Does mother know best? Range-wide narrowing of host preference in Aricia agestis confers fitness benefits but may incur long-term costs

Biological communities are responding to changes in climate through a combination of range shifts and evolution in situ. However, even in systems where we have genomic and phenotypic evidence of recent adaptation, the ecological and fitness consequences of such evolutionary change remain underexplored, especially in terms of consequences for interactions between species. European Lepidoptera are excellent systems for understanding such shifts, given our exceptional knowledge of their ecology, natural history, and past distributions. The UK Brown Argus butterfly (Aricia agestis) has expanded rapidly northward in response to climate change. This expansion has been associated with evolutionary shifts in maternal host preference, from the chalkland restricted but locally common perennial Rockrose (Helianthemum nummularium) to more broadly distributed, but locally rare annual Geraniums (Geraniaceae). Such a rapid range shift offers an excellent opportunity to test how the evolution of novel biotic interactions associated with climate adaptation affects individual fitness and population resilience. Using common garden assays of host preference on females sampled across the UK in 2013 and 2023, we show for the first time that female preference for Geranium molle or Rockrose hosts vary among individuals within as well as among sites and to different extents at different locations. This variation is consistent with genomic signals of recent selective sweeps on these traits. In particular, we show that in 2013, females in the new part of the range are more likely to oviposit on the locally dominant host, while females from the ancestral range will use both hosts, regardless of the locally dominant host. We also demonstrate a significant temporal shift in behaviour, with a narrowing of host preference towards Geranium molle recorded in A.agestis ancestral range in 2023. To evaluate the fitness consequences of these shifts in biotic interactions, we assessed larval performance of 1412 larvae from 49 families on both host plants. Irrespective of maternal preference, feeding on Geranium molle hosts conferred higher survival, faster development and larger adult body size under laboratory conditions. These fitness outcomes are likely linked to Geranium molles significantly higher protein: carbohydrate ratio when compared to Rockrose. Moreover, we also found that larvae fed Geranium molle produced adults with elevated mass-independent resting metabolic rates, a rare demonstration that larval diet has lasting metabolic consequences in a wild butterfly. Crucially, we found no evidence of trade-offs between preference and performance: even offspring of females that laid exclusively on Rockrose also performed better on Geranium molle. These results suggest that Geranium molle confers a general performance advantage, enhancing traits likely to promote population persistence where Geranium molle is reliably available. However, despite selection favouring Geranium molle in warming climates, we also show that a narrowing of host preference towards these more widespread host plants throughout the UK range is likely to make A.agestis populations more vulnerable under future warmer, drier conditions, when compared to Rockrose. Together these findings highlight a central challenge in evolutionary biology: what happens when traits that are adaptive in the short term spread rapidly through populations, but are likely to become maladaptive as environmental conditions continue to change?