Résumé
Experimental disturbance and fertilization in two Mediterranean old fields, three and nine years old, respectively, were used to identify functional groups for response to disturbance. Five morphological traits (canopy structure, height, lateral spread, habit and plasticity) and five regeneration traits (life cycle, seed mass, fecundity, dispersal mode and germination seasonality) were used for species classification. Correlation patterns of species composition and species attributes with disturbance treatments were analysed in order to characterize groups of response to disturbance. The classification based on morphological traits was repeatable across fields and reflected natural correlation patterns among attributes. Erect rosettes with low morphological plasticity and moderate lateral spread, mostly grasses, were intolerant of disturbance. Partial rosettes with low morphological plasticity and wide lateral spread, e.g. Asteraceae species, colonized disturbed locations. Flat proto-rosettes with a plastic architecture were indifferent to disturbance in the young plot but required disturbance to establish in the older plot. The classification based on regeneration traits repeatedly identified germination period to be correlated with disturbance response. Species with early germination were intolerant of disturbance while late-germinating species colonized disturbances. These groups were clearly distinct from groups based on natural attribute correlation patterns which related primarily to seed mass, and secondarily to dispersal mode and fecundity. Effects of fertilization were detected only within disturbed quadrats of the old plot. Fertilization favoured the colonization of disturbances by species with a partial rosette and low plasticity and by species with late germination and high fecundity. These results, complemented by direct analyses on individual traits, are mostly consistent with previous descriptions of the ruderal strategy. This study shows that additional understanding can be gained from a deductive approach that examines more specific traits such as detailed descriptors of plant architecture and seasonal germination dynamics. Our results are intended to contribute to the general discussion on the identification and use of functional classifications. In particular, it is argued that communities such as Mediterranean old fields, where the dynamics is already well understood, should serve as testing grounds for new methods and the development of theories to be later applied to less well known vegetation types.