The ability to grip on petal surfaces is of crucial importance for the interactions between bees and flowers. In this thesis, I explore the biomechanics of attachment and morphological diversity of bee attachment devices, linking this to the behavioural ecology of bee interactions with flowers.

Attachment devices come in two main kinds: claws or spines, and adhesive pads. Claw functioning is poorly described, particularly in terms of how their performance depends on body size, claw geometry, and surface roughness. Claw attachment performance was investigated using several insect species, each covering a large range of body masses. Weight-specific attachment forces decreased with body size, with claw sharpness seemingly playing a role. In bees there is considerable interspecific variation in tarsal claw morphology. This variation, and arolia presence/absence, was categorised for the large bee family Apidae. Cleft/bifid claws were shown to be present in the majority of the Apidae, often with differences between sexes and clades. Using $Bombus\; terrestris$, there was no evidence that cleft claws are important for pollen collection; however, I found that the inner tooth of cleft claws can act as a backup if the main tooth breaks. Although this may be one function of cleft claws, there are clearly other unresolved functions well worth further exploration.

Investigations were undertaken to explore how petal surface roughness affects bee foraging behaviour. Lab-based foraging trials on $B.\; terrestris$ visiting artificial flowers varying in slope, surface texture and sugar reward revealed a trade-off between the biomechanical difficulty of visiting and handling the ‘flowers’ and the quality of the reward offered. Flowers that were difficult to grip were often avoided even if they offered a higher reward. To further investigate reward preferences of bees, the effect of sucrose concentration on honey stomach offloading times was also explored.

Although the majority of petals do have a rough surface, some have slippery petals. In the field, bumblebees avoided landing on slippery hollyhock petals in favour of the easy-to-grip staminal column. In contrast, honey bees, which are smaller and have larger adhesive pads, landed on both the staminal column and the petals. Slippery petals may be an adaptation to increase contact with plant reproductive structures.

Grip is also important to allow the honey bee parasite $Varroa\; destructor$ to climb on to their host. Attachment forces experiments found that $V.\; destructor$ could support >300 times their body mass on honey bees, giving them strong attachment even when bees attempt to remove them through grooming. A grooming-based device for treating $V.\; destructor$ was tested in an apiary trial. The device was ineffective, providing valuable information for beekeepers considering using this product.

In summary, this thesis improves our understanding of the biomechanics of attachment as well as identifying several important aspects of grip in bee-plant interactions.