1. In many agricultural landscapes, it is important to restore networks of forests to provide habitat and stepping stones for forest specialist taxa. More knowledge is, however, needed on how to facilitate the immigration of such taxa in restored forest patches. Here, we present the first chronosequence study to quantify the dynamics of immigration credits of forest specialist plants in post-arable forest patches.

2. We studied the distribution of herbaceous forest specialist plant species in 54 post-arable broadleaved forest patches along gradients of age (20-140 years since forest establishment), distance from ancient forest (0-2600 m) and patch area (0.5-9.6 ha). With Linear Mixed Models we estimated the effects of these factors on species richness, patch means of four dispersal-related plant traits and with Generalized Linear Models on the occurrence of 20 individual species.

3. Post-arable forest patch age and spatial isolation from ancient forest, but not patch size, were important predictors for species richness of forest specialists, suggesting that also small patches are valuable for habitat connectivity. Compared to species richness in ancient forest stands, the immigration credit was reduced by more than 90% after 80 years in post-arable forest patches contiguous to ancient forest compared to 40% after 80 years and 60% after 140 years in isolated patches (at least 100 m to next forest). Tall-growing species with adaptations to long-distance dispersal were faster colonizers while species with heavy diaspores and clonal growth were slower to colonize.

4. Synthesis and applications: We show that post-arable oak plantations have a high potential for restoration of forest herb vegetation. Dispersal-related plant traits play a key role in explaining interspecific differences among forest specialists. To facilitate forest herb immigration across all functional groups in agricultural landscapes, we suggest to create clusters of relatively small new forest patches nearby older forest with source populations.

The data on the vascular plant flora were collected by the first author during field inventories between the years 2002 and 2018. Details on the inventories and collection of other data are given in the paper and the Supporting Information files.

Data on soil pH are missing for several post-arable forest patches. The missing values are indicated by n/a in excel file 1. However, soil pH was only used in preliminary data analysis and not in the final models.