Physiological novelty in the evolution of flowers

Background/Question/Methods

Since the early work of Kölreuter (1761) and Sprengel (1793) over 200 years ago, animal pollinators have long been thought to be the primary agents of selection responsible for variation in floral traits.  Case studies of narrowly defined clades (populations or congeneric species) have been showcased as exemplars of the strong influence pollinators can have on floral morphology.  Furthermore, animal pollinators are one of the major drivers of plant diversification; animal-pollinated lineages are more speciose than abiotically pollinated lineages.  Because pollination is so critical to reproduction, selection may favor reducing any physiological constraints that might limit pollination or oppose pollinator preference.  Reducing the physiological costs of floral display structures would enable greater flower production and allow floral morphology to covary with pollinator preference.  As a result, lineages with cheaper flowers, we predict, would have greater rates of diversification.  To test this hypothesis, we combined measurements of hydraulic and structural traits from ~175 species from across the angiosperm phylogeny with intensive measurements of the hydraulic efficiency of flowers from a few representative lineages and used phylogenetic comparative methods to test different models of trait evolution.

Results/Conclusions

We found strong support for our primary hypothesis that reduced investment in hydraulic and structural traits of flowers was associated with accelerated diversification, particularly early in the angiosperm radiation.  Flowers of speciose lineages generally have lower area-normalized biomass and water contents, lower minimum cuticular conductance to water vapor, lower Huber ratios, and are smaller than species-poor lineages.  Yet there was still substantial variation in traits among species.  In contrast, whole flower hydraulic conductance, a measure of the capacity for water transport, is very low in flowers and shows remarkably little variation among the magnoliids, monocots, and eudicots, compared to the Austrobaileyales, which have significantly higher hydraulic conductance.  This suggests that during early angiosperm evolution flowers underwent strong selection for lower hydraulic capacity (higher efficiency), yet there were many possible trait combinations capable of producing similar hydraulic capacities.  We posit that reducing the physiological costs of flowers relaxed biophysical constraints of water transport, allowing floral form to vary more widely.  This may be the underlying mechanism linking animal pollination, floral morphology, and lineage diversification.