age (abundance: F=3.75, p=0.054; richness: F=3.63, p=0.074). Moreover, herbivore abundance and richness showed a marginally significant response to prairie size (abundance: F=3.34, p=0.086; richness: F=3.90, p=0.060).

26

3.3 CA1: Arthropod community composition among cover types

The NMDS analysis indicated that dissimilarity in arthropod community composition was related to the cover type within a farm (k = 3, stress = 0.16, R² = 0.13, p = 0.001, Fig. 5). As a rule of thumb, values of stress <0.2 are traditionally used to indicate a good representation of data in multi-dimensional space (Kenkel & Orloci, 1993). Non-overlapping ellipses (represented by 95 % CI) indicate that the community composition is different between crop, prairie and forest cover. However, the closer distance between crop cover and prairie cover indicate that they are more similar in terms of arthropod communities than forest or crop cover, presumably due to plant structure or resources.

3.4 CA2: Assessment of arthropod-derived ecosystem service mediated by habitat

While average leaf defoliation was lower on the seven farms which had restored prairie adjacent to crop fields as compared to the three farms without restored prairie (10.8 % ± 1.5 % and 7.8 % ± 1.3 %, respectively), the difference was not significant (t = 1.53, p = 0.133, Fig. 6).

However, prairie presence influenced herbivore abundance (t = -2.23, p = 0.03) and pollinator abundance (t = -2.39, p = 0.02) in the crop field, but had no effect on other functional groups or arthropod richness (Fig. 7). Relatedly, prairie presence resulted in distance-decay effects on herbivore distribution (χ2 = 4.52, p = 0.033) and pollinator distribution (X2 = 3.82, p = 0.050), where abundance was greater near prairie borders and decreased further into the crop interior

(Fig. 8). Even though herbivore abundance was more concentrated near prairie borders, crop leaf damage was not significantly different from crop edge or interior plots sampled (X2 = 0.549, p =

0.458, Fig. 9).

27

3.5 CA3: Crop type differences for arthropod abundance and richness

Crop type had a significant effect on total arthropod abundance (Kruskal-Wallis X2 =

22.54, p < 0.001) and total arthropod richness (Kruskal-Wallis X2 = 32.97, p < 0.001). All focal functional groups (herbivore, predator, parasitoid, pollinator) had the greatest abundance and richness in the mixed organic crop type (Fig. 10A, Fig. 10B). This response was due to the disproportionately large number of herbivores caught which accounted for 50 % of the abundance caught and 32% of the richness caught. In other crop types herbivore abundance accounted for 16-38 % of the total arthropod abundance and 22-39 % of the richness. In terms of percent total caught, the apple orchard had the greatest percentage of beneficial group abundance, with 6.15 % parasitoid, 21.54 % pollinator and 36.92 % predator. Similarly, apple orchards had the greatest percentage of beneficial group richness, with 7.32 % parasitoid, 24.39

% pollinator and 29.27 % predator. In fact, apple orchards had the least percentage of herbivore abundance (22 %) and herbivore richness (16 %).

3.6 CA4: Regional family pool comparison

I found that the number of arthropod families on the farms I sampled (183) was similar to the number of families present regionally on non-farm sites (193) (Fig. 11). Nevertheless, this broad assessment between the family pool of arthropods in each site did show some interesting patterns. My results show that agricultural sites had more families of herbivores, more families of parasitoid wasps, less families of predators, and similar number of families of pollinators when compared to non-farm (iNaturalist) sites (Fig. 11). A total of 89 families were shared between non-farm (iNaturalist) sites and agricultural sites, while 94 families were found unique

28

to agricultural sites and 104 families were found unique to non-farm (iNaturalist) sites (Fig. 12; see Appendix 5: Table S9 for details of families).

There were key differences in which functional groups were absent from non-farm

(iNaturalist) sites and agricultural sites. Non-farm (iNaturalist) areas over-emphasized groups such as aquatic arthropods and under-emphasized small-bodied arthropods such as parasitic wasps. A total of 27 aquatic arthropod families were found in non-farm sites with only 5 aquatic families found in agricultural sites (Fig. 11). Of these 27 families found in non-farm sites, there were 25 uniquely found in non-farm sites and one family shared between the two sites

(Hydropsychidae (order: Trichoptera)). Only 8 parasitoid families were found in non-farm sites compared to 24 in farm sites, with 6 families shared between the two sites (Cynipidae,

Eulophidae, , , , (all order:

Hympenoptera)). Not only did farm sites have a greater number of parasitoid families, but also a greater number of herbivore families (Fig. 12). A total of 52 herbivore families were found in farm sites, and of this total, 27 were unique to farm sites.

4. Discussion

My study presents a detailed analysis of arthropod abundance and richness in an agricultural landscape of Southern Ontario. I tested two scales of habitat: landscape based on the isolation of suitable habitat and local based on farm cover type (crop, prairie, forest). I also examined what factors of prairie made this cover so important, testing overall plant species composition, and more specific resource-based factors of tissue quantity (standing biomass) and tissue quality (e.g., % N in foliage). Additionally, I tested how cover type affects community

29

composition, whether the addition of diverse prairie habitat affects crop damage (observable leaf defoliation), how plant crop type affects arthropod communities, and whether regional differences in family richness between farm and non-farm areas were apparent. These tests were derived from intensively surveying 22 farms across a 10,000 km2 study area, sampling 27,080 arthropod individuals from 223 families. My key finding is that local habitat cover type, rather than landscape-level distance-based factors, dictated arthropod abundance and richness in agricultural landscapes. More specifically, it was the presence of high-quality resource-rich local prairie that explained most of my findings, even though this habitat sometimes covered relatively small percentages of my farms. I found that plant composition mattered most, especially the percent cover of forbs and C3 grasses. In terms of potential service benefits against leaf defoliation, I found no positive or negative effects. Prairie cover had a significant positive impact for arthropod biodiversity including supporting larger numbers of herbivorous insects, many of which were crop pests. However, these large herbivore populations did not increase leaf defoliation in adjacent fields, possibly because beneficial predators were also highly abundant in restored prairies. Indeed, the richness and abundance of predators and parasitoids were strongly positively correlated with these same measures for herbivores. Lastly, I found that the family- level richness of arthropods in agricultural sites versus non-farm sites were generally similar, suggesting that even conventional farms can support arthropod biodiversity especially if high- quality habitat is present.

My landscape scale analysis largely suggested that “everything is everywhere”, in terms of family-level arthropod occurrences among farms. The only exception was parasitoid richness, which decreased as percent agricultural landcover increased. This result was only found at the lowest buffer radii of 500 m and is not surprising given that this arthropod group is heavily 30

dependent on woody nesting material and host density which can be limited at this small spatial scale (Chaplin-Kramer et al., 2011; Inclan, Cerretti, & Marini, 2015). The lack of an effect of landscape for other arthropod groups may be surprising, given that 92% of my study region supported intensive farming. One possible explanation could be some form of “ debt”, where remnant populations are ‘legacy populations’ from a time when these areas supported far more suitable habitat. If true, then the large numbers of arthropod groups that I detected could be doomed to disappear. This hypothesis is impossible to test with my data. However, given the vast levels of habitat loss in this system combined with the short generation times of arthropods – non-viable species would be likely to disappear relatively quickly. It might not have been surprising that I would have observed my farms to be relatively depauperate, based on increasing predictions of an “insect apocalypse”, especially on agricultural landscapes (Sánchez-Bayo &

Wyckhuys, 2019). Indeed, land clearing in my region started as far back as the late 19th century

(McCune et al. 2017), with habitat loss to crop fields even higher than today. However, it seems difficult to classify my farms as “depauperate”, given the large and diverse collection of arthropod families that I observed with 183 in agricultural sites and 193 families in the non-farm surrounding region (iNaturalist data).

Despite finding a similar number of pooled arthropod families on agricultural and non- farm sites, I found differences in functional group composition, family-specific feeding mode

(generalist vs. specialist), and tolerance to agricultural areas. I found farm sites to have a greater number of herbivore and parasitoid families, several which were unique to agricultural sites and considered having specialist feeding modes. For example, the family Agromyzidae is a soybean stem borer (Arnemann et al., 2016), Thripidae contains species of soybean and corn pests

(Godfrey et al., 2019), tend to be host-specific (Ouvrard, 2013), and many families of 31

parasitoid wasps are regarded as specialists (Gordh, Legner & Caltagirone, 1999). On one hand, this was not surprising given that farm sites are likely to be more easily colonized by crop- associated specialist arthropods (Jarvis, Padoch & Cooper, 2007). On the other hand, it was surprising to find a greater number of parasitoid families occurring in agricultural sites given my previous finding that parasitoid richness decreased with increasing agricultural landcover at the

500 m scale. It is important to note that for my [CA4] analysis I pooled arthropod families between agricultural sites and therefore effects of surrounding landcover or farm-specific differences could not be considered. Regardless of this fact, my iNaturalist comparisons do show that certain arthropod functional groups are absent from agricultural sites – those that may be less “tolerant” and with life history strategies less equipped to establish and persist in these areas.

This includes families associated with aquatic habitat, which are often not found near open farm areas (e.g., Baetidae, Caenidae, Ephemerellidae (all order: Ephemeroptera)), and groups such as butterflies (e.g., Geometridae, Hesperiidae, Noctuidae (order: )). Both butterflies and aquatic insects in the EPT community (orders: Ephemeroptera, Plecoptera and Trichoptera) are considered important environmental indicators for their intolerance to pollutants (Hamid &

Rawi, 2017) and rapid responses to loss of suitable habitat (Fleishman & Murphy, 2009).

As stated, the number of arthropod families on my farms was comparable to the non-farm areas. An alternative explanation to extinction debt is that my farms actually contain sufficient levels of habitat, that are adequately connected regionally such that many arthropod groups are being maintained. Indeed, an increasing occurrence of semi-natural habitats in farmland is likely to benefit biodiversity in all but the most heavily cleared landscapes (Tscharntke et al., 2004).

My agricultural sites were unique in that many possessed restored resource-rich tall-grass prairie which may be supporting arthropod populations while serving as stepping stones that increase 32

dispersal among farms, within and possibly across generations. Again, this does not preclude the possibility that the less tolerant arthropod communities have already been displaced regionally, with the remaining subset being those that ‘can make a living’ within heavily farmed regions.

The population viability and dispersal dynamics of the families I detected remain to be determined – my data are but a one-time snapshot of richness and abundance. However, my findings suggest that many arthropod groups appear to be persisting through time and, given the high degree of overlap among farms, may also be dispersing to some degree.

One potentially powerful factor thought to regulate arthropod communities in agricultural landscapes is management including pesticides, whose effects were not directly tested by this study (e.g., fully comparing conventional vs organic farms, comparing farm-level insect diversity before and after spraying). However, I did make two relevant observations. First, all the monoculture crop fields of corn, soy, and wheat were generally lacking large numbers of arthropods. These results are consistent with other studies on farm management practices, showing overall negative influences on arthropod abundance and richness (Weibull & Ostman,

2003; Ma et al. 2019). Second, although I only surveyed one organic farm, it supported the most arthropods of any non-prairie cover type, mostly via large populations of herbivores. It is impossible to know the exact cause of these patterns with my data, but this response could potentially be due to chemical spraying. This would have to be from the year prior (chemical residue) as no chemicals were applied before or during my sampling, or through dispersal from adjoining crop fields (chemical drift). Alternatively, this large population response could reflect the presence of high-quality habitat within the organic fields not present in other conventional managed cultivated crops. It has been demonstrated that organic cropping systems generally support greater levels of biodiversity, although this does not often translate into better provision 33

of ecosystem services because herbivores can be abundant (as I observed) and crop damage high

(Macfadyen et al., 2009). A clearer understanding of how farm management regulates arthropods, especially regarding issues of habitat loss versus chemical spraying and their impacts on crop damage remains an important research direction in agriculture (Larsen, Patton &

Martin, 2019).

Regardless of the causal factors driving low arthropod occurrences in crop fields, my work clearly shows the significant positive responses of arthropods to the construction of resource-rich prairie habitat on conventional farms, even if these areas occupy a relatively small percentage of the total farm area. These prairies had the highest levels of both richness and abundance, including many families of beneficial predators and parasitoids that prey on herbivores and crop pests. This contrasts with cultivated crop habitats, which contained a low abundance and richness of beneficial arthropods (predators, parasitoids, pollinators), a response that has been shown in other studies (Pfiffner & Luka, 2003; Koh & Holland, 2014).

There were interesting differences as well, between restored prairie versus forests, with the latter of interest given the recent work from European farm landscapes (Hallmann et al.,

2017). I found that prairies contained two times more abundance of arthropods than forest. My complementary analysis [CA1] showed that, in addition to differences in abundance, these two cover types had different arthropod community compositions, presumably related to vegetation structure of forest vs open herbaceous grassland (Prevedello &Vieira, 2010). For example, the permanent shade and high levels of nesting habitat in forest tends to support arthropod families that are less common or even absent in the more open habitats of prairie and crop fields, where desiccation risk in summer for arthropods can be high (Heinrich, 1996).

34

The addition of prairie as a strategy to combat arthropod biodiversity loss has not been widely studied, other than for pollinators groups (e.g., Hopwood, 2008; Hormon-Threatt &

Hendric, 2014; Paterson, 2014; Tonietto, Ascher & Larkin, 2017; Paterson et al., 2019). Rather, the common goal of prairie addition is to re-establish this formerly widespread plant community while providing a range of ecosystem services usually centered on plant biodiversity soil carbon storage (Wang, Vandenbygaart & Mcconkey, 2014). Clearly, however, my work shows that these areas can support large arthropod communities. One additional question I sought to test was why this occurs, focusing on four resource-based mechanisms that may attract arthropods to prairie: the high standing biomass of restored prairie (tissue quantity), the potential for prairie plants to provide high quality food resources especially for herbivores (tissue quality), factors relating to habitat structure, and micro-environment such as PAR, bare ground, and litter; or whether arthropods respond more to the presence of certain plant functional groups (plant composition). My results clearly support the latter, while also suggesting an influence of habitat structure. Generally, as the percentage of C3 grasses increased, so did the abundance of functional groups herbivore, predator and pollinators. An increasing percentage of forb cover increased the total arthropod community, as well herbivore and predator groups, but surprisingly was not influential for pollinator or parasitoid groups. Previous research on the assemblage of parasitoids and pollinators has shown that these groups are more sensitive to species identity within plant functional groups such as the presence of different species of forbs (Thomson et al.,

2010; Perdikis et al., 2011). The plant species that occurred within my prairie sites were most often generalists with a variety of nectar and pollen resources, not tailored to a specific functional group such as pollinators or parasitoids. Further investigation is needed on nectar and

35

pollen quality within prairies and how certain functional groups such as native pollinators may respond.

For habitat structure, there was a positive effect of PAR (light) on increasing abundance of the total arthropod community, herbivores and predators. Additionally, my predictor variable bare ground showed significant effects for beneficial functional groups. I found that an increasing bare ground percent cover increased predator abundance, but decreased pollinator abundance, and increased richness of predator and parasitoid groups. Higher PAR likely results in changes to microclimate conditions, as more light can reach the soil surfaces. This contrast in abundance responses between predators and pollinators could be attributed to the fact that pollinators (e.g., ground nesters) may not prefer nesting sites exposed to full sunlight or predator visibility. While predator groups could prefer an increasing amount of light at soil surface to improve prey searching ability (Letourneau et al., 2009). Overall, these results suggest that PAR and bare ground cover should be considered when monitoring prairie restoration efforts to target the needs of beneficial functional groups.

Relatedly, I found strong indirect influences of plants on my arthropod groups.

Specifically, plant composition had strong effects on herbivores that, in turn, increased predatory richness and abundance. These so-called “trophic dependencies” where prey affects predators via plants has been described before (Harvey & MacDougall, 2014). Alternative prey via plants can be important in attracting and retaining beneficial predatory groups which can sustain biological control in crop fields. Previous work by Gravesen and Toft (1987) showed that grasslands provided alternative prey to predators which increased populations to the point that populations were reduced in the adjacent field crops. Previous work has also shown that about

36

50% of herbivore species living in prairie grassland habitats were monophagous or oligophagous

(live only on a single host plant or a non-crop plant) and thus are not at risk of damaging cultivated plants but can serve as alternative prey for predator and parasitoids (Lethmayer, 1995;

Hausammann, 1996).

