Colour_UV_SignalWing phenotype data used in the paper "Temporal relationship between genetic and warning signal variation in the aposematic wood tiger moth (Parasemia plantaginis)" by Galarza et al. Shown are the reflectance values (hue) of Parasemia plantains male fore- and hindwings. The column headings indicate: fw col hue UV= Ultraviolet wavelength reflectance in the forewing fw col hue SW = Short wavelength reflectance in the forewing fw col hue MW = Middle wavelength reflectance in the forewing fw col hue LW = Long wavelength reflectance in the forewing fw blackratio = Forewing black ratio (in the manuscript referred to as signal size) hw col hue UV= Ultraviolet wavelength reflectance in the hindwing hw col hue SW = Short wavelength reflectance in the hindwing hw col hue MW = Middle wavelength reflectance in the hindwing hw col hue LW = Long wavelength reflectance in the hindwing hw blackratio = Hindwing black ratio (in the manuscript referred to as signal size)MtDNA_AlignmentMitochondrial (COI) alignment of Parasemia plantains used in the paper "Temporal relationship between genetic and warning signal variation in the aposematic wood tiger moth (Parasemia plantaginis)" by Galarza et al. The alignment is in MEGA format. Sample names are composed of the region, the colour morph (Y=yellow, W=white), the sex (M=male, F=Female), and the year of collection (09=2009, 10=2010, 11=2011).Microsat_GenotypesMicrosatellite data used in the paper "Temporal relationship between genetic and warning signal variation in the aposematic wood tiger moth (Parasemia plantaginis)" by Galarza et al. The first line indicates the name of the microsatellite, with alleles coded in a two-coloumn format. Missing data is represented by "-9". Each line represents one sample's multilocus genotype starting with the sample name. The name is composed of the region, the colour morph (Y=yellow, W=white), and the year of collection (09=2009, 10=2010, 11=2011)Wing_PatternForewing pattern grid data used in the paper "Temporal relationship between genetic and warning signal variation in the aposematic wood tiger moth (Parasemia plantaginis)" by Galarza et al. Each column represents one sample starting with the sample name. The title of each sheet indicates the location, the morph-type (Y=yellow, W=white) as well as the year of collection (9=2009, 10=2010, 11=2011).

Many plants and animals advertise unpalatability through warning signals in the form of colour and shape. Variation in warning signals within local populations is not expected because they are subject to directional selection. However, mounting evidence of warning signal variation within local populations suggests that other selective forces may be acting. Moreover, different selective pressures may act on the individual components of a warning signal. At present, we have a limited understanding about how multiple selection processes operate simultaneously on warning signal components, and even less about their temporal and spatial dynamics. Here, we examined temporal variation of several wing warning signal components (colour, UV reflectance, signal size, and pattern) of two co-occurring colour morphs of the aposematic wood tiger moth (Parasemia plantaginis). Sampling was done in four geographic regions over three consecutive years. We also evaluated each morph's temporal genetic structure by analysing mitochondrial sequence data and nuclear microsatellite markers. Our results revealed temporal differences between the morphs for most signal components measured. Moreover, variation occurred differently in the fore- and hindwings. We found no differences in the genetic structure between the morphs within years and regions, suggesting single local populations. However, local genetic structure fluctuated temporally. Negative correlations were found between variation produced by neutrally evolving genetic markers and those of the different signal components, indicating a non-neutral evolution for most warning signal components. Taken together, our results suggest that differential selection on warning signal components and fluctuating population structure can be one explanation for the maintenance of warning signal variation in this aposematic species.