In microeconomics, a standard framework is used for determining the optimal input mix for a two-input production process. Here we adapt this framework for understanding the way plants use water and nitrogen (N) in photosynthesis. The least-cost input mixture for generating a given output depends on the relative cost of procuring and using nitrogen versus water. This way of considering the issue integrates concepts such as water-use efficiency and photosynthetic nitrogen-use efficiency into the more inclusive objective of optimizing the input mix for a given situation. We explore the implications of deploying alternative combinations of leaf nitrogen concentration and stomatal conductance to water, focusing on comparing hypothetical species occurring in low- versus high-humidity habitats. We then present data from sites in both the United States and Australia and show that low-rainfall species operate with substantially higher leaf N concentration per unit leaf area. The extra protein reflected in higher leaf N concentration is associated with a greater drawdown of internal CO2, such that low-rainfall species achieve higher photosynthetic rates at a given stomatal conductance. This restraint of transpirational water use apparently counterbalances the multiple costs of deploying high-nitrogen leaves.