The Pennsylvania State University

The Graduate School

College of Agricultural Sciences

POLLEN NUTRITION,

THE FOUNDATION OF BUMBLE BEE FORAGING

A Dissertation in

Entomology

by

Anthony Damiano Vaudo

© 2016 Anthony Damiano Vaudo

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

December 2016

The dissertation of Anthony Damiano Vaudo was reviewed and approved* by the following:

Christina M. Grozinger Distinguished Professor of Entomology Dissertation Co-Adviser Chair of Committee

John F. Tooker Associate Professor of Entomology Dissertation Co-Adviser

Harland M. Patch Research Associate of Entomology Special Member

David A. Mortensen Professor of Weed and Applied Plant Ecology

Heather Hines Assistant Professor of Biology and Entomology

Gary W. Felton Professor and Department Head of Entomology

*Signatures are on file in the Graduate School

ii Abstract

Angiosperms and insect pollinators, especially bees, share a rich ecological and evolutionary history in which the radiation of the groups occurred through coevolutionary processes. This is because flowers facilitate reproduction through the transfer of pollen by attracting bees to flowers, and providing bees their entire source of nutrition. Historically, it was believed that bees were innately destined to visit flowers that provided specific or attractive morphologies, colors, or scents, known as pollination syndromes. However, individuals within bee species may visit and collect resources from different plant species during the day, season, and across years. Bee nutrition is partitioned between floral nectar which provides energy (carbohydrates) for foraging bees to collect nutritionally complex pollen (protein, lipids, and micronutrients). Because pollen quality differs between plant species and affects the health and development of bee larvae and adults, we expect that bee species forage to collect the right balance of pollen nutrients from their host-plant species. Therefore, if bees have species-specific nutritional needs, it may explain the foraging patterns among host-plant species observed. In this dissertation, I reveal that bumble bees (Bombus impatiens, Hymenoptera: Apidae) may actively select and balance their pollen diet among host-plant species to meet their optimum nutritional needs of protein and lipids, or protein:lipid ratio. This discovery was confirmed by 1) relating bumble bee pollen foraging preferences among host-plant species to pollen nutritional quality 2) assessing bumble bee preferences for pollen isolated from the host-plant species, 3) assessing bumble bee preferences to single source pollen modified to different protein:lipid ratios and concentrations, 4) assessing bumble bee regulation of synthetic diets modified to different protein:lipid ratios and concentrations, and finally 5) verifying these results in field studies by determining the nutritional value of pollen collected by free foraging colonies in which colonies defended their nutrient intake ratios independent of landscape. This dissertation presents a new theory that bees may be floral generalists, but nutritional specialists, adding a new dimension to previous assumptions of pollination syndromes. The particular floral traits involved in syndromes (i.e. color, morphology, scent) may have evolved as discriminative stimuli that bees associate with reward quality.

iii TABLE OF CONTENTS

List of Tables ...... vi List of Figures ...... vii Preface ...... ix References ...... xiii Acknowledgements ...... xv Chapter 1. Bee nutrition and floral resource restoration ...... 1 Highlights ...... 1 Abstract ...... 1 Introduction ...... 2 Bee Nutrition ...... 2 Floral Resource Nutritional Diversity and Bee Foraging Behavior ...... 5 Nectar ...... 5 Pollen ...... 7 The Importance of Plant Diversity for Bee Health ...... 9 Applying Bee Nutrition to Floral Resource Habitat Restoration ...... 10 References ...... 15 Chapter 2. Bumble Bees Exhibit Daily Behavioral Patterns in Pollen Foraging ...... 27 Abstract ...... 27 Introduction ...... 28 Materials and methods ...... 31 Insect and plant species ...... 31 Hoop house ...... 32 Foraging data collection ...... 33 Floral display ...... 35 Foraging data metrics ...... 36 Results ...... 37 Discussion ...... 38 The implications of daily foraging patterns for studies of plant-pollinator interactions ...... 38 Foraging patterns reveal interplay between pollen quality and abundance ...... 41 Conclusion ...... 44 References ...... 50 Chapter 3. Macronutrient ratios in pollen shape bumble bee (Bombus impatiens) foraging strategies and floral preferences ...... 54 Abstract ...... 54 Significance Statement ...... 55 Introduction ...... 56 Results ...... 59 Host-plant pollen foraging preferences ...... 59 Isolated pollen feeding preference assay ...... 60

iv Modified pollen feeding preference assay ...... 61 Discussion ...... 63 Materials and Methods ...... 68 Pollen nutritional analysis ...... 68 Assessing host-plant pollen foraging preferences ...... 71 Isolated pollen feeding preference assay ...... 72 Modified pollen feeding preference assay ...... 74 References ...... 84 Chapter 4. Bumble bees regulate their intake of the essential protein and lipid pollen macronutrients ...... 90 Abstract ...... 90 Introduction ...... 91 Methods...... 95 General bee rearing conditions ...... 95 Single P:L diet assay ...... 95 Paired P:L diets assay ...... 97 Statistical analysis ...... 97 Results ...... 99 Single P:L diet assay ...... 99 Paired P:L diets assay ...... 101 Discussion ...... 102 References ...... 121 Chapter 5: Bumble bees (Bombus impatiens) defend pollen nutritional preferences in the field while nutritional intake drives growth and reproduction ...... 125 Abstract ...... 125 Introduction ...... 126 Methods...... 129 Site selection and bumble bee placement: ...... 129 Data collection: ...... 130 Statistical analyses: ...... 131 Results and Discussion ...... 134 Pollen nutrition: ...... 134 Nutrition, behavior, growth, and reproduction: ...... 135 Case study of two ‘best’ field sites ...... 138 Conclusion ...... 140 Acknowledgments ...... 141 References ...... 150 Chapter 6: Discussion and Future Research ...... 154 Discussion ...... 154 Future Research ...... 156 References ...... 161

v List of Tables

Table 2-1. Plant species used for Bombus impatiens foraging observations ...... 46 Table 3-1. Host-plant species and associated pollen nutritional values used in “Host-plant pollen foraging preferences” and “Isolated pollen feeding preference assay” ...... 76 Table 4-1. Diet recipes...... 109 Table 4-2. Mean (± SE) daily consumption (mg) of nutrients for B. terrestris foragers in “Single P:L diet assay” ...... 110 Table 4-3. Cox – regression of survival for B. terrestris foragers in “Single P:L diet assay” ... 111 Table 4-4. Consumption (g; mean ± SE) by B. impatiens and B. terrestris foragers in the “Paired P:L diets assay” and protein:carbohydrate (P:C) and protein:lipid (P:L) intake ratios over seven days ...... 112 Table 5-1. Summary of pollen nutritional quality collected by B. impatiens colonies by habitat and reproductive success...... 142 Table 5-2. Summary of growth, reproduction, and foraging rates of B. impatiens colonies ...... 143

vi List of Figures

Figure 1-1. Conceptual schematic presenting a holistic framework relating basic research and landscape application for bee conservation and habitat restoration...... 14 Figure 2-1. Bombus impatiens foragers actively collecting pollen from Senna hebecarpa, the most preferred host-plant species in 2013...... 47 Figure 2-2. Design of B. impatiens foraging preference experiments...... 47 Figure 2-3. Bombus impatiens general community visitation rates and individual visit durations (means ± SE) (independent of time of day) by plant species in 2012 and 2013 (Table 2-1 defines plant codes)...... 48 Figure 2-4. Interactions of community visitation rates and individual visit durations by plant species and time of day in 2012 and 2013 (Table 2-1 defines plant codes)...... 49 Figure 3-1. The relationship between Bombus impatiens pollen foraging rates and pollen nutritional quality (Host-plant pollen foraging preferences)...... 77 Figure 3-2. Bombus impatiens pollen feeding preferences on isolated pollen is associated with pollen P:L ratios (Isolated pollen feeding preference assay)...... 78 Figure 3-3. In “Isolated pollen feeding assay,” the frequency that the pollen with the higher P:L ratio was preferentially selected by the bees relative to the pollen with the lower P:L pollen...... 79 Figure 3-4. Bombus impatiens pollen consumption of species based on “nutritional rank” for “Isolated pollen feeding preference assay”...... 80 Figure 3-5. Bombus impatiens pollen feeding events (A) and consumption (B) between nutritionally modified pollen diets (Modified pollen feeding preference assay)...... 81 Figure 3-6. Pairwise comparisons of feeding events (A) and consumption (B) of all pollen diet combinations in “Modified pollen feeding preference assay”...... 82 Figure 3-7. The relationship between protein (A) and lipid (B) concentration and amount of pollen consumed by B. impatiens (Modified pollen feeding preference assay)...... 83 Figure 4-1. Mean (± SE) daily consumption of diets across treatments for B. terrestris foragers in “Single P:L diet assay.” ...... 113 Figure 4-2. Nutritional arrays of B. terrestris foragers surviving seven days in “Single P:L diet assay.” ...... 114 Figure 4-3. Survival curve of B. terrestris foragers in “Single P:L diet assay.” ...... 115 Figure 4-4. Mean (± SE) cumulative consumption of nutrients by deceased (N = 11) and surviving (N = 4) B. terrestris foragers in 1:10 P:L treatment on Day 3 of “Single P:L diet assay”: 116 Figure 4-5. Mean (± SE) daily consumption of diets across treatments for a) B. impatiens and b) B. terrestris foragers in “Paired P:L diets assay.” ...... 117 Figure 4-6. Daily trajectories of B. impatiens (a-c) and B. terrestris (d-f) in “Paired P:L diets assay.” ...... 118 Figure 4-7. Mean (± SE) cumulative consumption nutrients of B. impatiens and B. terrestris foragers in “Paired P:L diets assay” that survived for seven days...... 119 Figure 4-8. Survival curve of B. impatiens and B. terrestris foragers in “Paired P:L diets assay.” ...... 120 Figure 5-1. Field sites of Bombus impatiens colonies in central Pennsylvania...... 144 Figure 5-2. Distributions of protein:lipid ratios (P:L) from individual Bombus impatiens forager corbiculate pollen loads...... 145

vii Figure 5-3. Principal component analysis for season long Bombus impatiens colony development, behavior, and nutrition...... 146 Figure 5-4. Colony level nutritional intake is highly correlated to colony (A) lifetime population and (B) maximum biomass gain...... 147 Figure 5-5. Nutritional intake positively affects colony reproduction...... 148 Figure 5-6. Colonies within sites show similarities in behavior and growth trajectories, but variation across sites...... 149

viii Preface

Foraging bees are faced with unique challenges when foraging; they must collect all nutrients for themselves and developing offspring from floral resources (pollen and nectar). Bee nutrition is nutrition is partitioned between floral nectar and pollen, requiring different handling and foraging techniques (1). Nectar is a primary carbohydrate source used to fuel bee foraging efforts (2).

Pollen, however is the primary source of all protein, lipid, vitamin, mineral, and other micronutrient needs and is essential for larval development and adult health (3-10). Plants on the other hand utilize many techniques to attract and repel pollinators (to ensure heterospecific pollen transfer for reproduction) by providing color, scent, and morphological cues, and variation in resource quality and composition (11). Bees must sift through these cues to assess and collect appropriate resources, and because floral resources vary in quality, quantity, and spatiotemporal availability among plant species, the bees are challenged to find the appropriate nutrients. Because variation in resource quality affects larval development and survival (4, 8, 12), at the fundamental level, bee foraging behavior is dedicated to collecting appropriate nutrition, particularly pollen, for reproductive success. Therefore, using the Common Eastern Bumble Bee, Bombus impatiens

Cresson (Hymenoptera:Apidae), this dissertation tests the hypothesis that bee foraging behavior is driven by pollen nutritional quality that matches their nutritional needs.

Chapter 1 is a literature review, synthesis, and conceptual framework for bee nutritional ecology and conservation (10). Here, we review the basic nutritional needs of bees, resource nutritional variability among plant species, and evidence and knowledge gaps in the literature regarding how bees forage for their nutritional needs. We present a framework addressing key research that should be conducted to determine bee species nutritional needs and how to select plant species that address

ix these needs in habitat restoration projects. This chapter provides an introduction to the research presented in the following chapters.

Chapter 2 is a research manuscript (13) in which we develop and test methodology to definitively determine B. impatiens pollen foraging preferences among host-plant species. By accounting for a number of factors influencing foraging behavior we determine that bumble bees have clear host- plant species preferences, but these can change throughout the day based on resource availability.

The data in this chapter is crucial for testing the nutritional basis of foraging behavior.

Chapter 3 is a research manuscript in which we relate the foraging data from Chapter 2 to pollen nutritional quality. We discover that B. impatiens foraging preferences are based on the relative protein and lipid concentration of pollen, or P:L ratio. We verify these preferences from isolated pollen from the plant species used in the study, and nutritionally modified pollen.

Chapter 4 is a research manuscript (in collaboration with Newcastle University) which we test B. terrestris (a common European bumble bee species) behavioral response and survival on defined synthetic diets of a range of P:L values within and beyond values expected from pollen. We also test the ability of both B. impatiens and B. terrestris to self-select their diet among different diets and regulate their protein and lipid intake to optimal P:L ratios. These data corroborate our findings from the previous chapter and indicate clear nutritional preferences of bees in the absence of all environmental and floral cues.

Finally, Chapter 5 is a research manuscript, in which we determine the nutrition of pollen collected

x by free foraging B. impatiens colonies placed in diverse landscapes in central Pennsylvania. In this study we find that bumble bees defend their nutritional preferences despite likely variation in nutritional resources in different landscapes. We discover that nutritional intake, a function of nutrition and foraging rates, is highly predictive of colony growth, reproductive success, and reproductive output.

In Chapter 6, I provide a synthesis of the research discoveries discuss the future research possibilities that this research will inspire for my own career and other researchers.

The research studies presented herein have implications for the fields of evolutionary ecology and conservation. By demonstrating that foraging behavior of bees is based on nutritional ecology, we can begin to understand how bees are distributed among plant communities in network studies (14-

18). We can explore if bees occupy different nutritional niches (i.e. have species-specific nutritional requirements) (19) and understand how bees compensate for competition by utilizing varied floral resources in these communities to achieve sufficient nutrition. Furthermore, we can use these studies to better understand how bee foraging influences the evolution of floral cues, where different bee species may preferentially choose different plant species for floral nutrition and thereby place higher selective pressure on floral cues and specific reward quality (20).

Understanding bee nutritional ecology also has direct applications for habitat restoration of altered landscapes and conservation of pollination services for agriculture. The majority of all plant species are animal pollinated (21) and bees are responsible for the majority of these services, making bees integral to the stability of ecosystems. However, the recent concern for global bee

xi decline has made us aware of the problems of land use intensification, particularly agriculture, that reduce the floral abundance and diversity on which bee species rely (22). To appropriately and efficiently address habitat restoration to promote stable plant-pollinator communities, or support crop pollinators in agricultural landscapes, we must provide bee communities with appropriate nutrition to ensure reproduction and healthy development over generations (10). Therefore, it is crucial to understand bee species’ specific nutritional needs, the plant species that provide them, and how the groups interact to create an effective conservation program.

xii References

1. Raine NE, Chittka L (2006) Pollen foraging: learning a complex motor skill by bumblebees (Bombus terrestris). Naturwissenschaften 94(6):459–464.

2. Nicolson SW, Nepi M, Pacini E (2007) Nectaries and Nectar eds Nicolson SW, Nepi M, Pacini E (Springer Science & Business Media, Dordrecht).

3. Roulston TH, Cane JH (2000) Pollen nutritional content and digestibility for animals. Plant Syst Evol 222(1):187–209.

4. Génissel A, Aupinel P, Bressac C, Tasei JN, Chevrier C (2002) Influence of pollen origin on performance of Bombus terrestris micro-colonies. Entomol Exper Applic 104(2-3):329– 336.

5. Tasei J-N, Aupinel P (2008) Nutritive value of 15 single pollens and pollen mixes tested on larvae produced by bumblebee workers (Bombus terrestris, Hymenoptera: Apidae). Apidologie 39(4):397–409.

6. Alaux C, Ducloz F, Crauser D, Le Conte Y (2010) Diet effects on honeybee immunocompetence. Biology Letters 6(4):562–565.

7. Di Pasquale G, et al. (2013) Influence of Pollen Nutrition on Honey Bee Health: Do Pollen Quality and Diversity Matter? PLoS ONE 8(8):e72016.

8. Vanderplanck M, et al. (2014) How does pollen chemistry impact development and feeding behaviour of polylectic bees? PLoS ONE 9(1):e86209.

9. Frias BED, Barbosa CD, Lourenço AP (2015) Pollen nutrition in honey bees (Apis mellifera): impact on adult health. Apidologie 47(1):15–25.

10. Vaudo AD, Tooker JF, Grozinger CM, Patch HM (2015) Bee nutrition and floral resource restoration. Current Opinion in Insect Science 10:133–141.

11. Willmer P (2011) Pollination and Floral Ecology (Princeton University Press).

12. Praz CJ, Müller A, Dorn S (2008) Specialized bees fail to develop on non-host pollen: do plants chemically protect their pollen? Ecology 89(3):795–804.

13. Vaudo AD, Patch HM, Mortensen DA, Grozinger CM, Tooker JF (2014) Bumble bees exhibit daily behavioral patterns in pollen foraging. Arthropod-Plant Interactions 8(4):273– 283.

14. Tiedeken EJ, Stout JC (2015) Insect-flower interaction network structure is resilient to a temporary pulse of floral resources from invasive Rhododendron ponticum. PLoS ONE 10(3):e0119733.

15. Waser NM, Chittka L, Price MV, Williams NM, Ollerton J (1996) Generalization in

xiii pollination systems, and why it matters. Ecology 77(4):1043-1060.

16. Baldock KCR, Memmott J, Ruiz-Guajardo JC, Roze D, Stone GN (2011) Daily temporal structure in African savanna flower visitation networks and consequences for network sampling. Ecology 92(3):687–698.

17. Petanidou T, Kallimanis AS, Tzanopoulos J, Sgardelis SP, Pantis JD (2008) Long-term observation of a pollination network: fluctuation in species and interactions, relative invariance of network structure and implications for estimates of specialization. Ecol Letters 11(6):564–575.

18. Russo L, DeBarros N, Yang S, Shea K, Mortensen D (2013) Supporting crop pollinators with floral resources: network-based phenological matching. Ecol Evol 3(9):3125–3140.

19. Behmer ST, Joern A (2008) Coexisting generalist herbivores occupy unique nutritional feeding niches. Proc Natl Acad Sci USA 105(6):1977–1982.

20. Johnson SD, Nicolson SW (2008) Evolutionary associations between nectar properties and specificity in bird pollination systems. Biology Letters 4(1):49–52.

21. Ollerton J, Winfree R, Tarrant S (2011) How many flowering plants are pollinated by animals? Oikos 120(3):321–326.

22. Biesmeijer JC, et al. (2006) Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 313(5785):351–354.

xiv Acknowledgements

I would like to thank my advisors Christina Grozinger and John Tooker for their guidance, advocacy, and support throughout my PhD program; you have been some of the best people, you make doing a job an art, you take advising and mentorship seriously, and you keep things so positive, I hope I can at least take a fraction of what you taught me into the future; thanks for keeping me afloat and on the path. I would like to thank Harland Patch for his philosophical and intellectual support. I thank my committee members David Mortensen and Heather Hines for their invaluable insight. I would like to thank past and present members of both the Grozinger and Tooker labs for their support and objective honesty in reviewing my research and writing. Thanks to Andrew Aschwanden, Philip Moore, Bernardo Niño, and Mario Padilla for being great lab managers, and Heike Betz for extending it into teaching me good science and real friendship. I would like to thank Jeri Wright for allowing me to work in her lab and her contribution to an important collaborative research project; and to Daniel Stabler for helping put my ideas to work. Special thanks to all my research assistants over the years: Bekki Waskovich, Kerry Simcock, Caitlin Jade Oliver, Victoria Bolden, Edwin Hochstedt, Simone Pope, Mackenzie Hodges and Liam Farrell. Thanks to Jan Getgood for rearing such wonderful and healthy plants, and Scott DiLoreto for giving them a winter home (and teaching me everything I never knew about growing plants). Thanks to Sebastien Jacob for providing so many bees to the cause. To Scott Smiles, you are truly missed. Special shout-out to the quintet: Sara Marzioli, Joche Gayles, Paul Ayayee (Papa Paul, the Professor, Alter Boy), and Michael Coccia; you held me together for so long and have been some of the best friends in my life. And yes, thanks to all the haters, where would I be without you. But really, thank you universe for providing the wonderful fantastically beautiful world of mystery for me to study. (“People call me rude, I wish we all were nude, I wish there was no black and white, I wish there were no rules”- Prince) Finally, to Mom and Dad, I love you.

xv

To All the Beautiful Things

xvi Chapter 1. Bee nutrition and floral resource restoration

Anthony D. Vaudo*, John F. Tooker, Christina M. Grozinger, Harland M. Patch

Department of Entomology, Center for Pollinator Research, The Pennsylvania State University, 501 ASI Building, University Park, PA 16802, USA

Highlights

• Bees’ specific nutritional requirements are partitioned between pollen and nectar

• Nectar and pollen provide carbohydrates, protein, lipids, and micronutrients

• Bee foraging behavior is influenced by nectar and pollen quality and quantity

• Diverse diets can improve bee health, reproduction, and resilience to stress

• Nutritionally balanced conservation plantings can support diverse bee populations

Abstract

Bee-population declines are linked to nutritional shortages caused by land-use intensification, which reduces diversity and abundance of host-plant species. Bees require nectar and pollen floral resources that provide necessary carbohydrates, proteins, lipids, and micronutrients for survival, reproduction, and resilience to stress. However, nectar and pollen nutritional quality varies widely among host-plant species, which in turn influences how bees forage to obtain their nutritionally appropriate diets. Unfortunately, we know little about the nutritional requirements of different bee species. Research must be conducted on bee species nutritional needs and host-plant species resource quality to develop diverse and nutritionally balanced plant communities. Restoring appropriate suites of plant species to landscapes can support diverse bee species populations and their associated pollination ecosystem services.

1 Introduction

A key factor driving pollinator declines is anthropogenic land-use intensification, which, among interacting factors such as pesticide use and introduced pests and pathogens, dramatically reduces the diversity and abundance of flowering plant species [1-7]. Bees (Hymenoptera: Apoidea:

Anthophila), as a monophyletic group of ~20,000 species [8], depend entirely on nutrition derived from floral resources (especially nectar and pollen) obtained from diverse plant species [9]. Bees therefore experience nutritional stress when limited in their choices of host-plant species or when only suboptimal floral resources are available, both of which could result in reduced population sizes and pollination efficiency [1-7]. We propose a rational approach for restoring and conserving pollinator habitat that focuses on bee nutrition by 1) determining the specific nutritional requirements of different bee species and how nutrition influences foraging behavior and host- plant species choice, and 2) determining the nutritional quality of pollen and nectar of host-plant species. Utilizing this information, we can then 3) generate targeted plant communities that are nutritionally optimized for pollinator resource restoration and conservation. Here, we review recent literature and knowledge gaps on how floral resource nutrition and diversity influences bee health and foraging behavior. We discuss how basic research can be applied to develop rationally designed conservation protocols that support bee populations.

Bee Nutrition

Adults and larvae of nearly all bee species depend on nutrients obtained from floral resources for development, reproduction, and health [9,10]. Adult foragers are challenged with seeking out appropriate nutrients from the environment for developing larvae and/or nurse bees and queens confined to a nest [9]. At the simplest level, bee nutrition is partitioned between nectar and pollen:

2 nectar provides bees’ main source of carbohydrates, whereas pollen provides proteins, lipids, and other micronutrients [11-13]. To obtain optimal nutrition, insects can balance their nutrient intake from complementary food sources, which is considered one of the most important factors shaping foraging behavior and insect fitness [14].

Bee species likely have different quantitative and qualitative nutritional requirements, which is suggested by their differences in life history, brood size, social structure, and different distributions amongst plant species. Whereas most bees are solitary and oligolectic (a single reproductive female lays eggs and provisions brood; specializes on one plant family or genus), the majority of literature studying the nutritional needs of bees have focused on two species of long-tongued bees: honey bees and bumble bees, both of which are generalists (foraging on a wide range of plant species in different families) and social (living in colonies with cooperative brood care and overlap of generations) [8,10,11,15]. The nutritional requirements of honey bees (colony, adults, and larvae) has been comprehensively reviewed [10], and even though this level of detail does not exist for other bee species, we can assume that other species have similar macronutrient demands; the proportions of macronutrients required may be species-specific (as exemplified in other closely related insect species that share the same host-plants [14,16]).

We can infer the general dietary requirements of bees from existing research. It is clear that both adults and larvae will starve without a constant carbohydrate, mainly nectar, source [10].

Relatively immobile larvae do not require the amounts of carbohydrate needed by foraging bees and their limited carbohydrate demands can be met by a blend of pollen, which contains digestible carbohydrates, and nectar [17-19]. Protein concentration of pollen is positively correlated with

3 larval development and adult reproduction (ovarian development and egg laying) in honey bees, bumble bees, and the sweat bee Lasioglossum zephyrum [20-27]. Lipids are crucial for a variety of physiological processes in bees (e.g. egg production, wax production, secondary energy source) and contribute to larval and adult health, ontogeny, and diapause/overwintering [10,27-30].

Linoleic acid (omega-6), an essential fatty acid for most insect species, in collected pollen has been associated with higher worker production in honey bee colonies [31]. A second essential fatty acid for insects, linolenic acid (omega-3), is also obtained from pollen, but its specific importance for bees is still not described [28]. Sterols obtained exclusively from pollen are the precursors for molting hormones, making pollen essential for larval development [10,27]. Recent research indicates that both honey bee and bumble bee foragers regulate their intake of carbohydrates and proteins to high ratios [32,33], and bumble bees can simultaneously regulate their intake of carbohydrates, proteins, and lipids (Vaudo et al., 2016). These studies reveal bees’ specific nutritional requirements, and potentially highlight how adults prioritize their foraging efforts between nectar and pollen for their nutritional components.

Information is lacking for the specific nutritional requirements of the vast majority of solitary oligolectic bee species, though bee taxa appear to have different requirements in nectar sugar composition (see “Nectar” discussion below). Even less is known of bees’ specific pollen nutritional requirements. For at least a few species of solitary bees, pollen quantity of brood provisions is linearly correlated to body size [34]. Additionally, some specialist bees do not survive well on non-host pollen [35], suggesting that either host-plant pollen is nutritionally optimal for specialists, or they cannot metabolize protective chemicals of non-host pollen. Because nectar and pollen quality varies considerably between host-plant species [11,12] and the bee community

4 exhibits different host-plant visitation patterns over time [36-38], we can assume that different bee species have specific nutritional demands that may influence their host-plant foraging patterns

[16].

Floral Resource Nutritional Diversity and Bee Foraging Behavior

Nectar

Nectar is the major carbohydrate source for most bee species [10,39,40]. Bee larvae require carbohydrates for normal development often in the form of brood food (pollen and nectar mixtures), but the greatest quantity of carbohydrate-rich nectar is required for adult foraging [10].

Nectar is an important floral reward and reinforcing stimulus for bee foragers, and profitable nectar sources can be learned and associated with floral characteristics such as scent and color [41-43].

Although nectar is a dynamic floral resource, varying by abiotic conditions and plant age

[12,25,44-48], there are three relatively constant characteristics that influence bee host-plant choice for nectar: sugar composition, nectar volume, and nectar concentration [18,39]. Other characteristics of nectar composition undoubtedly play a significant role in nectar choice, such as amino acids, lipids, minerals, and secondary plant compounds [46,49-59]; however, research on these characteristics, perhaps with exception of amino acids (recently reviewed in Nepi 2014 [60], has been limited and not systematic across bee species [59-63].

The three main sugars present in nectar are glucose and fructose (monosaccharide), and sucrose

(disaccharide) [12,64,65]. Flowers of a given taxa vary in the relative amounts of these sugars and plant families show a characteristic pattern of sugar composition [12,48,64,65]. Early research found that long-tongued bees prefer high sucrose nectars and short-tongued bees prefer nectars

5 with a higher percentage of monosaccharides [65]. Although the interpretation of these patterns has been questioned on many levels [12,66-68], it is likely that sugar composition of plant taxa is an important factor in determining pollinator host-plant choice [48,62,64,65,69-75].

Nectar concentration also determines patterns of pollinator host-plant visitation [12,76-79], limiting which pollinators can mechanically obtain the nectar, either by adhesion and capillary action or by suction. The rationale is that pollinators with long feeding apparatuses (long-tongue bees, moth/butterfly proboscis, long-tongue fly proboscis) will be limited to more dilute nectars.

Although overall viscosity is affected by temperature (and sugar concentration) [80], patterns of preference are evident (reviewed by Willmer [81]) and therefore likely play a role in the evolution of plant-pollinator communities. For example, honey bees (a long-tongued bee species) prefer a concentration of 30-50% whereas short-tongued bees utilize higher concentration nectars of 45%-

60% [82].

It has been proposed that nectar volume, a third characteristic of floral nectar, is the result of an evolutionary tradeoff [83] between high volumes that are energetically costly (potentially influencing vegetative growth and flower production) [84,85] and volumes that are too low to attract pollinators. Ideally, nectar volume of a given plant species should be high enough to attract pollinators, but low enough to ensure efficient visitation to other conspecific flowers. Nectar volume, therefore, should be strongly associated with the primary pollinators of plant taxa [86,87].

In a classic study of Costa Rican plants and their pollinators, flowers producing high volumes of nectar, which also had large floral mass, were visited by larger bees in contrast to smaller flowers with lower nectar volumes, which were visited by small bees and wasps [44].

6

Pollen

Bees obtain the majority of their protein, including free and protein bound essential amino acids, from pollen, but protein concentration varies considerably between plant species, ranging from

~2-60% [88,89]. Although preference for high protein pollen has not been clearly demonstrated for honey bees [90,91], significant decreases in pollen protein in the colony result in higher pollen foraging rates [91]. It has been suggested that honey bees may prefer pollen higher in essential amino acids [92], or obtain a balance of amino acids by collecting a diverse pollen diet [89].

Increasing evidence exists that bumble bees do prefer and will increase foraging rates to pollen sources higher in protein or essential amino acid concentration [25,93-96]. Indeed, when foraging in the same habitat amongst the same host-plant species, bumble bees collect pollen higher in protein concentration than honey bees, which may be linked to different foraging strategies; bumble bees may preferentially forage for pollen quality, where honey bees may forage for quantity to meet the vast demands of the colony [97]. This tradeoff between quantity and quality likely exists in other bee species.

Pollen serves as bees’ main lipid source (including essential fatty acids and sterols), and lipid concentrations from different plant species can range considerably, from 1-20% [11]. Furthermore, the lipid-rich oily exterior of entomophilous pollen, the pollenkitt, is an important discriminative stimulus, phagostimulus, and digestible component for pollen recognition and bee nutrition [98-

102]. Bees, therefore, may be cued by pollenkitt chemistry to recognize host-plant pollen quality, but research is sparse on how pollen lipid content and the pollenkitt influence bee foraging choice in the field.

7

Because protein and lipid concentrations between pollen species are variable and uncoupled

[11,13](Vaudo et al., 2016), foragers may selectively collect pollen amongst plant species to regulate their intake of these nutrients, or, alternatively, collect from a large array of host-plant species to passively achieve a nutritional balance (this may apply to generalist and oligolectic foragers alike). Research in other arthropod species, including beetles and spiders, indicates that they sense and regulate their intake of protein and lipids when choosing among food sources [103-

106]. Bumble bees, for instance, appear to collect pollen diets from the field that are both high in essential amino acid and sterol content [96]. Our recent research has demonstrated that ratio of protein:lipid concentration of pollen best predicted host-plant species preference of bumble bees; and when given multiple synthetic food sources, bumble bees indeed regulated their protein and lipid intake (Vaudo et al., 2016). These results suggest that bees potentially analyze pollen quality in multiple nutritional dimensions. Furthermore, because bees may not be able to taste protein directly [18], pollenkitt lipid and amino acid chemistry could convey information on pollen quality to bees.

Beyond proteins and lipids, pollen (and often nectar) is rich in micronutrients (e.g. vitamins and minerals) and phytochemicals (e.g. carotenoids, flavonoids, alkaloids and phenolics) that have antioxidant properties and antimicrobial activity [11,107-110]. High concentrations of secondary plant chemicals, however, as plant defenses, could be toxic to bees [54,111,112]. Some specialist bee species do not survive well on exclusive non-host pollen, potentially because they cannot metabolize these chemicals [35]. It has been suggested that oligoleges of the genus Colletes specialize on pollen of the plant subfamily Asteroideae, while generalists of Colletes do not,

8 possibly due to differences in their ability to cope with secondary plant chemicals of Asteroideae pollen [113]. A similar trend has been observed between larvae of closely related generalist Osmia species, having differing physiological abilities to survive on the same pollen diets due to pollen protective chemicals [114]. Therefore, bees could selectively collect or avoid host-plant pollen based on its phytochemical composition.

The Importance of Plant Diversity for Bee Health

Large scale land-use that reduces floral abundance and species richness will negatively affect bee species populations through nutritional shortage in both quantity and quality of resources [1-

7,115]. For example, the recorded population declines of bumble bee and other bee species in

Europe are associated with landscape-level reduction of host-plant availability [1-7,115,116].

Although farmland of bee-pollinated crops may provide a large quantity of floral resources, these habitats may be insufficient at maintaining healthy bees because they may only present single- source pollen or nectar. Also, when the crop is not blooming, the landscape may have few flowering plants, affecting all bee species whose foraging periods do not discretely overlap with crop bloom. Without diverse foraging options and diets during critical periods of reproduction and development, bees may suffer negative health consequences. Additional intensification, such as agrochemical use, can further exacerbate stress, negatively affecting bee foraging behavior

[4,7,117,118] and fitness [7,119-121].

