How important are large vs. small forest remnants for the conservation of the woodland flora in an urban context?

Aim To test whether, in an urban area, small forest fragments are more important than large ones, the present study aimed at firstly exploring the relationship between plant functional groups or individual species and urban woodlot characteristics such as patch area and isolation, and secondly investigating whether equal-sized combinations of islands tend to differ in species richness. Location The city of Brussels. Methods We considered the relationship between size, species richness and plant functional groups among one very large (1666 ha) and 11 rather small (2-123 ha) woodlots. Results The largest woodlot harboured species missing in the smaller ones. The species-area relationship plotted for these wooded patches fitted the semilog model very well. Twenty-three species had a significantly higher frequency in the main forest. Only six species had a significantly higher frequency in smaller woodlots. The occurrence of species groups with high conservation value (e.g. ancient forest species, rare species) was higher in large patches. However, a SLOSS analysis showed that habitat subdivision appeared to be associated with increased species counts. A marginally nonsignificant effect of distance to the main forest became significant when matrix species were removed from the patch samples. Conclusions Although overall the data showed a higher conservation value for large woods, some plant functional groups (e.g. woodland species vs. ancient forest species) responded differently to fragmentation. This illustrates that, for conservation strategies, studies considering the biotic characteristics of remnants should focus on the species number of particular plant functional groups, especially those with high conservation value. Furthermore, matrix species should be removed from the analysis in order not to mask underlying patch size and distance effects.