Body size is among the most important biological variables but despite much measurement of this trait, the factors driving its variation are not fully understood. Here, I describe variation in body size in the damselfly Calopteryx maculata to establish whether variations in growth and development observed in response to experimental manipulation of temperature and time stress in the laboratory can be scaled-up to variation among natural populations. Nine hundred and seven specimens of C. maculata males were collected from 34 sites across the species’ entire range in North America during the summer of 2010. A general measure of body size was derived from a series of wing and leg measurements. I compare the fit of models based on latitude (Bergmann’s rule), temperature (the temperature–size rule) and seasonal effects (a combination of temperature and time stress) using Akaike’s information criterion (AIC). The U-shaped relationship between size and latitude was best explained by a seasonality model containing both photoperiod and temperature. The presence of both these terms suggests that time stress dominates in the southern part of the range, reducing body size by accelerating development. However, the temperature–size rule dominates in the northern part of the range, increasing body size closer to the northern range margin. The best-fit model of geographic variation in size is in accordance with previous laboratory studies of temperature and photoperiod in damselflies and theoretical work, indicating that the findings from such studies can be applied to natural populations. These findings are likely applicable to any species with complex life histories inhabiting seasonal environments.