Bees should be given a range of diverse floral resources from which they can self-select their diet to meet their component nutrient requirements, which will sustain healthy populations that can endure disease and stress. For example, in bumble bees, the reproductive benefits of polyfloral

9 pollen diets surpassed those of monofloral diets, even when lower in protein concentration [24].

Polyfloral pollen diets can provide a balance of essential amino acids and fatty acids, whose concentrations differ between species [89]. Exposure to single pollen sources, such as Lupinus crops, that contain plant defensive chemicals can be detrimental to bumble bee colony fitness

[111]. Therefore, generalist bees may visit a variety of host-plant species to obtain pollen to dampen or nullify the harmful effects of pollen secondary metabolites [112]. Appropriate nutrition is necessary for bee immunity (DeGrandi-Hoffman & Chen, this issue); diverse pollen diets can enhance bees' immunocompetence and resistance to pathogens [122,123] and pesticides [124].

Applying Bee Nutrition to Floral Resource Habitat Restoration

To alleviate the negative effects of reduced floral resource availability and interacting stressors of agricultural intensification on bee population health and crop pollination services, selective foraging habitats should be restored in sufficient quantity surrounding areas of land-change

[7,125,126]. Thus, there is increasing demand and incentive based programs for farmers for application of agri-environmental schemes, including floral resource provisioning to support bee populations [7,115,127,128]. The development and design of these schemes have focused primarily on plant species that attract bee abundance and diversity. Because the bee community will visit different plant species throughout the day, season, and between years [36-38,129-131], floral diversity is the best way to attract and support multiple pollinator species over time.

Furthermore, farmland in proximity to natural habitat and/or supplemented with floral resources will attract a wider species richness and functional-group diversity of bees that can result in higher fruit yield [132-140], and economic benefit [139].

10 However, plant species diversity alone is not sufficient to ensure pollinator conservation and thus the aim should be to provide nutritionally optimized floral resources. Figure 1-1 provides a conceptual schematic relating research and application of criteria needed to support bee populations throughout their life cycle. While other factors (nesting habitat [141], structure of the pollinator community [131]) are important for developing pollinator plantings, for this review, we focus on the bee nutrition and the role it plays selecting appropriate plants that support a nutritionally balanced and diverse community. Foremost, plants should be chosen that present floral rewards in phenological succession throughout the day and season [129-131] spanning the active periods of bee species [139]. Then, 1) determine the nutritional value of the nectar and pollen of the agricultural crop, and commercially available native and, where advisable, non- invasive exotic host-plant species (exotic plants species should only be chosen that will not compete with endemic plant species and will promote plant-pollinator community stability [142]).

These studies include analyzing nectar composition, concentration, and volume, and pollen protein, lipid, and micronutrient quality. 2) Determine the nutritional needs of different bee species occupying the landscape, including those important for crop pollination. These studies can be conducted in field, semi-field, or laboratory settings correlating resource quality to nectar and pollen visitation data [94], or feeding assays using synthetic or supplemented diets [14].

Integrating this information will allow us to select plant-species that better meet bees' nutritional needs. Rich nectar sources diverse in their quality and quantity will provide the differing carbohydrate needs of bees and other pollinators. Further, plant species that are attractive, but whose pollen are complementary (to each other and the agricultural crop) in their protein, lipid, and micronutrient quality will allow bees to self-select their diet to balance their intake of these

11 nutrients to maximize their reproductive output and larval development/survival. Additionally, plant communities can be designed to match the changing nutritional needs of bees throughout the growing season. For example, with a strong understanding of pollen and nectar nutritional quality, we should be able to provide pollen sources early in the season to boost worker population growth for honey bee and bumble bee colonies [143,144], and late season nectar flow for honey bee overwintering and bumble bee gyne survival [115,145](SH Woodard, abstract 0406, Entomology

2014, Austin, TX). Finally, once pollen and nectar nutritional quality is better characterized, devised plantings should support wide generalists that collect diverse resources for quantity, or selectively for nutritional value. Because generalists visit the majority of host-plant species in local plant-pollinator communities [131], achieving diversity in our plant communities will also likely maximize attractiveness to solitary or specialist species that have limited foraging distances, shorter active periods, and narrower host-plant preferences.

Developing rationally designed floral provisioning schemes that optimize pollinator nutrition requires information about the nutritional requirements of pollinators, how these shape their foraging preferences, and the nutritional profiles of a range of the floral resources of native and agricultural plant species. Integrating this information will allow development of targeted, and simplified, plant communities, which can be used for conservation of a diverse range of bee species in a diversity of landscapes. These healthy and abundant bee populations will then sustain agricultural production in the face of increasing demands for food in a changing environment.

12 Acknowledgements

We would like to thank the Grozinger lab for their helpful discussions and critical insight to the preparation of this manuscript. Funding supporting the development of this review was provided by North American Pollinator Protection Campaign Bee Health Improvement Project Grant,

USDA AFRI NIFA Predoctoral Fellowships Grant number GRANT10359159, and from an anonymous donation to the Penn State Center for Pollinator Research.

13

Figure 1-1. Conceptual schematic presenting a holistic framework relating basic research and landscape application for bee conservation and habitat restoration. The essential research objectives are: 1) Seasonal and daily phenology of bee and plant species, 2) Bee nutritional requirements and the nutritional quality of nectar and pollen from commercially available host-plant species, and 3) Bee species nesting requirements. These research areas provide the environmental criterion necessary for supporting bees’ annual life cycle: 1) Timing of blooming that matches with bee active foraging periods, 2) Nectar characteristics necessary for bee energetic needs, especially during foraging, 3) Pollen characteristics necessary for bee reproduction and development, and 4) Nesting habitat for bees to rear brood and spend periods of time of inactivity and dormancy. We can then rationally design conservation plant communities by selecting host-plant species (and natural habitat) that meet these criteria. These plant communities constitute a diversity of host-plant species optimized for bee nutrition. The outcome of a comprehensive conservation effort is that we provide a diverse group of bee species appropriate nutrition and habitat that will stabilize their populations. Healthy and diverse bee populations will then be more effective pollinators of wild host-plant and crop species.

14 References (* represents references of interests followed by brief summary)

1. Biesmeijer JC, et al. (2006) Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 313(5785):351–354.

2. Carvell C, et al. (2006) Declines in forage availability for bumblebees at a national scale. Biol Conserv132(4):481–489.

3. Naug D (2009) Nutritional stress due to habitat loss may explain recent honeybee colony collapses. Biol Conserv 142(10):2369–2372.

4. Potts SG, et al. (2010) Global pollinator declines: trends, impacts and drivers. Trends Ecol Evol 25(6):345–353.

5. Winfree R, Bartomeus I, Cariveau DP (2011) Native Pollinators in Anthropogenic Habitats. Annu Rev Ecol Evol Syst 42(1):1–22.

6. * Ollerton J, Erenler H, Edwards M, Crockett R (2014) Pollinator declines. Extinctions of aculeate pollinators in Britain and the role of large-scale agricultural changes. Science 346(6215):1360–1362.

Using historical data from the mid-19th century until present, patterns of bee and wasp extinctions in the United Kingdom are highly correlated with periods of agricultural intensification policy and practice, including those that are associated with reductions in wildflower abundance and diversity.

7. * Goulson D, Nicholls E, Botías C, Rotheray EL (2015) Bee declines driven by combined stress from parasites, pesticides, and lack of flowers. Science 347(6229) doi: 10.1126/science.1255957.

This current review discusses the interacting and synergistic factors contributing to bee declines and potential pollination crisis, including loss of foraging and nesting habitat, monofloral diets, introduced parasites and diseases, pesticides, and phenological disruption; also included are suggestions for improvement such as restoring floral and nesting resources, reduction of pesticide use, and preventing the spread of non-native bees, parasites, and pathogens.

8. Danforth BN, Cardinal S, Praz C, Almeida EAB, Michez D (2013) The impact of molecular data on our understanding of bee phylogeny and evolution. Annu Rev Entomol 58:57–78.

9. Michener CD (2007) The Bees of the World (Johns Hopkins University Press).

10. * Brodschneider R, Crailsheim K (2010) Nutrition and health in honey bees. Apidologie 41(3):278–294.

This essential and comprehensive review of honey bee nutrition throughout the lifecycle describes the variety nutrients obtained from nectar and pollen and their important effects

15 on colony, worker, and larval behavior, development, and health; a key example of the detail needed for a variety of bee species to address their nutritional requirements.

11. Roulston TH, Cane JH (2000) Pollen nutritional content and digestibility for animals. Plant Syst Evol 222(1):187–209.

12. Nicolson SW, Thornburg RW (2007) Nectar chemistry. Nectaries and Nectar, eds Nicolson SW, Nepi M, Pacini E (Springer), pp 215–264.

13. Campos MGR, et al. (2008) Pollen composition and standardisation of analytical methods. J Api Res 47(2):154–161.

14. * Behmer ST (2009) Insect herbivore nutrient regulation. Annu Rev Entomol 54:165–187.

This review highlights the history, theory, methodology, and results of how insect foraging behavior is adapted to prioritize and regulate macronutrient intake to reach an optimum target for fitness, and how reaching or modifying this target can offset deleterious effects of plant secondary chemicals and other stressors; also described are specific differences in intake targets between related insect species.

15. Haydak MH (1970) Honey bee nutrition. Annu Rev Entomol 15(1):143–156.

16. Behmer ST, Joern A (2008) Coexisting generalist herbivores occupy unique nutritional feeding niches. Proc Natl Acad Sci USA 105(6):1977–1982.

17. Human H, Nicolson SW (2006) Nutritional content of fresh, bee-collected and stored pollen of Aloe greatheadii var. davyana (Asphodelaceae). Phytochemistry 67(14):1486– 1492.

18. Nicolson SW (2011) Bee Food: The chemistry and nutritional value of nectar, pollen and mixtures of the two. Afr Zool 46(2):197–204.

19. Human H, et al. (2013) Miscellaneous standard methods for Apis mellifera research. J Api Res 52(4):1–56.

20. Sutcliffe GH, Plowright RC (1990) The effects of pollen availability on development time in the bumble bee Bombus terricola K. (Hymenoptera: Apidae). Can J Zool 68(6):1120– 1123.

21. Génissel A, Aupinel P, Bressac C, Tasei JN, Chevrier C (2002) Influence of pollen origin on performance of Bombus terrestris micro-colonies. Entomol Exper Applic 104(2- 3):329–336.

22. Roulston TH, Cane JH (2002) The effect of pollen protein concentration on body size in the sweat bee Lasioglossum zephyrum (Hymenoptera: Apiformes). Evol Ecol 16(1):49– 65.

23. Human H, Nicolson SW, Strauss K, Pirk CWW, Dietemann V (2007) Influence of pollen

16 quality on ovarian development in honeybee workers (Apis mellifera scutellata). Journal of Insect Physiology 53(7):649–655.

24. Tasei J-N, Aupinel P (2008) Nutritive value of 15 single pollens and pollen mixes tested on larvae produced by bumblebee workers (Bombus terrestris, Hymenoptera: Apidae). Apidologie 39(4):397–409.

25. Cardoza YJ, Harris GK, Grozinger CM (2012) Effects of Soil Quality Enhancement on Pollinator-Plant Interactions. Psyche 2012(4):1–8.

26. Li C, Xu B, Wang Y, Feng Q, Yang W (2012) Effects of dietary crude protein levels on development, antioxidant status, and total midgut protease activity of honey bee (Apis mellifera ligustica). Apidologie 43(5):576–586.

27. * Vanderplanck M, et al. (2014) How does pollen chemistry impact development and feeding behaviour of polylectic bees? PLoS ONE 9(1):e86209.

Using bumble bee microcolonies confined to monofloral pollen diets of varying protein and sterol quality, bumble bees collected all pollen equally, but increases in protein/amino acid and sterols lead to higher larval weight, indicating healthier offspring that could lead to more productive adults.

28. Canavoso LE, Jouni ZE, Karnas KJ, Pennington JE, Wells MA (2001) Fat metabolism in insects. Annu Rev Nutr 21:23–46.

29. Toth AL, Kantarovich S, Meisel AF, Robinson GE (2005) Nutritional status influences socially regulated foraging ontogeny in honey bees. J Exp Biol 208(24):4641–4649.

30. Fliszkiewicz M, Wilkaniec Z (2007) Fatty acids and amino acids in the fat body of bumblebee Bombus terrestris (L) in diapausing and non-diapausing queens. J apic Sci 51(1):55–63.

31. Avni D, Hendriksma HP, Dag A, Uni Z, Shafir S (2014) Nutritional aspects of honey bee- collected pollen and constraints on colony development in the eastern Mediterranean. J Insect Physiol 69:65–73.

32. Paoli PP, et al. (2014) Nutritional balance of essential amino acids and carbohydrates of the adult worker honeybee depends on age. Amino Acids 46(6):1449–1458.

33. Stabler D, Paoli PP, Nicolson SW, Wright GA (2015) Nutrient balancing of the adult worker bumblebee (Bombus terrestris) depends on the dietary source of essential amino acids. J Exp Biol 218:793–802.

34. Müller A, et al. (2006) Quantitative pollen requirements of solitary bees: Implications for bee conservation and the evolution of bee–flower relationships. Biol Conserv 130(4):604– 615.

35. Praz CJ, Müller A, Dorn S (2008) Specialized bees fail to develop on non-host pollen: do

17 plants chemically protect their pollen? Ecology 89(3):795–804.

36. Olesen JM, Bascompte J, Elberling H, Jordano P (2008) Temporal dynamics in a pollination network. Ecology 89(6):1573–1582.

37. Petanidou T, Kallimanis AS, Tzanopoulos J, Sgardelis SP, Pantis JD (2008) Long-term observation of a pollination network: fluctuation in species and interactions, relative invariance of network structure and implications for estimates of specialization. Ecol Lett 11(6):564–575.

38. Baldock KCR, Memmott J, Ruiz-Guajardo JC, Roze D, Stone GN (2011) Daily temporal structure in African savanna flower visitation networks and consequences for network sampling. Ecology 92(3):687–698.

39. Cnaani J, Thomson JD, Papaj DR (2006) Flower choice and learning in foraging bumblebees: effects of variation in nectar volume and concentration. Ethology 112:278– 285.

40. Waser NM, Ollerton J (Eds) (2006) Plant-pollinator interactions: from specialization to generalization. (University of Chicago Press).

41. Raine NE, Ings TC, Dornhaus A, Saleh N, Chittka L (2006) Adaptation, genetic drift, pleiotropy, and history in the evolution of bee foraging behavior. Advances in the Study of Behavior 36:305–354.

42. Raine NE, Chittka L (2012) No trade-off between learning speed and associative flexibility in bumblebees: a reversal learning test with multiple colonies. PLoS ONE 7(9):e45096.

43. Knauer AC, Schiestl FP (2015) Bees use honest floral signals as indicators of reward when visiting flowers. Ecol Lett 18(2):135–143.

44. Opler PA (1983) Nectar production in a tropical ecosystem. The Biology of Nectaries, eds Bentley B, Elias T (Columbia University Press, New York), pp 30– 79.

45. Zimmermann M (1988) Nectar production, flowering phenology, and strategies for pollination. Plant reproductive ecology, patterns and strategies, eds Lovett Doust J, Lovett Doust L (Oxford University Press), pp 157–178.

46. Corbet SA (2003) Nectar sugar content: estimating standing crop and secretion rate in the field. Apidologie 34(1):1–10.

47. Peat J, Goulson D (2005) Effects of experience and weather on foraging rate and pollen versus nectar collection in the bumblebee, Bombus terrestris. Behav Ecol Sociobiol 58(2):152–156.

48. Petanidou T (2005) Sugars in Mediterranean floral nectars: an ecological and evolutionary approach. J Chem Ecol 31(5):1065–1088.

18 49. Deinzer ML, Thomson PA, Burgett DM, Isaacson DL (1977) Pyrrolizidine alkaloids: their occurrence in honey from tansy ragwort (Senecio jacobaea L.). Science 195(4277):497– 499.

50. Rodriguez-Arce AL, Diaz N: The stability of beta-carotene in mango nectar. J Agr U Puerto Rico, 1992 76:101–102.

51. Roshchina VV, Roshchina VD (1993) Excretory function of higher plants (Springer- Verlag).

52. Ecroyd CE, Franich RA, Kroese HW, Steward D (1995) Volatile constituents of Dactylanthus taylorii flower nectar in relation to flower pollination and browsing by animals. Phytochemistry 40(5):1387–1389.

53. Ferreres F, Andrade P, Gil MI, Tomás-Barberán FA (1996) Floral nectar phenolics as biochemical markers for the botanical origin of heather honey. Z Lebensm Unters Forch 202(1):40–44.

54. Adler LS (2000) The ecological significance of toxic nectar. Oikos 91(3):409–420.

55. Carter C, Thornburg RW (2004) Is the nectar redox cycle a floral defense against microbial attack? Trends Plant Sci 9(7):320–324.

56. Dötterl S, Schäffler I (2007) Flower scent of floral oil-producing Lysimachia punctata as attractant for the oil-bee Macropis fulvipes. J Chem Ecol 33(2):441–445.

57. Petanidou T (2007) Ecological and evolutionary aspects of floral nectars in Mediterranean habitats. Nectaries and Nectar, eds Nicolson SW, Nepi M, Pacini E (Springer), pp 343- 375.

58. González-Teuber M, Heil M (2009) Nectar chemistry is tailored for both attraction of mutualists and protection from exploiters. Plant Signal Behav 4(9):809–813.

59. Baker HG, Baker I (1973) Amino-acids in nectar and their evolutionary significance. Nature 241(5391):543–545.

60. Nepi M (2013) Beyond nectar sweetness: the hidden ecological role of non-protein amino acids in nectar. J Ecol 102(1):108–115.

61. Gottsberger G, Schrauwen J, Linskens HF (1984) Amino acids and sugars in nectar, and their putative evolutionary significance. Plant Syst Evol 145(1-2):55–77.

62. Petanidou T, Van Laere AJ, Smets E (1996) Change in floral nectar components from fresh to senescent flowers ofCapparis spinosa (Capparidaceae), a nocturnally flowering Mediterranean shrub. Plant Syst Evol 199(1-2):79–92.

63. Carter C, Shafir S, Yehonatan L, Palmer RG, Thornburg R (2006) A novel role for proline in plant floral nectars. Naturwissenschaften 93(2):72–79.

19 64. Percival MS (1961) Types of nectar in angiosperms. New Phytol 60(3):235–281.

65. Baker HG, Baker I: Floral nectar sugar constituents in relation to pollinator type. In Handbook of Experimental Pollination Ecology. Edited by Jones CE, Little RJ. Van Nostrand Reinhold; 1983:117– 41.

66. Wells H, Hill PS, Wells PH (1992) Nectarivore foraging ecology: rewards differing in sugar types. Ecol Entomol 17(3):280–288.

67. Willmer PG, Stone GN: Behavioral, ecological, and physiological determinants of the activity patterns of bees.]Advances in the Study of Behavior, eds Slater PJB, Rosenblatt JS, Snowdon CT, Roper TJ, Brockmann HJ, Naguib M (Elsevier) pp 2038-2043.

68. Johnson SD, Nicolson SW (2008) Evolutionary associations between nectar properties and specificity in bird pollination systems. Biol Lett 4(1):49–52.

69. Wykes GR (1952) The preferences of honeybees for solutions of various sugars which occur in nectar. J Exp Biol 29(4):511–519.

70. Percival M (1965) Floral Biology (Pergamon Press).

71. Baker HG, Baker I (1979) Sugar ratios in nectars. Phytochem Bull V 23:43-45.

72. Southwick EE, Loper GM, Sadwick SE (1981) Nectar production, composition, energetics and pollinator attractiveness in spring flowers of western New York. Am J Bot 68(7):994.

73. Baker HG, Baker I (1982) Chemical constituents in nectar in relation to pollination mechanisms and phylogeny. Biochemical Aspects of Evolutionary Biology, ed Nitecki MH (University of Chicago Press), pp 131-171.

74. Baker HG, Baker I (1990) The predictive value of nectar chemistry to the recognition of pollinator types. Israel J Bot 39:159–66.

75. Stiles FG, Freeman CE (1993) Patterns in floral nectar characteristics of some bird-visited plant species from costa rica. Biotropica 25(2):191.

76. Harder LD (1986) Effects of nectar concentration and flower depth on flower handling efficiency of bumble bees. Oecologia 69(2):309–315.

77. Pyke GH, Waser NM (1981) The production of dilute nectars by hummingbird and honeyeater flowers. Biotropica 13(4):260.

78. Roubik DW (Ed) (1995) Pollination of Cultivated Plants in the Tropics. Vol. 118 (Food & Agriculture Org.).

79. Borrell BJ (2006) Mechanics of nectar feeding in the orchid bee Euglossa imperialis: pressure, viscosity and flow. J of Exp Biol 209(24):4901–4907.

80. Mathlouthi M, Génotelle J (1995) Rheological properties of sucrose solutions and

20 suspensions. Sucrose, eds Mathlouthi M, Reiser P (Springer), pp 126-154.

81. Willmer P (2011) Pollination and Floral Ecology (Princeton University Press).

82. Roubik DW, Buchmann SL (1984) Nectar selection by Melipona and Apis mellifera (Hymenoptera: Apidae) and the ecology of nectar intake by bee colonies in a tropical forest. Oecologia 61(1):1–10.

83. Harder LD, Williams NM, Jordan CY, Nelson WA (20010 The effects of floral design and display on pollinator economics and pollen dispersal. Cognitive Ecology of Pollination, eds Chittka L, Thomson JD (Cambridge University Press), pp 297–317.

84. Harder LD, Cruzan MB (1990) An Evaluation of the Physiological and Evolutionary Influences of Inflorescence Size and Flower Depth on Nectar Production. Funct Ecol 4(4):559.

85. * Mu J, et al. (2014) Domesticated honey bees evolutionarily reduce flower nectar volume in a Tibetan lotus. Ecology 95(11):3161–3172.

By comparing local communities of alpine lotus at increasing distances from large honey bee apiaries and conducting a common garden experiment, the researchers found that honey bees place a significant selective pressure on plants causing adaptive trade-offs in floral characteristics; communities of plants nearer to apiaries exhibited reduced nectar volume, while increasing plant size, biomass, and number of flowers and seeds.

86. Cruden RW, Hermann SM, Peterson S (1983) Patterns of nectar production and plant- pollinator coevolution. The Biology of Nectaries, eds Bentley B, Elias TS (Columbia University Press), pp 80-125.

87. Faegri K, Van der Pijl L (2013) Principles of Pollination Ecology (Elsevier).

88. Roulston TH, Cane JH, Buchmann SL (2000) What governs protein content of pollen: pollinator preferences, pollen-pistil interactions, or phylogeny? Ecol Monogr 70(4):617– 643.

89. Weiner CN, Hilpert A, Werner M, Linsenmair KE, Blüthgen N (2010) Pollen amino acids and flower specialisation in solitary bees. Apidologie 41(4):476–487.

90. Van der Moezel PG, Delfs JC, Pate JS, Loneragan WA, Bell DT (1987) Pollen selection by honeybees in shrublands of the Northern Sandplains of Western Australia. J Api Res 26(4):224–232.

91. Pernal S, Currie R (2001) The influence of pollen quality on foraging behavior in honeybees (Apis mellifera L.). Behav Ecol Sociobiol 51(1):53–68.

92. Cook SM, Awmack CS, Murray DA, Williams IH (2003) Are honey bees' foraging preferences affected by pollen amino acid composition? Ecol Entomol 28(5):622–627.

21 93. Kitaoka TK, Nieh JC (2008) Bumble bee pollen foraging regulation: role of pollen quality, storage levels, and odor. Behav Ecol Sociobiol 63(4):501–510.

94. Hanley ME, Franco M, Pichon S, Darvill B, Goulson D (2008) Breeding system, pollinator choice and variation in pollen quality in British herbaceous plants. Functional Ecology 22(4):592–598.

95. * Konzmann S, Lunau K (2014) Divergent Rules for Pollen and Nectar Foraging Bumblebees – A Laboratory Study with Artificial Flowers Offering Diluted Nectar Substitute and Pollen Surrogate. PLoS ONE 9(3):e91900.

In a controlled study using artificial flowers differing in nectar (sugar concentration) and pollen (powder dilutions) quality and quantity, bumble bees show preferences for increased nectar and pollen concentration and volume, though will collect all pollen if available.

96. * Somme L, et al. (2014) Pollen and nectar quality drive the major and minor floral choices of bumble bees. Apidologie 46(1):92–106.

By identifying the source of pollen loads of four bumble bee species, and analyzing the amino acid and sterol quality of the host-plant species' pollen, bumble bees appear to prioritize the collection of high quality pollen (amino acid and sterol concentrations) while avoiding pollen chemical defenses; the bees also preferred plant species high in nectar volume and concentration.

97. Leonhardt SD, Blüthgen N (2011) The same, but different: pollen foraging in honeybee and bumblebee colonies. Apidologie 43(4):449–464.

98. Dobson HEM (1987) Role of flower and pollen aromas in host-plant recognition by solitary bees. Oecologia 72(4):618–623.

99. Dobson HEM, Bergström G (2000) The ecology and evolution of pollen odors. Plant Syst Evol 222:63–87.

100. Cook SM, et al. (2005) Could learning of pollen odours by honey bees (Apis mellifera) play a role in their foraging behaviour? Physiol Entomol 30(2):164–174.

101. Piskorski R, Kroder S, Dorn S (2011) Can pollen headspace volatiles and pollenkitt lipids serve as reliable chemical cues for bee pollinators? Chem Biodivers 8(4):577–586.

102. Pacini E, Hesse M (2005) Pollenkitt–its composition, forms and functions. Flora- Morphology 200(5):399–415.

103. Jensen K, Mayntz D, Toft S, Raubenheimer D, Simpson SJ (2011) Nutrient regulation in a predator, the wolf spider Pardosa prativaga. Anim Behav 81(5):993–999.

104. Lee KP, Simpson SJ, Raubenheimer D (2004) A comparison of nutrient regulation between solitarious and gregarious phases of the specialist caterpillar, Spodoptera

22 exempta (Walker). J Insect Physiol 50(12):1171–1180.

105. Mayntz D, Raubenheimer D, Salomon M, Toft S, Simpson SJ (2005) Nutrient-specific foraging in invertebrate predators. Science 307(5706):111–113.

106. Raubenheimer D, Mayntz D, Simpson SJ, Tøft S (2007) Nutrient-specific compensation following diapause in a predator: implications for intraguild predation. Ecology 88(10):2598–2608.

107. Campos MG, Webby RF, Markham KR, Mitchell KA, da Cunha AP (2003) Age-induced diminution of free radical scavenging capacity in bee pollens and the contribution of constituent flavonoids. J Agric Food Chem 51(3):742–745.

108. Basim E, Basim H, Özcan M (2006) Antibacterial activities of Turkish pollen and propolis extracts against plant bacterial pathogens. J Food Eng 77(4):992–996.

109. LeBlanc BW, Davis OK, Boue S, DeLucca A, Deeby T (2009) Antioxidant activity of Sonoran Desert bee pollen. Food Chem 115(4):1299–1305.

110. Morais M, Moreira L, Feás X, Estevinho LM (2011) Honeybee-collected pollen from five Portuguese Natural Parks: palynological origin, phenolic content, antioxidant properties and antimicrobial activity. Food Chem Toxicol 49(5):1096–1101.

111. Arnold SEJ, Idrovo MEP, Arias LJL, Belmain SR, Stevenson PC (2014) Herbivore defence compounds occur in pollen and reduce bumblebee colony fitness. J Chem Ecol doi:10.1007/s10886-014-0467-4.

112. * Eckhardt M, Haider M, Dorn S, Müller A (2014) Pollen mixing in pollen generalist solitary bees: a possible strategy to complement or mitigate unfavourable pollen properties? The Journal of Animal Ecology 83(3):588–597.

For the pollen generalist Osmis cornuta, larval survival, development time, and adult body mass are negatively affected by the proportion of unfavorable pollen in their diet, suggesting that generalists are able to dampen or nullify the toxic effects of chemical defenses by collecting pollen from a diversity of plant species.

113. Müller A, Kuhlmann M (2008) Pollen hosts of western palaearctic bees of the genus Colletes (Hymenoptera: Colletidae): the Asteraceae paradox. Biol J Linn Soc 95(4):719– 733.

114. Sedivy C, Müller A, Dorn S (2011) Closely related pollen generalist bees differ in their ability to develop on the same pollen diet: evidence for physiological adaptations to digest pollen. Funct Ecol 25(3):718–725.

115. * Potts S, et al. (2015) Status and Trends of European Pollinators. Key Findings of the STEP Project (Pensoft Publishers).

This report of the multi-institutional partnership STEP (Status and Trends of European

23 Pollinators) reviews the research and policy in Europe defining the impact of interacting factors of land-use on pollinator decline and associated agri-environmental schemes and ecological intensification used to support pollinator populations.

116. Goulson D, Lye GC, Darvill B (2008) Decline and conservation of bumble bees. Annu Rev Entomol 53(1):191–208.

117. Gill RJ, Ramos-Rodriguez O, Raine NE (2012) Combined pesticide exposure severely affects individual- and colony-level traits in bees. Nature 491(7422):105–108.

118. Gill RJ, Raine NE (2014) Chronic impairment of bumblebee natural foraging behaviour induced by sublethal pesticide exposure. Funct Ecol 28(6):1459–1471.

119. Whitehorn PR, O'Connor S, Wackers FL, Goulson D (2012) Neonicotinoid pesticide reduces bumble bee colony growth and queen production. Science 336(6079):351–352.

120. Sandrock C, et al. (2013) Sublethal neonicotinoid insecticide exposure reduces solitary bee reproductive success. Agr Forest Entomol 16(2):119–128.

121. Sandrock C, et al. (2014) Impact of chronic neonicotinoid exposure on honeybee colony performance and queen supersedure. PLoS ONE 9(8):e103592.

122. Alaux C, Ducloz F, Crauser D, Le Conte Y (2010) Diet effects on honeybee immunocompetence. Biol Lett 6(4):562–565.

123. * Di Pasquale G, et al. (2013) Influence of Pollen Nutrition on Honey Bee Health: Do Pollen Quality and Diversity Matter? PLoS ONE 8(8):e72016.

In a lab-based study, when infected with Nosema, pollen quality (protein and lipid concentration) of monofloral and polyfloral diets was positively correlated to nurse honey bee physiology, immunity, and survival.

124. Schmehl DR, Teal PEA, Frazier JL, Grozinger CM (2014) Genomic analysis of the interaction between pesticide exposure and nutrition in honey bees (Apis mellifera). Journal of Insect Physiology 71:177–190.

125. * Bommarco R, Kleijn D, Potts SG (2013) Ecological intensification: harnessing ecosystem services for food security. Trends Ecol Evol 28(4):230–238.

This review presents a conceptual framework for utilizing “ecological intensification” to protect ecosystem services, including pollination, to meet the increasing demands for agricultural productivity and global food security.

126. Scheper J, et al. (2013) Environmental factors driving the effectiveness of European agri- environmental measures in mitigating pollinator loss - a meta-analysis. Ecol Lett 16(7):912–920.

127. Potts SG, et al. (2011) Developing European conservation and mitigation tools for

24 pollination services: approaches of the STEP (Status and Trends of European Pollinators) project. J Api Res 50(2):152–164.

128. Vaughn M, Skinner M (2008) Using farm bill programs for pollinator conservation. USDA NRCS Technical Note 78:1-16.

129. Vaudo AD, Patch HM, Mortensen DA, Grozinger CM, Tooker JF (2014) Bumble bees exhibit daily behavioral patterns in pollen foraging. Arthropod-Plant Inte 8(4):273–283.

130. Basilio AM, Medan D, Torretta JP, Bartoloni NJ (2006) A year-long plant-pollinator network. Austral Ecol 31(8):975–983.

131. Russo L, DeBarros N, Yang S, Shea K, Mortensen D (2013) Supporting crop pollinators with floral resources: network-based phenological matching. Ecol Evol 3(9):3125–3140.

132. Kremen C, Williams NM, Thorp RW (2002) Crop pollination from native bees at risk from agricultural intensification. Proc Natl Acad Sci USA 99(26):16812–16816.

133. Westphal C, Steffan-Dewenter I, Tscharntke T (2003) Mass flowering crops enhance pollinator densities at a landscape scale. Ecol Lett 6(11):961–965.

134. Hoehn P, Tscharntke T, Tylianakis JM, Steffan-Dewenter I (2008) Functional group diversity of bee pollinators increases crop yield. Proc Roy Soc B-Biol Sci 275(1648):2283– 2291.

135. Ricketts TH, et al. (2008) Landscape effects on crop pollination services: are there general patterns? Ecol Lett 11(5):499–515.

136. Carvalheiro LG, Seymour CL, Veldtman R, Nicolson SW (2010) Pollination services decline with distance from natural habitat even in biodiversity-rich areas. J Appl Ecol 47(4):810–820.

137. Garibaldi LA, et al. (2011) Stability of pollination services decreases with isolation from natural areas despite honey bee visits. Ecol Lett 14(10):1062–1072.

138. Carvalheiro LG, Seymour CL, Nicolson SW, Veldtman R (2012) Creating patches of native flowers facilitates crop pollination in large agricultural fields: mango as a case study. J Appl Ecol 49(6):1373–1383.

139. * Blaauw BR, Isaacs R (2014) Flower plantings increase wild bee abundance and the pollination services provided to a pollination-dependent crop. J Appl Ecol 51(4):890–898.

