Evolutionary trends in leaf vein architecture across spatial scales

Link To Share This Poster: https://cdmcd.co/mXy84K
Live Discussion Link: https://cdmcd.co/QM8kAn

Hailey J. Park and Haley Grimmer, University of California Berkeley, Berkeley, CA, Natalie Vuong, University of Waterloo, Waterloo, ON, Canada, Ilaine S. Matos and Benjamin Blonder, ESPM, University of California Berkeley, Berkeley, CA

Background/Question/Methods
Leaf venation architecture is highly variable among plant species (e.g., single, branching or reticulate veins) and across spatial scales (e.g., lower-order/major veins or higher-order/minor veins). From an evolutionary perspective, some aspects of network architecture may be more labile than others. We test the hypothesis that some very large- and very small-scale patterns in venation reflect deep phylogenetic niche conservatism and/or limited functional variation, while intermediate-scale patterns may evolve rapidly and reflect adaptive responses. Due to the difficulty of collecting network architectural data for whole-leaves, prior studies have mostly focused on categorical descriptors of spatial scale (higher vs. lower vein orders), which has limited our ability to investigate evolutionary trends in leaf venation. In this study, we use machine learning algorithms and innovative statistical approaches to fully describe multiscale venation network properties (elongation ratio - ER, circularity ratio - CR, vein density - VD, and minimum spanning tree ratios - MST) of more than 160 whole-leaves from a evolutionarily and geographically diverse set of over 155 species (73 families). For each venation trait, we tested for detectable trends in phylogenetic conservatism across different network spatial scales.
Results/Conclusions
At larger vein sizes, there was a notable shift away from long loops with reduced infolding to finer and shorter loops with more tree-like branching, indicating a functional trade-off. Principal component analysis revealed that the four vein multiscale traits could be reduced to three axes of variation with VD and MST explaining the most variance. Phylogenetic conservatism across spatial scales was variable but mostly low, suggesting fast trait evolution and high lability at all vein sizes. Therefore, contrary to popular plant systematics literature, there appears to be no signs of slower trait evolution as we approach larger vein sizes at lower orders. Altogether, our results show that leaf venation architecture comprises labile responses across multiple spatial scales. Future work may identify the adaptive evolutionary processes driving this variation.