Despite this, an important consideration to growers is the implications the addition of semi-natural cover (e.g., prairie) has on crop fields. My work revealed two important effects prairie cover has on adjacent crop fields: (1) the distribution of functional groups was altered due to prairie grassland borders and (2) farms with prairie grasslands planted adjacent to crop fields did not incur greater crop damage despite more on-farm herbivore abundance. Results showed that farms with prairie present had the highest abundance of all focal functional groups

(herbivores, predators, parasitoids, pollinators) at the edges of crop fields closest to the prairie border. Farms without prairie present had the same abundance of all arthropod groups, including herbivores, at the edges and interior of crops. This arthropod movement between cultivated to uncultivated habitat can be related to natural dispersal of arthropods, lack of suitable food, host alteration, or major disturbance such as the use of pesticides (Gurr et al., 2017). Boetzl et al.

(2019) demonstrated that the proportion of multiple ground-dwelling predators declined from the field edge and this was attributed to the fact that semi-natural vegetation strips, which provided more suitable habitat and food resources, altered arthropod dispersal. Therefore, a strategic placement of prairie cover adjacent to crop fields has the potential to reduce herbivore spill-over into crop fields. This link between prairie cover and reduced herbivore distribution in the crop field with no increase in crop leaf defoliation should encourage farmers to include more habitat refuge areas.

37

Despite the striking results of my study, it is worth considering two limitations. First, my one-year arthropod sampling misses known temporal variation. Populations of arthropods can fluctuate strongly from year to year (Andrewartha & , 1954; Stange et al., 2011) and species turnover in habitat patches is high (Summerville et al., 2007). This highlights the importance of long-term monitoring (Haaland et al., 2011) versus the single-season work of this study. Indeed, my work has served as the basis for a long-term monitoring program set up by

BIO on my ALUS farms, using barcoding to sample arthropods every two weeks over multiple years (Fryxell, Betini, personal communication). This should provide a clearer view on temporal changes in arthropods, whether it be inter-annual climate differences, differing levels of prairie availability (larger prairie cover could provide temporal stability) or the impacts of the timing of pesticide applications vs the timing of insect phenology within and among years. Second, my arthropod field sampling methods have limitations to what functional groups I could collect. Pan traps and sweep netting occurred during the day, therefore omitting collection of important nocturnal arthropods such as robber or moths (Wickramasinghe et al., 2004). Future sampling should incorporate pitfall traps or malaise traps that will enable this nocturnal arthropod capture. Establishing a sampling protocol that is cost-effective and adequate in collecting whole arthropod communities in agricultural landscapes would be highly beneficial for long-term monitoring.

Regardless of these limitations, my study represents one of the most comprehensive arthropod surveys conducted in an agricultural landscape in Southern Ontario. From my results, I provide the following recommendations: (i) land managers should acknowledge that “semi- natural” areas are not all equal. For example, prairie grasslands produced different community composition and produced two-times more arthropod abundance than forest covers, indicating it 38

may be a faster management strategy to foster arthropod biodiversity, (ii) prairie monitoring efforts should consider the influence of plant composition, light penetration (related to plant structure), and bare ground cover, which all showed strong effects on beneficial arthropod groups (predators, parasitoids, pollinators), and (iii) prairie grasslands should be planted adjacent to crop fields to reduce herbivore abundance in crop fields, but farmers should refrain from spraying near crop edges to reduce detrimental effects to beneficial groups.

5. Conclusion

Overall, my findings revealed that distance-based constraints at the landscape scale were not detected, but rather on-farm cover type prairie grassland was the driving factor for arthropod abundance and richness. Local habitat addition on conventional farms is much more important than previously thought in conserving biodiversity and requires a management perspective and policy regulations to match.

With 40% of the world’s total land area dedicated to agriculture (Matson et al., 1997;

Ramankutty & Foley, 1999) agroecosystems have the potential to conserve a large proportion of arthropod biodiversity if high-quality habitat is present. Understanding how habitat-based factors positively influence agriculturally important arthropod functional groups will be essential to conserve arthropod-derived ecosystem services (Tilman et al., 2002; Foley et al., 2005; Scherr &

McNeely, 2008). Future research should focus on quantifying arthropod-derived ecosystem services provided by habitat (i.e, pollination, nutrient recycling, decomposition) which will provide better incentive for farmers and leverage policy to include more refuge areas on farms.

39

6. References

AAFC. (2017). Annual Crop Inventory 2017. [Data file]. Retrieved from:

https://open.canada.ca/data/en/dataset/ba2645d5-4458-414d-b196-6303ac06c1c9

AgMaps. (2017). Agricultural Information Atlas. [Geographic Information Portal] Retrieved

from: http://www.omafra.gov.on.ca/english/landuse/gis/portal.htm

AgriService BC. (2017). Spotted Wing Drosophila (Fruit ) pest alert. Retrieved from:

www2.gov.bc.ca/gov/content/industry/agriculture-seafood/-and-crops/plant-

health/insects-and-plant-diseases/tree-fruits/spotted-wing-drosophila

Albrecht, M., Duelli, P., Schmid, B., & Müller, C.B. (2007). Interaction diversity within

quantified arthropod food webs in restored and adjacent intensively managed meadows.

Journal of Ecology, 76(5), 1015–1025.

Alford, D.V. (2016). Pests of Fruit Crops: A colour Handbook, Second Edition. CRC Press.

Altieri, M.A. (1999). The ecological role of biodiversity in agroecosystems. Agriculture

Ecosystems & Environment, 74(1), 19–31.

ALUS Canada. (2019). What we do. Retrieved from: https://alus.ca/what-we-do/

Ambrose, J.D., Kevan P.C., & Gadawski, R.M. (1985). Hop tree (Ptelea trifoliata) in Canada:

Population and reproductive biology of a rare species. Canadian Journal of Botany,

63(11), 1928-1935.

40

Andrewartha, H.G., & Birch, L.C. (1954). The distribution and abundance of animals.

University of Chicago Press, Chicago, Illinois, USA.

Arnemann, J., Tay, W., Walsh, T., Brier, H., Gordon, K., Hickmann, F., . . . Guedes, J. (2016).

Soybean Stem Fly, Melanagromyza sojae (Diptera: Agromyzidae), in the New World:

Detection of high genetic diversity from soybean fields in Brazil. Genetics and Molecular

Research: GMR, 15(2), 1-13.

Arnett, R.H. (2000). False blister (Insecta: Coleoptera: Oedemeridae). Retrieved from:

https://edis.ifas.ufl.edu/pdffiles/IN/IN31100.pdf

Arnett, R.H., Thomas, M. C., Skelley, P.E., & Frank, J.H. (eds.). (2002). American Beetles,

Volume II: : Scarabaeoidea through Curculionoidea. CRC Press LLC, Boca

Raton, Florida.

Arnett, R. (2000). American insects: A handbook of the insects of America north of Mexico (2nd

ed.). Boca Raton, Fla.: CRC Press.

Ball, D.F. (1964). Loss-an-ignition as an estimate of organic matter and organic carbon in non-

calcareous soils. Journal of Soil Science, 15(1), 84-92.

Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models

using lme4. Journal of Statistical Software, 67(1), 1-48.

Beaman, R.S., Decker, P.J., & Beaman, J.H. (1988). Pollination of Rajjlesia (Rafflesiaceae).

American Journal of Botany, 75(8), 1148-1162.

Benton, T., Bryant, D., & Cole, L. (2002). Linking agricultural practice to insect and bird 41

populations: historical study over three decades. Journal of Applied Ecology, 39(4), 673-

687.

Benton, T.G., Vickery, J.A., & Wilson, J. D. (2003). Farmland biodiversity: Is habitat

heterogeneity the key? Trends in Ecology and Evolution, 18(4), 182–188.

Bianchi, F.J.J.A., Booij, C.J.H., & Tscharntke, T. (2006). Sustainable pest regulation in

agricultural landscapes: A review on landscape composition, biodiversity and natural pest

control. Proceedings of the Royal Society, 273(1595), 1715–1727.

Boetzl, F., Krimmer, E., Krauss, J., Steffan-Dewenter, I., & Lewis, O. (2019). Agri-

environmental schemes promote ground-dwelling predators in adjacent oilseed rape

fields: Diversity, species traits and distance-decay functions. Journal of Applied

Ecology, 56(1), 10-20.

Bradley, R., Buchanan, S., & American Arachnological Society. (2013). Common spiders of

North America. Berkeley: University of California Press.

Bray, R.J., & Curtis, J.I. (1957). An ordination of upland forest communities of southern

Wisconsin. Ecological Monographs, 27(4), 325-349.

Burgin, S.G., & Hunter, F.F. (1997). Nectar versus honeydew as sources of sugar for male and

female black flies (Diptera: Simuliidae). Journal of Medical Entomology, 34(6), 605-608.

Bystrak, P.C., & Wirth, W.W. (1978). The North American species of Forcipomyia, subgenus

Euprojoannisia (Diptera: Ceratopogonidae). Retrieved from:

https://pdfs.semanticscholar.org/d1ee/c4b688abc19fe564d122abbb85df2cc1863a.pdf

42

Cardoso, P., Silva, I., Oliveira, N.G., & Serrano, A.R.M. (2004) Higher taxa surrogates of spider

(Araneae) diversity and their efficiency in conservation. Biological Conservation, 117(4),

453-459.

Chad, A. (2006). Stick insects of the continental United States and Canada. Coachwhip

Publications.

Chaplin-Kramer, R., O’Rourke, M.E., Blitzer, E.J.M., & Kremen, C. (2011). A meta-analysis of

crop pest and natural enemy response to landscape complexity. Ecology Letters, 14(9),

922–932.

Cheung, D.K.B., Marshall, S.A., & Webb, D.W. (2006). of Ontario. Canadian Journal

of Arthropod Identification, 1(1), 1-13.

Crocker, R.L., & Whitcomb, W.H. (1980). Feeding niches of the big-eyed bugs Geocoris

bullatus, G. punctipes, and G. uliginosus (Hemiptera: : Geocorinae).

Environmental Entomology, 9(5), 508-513.

Dailey, T.B., & Scott, P.E. (2006). Spring nectar sources for solitary bees and flies in a landscape

of forest and agricultural fields: production, variability, and consumption. The

Journal of the Torrey Botanical Society, 133(4), 535-548.

D'Arcy-Burt, S., & Blackshaw, R.P. (1991). Bibionids (Diptera: Bibionidae) in agricultural land:

A review of damage, benefits, natural enemies and control. Annals of Applied Biology,

118(3), 695-708.

43

Davies, D.M., & Petersen, B.V. (1956). Observations on the mating, feeding, ovarian

development, and oviposition of adult black flies (Simuliidae, Diptera). Canadian Journal

of Zoology, 34(6), 615-655.

De Clerck-Floate, R., & Cárcamo, H. (2011). Biocontrol arthropods: New denizens of Canada’s

grassland agroecosystems. In K.D. Floate (Eds), Arthropods of Canadian grasslands

(Volume 2): Inhabitants of a Changing Landscape (pp. 291-321). Ottawa, Ontario.

Agriculture Canada.

Deguines, N., Jono, C., Baude, M., Henry, M., Julliard, R., & Fontaine, C. (2014). Large-scale

trade-off between agricultural intensification and crop pollination services. Frontier in

Ecology and the Environment, 12(4), 212–217.

DeWalt, R.E., Maehr, M.D., Neu-Becker, U., & Stueber, G. (2013). Plecoptera species file

online. Version 5.0. Retrieved from: http://plecoptera.speciesFile.org.

Dillon, E. (1961). A manual of common beetles of eastern North America. Dover Publications.

Dirzo, R., Young, H., Galetti, M., Ceballos, G., Isaac, N., & Collen, B. (2014). Defaunation in the Anthropocene. Science, 345(6195), 401-406.

Disney, R.H.L. (1983). Scuttle flies (Diptera: Phoridae) in the British Isles. Retrieved from:

http://people.ds.cam.ac.uk/mw245/Disney,%201983a.pdf

Donald, E.B. (1987). The metallic wood-boring beetles of Canada and Alaska: Coleoptera:

Buprestidae (The Insects and arachnids of Canada) (Publication 1810). Ottawa, Ontario.

Agriculture Canada.

44

Dondale, C.D., & Redner, J.H. (1978). The insects and arachnids of Canada, Part 5. The crab

spiders of Canada and Alaska, Araneae: Philodromidae and Thomisidae (Publication

1663). Ottawa, Ontario. Agriculture Canada.

Dondale, C.D., & Redner, J.H. (1982). The insects and arachnids of Canada, Part 9. The sac

spiders of Canada and Alaska, Araneae: Clubionidae and Anyphaenidae (Pulication

1724). Ottawa, Ontario. Agriculture Canada.

Dondale, C.D., & Redner, J.H. (1990). The insects and arachnids of Canada, Part 17. The wolf

spiders, nurseryweb spiders, and lynx spiders of Canada and Alaska, Araneae:

Lycosidae, Pisauridae, and Oxyopidae (Publication 1856). Ottawa, Ontario. Agriculture

Canada.

Dondale, C.D., Redner, J.H., Paquin, P., & Levi, H.W. (2003). The insects and arachnids of

Canada. Part 23. The orb-weaving spiders of Canada and Alaska (Araneae: Uloboridae,

Tetragnathidae, Araneidae, Theridiosomatidae). Ottawa: NRC Research Press.

Downes, J.A., & Smith, S.M. (1969). New or little-known feeding habits in Empididae (Diptera).

The Canadian Entomologist, 101(4), 404-408.

Downie, N.M., & Arnett, R.H. (1996). The Beetles of Northeastern North America. The Sandhill

Crane Press, Gainesville, Florida.

Duelli, P., & Obrist, M.K. (2003). Regional biodiversity in an agricultural landscape: the

contribution of seminatural habitat islands. Basic and applied ecology, 4(2), 129-138.

45

Duflot, R., Aviron, S., Ernoult, A., Fahrig, L., & Burel, F. (2015). Reconsidering the role of

‘semi-natural habitat’ in agricultural landscape biodiversity: A case study. Ecological

Research, 30(1), 75-83.

Dunford, J.C., & Somma, L.A. (2008). Encyclopedia of Entomology: Scorpionflies (Mecoptera).

Springer Science & Business Media.

Eagalle, T., & Smith, M. (2017). Diversity of parasitoid and parasitic wasps across a latitudinal

gradient: Using public DNA records to work within a taxonomic

impediment. FACETS, 2(2), 937-954.

Environmental Systems Research Institute (ESRI). (2014) ArcGIS Release 10.3.1. Redlands,

California, USA.

Evans, H.E. (1978). The of America north of Mexico (Memoirs of the American

Entomological Institute). American Entomological Institute.

Fick, S.E., & Hijmans, R.J. (2017). Worldclim 2: New 1-km spatial resolution climate surfaces

for global land areas. International Journal of Climatology. 37(12), 4302-4315.

Fitzgerald, S.J. (2005). Evolution and classification of Bibionidae (Diptera: Bibionomorpha).

(Ph.D. Thesis). Oregon State University.

Fleishman, E., & Murphy, D. (2009). A realistic assessment of the indicator potential of

butterflies and other charismatic taxonomic groups. Conservation Biology, 23(5), 1109-

1116.

46

Foelix, R.F. (2011). Biology of spiders, Third Edition. Oxford University Press, New York.

Ford E.J., & Jackman, J.A. (1996) New larval host plant associations of tumbling flower beetles

(Coleoptera: Mordellidae) in North America. The Coleopterists Bulletin, 50(4), 361-368.

Fox, J., & Weisberg, S. (2011). An R Companion to Applied Regression, Second Edition.

Thousand CA: Sage.

Funderburk, J.E., Pedigo, L.P., & Berry E.C. (1983). Seedcorn maggot (Diptera: Anthomyiidae)

emergence in conventional and reduced- tillage soybean systems in Iowa. Journal of

Economic Entomology, 76(1), 131-134.

Geiger, F., Bengtsson, J., Berendse, F., Weisser, W.W., Emmerson, M., Morales, M.B.. . .

Inchausti, P. (2010). Persistent negative effects of pesticides on biodiversity and

biological control potential on European farmland. Basic and Applied Ecology, 11(2), 97-

105.

Gibson A.P. (1993). Superfamilies and Chalcidoidea (pp. 570-655). In

Goulet, H. & Huber, J. (eds). Hymenoptera of the World: an identification guide to

families. Ottawa, Canada. Agriculture Canada.

Gimmel, M. (2013). Genus-level revision of the family Phalacridae (Coleoptera: Cucujoidea).

Zootaxa, 3605(1), 1–147.

