Conflicts of interests between males and females over reproduction is a universal feature of sexually reproducing organisms and has driven the evolution of intersexual mimicry, mating behaviors and reproductive polymorphisms. Here we show how temperature drives pre-reproductive selection in a female colour polymorphic insect that is subject to strong sexual conflict. This species has three female colour morphs, one of which is a male mimic. This polymorphism is maintained by frequency-dependent sexual conflict caused by male mating harassment. The frequency of female morphs varies geographically, with higher frequency of the male mimic at higher latitudes. We show that differential temperature sensitivity of the female morphs and faster sexual maturation of the male mimic increases the frequency of this morph in the north. These results suggest that sexual conflict during the adult stage is shaped by abiotic factors and frequency-independent pre-reproductive selection that operate earlier during ontogeny of these female morphs.

This dataset has been collected in the field in various countries in Europe and in Sweden, where also mesocosm experiments were carried out, as well as pre-reproductive selection on female morphs was estimated in a series of populations outside Lund, province of Scania (Skåne) as described in the Metods-section of our paper. Our dataset contains both rawdata-files and code (R-scripts) that should make it possible to re-run our statistical analyses and re-generate our graphs.

This data package contains six files:

1. DryadScriptBiogeographicVariation: a file containing code (R-script) needed to re-generate Fig. 2A-C, Tables S1-S2 and Fig. S2,  and all associated test-statistic in combination with the datafile below. The code is annotated, explaining in a step-by-step the analyses and the meaning of the different variables, referring to each figure or table when it is generated.

2. Europe128populations.csv: a datafile in csv-format that should be read in to the script above (1), and which contains data on female colour morph frequencies from 128 populations across Europe, their locations (countries) and latitudinal positions and bioclimatic variables (Bioclim-variables). 

3. DryadScriptFieldSelectionCoefficients: a file containing code (R-script) needed to re-generate Fig. 4A-4B  and all associated test-statistic (posthoc-tests) in combination with the datafile below. The code is annotated, explaining in a step-by-step the analyses and the meaning of the different variables, referring to each figure when it is generated. We estimate relative fitnesses and selection coefficients on the three female morphs using a Bayesian approach, as described in the Methods-section of our paper.

4. FieldSelection.csv: a datafile in csv-format that should be read in to the script above (3), and which contains data on the frequencies of the three different female morphs in two different age classes (immature and sexually mature), from which we estimated cross-sectional relative fitnesses and selection coefficients on pre-reproductive survival.

5. DryadScriptMesocosmExperiments: a file containing code (R-script) needed to re-generate Fig. 3A-2B and Tables 1-2  and all associated test-statistic (posthoc-tests) in combination with the datafile below. This will re-generate all the analyses from the mesocosm experiments, i. e. morph differences in development time and probability of surviving to maturity in relation to morph, temperature and the interaction between these variables (Table 2). The code is annotated, explaining in a step-by-step the analyses and the meaning of the different variables, referring to each figure when it is generated.

6. MesocosmExperimentsData.csv: a datafile in csv-format that should be read in to the script above (5), and which contains data on female development time in relation to colour morph and probability of reaching maturity in relation to temperature experienced during development and three days prior to capture (reflecting acclimation effects). Rationale and meaning of the different variables are explained in the Methods-section and in the annotated script (5) above.