Over four years, by planting large patches of native perennial wildflowers that bloom throughout the season adjacent to blueberry farms, wild bee and hoverfly abundance in blueberry fields increased annually, leading to higher blueberry fruit yield that lead to exceeding the cost of establishment and maintenance of the supplementary plantings.

140. Blaauw BR, Isaacs R (2014) Larger patches of diverse floral resources increase insect

25 pollinator density, diversity, and their pollination of native wildflowers. Basic Appl Ecol 15(8):701–711.

141. Williams NM, et al. (2010) Ecological and life-history traits predict bee species responses to environmental disturbances. Biol Conserv 143(10):2280–2291.

142. Stouffer DB, Cirtwill AR, Bascompte J, Bartomeus I (2014) How exotic plants integrate into pollination networks. J Ecol 102(6):1442–1450.

143. Riedinger V, Renner M, Rundlöf M, Steffan-Dewenter I, Holzschuh A (2013) Early mass- flowering crops mitigate pollinator dilution in late-flowering crops. Landscape Ecol 29(3):425–435.

144. Williams NM, Regetz J, Kremen C (2012) Landscape-scale resources promote colony growth but not reproductive performance of bumble bees. Ecology 93(5):1049–1058.

145. Röseler PF, Röseler I (1986) Caste specific differences in fat body glycogen metabolism of the bumblebee, Bombus terrestris. Insect Biochem 16(3):501–508.

26

Chapter 2. Bumble Bees Exhibit Daily Behavioral Patterns in Pollen Foraging

Anthony D. Vaudo* a, Harland M. Patch a, David A. Mortensen b, Christina M. Grozingera, and John F. Tooker a a Department of Entomology, Center for Pollinator Research, The Pennsylvania State University, 501 ASI Building, University Park, PA 16802, USA b Department of Plant Science, Center for Pollinator Research, The Pennsylvania State University, 422 ASI Building, University Park, PA 16802, USA

Abstract

In response to global declines in bee populations, several studies have focused on floral resource provisioning schemes to support bee communities and maintain their pollination services.

Optimizing host-plant selection for supplemental floral provisioning requires an understanding of bee foraging behavior and preferences for host-plant species. However, fully characterizing these preferences is challenging due to multiple factors influencing foraging, including the large degree of spatiotemporal variability in floral resources. To understand bee pollen-foraging patterns, we developed a highly controlled mechanistic framework to measure pollen foraging preferences of the bumble bee Bombus impatiens to nine plant species native to Pennsylvania. We recorded continuous observations of foraging behavior of the experimental bee community and individual bees, while simultaneously standardizing for the number of foragers in the environment and differences in floral display of each plant species, while controlling for flowering phenology such that bees only foraged when all plant species’ flowers were open. Our results demonstrate that B. impatiens exhibit predictable daily patterns in their pollen foraging choices and their preferences are dominated by the host-plants they visit first. We hypothesize that these patterns at the community and individual levels are driven by the interplay between pollen abundance and quality.

We recommend that daily cycles of host-plant visitation be considered in future studies to ensure

27

precise and accurate interpretations of host-plant preference. Such precision is critical for comprehensive analyses of the proximate and ultimate mechanisms driving bee foraging behavior and the selection of host-plant species to use in habitat restoration protocols.

Keywords: Bombus impatiens; daily phenology; foraging preferences; native bee conservation; pollination ecology

Introduction

Global bee declines have been linked to agricultural intensification, which decreases nesting habitat and the diversity and abundance of flowering plant species on which bees rely (1, 2). As bee populations decrease, there is concern that pollination services to ecosystems and agricultural crops will diminish (3-7). To mitigate loss of pollination services in agricultural ecosystems and maintain crop productivity, there is an increased interest in developing approaches to conserve natural habitat in proximity to crop fields, including establishing floral-resource provisioning systems to support native bee communities (8-11).

The primary challenge of designing floral resource provisioning schemes is including plant species that will provide pollen and nectar to a diverse bee community. Practitioners and researchers typically select plants based on apparent preferences of bees to particular plant species in the field, and these “preferences” are often determined by summing the number of bees visiting a particular plant species during a sampling period (12-14). There is, however, a lack of consensus on appropriate sampling methodology to evaluate preferences, because different methods can lead to substantially different results (15). Furthermore, simply summing bee abundance on different plant

28

species overlooks many factors that influence bee foraging behavior, possibly leading researchers to misread preferences for certain plant species. These factors can include flower color (16, 17), scent (18), morphology (19), size of floral display (20, 21), competition with other floral visitors

(22), learning and habitual behavior of individual bees (23, 24), spatiotemporal availability of resources (25, 26), and what resource bees collect (pollen and/or nectar) at a particular time (27).

Because these factors are often intertwined, it is difficult to resolve which are most proximately influencing foraging behavior in the field. Furthermore, foraging-preference studies that do not discriminate between nectar and pollen foraging tend to overlook an important factor that should be considered when designating nutritional resources for bees, namely that bees use nectar for fueling their activities but rely on pollen for rearing larvae (28). Focusing on pollen-foraging preferences should provide more robust data on the plant species that best support population growth of individual bee species, whereas selecting plants for floral provisioning based on simple measurements of abundance may result in a community of plants that do not necessarily optimally support bee communities.

Temporal variation in pollinator-plant interactions can also confound assessments of pollinator preference. Long-term studies have demonstrated that patterns of bee visitation to the same plant species can vary through space and time, both within and among seasons (29-32). Network studies reveal that interactions between pollinator and plant communities change within flowering seasons

(30) and even throughout a day (31), probably due to the differences between pollinator- and plant- community phenologies. Moreover, different plant species vary in the time of day that they present pollen and/or nectar, which may serve as a mechanism to reduce competition between plant species relying on the same pollinator species (33-36). Such asynchronous blooming of host-plant species

29

may greatly skew results of field-based pollinator preference assessments, because the resources of the most preferred plant species may be depleted before observations begin. Thus, studies that only sample bee visitation in one or a few observation periods during the day may misrepresent floral preferences for particular bee species and provide a poor basis for conservation recommendations, highlighting the importance of considering the influence of time of day on foraging preferences of individual pollinator species.

In this study, we developed a highly controlled, mechanistic framework, evaluating visitation rates and visitation durations, to assess bee pollen-foraging preferences for their host-plant species. We addressed whether the eastern bumble bee Bombus impatiens Cresson (Figure 2-1; Hymenoptera:

Apidae) 1) displays distinct pollen foraging preferences among different plant species, and 2) exhibits daily patterns of pollen foraging preferences on these plant species. To address these questions, we tracked pollen foraging by B. impatiens to nine perennial plant species that are native to central Pennsylvania, USA and commonly recommended for floral resource provisioning protocols; both B. impatiens and the plant species share habitats and seasonal and daily phenologies in Pennsylvania. We conducted our studies in a hoop house with managed colonies of bumble bees, thereby controlling for competition with other pollinator species and competition with other flowering plant species. This approach also allowed us to simultaneously evaluate relative preferences for plant species that present their floral resources at different times of day in the field: we controlled the timing of initiation of foraging, such that bees foraged when resources from all plants were simultaneously available. To measure community-level foraging efforts, we tracked how frequently B. impatiens workers collected pollen among the plant species.

Additionally, we timed how long individual bees collected pollen from each plant species as an

30

indicator or individual-level foraging efforts. If in agreement, these two data sets would indicate that the most preferred plant species were those that were visited most frequently and for longer durations. Finally, to reduce variation in the data, we standardized our assessments by the size of the floral display of each plant species, reflecting flower patch size and relative pollen quantity, and numbers of foraging bees present. Thus, our approach provides a highly controlled, relativized, and standardized system for evaluating foraging preferences. Because we tested B. impatiens foraging behavior among plant species directly, we did not control for flower color, scent, and morphology; however, our methods and results will allow us to further test the influence of other factors, including pollen quality and quantity, on bee foraging behavior and make recommendations for host-plant restoration protocols targeting particular bee species.

Materials and methods

Insect and plant species

Bombus impatiens (Figure 2-1) is a generalist foraging bumble bee species native to the eastern

United States (37, 38), but is also commercially available to pollinate fruit and vegetable crops

(39). It is common in central Pennsylvania and active from spring through fall (40). Bombus impatiens is primitively eusocial and its annual colonies produce as many as 500 workers (28, 41).

In recent observations in central Pennsylvania that informed our selection of plant species for this study, B. impatiens was the most abundant bee species (14, 32).

Each year, we purchased two B. impatiens research colonies (Koppert Biological Systems, Inc.,

Howell, MI, USA), comprising one queen and approximately 30 workers. Each colony box includes a large bag of sugar water for bees to obtain an ad libitum carbohydrate source. Weekly

31

throughout the study, we counted numbers of workers in each colony to estimate the total bee, or potential foraging, population. To correct for the foraging population and growth or decay of each colony and its influence on foraging rates, each week we standardized foraging rates to 100 bees for our analyses of foraging preferences.

To build upon findings from previous work in central Pennsylvania, we chose nine native plant species that spanned a range of observed visitation rates by B. impatiens (14). Bombus impatiens visited three of these plant species frequently (> 40 individuals collected), three moderately (13-

19 individuals collected), and three infrequently (1-6 individuals collected; (14). Note that due to a lack of flowering, three plant species were only used in one of the two years of the current study

(see Table 2-1). In addition to the frequency of visitation by B. impatiens, we chose the nine plant species based on their synchrony of flowering between July and August. Information regarding each plant species can be found in the United States Department of Agriculture Plants Database

(www.plants.usda.gov). We ordered sixteen individuals (7.6-L pots) of each plant species

(Meadowood Nursery, Hummelstown, PA, USA); the plants had been grown outdoors and were at least two-years old. When plants were not being used for foraging observations, they were stored in field cages to prevent unwanted floral visitation and herbivory.

Hoop house

We collected all foraging observation data inside a hoop house, a large flight arena constructed as a semi-cylinder tunnel (11 x 6.1 m, 3.05 m height), covered with a 70% shade cloth (Figure 2-2).

With the mesh fabric, the interior of the hoop house was subject to the weather but sealed so that no foraging bumble bees could escape, nor any other floral visitors enter. Therefore, by introducing

32

purchased colonies, only B. impatiens foragers visited the plants we provided. We placed two B. impatiens colonies at one end of the hoop house. At the other end, we placed equal numbers of individual plants from each species that were in bloom (either three or four individuals from each species). We randomly arranged the individual plants to positions within a grid (1 m spacing;

Figure 2-2). Each week, we replaced individual plants with those stored in field cages, but the mixture of the group remained the same (unless a species began or ceased flowering). Each day of observations, we randomized the position of the individual plants in the grid to prevent bee forager learning, traplining, and foot-printing the location of plants (24, 42, 43). Therefore, the bees would always forage from a diversity of plant species, but the locations of the individual plants would change daily and each week bees would forage from new individual plants.

Foraging data collection

To ensure bees acclimated to the hoop house and learned how collect pollen from all plant species, all colonies were given three days to forage among the plant species prior to data collection.

Because this study focused on pollen-foraging behavior, we only collected data for bees that were collecting pollen. Even though specific pollen-foraging behavior differed between plant species of different flower types, pollen foraging was easily distinguished on all plant species. Generally, pollen foraging included bees actively scraping pollen off anthers with their legs, running in circles collecting pollen on their bodies and legs around open floral displays, or “buzz” pollinating. Nectar collection occurred when bees extended their tongues into floral nectaries. During our observations, pollen collection was the primary behavior until pollen resources were exhausted.

Furthermore, to encourage bees to focus on pollen foraging only, the bees continued to have access to the sugar water source that came with the colonies.

33

Each year, we collected 18 days of bumble bee pollen-foraging observations (2012: between 26

June and 30 August; 2013: 5 July – 12 August). To control for variation in foraging behavior due to weather (27), data were only collected on warm and partly cloudy to sunny days. We collected data continuously and divided observations into six discrete collection periods for analysis: 0930-

1000, 1000-1030, 1030-1100, 1115-1145, 1200-1230, and 1245-1315 DST. All the plant species had pollen available for collection by 0930, allowing us to directly compare preferences among the plant species (note that Senna hebecarpa and Tradescantia ohiensis bloomed as early as 0630-

0700). We did not collect observation data beyond 1315 because nearly all pollen had been collected from the plants by this time and the bees started collecting mainly nectar. We determined that pollen was depleted on flowers by visually inspecting the anthers or by brushing anthers with a paintbrush. Colony boxes had entrances with three settings that allowed bees 1) free access to fly in and out of the colony, 2) only to fly in, or 3) no movement in or out of the colony. We opened the entrances for free flight five minutes prior to data collection to prevent bees from collecting pollen before we could make observations. After foraging observations finished for the day, we closed the entrances to only allow bees to return from foraging; thus, all foragers were trapped inside the colony (rarely, a forager or two would not return).

Within each collection period, we observed each individual plant for one minute and recorded the number of pollen foraging visitors ("community visitation rate"). We collected 2,286 community visitation rate data points in 2012 and 2,104 data points in 2013. Concurrently, we recorded the time individuals spent pollen foraging from each plant species ("individual visit duration"). In

2012, we followed the flight paths of haphazardly selected individual foragers to determine

34

durations of individual visits. We recorded the order of plant species visited and how long foragers spent collecting pollen on each plant. In 2013, rather than following individual bees to determine the visit duration, we observed each individual plant within each collection period and recorded the length of time individual foragers spent collecting pollen from those plants; this approach ensured that we collected data on duration of individual visits for all the plant species. If we observed no foragers on a plant species, that species received a time of zero allowing us to track how visitation to each plant species changes throughout the day. These data, however, were excluded from the analysis of average visit duration, which was independent of time of day. We timed 1,675 individual visits in 2012 and 1,302 data points in 2013. To ensure that we collected continuous foraging observations, if time allowed, we repeated for each individual plant all observations within a collection period.

Floral display

To standardize our foraging observations for the influence of floral display on visitation patterns, each day for each individual plant observed, we measured the area of floral display of the blooming flowers, including only flowers that were presenting pollen. Because the plant species differed in individual flower sizes and types (single or composite), which corresponded to differences in the amount of pollen available (eg, single large flowers produced equivalent pollen to many small flowers in a composite display), measuring the area of floral display most accurately accounted for these differences and allowed us to estimate relative pollen quantities between plant species.

Using a digital camera (Cannon PowerShot G9; Cannon Inc., Tokyo, Japan), we photographed each flower or cluster of flowers from each individual plant; for reference, we included a ruler in each image. We analyzed images with ImageJ 1.46r software (National Institutes of Health 2012)

35

to calculate the area (cm2) of the floral display. We analyzed 382 photographs in 2012 and 616 photographs in 2013.

Foraging data metrics

Because we wanted to compare bee pollen foraging preferences among the plant species directly, we created single metrics to analyze B. impatiens foraging efforts at both the community and individual bees. These metrics reduce the variation in the data caused by the number of bees in the environment and size of floral display. To analyze bee foraging data, we used two metrics:

“community visitation rate” and “individual visit durations.” Community visitation rates indicate how frequently bees visit a particular plant species to collect pollen while individual visit durations indicate how much time individuals spent collecting pollen from each plant species. These two metrics should complement each other to reveal “preferred” (and therefore profitable) plant species if they were visited more frequently and for longer periods of time. We calculated community visitation rate as the number of pollen foraging visits to a plant species per minute per cm2 of floral display multiplied by a conversion factor (conversion factor = 100 / # bees in each colony) to standardize the data for 100 bees in the environment (visits/min/cm2/100 bees). We calculated individual visit duration as the time in seconds spent by a foraging bee collecting pollen at a plant species per cm2 of floral display (sec/cm2). To determine if there were general differences in foraging rates to each plant species, we first used ANOVA to analyze community visitation rate and individual visit duration data independent of time of day, followed by post-hoc analyses to determine differences in foraging rates between each pair of plant species. We then used a two- way ANOVA to analyze community visitation rates and individual visit duration to determine if there was an interaction between plant species and time of day, which would indicate that B.

36

impatiens foragers change their foraging efforts to the plant species in different collection periods.

Data from 2012 and 2013 were analyzed separately. All statistical analyses were conducted using

JMP v.10.0.0 (SAS Institute 2012).

Results

In both years, the community visitation rates of B. impatiens foragers differed significantly among plant species, independent of time of day (Figure 2-3a,c; 2012: F6,1219 = 29.7, P < 0.0001; 2013:

F7,2095 = 53.0, P < 0.0001). In 2012, Tradescantia ohiensis and Veronicastrum virginicum were visited most frequently. In 2013, Senna hebecarpa dominated the community visitation rates

(Figure 2-3c); however, exclusion of S. hebecarpa from the 2013 analysis (since it was not available in 2012) produced similar results as 2012, with T. ohiensis and V. virginicum visited more frequently than the remaining plant species (Figure 2-3e; F6,1754 = 28.73, P < 0.0001).

When we analyzed community visitation rate data for 2012 by time of day, we detected a significant interaction between plant species and time of day (Figure 2-4a; F30,1184 = 3.7, P <

0.0001) indicating that visitation among plant species changed during the course of the day. In

2013, when S. hebecarpa bloomed, we found no significant interaction between plant species and time of day (Figure 2-4c; F35,2055 = 0.69, P = 0.92), presumably because foragers spent so much time on this particular plant species. When we excluded S. hebecarpa from the analysis, a significant ‘plant species x time of day’ interaction again emerged (Figure 2-4e; F30,1719 = 3.5, P <

0.0001).

When we considered individual visit duration of B. impatiens foragers, rather than their community

37

visitation rates (above), we found that for both years individual visit duration also exhibited significant differences among plant species, independent of time of day (Figure 2-3b, d, f; 2012:

F5,680 = 16.84, P < 0.0001 [Eupatorium perfoliatum excluded due to insufficient observations];

2013: F6,1178 = 27.08, P < 0.0001 [S. hebecarpa included, Monarda fistulosa excluded due to lack of sufficient observations]; F5,876 = 7.40, P < 0.0001 [S. hebecarpa excluded, and M. fistulosa excluded due to lack of sufficient observations]).

When analyzed by time of day, we found for 2012 a significant effect of the ‘plant species x time of day’ interaction on individual visit durations (Figure 2-4b; F20,631 = 2.80, P < 0.0001 [E. perfoliatum and Symphyotrichum novae-angliae excluded due to insufficient observations]). In

2013, when we included S. hebecarpa in the analysis, we did not detect a significant ‘plant species x time of day’ interaction (Figure 2-4d; F30,1236 = 1.11, P = 0.31) likely due again to the dominant visitation to S. hebecarpa. Again however, when we excluded S. hebecarpa from the analysis the interaction was evident (Figure 2-4f; F25,939 = 1.72, P < 0.02 [M. fistulosa excluded due to insufficient observations]).

Discussion

The implications of daily foraging patterns for studies of plant-pollinator interactions

By restricting bee foraging inside a hoop house and limiting the variability typically associated with field-based studies, our experiments provided a controlled setting in which to accurately determine relative pollen foraging preferences among the plant species tested. Though we did not control for every potential factor influencing bee foraging behavior, we measured foraging preferences by controlling and standardizing for important factors such as flower patch size,

38

specific resource foraging, interspecies competition, and spatial memory and marking of the location of resources. Importantly, by only giving B. impatiens foragers access to flowers when all species were presenting their floral resources, we were able to directly evaluate foraging preferences for plant species that otherwise bloom at different times of the day. Our results indicate that B. impatiens foragers exhibited observable and habitual patterns of pollen foraging on the plant species we offered. The foragers showed distinct pollen-foraging rates at the community and individual levels, regardless of the time of day (Figure 2-3), but notably, showed daily patterns in pollen foraging (Figure 2-4), exhibiting different foraging rates to the flowering-plant species at different times of day.

With our controlled and continuous foraging observations, we revealed often-overlooked patterns in bee foraging behavior. Our data suggest that future assessments of pollinator preference need to be mindful of daily cycles in bee foraging behavior and host-plant resource presentation (33-36,

44). For instance, from the current study if we only collected foraging preference data at 1300, we would interpret bee foraging preference data much differently than if we only collected the data at

0930. Indeed, previous data from our area indicated that B. impatiens infrequently visited Senna hebecarpa and Tradescantia ohiensis while frequently visiting Eupatorium perfoliatum (14, 32); however, our data show the opposite. This disparity likely resulted from three factors: 1) aggregating daily observations, 2) the associated assumption that patterns of foraging are consistent throughout the day (and year) (45, 46), and 3) by having missed foraging that occurred prior to mid-day observation periods. Moreover, the disparity must also have been influenced by our methodology of controlling for flowering phenology by limiting access to flowers until the same time each day and tracking a single pollinator species, approaches that differ from typical

39

community network studies. Field studies, of course, provide a realistic view of community-level distribution of bee foraging for host-plant species in a particular area. But ecological differences between field sites could perhaps obscure what are the underlying mechanisms that shape pollinator preferences for certain plant species. These factors can include overall plant-pollinator community composition, daily and seasonal plant species blooming phenologies, daily and seasonal bee foraging activity, and species interactions, all of which combine to shape pollinator communities throughout the day (31). Our methodology therefore provides a framework to test the primary mechanisms (such as resource quantity and quality) that drive pollinator host-plant choice, and our results reveal patterns of host-plant visitation that beg a mechanistic explanation.

Nectar is the primary carbohydrate source for bees and is often used to fuel foraging, whereas pollen is their primary source of proteins, lipids, and micronutrients and is essential for rearing offspring and is often presented by flowers in limited quantities (reviewed in [47]). Studying pollen foraging preferences, therefore, will provide insight on host-plant species that may best support future generations of specific bee species. In this study, the behavioral differences between pollen and nectar foraging were easily observed, and in the future it will be valuable to distinguish between the two types of foraging when collecting data to accurately report bee foraging efforts and determine host plant preferences. By differentiating between pollen and nectar foraging, we resolved differences in pollen-foraging preferences between host-plant species. For example, B. impatiens visits to Echinacea purpurea and Monarda fistulosa especially were dominated by nectar collection (data not shown); therefore, if we merely summed all bee visits (for both pollen and nectar) to those plant species, then we may not have been able to quantify differences in community and individual pollen visitation rates among the plant species. This level of resolution

40

will allow us to subsequently test the mechanisms driving the pollen visitation rates we observed.

Foraging patterns reveal interplay between pollen quality and abundance

Observing bumble bees at both community and individual levels revealed interesting patterns in pollen-foraging behavior. We observed that, generally, the community visitation rates and individual visit durations to each plant species were similar independent of time of day. But importantly, B. impatiens foragers appear to visit their preferred pollen host-plant species early in the day and then, after exhausting those host-plant resources, move on to less preferred species.

We hypothesize that these interactions may be driven by tradeoff between resource (pollen) quality and abundance.

When B. impatiens foragers visited Veronicastrum virginicum in 2012 and S. hebecarpa in 2013

(their most “preferred” host-plants in each year independent of time of day; Figure 2-3a,b,c,d), their community and individual level foraging behavior did not change significantly through the day (unlike the remaining plant species, Figure 2-4a,b,c,d), indicating their preferred status. These results suggest that these host-plant species were the most rewarding for B. impatiens, which consistently tried to collect their resources even when we observed that their pollen stores had been depleted. We hypothesize that these plant species produced high-quality pollen that evoked consistent foraging behavior by B. impatiens.

For the remaining plant species, B. impatiens still visited their next most preferred plant species early, moving on to less preferred species as the day progressed. For example, in both 2012 and

2013, independent of time of day, T. ohiensis received the second highest community visitation

41

rates, but relatively lower individual visit durations (Figure 2-3a vs. Figure 2-3b; Figure 2-3c vs.

Figure 2-3d). But with respect to time of day, at both the community and individual levels, B. impatiens collected pollen from T. ohiensis most frequently early in the day and then visits decreased as the day progressed (Figure 2-4). We hypothesize that T. ohiensis provided high- quality pollen that was not very abundant (indeed, its floral display was relatively small); therefore, the many foragers that collected pollen from T. ohiensis early in the day exhausted its pollen supply, then moved to a different pollen source. Nevertheless, a few B. impatiens still attempted to collect pollen from T. ohiensis later in the day. Surprisingly, these continued visits to T. ohiensis and also S. hebecarpa, even when their pollen was depleted and pollen was available from the other plant species, included pollen robbing. Foragers would use their mandibles to cut into unopened flowers to collect pollen (48, 49). Senna hebecarpa and T. ohiensis do not provide floral nectar (50, 51), and therefore we hypothesize that these plant species produce high-quality pollen as their only reward. And in a setting with limited resources such as a hoop house, this can lead to pollen robbing behavior if the remaining pollen available from other plant species is of lesser quality.

It is important to consider the relationship between host-plant flowering phenology and time of bee activity. In some settings, few plant species bloom and few pollinator species are active early in the morning (31), but, bumble bees forage earlier in the day and under cooler conditions than most other bee species in temperate environments (52). It is interesting, therefore, that the most preferred host-plant species of B. impatiens in our work (S. hebecarpa, T. ohiensis, and V. virginicum) were the species that bloomed earliest in the day (personal observation). These apparently corresponding phenologies may reflect a coevolutionary relationship between B.

42

impatiens and these host-plant species, and their pollen quality may be particularly suitable for B. impatiens.

As the day progressed, B. impatiens increased their community and individual visitation to

Eupatorium perfoliatum, Eutriochium purpureum, and Symphyotrichum novae-angliae (Figure 2-

4a,b,e,f), suggesting that these lesser preferred species were only visited after the pollen of more preferred species had been collected. We again hypothesize that pollen quality and abundance drove this pattern. Community visitation rates to these host-plant species, which may have lower quality pollen, appeared to increase only after the most preferred plant species were depleted; individual level visit durations to these plant species also increased as the day progressed, suggesting that these species produced high quantities of pollen that sustained longer individual visits. B. impatiens foraging behavior to Eu. purpureum may have been an exception to the overall consistency in the community and individual visit durations; but this discrepancy only appeared when considering the behavior independent of time of day (its community visitation rates were low compared to the large amount of time individuals collected its pollen; Figure 2-3a,e vs. Figure

2-3b,f). This pattern could arise if Eu. purpureum produces low quality pollen, therefore eliciting low community visitation rates, but high quantities of pollen, therefore ensuring long individual visit durations. Notably, the floral displays of Eu. purpureum were of the largest of the plant species tested (data not shown) and also produced a vast amount of pollen (determined by our efforts to collect fresh pollen for a separate study); therefore, it is conceivable that any bee collecting pollen would need to spend more time to adequately cover an individual plant (20).

Flower handling time is one potential constraint to our interpretation of pollen foraging preferences

43

at the individual level. Indeed, for some plant species, bumble bees must invest considerable time to learn how to effectively obtain floral resources (53). Among our plant species, S. hebecarpa had the most complex flower, which required “buzz-pollination” (54), while flowers of the other plant species were “open” with easily accessed resources. Despite the fact that S. herbecarpa resources may be more challenging to obtain, our results indicate S. hebecarpa was strongly preferred by B. impatiens for three reasons: 1) S. hebecarpa received both the vast highest community visitation rates and individual visit durations; 2) individual visit durations did not change to S. hebecarpa throughout the day, suggesting that even as bees learned how to handle the flowers they still spent considerable time collecting resources from these plants; and 3) prior to beginning observations, we allowed B. impatiens to forage from the plant species so that they had the opportunity to learn handling techniques for all flowers.

Conclusion

Our study explored relative bee-foraging preferences for host-plant species through a fine-scale and controlled approach that is unlike many previous studies, including field studies typically used for characterizing plant-pollinator networks. But importantly, our focus differed from that of typical community level plant-pollinator networks because our methodology provides a fundamental framework to address the mechanisms that drive foraging behavior of a single bee species over time, whereas community level studies provide information on the outcome of interactions of these mechanisms. By examining bees at both community and individual levels in a controlled setting that considers and standardizes for differences in flowering phenology between host-plant species, our data support the hypothesis that daily patterns in foraging behavior may be driven by the interplay of resource quality and quantity. In turn, these factors may shape overall

44

plant-pollinator community network interactions over time. Our results demonstrate the importance of considering daily foraging patterns and the resource that pollinators are collecting

(pollen vs. nectar) when evaluating floral preferences. Differentiating between pollen and nectar sources, considering foraging timing, and scrutinizing plant species based on the quality and abundance of their floral resources will allow us to recommend host-plant species for floral resource provisioning schemes that better support larval development and future generations of bees.

Acknowledgements

We would like to thank to Nelson DeBarros and Laura Russo for sharing their data and insight,

Bernardo Niño for helping construct the hoop house, Bekki Waskovich for invaluable help with the field observations, Michael Coccia for help with data analysis, Scott Diloreto for providing space and advice for overwintering our plants, and the Tooker and Grozinger labs for helpful discussions and critical reading of the manuscript. This work was supported by generous funding from an anonymous donation to the Penn State Center for Pollinator Research.

45

Infrequent Moderate High 2010) (DeBarros frequency Visitation p species plant each whether rewards or pollen only. “Years bloomed” representsrepresents years current of study “Reward” graphs. in identification for used code” “Plant 2010). (DeBarros area our in Table 2

- 1. Plant species used for

Tradescantiaohiensis Barneby hebecarpa Senna fistulosaMonarda Veronicastrum virginicum Lamont purpureum Eutrochium purpureaEchinacea Nesom G.L. novae Symphyotrichum Pycnant perfoliatum Eupatorium Species

hemum tenuifolium tenuifolium hemum

Bombus impatiens Bombus

(Fernald) Irwin & Irwin (Fernald) L. (L.) Moench (L.)

Raf. -

angliae foraging observations. L. (L.) (L.) Farw. E.E. (L.)

Scrad. (L.)

Commelinaceae Fabaceae Lamiaceae Scrophulariaceae Asteraceae Asteraceae Asteraceae Lamiaceae Asteraceae Family

Species Species were classified for expected visitation

in which plant species bloomed or were available for data collection. data for available were or bloomed species plant which in

Spiderwort American senna bergamot Wild Culver'sroot Joe coneflower Purple aster England New Mountainmint Boneset name Common - Pye weed Pye

To Sh Mf Vv Jp Ech Ast Pt Bon code Plant

freque pollen pollen pollen/nectar pollen/nectar pollen/nectar pollen/nectar pollen/nectar pollen/nectar pollen/nectar Reward resents both pollen and nectar ncy based on previous work

2012/2013 2013 2012/2013 2012/2013 2012/2013 2012/2013 2012 2013 2012/2013 bloomed Years

46

Figure 2-1. Bombus impatiens foragers actively collecting pollen from Senna hebecarpa, the most preferred host-plant species in 2013. Each forager is recorded as a single pollen foraging visit, and all visits are summed for “community visitation rate”. The time that each forager spends collecting pollen on the individual plant is recorded as “individual visit duration”. Image by Anthony Vaudo.

Figure 2-2. Design of B. impatiens foraging preference experiments. a) An image of the hoop house including the grid of flowering plant species and the tent housing bumble bee colonies in the background. Image by Anthony Vaudo. b) A schematic of the hoop house (11 m length x 6.1 m width) and experimental design. Black circles represent individual plants in 7.6 L pots. Individual plants from each plant species were randomized in the grid each day.

47

Figure 2-3. Bombus impatiens general community visitation rates and individual visit durations (means ± SE) (independent of time of day) by plant species in 2012 and 2013 (Table 2-1 defines plant codes). Bars within graphs labeled with different letters are statistically different (P < 0.05); see text for details on statistics. Figures: a) community visitation rates in 2012; b) individual visit durations in 2012; c) visitation rates in 2013; d) visit durations in 2013; e) visitation rates in 2013 excluding Senna hebecarpa; f) visit durations in 2013 excluding S. hebecarpa.

48

Figure 2-4. Interactions of community visitation rates and individual visit durations by plant species and time of day in 2012 and 2013 (Table 2-1 defines plant codes). Data are represented as relative means totaling 100% at each time period. Figures: a) community visitation rates in 2012; b) individual visit durations in 2012; c) visitation rates in 2013; d) visit durations in 2013; e) visitation rates in 2013 excluding Senna hebecarpa; f) visit durations in 2013 excluding S. hebecarpa.

49

References

1. Ricketts TH, et al. (2008) Landscape effects on crop pollination services: are there general patterns? Ecol Letters 11(5):499–515.

2. Potts SG, et al. (2010) Global pollinator declines: trends, impacts and drivers. Trends Ecol Evol 25(6):345–353.

3. Kremen C, Williams NM, Thorp RW (2002) Crop pollination from native bees at risk from agricultural intensification. Proc Natl Acad Sci USA 99(26):16812–16816.

4. Klein AM, et al. (2007) Importance of pollinators in changing landscapes for world crops. Proc Roy Soc B 274(1608):303–313.

5. Aizen MA, Garibaldi LA, Cunningham SA, Klein AM (2009) How much does agriculture depend on pollinators? Lessons from long-term trends in crop production. Ann Bot 103(9):1579–1588.

6. Carvalheiro LG, Seymour CL, Veldtman R, Nicolson SW (2010) Pollination services decline with distance from natural habitat even in biodiversity-rich areas. J Appl Ecol 47(4):810–820.

7. Eilers EJ, Kremen C, Smith Greenleaf S, Garber AK, Klein A-M (2011) Contribution of Pollinator-Mediated Crops to Nutrients in the Human Food Supply. PLoS ONE 6(6):e21363.

8. Decourtye A, Mader E, Desneux N (2010) Landscape enhancement of floral resources for honey bees in agro-ecosystems. Apidologie 41(3):264–277.

9. Carvalheiro LG, et al. (2011) Natural and within-farmland biodiversity enhances crop productivity. Ecol Lett 14(3):251–259.

10. Carvalheiro LG, Seymour CL, Nicolson SW, Veldtman R (2012) Creating patches of native flowers facilitates crop pollination in large agricultural fields: mango as a case study. J Appl Ecol 49(6):1373–1383.

