1 INTRODUCTION

Wild bees play a prominent role in global crop production (Reilly et al., 2020; Winfree et al., 2008), supplementing or even exceeding the contributions of the domesticated honeybee, Apis mellifera (Boyle et al., 2020; Garibaldi et al., 2019; Koh et al., 2018; McKerchar et al., 2020). In North America, wild bees are as commonly found on crop flowers as honeybees (Reilly et al., 2020), and they can be more effective pollinators of some fruit crops (Garibaldi et al., 2013; MacInnis & Forrest, 2019; Pérez-Méndez et al., 2020). However, although wild pollinators support crop production, they are also heavily impacted by agricultural practices (Potts et al., 2016). The conversion of natural habitat for human use often leads to decreased availability of floral resources, as well as, for many species, reductions in nesting resources (Dicks et al., 2021; Roulston & Goodell, 2011). The latter are an often overlooked yet essential component of wild bee habitat (Belsky, & Joshi, 2019; Roulston & Goodell, 2011). Because high floral visitor densities are the greatest contributor to the yields of pollinator-dependent crops, declines in pollinator abundance and diversity have been associated with reduced crop yield in agricultural landscapes (Potts et al., 2016), showcasing the delicate balance between agricultural productivity and pollinator conservation.

Identifying the resources that limit bee populations is essential for maintaining pollination services and for bee conservation. Floral resources (particularly in agroecosystems) have long been the primary focus in studies of resource limitation in bees (Bloom et al., 2019; Ebeling et al., 2008; Fornoff et al., 2017; Ghazoul, 2004; Theodorou et al., 2020). This focus on floral resources has left the role of nesting resources in regulating bee populations largely unexplored (Roulston & Goodell, 2011). Several studies have shown that the abundance and richness of bees are correlated, to some degree, with the availability of nesting resources (i.e. bare ground, dead wood, soil burrows and cavities) in the local environment (Bartholomée et al., 2020; Caudill et al., 2017; Rodríguez & Kouki, 2017; Westerfelt et al., 2018). However, the existence of nesting resource limitation has so far never been tested in a controlled, replicated experiment, even though there is reason to suspect that bee populations could be nest-site limited in certain situations. In particular, bees that nest in pre-existing above-ground tunnels in wood or hollow stems (henceforth ‘cavity-nesting bees’) tend to be rare in agroecosystems compared to their ground-nesting counterparts (Bartholomée et al., 2020; Bihaly et al., 2020; Forrest et al., 2015; Williams et al., 2010), raising the question whether availability of nesting cavities in such habitats limits their abundance.

Apple (Malus domestica Borkh) is a common insect-pollinated crop in temperate regions of the world including eastern Canada (Ontario Apple Growers, 2019). It is typically self-incompatible and often pollen limited (Osterman et al., 2021; Reilly et al., 2020). Because apple blooms in spring, it depends for pollination on spring-emerging bees like honeybees, bumblebee queens (Bombus spp.), and wild solitary bees such as Andrena spp. and Osmia spp. (Carvell et al., 2015; Delaplane & Mayer, 2004; Földesi et al., 2016; Márquez et al., 1994; Martins et al., 2014; Park et al., 2016; Pérez-Méndez et al., 2020; Sekita, 2001). Some wild bees, including Osmia spp., can have synergistic effects on orchard fruit yield by altering the foraging behaviour of honeybees, causing them to move more frequently between crop rows, which often comprise different crop cultivars. The increased between-row movement increases transfer of compatible pollen, fruit set and, ultimately, crop yield (Brittain et al., 2013; Martins et al., 2014). In North America, Osmia lignaria (the ‘blue orchard bee’) in particular distinguishes itself from honeybees by its ability to fly at cooler temperatures (Boyle & Pitts-Singer, 2019; Isaacs et al., 2017) and deposit larger pollen loads on a wide range of flowers (Földesi et al., 2016). Unlike other apple pollinators, many Osmia spp. are cavity nesters; thus, they may be less abundant in apple orchards than they would be if orchards provided nesting cavities (i.e. reeds, hollow twigs and dead wood).

The primary goal of this study was to understand the degree to which cavity-nesting bee populations are limited by nesting resources in orchards. We conducted a controlled, replicated field experiment to test, first, whether cavity-nesting bees (Osmia) in apple orchards are limited by nesting habitat such that the addition of nesting structures increases Osmia abundance in orchards. If cavity-nesting bee populations increase due to the addition of nesting structures, populations of other pollinators could decrease in response, or they might alter their activity patterns in ways that increase flower visitation and pollination. We therefore also examined whether the addition of nesting structures affects overall pollinator visitation to flowers in apple orchards. Finally, we tested whether the addition of nesting structures affects apple yield and quality.