Appendix 1. The pattern of voltinism of Chiasmia clathrata in Finland.

Temporal distribution of C. clathrata observations recorded in the Finnish Entomological Database ( in the appropriate biogeographical provinces in relation to sampling sites (solid red circles) representing either univoltine (Obb) or bivoltine (Ab / N) regions. Observations from the southern provinces N (n = 220) and Ab (n = 205) cover the period 2009–2012, whereas all recorded observations from the northern province Obb were included (1974–2011) to obtain a reasonable sample size (n = 22).

Appendix 2. Schematic view of the full-factorial split-brood design of the experiment with notes on the developmental pathways (direct vs. diapause) induced.

Appendix 3. Family-specific thermal reaction norms for induction of the direct developmental pathway at high (left column) and low (right column) larval densities in bivoltine (upper row) and univoltine (lower row) Chiasmia clathrata populations. The number of families per population is 12 in each case. The values at 20°C have been adjusted by 0.02 units in the case of totally overlapping reaction norms except for the univoltine populations at high density, where none of the individuals developed directly into adults. Note the different scales on the y-axes for bivoltine and univoltine populations.

Appendix 4. Family-specific thermal reaction norms of larval development time (left column), standardized growth rate (middle column) and pupal mass (right column) in bivoltine (Halikko and Somero) and univoltine (Tornio and Pello) Chiasmia clathrata populations. For simplicity, only individuals that entered the diapause pathway are taken into account.

Appendix 5. Family-specific reaction norms for larval development time (upper row), standardized growth rate (middle row) and attained pupal mass (lower row) in relation to developmental pathway at 20°C in bivoltine (left column) and univoltine (right column) Chiasmia clathrata populations.

Appendix 6. Fixed effects of the final LMMs fitted by REML explaining variation in larval growth rate in Chiasmia clathrata calculated with varying values of b (Tammaru and Esperk 2007). For parameter estimates of the final models, we derived mean values and 95 % HPD credibility intervals with the MCMC method by using the function MCMCglmm (Hadfield 2012).

Trait / Parameter / Linear mixed-effect model / MCMC simulation
Estimate / SE / t / Posterior mean / Lower bound / Upper bound
Growth rate (b = 0.75) / Intercept / -0.936 / 0.028 / -33.9 / -0.943 / -1.01 / -0.867
Temperature (20°C) / 1.636 / 0.067 / 24.31 / 1.653 / 1.518 / 1.762
Region (bivoltine) / -0.004 / 0.041 / -0.1 / -0.008 / -0.112 / 0.099
Pathway (direct) / -0.002 / 0.146 / -0.02 / 0.035 / -0.29 / 0.324
Temperature (20°C) × region (bivoltine) / 0.098 / 0.113 / 0.87 / 0.077 / -0.091 / 0.323
Temperature (20°C) × pathway (direct) / 0.028 / 0.174 / 0.16 / -0.071 / -0.358 / 0.366
Region (bivoltine) × pathway (direct) / 0.002 / 0.16 / 0.01 / -0.04 / -0.367 / 0.301
Temperature (20°C) × region (bivoltine) × pathway (direct) / 0.399 / 0.201 / 1.99 / 0.422 / 0.03 / 0.861
Growth rate (b = 0.85) / Intercept / -0.854 / 0.039 / -21.664 / -0.866 / -0.958 / -0.756
Temperature (20°C) / 1.389 / 0.096 / 14.532 / 1.4 / 1.218 / 1.574
Region (bivoltine) / -0.001 / 0.059 / -0.013 / -0.003 / -0.136 / 0.151
Pathway (direct) / 2.91 × 10-4 / 0.208 / -0.001 / 0.06 / -0.398 / 0.493
Temperature (20°C) × region (bivoltine) / 0.036 / 0.161 / 0.225 / 0.056 / -0.268 / 0.34
Temperature (20°C) × pathway (direct) / 0.032 / 0.248 / 0.131 / -0.001 / -0.507 / 0.541
Region (bivoltine) × pathway (direct) / 2.17 × 10-4 / 0.229 / 0.001 / -0.04 / -0.525 / 0.455
Temperature (20°C) × region (bivoltine) × pathway (direct) / 0.685 / 0.286 / 2.392 / 0.722 / 0.074 / 1.298

Random effects of the models included random intercepts for populations and families combined with family-specific random slopes for temperature.

Tammaru T, Esperk T (2007) Growth allometry in immature insects: larvae do not grow exponentially. Funct Ecol 21:1099–1105

Hadfield J (2012) MCMCglmm: MCMC generalised linear mixed models. R Package, Version 2.17. Available at

Appendix 7. Family-specific reaction norms of attained pupal mass (upper row) and larval development time (lower row) in relation to larval density at 20°C in bivoltine (left column) and univoltine (right column) Chiasmia clathrata populations. Solid black markers and whiskers denote region level means and their 95 % confidence intervals, respectively. For simplicity, only individuals that entered the diapause pathway are taken into account.

Appendix 8. Visualization of unbiased coefficients of variation (ĉv*) (Sokal and Rohlf 1995) for phenotypic life history variation among genotypes under diapause and direct development at 20°C in bivoltine (dashed red lines) and univoltine (dashed blue lines) Chiasmia clathrata populations. Solid red and blue lines represent region-level values.

Sokal RR, Rohlf FJ (1995) Biometry. Freeman, New York