Testing adaptive hypotheses about how continuous traits evolve in association with developmentally-structured discrete traits, while accounting for the confounding influence of other, hidden, evolutionary forces, remains a challenge in evolutionary biology. In one example of this, geophytes are herbaceous plants capable of retreating underground and use underground storage organs (USOs) to survive extended periods of unfavorable conditions. Such plants have evolved multiple times independently across all major vascular plant lineages. Even within closely related lineages, however, geophytes show impressive variation in the morphological modifications and structures (i.e., “types” of USOs) that allow them to survive underground. Despite the developmental and structural complexity of USOs, the prevailing hypothesis is that they represent convergent evolutionary “solutions” to a common ecological problem, though some recent research has drawn this into question. We extend existing phylogenetic comparative methods to test for links between geophytes’ hierarchical discrete morphological traits associated with USOs and adaptation to environmental variables, using a phylogeny of 621 species in Liliales. We found that plants with different USO type do not differ in climatic niche more than expected by chance, with the exception of root morphology, where modified roots are associated with lower temperature seasonality. These findings suggest that root tubers may reflect adaptations to different climatic conditions than those represented by other types of USOs. Thus, the tissue type and developmental origin of the USO structure may influence the way it mediates ecological relationships, which draws into question the appropriateness of ascribing broad ecological patterns uniformly across geophytic taxa. This work provides a new framework for testing adaptive hypotheses and for linking ecological patterns across morphologically varying taxa while accounting for developmental (non-independent) relationships in morphological data.

All data files can be opened with an open source text editor such as Sublime Text. For alignment files (.nexus) we recommend an open source alignment viewer such as AliView. CSV files may be opened using the R programming language or visualised using a spreadsheet program such as google sheets.