Variation in insect size is often related to temperature during development, and may affect the persistence of populations under future climate warming if smaller individuals have reduced fitness. Montane species are particularly vulnerable to climate-driven local extinctions due to range retractions at their warm range margins, and so we examined spatial and temporal variation in body size in the butterfly Erebia epiphron in the UK, where it is restricted to two montane regions in England and Scotland. We examine spatial and temporal variation in body size in relation to temperature.

We sampled 19 populations (6-15 individuals per population) in England and Scotland between 2018 and 2019 spanning elevations from 380-720 m, and examined museum specimens collected between 1890 and 1980. We examined individual body size (forewing length) and its relationship with the local temperature of sites, as well as temporal variation in body size over the last century in relation to the temperature during larval development.

The forewing lengths of field-collected individuals in England were on average 7-8% smaller than in Scotland (England, mean = 14.9 mm, Scotland, mean = 15.9 mm), and warmer sites also had smaller individuals (0.13mm reduction in wing length per 1^oC increase in local site mean temperature). However, we found no effect of temporal temperature variation on body size changes during larval development.

E. epiphron were smaller in England than Scotland, and at warm range edge populations, which could have impacts on fecundity and dispersal ability. Future work should seek to understand the life-cycle lengths, genetics and phenotypic plasticity of these two populations to evaluate potential explanations for regional differences.

In 2018 and 2019, we collected 6-15 male E. epiphron from each of 9 populations in England and 10 populations in Scotland representing a wide range of local elevation (380-780 m above sea level) and temperature gradients (5-7.6^oC mean annual temperature). Wings were removed from individuals and electronically scanned.

Photographs of specimens held at the Natural History Museum, London (2014) were downloaded (https://data.nhm.ac.uk/), and up to 5 males collected between 1890 and 1987 were measured (England n = 127 individuals, Scotland n = 100).

We measured forewing length (distance between cell base and v10 wing margin veins; see Supplementary Materials SM1), a widely accepted proxy for body size, using the ‘draw line’ tool in Image J (https://imagej.nih.gov/ij/index.html). The length in pixels was converted into millimetres using the scale bar. Only wings where all veins and cell bases were visible were used for measurements. For each individual, both left and right forewings were measured and averaged. A measurement accuracy of ±0.3mm was estimated based on three replicate measurements of 10 random individuals.