Supporting Information
Additional supporting information can be found online at https://doi.org/.
Descriptions
Data S1. An excel file with data sheets: 1 (Climate): code of the third mesh (Japan standardized regional mesh code: JIS X0410), where populations are located, and climatic variables (climatic value mesh data, Japan Meteorological Agency 1987) in each mesh, 2 (LeafSize): measurements of three leaves of each individual, 3 (ESTSSR): nuclear EST-SSR genotypes of 693 individuals in 24 populations, 4 (HWEchisq): estimated frequency of null alleles in each locus and P value of χ2 tests for deviation from the Hardy-Weinberg equilibrium in each locus and each population, 5 (Diversity): locations and genetic diversity indices in populations, 6 (Pairwise): PairwiseF ST values (upper diagonal) and Nei’s genetic distances (D A; lower diagonal) between populations. 7 (StructureL): Log-likelihood values of runs with 1–5 clusters in STRUCTURE (Pritchard et al . 2000) analysis, 8 (StructureQ): ancestry proportions of 2–4 clusters in STRUCTURE analysis in each individual.
Figure S1. Three hypotheses (a–b) explaining genetic divergence among populations of sub-alpine (A) and mountain (M) ecotypes in northern (1) and southern (2) mountain ranges.
Figure S2. Chloroplast DNA haplotypes discriminated by substitutions at four nucleotide sites in rp L32–trn L(UAG) region. Nucleotide variation (a) and haplotype network (b) are shown.
Figure S3. Allelic richness of 32 alleles (a, b) and expected heterozygosity (c, d) of populations along latitudinal (a, c) and elevational (b, d) gradients. Colors of circles indicate taxonomic and genetic categories (red: Qc , blue: Qch , and green: populations identified as Qch in the field observation but grouped to Qc in genetic analysis).
Figure S4. Log-likelihood values of replications with 1–5 clusters in STRUCTURE analysis.