A predicted developmental and evolutionary morphospace for grapevine leaves

Abstract

The shapes of leaves have been extensively measured and modeled. Using conventional statistical approaches there exist powerful methods to discern between groups and classify shapes. Embedded in calculated morphospaces is information that allows us to visualize theoretical leaves. These unmeasured leaves, that may or may not exist, are never considered in their relation to empirically observable leaves nor how the negative morphospace can inform us about the forces responsible for shaping leaf morphology. Using landmark data for 8465 grapevine leaves, we model leaf shape using an allometric indicator of leaf size, the ratio of vein to blade areas. The borders of the observable morphospace are restricted by the constraint of the subregions of the leaf reducing to nothing. Assuming that self-intersecting polygonal regions in leaves are a developmental constraint, we define a region of morphospace that represents leaf shapes that are possible but not observed. We end with a discussion of how we choose not to observe what seemingly does not exist and the ramifications of this perspective to survivorship bias in the sciences in general.

A high resolution model of the grapevine leaf morphospace predicts synthetic leaves

“…were sampled from the Teaching Vineyard at the Robert Mondavi Institute for Wine and Food Science at UC Davis. Original scans with scale for each leaf sampled from the UC Davis teaching vineyard are available (Chitwood, 2024). Leaves from the USDA Wolfskill, CA and Geneva, NY collections were measured by selecting leaves from previously published data (Chitwood, 2020;Chitwood et al, 2020).…”