Cell and chloroplast anatomical features are poorly estimated from 2D cross-sections.

Leaf function is intimately related to the size, shape, abundance and position of cells and chloroplasts. Anatomy has long been assessed and quantified in two dimensions with 3D structure inferred from 2D micrographs. Serial block face scanning electron microscopy (SBF-SEM) was used to reconstruct 95 cells and 1173 chloroplasts from three wheat and nine chickpea leaves (three samples each from three chickpea genotypes). Wheat chloroplast volume was underestimated by 61% in mesophyll cells and 45% in bundle sheath cells from 2D micrographs, whereas chickpea mesophyll chloroplast volume was underestimated by 60% using simple geometrical models. Models of chickpea spongy and palisade cells both under- and overestimated surface area and volume by varying degrees. These models did not adequately capture irregular shapes such as flattening of chloroplasts or lobed spongy mesophyll cells. It is concluded that simple geometrical models to estimate chloroplast and cell 3D volume and surface area from 2D micrographs are inadequate, and that SBF-SEM has strong potential to contribute to improved understanding of leaf form and function.