Godfrey, L.D., Wright, S.D., Summers, C.G. & Frate, C.A. (2019). UC IPM Pest Management

Guidelines: Corn (Publication 3443). Retrieved from:

http://ipm.ucanr.edu/PMG/r113300711.html

47

Gordh, G., Legner, E.F., & Caltagirone, L.E. (1999). Biology of Parasitic Hymenoptera-Chapter

15. In T.S. Bellows, & T.W. Fisher (Eds), Handbook of Biological Control (pp. 355-381).

Academic Press.

Goulet, H., & Huber, J.Y. (1993). Hymenoptera of the world: an identification guide to families.

Ottawa, Ontario. Agriculture Canada.

Goulson, D., & Ebrary, Inc. (2010). : Behaviour, ecology, and conservation (Oxford

biology). Oxford; New York: Oxford University Press.

Government of Canada. (2017). Annual Crop Inventory 2017. [Data file]. Retrieved from:

https://open.canada.ca/data/en/dataset/ba2645d5-4458-414d-b196-6303ac06c1c9

Government of Canada. (2019). Canadian climate normals 1981-2010 station data. [Data file]

Retrieved from: http://climate.weather.gc.ca/climate_normals/results_1981_2010_e.html?

searchType=stnProv&lstProvince=ON&txtCentralLatMin=0&txtCentralLatSec=0&txtCe

ntralLongMin=0&txtCentralLongSec=0&stnID=4624&dispBack=0

Goyal, G., Nuessly, G., Seal, D., Steck, G., Capinera, J., & Meagher, R. (2012). Examination of

the pest status of corn-infesting Ulidiidae (Dipera). Environmental Entomology, 41(5),

1131-1138.

Grainger, T.N., & Gilbert, B. (2017). Multi-scale responses to warming in an experimental insect

metacommunity. Global Change Biology, 23(12), 5151-5163.

Graveson, E., & Toft, S. (1987). Grass fields as reservoirs for polyphagous predators

(Arthropoda) of (Homoptera: ). Journal of Applied Entomology, 104(5),

48

461-473.

Grissell, E.E., & Schauff, M. E. (1997). A handbook of the families of Nearctic Chalcidoidea

(Hymenoptera). 2nd Edition. The Entomological Society of Washington.

Gullan, P.J., & Cranston, P.S. (2014). Chapter 11: Pollination. In Insects: An outline of

entomology (Fifth ed.). Chichester, West Sussex: Wiley Blackwell.

Gurr, G., Wratten, S., Landis, D., & You, M. (2017). Habitat Management to Suppress Pest

Populations: Progress and Prospects. Annual Review of Entomology, 62(1), 91-109.

Haaland, C., Naisbit, R.E., & Bersier, L-F. (2011) Sown wildflower strips for insect

conservation: a review. Insect Conservation and Diversity, 4(1), 60–80.

Hallmann, C.A., Sorg, M., Jongejans, E., Siepel, H., Hofland, N., Schwan, H., …… De Kroon,

H. (2017). More than 75 percent decline over 27 years in total flying insect biomass in

protected areas. PLoS ONE, 12(10): E0185809.

Hamid, S., & Rawi, C. S. (2017). Application of Aquatic Insects (Ephemeroptera, Plecoptera

And Trichoptera) In Water Quality Assessment of Malaysian Headwater. Tropical life

sciences research, 28(2), 143–162.

Hanski, I., & Ovaskainen, O. (2000). The metapopulation capacity of a fragmented landscape.

Nature, 404(6779), 755-8.

Hanson, H. (1939). Ecology in Agriculture. Ecology, 20(2), 111-117.

49

Harmon-Threatt, A.N., & Hendrix, S.D. (2014). Prairie restorations and bees: the potential

ability of seed mixes to foster native bee communities. Basic and Applied Ecology, 16(1),

64–72.

Harvey, E., & MacDougall, A. S. (2015). Spatially heterogeneous perturbations homogenize the

regulation of insect herbivores. The American Naturalist, 186(5), 623-633.

Harvey, E., & MacDougall, A.S. (2014). Trophic island biogeography drives spatial divergence

of community establishment. Ecology, 95(10), 2870-2878.

Hausammann, A. (1996). Strip-management in rape crop: Is winter rape endangered by negative

impacts of sown weed strips? Journal of Applied Entomology, 120(8), 505-512.

Hawkeswood, T.J., & Turner, J.R. (2007). Record of pollination of Lomatia silaifolia (Sm.)

R.Br. (Proteaceae) by the longicorn Uracanthus triangularis (Hope, 1833)

(Coleoptera: Cerambycidae). Calodema Supplementary Paper, 53:1–3.

Heinrich, B. (1996). The Thermal Warriors: Strategies of Insect Survival. Harvard University

Press, Cambridge.

Henry, T.J. (2009). Biodiversity of the In: Foottit R.G., Adler P.H., eds. Insect

biodiversity: Science and society. Oxford: Wiley-Blackwell.

Hill, D.S. (2008). Pests of Crops in Warmer Climates and Their Control. Dordrecht: Springer

Netherlands.

Hocking, B. (1968). Insect-flower associations in the High Arctic with special reference to

nectar. Oikos, 19(2), 359-88. 50

Hoebeke, E.R., & Carter, M.E. (2003). Halyomorpha halys (Stål) (Heteroptera: ):

A polyphagous plant pest from Asia newly detected in North America. Proceedings of the

Entomological Society of Washington, 105(1), 225-237.

Holzenthal, R.W., Thomson, R.E., & Ríos-Touma, B. (2015). Order Trichoptera. In Thorp J.,

Rogers D.C. (Eds.) Ecology and general biology: Thorp and Covich's freshwater

invertebrates (pp. 965-1002). Academic Press.

Hopkin, S.P. (1997). Biology of springtails (Insecta: Collembola). Oxford University Press.

Hopwood, J.L. (2008). The contribution of roadside grassland restorations to native bee

conservation. Biological Conservation, 141(10), 2632–2640.

Horton, D.R. (2008) Minute Pirate Bugs (Hemiptera: ). Encyclopedia of

Entomology. Springer Science.

Hunt, J.H., Baker, I., & Baker, H.G. (1982). Similarity of amino acids in nectar and larval saliva:

the nutritional basis for trophallaxis in social wasps. Evolution, 36(6), 1318-1322.

Inclan, D.J., Cerretti, P., & Marini, L. (2015). Landscape composition affects parasitoid

spillover. Agriculture, Ecosystems and Environment, 208(1), 48-54.

Jackson, M., & Jewiss-Gaines, A. (2012). Field guide to the jewel beetles: (coleoptera:

Buprestidae) of Northeastern North America. Victoria, B.C: First Choice Books.

Jarvis, D., Padoch, C., & Cooper, H. (2007). Managing biodiversity in agricultural ecosystems.

New York: Columbia University Press: Biodiversity International.

51

Jaskula, R. (2013). Feeding behaviours of a predatory Calomera littoralis nemoralis

(Olivier, 1790) (Coleoptera: Cicindelidae). Journal of Entomological Research Society,

15(1), 1-6.

Johnson, N.F., & Triplehorn, C.A (2004). Borror and DeLong's introduction to the study of

insects. Brooks Cole.

Kearns, C.A. (2001). North American dipteran pollinators: assessing their value and

conservation status. Conservation Ecology, 5(1), 1-5.

Kelton, L.A. (1982). Plant bugs on fruit crops in Canada: Heteroptera: . Ottawa,

Ontario. Agriculture Canada.

Kevan, P. G., & Baker, H. G. (1983). Insects as flower visitors and pollinators. Annual Review

of Entomology, 281(1), 407-453.

Kevan, P.G. (1970). High arctic insect-flower relations: the inter-relationships of arthropods and

flowers at Lake Hazen, Ellesmere Island, Northwest Territories, Canada. (Ph.D. thesis),

University of Alberta.

Kirk, W.D.J. (1984). Ecologically selective coloured traps. Ecological Entomology, 9(1), 35–41.

Klymko, J., & Marshall, S.A. (2008). Review of the Nearctic Lonchopteridae (Diptera),

including descriptions of three new species. The Canadian Entomologist, 140(6), 649-

673.

52

Knuth, P. (1908). Handbook of flower pollination (translated by J.R. Ainsworth Davis). Oxford:

Oxford University Press.

Koenig, W., & Knops, J. (1998). Testing for spatial autocorrelation in ecological

studies. Ecography, 21(4), 423-429.

Koh, I., & Holland, J.D. (2014). Grassland plantings and landscape natural areas both influence

insect natural enemies. Agriculture, Ecosystems and Environment, 199(2015), 190–199.

Larsen, A.E., Patton, M., & Martin, E.A. (2019). High highs and low lows: Elucidating striking

seasonal variability in pesticide use and its environmental implications. Science of the

Total Environment, 651(2019), 828–837.

Larson, B.M.H., Kevan, P.G., & Inouye, D.W. (2001). Flies and flowers: taxonomic diversity of

anthophiles and pollinators. The Canadian Entomologist, 133(4), 439–465.

Lawton, J.H. (1983) Plant architecture and the diversity of phytophagous insects. Annual Review

of Entomology, 281(1), 23-39.

Leong, J.M., & Thorp, R.W. (1999). Colour-coded sampling: the pan colour preferences of

oligolectic and nonoligolectic bees associated with a vernal scale plant. Ecological

Entomology, 24(3), 329-335.

Lerdau, M.T., & Throop, H.L. (2004). Effects of nitrogen deposition on insect herbivory:

Implications for community and ecosystem processes. Ecosystems, 7(2), 109–133.

Lethmayer, C. (1995). Effects of sown weed strips on pest insects. (Ph.D. thesis). University of

Bern Switzerland.

53

Letourneau, D.K., Jedlicka, J.A., Bothwell, S.G., & Moreno, C.R. (2009). Effects of natural

enemy biodiversity on the suppression of arthropod herbivores in terrestrial

ecosystems. Annual Review of Ecology, Evolution, and Systematics, 40, 573-592.

Losey, J.E., & Vaughan, M. (2006). The economic value of ecological services provided by

insects. Bioscience, 56(4), 331–323.

Ma, A., Lu, X., Gray, C., Raybould, A., Tamaddoni-Nezhad, A., Woodward, G., & Bohan, D.

(2019). Ecological networks reveal resilience of agro-ecosystems to changes in farming

management. Nature Ecology & Evolution, 3(2), 260-264.

Macfadyen, S., Gibson, R., Polaszek, A., Morris, R., Craze, P., Planqué, R., . . . Memmott, J.

(2009). Do differences in food web structure between organic and conventional farms

affect the ecosystem service of pest control? Ecology Letters, 12(3), 229-238.

Macfadyen, S., Gibson, R., Polaszek, A., Morris, R., Craze, P., Planqué, R., ….Memmott, J.

(2009). Do differences in food web structure between organic and conventional farms

affect the ecosystem service of pest control? Ecology Letters, 12(3), 229-238.

Mader, E., Shepherd, M., Vaughan, M., Black, S. H., & LeBuhn, G. (2011). Attracting Native

Pollinators. The Xerces Society. Storey Publishing.

Magurran, A.E., Baillie, S.R., Buckland, S.T., Dick, J., Elston, D.A., Scott, E.M. . . . Watt, A.D.

(2010). Long-term datasets in biodiversity research and monitoring: Assessing change in

ecological communities through time. Trends in Ecology & Evolution, 25(10), 574-582.

54

Marshall, S.A. (2006). Insects: Their Natural History and Diversity: With a Photographic Guide

to Insects of Eastern North America. Firefly Books Ltd.

Masner, L. (1993). Superfamily (pp. 566-569). In Goulet, H. & Huber, J. (eds).

Hymenoptera of the World: an identification guide to families. Ottawa, Ontario.

Agriculture Canada.

Masner, L. (1993). Superfamily (pp. 537-557). In Goulet, H. & Huber, J. (eds).

Hymenoptera of the World: an identification guide to families. Ottawa, Canada.

Agriculture Canada.

Matesco, V.C., & Grazia, J. (2015) Negro Bugs (). In: Panizzi A., Grazia J. (eds)

True Bugs (Heteroptera) of the Neotropics. Entomology in Focus, vol 2. Springer,

Dordrecht.

Mathis, W.N., & Rung, A. (2011). World Catalog and Conspectus on the Family Dryomyzidae

(Diptera: Schizophora). Retrieved from: https://repository.si.edu/bitstream/handle/10088/

18926/ent_MYIA12_Dryomyzidae.pdf?sequence=1&isAllowed=y

Matson, P.A., Parton, W.J., Power, A.G., & Swift, M.J. (1997). Agricultural intensification and

ecosystem properties. Science, 277(5325), 504–509.

Maw, H.E.L., Foottit, R.G., Hamilton, K.G.A., & Scudder, G.G.E. (2000). Checklist of the

Hemiptera of Canada and Alaska. Ottawa, National Research Council of Canada, NRC

Research Press.

55

McCoy, E.D., & Bell, S.S. (1991) Habitat structure: the evolution and diversification of a

complex topic. In S.S. Bell, E.D. McCoy and H.R. Mushinsky (Eds), Habitat Structure:

The Physical Arrangement of Objects in Space (pp. 3–27). Chapman and Hall, London,

UK.

McCune, B., & Grace, J. (2002). Analysis of ecological communities. Gleneden Beach, OR:

MjM Software Design.

McCune, J. L., Van Natto, A., & MacDougall, A.S. (2017). The efficacy of protected areas and

private land for plant conservation in a fragmented landscape. Landscape ecology, 32(4),

871-882.

McWilliams, D.A., Berglund, D.R., & Endred, G.J. (1999). Soybean Growth and Management

Quick Guide. North Dakota State University Extension (Publication A1174). Retrieved

from: https://www.ag.ndsu.edu/pubs/plantsci/rowcrops/a1174.pdf

Michener, C.D. (2000). The bees of the world. The Johns Hopkins University Press, Baltimore,

Maryland, USA.

Moore, K.J., & Jung, H.J.G. (2001). ADF and Fiber Digestion. Journal of Range Management,

54(4), 420.

Ohnesorg, W.J., & Hunt, T.E. (2015). Managing soybean defoliators (Publication G2259).

Retrieved from: http://extensionpublications.unl.edu/assets/pdf/g2259.pdf

Oksanen, J. (2014). Multivariate Analysis of Ecological Communities in R: vegan tutorial.

Retrieved from: http://cc.oulu.fi/~jarioksa/popular.htm

56

Olson, M.D., Takasu, K., & Lewis, W.J. (2005). Food needs of adult parasitoids: behavioural

adaptations and consequences. In: Wäckers, F.L., van Rijn, P.C.J., Bruin, J. (Eds.), Plant-

Provided Food for Carnivorous Insects: A Protective Mutualism and its Applications (pp.

137-147). Cambridge University Press.

OMAFRA Publication 811. (2017). Agronomy Guide for Field Crops. Retrieved from:

http://www.omafra.gov.on.ca/english/crops/pub811/pub811.pdf

OMAFRA Publication 812. (2018). Field Crop Protection Guide, 2018-2019. Retrieved from:

http://www.omafra.gov.on.ca/english/crops/pub812/pub812ch4.pdf

O'Toole, C., & Raw, A. (1999). Bees of the world. Blandford Book Publishing.

Ouvrard, D. (2013) Psyl'list - The World Database. Retrieved from:

http://www.hemiptera-databases.com/psyllist

Owen, L.N., Catchot, A.L., Musser, F.R., Gore, J., Cook, D.C., Jackson, R., & Allen, C. (2013).

Impact of defoliation on yield of group IV soybeans in Mississippi. Crop Protection, 54

(2013), 206-212.

Pape, T., Dahlem, G., Mello Patiu, C.A. de & Giroux, M. (2012). World of Flesh Flies (Diptera:

Sarcophagidae). Retrieved from:

http://www.zmuc.dk/entoweb/sarcoweb/sarcweb/sarc_web.htm.

Paradis, E., & Schliep, K. (2018). ape 5.0: an environment for modern phylogenetics and

evolutionary analyses in R. Bioinformatics.

57

Paterson, C. (2014). Small Restoration, Big Impacts: How Habitat Influences Native Pollinators

in Intensive Agricultural Landscapes (Master’s thesis). University of Guelph Atrium.

Paterson, C., Cottenie, K., & MacDougall, A.S. (2019). Restored native prairie supports

abundant and species-rich native bee communities on conventional farms. Restoration

Ecology (in press).