11. Holzschuh A, Dudenhöffer J-H, Tscharntke T (2012) Biological Conservation. Biol Conserv 153(C):101–107.

12. Pontin DR, Wade MR, Kehrli P, Wratten SD (2006) Attractiveness of single and multiple species flower patches to beneficial insects in agroecosystems. Ann Appl Biol 148(1):39– 47.

13. Tuell JK, Fiedler AK, Landis D, Isaacs R (2008) Visitation by wild and managed bees (Hymenoptera: Apoidea) to eastern U.S. native plants for use in conservation programs. Environ Entomol 37(3):707–718.

14. DeBarros NB (2010) Floral Resource Provisioning for Bees in Pennsylvania and the Mid-

50

Atlantic Region. Master's thesis, The Pennsylvania State University.

15. Gibson RH, Knott B, Eberlein T, Memmott J (2010) Sampling method influences the structure of plant-pollinator networks. Oikos 120(6):822–831.

16. Lunau K, Wacht S, Chittka L (1996) Colour choices of naive bumble bees and their implications for colour perception. J Comp Physiol A 178(4):477–489.

17. Raine NE, Ings TC, Dornhaus A, Saleh N, Chittka L (2006) Adaptation, genetic drift, pleiotropy, and history in the evolution of bee foraging behavior. Adv Stud Behav 36:305- 354.

18. Dobson HEM (1987) Role of flower and pollen aromas in host-plant recognition by solitary bees. Oecologia 72(4):618–623.

19. Johnson S, Linder H, Steiner K (1998) Phylogeny and radiation of pollination systems in DISA (Orchidaceae). Am J Bot 85(3):402–402.

20. Mitchell RJ, Karron JD, Holmquist KG, Bell JM (2004) The influence of Mimulus ringens floral display size on pollinator visitation patterns. Funct Ecol 18(1):116–124.

21. Nattero J, Malerba R, Medel R, Cocucci A (2011) Factors affecting pollinator movement and plant fitness in a specialized pollination system. Plant Syst Evol 296(1-2):77–85.

22. Greenleaf SS, Kremen C (2006) Wild bees enhance honey bees' pollination of hybrid sunflower. Proc Natl Acad Sci USA 103(37):13890–13895.

23. Waser NM (1986) Flower constancy: definition, cause, and measurement. Am Nat 127(5):593-603

24. Saleh N, Chittka L (2006) Traplining in bumblebees (Bombus impatiens): a foraging strategy’s ontogeny and the importance of spatial reference memory in short-range foraging. Oecologia 151(4):719–730.

25. Burkle LA, Alarcon R (2011) The future of plant-pollinator diversity: Understanding interaction networks across time, space, and global change. Am J Bot 98(3):528–538.

26. Devoto M, Bailey S, Memmott J (2013) Ecological meta-networks integrate spatial and temporal dynamics of plant-bumble bee interactions. Oikos 123(6):714–720.

27. Peat J, Goulson D (2005) Effects of experience and weather on foraging rate and pollen versus nectar collection in the bumblebee, Bombus terrestris. Behav Ecol Sociobiol 58(2):152–156.

28. Michener CD (2007) The Bees of the World (Johns Hopkins University Press).

29. Petanidou T, Kallimanis AS, Tzanopoulos J, Sgardelis SP, Pantis JD (2008) Long-term observation of a pollination network: fluctuation in species and interactions, relative

51

invariance of network structure and implications for estimates of specialization. Ecol Lett 11(6):564–575.

30. Basilio AM, Medan D, Torretta JP, Bartoloni NJ (2006) A year-long plant-pollinator network. Austral Ecol 31(8):975–983.

31. Baldock KCR, Memmott J, Ruiz-Guajardo JC, Roze D, Stone GN (2011) Daily temporal structure in African savanna flower visitation networks and consequences for network sampling. Ecology 92(3):687–698.

32. Russo L, DeBarros N, Yang S, Shea K, Mortensen D (2013) Supporting crop pollinators with floral resources: network-based phenological matching. Ecol Evol 3(9):3125–3140.

33. Mosquin T (1971) Competition for pollinators as a stimulus for the evolution of flowering time. Oikos 22(3):398.

34. Lack AJ (1982) Competition for pollinators in the ecology of Centaurea scabiosa L. and Centaurea nigra L. New Phytol 91(2):297-308.

35. Stone GN, Willmer P, Rowe JA (1998) Partitioning of pollinators during flowering in an African Acacia community. Ecology 79(8):2808–2827.

36. Raine NE, Pierson AS, Stone GN (2007) Plant–pollinator interactions in a Mexican Acacia community. Arthropod-Plant Inte 1(2):101–117.

37. Mitchell TB (1962) The Bees of the Eastern United States. II. Technical bulletin North Carolina Agricultural Experiment Station 152:1–557.

38. Lozier JD, Strange JP, Stewart IJ, Cameron SA (2011) Patterns of range-wide genetic variation in six North American bumble bee (Apidae: Bombus) species. Mol Ecol 20(23):4870–4888.

39. Velthuis HHW, van Doorn A (2006) A century of advances in bumblebee domestication and the economic and environmental aspects of its commercialization for pollination. Apidologie 37(4):421–451.

40. Szabo TI, Pengelly DH (1973) The over-wintering and emergence of Bombus (Pyrobombus) impatiens (Cresson) (Hymenoptera: Apidæ) in southern Ontario. Insect Soc 20(2):125–132.

41. Cnaani J, Schmid-Hempel R, Schmidt JO (2002) Colony development, larval development and worker reproduction in Bombus impatiens Cresson. Insect Soc 49(2):164–170.

42. Saleh N, Scott AG, Bryning GP, Chittka L (2007) Distinguishing signals and cues: bumblebees use general footprints to generate adaptive behaviour at flowers and nest. Arthropod-Plant Inte 1(2):119–127.

43. Ohashi K, Thomson JD (2009) Trapline foraging by pollinators: its ontogeny, economics

52

and possible consequences for plants. Ann Bot 103(9):1365–1378.

44. Stone GN, et al. (1999) Windows of opportunity and the temporal structuring of foraging activity in a desert solitary bee. Ecol Entomol 24(2):208–221.

45. Vázquez DP, Aizen MA (2004) Asymmetric specialization: a pervasive feature of plant- pollinator interactions. Ecology 85(5):1251–1257.

46. Jordán F, Osváth G (2009) The sensitivity of food web topology to temporal data aggregation. Ecol Model 220(22):3141–3146.

47. Nicolson SW (2011) Bee Food: The chemistry and nutritional value of nectar, pollen and mixtures of the two. Afr Zool 46(2):197–204.

48. Hargreaves AL, Harder LD, Johnson SD (2009) Consumptive emasculation: the ecological and evolutionary consequences of pollen theft. Biol Rev 84(2):259–276.

49. Irwin RE, Bronstein JL, Manson JS, Richardson L (2010) Nectar robbing: ecological and evolutionary perspectives. Annu Rev Ecol Evol Syst 41(1):271–292.

50. Faden RB (1992) Floral attraction and floral hairs in the Commelinaceae. Ann MO Bot Gard 79(1):46.

51. Marazzi B, Sanderson MJ (2010) Large-scale patterns of diversification in the widespread legume genus Senna and the evolutionary role of extrafloral nectaries. Evolution 64(12):3570–3592.

52. Goulson D (2010) Bumblebees: behaviour, ecology, and conservation. (Oxford University Press).

53. Raine NE, Chittka L (2006) Pollen foraging: learning a complex motor skill by bumblebees (Bombus terrestris). Naturwissenschaften 94(6):459–464.

54. Marazzi B, Conti E, Endress PK (2007) Diversity in anthers and stigmas in the buzz- pollinated genus Senna (Leguminosae, Cassiinae). Int J Plant Sci 168(4):371–391.

53

Chapter 3. Macronutrient ratios in pollen shape bumble bee (Bombus impatiens) foraging strategies and floral preferences

Anthony D. Vaudo *a, Harland M. Patcha, David A. Mortensenb, John F. Tookera, Christina M.

Grozingera

a Department of Entomology, Center for Pollinator Research, The Pennsylvania State University,

501 ASI Building, University Park, PA 16802, USA

b Department of Plant Science, Center for Pollinator Research, The Pennsylvania State University,

422 ASI Building, University Park, PA 16802, USA

Abstract

To fuel their activities and rear their offspring, foraging bees must obtain a sufficient quality and quantity of nutritional resources from a diverse plant community. Pollen is the primary source of proteins and lipids for bees, and among host-plant species, it can vary widely in concentrations of these nutrients. Therefore, we hypothesized that foraging decisions of bumble bees are driven by both the protein and lipid content of pollen. By successively reducing environmental and floral cues, we analyzed pollen foraging preferences of Bombus impatiens to 1) host-plant species, 2) pollen isolated from these host-plant species, and 3) nutritionally modified single-source pollen diets encompassing a range of protein and lipid concentrations. In our semi-field experiments, B. impatiens foragers exponentially increased their foraging rates of pollen from plant species with high protein:lipid (P:L) ratios, where the most preferred plant species had the highest ratio (~4.6:1).

54

These preferences were confirmed in cage studies where, in absence of other floral cues, B. impatiens workers still preferred in pairwise comparisons pollen with higher P:L ratios. Finally, when presented with nutritionally modified pollen, workers were most attracted to pollen with 5:1 and 10:1 P:L ratios, but increasing the protein and/or lipid concentrations (while leaving ratios intact) reduced attraction. Thus, macronutritional ratios appear to be a primary factor driving bee pollen foraging behavior and may explain observed patterns of host-plant visitation across the landscape. Nutritional quality of pollen resources should be taken into consideration when designing conservation habitats supporting bee populations.

Keywords: behavior, foraging, nutrition, nutritional ecology, pollen, pollinator, preferences

Significance Statement

Bees pollinate the majority of flowering plant species, including agricultural crops. The pollen they obtain is their main protein and lipid source and fuels development and reproduction. Bee populations are declining globally, due in large part to landscape-level loss of host-plant species, contributing to a nutritional shortage. To mitigate declines, we must understand how nutritional requirements of bees influence foraging behavior. We demonstrate that bumble bees selectively collect pollen from host-plant species based on the protein:lipid ratios of pollen. Our research indicates that bees evaluate pollen quality and adjust foraging decisions to meet their nutritional needs. To be effective, conservation initiatives must include host-plant species that provide pollen that satisfies the nutritional demands of bees to support their populations.

55

Introduction

Foraging animals must obtain from their environments appropriate nutrients for growth, development, and reproduction. Bees forage in a very complex and changing environment, where floral nutritional resources (nectar and pollen) vary widely in quality and quantity among plant species (1). These resources are accompanied by myriad floral cues, including floral odors, color, morphology, and display area, and can vary dramatically in spatiotemporal availability, all of which may influence and reinforce foraging decisions (2, 3). Worldwide declines in populations of bees and other pollinators have been linked to reduced diversity and abundance of host-plant species, likely placing bees under nutritional stress (4, 5). To develop strategic conservation protocols that preserve or restore foraging habitat that supports healthy pollinator populations, we must understand how bees forage in their environments to meet their nutritional needs. It is well established that solitary and social insects can selectively forage and regulate their intake of synthetic diets spanning a range of macronutrient nutritional qualities to reach their optimal, species-specific, nutritional intake (6-8). Here, we examine whether the generalist bumble bee species Bombus impatiens Cresson forages selectively among different plant species and pollen sources for specific macronutrient ratios.

Pollen is the primary source of proteins, lipids and other micronutrients for bees, and is necessary for brood rearing, reproduction, and health (1, 9-17). However, pollen nutritional quality varies widely among plant species, ranging from 2-60% protein and 1-20% lipids by weight (10, 18); thus, it is likely critical that bees selectively collect pollen species with the necessary nutritional quality to support their needs (1). Protein and amino acid concentrations of pollen modulate immunocompetence in honey bees (16, 19) and reproduction in bumble bees (ovary activation and

56

larval development; [12, 14, 20-22]). Furthermore, lipids are key to a variety of physiological processes in insects, including molting hormone production (23), and high sterol content in pollen may increase bumble bee larval size and growth (21). Recently, deficiency in linolenic fatty acid

(an essential fatty acid) in honey bees has been linked to reduced learning and development of brood-food producing glands (24).

There is some evidence that foraging bees can select host-plant species based on pollen protein content. While foraging in the same landscape, bumble bees foraged preferentially on plant species with higher protein content than honey bees (25), suggesting species-specific differences in protein acquisition. Bumble bee workers can taste and discriminate among diets of different protein or pollen concentrations (26) and their foraging activity has been positively correlated with pollen protein content using modified (diluting with cellulose powder) single-source pollen diets (27, 28) or a single plant species where pollen protein content varied with soil conditions (20). (Note that in field studies honey bees do not appear to forage preferentially on pollen of higher protein concentrations [29, 30]). However, diluting pollen with cellulose powder may simply make diets less attractive by reducing all pollen cues, while modifying soil conditions may alter factors other than pollen protein that may influence bee choice. Only two studies have demonstrated a correlation between bumble bee foraging preference and pollen protein content in landscapes with multiple plant species (31, 32). Therefore, it is uncertain if bumble bees truly seek out the host- plant species with higher pollen protein or if their host-plant species choice is driven by other factors.

57

From these previous studies, it is also unclear if bumble bees selectively forage to meet multiple macronutrient needs, or if they forage to maximize the quantity of a single macronutrient (e.g., protein). Other studies have demonstrated that the pollenkitt, the lipid-dominated, oily outer surface of entomophilous pollen, contains important discriminative stimuli for bees (33-36).

Furthermore, given the nutritional importance of lipids, bees may assess the ratios of proteins and lipids when they are foraging for pollen. Indeed, other arthropod species (e.g., beetles, spiders) can regulate their dietary intake and forage selectively to meet specific ratios of lipids and proteins

(37-39).

It is unknown, therefore, what are the preferred pollen nutritional qualities for bee species, if bee host-plant species choice is driven by optimal pollen quality on multiple nutritional dimensions

(i.e., not specifically maximizing nutrient acquisition), or if the preference for optimal pollen nutrition is maintained in the absence of external floral cues. Our previous work demonstrated that foragers of Bombus impatiens Cresson (Hymenoptera: Apidae), the Common Eastern Bumble Bee, exhibited distinct pollen foraging preferences among nine host-plant species (40). In the current study, we tested the mechanistic basis for these foraging preferences. First, we determined if B. impatiens pollen-foraging preferences and visitation rates to different host-plant species related to pollen nutritional composition, and whether B. impatiens evaluates macronutrient (protein, lipid and carbohydrate) levels individually or in ratios. Second, to determine if B. impatiens workers maintained plant species preferences in absence of external floral cues, we evaluated bumble bee foraging preferences to isolated pollen. Finally, we modified the protein and lipid concentrations of single-source pollen to determine whether B. impatiens worker preferences were related to P:L ratios or increased nutrient concentration.

58

Results

Host-plant pollen foraging preferences

To test if foraging preferences among host-plant species were associated with pollen nutritional quality, we collected extensive simultaneous foraging visitation data of B. impatiens colonies to multiple plant species (controlling for phenology and floral area/resource availability) and analyzed each host-plant species’ pollen nutritional content (carbohydrate, protein, and lipid concentrations; Table 1). All nutritional values were within the expected range from pollen (18).

Foraging data and methodology were previously published in (40) and briefly described in the methods section below.

We used multiple regression analysis to quantify the relationship between nutrient concentration and visitation rate because each nutritional component of pollen is not independent from one another. Carbohydrate concentration and the interaction of protein and carbohydrate did not influence foraging rates (carbohydrate: P = 0.47; protein x carbohydrate: P = 0.63). Interestingly, protein concentration, lipid concentration, and their interaction were significantly associated with foraging rate (protein: P < 0.01; lipid: P < 0.01; protein x lipid: P < 0.01; Model: F5,1776 = 57.53,

P < 0.01, R2 = 0.14). We explored this interaction further, and found that bumble bees exponentially increased their visitation rates to the plant species as the protein:lipid ratio (P:L) of the pollen increased (Y = 0.0025e1.03*X, R2 = 0.96; Figure 3-1). Importantly, protein:carbohydrate ratios, often considered the main nutritional drivers in arthropod herbivore foraging behavior and

2 nutrient regulation (6), did not influence bumble bee host-plant choice (F1,4 = 0.02, P = 0.89, R =

0.006). Because this trend may have been driven by the bumble bees’ most preferred plant species, we reanalyzed the data excluding Senna hebecarpa and found the same trend; as P:L ratio of pollen

59

increased, so too did visitation rate (Y = 0.013e0.46*X, R2 = 0.67). Furthermore, the peak foraging rate to each plant species throughout the day (see (40) for details), followed the order of highest to lowest P:L value (Figure 3-1), suggesting that bees visit the plant species with the highest P:L ratios first, and, once that pollen has been depleted, move on to the next highest P:L ratio plant species. Overall, these data indicate that P:L ratio drives bumble bee preferences for host-plant species. Notably, Senna hebecarpa (with poricidal anthers) and Tradescantia ohiensis, the two most preferred host plant species and those with the highest P:L ratios we tested, were “buzz- pollinated” by the bumble bees (indicating the specialized behavior needed to extract their rewards) and do not produce nectar, suggesting that these plant species may have evolved to produce “high quality” pollen as their only reward.

Isolated pollen feeding preference assay

To determine pollen preferences independent of other floral and environmental cues and factors influencing foraging decisions, we hand-collected pollen from six host-plant species (Table 3-1).

Two different pollen species were then presented to caged B. impatiens workers in a pairwise choice test (Figure 3-2a). In each trial, we evaluated caged bees for their preferences in all possible combinations of 4 pollen species. Across the five trials, six possible pollen-species were used (it was not possible to compare pairwise combinations across all six pollen species in each trial due to limited amounts of pollen). To integrate the data across all the trials for analysis, within each trial we ranked the pollen species according to P:L ratios (with the pollen with the highest P:L ratio given a rank of "1" and the pollen with the lowest P:L ratio in the trial being given a rank of

"4"), allowing us to compare preferences based on relative P:L ratios rather than species of origin.

60

When evaluating the preferences of bees within each cage, the pollen with the higher P:L ratio of each paired choice, independent of plant species, was significantly more preferred in terms of the

2 number of feeding events, or attractiveness (in 73% of the cages, � (2, N= 30) = 20.03, P < 0.01,

2 Figure 3-3) and mg pollen consumed (in 63% of the cages, � (2, N= 30) = 13.6, P = 0.011; Figure 3-

3).

Across all pairwise combinations and trials, nutritional rank positively influenced the number of feeding events observed, or attractiveness of pollen (F3,52 = 5.52, P = 0.002; Figure 3-2b,c): the frequency of feeding events decreased as the P:L ratio of the pollen decreased (Figure 3-2b,c).

Nutritional rank also positively influenced the amount of the pollen consumed with the highest

P:L pollen eaten the most, and lowest P:L pollen eaten the least (F3,52 = 19.18, P < 0.0001; Figure

3-4). Thus, in the absence of floral cues, bumble bees consumed pollen more frequently and in larger quantities as its P:L ratio increased, consistent with their preferences observed in our host- plant foraging study. These results all the more striking given the bees in this experiment were never exposed to these plant species yet still made choices based on P:L ratio.

Modified pollen feeding preference assay

Using the same experimental design as used in the “Isolated pollen feeding preference assay”

(Figure 3-2a), we presented Bombus impatiens workers with paired choices of nutritionally modified honey-bee-collected (HB) pollen with P:L ratios and protein and lipid concentrations similar to or greater than our fresh collected pollen (Figure 3-5). HB pollen alone had a P:L ratio of 1.6:1. We found that the P:L ratio of the diets influenced the number of feeding events observed

(see Figure 3-6 for details about results of individual pairwise comparisons of diets, and Figure 3-

61

5 for compiled results across all pairwise comparisons). Diets of 5:1 and 10:1 P:L were more attractive (visited more frequently) than 1.6:1 (HB pollen) or 25:1 diets (F6,157 = 9.43, P < 0.0001;

Figure 3-5a). However, B. impatiens consumed more of the 1.6:1 (HB pollen) and 5:1 than 10:1 or 25:1 (F6,157 = 24.20, P < 0.0001; Figure 3-5b). These data indicate that 5:1 and 10:1 P:L diets are the most attractive and higher concentrations of protein in the diet lead to reduced consumption.

When we held the P:L ratio constant, concentrations of proteins and lipids above those found in

HB pollen reduced feeding events and the amount of pollen consumed (Figure 3-5) suggesting that simply increased nutrient concentrations can actually be unattractive to bumble bees. Overall, when considering both attraction and feeding, we infer that B. impatiens preferred the 5:1 P:L diet, confirming the results of our previous experiments.

Because pollen consumption of B. impatiens appeared to be significantly influenced by nutrient concentration (Figure 3-5b), we analyzed effects of protein and lipid concentration on amounts of each diet consumed. As absolute protein concentration increased across all diets, the consumption

2 of the diet decreased linearly (F1,166 = 108.17, P < 0.0001, R = 0.44; Figure 3-7a), suggesting that

B. impatiens could obtain similar levels of proteins by eating larger amounts of low-concentration diets or small amounts of high-concentration diets. However, lipid concentration had a biexponential effect on amount of pollen consumed: bees increased their pollen consumption as lipid increased on low concentration diets, but ate less on high lipid concentration diets (Y =

1558*e-26.8*X – 1600*e-28.0*X, R2 = 0.44; Figure 3-7b). These data indicate that lipid concentration is responsible for attractiveness and phagostimulation at low concentrations, but increased concentrations could lead to satiation similar to the protein concentration data.

62

Discussion

Our results demonstrate that the macronutrient ratios in pollen are a key factor determining bee foraging behavior. Bombus impatiens discriminated among plant species based on pollen nutritional quality, and in field and lab assays exhibited preference for species or isolated pollen with the highest P:L ratios (~4.6:1; Figure 3-1 and 2). When presented pollen with altered P:L ratios or protein and lipid concentrations (including a 25:1 ratio, which was higher than what was produced by the plant species in our study, or pollen with the same ratio but increasing protein and lipid concentrations), B. impatiens still preferred the 5:1 and 10:1 diets, suggesting that B. impatiens workers seek to optimize their nutritional intake, and do not simply try to maximize the amount of protein in their diets. Indeed, increased nutrient concentrations of both proteins and lipids actually lead to decreased attraction and consumption of diets overall (Figure 3-5 and 3-7), though at lower concentrations lipids appeared to have had a phagostimulatory effect. Thus, B. impatiens, and likely other bumble bee and generalist bee species, appear to have a sophisticated ability to assess pollen nutritional quality and selectively forage to reach nutritional intake targets

(6, 41-43).

Comparing across plant species, isolated pollen, and manipulating P:L ratios in pollen, allowed us to demonstrate that bumble bee foraging preferences were driven by simultaneous assessments of multiple nutritional components of pollen, specifically the P:L ratio. Among the host-plant species we tested, carbohydrates, proteins, and lipids independently were insufficient to explain bumble bee foraging behavior. Protein and lipid concentrations were significantly associated with foraging rates, but also showed a significant interaction. Exploring this interaction further revealed that the ratio of these two components was the major driver of foraging rates. Furthermore, though

63

protein:carbohydrate ratios are commonly associated with nutrient regulation in herbivores (6), these ratios also did not correlate to host-plant species pollen preferences. Interestingly, higher protein concentrations in modified diets led to fewer feeding events and reduced pollen consumption. It is important to note that in previous experiments that found increasing protein concentration increased attraction (27, 28), the “protein composition” of pollen was modified by diluting it with cellulose powder, thereby reducing all nutrient concentrations and sensory cues.

Furthermore, our studies demonstrate that B. impatiens are not simply avoiding pollen (or diets) of high lipid content: one of their most preferred plant species, T. ohiensis, had relatively high lipid concentrations (Table 1), and they preferentially consumed modified diets with moderate lipid concentrations. Thus, similar to other insects (caterpillars, predators (37, 38, 44)), B. impatiens appears to regulate intake of dietary P:L ratios, which may at least partially drive feeding behavior in bees.

Despite importance of P:L ratios, other components in pollen, such as micronutrients, or individual amino or fatty acids, may still influence bee foraging decisions (45-47). Indeed, pollenkitt, which includes free fatty and amino acids that vary among pollen species, is critical for pollen recognition and phagostimulation (34-36, 48, 49). Furthermore, as generalist foragers, bumble bees may avoid or dilute negative effects of toxic phytochemicals by collecting pollen from multiple host-plant species (50, 51). It is unclear, however, whether micronutrient variation in pollen or secondary plant metabolites can alter bee foraging for macronutrients, or whether concentrations of these compounds are somehow associated with the macronutrient levels we measured. Because we manipulated P:L ratios in our study using the same honey bee-collected pollen samples, the

64

amounts of micronutrients and other chemicals remained constant (or were reduced in concentration) and thus were likely not a factor in this analysis.

Our results suggest that foraging bumble bees may assess pollen nutritional quality via both pre- and post-ingestive processes, but additional studies are needed to fully evaluate the proximate mechanisms underlying the preferences we observed. In the field, we observed bees antennating pollen, potentially assessing its quality. Bumble bees appear to determine pollen quality by its protein content through tactile chemoreceptors and show preferences for high protein pollen (26,

52), whereas honey bees do not appear to share the same preference (25, 29, 30). Although both species may be sensitive to protein quality, the preferences observed in these previous studies may be due to species specific differences in nutritional requirements for protein and lipids. Bees can also discriminate between pollen types and may be able to assess pollen quality via pollenkitt, or volatile chemical profiles (34-36, 48, 49). Pollenkitt contains lipids, proteins, and carbohydrates

(35), which bees could detect directly prior to ingesting or collecting. The diverse and large number of odor receptors and glomeruli of bees may allow them to decipher the chemical composition of the pollenkitt (53). Bumble bees may ingest some pollen while foraging (potentially while grooming and packing pollen) or when returning to the hive, and the nutritional quality of the pollen may then influence sensory physiology via post-ingestive effects (54, 55). Further evidence for post-ingestive effects comes from our studies using modified pollen diets, where B. impatiens workers most frequently ate pollen with high P:L ratios (5:1 and 10:1), but consumed more of the

1.6:1 and 5:1 diets than the 10:1 and 25:1 diets. Reduced consumption of diets with higher protein and/or lipid concentrations indicate post-ingestive effects, because eating less food of a higher concentration may result in the same quantity of nutrients ingested. Post-ingestive behavioral

65

responses to pollen quality may involve changes in chemo- or gustatory receptor sensitivity based on haemolymph concentration of nutrients after consumption (56). Finally, bumble bees may receive social information about pollen quality and availability when pollen is stored in pollen pots by returning foragers (27, 55). Regardless of whether foragers use pre- or post-ingestive effects to evaluate pollen nutritional quality, there is ample evidence that bumble bees can learn rewarding flowers after only a handful of foraging bouts, quickly establishing traplines to the highest quality flowers (57-59). Overall, if foragers themselves are able to assess and compare nutrient ratios among available types of pollen, it would be the most efficient way for the colony to regulate nutritional intake.

Though the nutritional requirements of bees vary throughout their lifetime and differ substantially between developing brood and adult foragers (9, 42), foraging preferences of B. impatiens workers in our study remained consistent in presence and absence of their mother colony. In our host-plant- pollen foraging experiment, workers collected pollen for their colonies, which had queens and developing larvae. In the caged feeding assay, workers were kept separate from brood, but they still exhibited the same plant-species preferences. Here, they may have had social influence on each other’s feeding behavior; however, the results were replicated across 30 cages and the group preference was clearly consistent. Thus, foraging preferences and nutritional needs of foragers may be closely matched to larvae, though additional studies are needed to comprehensively examine the foraging preferences of colonies with and without brood. Furthermore, though we demonstrated that B. impatiens were consistently attracted to their preferred P:L ratio, ~5-10:1 P:L

(Figures 1, 2bc, 3a), increasing concentration of protein and lipids had different effects on feeding behavior (Figure 3-7). Additional studies are needed to determine if these ratios are optimal to

66

support bumble bee fitness (6, 42, 43). Indeed, in studies using the geometric framework for nutrition approach (6, 60, 61), bees can regulate their protein and carbohydrate intake (42,43), and other arthropod species can regulate their protein and lipid intake to support fitness (37-39).

The results of these studies can readily be applied to management and conservation of pollinator populations. Bumble bees are critical pollinators of many agricultural crops (62). As generalists foragers, covering large geographic ranges, supporting colonies over entire growing seasons, and populations of up to 500 workers, they forage on a wide array of plant species (63-68).

Unfortunately, populations of approximately half of the bumble bee species in North America and

Europe are declining (69-71). One of the key factors driving this decline (and those of other bee species) is loss of habitat and the associated nutritional resources provided by a diversity of flowering plant species (72, 73). In depauperate landscapes, bees likely do not have access to the diversity of host-plant species needed to adequately self-select their diet and balance their nutritional intake. This lack of resources may reduce colony growth, health, and reproduction, negatively influencing long-term bee populations. Once we better understand pollen nutritional values across diverse, commercially available plant species, floral provisioning protocols could address nutritional shortcomings by restoring pollen sources that allow bees to balance macro- and micronutrients and phytochemical pollen components that differ among plant species (1). By selecting plant species that better satisfy nutritional requirements of pollinators, the effectiveness of these schemes can undoubtedly be greatly improved and optimized.

Conclusion

67

In this study, a generalist pollinator species discriminated among host-plant species according to nutritional quality. Notably, preference of B. impatiens for P:L ratios remained remarkably stable across different conditions: 1) foraging among host-plant species, pollen isolated from flower species, and nutritionally modified single-source pollen, and 2) foraging in the presence or absence of a colony with developing brood. These results suggest that bees may consistently navigate through a variety of environmental influences to find optimal pollen resources. Furthermore, optimizing P:L intake may improve fitness of B. impatiens, as is the case in other insect species

(6, 37-39), though additional research must examine effects on individual, larval, and colony health and productivity. The species-specific nutritional needs and preferences of bees should be considered when designing protocols and policies for conservation and management of bee populations.

Materials and Methods

Pollen nutritional analysis

We collected pollen from 16 individuals of seven perennial pollinator host-plant species native to

Pennsylvania (Table 1). The individual plants were reared in pots outdoors and used to study bumble bee foraging preferences (40). When not in use for collecting foraging data, they were stored in outdoor field cages to exclude any floral visitors and allow efficient pollen collection.

Fresh pollen was collected from the plant species by gently brushing the pollen off of flowers into a glass container. Because S. hebecarpa has poricidal anthers, we collected whole anthers into a glass container and vortexed them to release the pollen. All pollen was stored at -20°C until analysis or use in experiments. Pollen was dried for ~24 hours at 36°C for analysis. To analyze the

68

protein, lipid, and carbohydrate concentrations of pollen, we divided the pollen into three 1mg replicates for protein analysis and three 1mg replicates for lipid and carbohydrate analysis.

We analyzed the protein concentration of pollen using the Bradford assay. To prepare the samples for analysis, we dried the pollen for 24 hours at 36°C. Then we divided the pollen into three 1mg replications for each individual plant species in 1.5mL Eppendorf microcentrifuge tubes

(Eppendorf North America, Hauppauge, NY). To facilitate breaking of the pollen wall, three drops of 0.1M NaOH were added to each sample and then ground with a microcentrifuge pestle. After grinding, the sample was filled to 1.5mL of 0.1M NaOH and vortexed. All samples were allowed to sit for 24 hours and centrifuged to precipitate all debris/solids. We conducted the Bradford assay with the Bio-Rad Protein Assay Kit microassay 300 µL microplate protocol using bovine γ- globulin as the protein standard (Bio-Rad Laboratories, Inc., Hercules, CA). Due to the high protein concentration of the pollen, we diluted 50 µL of each replicate into 100 µL 0.1M NaOH in each well of a BD Falcon 300 µL sterile non-tissue culture treated 96 well plate (BD, Franklin

Lakes, NJ). Absorbance readings at 595nm were measured using a SpectraMax 190 spectrophotometer (Molecular Devices, LLC, Sunnyvale, CA) and protein concentrations calculated using simple linear regression analysis from the protein standards using SoftMax Pro v.4.0 software (Molecular Devices, LLC 2001).

Pollen lipid and carbohydrate concentrations were determined using a modified protocol from Van

Handel and Day 1988 (74). To prepare the samples for analysis, we divided dried pollen into three

1mg replications for each individual plant species in 1.5mL Eppendorf microcentrifuge tubes. We added 0.2mL 2% sodium sulfate into each tube and homogenized the samples with a

69

microcentrifuge pestle. We washed each sample into a glass tube with 1.6mL chloroform/methanol

(1:1 v:v) and centrifuged the samples at 3000 rpm for 5 min, separating all solids, including the indigestible pollen exines and intines (including cellulose), from the lipid and sugar extract. We transferred the supernatant to a clean glass tube, added 600 µL DI water, and centrifuged the sample at 3000rpm for 5 min. We separated the top carbohydrate/water/methanol fraction for sugar analysis and remaining chloroform fraction was used for lipid analysis. For carbohydrate analysis, we heated each sample at 100°C to evaporate the solvent to ~100µL. We added anthrone/sulfuric acid reagent to equal 5mL and heated at 100°C for 17 min. Each sample was removed from the heat and allowed to cool. We used two technical replications for each biological replication and measured absorbance at 625nm using a SpectraMax 190 spectrophotometer. Carbohydrate concentrations were calculated using simple linear regression analysis from anhydrous glucose standards using SoftMax Pro v.4.0 software. The lipid/chloroform fraction was heated at 100°C to evaporate the solvent. We added 0.2mL sulfuric acid to the sample and heated at 100°C for 10min and then added vanillin/phosphoric acid reagent to equal 5mL, removed from heat, and allowed to cool. We used two technical replications for each biological replication and measured absorbance at 525nm using a SpectraMax 190 spectrophotometer. Lipid concentrations were calculated using simple linear regression analysis from vegetable oil (soybean based; Crisco®, The J. M. Smucker

Company, Orville, OH) standards using SoftMax Pro v.4.0 software. Pollen concentrations of protein, carbohydrate, or lipids is reported as µg nutrient/mg pollen and subsequent protein:carbohydrate (P:C) and protein:lipid ratios (P:L) were determined for each plant species

(Table 1). All nutritional values were within the expected range from pollen (10). Pollen concentrations of protein, lipids, or carbohydrate are reported as µg nutrient/mg pollen and

70

subsequent protein:carbohydrate (P:C) and protein:lipid ratios (P:L) were determined for each plant species (Table 1).

