Evaluating trait variation and covariation in perennial, herbaceous crop candidate species using herbarium specimens

Abstract

Contemporary agriculture in many parts of the world consists of annual plant species grown in monoculture, but recently interest has focused on developing perennial, herbaceous species for use in perennial polycultures that offer greater ecosystem services relative to annual monocultures. Because few perennial, herbaceous species were domesticated by early farmers, one of the initial steps in identifying wild, perennial, herbaceous species for potential crop development is understanding patterns of variation and covariation among reproductive traits that underpin yield and vegetative traits. Herbaria can serve as a valuable resource for taxonomically broad surveys of phenotypic variation and covariation for pre-breeding. In this study, we used herbarium specimens to investigate trait variation and covariation in five genera of the legume family (Fabaceae). We addressed two questions: (1) do vegetative and reproductive traits vary as a function of life span (annual vs. perennial species), and are these differences consistent across genera?; and (2) how do vegetative and reproductive traits correlate with one another across lifespan and/or within genera? Reproductive traits and vegetative traits were measured on 535 herbarium specimens representing annual and perennial species of Astragalus, Lupinus, Phaseolus, Strophostyles, and Vigna. While species exhibited intra- and interspecific variation, few morphological differences were observed among annual and perennial species. Correlations among vegetative traits and among reproductive traits were mostly positive when significant, and perennial species generally exhibited more positive correlations between vegetative and reproductive traits than annual species. These data have potential relevance for perennial crop development, and provide further support that herbaria represent a rich, relatively untapped resource of information about plant trait variation and covariation.