Comparative analyses of flower scent in Sileneae reveal a contrasting phylogenetic signal between night and day emissions

Abstract

Abstract Flower scents are complex blends of volatile compounds often shaped by selection pressures exerted by mutualistic and antagonistic interaction partners, but also by phylogenetic constraints. So far, little is known about the relative effect of selection and phylogenetic signal on scent patterns, and no study to date analyzed the phylogenetic signal in multivariate semiquantitative scent patterns. We analyzed the phylogenetic signal in qualitative and semi‐quantitative patterns of flower scents in 47 Sileneae (Caryophyllaceae) species using phylogenetic principal component analysis (pPCA) and several indices of univariate and multivariate phylogenetic signal. As previous results showed that Sileneae species are visited by diurnal and nocturnal pollinators and flower scents also vary along the day, we compared the phylogenetic signal between night and day. Multivariate pPCA analyses identified compounds that correlate with the phylogeny at both night and day; however, multivariate Bloomberg′s K detected phylogenetic signal in the dataset of night scents, but not of day scents. In multivariate qualitative datasets, phylogenetic signal was neither found for day nor for night scents. In univariate analyses, phylogenetic signal was detected for some compounds both for day and night scents. Overall, we found that the phylogenetic signal is stronger in night compared to day scents, which might be owed to the different guilds of pollinators at day and night. At day, the phylogenetic resemblance of Sileneae scents might be masked by or disappear due to divergent selective pressures exerted by a diverse guild of pollinators on the different species. In contrast, we hypothesize that the nocturnal moth pollinators exert similar selective pressures; thus, the phylogenetic similarity of scent profiles might be conserved. Future studies of scent phylogenetic signal must consider not only the usage of qualitative measures but also semiquantitative analyses.

Keywords

• Abouheif's Cmean
• D statistic
• diel scent variation
• floral scent evolution
• multivariate Bloomberg′s K
• Pagel's λ