Assessing host-plant pollen foraging preferences

Utilizing “community visitation rate” data we collected in 2013 (40), we correlated pollen foraging rates of B. impatiens to different host-plant species to measures of pollen nutritional quality. In a controlled foraging arena or hoop house (11 x 6.1 m, 3.05m height), we confined two B. impatiens colonies to perennial host-plant species (3-4 individuals of eight species; Table 1). On 18 separate days over the course of 5 weeks, we continuously recorded (from 0930-1315 DST) the frequency that bumble bee foragers collected pollen from each plant species (visits/min). Note that colonies were allowed to acclimate and learn pollen handling techniques to flowers three days prior to data collection (75), and each day of data collection, colony entrances were opened at 0930, ensuring that foragers were active only when all species' flowers were presenting pollen and an observer was present. We also measured the number of workers in each colony each week to standardize the data for the bumble bee foraging population. Each day of data collection, we measured the area of floral display of flowers presenting pollen of each plant species to standardize the foraging data for the potential influence of relative floral patch size, difference in number of flowers (single vs composite flowers), and amount of pollen per plant species on foraging behavior. Therefore,

“community visitation rate” is presented as a single metric: # visits/min/cm2 floral area/100 bees.

Two species that were used in the study, Monarda fistulosa and Pycnanthemum tenuifolium, were predominately nectar rewarding, were almost never visited for pollen collection by the bumble bees (40), and did not produce enough pollen for nutritional analysis and were therefore excluded.

For full detailed discussion of the methodology, please see Vaudo et al. (2014) (40).

71

Because each nutritional component of pollen is not independent from one another, we conducted a multiple regression analysis of nutrient concentration (protein, lipid, carbohydrate, the interaction of protein and carbohydrate and interaction of protein and lipid) and visitation rate. We followed with regression analysis to determine if the pollen protein:carbohydrate (P:C) ratio influences visitation rate (considered an essential nutritional ratio in arthropod herbivore foraging behavior and nutrient regulation [6])). The interaction of protein and lipid was significant, therefore we conducted nonlinear regression of protein:lipid (P:L) ratio of pollen and average visitation rates to each plant species. Additionally, we performed the analysis with log-transformed

P:L ratios, which did not influence the results, and therefore we reported actual P:L values.

Isolated pollen feeding preference assay

We confined B. impatiens workers to cages to assess their preferences for pollen collected from different plant species in the absence of other floral cues. We purchased four research colonies of

B. impatiens from Koppert Biological Systems (Howell, MI, USA). Each colony contained approximately 100 workers and the natal queen. Three B. impatiens workers from the same mother colony were placed in 6.6 x 8.3 x 9.5 cm Plexiglas cages. Pollen was presented to bees in small plastic thimbles. 0.02g of pollen was placed in each thimble and four drops of water were added to prevent the pollen from being spilled by the bees. The bees were kept in a dark room at ~30ºC and ~30% humidity.

In each trial, we utilized four pollen species. We presented each cage of three bees with a choice between two host-plant pollen species and monitored 6 cages/trial so that there were all pairwise

72

combinations of the four pollen species in each trial (Figure 3-2a). Because we had limited amounts of fresh collected pollen from six plant species, we were only able to test 4 pollen species in each trial for a total of 5 trials. In each trial, we assigned each of the four pollen species a nutritional rank based on P:L ratios, with the highest ratio considered "1" and the lowest "4" (nutritional rank was based on the host-plant foraging preferences assessed earlier, where the highest P:L ratio was most attractive). We used this nutritional rank for our statistical analyses. Thus, we tested if the bees could differentiate and choose between pollen types based on nutritional value, not simply between species. The nutritional ranks and plant species used were 1: S. hebecarpa (trial 1-5); 2:

T. ohiensis (trial 1-5); 3: E. purpureum (trial 1-2), E. purpurea (trial 3-5); 4: E. perfoliatum (trial

1-2), E. purpureum (trial 3-4), S. novae-angliae (trial 5).

We used two behavioral metrics to assess the bees’ pollen preferences. We continuously monitored the bees for three hours and counted the number of times the bees fed from each of the pollen species with their proboscis or mandibles, named “feeding events”. Secondly, we measured the amount of pollen consumed of each pollen species, named “pollen consumed,” by determining the difference in starting (0.02g) and end weight (weight after drying) of each pollen sample. To determine if bumble bees consistently chose the higher P:L pollen of each paired choice independent of plant species, we assigned each cage/replicate of the study (N = 30 cages) a category of “win,” “tie” or “loss” based on if the higher P:L pollen received more feeding events or was more consumed. We used a Chi-square test to analyze if the frequency that the higher P:L pollen “won” (i.e. received more feeding events or was more consumed) was greater than random choice (Figure 3-3). We also conducted independent t-tests to compare average feeding events and pollen consumed for each paired choice between nutritional ranks (i.e. each cage, Figure 3-2b,

73

Figure 3-4a). Finally, to determine relative preferences between nutritional ranks for both feeding events and pollen consumed, we used ANOVA with pollen nutritional rank 1-4 as the independent variable and trial as blocking variable. Post hoc pairwise analyses were used to determine differences between individual pollen nutritional ranks.

Modified pollen feeding preference assay

In this experiment, we manipulated homogenized honey bee collected pollen (HB; multifloral pollen obtained from Brushy Mountain Bee Farm, Moravian Falls, NC) to a range of different protein and lipid concentrations and P:L ratios. Because honey bees collect pollen from multiple floral sources, and it was trapped as corbiculate pollen balls, we first lightly ground ~20g of the pollen with mortar and pestle to break apart the balls and sifted it through a strainer, then stirred it to create a homogenous mix. We repeatedly tested the protein and lipid concentration of the pollen mix (see “Pollen nutritional analysis” above and Supplemental Information) averaging a P:L ratio of 1.6:1 (Figure 3-5, Diet: HB).

We then added purified casein from bovine milk (Sigma-Aldrich, St. Louis, MO) as a protein source to create diets of 5:1, 10:1, and 25:1 P:L (Figure 3-5, Diets: 2-4), which are above the range of P:L ratios of fresh pollen we observed in the previous experiments (Table 1). These ratios were used to test the findings that bumble bees prefer higher P:L ratio pollen. We then added the same protein amounts to three more diets, also adding canola oil (which contains bees’ essential omega-

3 and omega-6 fatty acids) to maintain the same 1.6 P:L ratio as the honey bee pollen (Figure 3-5,

Diets: 5-7). These diets were used to test the hypothesis that bumble bees prefer higher total nutrient concentrations.

74

We then followed the experimental protocol as “Isolated pollen feeding assay”, providing caged bees all 21 paired choices of the seven different diets, recording feeding events and pollen consumed over 4 trials (See Figure 3-2a and Figure 3-6 for design). We provided 0.03g of each diet plus four drops of water in the feeding thimbles and repeated the assay four times with four different B. impatiens colonies. We analyzed the response variables “feeding events” and “pollen consumed” between all diets with ANCOVA using colony/trial and the mean weight of the three bees in each cage as a covariate (to control for inter-cage variation; larger bees tend to eat more) and post hoc pairwise analyses to determine differences between individual diets. Because nutrient concentration appeared to significantly affect bumble bee consumption of each diet, we used regression analyses of protein and lipid concentration to determine the influence of absolute nutrient concentration on pollen consumed. All data were analyzed with JMP Pro 12.1.0 software

(SAS Institute 2015).

Acknowledgements

We would like to thank the Tooker and Grozinger labs for their helpful discussions and critical insight to the preparation of this manuscript, and Heike Betz, Bekki Waskovich, Edwin Hochstedt,

Victoria Bolden, and Liam Farrell for their assistance in experiment preparation and data collection. This work was supported by NAPPC Bee Health Improvement Project Grant to A.D.V., a USDA AFRI NIFA grant to C.M.G. (2009-05207), a USDA AFRI NIFA Predoctoral

Fellowships Grant (10359159) to A.D.V., and generous funding from an anonymous donation to the Penn State Center for Pollinator Research.

75

perfoliatum Eupatorium purpureum Eutrochium novae Symphyotrichum purpurea Echinacea virginicum Veronicastrum ohiensis Tradescantia hebecarpa Senna Species o in listed are species Plant provided. are feedingassay". Nutrientconcentrations are reportedas collection“Hostdatabloomduringforperiods freshlyfromhanddehisced flowers.pollen collectedNote feeding preferencewasby assay.”pollen All 3 Table -

angliae - 1. Host 1.

-

plant species and associated pollen nutritionalassociatedpollenvalue and speciesplant

Asteraceae Asteraceae Asteraceae Asteraceae Scrophulariaceae Commelinaceae Fabaceae Family

rder of highest to lowest P:L values (nutritional rank). Boneset Joe New Purple Culver's Spiderwort American Common - - plantforag Pye

England

coneflower

weed root

µ senna Name

g nutrient/mg pollen.Protein:carbohydrate (P:C)and protein:lipid (P:L) ratios

preference pollen "Isolated the for pollen the used we but preferences" ing aster

Carbohydrate s useds “Hostin 112.66 101.76 118.47 87.20 91.35 55.32 70.43 -

plant pollen foraging preferences” and “Isolated andpreferences”foraging pollen plant Protein 146.48 171.43 186.72 358.25 237.28 78.00 91.04

108.01 158.76 103 that that Lipid 84.03 95.03 83.71 51.72 .74 S. novaeS.

- 0.89 1.30 1.00 1.68 3.38 5.09 2.00 P:C angliae

did 0.72 0.92 1.08 1.80 2.23 3.45 4.59 P:L

not not

76

Figure 3-1. The relationship between Bombus impatiens pollen foraging rates and pollen nutritional quality (Host-plant pollen foraging preferences). "Community visitation rates" are the average number of pollen foraging visits/min/cm2 floral area/100 bees across the season, see methods and Vaudo et al (2014) for more information. Legends and symbols represent the plant species used in each analysis. Data are mean ± SEM. Pollen foraging rates are exponentially related to protein:lipid ratio (P:L) of pollen (Y = 0.0025e1.03*X, R2 = 0.96). Numbers next to each symbol represent the order during the day in which each plant species was most frequently visited, with the exception of S. hebecarpa which experienced consistently higher visitation rates throughout the day.

77

Figure 3-2. Bombus impatiens pollen feeding preferences on isolated pollen is associated with pollen P:L ratios (Isolated pollen feeding preference assay). (A) Experimental design for "Isolated pollen feeding preference assay”. (B) Results of pairwise choice tests according to nutritional rank. (C) Average feeding events independent of paired comparison across trials. In each trial of the experiment, separate cages of three bees were presented one of the six possible pairs of four pollen types. Results represent the results of five trials. Pollen nutritional rank represent pollen species ranked by highest to lowest, “1-4,” protein:lipid ratio (P:L). Data are represented by mean ± SEM. Asterisks in (B) indicate significant differences within each pair (P < 0.05). Bars within (C) labeled with different letters are statistically different (P < 0.05). Arrow under (C) x-axis indicate increasing P:L ratio. The nutritional ranks and plant species used were 1: S. hebecarpa (trial 1-5); 2: T. ohiensis (trial 1-5); 3: E. purpureum (trial 1-2), E. purpurea (trial 3-5); 4: E. perfoliatum (trial 1-2), E. purpureum (trial 3-4), S. novae-angliae (trial 5).

78

Figure 3-3. In “Isolated pollen feeding assay,” the frequency that the pollen with the higher P:L ratio was preferentially selected by the bees relative to the pollen with the lower P:L pollen. For each cage (N = 30 paired comparisons), we assessed whether the pollen with the higher P:L ratio was preferred or "won" relative to the pollen with the lower P:L ratio (for feeding events or amounts consumed), if the two pollens were equally preferred ("tied"), or if the pollen with the higher P:L ratio was less preferred than the pollen with the lower ratio ("lost"). The pollen of higher P:L ratio of each paired choice - independent of plant species - was more frequently 2 preferred for both number of feeding events, or attractiveness (73%, � (2, N = 30) = 20.03, P < 0.01) 2 and mg pollen consumed (63%, � (2, N = 30) = 13.6, P = 0.011).

79

Figure 3-4. Bombus impatiens pollen consumption of species based on “nutritional rank” for “Isolated pollen feeding preference assay”. (A) Results of pairwise choice tests according to nutritional rank. (B) Average pollen consumed independent of paired comparison across trials. In each trial of the experiment, separate cages of three bees were presented one of the six possible pairs of four pollen types. Pollen nutritional rank represents pollen species ranked by highest to lowest, “1-4,” protein:lipid ratio (P:L). Data are represented by mean ± SEM. Asterisks in (A) indicate significant differences within each pair (P < 0.05). Bars within (B) labeled with different letters are statistically different (P < 0.05). Arrow under (B) x-axis indicate increasing P:L ratio. The nutritional ranks and plant species used were 1: S. hebecarpa (trial 1-5); 2: T. ohiensis (trial 1-5); 3: E. purpureum (trial 1-2), E. purpurea (trial 3-5); 4: E. perfoliatum (trial 1-2), E. purpureum (trial 3-4), S. novae-angliae (trial 5).

80

Figure 3-5. Bombus impatiens pollen feeding events (A) and consumption (B) between nutritionally modified pollen diets (Modified pollen feeding preference assay). All diets were modified from honey bee collected pollen (HB, 1:6 P:L ratio). Diets were either modified by adding protein, therefore increasing P:L ratio (Diets 2, 3, and 4), or adding protein and lipid, therefore increasing nutrient concentrations but maintaining 1:6 P:L ratio (Diets 5, 6, and 7). Legend and bar pattern represent specific diet, nutrient concentration, and protein:lipid (P:L) ratios. All diets were modified honey bee collected pollen (HB) with casein (protein) and canola oil (lipid) to alter concentrations. X-axes represent the protein:lipid (P:L) ratio of diets. HB represents unmodified honey bee collected pollen (1.6 P:L ratio). Arrows under x-axis indicate increasing protein (P) and lipid (L) concentration of modified pollen diets. Data are represented by mean ± SEM. Bars within graphs labeled with different letters are statistically different (P < 0.05). Note that experimental design is that of “Isolated pollen feeding preference assay” (Figure 3-2a) where in each trial of the experiment, cages of three bees were presented one of the 21 pairs of seven pollen diets.

81

Figure 3-6. Pairwise comparisons of feeding events (A) and consumption (B) of all pollen diet combinations in “Modified pollen feeding preference assay”. Highlighted cells indicate that B. impatiens workers ate the diet in the heading significantly more (blue), similarly (grey), or less (orange) than the diet type written in the cell (P < 0.05). Cells indicate header diet versus (“v”) paired diet.

82

Figure 3-7. The relationship between protein (A) and lipid (B) concentration and amount of pollen consumed by B. impatiens (Modified pollen feeding preference assay). Data are mean and 95% CI. Legend and symbols represent specific diet. Diets were modified from honey bee collected pollen (HB). See Figure 3 for specific nutrient concentrations. There was a negative linear relationship between protein concentration and pollen consumption (Y = 15.129607 - 30.761651*X, R2 = 0.44). However, the relationship of lipid concentration to consumption was biexponential, showing increased consumption at low concentrations, but then decreasing at higher concentrations (Y = 1558*e-26.8*X – 1600*e-28.0*X, R2 = 0.44).

83

References

1. Vaudo AD, Tooker JF, Grozinger CM, Patch HM (2015) Bee nutrition and floral resource restoration. Curr Opin Insect Sci 10:133–141.

2. Willmer P (2011) Pollination and Floral Ecology (Princeton University Press).

3. Brunet J, Thairu MW, Henss JM, Link RI, Kluever JA (2015) The effects of flower, floral display, and reward sizes on bumblebee foraging behavior when pollen is the reward and plants are dichogamous. Int J of Plant Sci 176(9):811–819.

4. Ollerton J, Erenler H, Edwards M, Crockett R (2014) Pollinator declines. Extinctions of aculeate pollinators in Britain and the role of large-scale agricultural changes. Science 346(6215):1360–1362.

5. Goulson D, Nicholls E, Botías C, Rotheray EL (2015) Bee declines driven by combined stress from parasites, pesticides, and lack of flowers. Science 347(6229):1255957.

6. Behmer ST (2009) Insect herbivore nutrient regulation. Annu Rev Entomol 54:165–187.

7. Behmer ST, Joern A (2008) Coexisting generalist herbivores occupy unique nutritional feeding niches. Proc Natl Acad Sci USA 105(6):1977–1982.

8. Dussutour A, Simpson SJ (2009) Communal nutrition in ants. Curr Biol 19(9):740–744.

9. Brodschneider R, Crailsheim K (2010) Nutrition and health in honey bees. Apidologie 41(3):278–294.

10. Roulston TH, Cane JH (2000) Pollen nutritional content and digestibility for animals. Plant Syst Evol 222(1):187–209.

11. Haydak MH (1970) Honey bee nutrition. Annu Rev Entomol 15(1):143–156.

12. Génissel A, Aupinel P, Bressac C, Tasei JN, Chevrier C (2002) Influence of pollen origin on performance of Bombus terrestris micro-colonies. Entomol Exper Applic 104(2-3):329– 336.

13. Human H, Nicolson SW, Strauss K, Pirk CWW, Dietemann V (2007) Influence of pollen quality on ovarian development in honeybee workers (Apis mellifera scutellata). J Insect Physiol 53(7):649–655.

14. Tasei J-N, Aupinel P (2008) Nutritive value of 15 single pollens and pollen mixes tested on larvae produced by bumblebee workers (Bombus terrestris, Hymenoptera: Apidae). Apidologie 39(4):397–409.

15. Li C, Xu B, Wang Y, Feng Q, Yang W (2012) Effects of dietary crude protein levels on development, antioxidant status, and total midgut protease activity of honey bee (Apis mellifera ligustica). Apidologie 43(5):576–586.

84

16. Alaux C, Ducloz F, Crauser D, Le Conte Y (2010) Diet effects on honeybee immunocompetence. Biol Letters 6(4):562–565.

17. LeBlanc BW, Davis OK, Boue S, DeLucca A, Deeby T (2009) Antioxidant activity of Sonoran Desert bee pollen. Food Chem 115(4):1299–1305.

18. Roulston TH, Cane JH, Buchmann SL (2000) What governs protein content of pollen: pollinator preferences, pollen-pistil interactions, or phylogeny? Ecological Monographs 70(4):617–643.

19. Di Pasquale G, et al. (2013) Influence of pollen nutrition on honey bee health: do pollen quality and diversity matter? PLoS ONE 8(8):e72016.

20. Cardoza YJ, Harris GK, Grozinger CM (2012) Effects of soil quality enhancement on pollinator-plant interactions. Psyche 2012(4):1–8.

21. Vanderplanck M, et al. (2014) How does pollen chemistry impact development and feeding behaviour of polylectic bees? PLoS ONE 9(1):e86209.

22. Moerman R, et al. (2015) Growth rate of bumblebee larvae is related to pollen amino acids. J Econ Entomol 109(1):25-30.

23. Canavoso LE, Jouni ZE, Karnas KJ, Pennington JE, Wells MA (2001) Fat metabolism in insects. Ann Rev Nutr 21:23–46.

24. Arien Y, Dag A, Zarchin S, Masci T, Shafir S (2015) Omega-3 deficiency impairs honey bee learning. Proc Natl Acad Sci USA 112(51):15761–15766.

25. Leonhardt SD, Blüthgen N (2011) The same, but different: pollen foraging in honeybee and bumblebee colonies. Apidologie 43(4):449–464.

26. Ruedenauer FA, Spaethe J, Leonhardt SD (2015) How to know which food is good for you: bumblebees use taste to discriminate between different concentrations of food differing in nutrient content. J Exp Biol 218(Pt 14):2233–2240.

27. Kitaoka TK, Nieh JC (2008) Bumble bee pollen foraging regulation: role of pollen quality, storage levels, and odor. Behav Ecol Sociobiol 63(4):501–510.

28. Konzmann S, Lunau K (2014) Divergent rules for pollen and nectar foraging bumblebees – a laboratory study with artificial flowers offering diluted nectar substitute and pollen surrogate. PLoS ONE 9(3):e91900.

29. van der Moezel PG, Delfs JC, Pate JS, Loneragan WA, Bell DT (1987) Pollen selection by honeybees in shrublands of the Northern Sandplains of Western Australia. J of Api Res 26(4):224–232.

30. Pernal S, Currie R (2001) The influence of pollen quality on foraging behavior in honeybees (Apis mellifera L.). Behav Ecol Sociobiol 51(1):53–68.

85

31. Hanley ME, Franco M, Pichon S, Darvill B, Goulson D (2008) Breeding system, pollinator choice and variation in pollen quality in British herbaceous plants. Func Ecol 22(4):592– 598.

32. Somme L, et al. (2014) Pollen and nectar quality drive the major and minor floral choices of bumble bees. Apidologie 46(1):92–106.

33. Dobson HEM (1987) Role of flower and pollen aromas in host-plant recognition by solitary bees. Oecologia 72(4):618–623.

34. Dobson HEM, Bergström G (2000) The ecology and evolution of pollen odors. Plant Syst Evol 222(1-4):63–87.

35. Pacini E, Hesse M (2005) Pollenkitt–its composition, forms and functions. Flora 200(5):399–415.

36. Piskorski R, Kroder S, Dorn S (2011) Can pollen headspace volatiles and pollenkitt lipids serve as reliable chemical cues for bee pollinators? Chem Biodivers 8(4):577–586.

37. Mayntz D, Raubenheimer D, Salomon M, Toft S, Simpson SJ (2005) Nutrient-specific foraging in invertebrate predators. Science 307(5706):111–113.

38. Raubenheimer D, Mayntz D, Simpson SJ, Tøft S (2007) Nutrient-specific compensation following diapause in a predator: implications for intraguild predation. Ecology 88(10):2598–2608.

39. Jensen K, Mayntz D, Toft S, Raubenheimer D, Simpson SJ (2011) Nutrient regulation in a predator, the wolf spider Pardosa prativaga. Anim Behav 81(5):993–999.

40. Vaudo AD, Patch HM, Mortensen DA, Grozinger CM, Tooker JF (2014) Bumble bees exhibit daily behavioral patterns in pollen foraging. Arthropod-Plant Inte 8(4):273–283.

41. Altaye SZ, Pirk CWW, Crewe RM, Nicolson SW (2010) Convergence of carbohydrate- biased intake targets in caged worker honeybees fed different protein sources. J Exp Biol 213(19):3311–3318.

42. Paoli PP, et al. (2014) Nutritional balance of essential amino acids and carbohydrates of the adult worker honeybee depends on age. Amino Acids 46(6):1449–1458.

43. Stabler D, Paoli PP, Nicolson SW, Wright GA (2015) Nutrient balancing of the adult worker bumblebee (Bombus terrestris) depends on the dietary source of essential amino acids. J Exp Biol 218:793–802.

44. Stockhoff BA (1993) Ontogenetic change in dietary selection for protein and lipid by gypsy moth larvae. J Insect Physiol 39(8):677–686.

45. Cook SM, Awmack CS, Murray DA, Williams IH (2003) Are honey bees' foraging preferences affected by pollen amino acid composition? Ecol Entomol 28(5):622–627.

86

46. Weiner CN, Hilpert A, Werner M, Linsenmair KE, Blüthgen N (2010) Pollen amino acids and flower specialisation in solitary bees. Apidologie 41(4):476–487.

47. Hendriksma HP, Shafir S (2016) Honey bee foragers balance colony nutritional deficiencies. Behav Ecol Sociobiol 70(4):509–517.

48. Linander N, Hempel de Ibarra N, Laska M (2012) Olfactory detectability of L-amino acids in the European honeybee (Apis mellifera). Chem Senses 37(7):631–638.

49. Cook SM, et al. (2005) Could learning of pollen odours by honey bees (Apis mellifera) play a role in their foraging behaviour? Physiol Entomol 30(2):164–174.

50. Eckhardt M, Haider M, Dorn S, Müller A (2014) Pollen mixing in pollen generalist solitary bees: a possible strategy to complement or mitigate unfavourable pollen properties? J Anim Ecol 83(3):588–597.

51. Müller A, Kuhlmann M (2008) Pollen hosts of western palaearctic bees of the genus Colletes (Hymenoptera: Colletidae): the Asteraceae paradox. Biol J Linn Soc 95(4):719– 733.

52. Ruedenauer FA, Spaethe J, Leonhardt SD (2016) Hungry for quality—individual bumblebees forage flexibly to collect high-quality pollen. Behav Ecol Sociobiol:1–9.

53. Robertson HM, Wanner KW (2006) The chemoreceptor superfamily in the honey bee, Apis mellifera: Expansion of the odorant, but not gustatory, receptor family. Genome Research 16(11):1395–1403.

54. Simpson SJ, Simpson CL (1992) Mechanisms controlling modulation by haemolymph amino acids of gustatory responsiveness in the locust. J Exp Biol 168(1):269–287.

55. Dornhaus A, Chittka L (2005) Bumble bees (Bombus terrestris) store both food and information in honeypots. Behav Ecol 16(3):661–666.

56. Abisgold JD, Simpson SJ (1988) The effect of dietary protein levels and haemolymph composition on the sensitivity of the maxillary palp chemoreceptors of locusts. J Exp Biol 135(1):215–229.

57. Lihoreau M, et al. (2014) Unravelling the mechanisms of trapline foraging in bees. Commun Integr Biol 6(1):e22701.

58. Ohashi K, Thomson JD (2009) Trapline foraging by pollinators: its ontogeny, economics and possible consequences for plants. Ann Bot 103(9):1365–1378.

59. Raine NE, Chittka L (2012) No trade-off between learning speed and associative flexibility in bumblebees: a reversal learning test with multiple colonies. PLoS ONE 7(9):e45096.

60. Simpson SJ, Raubenheimer D (1993) A multi-level analysis of feeding behaviour: the geometry of nutritional decisions. Philos T R Soc B 342(1302):381–402.

87

61. Raubenheimer D, Simpson SJ (1999) Integrating nutrition: a geometrical approach. Entomol Exper Applic 91(1):67–82.

62. Velthuis HHW, van Doorn A (2006) A century of advances in bumblebee domestication and the economic and environmental aspects of its commercialization for pollination. Apidologie 37(4):421–451.

63. Robertson C (1928) Flowers and insects (The Science Press Printing Company, Lancaster, PA).

64. Szabo TI, Pengelly DH (1973) The over-wintering and emergence of Bombus (Pyrobombus) impatiens (Cresson) (Hymenoptera: Apidae) in southern Ontario. Insect Soc 20(2):125–132.

65. Cnaani J, Schmid-Hempel R, Schmidt JO (2002) Colony development, larval development and worker reproduction in Bombus impatiens Cresson. Insect Soc 49(2):164–170.

66. Rasmont P, Coppee A, Michez D, De Meulemeester T (2008) An overview of the Bombus terrestris (L. 1758) subspecies (Hymenoptera: Apidae). Ann Soc Entomol Fr 44(2):243– 250.

67. Lozier JD, Strange JP, Stewart IJ, Cameron SA (2011) Patterns of range-wide genetic variation in six North American bumble bee (Apidae: Bombus) species. Mol Ecol 20(23):4870–4888.

68. Amsalem E, Grozinger CM, Padilla M, Hefetz A (2015) Bumble bee sociobiology: the physiological and genomic bases of bumble bee social behaviour. Adv Insect Physiol doi: 10.1016/bs.aiip.2015.01.001.

69. Goulson D, Lye GC, Darvill B (2008) Decline and conservation of bumble bees. Annu Rev Entomol 53(1):191–208.

70. Cameron SA, et al. (2011) Patterns of widespread decline in North American bumble bees. Proc Natl Acad Sci USA 108(2):662–667.

71. Potts S (2015) STATUS and TRENDS of EUROPEAN POLLINATORS. key findings of the STEP project. 1–74.

72. Carvell C, et al. (2006) Declines in forage availability for bumblebees at a national scale. Biol Conserv 132(4):481–489.

73. Potts SG, et al. (2010) Global pollinator declines: trends, impacts and drivers. Trends Ecol Evol 25(6):345–353.

74. Van Handel E, Day JF (1988) Assay of lipids, glycogen and sugars in individual mosquitoes: correlations with wing length in field-collected Aedes vexans. J Am Mosq Control Assoc 4(4):549–550.

88

75. Raine NE, Chittka L (2006) Pollen foraging: learning a complex motor skill by bumblebees (Bombus terrestris). Naturwissenschaften 94(6):459–464.

89

Chapter 4. Bumble bees regulate their intake of the essential protein and lipid pollen macronutrients

Anthony D. Vaudo*a, Daniel Stablerb, Harland M. Patcha, John F. Tookera, Christina M.

Grozingera, Geraldine A. Wrightb

a Department of Entomology, Center for Pollinator Research, The Pennsylvania State University,

501 ASI Building, University Park, PA 16802, USA

b Center for Behavior and Evolution, Institute of Neuroscience, Henry Wellcome Building for

Neuroecology, Newcastle University, Framlington Place, Newcastle Upon Tyne NE2 4HH,

England

Abstract

Bee population declines are linked to reduction of nutritional resources due to land-use intensification, yet we know little about the specific nutritional needs of many bee species. Pollen provides bees their primary source of protein and lipids, but nutritional quality varies widely among host-plant species. Therefore, bees may be adapted to assess resource quality and adjust their foraging behavior to balance nutrition from multiple food sources. We tested the ability of two bumble bee species, Bombus terrestris and B. impatiens, to regulate protein and lipid intake.

We restricted B. terrestris adults to single synthetic diets varying in protein:lipid ratios (P:L). The bees overate protein on low fat diets and overate lipid on high fat diets to reach their targets of lipid and protein respectively. The bees survived best on a 10:1 P:L diet; the risk of dying increased

90

as a function of dietary lipid when bees ate diets with lipid contents greater than 5:1 P:L.

Hypothesizing that P:L intake target of adult worker bumble bees was between 25:1-5:1, we presented workers from both species unbalanced but complementary paired diets to determine if they self-select their diet to reach a specific intake target. Bees consumed similar amounts of proteins and lipids in each treatment and averaged a 14:1 P:L for B. terrestris and 12:1 P:L for B. impatiens. These results demonstrate that adult worker bumble bees likely select foods that provide them with a specific ratio of P:L. These P:L intake targets could affect pollen foraging in the field and help explain patterns of host-plant species choice by bumble bees.

Keywords: Bumble Bee, Geometric Framework, Lipid, Nutritional Ecology, Pollen, Protein

Introduction

Bee population declines are linked with many interacting factors associated with anthropogenic land-use intensification (1, 2), including the reduction of host-plant abundance and diversity, which may lead to nutritional stress for some bee species (3-5). Differences in resource quality can have direct effects on bee development, reproduction, immunocompetence, resilience to stress, and survival (6). Therefore, to address the problem of nutritional deprivation in the landscape, it is crucial to develop a comprehensive understanding of the nutritional requirements of bees.

Bees obtain their macronutrients (carbohydrates, proteins, and lipids) from floral nectar and pollen.

Bees primarily obtain carbohydrates from nectar to fuel energetically costly foraging efforts, and adults cannot survive without a continuous carbohydrate source (7). Bees obtain proteins and lipids from pollen. Differences in protein in bee diets can influence adult reproduction, physiology, and

91

immunity, and larval development (8-14). For bees, lipids play important roles in production of cuticular hydrocarbons and wax, behavioral maturation in adults (through the reduction in lipid stores), diapause, learning, and development of glands that produce brood food (15-17). Essential sterols obtained exclusively from pollen are precursors for molting hormone, which is essential for larval development (18-20). Moreover, the lipid-dominant pollenkitt on the exterior of pollen is an important discriminative stimulus and phagostimulus of pollen for bees (21, 22).

Although bees can obtain protein and lipids from most pollen sources, pollen protein (including essential amino acids) and lipid (including essential fatty acids and sterols) concentrations vary considerably among plant species (pollen contains ~2-60% protein and ~2-20% lipid; (19).

Inequality of nutrients among plant species implies that bees may selectively forage for pollen to meet their nutritional demands. Generalist bee species, such as Bombus terrestris (Hymenoptera:

Apidae) in Europe, North Africa, and the Middle East, and B. impatiens in North America, forage on a variety of different plant species during their lives. A handful of studies have suggested that bumble bees preferentially forage on flowers that have high sugar concentrations in nectar (23,

24), and high protein (12, 25, 26) or amino acid and sterol content in pollen (24). A recent study demonstrated that B. impatiens – both when foraging for colonies with brood or isolated from brood – preferentially forage for pollen with high protein:lipid ratios and their consumption of pollen diets depended on protein and lipid concentrations (27). This indicates that bees are sensitive to both protein and lipids in diet and are likely to exhibit nutrient regulation that affects their feeding behavior.

92

Although foraging bumble bees collect pollen mainly to feed developing larvae, adult workers eat pollen as well (7, 19), when they assess nutritional stores in pollen pots (28), while they feed pollen to larvae (29, 30), or when they eat pollen to develop their own ovaries for male-egg laying (10,

31). Note that in three-worker queenless microcolonies, workers ate between 0.4-0.9g of pollen in the five days prior to egg laying, which would average ~25-60mg/pollen/day by worker egg-layers

(10, 32).

