Leaf architecture and functional traits for 122 species at the University of California at Berkeley botanical garden

The data set contains leaf venation architecture and functional traits for a phylogenetically diverse set of 122 plant species (including ferns, basal angiosperms, monocots, basal eudicots, asterids, and rosids) collected from the living collections of the University of California Botanical Garden at Berkeley (37.87 °N, 122.23 °W; CA, USA) from February to September, 2021. The sampled species originated from all continents, except Antarctica, and are distributed in different growth forms (aquatic, herb, climbing, tree, shrub). The functional dataset comprises 31 (mechanical, hydraulic, anatomical, physiological, economical, and chemical) traits and describes six main leaf functional axes (hydraulic conductance, resistance and resilience to damages caused by drought and herbivory, resilience to damages, mechanical support, and construction cost), as well as how architecture features vary across venation networks. Our trait dataset is suitable for (1) functional and architectural characterization of plant species; (2) identification of venation architecture-function trade-offs; (3) investigation of evolutionary trends in leaf venation networks, and (4) mechanistic modeling of leaf function.  Data is made available under the Open Data Commons Attribution License. You are free to copy, distribute, and use the database; to produce works from the database; and to modify, transform, and build upon the database. You must attribute any public use of the database, or works produced from the database, in the manner specified in the license. For any use or redistribution of the database, or works produced from it, you must make clear to others the license of the database and keep intact any notices on the original database.

Methods
Leaf venation architecture and functional traits were measured on 122 species of ferns and angiosperms species collected from the UCBG living collections. The sampled species are distributed among 112 families and 56 orders, and include species in different plant clades (ferns, basal angiosperms (including ANITA and magnoliids), monocots, basal eudicots, rosids, and asterids), with different growth forms (aquatic, herb, climbing, tree, shrub) and from all continents, except Antarctica. Our sampling approach aimed to maximize both the phylogenetic coverage and the evolutionary distinctiveness.

For each of the 122 species, we either sampled branches (> 1 meter long branch) from a single individual per species in the case of woody species, or from 1-5 whole individuals in the case of herbaceous species. Samples were collected in the morning (9-11 AM), brought to a UCBG research greenhouse, re-cut under water, re-hydrated overnight, and then used for the measurement of leaf architecture and functional traits. As most of the leaf traits measured in this study are destructive, we used different leaves for the measurement of each trait.

To describe how efficiently water flows through the leaf venation network, we measure the maximum leaf hydraulic conductance (Kleaf max, mmol m-2 s-1 MPa-1). Kleaf max measurements were performed on 3-10 fully re-hydrated (leaf water potential > -0.5 MPa) and mature leaves per species (N = 122 species) using the evaporative flux method with a pressure-drop flow meter. In addition to Kleafmax, we also measured maximum assimilation rate (Amax, µmol m-2 s-1) and maximum stomatal conductance rate to water vapor (gsmax, mmol m-2 s-1) as proxies for flow efficiency. Amax and gsmax were measured on 3-4 intact leaves per species (N = 34 species) using a portable photosynthesis analyzer (LI-COR 6800, LI-COR, Lincoln, NE, USA). Measurements were conducted between 10 AM and 3 PM during sunny and cool days in November 2021. To obtain the maximum values of photosynthesis and stomatal conductance, we set the LI-COR environmental parameters as follow: 1,200 µmol m-2 s-1 photosynthetic photon flux density, 25 °C leaf temperature, 0.7 kPa vapor pressure deficit, and 412 mmol mol-1 CO2 concentration. Measurements were initiated by selecting mature and sunny exposed leaves, which were then sealed in the LI-COR leaf chamber. 

To describe the network ability to avoid physical damage caused by embolisms (i.e. formation and propagation of air bubbles inside the xylem vessels) we measured the leaf water potentials inducing 50% (P50, MPa) and 88% loss (P88, MPa) of leaf hydraulic conductance. To obtain P50 and P88 values, we constructed vulnerability curves using the EFM method. To describe the xylem resistance to implosion (i.e. resistance to collapse of xylem cell walls under negative pressure), we measured the cell implosion safety index (ISI, dimensionless) from leaf cross-sectional anatomical images of all 122 species. To describe the network ability to avoid physical damage caused by herbivores, we measured punch strength (PS, kN m-2), punch specific strength (PSS, kN m-2 m-1), work to punch (WP, J m-2), specific work to punch (SWP, kJ m-2 m-1), shear strength (SS, MN m-2), work to shear (WS, J m-1 ), and specific work to shear (SWS, kJ m-2). While the shear and punch strength describes the maximum stress at which the leaf cracks (due to the shearing or punching forces), the specific work to shear and specific work to punch, indicates the amount of work done, or energy required, to shear or to punch a leaf per unit leaf thickness. To quantify species capacity to resist herbivory via chemical defenses (i.e. via production of secondary metabolites) we also quantified the total phenols (Phe) in dried leaves using the Folin–Ciocalteu (F–C) method. 

To describe the network's ability to maintain flow after physical damages have occurred, we measured the change in Kleaf (ΔKleaf) after simulated herbivory. To simulate herbivory, we performed two different treatments (midrib and lamina) by either cutting the leaf midrib (and keeping the lamina intact) or by cutting the leaf lamina. 48 hours after the damage, we excised the damaged leaves and measured Kleaf using the EFM method.

To describe venation network capacity to support the leaf upright against gravity, wind and other bending forces, we measured the leaf flexural stiffness (Σ, mN m-2 m-1) and the leaf modulus of elasticity (ε, MN m-2) by conducting 3-point bending tests using the UTM machine (Test stand ES30, force gauge M5-20, bending fixtures G1095-97, Mark-10, Copiague, NY, USA).

To describe the total amount of resources invested in constructing the leaf, we measured leaf mass per area (LMA, g m-2), specific leaf area (SLA, m-2 kg-1), leaf dry mass content (LDMC, mg g-1) following standard protocols.

To obtain leaf venation architecture traits, we first pressed fresh leaves until flat and gently removed trichomes and spores (if present) under a dissecting microscope.  Next, we placed the pressed and cleaned leaf samples inside a labeled mesh envelope and subjected them to standard chemical processes of clearing and staining. The cleared leaf samples were imaged using a 100 mm macro objective lens (Tokina, Huntington Beach, CA, USA) and color camera (EOS 6D, Canon, Southend-on-Sea, UK). The leaf cleared images were analyzed in Matlab (Mathworks, Natick, MA, USA) using LeafVeinCNN software versions 1.0.7 and newer. LeafVeinCNN relies on an ensemble of three convolutional neural networks (CNNs) to automatically segment veins and produce a spatial graph representation of the entire venation network. After segmenting and producing the spatial graphs of the entire leaf venation network, LeafVeinCNN calculates several single-scale venation architecture traits to describe the architectural characteristics of all veins (edges), nodes (point of connection between two or more veins), areoles or loops (regions wholly enclosed by veins, or enclosed by veins and leaf boundary), and polygonal areas (area inside the areole) in the network. LeafVeinCNN also uses hierarchical loop decomposition (HLD) algorithms to extract multiscale venation statistics, which describe how the venation network architecture vary across vein spatial scales (i.e. across vein sizes).