Summary Photosynthetic ‘least‐cost’ theory posits that the optimal trait combination for a given environment is that where the summed costs of photosynthetic water and nutrient acquisition/use are minimised. The effects of soil water and nutrient availability on photosynthesis should be stronger as climate‐related costs for both resources increase. Two independent datasets of photosynthetic traits, Globamax (1509 species, 288 sites) and Glob13C (3645 species, 594 sites), were used to quantify biophysical and biochemical limitations of photosynthesis and the key variable Ci/Ca (CO2 drawdown during photosynthesis). Climate and soil variables were associated with both datasets. The biochemical photosynthetic capacity was higher on alkaline soils. This effect was strongest at more arid sites, where water unit‐costs are presumably higher. Higher values of soil silt and depth increased Ci/Ca, likely by providing greater H2O supply, alleviating biophysical photosynthetic limitation when soil water is scarce. Climate is important in controlling the optimal balance of H2O and N costs for photosynthesis, but soil properties change these costs, both directly and indirectly. In total, soil properties modify the climate‐demand driven predictions of Ci/Ca by up to 30% at a global scale.

Least-cost theory" posits that C 3 plants should balance rates of photosynthetic water loss and carboxylation in relation to the relative acquisition and maintenance costs of resources required for these activities. Here we investigated the dependency of photosynthetic traits on climate and soil properties using a new Australia-wide trait dataset spanning 528 species from 67 sites. We tested the hypotheses that plants on relatively cold or dry sites, or on relatively more fertile sites, would typically operate at greater CO 2 drawdown (lower ratio of leaf internal to ambient CO 2 , C i :C a ) during light-saturated photosynthesis, and at higher leaf N per area (N area ) and higher carboxylation capacity (V cmax 25 ) for a given rate of stomatal conductance to water vapour, g sw . These results would be indicative of plants having relatively higher water costs than nutrient costs.In general, our hypotheses were supported. Soil total phosphorus (P) concentration and (more weakly) soil pH exerted positive effects on the N area -g sw and V cmax 25 -g sw slopes, and negative effects on C i :C a . The P effect strengthened when the effect of climate was removed via partial regression. We observed similar trends with increasing soil cation exchange capacity and clay content, which affect soil nutrient availability, and found

No abstract