The effect of revegetation with different combinations of native species in the recovery dynamic of a shrubland dominated by Erica australis subjected to experimental fire was studied. The herbs sown had very low cover throughout the study period. Amongst the species of shrubs, Cytisus spp. develops cover above 50% from the fourth year after sowing in the plots where it was sown and colonisation is observed in the unsown ones from the ninth year. Quercus pyrenaica seedlings have a survival rate of 50%, but their mean height does not reach 40 cm after nine years. The shrub and herbaceous species present in the area before experimental burning are characterised by their great resilience, resprouting vegetatively or recovering from seed. It is therefore not possible to detect significant differences among the plots subjected to the different revegetation treatments and the controls regarding woody, herbaceous or total cover, nor are any significant differences observed in the diversity counted as number of species (richness). The good development of Cytisus spp. points to an important physiognomic change and this, together with the high survival rate, although not growth so far, of Q. pyrenaica may influence the long-term dynamic of the community. # 2004 Published by Elsevier B.V.

Erica australis heathlands in León province (NW Spain) have high resilience to disturbances and their post-fire recovery is very fast. The risk of soil erosion is high in the first few months after fire. The aim of this study is to investigate the effects on post-fire succession of sowing grass (Agrostis capillaris and Festuca rubra) and legume (Lotus corniculatus) seeds in a heathland burned by a summer wildfire, and to determine the most suitable native herbaceous species combination for protecting the soil in the first few phases of recovery. Fifteen permanent 4 m 2 plots are established in the burned area; four treatments and a control (unsown) are applied, each with three replicates. Three similar unburned plots are also considered (unburned control). Total cover is significantly higher in the sown plots in relation to the control in the first few months after sowing, but there are no differences after 18 months. Lotus corniculatus appears only in the first year and has no effect on the total cover. F. rubra appears earlier than Agrostis capillaris, but decreases significantly in cover after 18 months. Shrub species have the highest cover in the control plots and the lowest in the Agrostis plots. The correspondence analysis shows that the trend for vegetation in all plots reaches similar species composition by the time of final sampling. The last sampling of sown plots shows greater similarity to the control plots than the sampling of these plots within the first year. The fast initial growth of F. rubra, together with its decrease and subsequent low cover from the second year, make it more preferable than Agrostis capillaris for purposes of soil protection. However, additional research, both species-and site-specific, is necessary, as different responses due to different post-fire conditions and pre-fire species composition can have important implications on community dynamics.

No abstract