Over the past 15 years, the question of how and why animals integrate multimodal stimuli has received considerable attention from animal behaviorists. Even outside our field, piecing together the puzzle of multisensory integration has garnered the attention of neuroscientists interested in the mechanisms of perception working with both humans and model organisms such as Drosophila. Against this backdrop, a century of work on the sensory worlds of bees has led to their emergence as a highly tractable system for the study of multimodal integration from both proximate and ultimate perspectives. Not only are their visual and olfactory worlds relatively well-characterized, but we can also straightforwardly manipulate features of bees' developmental environment relevant to sensory processing. Likewise, by framing research questions in the context of a mutualism, we can assess the fitness consequences of sensory complexity from the perspectives of both sender (plant) and receiver (pollinator). Using bees as a case study, we adapt a classic ethological framework to answer the question of what makes multimodal signals unique from ontogenetic, mechanistic, functional, and (co-)evolutionary perspectives. Highlighting a series of findings from our lab-based experiments on bumblebees (Bombus impatiens) using both free-flying behavioral assays and classical conditioning of the proboscis extension response (PER), we explore how multimodality enhances the efficacy of signal transmission across multiple temporal and spatial scales. Building from recent work showing that complex interactions between scent and color guide discrimination learning in a nectar-foraging context, we outline future prospects relating to developing a multisensory perceptual space and untangling the role of multimodality in mediating the difficulty of floral choice.