Interactive effects of elevated CO2, P availability and legume presence oncalcareous grassland: results of a glasshouse experiment

1, We investigated the interactive effects of elevated CO2, supply of phosp horus (P) and legume presence in model communities of calcareous grassland. Half of the communities contained six graminoids and eight nan-legume forb species, In the other half, four non-legume forbs were replaced by legumes . 2, Ecosystem responses. Above-ground phytomass (>5 cm) hardly responded to elevated CO2 alone. However, when P and legumes were combined. the CO2 effe ct on above-ground community phytomass in year two was a stimulation of 45% (P < 0.001). Below-ground community dry matter was stimulated by elevated CO2 alone by + 36% (P < 0.01), but was only + 20% (P < 0.05) when legumes w ere present and P was added, At the final (late season) harvest the mean ef fect of elevated CO2 on total above- and below-ground phytomass was + 23% ( P < 0.001) and revealed no significant interactions among treatment combina tions, because above- and belowground effects of CO2 enrichment had opposit e directions. 3, Functional group responses. When legumes were absent, graminoids increas ed their total above- and below-ground phytomass in elevated CO2 by 60% (P < 0.001) but there was no increase when legumes were present. The response of forbs to CO2 was nor significant, irrespective of co-treatment. Legumes. however, were significantly stimulated by P supply and their CO2 response was much larger when P was added (+ 55%, P < 0.01 vs + 25%, NS). 4, Species responses. CO2 effects on species ranged from highly positive ( 143%) to moderately negative (- 43%). 5, Our results demonstrate that the effect of CO2 enrichment in such natura l grassland communities will be low on above-ground phytomass and largely b elow-ground if no additional nutrients are provided. N-2-fixing legumes app ear to be crucial for the community response to elevated CO2 but legume res ponsiveness is largely controlled by P availability.