Pedersen, P. (2007). Soybean Growth and Development. Department of Agronomy. Iowa State

University Extension. Retrieved from: http://extension.agron.iastate.edu

Pedersen, P., Elbert, B., & Iowa State University. University Extension. (2014). Soybean growth

and development (Pm (Series) (Ames, Iowa); 1945). Ames, Iowa: Iowa State University,

University Extension.

Peres-Neto, P., & Legendre, P. (2010). Estimating and controlling for spatial structure in the

study of ecological communities. Global Ecology and Biogeography, 19(2), 174-184.

Pfiffner, L., & Luka, H. (2003). Effects of low-input farming systems on carabids and epigeal

spiders in cereal crops – a paired farm approach in NW-Switzerland. Basic and Applied

Ecology, 4(2), 117–127.

Phalan, B., Balmford, A., Green, R.E., & Scharlemann, J.P.W. (2011). Minimizing the harm to

biodiversity of producing more food globally. Food Policy, 36(1), 62–71.

Philip, H. (2015). Field Crop and Forage Pest and their Natural Enemies in Western Canada:

Identification and Management Field Guide. Minister of Agriculture and Agri-Food

Canada.

58

Pont, A.C. (1993). Observations on anthophilous Muscidae and other Diptera (Insecta) in Abisko

National Park, Sweden. Journal of Natural History, 27(3), 631-643.

Pratt, G.K., & Pratt, H.D. (1972). Records of Tabanidae (Diptera) collected on flowers.

Retrieved from: http://agris.fao.org/agris-search/search.do?recordID=US201303217049

Pratt, H.D. (2003). The winter crane flies of North America north of Mexico (Diptera:

Trichoceridae). Proceedings of the Entomological Society of Washington, 105(4), 901–

914.

Presant, E., Acton, C., Land Resource Research Institute, & Ontario Ministry of Agriculture

Food. (1984). The soils of the Regional Municipality of Haldimand-Norfolk (Report

(Ontario Institute of Pedology) (Publication 57). Guelph: Agriculture Canada Research

Branch: Ontario Ministry of Agriculture and Food.

Prevedello, J.A., & Vieira, M.V. (2010) Does the type of matrix matter? A quantitative review of

the evidence. Biodiversity and Conservation, 19(5), 1205–1223.

Proctor, M., &Yeo, P. (1973). The pollination of flowers (The New naturalist). London: Collins.

Proctor, M., Yeo, P., & Lack, A. (1996). The natural history of pollination. Portland, Oregon.

Timber Press.

Punchihewa, R.W.K. (1984). Anthophilous insects and the pollination ecology of Asclepias

syriaca L., and Asclepias incarnata L. in southern Ontario. (M.Sc. thesis). University of

Guelph Atrium.

59

R Core Team (2019). R: A language an environment for statistical computing. R foundation for

statistical computing, Vienna, Austria. Retrieved from: https://www.R-project.org/

Rader, R., Bartomeus, I., Garibaldi, L. A., Garratt, M. P. D., Howlett, B. G., Winfree, R., …

Woyciechowski, M. (2016). Non-bee insects are important contributors to global crop

pollination. Proceedings of the National Academy of Sciences, 113(1), 146–151.

Ramsden, M., Menendez, R., Leather, S., & Wäckers, F. (2017). Do natural enemies really make

a difference? Field scale impacts of parasitoid wasps and hoverfly larvae on cereal aphid

populations. Agricultural and Forest Entomology, 19(2), 139-145.

Robinson, R.A., & Sutherland, W.J. (2002) Post-war changes in arable farming and biodiversity

in Great Britain. Journal of Applied Ecology, 39(1), 157–176.

Rodrigues, P.F.M., & Calhau, J. (2016). Family Therevidae. Zootaxa 4122(1), 387–392.

RStudio Team (2015). RStudio: Integrated Development for R. RStudio, Inc., Boston, MA.

Retrieved from: http://www.rstudio.com/.

Sánchez-Bayo, F., & Wyckhuys, K.A.G. (2019). Worldwide decline of the entomofauna: A

review of its drivers. Biological Conservation, 232(2019), 8-27.

Sandholm, H.A., & Price, R.D. (1962). Field observations on the nectar feeding habits of some

Minnesota mosquitoes. Mosquito News, 22(4), 346-9.

Savely, H.E. (1939). Ecological relations of certain animals in dead and logs.

Ecological Monograph, 9(3), 321-385.

60

Schädler, M., Jung, G., Auge, H., & Brandl, R. (2003). Palatability, decomposition and insect

herbivory: Patterns in a successional old‐field plant community. Oikos, 103(1), 121-132.

Schellhorn, N.A., Bianchi, F.J.J.A., & Hsu, C.L. (2014). Movement of entomophagous

arthropods in agricultural landscapes: Links to pest suppression. Annual Review of

Entomology, 59(1), 559–581.

Schellhorn, N.A., Gagic, V., & Bommarco, R. (2015). Time will tell: resource continuity bolsters

ecosystem services. Trends in Ecology & Evolution, 30(9), 524–530.

Scherr, S., & McNeely, J. (2008). Biodiversity conservation and agricultural sustainability:

Towards a new paradigm of 'ecoagriculture' landscapes. Philosophical Transactions of the

Royal Society of London. Series B, Biological Sciences, 363(1491), 477-494.

Schmidt, S., & Smith D.R. (2016) of the World – an online catalogue of the

family Pergidae (Insecta, Hymenoptera, Symphyta) (World Wide Web electronic

publication). Retrieved from: http://pergidae.snsb-zsm.de

Schowalter, T. (2006). Insect ecology: An ecosystem approach (2nd ed.). Elsevier Science.

Shapiro, S.S., & Wilk, M.B. (1965). An analysis of variance test for normality (complete

samples). Biometrika, 52(3/ 4), 591−611.

Shea, F., & Watts, C. E. (1939). Dumas method for organic nitrogen. Industrial & Engineering

Chemistry Analytical Edition, 11(6), 333-334.

Snedecor, G.W., & Cochran, W. G. (1989). Statistical methods, 8th Ed. Iowa State University

Press, Ames, IA.

61

Soberón, J. (2007). Grinnellian and Eltonian niches and geographic distributions of species.

Ecology Letters, 10(12), 1115-1123.

Southwood, T.R.E. (1977). Habitat, the templet for ecological strategies? Journal of Animal

Ecology, 46(2), 337–365.

Speight, M.C.D. (1978). Flower-visiting flies. In A. Stubbs, P. Chandler (Eds), A dipterist's

handbook (pp. 229-236). Hanworth: The Amateur Entomologists' Society.

Stange, E.E., Ayres, M.P., & Bess, J.A. (2011). Concordant population dynamics of Lepidoptera

herbivores in a forest ecosystem. Ecography, 34(5), 772–779.

Taiz, L., Zeiger, E., Møller, I., Murphy, A., & Møller, I.M. (2015). Plant physiology and

development (Sixth ed.). Sunderland, Massachusetts: Sinauer Associates.

Tilman, D., Balzer, C., Hill, J., & Befort, B.L. (2011). Global food demand and the sustainable

intensification of agriculture. Proceedings of the National Academy of Sciences of the

United States of America, 108(50), 20260–20264.

Timms, L.L., Bowden, J.J., Summerville, K.S., & Buddle, C.M. (2013) Does species-level

resolution matter? Taxonomic sufficiency in terrestrial arthropod biodiversity studies. Insect

Conservation and Diversity, 6(4), 453-462.

Tonietto, R., Ascher, J., & Larkin, D. (2017). Bee communities along a prairie restoration

chronosequence: Similar abundance and diversity, distinct composition. Ecological

Applications, 27(3), 705-717.

62

Tscharntke, T., & Brandl, R. (2004). Plant–insect interactions in fragmented landscapes. Annual

Review of Entomology, 49(1), 405–430.

Tscharntke, T., Bommarco, R., Clough, Y., Crist, T., Kleijn, D.R., Tylianakis, J., & Vidal, S.

(2007). Conservation biological control and enemy diversity on a landscape scale.

Biological Control, 43(3), 295–309.

Tscharntke, T., Tylianakis, J.M., Rand, T.A., Didham, R.K., Fahrig, L., Batary, P.. . . Westphal,

C. (2012). Landscape moderation of biodiversity patterns and processes ‐ eight

hypotheses. Biological Reviews, 87(3), 661-685.

Universal Chalcidoidea Database, 2017. (2017). The natural history museum universal

Chalcidoidea database. Retrieved from: www.nhm.ac.uk/our-

science/data/chalcidoids/introduction.html

University of Kentucky Entomology. (2017). University of Kentucky college of agricultural:

Cooperative extension service. Retrieved from:

www.uky.edu/Ag/CritterFiles/casefile/spiders/spiderfile.htm

Venables, W., & Ripley, B. (2002). Modern applied statistics with S (4th ed., Statistics and

computing). New York: Springer.

Wang, Q. (2017). Cerambycidae of the world: Biology and pest management (Contemporary

topics in entomology series). Boca Raton, FL: CRC Press, Taylor & Francis Group.

63

Wang, X., Vandenbygaart, A., & Mcconkey, B. (2014). Land management history of Canadian

grasslands and the impact on soil carbon storage. Rangeland Ecology &

Management, 67(4), 333-343.

Weibull, A.C., & Ostman, O. (2003). Species composition in agroecosystems: The effect of

landscape, habitat, and farm management. Basic and Applied Ecology, 4(4), 349–361.

Welch, K.D., & Harwood, J.D. (2014). Temporal dynamics of natural enemy–pest interactions in

a changing environment. Biological Control, 75(2014), 18-27.

White, R.E., & Borror, D.J. (1970). A field guide to insects. Peterson Field Guides. Houghton

Mifflin Co. Boston, New York.

Wickramasinghe, L.P., Harris, S., Jones, G., & Jennings, N.V. (2004). Abundance and species

richness of nocturnal insects on organic and conventional farms: effects of agricultural

intensification on bat foraging. Conservation Biology, 18(5), 1283–92.

Wilcox, J.A. (1965). A synopsis of the North American Galerucinae (Coleoptera:

Chrysomelidae). New York State Museum and Science Service.

Williams, N., & Kremen, C. (2007). Resource distributions among habitats determine solitary

bee offspring production in a mosaic landscape. Ecological Applications, 17(3), 910-921.

Wilson, J.S., & Carril, O.J. (2015). The Bees in your backyard: A guide to North American's

bees. Princeton University Press.

Xerces Society. (2014). Farming with native beneficial insects: Ecological pest control.

Solutions, Storey Publishing, LLC. 64

Yafuso, M. (1993). Thennogenesis of Alocasia odora (Araceae) and the role of Colocasiomyia

flies (Diptera: Drosophilidae) as cross-pollinators. Environmental Entomology, 22(3),

601-606.

Yong, T-H. (2005). Prey capture by a generalist predator on flowering and nonflowering ambush

sites: are inflorescences higher quality hunting sites? Environmental Entomology, 34(4),

969-976.

Yoshimoto, C.M. (1990). A review of the genera of New World Mymaridae (Hymenoptera;

Chalcidoidea). Sandhill Crane Press, Gainesville, Fla.

Zhang, W., Ricketts, T.H., Kremen, C., Carney, K., & Swinton, S.M. (2007). Ecosystem services

and dis-services to agriculture. Ecological Economics, 64(2), 253-260.

65

LIST OF TABLES

Table 1: Results of linear models between percentage of landcover type of buffer radii (500, 1000, 1500, 2000 m) and total arthropod abundance and richness for 13 farms in Southern Ontario, Canada.

Total Herbivore Predator Parasitoid Pollinator Response Buffer F- p - F- p - F- p - F- p - F- p - distance model value model value model value model value model value (m) Abundance % semi-natural 500 1.27 0.321 3.83 0.121 0.87 0.401 0.00 0.948 0.98 0.381 landcover 1000 1.94 0.235 2.09 0.221 0.07 0.796 0.69 0.450 0.17 0.696

1500 1.18 0.337 1.19 0.335 0.20 0.675 1.27 0.321 0.05 0.822

2000 0.89 0.398 0.79 0.422 0.36 0.579 1.94 0.235 0.18 0.688

% agricultural 500 0.87 0.403 3.11 0.152 1.09 0.353 0.00 0.989 1.10 0.352 landcover 1000 2.18 0.213 2.63 0.179 0.15 0.710 0.99 0.374 0.14 0.719

1500 0.00 0.988 0.00 0.947 0.53 0.505 3.15 0.150 0.10 0.761

2000 1.05 0.361 0.86 0.403 0.24 0.645 1.72 0.259 0.17 0.696

Richness % semi-natural 500 0.03 0.869 0.12 0.746 0.33 0.592 0.11 0.757 0.12 0.739 landcover 1000 0.42 0.548 0.01 0.921 0.29 0.618 0.19 0.678 0.00 0.950

1500 0.15 0.711 0.14 0.726 0.26 0.636 0.17 0.698 0.01 0.911

2000 1.19 0.336 0.73 0.440 0.04 0.843 0.00 0.970 0.55 0.498

% agricultural 500 0.00 0.988 0.06 0.814 0.72 0.442 9.21 0.01 0.04 0.836 landcover 1000 0.21 0.670 0.09 0.778 0.34 0.588 0.13 0.729 0.01 0.905

1500 0.31 0.605 0.94 0.385 0.03 0.863 0.01 0.901 0.50 0.517

2000 1.14 0.345 0.49 0.521 0.09 0.769 0.00 0.959 0.49 0.519

Observations 13 13 13 13 13

66

Table 2: Mantel test r-values (Pearson’s product-moment correlation) for the correlation between similarity matrices for functional group abundance and richness (family-level) in conventional farms and similarity matrices for spatial connectivity of farms. Total Herbivore Predator Parasitoid Pollinator Response r p r p r p r p r p Abundance 0.067 0.242 -0.025 0.455 -0.144 0.853 - 0.609 -0.029 0.451 0.104 Richness 0.095 0.283 -0.000 0.275 -0.073 0.500 0.045 0.222 0.149 0.182 (family-level)

Observations 13 13 13 13 13

67

Table 3: Summary of F- and p- values from analysis of arthropod abundance and family-level richness across four focal functional groups in response to cover type. Significant effects (p ≤ 0.05) in bold.

Functional Abundance Richness (family-level) Group F - value p - value F - value p - value Total 19.56 <0.001 41.84 <0.001 Herbivore 26.11 <0.001 17.70 <0.001 Predator 26.98 <0.001 59.43 <0.001 Parasitoid 6.92 0.031 15.30 <0.001 Pollinator 50.87 <0.001 44.86 <0.001

68

Table 4: Results of Generalized Linear Mixed Models with negative binomial function of field observations evaluating the effect of plant-related variables on arthropod abundance.

Total Herbivore Predator Parasitoid Pollinator

Abundance Abundance Abundance Abundance Abundance Z- p- Z- p- Z- p- Z- p- Z- p- Predictors statistic value statistic value statistic value statistic value statistic value (Intercept) 2.34 0.019 -0.81 0.418 -2.11 0.035 -1.73 0.084 2.68 0.007

C3 % cover 1.80 0.072 3.34 0.001 3.06 0.002 1.59 0.113 -2.67 0.008

C4 % cover 1.01 0.314 2.14 0.033 0.30 0.766 -1.00 0.316 -1.66 0.097

Forb % cover 2.22 0.026 2.91 0.004 3.35 0.001 0.55 0.584 -0.85 0.396

Richness 0.18 0.857 -0.86 0.388 1.82 0.068 0.96 0.337 0.27 0.784

C:N 0.21 0.836 -0.42 0.674 1.39 0.166 1.40 0.161 0.57 0.569

Phosphorous -0.39 0.695 1.08 0.280 -1.96 0.051 0.22 0.824 -1.20 0.229

ADF -0.76 0.448 1.26 0.208 1.79 0.074 -1.22 0.222 -1.48 0.140

Organic 0.84 0.403 1.14 0.254 -0.91 0.362 -1.06 0.289 0.44 0.657 carbon

Nitrogen -0.66 0.507 -0.49 0.622 -0.32 0.751 0.15 0.885 -0.39 0.695

Biomass -0.63 0.530 0.32 0.748 -0.31 0.758 1.48 0.140 -1.59 0.113

Litter 1.41 0.159 1.13 0.259 1.87 0.062 2.16 0.031 0.50 0.615

Bare ground 0.93 0.354 1.91 0.056 2.36 0.018 1.46 0.146 -2.15 0.032

PAR 2.09 0.036 1.98 0.048 2.06 0.039 -0.10 0.918 0.72 0.470

Herbivore - - - - 3.95 <0.001 2.99 0.003 - -

Random Effects σ2 0.32 0.42 0.49 1.01 0.63

τ00 0.22 Lot 0.63 Lot 0.57 Lot 0.64 Lot 0.17 Lot

ICC 0.41 Lot 0.60 Lot 0.54 Lot 0.39 Lot 0.21 Lot Observations 90 90 90 90 90 Marginal R2 / 0.165 / 0.503 0.165 / 0.667 0.517 / 0.776 0.394 / 0.629 0.267 / 0.423 Conditional R2

69

Table 5: Results of Generalized Linear Mixed Models with negative binomial function of field observations evaluating the effect of plant-related variables on arthropod family-level richness.

