Abstract
Abstract
Background and aims
Pollinator-mediated selection is a primary driver of floral trait evolution, yet how far phylogenetic history constrains this process remains poorly understood in aerial parasitic plants, whose reproduction depends entirely on animal mutualists. We studied floral divergence in four southern South American Loranthaceae mistletoes: Tristerix aphyllus, T. corymbosus, T. verticillatus, and Desmaria mutabilis, spanning three nested divergence times (Eocene–Pleistocene): an Eocene intergeneric split, a Miocene intrageneric divergence, and a Pleistocene sister-species speciation event.
Methods
We measured five floral traits (ovary length, tepal length, pistil length, stamen length, and herkogamy) in 78 individuals (15–22 per species) and quantified divergence using linear discriminant analysis with leave-one-out cross-validation, Mahalanobis distances, and effect sizes (η2).
Key results
All five traits differed among species overall, although not every species pair was separated by every trait, and the discriminant analysis assigned 82.1% of individuals correctly. Desmaria mutabilis was the most distinct species (Mahalanobis distances: 6.83–7.46 from all Tristerix species), consistent with its position as the intergeneric outgroup and with its pollination biology: it flowers in summer and receives far more insect than hummingbird visits. Tristerix verticillatus was intermediate, while Pleistocene sister species T. aphyllus and T. corymbosus, which share Sephanoides sephaniodes as pollinator despite radical ecological divergence, were nearly morphologically indistinguishable (d = 1.174). Ovary length was the most informative trait (η2 = 0.805) and herkogamy most clearly separated D. mutabilis from the ornithophilous Tristerix clade.
Conclusions
Floral similarity among the three Tristerix species reflects retention of a shared ornithophilous morphology rather than convergence, since they form a clade. Morphological distance is ordered as their divergence times are, a nested pattern consistent with the Andean orogeny-driven history of the group. Four species sampled at two sites cannot, however, separate divergence time from the differences that accumulate alongside it.