Predator and floral traits change pollinator behaviour, with consequences for plant fitness

Abstract

1. Flower-dwelling predators may interfere in plant-pollinator interactions through changes in pollinator behaviour, leading to a reduction in pollination services. Although the context-dependency of tri-trophic interactions is often acknowledged, the relative contribution of predator, pollinator and flower traits in shaping pollinator behaviour has been rarely studied simultaneously.2. Pollinators' responses to predation risk were evaluated using artificial spiders on Alstroemeria aurea flowers in order to experimentally test whether predator traits (colour and size) and nectar availability affect their behaviour and plant fitness.3. Regarding the predator's traits, both sizes of artificial spiders used were similarly rejected, but spider colour was detected differently. In particular, red and black artificial spiders were more strongly rejected than the rest. In turn, nectar availability increased the time spent in the inflorescences, yet most pollinators rejected flowers with artificial spiders independently of nectar availability. Furthermore, responses to artificial spider colour and nectar availability varied significantly between dipterans and hymenopterans. Finally, artificial spiders reduced seed set and fruit weight of plants owing to changes in the behaviour of the most efficient pollinators.4. It was shown that poorly studied predator traits such as colour and size may affect pollinator behaviour in different ways among taxa. In addition, feeding necessity in a rich-resource environment was a weaker selecting force than predation risk. Hence, it is argued here that trade-offs between predation and flower reward may arise depending on predator detection, relationship between pollinator and predator size, and resource availability in the ecosystem with consequences for plant fitness.

Intraspecific variation in body size of bumblebee workers influences anti‐predator behaviour

“…The larger spider models were made to determine whether pollinators perceived and respond to larger objects as a greater predation risk, since preliminary field observations suggested that real spiders were often not detected by pollinators. However, this is possibly due to spider hunting strategies such as background colour matching (i.e., body colour similar to the background colour) or behavioural strategies such as remaining still and hiding within flowers, either between petals or below stamens (Gavini et al, ). Yet, bumblebee corpses are found in the field, indicating that they are subject to some level of predation risk.…”

“…Pollinator traits such as body size, taxon and sociality are crucial traits for explaining the differences in pollinator behaviour in the presence of predators (reviewed by Romero et al, ). In turn, pollinator behaviour in response to predation risk is also known to vary due to a combination of traits among all interacting partners (Gavini, Quintero, & Tadey, ; Llandres, Mas, & Rodríguez‐Gironés, ). In particular, specific traits of predators such as colour, size and hunting mode (Bray & Nieh, ; Kersch‐Becker, Grisolia, Campos, & Romero, ; Llandres & Rodríguez‐Gironés, ; Miller, Ament, & Schmitz, ; Wang, Ings, Proulx, & Chittka, ) or plant traits such as reward availability (Abbot, ; Jones, ; Jones & Dornhaus, ; Llandres et al, ) can influence pollinator response to predation risk.…”

“…However, predation risk can also influence these optimal foraging choices (Higginson & Gilbert, ), potentially resulting in patch avoidance, increasing flower inspection times, decreasing flower visitation rates and, ultimately, decreasing pollinator foraging efficiency (Ings & Chittka, ). Therefore, a trade‐off arises for pollinators choosing between maximizing resource intake and minimizing predation risk (Gavini et al, ; Jones, ; Jones & Dornhaus, ; Llandres et al, ).…”

“…The use of artificial predators in natural conditions allows us to simplify and evaluate the influence of different traits across trophic levels (spider-bumblebee-flower) on pollinator behaviour (Gavini et al, 2018;Gonçalves-Souza, Omena, Souza, & Romero, 2008). This system is ideal, since B. terrestris is currently the most abundant and effective pollinator of A. aurea (Morales, Arbetman, Cameron, & Aizen, 2013), and in our study site, several sit-andwait spiders (e.g., Thomisidae and Anyphaenidae) were observed hunting on A. aurea's flowers (~25% of the plants harbour a spider) (Gavini et al, 2018), preying on a wide range of pollinators, including B. terrestris. Although bumblebees are large-sized social bees, the body size of B. terrestris workers may display up to 10fold variation in body mass (Goulson, 2003;Goulson et al, 2002), comparable to variation at the interspecific level.…”

“…increasing flower inspection times, decreasing flower visitation rates and, ultimately, decreasing pollinator foraging efficiency (Ings & Chittka, 2008). Therefore, a trade-off arises for pollinators choosing between maximizing resource intake and minimizing predation risk (Gavini et al, 2018;Jones, 2010;Jones & Dornhaus, 2011;Llandres et al, 2012).…”