Total Herbivore Predator Parasitoid Pollinator

Richness Richness Richness Richness Richness Z- p- Z- p- Z- p- Z- p- Z- p- Predictors statistic value statistic value statistic value statistic value statistic value (Intercept) 3.49 <0.001 0.44 0.660 -2.27 0.023 -2.03 0.043 2.05 0.040

C3 % cover 2.47 0.014 1.94 0.053 2.69 0.007 2.10 0.036 -1.16 0.248

C4 % cover 0.87 0.383 0.48 0.634 0.37 0.712 -0.28 0.782 -1.10 0.269

Forb % cover 3.00 0.003 1.29 0.196 2.75 0.006 0.48 0.632 0.70 0.487

Richness -0.49 0.622 -0.51 0.608 0.94 0.346 0.76 0.447 -1.60 0.109

C:N 0.92 0.359 0.89 0.373 0.15 0.884 -0.87 0.386 0.57 0.568

Phosphorous -0.64 0.522 -0.47 0.637 -1.48 0.138 -0.18 0.857 -0.35 0.725

ADF 0.08 0.935 0.57 0.568 1.21 0.228 0.60 0.548 0.00 0.996

Organic carbon -1.90 0.058 -1.57 0.116 -0.42 0.677 -1.04 0.300 -0.89 0.374

Nitrogen -0.11 0.910 0.58 0.560 -0.17 0.868 -0.19 0.852 -0.57 0.570

Biomass -0.35 0.723 -0.84 0.402 -0.04 0.967 1.54 0.124 -1.30 0.193

Litter 2.20 0.028 1.39 0.164 1.39 0.164 0.65 0.517 0.85 0.396

Bare ground 1.16 0.246 -0.25 0.801 2.17 0.030 2.00 0.045 -1.07 0.284

PAR 2.80 0.005 2.53 0.011 1.83 0.067 0.17 0.863 0.89 0.376

Herbivore - - - - 2.02 0.044 3.22 0.001 - -

Random Effects σ2 0.12 0.31 0.53 1.06 0.50

τ00 0.09 Lot 0.03 Lot 0.19 Lot 0.01 Lot 0.01 Lot

ICC 0.44 Lot 0.09 Lot 0.26 Lot 0.01 Lot 0.02 Lot Observations 90 90 90 90 90 Marginal R2 / 0.332 / 0.628 0.290 / 0.357 0.546 / 0.664 0.334 / 0.459 0.215 / 0.233 Conditional R2

70

Table 6: Results from ANOVA summary statistics analyzing factors prairie size and prairie age for arthropod abundance and richness (family-level) response.

Prairie size Prairie age Response F-value p - value F-value p - value Abundance Total 3.44 0.082 0.855 0.367 Herbivore 3.34 0.086 0.155 0.698 Predator 0.19 0.662 3.75 0.054 Parasitoid 0.08 0.780 1.88 0.189 Pollinator 1.05 0.319 0.62 0.439 Richness (family-level) Total 1.14 0.299 1.80 0.197 Herbivore 3.90 0.06 0.92 0.349 Predator 0.02 0.902 3.63 0.074 Parasitoid 0.36 0.554 0.003 0.957 Pollinator 0.04 0.842 3.06 0.098

71

LIST OF FIGURES

A) 22 total farm profile B) 13 “main” farms

1.001.00

0.33 0.33

0.750.75 FunctionalFunctional

y 0.56

y 0.56

t

t i

i GroupGroup

n

n u

u herbivherbivore ore m

m predatorpredator

m

m o

o 0.25 0.25 parasitoidparasitoid

C

C

0.50 0.50 f

f pollinatorpollinator

o

o

anthophilousanthophilous

n

n

o o

i kleptoparasite

i kleptoparasite t 0.13

t 0.13 c

c 0.13 0.13 aquaticaquatic

a

a r

r decomposer

F 0.04 decomposer F 0.04 uncommon 0.250.25 0.09 uncommon 0.08 0.08 0.09

0.08 0.08 0.17 0.17 0.02 0.04 0.02 0.04 0.00 0.04 0.00 0.04 Abundance Richness Abundance Richness Figure 1: Total arthropod abundance and richness (family-level) sampled across (A) 22 total farms and (B) 13 farms as they relate to 9 key functional groups. The focal functional groups are given in colour (herbivores = purple, predators = yellow, parasitoids = blue, pollinators = green) while all other functional groups are in grey. Bars are annotated to show the fraction of the total arthropod community comprised by each functional group when the fraction was > 0.1.

72

20

s

s e

n 15

h

c

i

r

d

i

o

t i

s 10

a

r

a P

5

20 40 60 80 Agricultural land cover within 500 m buffer (%)

Figure 2: Parasitoid richness (family-level) as related to percent agricultural land use at a buffer radius of 500 m.

73

A. b B. b

600 60

level) b

-

(family

e s

c 40 s

n 400 a

e

a

n

d h

n a

a c

u

i

b

R A 200 20

0 0 Crop Prairie Forest Crop Prairie Forest Cover type

Figure 3: (A) Abundance (B) and richness (family-level) of arthropods across three different cover types (crop = green, prairie = yellow, forest = brown). Error bars show ± 1 SE, and lower- case letters indicate significance at p<0.050 as determined by post-hoc tests.

74

A. b B. b

20

400 ab b

level) - 15 c Cover

300 (family

e s

c a s

n type

e

a

n d

h b

n 10 Crop

c u 200 i

b a b Prairie a R

A ab a a Forest a a 100 b b bc 5 a b ab a a a 0 0

r r r r re id o o re id o o o to t t o to t t iv i a a iv i a a b s in d b s in d r ra ll re r ra ll re e a o p e a o p h p p h p p Functional Group 1 Figure 4: (A) Abundance (B) and richness (family-level) of arthropods belonging to the 4 focal 2 functional groups across three different cover types (crop = green, prairie = yellow, forest = brown). Error bars show ± 1 SE, and lower-case letters indicate significance at p<0.050 as determined by post-hoc tests.

75

Figure 5: Ordination of arthropod community composition using non-metric multidimensional scaling (NMDS) showing variability between cover types on farm sites (n=13). Farm sites near each other have a community composition that is more similar than farms that are separated by greater distance. Ellipses represent 95% CI around the centroid.

76

12.5

10.0

)

%

(

n

o 7.5

i

t

a

i

l

o f

e 5.0

d

f

a e L 2.5

0.0 Prairie absent Prairie present

Figure 6: Average soybean leaf defoliation for farms which had prairie adjacent to the crop field (n=7) and those without adjacent prairie (n=3). Error bars show ± 1 SE.

77

A B

C D A

Figure 7: Relationship between abundance and richness of arthropod functional groups in farms with prairie present (n=7) and farms with prairie absent (n=3). (A) Mean abundance of arthropod functional groups in prairie absent farms. (B) Mean abundance of arthropod functional groups in prairie present farms. (C) Mean richness (family-level) of arthropod functional groups in prairie absent farms. (D) Mean richness (family-level) of arthropod functional groups in prairie present

78

A B

Figure 8: Relationship between abundance of arthropod functional groups in crop fields and distance from edge from farms with (A) farms with prairie present adjacent to crop (n=7) and (B) farms with no prairie present (n=3).

79

A B

Figure 9: Relationship between leaf defoliation and distance from edge on farms with prairie present adjacent to crop (n=7) and farms with no prairie present (n=3).

80

N=1 A

N=1

N=2

N=10 N=1 N=5

(organic)

N=1 B

level)

- (family

N=1

N=1 N=2 N=2 N=13 N=13

N=10 N=5 N=13

(organic) Figure 10: (A) Mean abundance of arthropod functional groups for different crop types. (B) Mean richness (family-level) of arthropod functional groups for different crop types.

81

Figure 11: Data from non-farm iNaturalist observations in Southern Ontario compared to 13 agricultural field sites. Each bar in the graph represents the summed count of total arthropod families in each dataset and segments in the graph represent the functional group composition of that total.

82

)

shared or unique or shared Number of families ( families of Number

Figure 12: Number of arthropod families unique or shared from data of non-farm iNaturalist observations in Southern Ontario compared to 13 agricultural field sites. Each bar in the graph represents the summed count of total arthropod families in each dataset and segments in the graph represent the functional group composition of that total.

83

Appendix 1: Conceptual figures

Figure S1: Hierarchical framework of habitat-based drivers of arthropod abundance is shown. Habitat occurs at multiple scales: landscape distance-based and local farm cover based. A specialized plot-level analysis was conducted to explain my local-scale prairie cover which consisted of plant-related factors of plant quantity, plant tissue quality, and plant composition.

84

Appendix 2: Site selection and description

Table S1: List of study sites and cover type variables. Farms located in Southern Ontario, field research taken from May – July 2017. Annual mean temperature and annual mean precipitation values are taken from WorldClim 1.4: current conditions data at 30 arc-seconds (~1 km).

Farm Location Prairie Forest Cover Crop Julian Sampling Annual Annual size size type size date date mean mean (acre) (acre) (acre) temp. temp. prec. (℃) (℃) (mm)

1 Waterford - 7.55 Corn 31.94 158 15 7.75 950 2 Waterford 12.3 0.58 Soybean 20.51 149 20 7.71 950 3 Langton 1.00 6.89 Soybean 23.81 135 15 8.02 961 4 Straffordville 0.51 0.069 Soybean 1.8 143 15 8.09 930 e 5 Straffordville 1.08 0.995 Wheat 25.45 144 16 8.09 930 6 Norfolk 5.47 0.87 Corn 29.29 138 24 7.67 951 7 Simcoe 14.21 1.72 Wheat 12.34 150 18 7.71 951 8 Simcoe 9.9 2.67 Grasslan - 137 20 7.83 949 9 Delhi 2.50 1.97 Soybeands 15.68 153 15 7.72 951 10 Norwich 3.91 0.61 Soybean 37.61 163 26 7.78 949 11 Delhi 7.9 1.683 Corn 13.92 152 15 7.54 946 12 Port Dover 4.87 2.6 Soybean 10.12 146 16 7.71 947 13 Dunwich - 2.65 Soybean 30.08 170 20 7.76 960 14 Dunwich - 2.19 Soybean 40.25 170 20 7.76 950 15 Aylmer 23.37 1.95 Corn 14.39 185 19 7.67 951 16 Aylmer 13.21 3.25 Corn 34.41 179 17 7.97 959 17 Eden - 1.68 Soybean 30.88 180 22 8.12 951 18 Tillsonburg 15.9 100 Sorghu 15.04 172 17 7.97 954 m 19 Simcoe 20.49 6.22 Grasslan - 159 16 7.66 947 20 Rodney 3.20 1.91 Soybeands 30.96 171 18 7.8 943 21 Tillsonburg 4.27 0.85 Apple 11.37 165 18 5.85 922 orchard 22 Hillsburgh 3.84 0.96 Organic 5.7 167 20 8.45 915 mixed

85

Table S2: Total plant species and associated functional groupings found within prairie cover.

Common name Family Genus Species Functional group redtop Poaceae Agrostis Agrostis gigantea C3 grass smooth brome Poaceae Bromus Bromus inermis C3 grass drooping brome Poaceae Bromus Bromus tectorum C3 grass orchard grass Poaceae Dactylis Dactylis glomerata C3 grass couch grass Poaceae Elymus Elymus repens C3 grass

Timothy-grass Poaceae Phleum Phleum pratense C3 grass annual bluegrass Poaceae Poa Poa annua C3 grass

Kentucky C3 grass bluegrass Poaceae Poa Poa pratensis

Indian grass Poaceae Sorghastrum Sorghastrum nutans C4 grass big bluestem Poaceae Andropogon Andropogon gerardii C4 grass hairy crabgrass Poaceae Digitaria Digitaria sanguinalis C4 grass ryegrass Poaceae Lolium Lolium spp. C4 grass switchgrass Poaceae Panicum Panicum virgatum C4 grass common yarrow Asteraceae Achillea Achillea millefolium Forb redroot pigweed Amaranthaceae Amaranthaceae Amaranthus retroflexus Forb common ragweed Asteraceae Ambrosia Ambrosia artemisiifolia Forb swamp milkweed Apocynaceae Asclepias Asclepias incarnata Forb hoary alyssum Brassicaceae Berteroa Berteroa incana Forb mouse-ear chickweed Caryophyllaceae Caryophyllaceae Cerastium vulgatum Forb creeping thistle Asteraceae Cirsium Cirsium arvense Forb pigweed Amaranthaceae Chenopodium Chenopodium album Forb bird's-foot trefoil Apiaceae Daucus Daucus carota Forb showy tick trefoil Fabaceae Desmodium Desmodium canadense Forb common horsetail Equisetum Equisetum Equisetum arvense Forb fleabane Asteraceae Erigeron Erigeron annuus Forb Canadian horseweed Asteraceae Erigeron Erigeron canadensis Forb Philadelphia fleabane Asteraceae Erigeron Erigeron philadelphicus Forb wild strawberry Rosaceae Fragaria Fragaria vesca Forb rough oxeye Asteraceae Heliopsis Heliopsis helianthoides Forb meadow hawkweed Asteraceae Hieracium Hieracium pratense Forb

86

honeysuckle Caprifoliaceae Lonicera Lonicera spp. Forb bird's-foot trefoil Fabaceae Lotus Lotus corniculatus Forb black medick Fabaceae Medicago Medicago lupulina Forb sweet clover Fabaceae Melilotus Melilotus alba Forb common mallow Malvaceae Malva Malva neglecta Forb wild bergamot Lamiaceae Monarda Monarda fistulosa Forb narrow leaf plantain Plantaginaceae Plantago Plantago lanceolata Forb broadleaf plantain Plantaginaceae Plantago Plantago major Forb Virginia knotweed Polygonaceae Persicaria Persicaria virginiana Forb common cinquefoil Rosaceae Potentilla Potentilla simplex Forb common bracken Dennstaedtiaceae Pteridium Pteridium spp. Forb wild radish Brassicaceae Raphanus Raphanus raphanistrum Forb black-eyed Susan Asteraceae Rudbeckia Rudbeckia hirta Forb trailing pearlwort Caryophyllaceae Sagina Sagina decumbens Forb green foxtail Poaceae Setaria Setaria viridis Forb bladder campion Caryophyllaceae Silene Silene vulgaris Forb Canadian goldenrod Asteraceae Solidago Solidago canadensis Forb chickweed Caryophyllaceae Stellaria Stellaria media Forb Symphyotrichum white heath aster Asteraceae Symphyotrichum ericoides Forb Symphyotrichum novae- New England aster Asteraceae Symphyotrichum angliae Forb purple-stemmed Symphyotrichum aster Asteraceae Symphyotrichum puniceum Forb common dandelion Asteraceae Taraxacum Taraxacum officinale Forb red clover Fabaceae Trifolium Trifolium pratense Forb great mullein Crophulariaceae Verbascum Verbascum thapsus Forb cow vetch Fabaceae Vicia Vicia cracca Forb riverbank grape Vitaceae Vitis Vitis riparia Forb field pansy Violaceae Viola Viola arvensis Forb

87

Hillsburgh

Prairie

Norfolk County Elgin County

Figure S2: Location of the 22 study sites within the Southern Ontario, Canada region. Farms strictly used in my main analysis are coloured in green. Farms used in complementary analyses include farms coloured in black.

88

Figure S3: Monthly temperature and precipitation data from historical records (in red) and within my sampling year (in blue) during sampling months May-July 2017.

89

Appendix 3: Sampling methods

Figure S4: Quadrant sampling design for arthropod and plant measurements. All sampling occurred from May-July 2017 on farms located in Southern Ontario, Canada.

90

Figure S5: Conceptual figure of sampling design used. For the majority of cases, a 5x 5 grid would be overlaid onto the satellite image of each farm cover (crop, prairie, wood). In cases where covers were unsymmetrical or where a 5 x 5 grid could not be easily overlaid, instead 25 grid squares that spanned the whole cover were overlaid. In all cases, five sampling locations were chosen using a random number generator.