Many studies have demonstrated that insects regulate their consumption of food around optimal proportions of macronutrients in ways that reflect their age, somatic needs, and reproductive status

(33-35). The geometric framework (GF) for nutrition is a method for examining the mechanisms and constraints that govern how animals regulate feeding to achieve specific macronutrient optima, or “intake targets”. It employs an approach wherein individuals self-select diets or alter food consumption when confined to diets comprising specific ratios of macronutrients (33, 36). The GF has been successfully used to characterize nutrient balancing for protein and carbohydrate in worker honey bees (37-39) and bumble bees (40). Workers, especially foragers, have a high demand for carbohydrates, as reflected in their measured intake targets (for bumble bees, this is

~1:150 protein:carbohydrate or P:C ratio). Moreover, their tolerance of dietary protein (or essential amino acids) is relatively low, as they have reduced survival when forced to ingest surplus protein

(37-40). This has also been observed in ants (41) and for the over-ingestion of specific amino acids in fruit flies (42).

None of the previous studies using the GF have tested whether bees or other social insects regulate their dietary intake of fats. The few studies that have investigated protein and fat regulation in

93

insect herbivores have been limited to lepidopteran larvae, but were not clear assessments using the GF of simultaneous regulation of both nutrients (43, 44). In contrast, arthropod predators clearly regulate both protein and fat simultaneously. For example, the ground beetle Agonum dorsale, adjusts its consumption of complementary foods to meet an intake target of proteins and fat (45). Similarly, the wolf spider Pardosa prativaga regulated its diet by eating flies that complemented a previous diet higher in protein or fat (45), and overate protein on lipid poor diets to reach an intake target for lipid (46).

Here, we use the GF methodology to test and measure regulation of protein and lipid intake in bumble bee foragers of two species, B. terrestris and B. impatiens (both important crop pollinators and commercially available in their respective geographic range, (31, 47)). In our first experiment, we restricted B. terrestris individuals to single synthetic diets differing in P:L ratios that spanned the realistic and extreme possibilities found in pollen, and measured their food consumption and survival. Then, using the results of the first experiments to select appropriate diets, we presented

B. terrestris and B. impatiens individuals two diets differing in their P:L ratios to determine if the two species indeed regulate protein and lipids to a specific intake target. We expected that the species would regulate their P:L intake to a target at which they survived best. We also expected that the bumble bees would defend a carbohydrate target, given the importance of carbohydrates for bees. Our results characterize the specific macronutrient requirements of these two species and provide insights into the ability of bumble bees to regulate lipids in their diet, suggesting nutritional quality may drive pollen foraging preferences.

94

Methods

General bee rearing conditions

We purchased mature research colonies of Bombus terrestris (“Single P:L diet assay” and “Paired

P:L diets assay”) and B. impatiens (“Paired P:L diets assay”) from Koppert Biological Systems

(Havervill, Suffolk, UK for B. terrestris; Howell, MI, USA for B. impatiens). Each colony contained approximately 100 workers and the natal queen. During the course of the study, we stored colonies at ambient temperatures and provided them sugar water ad libitum. For each assay, we collected foragers as they exited their colonies and placed individual bees in their own 11 × 11

× 10-cm plastic cages kept in a 24-hr dark incubator at 28°C and 40% humidity. We provided all diets to bees in 2-mL microcentrifuge tubes with four holes drilled in the tube from which the bees could feed. The tubes were suspended halfway up and at opposite sides of each cage such that the bees could perch on the tube and feed through the holes. We first performed the “Single P:L diet assay” with B. terrestris in the UK. Based on the results of this assay, we designed the “Paired P:L diets assay” to be sensitive for both bumble bee species as we expected their intake targets are not radically different. We conducted the “Paired P:L diets assay” for B. terrestris in the UK, and B. impatiens in the USA.

Single P:L diet assay

Individual forager B. terrestris bees (15 bees/treatment, 4 colonies) were given access to food tubes containing 0.5 M sucrose solution or 0.5 M sucrose solution containing a specific protein:lipid ratio (P:L). We tested eight different dietary ratios of P:L (Protein-only, 50:1, 25:1, 10:1, 5:1, 1:1,

1:5, and 1:10; Table 4-1). The sucrose-only food source was necessary to allow bees to reach their high carbohydrate demand and needed to be separate for bees to freely consume it without

95

consuming proteins and lipids. This also provided the simulation of what bees actually experience by providing a carbohydrate-only source or “nectar” and a fixed protein/lipid/sugar source or

“pollen.” Protein was held constant while we adjusted the lipid concentration. We chose these particular P:L diets to include possible ranges of P:L ratios in pollen (19) as well as values outside of the reported range of P:L in pollen. Nutrient sources were sucrose (Sigma-Aldrich, St. Louis,

MO, USA) for carbohydrates, casein sodium salt from bovine milk (Sigma-Aldrich) for protein, and 100% soy lecithin (Optima Health & Nutrition, Bradford, UK) for lipids (>91% fat), which contains essential fatty acids (32% ⍵-6/linoleic acid, 4% ⍵-3/alpha-linolenic acid). Soy lecithin was chosen as the lipid source because it is an emulsifier and can be used for liquid diets. To prepare the diets, we mixed the lecithin into solution using a stir plate for ~1-2 hours under low heat. Liquid diets were used because they are easy for the bees to ingest and allow accurate measurement of consumption.

Experiments lasted seven days, and we replaced each food tube daily. We weighed food tubes each day prior to placement in the cage and 24 hr later. Cages with three tubes of each diet (replaced daily) with no bees served as controls to measure the daily evaporation rate for each diet. Amounts of solution (g) consumed by bees were adjusted by the daily mean amount of solution that had evaporated from the “control” cages prior to analysis. We calculated the mass of each nutrient

(carbohydrate, protein, lipid) consumed from the total mass consumed from each diet tube each day. We measured the thorax width of each individual bee as a covariate in data analyses to control for the effect of size on diet consumption. We recorded the number of days each bee survived in the assay with a maximum of seven days.

96

Paired P:L diets assay

To test our hypothesis that bumble bee intake targets lie within the 25:1-5:1 P:L range (see

“Results-Single P:L diet assay”), we measured survival and nutrient consumption of B. impatiens and B. terrestris foragers presented with paired P:L diets encompassing this range. As in the

“Single P:L diet assay,” we collected B. impatiens and B. terrestris foragers as they exited their colonies and caged them individually (20 bees/treatment; 2 colonies for each species).

For each treatment, we provided a bee with one of four paired P:L diets and with a sucrose-only food tube. These diet pairings were: 1) 25:1 and 5:1, 2) 50:1 and 5:1, 3) 75:1 and 5:1, and 4) 100:1 and 5:1 P:L (diets prepared as above; Table 4-1). We measured daily consumption of each diet and nutrient (accounting for evaporation rate) and survival of bees over seven days (see “Single P:L diet assay”). Prior to placement in cages, we cold anaesthetized and weighed foragers to use their weight as a covariate in data analyses to control for effects of size on diet consumption (note thorax width and bee weight are correlated (40), and we measured thorax width in the “Single P:L diet assay”).

Statistical analysis

Single P:L diet assay

We conducted survival analyses with Cox-regression proportional hazards, and used the Protein- only treatment as reference or control to determine the effect of adding lipid to the diet on bee survival. To determine whether bumble bees ate randomly among diet sources or if particular treatment diets caused differential feeding behavior, we analyzed differences in daily consumption of diet sources among treatments by 2-way ANOVA and post-hoc Tukey-HSD pairwise

97

comparisons with treatment, diet source (treatment diet or sucrose-only), and the interaction of treatment and diet source as independent variables and thorax width as a covariate. To analyze differences in daily consumption of nutrients among treatments, we used MANCOVA with post- hoc Tukey-HSD pairwise comparisons with nutrient (carbohydrate, protein, or lipid) as the dependent variable and thorax width as a covariate. Finally, for bees that survived on the diets for all seven days, we analyzed differences in cumulative consumption of carbohydrate, protein, and lipid with MANCOVA and post-hoc Tukey-HSD pairwise comparisons with nutrient

(carbohydrate, protein, or lipid) as the dependent variable and thorax width as a covariate. After reviewing the data, it was apparent that there were differences in amounts of nutrients consumed between bees that died and survived in the 1:10 P:L treatment. We compared their cumulative consumption of nutrients on day three, using MANOVA and post-hoc t-tests for each nutrient.

Paired P:L diets assay

Bombus terrestris and Bombus impatiens were analyzed separately. We analyzed differences in survival among treatments with the Kaplan-Meier test (because there was no reference group as above for Cox-regression). To determine daily differences in mass of diets consumed among treatments, we conducted 2-way ANOVA and post-hoc Tukey-HSD pairwise comparisons, using treatment, diet source (5:1, treatment diet, and sucrose-only), and the interaction of treatment and diet source as independent variables with colony and bee weight as covariates. Note that bee weight was used as a measure of size for this assay while thorax width was used in the “Single P:L diet assay”. These are correlated metrics of bee size used as covariates for consumption/bee (40).

Finally, for bees that survived all seven days, we analyzed cumulative nutrient consumption among treatments with MANCOVA with post-hoc Tukey-HSD pairwise comparisons with nutrient

98

(carbohydrate, protein, or lipid) as the dependent variable and colony and bee weight as covariates.

If consumption of each nutrient among treatments was similar, we could conclude that the bumble bees were regulating their nutrients equally. We determined P:C and P:L ratios consumed by bees using the average cumulative consumption of each treatment. All statistical analyses were conducted with JMP Pro v.12 (SAS Institute; SPSS Statistics [IBM] was used for Cox-regression).

Results

Single P:L diet assay

For seven days, we fed B. terrestris foragers with sucrose only and one of the P:L diets. The total quantities of food the bees consumed each day did not differ significantly across treatments (F7,1321

= 1.99, P = 0.053); the only pairwise difference was that foragers in the “protein only” treatment ate more each day than bees on the high fat 1:5 P:L treatment at P < 0.05 (Figure 4-1). Bees differed in the relative amounts of each diet (treatment diet versus sucrose only) consumed (treatment x solution; F7,1321 = 16.0, P < 0.001) (Figure 4-1). Notably, bees consumed much less of the treatment diet than sucrose-only diet in the highest lipid treatments (1:5, 1:10 P:L; Figure 4-1).

The only significant difference in daily consumption of carbohydrates was between protein-only and 1:5 treatments (F8,666 = 5.32, P < 0.001; Table 4-2), but bees across treatments differed significantly in amounts of protein and lipid consumed (MANCOVA: F21,1640 = 13.7, P < 0.001).

Bees on the highest fat diets (1:5 and 1:10 P:L) consumed much less protein than the other treatments (F8,663 = 14.7, P < 0.001; Table 4-2), suggesting that they ceased eating the diet after having reached or exceeded their lipid intake target, and therefore did not reach their protein target.

Finally, bees across treatments differed significantly in amounts of lipids consumed; specifically,

99

bees consumed more lipids as lipid content of the treatment diet increased (F7,573 = 20.4, P < 0.01;

Table 4-2).

For the bees that survived all seven days of the experiment, there were significant differences among treatments in cumulative amount of nutrients consumed (MANCOVA: F21,164 = 5.03, P <

0.001; Figure 4-2). Though there were no differences in cumulative carbohydrates consumed across treatments (F7,59 = 1.13, P = 0.36; Figure 4-2a,c), bees on different diets consumed significantly different amounts of cumulative protein and lipids over seven days. Similar to the daily consumption data, bees on the highest lipid treatments (1:5 and 1:10 P:L) consumed significantly less protein (F7,59 = 3.86, P = 0.002; Figure 4-2a,b).

For cumulative lipids consumed, surviving bees in the 1:10, 1:5, and 1:1 treatments consumed significantly more lipids than bees on the remaining treatments (F7,59 = 10.2, P < 0.001, Figure 4-

2b,c). Furthermore, bumble bee foragers consumed on average ~3.5mg protein on 1:1, 5:1, 10:1 and 25:1 P:L diets, while consuming ~5.1mg protein on the 50:1 P:L diet (F1,59 = 2.86, P < 0.1), suggesting that bees compensated for low lipids by overeating the 50:1 diet to reach an intake target for lipid (Figure 4-2b). These data also indicate that B. terrestris foragers regulated their protein intake eating similar amounts of proteins (~4.0mg) except on the highest lipid diets of 1:5 and 1:10 (~0.6mg).

Bombus terrestris foragers had a greater risk of mortality when they consumed diets high in lipid

(Table 4-3). Specifically, the mortality risk was lowest for the bees fed the 10:1 and 5:1 diets, whereas bees fed diets with proportionally greater quantities of lipids had increased risk of dying

100

over seven days (Table 4-3). Although bees in the high fat treatment (1:5 P:L) appeared to survive well in the first days of the study, their mortality increased sharply over the remainder of the week and ended with the second highest mortality and a nearly equal hazard ratio (Figure 4-1, Figure 4-

3). Interestingly, by day three on the 1:10 P:L diet, surviving bees had eaten significantly less of their treatment diet (protein and lipid) than those bees that died (t14 = 2.29, P < 0.02), but living and dead bees ate equal amounts of carbohydrates (t14 = 0.64, P = 0.27; Figure 4-4). These data suggest that high lipid consumption leads to toxicity and increased mortality.

Bombus terrestris foragers 1) overate lipids to defend their protein intake, 2) had increased mortality as lipid content of diets increased or decreased away from 10:1 P:L), and 3) increased protein consumption on the 50:1 P:L diet to potentially defend a lipid target. Therefore, we hypothesized that the bumble bees’ P:L intake target lies within the 25:1 – 5:1 range. We performed a “Paired P:L diets assay” to identify the actual intake target for P:L of B. terrestris and

B. impatiens.

Paired P:L diets assay

For seven days, we fed Bombus impatiens and B. terrestris workers a single sucrose-only diet, a

5:1 P:L diet, and a complementary treatment P:L diet (25:1, 50:1, 75:1, or 100:1). Each diet pairing of 5:1 P:L and treatment P:L created a protein and lipid nutrient space encompassing the hypothesized P:L intake target. The bees consumed significantly different amounts of total food across treatments (B. impatiens: F3,1446 = 5.65, P < 0.001; B. terrestris: F3,1178 = 4.75, P < 0.003), diet sources (B. impatiens: F2,1446 = 23.7, P < 0.01; B. terrestris: F2,1178 = 30.7, P < 0.001), and the relative amounts of each diet source consumed among treatments (treatment × diet source

101

interaction: B. impatiens: F6,1446 = 3.55, P = 0.0017; B. terrestris: F6,1178 = 3.31, P = 0.003; Figure

4-5). Importantly, daily consumption differed between the treatment diet (25:1, 50:1, 75:1, 100:1) and the 5:1 diet for both B. impatiens and B. terrestris, indicating that these diets were not being consumed randomly (Figure 4-5).

Surviving B. impatiens and B. terrestris foragers, analyzed separately, regulated their carbohydrate, protein, and lipid intake. Consumption of the three macronutrients and total nutrients across treatments was not significantly different within each species (carbohydrate: B. impatiens:

F3,52 = 2.20, P = 0.10; B. terrestris: F3,47 = 1.50, P = 0.23; protein: B. impatiens: F3,52 = 2.63, P =

0.06; B. terrestris: F3,47 = 1.02, P = 0.39; lipid: B. impatiens: F3,52 = 1.78, P = 0.16; B. terrestris:

F3,47 = 0.02, P = 0.99; total nutrients: B. impatiens: MANCOVA: F9,122 = 1.35, P = 0.22; B. terrestris: MANCOVA: F9,110 = 1.07, P = 0.39; Table 4-4, Figure 4-6, Figure 4-7). Therefore, B. impatiens and B. terrestris, foragers regulated their P:L intake to within our hypothesized range, averaging 12:1 P:L for B. impatiens and 14:1 P:L for B. terrestris (Table 4-4, Figure 4-6, Figure

4-7). The P:C intake targets regulated by both species averaged 1:85 P:C for B. impatiens and 1:67

P:C for B. terrestris (Table 4-4, Figure 4-6, Figure 4-7). Both bee species survived equally well on the various diets (B. impatiens: c2 = 3.98, df = 3, P = 0.26; B. terrestris: c2 = 0.39, df = 3, P =

0.94; Figure 4-8).

Discussion

Our experiments revealed that B. terrestris and B. impatiens regulated their protein and lipid intake to an average of 14:1 and 12:1, respectively, with B. terrestris preferring a diet slightly lower in fat than B. impatiens. Also, bees limited to diets high in lipids had increased risk of mortality

102

(Table 4-3, Figure 4-3). Taken together, this study provides the first evidence that pollinators

(specifically Bombus spp. bees) regulate fat intake. Coupled with our previous study that demonstrated that bumble bees foraging preferences were significantly correlated with protein:lipid ratios in pollen (27), these results suggest that pollinators adjust their foraging to achieve specific macronutrient targets.

The protein and lipid regulation of bumble bee adults appears more similar to predaceous arthropods than herbivorous ones. Manduca sexta caterpillars, within a similar design as our

“Paired P:L diets assay,” failed to regulate lipid intake but preferred diets high in fat (44). In contrast, both B. terrestris and B. impatiens workers regulated their intake of fat, and preferred diets with specific P:L ratios. This difference is likely due to the vastly different life histories between lepidopteran larvae, which are typically constrained to specific food sources, and hymenopteran adults, which can forage among many sources. Both predaceous species (i.e., the wolf spider and ground beetle) ate protein excessively on low fat diets, apparently to reach a lipid intake target (~4:1 P:L for wolf spider; or ~2:1 P:L in for ground beetle; (45, 46, 48). In our work,

B. terrestris generally ate more protein on the low-fat diet (50:1 P:L) than the other treatments, including those that provided only protein. This behavior indicates that workers may also overeat protein to reach their lipid intake; indeed, lipid intake did not differ across the groups fed 50:1,

25:1, 10:1 and 5:1 diets. Finally, the web building spider Stegodyphus lineatus, having no control over the nutrient composition of prey captured in its web, selectively extracts dietary protein from prey based on previous feeding history (45). Bee larvae assimilate pollen protein and lipids efficiently (19), but it remains to be tested if the sedentary and dependent bee larvae can differentially assimilate these nutrients to reach their intake targets or if they are completely

103

dependent upon adults to sense and select an appropriate diet for them.

In contrast to A. dorsale, the predatory ground beetle, which stopped eating when it reached its lipid intake target in high fat diets (48), B. terrestris overate lipid in high-fat diets (1:1, 1:5, and

1:10 P:L), potentially to reach their protein target. This overconsumption of lipid to reach a protein target may have led to increased mortality. For example, bees survived when they ate less of the high fat diet 1:10 P:L (Figure 4-4). And although the bees in the 1:5 P:L treatment ate significantly less of the treatment diet than the sucrose-only diet, their high lipid consumption in the first days of the study likely lead to their rapid death (Figure 4-1-3). Thus, it appears that the surviving bees were able to eat enough to meet their nutritional needs, sense the toxicity of the diet, and cease feeding, while the others did not. What caused this individual variation in behavior remains to be determined; the bees used in this study were not age-controlled, and thus there may have been physiological differences associated with age, social status, or behavioral task. Additionally, in attempt to regulate nutritional intake, the trend of over-ingesting diets at the cost of mortality has also been observed in Spodoptera littoralis caterpillars overeating carbohydrates on high- carbohydrate, low-protein diets (49).

Although feeding behavior may be affected by total nutrient concentration of the diets, we show that it was fat concentration or P:L ratios of the diets that influenced bee regulation of protein and lipid intake. In nearly all treatments in the “Single P:L diet assay” the bees consumed similar quantities of total food. Thus by fixing protein and adjusting lipid concentration in the diet, we demonstrated that the bees changed their feeding behavior to compensate for low fat in the diet, or

104

suffered mortality attempting to reach a protein target. Combining this information with that of the paired diets, the bees indeed regulated to a particular P:L ratio and concentration of nutrients.

The exact mechanism underlying the toxicity of high-fat diet consumption is unclear. One possibility is a deficiency in protein intake, though this seems unlikely because adult bees can survive quite well on sugar diets alone (7, 39). Another possibility is that high intracellular concentrations of lipids is toxic; with too much fat in the diet, insufficient amounts could be converted into storage triacylglycerols or expelled from the body (15). The ratio of the essential fatty acids ⍵-6:⍵-3 in our diets was 8:1. Excessive amount of ⍵-6 in diets (i.e., ⍵-3 deficiency) has been linked to chronic diseases in humans (50, 51), and impaired learning and physiology in honey bees (52). Moreover, high polyunsaturated fatty acids (including essential fatty acids) in the diet may lead to lipid peroxidation and cell damage, and cell membrane composition has been linked to the vast difference in maximum lifespan between honey bee queens (highly monounsaturated) and workers (highly polyunsaturated) (53).

Although not the focal test of the study, bees consistently ate similar amounts of carbohydrates across all treatments in both the single and paired diets assays. The protein:carbohydrate ratio

(P:C) intake target averaged 1:69 P:C for B. terrestris and 1:85 for B. impatiens. These intake targets are carbohydrate-biased as expected, but significantly lower than previously found for B. terrestris in studies that did not include lipid intake (40). It may be that the energy otherwise obtained from carbohydrates (e.g., for flight) was metabolized from the lipids ingested in our study, resulting in reduced feeding from the sucrose only solution (15).

105

The results of this study may provide insights into the nutritional ecology of foraging bees. First, the high requirement of carbohydrates for bumble bees is likely met by nectar foraging, which explains the attraction of bees to flowering species with high volumes and high sugar concentrations of nectar (23, 24). Because carbohydrate concentrations in pollen are fairly low, bees appear to forage on pollen to meet their protein and lipid needs. Our results suggest that bumble bees forage to obtain pollen that allows them to achieve a dietary ratio of 12:1 - 14:1 P:L.

Notably, in previous work, B. impatiens exponentially increased their foraging rates to the plant species with the 5:1 P:L ratio; moreover, using assays with caged bees and nutritionally modified pollen, B. impatiens was most attracted to 5:1 and 10:1 P:L diets (27). These preferred diets matched the results from the current study, which found that bumble bee workers survive best on, and regulate their diets to, approximately 10:1 P:L. Because the pollen P:L ratio in the previous work had an upper limit of 5:1 (27), it is unclear whether bumble bees can reach 10:1 P:L from pollen in the field. Even if the target P:L ratio cannot be met, the predisposition of bumble bees to prefer protein-biased pollen may explain host-plant preferences in natural environments (12, 24,

25, 27).

It must be noted that in the current study, we evaluated feeding preferences of isolated bumble bee workers. It is unknown whether bumble bee foragers adjust their nutritional and foraging preferences depending on the colony needs, and specifically presence of larvae (54). Information on pollen quality and its availability in the colony may be accessible to workers via pollen pots

(28, 55) allowing the colony to make informed foraging decisions. In our other studies, attraction of bumble bees to pollen with 5:1 and 10:1 P:L ratios remained intact for both bees foraging for

106

colonies or foraging in cages (in the absence of brood), suggesting that these dietary preferences are conserved across a variety of scenarios (27).

Our study demonstrated that two bumble bee species, which occupy separate geographic ranges, regulate their protein to fat intake and exhibit similar intake targets, likely due to their relatedness, similar life histories, and foraging behavior (31). Notably, their ability to regulate protein and lipids is more similar to arthropod predators than herbivores, perhaps because pollen is more nutritionally similar to prey (versus leaf tissue) with high protein and lipid concentrations (46, 48). Because bees are a monophyletic group evolved from predatory wasps (56), it is likely that bees maintained their protein and lipid biases when making the transition to pollen feeding. There may be taxa- specific P:L intake targets across bee families, genera, or species that could explain the patterns of foraging behavior and pollen preferences observed among host-plant species in field-based studies

(57). Knowing these particular intake targets can guide decisions for targeted habitat restoration protocols by matching nutritional intake targets of bee species to pollen quality of host-plant species (6).

List of Symbols and Abbreviations

GF – Geometric framework for nutrition

P:C – Protein to carbohydrate ratio

P:L – Protein to lipid ratio

Acknowledgements

We would like to thank the Grozinger, Tooker, and Wright labs for their helpful discussions and

107

critical insight to the preparation of this manuscript, Edwin Hochstedt, Victoria Bolden, Kerry

Simcock, and Caitlin Jade Oliver for their assistance in experiment preparation and data collection, and Michael Coccia for help with data analysis. This work was supported by a BBSRC transatlantic partnering award (grant number BB/I025220/1) to GMG, ADV, and GAW, NAPPC Bee Health

Improvement Project Grant to ADV, USDA AFRI NIFA Predoctoral Fellowships Grant 2014-

02219 to ADV, a NERC funded PhD studentship to DS, and generous funding from an anonymous donation to the Penn State Center for Pollinator Research.

108

H Lecithin Casein Sucrose source Nutrient source. protein a as casein and source, lipid the as used was lecithin soy source, 4 Table 2 O

-

1. Diet recipes. Diet 1.

1L - - 171g only Sucrose

-

Dietsare represented theirby protein:lipid (P:L) ratios sucroseor 100mL - 0.342g 17.1g only Protein

- 100mL 0.00342g 0.342g 17.1g 100:1

100mL 0.00456g 0.342g 17.1g 75:1

100mL 0.0068 0.342g 17.1g 50:1

5

g

100 0.013 0.342g 17.1g 25:1 mL

-

7 only and proteinand only

g

100mL 0.0342g 0.342g 17.1g 10:1

- only diets. Sucrose was used as the carbohydrate the as used was Sucrose diets. only 100mL 0.068 0.342g 17.1g 5:1

5

g

100mL 0.342g 0.342g 17.1g 1:1

100mL 1.71g 0.342g 17.1g 1:5

100mL 3.42g 0.342g 17.1g 1:10

109

Table 4-2. Mean (± SE) daily consumption (mg) of nutrients for B. terrestris foragers in “Single P:L diet assay.” Treatments are represented by their protein:lipid (P:L) diet ratio, including protein-only diet. Means marked with different letters within each column are statistically different (P < 0.05). Treatment Carbohydrate Protein Lipid 1:10 50 ± 7 ab 0.12 ± 0.02 b 1.20 ± 0.23 a 1:5 44 ± 3 b 0.11 ± 0.02 b 0.57 ± 0.10 b 1:1 50 ± 4 ab 0.44 ± 0.06 a 0.44 ± 0.06 bc 5:1 47 ± 3 ab 0.50 ± 0.06 a 0.11 ± 0.012 cd 10:1 47 ± 3 ab 0.49 ± 0.05 a 0.05 ± 0.005 d 25:1 50 ± 3 ab 0.47 ± 0.05 a 0.02 ± 0.002 d 50:1 57 ± 5 ab 0.66 ± 0.11 a 0.01 ± 0.002 d Protein-only 60 ± 4 a 0.60 ± 0.05 a -

110

Table 4-3. Cox – regression of survival for B. terrestris foragers in “Single P:L diet assay.” Treatments are represented by their protein:lipid (P:L) diet ratio, including protein-only diet. Protein-only diet (no lipid) was used as reference to test the effect of adding lipids to the diet. Note that likelihood of mortality (B) decreased for 10:1 treatment, and increased as the lipid content of the diet increased. Model: c2 = 10.52, df = 7, p = 0.161 95.0% CI for Exp(B) Treatment B SE c2 df Sig. Exp(B) Lower Upper Protein 9.667 7 0.208 50:1 0.266 0.606 0.193 1 0.661 1.305 0.398 4.275 25:1 0.186 0.606 0.094 1 0.759 1.204 0.367 3.946 10:1 -0.256 0.671 0.146 1 0.703 0.774 0.208 2.884 5:1 -0.019 0.632 0.001 1 0.976 0.981 0.284 3.389 1:1 0.375 0.586 0.410 1 0.522 1.455 0.462 4.584 1:5 0.372 0.570 0.425 1 0.514 1.451 0.474 4.436 1:10 1.136 0.540 4.424 1 0.035 3.113 1.080 8.970

111

B. terrestris impatiens B. no statistical differencesin total carbohydrate, protein, or lipid consumed. (P:C) and protein:lipid (P:L) intake ratios over seven days. Table 4 - 4. Consumption (g; mean

100:1 75:1 5 25:1 100:1 75:1 50:1 25:1 Treatment 0:1

± SE) by 335 264 248 199 398 344 470 475 Carbohydrate B. B. impatiens ± ± ± ± ± ± ± ±

39.5 65.4 36.1 29.5 51.9 46.7 70.2 58.5

and and

Each treatment was pa B. terrestris B. terrestris 5.01 5.01 4.09 3.47 2.74 4.34 3.84 6.62 5.46 Protein ± ± ± ± ± ± ± ±

0.76 1.32 0.62 0.41 0.66 0.90 1.29 0.90 foragers in the “Paired P:L diets assay” and protein:carbohydrate

ired with a 5:1 P:L diet. Within each species, there were were there species, each Within diet. P:L 5:1 a with ired 0.27 0.27 0.32 0.26 0.25 0.29 0.37 0.54 0.56 Lipid ± ± ± ± ± ± ± ±

0.05 0.13 0.09 0.05 0.06 0.15 0.12 0.11

1:66.86 1:64.61 1:71.39 1:72.41 1:91.69 1:89.55 1:71.05 1:87.01 P:C

18.40 12.98 13.29 10.83 14.83 10.49 12.22 9. P:L 84

112

Figure 4-1. Mean (± SE) daily consumption of diets across treatments for B. terrestris foragers in “Single P:L diet assay.” Treatments are represented by their protein:lipid (P:L) treatment diet ratio, including protein-only diets. Diets are represented as sucrose-only and diet associated with each treatment. Asterisks represent significant differences (P < 0.05) in diet consumed within treatment (N = 15 bees/treatment).

113

Figure 4-2. Nutritional arrays of B. terrestris foragers surviving seven days in “Single P:L diet assay.” Treatments are represented by their protein:lipid (P:L) diet ratio, including protein-only diet. Markers of each treatment represent mean cumulative consumption of each nutrient for each successive day up to seven days forming daily trajectories. a) carbohydrate and protein array, b) protein and lipid array, c) carbohydrate and lipid array (NProtein = 10, N50:1 = 9, N25:1 = 9, N10:1 = 11, N5:1 =10, N1:1 = 8, N1:5 = 7, N1:10 = 4).

114

Figure 4-3. Survival curve of B. terrestris foragers in “Single P:L diet assay.” Treatments are represented by their protein:lipid (P:L) treatment diet ratio, including protein-only diet. Note that mortality increased as the lipid content of the diets increased (N = 15 bees/treatment).

115

Figure 4-4. Mean (± SE) cumulative consumption of nutrients by deceased (N = 11) and surviving (N = 4) B. terrestris foragers in 1:10 P:L treatment on Day 3 of “Single P:L diet assay”: a) carbohydrate and protein, b) protein and lipid, c) carbohydrate and protein. Note that surviving bees ate significantly less protein and lipid than the deceased bees.

116

Figure 4-5. Mean (± SE) daily consumption of diets across treatments for a) B. impatiens and b) B. terrestris foragers in “Paired P:L diets assay.” Diets are represented as 5:1 P:L, sucrose-only, and the treatment P:L diet (25:1, 50:1, 75:1, and 100:1). Bars marked with different letters are statistically different (P < 0.05) within treatment (N = 20 bees/treatment).

117

Figure 4-6. Daily trajectories of B. impatiens (a-c) and B. terrestris (d-f) in “Paired P:L diets assay.” Treatments are represented by their protein:lipid diet ratio (P:L) paired with 5:1 P:L diet. Markers within each diet represent mean cumulative consumption of each nutrient for each successive day up to seven days: a,d) carbohydrate and protein trajectories, b,e) protein and lipid trajectories, c,f) carbohydrate and lipid trajectories (B. impatiens: N25:1 = 16, N50:1 = 16, N75:1 = 12, N100:1 = 16; B. terrestris: N25:1 = 12, N50:1 = 16, N75:1 = 14, N100:1 = 14).

118

Figure 4-7. Mean (± SE) cumulative consumption nutrients of B. impatiens and B. terrestris foragers in “Paired P:L diets assay” that survived for seven days. Note for both species there were no significant differences in carbohydrate, protein, or lipid consumption across treatments. Treatments are represented by protein:lipid diet ratio (P:L) paired with 5:1 P:L diet: a) carbohydrate and protein, b) protein and lipid. Lines represent the different diet rails, emphasizing that across treatments all P:L intake targets lie within our expected 25:1-5:1 P:L range, c) carbohydrate and protein (B. impatiens: N25:1 = 16, N50:1 = 16, N75:1 = 12, N100:1 = 16; B. terrestris: N25:1 = 12, N50:1 = 16, N75:1 = 14, N100:1 = 14).

119

Figure 4-8. Survival curve of B. impatiens and B. terrestris foragers in “Paired P:L diets assay.” Treatments are represented by their species and protein:lipid diet ratio (P:L) paired with 5:1 P:L diet (N = 20 bees/treatment).

120

References

1. Ollerton J, Erenler H, Edwards M, Crockett R (2014) Pollinator declines. Extinctions of aculeate pollinators in Britain and the role of large-scale agricultural changes. Science 346(6215):1360–1362.

2. Goulson D, Nicholls E, Botías C, Rotheray EL (2015) Bee declines driven by combined stress from parasites, pesticides, and lack of flowers. Science 347(6229) doi: 10.1126/science.1255957.

3. Biesmeijer JC, et al. (2006) Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 313(5785):351–354.

4. Carvell C, et al. (2006) Declines in forage availability for bumblebees at a national scale. Biol Conserv 132(4):481–489.

