Heliconiini butterflies can learn time-dependent reward associations

Abstract

4For many pollinators, flowers provide predictable temporal schedules of resource availability, 1 5 meaning an ability to learn time-dependent information could be widely beneficial. However, 1 6 this ability has only been demonstrated in a handful of species. Observational studies of 1 7Heliconius butterflies suggest that they may have an ability to form time-dependent foraging 1 8 preferences. Heliconius are unique among butterflies in actively collecting and digesting pollen, 1 9 a dietary behaviour linked to spatiotemporally faithful 'trap-line' foraging. Time-dependency of 2 0foraging preferences is hypothesised to allow Heliconius to exploit temporal predictability in 2 1 alternative pollen resources, as well as contributing to optimal use of learnt foraging routes. 2Here, we provide the first experimental evidence in support of this hypothesis, demonstrating 2 3 that Heliconius hecale can learn opposing colour preferences in two time periods. This shift in 2 4preference is robust to the order of presentation, suggesting that preference is tied to the time of 2 5 day and not due to ordinal learning. However, we also show that this ability is not limited to 2 6Heliconius, as previously hypothesised, but is also present in a related genus of non-pollen 2 7 feeding butterflies. This demonstrates that time learning pre-dates the origin of pollen-feeding 2 8and may be prevalent across butterflies with less specialized foraging behaviours. 2 9 3 0 2

Wing-pattern-specific effects of experience on mating behavior inHeliconius melpomenebutterflies

“…Heliconius butterflies have many of the characteristics often found in species where past experience informs future social behavior; thus, our finding that H. melpomene males modify their mating behavior in response to experience, while novel, may not be unexpected. H. melpomene butterflies are relatively long-lived (up to 6 months in nature) (Gilbert, 1972), highly social (they roost in groups at night) (Mallet and Gilbert, 1995), and learn food sources and color cues (Toure et al, 2020). They have large brains (Montgomery et al, 2016) and are both physically larger, and longer lived than the butterfly Bicyclus anynana , which also uses past experience to inform current mating behavior (Dion et al, 2020; Westerman et al, 2012; Westerman et al, 2014).…”

What is the state space of the world for real animals?

“…Similarly, for many animals, cues reflecting the end of winter are predictive of an increased availability of food reward. It is then perhaps not surprising that even insects show evidence of such learning (Chittka et al, 1995; Dylla et al, 2013; Menzel, 2012; Toure et al, 2020; Wystrach et al, 2019a). We will first discuss the common mathematical formulation for representing state space in this task, before discussing implicit assumptions and their problematic implications.…”