Plants produce a great number of phytochemical compounds mediating a variety of different functions. Recently, phytochemical diversity (chemodiversity), a way which to quantify the complex phenotype formed by sets of phytochemicals, has been suggested to be important for function. However, no study has systematically examined the potential (in)direct functional importance of chemodiversity on a general level, partly due to a lack of an agreement on how to quantify this aspect of the plant phenotype. This paper has four aims: 1) We discuss how chemodiversity (deconstructed into components of richness, evenness and disparity) may quantify different aspects of the phenotype that are ecologically relevant. 2) We systematically review the literature on chemodiversity to examine methodological practices, explore ecological patterns of variability in diversity across different levels of biological organization, and investigate the functional role of this diversity in interactions between plants and other organisms. 3) We provide a framework facilitating decisions on which measure of chemodiversity is best used in different contexts. 4) We outline open questions and avenues for future research in this area. A more thorough understanding of phytochemical diversity will increase our knowledge on the functional role phytochemical compounds, and how they shape ecological interactions between plants and their environment.

Plants harbour a great chemodiversity, i.e., diversity of specialized metabolites (SMs), at different scales. For instance, individuals can produce a large number of SMs and populations can differ in their metabolite composition. Given the ecological and economic importance of plant chemodiversity, it is important to understand how it arises and is maintained over evolutionary time. For other dimensions of biodiversity, i.e., species diversity and genetic diversity, quantitative models play an important role in addressing such questions. Here we provide a synthesis of existing hypotheses and quantitative models, i.e. mathematical models and computer simulations, for the evolution of plant chemodiversity. We describe each model’s ingredients, i.e., the biological processes that shape chemodiversity, the scales it considers, and whether it has been formalized as a quantitative model. Although we identify several quantitative models, not all are dynamic and many influential models have remained verbal. To fill these gaps, we identify quantitative models used for genetic variation that may be adapted for chemodiversity. We end by outlining our vision for the future of chemodiversity modeling, presenting a flexible framework for the creation of individual-based models that address different scales of chemodiversity and combine different ingredients that bring this chemodiversity about.

Plants harbour an astonishing amount of chemodiversity, i.e., diversity of specialized metabolites, at different scales. For instance, individual plants can produce a large number of different specialized metabolites and individuals in a population can differ in their metabolite composition. Given the ecological and economic importance of plant chemodiversity, it is important to understand how it arises and is maintained over evolutionary time. For other types of biodiversity, i.e., species diversity and genetic diversity, quantitative models, that is, mathematical models and computer simulations, have long played an important role in addressing such questions. Here we review models and hypotheses for the evolution of plant chemodiversity and, in particular, explore what quantitative models have been proposed so far and what gaps there are in quantitative modeling of chemodiversity. For each model or hypothesis we review its ingredients, i.e., the biological processes that are assumed to shape chemodiversity, the scales at which the model explains or claims to explain chemodiversity, and the extent to which the model has been formalized as a mathematical or simulation model. From this review, a mixed picture emerges. We identified a small number of quantitative models for the evolutionary dynamics of plant chemodiversity. In addition we found a number of models that use equations to derive an optimal defense, but are not dynamic. Many influential models, however, have remained verbal so far. Furthermore, we identify some quantitative models used for genetic variation that have not been used for chemodiversity so far, but could be easily extended to do so. We end by outlining our vision for future model building for the evolution of plant chemodiversity.