5. Potts SG, et al. (2010) Global pollinator declines: trends, impacts and drivers. Trends Ecol Evol 25(6):345–353.

6. Vaudo AD, Tooker JF, Grozinger CM, Patch HM (2015) Bee nutrition and floral resource restoration. Curr Opin Insect Sci 10:133–141.

7. Brodschneider R, Crailsheim K (2010) Nutrition and health in honey bees. Apidologie 41(3):278–294.

8. Génissel A, Aupinel P, Bressac C, Tasei JN, Chevrier C (2002) Influence of pollen origin on performance of Bombus terrestris micro-colonies. Entomol Exper Applic 104(2-3):329– 336.

9. Human H, Nicolson SW, Strauss K, Pirk CWW, Dietemann V (2007) Influence of pollen quality on ovarian development in honeybee workers (Apis mellifera scutellata). J Insect Physiol 53(7):649–655.

10. Tasei J-N, Aupinel P (2008) Nutritive value of 15 single pollens and pollen mixes tested on larvae produced by bumblebee workers (Bombus terrestris, Hymenoptera: Apidae). Apidologie 39(4):397–409.

11. Alaux C, Ducloz F, Crauser D, Le Conte Y (2010) Diet effects on honeybee immunocompetence. Biol Lett 6(4):562–565.

12. Cardoza YJ, Harris GK, Grozinger CM (2012) Effects of Soil Quality Enhancement on Pollinator-Plant Interactions. Psyche 2012(4):1–8.

13. Li C, Xu B, Wang Y, Feng Q, Yang W (2012) Effects of dietary crude protein levels on development, antioxidant status, and total midgut protease activity of honey bee (Apis mellifera ligustica). Apidologie 43(5):576–586.

14. Di Pasquale G, et al. (2013) Influence of pollen nutrition on honey bee health: go pollen

121

quality and diversity matter? PLoS ONE 8(8):e72016.

15. Canavoso LE, Jouni ZE, Karnas KJ, Pennington JE, Wells MA (2001) Fat metabolism in insects. Ann Rev Nutr 21:23–46.

16. Toth AL, Kantarovich S, Meisel AF, Robinson GE (2005) Nutritional status influences socially regulated foraging ontogeny in honey bees. J Exp Biol 208(24):4641–4649.

17. Fliszkiewicz M, Wilkaniec Z (2007) Fatty acids and amino acids in the fat body of bumblebee Bombus terrestris (L) in diapausing and non-diapausing queens. J apic Sci 51(1):55–63.

18. Feldlaufer MF, Svoboda JA, Herbert EW Jr (1986) Makisterone A and 24- methylenecholesterol from the ovaries of the honey bee, Apis mellifera L. Experientia 42(2):200–201.

19. Roulston TH, Cane JH (2000) Pollen nutritional content and digestibility for animals. Plant Syst Evol 222(1):187–209.

20. Vanderplanck M, et al. (2014) How Does Pollen Chemistry Impact Development and Feeding Behaviour of Polylectic Bees? PLoS ONE 9(1):e86209.

21. Dobson HEM, Bergström G (2000) The ecology and evolution of pollen odors. Plant Syst Evol 222(1-4):63–87.

22. Pacini E, Hesse M (2005) Pollenkitt–its composition, forms and functions. Flora- Morphology 200(5):399–415.

23. Cnaani J, Thomson JD, Papaj DR (2006) Flower choice and learning in foraging bumblebees: effects of variation in nectar volume and concentration. Ethology 112:278– 285.

24. Somme L, et al. (2014) Pollen and nectar quality drive the major and minor floral choices of bumble bees. Apidologie 46(1):92–106.

25. Hanley ME, Franco M, Pichon S, Darvill B, Goulson D (2008) Breeding system, pollinator choice and variation in pollen quality in British herbaceous plants. Funct Ecol 22(4):592– 598.

26. Konzmann S, Lunau K (2014) Divergent rules for pollen and nectar foraging bumblebees – a laboratory study with artificial flowers offering diluted nectar substitute and pollen surrogate. PLoS ONE 9(3):e91900.

27. Vaudo AD, Patch HM, Mortensen DA, Tooker JF, Grozinger CM (2016) Macronutrient ratios in pollen shape bumble bee (Bombus impatiens) foraging strategies and floral preferences. Proc Natl Acad Sci USA 10.1073/pnas.1606101113.

28. Dornhaus A, Chittka L (2005) Bumble bees (Bombus terrestris) store both food and

122

information in honeypots. Behav Ecol 16(3):661–666.

29. Pereboom JJM (2000) The composition of larval food and the significance of exocrine secretions in the bumblebee Bombus terrestris. Insect Soc 47(1):11–20.

30. Pereboom JJM, Duchateau MJ, Velthuis HHW (2003) The organisation of larval feeding in bumblebees (Hymenoptera, Apidae) and its significance to caste differentiation. Insect Soc 50(2):127–133.

31. Amsalem E, Grozinger CM, Padilla M, Hefetz A (2015) The Physiological and Genomic Bases of Bumble Bee Social Behaviour. Genomics, Physiology and Behaviour of Social Insects, Advances in Insect Physiology. (Elsevier), pp 37–93.

32. Tasei J-N, Aupinel P (2008) Validation of a method using queenless Bombus terrestris micro-colonies for testing the nutritive value of commercial pollen mixes by comparison with queenright colonies. J Econ Entomol 101(6):1737–1742.

33. Simpson SJ, Raubenheimer D (1993) A multi-level analysis of feeding behaviour: the geometry of nutritional decisions. Philos Trans Roy Soc Bs 342(1302):381–402.

34. Simpson SJ, Sibly RM, Lee KP, Behmer ST, Raubenheimer D (2004) Optimal foraging when regulating intake of multiple nutrients. Anim Behav 68(6):1299–1311.

35. Behmer ST (2009) Insect herbivore nutrient regulation. Annu Rev Entomol 54:165–187.

36. Raubenheimer D, Simpson SJ (1999) Integrating nutrition: a geometrical approach. Entomol Exper Applic 91(1):67–82.

37. Pirk CWW, Boodhoo C, Human H, Nicolson SW (2009) The importance of protein type and protein to carbohydrate ratio for survival and ovarian activation of caged honeybees ( Apis mellifera scutellata). Apidologie 41(1):62–72.

38. Altaye SZ, Pirk CWW, Crewe RM, Nicolson SW (2010) Convergence of carbohydrate- biased intake targets in caged worker honeybees fed different protein sources. J Exp Biol 213(19):3311–3318.

39. Paoli PP, et al. (2014) Nutritional balance of essential amino acids and carbohydrates of the adult worker honeybee depends on age. Amino Acids 46(6):1449–1458.

40. Stabler D, Paoli PP, Nicolson SW, Wright GA (2015) Nutrient balancing of the adult worker bumblebee (Bombus terrestris) depends on the dietary source of essential amino acids. J Exp Biol 218:793–802.

41. Dussutour A, Simpson SJ (2012) Ant workers die young and colonies collapse when fed a high-protein diet. Proc Roy Soc B 279(1737):2402–2408.

42. Grandison RC, Piper MDW, Partridge L (2009) Amino-acid imbalance explains extension of lifespan by dietary restriction in Drosophila. Nature 462(7276):1061–1064.

123

43. Stockhoff BA (1993) Ontogenetic change in dietary selection for protein and lipid by gypsy moth larvae. J Insect Physiol 39(8):677–686.

44. Thompson SN, Redak RA (2005) Feeding behaviour and nutrient selection in an insect Manduca sexta L. and alterations induced by parasitism. J Comp Physiol 191(10):909–923.

45. Mayntz D, Raubenheimer D, Salomon M, Toft S, Simpson SJ (2005) Nutrient-specific foraging in invertebrate predators. Science 307(5706):111–113.

46. Jensen K, Mayntz D, Toft S, Raubenheimer D, Simpson SJ (2011) Nutrient regulation in a predator, the wolf spider Pardosa prativaga. Anim Behav 81(5):993–999.

47. Velthuis HHW, van Doorn A (2006) A century of advances in bumblebee domestication and the economic and environmental aspects of its commercialization for pollination. Apidologie 37(4):421–451.

48. Raubenheimer D, Mayntz D, Simpson SJ, Tøft S (2007) Nutrient-specific compensation following diapause in a predator: implications for intraguild predation. Ecology 88(10):2598–2608.

49. Raubenheimer D, Lee KP, Simpson SJ (2005) Does Bertrand's rule apply to macronutrients? Proc Roy Soc B 272(1579):2429–2434.

50. Simopoulos AP (2002) The importance of the ratio of omega-6/omega-3 essential fatty acids. Biomed Pharmacother 56(8):365–379.

51. Simopoulos AP (2008) The importance of the omega-6/omega-3 fatty acid ratio in cardiovascular disease and other chronic diseases. Exp Biol Med 233(6):674–688.

52. Arien Y, Dag A, Zarchin S, Masci T, Shafir S (2015) Omega-3 deficiency impairs honey bee learning. Proc Natl Acad Sci USA 112(51):15761–15766.

53. Haddad LS, Kelbert L, Hulbert AJ (2007) Extended longevity of queen honey bees compared to workers is associated with peroxidation-resistant membranes. Exp Gerontol 42(7):601–609.

54. Hendriksma HP, Shafir S (2016) Honey bee foragers balance colony nutritional deficiencies. Behav Ecol Sociobiol 70(4):509–517.

55. Kitaoka TK, Nieh JC (2008) Bumble bee pollen foraging regulation: role of pollen quality, storage levels, and odor. Behav Ecol Sociobiol 63(4):501–510.

56. Danforth BN, Cardinal S, Praz C, Almeida EAB, Michez D (2013) The impact of molecular data on our understanding of bee phylogeny and evolution. Annu Rev Entomol 58:57–78.

57. Behmer ST, Joern A (2008) Coexisting generalist herbivores occupy unique nutritional feeding niches. Proc Natl Acad Sci USA 105(6):1977–1982.

124

Chapter 5: Bumble bees (Bombus impatiens) defend pollen nutritional preferences in the field while nutritional intake drives growth and reproduction

Anthony D. Vaudo, Liam M. Farrell, Harland M. Patch, Christina M. Grozinger, John F. Tooker

Department of Entomology, Center for Pollinator Research, The Pennsylvania State University, 501 ASI Building, University Park, PA 16802, USA

Abstract

Adaptation of foraging behavior to obtain appropriate nutrients from the environment is critical to insect development and fitness. Bumble bees (Bombus spp.) form annual colonies which must rapidly increase their worker populations to support rearing reproductive individuals prior to the end of the season. Therefore, colony growth and reproduction should be dependent on the quality and quantity of pollen resources in the surrounding landscape. Our previous research found that B. impatiens foraging rates to different plant species is shaped by pollen protein:lipid nutritional ratios

(P:L), with foragers preferring pollen with a ~5:1 P:L. In this study, we placed B. impatiens colonies in three different types of landscapes (forest, forest edge, and valley) to determine if the nutritional quality of pollen collected by the colonies differed between landscape types. We found that landscape type did not influence the nutritional quality of the collected pollen, with colonies collecting pollen averaging 4:1 P:L. Furthermore, there was no difference in the nutritional quality of the pollen collected by colonies that successfully reared reproductives and those that did not.

We found however, that foraging rates, pollen corbiculate mass, colony growth in biomass, and lifetime population were related to colony reproduction. Overall, ‘nutrient intake’, calculated as the colony-level intake rate of nutrients (protein, lipid, and sugar), was strongly related to colony growth and reproductive output. Therefore, we conclude that although B. impatiens is a generalist

125

pollinator, they are pollen nutritional specialists, and colony development is dependent on the preferred resource abundance in the surrounding landscape.

Key words: colony development, foraging preferences, nutritional ecology, pollen quality, pollination

Introduction

Foraging behavior to obtain appropriate nutrition from the environment is considered a critical evolutionary adaptation to maximize fitness for animals, including insects (1). At a fundamental level, every organism needs a particular quantity of macronutrients for development and reproduction. Nutrient concentrations and ratios, or ‘intake targets’, that are optimal for fitness can be taxa specific (2, 3). These nutritional needs are anticipated to drive foraging behavior, ensuring that an animal obtains its correct intake target from varied environments where the nutritional qualities of particular resources may be unbalanced respective to the animal’s optimal needs.

Indeed, several studies have demonstrated that different species can forage preferentially amongst multiple unbalanced food sources to meet their species-specific macronutrient (carbohydrate, protein, and lipids) intake targets (4-8).

Foraging bees obtain all their nutrients from floral pollen and nectar that vary between plant species in quality, quantity, and availability throughout space and time. Foraging bees must select amongst these resources support their own homeostasis and reproduction, and provide nutrients for larvae confined to brood cells (9-12). In landscapes where floral abundance and diversity are reduced, it may be particularly challenging for bees to meet their intake targets. Indeed, global declines in bee populations have been linked to habitat degradation that reduces floral abundance

126

and diversity (13, 14), though whether these declines are truly due to the inability of bees to meet their intake targets remains to be determined.

Bumble bee species (Bombus spp.) may be especially sensitive to declines in the nutritional quality and diversity of floral resources in the landscape. Bumble bees form annual colonies, initiated by a single foundress queen, that ultimately reach sizes of several hundred individuals before producing the next generation of reproductives (female gynes and males) by the end of the growing season (15-19). Thus, these species must have continuous floral resources available for months to continuously grow to reach a reproductive stage. In this study, we use the Common Eastern

Bumble Bee, B. impatiens, as our model system. Bombus impatiens is active throughout the vast majority of the growing season in central Pennsylvania, where foundresses initiate colonies in

April and gyne and males are active throughout September and October (20).

Only a handful of studies have monitored how bumble bee colony dynamics and fitness

(production of reproductives) are affected by landscape and floral resource availability (18, 21-

24), and none using B. impatiens. Colony growth, measured by total colony population (total brood cells), or biomass, appears to be strongly correlated to early season resource availability (18, 23), although growth has not been found to be a predictor of reproductive output, in terms of males or gynes (18, 19, 21). In areas with higher density and diversity of floral resources, colonies grew more quickly and larger, yet did not differ in their ultimate reproductive output (21). This lack of correlation is surprising, since late season resources are clearly needed to maintain colony growth to produce colonies large enough to support the rearing of the next generation of reproductives. It is possible that though environmental differences led to differences in growth rates, ultimately all

127

the colonies were able to reach the sizes necessary to produced equivalent numbers of reproductives (22, 25). Another unexplored possibility is that the switch to gyne and drone production is environmentally driven, where bees respond to depleting resource availability by producing reproductive individuals earlier (26). Nonetheless, previous studies have not examined an important aspect of colony dynamics and landscape resources: how landscape influences colony-level foraging behavior and ability to obtain high quality nutrition.

Bumble bees, specifically B. impatiens in this study, are considered to be generalists because they visit a large variety of plant species (27). However, mounting evidence has indicated that bumble bees, especially B. impatiens and B. terrestris (found in Europe), show preferences for plant species based on their pollen's nutritional quality (28-33). Pollen provides key nutrition for larval development and suboptimal pollen quality can lead to reproductive deficit, egg cannibalism, and larval ejection (34, 35). Bumble bees show preferences in the lab for pollen diets higher in protein concentration (protein concentration is generally reduced with the addition of cellulose powder;

(30, 33, 36, 37), and these preferences extend to the field amongst plant species or within the same species (28, 29). Furthermore, bumble bee colonies will increase their foraging efforts to higher quality pollen (or nectar), or reduce foraging efforts to low quality pollen, even if no alternative is available (30, 38, 39).

Our previous research revealed that B. impatiens, when foraging for their colony in a foraging arena, preferred host-plant species with pollen of high (5:1, the maximum available in our study) protein:lipid ratios (P:L) (40). Furthermore, foragers nearly ignored plant species offering the lowest P:L pollen, even when abundant pollen was available for collection (40). These P:L

128

preferences were maintained in the lab, where in the absence of external floral cues and brood, adult bees consistently preferred higher P:L pollen when offered choices of pollen from two different plant species, and preferred 5:1 and 10:1 P:L of nutritionally modified pollen (40). Using synthetic diets, isolated B. impatiens workers regulated their dietary intake of proteins and lipids to ~10:1 P:L intake target (41).

Here, we observed B. impatiens colonies throughout their development in different types of landscapes that vary in floral diversity and abundance. We hypothesize that differences in landscapes floral diversity and abundance would lead to differences in the nutritional quality and quantity available to foraging bees, which would ultimately lead to differences in colony growth and reproduction. We 1) determined if the nutritional quality of the pollen collected by the colonies varied among landscapes or if bees defended their intake targets regardless of landscape type, 2) determined whether behavioral and nutritional factors (foraging rates, pollen quantity, and pollen quality) influence colony growth and reproduction, and 3) compared foraging behavior and growth trajectories between exemplary colonies in different landscapes. Overall, these data provide critical information about how bumble bees’ nutritional intake and foraging behavior is affected by landscape resources, and how these three factors interact to impact colony dynamics and fitness.

Methods

Site selection and bumble bee placement:

We placed 24 bumble bee colonies (Biobest Canada Ltd., Leamington, ON) in three typical landscape types of central Pennsylvania, USA. The landscape types were: 1) “Valley”, a landscape composed of agricultural and residential land use, 2) “Edge”, the border between agriculture and

129

forest landscapes, and 3) “Forest”, a completely wooded landscape. Twelve sites were chosen (4 sites/habitat), arranged in a grid, such that each site was approximately 1km apart from the others

(Figure 5-1). Valley and Edge habitats were located on Penn State’s Russell E. Larson Agricultural

Research Center, while the Forest habitats were located in the adjacent section of Rothrock State

Forest (Pennsylvania DCNR Bureau of Forestry research reference #SFRA-1511). We placed two bumble bee colonies at each site with colony entrances facing opposite directions from each other.

The colonies were elevated off the ground on cinder blocks and were secured to the blocks with twine. We tented heavy-duty tarp above the colonies to protect them from direct sun and rain. At the start of the experiment, the colonies contained one queen and averaged 50 ± 2 workers and 73

± 5 brood cells (mean ± SEM) and did not differ between sites (initial workers: F11,23 = 0.60, P =

0.79, initial brood cells: F11,23 = 0.96, P = 0.52). All colonies were deployed on June 3, 2015.

Data collection:

Each week of the study, on non-rainy days, we monitored the foraging rate each colony for one hour from ~0900-1600, randomizing the order of sites observed each week. We recorded the number of foragers returning to each colony with pollen loads (‘pollen foraging rate’) and those without (‘non-pollen foraging rate’). During the hour, we collected corbiculate pollen loads from a maximum of five pollen foragers during the one-hour observation periods (unfortunately in many cases, our pollen quota could not be met due to low foraging rates). As the pollen forager landed in the colony entrance, we held her by her mid- or hind-leg with forceps and scraped the pollen loads off the corbiculae with soft forceps into 1.7mL microcentrifuge vials. Pollen samples were placed on ice in the field and stored in -80°C until analysis. We analyzed each corbiculate pollen load for its dry mass (mg) and recorded corbiculate load mass as mg per forager. We analyzed the

130

pollen load protein, lipid, and sugar concentrations and protein:lipid ratio (P:L) as in Vaudo, 2016

(40).

Each week, we visited the colonies once at night between 2300-0300 to weigh each colony. We weighed (g) the plastic box containing each colony using a Tree KHR3000 High Resolution

Kitchen Scale (LW Measurements, LLC, Rohnert Park, CA). To obtain the biomass of the colony, we subtracted the average weight of 5 clean Biobest boxes (provided by Biobest) from the measured weight. To most accurately measure the biomass influenced completely by the field, we calculated the actual ‘maximum biomass’ of each colony after the third week of the study. Because colonies had different initial masses, we also calculated ‘maximum biomass gain’ of each colony as the initial mass subtracted from the ‘maximum biomass’ after the third week of the study (such that all late instar larvae and pupae reared prior to delivery had emerged).

We measured colony biomass until the colony failed to produce any more workers or reproductives

(males or gynes) or August 7, 2015, whichever came sooner. At the time of colony termination, any remaining pupae were allowed to emerge in the lab and we counted the number of total cells as a measure of ‘lifetime population’ (old cells are not reused to rear larvae). We counted the total number of reproductives (males or gynes) produced by each colony (only colony #18 produced gynes), and subsequently named any colony that produced reproductive individuals as reproductively “successful” for our analyses.

Statistical analyses:

All statistical analyses were conducted with JMP Pro 12.2.0 (SAS Institutes, Inc. 2015).

131

Pollen nutrition: We analyzed the distribution of all pollen loads by their protein, lipid, and sugar concentrations (�g nutrient/mg pollen) and protein:lipid (P:L) ratios. We determined if pollen nutritional value differed 1) between landscape types and 2) between reproductively successful or unsuccessful colonies using ANOVA. We analyzed differences in pollen protein and lipid concentrations, and P:L ratios by habitat and week with two-way ANOVA.

Nutrition, behavior, growth, and reproduction: To determine the influence of landscape type on colony behavior, growth, and reproduction, we 1) analyzed colony maximum biomass, maximum biomass gain, lifetime population, number of reproductives produced, and colony-average non- pollen and pollen foraging rates with ANOVA and 2) determined if colonies differed in reproductive success (whether they produced one reproductive individual or not) between landscape types with contingency analysis.

To determine the correlated factors that influence colony growth and reproduction independent of landscape type, we conducted a principal components analysis (PCA) with the following variables related to 1) pollen nutrition: colony average pollen protein, lipid, and sugar concentrations and

P:L values; 2) foraging behavior and resource availability: colony average ‘pollen foraging rate’ and ‘non-pollen foraging rate’ and colony average corbiculate pollen load mass; 3) colony growth: lifetime population, maximum biomass, maximum biomass gain; and 4) reproduction: reproductive success (which we ordered 1 = “successful”, 0 = not reproductive) and total number of reproductive individuals produced.

132

Because nutritional quality and quantity are integral to development and reproduction, and the unit of reproduction for bumble bees is the colony itself, we created a metric to evaluate colony-level nutritional intake. ‘Nutritional intake’ or “mg nutrient/hour” is calculated as colony-average corbiculate pollen nutritional concentration multiplied by the colony-average corbiculate pollen mass multiplied by the colony average pollen foraging rate (averaged across weeks). Nutritional intake is calculated for protein, lipid, and carbohydrate separately:

mg nutrient ug nutrient # pollen foragers nutritional intake = = × mg corbiculate pollen × hr mg pollen hr

This metric provides a rate of intake of nutrients by the colony, or measure of quantity of nutrients foraged by each colony on an hourly basis. To determine the relationship between nutritional intake and colony growth, we used regression analyses between nutrient intake and colony lifetime population and maximum biomass gain. To determine the relationship between nutritional intake and reproduction, we used t-tests to evaluate the difference in intake between reproductively successful and unsuccessful colonies, and used regression analysis between intake and the number of reproductives produced by successful colonies.

Case study of two ‘best’ field sites: Finally, we examined the similarities and differences of growth trajectories and pollen foraging behavior between the four most reproductively successful colonies, located at sites Valley 2 and Edge 4. We plotted the weekly biomass trajectories of each colony and weekly pollen foraging rates. We also overlaid the estimated occurrence of the switching point of each colony, or time male or gyne eggs would have been laid, back-calculated from the time of the emergence of adult males and gynes (16).

133

Results and Discussion

Pollen nutrition:

We collected corbiculate pollen loads from 301 bumble bee foragers as they returned to their colonies and analyzed each for nutritional quality (Table 5-1, Figure 5-2). The P:L concentrations of the pollen averaged 4:1 ± 0.14 (Table 5-1, Figure 5-2a). Our previous findings showed that B. impatiens preferred pollen of 5:1 (the maximum available for the plant species used in our study) and regulated their P:L intake from synthetic diets to ~10:1(40, 41). The distribution of pollen across the colonies and weeks was both above and below 5:1, averaging to 4:1 (Figure 5-2a). This indicates that there was substantial variation in the P:L ratios of the available pollen in the landscapes, and the bumble bees balanced collections across many plant species to reach this ratio, exhibiting flexible foraging behavior to obtain quality pollen (33, 42). It is possible that even though B. impatiens have preferred nutritional ratios that drive their pollen foraging behavior, P:L ratios such as 10:1 may be rare or limited in quantity in the environment as it may be costly for plants to produce. For example, doubling the protein concentration or halving the lipid concentration of our highest (~5:1) P:L pollen in previous experiments would put the concentrations at the highest or lowest extreme respectively found in nature (see ref 9, 40, 43 for pollen nutritional concentrations).

Protein, lipid, sugar, and P:L ratios did not differ between colonies in different landscape types

(Table 5-1). P:L ratios did vary throughout the season by week but not between landscape type, indicating that bumble bee colonies were collecting similar resources throughout the season in different landscapes (Figure 5-2b). Interestingly, pollen protein concentration varied by week where lipid concentration did not (protein: F8,300 = 15, P < 0.01; lipid: F8,296 = 1.2, P = 0.09). This

134

suggests that bumble bees prioritized and consistently collected pollen with low lipid concentrations, and P:L concentrations were driven by protein content that changed as pollen resources turned over throughout season. B. impatiens therefore may be bound to collect only quality pollen, searching for only diverse or quality patches of flowers (33, 42), because sacrificing pollen quality could result in egg cannibalism, larval ejection, or poor larval development (34, 35).

To evaluate environmental factors may influence pollen protein content, we analyzed the influence of daily temperature and time of day on the protein content. The interaction of daily temperature

2 and time of day was significant (F3,98 = 10.93, P < 0.01, R = 0.26), where protein content of pollen was higher on warmer days, but decreased as the day progressed. These results are similar to our previous findings that more preferred pollen was collected by bumble bees earlier in the day (40,

44). This trend was not observed with lipid concentrations however (F3,97 = 0.52, P = 0.67), again indicating bumble bees’ prioritization of pollen lipid quality.

Nutrition, behavior, growth, and reproduction:

There were no statistical differences in colony reproductive success between landscape types (�2

= 5.2, P = 0.07) though there were differences in number of successful colonies (# successful:

Forest: 1, Edge: 3, Valley: 5). Forest colonies had the lowest rates of reproductive success likely because they were damaged by mold and large mammals. Furthermore, there were no statistical differences between landscape types in colony maximum biomass, maximum biomass gain, number of reproductives and non-pollen foraging rates (Table 5-2). However, there were differences between colony lifetime population and colony average pollen foraging rates between landscape types with Valley colonies exhibiting the highest of both categories (Table 5-2). This

135

indicates that Valley colonies may have had access to more preferred pollen resources increasing foraging rates and worker populations in the agricultural landscape (45). There was a numerical trend that colonies in the Valley had the highest values of all parameters, followed by Edge, and lastly Forest (Table 5-2). But these values are largely driven by a few colonies within each landscape type (See ‘Case study’ below).

Although there were no differences in the nutritional quality of the pollen obtained by the foraging bees between landscape types, and only some differences between colony growth and behavior between landscape types, there were differences between the colonies’ growth and reproductive success. Therefore, we determined the nutritional and behavioral factors related to colony growth and reproduction using PCA. Factor analysis divided the data into two factors accounting for

69.1% of the variance (Figure 5-3). Factor 1 (49.4% of the variance) included the positively correlated variables related to behavior, growth, and reproduction: average pollen foraging rate, average non-pollen foraging rate, corbiculate pollen mass, colony maximum biomass gain, colony maximum biomass, lifetime population, colony success, and number of reproductives produced by each colony. Factor 2 (19.7% of the variance) included pollen nutritional variables with pollen protein and P:L negatively correlated with pollen sugar and lipid concentrations (Figure 5-3). As in our nutritional results above, these data indicate that pollen nutritional quality is consistent across colonies and independent of the outcome of each colony.

Thus, colony weight, total population, were all correlated with colony reproductive success and the number of reproductives (Figure 5-3), supporting the model that increasing colony size is critical for colonies to switch to producing new reproductives (23). All of these factors are

136

positively correlated with pollen and non-pollen foraging rates and corbicular pollen size, indicating that nutritional quality is held consistent by the foraging bees while nutritional quantity through preferred resource abundance is critical for colony development (18). Furthermore, the analysis suggests a positive feedback loop between environmental availability of resources

(nutritional quantity) and pollen foraging rates and capacity (30, 38, 39, 46, 47): increase resources lead to higher worker populations and increased foraging (26).

By creating a metric of colony-level nutritional intake, we were able to integrate nutrition and behavior to evaluate the rate of intake of nutrients by the colonies and how this correlated with colony growth and reproduction. Nutritional intake for each macronutrient was significantly

2 correlated to total colony population (protein intake: F1,15 = 92.3, P < 0.01, R = 0.86, lipid intake:

2 2 F1,15 = 90.5, P < 0.01, R = 0.86, carbohydrate intake: F1,17 = 78.6, P < 0.01, R = 0.84; Figure 5-

2 4a) and colony maximum biomass gain (protein intake: F1,17 = 62.8, P < 0.01, R = 0.88, lipid

2 2 intake: F1,17 = 97.3, P < 0.01, R = 0.92, carbohydrate intake: F1,17 = 90.4, P < 0.01, R = 0.91;

Figure 5-4b). Therefore, between all three macronutrients, colonies exhibiting higher nutritional intake were able to outgrow other colonies. Furthermore, nutritional intake was significantly higher in colonies that produced reproductive individuals than those that did not (protein intake: t19 = 2.3, P = 0.02; lipid: t19 = 0.03, P = 0.01, sugar: t22 = 2.2, P = 0.02; Figure 5-5a). Finally, nutritional intake of all the macronutrients found in pollen was strongly positively correlated to the number of total reproductive individuals produced by a successful colony (protein intake: F2,6

2 2 = 26.1, P < 0.01, R = 0.90, lipid intake: F1,8 = 56.1, P < 0.01, R = 0.95, carbohydrate intake:

2 F2,6= 43.7, P < 0.01, R = 0.94; Figure 5-5b). Colony #18, which was the only one to produce

137

gynes (all the other reproductive colonies produced males), had the highest nutritional intake amongst all colonies for protein, lipid, and sugar (Figure 5-4, Figure 5-5).

Thus, B. impatiens foragers appear to collect pollen that matches the requirements of developing larvae (eg, a 4:1 P:L ratio) and then the quantity of the pollen that the foragers collect determines both the colony growth rate and its ability to reach the threshold where it can successfully produce reproductives (18, 21, 23, 24, 32, 33, 40, 42). Further studies are needed to 1) explore the level of nutritional surplus to determine maximum colony growth and 2) determine the environmental or nutritional threshold between male and/or gyne reproductive output, a still elusive question in bumble bee field studies (18, 19).

Case study of two ‘best’ field sites

The sites Valley 2 (colonies 8 and 18) and Edge 4 (colonies 17 and 21; Figure 1) had the highest reproductive output (Colony 8: 56 males; Colony 18: 60 males, 18 gynes; Colony 17: 27 males;

Colony 21: 27 males). The colonies within each site had parallel growth trajectories, but the growth trajectories differed substantially between sites (Figure 5-6). Correlated with these growth trajectories, pollen foraging rates at each site were similar between colonies, but differed between sites (Figure 5-6). This results indicate clear environmental effects on colony growth and reproductive outcome, but demonstrates that colonies can still growth and reproduce under very different conditions.

Valley 2 colonies experience constant colony growth, with rather stable pollen foraging rates.

Pollen foraging rates spiked between weeks six and seven, and this spike in pollen foraging

138

correlated to the switching point of the colony and the generation of reproductives (Figure 5-6a).

At Edge 4, there was an initial pollen foraging spike at week three that correlated with a loss in colony weight, and then subsequent weight gain for the following two weeks. There was a second increase in pollen foraging rates at Edge 4 corresponding with the colony switching points (Figure

5-6b).

It remains to be determined if the production of reproductive individuals caused the spike in pollen foraging rates at both sites, or if the spike in foraging rates triggered the production of reproductive individuals (30, 46, 48). In the case of Edge 4 colonies, it appears that pollen foraging efforts change in response to colony status. This foraging response could be that bumble bees have a reserve of unspecialized (pollen or nectar) foragers that could change roles when in need (49).

In this study, we expected to see differences in nutritional quality or quantity between landscape types, but bumble bees collected similar nutritional resources no matter where they nested. As exemplified by Valley 2 and Edge 4 sites, it appears that nutritional quantity, and subsequently colony nutritional intake, was more fine grained and differed by site. Valley 2 colonies were placed in an agricultural area where hedgerows, neighborhoods, and flowering crops had consistent floral resource availability (21); in opposition, Valley 4 colonies were placed in wheat and corn fields without any obvious wildflower habitat. However, Edge 4 colonies were placed in the only perceivable area with wildflower abundance along the Forest edge, where the other sites were regularly mowed to the tree line. In the Forest, we found little wildflower abundance. Therefore, assessing floral resources in proximity to colonies would likely be more predictive of resource abundance than assumptions by general landscape type.

139

Conclusion

Our study represents one of the few studies to examine the relationship between bumble bee nutrition, foraging behavior, and colony growth and reproduction under field conditions, and the only for B. impatiens to date (for related studies in B. terrestris, see ref 21,23,24; for B. vosnesenskii, see 18,19; for B. apositis, see 25). Importantly, we found that B. impatiens defend their nutritional preferences for pollen (particularly lipid and P:L ratios) independent of where they are nesting or the time of year (24, 32). Across the entire season and across all sites, B. impatiens collected an average of 4:1 P:L, remarkably similar to their 5:1 preferences in our controlled experiments (40). Furthermore, though the nutritional values of the collected pollen did not differ between reproductively successful and unsuccessful colonies, the foraging rates, corbiculate loads, and overall macronutrient quantities were strongly positive correlated with colony growth and reproduction.