91

level)

- Richness(family

Figure S6: Sample-based rarefaction curves of arthropod richness were drawn three different cover types (prairie, wood, crop). Solid lines represent the estimated number of species for each cover and the shaded area between the lines indicates the estimated 95% confidence interval.

92

A

undance

Ab

level)

- B Richness (family Richness

Figure S7: (A) Mean abundance (± SE) of arthropod community per cover type and (B) Mean richness (family-level) (± SE) of arthropod community per cover type. Bars with the same letter are not significantly different (ANOVA, followed by comparisons between cover type using Tukey’s HSD test, α = 0.05).

93

Appendix 4: Functional group categorization and definitions

Table S3: Definitions of arthropod functional groups. Functional groups Definition Primary groups Herbivore plant feeder, economically important pests lumped here because all pests are herbivores but not all herbivores are pests Predator beneficial arthropod; eats other arthropods Parasitoid beneficial arthropod; larvae are parasites that kill their hosts Pollinator beneficial arthropod; bee and non-bee that are effective pollinators Secondary groups Anthophilous flower visiting, or flower eating but are not as effective in spreading pollen to gametes of flowers Aquatic has an aquatic life cycle stage to become an adult; usually non-feeding adult stage. Decomposer decomposes organic material; often found in soil habitats or within ground Kleptoparasite arthropod that robs other species of food, not a beneficial arthropod

Uncommon less than three total abundance found on farm; or unresolved feeding guild

94

Table S4: Functional group categorization and literature search. Order/Family Functional Feeding description References group

Tabanidae anthophilous Often visits flowers. Both males and females Larson & Inouye, collect nectar from flowers females later 2001; Pratt & Pratt, switch to blood meals after mating. 1972 Lonchopteridae anthophilous Adults feed on nectar and are associated with Larson & Inouye, flower visitation. 2001; Proctor, Yeo & Lack, 1996; Klymko & Marshall, 2008 Mycetophilidae anthophilous The small gnat-like adults occasionally visit Marhsall, 2006; Larson flowers with exposed nectaries, but they are & Inouye, 2001 not important pollinators. Phalacridae anthophilous Larvae feed on fluid material within flower Gimmel, 2013; heads of Asteraceae and pollen feeders as Marhsall, 2006 adults. Ceratopogonidae anthophilous Feed on flower nectar or other sugar source. Larson & Inouye, Females and males commonly visit flowers 2001; Bystrak, 1978; with exposed nectar, All males feed only on Kearns, 2001 sugars. Chironomidae anthophilous Adults are short-lived but are ranging from Larson & Inouye, opportunistic nectar and pollen consumers. 2001; Kevan & Baker, 1983; Kearns, 2001 Mordellidae anthophilous Common on flowers and foliage. Adults feed Ford & Jackson, 1996; on pollen and nectar from a variety of plants. Dillon, 1961 Phoridae anthophilous Of the Cyclorrapha, recent reviews indicate Larson & Inouye, that they are flower visiting. 2001; Disney, 1983 Tipulidae anthophilous Long-beaked members can be found Larson & Inouye, abundantly on flowers of some Lamiaceae in 2001; Marhsall, 2006 Ontario. Bibionidae anthophilous Observed early in the year on flowers Larson & Inouye, because mating swarms often forage on 2001; Fitzgerald, 2005; nearby flowering plants for sustenance. Their D'arcy-Burt & effectiveness needs to be quantified. Blackshaw, 1991; Kearns, 2001 Scatopsidae anthophilous Observed on flowers with exposed nectar or Proctor, Yeo & Lack, pollen. 1996; Speight, 1978 Conopidae anthophilous Often regarded as flower visitors, adults take Proctor, Yeo & Lack, nectar. 1996; Proctor & Yeo, 1973 Culicidae anthophilous The abundance of these common flower Downes, 1958; visitors may compensate for their low Stockholm & Price, individual effectiveness; mosquitoes often 1962; Hocking, 1968; visit flowers at night. Kevan, 1970; Kearns, 2001 95

Simuliidae anthophilous Observed on flowers with exposed nectar, Davies & Peterson such as umbellifers (Apiacaea). 1956; Burgin & Hunter 1997 Calliphoridae anthophilous Common visitors to flowers of Asclepias spp. Punchihewa, 1984; (Asclepiadaceae) Well known as anthophiles Ambrose et al., 1985; and nectarophages. Beaman, Decker & Beaman 1988 Formicidae anthophilous . are commonly anthophilous (flower Gullan & Cranston, loving), but rarely pollinate the plants that 2014 they visit. Muscidae anthophilous Muscidae are often considered the most Pont, 1993; Speight, significant anthophiles after syrphids. 1978; Proctor, Yeo & Lack, 1996; Larson & Inouye, 2001 Sarcophaginae anthophilous Flesh flies (Sarcophagidae) and dung flies Larson & Inouye, 2001 (Scathophagidae) are occasionally reported as anthophilous. Some sarcophagids and scathophagids may be pollinators, but their true roles have not been well elucidated. Sepsidae anthophilous Feed on nectar at umbels of Apiaceae but are Larson & Inouye, not likely pollinators. 2001; Marshall, 2006 Sphaeroceridae anthophilous Found in trap flowers and the spathes of Yafuso 1993; Proctor, Araceae; sparse records of anthophily. Yeo & Lack, 1996 Sarcophagidae anthophilous Flesh flies (Sarcophagidae) and dung flies Larson & Inouye, (Scathophagidae) are occasionally reported as 2001; Kevan, 1970; anthophilous but may be more prevalent than Proctor, Yeo & Lack, presently thought. Some sarcophagids and 1996 scathophagids may be pollinators, but their true roles have not been well elucidated. Stratiomyidae anthophilous Adults have been recorded from the flowers Knuth, 1908; Ambrose of 22 plant families: they were observed in et al., 1985 low numbers but often on the flowers. Cerambycidae anthophilous Many adults (especially the brightly coloured Wang, 2017; ones) feed on flowers. Many of the flower- Hawkeswood & feeding cerambycids pollinate their food Turner, 2007 plants as they feed on pollen and nectar. Phlaeothripidae anthophilous Includes many common species, commonly Marshall, 2006 found on daisy flowers. anthophilous Larvae feed on stored pollen or prey of host's Marshall, 2006 nest (usually bees or wasps), adults feed on nectar of primarily carrot family (Apiaceae). Plecoptera aquatic Nymphs feed on algae, diatoms, mosses, and White & Borror, 1970 immature aquatic invertebrates, including mayflies and midges; most spring and summer adults do not feed, and are nocturnal;

96

winter stoneflies are day-flying, and feed on blue-green algae and foliage. Ephemeroptera aquatic Adults have no functional mouthparts and do Arnett, 2000 not feed. Trichoptera aquatic Some adults take liquid food, such as nectar, Holzenthal et al., 2015; others do not feed. Marshall, 2006 Ephydridae aquatic Larvae of most spp. filter microorganisms Arnett, 2000 (bacteria, unicellular algae, yeasts) from the surrounding semiliquid medium. Associated with semi aquatic or aquatic habitats. Only 3 instances of this family were found. Hydropsychidae aquatic non-feeding adult stage Holzenthal et al., 2015 Nemouridae aquatic see Pecoptera. non -feeding adult stage. White & Borror ,1970 Brachyptera aquatic see Plecoptera. non -feeding adult stage. White & Borror ,1970 Entomobryidae decomposer Springtails are "decomposers" that thrive Marshall, 2006; mostly on decaying organic matter, especially Hopkin, 1997 vegetable matter. Sminthuridae decomposer Springtails are "decomposers" that thrive Marshall, 2006; mostly on decaying organic matter, especially Hopkin, 1997 vegetable matter. Dryomyzidae decomposer North American Dryomizids are stickly Mathis & Rung, 2011; microbioal grazers in decomposing material Marshall, 2006 such as rotting fungus and dead invertebrates. Tetratomidae decomposer mostly fruiting bodies of hymenomycete Arnett et al., 2002 fungi, esp. Polyporaceae and Tricholomataceae; adults usually on the surface, larvae bore inside. Fanniidae decomposer Larvae live as scavengers in various kinds of Marshall, 2006 decaying organic matter. Lauxaniidae decomposer Associated with mining dead or decaying Marshall, 2006 leaves. Their habitat includes forests and forests. Leiodidae decomposer Their habitat includes litter, various decayed Marshall, 2006 matter, vertebrate and nests/burrows. Most common encountered beetle in this family is attracted to dead bodies. Lonchaeidae decomposer The larvae of most Lonchaeidae are invaders Savely, 1939 in diseased or injured plants. Known to aid in decomposition. Trichoceridae decomposer Larvae occur in moist habitats where they Pratt, 2003 feed on decaying vegetable matter. They are

of no economic importance. herbivore Most are phytophagous, many may be host- Arnett 2000 specific, often associated with plants with glandular trichomes in Geraniaceae, 97

Onagraceae, Scrophulariaceae, and Solanaceae. Heteronemiidae herbivore All Phasmids feed on foliage, and possess Chad 2006 chewing mouthparts designed for breaking down plant material. Membracidae herbivore Most species are host-specific and feed on Arnett, 2000 trees and shrubs, some on herbaceous plants. Only a few species are considered pests; most of the damage is caused by egg-laying. Nitidulidae herbivore They feed mainly on decaying vegetable Downie & Arnett, matter, over-ripe fruit, and sap. 1996 Buprestidae herbivore The larvae burrow through roots and logs, Jackson & Jewiss- from within the bark to within the cambium Gaines, 2012; Donald layers, or are leaf and stem miners of & Bright, 1987 herbaceous and woody plants, including grasses. herbivore Herbivore, sapsucking, but not seen as pest Arnett, 2000 on crops Languriini herbivore Observed as herbivorous (unlike other Ernett et al., 2002 erotylid lineages). Lygaeidae herbivore Mostly seeds, although some species are Arnett, 2000 considered crop pests. Pergidae herbivore Food plant diversity exceeds that found in Schmidt & Smith, any other sawfly family. No observations of 2016 Pergidae consuming crop plants. Psyllidae herbivore Associated almost exclusively with Hill, 2008; Ouvrard, dicotyledons, with a few species developing 2013 on monocots and on . Known pest of tomato and potatoes and occasionally serious pest on citrus. herbivore Seeds of herbaceous plants, but some are MacGavin, 1999; arboreal. Johnson & Triplehorn, 2004 herbivore Eat seeds, formerly treated as subfamily of Arnett, 2000; Henry, Lygaeidae. 2009 herbivore Both adults and larvae feed mainly on Marshall, 2006 staminate pine flowers, buds or developing cones, although some are known to feed on deciduous trees. herbivore Herbaceous plants and seeds, primarily that Marshall, 2006 of legumes (Fabaceae). Agridae herbivore Feeds on wide range of trees and shrubs, Marshall, 2006 including poison ivy by some species. Siricidae herbivore Larvae feed on fungi inside the decaying tree Marshall, 2006 trunk, adults feed only on nectar and water 98

Oedemeridae herbivore Adults feed on pollen of variety of plants, Arnett, 2000 such as Asteraceae, Poaceae, and Rosaceae. herbivore Feeds on plant seeds and sap. Marshall, 2006 Anobiidae herbivore Larvae feed inside the trunk of decaying oak, Marshall, 2006 beech, or pine trees, adults do not eat. herbivore Most adults feed on plant sap, nymphs of Marshall, 2006 some species are known to feed on fungi. herbivore Feeds on variety of trees, shrubs and Marshall, 2006 (unknown family) herbaceous plants, usually young foliage. Cymidae herbivore Feeds on the seeds of sedges (Cyperaceae), Marshall, 2006 rushes (Juncaceae), and occasionally grasses (Poaceae). Oedemeridae herbivore Pollen feeders often found on flowers in early Marshall, 2006 spring. Passalidae herbivore Both adult and larvae feed on Marshall, 2006 microorganism-rich rotting wood to digest cellulose. herbivore Adults and larvae of most species feed on Marshall, 2006 tree or shrub foliage, adults of some species feed on pollen and nectar. herbivore Adults and larvae of most species feed on Marshall, 2006 tree or shrub foliage, adults of some species feed on pollen and nectar. Thyreocoridae herbivore Feed on flowers and developing seeds. They Marshall, 2006; are exclusively herbivorous. Matesco & Grazia, 2015 Taeniopterygidae herbivore Adults and nymphs are plant feeders. DeWalt 2013 herbivore Feed mainly on leaves of trees and shrubs, Johnson & Triplehorn, causing yellow spotting and sometimes 2004; Hill, 2008 browning and death of the leaves. Known as a sap-sucker with toxic saliva. Acrididae herbivore Typically eat foliage of forbs, grasses. Some Hill, 2008 species take a variety of plants, while others are restricted to a few species of closely related plants. Cited as major maize pest. Agromyzidae herbivore Larvae are leaf miners, some are stem or seed Hill, 2008 borers, cited as minor pest of soybean. Anthomyiidae herbivore It is an important pest of germinating Hill, 2008; Philip, soybeans and corn. It also attacks a wide 2015; Funderburk et range of horticultural crops including beans, al., 1983 peas, cucumber, melon, onion, pepper, potato, and other vegetables. Aphididae herbivore Aphids suck juices from plants and may be Hill, 2008; Philip, quite damaging. Cited as major pest of wheat 2015 in North America, cited as major maize pest. 99

Cecidomyiidae herbivore Most are gall makers. Others feed on plants. Hill, 2008 Cited as major pest of wheat (including barley and oats) in Europe and North America. herbivore Grass-feeders (xylem). Cited as minor Hill, 2008 agricultural pests of forage crops. Chloropidae herbivore In most species, larvae feed on grass stems of Hill, 2008 the families Poaceae, Cyperaceae, and Typhaceae. Some are serious pests of cereals. Cicadellidae herbivore Nymphs and adults feed on sap of above- Hill, 2008 ground stems or leaves of plants; some species are more host-specific than others, cited as major pest of maize. herbivore Nymphs subterranean, feeding on roots (and Hill, 2008 possibly fungi). Many are presumed to be polyphagous as adults, most often on woody plants. herbivore All are plant feeders. Hill, 2008 Curculionidae herbivore Most larvae and adults occur and feed on all Hill, 2008; Philip, parts of plants, and many species are 2015 important pests because they chew holes in fruits, nuts, and other parts of cultivated plants, cited as major crop pest of wheat in North America. herbivore Mostly grass feeders, cited as minor pest of Hill, 2008 wheat in North America. Elateridae herbivore Adults usually eat plants. Larvae eat newly Hill, 2008 planted seeds, roots, cited as minor pest of wheat. Gryllidae herbivore Cited as minor pest of maize. Hill, 2008 Meloidae herbivore Adults feed on leaves and flowers of several Hill, 2008 families of plants, particularly Asteraceae, Fabaceae, and Solanaceae. Pentatomidae herbivore The majority are herbivorous, but members Hill, 2008 of one subfamily (Asopinae) are predaceous on other arthropods. Both adults and nymphs of plant-feeding species may damage plants, mostly by piercing the plant tissues and thus opening a path for pathogens to enter the plant, cited as wheat crop pest. Scarabaeidae herbivore Adults take a variety of foods, some of these Hill, 2008 are agricultural pests., cited as major crop pest of wheat in North America. Scathophagidae herbivore Larvae of Delininae are leaf miners in Hill, 2008 Liliaceae, Orchidaceae, and Commelinaceae.

