International audienceBackground and Aims Seeds buried in the soil detect burial depth through light and diurnally fluctuating temperatures (DFT) and in this way limit losses due to germination too deep in the soil. DFTs and germination also increase in vegetation gaps. However, dry open environments with high DFTs can also increase seedling mortality, creating conflicting selection pressures for reaction to DFTs. Since this questions the general function of DFT detection, we therefore tested if interspecific differences in DFT detection are related to mortality in different soil depths. Methods We buried seeds of ten annual plants including species pairs of increasing and decreasing germination in response to DFTs. Seeds were buried in 5, 10 and 25cm soil depth and exhumed after two different burial times. Seed viability was tested using germination in growth chambers and tetrazolium. We also measured DFTs at these depths using temperature data loggers. Results DFT detection was not related to differences in mortality at three burial depths. Three species showed a clear pattern of depth dependent mortality, however inconsistently related to DFT detection. Conclusions Depth detection mechanisms are more species-specific than expected. Hence, interspecific differences in seed mortalities are difficult to predict by DFT detection alone and alternative soil depth sensing mechanisms should be explored in future

No abstract

The fragmentation of natural habitats is a major threat for biodiversity. However, the impact and spatial scale of natural isolation mechanisms leading to species loss, compared to anthropogenic fragmentation, are not clear, mainly due to differences between fragments and islands, such as matrix permeability. We studied a 500 km2 Mediterranean region in France, including urban habitat fragments, continuous habitat, and continental‐shelf islands. On the basis of 295 floristic relevés, we built species–area relationships to compare isolation in fragments after urbanization, with continuous habitat and continental‐shelf islands. We assumed either no dispersal, infinite dispersal, or estimated intermediate levels of habitat reachability through graph theory. Isolation mechanisms occurred in fragments but with a lower strength than in near‐shore islands, and most importantly affected perennial plants. Annual plants were less affected, probably due to their smaller size and shorter life cycle. Isolation occurred at landscape level in fragments and at patch level in islands. The amount of reachable habitat (accounting for spatial configuration) explained local species richness in both systems, but the amount of habitat (no consideration of spatial configuration) was already a good predictor. These results suggest an important role of habitat amount around fragments in mitigating the isolation effects observed in near‐shore islands, and the importance of carefully considering different functional groups.