Using data from several published studies [38,65,83–85], we estimated the mean percent variance in genome-wide gene expression levels explained (PVE) by tissue composition, demographic effects (age and sex), local genetic variation, and a range of social and physical environmental inputs (grey wolf = mange and social status; rhesus macaque = social status; yellow baboon = maternal social connectedness; human = smoking). All studies were conducted in blood-derived samples, and mean PVE was taken from the text, supplementary information, or calculated using publicly available effect size estimates and data files. We note that PVE estimates are strongly influenced by the covariates included in models to detect environmental effects, and by the amount of variation in the environmental variable itself. However, across studies it is clear that environmental effects rival or exceed other widely accepted drivers of gene regulatory variation.