Estimation of stomatal density of leaves with hierarchical reticulate venation

ABSTRACT Stomatal density (SD) is important to photosynthetic rates. However, it is time-consuming to measure SD. Here, we provide a method for estimating SD based on the scaling relationship between SD and mean nearest neighbour distance (MNND) of sampled stomatal centres. 397 leaves from eight Magnoliaceae species were used for this study. For each leaf, three 1.2 mm × 0.9 mm lamina sections, positioned equidistantly between leaf left margin and midrib, were examined (in total 1189 sections). SD and MNND were calculated for each section. Regression protocols were used to test for a negative SD vs. MNND scaling relationship at the species and family levels. Additionally, 10 to 200 stomata from each section were randomly sampled to check for the prediction accuracy of SD using the SD vs MNND scaling relationship. There were significant differences in SD among the different lamina positions for 7 of 8 species. The inverse SD vs MNND scaling relationship was validated at the species and family levels. For the pooled data, the MNND values using 14, 25 and 50 stomata accounted for >80%, 85% and 90% of the variance in SD on a log-log scale, respectively. SD was characterized by high interspecific variability, and within-leaf variability, decreasing from the position near the midrib to that near the leaf margin. SD scaled inversely with MNND for the eight species. Thus, using the rapidly estimated trait MNND significantly simplifies and expedites the estimation of SD.