Data are available in five Appendices that are presented in Word files. APPENDIX S1.Figure S1. Crossing scheme for construction of scent-manipulated genotypes (CNL1/TPS2, CNL1/tps2, cnl1/TPS2 and cnl1/tps2). APPENDIX S2. Table S1. Primers for scent-manipulated individual genotyping and haplotype amplification. APPENDIX S3. Table S2. Peak area of BAld and β-ocimene in 3-hour floral scent and air headspace sampling. The amount of BAld and β-ocimene emitted by the flowers in flower scent samples (net emission) was calculated by subtracting the respective volatile peak area in the air samples collected on the same collection device on the same collecting date from the volatile peak area in the floral scent headspace sample. The emission rates are calculated as their net emissions divided by the number of flowers enclosed in the headspace collection chamber and the sampling time (3 hours). See more details in “Headspace floral scent collection and analysis”. APPENDIX S4. Table S3. Number of visits by different pollinator group to flowers of the four scent-manipulated genotypes and SC Capsella grandiflora in forty 15-min observation periods. APPENDIX S5. Figure S2. Haplotype-frequency dot plots for each amplicon. Dots with the same color between different individuals indicate the same haplotype sharing. CgCg: CNL1/TPS2, CgCr: cnl1/TPS2, CrCg: CNL1/tps2, CrCr: cnl1/tps2 and SC: SC C. grandiflora. Numbers after genotype names indicate their individual and sampling plot number, for example, CgCg2_1 refers the individual No. 1 of CNL1/TPS2 genotype in sampling plot No. 2. Individuals with number 48/418 refer to the ones whose leaf materials were harvested to provide parental haplotype information.