Significant proportion of crop lands have been abandoned as management strategies have changed in Central and Eastern Europe in the past decades. The study of insect versus plant communities in such areas could help us understand how these processes take place, and whether these communities return to a semi-natural state maintained by human activities. Amongst insects ants, as ecosystem engineers, are a perfect target group in this respect. We studied epigaeic ant and plant communities of abandoned old-fields in Romania. Contrary to our expectations, the total number of ant species did not increase with time during succession on old-fields contrary to plants, where an increase was registered in the total number. Disturbancetolerant ant species dominated the ant communities throughout the successional gradient, while in the case of plants a transition was found from weed-dominated to semi-natural communities. The diversity of both ant and plant communities increased after the 1-year stage, but the patterns were different. While a return to semi-natural state could be observed in plants during old-field succession, such a definite change did not occur in ants. This might be caused by the landscape context: the lack of connectivity of old-fields to larger natural areas. While plant propagules of semi-natural and natural habitat species can still successfully colonize the old fields even under such conditions, ant colonizers are mainly disturbance-tolerant species typical for agricultural areas, which can be hardly replaced by typical grassland species. Our findings underline the existence of important discrepancies between plant and ant community succession, mostly treated as paralleling each other. This is the first study to handle the effect of abandonment on ant and plant communities simultaneously in Eastern Europe.

Studies addressing the question of how communities develop reported contrasting temporal patterns of species associations during succession. Several hypotheses were formulated about succession, but a general explanation of community assembly is missing. We analysed trends of species associations during old-field succession in two contrasting habitats: the first with chernozemic brown forest soil and temperate climate, and the second with sand soil and dryer climate. Significant pair-wise associations were calculated across a range of spatial scales. Comparing the two succession seres, one under harsh and the other under favourable environment, we attempted to make generalisations about species relation patterns. We found no trend but fluctuation in the level of community organization during succession. None of the existing succession models explained our results about changes in spatial structure of grassland communities during succession. Fluctuation in the number of significant associations was more intense and took longer under less favourable environmental conditions. Our results suggest that the stressed habitat type posed stronger constraints on species coexistence during succession than the favourable habitat did, but validating this hypothesis needs further investigations.

Decreasing diversity and plant cover, as well as increasing variability of these characteristics with increasing aridity are expected in grasslands due to climate change. These predictions were tested in perennial sand grasslands in Hungary. Two sites were chosen in different positions on an aridity gradient and two stands in each site were monitored for 9 years. Presence of plant species were recorded along 52 m long circular belt transects of 1040 units of 5 cm × 5 cm contiguous microquadrats. This sampling procedure — a version of line-intercept methods — enabled us to monitor diversity and total abundance in a sensitive, precise and non-destructive way. We found no trend but fluctuation in most community level attributes and in species composition. Contrary to fluctuations, between-site differences in diversity did not change and diversity remained lower in the more arid site during our 9-year-study. Compositional diversity performed better than species diversity because allowed us to detect vegetation changes that would have remained hidden if monitoring would be based only on the species richness. Comparing the magnitudes of fluctuations, five times higher relative interannual variability (CV%) was found for compositional diversity at the more arid site, while the relative temporal variability of total abundance and species richness did not show consistent patterns. We conclude that a 9 year-long study was too short to identify trends caused by the changing climate. However, the larger temporal variability of species combinations found in the more arid site suggests larger vulnerability and highlights the importance of non-linear dynamics during climate changes.

In the present paper we report original thousand-seed weight data for the flora of the Pannonian Basin. Our goal was to demonstrate the usefulness of seed weight databases by analysing seed weight data in relation to social behaviour types and life forms. We specifically asked the following questions: (i) how the seed weights are related to social behaviour type categories; (ii) how the life form of the species influences seed weight differences between respective social behaviour types? Own weight measurements are provided for 1,405 taxa; and for 187 taxa we published seed weight data for the first time: these were mostly endemics, orchids and/or species with Pontic, Caspian or continental distribution. Several taxonomic or functional groups are underrepresented in our database, like aquatic plants, rare arable weeds and sub-Mediterranean species. Problematic taxa, some difficult-to-harvest species or species with low seed production and cultivated adventives are also underrepresented. We found that the plant strategies expressed by social behaviour types were significantly different in terms of seed weights. The lowest seed weight scores were found for natural pioneers, whereas the highest ones were found for adventives and introduced cultivated plants. Short-lived herbaceous species had significantly higher seed weight scores than herbaceous perennials. No significant differences were found between specialists and generalists within the stress tolerant group. We found that short-lived graminoids possess heavier seeds than perennial graminoids, perennial and annual forbs. Naturalness scores were negatively correlated with seed weights. Our findings showed that seed collections and databases are not only for storing plant material and seed weight data, but can be effectively used for understanding ecological trends and testing plant trait-based hypotheses. Even the identified gaps underline the necessity of further seed collection and measurements.