Intraspecific variation of specialized metabolites in plants, such as terpenoids, are used to determine chemotypes. Tansy (Tanacetum vulgare L.) exhibits diverse terpenoid profiles that affect insect communities. However, it is not fully known whether patterns of their chemical composition and associated insects vary beyond the community scale. Here, we investigated the geographic distribution of mono- and sesquiterpenoid chemotypes in tansy leaves and their relationships with specific insect communities across Germany. We sampled tansy leaves from ten plants with and five plants without aphids in each of 26 sites along a north-south and west-east transect in Germany. Hexane-extracted metabolites from leaf tissues were analyzed by gas chromatography-mass spectrometry (GC-MS). Plant morphological traits, aphid occurrence and abundance, and occurrence of ants were recorded locally. The effect of plant chemotype, plant morphological parameters, and abiotic site parameters such as soil types, temperature and precipitation on insect occurrences were analyzed. Plants clustered into four monoterpenoid and four sesquiterpenoid chemotype classes. Monoterpene classes differed in their latitudinal distribution, whereas sesquiterpenes were more evenly distributed across the transect. Aphid and ant occurrence was influenced by monoterpenoids. Plants of monoterpenoid class 1 were colonized by aphids and ants significantly more often than expected by chance, whereas in other classes there were no significant differences. Aphid abundance was affected by soil type, and average annual temperature positively correlated with the occurrence of ants. We found significant geographic patterns in the distribution of tansy chemodiversity and show that monoterpenoids affect aphid and ant occurrence, while the soil type can influence aphid abundance. We show that geographic variation in plant chemistry influences insect community assembly on tansy plants.

This data was curated by collecting tansy (Tanacetum vulgare) plants from 26 sites along northwest-southeast transects in Germany. During the survey, GPS coordinates, abiotic site parameters, plants with and without aphids, ant species, and plant morphological data were recorded. Plants were then processed for chemical analysis with n-hexane extraction protocol and GC-MS analysis as described by Clancy et al., 2016. The plants were then clustered into chemotypes based on their mono- and sesquiterpenoid content by using "hclust" function in "factoextra" package in R. Different statistical analyses and tests were performed: The Discriminant Analysis of Principal Components (DAPC), generalized linear models, posthoc Tukey tests, and tanglegram of mono- and sesquiterpenoid chemotype trees. 

All the statistical analyses were conducted in R, using the "hclust", "vegan", "adegenet", "chemodiv", "emmeans" and "car" packages.