We propose that bumble bees will expend their effort to obtain suitable resources rather than feeding their larvae subpar resources (35). Additionally, they appear to increase foraging rates to sufficient resources if and when found (30, 35, 38, 39). Our metric of ‘nutritional intake’ provides a means of measuring colony-level consumption rate of nutrients its relationship to colony dynamics by including pollen nutritional quality, colony behavior, and indicators of resource availability (pollen foraging rates and corbiculate weight). Therefore, we conclude that B. impatiens are nutritional specialists, consistently defending their nutritional intake targets, and growth and reproduction are limited by abundance of preferred pollen species (22). The

140

reproduction of colonies and global decline of bumble bee populations may be linked not just to floral resource abundance, but the quality of resources throughout the landscape (13).

Acknowledgments

We thank the J.F.T. and C.M.G. laboratories for helpful discussions and critical insight in the preparation of this manuscript. We are grateful to Scott Smiles for assistance with field site selection. We thank Biobest for donating bumble bee research colonies for the research. Work in

Rothrock State Forest was approved by Pennsylvania DCNR Bureau of Forestry research reference

#SFRA-1511. This work was supported by a North American Pollinator Protection Campaign Bee

Health Improvement Project Grant (to A.D.V.); USDA AFRI NIFA Predoctoral Fellowship Grant

2014- 02219 (to A.D.V.); Apes Valentes Undergraduate Research Award (to L.M.F.); Dutch Gold

Honey Scholarship (to L.M.F.); generous funding from an anonymous donation to the Center for

Pollinator Research at Pennsylvania State University.

141

Reproduction Habitat ave colony by analyzed ANOVA but shown, are loads o distributionsactualsuccess, reproductive For status.reproductive or habitatssignificantly differbetween not did values Table

5 -

1. 1. Summary of pollen nutritional quality collected by

Successful Unsuccessful Valley Edge Forest

Protein ( Protein 181 120 156 100 45 N

211.22 187.52 213.31 192.74 181.86 � Mean g/mgpollen) F F 2,298 P 1,20

P = 0.054 = 0.49

= 2.9; = 2.9; = 0.5; = 0.5; 15.46 rage concentrationsnutrient of collected pollen. 6.36 8.59 6.50 9.39 SE

B. B. impatiens Lipid ( Lipid 177 120 153 99 45 N

� g/mg pollen) Mean 56.16 53.17 55.15 53.42 57.65 Pollen Nutritional Content Nutritional Pollen

F colo F 1,20 2,294 P P = 0.47

nies by habitat and reproductive success. Note that nutritional

= 0.03; = 0.03; = 0.90 = 0.7; = 0.7; 1.48 1.74 1.54 1.40 4 .34 SE

Sugar Sugar ( 178 120 154 99 45 N

� 388.70 358.17 362.94 395.62 380.21 g/mg pollen) Mean

F F 2,295 1,20

P P = 0.12 = 0.14

= 2.0; = 2.0; = 2.7; = 2.7; 10.22 10.64 10.21 13.45 18.73 SE

Protein:Lipid Ratio Protein:Lipid 177 120 153 99 45 N

f all corbiculate all f

Mean 4.18 3.97 4.27 4.00 3.72 F F 2,294 1,20 P P

= 0.45 = 0.36

= 1.0; = 1.0; = 0.6; = 0.6;

0.17 0.23 0.17 0.26 0.43 S E

142

Valley Edge Forest Habitat (seeResults and Discussion) (* significantdifferences only the were ratesforaging Table 5

-

2. Summary of growth, reproduction, and foraging rates of

Max biomass (g) 106. Mean 96.19 75.33 68 F

2,22 P = 0.47

29.29 11.65 = 0.7 = 0.7 9.57 SE .

Biomass Mean 48.15 16.40 9.51 Growth and Reproduction and Growth F 2,22 P

gain (g)

= 0.25

= 1.5 = 1.5 29.5 6.27 5.21 SE 0

P Lifetime population Lifetime <

0.05)in colony dynamics between habitats, likely due to site and colony specific differences 162.38 Mean 93.60 56.60

F *

2,17 P B. impat = 0.02

26.32 29.92 = 4.8 8.36 SE

i ens

colonies colonies # reproductives Mean 21.63 7.00 0.11 F

(N=24). Note that colony lifetime population and average pollen 2,22 P = 0.07

10.77 = 2.9 4.37 0.11 SE

Non - pollen foraging rate Foraging Rates Foraging Mean 5.84 4.10 2.79 F 2,22 P

= 0.23

= 1.6 1.42 1.17 1.12 SE

(avg/colony)

P ollen foraging rate Mean 12.70 7.17 6.97

* F P = 2,22

0.007 = 6.2 1.04 1.23 1.57 SE

143

Figure 5-1. Field sites of Bombus impatiens colonies in central Pennsylvania. Two colonies were placed at each site facing opposite directions from each other. The four sites in the Valley were along the field border of The Pennsylvania State University Russell E. Larson Agricultural Research Center and were in a predominately agricultural landscape with two small residential neighborhoods. The four Edge sites were along the border of the research center and Rothrock State Forest. The Forest sites were placed approximately 5m into the forest off Kepler Rd. in Rothrock State Forest. Photograph generated by Google Earth Pro v.7.1.5.1557

144

Figure 5-2. Distributions of protein:lipid ratios (P:L) from individual Bombus impatiens forager corbiculate pollen loads. (A) Distribution of P:L ratios of all individual forager pollen loads collected (N = 297). Bars represent total numbers of corbiculate loads found within a given P:L range. (B) Mean P:L ratios (± SE) of pollen loads for each week of the study by habitat. P:L ratios did not differ between habitats each week (F13,273 = 1.2, P = 0.29). The smooth line was added to show similar trends throughout the season in each habitat. Note that protein concentration of pollen differed by week while lipid concentration did not, suggesting that P:L ratios were driven by protein concentration of pollen collected by bumble bees (see Results).

145

Figure 5-3. Principal component analysis for season long Bombus impatiens colony development, behavior, and nutrition. Factor analysis grouped factors associated with resource abundance and foraging behavior (pollen and non-pollen foraging rates and corbiculate pollen mass) with colony growth and reproduction (maximum biomass and biomass gain of colonies, lifetime population, reproductive success, and number of reproductives). Pollen nutrition was not correlated to colony dynamics because it did not differ between colonies (see Results) and was assigned to a second factor where lipid and sugar concentrations were negatively correlated to protein concentration and P:L ratios.

146

Figure 5-4. Colony level nutritional intake is highly correlated to colony (A) lifetime population and (B) maximum biomass gain. Nutritional intake is calculated as colony-level average mg lipid, protein, and sugar foraged per hour (see Methods for formula).

147

Figure 5-5. Nutritional intake positively affects colony reproduction. (A) Nutritional intake (mean ± SE) for lipid, protein, and sugar is higher in colonies that were reproductively “successful,” producing at least one productive individual. Asterisks represent statistical difference at P < 0.05. (B) Nutritional intake of lipid, protein, and sugar is linearly correlated to the number of reproductive individuals produced by “successful” colonies. Nutritional intake is calculated as colony-level average mg lipid, protein, and sugar foraged per hour (see Methods for formula).

148

Figure 5-6. Colonies within sites show similarities in behavior and growth trajectories, but variation across sites. (A) Valley 2 (colonies 8 and 18) and (B) Edge 4 (colonies 17 and 21), the two sites that produced the highest total number of reproductives in both colonies. Graphs show colony growth trajectories and pollen foraging rates by week. Maps show local habitat of each site location with 1km approximate foraging radius surrounding each colony. Vertical lines on each graph represent hypothesized switching point (back calculated from emergence of reproductives) showing the synchrony switching point and pollen foraging spikes.

149

References

1. Simpson SJ, Raubenheimer D (2012) The Nature of Nutrition (Princeton University Press) pp 239.

2. Behmer ST, Joern A (2008) Coexisting generalist herbivores occupy unique nutritional feeding niches. P Natl Acad Sci USA 105(6):1977–1982.

3. Behmer ST (2009) Insect herbivore nutrient regulation. Annu Rev Entomol 54:165–187.

4. Simpson SJ, Raubenheimer D (1993) A multi-level analysis of feeding behaviour: the geometry of nutritional decisions. Philos T R Soc B 342(1302):381–402.

5. Raubenheimer D, Simpson SJ (1999) Integrating nutrition: a geometrical approach. Entomol Exper Applic 91(1):67–82.

6. Mayntz D, Raubenheimer D, Salomon M, Toft S, Simpson SJ (2005) Nutrient-specific foraging in invertebrate predators. Science 307(5706):111–113.

7. Raubenheimer D, Mayntz D, Simpson SJ, Tøft S (2007) Nutrient-specific compensation following diapause in a predator: implications for intraguild predation. Ecology 88(10):2598–2608.

8. Jensen K, Mayntz D, Toft S, Raubenheimer D, Simpson SJ (2011) Nutrient regulation in a predator, the wolf spider Pardosa prativaga. Anim Behav 81(5):993–999.

9. Roulston TH, Cane JH (2000) Pollen nutritional content and digestibility for animals. Plant Syst Evol 222(1):187–209.

10. Nicolson SW, Nepi M, Pacini E (2007) Nectaries and Nectar eds Nicolson SW, Nepi M, Pacini E (Springer Science & Business Media, Dordrecht) pp 414.

11. Brodschneider R, Crailsheim K (2010) Nutrition and health in honey bees. Apidologie 41(3):278–294.

12. Cane JH (2016) Adult pollen diet essential for egg maturation by a solitary Osmia bee. J Insect Physiol 95:105–109.

13. Biesmeijer JC, et al. (2006) Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 313(5785):351–354.

14. Goulson D, Nicholls E, Botías C, Rotheray EL (2015) Bee declines driven by combined stress from parasites, pesticides, and lack of flowers. Science 347(6229):1255957– 1255957.

15. Duchateau MJ, Velthuis H (1988) Development and reproductive strategies in Bombus terrestris colonies. Behaviour 107(3):186–207.

150

16. Cnaani J, Schmid-Hempel R, Schmidt JO (2002) Colony development, larval development and worker reproduction in Bombus impatiens Cresson. Insect Soc 49(2):164–170.

17. Goulson D (2010) Bumblebees: behaviour, ecology, and conservation. (Oxford University Press, Oxford).

18. Williams NM, Regetz J, Kremen C (2012) Landscape-scale resources promote colony growth but not reproductive performance of bumble bees. Ecology 93(5):1049–1058.

19. Crone EE, Williams NM (2016) Bumble bee colony dynamics: quantifying the importance of land use and floral resources for colony growth and queen production. Ecol Letter 19(4):460-468.

20. Szabo TI, Pengelly DH (1973) The over-wintering and emergence of Bombus (Pyrobombus) impatiens (Cresson) (Hymenoptera: Apidæ) in southern Ontario. Insect Soc 20(2):125–132.

21. Goulson D, Hughes W, Derwent L, Stout J (2002) Colony growth of the bumblebee, Bombus terrestris, in improved and conventional agricultural and suburban habitats. Oecologia 130(2):267–273.

22. Elliott SE (2009) Surplus nectar available for subalpine bumble bee colony growth. Environ Entomol 38(6):1680–1689.

23. Westphal C, Steffan-Dewenter I, Tscharntke T (2009) Mass flowering oilseed rape improves early colony growth but not sexual reproduction of bumblebees. J Appl Ecol 46(1):187–193.

24. Kämper W, et al. (2016) How landscape, pollen intake and pollen quality affect colony growth in Bombus terrestris. Landscape Ecol 30(10):2245-2258.

25. Elliott SE (2009) Subalpine bumble bee foraging distances and densities in relation to flower availability. Environ Entomol 38(3):748–756.

26. Amsalem E, Grozinger CM, Padilla M, Hefetz A (2015) The Physiological and Genomic Bases of Bumble Bee Social Behaviour. Genomics, Physiology and Behaviour of Social Insects, Advances in Insect Physiology. (Elsevier), pp 37–93.

27. Robertson C (1928) Flowers and insects (The Science Press Printing Company, Lancaster, PA)

28. Hanley ME, Franco M, Pichon S, Darvill B, Goulson D (2008) Breeding system, pollinator choice and variation in pollen quality in British herbaceous plants. Funct Ecol 22(4):592– 598.

29. Cardoza YJ, Harris GK, Grozinger CM (2012) Effects of Soil Quality Enhancement on Pollinator-Plant Interactions. Psyche 2012(4):1–8.

151

30. Kitaoka TK, Nieh JC (2008) Bumble bee pollen foraging regulation: role of pollen quality, storage levels, and odor. Behav Ecol Sociobiol 63(4):501–510.

31. Leonhardt SD, Blüthgen N (2011) The same, but different: pollen foraging in honeybee and bumblebee colonies. Apidologie 43(4):449–464.

32. Kriesell L, Hilpert A, Leonhardt SD (2016) Different but the same: bumblebee species collect pollen of different plant sources but similar amino acid profiles. Apidologie. doi:10.1007/s13592-016-0454-6.

33. Ruedenauer FA, Spaethe J, Leonhardt SD (2016) Hungry for quality—individual bumblebees forage flexibly to collect high-quality pollen. Behav Ecol Sociobiol:1–9.

34. Tasei J-N, Aupinel P (2008) Nutritive value of 15 single pollens and pollen mixes tested on larvae produced by bumblebee workers (Bombus terrestris, Hymenoptera: Apidae). Apidologie 39(4):397–409.

35. Génissel A, Aupinel P, Bressac C, Tasei JN, Chevrier C (2002) Influence of pollen origin on performance of Bombus terrestris micro-colonies. Entomol Exper Applic 104(2-3):329– 336.

36. Ruedenauer FA, Spaethe J, Leonhardt SD (2015) How to know which food is good for you: bumblebees use taste to discriminate between different concentrations of food differing in nutrient content. J Exp Biol 218(14):2233–2240.

37. Konzmann S, Lunau K (2014) Divergent rules for pollen and nectar foraging bumblebees – a laboratory study with artificial flowers offering diluted nectar substitute and pollen surrogate. PLoS ONE 9(3):e91900.

38. Dornhaus A, Chittka L (2001) Food alert in bumblebees (Bombus terrestris): possible mechanisms and evolutionary implications. Behav Ecol Sociobiol 50(6):570–576.

39. Dornhaus A, Chittka L (2004) Information flow and regulation of foraging activity in bumble bees (Bombus spp.). Apidologie 35(2):183–192.

40. Vaudo AD, Patch HM, Mortensen DA, Tooker JF, Grozinger CM (2016) Macronutrient ratios in pollen shape bumble bee (Bombus impatiens) foraging strategies and floral preferences. P Natl Acad Sci USA 113(28):E4035–42.

41. Vaudo AD, et al. (2016) Bumble bees regulate their intake of the essential protein and lipid pollen macronutrients. J Exp Biol doi: 10.1242/jeb.140772.

42. Jha S, Kremen C (2013) Resource diversity and landscape-level homogeneity drive native bee foraging. P Natl Aca Sci USA 110(2):555-558

43. Roulston TH, Cane JH, Buchmann SL (2000) What governs protein content of pollen: pollinator preferences, pollen-pistil interactions, or phylogeny? Ecol Monogr 70(4):617– 643.

152

44. Vaudo AD, Patch HM, Mortensen DA, Grozinger CM, Tooker JF (2014) Bumble bees exhibit daily behavioral patterns in pollen foraging. Arthropod-Plant Inte 8(4):273–283.

45. Requier F, et al. (2015) Honey bee diet in intensive farmland habitats reveals an unexpectedly high flower richness and a major role of weeds. Ecol Appl 25(4):881–890.

46. Dornhaus A, Brockmann A, Chittka L (2003) Bumble bees alert to food with pheromone from tergal gland. J Comp Physiol A 189(1):47–51.

46. Hendriksma HP, Shafir S (2016) Honey bee foragers balance colony nutritional deficiencies. Behav Ecol Sociobiol 70(4):509–517.

47. Dornhaus A, Chittka L (2005) Bumble bees (Bombus terrestris) store both food and information in honeypots. Behav Ecol 16(3):661–666.

48. Hagbery J, Nieh JC (2012) Individual lifetime pollen and nectar foraging preferences in bumble bees. Naturwissenschaften 99(10):821–832.

153

Chapter 6: Discussion and Future Research

Discussion

Foraging bees are challenged to obtain the appropriate nutrients for themselves and developing offspring from pollen resources that vary in quality between host-plant species. In this dissertation, we demonstrate that bumble bees, Bombus impatiens, assess pollen quality when making foraging decisions and are selective among plant-species for pollen nutrition. Importantly, their pollen preferences are linked to their optimum protein:lipid (P:L) dietary ratio. Among host-plant species, these foraging preferences were correlated to increasing P:L values (maximum 5:1). The bees maintained these preferences for isolated pollen of the same plant species, and were more attracted to nutritionally modified pollen of 5:1 and 10:1 P:L ratios. However, attraction alone did not explain how the adults regulated their nutritional intake. When restricted to individual synthetic diets of a range of P:L values, the bees survived best on 5:1 and 10:1 diets. Furthermore, when given the opportunity to regulate their nutritional intake from two synthetic diets, they regulated their P:L intake ratio close to 10:1. Therefore, we show that the bumble bees are sensitive to and prefer pollen nutritional quality that nears their preferred P:L intake targets. Finally, in the field, when foraging among varied habitats, colonies appear to balance pollen nutrition from multiple plant species to an average of 4:1 overall. This indicates, even as generalists, bumble bees prioritize and defend their nutritional preferences.

A remarkable finding of these studies is that adult bumble bees appear to have the same nutritional preferences when foraging for a colony with brood and when isolated away from brood and social interactions (1). Different than honey bees, bumble bees do not consume pollen and convert it to brood-food through glandular secretions and foragers and nurses are not completely distinct castes

154

(2-4). Therefore, foragers are likely more directly in tune to larval nutritional needs when making foraging decisions. This may be equally true for all other bee species that feed pollen to larvae directly. Bumble bees, even as generalists, are selective foragers, and tend to forage individually, but colony level foraging rates do increase when quality resources return to the colony (5-7), indicating that bumble bee colonies do not spend as much effort collecting subpar resources. We observed evidence of this in the field where colonies exhibited similar nutritional ratios of pollen independent of colony status, but differed in their foraging rates, likely due to low preferred resource availability. Colony growth and reproductive output then was dependent on preferred resource abundance more than total resource abundance.

There were some differences however in the preferred pollen ratios found in our controlled experiments (Chapters 3 and 4) and field experiment (Chapter 5). Bumble bees preferred pollen species and modified pollen at 5:1 P:L, but among synthetic diets, they regulated their diet to >10:1

P:L. Of the host-plant species tested, Senna hebecarpa had the highest P:L ratio which was close to 5:1, and lowest lipid concentration. To achieve a 10:1 ratio, if protein content stayed the same, lipid content would have to decrease by at least half, which would place the concentration at the lowest end of the spectrum found in nature (8). In contrast, if lipid content stayed the same, protein content would have to at least double, placing it on the highest end of protein concentrations found in nature (9). Therefore, 10:1 P:L can represent an extreme case for pollen nutritional ratio, and could be costly for plants, considering lipid content is probably very important for attracting bees, as a phagostimulant and source of volatile compounds, and also practical for sticking to bee hair

(10-15). Thus, even though 10:1 P:L may be optimal for bumble bees, its availability in nature may be limited to few plant species or less abundant resources. Nonetheless, this inherent nutrient

155

desire by bumble bees may drive their pollen collections in the field to plant species that most closely resemble this need. Additionally, bumble bees may have to compensate macronutrient needs at times to balance other micronutrients in their diet and thereby not collect pollen from their most preferred species at all times.

Future Research

This research is the first to describe the phenomenon that bees’ foraging behavior is driven by host-plant pollen nutritional ratios. Now that the phenomenon is established with its known limitations, there are four research areas that can be readily studied in response: 1) how nutrition is perceived and regulated at the organismal level, 2) how it extends to the bumble bee colony, 3) how nutritional ecology extends to other bee species and plant-pollinator communities, and 4) the application to conservation.

We show that bumble bees can and do assess protein and lipid components of their food, and can select and adjust their foraging and feeding behavior to meet their nutritional needs. Therefore, research should now explore the mechanistic and physiological basis for how bees make this assessment. First, there can be volatile or tactile chemical receptors on bee antennae or mouthparts to assess pollen quality (16, 17). This would allow bees to rapidly evaluate the chemical profile of pollen. The pollenkitt contains many chemical compounds including lipids, carbohydrates, and free fatty and amino acids (12-14). Because free amino acid concentrations are correlated to protein content of pollen, and there is evidence that bees can taste amino acids directly (1, 16), it is likely that bees can assess protein content by surface amino acids. Likewise, this may be true for lipid content; and in this dissertation, we show that fats at low concentrations can be phagostimulatory

156

for bumble bees. Bees could also ingest some pollen while foraging or when returning to the colony. The nutritional profile, if rewarding or punishing may be committed to memory and therefore associated with other discriminatory floral cues (scent, color, blooming time, location) for rapid revisiting and handling. Therefore, we can test bees’ ability to learn, discriminate, and associate rewards with other cues. Furthermore, we show that bees regulate their nutritional intake over longer lengths of time (7 days). This regulatory ability was likely due to post-ingestive responses to nutrient intake. Therefore, we can test whether bees have different receptor responses to nutrients after satiation or starvation, and relate these to hemolymph levels, fat body storage, or other tissue levels of nutrients (and even extend them to different life stages, castes, or reproductive status) (18-21).

Following research should extend from the findings in Chapter 5 and further test the effect of pollen nutrition, diversity, and resource availability on colony dynamics. Because the bees in our study appeared dedicated to particular pollen quality, this can be studied in more controlled conditions where colonies are restricted to predefined diets. Using single source pollen or nutritionally modified pollen, colonies can be studied in a geometric framework type of study as exemplified in Chapter 4, where colonies can be restricted to single pollen types or allowed to choose and regulate their intake between nutritionally balanced diets. Then we can very accurately measure colony nutrient intake, growth, reproduction and other dynamics throughout the entire life cycle. Furthermore, we can also test how nutritional preferences of colonies may change at different stages, especially when rearing reproductive individuals, as well as the nutritional needs of foundresses prior to diapause and upon nest formation. Studies have been conducted on microcolonies (three workers and male larval rearing) for the effect of pollen protein or sterols

157

(22-24), but incorporating whole colonies and additional nutritional elements would be more informative. A more interesting subject would be to conduct this research while allowing bees to foraging on host-plant species to emulate a more realistic situation. Host-plant species composition and abundance can be altered in foraging arenas such as the hoop-house used in Chapter 2. It is most interesting to understand the dynamics of nutrition, foraging behavior, and colony performance in real life, and therefore, the study in Chapter 5 can be replicated across multiple seasons where the floral community is identified and quantified (25). Furthermore, using modern pollen metabarcoding techniques, we can identify the plant species and relative quantities of pollen collected by the bumble bees to determine how they utilize different plant species to balance their nutrition among the floral community (26-28).

The above studies continue to focus on bumble bee physiology and colony dynamics. However, research should be extended to other bee species and plant-pollinator communities. Simply, we can conduct similar studies to Chapters 2 - 4 with other bee species. It may be difficult to conduct these studies with unmanaged solitary bees, but if we can bait different species with nest sites, perhaps we can obtain quantity of nests to allow for acceptable replication. With multiple species, we can study if there are species-specific pollen nutritional needs and infer different nutritional niches (29). We can study how species change foraging behavior and pollen collection in competitive situations among a controlled floral community in the hoop-house (30). The most informative study would be to understand the nutritional basis behind plant-pollinator community networks in undisturbed communities. This can be explored by collecting interaction data carefully so that we have accurate measurements of bee host-plant utilization (as discussed in Chapter 2), focusing on dedicated pollen collection behavior, and accounting for floral patch size and floral

158

phenology. By quantifying resource quantity and nutritional quality of host-plant species, we can actually calculate a metric of nutritional intake of different bee species directly from the interaction data; i.e. we can use nutrition as an explanatory variable of plant-pollinator interaction data. By collecting pollen directly from pollen visitors, we can verify if nutritional quality is consistent among different individuals of the same species. And again, using pollen metabarcoding techniques (28, 31), we can create interaction networks based on species composition of the pollen collections and determine if bees accomplish nutritional balancing across a wide or narrow group of plant species.

Finally, understanding the nutritional ecology of bees provides direct and practical information for the design of habitat restoration and resource provisioning for bees. Making sure to consider diversity for nutritional balancing and phenological matching, we can restore plant species that provide the optimal pollen nutritional quality for targeted bee species (32). This would already provide a more informed protocol than those whose goal is to maximize multiple bee species

“attraction” based on only interaction data(33, 34). Instead we would expect that bee populations of the targeted species should respond positively over time because their nutritional needs would be met. When attempting to provide resources for diverse bee communities in large scale restoration projects, we can not necessarily study all bee species nutritional needs. However, we clearly must provide a diversity of plant species to provide sufficient resource abundance.

However, instead of simple floral diversity, such as selecting multiple plant species from the same genus or family, we can select plant species from an array of resource nutritional values to provide nutritional diversity. Although logistically demanding, we can conduct large scale projects in which we measure the nutritional quality of floral resources of commercially grown plant species,

159

preferably those that are native and locally adapted to the region they will be utilized. Therefore, from this database, we can select and create a nutritionally balanced environment in which to test bee population response, and home our target of restoring diverse and stable plant-pollinator communities.

160

References

1. Ruedenauer FA, Spaethe J, Leonhardt SD (2016) Hungry for quality—individual bumblebees forage flexibly to collect high-quality pollen. Behav Ecol Sociobiol:1–9.

2. Pernal S, Currie R (2000) Pollen quality of fresh and 1-year-old single pollen diets for worker honey bees (Apis mellifera L.). Apidologie 31(3):387–409.

3. Pereboom JJM (2000) The composition of larval food and the significance of exocrine secretions in the bumblebee Bombus terrestris. Insect Soc 47(1):11–20.

4. Jandt JM, Huang E, Dornhaus A (2009) Weak specialization of workers inside a bumble bee (Bombus impatiens) nest. Behav Ecol Sociobiol 63(12):1829–1836.

5. Dornhaus A, Chittka L (2001) Food alert in bumblebees (Bombus terrestris): possible mechanisms and evolutionary implications. Behav Ecol Sociobiol 50(6):570–576.

6. Dornhaus A, Chittka L (2004) Information flow and regulation of foraging activity in bumble bees ( Bombusspp.). Apidologie 35(2):183–192.

7. Kitaoka TK, Nieh JC (2008) Bumble bee pollen foraging regulation: role of pollen quality, storage levels, and odor. Behav Ecol Sociobiol 63(4):501–510.

8. Roulston TH, Cane JH (2000) Pollen nutritional content and digestibility for animals. Plant Syst Evol 222(1):187–209.

9. Roulston TH, Cane JH, Buchmann SL (2000) What governs protein content of pollen: pollinator preferences, pollen-pistil interactions, or phylogeny? Ecol Monogr 70(4):617– 643.

10. Dobson HEM (1987) Role of flower and pollen aromas in host-plant recognition by solitary bees. Oecologia 72(4):618–623.

11. Dobson HEM, Bergström G (2000) The ecology and evolution of pollen odors. Plant Syst Evol 222(1-4):63–87.

12. Pacini E, Hesse M (2005) Pollenkitt–its composition, forms and functions. Flora- Morphology 200(5):399–415.

13. Piskorski R, Kroder S, Dorn S (2011) Can pollen headspace volatiles and pollenkitt lipids serve as reliable chemical cues for bee pollinators? Chem Biodivers 8(4):577–586.

14. Dobson H (1988) Survey of pollen and pollenkitt lipids--chemical cues to flower visitors? Am J Bot 75(2):170.

161

15. Flamini G, Cioni PL, Morelli I (2003) Use of solid-phase micro-extraction as a sampling technique in the determination of volatiles emitted by flowers, isolated flower parts and pollen. J Chromatogr A 998(1-2):229–233.

16. Ruedenauer FA, Spaethe J, Leonhardt SD (2015) How to know which food is good for you: bumblebees use taste to discriminate between different concentrations of food differing in nutrient content. J Exp Biol 218(Pt 14):2233–2240.

17. Robertson HM, Wanner KW (2006) The chemoreceptor superfamily in the honey bee, Apis mellifera: Expansion of the odorant, but not gustatory, receptor family. Genome Res 16(11):1395–1403.

18. Abisgold JD, Simpson SJ (1988) The Effect of Dietary Protein Levels and Haemolymph Composition on the Sensitivity of the Maxillary Palp Chemoreceptors of Locusts. J Exp Biol 135(1):215–229.

19. Jensen K, Mayntz D, Toft S, Raubenheimer D, Simpson SJ (2011) Nutrient regulation in a predator, the wolf spider Pardosa prativaga. Anim Behav 81(5):993–999.

20. Raubenheimer D, Mayntz D, Simpson SJ, Tøft S (2007) Nutrient-specific compensation following diapause in a predator: implications for intraguild predation. Ecology 88(10):2598–2608.

21. Mayntz D, Raubenheimer D, Salomon M, Toft S, Simpson SJ (2005) Nutrient-specific foraging in invertebrate predators. Science 307(5706):111–113.

22. Vanderplanck M, et al. (2014) How Does Pollen Chemistry Impact Development and Feeding Behaviour of Polylectic Bees? PLoS ONE 9(1):e86209.

23. Tasei J-N, Aupinel P (2008) Nutritive value of 15 single pollens and pollen mixes tested on larvae produced by bumblebee workers (Bombus terrestris, Hymenoptera: Apidae). Apidologie 39(4):397–409.

24. Génissel A, Aupinel P, Bressac C, Tasei JN, Chevrier C (2002) Influence of pollen origin on performance of Bombus terrestris micro-colonies. Entomol Exper Applic 104(2-3):329– 336.

25. Williams NM, Regetz J, Kremen C (2012) Landscape-scale resources promote colony growth but not reproductive performance of bumble bees. Ecology 93(5):1049–1058.

26. Keller A, et al. (2015) Evaluating multiplexed next-generation sequencing as a method in palynology for mixed pollen samples. Plant Biol 17(2):558–566.

27. Sickel W, et al. (2015) Increased efficiency in identifying mixed pollen samples by meta- barcoding with a dual-indexing approach. BMC Ecol 15(1):20.

28. Richardson RT, et al. (2015) Application of ITS2 Metabarcoding to Determine the Provenance of Pollen Collected by Honey Bees in an Agroecosystem. Appl Plant Sci 3(1)

162

doi: 10.3732/apps.1400066

29. Behmer ST, Joern A (2008) Coexisting generalist herbivores occupy unique nutritional feeding niches. Proc Natl Acad Sci USA 105(6):1977–1982.

30. Brosi BJ, Briggs HM (2013) Single pollinator species losses reduce floral fidelity and plant reproductive function. Proc Natl Acad Sci USA 110(32):13044–13048.

31. Richardson RT, et al. (2015) Rank-Based Characterization of Pollen Assemblages Collected by Honey Bees Using a Multi-Locus Metabarcoding Approach. Appl Plant Sci 3(11):1500043–9.

32. Vaudo AD, Tooker JF, Grozinger CM, Patch HM (2015) Bee nutrition and floral resource restoration. Curr Opin Insect Sci 10:133–141.

33. Russo L, DeBarros N, Yang S, Shea K, Mortensen D (2013) Supporting crop pollinators with floral resources: network-based phenological matching. Ecol Evol 3(9):3125–3140.

34. M'Gonigle LK, Williams NM, Lonsdorf E, Kremen C (2016) A tool for selecting plants when restoring habitat for pollinators. Conserv Lett doi:10.1111/conl.12261.

163

Anthony Damiano Vaudo Vita

Throughout my Master’s and PhD career, I have focused on how landscape, foraging behavior, and nutrition affects honey bee and bumble bee colony heath. For my Master’s degree, in South Africa, I studied wild honey bee population density, nest site selection, and colony strength and linked these parameters different habitat types. These studies led to two published manuscripts (J Insect Conservation and Insectes Sociaux) and a methods section in the COLOSS Beebook. For my PhD research at Penn State, I have conducted detailed projects in landscape, semi-field, and lab conditions to understand bumble bee (Bombus impatiens) pollen foraging preferences for host- plant species and pollen nutrition. My work has lead to an important and novel finding that bumble bees do have specific pollen nutritional preferences and these are the interactions of both the protein and lipid content of pollen.

My PhD program so far has led to one published review article on bee nutrition and conservation (Chapter 1: Current Opinion in Insect Science), three published research manuscripts (Chapters 2- 4: Arthropod-Plant Interactions, Proceedings of the National Academy of Sciences, and Journal of Experimental Biology), and a manuscript currently in preparation for submission (Chapter 5). In addition to the publication of my research, I have disseminated this knowledge through both scientific and extension/outreach presentations, including ESA conferences (in which I won the student competition in 2014 and am an invited speaker in 2016), courses (Penn State’s “Honey Bees and Humans”), and other organizations (Master Gardeners and Wyoming Bee College). I have mentored five undergraduate students, including two winners of Dutch Gold Scholarships (Victoria Bolden and Liam Farrell) and one Apes Valentes Undergraduate Research Award (Liam Farrell). My research has been supported largely through self-funded grants and fellowships (AFRI NIFA Predoctoral Fellowship, NAPPC Bee Improvement Project Grant) and included a funded collaboration with Dr. Geraldine Wright from Newcastle University (BBSRC). The merit of this dissertation has been recognized recently; I have received a departmental award for outstanding achievement (Ralph O. Mumma Graduate Award).

After graduation, my interest is to understand how nutritional ecology applies to plant-pollinator communities in natural and/or undisturbed habitats both domestically and in the tropics. My next line of research will involve modeling community networks by incorporating phenology, chemical communication, and nutrition as explanatory variables, and how these interactions can be applied for conservation efforts.