We propose that the response of ecosystems to global change, in particular increasing C02, will depend mainly on the extent to which plant species adjust to new conditions by plasticity or evolutionary change, which depends on genetic variation. We show that the possibility of extrapolating results of short-term experiments to long-term prediction depends on the relative importance of these two processes. Moreover, if genetic variation exists within species in the response to elevated C02, it will not be possible to predict evolutionary changes from heritability estimates made under present atmospheric conditions.

In a field experiment we grew genotypes of two closely related grassland perennials without competition and with competition at three levels of biodiversity under ambient and elevated levels of C02. The generalist P. vulgaris was genetically more variable but not more plastic than the specialist P. grandiflora and was at a relative disadvantage under elevated C02. Thus, the hypothesis that specialists should be replaced by generalists when the environment changes was not supported in the short term in this exemplary case study. Long-term predictions about the response to environmental change will, however, be difficult according to our result because P. grandiflora showed significant genetic variation in its response to increased C02, indicating that its long-term response cannot be predicted from heritability estimates made under present conditions. Furthermore, an effect of biodiversity on the growth of the two Prunella species was only visible at elevated C02, indicating that biodiversity may become more important as the environment changes.

The existence of large genetic variation within populations in ecologically relevant characters demonstrates that species cannot be considered as constants in ecosystem studies and global change models. Characters measured at the level of the individual may therefore be more promising indicators of ecosystem change than higher level characters.

0 0

Post a comment