1. In this paper we develop a theoretical framework for describing the composition and morphology of leaves on a volumetric basis which is inclusive of both the liquids and internal air space. That framework is then used to link the composition and morphology with the function of leaves.
2. Leaves are segmented into a functional scheme named a–u–s–V (air space–solution–structure–Volume basis) and a measurement scheme named a–q–d–M (air space–liquid–dry matter–Mass basis). Measurements of mass, liquid mass, volume and fractional air space are necessary to (partially) link the two schemes.
3. The liquid content of leaves can be used to infer the mode of mechanical support and, hence, the likely values of air space, solution and structure. For example, if a leaf has a high liquid content, then it cannot have a large volumetric fraction of structure (which is largely made of dry matter) and is unlikely to have large internal air spaces.
4. Based on measurements of individual leaf components, the specific gravity of the non-gaseous fraction in leaves would usually be in the range 1–1·3. Consequently, the specific gravity of leaves would be most sensitive to variation in fractional air space.
5. For the dimensions typical of most leaves, the surface area to volume ratio of a leaf is determined by leaf thickness (or diameter).
6. Leaf design is considered in terms of light interception and gas exchange. The requirements for each are essentially incompatible and lead to a trade-off on leaf thickness. However, that relationship is complicated because variation in fractional air space can potentially lead to broad optima for gas exchange at any given leaf thickness.
7. The theoretical predictions have not been assessed owing to a lack of measurements of leaf density, fractional air space and liquid content. However, published data were used to develop the following hypotheses: (1) the mass of nitrogen per unit mass of liquid is relatively constant within leaves; (2) the surface area to volume ratio of leaves is proportional to leaf liquid content. These hypotheses are tested in a subsequent paper.