100

Larvae of Scathophaginae have been reported as phytophagous. herbivore Though most jewel bugs do little harm to Hill, 2008 crop plants, a few members of Scutelleridae are considered major agricultural pests. herbivore Most of the larvae feed on foliage, but a few Hill, 2008 are leaf miners or stem borers, observed study of crop plants and leaf feeding damage found laminae sclerotized. Tettigoniidae herbivore Most common form of leaf damage by Hill, 2008 defoliating arthropods is by chewing insects such as grasshoppers. Grasshoppers cited as major maize pest. Ullidiidae herbivore All three species studied here should be Goyal et al., 2012 considered primary pests that can render unprotected sweet corn ears unmarketable. Drosophilidae herbivore A serious fruit fly pest of soft fruit and AgriService BC, 2017 berries. Miridae herbivore In Canada, the plant bug genus Lygus Kelton, 1982; Maw et (Heteroptera: Miridae) contains 27 native al., 2000 species, of which 14 are recorded as agricultural field pests. Tephritidae herbivore Leaf mining fruit fly which attacks many Philip, 2015; Hill, crops in Canada, both larval and adult stages 2008 damages crops. Tenebrionidae herbivore Cited a minor crop pest of wheat in North Hill, 2008 America and Europe. Opomyzidae herbivore Food includes grasses and cereal crops. Arnett, 2000; Hill, bore shoots. 2008 Galerucinae herbivore Known as leaf skeletonizes. This family has Wilcox, 1965; Hill, some serious agricultural pests, causing 2008 damage directly by plant feeding or indirectly by transmitting viruses. Bruchidae herbivore Cited as minor pest of soybean. Hill, 2008 Psilidae herbivore Some species are serious crop pests. Hill, 2008 herbivore All specimens were Brown Marmorated Hoebeke & Carter, Stink Bug, considered serious crop pest. All 2003 adults are phytophagous. Lasiocampidae herbivore Family of tent caterpillar. Larvae feed on Alford, 2016 foliage of many fruit crops in the Rosaceae family, such as apple, pear, cherry, peach, and plum. Oedipodinae herbivore Most families of grasshoppers seen as general Hill, 2008 pests. Noted as leaf chewers.

101

Lygaeidae herbivore Cited as minor pest of wheat in North Hill, 2008 America. Thripidae herbivore Plant feeding common thrips in this family Marshall, 2006 are frequent problems in greenhouses and are considered pests, also because they carry and spread disease. Chrysomelidae herbivore Chrysomelids are phytophagous and are Hill, 2008; Philip, among the most diverse and conspicuous 2015 arthropod families on plants. The adults feed on living plant material, usually consuming leaves or sometimes various flower parts including pollen. Cited as minor crop pest of wheat in North America. Bostrichidae herbivore Most species attack wood, either living, or in Marshall, 2006 some cases, dead, including seasoned lumber. Miltogramminae klepto- Kleptoparasites of Digger Wasps; the prey Pape et al., 2012 parasite may include arthropods from a variety of families and orders. Chrysididae kepto- Some species are parasitoids and others Xerces Society, 2014; parasite kleptoparasites. Either way the host larva Goulet & Huber, 2001; dies. All specimens found were from Wilson & Carril, 2015 subfamily Chrysidinae which are generally kleptoparasites. klepto- All species feed on nectar and pollen, but a Goulson & Ebrary, parasite few are kleptoparasites, parasitic species do 2010; O'Toole et al., not possess scopae and all specimens found 1999; Mader et al., in this study had one present. The genera 2011; Wilson & Carril, found in this study were majority Osmia, 2015 Megachile, and Hoplitis. parasitoid They are almost exclusively phytophagous Universal arthropods, forming galls in plant stems, Chalcidoidea leaves, or seeds. Database, 2017, 2017 Tiphiinae parasitoid Paralyze grubs of scarab beetles, lay an egg Xerces Society, 2014; on it and leave the grub for the young to eat. Bethylidae parasitoid Parasites of coleopteran and lepidopteran Evans, 1978 larvae. parasitoid Many life histories adapted to parasitizing Xerces Society, 2014; hosts as diverse as aphids, bark beetles, and Goulet & Huber, 1993 foliage-feeding . Many species are egg-larval parasitoids, laying eggs within host eggs and then not developing until the host is in the larval stage. Pipunculidae parasitoid Adults feed on honeydew secretions; larvae Xerces Society, 2014; mostly parasitize and Goulet & Huber, 1993 .

102

Tiphiidae parasitoid Parasitoids of subterranean beetle larvae Xerces Society, 2014; (especially of Scarabaeoidea, Tenebrionidae, Goulet & Huber, 1993 and Cicindelinae) and various bees/wasps occurring in soil or rotten wood. Ichneumonidae parasitoid A great variety of hosts (mostly immature Xerces Society, 2014; stages) is used, though most species attack Goulet & Huber, 1993 only a few host types; some infest spiders and other non-arthropod arthropods. parasitoid Larvae are parasites of xylobiontic arthropod Goulet & Huber, 1993 larvae, primarily of Buprestidae, but also of Cerambycidae, and Siricidae. Ripiphoridae parasitoid Lays eggs on families of Hymenoptera, Arnett et al., 2002 wood-boring beetle larvae. Eurytomidae parasitoid The Eurytomidae contains species which Goulet & Huber, 1993; exhibit a wide range of biologies, but the Universal majority seem to be endophytic, either as Chalcidoidea phytophages or as parasitoids of Database, 2017, 2017 phytophagous arthropods. parasitoid Parasites on a wide range of species of Goulet & Huber, 1993; different orders (Lepidoptera, Hemiptera, Universal Hymenoptera, Coleoptera, Neuroptera, Chalcidoidea Orthoptera). Database, 2017, 2017 parasitoid Typically, parasites of fungus gnats and other Goulet & Huber, 1993 dipterans. Eucoilidae parasitoid Parasitoids of fly . Goulet & Huber, 1993 parasitoid Mostly parasitoids of fly pupae. Goulet & Huber, 1993 Scelionidae parasitoid They generally attack the eggs of many Goulet & Huber, 1993 different types of arthropods, spiders, butterflies. Many are important in biological control. parasitoid Hosts: mostly Lepidoptera and Diptera, Goulet & Huber, 1993; though a few attack Hymenoptera, Universal Coleoptera or Neuroptera. Chalcidoidea Database, 2017, 2017 parasitoid Majority are primary parasitoids on Goulet & Huber, 1993; Hemiptera, though other hosts are attacked, Universal and details of the life history can be variable Chalcidoidea (e.g., some attack eggs, some attack pupae). Database, 2017, 2017 They are found throughout the world in virtually all habitats and are extremely important as biological control agents. parasitoid Parasites of an extremely wide range of Goulet & Huber, 1993; species. Some attack eggs, larvae, adults or Grissell & Schauff, they are hyperparasitoids. 1997; Universal Chalcidoidea Database, 2017 103

Eulophidae parasitoid The majority are primary parasitoids on a Goulet & Huber, 1993; huge range of arthropods at all stages of Universal development. Chalcidoidea Database, 2017 Mymaridae parasitoid are tiny wasps that are egg Goulet & Huber, 1993; parasitoids belonging to the Chalcidoidea, Yoshimoto, 1990; and have a global distribution. Universal Chalcidoidea Database, 2017 parasitoid They are parasitoids of phytophagous Goulet & Huber, 1993; arthropods, primarily flies. Universal Chalcidoidea Database, 2017 Perilampidae parasitoid Most are hyperparasitoids. Some are primary Goulet & Huber, 1993; parasitoids of wood-boring beetle larvae. Universal Chalcidoidea Database, 2017 parasitoid Many are parasitoids on gall-forming Goulet & Huber, 1993; arthropods. Universal Chalcidoidea Database, 2017 Platygastridae parasitoid They are generally idiobionts, attacking the Goulet & Huber, 1993 eggs of either beetles or Hemiptera. parasitoid These wasps’ larvae are parasitoids of beetle Xerces Society, 2014 larva, particularly scarab beetles, for example, Japanese beetles. parasitoid Endoparasitoids of beetle larva (Coleoptera) Goulet & Huber, 1993; or less commonly fly larva (Mycetophilidae: Masner, 1993 Diptera). parasitoid It is a poorly known group, though most are Goulet & Huber, 1993; believed to be parasitoids Masner, 1993 parasitoid Parasitoids of gall-forming Hymenoptera or Goulet & Huber, 1993; possibly primary gall-formers. Worldwide Gibson, 1993 distribution Bombyliidae pollinator Adults take nectar/pollen. Larson & Inoyue 2001 pollinator Many are valuable pollinators in natural Goulson & Ebrary, habitats and for agricultural crops. The 2010; O'Toole et al., genera found in this study are Apis (Apis 1999; Mader et al., Melifera), Bombus, Ceratina which were 2011; Wilson & Carril, classified as pollinators and which 2015 was classified as a kleptoparasite. pollinator These relatively hairless bees lack an external Goulson & Ebrary, structure to carry pollen and resemble wasps. 2010; O’Toole et al., The common genus found in this study was 1999; Mader et al., Hylaeus. 2011; Wilson & Carril, 2015 104

Papilionidae pollinator Adults of all species visit flowers for nectar. Marshall, 2006 Syrphidae pollinator Flower flies play a particularly important role Rader et al. 2016; in pollination. Marshall, 2006; Xerces Society, 2014 Andrenidae pollinator The Andrenidae (commonly known as Goulson & Ebrary, mining bees) are a large, nearly cosmopolitan 2010; O’Toole et al., family of solitary, ground-nesting bees. This 1999; Mader et al., family had genus Andrena. 2011; Wilson & Carril, 2015 pollinator All species are pollen feeders and may be Goulson & Ebrary, important pollinators. The genus Sphecodes 2010; O’Toole et al., is considered a kleptoparasite in this study. 1999; Mader et al., Other genera found was Lasioglossum, 2011; Wilson & Carril, Halictus, Augochlora, Augochloropsis and 2015 Agapostomon which were all considered pollinators. pollinator Melittids are typically small to moderate- Goulson & Ebrary, sized bees, which are well known for their 2010; O’Toole et al., specialist and oligolectic foraging habits. 1999; Mader et al., 2011; Wilson & Carril, 2015; Michener, 2000 Drepanidae pollinator Nectar feeding as adult and notes as effective Marshall, 2006 pollinators. Pieridae pollinator Most larva feed on leaves of Brassicaceae, Marshall, 2006 while some feed on Fabaceae. Nymphalidae pollinator Adults show wide range of feeding behavior; Marshall, 2006 some feed on nectar, sap fluid, or rotting fruits. Erebidae pollinator Nectar feeding on common prairie plant Marshall, 2006 species including the Asteraceae family. Hesperidae pollinator Nectar feeding on common prairie plant Marshall, 2006 species including the Asteraceae family. Lycaenidae pollinator Adults of most species are avid flower Marshall, 2006 visitors and derive most of their nutrition in this way. Empididae predator Adults have long piercing mouthparts that Downes, 1969; they use for predaceous feeding. They feed Marshall, 2006; Arnett, on other small flies, beetles, and moths. 2000 Phytoseiidae predator Feed on spider and small arthropods Xerces Society, 2014 such as thrips. They are important for biocontrol Anthicidae predator Adult beetles are omnivorous, being known Marshall, 2006 to consume small arthropods, pollen, fungi, and whatever else they can find. Some species are of interest as biological control

105

agents, as they can eat the eggs or larvae of pests. Anthocoridae predator All known to eat small arthropods. This Xerces Society, 2014; family is used for biocontrol purposes. Horton, 2008 Asilidae predator Diurnal ambush predators, feed on variety of Marshall, 2006 arthropods depending on the size. Larger species feed on beetles, wasps or dragonflies, smaller species feed on springtails or other small arthropods. Some feed preferentially on honey bees. Carabidae predator Most adults rapidly pursue their prey (other Xerces Society, 2014 arthropods) at night. A few eat pollen, berries, and seeds. Most larvae are predators. Chrysopidae predator Some adults are predators, others take liquids Xerces Society, 2014 such as honeydew and some feed on pollen. Larvae are predatory on other arthropods, especially aphids. Cleridae predator Larvae feeds on larvae of wood-burrowing Marshall, 2006 arthropods, adults feeds on adults of those wood-burrowing arthropods. Coenagrionidae predator Both nymphs and adults are predators; Marshall, 2006 nymphs feed on small aquatic arthropods, tadpoles and fish, while adults feed on flies, aphids, mosquitoes. predator Feeds on variety of arthropods such as Marshall, 2006 aphids, bees, butterflies, moths, , cockroaches, crickets, flies, and grasshoppers. Adults hunt by stinging and paralyzing, larvae feed on the prey captured by the adults. Lampyridae predator They feed on snails, earthworms, and larvae Xerces Society, 2014 of other arthropods. Some may consume nectar or pollen. predator Many of these mites are not parasitic but Marshall, 2006 free-living and predatory. Some species feed on Nematoda and Collembola. Odonata predator Dragonflies are generalists. According to Marshall, 2006 most studies, the main diet of adult Odonates consists of small arthropods, especially Diptera (flies). Opiliones predator Many species are omnivorous, eating Marshall, 2006 primarily small arthropods and all kinds of plant material and fungi. Pompilidae predator Larvae feed on spiders. Adults usually found Marshall, 2006 on flowers or on the ground searching for prey. 106

Reduviidae predator Most prey on arthropods. Their saliva is Marshall, 2006 commonly effective at killing prey substantially larger than itself. Rhagionidae predator Both adults and larvae are predaceous on a Marshall, 2006 variety of small arthropods. Scenopinidae predator Prey are small arthropods; adults take nectar Marshall, 2006 from flowers, and feed on honeydew. Sciomyzidae predator Prey on freshwater or terrestrial snails which Marshall, 2006 are pests to many crops. predator Larvae feed on paralyzed arthropods (the host Marshall, 2006 varies according to species) provided by adult; common hosts include spiders, grasshoppers, and caterpillars. Staphylinidae predator Most adults and larvae are predatory on other Marshall, 2006 invertebrates. Some larvae feed on decaying vegetation. predator Feed on arthropods including millipedes, Marshall, 2006 spiders, scorpions. Adults may take nectar. Therevidae predator Feed on beetles, butterflies, moths and flies Marshall, 2006; that live either on ground or in leaf litter. Rodrigues & Calhau, 2016 predator Adults feed masticated small arthropods and Hunt et al., 1982; caterpillars to larvae by trophallaxis. Adults Xerces Society, 2014 primarily feed on larval saliva which is rich in amino acids. Xylophagidae predator Larvae are scavengers or predators, feeding Marshall, 2006 on arthropod larvae. Cantharidae predator Adults eat nectar, pollen, other arthropods; Xerces Society, 2014 larvae are fluid-feeding predators, feed on arthropod eggs and larvae. Dolichopodidae predator Mouthparts are for piercing (with a short Marshall, 2006 proboscis). Adults and larvae prey on small arthropods. Hybotidae predator Known as predators from other studies. They Marshall, 2006 have piercing mouthparts to capture the prey. predator Both adults and nymphs are predaceous. Marshall, 2006 They feed on mites, aphids, caterpillars, as well as other arthropods eggs and nymphs. Adults both ambush and actively pursue the prey. Melyridae predator Adults evidently feed on flower-visiting Xerces Society, 2014 arthropods and pollen, larvae are primarily predators of other arthropods. Mantidae predator Almost exclusively predatory and Xerces Society, 2014 carnivorous. They feed on variety of 107

arthropods such as grasshoppers, moths, bees, wasps, katydids, and flies. Cicindelidae predator Both larvae and adults are predaceous and Jaskula, 2013; feed on other small arthropods and spiders. Marshall, 2006 Athericidae predator Larvae are aquatic and can be predaceous on Marshall, 2006 other invertebrates. Adults noted as predaceous. Histeridae predator Adults and larvae feed on other arthropods Marshall, 2006 (such as maggots), and other small invertebrates. Platystomatidae predator Some species are predators of other Marshall, 2006 arthropods, others feed on decaying vegetations and animals. Micropezidae predator Adults feed on small arthropods, while some Marshall, 2006 species are attracted to feces or decaying fruit. Geocoridae predator Opportunistic predators that show omnivory; Crocker & Whitcomb, feeds on variety of plants (grasses, and 1980; Marshall, 2006 soybeans) and arthropods (primarily mites, thrips, and aphids). Chamaemyiidae predator Larvae feed on aphids and scale arthropods, Marshall, 2006 and some species have been used for pest controls. Phymatidae predator They hunt variety of small arthropods and Yong, 2005; Marshall, other arthropods, but primarily hunt for 2006 winged adults of Diptera, Homoptera, and Hymenoptera. Libellulidae predator Both nymphs and adults are predators. Marshall, 2006 Nymphs feed on small aquatic arthropods, adults feed on variety of winged arthropods such as flies, moths, mayflies, and sometimes bees. Forficulidae predator They feed on wide range of prey including Marshall, 2006 caterpillars, foliage feeds, small arthropods. Aeshnidae predator Nymphs are aquatic predators, feeding on Marshall, 2006 aquatic arthropods and small fish. Adults feed on flying arthropods such as flies, mosquitoes, grasshoppers, and moths. Agelenidae predator Opportunistic hunters. Captures mainly Foelix, 2011; Bradley arthropods such as grasshoppers, although & Buchanan, 2013 they have been noted of preying upon other funnel weavers. Araneidae predator Any small arthropods that they are capable of Dondale et al., 2003; capturing. Larger Araneidae members in Foelix, 2011; Bradley genus Argiope is known to capture small & Buchanan, 2013 frogs and humming birds. 108

