Thirty isofemale lines collected in three different years from the same wild French population were grown at seven different temperatures (12–31°C). Two linear measures, wing and thorax length, were taken on 10 females and 10 males of each line at each temperature, also enabling the calculation of the wing/thorax (W/T) ratio, a shape index related to wing loading. Genetic correlations were calculated using family means. The W–T correlation was independent of temperature and on average, 0.75. For each line, characteristic values of the temperature reaction norm were calculated, i.e. maximum value, temperature of maximum value and curvature. Significant negative correlations were found between curvature and maximum value or temperature of maximum value. Sexual dimorphism was analysed by considering either the correlation between sexes or the female/male ratio. Female–male correlation was on average 0.75 at the within line, within temperature level but increased up to 0.90 when all temperatures were averaged for each line. The female/male ratio was genetically variable among lines but without any temperature effect. For the female/male ratio, heritability (intraclass correlation) was about 0.20 and evolvability (genetic coefficient of variation) close to 1. Although significant, these values are much less than for the traits themselves. Phenotypic plasticity of sexual dimorphism revealed very similar reaction norms for wing and thorax length, i.e. a monotonically increasing sigmoid curve from about 1.11 up to 1.17. This shows that the males are more sensitive to a thermal increase than females. In contrast, the W/T ratio was almost identical in both sexes, with only a very slight temperature effect.