Fine-scale spatial variation in eastern grey squirrel melanism across a complex urban landscape.

Urban environments impose novel selective pressures that can drive rapid phenotypic and genetic change in wild populations. While previous studies have focused on broad urban-rural contrasts or cross-city comparisons, less is known about how phenotypic traits vary across fine spatial scales within cities, where environmental gradients and geographic features may create localized heterogeneity. We studied spatial variation in coat colour morph of the eastern grey squirrel (Sciurus carolinensis) across the Montreal Metropolitan Area, focusing on grey and melanic morphs, the latter of which is caused by an incompletely dominant MC1R mutation. We conducted standardized scan sampling surveys across 197 parks and supplemented this data with citizen science observations from iNaturalist and SquirrelMapper. As an archipelago segmented by major waterways, Montreal exhibits a spatial mosaic where phenotypic gradients are partitioned by physical barriers. We examined variation in melanism in relation to urbanization, local environmental conditions (e.g. temperature and canopy cover) and geographic structure. Unlike patterns in many cities where melanism peaks in urban cores, melanism in Montreal was generally negatively correlated with urbanization, but the strength and direction of urban-rural clines varied sharply across the city's complex island landscape. Abrupt shifts in morph frequency corresponded with major waterways, including the Saint-Lawrence River and its tributaries, suggesting that rivers might act as partial barriers to dispersal which partition the landscape into distinct phenotypic regions. Consistent with the thermoregulatory advantages of melanism, melanic morph frequency increased most in areas with cooler winter temperatures and higher canopy cover. Structural equation modelling revealed that urbanization indirectly suppresses melanism by altering these two variables. Our results suggest that phenotypic patterns within cities are not uniform but are instead shaped by fine-scale environmental heterogeneity and geographic barriers, highlighting the need to move beyond simple urban-rural contrasts when predicting phenotypic variation in urban populations.