Plant trait responses to soil moisture and grazing: are dominants and subordinates different?

CINGOLANI, A.M.; CABIDO, M.; GURVICH, D.E.; RENISON, D.; DÍAZ, S.

JOURNAL OF VEGETATION SCIENCE

WILEY-BLACKWELL PUBLISHING, INC

Año: 2007 vol. 18 p. 911 - 920

1100-9233

(1) In the assembly of plant communities, are plant traits filtered differently for species presence and for species abundance? (1) In the assembly of plant communities, are plant traits filtered differently for subordinate and dominant species? (1) Is the response of plant traits to environment at the community level similar when considering their abundance (trait weighted-average) than when not considering it? (trait simple average) (2) How does the magnitude of the weighted average differ from the magnitude of the simple average along environmental gradients? Location: Mountain grasslands, Central Argentina. Methods: We used data from 57 floristic samples ordinated through DCCA along moisture and grazing gradients, combined with trait values from 85 species (SLA, leaf toughness, leaf area, leaf thickness and plant height). For each sample and trait we calculated the simple and weighted averages of component species trait-values, as representatives of the probability of species of being present and of becoming dominant in a community, respectively. We analyzed the relationship between both gradients (DCCA scores) as independent variables and simple and weighted average trait values as dependent variables, through multiple regressions. Results: Weighted averages of trait values per sample were significantly associated to both moisture and grazing in all traits, while simple averages not always responded. When responding, in some cases the responses of weighted averages followed similar but stronger trends than the responses of simple averages, but in other cases these responses were qualitatively different. Response curves of traits to environment showed that differences between both averages were not constant along the gradients, and in some cases were inverted.Conclusion: The trait values and combinations which determine the probability of presence of a species in a community are not necessary the same than those which determine their dominance. Thus, when interpreting the filtering process and predicting assembly rules both filtering levels should be taken into account as different, although closely linked processes.