Dictynidae predator Web builder; primarily captures flies, Foelix, 2011; Bradley occasionally grasshoppers, moths, and other & Buchanan, 2013 small spiders. Hahniidae predator Web builder; they build webs on the small Foelix, 2011; Bradley depressions in the soil. They eat small & Buchanan, 2013 arthropods or other soil arthropods. Linyphiidae predator Webs are built in trees, grasses, fences, or Foelix, 2011; Bradley directly on the soil. They are known to eat & Buchanan, 2013 small arthropods such as flies, grasshoppers, small wasps, as well as soil arthropods (mainly springtails) and other spiders. Tetragnathidae predator Builds a web usually near the bodies of water Dondale et al., 2003; (rivers, ponds, lakes, swamps) and captures Foelix, 2011; Bradley flies, moths, and other arthropods. & Buchanan, 2013 Theridiidae predator They will eat any arthropods such as flies, Foelix, 2011; Bradley moths, and grasshoppers that are small & Buchanan, 2013 enough to be caught in the web. Theridisomatidae predator Web builder; they build a conical orb web Dondale et al., 2003; and captures any small arthropods that fall Foelix, 2011; Bradley onto or attempts to fly through the web. & Buchanan, 2013 Amaurobiidae predator Ambush predator; feeds on small arthropods. Foelix, 2011; Bradley & Buchanan, 2013 Anyphaenidae predator Ambush predator; mostly eats small Dondale & Rener, arthropods, although some are known to eat 1982; Foelix, 2011; flower nectar. They play a role in pest control Bradley & Buchanan, such as aphids in tree crops. 2013 Clubionidae predator Ambush predator; they hunt for small Dondale & Rener, arthropods and soil arthropods in the night. 1982; Foelix, 2011; Bradley & Buchanan, 2013 Thomisidae predator Ambush predator; they wait motionless on Dondale & Rener the leave or flowers until the prey 1978; Foelix, 2011; approaches. They eat small arthropods such Bradley & Buchanan, as flies or moths. 2013 Lycosidae predator Active predator that hunts without the web. Dondale & Rener Mainly consumes any small arthropods, as 1990; Foelix, 2011; well as sometimes other small spiders. Bradley & Buchanan, 2013 Philodromidae predator Active predator; they hunt similar prey as Dondale & Rener Thomisidae (small arthropods such as flies or 1978; Foelix, 2011; moths), and they can also pursue escaping Bradley & Buchanan, prey. 2013 Salticidae predator Active predators as they are excellent at Foelix, 2011; Bradley climbing and jumping. They eat mainly small & Buchanan, 2013 arthropods such as flies, beetles and crickets.

109

Phalangiidae predator Active predator. Common in disturbed, Marshall, 2006; Foelix, anthropogenic habitats (e.g., agricultural 2011; Bradley & fields, urban areas). Buchanan, 2013 Panorpidae uncommon Adult mecopterans are mostly scavengers, Cheung et al., 2006; feeding on decaying vegetation and the soft Dunford et al., 2008 bodies of dying invertebrates. Ixodidae uncommon Feed on blood of medium- to large-sized Marshall, 2006 mammals by attaching themselves. Clusiidae uncommon Most species feed on nectar, sap, rotting Marshall, 2006 vegetation, and feces of birds and mammals, while others are known to feed on other arthropods such as beetle larvae. Pscoptera uncommon This order is common on bark and on foliage Marshall, 2006 but are not an economically important group. Little is known about any damage to crops in this group. Hippoboscidae uncommon Feed on blood of birds or mammals, often Marshall, 2006 associated with road-killed birds. Erotylidae uncommon Larvae and adults feed on the fruiting bodies Arnett et al., 2002 of fungi growing in decaying wood. Piophilidae uncommon Both adult and larvae feed on decaying Marshall, 2006 materials ranging from fungi to mammal. Heleomyzidae uncommon Usually in shaded areas, typically near rotting Marshall, 2006 matter (compost, dung, carrion etc.) or fungi. Acartophthalmidae uncommon The biology of Acartophthalmidae is almost Marshall, 2006 unknown. The adults have mainly been found

in forests. Chyromyidae uncommon A small but widely distributed group of Marshall, 2006 rather uncommon flies.

110

Appendix 5: Supplementary methods and results

Table S5: Proportion of agricultural and semi-natural landcover at four buffer radii (500 m, 1000 m, 1500 m, and 2000 m) of 13 farms in Southern Ontario, Canada. Landcover data was extracted from Agri-Canada Annual Crop Inventory.

Farm ag500 ag1000 ag1500 ag2000 nat500 nat1000 nat1500 nat2000 1 57.00 37.41 36.99 26.20 36.12 58.69 60.27 71.53 2 42.75 45.86 45.86 49.98 54.05 52.49 52.49 46.45 3 59.26 44.44 44.44 48.42 35.06 50.74 50.74 47.34 4 45.39 65.48 65.48 74.95 49.38 30.95 30.95 21.28 5 38.92 49.32 49.32 49.13 42.86 26.36 26.36 35.75 6 10.19 14.10 34.22 31.25 83.33 63.44 59.14 65.40 7 24.69 39.43 39.43 41.46 72.10 58.40 58.40 55.93 8 84.84 74.63 74.63 73.64 8.31 21.49 21.49 22.35 9 69.23 39.79 39.79 50.76 26.55 55.52 55.52 43.10 10 27.59 37.95 37.95 54.57 59.11 52.87 52.87 34.64 11 54.63 58.01 58.01 54.50 41.46 38.35 38.35 42.11 12 66.09 77.44 77.44 77.14 16.71 14.32 14.32 17.61 13 64.52 60.04 60.04 64.84 32.51 36.28 36.28 32.17

*Note: ag= agricultural landcover, nat= semi-natural cover

111

Table S6: Lab methods for plant tissue quality analysis as provided by SGS Labs, Guelph, Ontario.

Test Analysis Description Purpose Nitrogen % Dumas Sample is weighed than heated Nitrogen (N) plays a key role in the method at temperatures ranging plant life cycle. It is the main plant between 800–1,000 °C. Next is mineral nutrient needed for a water removal procedure chlorophyll production and other followed by gas separation. The plant cell components (proteins, last step is to measure gaseous nucleic acids, amino acids). A strong nitrogen concentration. A positive relationship between foliar N thermal conductivity detector is concentration and arthropod used in which differences in survivorship, development, growth, thermal conductivity of the and reproductive (Lerdau & Throop, gases are detected (Shea & 2004) rates has been demonstrated for Watts, 1939). many arthropod species. Phosphorus Inductively Metals and other elements in Phosphorus participates in a number % coupled plant tissues by ICP – Sample is of processes determining the growth, plasma (ICP) dry ashed, followed by acid development and the productivity of spectrometry digestion with HCl. Minerals the plant: formation of cell nucleus are then analyzed by and cell multiplication, synthesis of inductively couple plasma after lipids and specific proteins, appropriate dilution. transmission of hereditary properties, breathing and photosynthesis, energy transmission from richer to poorer energetic compounds (Taiz et al., 2015) Organic Loss on The percentage organic matter Including the effects of herbivory in carbon % ignition is the ratio of the weight loss ecosystem studies, particularly in during ignition to the weight of systems where rates of herbivory are plant tissue dried at 10SoC X high and linked to plant C:N. 100. The ignition method was Deposition induced shifts in plant adapted from Ball (1964) C:N may therefore strongly affect plant–herbivore interactions (Lerdau & Throop, 2004) ADF % ANKOM AOAC 973.18 – Fiber (Acid There are large differences in Technology Detergent) and ADF in Animal lignification between grasses and Method 12. Feed. It is the residue remaining legumes and differences in the impact Filter Bag after boiling a forage of ADF per se on their forage quality. Technique sample acid detergent solution. This can influence herbivore and ADF. contains cellulose, lignin, arthropod foraging. Carbon-based and silica, not hemicellulose. It secondary compounds such as often is used to calculate phenolics and ADF are important digestibility. ADF is the controls of palatability (Schädler et percentage of plant material al., 2014) which is insoluble (Moore & Jung, 2001)

112

Table S7: Summary of arthropod functional group abundance data. All data is presented as the pooled number of individuals caught per site for both yellow pan traps and sweep nets samples.

Farm Anthophilous Decomposer Kleptoparasite Aquatic Uncommon Herbivore Parasitoid Pollinator Predator

1 223 11 8 1 0 495 57 43 175

2 111 3 2 0 0 237 23 138 69

3 85 3 1 3 0 355 53 141 52

4 61 2 1 0 1 113 29 21 44

5 57 3 1 0 1 90 12 27 42

6 215 11 3 1 3 924 17 76 77

7 468 51 2 1 7 1707 141 158 353

8 172 3 1 0 1 1083 40 121 226

9 193 24 4 0 1 1261 72 85 177

10 146 48 2 0 0 793 46 121 214

11 51 14 1 0 9 799 28 74 150

12 135 13 0 0 2 1332 100 190 316

13 255 7 4 0 1 1628 96 122 218

14 176 224 1 0 0 433 43 73 184

15 77 6 3 0 0 195 21 36 56

16 166 10 0 2 0 196 26 72 169

17 443 14 2 0 1 206 142 25 81

18 100 11 0 2 4 363 36 9 111

19 311 6 4 8 5 433 41 64 75

20 62 4 0 0 0 114 8 15 23

21 243 20 11 1 3 2421 200 149 290

22 177 10 1 5 0 836 54 104 275

113

Table S8: Summary of arthropod functional group richness (family-level) data. All data is presented as the pooled number of individuals caught per site for both yellow pan traps and sweep nets samples.

Farm Anthophilous Decomposer Kleptoparasite Aquatic Uncommon Herbivore Parasitoid Pollinator Predator

1 17 4 3 1 0 29 10 8 22

2 12 2 2 0 0 18 9 10 15

3 9 1 1 1 0 25 9 13 20

4 10 2 1 0 1 21 13 9 15

5 11 2 1 0 1 20 9 9 13

6 17 4 3 1 3 24 6 11 23

7 17 6 1 1 2 34 19 9 36

8 12 2 1 0 1 43 16 9 29

9 14 3 2 0 1 31 14 13 25

10 12 4 1 0 0 38 15 9 37

11 13 3 1 0 1 29 9 9 24

12 15 2 0 0 2 30 11 11 35

13 18 6 2 0 1 35 11 10 39

14 12 2 1 0 0 25 15 3 26

15 9 2 3 0 0 17 8 6 17

16 14 4 0 1 0 22 5 8 19

17 15 2 2 0 1 22 16 7 22

18 12 5 0 1 3 25 10 4 19

19 13 4 3 2 1 27 11 16 20

20 8 2 0 0 0 14 3 5 11

21 22 4 2 1 3 48 23 15 43

22 17 3 1 2 0 30 12 14 29

114

Table S9: Results of arthropod family identity in agricultural and non-farm (Naturalist) sites and the shared or unique families between them.

Agricultural (unique) iNaturalist (unique) (n=94) (n=104) Shared (n=89) Anthophilous Anthophilous Anthophilous Bibionidae Gasteruptiidae Calliphoridae Conopidae Limoniidae Ceratopogonidae Lonchopteridae Platypezidae Chironomidae Mycetophilidae Psychodidae Culicidae Phalacridae Pterophoridae Formicidae Phlaeothripidae Stratiomyidae Mordellidae Phoridae Aquatic Muscidae Sarcophaginae Baetidae Sarcophagidae Scatopsidae Baetiscidae Sepsidae Sphaeroceridae Bdellidae Simuliidae Aquatic Tabanidae Ephydridae Blastobasidae Tipulidae Nemouridae Bolboceratidae Aquatic Perlidae Caenidae Hydropsychidae Taeniopterygidae Decomposer Decomposer Corydalidae Elateridae Clusiidae Dytiscidae Fanniidae Dryomyzidae Elmidae Lauxaniidae Entomobryidae Ephemerellidae Herbivore Leiodidae Ephemeridae Acrididae Lonchaeidae Anthomyiidae Sminthuridae Gyrinidae Aphididae Herbivore Haliplidae Berytidae Agridae Heptageniidae Buprestidae Agromyzidae Hydrophilidae Cecidomyiidae Alydidae Leptoceridae Cerambycidae Anobiidae Cercopidae Chloropidae Chrysomelidae Cimbicidae Cicadellidae Delphacidae Phryganeidae Cixiidae Drosophilidae Psephenidae Coreidae Galerucie Curculionidae Issidae Decomposer Gryllidae Languriidae Lygaeidae Largidae Silphidae Meloidae Nitidulidae Herbivore Membracidae 115

Oedemeridae Acanthosomatidae Miridae Pergidae Pentatomidae Phaleothripidae Cicadidae Rhopalidae Phasmatodea Scarabaeidae Psilidae Crambidae Tenthredinidae Psyllidae Tettigoniidae Rhyparochromidae Thyreocoridae Scathophagidae Tingidae Scutelleridae Elachistidae Kleptoparasite Siricidae Apidae Tephritidae Gelechiidae Chrysididae Thripidae Gracillariidae Parasitoid Ullidiidae Lasiocampidae Cynipidae Xyelidae Lycidae Eulophidae Kleptoparasite Eurytomidae Miltogramminae Pseudococcidae Ichneumonidae Parasitoid Pyralidae Perilampidae Bethylidae Rhaphidophoridae Platygastridae Braconidae Tetranychidae Pollinator Tetrigidae Bombyliidae Chalcididae Tortricidae Halictidae Diapriidae Parasitoid Lycaenidae Encyrtidae Carnidae Nymphalidae Eucoilidae Pelecinidae Papilionidae Eupelmidae Uropodidae Syrphidae Figitidae Pollinator Predator Megaspilidae Erebidae Araneidae Mymaridae Geometridae Asilidae Orussidae Hesperiidae Cantharidae Pipunculidae Noctuidae Carabidae Proctotrupidae Notodontidae Chrysopidae Sphingidae Clubionidae Scelionidae Predator Coccinellidae Aeshnidae Coenagrionidae Torymidae Ascidae Crabronidae Pollinator Coniopterygidae Dictynidae Andrenidae Damaeidae Dolichopodidae Colletidae Eremaeidae Empididae Megachilidae Euphthiracaridae Lampyridae Pieridae Eupodidae Linyphiidae Predator Gnaphosidae Lycosidae Agelenidae Gomphidae Phalangiidae 116

Amaurobiidae Hemerobiidae Philodromidae Anthicidae Histeridae Pholcidae Anthocoridae Hypochthoniidae Pompilidae Anyphaenidae Lestidae Bittacidae Libellulidae Rhagionidae Cleridae Macromiidae Salticidae Coegrionidae Mimetidae Sciomyzidae Forficulidae Myrmeleontidae Sphecidae Hybotidae Oribotritiidae Staphylinidae Liocranidae Pachylaelapidae Tachinidae Melyridae Panorpidae Tetragnathidae Mesostigmata Parasitidae Theridiidae Micropezidae Thomisidae Nabidae Phytoseiidae Vespidae Phymatidae Pisauridae Uncommon Platystomatidae Chyromyidae Scenopinidae Scheloribatidae Heleomyzidae Tetragthidae Scutacaridae Ixodidae Therevidae Siteroptidae Tenebrionidae Theridisomatidae Sperchontidae Xylophagidae Trematuridae Uncommon Trombidiidae Bostrichidae Tydeidae Erotylidae Uropodellidae Veigaiidae Uncommon Astegistidae Coleophoridae Eniochthoniidae Eriophyidae Psocidae Psychidae

117

Figure S8: Correlation matrix of landcover data carried out to assess the relationship among predictor variables. Squares in blue indicate a strong negative correlation, squares in red indicate a strong positive correlation.

118

Figure S9: Correlation matrix of plant resource data carried out to assess the relationship among predictor variables. Squares in blue indicate a strong negative correlation, squares in red indicate a strong positive correlation.

119

Southern Ontario, Canada

10,000 km²

Figure S10: Locations of insect and arachnid iNaturalist observations logged from 05 May - 17 July 2017 spanning a 10,000 km² region of Southern Ontario, Canada.

120

A

Totalarthropod abundance

B

level)

-

(family

richness Totalarthropod

Farm site Figure S11: Summary of arthropod (A) abundance and (B) richness (family-level) per farm from a total of 22 farms in Southern Ontario, Canada.

121